The recruitment of effector CD8⁺ T cells into mouse flank skin via a viral infection (e.g. Vaccinia virus) or through an inflammatory stimulus (e.g. ‘DNFB-pull’) results in comparable numbers of long-lived CD103⁺ epidermal T_RM and is independent of the presence or absence of cognate antigen in the skin (Mackay et al., 2012; Hirai et al., 2019; Davies et al., 2017). This, however, has only been demonstrated using HSV-specific (gBT-I) or OVA-specific (OT-I) TCR transgenic T cells. To determine if endogenous polyclonal CD8⁺ T cells share this phenotype, we employed a dual VV infection ‘DNFB-pull’ model. Cohorts of wild-type C57BL/6 mice were infected on the left flank with VV by skin scarification (Figure 1A). The infection expanded CD8⁺ effectors specific for VV antigens, which were then recruited into the left flank in response to the ongoing infection-induced inflammation. DNFB is applied to the right flank 5 days post-infection to recruit into the skin circulating CD8 effectors expanded by infection. Mice were then rested for 50+ days to allow for the formation of T_RM. We have previously found that evaluation of T_RM numbers in the epidermis by flow cytometry is much less accurate than direct immunofluorescent visualization (Hirai et al., 2021). Evaluation of whole-mounted epidermal sheets stained with anti-CD8 revealed comparable numbers of CD8⁺ T cells at the VV-infected (left flank) and DNFB-treated (right flank) sites (Figure 1B and C). Expression of CD103 on T_RM as evaluated by flow cytometry was also similar at both sites (Figure 1D and Figure 1—figure supplement 1A). To evaluate the frequency of antigen-specific T cells at each site, skin from a separate cohort was examined by flow cytometry using the B8R tetramer which recognizes an immunodominant epitope of VV in H2-K^b (Moutaftsi et al., 2006). We observed equivalent numbers of B8R⁺ T_RM at the VV-infected and DNFB-treated sites (Figure 1E and F). From these data, we conclude that polyclonal CD8 T cells form epidermal T_RM with comparable efficiency when recruited into the skin by viral infection or sterile inflammation and that the presence or absence of cognate antigen in the skin has no effect on the number or frequency of antigen-specific T_RM.

Although T_RM efficiently populate the epidermis independently of cutaneous cognate antigen, we have previously demonstrated that T_RM that form at skin sites containing cognate antigen are functionally distinct from bystander T_RM that form in its absence (Hirai et al., 2021). To analyze potential transcriptional differences between local antigen experienced and bystander T_RM, we performed single-cell RNA-seq in collaboration with the ImmgenT consortium (Zemmour et al., 2022). The consortium consists of multiple groups that isolated populations of murine T cells from multiple tissues and contexts and then subjected them to single-cell RNA-seq (Zemmour et al., 2022)(see Methods). 627,692 mature T cells from 80 different experiments and 703 samples were integrated and batch-corrected using totalVI (Korsunsky et al., 2019) (see Methods). Cells were projected in two dimensions using Minimal-Distorted Embedding (see Methods), revealing the expected transcriptional clustering of distinct types of T cells (Figure 1G). As part of the consortium, we isolated cells from the flanks of rested (>65 days) VV-infected or DNFB-treated flank skin by enzymatic digestion. Cells were pooled from 10 mice and purified based on expression of CD90 and CD8 (Figure 1—figure supplement 1B). As expected, most cells expressed CD69 and CD103 (Figure 1—figure supplement 1C). Skin cells from each group and naïve spleen cells were hash tagged and analyzed by single-cell RNA-seq. Analysis of transcripts revealed a total of 6 clusters (0–5) which are shown overlayed on the complete ImmgenT cell embedding (Figure 1H, Figure 1—figure supplement 1D). As expected, most cells clustered with CD8⍺β T cells (Figure 1G, black). There was, however, a high degree of heterogeneity, and some possibly contaminating cells fell outside of this region. A comparison of the relative frequency of cells in each of the clusters (0–5) isolated from the VV or DNFB sites revealed a strong enrichment of cluster 3 in the VV group and of clusters 0 and 2 in the DNFB group (Figure 1I, magenta). This suggested that cluster 3 cells were manifesting a distinct transcriptional state likely induced by antigen encounter in the skin.

