elifeeLifeeLife2050-084XeLife Sciences Publications, Ltd6822110.7554/eLife.68221Research ArticleCancer BiologyDefining function of wild-type and three patient-specific TP53 mutations in a zebrafish model of embryonal rhabdomyosarcomaChenJiangfei12†BaxiKunal23†LipsittAmanda Ehttps://orcid.org/0000-0002-3757-849324HenschNicole Raehttps://orcid.org/0000-0001-9946-099523WangLonghttps://orcid.org/0000-0002-7935-414823SreenivasPrethish23ModiPaulomi23ZhaoXiang Ru23BaudinAntoine25RobledoDaniel G2BandyopadhyayAbhik2SugalskiAaron4ChallaAnil K26KurmashevDias2GilbertAndrea R7TomlinsonGail E24HoughtonPeter23ChenYidong8HayesMadeline N9ChenEleanor Y10LibichDavid Shttps://orcid.org/0000-0001-6492-280325IgnatiusMyron Shttps://orcid.org/0000-0001-6639-7707ignatius@uthscsa.edu23https://ror.org/00rd5t069Institute of Environmental Safety and Human Health, Wenzhou Medical UniversityWenzhouChinaGreehey Children's Cancer Research Institute (GCCRI), UT Health Sciences CenterSan AntonioUnited StatesDepartment of Molecular Medicine, UT Health Sciences CenterSan AntonioUnited StatesDepartment of Pediatrics, Division of Hematology Oncology, UT Health Sciences CenterSan AntonioUnited StatesDepartment of Biochemistry and Structural Biology, UT Health Sciences CenterSan AntonioUnited Stateshttps://ror.org/008s83205Department of Biology, University of Alabama at BirminghamBirminghamUnited StatesDepartment of Pathology and Laboratory Medicine, UT Health Sciences CenterSan AntonioUnited StatesDepartment of Population Health Sciences, UT Health Sciences CenterSan AntonioUnited Stateshttps://ror.org/057q4rt57Developmental and Stem Cell Biology, Hospital for Sick ChildrenTorontoCanadahttps://ror.org/00cvxb145Department of Laboratory Medicine and Pathology, University of WashingtonSeattleUnited StatesEkkerStephen CReviewing Editorhttps://ror.org/03zzw1w08Mayo ClinicUnited StatesWhiteRichard MSenior Editorhttps://ror.org/052gg0110Ludwig Institute for Cancer Research, University of OxfordUnited Kingdom
In embryonal rhabdomyosarcoma (ERMS) and generally in sarcomas, the role of wild-type and loss- or gain-of-function TP53 mutations remains largely undefined. Eliminating mutant or restoring wild-type p53 is challenging; nevertheless, understanding p53 variant effects on tumorigenesis remains central to realizing better treatment outcomes. In ERMS, >70% of patients retain wild-type TP53, yet mutations when present are associated with worse prognosis. Employing a kRASG12D-driven ERMS tumor model and tp53 null (tp53-/-) zebrafish, we define wild-type and patient-specific TP53 mutant effects on tumorigenesis. We demonstrate that tp53 is a major suppressor of tumorigenesis, where tp53 loss expands tumor initiation from <35% to >97% of animals. Characterizing three patient-specific alleles reveals that TP53C176F partially retains wild-type p53 apoptotic activity that can be exploited, whereas TP53P153Δ and TP53Y220C encode two structurally related proteins with gain-of-function effects that predispose to head musculature ERMS. TP53P153Δ unexpectedly also predisposes to hedgehog-expressing medulloblastomas in the kRASG12D-driven ERMS-model.
zebrafishTP53rhabdomyosarcomatumor initiationmutant tp53rare varientsTP53Y220CTP53C176FTP53P153ΔResearch organismZebrafishhttp://dx.doi.org/10.13039/100000002National Institutes of HealthR00CA175184HoughtonPeterIgnatiusMyron Shttp://dx.doi.org/10.13039/100004917Cancer Prevention and Research Institute of TexasScholar Grant RR160062IgnatiusMyron Shttp://dx.doi.org/10.13039/100007835Wenzhou Medical UniversityYoung Scientist Training Program (2019)ChenJiangfeihttp://dx.doi.org/10.13039/100007835Wenzhou Medical UniversityKYYW202203ChenJiangfeihttp://dx.doi.org/10.13039/100006058St. Baldrick's FoundationLibichDavid Shttp://dx.doi.org/10.13039/100000928Welch FoundationLibichDavid Shttp://dx.doi.org/10.13039/100006772Max and Minnie Tomerlin Voelcker FundYoung Investigator AwardLibichDavid SIgnatiusMyron Shttp://dx.doi.org/10.13039/100008635University of Texas Health Science Center at San AntonioT32CA148724BaxiKunalhttp://dx.doi.org/10.13039/100008635University of Texas Health Science Center at San AntonioTL1TR002647BaxiKunalhttp://dx.doi.org/10.13039/100008635University of Texas Health Science Center at San AntonioGreehey Graduate Fellowship in Children's HealthHenschNicole RaeModiPaulomihttp://dx.doi.org/10.13039/100004917Cancer Prevention and Research Institute of TexasTraining Award RP170345HenschNicole RaeLipsittAmanda Ehttp://dx.doi.org/10.13039/100008258Hyundai Hope On WheelsYoung Investigator GrantLipsittAmanda EThe funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.Author impact statementZebrafish are an optimal model organism to study rare TP53 mutations whose functions are not readily understood.Introduction
TP53 is the best-known tumor suppressor protein that is mutated or functionally disrupted in more than 50% of human tumors (Kastenhuber and Lowe, 2017; Muller and Vousden, 2014). Germline mutations in TP53 are responsible for Li–Fraumeni syndrome that predisposes to a wide but distinct spectrum of tumors that varies with age (Malkin, 2011). Comprehensive analyses of TP53 function in vivo and in vitro have revealed three different ways by which TP53 can modulate tumorigenesis. These include effects caused by loss- or gain-of-function or dominant-negative mutations that disrupt the function of wild-type protein (Ko and Prives, 1996; Levine et al., 1991). Furthermore, the TP53 mutation spectrum is likely tumor-specific (e.g., G>T transversions in lung cancer that do not correspond to the classic TP53 hotspot mutations) (Olive et al., 2004; Petitjean et al., 2007). Recent studies looking at TP53 alterations in rhabdomyosarcoma (RMS) patients show that TP53 mutations are correlated with an increased risk of developing RMS and a worse prognosis (Casey et al., 2021; Shern et al., 2021); however, the role for specific TP53 variants in sarcoma progression remains to be defined.
TP53 mutations are present in >40% of adult carcinomas and thought to play a major role in tumorigenesis (Gröbner et al., 2018). In contrast, mutations in TP53 are found in less than 6% of pediatric cancers (Chen et al., 2014a; Chen, 2013; Gröbner et al., 2018; Seki et al., 2015; Shern et al., 2014; Tirode et al., 2014). One major exception is Li–Fraumeni syndrome, where germline TP53 mutations predispose individuals to a unique tumor spectrum that includes soft tissue sarcomas, such as RMS, and bone tumors (Guha and Malkin, 2017; Malkin, 2011). It is important to also note that the TP53 pathway is often suppressed in sarcomas. For example, in human embryonal rhabdomyosarcoma (ERMS), a common pediatric cancer of muscle, the TP53 locus is mutated or deleted in 16% of tumors, while transcriptional activity is altered in >30% of tumors, either through direct locus disruption or MDM2 amplification (Chen, 2013; Seki et al., 2015; Shern et al., 2014; Taylor et al., 2000). TP53 mutations are also detected in ERMS at relapse, suggesting a role in tumor progression and/or resistance to therapy (Chen, 2013).
Mouse models have led the way in understanding Tp53 function in vivo. Several murine genetic models were developed to assess the effects of both loss- and gain-of-function Trp53 mutations (Attardi and Donehower, 2005; Garcia and Attardi, 2014) with mutant and null alleles spontaneously developing cancer in multiple tissues (Lozano, 2010). Of note, the tumor spectrum in mice varies depending on the mutant allele and genetic background; however, most in vivo studies have focused on a subset of hotspot mutations compared to a null or heterozygous background, and are seen only in a subset of sarcoma patients (Attardi and Donehower, 2005; Garcia and Attardi, 2014; Guha and Malkin, 2017). The vast majority of mutations observed in patients have not been interrogated in animal models, but rather function is inferred from these commonly studied hotspot mutants. TP53 variants likely play different roles depending on tumor type, mutation, or the genetic background, presenting a significant challenge when defining pathogenicity and potential impacts on therapeutic approaches.
To define p53 biology in vivo, we recently generated a complete loss-of-function tp53 deletion allele in syngeneic CG1-strain zebrafish using TALEN endonucleases (Ignatius et al., 2018). tp53del/del (tp53-/-) zebrafish spontaneously developed a spectrum of tumors that includes malignant peripheral nerve-sheath tumors (MPNSTs), angiosarcomas, germ cell tumors, and an aggressive natural killer cell-like leukemia (Ignatius et al., 2018). The role for tp53 in self-renewal and metastasis of kRASG12D-induced ERMS tumors was assessed using cell transplantation assays and revealed that tp53 loss does not change the overall frequency of ERMS self-renewing cancer stem cells compared to tumors expressing wild-type tp53 (Ignatius et al., 2018). In contrast, tp53-/- ERMS were more invasive and metastatic compared to tp53 wild-type tumors, providing new insights into how tp53 suppresses ERMS progression in vivo (Ignatius et al., 2018). In the present study, we take advantage of the fact that in zebrafish, similar to humans, p53 loss of function is not required for tumor initiation. We employ tp53-/- zebrafish to further assess the role for p53 on kRASG12D-driven ERMS and find that wild-type tp53 is a major tumor suppressor in ERMS, affecting proliferation with a smaller effect on apoptosis. Next, we find that wild-type human TP53 is functional in zebrafish and potently suppresses ERMS initiation in vivo. We also define the pathogenicity of three TP53 mutations: (1) a TP53C176F variant found in ERMS patients (Chen, 2013; Seki et al., 2015; Shern et al., 2014); (2) a TP53P153Δ variant of unknown significance, found in a teenager with an aggressive osteosarcoma at our clinic; and (3) a p53Y220C mutant that is structurally related to p53P153Δ and may share some aspects of its function. We find that these three TP53 mutants have different effects on tumor initiation, location, proliferation, and apoptosis in our zebrafish model. Taken together, these analyses highlighting a role for the zebrafish ERMS model that can be used to characterize the spectrum of both common and rare TP53 mutations in sarcoma patients.
Results<italic>tp53</italic> is a potent suppressor of ERMS initiation, growth, and invasion
Whole-genome/-exome/-transcriptome sequencing analysis revealed that a majority of primary ERMS tumors are wild type for TP53. However, TP53 loss, mutation, or MDM2 amplification accounts for pathway disruption in approximately 30% of ERMS (Chen, 2013; Seki et al., 2015; Shern et al., 2014). Furthermore, TP53 mutations are associated with poor prognosis (Casey et al., 2021; Shern et al., 2021). In the zebrafish ERMS model, expressing the human kRASG12D oncogene in muscle progenitor cells is sufficient to generate ERMS with morphological and molecular characteristics that recapitulate the human disease (Langenau et al., 2007). Although robust, in a tp53 wild-type genetic background, the upper limit of tumor formation observed is 40% by 50 d post injection/fertilization (Langenau et al., 2007). To test whether tp53 suppresses ERMS initiation in vivo, we generated tumors using rag2:KRASG12D in tp53 wild-type (tp53+/+) and tp53-/- mutant zebrafish (Figure 1—figure supplement 1). Co-injection of rag2:kRASG12D and rag2:DsRed resulted in approximately 34% of tp53+/+ animals (n = 49/143 animals) with DsRed-positive tumors by 60 d post injection. In stark contrast, up to 97% of tp53-/- animals displayed tumors (n = 139/142 animals) by 60 d. kRASG12D-induced tp53-/- ERMS displayed a very rapid onset, with the majority of arising within the first 20 d of life (Figure 1A; p<0.0001). We assessed tumor size and the number of tumors initiated and found that tp53-/- mutant ERMS displayed a significant increase in size (Figure 1B and D; p<0.0001) and in the number of tumors per animal (Figure 1B and C; p<0.0001) compared to tp53+/+ animals. Histological assessment showed no major differences between tp53+/+ and tp53-/- ERMS, and tumors formed in the head, trunk, or tail similarly in both backgrounds (Figure 1E and F, Figure 1—figure supplement 2B). We assessed relative expression of kRASG12D and its downstream effector dusp4 in tumors from tp53-/- and tp53+/+ zebrafish and confirmed similar expression levels of kRASG12D and dusp4 using quantitative qPCR (Figure 1—figure supplement 2A; n = 3 tumors each, p=0.8966 unpaired t-test with Welch’s correction). Taken together, our analysis suggests that tp53 is a major suppressor of tumor initiation in RAS-driven zebrafish ERMS.
(A) Kaplan–Meier plot showing ERMS tumor initiation in tp53-/- and tp53+/+ fish. (B) Representative images of DsRed-positive zebrafish ERMS. Arrows show tumor location for each fish. All tumor-burdened zebrafish are 10 days old. Scale bar = 0.5 mm. (C) Tumor numbers per zebrafish in tp53-/- and tp53+/+ fish. n = 44 (tp53+/+), n=130 (tp53-/-). (D) Ratio of tumor area to total body area in in tp53-/- and tp53+/+ fish. n = 10. (E) Pie chart showing percentage of tumors found in varying regions of tp53-/- and tp53+/+ fish, showing no significant differences in tumor localization. Head – p=0.25848, trunk – p=0.39532, tail – p=0.92034 (two-tailed two proportions Z-test). (F) Representative H&E staining of zebrafish ERMS tumors. Scale bar = 100 µm.
Schematic of the experimental setup to generate embryonal rhabdomyosarcoma (ERMS) tumors in zebrafish via microinjection of indicated linearized DNA constructs into the one-cell-stage zebrafish embryos generated from in crosses of <italic>tp53<sup>-/-</sup></italic> or <italic>tp53<sup>+/+</sup></italic> zebrafish.
(A) mRNA expression of kRASG12D and dusp4 in compared to 18s expression level. (B) Representative H&E and immunohistochemistry (anti-DsRed) staining in embryonal rhabdomyosarcoma (ERMS) tumor fish. Arrows denote location of tumors. Zebrafish shown are 35 days old. Scale bar = 100 µm.
<italic>tp53</italic> suppresses proliferation and to a lesser extent apoptosis in ERMS tumors
Given that tp53-/- tumors are larger than equivalent staged tp53+/+ tumors, we performed EdU and phospho-histone H3 staining to assess proliferation and Annexin V staining to assess apoptosis in primary ERMS. Wild-type and tp53-/- tumor-burdened animals were treated with a 6 hr pulse of EdU, euthanized, sectioned, and stained for EdU-positive cells, a marker for proliferation. EdU analyses revealed that tp53-/- tumors displayed a significant increase in EdU-positive cells (Figure 2A; p<0.0001 Student’s t-test) and significantly more phospho-histone H3-expressing cells when compared to tp53 wild-type ERMS sections (Figure 2B; p<0.01 Student’s t-test), indicating increased proliferation. Annexin V staining, that can distinguish cells beginning to undergo apoptosis (early apoptosis; low Annexin V/ propidium iodide [PI]-positive), or that are undergoing apoptosis (late apoptosis; high Annexin V/ high PI) or necrosis (high Annexin V/ low PI), was performed on live single-cell suspensions of ERMS tumor cells extracted post euthanasia. Annexin V staining showed no significant difference in the rate of late apoptosis (Q2) (Figure 2C–E; p=0.592 Student’s t-test), but showed a small yet significant decrease in early apoptosis (Q4) in tp53-/- mutant ERMS compared to wild-type (Figure 2C–E; p=0.0073 Student’s t-test). Taken together, our data suggests that tp53 is a potent suppressor of kRAS-induced ERMS tumor cell proliferation, with only a moderate effect on apoptosis, affecting ERMS initiation.
<italic>tp53</italic> is a potent suppressor of proliferation and to a lesser extent of apoptosis.
(A) Representative confocal microscopy images of EdU staining on embryonal rhabdomyosarcoma (ERMS) tumor sections and a plot quantifying the percentage of EdU-positive cells. Average of n = 8–11/primary tumors. White arrows show EdU-positive cells. (B) Representative confocal microscopy images of phospho-histone H3 staining on ERMS tumor sections (scale bar = 100 µm). Total number pHH3-positive cells per single ERMS tumor ×200 confocal image section assessed from n7-9 primary tumors. One the right-most panel is a plot quantifying the total number of pHH3-positive cells per single ERMS section. White arrows show pHH3-positive cells. (C, D) Representative flow cytometry analysis of Annexin V staining of tp53+/+and tp53-/- ERMS tumors, respectively. (E) Quantification of flow cytometry analysis of Annexin V staining. Q1 = pre-necrotic cells, Q2 = late apoptosis + necrotic cells, Q3 = living cells, Q4 = early apoptotic cells. n = 7. ns, not significant, p=0.5926, unpaired t-test.
Reintroducing human <italic>TP53</italic> in <italic>tp53<sup>-/-</sup></italic> zebrafish blocks tumor initiation, growth, proliferation, and increases apoptosis
Zebrafish and human p53 are functionally similar and share 56% identity with respect to amino acid sequence (67% positives, Figure 3—figure supplement 1; Berghmans et al., 2005; Ignatius et al., 2018; Parant et al., 2010; Storer and Zon, 2010). Within the core DNA-binding region where a majority of mutations occur in patients, 72% conservation exists (79% positives, Figure 3—figure supplements 1 and 2). To assess functional conservation in vivo, we co-expressed wild-type human TP53 (TP53WT) in the cells from which ERMS tumors initiate in tp53-/- animals (Figure 3—figure supplement 3). Importantly, co-expression of kRASG12D along with TP53WT resulted in the expression of wild-type TP53 from the very beginning in cells from which tumors initiate and also in the resulting tumors. Co-expression of kRASG12D along with TP53WT significantly suppressed tumor initiation (Figure 3A; p<0.0001, Student’s t-test). ERMS in TP53WT-expressing animals were smaller than tumors in age-matched tp53-/- zebrafish (Figure 3B and C; p<0.0001 Student’s t-test); however, we observed no significant difference in the number or distribution of ERMS initiated per zebrafish between the two groups (Figure 3D and E; p=0.065 for Figure 3D, Student’s t-test). Analysis of EdU staining of ERMS tumors determined that co-expressing TP53WT significantly inhibited proliferation (Figure 3F; p=0.0007 Student’s t-test). Also, we observed an overall increase in the number of apoptotic cells in TP53WT-expressing tumors, as seen by Annexin V staining (Figure 3G; p<0.0001 Student’s t-test). Similarly, wild-type zebrafish tp53 when co-expressed with kRASG12D in tp53-/- zebrafish significantly suppressed tumor initiation, proliferation, and apoptosis (Figure 3—figure supplement 4; p<0.0001 Student’s t-test). Importantly, expression of human TP53WT or wild-type zebrafish tp53 selectively in ERMS using the rag2 promoter had no effect on the overall viability of zebrafish embryos or larvae. Co-expression of human TP53 or zebrafish tp53 in tp53-/- zebrafish did result in increased apoptosis compared to tumors generated in tp53+/+ zebrafish. However, it is important to point out that in this assay human p53 was expressed at significantly lower levels than is present in an RMS cell line expressing mutant p53 (see Figure 3A).
Human <italic>TP53</italic> blocks tumor initiation, growth, and proliferation and increases apoptosis in <italic>tp53<sup>-/-</sup></italic> zebrafish.
(A) Kaplan–Meier plot showing embryonal rhabdomyosarcoma (ERMS) tumor initiation in tp53-/- fish with or without p53WT expression. Western blot analysis was performed to assess p53WT expression level in tumors. (B) Representative images of ERMS tumors in tp53-/- fish with or without human TP53WT expression. Tumor-burdened zebrafish are between 15 and 20 days old. Scale bar = 1 mm. (C) Ratio of tumor area to total body area in tp53-/- fish with or without expression of TP53WT. n = 18. (D) Number of tumors per tp53-/- zebrafish with or without expression of TP53WT. ns, not significant. n = 36 (tp53-/-), n = 28 (TP53WT). (E) Pie chart showing site of tumor localization in tp53-/- fish with or without expression of TP53WT showing no statistical differences. Head – p=0.20045, trunk – p=0.42858, tail – p=0.3336. Quantification of proliferation (F) and apoptosis (G) via EdU staining (n = 10) and Annexin V staining (n = 3), respectively, for tumors arising in tp53-/- fish with or without expression of TP53WT.
Human TP53 blocks tumor initiation, growth, and proliferation and increases apoptosis in tp53<sup>-/-</sup> zebrafish.
(A) Schematic of the protein structure of p53 with the different domains highlighted with different colors. (B) The amino acid sequence alignment for p53 between human (Hs) and zebrafish (Dr-Danio rerio). The colors highlight domains in (A); - indicates gaps; : indicates positive substitutions; * indicates conserved amino acids,. (C) Table showing amino acid conservation, positive substitutions, and gaps in the full-length and DNA-binding domain.
Human TP53 blocks tumor initiation, growth, and proliferation and increases apoptosis in tp53<sup>-/-</sup> zebrafish.
(A) Lollipop plot showing common p53 mutations in rhabdomyosarcoma patients as well as the seven most common p53 hot spot mutations across all cancers. (B) The amino acid sequence alignment for p53P153 with other species. The yellow highlight shows the amino acid of P153, the top panel shows the conserved ones, and the bottom five species show the non-conserved ones.
Schematic of the experimental setup to generate embryonal rhabdomyosarcoma (ERMS) tumors in zebrafish via microinjection of indicated linearized DNA constructs into the one-cell-stage zebrafish embryos generated from in crosses of <italic>tp53<sup>-/-</sup></italic> zebrafish.
