CTRP10 protein is highly conserved from zebrafish to human (Figure 1A), with an amino acid identity of 67%, 71%, 77%, and 94% between the full-length human protein and the fish, frog, chicken, and mouse orthologs, respectively. The conservation is much higher at the C-terminal globular C1q domain (93–100% identity) between the orthologs. Among the 12 different mouse tissues examined, the brain had the highest expression of Ctrp10 (Figure 1B), consistent with previous findings (Iijima et al., 2010). Expression of Ctrp10 in most peripheral tissues was lower than the brain (Figure 1B). We first determined whether Ctrp10 expression is modulated by nutrition and metabolic state. Male mice (~10 weeks old) were subjected to fasting and refeeding. In the refed period after an overnight fast, we observed a significant downregulation of Ctrp10 in the visceral (gonadal) white adipose tissue (gWAT), liver, skeletal muscle, kidney, cerebellum, cortex, and hypothalamus relative to the fasted state (Figure 1C). Next, we examined whether an obesogenic diet alters the expression of Ctrp10. Male mice fed a high-fat diet (HFD) for 12 weeks had a modest increase in Ctrp10 expression in brown adipose tissue (BAT) and heart and decreased expression in skeletal muscle relative to mice fed a control low-fat diet (LFD; Figure 1D). These data indicate that Ctrp10 expression is dynamically regulated by acute alterations in energy balance, and perhaps to a lesser extent to chronic changes in nutritional state on an obesogenic diet.

We used mice lacking CTRP10 to address whether this secreted protein has a metabolic role in vivo. The mouse Ctrp10 gene consists of two exons (Figure 1E). The CRISPR-Cas9 method was used to remove the entire protein coding region spanning exon 1 and 2, thus ensuring a complete null allele (Figure 1E–F). The targeted allele was confirmed by sequencing. As expected, based on the gene deletion strategy, the Ctrp10 transcript was absent from KO mice (Figure 1G).

We first addressed the requirement of CTRP10 in maintaining baseline metabolic homeostasis by assessing metabolic parameters of mice fed a LFD. The body weight and body composition of male mice fed a LFD were not different between genotypes (Figure 2A–B). By 20 weeks of age, female KO mice fed LFD had a modestly higher body weight relative to WT controls (Figure 2C), although the body composition was not different between genotypes (Figure 2D). Food intake, physical activity, and energy expenditure as measured by indirect calorimetry were also not different between genotypes of either sex across the circadian cycle (light and dark) and metabolic states (ad libitum fed, fasted, refed; Figure 2E–J). Because Ctrp10 expression is regulated by nutritional states (Figure 1C), we assessed serum metabolite levels in WT and KO mice in response to fasting and refeeding. No significant differences in fasting and refeeding blood glucose, serum insulin, triglyceride (TG), cholesterol, non-esterified free fatty acids (NEFA), and β-hydroxybutyrate levels were observed between genotypes of either sex, except the female KO mice had slightly lower fasting β-hydroxybutyrate levels (Figure 3A–B). We performed glucose and insulin tolerance tests to determine any potential differences in glucose handling capacity and insulin sensitivity. No significant differences in glucose and insulin tolerance were noted between genotypes of either sex (Figure 3C–F). Together, these data indicate that CTRP10 is dispensable for metabolic homeostasis when mice are young (<20 weeks old) and fed a LFD.