To test the hypothesis that cells in cluster 3 could be T_RM that have encountered antigen in the skin during development, we performed signature score enrichment analysis comparing the transcriptional profile of our cells to published core genes of well-defined CD8 T cell states. Signature scores were calculated based on normalized differential gene expression compared to core genes of T_RM generated by acute infection models (GSE47045), activated T cells (T_ACT, GSE10239), circulating memory T cells (T_MEM, GSE41867), and exhausted T cells (T_EX, GS41867) (Jaiswal et al., 2022; Sarkar et al., 2008; Doering et al., 2012). Notably, cells in cluster 3 showed enriched expression of T_RM core genes compared to all other clusters (Figure 2A). We performed a similar analysis using pseudo-bulk analysis of all cells isolated from the VV or DNFB sites. Consistent with the enrichment of cluster 3 in cells from the VV site, we found a strong enrichment of T_RM core genes in the VV group (Figure 2B). These data suggest that cells in cluster 3 isolated from the VV site represent fully differentiated T_RM.

To further test the hypothesis that cluster 3 represents fully differentiated T_RM, we performed signature score enrichment analysis of clusters 0–5 against a dataset of skin T cells isolated at different times following skin VV infection (GSE79805) (Pan et al., 2017). Cells from clusters 0, 2 and 3 showed similar signature scores in comparison with skin T cells up to day 5 post-infection; however, clusters 0 and 2 then plateaued suggesting a lack of continued differentiation (Figure 2C). In contrast, cluster 3 transcripts scored higher at later time points in T_RM differentiation. Signature scores for cells isolated from the DNFB and VV sites were similar at early time points but diverged at later time points with increased scores observed in cells from the VV site (Figure 2D). Taken together, these data suggest that T_RM isolated from skin sites where they can encounter cognate antigen are transcriptionally more fully differentiated T_RM while bystander T_RM isolated from a site lacking cognate antigen remain in a less differentiated state.

Analysis of the 10 highest differentially expressed genes (DEGs) in cluster 3 (Figure 1—figure supplement 1D) showed increased expression of genes associated with T cell activation, including Dusp1 and Nr4a1 as well as the AP-1 family members Junb and Fos (Papavassiliou and Musti, 2020; Owens and Keyse, 2007; Gazon et al., 2017; Schnoegl et al., 2023; Sun et al., 2021). Given the importance of AP-1 family members in T cell proliferation and differentiation, we next tested whether this corresponded to functional differences in T_RM activity (Schnoegl et al., 2023; Atsaves et al., 2019; Shaulian and Karin, 2002; Buquicchio et al., 2024). An important functional property of T_RM is their capacity to rapidly expand following re-encounter with cognate antigen (Szabo et al., 2019; Schenkel and Masopust, 2014). To determine whether local antigen experienced and bystander T_RM have different recall responses, we repeated our dual VV infection and DNFB ‘pull’ model in combination with an OT-I adoptive transfer model to allow for antigen recall using the SIINFEKL peptide. CD90.1⁺ OT-I cells were adoptively transferred into naïve C57BL/6 mice followed by infection with a vaccinia virus expressing SIINFEKL peptide, OVA_257-264 (VV-OVA) on the left flank. On day 5 post-infection, the right flank was treated with DNFB to recruit expanded OT-I effectors to the site (Figure 3A). On day 50+ after infection, topical SIINFEKL peptide or control PBS was painted on both flanks in a single application (1° recall). The total number of OT-I cells in the epidermis 6 days later was determined by immunofluorescent microscopic evaluation of the CD90.1 congenic marker in epidermal whole mounts (Figure 3B–C). Restimulation with peptide led to an expansion of T_RM at both sites compared to PBS-treated controls. Notably, local antigen experienced OT-I T_RM cells in the VV-OVA treated flanks expanded to a larger extent compared to bystander T_RM in DNFB treated flanks. To inhibit potential recruitment of circulating effector OT-I cells, we repeated these experiments administering FTY720 to block migration of T cells from lymph node or titrated anti-Thy1.1 antibody to ablate circulating OT-I but not T_RM (Figure 3D). Both methods successfully depleted OT-I from the blood (Figure 3—figure supplement 1A–D), without affecting T_RM in the skin (Figure 3—figure supplement 1E). Administration of FTY720 (Figure 3E–F) or anti-Thy1.1 treatment (Figure 3G) had no effect on the number of OT-I T_RM in the epidermis after a 1° recall response and local antigen experienced T_RM from the VV-OVA treated flank still expanded to a greater degree than bystander T_RM from the DNFB treated flank. Thus, the increased number of epidermal OT-I at the local antigen experienced VV site during a 1° recall does not appear to result from recruitment of circulating cells.