Zebrafish tp53 (italics) blocks tumor initiation in tp53<sup>-/-</sup> zebrafish.
(A) Kaplan–Meier plot showing embryonal rhabdomyosarcoma (ERMS) tumor initiation in tp53-/- fish (AB WT strain) with or without tp53. (B) Zebrafish p53 expression levels in three tp53-/- tumors expressed rag2:tp53 along with rag2:kRASG12D, using AB WT tumor driven by rag2:kRASG12D as control.
<italic>TP53<sup>C176F</sup></italic> is a hypomorphic allele while <italic>TP53<sup>P153Δ</sup></italic> has gain-of-function effects in ERMS
The effects of specific TP53 mutations on ERMS are not well defined. Analysis of TP53 mutations in RMS patients found that a majority of mutations lie outside the well-studied hotspot locations and are mostly uncharacterized (Figure 3—figure supplement 2). To determine whether ERMS-expressing patient-specific TP53 point mutations differ from TP53 deletion in vivo, we assessed mutant activity in tp53-/- (null) ERMS. The first selected mutation was a TP53C176F allele that is expressed by at least two ERMS patients (Chen, 2013; Seki et al., 2015). In one of these patients, the second TP53 allele was deleted in the tumor and TP53C176F is present in both the primary and relapsed tumor (Chen, 2013). Another mutation we selected to model was a rare TP53P153Δ (deletion of Proline 153, P153Δ) allele present in a patient in our clinic with an aggressive osteosarcoma. This patient developed an aggressive refractory osteosarcoma as a teenager, while her mother developed osteosarcoma in her early twenties; both eventually succumbed to their disease. Based on the patient’s family history, tumor type, and TP53 mutation, a diagnosis of Li–Fraumeni syndrome was made. However, since the mutation is rare, pathogenicity could not be assigned in this particular allele.
Analysis of amino acid sequence conservation showed that p53 residue C176 is conserved across humans, mice, and zebrafish. P153 is present only in humans; however, it is located in a region that is highly conserved across all three species and P153 is conserved across other closely related mammalian species (Figure 4A, Figure 3—figure supplement 2). We confirmed TP53P153Δ mutation by sequencing the DNA from the first passage murine patient-derived xenograft (PDX) generated from the patient at autopsy. Tumor cells obtained from the patient also harbored a missense mutation at c.476C>T (p.A159V) (Figure 4B). We next assessed whether p53C176F and p53P153Δ proteins were expressed in the ERMS murine PDX SJRHB00011 that harbors the TP53C176F mutation and in the primary osteosarcoma that harbors the TP53P153Δ mutation using immunohistochemistry on tumor sections (Chen, 2013). Both the SJRHB00011 PDX tumor and the primary osteosarcoma tumor expressed p53 protein, as evidenced by strong positive nuclear staining in the majority of tumor cells (Figure 4C and D). H&E staining confirmed RMS (heterogeneous population of ovoid to slightly spindled cells; Figure 4E) and osteosarcoma diagnosis (pleomorphic neoplastic tumor cells with irregular disorganized trabeculae of unmineralized malignant osteoid; Figure 4F).
Assigning pathogenicity to two human <italic>TP53</italic> sarcoma mutations in the kRAS<sup>G12D</sup>-driven embryonal rhabdomyosarcoma (ERMS) model.
(A) Lollipop plot showing the two novel, human p53 mutations P153Δ and C176F, as well as the amino acid sequence alignment for human, mouse, and zebrafish protein. (B) DNA sequencing data from osteosarcoma patient confirming the germline P153Δ mutation, as well as somatic A159V mutation. (C, D) p53 immunohistochemistry staining of p53 in ERMS patient-derived xenograft (PDX) SJRHB00011 expressing p53C176F and osteosarcoma expressing p53P153Δ. (E, F) Representative H&E staining of ERMS PDX expressing the C176F mutation and diagnostic biopsy of osteosarcoma tumor expressing osteosarcoma expressing p53P153Δ showing neoplastic tumor cells with pleomorphic nuclei, irregular chromatin pattern, as well as irregular disorganized trabeculae of unmineralized malignant osteoid (stars). (G) Protein expression of mutant p53 in zebrafish ERMS tumors, with rhabdomyosarcoma (RMS) cell line, Rh30, as a control. (H) Kaplan–Meier plot showing tumor initiation in tp53-/- fish with or without expression of mutant TP53. (I) Representative images of tumor localization in tp53-/- fish with or without expression of mutant TP53. Age of zebrafish in panels is 37 d. Scale bar = 1 mm (J) Pie chart showing percentage of tumors found in varying regions of tp53-/- fish with or without expression of mutant TP53. Percentages in red indicate a significant difference to tp53-/- (p=0.0096, two-tailed two proportions Z-test). (K) Quantification of Annexin V staining in tumors arising in tp53-/- fish with or without expression of mutant TP53. n = 6–7. (L) Representative H&E staining of tumors arising in tp53-/- fish with or without expression of mutant TP53. Scale bar = 100 µm.
Schematic of the experimental setup to generate embryonal rhabdomyosarcoma (ERMS) tumors in zebrafish via microinjection of indicated linearized DNA constructs into the one-cell-stage zebrafish embryos generated from in crosses of <italic>tp53<sup>-/-</sup></italic> zebrafish.
Assigning pathogenicity to two human TP53 sarcoma mutations in the kRASG12D-driven embryonal rhabdomyosarcoma (ERMS) model.
(A) Tumor numbers per zebrafish expressing no TP53 (tp53-/-), or TP53C176F, or TP53P153Δ.n = 36 (tp53-/-), n = 58 (TP53C176F), n = 81 (TP53P153Δ). (B) Percentage of EdU-positive cells in representative areas of fish tumors expressing either no TP53 (tp53-/-), TP53C176F, or TP53P153Δ. n = 8–10. (C) Total number of pHH3-positive cells per single ERMS tumor ×200 confocal image section assessed from tumors expressing WT TP53, TP53C176F, or TP53P153Δ intp53-/- zebrafish; n = 3–5 primary tumors. (D) Representative images of tumor-burdened adult zebrafish (4–6 months of age) with tumors expressing kRASG12D + TP53C176F (DsRed+) treated with either DMSO or ZMC1 over a span of 3 wk. (E) Western blot showing p53 expression levels in fish tumors expressing kRASG12D + TP53C176F and treated with either DMSO or ZMC1. (F) Percentage of Annexin V-positive cells in tumors expressing kRASG12D + TP53C176F that were treated with either DMSO or ZMC1. n = 4. (G) Ratio of tumor size to total body for in fish tumors expressing kRASG12D + TP53C176F, treated with either DMSO or ZMC1. ns, not significant, n = 4. (H) Representative images of tumor-burdened zebrafish expressing kRASG12D (DsRed+) treated with either DMSO or ZMC1 over a span of 3 wk. Scale bar = 2 mm (I) Western blot showing p53 expression levels in Rh30 cells (control) and fish tumors expressing kRASG12D, treated with either DMSO or ZMC1. (J) Percentage of Annexin V staining of fish tumors expressing kRASG12D, treated with either DMSO or ZMC1. ns, not significant. n = 6–7. (K) Ratio of tumor size to total body in tumors expressing kRASG12D, treated with either DMSO or ZMC1. ns, not significant, n = 4.
Semi-quantitative qPCR analyses comparing expression of known p53 direct regulated genes <italic>baxa, bbc3</italic>, <italic>cdkn1a, gadd45a,</italic> and <italic>noxa</italic> in primary zebrafish embryonal rhabdomyosarcoma (ERMS) expressing wt <italic>tp53</italic> (blue, n = 3), <italic>tp53<sup>-/-</sup></italic> (red, n = 3), <italic>tp53<sup>-/-</sup> + TP53</italic><sup><italic>C176F</italic></sup> (purple, n = 4), or <italic>tp53<sup>-/-</sup> + TP53<sup>P153Δ</sup></italic> (green, n = 4).
Also shown is expression of zebrafish tp53 across the tumors.
We generated ERMS in tp53-/- zebrafish using the same approach as earlier described and co-expressed TP53C176F or TP53P153Δ along with kRASG12D in tp53-/- embryos in cells from which tumors initiate (Figure 4—figure supplement 1). Expression of mutant p53 protein in ERMS tumors was confirmed by western blot analysis using a human-specific p53 antibody. Mutant p53 expression in Rh30 RMS cells was used as a positive control while ERMS cells from tp53-/- zebrafish were used as a negative control (Figure 4G). Expression of TP53C176F with kRASG12D in tp53-/- zebrafish resulted in a significant reduction in tumor initiation compared to expressing kRASG12D alone (Figure 4H, p=0.0005). By assessing TP53C176F-expressing zebrafish for ERMS incidence, location, proliferation, and apoptosis, we found that the number of primary ERMS per fish was reduced but not significant between tp53-/- and tp53-/- + TP53C176F-expressing animals (Figure 4—figure supplement 2A, tp53-/- vs. tp53-/- + TP53C176F, p=0.065 (TP53C176F) one-way ANOVA with Tukey’s multiple-comparisons test). tp53-/- + TP53C176F-expressing ERMS had fewer EdU-positive cells but similar levels of cells undergoing mitosis compared to tp53-/- tumors (Figure 4—figure supplement 2B and C). In contrast, TP53C176F tumors showed significantly higher rates of apoptosis (Figure 4K, p<0.0001). There were no histological differences between the two groups (Figure 4L).
In contrast to TP53C176F-expressing ERMS, expression of TP53P153Δ with kRASG12D in tp53-/- zebrafish did not affect tumor initiation compared to kRASG12D alone (Figure 4H, p=0.774). However, unexpectedly, tp53-/- + TP53P153Δ-expressing zebrafish had developed twice as many ERMS in the head musculature (Figure 4I and J, p=0.0096, two proportions Z-test), suggesting a gain-of-function effect with respect to initiation site. Additionally, while we did not observe a difference in apoptosis, we did find that tp53-/- + TP53P153Δ tumors were less proliferative compared to tp53-/- (Figure 4—figure supplement 2B and C).Tumor histology remained unchanged across both groups (Figure 4L).
To assess the effects of p53 variant expression on transcription, we performed qPCR analyses comparing tp53+/+, tp53-/- and tp53-/- + TP53C176F and tp53-/- + TP53P153Δ ERMS and p53 target genes including baxa, cdkn1a, gadd45a, noxa, and puma/bbc3. We found that both tp53-/- + TP53C176F and tp53-/- + TP53P153Δ tumors displayed increased bbc3 expression compared to tp53-/- or tp53+/+ tumors, but had no difference with respect to cdkn1a, gadd45a and noxa expression. We also found that tp53-/- + TP53C176F and tp53-/- + TP53P153Δ tumors expressed higher levels of cdkn1a compared to tp53+/+ ERMS; however, rather unexpectedly tp53-/- tumors also express higher levels of cdkn1a compared to tp53+/+ controls (Figure 4—figure supplement 3).
Given that the TP53C176F mutation retains the ability to induce apoptosis, we next tested whether this activity could be augmented to inhibit tumor growth in vivo. It has been shown previously that ZMC1, a synthetic metallochaperone that transports zinc into cells as an ionophore, can restore p53 activity by stabilizing mutant p53 proteins such as p53C176F (Blanden et al., 2015). To test this, we generated ERMS in the syngeneic CG1 strain zebrafish that were either tp53-/- or tp53-/- +TP53C176F. Next, we expanded tumors from both groups in recipient wild-type CG1 animals and treated tumors in recipient host animals for 2 wk with either DMSO or ZMC1. Compared to DMSO-only control treatment, tp53-/- + TP53C176F-expressing ERMS treated with ZMC1 showed a significant reduction in tumor growth over time (Figure 4—figure supplement 2D and G; p=0.0116 at week 3). We next assessed effects of ZMC1 on apoptosis and found increased apoptosis in tp53-/- + TP53C176F-expressing ERMS treated with ZMC1 (Figure 4—figure supplement 2F; p=0.0008). ZMC1 treatment increased p53C176F expression at the protein level, suggesting increased stability of p53C176F protein in response to drug (Figure 4—figure supplement 2E). In contrast, ZMC1 treatment in tp53-/- ERMS cohorts did not affect tumor growth (Figure 4—figure supplement 2H and K, p=0.961) or apoptosis (Figure 4—figure supplement 2I and J, p=0.583).
Altogether, TP53C176F appears to retain some wild-type function that partially prevents ERMS initiation in tp53-/- zebrafish, as well as trigger apoptosis in vivo. However, this variant appears not to have a significant effect on cell proliferation. Interestingly, the activity of p53C176F can be further enhanced by tp53 reactivators, such as ZMC1. In contrast, we found that the TP53P153Δ mutation functions as a gain-of function allele with respect to the site of tumor initiation, but has no effect on overall tumor initiation. However, tumors expressing the TP53P153Δ while having no effects on apoptosis compared to tp53-/- have significantly fewer proliferating cells, suggesting an allele with gain of function and also some wild-type activity (Supplementary file 1).
Expression of <italic>TP53<sup>P153Δ</sup></italic> with <italic>kRAS<sup>G12D</sup></italic> in <italic>tp53<sup>-/-</sup></italic> zebrafish results in the initiation of medulloblastomas with a shh gene signature
The head ERMS tumors that are formed in zebrafish expressing kRASG12D and TP53P153Δ in the tp53-/- background initiate tumors in the head musculature around the eye and in the jaw; however, these tumors are overall less proliferative compared to the trunk tumors (Figure 4—figure supplement 2B and C). We generated tumors expressing kRASG12D and GFP or kRASG12D, TP53P153Δ, and GFP in syngeneic tp53-/- CG1 strain zebrafish and expanded three primary head tumors from kRASG12D; TP53P153Δ;tp53-/-; GFP-expressing animals and one head and two trunk tumors expressing control kRASG12D; tp53-/-; GFP in syngeneic CG1 strain recipients that are wild-type for tp53. We find no difference in labeling tumors with different fluorescent reporter proteins, so we use these reporters interchangeably (Ignatius et al., 2012; Ignatius et al., 2018; Langenau et al., 2008). Secondary tumor cells were enriched by flow cytometry and RNA extracted and processed for RNA sequencing analyses. RNAseq unexpectedly revealed that all three kRASG12D;TP53P153Δ; tp53-/-;GFP-positive tumors displayed RNA expression signatures associated with neuronal differentiation, neurotransmitter secretion, synapse formation, and neurogenesis; while kRASG12D; tp53-/-;GFP-positive-only tumors as expected displayed signatures associated with myogenesis, skeletal muscle structure, and muscle differentiation (Figure 5A and B, Supplementary file 2). Unbiased gene set enrichment analysis (GSEA) found that TP53P153Δ tumors had gene signatures associated with pro-neural glioblastoma, neuronal markers, and the downregulation of negative regulators of hedgehog signaling. These tumors also displayed negative correlation with soft tissue tumors and glioblastomas of mesenchymal origin (Figure 5—figure supplement 1, Supplementary file 2). Together, these data suggest that at least some of the head tumors are more likely brain tumors with neuronal gene expression signatures, rather than ERMS.
Expression of <italic>TP53<sup>P153Δ</sup></italic> with <italic>kRAS<sup>G12D</sup></italic> in <italic>tp53<sup>-/-</sup></italic> zebrafish results in the initiation of medulloblastomas with a shh gene signature.
(A) Heatmap from RNAseq analyses comparing tumors expressing kRASG12D;tp53-/- to kRASG12D; tp53-/-; TP53P153Δ (n = 3/group). A total of 643 genes were selected for the heatmap, with adjusted p-value<0.01 and fold-change >10. (B) Enriched Gene Ontology (GO) Biological Processes (BP) in upregulated genes in p53-/- group (left panel with pink bars) consistent with the expected tissue of origin for kRASG12D;tp53-/- to kRASG12D; tp53-/-; TP53P153Δ (right panel with light blue bars). (C) Representative images of medulloblastoma tumors expressing kRASG12D and DsRED (top panel) or GFP (bottom two panels) in tp53-/- fish with or without human TP53P153Δ expression. Tumor-burdened zebrafish are between 30 (Fish 1) and 70 (Fish 2, 3) days old. Scale bar = 1 mm. (D) Representative H&E staining of tumors arising in the head region of tp53-/- fish with or without expression of human TP53P153Δ. Images between 40 and 60 times magnification. (E–G) Representative sections of tumor-burdened zebrafish showing H&E staining (D, E) and IHC staining for Sox10 (F), and Gfap (G) in head tumors expressing kRASG12D in tp53-/- fish or kRASG12D andTP53P153Δ in tp53-/- fish. Scale bar = 60 µm (H) Gene set enrichment analysis (GSEA) showing the enrichment of medulloblastoma gene signatures from Pomeroy et al., and Kool et al., with our zebrafish brain tumors. Log-fold-change derived from the differential expression analysis between tp53-/- and tp53-/-; TP53P153Δ was used with GSEA pre-ranked function (GSEA, v4.0.3, Broad Institute, MA).
Unbiased gene set enrichment analysis (GSEA) (GSEApreranked function) comparing gene expression between <italic>tp53<sup>-</sup></italic><sup>/-</sup> to <italic>tp53<sup>-/-</sup>; TP53<sup>P153Δ</sup></italic> tumors (log-fold-change) were performed with chemical and genetic perturbation (CGP) gene sets, part of MSigDB C2 curated gene set collection.
Schematic of the experimental setup to generate embryonal rhabdomyosarcoma (ERMS) tumors in zebrafish via microinjection of indicated linearized DNA constructs into the one-cell-stage zebrafish embryos generated from in crosses of <italic>tp53<sup>-/-</sup></italic> zebrafish.
Expression of <italic>TP53</italic><sup><italic>P153Δ</italic></sup> with kRAS<sup>G12D</sup> in tp53<sup>-/-</sup> zebrafish results in the initiation of medulloblastomas with a shh gene signature. (<italic>TP53</italic><sup><italic>P153Δ</italic></sup>).
(A) Representative images of tp53-/- fish with head embryonal rhabdomyosarcoma (ERMS) tumors, expressing TP53P153Δ (DsRED positive). The zebrafish in (A) are between 25 and 50 d. Scale bar = 1 mm. (B) H&E of the head regions of tp53-/- fish with head ERMS tumors, expressing TP53P153Δ. (C) H&E stained sections at ×20 magnification of tp53-/- fish with head ERMS tumors, expressing TP53P153Δ. (D) IHC for myosin heavy chain at ×20 magnification of tp53-/- fish with head ERMS tumors, expressing TP53P153Δ. Scale bar = 100 µm.
Gene set enrichment analysis (GSEA) showing the enrichment of medulloblastoma gene signatures from Pomeroy et al., with our zebrafish brain tumors.
Log-fold-change derived from the differential expression analysis between tp53-/- to tp53-/-; TP53P153Δ was used with GSEApreranked function (GSEA, v4.0.3, Broad Institute, MA).
To further assess the unexpected finding of tumors of neural origin in the central nervous system in our models, we expressed kRASG12D with either DsRED or GFP in tp53-/-, tp53-/- + TP53C176F and tp53-/- + TP53P153Δ backgrounds (Figure 5—figure supplement 2). As previously observed, head tumors were enriched in the tp53-/- + TP53P153Δ group and histological analyses revealed that the majority of these tumors had histology consistent with ERMS (Figure 5—figure supplement 3). However, we did notice that approximately 20% (5 of 24) displayed histology consistent with medulloblastoma (Figure 5C–E). In contrast, the tp53-/- group had only one zebrafish (1 of 20 tumors) with a DsRED-positive tumor in the head with histology consistent with medulloblastoma (Figure 5C–E). Non-ERMS head tumors highly expressed sox10 and gfap (Supplementary file 2) and immunohistochemistry staining confirmed that these brain tumors express high levels of Sox10 and Gfap (Figure 5F and G). None of the zebrafish expressing tp53-/- + TP53C176F initiated head tumors. Finally, by comparing our zebrafish brain tumors with medulloblastoma gene signatures outlined in Pomeroy et al., and Kool et al., we found that our brain tumors expressing tp53-/- + TP53P153Δ were consistent with the sonic hedgehog subgroup of medulloblastomas (Figure 5H, Figure 5—figure supplement 4; Kool et al., 2014; Pomeroy et al., 2002).