Because female KO mice were slightly heavier at 20 weeks of age (Figure 2C), we suspected the weight may diverge further with age. Consequently, we monitored the body weight of female mice fed LFD over an extended period. Indeed, the female KO mice gained significantly more weight and adiposity with age (Figure 4A–C). By the time the mice reached 40 weeks of age, female KO mice weighed ~6 g (20 %) heavier than the WT controls. Consistent with greater adiposity, the adipocyte cell size (cross-sectional area) was also significantly larger in both gonadal (visceral) white adipose tissue (gWAT) and inguinal (subcutaneous) white adipose tissue (iWAT) (Figure 4D–E). Increased weight gain over time was not attributed to differences in food intake, as measured manually over a 24 hr period (Figure 4F). Fecal output, frequency, and energy content were also not different between genotypes (Figure 4G), suggesting that weight gain was not due to greater nutrient absorption. Deep colon temperatures during both light and dark cycle were also not different between genotypes (Figure 4H). Indirectly calorimetry analyses also revealed no significant differences between genotypes in food intake, physical activity, and energy expenditure across the circadian cycle and metabolic states (ad libitum fed, fasted, refed; Figure 4I–K). Despite significantly greater body weight and adiposity, female KO mice had the same metabolic profile as the lean WT controls. Given the important role of sex hormone in systemic metabolism, we measured serum estradiol levels and they were not significantly different between WT and KO female mice (Figure 4—figure supplement 1). There were no differences in fasting blood glucose, serum insulin, TG, cholesterol, NEFA, and β-hydroxybutyrate levels between genotypes (Figure 4L). Interestingly, VLDL-TG levels were lower in female KO mice whereas HDL-cholesterol level was not different between genotypes (Figure 4M). Direct assessments of glucose handling capacity and insulin sensitivity by glucose and insulin tolerance tests, respectively, also revealed no differences between genotypes (Figure 4N–O). Measurements of mitochondrial respiration in skeletal muscle (gastrocnemius) showed no significant differences in respiration through complex I (CI), CII, and CIV between genotypes; in contrast, Ctrp10 KO female mice had higher respiration at CIV relative to WT controls (Figure 4—figure supplement 2). Together, these data indicate that Ctrp10-KO female mice fed LFD develop obesity, but preserve a largely healthy metabolic profile similar to the much leaner WT female mice.

Next, we challenged the mice with a HFD to determine if the sex-dependent effects on body weight become more pronounced. When fed a HFD, body weight gain and body composition were not different between genotypes in male mice (Figure 5A–B). Food intake, physical activity, and energy expenditure were also not different between genotypes in male mice across the circadian cycle (light and dark) and metabolic states (ad libitum fed, fasted, refed; Figure 5C–E). In striking contrast, female KO mice gained weight rapidly on HFD (~9 g or 28% heavier) and had greater adiposity than the WT controls (Figure 5F–H). Weight gain in female mice was not due to differences in food intake, body temperature, or changes in fecal output, frequency, and energy content between genotypes (Figure 5—figure supplement 1). We measured serum estradiol levels and they were not significantly different between WT and KO female mice on HFD (Figure 5—figure supplement 2). Metabolic cage analysis also revealed no differences in caloric intake, physical activity, and energy expenditure between genotypes in female mice (Figure 5I–K). The ANCOVA analysis of energy expenditure using body weight as a covariate also did not reveal any differences between genotypes in female mice (Figure 5L). Interestingly, the respiratory quotient (RER) was significantly lower in female KO mice relative to WT controls, especially during fasting and refeeding (Figure 5M), suggesting a greater reliance on lipid substrates for energy metabolism during those periods. Interestingly, hepatic, but not skeletal muscle, mitochondrial respiration through CII and CIV was lower in Ctrp10 KO female mice relative to WT controls on HFD (Figure 5—figure supplement 3). Together, these data indicate that CTRP10 is required for female-specific body weight control in response to caloric surplus, but neither food intake, physical activity level, nor whole-body energy expenditure could account for the marked increase in body weight and adiposity.

To evaluate whether the increased adiposity in KO females led to metabolic dysregulation, we again measured fasting and refeeding responses in WT and Ctrp10-KO mice fed a HFD. No differences in fasting and refeeding blood glucose, serum insulin, TG, cholesterol, NEFA, and β-hydroxybutyrate levels were noted between genotypes in male mice (Figure 6A). Female KO mice, however, had higher fasting blood glucose and serum insulin levels, and lower β-hydroxybutyrate levels compared to the WT controls (Figure 6B). In the refed state, serum insulin levels continued to be significantly higher in female KO mice. Unlike the WT female mice where refeeding markedly lowered serum β-hydroxybutyrate (ketone) levels as expected, KO female mice appeared unable to suppress serum β-hydroxybutyrate levels in response to refeeding (Figure 6B).