Finally, to determine whether increased expansion by local antigen experienced T_RM persists with multiple rounds of stimulation, we repeated this experiment but allowed mice to rest for 120 days after the 1° recall response. The number of T_RM at both the VV-OVA and DNFB sites contracted down to equivalent numbers that were increased compared to PBS-treated, control mice (Figure 3C). Mice were then given a 2° recall by a second application of peptide at day +176. The number of T_RM expanded but T_RM at the VV-OVA site increased by a larger amount than T_RM at the DNFB site (Figure 3B–C). Based on these data, we conclude that T_RM encountering antigen in the skin during development expand more efficiently in response to antigen re-encounter and this phenotype persists long-term with subsequent antigen encounters.

We next hypothesized that the increased expansion of local antigen experienced T_RM during a recall response resulted from increased proliferation. To test this, as above, WT mice adoptively transferred with OT-I cells were infected with VV-OVA on the left flank and DNFB on the right flank 5 days later. After at least 50 days of rest, both sides were painted with topical SIINFEKL peptide and mice were administered 2 mg of BrdU i.p. daily (Figure 4A). Epidermal sheets for immunofluorescent visualization were harvested 2 days later. We noted a clear increase in the percent and total number of dividing cells as well as the total number of OT-I T cells in the epidermis at the VV-OVA site compared with the DNFB site (Figure 4B–D). Similar results were obtained by flow cytometry, with an increase in the percentage of BrdU⁺ in T_RM isolated from the VV-OVA site (Figure 4E–G). Staining for Annexin V was equivalent in both groups, suggesting that an altered rate of apoptosis does not contribute to the observed changes in expansion (Figure 4H–I). In sum, local antigen experienced T_RM show augmented proliferation during an antigen recall response compared with bystander T_RM. Moreover, an enhanced proliferative capacity following antigen restimulation can now be added to other parameters of T_RM fitness, including enhanced epidermal persistence when active TGFβ is limited.

We next hypothesized that the strength of TCR signal provided in the skin during T_RM differentiation should correlate with T_RM phenotype at late time points. To generate T_RM with a range of local antigen encounter signal strengths, we wanted to expose newly-recruited T cells to either topical SIINFEKL peptide (N4), or a topical altered peptide ligand (APL) SIYNFEKL (Y3) that has ¼ the avidity for the TCR receptor (Zehn et al., 2009). We hypothesized that encountering an APL would induce an intermediate functional phenotype between full-strength SIINFEKL and no local antigen encounter. To test this hypothesis, WT mice were adoptively transferred with OT-I cells, then infected with VV-OVA on the left flank and 5 days later were challenged with DNFB on the right flank (Figure 5A). One day later, the DNFB site was painted with PBS, SIINFEKL (N4), or an APL SIYNFEKL (Y3). After 50+ days of rest, all mice were challenged in a 1° recall response with topical N4 or control PBS at the DNFB sides (Figure 5A). Epidermal sheets for immunofluorescence were harvested 6 days later. As expected, mice restimulated with PBS showed equivalent numbers of T_RM regardless of which peptide was provided during development (Figure 5B–C). Following a 1° recall response, T_RM exposed to N4 (DNFB +N4) expanded to a level equivalent to T_RM from the VV-OVA site. In contrast, T_RM exposed to Y3 (DNFB +Y3) showed a degree of expansion that was intermediate compared to T_RM that did not experience antigen in the skin (DNFB +PBS) and those that experienced SIINFEKL (DNFB +N4). Thus, the strength of TCR engagement determines the fitness of T_RM based on their capacity to expand during a 1° recall response.