Expression of <italic>TP53<sup>Y220C</sup></italic> predisposes to head ERMS in zebrafish
The location of the deleted proline in p53P153Δ suggested a structural change. We sought to understand the effect of the P153Δ mutation on function by assessing protein structure and stability using in silico modeling. Homology models of p53P153Δ generated by SWISS-MODEL (Waterhouse et al., 2018) indicate that deletion of P153 causes a partial narrowing of a small pocket on the surface of p53WT (Figure 6A). P153 is the C-terminal residue of a tri-proline surfaced-exposed loop that retains flexibility and mobility. Conversely, another known p53 mutation, the p53Y220C mutation, causes expansion and deepening of this pocket by forming a cleft bounded by L145, V147, T150-P153, P222, and P223 (Figure 6A). We searched the TCGA database for other TP53 mutants sequentially and structurally close to P153 and identified proline residues P151 and P152 that are also mutated in cancers. Interestingly, mutations in P151 and P152 are present in a subset of tumors overlapping with the Y220C mutation (TCGA Research Network data; http://cancergenome.nih.gov). Previous molecular dynamics simulations of p53Y220X (X = C, H, N, or S) mutations found that the tri-proline loop (P151-153) is mobile and can precipitate a concerted collapse of the pocket, forming a frequently populated closed state and contributing to p53 instability (Bauer et al., 2020). Since the binding of small molecules in the pocket increases the stability of p53Y220X mutants, it is not inconceivable that mutations lining this pocket such as the tri-proline loop could be involved in fluctuations that decrease structural stability. Similarly, in our static models, the loss of P153 caused this pocket to be slightly occluded, suggesting further destabilization and abnormal p53 function. (Figure 6A). The TP53Y220C mutation is the ninth most frequent TP53 missense mutation and is the most common ‘conformational’ TP53 mutation in cancer (Baud et al., 2018). Interestingly, the TP53Y220C allele is frequently associated with sarcomas and head and neck carcinomas, and has been reported in patients with osteosarcoma and RMS (Castresana et al., 1995; Overholtzer et al., 2003). We therefore hypothesized that loss of P153 via an in-frame deletion might result in similar structural defects as TP53Y220C by contributing to the overall structural instability (Wang and Fersht, 2017). Therefore, we predicted that ERMS tumors expressing TP53Y220C may phenocopy TP53P153Δ with increased ERMS in the head region. To test this hypothesis, we generated ERMS that expressed TP53Y220C with kRASG12D and GFP in tp53-/- embryos (Figure 6—figure supplement 1). Western blot analyses confirmed that mutant p53Y220C protein was expressed (Figure 6B). Kaplan–Meier analyses indicated that while TP53Y220C expression inhibited ERMS initiation (Figure 6C, p<0.0001) and decreased the number of primary ERMS per fish (Figure 6—figure supplement 2A, p<0.0001), TP53Y220C expression also led to a significant increase in head ERMS tumors compared to tp53-/- animals, reminiscent of tp53-/- animals expressing TP53P153Δ (Figure 6D, F and H and Figure 6—figure supplement 2B; 30.8% of ERMS; n = 150). We confirmed ERMS pathology on tumor sections (Figure 6E and G). Histological analyses confirmed ERMS in 11 of 12 tumor burdened animals, with one tumor displaying characteristics of round blue cell tumor (Figure 6—figure supplement 3). We also assessed TP53Y220C-expressing tumors for effects on proliferation and apoptosis and similar to tp53-/-; TP53P153Δ ERMS, expression TP53Y220C decreased tumor cell proliferation (Figure 6J, p<0.0001). However, unlike TP53P153Δ-expressing ERMS, TP53Y220C expression led to a significant increase in apoptosis (Figure 6I, p=0.001). Lastly, we performed qPCR analyses on bulk tumors to assess whether TP53Y220C differentially regulated known targets of wild-type p53, including baxa, cdkn1a, gadd45a, noxa, and puma/bbc3. Compared to tp53+/+ and tp53-/- tumors, we found that tp53-/- + TP53Y220C tumors expressed significantly higher gadd45a, and two of three tumors expressed higher noxa and baxa but not bbc3 or cdkn1a/p21 (Figure 6—figure supplement 4).
<italic>TP53<sup>Y220C</sup></italic> predisposes to head embryonal rhabdomyosarcoma (ERMS) in zebrafish.
(A) Surface representation of p53WT (PDB 2XWR) and p53P153Δ (homology model) showing key residues lining a surface exposed pocket (sticks). The green ovals compare the size and shape of the pocket between the two structures. (B) p53 protein expression levels in tp53-/- fish tumors with or without TP53Y220C, with rhabdomyosarcoma (RMS) cell line, Rh30, as a control. (C) Kaplan–Meier plot showing tumor initiation in tp53-/- fish, with or without TP53Y220C. (D, F) Representative images of tp53-/- fish with ERMS tumors, with or without TP53Y220C (GFP-positive). Dashed region outlines the tumor. The zebrafish in (F) are 35 d. Scale bar in (F) 1 mm. (E, G) Representative H&E staining of tumors in tp53-/- fish, with or without TP53Y220C. Scale bar = 100 µm. (H) Pie chart showing localization of tumors expressed as a percentage found in varying regions of in tp53-/- fish with and without TP53Y220C. Percentage in red indicates a significant difference to tp53-/- (p=0.01928, two-tailed two proportions Z-test). (I) Quantification of Annexin V staining in tumors of tp53-/- fish with or without expression of TP53Y220C. n = 3–4. (J) Quantification of EdU staining in tumors of tp53-/- fish with or without expression of TP53Y220C. n = 4–9.
Schematic of the experimental setup to generate embryonal rhabdomyosarcoma (ERMS) tumors in zebrafish via microinjection of indicated linearized DNA constructs into the one-cell-stage zebrafish embryos generated from in crosses of <italic>tp53<sup>-/-</sup></italic> zebrafish.
<italic>TP53</italic><sup><italic>Y220C</italic></sup> predisposes to head embryonal rhabdomyosarcoma (ERMS) in zebrafish.
(A) Tumor numbers per zebrafish tumor expressing either no TP53 (tp53-/-) or TP53Y220C. n = 89 (tp53-/-), n = 54 (TP53Y220C). (B) Representative images showing variation in tumor localization in fish expressing mutant TP53 (either DsRed positive or GFP positive). Scale bar = 1 mm. Age of the fish 35–45 d.
TP53<sup>Y220C</sup> predisposes to a round blue cell tumor in zebrafish.
(A) H&E section of the head and trunk regions of a tp53-/-;TP53Y220C zebrafish with a round blue cell tumor (×40–60 magnification). (B, C) H&E stained sections at ×20 magnification of tp53-/-;TP53Y220C zebrafish with a round blue cell tumor. Scale bar = 100 µm.
Semi-quantitative qPCR analyses comparing expression of known p53 direct regulated genes <italic>baxa, bbc3</italic>, <italic>cdkn1a, gadd45a,</italic> and <italic>noxa</italic> in primary zebrafish embryonal rhabdomyosarcoma (ERMS) expressing wt <italic>tp53</italic> (blue, n = 3), <italic>tp53<sup>-/-</sup></italic> (red, n = 3), and <italic>tp53<sup>-/-</sup> + TP53<sup>Y220C</sup></italic> (green, n = 3).
Also shown is expression of zebrafish tp53 across the tumors. The wt and tp53-/- tumors are the same controls from Figure 4—figure supplement 3.
Altogether, our data reveal that both the TP53P153Δand TP53Y220C mutations predispose to head ERMS tumors in zebrafish but differ in their effects on tumor initiation, apoptosis and expression of p53 target genes noxa, baxa, bbc3 and gadd45a (SupplementarySupplementary file 1 File 1).
<italic>kdr</italic> downstream of <italic>TP53<sup>P153Δ</sup></italic> predisposes to head ERMS in tp53<sup>-/-</sup> zebrafish
In a mouse genetic model, activated hedgehog signaling can initiate head and neck ERMS from a potential bipotent Kdr-positive progenitor cell that has endothelial and myogenic potential (Drummond et al., 2018). To test whether head ERMS in zebrafish tumors have molecular features distinct from tumors arising in the trunk, we used RNAseq analyses to compare gene expression between sorted secondary head and trunk tumors generated in tp53-/- CG1 syngeneic animals (Supplementary file 3). The head ERMS tumor, while having common shared myogenic features, was distinguishable from trunk tumors and expressed several genes that are normally expressed in the head musculature, including tbx1, dlx3b, dlx4b, and ptch2 (Figure 7A, Supplementary file 3), suggesting that head ERMS are molecularly different from trunk tumors. Mutant TP53 has been previously shown to activate KDR expression in breast cancer cell lines (Pfister et al., 2015) and high KDR expression in a subset of osteosarcoma tumors is associated with poor outcome (Negri et al., 2019). Clinical data from our TP53P153Δ osteosarcoma patient revealed KDR amplification. From the patient-derived xenograft, we confirmed KDR expression using IHC analysis and using western blot found that while commonly used TP53 null osteosarcoma SaOS2 cells do not express KDR, our TP53P153Δ osteosarcoma PDX robustly expressed KDR (Figure 7B and C). Therefore, we hypothesized that loss of kdr expression in tp53-/- zebrafish co-expressing rag2:kRASG12D and rag2:TP53P153Δ would inhibit ERMS initiation in the head. To test this, we injected rag2:kRASG12D, rag2:TP53P153Δ, rag2:GFP together with Cas9 protein and guide RNAs (gRNAs) targeting either kdr (catalytic domain) or mitfa (control) in tp53-/- embryos (Figure 7D and E). We expected that zebrafish would tolerate mosaic loss of kdr due to the presence of kdrl (Bussmann et al., 2008), while mitfa would be required for melanophore specification, but not viability in zebrafish (Lister et al., 2001). We first assessed tumor initiation and found that mosaic targeting of kdr resulted in an overall reduction in tumor initiation compared to mitfa (Figure 7F). Importantly, kdr CRISPR/Cas9 resulted in a significant reduction in head ERMS tumors (Figure 7G–I), while targeting mitfa did not suppress head ERMS initiation. We confirmed efficient CRISPR/Cas9 targeting using PCR, identifying deletions in exons 12 and 13, which contain the kdr catalytic domain (Figure 7J, Figure 7—figure supplement 1). Altogether, our data supports a role for mutant TP53P153Δ functioning in part through kdr to influence the formation of ERMS tumors in head musculature. However, it is important to state that our mosaic analyses are unable to distinguish whether TP53P153Δ regulation of kdr expression is required cell autonomously or in the tumor vasculature.
<italic>kdr</italic> downstream of <italic>TP53<sup>P153Δ</sup></italic> predisposes to head embryonal rhabdomyosarcoma (ERMS) in tp53<sup>-/-</sup> zebrafish.
(A) Volcano plot comparing head to trunk ERMS tumors. Genes with adj. p-value<0.01 and fold-change >2 are colored in red (upregulated in head tumor) or blue (downregulated in head tumor). Table adjacent the plot shows the top differentially expressed genes. (B) Western blot showing p53, KDR, and RUNX2 protein expression in bone marrow mesenchymal cells, the osteosarcoma PDX expressing TP53P153Δ, and osteosarcoma cell lines SaOS2 and 143B. (C) Representative KDR IHC staining of osteosarcoma PDX expressing p53P153Δ. Scale bar = 60 µm. (E) Schematic demonstrating microinjection into the one-cell-stage zebrafish embryos with the indicated constructs to generate GFP-expressing tumors with either kdr or mitfa mosaically deleted in vivo. (F) Kaplan–Meier plot showing ERMS tumor initiation in tp53-/- fish expression TP53P153Δ with either mitfa or kdr ablation. (H) Representative H&E staining of tumors in tp53-/- fish or tp53-/- fish expressingTP53P153Δ with either mitfa or kdr ablation. Scale bar = 60 µm (G) Representative images of tumor localization in tp53-/- fish with expression of TP53P153Δ with either mitfa or kdr deletion. Fish in the panels are 30 days old. (I) Pie chart showing the percentage of tumors found in varying regions of tp53-/- fish with expression of TP53P153Δ with either mitfa or kdr deletion, showing significant differences in the head region. *p=0.0088 (two-tailed two proportions Z-test). (J) Example of sequencing from an ERMS tumor with a deletion in a portion of kdr exon 13. Five of five of ERMS tumors analyzed have deletions in kdr.
kdr downstream of TP53<sup>P153Δ</sup> predisposes to head embryonal rhabdomyosarcoma (ERMS) in tp53-/- zebrafish.
(A) Schematic of the strategy to target exons 12 and 13 in kdr that contains the enzymatic catalytic domain. (B) Sequence of the targeting guide RNAs targeting exon 12/13 in zebrafish kdr. (C) Polyacrylamide gel showing wild type, mutant heteroduplex, and complete deletion DNA bands obtained from a 641 PCR amplicon of the region of exons 12 and 13 of kdr that was targeted for ablation by three guide RNAs from five independent embryonal rhabdomyosarcoma (ERMS) tumors (T1–T5) from the tp53-/- + TP53P153Δ + Cas9/gRNA targeting kdr experimental group. The presence of deletions as visualized by small PCR amplicons corresponding to a loss of 334 bp in a 641 bp wild-type PCR product (deletion mutant amplicon = 212 bp). In addition, presence of heteroduplex products can be seen via PAGE, indicating the presence of indels. (D) Trace DNA sequencing reads from four independent tumors showing that the gRNA used are active and can introduce deletions in exons 12 and 13 of kdr. Indels were confirmed by cloning PCR products and sequencing individual clones. Please see indels +21 bp, –7 bp at the g1 site, 7 bp deletion at the g2 site, and 25 bp deletion at the g3 site. (E) Schematic of the strategy to target exon 2 in mitfa. (F) Sequence of the targeting guide RNAs targeting exon 2 in zebrafish mitfa.
Discussion
The TP53 tumor-suppressor gene is mutated in >40% of human tumors, and patients with Li–Fraumeni syndrome with germline mutations are predisposed to a spectrum of tumors that includes several lethal childhood sarcomas, such as rhabdomyosarcoma and osteosarcoma (Gröbner et al., 2018; Guha and Malkin, 2017). Here, we addressed three poorly understood aspects of TP53 function in ERMS, a devastating pediatric malignancy of the muscle. First, we found that the tp53 pathway is a major suppressor of tumor initiation in RAS-driven ERMS. Second, we establish that human TP53 can complement zebrafish tp53 function. Third, we utilized our robust zebrafish ERMS model to assign function to three mutations whose effects in ERMS were previously unknown. We defined TP53C176F as a hypomorphic allele and TP53P153Δ as pathogenic with gain-of-function activity. We found that the structurally related mutants p53P153Δand p53Y220C predispose to head musculature ERMS (Supplementary file 1). Finally, p53P153Δ also predisposes to hedgehog driven medulloblastomas.
Patients with mutant RAS-driven ERMS have high-risk disease and are a challenge to treat (Chen, 2013; Shern et al., 2014). Murine RAS mutant ERMS models require loss of Trp53 and/or p16 (Kashi et al., 2015), making the zebrafish ERMS model more similar to human in that RAS activation is sufficient to drive tumor formation. However, a maximum of 40% of wild-type kRASG12D-expressing zebrafish initiate tumors by 50 days of life (Langenau et al., 2007). Interestingly, both human and zebrafish ERMS with wild-type TP53 display pathway inhibition with MDM2 amplification and/or overexpression (Chen, 2013; Langenau et al., 2007; Seki et al., 2015; Shern et al., 2014), suggesting a role for p53 pathway suppression during ERMS initiation. Additionally, genes such as TWIST1 have been shown to promote tumor initiation in sarcomas by inhibiting p53 expression (Piccinin et al., 2012). Of note, mutant kRAS-induced ERMS formation increases to approximately 70% in tp53M214K/M214K mutant zebrafish (Berghmans et al., 2005; Langenau et al., 2007), suggesting that tp53 plays an important role in suppressing ERMS initiation.
The tp53-/- complete loss-of-function mutant zebrafish line (Ignatius et al., 2018) spontaneously forms a broader tumor spectrum than tp53M214K/M214K, including MPNSTs, angiosarcomas, NK-cell leukemias, as well as germ cell tumors (Berghmans et al., 2005; Ignatius et al., 2018), suggesting the possibility of differential effects in RAS-driven ERMS. Thus, our study tested the role for loss of tp53 in ERMS initiation by generating kRASG12D-induced ERMS in the tp53-/- background. We found that >97% of animals form ERMS, revealing that tp53 pathway suppression is required for full tumor penetrance. Moreover, when compared to tp53 wild-type, tp53-/- animals display significantly more tumors per animal, increased tumor cell proliferation, and minor effects on tumor cell apoptosis. These effects on apoptosis can be enhanced by wild-type human TP53 or zebrafish tp53 expression specifically in ERMS tumor cells using the zebrafish rag2 promoter. Our previous work showed that tp53 loss specifically enhanced invasion and metastasis in vivo (Ignatius et al., 2018), suggesting that while TP53 is shown to suppress tumor progression by multiple mechanisms depending on tumor type, loss of TP53 in human ERMS may influence clinical outcomes through a combination of increased tumor cell proliferation, invasion, and metastasis (Ignatius et al., 2018).
A second important finding is that human TP53 can complement wild-type tp53 function in zebrafish ERMS, which allows for direct study of different human variants in vivo. Overall human and zebrafish TP53 show a high level of similarity (56% identity), with the DNA-binding domain displaying especially high homology with 72% amino acid identity (79% positive substitutions) and no gaps in the sequence (Figure 3—figure supplements 1 and 2). We found that both wild-type human and zebrafish Tp53 suppresses ERMS initiation and decreases proliferation in the tp53-/- background. However, we observed a significant increase in apoptotic cells only with overexpression, suggesting that higher levels of p53 induction may be required to induce apoptosis in a wild-type background. Our co-expression approach has been previously used to study multiple aspects of tumorigenesis in ERMS, T-cell acute lymphoblastic leukemia (T-ALL), melanoma, liver cancer, and neuroblastoma (Blackburn et al., 2014; Langenau et al., 2007; Lobbardi et al., 2017; White et al., 2013), and results in protein expression comparable to or lower than p53 in human Rh30 RMS cells that endogenously express mutant p53R273C (Gibson et al., 1998). Co-expression has been an effective method to assess the role for multiple signaling pathways in ERMS, including Notch, canonical and non-canonical Wnt, Myf5, and MyoD, with pathway validation in human cells using in vitro and in vivo methods (Chen et al., 2014b; Hayes et al., 2018; Ignatius et al., 2017; Tenente et al., 2017).
Having established that wild-type human TP53 functions similarly to zebrafish tp53 in ERMS, we are now able to assess patient variants identified in sarcomas and define their roles or activity during tumorigenesis (Supplementary file 1). A TP53C176F mutation was identified in a patient with ERMS that displayed functional loss of the second TP53 allele (Chen, 2013). A second TP53P153Δ mutation was found as a germline variant in a patient with osteosarcoma. Finally, the TP53Y220C mutation is commonly found in sarcomas, including RMS and osteosarcoma (Castresana et al., 1995; Overholtzer et al., 2003). Using our in vivo assays, we were able to define TP53C176F as a hypomorph with respect to ERMS initiation that retains the ability to induce apoptosis. These findings are consistent with studies showing that TP53C176F does retain some aspects of wild-type p53 function, can form tetramers with wild-type p53, and can differentially induce TP53 target gene transcription (Hoffman-Luca et al., 2015; Kato et al., 2003). Compounds like ZMC1 that stabilize mutant protein to enhance p53-mediated tumor cell apoptosis and decrease growth (Blanden et al., 2020; Blanden et al., 2015; Yu et al., 2012) selectively increased ERMS apoptosis in tp53-/- + TP53C176F but not in tp53-/- tumors, further highlighting our assay as an effective in vivo drug efficacy screening tool. Altogether, our data shows that TP53C176F is likely hypomorphic and chemical stabilization of the resulting mutant protein could be an effective therapeutic strategy for patients with this allele.
In contrast to TP53C176F or TP53Y220C, the TP53P153Δ variant is rare, with only one other known patient reported (Michalarea et al., 2014). This patient was diagnosed with multiple tumors over several decades, including bilateral breast cancer, malignant fibrous histocytoma, and an EGFR mutant lung adenocarcinoma. The patient’s mother, maternal aunts, and maternal grandmother all experienced early-onset cancers, meeting the criteria for a Li–Fraumeni diagnosis (Michalarea et al., 2014). The specific Proline 153 residue, while in a region that is highly conserved across mammalian species (Figure 3—figure supplement 2), is not conserved in mice or zebrafish, creating a challenge for modeling endogenous p53153 expression in vivo. Due to its rarity in the literature and lack of animal models, genetic testing done through Invitae, a medical genetic testing provider, could not assign pathogenicity to TP53P153Δ; however, the family experience strongly suggested that this is a Li–Fraumeni variant. Clinically, the patient’s osteosarcoma was extremely aggressive and refractory to multiple anticancer agents. Using our in vivo model, we found that TP53P153Δ predisposes to head ERMS that are relatively resistant to apoptosis. Currently, no other gene or pathway has been identified in the zebrafish model that predisposes to head tumors, with both TP53P153Δ and TP53Y220C expression in zebrafish leading to an increase in head ERMS. This reveals two TP53 variants that may have shared gain-of-function effects in ERMS with respect to the initiation of head tumors. Unexpectedly, approximately 20% of tp53-/- zebrafish expressing TP53P153Δ also develop medulloblastoma-like tumors with gene signatures consistent with the sonic hedgehog subgroup. Brain tumors have not been observed in rag2:kRASG12D –induced ERMS model, and we did not identify any brain tumors in our TP53Y220C or TP53C176F experimental groups. We did identify however one medulloblastoma-like tumor in tp53-/- zebrafish. Given the rapid onset of ERMS is our model, it is possible that medulloblastoma requires more time to initiate and that removal of ERMS burdened animals from the experiment before 30 days precludes study of additional tumor types in tp53-/- animals. Tumors in the tp53-/- + TP53C176F group grew slowly and experimental animals were followed for up to 90 days, yet no medulloblastoma tumors were observed, suggesting that hypomorphic TP53P153Δ activity is likely not sufficient to initiate medulloblastoma tumors, in addition to head and neck ERMS.
Head and neck ERMS represent a significant proportion of all ERMS diagnoses, with the basis of regional predispositions not fully understood. Expression of activated Smoothened protein in mice under the control of the Ap2 promoter results primarily in head and neck ERMS that arise from Kdr-positive endothelial cells (Hatley et al., 2012). Smoothened is a key component of the hedgehog signaling pathway that is commonly activated in ERMS (Drummond et al., 2018; Satheesha et al., 2016). Modeling in mice indicates that the cell of origin plays a major role in initiation, given that activated Smoothened under the control of a more ubiquitously expressed promoter leads to ERMS tumors in other skeletal muscle populations (Mao et al., 2006). Importantly, a recent analysis of TP53 in RMS found a higher proportion of head and neck RMS tumors have TP53 mutations (Shern et al., 2021). Our data suggests that differential features of TP53P153Δ and TP53Y220C may predispose patients to head ERMS via regulation of kdr expression, providing additional mechanistic insights into anatomical differences in sarcoma initiation. Therefore, certain TP53 variants may predispose individuals to head and neck rhabdomyosarcoma tumors; however, the cell of origin for zebrafish head ERMS remains to be identified, as well as whether mutant TP53 interacts with hedgehog pathway signaling in head ERMS and medulloblastoma.