Consistent with the fasting blood glucose and insulin data, glucose handling capacity and insulin sensitivity assessments by glucose and insulin tolerance in HFD-fed male mice also revealed no differences between genotypes (Figure 6C–D). Female KO mice, however, had higher fasting blood glucose and serum insulin levels suggesting the presence of mild insulin resistance (Figure 6B). We therefore expected to see differences in either glucose and/or insulin tolerance tests. To our surprise, the rate of glucose clearance in response to glucose or insulin injection was virtually identical between WT and KO female mice (Figure 6E–F), suggesting no difference in insulin sensitivity between genotypes. VLDL-TG and HDL-cholesterol profiles were also indistinguishable between WT and KO female mice (Figure 6G). Altogether, these data indicate that CTRP10 is not required for metabolic homeostasis in male mice challenged with a HFD. In female mice, however, loss of CTRP10 markedly promotes weight gain in the face of caloric surplus, but, paradoxically, the excess adiposity is largely uncoupled from obesity-linked insulin resistance and dysregulated glucose and lipid metabolism.

Consistent with greater fat mass in visceral (gonadal) fat depot of Ctrp10-KO female mice (Figure 7A), histological analysis and quantification also indicated significantly larger adipocyte cell size (Figure 7B). Although the subcutaneous (inguinal) fat pad weight was also significantly heavier in female KO mice (Figure 7C), the adipocyte cell size was marginally bigger but not significant (Figure 7D). A bigger fat pad with only marginally larger cell size suggests greater adipocyte hyperplasia in the subcutaneous depot. Increased adipogenesis in response to caloric surfeit is known to be associated with improved systemic metabolic profile (Vishvanath and Gupta, 2019).

Obesity is known to be associated with low-grade inflammation (Hotamisligil, 2006), fibrosis (Sun et al., 2013a), and ER and oxidative stress (Masschelin et al., 2019; Hotamisligil, 2010). Despite marked differences in body weight and adiposity, the expression of genes associated with inflammation (except for Ccr2), fibrosis, oxidative stress in gWAT and iWAT were not significantly different between female KO mice and WT controls (Figure 7E). The expression of some genes associated with ER stress (e.g. Ddit3/CHOP, Atf4, Xbp1) in iWAT were actually lower in female KO mice (Figure 7E). Corroborating the gene expression data, quantification of hydroxyproline (marker of fibrosis), and malondialdehyde (marker of oxidative stress) revealed no significant differences between genotypes (Figure 7F–G).

The liver weight of female KO mice was modestly increased (Figure 7H), but when normalized to body weight, it was not significantly different from WT controls (2.76% in WT and 2.60% in KO, p=0.24). Histological analysis and quantification revealed no differences in hepatic lipid content (% lipid area) between genotypes (Figure 7I). Interestingly, although hepatic fat content was similar between genotypes, the expression of lipogenic genes (e.g. Fasn) was lower and fat catabolism genes (e.g. Cpt2, Ppara, Acadl, Acadm, Acad11, Acadvl) was higher in female KO mice (Figure 7J). The expression of genes associated with inflammation, fibrosis, ER and oxidative stress in liver were not significantly different between genotypes (Figure 7J). Consistent with the gene expression data, quantification of hydroxyproline (marker of fibrosis) in the liver revealed no significant difference between genotypes (Figure 7K). The Ctrp10 KO female mice, however, had higher levels of malondialdehyde (a marker of oxidative stress) in the liver, suggesting a modest increase in oxidative stress (Figure 7L). Altogether, these data indicate that obesity is largely uncoupled from inflammation, fibrosis, ER, and oxidative stress in Ctrp10 KO female mice.