A feature of local antigen experienced T_RM is that they are better able to persist in the epidermis than bystanders when levels of activated TGFβ are limited, either artificially or during competition with newly recruited T_EFF (Hirai et al., 2021). We have previously found that expression of the canonical TGFβ receptors, Tgfbr1 and Tgfbr2 have equivalent expression in local antigen experienced and bystander T_RM (Hirai et al., 2021). However, there is a third TGFβ receptor, TGFβRIII, that lacks a signaling component but functions as a reservoir of ligand for TGFβ, increasing avidity of the TGFβ receptor (Miyazono, 1997; Heldin and Moustakas, 2016). Notably, expression of TGFβRIII has been reported to increase in T cells following TCR ligation (Ortega-Francisco et al., 2017). This was confirmed in vitro with anti-CD3 anti-CD28 stimulated OT-I splenocytes (Figure 6A). We also observed increased expression of TGFβRIII in vivo, comparing local antigen experienced OT-I T_RM with bystanders (Figure 6B–C). Based on this data, we hypothesized that increased expression of TGFβRIII on antigen experienced T_RM could explain their capacity to maintain epidermal residence when available TGFβ is limiting.

To test this, we generated OT-I Thy1.1⁺ E8i-cre^ERT2 huNGFR Tgfbr3^fl/fl (Tgfbr3^∆CD8) mice which allow for inducible ablation of Tgfbr3 from T_RM as well as control OT-I Thy1.1⁺ E8i-cre^ERT2 huNGFR Tgfbr3^WT/WT (Tgfbr3^WT) mice (Figure 6D). To validate these mice, Tgfbr3^∆CD8 mice were treated with 0.05 mg/g tamoxifen i.p. for 5 days. CD8⁺ T cells from tamoxifen-treated Tgfbr3^∆CD8 mice were then isolated from spleen and lymph nodes and cultured for 2 days with anti-CD3 anti-CD28. Cells were evaluated by flow cytometry for expression of TGFβRIII and huNGFR, an indicator of successful induction of cre-mediated excision by tamoxifen. In the huNGFR-negative population where cre was not activated, approximately 50% of the cells expressed TGFβRIII (Figure 6E–F). In contrast, TGFβRIII was largely absent from cells co-expressing huNGFR, indicative of efficient ablation of Tgfbr3.

To ablate TGFβRIII from T_RM once they have fully differentiated in the epidermis, we next adoptively transferred CD8⁺ T cells isolated from either Tgfbr3^WT or Tgfbr3^∆CD8 mice into naïve WT mice followed by skin VV-OVA infection on the left flank. On day 5 post-infection, the right flank was treated with DNFB. After at least 50 days of rest to allow for T_RM formation, both cohorts were treated daily with 0.05 mg/g tamoxifen i.p. to transform approximately 50% of adoptively transferred cells, generating populations of both control huNGFR negative and huNGFR positive OT-I cells. To test the effects of TGFβRIII ablation, TGFβ activation was inhibited for 10 days by i.p. administration of CWHM12, a small molecule inhibitor of integrins ⍺vβ6 and ⍺vβ8 (Figure 6G–H). Analysis of epidermal whole mounts by immunofluorescence revealed that blockade of TGFβ activation reduced the number of bystander Tgfbr3^WT T_RM at the DNFB site, but not local antigen experienced T_RM at the VV-OVA site, consistent with our earlier results (Figure 6I–J). In contrast, local antigen experienced huNGFR⁺ Tgfbr3^∆CD8 T_RM from the VV-OVA site that lack TGFβRIII were efficiently depleted from the epidermis by CWHM12 treatment. The depletion of bystander huNGFR⁺ Tgfbr3^∆CD8 T_RM at the DNFB site was augmented. Notably, numbers of Tgfbr3^∆CD8 T_RM in cohorts treated with vehicle did not induce a statistically significant reduction in steady-state T_RM, indicating that loss of TGFβRIII is not an absolute requirement for T_RM epidermal residence in the steady state. Thus, expression of TGFβRIII by local antigen experienced T_RM which is downstream of TCR ligation is required for their capacity to remain in the epidermis when activated TGFβ is limiting.

We generated Thy1.1⁺Rag1^−/−OT-I mice by crossing OT-I mice with Rag1^−/− and Thy1.1 mice. E8I-creER^T2 and ROSA26.LSL.hNGFR reporter mice were developed by Dario A.A. Vignali (University of Pittsburgh) (Hirai et al., 2019). E8I-creER^T2 mice and Thy1.1⁺Rag1^-/-OT-I mice were crossed with ROSA26.LSL.hNGFR reporter mice to generate Tgfbr3^WT mice and additionally with Tgfbr3^fl/fl mice to obtain Tgfbr3^ΔCD8 mice. Tgfbr3^fl/fl mice were developed by Herbert Y Lin (Program in Membrane Biology/Nephrology, Department of Medicine, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts) (Li et al., 2018). We used age- and sex-matched female mice that were between 6 and 12 weeks of age at the start of all experiments. All mice were maintained under specific-pathogen-free conditions and all animal experiments were approved by the University of Pittsburgh Institutional Animal Care and Use Committee under the Animal Welfare Assurance Number D16-00118 (A3187-01).