Finally, we would like to state the limitation of our zebrafish ERMS model. In Li–Fraumeni patients, TP53 mutations are germline with somatic modifier mutations occurring secondarily. In our experiments, while our starting point is a tp53-/- null background, kRASG12D and the human TP53 variant allele are introduced at the same time under the control of the rag2 promoter. A second limitation is that effects of mutant TP53 in the tumor microenvironment are missed in our assays. However, our assay is rapid, given that the generation of zebrafish strains with germline knock-in of the equivalent mutations takes an average of 2 years for full assessment. In mice, only the most common Tp53 variants have been fully characterized, and prioritizing which TP53 mutants to model is a challenge given cost and time restraints. Our mosaic assay bridges a gap, especially for sarcomas where the majority of TP53 mutations remain uncharacterized. Our finding that expression of three different patient TP53 variants leads to very different effects on tumor formation highlights the importance of in vivo precision modeling, with our model promising to help further define patient-specific p53 biology. In summary, we highlight that the zebrafish ERMS model can be effective in defining multiple aspects of p53 tumor suppressor function and delineating a spectrum of null, partial loss- and gain-of-function mutational effects in vivo.
Materials and methodsAnimals
Animal studies were approved by the UT Health San Antonio Institutional Animal Care and Use Committee (IACUC) under protocol #20150015AR (mice) and 20170101AR (zebrafish). Zebrafish strains used in this work include AB wild type, AB/tp53-/-, CG1 wild type, CG1 /tp53-/-, Tg casper; myf5:GFP. Zebrafish were housed in a facility on a continuous flow system (Aquarius) with temperature-regulated water (~28.5°C) and a 14 hr light–10 hr dark cycle. Adult fish were fed twice daily with brine shrimp, supplemented with solid food (Gemma). Larval fish were kept off the continuous flow system until 15 days post-fertilization and supplemented with a paramecium/algae culture before transferred online.
Generation of zebrafish rhabdomyosarcoma tumors
rag2:kRASG12D and rag2:DsRed plasmids were linearized with XhoI, followed by phenol–chloroform extraction and ethanol precipitation. The purified DNA was resuspended in nuclease-free water (AM9916, Thermo Fisher) and injected into embryos at the one-cell stage of development (40 ng/μl of rag2:kRASG12D and 20 ng/μl of rag2:DsRed). For expression of zebrafish tp53 or human TP53 in ERMS tumors in zebrafish, XhoI-linearized rag2:kRASG12D and rag2:DsRed plasmids were injected along with Xmn1-linearized rag2:TP53 (human) or rag2:tp53 (zebrafish) constructs. For expression of mutant human TP53 experiments, tumors were generated by injecting XhoI-linearized rag2:kRASG12D and rag2:DsRed or rag2:GFP along with Xmn1-linearized rag2:TP53C176F, or TP53P153Δ, or TP53Y220C (35 ng/μl of rag2:kRASG12D, 15 ng/μl rag2:TP53 wild-type or mutant TP53 and 10 ng/μl of rag2:DsRed) into the one-cell stage of zebrafish embryos <1 hr post-fertilization (Ignatius et al., 2012; Langenau et al., 2007). For testing the role of kdr on head ERMS initiation TP53 experiments, tumors were generated by injecting XhoI-linearized rag2:kRASG12D, rag2:GFP and TP53P153Δ along with a CAS9 protein + guide RNA complex (35 ng/μl of rag2:kRASG12D, 15 ng/μl rag2:TP53P153Δ, 10 ng/μl of rag2:DsRed and 200 ng/μl CAS9 + 100 ng/μl gRNA1 kdr + 100 ng/μl gRNA2 kdr + 100 ng/μl gRNA3 kdr or 100 ng/μl mitf) into the one-cell stage of zebrafish embryos <1 hr post-fertilization. Animals were allocated into experimental groups based on which injection cocktail they received during microinjection. Animals were monitored for tumor onset beginning at 10 d post-fertilization by scoring for DsRed or GFP fluorescence under an Olympus MVX10 stereomicroscope with an X-Cite series 120Q fluorescence illuminator. Scoring for tumor initiation was conducted for 60 days. The tumor onset was visualized by Kaplan–Meier plots, and the significance was analyzed by log-rank (Mantel–Cox) test for each two groups at a time. Sample size was determined taking into account the observed tumor initiation rates from previous studies using the zebrafish ERMS model (Langenau et al., 2008; Tenente et al., 2017).
Tumor size (ratio) measurements
Tumor size was measured using FIJI image analysis software (https://fiji.sc/) by comparing area occupied by the tumor region to the total body area of the fish. Specifically, all tumor-burdened fish were imaged in similar lateral position under both bright-field (white) and fluorescent light. The final representative images were generated by superimposing bright-field and fluorescence images using Adobe Photoshop. Significance was assessed by Student’s t-test.
EdU staining
5-ethynyl-2′-deoxyuridine (EdU, Molecular Probes, Life Technologies) was dissolved in DMSO to make a 10 mM stock solution. This stock solution was further diluted 50 times using PBS to 200 μM, and 0.15 μl and 0.3 μl were injected into juvenile and older zebrafish, respectively. After 6 hr EdU treatment, tumor fish were euthanized with an overdose of MS-222 (Tricaine) and fixed in 4% paraformaldehyde (PFA) at 4°C overnight. Then, the fixed samples were soaked using 25% sucrose (in PBS) at 4°C overnight. Finally, the samples were embedded in OCT medium and the medium was allowed to solidify on dry ice. Tissue was sectioned using a Leica CM1510 S Cryostat and tissue sections were placed on plus gold microscope slide (Fisherbrand). The slides were post-fixed in 4% PFA for 15 min and permeabilized in 0.5% Trition x-100 in BPS for 20 min. Then, freshly made Click-iT Plus reaction cocktail was added to each slide and allowed to incubate for 30 min. A small amount of Vectashield mounting medium with DAPI was added to the slide and covered with a coverslip. The EdU Click-iT Plus EdU Alexa Fluor 647 Imaging kit (Molecular Probes, Life Technologies) was used for EdU staining. Tissue sections (one section per tumor) were subsequently analyzed using an Olympus FV3000 confocal microscope and percentage of Edu-positive determined by counting the total number of DAPI-positive nuclei in tumor section. Significance was assessed by Student’s t-test. This assay was completed using at least n ≥ 3 biological replicates (multiple primary tumors or transplanted syngeneic zebrafish tumors).
qPCR analysis
Tumors were extracted from euthanized fish and homogenized inside a 1.5 ml Eppendorf tube using a pestle connected to a Pellet Pestle Cordless Motor (DWK Life Sciences Kimble Kontes). Tumor RNA was extracted using NEB’s Monarch RNA mini prep kit. cDNA was synthesized from the extracted RNA using Invitrogen’s First Strand Synthesis cDNA kit. Following cDNA synthesis, real-time qPCR was performed in 384-well plates using Applied Biosystem’s Sybr Green qPCR MasterMix (Comparative Cт (ΔΔCт) method). Results were compiled and analyzed using the QuantStudio 7 Flex system (Applied Biosciences). PCR primers are provided in Supplementary file 4. All assays were completed using technical replicates, with at least three tumors tested per experimental group.
Phospho-histone H3 staining
Fish were fixed in 4% PFA at 4°C overnight. The fixed fish were subsequently soaked in 25% sucrose overnight and then embedded in OCT medium before being sectioned at 10 μm with a Leica CM1510 S cryostat. After being washed three times in PBST (0.1% Triton X-100 and 0.1% Tween 20 in PBS), the sections were incubated in blocking solution (2% horse serum, 10% FBS, 0.1% Triton X-100, 0.1% Tween 20, 10% DMSO in PBS) for 60 min. The sections were then incubated with rabbit anti-phospho-histone H3 (Ser10) primary antibody (1:500 dilution) at 4°C overnight. The following day, sections were washed three times in PBST and incubated with Alexa Fluor 647 conjugated anti-rabbit secondary antibody at room temperature for 2 hr. Vectashield mounting medium with DAPI was added to the slide and then a coverslip was placed over the sample. The slides were dried in the dark and sealed by nail polish. Sections were imaged using an Olympus FV3000 confocal microscope. Significance was assessed by Student’s t-test. This assay was completed using at least three biological replicates and counting total number of positive cells per a single confocal image taken at ×200 magnification of a tumor section (n ≥ 3 primary tumors).
Annexin V-FITC/PI staining
Fish were euthanized and tumor was isolated, following which the tumor was homogenized in 0.9× PBS + 5% FBS manually using a razor blade and made into single suspensions using 45 micron filters. Tumor cells were washed with 0.9× PBS + 5% FBS and resuspended in the binding buffer containing Annexin V-FITC and propidium iodide for 15 min in the dark at room temperature. Then the cells were detected by flow cytometry (FCM, FACS Canto, BD, CA). Significance was assessed by Student’s t-test. This assay was completed using at least three biological replicates (n ≥ 3).
Histology and immunohistochemistry
Euthanized zebrafish were fixed in 4% PFA overnight at 4°C. Embedding, sectioning, and immunohistochemical analysis of zebrafish sections were performed as previously described (Chen et al., 2014b; Ignatius et al., 2012). H&E staining was performed at the Greehey CCRI histology core. Slides were imaged using a Motic EasyScan Pro slide scanner. Pathology review and staging were completed by board-certified pathologists (EYC and ARG).
Cloning <italic>TP53</italic> wild-type and mutants constructs
Wild-type TP53 from both human (Addgene plasmid #69003) or zebrafish (3-day-old embryos cDNA) were amplified by PCR and cloned into pENTR-D-TOPO vector, which was verified by DNA sequencing. rag2:TP53 (human) or rag2:tp53 (zebrafish) plasmids were generated by one-step Gateway reaction between a Gateway-compatible plasmid with the zebrafish rag2 promoter flanked by attR sites and the respective pENTR-D-TOPO plasmid. TP53C176F, TP53P153Δ, or TP53Y220C fragments were constructed by amplifying human TP53 as two separate fragments with the respective mutations, with the 3′ end of the first fragment possessing 60 bp of homology with the 5′ end of the second fragment. These two fragments were purified from a 1% agarose gel using a Macherey–Nagel purification kit and spliced together using overlap extension PCR with Phusion high fidelity DNA polymerase (Szymczak-Workman et al., 2012). The entire spliced fragment was then blunt-ligated into a pENTR-D-TOPO vector. The TP53 insert was sequenced after which it was cloned into a Gateway-compatible plasmid with the zebrafish rag2 promoter flanked by attR sites using a one-step Gateway reaction using Gateway LR Clonase Enzyme mix. All other PCR amplification was carried out using Q5 high-fidelity DNA polymerase.
Western blot analyses
Western blot analysis on fish tumors was performed by first extracting tumor from fish and homogenizing tumor cells suspended in SDS lysis buffer inside a 1.5 ml Eppendorf tube using a pestle connected to a cordless motor. The total protein concentration for each lysate solution was normalized with a BCA assay kit (Thermo Fisher Scientific, Carlsbad, CA). Then, 40 μg of total protein was run on a 10% SDS/PAGE gel. The protein transferred membrane were blocked using 5% fat-free milk in TBST, followed by incubation in the appropriate antibody. The list of antibodies and where they were obtained from is provided in Supplementary file 5.
Identifying <italic>TP53</italic> mutation status in osteosarcoma PDX sample
DNA sample from osteosarcoma PDX was isolated using the QIAGEN Puregene Core Kit A. PDX sample was first homogenized in lysis buffer, followed by heating at 65°C for 30 min. Thereafter, RNAse A was added to the sample and incubated for 30 min at 37°C. Protein precipitation buffer was used to precipitate protein, and the sample was vortexed and then centrifuged at 15,000 RPM for 3 min. The supernatant was then removed, and isopropanol was added to precipitate genomic DNA, following which the sample was centrifuged at 15,000 RPM for 2 min at 4°C. The supernatant was then drained, and the DNA pellet was washed with 70% ethanol, followed by centrifugation at 15,000 RPM for 1 min. The DNA pellet was then resuspended in DNA hydration solution, incubated at 65°C for 1 hr to dissolve DNA, and then incubated at 22°C for 1 hr. This precipitated DNA was used as a template in a PCR reaction using TP53-specific primers flanking the A159 and P153 residues. The PCR amplicon was ligated into a pCR4-TOPO blunt cloning vector, and the ligation mix was transformed into DH5α chemically competent cells. Plasmid DNA was isolated from transformed colonies and sequenced using M13 primers.
ZMC1 treatment
Tumor-burdended fish were incubated in fish water containing ZMC1 at 70 nM concentration with another batch of tumor-bound fish incubated in 0.1% v/v DMSO (diluted in fish water) as control (Supplementary file 6). Fish were imaged every week on an Olympus MVX10 stereomicroscope with an X-Cite series 120Q fluorescence illuminator, and size of tumors was measured from the acquired images using FIJI software.
FACS sorting
FACSorting was completed essentially as previously described (Chen et al., 2014a; Ignatius et al., 2012; Ignatius et al., 2018). Tumor-burdened syngeneic CG1 zebrafish with secondary head tumors expressing kRASG12D; TP53P153Δ; GFP were euthanized and tumor isolated and made into single-cell suspensions in 0.9× PBS containing 5% FBS (Chen et al., 2014a; Ignatius et al., 2012; Ignatius et al., 2018). Live tumor cells were stained with DAPI to exclude dead cells and sorted twice using a Laser BD FACSAria II Cell Sorter. Sort purity and viability were assessed after two rounds of sorting, exceeding 85% respectively. GFP+ tumor cells were isolated by FACS from secondary tumor transplanted fish, cells sorted and fixed in RNAlater and/or Trizol and RNA isolated for RNA sequencing.
RNAseq and bioinformatic analyses
Total RNA was isolated from the six zebrafish tumors using NEB’s Monarch RNA mini prep kit (NEB Inc, MA) according to the manufacturer’s instructions. DNase (Thermo Scientific, Rockford, IL, Cat# EN0521) was added to the first wash solution at 10 µg/70 µl and incubated for 15 min at room temperature to remove genomic DNA contamination. The quality of RNA samples was analyzed with a Bioanalyzer (Agilent 2100 Bioanalyzer, Agilent Technologies, Santa Clara, CA) by the GCCRI Genome Sequencing Facility. Samples with RNA Integrity Number (RIN) ≥ 7 were used for RNA sequencing library construction using TruSeq Stranded mRNA Library Prep kit according to the manufacturer’s protocol (Cat# RS122-2002; Illumina, Inc). Samples were sequenced in the Illumina HiSeq 3000 (Illumina, Inc) using a 50 bp single-read sequencing protocol. All sequence reads were aligned to the UCSC zebrafish genome build danRer11 using TopHat2 and expression quantification with HTSeq-count (Anders et al., 2015) to obtain the read counts per gene in all samples. The RNAseq data was deposited on GEO (GSE213869). Differential expression analysis was performed using R package DESeq (Anders and Huber, 2010). Statistically and biologically significant DE genes (DEGs) were defined by applying the following stipulations: adjusted p-values<0.01 and fold change ≥2. Gene Ontology enrichment from the DEG list was performed using TopGO package (Alexa et al., 2006), and significantly enriched functions were selected based on classic Fisher’s exact test p-value<0.05. Gene set enrichment analysis (GSEA) was performed using standard alone algorithm (http://www.gsea-msigdb.org/gsea/index.jsp, Subramanian et al., 2005) or R package (R/fgsea, Korotkevich et al., 2019). Differential expression was visualized, such as the volcano plot, using R (https://www.r-project.org).
Homology modeling
The p53P153Δ mutant was modeled using the SWISS-MODEL (Schwede et al., 2003) homology modeling server. The mutant p53 sequence and the p53WT crystal DNA-binding domain crystal structure 2XWR (Natan et al., 2011) were used as input files. The homology models are built, scored, and selected using statistical potentials of mean force scoring methods. Side-chain rotomers for non-conserved residues are selected through a local energy minimization function, and the final models are subjected to global energy minimization using the CHARM22/CMAP forcefield (Waterhouse et al., 2018).
Strategy and quantification of mosaic ablation of <italic>kdr</italic> and <italic>mitfa</italic> in zebrafish ERMS tumors using CRISPR/Cas9 reagents
Using Benchling, three high-scoring CRISPR guides (with minimal off-target scores) were selected to target DNA sequences in exons 12 and 13 of kdr that encode parts of two Ig repeats in the extracellular domain (aa 535–647; InterPro domain IPR003599). Efficient nuclease activity of guides 1 and 3 would result in a 334 bp deletion, leading to a predicted frameshift and premature termination of the polypeptide sequence. When this larger deletion does not result from synchronous CRIPSR nuclease activity of guides 1 and 3, similar consequences are likely to result due to independent indels generated from nuclease activity and NHEJ repair at each of the three individual target sites of guides 1, 2, and 3. This region being upstream of the transmembrane domain (aa 773–795), we predict that indels or deletions leading to frameshift mutations would, at best, produce a partial extracellular domain containing polypeptide incapable of transducing signals into the cell.
Guide
Target sequence (PAM)
Benchling (Hsu et al., 2013) off-target score
g1
GGTAGCGATGCACCTGTATA (GGG)
95.9
g2
CTATAACTTGCGCTGGTATC (GGG)
96.4
g3
CCTGGCAAGCGTACAAGCCT (TGG)
98.4
Using PCR and TBE-PAGE, the presence of deletions was visualized by small PCR amplicons corresponding to a loss of 334 bp in a 641 bp wild-type PCR product (deletion mutant amplicon = 212 bp). In addition, presence of heteroduplex products was observed via PAGE, indicating the presence of indels. Indels were confirmed by cloning PCR products and sequencing individual clones. Shown in Figure 7—figure supplement 1D are indels +21 bp, –7 bp at the g1 site, 7 bp deletion at the g2 site, and 25 bp deletion at the g3 site. The effectiveness of the three sgRNA/CRISPR guides by way of the spectrum of mutations they cause suggests that despite the mosaicism of the mutations in sgRNA-Cas9-injected embryos, tumors that arise harbor a relatively high-frequency loss-of-function mutations in the kdr gene in all tumors. mitfa was selected as a control gene because loss-of-function mutations can be easily screened by looking for the loss of pigmentation in developing larvae using brightfield microscopy. In addition to not having an effect on embryo survival and development, it obviates the need to obtain sequence data to confirm the presence of loss-of-function mutations in the embryos.
Additional informationCompeting interests
No competing interests declared
Author contributions
Data curation, Formal analysis, Validation, Investigation, Visualization, Methodology, Writing – review and editing
Data curation, Formal analysis, Validation, Investigation, Visualization, Methodology, Writing – review and editing
Data curation, Investigation, Writing – review and editing
Formal analysis, Validation, Visualization, Methodology, Writing – review and editing
Data curation, Formal analysis, Investigation, Writing – review and editing
Conceptualization, Data curation, Software, Formal analysis, Validation, Investigation, Methodology
Data curation, Software, Formal analysis, Validation, Investigation, Visualization, Writing – review and editing
Data curation, Formal analysis, Investigation, Methodology
Data curation, Formal analysis, Visualization, Methodology, Writing – review and editing
Data curation, Writing – review and editing
Data curation, Methodology, Writing – review and editing
Supervision, Writing – review and editing
Validation, Investigation, Visualization, Writing – review and editing
Data curation, Formal analysis, Methodology
Resources, Data curation, Investigation, Writing – review and editing
Supervision, Writing – review and editing
Resources, Supervision, Writing – review and editing
Resources, Data curation, Software, Formal analysis, Investigation, Methodology, Writing – review and editing
Data curation, Formal analysis, Investigation, Methodology, Writing – review and editing
Formal analysis, Validation, Writing – review and editing
Resources, Formal analysis, Validation, Visualization, Writing – review and editing
Conceptualization, Formal analysis, Supervision, Funding acquisition, Writing – original draft, Project administration, Writing – review and editing
Ethics
Human subjects: Patient presenting with osteosarcoma signed a Consent to be part of a Repository, Epidemiology of Cancer in Children, Adolescents and Adults. In brief, this allowed for the storage of tissue, cataloging of medical information, and for research to be conducted from collected samples. The study's IRB number is HSC20080057H. Patient was informed of the risks and benefits. The umbrella study covering epidemiological study and patient-derived xenograft generation is IRB approved through UT Health San Antonio.
Animal studies were approved by the UT Health San Antonio Institutional Animal Care and Use Committee (IACUC) under protocol #20150015AR (mice) and #20170101AR (zebrafish). Zebrafish images were taken with specimens under tricaine anesthesia. Zebrafish tumor extraction was performed by administering high dose tricaine to minimize suffering.
Additional files
Table summarizing the main findings for wild-type and mutant <italic>TP53</italic> on tumor initiation, localization, proliferation, apoptosis, and ability to transactivate gene expression.
Table with RNAseq data for gene expression comparing three <italic>kRAS<sup>G12D</sup>;tp53<sup>-/-</sup>TP53<sup>P153Δ</sup>;GFP-positive</italic> tumors to three <italic>kRAS<sup>G12D</sup>; tp53<sup>-/-</sup>;GFP-positive</italic> tumors.
Table with RNAseq data for gene expression comparing one head to two trunks ERMS expressing <italic>kRAS<sup>G12D</sup>; tp53<sup>-/-</sup>;GFP.</italic>
List of primers and guide RNAs used in this study.
List of antibodies used in this study.
List of reagents used in this study.
Source data containing full immunoblots for the data presented in the figures.
Data availability
Data sets were submitted to Dryad, available here: https://doi.org/10.5061/dryad.zgmsbccb6.