To define the specific mechanisms mediating the female-specific effects of Ctrp10 ablation on favorable metabolic outcomes, four major metabolic tissues (gWAT, iWAT, liver, and skeletal muscle) from female WT and KO mice fed a HFD were subjected to RNA-sequencing (Figure 8—source data 1–8). Comparison of differentially expressed genes (DEGs) via limma (Ritchie et al., 2015) showed robust changes across tissues, with the largest changes seen in the liver (Figure 8A–D). In liver, gWAT, and muscle, we observed comparable numbers of DEGs that were up- and down-regulated, whereas more genes were transcriptionally suppressed in the iWAT of Ctrp10 KO female mice (Figure 8E, top panel). While significant DEGs were identified in all four tissues, only limited overlap was observed between the DEGs in each tissue (Figure 8E, bottom panel). Gene set enrichment analyses of the DEGs highlighted distinct and shared processes up- or down-regulated across the four tissues (Figure 8F). Pathways and processes related to lipid metabolism and estrogen receptor were the top-ranked up-regulated enrichments across tissues (Figure 8F, top panel), whereas processes related to blood clotting and lipoprotein metabolism were the top-ranked down-regulated enrichments (Figure 8F, bottom panel).

Of the DEGs, we found significant changes across tissues in relevant classes of genes that encode proteins involved in gene expression (e.g. transcription factors), signaling (e.g. receptors), tissue crosstalk (e.g. secreted proteins), and metabolism (Figure 9). Notably, the nuclear receptor, Nr1d1 (also known as Rev-Erbα), is the only gene consistently suppressed across all four tissues (liver, gWAT, iWAT, and muscle) of Ctrp10 KO female mice (Figure 8E lower panel and Figure 9). Interestingly, global deletion of Nr1d1 promotes lipogenesis, adipose tissue expansion, and obesity (Delezie et al., 2012; Hand et al., 2015). Although the whole-body and adipose-specific Nr1d1 KO mice fed with HFD become markedly obese, the obesity is not accompanied by insulin resistance, adipose tissue inflammation, and fibrosis (Hand et al., 2015; Hunter et al., 2021). Like the Ctrp10 KO female mice, HFD-fed mice lacking NR1D1 can maintain a relatively healthy metabolic profile despite being strikingly obese. Since only male mice were used in these previous studies, we do not know whether female mice lacking NR1D1 would also exhibit similar insulin-sensitive obesity phenotype. In WT mice, NR1D1 acts as a transcriptional repressor of metabolic genes whose expression are upregulated by high-fat feeding; loss of NR1D1 is thought to result in the de-repression of these genes, leading to greater lipid synthesis and fat mass accrual in response to caloric excess (Hunter et al., 2021). Thus, the suppression of Nr1d1 expression—mimicking NR1D1 deficiency—across tissues in HFD-fed Ctrp10 KO female mice may contribute to benign fat mass expansion without the accompanying adipose tissue fibrosis, inflammation, and oxidative stress.

Among the upregulated genes, Fgf21 and Fgf1 were significantly elevated in the liver of Ctrp10 KO female mice (Figure 9). FGF21 is an hepatokine known to improve systemic insulin sensitivity and to promote a favorable metabolic profile in diet-induced obese mice (Xu et al., 2009). While hepatic Fgf21 expression is significantly elevated, its systemic circulating levels were only found to be higher (p=0.051) in LFD-fed, but not HFD-fed, KO female mice (Figure 9—figure supplement 1); this suggests liver-derived FGF21 may be retained and act locally within the liver parenchyma. FGF1 is another secreted protein that has been shown to dampen hepatic glucose output by suppressing adipose lipolysis (Sancar et al., 2022), improve systemic insulin sensitivity by reducing adipose inflammation (Zhao et al., 2020), and alleviate hepatic steatosis, inflammation, and insulin resistance (Fan et al., 2019). Thus, upregulated expression of Fgf21 and Fgf1 in Ctrp10 KO female mice could contribute to the MHO phenotype. In addition, the upregulated hepatic expression of IL-22 receptor (Il22ra1) in Ctrp10 KO female mice (Figure 9) may confer protection against obesity-associated fatty liver, inflammation and fibrosis (Kong et al., 2012; Yang et al., 2010; Wang et al., 2014). Further, a marked increase in uncoupling protein 3 (Ucp3) and Krüppel-like factor 15 (Klf15) expression in the skeletal muscle (Figure 9) may promote lipid utilization and help mitigate lipid-induced insulin resistance in Ctrp10 KO female mice (Choi et al., 2007; Fan et al., 2021). Taken together, these combined changes—at the level of gene expression and biological pathways and processes across tissues—acting in concert likely contribute to the apparently healthy obesity phenotype seen in the KO female mice.