Mice were infected by skin scarification (skin infection) with 3 × 10⁶ plaque-forming units recombinant vaccinia virus expressing the SIINFEKL peptide of ovalbumin (VV-OVA), or vaccinia virus without recombinant peptide expression (VV). For skin scarification, 45 μL of VV-OVA or VV was applied to shaved left flank (4–5 cm²) and the skin were gently scratched 100 times with 27 G needle under anesthesia. VV or VV-OVA infected mice were further treated with 0.15% 1-Fluoro-2,4-dinitrobenzene (DNFB, D1529: Sigma-Aldrich) in 4:1 acetone: olive oil (O1514, Sigma-Aldrich) on the flank opposite the site of infection (40 μL) 5 days post infection. In some experiments, 4:1 acetone and olive oil was added to the stock solution of OVA_257–264 (SIINFEKL or SIYNFEKL, RP101611, Genscript/Fisher) in DMSO (D2650, Sigma-Aldrich) (10 mg/mL) for a final concentration of OVA_257–264 at 0.2 mg/mL and 50 μL was painted to the DNFB-treated skin 1 day after DNFB treatment.

50+days after VV-OVA infection. 4:1 acetone and olive oil was added to the stock solution of OVA_257–264 (SIINFEKL, RP101611, Genscript/Fisher) in DMSO (D2650, Sigma-Aldrich) (10 mg/mL) for a final concentration of OVA_257–264 at 0.2 mg/mL and 50 μL was painted to both flanks.

For blocking TGFβ activation with a small molecule integrin inhibitor, compound CWHM12 (kindly provided by Indalo Therapeutics, Cambridge, MA) was solubilized in 100% DMSO. Further dilution to 50% DMSO was made in sterile 100% PBS and dosed to 100 mg per kg body weight per day delivered by i.p. injections once a day for 10 days.

OT-I cells were purified from spleen and lymph nodes by MojoSort Mouse CD8 T Cell Isolation Kit (48007, BioLegend) according to the manufacturer’s instructions and 1 × 10⁵ OT-I cells were intravenously transferred in all experiments. All mice were allowed to rest for 1 day before further experimentation.

Tamoxifen (T5648; Sigma-Aldrich) was dissolved in 1/10th volume of 200 proof ethanol following incubation at 37°C for 15–30 min with 300 rpm shaking. Corn oil (C8267, Sigma-Aldrich) was added for a final concentration of Tamoxifen at 10 mg/ml and was administered to mice for 5 consecutive days by intraperitoneal injection at 0.05 mg per g body weight, in order to transform approximately 50% of adoptively transferred cells.

BrdU (B5002, Sigma-Aldrich) was dissolved in sterile PBS. 2 mg in 200 uL of PBS was injected i.p. daily into each mouse for 2 consecutive days before harvest.

Epidermal sheets were prepared as previously described (Mohammed et al., 2016). Briefly, 2 square cm sections were harvested 2 cm distal from the mid-clavicular line for consistency, and then the epidermal side of shaved defatted flank skin was affixed to slides with double-sided adhesive (3M, St. Paul, MN). Slides were incubated in 10 mM EDTA in PBS for 90 min at 37°C, followed by physical removal of the dermis. The epidermal sheets were fixed in 4% PFA at RT for 15 min. The epidermal sheets were blocked with PBS containing 0.1% Tween-20, 2% BSA and 2% rat serum for 1 hr at RT before staining for 1 hour with antibodies at RT in PBS containing 0.1% Tween-20 and 0.5% BSA. The slides were mounted with ProLong Gold Antifade Mount with DNA stains DAPI (P36931, Thermofisher). Images were captured on a Keyence BZX800 fluorescent microscope (Keyence, Osaka, Japan). Analysis was performed using BZ-H4A Advanced Analysis Software (Keyence, Osaka, Japan). For the enumeration of cells, three images from distant sites randomly determined within an epidermal sheet from a mouse were counted (total 3 mm²) manually after image processing by Adobe Photoshop (version 6) and the average number per mm² epidermis was calculated as representative of the epidermal sheet. Anti-CD8α (53–6.7), Thy1.1 (OX-7) and huNGFR (ME20.4) were purchased from BioLegend.