HenschNRChenJBaxiKLipsittAEWangLBaudinARobledoDGBandyopadhyayAChallaAGilbertARTomlinsonGEHoughtonPChenEYLibichDSIgnatiusMSModiPZhaoXRKurmashevDChenYHayesMNSreenivasP2023Defining function of wild-type and three patient specific TP53 mutations in a zebrafish model of embryonal rhabdomyosarcomaDryad Digital Repository10.5061/dryad.zgmsbccb6
Acknowledgements
This project has been funded with federal funds from NIH grants MI and PH (R00CA175184), Cancer Prevention & Research Institute of Texas (CPRIT)-funded Scholar grant to MI (RR160062). JC was supported by Wenzhou Medical University young scientist training program (2019) and the research program (KYYW202203). DSL was supported by the St. Baldricks Foundation and the Welch Foundation. DSL and MI are each recipients of the Max and Minnie Tomerlin Voelcker Fund Young Investigator Awards. KB is a T32 and TL1 fellow (T32CA148724, TL1TR002647). NH was supported by the Greehey CCRI Graduate Student Fellowship and the Cancer Prevention & Research Institute of Texas (CPRIT)-funded Research Training Award (RP 170345). AL was supported by the Cancer Prevention & Research Institute of Texas (CPRIT)-funded Research Training Award (RP 170345) and by a Hyundai Hope On Wheels Young Investigator Grant. RNA sequencing data was generated in the GCCRI Genome Sequencing Facility, which is supported by GCCRI, NIH-NCI P30 CA054174 (NCI Cancer Center Support Grant UT Health San Antonio), NIH Shared Instrument grant 1S10OD030311-01, and CPRIT Core Facility Award RP160732.
ReferencesAlexaARahnenführerJLengauerT2006Improved scoring of functional groups from gene expression data by Decorrelating GO graph structureBioinformatics221600160710.1093/bioinformatics/btl14016606683AndersSHuberW2010Differential expression analysis for sequence count dataGenome Biology11R10610.1186/gb-2010-11-10-r10620979621AndersSPylPTHuberW2015Htseq--a python framework to work with high-throughput sequencing dataBioinformatics3116616910.1093/bioinformatics/btu638AttardiLDDonehowerLA2005Probing P53 biological functions through the use of genetically engineered mouse modelsMutation Research/Fundamental and Molecular Mechanisms of Mutagenesis57642110.1016/j.mrfmmm.2004.08.022BaudMGJBauerMRVerduciLDinglerFAPatelKJHoril RoyDJoergerACFershtAR2018Aminobenzothiazole derivatives stabilize the Thermolabile P53 cancer mutant Y220C and show anticancer activity in P53-Y220C cell linesEuropean Journal of Medicinal Chemistry15210111410.1016/j.ejmech.2018.04.03529702446BauerMRKrämerASettanniGJonesRNNiXKhan TarequeRFershtARSpencerJJoergerAC2020Targeting cavity-creating P53 cancer mutations with small-molecule stabilizers: the Y220X paradigmACS Chemical Biology1565766810.1021/acschembio.9b0074831990523BerghmansSMurpheyRDWienholdsENeubergDKutokJLFletcherCDMMorrisJPLiuTXSchulte-MerkerSKankiJPPlasterkRZonLILookAT2005Tp53 mutant Zebrafish develop malignant peripheral nerve sheath tumorsPNAS10240741210.1073/pnas.0406252102BlackburnJSLiuSWilderJLDobrinskiKPLobbardiRMooreFEMartinezSAChenEYLeeCLangenauDM2014Clonal evolution enhances leukemia-Propagating cell frequency in T cell acute Lymphoblastic leukemia through AKT/Mtorc1 pathway activationCancer Cell2536637810.1016/j.ccr.2014.01.03224613413BlandenARYuXWolfeAJGilleranJAAugeriDJO’DellRSOlsonECKimballSDEmgeTJMovileanuLCarpizoDRLohSN2015Synthetic Metallochaperone Zmc1 Rescues mutant P53 conformation by transporting zinc into cells as an IonophoreMolecular Pharmacology8782583110.1124/mol.114.09755025710967BlandenARYuXBlayneyAJDemasCHaJ-HLiuYWithersTCarpizoDRLohSN2020Zinc shapes the folding landscape of P53 and establishes a pathway for Reactivating structurally diverse cancer MutantseLife9e6148710.7554/eLife.6148733263541BussmannJLawsonNZonLSchulte-MerkerSZebrafish NomenclatureC2008Zebrafish VEGF receptors: a guideline to nomenclaturePLOS Genetics4e100006410.1371/journal.pgen.100006418516225CaseyDLPitterKLWexlerLHSlotkinEKGuptaGPWoldenSL2021Tp53 mutations increase Radioresistance in Rhabdomyosarcoma and Ewing sarcomaBritish Journal of Cancer12557658110.1038/s41416-021-01438-234017087CastresanaJSRubioMPGómezLKreicbergsAZetterbergABarriosC1995Detection of Tp53 Gene mutations in human SarcomasEuropean Journal of Cancer31A73573810.1016/0959-8049(95)00121-x7640047ChenX2013Jude children’s research hospital-Washington University pediatric cancer genome, PCancer Cell2471072410.1016/j.ccr.2013.11.002ChenXBahramiAPappoAEastonJDaltonJHedlundE2014aJude children’s research hospital-Washington University pediatric cancer genomeCell Reports710411210.1016/j.celrep.2014.03.003ChenEYDeRanMTIgnatiusMSGrandinettiKBClaggRMcCarthyKMLobbardiRMBrockmannJKellerCWuXLangenauDM2014bGlycogen synthase kinase 3 inhibitors induce the Canonical WNT/beta-Catenin pathway to suppress growth and self-renewal in embryonal RhabdomyosarcomaPNAS1115349535410.1073/pnas.1317731111DrummondCJHannaJAGarciaMRDevineDJHeyranaAJFinkelsteinDRehgJEHatleyME2018Hedgehog pathway drives fusion-negative Rhabdomyosarcoma initiated from non-Myogenic endothelial ProgenitorsCancer Cell3310812410.1016/j.ccell.2017.12.00129316425GarciaPBAttardiLD2014Illuminating P53 function in cancer with genetically engineered mouse modelsSeminars in Cell & Developmental Biology27748510.1016/j.semcdb.2013.12.01424394915GibsonAAHarwoodFGTillmanDMHoughtonJA1998Selective sensitization to DNA-damaging agents in a human Rhabdomyosarcoma cell line with inducible wild-type P53 overexpressionClinical Cancer Research41451529516963GröbnerSNWorstBCWeischenfeldtJBuchhalterIKleinheinzKRudnevaVAJohannPDBalasubramanianGPSegura-WangMBrabetzSBenderSHutterBSturmDPfaffEHübschmannDZipprichGHeinoldMEilsJLawerenzCErkekSLamboSWaszakSBlattmannCBorkhardtAKuhlenMEggertAFuldaSGesslerMWegertJKapplerRBaumhoerDBurdachSKirschner-SchwabeRKontnyUKulozikAELohmannDHettmerSEckertCBielackSNathrathMNiemeyerCRichterGHSchulteJSiebertRWestermannFMolenaarJJVassalGWittHBurkhardtBKratzCPWittOvan TilburgCMKrammCMFleischhackGDirksenURutkowskiSFrühwaldMvon HoffKWolfSKlingebielTKoscielniakELandgrafPKosterJResnickACZhangJLiuYZhouXWaandersAJZwijnenburgDARamanPBrorsBWeberUDNorthcottPAPajtlerKWKoolMPiroRMKorbelJOSchlesnerMEilsRJonesDTWLichterPChavezLZapatkaMPfisterSMICGC PedBrain-Seq ProjectICGC MMML-Seq Project2018Author correction: the landscape of Genomic alterations across childhood cancersNature559E1010.1038/s41586-018-0167-229875405GuhaTMalkinD2017Inherited Tp53 mutations and the Li-Fraumeni syndromeCold Spring Harbor Perspectives in Medicine7a02618710.1101/cshperspect.a026187HatleyMETangWGarciaMRFinkelsteinDMillayDPLiuNGraffJGalindoRLOlsonEN2012A mouse model of Rhabdomyosarcoma originating from the Adipocyte lineageCancer Cell2253654610.1016/j.ccr.2012.09.00423079662HayesMNMcCarthyKJinAOliveiraMLIyerSGarciaSPSindiriSGryderBMotalaZNielsenGPBorgJ-Pvan de RijnMMalkinDKhanJIgnatiusMSLangenauDM2018Vangl2/Rhoa signaling pathway regulates stem cell self-renewal programs and growth in RhabdomyosarcomaStem Cell2241442710.1016/j.stem.2018.02.00229499154Hoffman-LucaCGYangCYLuJZiazadehDMcEachernDDebusscheLWangS2015Significant differences in the development of acquired resistance to the Mdm2 inhibitor Sar405838 between in vitro and in vivo drug treatmentPLOS ONE10e012880710.1371/journal.pone.012880726070072HsuPDScottDAWeinsteinJARanFAKonermannSAgarwalaVLiYFineEJWuXShalemOCradickTJMarraffiniLABaoGZhangF2013DNA targeting specificity of RNA-guided Cas9 NucleasesNature Biotechnology3182783210.1038/nbt.264723873081IgnatiusMSChenEElpekNMFullerAZTenenteIMClaggRLiuSBlackburnJSLinardicCMRosenbergAENielsenPGMempelTRLangenauDM2012In vivo imaging of tumor-Propagating cells, regional tumor heterogeneity, and dynamic cell movements in embryonal RhabdomyosarcomaCancer Cell2168069310.1016/j.ccr.2012.03.04322624717IgnatiusMSHayesMNLobbardiRChenEYMcCarthyKMSreenivasPMotalaZDurbinADMolodtsovAReederSJinASindiriSBeleyeaBCBhereDAlexanderMSShahKKellerCLinardicCMNielsenPGMalkinDKhanJLangenauDM2017The Notch1/Snail1/Mef2C pathway regulates growth and self-renewal in embryonal RhabdomyosarcomaCell Reports192304231810.1016/j.celrep.2017.05.06128614716IgnatiusMSHayesMNMooreFETangQGarciaSPBlackburnPRBaxiKWangLJinARamakrishnanAReederSChenYNielsenGPChenEYHasserjianRPTirodeFEkkerSCLangenauDM2018Tp53 deficiency causes a wide tumor spectrum and increases embryonal Rhabdomyosarcoma metastasis in ZebrafisheLife7e3720210.7554/eLife.3720230192230KashiVPHatleyMEGalindoRL2015Probing for a deeper understanding of Rhabdomyosarcoma: insights from complementary model systemsNature Reviews. Cancer1542643910.1038/nrc396126105539KastenhuberERLoweSW2017Putting P53 in contextCell1701062107810.1016/j.cell.2017.08.02828886379KatoSHanSYLiuWOtsukaKShibataHKanamaruRIshiokaC2003Understanding the function-structure and function-Mutation relationships of P53 tumor Suppressor protein by high-resolution Missense Mutation analysisPNAS1008424842910.1073/pnas.1431692100KoLJPrivesC1996P53: puzzle and paradigmGenes & Development101054107210.1101/gad.10.9.1054KoolMJonesDTWJägerNNorthcottPAPughTJHovestadtVPiroRMEsparzaLAMarkantSLRemkeMMildeTBourdeautFRyzhovaMSturmDPfaffEStarkSHutterSŞeker-CinHJohannPBenderSSchmidtCRauschTShihDReimandJSieberLWittmannALinkeLWittHWeberUDZapatkaMKönigRBeroukhimRBergtholdGvan SluisPVolckmannRKosterJVersteegRSchmidtSWolfSLawerenzCBartholomaeCCvon KalleCUnterbergAHerold-MendeCHoferSKulozikAEvon DeimlingAScheurlenWFelsbergJReifenbergerGHasselblattMCrawfordJRGrantGAJabadoNPerryACowdreyCCroulSZadehGKorbelJODozFDelattreOBaderGDMcCabeMGCollinsVPKieranMWChoY-JPomeroySLWittOBrorsBTaylorMDSchüllerUKorshunovAEilsRWechsler-ReyaRJLichterPPfisterSM2014Genome sequencing of SHH Medulloblastoma predicts genotype-related response to Smoothened inhibitionCancer Cell2539340510.1016/j.ccr.2014.02.004KorotkevichGSukhovVBudinNShpakBArtyomovMNSergushichevA2019Fast Gene Set Enrichment AnalysisbioRxiv10.1101/060012LangenauDMKeefeMDStorerNYGuyonJRKutokJLLeXGoesslingWNeubergDSKunkelLMZonLI2007Effects of RAS on the genesis of embryonal RhabdomyosarcomaGenes & Development211382139510.1101/gad.154500717510286LangenauDMKeefeMDStorerNYJetteCASmithACHCeolCJBourqueCLookATZonLI2008Co-injection strategies to modify radiation sensitivity and tumor initiation in transgenic ZebrafishOncogene274242424810.1038/onc.2008.5618345029LevineAJMomandJFinlayCA1991The P53 tumour suppressor geneNature35145345610.1038/351453a02046748ListerJACloseJRaibleDW2001Duplicate Mitf genes in Zebrafish: complementary expression and conservation of Melanogenic potentialDevelopmental Biology23733334410.1006/dbio.2001.037911543618LobbardiRPinderJMartinez-PastorBTheodorouMBlackburnJSAbrahamBJNamikiYMansourMAbdelfattahNSMolodtsovAAlexeGToiberDde WaardMJainEBoukhaliMLionMBhereDShahKGutierrezAStegmaierKSilvermanLBSadreyevRIAsaraJMOettingerMAHaasWLookATYoungRAMostoslavskyRDellaireGLangenauDM2017TOX regulates growth, DNA repair, and Genomic instability in T-cell acute Lymphoblastic leukemiaCancer Discovery71336135310.1158/2159-8290.CD-17-026728974511LozanoG2010Mouse models of P53 functionsCold Spring Harbor Perspectives in Biology2a00111510.1101/cshperspect.a00111520452944MalkinD2011Li-Fraumeni syndromeGenes & Cancer247548410.1177/194760191141346621779515MaoJLigonKLRakhlinEYThayerSPBronsonRTRowitchDMcMahonAP2006A novel somatic mouse model to survey Tumorigenic potential applied to the Hedgehog pathwayCancer Research66101711017810.1158/0008-5472.CAN-06-065717047082MichalareaVCalcasolaMCanePTobalKIzattLSpicerJ2014EGFR-Mutated lung cancer in Li-Fraumeni syndromeLung Cancer8548548710.1016/j.lungcan.2014.06.01725047674MullerPAJVousdenKH2014Mutant P53 in cancer: new functions and therapeutic opportunitiesCancer Cell2530431710.1016/j.ccr.2014.01.02124651012NatanEBalogluCPagelKFreundSMVMorgnerNRobinsonCVFershtARJoergerAC2011Interaction of the P53 DNA-binding domain with its N-terminal extension modulates the stability of the P53 TetramerJournal of Molecular Biology40935836810.1016/j.jmb.2011.03.047NegriGLGrandeBMDelaidelliAEl-NaggarACochraneDLauCCTricheTJMooreRAJonesSJMontpetitAMarraMAMalkinDMorinRDSorensenPH2019Integrative Genomic analysis of matched primary and metastatic pediatric OsteosarcomaThe Journal of Pathology24931933110.1002/path.531931236944OliveKPTuvesonDARuheZCYinBWillisNABronsonRTCrowleyDJacksT2004Mutant P53 gain of function in two mouse models of Li-Fraumeni syndromeCell11984786010.1016/j.cell.2004.11.00415607980OverholtzerMRaoPHFavisRLuX-YElowitzMBBaranyFLadanyiMGorlickRLevineAJ2003The presence of P53 mutations in human Osteosarcomas correlates with high levels of Genomic instabilityPNAS100115471155210.1073/pnas.193485210012972634ParantJMGeorgeSAHoldenJAYostHJ2010Genetic modeling of Li-Fraumeni syndrome in ZebrafishDisease Models & Mechanisms3455610.1242/dmm.00374920075382PetitjeanAAchatzMIWBorresen-DaleALHainautPOlivierM2007Tp53 mutations in human cancers: functional selection and impact on cancer prognosis and outcomesOncogene262157216510.1038/sj.onc.121030217401424PfisterNTFominVRegunathKZhouJYZhouWSilwal-PanditLFreed-PastorWALaptenkoONeoSPBargonettiJHoqueMTianBGunaratneJEngebraatenOManleyJLBørresen-DaleA-LNeilsenPMPrivesC2015Mutant P53 cooperates with the SWI/SNF Chromatin remodeling complex to regulate Vegfr2 in breast cancer cellsGenes & Development291298131510.1101/gad.263202.11526080815PiccininSToninESessaSDemontisSRossiSPecciariniLZanattaLPivettaFGrizzoASonegoMRosanoCDei TosAPDoglioniCMaestroR2012A "twist box" code of P53 inactivation: twist box: P53 interaction promotes P53 degradationCancer Cell2240441510.1016/j.ccr.2012.08.00322975381PomeroySLTamayoPGaasenbeekMSturlaLMAngeloMMcLaughlinMEKimJYHGoumnerovaLCBlackPMLauCAllenJCZagzagDOlsonJMCurranTWetmoreCBiegelJAPoggioTMukherjeeSRifkinRCalifanoAStolovitzkyGLouisDNMesirovJPLanderESGolubTR2002Prediction of central nervous system embryonal tumour outcome based on gene expressionNature41543644210.1038/415436a11807556SatheeshaSManzellaGBovayACasanovaEABodePKBelleRFeuchtgruberSJaaksPDoganNKoscielniakESchäferBW2016Targeting Hedgehog signaling reduces self-renewal in embryonal RhabdomyosarcomaOncogene352020203010.1038/onc.2015.26726189795SchwedeTKoppJGuexNPeitschMC2003SWISS-MODEL: an automated protein Homology-modeling serverNucleic Acids Research313381338510.1093/nar/gkg52012824332SekiMNishimuraRYoshidaKShimamuraTShiraishiYSatoYKatoMChibaKTanakaHHoshinoNNagaeGShiozawaYOkunoYHosoiHTanakaYOkitaHMiyachiMSouzakiRTaguchiTKohKHanadaRKatoKNomuraYAkiyamaMOkaAIgarashiTMiyanoSAburataniHHayashiYOgawaSTakitaJ2015Integrated genetic and epigenetic analysis defines novel molecular subgroups in RhabdomyosarcomaNature Communications6755710.1038/ncomms855726138366ShernJFChenLChmieleckiJWeiJSPatidarRRosenbergMAmbrogioLAuclairDWangJSongYKTolmanCHurdLLiaoHZhangSBogenDBrohlASSindiriSCatchpooleDBadgettTGetzGMoraJAndersonJRSkapekSXBarrFGMeyersonMHawkinsDSKhanJ2014Comprehensive Genomic analysis of Rhabdomyosarcoma reveals a landscape of alterations affecting a common genetic axis in fusion-positive and fusion-negative tumorsCancer Discovery421623110.1158/2159-8290.CD-13-063924436047ShernJFSelfeJIzquierdoEPatidarRChouH-CSongYKYoheMESindiriSWeiJWenXRudzinskiERBarkauskasDALoTHallDLinardicCMHughesDJamalSJenneyMChisholmJBrownRJonesKHicksBAngeliniPGeorgeSCheslerLHubankMKelseyAGatzSASkapekSXHawkinsDSShipleyJMKhanJ2021Genomic classification and clinical outcome in Rhabdomyosarcoma: A report from an international consortiumJournal of Clinical Oncology392859287110.1200/JCO.20.0306034166060StorerNYZonLI2010Zebrafish models of P53 functionsCold Spring Harbor Perspectives in Biology2a00112310.1101/cshperspect.a00112320679337SubramanianATamayoPMoothaVKMukherjeeSEbertBLGilletteMAPaulovichAPomeroySLGolubTRLanderESMesirovJP2005Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profilesPNAS102155451555010.1073/pnas.0506580102Szymczak-WorkmanALVignaliKMVignaliDAA2012Generation of 2A-linked Multicistronic cassettes by recombinant PCRCold Spring Harbor Protocols125125410.1101/pdb.prot06788422301657TaylorACShuLDanksMKPoquetteCAShettySThayerMJHoughtonPJHarrisLC2000P53 Mutation and Mdm2 amplification frequency in pediatric Rhabdomyosarcoma tumors and cell linesMedical and Pediatric Oncology359610310.1002/1096-911x(200008)35:2<96::aid-mpo2>3.0.co;2-z10918230TenenteIMHayesMNIgnatiusMSMcCarthyKYoheMSindiriSGryderBOliveiraMLRamakrishnanATangQChenEYPetur NielsenGKhanJLangenauDM2017Myogenic regulatory transcription factors regulate growth in RhabdomyosarcomaeLife6e1921410.7554/eLife.1921428080960TirodeFSurdezDMaXParkerMLe DeleyMCBahramiAZhangZLapoubleEGrossetête-LalamiSRuschMReynaudSRio-FrioTHedlundEWuGChenXPierronGOberlinOZaidiSLemmonGGuptaPVadodariaBEastonJGutMDingLMardisERWilsonRKShurtleffSLaurenceVMichonJMarec-BérardPGutIDowningJDyerMZhangJDelattreO2014Genomic landscape of Ewing sarcoma defines an aggressive subtype with Co-Association of Stag2 and Tp53 mutationsCancer Discovery41342135310.1158/2159-8290.CD-14-062225223734WangGFershtAR2017Multisite aggregation of P53 and implications for drug rescuePNAS114E2634E264310.1073/pnas.170030811428292898WaterhouseABertoniMBienertSStuderGTaurielloGGumiennyRHeerFTde BeerTAPRempferCBordoliLLeporeRSchwedeT2018SWISS-MODEL: Homology Modelling of protein structures and complexesNucleic Acids Research46W296W30310.1093/nar/gky42729788355WhiteRRoseKZonL2013Zebrafish cancer: the state of the art and the path forwardNature Reviews. Cancer1362463610.1038/nrc358923969693YuXVazquezALevineAJCarpizoDR2012Allele-specific P53 mutant reactivationCancer Cell2161462510.1016/j.ccr.2012.03.0422262471210.7554/eLife.68221.sa0Editor's evaluationEkkerStephen CReviewing Editorhttps://ror.org/03zzw1w08Mayo ClinicUnited States
This paper uses the zebrafish as an in vivo model for exploring cancer genetics. The work on patient-specific alleles of the key oncogene TP53 enables new insights. The focus on embryonic stages enables a new understanding of the mechanism underlying this pediatric cancer.