Next, we asked whether the female-specific transcriptomic effects across tissues were conserved in humans. To address this, we analyzed transcriptional co-correlation of mouse DEG (Figure 10) orthologues in GTEx (GTEx Consortium, 2013), consisting of 210 males and 100 females filtered for comparison of gene expression across tissues (Velez et al., 2022; Sarver et al., 2023). Hierarchical clustering of transcriptional correlation of the orthologous DEGs among four metabolic tissues—subcutaneous and visceral white adipose tissue, liver, and skeletal muscle—showed differing patterns of gene connectivity between females (Figure 10A) and males (Figure 10B). When grouped according to sex in each tissue, the degree of sex-specific gene correlation pairs of DEGs orthologues showed the most significant differences in subcutaneous adipose tissue (Figure 10C). Given the whole-body metabolic effects of Ctrp10 ablation in mice, we further examined the degree of sex-dependent DEG co-correlation across metabolic tissues. This analysis showed that human orthologous genes in subcutaneous adipose tissue (Figure 10D, top row) and liver (Figure 10D, third row) also exhibited highly significant sex differences in their transcriptional correlation with other DEGs across key metabolic tissues (Figure 10D). These analyses highlight the sex-specificity of CTRP10 DEG orthologues in humans, suggest possible sex-biased mechanisms of tissue crosstalk, and overall underscores the conservation of the sex-dependent metabolic function of CTRP10.

Eight-week-old mouse tissues (gonadal and inguinal white adipose tissues, interscapular BAT, liver, heart, skeletal muscle, kidney, pancreas, cerebellum, cortex, hippocampus, hindbrain, and hypothalamus) from C57BL/6J male mice (The Jackson Laboratory, Bar Harbor, ME) were collected from fasted and refed experiments as we have previously described (Rodriguez et al., 2019). For the fasted group (~10 weeks old), food was removed for 16 hr (beginning 10 hr into the light cycle), and mice were euthanized 2 hr into the light cycle. For the refed group (~10 weeks old), mice were fasted for 16 hr and refed with chow pellets for 2 hr before being euthanized. Tissues (white and brown adipose tissues, liver, whole brain, kidney, spleen, heart, skeletal muscle, pancreas, small intestine, and colon) from C57BL/6 J male mice fed a LFD or a HFD for 12 weeks were also collected as we have previously described (Rodriguez et al., 2019).

The Ctrp10/C1ql2-KO mice (C57BL/6NCrl-C1ql2em1(IMPC)Mbp/Mmucd; stock number 050587-UCD) were generated using the CRISPR-cas9 method at UC Davis. The two guide RNAs (gRNA) used were 5’-CCGGCGCCGCTCCACCATTACCT-3’ and 5’-TCAGGCCACCCCATCCCCATCGG-3’. The Ctrp10 gene consists of two exons. The entire protein coding region spanning exon 1 and 2 was deleted. This KO strategy ensures a complete null allele for Ctrp10. The KO mice were maintained on a C57BL/6J genetic background. Genotyping primers for WT allele were forward (m10-Com-F) 5’-TGTCGGGCTCTTCGACTCTCCA-3’ and reverse (m10-WT-R) 5’-GCATCTCGTAGTGAGCCGCTCC-3’. The size of the WT band was 360 bp. Genotyping primers for the Ctrp10 KO allele were forward (m10-Com-F) 5’-TGTCGGGCTCTTCGACTCTCCA-3’ and reverse (m10-Mut-R1) 5’-GTCCAATCAGCTTTCTCAAGTCTGG-3’. The size of the KO band was 422 bp. The genotyping PCR parameters were as follows: 94 °C for 5 min, followed by 10 cycles of (94 °C for 10 s, 65 °C for 15 s, 72 °C for 30 s), then 25 cycles of (94 °C for 10 s, 55 °C for 15 s, 72 °C for 30 s), and lastly 72 °C for 5 min. Due to the presence of GC-rich sequences, 7% DMSO was included in the PCR genotyping reaction. Mice were generated by intercrossing Ctrp10 heterozygous (+/-) mice, supplemented with intercrossing WT or KO mice. Ctrp10 KO (-/-) and WT (+/+) controls were housed in polycarbonate cages on a 12 hr light–dark photocycle with ad libitum access to water and food. Mice were fed either a control LFD (10% kcal derived from fat; # D12450B; Research Diets, New Brunswick, NJ) or a HFD (60% kcal derived from fat; #D12492, Research Diets). LFD was provided for the duration of the study, beginning at 5 weeks of age; HFD was provided for 14 weeks, beginning at 6–7 weeks of age. At termination of study, all mice were fasted for 2 hr and euthanized. Tissues were collected, snap-frozen in liquid nitrogen, and kept at –80 °C until analysis. All mouse protocols (protocol # MO22M367) were approved by the Institutional Animal Care and Use Committee of the Johns Hopkins University School of Medicine. All animal experiments were conducted in accordance with the National Institute of Health guidelines and followed the standards established by the Animal Welfare Acts.