For depletion of circulating memory OT-I cells, OT-I adoptive transfer VV-OVA and DNFB treated mice were injected i.p. with 0.3–1 μg anti-Thy1.1 (HIS51, ThermoFisher) in 200 μL PBS for 10 consecutive days. Depletion of OT-I T cells (<0.5% of total CD8⁺ T cells and <1%) were confirmed by staining with anti-Thy1.1 (OX-7) using blood 3 days after depletion. For FTY720 treatment, OT-I adoptive transfer VV-OVA and DNFB-treated mice were injected i.p. with FTY720 (10006292, Fisher/Cayan Chemical) at 1 μg/g body weight in sterile PBS with 0.5% DMSO (D2650, Sigma-Aldrich) daily for 6 days. Depletion of circulating cells was confirmed with staining of anti-Thy1.1 (OX-7) and anti-CD8α (53–6.7) using blood 3 days after the end of treatment.

Preparation of single cell suspension from skin, shaved skin was harvested, and fat tissues was removed by forceps mechanically. The skin was then mechanically digested with either scissors or a gentleMACS Dissociator (Miltenyi Biotec, Bergisch Gladbach, Germany), and then resuspended in RPMI 1640 media (Gibco, Grand Island, NY) containing 2.5 mg/ml Collagenase XI (9001-12-1, Sigma-Aldrich), 0.1 mg/ml DNase (04536282001, Sigma-Aldrich), 0.01 M 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (7365-45-9, Sigma-Aldrich), and 10% fetal bovine serum, followed by incubation in a shaking incubator for 30 min at 37°C. The resulting cell mesh was filtered through a 40-μm cell strainer (BD Biosciences, San Jose, CA). The single-cell suspension was then resuspended in 44% Percoll (89428-524, VWR) in RPMI and layered over 67% Percoll in PBS. After 20 min of centrifugation at 931 x g at RT, the interface was isolated and resuspended in FACS buffer. LNs (axillary and inguinal) were incubated in 400 U/ml Collagenase D (Roche Applied Science, Penzberg, Germany) and 0.1 mg/ml DNase in RPMI 1640 with 10% fetal bovine serum for 40 min at 37°C and then minced through a 40-μm cell strainer. In some experiments, LNs were heat-treated at 60°C in a hot water bath for 1 hour. Blood was collected with heparin (H3393, Sigma-Aldrich) and treated with red blood cell lysing buffer (R7757, Sigma-Aldrich). Single-cell suspensions were blocked with 2.4G2 culture supernatant (ATCC, Manassas, VA). Surface staining was performed in standard FACS buffer for 30 min at 4°C. Anti-CD8α (53–6.7), CD44 (IM7), CD3 (17A2), Thy1.2 (30-H12), Thy1.1 (OX-7), CD69 (H1.2F3), CD103 (2E7), BrdU (3D4), MHC-II (M5/114.15.2), CD45.2 (104), and Annexin V were purchased from BioLegend. Anti-TGFßRIII (1C5H11) was purchased from Sigma-Aldrich. Soluble tetrameric B8R_20–27/H2-K^b complex (made by the NIH Tetramer Core Facility) was conjugated to PE-labeled streptavidin. A BD LSRFORTESSA (BD Biosciences) and FlowJo software (TreeStar, Ashland, OR) were used for analysis.

For each sample, 500–5000K DAPI^- CD45⁺ Thy1.2⁺ CD8a⁺ (skin samples) or DAPI^- CD45⁺ Thy1.2⁺ CD8a⁺ CD44⁺ CD103⁺ (LN samples) T cells were sorted on the FACSAria III and pooled in a single collection tube at 4C. DAPI was added just before the sort.

The pooled single-cell suspension was stained with the ImmGenT TotalSeq-C custom mouse panel, containing 128 antibodies (BioLegend Part no. 900004815) and FcBlock (Bio X Cell #BE0307). Because 500,000 cells are required for staining, unstained splenocytes were spiked in to reach a total of 500,000 cells.

Cells were subsequently sorted a second time with the addition of DAPI to select for live T cells, and include 5000 total splenocyte standard cells that were sorted into a single collection tube.