10.7554/eLife.68221.sa1Decision letterEkkerStephen CReviewing Editorhttps://ror.org/03zzw1w08Mayo ClinicUnited StatesMcGrailMauraReviewerhttps://ror.org/04rswrd78Iowa State UniversityUnited States
Our editorial process produces two outputs: (i) public reviews designed to be posted alongside the preprint for the benefit of readers; (ii) feedback on the manuscript for the authors, including requests for revisions, shown below. We also include an acceptance summary that explains what the editors found interesting or important about the work.
Decision letter after peer review:
Thank you for submitting your article "Defining function of wild-type and patient specific TP53 mutations in a zebrafish model of embryonal rhabdomyosarcoma" for consideration by eLife. Your article has been reviewed by 3 peer reviewers, and the evaluation has been overseen by a Reviewing Editor and Richard White as the Senior Editor. The following individuals involved in review of your submission have agreed to reveal their identity: Maura McGrail (Reviewer #1).
The reviewers have discussed their reviews with one another, and the Reviewing Editor has drafted this to help you prepare a revised submission.
Essential revisions:
1) Are the TP53 alleles really hypomorphic? The cell line work does not address this issue. It should be demonstrated that the proteins retain some level of transcriptional activity. qPCR of P53 target genes (BAX, BBC3/PUMA, PMAIP1/NOXA) in the zebrafish tumor model would address the mechanism by which the human variants can induce apoptosis.
2) The osteosarcoma cell line work should be removed as it is out of place in this manuscript, given the focus on embryonal rhabdomyosarcoma. These should be removed and if any in vitro studies are included, these should be in in an ERM cell line.
3) Language: For example, the authors overstate their claim in the Introduction, "highlight the zebrafish ERMS model as a powerful and high-throughput system" when they describe examining only 3 specific mutations in this report. This should be addressed throughout to reflect the focus of this work on embryonal rhabdomyosarcoma. Such focus will improve the communication of this story.
Reviewer #1 (Recommendations for the authors):
Results and Discussion:
1. The Langenau et al., 2007 rag2-kRASG12D ERMS model is well established and the results briefly describe the model, but it would be helpful to provide a more thorough description of the model in the first paragraph, particularly since this is a somatic model created by injecting linear DNA expressing rag2 promoter driven kRASG12D and a rag2 promoter driven fluorescent reporter. Data presented later in the manuscript underscores the considerable variability in generating tumors or driving gene expression using this approach. The concentration of DNA injected into single cell embryos to create tumors was indicated in the Methods, but the amount of DNA injected wasn't recorded. Please include. A diagram in figure 1 showing how the model is generated would also be very helpful.
2. Figure 1. Include the age of the individual fish shown in Figure 1. Some of the arrows point to tissue with DsRed expression, which may or may not be transformed by expression of kRASG12D. It would be useful to explain the criteria used to determine a tumor vs. simply expression of the co-injection marker.
3. Figure S1C-E The impact of loss of tp53 on the rate of ERMS tumor growth was assessed over a 3 week period using age matched fish. The age of the fish wasn't indicated, but they appear to be adults, not juveniles as shown in Figure 1. Overall, a more complete description of procedures and analyses would strengthen the paper.
4. Figure S1C-E The text states that metastases were observed in ERMS ; tp53-/- fish #3. A second site tumor isn't visible in the image of fish #3. To support this claim in the text, evidence for metastasis vs. occurrence of a second site primary tumor is required.
5. Figure 2 A and B. Cellular characterization of the ERMS; tp53-/- tumors reveals increased cellular proliferation compared to ERMS tumors. The numbers for panels A (Edu+ cells) and B (pHH3+ cells) appear to be switched: 2.3 vs. 13.3 should be in B, 21 vs 65.5 should be in A. Does n refer to individual tumors, or individual sections? The data is very strong, but quantification plots should be included with p values for both markers of proliferation.
6. Figure 2 C-E and Figure 3. The analysis of apoptosis in the ERMS vs. ERMS; tp53 tumors indicates a lack of apoptosis in the tumor model. Figure 3 presents data to show that expression of human Tp53 significantly suppressed tumor initiation and tumor size in ERMS; tp53-/- animals. The data is significant, and indicates a reduction in the number of EdU positive proliferating cells. Given ERMS tumors in a wildtype zebrafish tp53 background have very little apoptosis, the observation that there is an increase in apoptosis could be further clarified.
7. The comparison of ERMS;tp53-/- with and without expression of human TP53 showed that there isn't a significant difference in the number of tumors per fish. Data is needed to support the claim that expression of human tp53 prevents spreading and/or metastasis.
8. Figure S3: An experiment testing the dose dependency of tumor suppression was performed by injection of different concentrations of rag:TP53. 15ng/ul vs. 20ng/ul of linear construct was injected. Details on the volume injected, whether this was human tp53 or zebrafish tp53, were not included. In order to demonstrate a dose response, the data should show that there is a significant difference between 15 ng/ul vs. 20 ng/ul, and the experiment completed with at least 3 biological replicates. The data in c shows high variability in tp53 protein levels in individual tumors, without details on the source of the tumors – were they from fish injected with the same dose, or different doses. Overall this data suggests the method of the tumor model and expression of tp53 variants leads to significant inconsistencies. A more thorough analysis is necessary to draw a conclusion about dose dependency.
9. Figure 4 and Figure 5. The activity of tp53-C176F and tp53-P153∆ were analyzed in osteosarcoma cells, and it was observed neither inhibits cell growth, suppresses colony formation in vitro, drives luciferase expression, or induces apoptosis or expression of p21. This indicates neither mutant has wild type function, transcriptional activation activity, or is a gain of function.
When expressed in the zebrafish ERMS; tp53-/- model, tp53-C176F can suppress tumor formation (Figure 5B) and lead to increased apoptosis (Figure 5F) – a p value to demonstrate significance should be included. This doesn't match with the lack of apoptosis in ERMS; tp53+/+ data referenced in point 6. Including qPCR analysis of tp53 target gene expression in the zebrafish tumors would address whether tp53-C176F retains some level of wild type activity.
10. Section heading starting on page 9. Figure 5. The analyses of tp53-C176F and tp53-P153∆ could be placed in separate sections or described in separate paragraphs, so the results documenting effect on tumor incidence, location, effects on proliferation and apoptosis are clearly and systematically described for each mutant in comparison to wild type. In the text tp53-C176F is described as a hypomorph. tp53-P153∆ is described as a gain of function. But the figure legend for Figure 5 indicates both are gain of function.
Figure 5D. Does the incidence of head tumors occur in relatively young fish? The image in Figure 5D bottom panel appears to be a juvenile fish, in contrast to the adult fish shown in the top and middle panels. The scale bar isn't indicated in the figure legend. This could be an interesting result to follow up – a correlation between age of onset and tumor location.
11. Figure S5. The critical experiment to demonstrate tp53-C176F is a hypomorph that retains some wildtype function was designed as follows: tp53-C176F was expressed in the ERMS; tp53-/- model, then tumors from control and tp53-C176F were explanted into hosts and the hosts were treated with or without the tp53 stabilizer ZMC1. The representative images in panels D (ERMS; tp53-/- + tp53-C176F) and H (ERMS; tp53 -/-) aren't equivalent – both control groups treated with DMSO should show the same level of tumor growth/expansion; both ZMC1 treatment groups appear to impact tumor expansion equally.
12. In the Results section "Expression of TP53Y220C predisposes to head ERMS in zebrafish" starting on page 10 there is extensive discussion of the co-occurrence of tp53-P153∆ with tp53-Y220C. This would fit better in the introduction and/or in the discussion.
13. Figure 6 shows results examining the tp53-Y220C variant on tumor incidence, location, level of apoptosis and proliferation. The switch from co-expressing a DsRed marker to a GFP marker wasn't described.
Similar to the analysis of tp53-C176F and tp53-P153∆, the tp53-Y220C variant did not demonstrate wild type tp53 activity in in vitro assays, but had a suppressive effect on tumor incidence in the zebrafish ERMS;tp53-/- model. It also appeared to lead to increased incidence of head tumors. Figure 6 D and F doesn't indicate age or size of the fish shown.
14. Inclusion of a summary diagram or table that maps activities to each tp53 mutant/variant analyzed would help to clarify the key findings of the study.
15. Overall, there was an absence of detail in the figure legends, number of biological replicates for key experiments, age of fish shown in the images, and scale bar measurements.
16. Discussion:
The discussion could better address why overexpression of tp53 variants in the zebrafish ERMS; tp53-/- model could suppress tumor initiation, with or without impact on proliferation and apoptosis, when the same tp53 variants did not have an impact on osteosarcoma cell proliferation, viability, and colony formation in vitro. The conclusion that the variants are either hypomorphs or gain of function alleles would be better supported by demonstrating their effect on gene expression or other cellular activity.
A number of statements were made that are vague or not supported by the data:
"when compared to wild-type animals, tp53-/- animals … exhibit a relatively less pronounced effect on tumor cell apoptosis." "Thus, while TP53 is shown to suppress tumor progression by multiple mechanisms depending on tumor type, our results suggest that loss of TP53 in human ERMS may increase aggressiveness through enhanced proliferation, invasion, and metastasis." Evidence of invasion or metastasis in the tumor model aren't presented.
Reviewer #2 (Recommendations for the authors):
1. The authors overstate their claim in the Introduction, "highlight the zebrafish ERMS model as a powerful and high-throughput system" when they describe examining only 3 specific mutations in this report.
2. More details regarding the transgenic generation approach should be described and the transgenic constructs clearly articulated throughout the Results section. The authors shouldn't assume that readers are familiar with prior strategies undertaken by these authors or by the Langenau lab.
3. Page 6: in the phrase "Treatment of wild-type and tp53-/- tumors with a 6-hour pulse of EdU" – it needs to be clarified that this is being undertaken in whole larval context.
4. The concepts of early and late apoptosis need to be properly defined and explained.
5. Page 6: the phrase "introducing wild-type human TP53 (TP53WT) in ERMS tumors in tp53-/- animals" is not an accurate reflection of what is being undertaken, which is introducing WT TP53 along with the KRAS transgene at the outset rather than adding the gene to already developed tumors, which is how the text currently reads.
6. The authors should comment on the amino acid conservation between human and zebrafish p53 and how this may impact interactions with other endogenous zebrafish proteins.
7. Page 8: the statement that the TP53 mutants tested "do not show any gain-of function activity in the SaOS2 TP53-/- background" appears a generalization not justified by the experimental data presented.
8. The Discussion should reference that the order in which oncogenic mutations occur may impact tumor phenotype and as such, may not be fully reflected in the zebrafish models as presented.
Reviewer #3 (Recommendations for the authors):
1. In many figures, including Figures 1D, 3C, S1E, and S5G, the plots of Tumor/body size or Tumor area ratio showed very high tumor burden in certain groups, the ratios in some plots were even higher than 0.6. Please make sure your studies complied with the IACUC regulation on the tumor size/weight per animal.
2. Although the SaOS2 cell line is a commonly used assay system to study effects of TP53 variants, the data generated in this cell line were not consistent at all with those obtained from the in vivo zebrafish studies. Could authors discuss this inconsistency? Can authors use ERM cell line(s) instead of the SaOS2 osteosarcoma cell line to assess the activities of the TP53P153△, TP53C176F and TP53Y220C variants? In addition, engineering the cell lines with these exact mutations, not overexpression, would be helpful to accurately characterize their activities in vitro.
3. Coinjection of either TP53P153△ or TP53Y220C mutant with kRASG12D can increase development of ERMS in the head musculature. However, these two mutants behave very differently in many aspects, such as in kRASG12D-induced tumor initiation (Figures 5B vs. 6C) or in cell survival of ERMs (Figures 5F vs. 6I). Can authors provide clearer explanation on how these two different mutants promote more ERM tumorigenesis in the head musculature? Since it has been hypothesized that both TP53P153△ and TP53Y220C mutants might cause p53 instability, it would be interesting to test whether ZMC1 can stabilize these mutants leading to inhibition of head tumor development. In addition, whether Hedgehog signaling pathway could be specifically activated in the head tumors with overexpression of kRASG12D and TP53P153△ or TP53Y220C?
4. TP53C176F coinjection with kRASG12D in p53-/- fish can reduce kRASG12D-induced tumorigenesis (Figure 5B) and increase apoptosis (Figure 5F), suggesting TP53C176F functions as a hypomorphic mutant. Can authors examine whether the p21 expression is upregulated in the ERMs with TP53C176F overexpression, which would provide further mechanistic insight on the hypomorphic effect of the TP53C176F mutant.
[Editors' note: further revisions were suggested prior to acceptance, as described below.]
Thank you for resubmitting your work entitled "Defining function of wild-type and three patient specific TP53mutationsin a zebrafish model of embryonal rhabdomyosarcoma" for further consideration by eLife. Your revised article has been evaluated by Richard White (Senior Editor) and a Reviewing Editor.
There is strong consensus that the manuscript has been substantially improved but there are some remaining issues that need to be addressed, as outlined below:
First, there appears to be a missing item – Supplemental Table I. Second, the new findings about the occurrence of medulloblastoma in the tp53P153 model should be mentioned in the Abstract.
Reviewer #1 (Recommendations for the authors):
Overall, the revised manuscript addressed the essential revisions and the authors have responded to the reviewer's requests. The revised manuscript also presents two new analyses, which weren't mentioned in the abstract or introduction. A new Figure 5 and supporting figures show identification of medulloblastoma in addition to head ERMS arising in the tp53P153delta ERMS model. A new Figure 7 shows experiments to test the dependence of the tp53P153delta ERMS tumor induction on Kdr, which is overexpressed in patient osteosarcoma and the zebrafish model. The Kdr gene and function of the protein could be better introduced, and the experiment have stronger experimental validation as described below. But the kdr experiment illustrates how this model can be used to investigate the in vivo mechanisms by which patient tp53 variants drive ERMS tumorigenesis.
Some issues were not adequately addressed: the statement that the model is high throughput has not been removed from the discussion, Figure 2 A, B quantification data was not included and it's still unclear what the numbers in those panels reflect, and new RT-qPCR on p53 target gene expression contained contradictory or inconsistent results. Different tp53 variants did not lead to consistent elevation of apoptotic genes bax, bbc3, noxa, – tp53C176F and tp53P153delta induced expression of bbc3, but not noxa or bax; tp53Y220C increased noxa and cell growth arrest gene gadd54a; cell cycle arrest gene cdkn1a was elevated in all variant models including the tp53 null. A discussion of this data, in the context of demonstrating that the tp53 variants retain some level of transcriptional activity, was not included.
The experiment to demonstrate ERMS tp53P153delta is dependent on elevated kdr expression via somatic CRISPR RNP kdr targeting did not contain much detail. A more thorough analysis of kdr deletion/mutagenesis in the tumors, kdr transcript knockdown measured by RT-qPCR, would validate the conclusion that Kdr drives oncogenesis in the ERMS tp53P153delta model. Figure 7 Supplement 1 does not contain enough detail explaining how targeted mutagenesis was induced, documented and measured.
Reviewer #2 (Recommendations for the authors):
The authors have done an admirable job responding to previous reviews, resulting in a manuscript that is more logically constructed and better focused in the data presented.
A few residual comments:
1. I was not able to find the Supplemental Table that was suggested by reviewers and referred to in the Response to Reviewers that summarizes and highlights the main findings and differences between wild type and various tp53 zebrafish mutants described in the manuscript.
2. The choice of analyzing the TP53 P153 mutation remains a bit unusual given its rarity and identification as a germline mutation in an aggressive osteosarcoma, given the focus of this manuscript on ERMS (especially now with removal of the osteosarcoma cell line data which contrasted with the zebrafish data).
3. Moreover this P153 mutation leads to Shh medulloblastoma tumors in a subset of zebrafish. This is new data that is interesting, but a bit of a distraction from the focus of the manuscript on ERMS.
10.7554/eLife.68221.sa2Author response
Essential revisions:
1) Are the TP53 alleles really hypomorphic? The cell line work does not address this issue. It should be demonstrated that the proteins retain some level of transcriptional activity. qPCR of P53 target genes (BAX, BBC3/PUMA, PMAIP1/NOXA) in the zebrafish tumor model would address the mechanism by which the human variants can induce apoptosis.
As requested, qPCR was performed on newly generated primary tumors expressing tp53+/+, tp53-/- and tp53-/- + TP53C176F, tp53-/- + TP53P153Δ and tp53-/- + TP53Y220C and p53 target genes including baxa, cdkn1a, gadd45a, noxa and puma/bbc3 were assessed. We report that in the TP53C176F and TP53P153Δ expressing tumors bbc3 is induced but not noxa, baxa or gadd45a. While in the TP53Y220C expressing tumors gadd45a is induced and in 2 of the 3 primary tumors noxa is induced. Rather unexpectedly in the tp53-/-, tp53-/- + TP53C176F, tp53-/- + TP53P153Δ and tp53-/- + TP53Y220C tumors cdkn1a is induced compared to tp53+/+ tumors. We summarize our major findings in Supplemental table 1. Similarly, we find in human RMS cell lines high p21/CDKN1A is expressed in multiple cell lines which are both wild-type or mutant for p53 (see Reviewer figure 3). Suggesting that in RMS CDKN1A expression can also be independent of p53.
We also generated new tumors to address reviewer comments. Unexpected, we found that the tp53-/- + TP53P153Δ head tumors also included potential brain tumors which resembled medulloblastomas (new Figure 5). These tumors have previously never been reported in the kRASG12D ERMS model in zebrafish but can occur in Li Fraumeni patients. As this finding could confound our reported results, we assessed HandE of tumors in our analyses and also generated new tumors to assess the contribution of the medulloblastoma to our analyses and show that a majority of (16 of 21) of head tumors are ERMS. Lastly, we characterize the head ERMS tumors and we show that (i) Head ERMS tumors are molecularly distinct from trunk tumors and (ii) kdr downstream of TP53P153Δ predisposes to head ERMS in tp53-/- zebrafish.
2) The osteosarcoma cell line work should be removed as it is out of place in this manuscript, given the focus on embryonal rhabdomyosarcoma. These should be removed and if any in vitro studies are included, these should be in in an ERM cell line.
We agree with the reviewers and have removed in vitro studies in the SaOS2 cells.
3) Language: For example, the authors overstate their claim in the Introduction, "highlight the zebrafish ERMS model as a powerful and high-throughput system" when they describe examining only 3 specific mutations in this report. This should be addressed throughout to reflect the focus of this work on embryonal rhabdomyosarcoma. Such focus will improve the communication of this story.
We have made changes throughout the manuscript to reflect only what our results find and have made changes recommended to the text and Results section (Please see reviewer comments that address specific statements).
Reviewer #1 (Recommendations for the authors):
Results and Discussion:
1. The Langenau et al., 2007 rag2-kRASG12D ERMS model is well established and the results briefly describe the model, but it would be helpful to provide a more thorough description of the model in the first paragraph, particularly since this is a somatic model created by injecting linear DNA expressing rag2 promoter driven kRASG12D and a rag2 promoter driven fluorescent reporter. Data presented later in the manuscript underscores the considerable variability in generating tumors or driving gene expression using this approach. The concentration of DNA injected into single cell embryos to create tumors was indicated in the Methods, but the amount of DNA injected wasn't recorded. Please include. A diagram in figure 1 showing how the model is generated would also be very helpful.
We thank the reviewer for this suggestion and agree that schematics setting up the experiments described in the zebrafish ERMS model are important. We have therefore included schematics for the main figures to better explain the experimental set up. We also have included in the Methods section the amount of DNA injected in addition to the concentration of the DNA cocktails used in different experiments.
Amount of DNA in experiments. Wild-type AB, tp53-/- AB and tp53-/- AB/CG1 mixed strain zebrafish can tolerate a maximum of approximately 25-30 pg of DNA, while CG1, tp53-/-; CG1 strain syngeneic zebrafish can tolerate half that amount (15 pg of DNA). Second, the amount of kRASG12D or TP53 wild-type or mutant total DNA injected is kept constant with a mix of 40 ng rag2-kRASG12D/ 20 ng rag2-dsRED/GFP (60 ng/μL) in experiments in Figure 1 and for the rest of the manuscript 35 ng rag2-kRASG12D/15 ng rag2-TP53 wild-type or mutant/ 10 ng rag2-GFP/dsRED (60 ng/μL).
Addressing variability in generating tumors. As ERMS tumors arise very quickly within 60 days (2 months) and animals can start breeding on average between 4-6 months, it is not possible to maintain stable rag2:kRASG12D stable transgenic lines. We have attempted to generate conditional transgenic lines (UAS-Gal4 transgenics) in the past and while we have shown that the method works in principle. However, we were unable isolate stable lines that generate tumors possibly due to silencing of the kRASG12D in the germline. To account for variability in our experiments in each figure we perform both experimental and control arms of the experiment at the same time. i.e. on any given day we inject zebrafish (100 to >500) embryos from zebrafish mattings for both experimental groups and raise animals in the exact same conditions. We also perform at least 2 or more independent micro-injection experiments for each analyses. As there are loses of larval zebrafish due to normal animal husbandry, we combine data from experiments that were done close together i.e. within the same 4-8 week period. Thus, while the ERMS in zebrafish model is relatively easy use, each Kaplan Meijer analyses in the figure panel takes on average 3-6 months to complete and for each Kaplan Meijer analyses we have > 100 larval zebrafish per group except for Figure 7F (n>45 in both experimental arms). Additionally, because tumor initiation in the tp53-/- background results in >75-97% of the experimental zebrafish initiating tumors it is possible to study more subtle differences between human TP53 mutants. However, in the zebrafish tp53 wild-type setting, where tumor initiation can vary to about 15-35% of zebrafish getting tumors, it is extremely difficult to discern subtle differences in ERMS initiation. We would like to state that our protocol is a similar to protocols used by the laboratories of Dr. David Langenau, Dr. Thomas Look, Dr. Leonard Zon and former members from these laboratories to generate ERMS, leukemia, neuroblastoma, melanoma and other tumors. Lastly, in murine or zebrafish in vivo tumor models where tumors can take 10 days to a year or more to initiate, biological replicates are not the norm but rather a power calculation used to assess number of animals needed to achieve a p value of <0.05 compared to experimental and the control group.