Body composition analyses for total fat, lean mass, and water content were determined using a quantitative magnetic resonance instrument (Echo-MRI-100, Echo Medical Systems, Waco, TX) at the Mouse Phenotyping Core facility at Johns Hopkins University School of Medicine.

LFD- or HFD-fed WT and Ctrp10 KO male and female mice were used for simultaneous assessments of daily body weight change, food intake (corrected for spillage), physical activity, and whole-body metabolic profile in an open flow indirect calorimeter (Comprehensive Laboratory Animal Monitoring System, CLAMS; Columbus Instruments, Columbus, OH) as previously described (Sarver et al., 2020). In brief, data were collected for 3 days to confirm mice were acclimatized to the calorimetry chambers (indicated by stable body weights, food intakes, and diurnal metabolic patterns), and data were analyzed from the fourth day. Rates of oxygen consumption (V˙_O2; mL·kg^–1·h^–1) and carbon dioxide production (V˙_CO2; mL·kg^-1·h^–1) in each chamber were measured every 24 min throughout the studies. Respiratory exchange ratio (RER = V˙_CO2/V˙_O2) was calculated by CLAMS software (version 5.66) to estimate relative oxidation of carbohydrates (RER = 1.0) versus fats (RER = 0.7), not accounting for protein oxidation. Energy expenditure (EE) was calculated as EE = V˙_O2× [3.815 + (1.232×RER)] and normalized to lean mass. Because normalizing to lean mass can potentially lead to overestimation of EE, we also performed ANCOVA analysis on EE using body weight as a covariate (Tschöp et al., 2011). Physical activities were measured by infrared beam breaks in the metabolic chamber.

To independently confirm the food intake data collected in the metabolic cage (CLAMS), we also performed 24 hr food intake measurements manually. All mice were singly housed, with wire mesh flooring inserts over a piece of cage paper on the bottom of the cage. A known weight of food pellets was given to each mouse. Twenty-four hours later, the leftover food pellets remaining on the flooring insert, along with any spilled crumbs on the cage paper, were collected and weighed. Thus, food intake was corrected for spillage.

All tolerance tests were conducted as previously described (Lei and Wong, 2019; Rodriguez et al., 2016; Tan et al., 2016). For glucose tolerance tests (GTTs), mice were fasted for 6 hr before glucose injection. Glucose (Sigma, St. Louis, MO) was reconstituted in saline (0.9 g NaCl/L), sterile-filtered, and injected intraperitoneally (i.p.) at 1 mg/g body weight (i.e. 10 μL/g body weight). Blood glucose was measured at 0, 15, 30, 60, and 120 min after glucose injection using a glucometer (NovaMax Plus, Billerica, MA). For insulin tolerance tests (ITTs), food was removed 2 hr before insulin injection. 6.5 μL of insulin stock (4 mg/mL; Gibco) was diluted in 10 mL of saline, sterile-filtered, and injected i.p. at 0.75 U/kg body weight (i.e. 10 μL/g body weight). Blood glucose was measured at 0, 15, 30, 60, and 90 min after insulin injection using a glucometer (NovaMax Plus).