2. Figure 1. Include the age of the individual fish shown in Figure 1. Some of the arrows point to tissue with DsRed expression, which may or may not be transformed by expression of kRASG12D. It would be useful to explain the criteria used to determine a tumor vs. simply expression of the co-injection marker.
We have updated all our figures and provide age of animals. We have previously shown that rag2:dsRED is highly expressed in all ERMS tumors, including in tumors in zebrafish that are < 10 days old. In the muscle, rag2:dsRED is expressed at very low levels which is not easily observable in non-tumor muscle cells even in the Tg(rag2-GFP) or Tg (rag2-dsRED) transgenic animals and hence easy to exclude from our analyses. We have previously confirmed tumors and developed methods to stage our ERMS tumors in experiments and show by co-labeling that even the earliest rag2:DsRED tumors that occur in one half or single somites express high levels of tumor propagating cell marker myf5-GFP and these early stage tumors were also confirmed via histology (Ignatius et al., Cancer Cell 2012). In our manuscript, we are being conservative and calling only stage 2 and above tumors, we also confirm tumors by performing HandE staining on a subset of all our tumors in all figures/experiments and the histology has been confirmed by soft tissue expert pathologist Dr. Eleanor Chen.
3. Figure S1C-E The impact of loss of tp53 on the rate of ERMS tumor growth was assessed over a 3 week period using age matched fish. The age of the fish wasn't indicated, but they appear to be adults, not juveniles as shown in Figure 1. Overall, a more complete description of procedures and analyses would strengthen the paper.
Please see below.
4. Figure S1C-E The text states that metastases were observed in ERMS ; tp53-/- fish #3. A second site tumor isn't visible in the image of fish #3. To support this claim in the text, evidence for metastasis vs. occurrence of a second site primary tumor is required.
We agree with the reviewer, the experiments we have done in the primary tumor setting cannot exclude the possibility that tumors occurring at secondary sites are not a primary tumor at the secondary site. Since suppression of invasion and metastasis effects of wild-type tp53 have already been previously established in a transplant model using syngeneic zebrafish (Ignatius et al., eLife 2018) and invasion and metastasis is not directly relevant to our current study, we have removed this analyses from our revised manuscript. Secondly, as tumors in tp53-/- mutants initiate in younger animals and grow rapidly we had to pick from the subset that occur later and are smaller slower growing in order to carry out this analyses.
5. Figure 2 A and B. Cellular characterization of the ERMS; tp53-/- tumors reveals increased cellular proliferation compared to ERMS tumors. The numbers for panels A (Edu+ cells) and B (pHH3+ cells) appear to be switched: 2.3 vs. 13.3 should be in B, 21 vs 65.5 should be in A. Does n refer to individual tumors, or individual sections? The data is very strong, but quantification plots should be included with p values for both markers of proliferation.
We thank the reviewer in pointing out this discrepancy. We have updated our figures and figure legends to point out how these counts were performed. In panels A and B, we are estimating the percentage of Edu+ cells compared to the total number of DAPI-positive nuclei which are also dsRED-positive (Tumor). In B we calculate the total number of pHH3-positive nuclei in 6 confocal sections. Since much fewer cells are actively undergoing mitosis compared to those in S phase (Edu labeling), this sampling is able to overcome the variability of performing counts on single sections. In the figure we show one section.
6. Figure 2 C-E and Figure 3. The analysis of apoptosis in the ERMS vs. ERMS; tp53 tumors indicates a lack of apoptosis in the tumor model. Figure 3 presents data to show that expression of human Tp53 significantly suppressed tumor initiation and tumor size in ERMS; tp53-/- animals. The data is significant, and indicates a reduction in the number of EdU positive proliferating cells. Given ERMS tumors in a wildtype zebrafish tp53 background have very little apoptosis, the observation that there is an increase in apoptosis could be further clarified.
We thank the reviewer for pointing this out. We have included a statement in the text to clarify this difference. We believe the increased apoptosis seen in tumors co-expressing zebrafish or human WT TP53 is due to higher levels of the wild-type protein. This is a limitation of our experimental set up, but it is also the approach that is commonly used to study effects of oncogenes, for example, Myc, Notch, Akt2, or tumor suppressors for example, bbc3/ puma, noxa on tumor biology in zebrafish cancer models (Blackburn et al., 2014; Langenau et al., 2007; Lobbardi et al., 2017; White, Rose, and Zon, 2013). In order to ensure we are not expressing high levels of wild-type and mutant p53 proteins, we compared the amount of mutant or wild-type protein expressed to mutant p53 in the Rh30 rhabdomyosarcoma cell line. Rh30 cells express the p53R273C hotspot mutant protein. It is important to note that mutant p53 proteins can be highly expressed in tumors and the p53C176F and p53P153Δ mutant proteins studied here are highly expressed via IHC and western blot analyses (see Figure 4C, D). Further, our data also show that methods that can reactivate wild-type or mutant protein can also effect apoptosis if induced sufficiently.
7. The comparison of ERMS;tp53-/- with and without expression of human TP53 showed that there isn't a significant difference in the number of tumors per fish. Data is needed to support the claim that expression of human tp53 prevents spreading and/or metastasis.
Since in primary tumors we are unable to study metastasis and the data we have is mostly corelative, we have removed invasion and metastasis from the manuscript and focus on effects on tumor initiation, tumor number, growth, proliferation, apoptosis and site of tumor initiation. We do find that compared to tumors that initiate in the tp53-/- background, overall fewer TP53 or Tp53 expressing animals initiate tumors. However, in the tp53-/- background in zebrafish expressing rag2:kRASG12D, rag2:TP53 and rag2:dsRED in tumor burdened zebrafish multiple tumors can initiate per animal suggesting possibly non-cell autonomous roles for tp53 in tumor initiation.
8. Figure S3: An experiment testing the dose dependency of tumor suppression was performed by injection of different concentrations of rag:TP53. 15ng/ul vs. 20ng/ul of linear construct was injected. Details on the volume injected, whether this was human tp53 or zebrafish tp53, were not included. In order to demonstrate a dose response, the data should show that there is a significant difference between 15 ng/ul vs. 20 ng/ul, and the experiment completed with at least 3 biological replicates. The data in c shows high variability in tp53 protein levels in individual tumors, without details on the source of the tumors – were they from fish injected with the same dose, or different doses. Overall this data suggests the method of the tumor model and expression of tp53 variants leads to significant inconsistencies. A more thorough analysis is necessary to draw a conclusion about dose dependency.
We agree with the reviewer and have removed the 20 ng/μL arm from our analyses and include data only for the 15 ng/μL arm as this corresponds to the amount of TP53 wild-type and mutant DNA we use in all our analyses. All western blots analyses were from the 15 ng/μL injections of the small tumors that arise from this injection. The zebrafish Tp53 in #2 tumor is higher than others, however this in outlier in expression of p53 compared to all our experiments. please see figure 3A, 4G and 6B.
9. Figure 4 and Figure 5. The activity of tp53-C176F and tp53-P153∆ were analyzed in osteosarcoma cells, and it was observed neither inhibits cell growth, suppresses colony formation in vitro, drives luciferase expression, or induces apoptosis or expression of p21. This indicates neither mutant has wild type function, transcriptional activation activity, or is a gain of function.
We agree with the reviewers that the studies in the SaOS2 cells which are TP53 deficient osteosarcoma cells does not agree in vivo studies in our zebrafish embryonal rhabdomyosarcoma model. We would like to note that while the SaOS2 and other TP53 deficient cells have traditionally used to characterize mutant TP53 function a significant subset of mutants do not show any phenotypes or show mutant p53 effects only in some and not in other assay systems. Additionally, whether TP53 mutants can have GOF effects remains controversial in some tumors for example leukemia (Boettcher et al., 2019). We initially added the SaOS2 data to highlight that the SaOS2 cell based system may not be optimal for the mutants we study. However, as we have not tested other cell systems and have not generated knock-in alleles for these mutants in ERMS cells lines, we cannot fully support this assertion and have accordingly removed the SaOS2 data as suggested.
When expressed in the zebrafish ERMS; tp53-/- model, tp53-C176F can suppress tumor formation (Figure 5B) and lead to increased apoptosis (Figure 5F) – a p value to demonstrate significance should be included. This doesn't match with the lack of apoptosis in ERMS; tp53+/+ data referenced in point 6. Including qPCR analysis of tp53 target gene expression in the zebrafish tumors would address whether tp53-C176F retains some level of wild type activity.
We have included this data in Figure 5F (now Figure 4K) and show that tumors expressing TP53C176F in the tp53-/- background compared to tp53-/- tumors. However, there is no difference in the rate of apoptosis between tp53-/- and TP53P153Δ; tp53-/- expressing tumors. Mutant p53 proteins can often be highly expressed in tumors compared to the WT p53 which is usually poorly expressed but can be induced after damage. Additionally, we do show that the WT human p53 can complement the zebrafish protein and that the WT and the three mutants can show very different but reproducible effects on different aspects of TP53 function.
As requested qPCR was performed on newly generated primary tumors. “To assess the effects of p53 variant expression on transcription, we performed qPCR analyses comparing tp53+/+, tp53-/- and tp53-/- + TP53C176F and tp53-/- + TP53P153Δ ERMS and p53 target genes including baxa, cdkn1a, gadd45a, noxa and puma/bbc3. We found that both tp53-/-+TP53C176F and tp53-/- + TP53P153Δ tumors displayed increased bbc3 expression compared to tp53-/- or tp53+/+ tumors, but had no difference with respect to cdkn1a, gadd45a and noxa expression. We also found that tp53-/- + TP53C176F and tp53-/- + TP53P153Δ tumors expressed higher levels of cdkn1a compared to tp53+/+ ERMS; however, rather unexpectedly tp53-/- tumors also express higher levels of cdkn1a compared to tp53+/+ controls (Figure 4 Supplement 3).”
10. Section heading starting on page 9. Figure 5. The analyses of tp53-C176F and tp53-P153∆ could be placed in separate sections or described in separate paragraphs, so the results documenting effect on tumor incidence, location, effects on proliferation and apoptosis are clearly and systematically described for each mutant in comparison to wild type.
We have rewritten this section and have separated out effects of the TP53C176Fand TP53P153Δ mutants in the text. Additionally, we provide a new table (Supplemental Table 1) to compare and contrast the cellular effects of the wild-type and mutant TP53 genes when expressed in the zebrafish ERMS model.
In the text tp53-C176F is described as a hypomorph. tp53-P153∆ is described as a gain of function. But the figure legend for Figure 5 indicates both are gain of function.
We thank the reviewer for pointing out this oversight and have changed the figure legend to “TP53C176F is a hypomorphic allele while TP53P153Δ has gain-of-function effects in ERMS”.
Figure 5D. Does the incidence of head tumors occur in relatively young fish? The image in Figure 5D bottom panel appears to be a juvenile fish, in contrast to the adult fish shown in the top and middle panels.
We went back to our analyses and find that the head tumors do not initiate later. However, we do find that the head tumors grow slightly more slowly. To clearly reveal the head tumor phenotypes, we waited until the tumors were more prominent. This is consistent with our finding that the tumors in the tp53-/- background come up quickly and grow rapidly. We also generated head ERMS tumors expressing TP53P153Δ in the syngeneic CG1; tp53-/- background. We expanded 3 head tumors expressing TP53P153Δ and also 3 tp53-/- tumors (1 head and 2 trunk tumors), sorted pure secondary tumors and performed RNAseq analyses across both groups. Rather unexpectedly, the head tumors we isolated were all pro-neural brain tumors with gene signatures consistent with the sonic hedgehog sub-group of medulloblastoma (See new Figure 5). To address how frequent the medulloblastoma tumors were, we generated additional tumors with the following genotypes kRASG12D/ tp53-/-, kRASG12D/ TP53P153Δ, tp53-/- and kRASG12D/ TP53C176F, tp53-/-. We identified 5 out 21 head tumors had histology consistent with medulloblastoma in the kRASG12D/ TP53P153Δ, tp53-/- group, the rest were all head ERMS tumors (>76% of all head tumors). In contrast only 1 out of 20+ kRASG12D/ tp53-/- tumors had a medulloblastoma tumor and we did not see any head tumors in the kRASG12D/ TP53C176F/tp53-/- group (0 of 20 ERMS). Finally, we compared the CG1; tp53-/- ERMS head tumor with the two trunk tumors and show that the head tumors express genes consistent with head musculature including tbx1, dlx3b, dlx4b and ptch2, suggesting that there are molecular differences between head and trunk ERMS tumors consistent with analyses in a hedgehog driven mouse tumor model of head and neck rhabdomyosarcoma (Drummond et al., 2018).
The scale bar isn't indicated in the figure legend. This could be an interesting result to follow up – a correlation between age of onset and tumor location.
We have added scale bars and the age of the zebrafish in every figure or in the figure legend.
11. Figure S5. The critical experiment to demonstrate tp53-C176F is a hypomorph that retains some wildtype function was designed as follows: tp53-C176F was expressed in the ERMS; tp53-/- model, then tumors from control and tp53-C176F were explanted into hosts and the hosts were treated with or without the tp53 stabilizer ZMC1. The representative images in panels D (ERMS; tp53-/- + tp53-C176F) and H (ERMS; tp53 -/-) aren't equivalent – both control groups treated with DMSO should show the same level of tumor growth/expansion; both ZMC1 treatment groups appear to impact tumor expansion equally.
We have found that transplanted secondary tumors can grow at different rates. Taking this into account the variability in this experiment, we compare the effect of ZMC1 vs DMSO in tp53-/-; TP53C176F and tp53-/- tumors. Relative to the DMSO treatment there the ZMC1 tp53-/-; TP53C176F are smaller, have p53 stabilized and also have increased Annexin V-positive cells. In contrast, there is no difference in tumor size and apoptosis between DMSO and ZMC1 treated tp53-/- tumors.
12. In the Results section "Expression of TP53Y220C predisposes to head ERMS in zebrafish" starting on page 10 there is extensive discussion of the co-occurrence of tp53-P153∆ with tp53-Y220C. This would fit better in the introduction and/or in the discussion.
We have edited the text to briefly introduce why we chose to study the TP53Y220C mutant in zebrafish. We have moved the more detailed description to the Discussion section as suggested.
13. Figure 6 shows results examining the tp53-Y220C variant on tumor incidence, location, level of apoptosis and proliferation. The switch from co-expressing a DsRed marker to a GFP marker wasn't described.
We include a statement in the main text and also in the methods section addressing this change. We and other have interchangeably use dsRed and or GFP to label ERMS tumors (Ignatius et al., 2012; Ignatius et al., 2018; Langenau et al., 2008).
Similar to the analysis of tp53-C176F and tp53-P153∆, the tp53-Y220C variant did not demonstrate wild type tp53 activity in in vitro assays, but had a suppressive effect on tumor incidence in the zebrafish ERMS;tp53-/- model. It also appeared to lead to increased incidence of head tumors. Figure 6 D and F doesn't indicate age or size of the fish shown.
Please see Section 9.
We have added age of the zebrafish in Figure 6 D and F.
14. Inclusion of a summary diagram or table that maps activities to each tp53 mutant/variant analyzed would help to clarify the key findings of the study.
We thank the reviewer for this suggestion and have included a supplementary table (Supplemental Table 1) highlighting the main findings and differences between the wild-type and mutant TP53 highlighted.
15. Overall, there was an absence of detail in the figure legends, number of biological replicates for key experiments, age of fish shown in the images, and scale bar measurements.
We thank the reviewer for this comment and have gone through each section of the manuscript and have included additional details including schematics for the experimental set up, included more details of how the experiments were performed, age of the fish in experiments, scale bar measurements and details of biological replicates. We have also highlighted these details in the response to reviewer points.
16. Discussion:
The discussion could better address why overexpression of tp53 variants in the zebrafish ERMS; tp53-/- model could suppress tumor initiation, with or without impact on proliferation and apoptosis, when the same tp53 variants did not have an impact on osteosarcoma cell proliferation, viability, and colony formation in vitro. The conclusion that the variants are either hypomorphs or gain of function alleles would be better supported by demonstrating their effect on gene expression or other cellular activity.
A number of statements were made that are vague or not supported by the data:
"when compared to wild-type animals, tp53-/- animals … exhibit a relatively less pronounced effect on tumor cell apoptosis." "Thus, while TP53 is shown to suppress tumor progression by multiple mechanisms depending on tumor type, our results suggest that loss of TP53 in human ERMS may increase aggressiveness through enhanced proliferation, invasion, and metastasis." Evidence of invasion or metastasis in the tumor model aren't presented.
We have reworked/reworded the Results section and the discussion to include the following as recommended by the reviewers. (1) We have removed all data in SaOS2 p53 deficient cells. (2) We have removed data pertaining to invasion and metastasis as this data is corelative and not related to our main findings. We instead cite our previous publication where we addressed the role of tp53 in suppressing invasion and metastasis and having no effect on tumor self-renewal (Ignatius et al., 2018). (3) We add a section about the limitations of our study.
Reviewer #2 (Recommendations for the authors):
1. The authors overstate their claim in the Introduction, "highlight the zebrafish ERMS model as a powerful and high-throughput system" when they describe examining only 3 specific mutations in this report.
We thank the reviewer from their comment and agree. We have changed the title and the text to reflect the results presented to remove any statement that overrepresents our significant findings.
Title: Defining function of wild-type and three patient specific TP53 mutations in a zebrafish model of embryonal rhabdomyosarcoma.
2. More details regarding the transgenic generation approach should be described and the transgenic constructs clearly articulated throughout the Results section. The authors shouldn't assume that readers are familiar with prior strategies undertaken by these authors or by the Langenau lab.
We thank the reviewers for this suggestion and agree that schematics setting up the experiments described in the zebrafish ERMS model is important. We have therefore included schematics for the main figures to better explain the experimental set up. We also have included in the Methods section the amount of DNA injected in addition to the concentration of the DNA cocktails used in different experiments.
3. Page 6: in the phrase "Treatment of wild-type and tp53-/- tumors with a 6-hour pulse of EdU" – it needs to be clarified that this is being undertaken in whole larval context.
We thank the reviewer for this suggestion and have clarified the experiment being performed.
“Wild-type and tp53-/- tumor burdened animals were treated with a 6-hour pulse of EdU and then animals were euthanized and sectioned, and stained for Edu-positive cells, a marker for proliferation.”
4. The concepts of early and late apoptosis need to be properly defined and explained.
We have added more details and a reference to introduce how Annexin V in combination with Propidium iodide can be used to label live cells that are progressing through apoptosis.
“Annexin V staining that can distinguish cells beginning to undergo apoptosis (early apoptosis; low Annexin V/Propidium Iodide (PI)-positive), or that are either undergoing apoptosis (late apoptosis; high Annexin V/high PI) or necrosis (high Annexin V/ low PI), was performed on live single cells suspensions of ERMS tumor cells extracted post euthanasia.”
5. Page 6: the phrase "introducing wild-type human TP53 (TP53WT) in ERMS tumors in tp53-/- animals" is not an accurate reflection of what is being undertaken, which is introducing WT TP53 along with the KRAS transgene at the outset rather than adding the gene to already developed tumors, which is how the text currently reads.
We agree with the reviewer comment and have amended the text to reflect what the experiment is testing.
“We next assessed the consequence of co-expressing wild-type human TP53 (TP53WT) in the cells from which ERMS tumors initiate in tp53-/- animals. Co-expression of kRASG12D along with TP53WT results in the expression of WT TP53 from the very beginning in cells from which tumors initiate and also in the resulting tumors.”
6. The authors should comment on the amino acid conservation between human and zebrafish p53 and how this may impact interactions with other endogenous zebrafish proteins.
We have included the following statement and additional Supplemental figure in the Results section to address this comment by the reviewer.
“Zebrafish and human p53 are functionally similar and share 56% identity with respect to amino acid sequence (67% positives, Figure 3 Supplement 1) (Berghmans et al., 2005; Ignatius et al., 2018; Parant, George, Holden, and Yost, 2010; Storer and Zon, 2010). Within the core DNA-binding region where a majority of mutations occur in patients, 72% conservation exists (79% positives, Figure 3 Supplement 1, 2). To assess functional conservation in vivo, we co-expressed wild-type human TP53 (TP53WT) in the cells from which ERMS tumors initiate in tp53-/- animals (Figure 3 Supplement 3). Importantly, co-expression of kRASG12D along with TP53WT resulted in the expression of wild type TP53 from the very beginning in cells from which tumors initiate and also in the resulting tumors.”
Secondly, in the Discussion section we have added a statement on the limitation of our in vivo model.
“Finally, we would like to state the limitation of our zebrafish ERMS model. In Li Fraumeni patients, TP53 mutations are germline with somatic modifier mutations occurring secondarily. In our experiments, while our starting point is a tp53-/- null background, kRASG12D and the human TP53 variant allele are introduced at the same time, under the control of the rag2 promoter. A second limitation is that effects of mutant TP53 in the tumor microenvironment are missed in our assays. However, our assay is very rapid, given that the generation of zebrafish strains with germline knock-in of the equivalent mutations takes an average of two years for full assessment. In mice, only the most common Tp53 variants have been fully characterized and prioritizing which TP53 mutants to model is a challenge given cost and time restraints.”
7. Page 8: the statement that the TP53 mutants tested "do not show any gain-of function activity in the SaOS2 TP53-/- background" appears a generalization not justified by the experimental data presented.
We agree with the reviewers that the studies in the SaOS2 cells which are TP53 deficient osteosarcoma cells does not agree in vivo studies in our zebrafish embryonal rhabdomyosarcoma model. We would like to note that while the SaOS2 and other TP53 deficient cells have traditionally used to characterize mutant TP53 function a significant subset of mutants do not show any phenotypes or show mutant p53 effects only in some and not in other assay systems. Additionally, whether TP53 mutants can have GOF effects remains controversial and may be tumor type specific. We initially added the SaOS2 data to highlight that the SaOS2 cell based system may not be optimal for the mutants we tested. However, as we have not tested other systems and have not generated knockin alleles for these mutants in ERMS cells lines, we cannot support this assertion and have accordingly removed the SaOS2 data as suggested.
8. The Discussion should reference that the order in which oncogenic mutations occur may impact tumor phenotype and as such, may not be fully reflected in the zebrafish models as presented.
We thank the reviewer for this comment and have added a section to our discussion to address order of the oncogenic mutations and on the limitations of the zebrafish in vivo experimental system we employ. We also include a brief statement of where our studies would fit in addressing an important yet unaddressed need in understanding the function of mutant and wild-type p53 in patients with sarcoma.
Reviewer #3 (Recommendations for the authors):
1. In many figures, including Figures 1D, 3C, S1E, and S5G, the plots of Tumor/body size or Tumor area ratio showed very high tumor burden in certain groups, the ratios in some plots were even higher than 0.6. Please make sure your studies complied with the IACUC regulation on the tumor size/weight per animal.
We thank the reviewer for pointing this out. Our studies comply with our IACUC protocol. In the tumor initiation studies in Figure 1B, 1D, 3 B, 3C we are scoring for tumors at the earliest time point 10-20 days at which time in the tp53-/- animals that initiate tumors are removed from the analyses. An important criteria is if the animals are experiencing any distress or are showing differences in their ability to swim. Since the muscle is the major tissue in the zebrafish at 10-20 days most of the tumor burdened animals are in-distinguishable from the no tumor animals under the naked eye and fluorescence is the only way to distinguish the difference. We also present new figure 5 to follow medulloblastomas that occur in the zebrafish brain. In these experiments as the tumors are slower growing and unexpected, we allow the tumors to grow a little bit to help with the histology before we euthanize the animals. In the adult animals (Figure S5G now S4D,H), we again are using fluorescence area and intensity to estimate relative size and growth rates. In these animals the tumors are again on a subset of the normal tissue and externally visible only if they grow locally. In all our experiments death is not an end point and any animal experiencing distress is immediately euthanized.
2. Although the SaOS2 cell line is a commonly used assay system to study effects of TP53 variants, the data generated in this cell line were not consistent at all with those obtained from the in vivo zebrafish studies. Could authors discuss this inconsistency? Can authors use ERM cell line(s) instead of the SaOS2 osteosarcoma cell line to assess the activities of the TP53P153△, TP53C176F and TP53Y220C variants? In addition, engineering the cell lines with these exact mutations, not overexpression, would be helpful to accurately characterize their activities in vitro.
We agree with the reviewers that the studies in the SaOS2 cells which are TP53 deficient osteosarcoma cells do not agree with in vivo studies in our zebrafish embryonal rhabdomyosarcoma model. We would like to note that while the SaOS2 and other TP53 deficient cells have traditionally used to characterize mutant TP53 function a significant subset of mutants do not show any phenotypes or show mutant p53 effects only in some and not in other assay systems. Additionally, whether TP53 mutants can have GOF effects remains controversial in some tumors for example leukemia (Boettcher et al., 2019). However, as we have not tested other cell systems and have not generated knockin alleles for these mutants in ERMS cells lines, we cannot fully support this assertion and have accordingly removed the SaOS2 data as suggested.
3. Coinjection of either TP53P153△ or TP53Y220C mutant with kRASG12D can increase development of ERMS in the head musculature. However, these two mutants behave very differently in many aspects, such as in kRASG12D-induced tumor initiation (Figures 5B vs. 6C) or in cell survival of ERMs (Figures 5F vs. 6I). Can authors provide clearer explanation on how these two different mutants promote more ERM tumorigenesis in the head musculature? Since it has been hypothesized that both TP53P153△ and TP53Y220C mutants might cause p53 instability, it would be interesting to test whether ZMC1 can stabilize these mutants leading to inhibition of head tumor development. In addition, whether Hedgehog signaling pathway could be specifically activated in the head tumors with overexpression of kRASG12D and TP53P153△ or TP53Y220C?
We thank the reviewer for these comments and have addressed them with the following four sets of experiments.
We tried to generate TP53P153Δ in the syngeneic CG1; tp53-/- background but unexpectedly discovered that TP53P153Δ can collaborate with kRASG12D and tp53 loss to initiate medulloblastomas with a sonic hedgehog signature. See Text and new Figure 5 “We generated head ERMS tumors expressing TP53P153D in the syngeneic CG1; tp53-/- background. We expanded 3 head tumors expressing TP53P153Δ and also 3 tp53-/- tumors (1 head and 2 trunk tumors), sorted pure secondary tumors and performed RNAseq analyses across both groups. Rather unexpectedly, the head tumors we isolated were all pro-neural brain tumors with gene signatures consistent with the sonic hedgehog group medulloblastoma. To address how frequent the medulloblastoma tumors were, we generated additional tumors with the following genotypes kRASG12D/ tp53-/-, kRASG12D/ TP53P153Δ, tp53-/- and kRASG12D/ TP53C176F, tp53-/-. We identified 5 out 21 head tumors had histology consistent with medulloblastoma in the kRASG12D/ TP53P153Δ ,tp53-/- group, the rest were all head ERMS tumors (<24% of all head tumors). In contrast only 1 out of 20+ kRASG12D/ tp53-/- tumor had a medulloblastoma tumor and we did not see any head tumors in the kRASG12D/ TP53C176F/tp53-/- group.
In a second set of experiments (See Figure 7A) we asked if similar to the mouse model of hedgehog driven head and neck rhabdomyosarcoma if the head and trunk tumors are molecularly different. RNAseq analyses comparing the CG1; tp53-/- ERMS head tumor with the two trunk tumors and show that the head tumors express genes consistent with head musculature including tbx1, dlx3b, dlx4b and ptch2 and lower expression of other myogenic genes including.
In the mouse model of hedgehog driven head and neck RMS, the authors suggest that head and neck tumor arise from Kdr-positive endothelial precursors (Drummond et al., 2018). The cell of origin of zebrafish ERMS tumors is not known but we show that zebrafish ERMS show tumor propagating gene expression signatures consistent with an activated satellite cell (Ignatius et al., 2012). Mutant p53 can activate KDR expression.
Genomic analyses of the osteosarcoma from the patient with the germline TP53P153Δ mutation identified that KDR was amplified. Further, we confirmed if KDR was expressed via western blot and IHC analyses on PDX that we generated from the patient at autopsy (see Figure 7B, C). Lastly using the zebrafish ERMS model we tested if ablation of kdr in zebrafish would significantly reduce or eliminate head ERMS tumors. Our results show that ablating kdr and not control gene mitfa resulted in loss of overall tumor initiation and even more significantly head ERMS tumors (Figure 7D-J).
We tried to generate a cell line from the Osteosarcoma PDX expressing TP53P153Δ and have failed thus far. However, we were able to identify that the ewing sarcoma cell line EW8 expressed the TP53Y220C mutation. We therefore determined if there were differences in the effects on p53 and p21 protein expression of published dosing of ZMC1 and Pikan083 on EW8 cell that has the TP53Y220C mutation and show that Pikan083 can increase p53 and p21 expression (Author response image 1). We also compared EW8 cells to RMS RD cells that harbors the TP53R248W mutation and with RMS SMS-CTR cells as controls as they are TP53 deficient. Our results show that ZMC1 can stabilize p53R248W but to a much lesser extent p53Y220C, while Pikan083 stabilizes p53Y220C but not p53 R248W. We also assessed if stabilization of p53 resulted in induction of p21 expression. And while our data holds true for RD and EW8 cells, we unexpectedly find that p21 is induced in SMC-CTR cells treated with both ZMC1 or Pikan083 indicating that p21 can also be induced via p53 independent mechanisms in RMS cells.
We also added the data from a recent publication that showed head and neck RMS tumors have a higher proportion of TP53 mutations compared to other tumor sites (Shern et al., 2021).
(Author response images 2, 3). This data is consistent with our data in our zebrafish ERMS model where p21 is induced tp53-/- ERMS compared to tumors expressing wt tp53. These data also indicate that ZMC1 and Pikan083 can have different effects on p53Y220C and p53R248W mutant protein expression.
Addition of Pikan083 to EW8 (p53<sup>Y220C</sup>) ewing sarcoma cells results in increased p53 and p21 expression.
p53 and p21 expression in EW8 cells treated DMSO or with 1 and 5 μM ZMC1 or 50 and 100 μM Pikan083. Β-tubulin was used as a loading control.
Addition of ZMC1 and Pikan083 has differential effects on p53 and p21 expression in EW8, RD and SMC-CTR cells.
EW8 (p53Y220C), RD (p53R248W) and SMS-CTR (p53 null) cells treated with 5 µM ZMC1 (top panel) or 100 µM Pikan083 (bottom panel). GAPDH was used as a loading control.
p21 expression across commonly used rhabdomyosarcoma cell lines.
Rh18, Rh36 are WT for p53. RD, JR1, Rh28, Rh30, Rh41 are p53 mutant and SMS-CTR are p53 deficient (Hinson et al., 2013). Β-tubulin was used as a loading control.
e) We also added the data from a recent publication that showed head and neck RMS tumors have a higher proportion of TP53 mutations compared to other tumor sites (Shern et al., 2021).
4. TP53C176F coinjection with kRASG12D in p53-/- fish can reduce kRASG12D-induced tumorigenesis (Figure 5B) and increase apoptosis (Figure 5F), suggesting TP53C176F functions as a hypomorphic mutant. Can authors examine whether the p21 expression is upregulated in the ERMs with TP53C176F overexpression, which would provide further mechanistic insight on the hypomorphic effect of the TP53C176F mutant.
As requested qPCR was performed on newly generated primary tumors. “To assess the effects of p53 variant expression on transcription, we performed qPCR analyses comparing tp53+/+, tp53-/- and tp53-/- + TP53C176F and tp53-/- + TP53P153Δ ERMS and p53 target genes including baxa, cdkn1a, gadd45a, noxa and puma/bbc3. We found that both tp53-/-+TP53C176F and tp53-/- + TP53P153Δ tumors displayed increased bbc3 expression compared to tp53-/- or tp53+/+ tumors, but had no difference with respect to cdkn1a, gadd45a and noxa expression. We also found that tp53-/- + TP53C176F and tp53-/- + TP53P153Δ tumors expressed higher levels of cdkn1a compared to tp53+/+ ERMS; however, rather unexpectedly tp53-/- tumors also express higher levels of cdkn1a compared to tp53+/+ controls (Figure 4 Supplement 3).”
[Editors' note: further revisions were suggested prior to acceptance, as described below.]
There is strong consensus that the manuscript has been substantially improved but there are some remaining issues that need to be addressed, as outlined below:
First, there appears to be a missing item – Supplemental Table I.
We thank the reviewer for pointing this out and apologize for this oversight on our part. While the two versions of the manuscript we submitted have Supplemental Table1. The version generated by the eLife server and that was sent out for review did not have Supplemental Table 1.
All tumors were generated by co-expressing Human TP53 wild-type or mutant alleles along with kRASG12D and a fluorescent reporter in the tp53-/- zebrafish background.
Second, the new findings about the occurrence of medulloblastoma in the tp53P153 model should be mentioned in the Abstract.
We have edited our abstract to include changes requested by the reviewers/editors and also added a line stating that TP53P153Δ when expressed in the kRASG12D-driven ERMS-model can also initiate medulloblastoma tumors.
“TP53P153Δ unexpectedly also predisposes to hedgehog expressing medulloblastomas in the kRASG12D-driven ERMS-model.”
Reviewer #1 (Recommendations for the authors):
Overall, the revised manuscript addressed the essential revisions and the authors have responded to the reviewer's requests. The revised manuscript also presents two new analyses, which weren't mentioned in the abstract or introduction. A new Figure 5 and supporting figures show identification of medulloblastoma in addition to head ERMS arising in the tp53P153delta ERMS model. A new Figure 7 shows experiments to test the dependence of the tp53P153delta ERMS tumor induction on Kdr, which is overexpressed in patient osteosarcoma and the zebrafish model. The Kdr gene and function of the protein could be better introduced, and the experiment have stronger experimental validation as described below. But the kdr experiment illustrates how this model can be used to investigate the in vivo mechanisms by which patient tp53 variants drive ERMS tumorigenesis.
Some issues were not adequately addressed: the statement that the model is high throughput has not been removed from the discussion.
We thank the reviewer for pointing that oversight and have removed all references to the zebrafish model being a high-throughput platform. We instead highlight the fact that we characterize zebrafish wild-type tp53 function and also the human TP53 wild-type allele and three patient specific alleles using the zebrafish ERMS in vivo model.
“Third, we utilized our robust zebrafish ERMS model to assign function to three TP53 mutants whose effects in ERMS were previously unknown.”
Figure 2 A, B quantification data was not included and it's still unclear what the numbers in those panels reflect,
Figure 2 A. For this analysis we counted Edu+ cells from an entire tumor section (dsRED or GFP+ cells) and normalized Edu+ counts to the total number of DAPI positive cells. Each section represents one ERMS tumor so for this analysis we assessed tumors from multiple tumor burdened zebrafish and these counts were used for statistical analyses. Edu and DAPI counts are available in Supplementary excel files of all the raw data which we have uploaded on the Dryad site. In Figure 2A we show one representative tumor each of Edu+ labeled cells in tumors generated in tp53+/+ or tp53-/- (null) zebrafish. We also added a third panel with a plot showing % of Edu-positive cells in the different tumor sections. We have also updated our figure legend and methods section to clearly state how the analysis was performed.
Figure 2 B. Our earlier analyses grouped counts from multiple sections. We agree that these counts can be confusing to interpret the data in the images provided. Accordingly, we have updated our counts to reflect the counts in the image shown. For each tumor, we imaged a single confocal section at 200x magnification. We then counted pHH3-positive cells in each confocal section. Counts (sections) from multiple tumor burdened animals per genotype were then used to calculate absolute numbers and for statistical analyses of cells per fixed area undergoing mitosis. We also added a third panel with a plot showing the total number of pHH3-positive cells in the different tumor sections. The counts are available in Supplementary excel files 2A and 2B which we have uploaded on Dryad site.
and new RT-qPCR on p53 target gene expression contained contradictory or inconsistent results. Different tp53 variants did not lead to consistent elevation of apoptotic genes bax, bbc3, noxa, – tp53C176F and tp53P153delta induced expression of bbc3, but not noxa or bax; tp53Y220C increased noxa and cell growth arrest gene gadd54a; cell cycle arrest gene cdkn1a was elevated in all variant models including the tp53 null. A discussion of this data, in the context of demonstrating that the tp53 variants retain some level of transcriptional activity, was not included.
We thank the reviewer for this feedback. We have included a new section in our discussion to address differences in p53 target genes expressed in the tumors with the different TP53 mutants in the tp53-/- background.
The experiment to demonstrate ERMS tp53P153delta is dependent on elevated kdr expression via somatic CRISPR RNP kdr targeting did not contain much detail. A more thorough analysis of kdr deletion/mutagenesis in the tumors.
We thank the reviewer for this feedback. We have added more details to introduce this experiment including the design of the targeting guide RNAs and why we picked mitfa as an experimental control. We also show that we assessed cutting of DNA in 5 of the 12 kdr group tumors obtained using DNA heteroduplex assays to show that we get the expected large deletion and also multiple other indels and an almost complete loss of the wild-type band (see Figure 7 Supplement 1). Finally, we cloned the PCR products from kdr ablated tumors and sequenced several clones and our analyses shows that our guide RNAs are highly active.
Figure 7 Supplement 1 does not contain enough detail explaining how targeted mutagenesis was induced, documented and measured.
We have added the following to the Methods section and also better annotated Figure 7 Supplement 1.
“Strategy and quantification of mosaic ablation of kdr and mitfa in zebrafish ERMS tumors using CRISPR/Cas9 reagents.
Using Benchling, three high scoring CRISPR guides (with minimal off-target scores) were selected to target DNA sequences in exons 12 and 13 of kdr that encode parts of two Ig repeats in the extracellular domain (aa 535 to 647; InterPro domain IPR003599). Efficient nuclease activity of guides 1 and 3 would result in a 334 bp deletion leading to a predicted frameshift and premature termination of the polypeptide sequence. When this larger deletion does not result from synchronous CRIPSR nuclease activity of guides 1 and 3, similar consequences are likely to result due to independent indels generated from nuclease activity and NHEJ repair at each of the three individual target sites of guides 1, 2 and 3. Since this region is upstream of the transmembrane domain (aa 773-795), we predict that indels or deletions leading to frameshift mutations would, at best, produce a partial extracellular domain containing polypeptide incapable of transducing signals into the cell.
Using PCR and TBE-PAGE, the presence of deletions were visualized by small PCR amplicons corresponding to a loss of 334 bp in a 641 bp wildtype PCR product (deletion mutant amplicon = 212 bp). In addition, presence of heteroduplex products were observed via PAGE indicating the presence of indels. Indels were confirmed by cloning PCR products and sequencing individual clones. Shown in Figure 7 Supplement 1D are indels +21bp, -7 bp at the g1 site, 7 bp deletion at the g2 site and 25 bp deletion at the g3 site. The effectiveness of the three sgRNA/CRISPR guides by way of the spectrum of mutations they cause suggest that despite the mosaicism of the mutations in sgRNA-Cas9 injected embryos, tumors that arise harbor a relatively high frequency loss of function mutations in the kdr gene, in all tumors. mitfa was selected as a control gene because loss-of-function mutations can be easily screened by looking for the loss of pigmentation in developing larvae using brightfield microscopy. In addition to not having an effect on embryo survival and development, it obviates the need to obtain sequence data to confirm the presence of loss of function mutations in the embryos.”
kdr transcript knockdown measured by RT-qPCR, would validate the conclusion that Kdr drives oncogenesis in the ERMS tp53P153delta model.
Given that kdr functions cell autonomously as suggested by (Drummond et al., 2018), where upon trans-differentiation of endothelial cells to rhabdomyosarcoma tumors there is a loss of kdr expression. Accordingly, analyses of kdr mRNA expression in sorted tumors would be required to address cell autonomous requirements. This is technically challenging due difficulties sorting small ERMS tumors and limited number of head tumors obtained in the kdr mutant group. Furthermore, the expected result would be that there is no difference in kdr expression in control or kdr mutant/loss of function tumors. Our experiments are not designed to address whether the effect we observe is cell autonomous or in the supporting vasculature i.e. effects on tumor angiogenesis. While this is an important question, we believe that this question is beyond the scope of the current study. We incorporate a statement in the manuscript stating the limitation of our analyses.
Reviewer #2 (Recommendations for the authors):
The authors have done an admirable job responding to previous reviews, resulting in a manuscript that is more logically constructed and better focused in the data presented.
A few residual comments:
1. I was not able to find the Supplemental Table that was suggested by reviewers and referred to in the Response to Reviewers that summarizes and highlights the main findings and differences between wild type and various tp53 zebrafish mutants described in the manuscript.
We thank the reviewer for pointing this out and apologize for this oversight on our part.
All tumors were generated by co-expressing Human TP53 wild-type or mutant alleles along with kRASG12D and a fluorescent reporter in the tp53-/- zebrafish background.
2. The choice of analyzing the TP53 P153 mutation remains a bit unusual given its rarity and identification as a germline mutation in an aggressive osteosarcoma, given the focus of this manuscript on ERMS (especially now with removal of the osteosarcoma cell line data which contrasted with the zebrafish data).
3. Moreover this P153 mutation leads to Shh medulloblastoma tumors in a subset of zebrafish. This is new data that is interesting, but a bit of a distraction from the focus of the manuscript on ERMS.
We thank the reviewer for this comment but would like to state our reasoning for including the TP53P153Δ mutant analyses. This whole project was initiated because of a patient in our clinic with an aggressive osteosarcoma who responded very poorly to therapy. The patient had a rare germline TP53 variant of unknown significance (Dr. Amanda Lipsitt attending physician-Second author). The rare TP53 variant and the patient history where the mom also succumbed to osteosarcoma suggested to us that this may be a gain of function mutant. Assessing the literature, we were surprised that for such a well-known gene, the majority of mutations in sarcoma (osteosarcoma and rhabdomyosarcoma) were uncharacterized and had a different spectrum from the mutations in adult cancers and from those modelled in mice. Since a robust osteosarcoma model in zebrafish is lacking, we decided to test if the TP53 variant was able to show different effects from the tp53 null background in our zebrafish ERMS model. The advantage of this model is that tumor initiation occurs as early as 7-10 days and in most analyses majority of the tumors initiate by 60 days. Finally, we unexpectedly found that >97% of animals initiate tumors in the tp53 null background. This enabled us to assess differences in tumor initiation and the retention of wild-type function in TP53 variants. However, to understand if our system worked, we had to define human wild-type TP53 and at least one other mutant that was present in a patient with ERMS. These were our controls.
Our unbiased approach has revealed that the zebrafish ERMS model can be used to assess effects on tumor initiation, apoptosis, proliferation, site of tumor initiation and also in the rag2-kRASG12D ERMS model the initiation of medulloblastoma tumors. We also find that the different TP53 mutants/variants can differentially regulate the expression of well-known p53 direct transcriptionally regulated genes puma, noxa, baxa and gadd45a. Specifically, we find that the TP53153Δ is a pathogenic gain of function mutant that predisposes to head ERMS and also the hedgehog-positive sub-type of medulloblastoma tumors. TP53153Δ also directly or indirectly regulates kdr expression to promote head ERMS formation. Medulloblastoma initiation in the rag2-kRASG12D ERMS model has never been reported and strongly supports our finding that TP53153Δ displays gain-of-function effects with respect to site of ERMS initiation and the initiation of medulloblastoma tumors.