Mice were fasted overnight (~16 hr) then reintroduced to food as described (Sarver et al., 2022a). Blood glucose was monitored at the 16 hr fast time point (time = 0 hr refed) and at 2 hr into the refeeding process. Serum was collected at the 16 hr fast and 2 hr refed time points for insulin ELISA, as well as for the quantification of TG, cholesterol, NEFAs, and β-hydroxybutyrate levels.

Tail vein blood samples were allowed to clot on ice and then centrifuged for 10 min at 10,000 x g. Serum samples were stored at –80 °C until analyzed. Serum TGs and cholesterol levels were measured according to manufacturer’s instructions using an Infinity kit (Thermo Fisher Scientific, Middletown, VA). NEFAs were measured using a Wako kit (Wako Chemicals, Richmond, VA). Serum β-hydroxybutyrate (ketone) concentrations were measured with a StanBio Liquicolor kit (StanBio Laboratory, Boerne, TX). ELISA kits were used to measure serum insulin (Crystal Chem, Elk Grove Village, IL; cat # 90080), estradiol (Calbiotech, El Cajon, CA; cat # ES380S), and FGF21 (R&D Systems, Minneapolis, MN; cat # MF2100) according to the manufacturer’s instructions. Hydroxyproline assay (Sigma Aldrich, MAK008) was used to quantify total collagen content in liver and adipose tissues according to the manufacturer’s instructions. Lipid peroxidation levels (marker of oxidative stress) in the liver and adipose tissues were assessed by the quantification of malondialdehyde (MDA) via Thiobarbituric Acid Reactive Substances (TBARS) assay (Cayman Chemical, 700870) according to the manufacturer’s instructions.

Food was removed for 2–4 hr (in the light cycle) prior to blood collection. Sera collected from mice were pooled (n=6–7/genotype) and sent to the Mouse Metabolism Core at Baylor College of Medicine for analysis. Serum samples were first fractionated by fast protein liquid chromatography (FPLC). A total of 45 fractions were collected, and TG and cholesterol in each fraction was quantified.

Inguinal (subcutaneous) white adipose tissue (iWAT), gonadal (visceral) white adipose tissue (gWAT), and liver were dissected and fixed in formalin. Paraffin embedding, tissue sectioning, and staining with hematoxylin and eosin were performed at the Pathology Core facility at Johns Hopkins University School of Medicine. Images were captured with a Keyence BZ-X700 All-in-One fluorescence microscope (Keyence Corp., Itasca, IL). Adipocyte (gWAT and iWAT) cross-sectional area (CSA), as well as the total area covered by lipid droplets in hepatocytes were measured on hematoxylin and eosin-stained slides using ImageJ software (Schneider et al., 2012). For CSA measurements, all cells in one field of view at ×100 magnification per tissue section per mouse were analyzed. Image capturing and quantifications were carried out blinded to genotype.

Fecal pellet frequency and average fecal pellet weight were monitored by housing each mouse singly in clean cages with a wire mesh sitting on top of a cutout cardboard that lay at the bottom of the cage for fecal collection. The number of fecal pellets and their total weight was recorded at the end of 24 hr period. Additional fecal pellets collected for 3 full days were combined and shipped to the University of Michigan Animal Phenotyping Core for fecal bomb calorimetry. Briefly, fecal samples were dried overnight at 50 °C prior to weighing and grinding them to powder. Each sample was mixed with wheat flour (90% wheat flour, 10% sample) and formed into 1.0 g pellet, which was then secured into the firing platform and surrounded by 100% oxygen. The bomb was lowered into a water reservoir and ignited to release heat into the surrounding water. These data were used to calculate fecal pellet frequency (bowel movements/day), average fecal pellet weight (g/bowel movement), fecal energy (cal/g feces), and total fecal energy (kcal/day).

High-resolution respirometry was conducted on previously frozen tissue samples (liver and gastrocnemius muscle) to assay for mitochondrial activity as we have previously described (Sarver et al., 2024). Briefly, all tissues were dissected, snap frozen in liquid nitrogen, and stored at –80 °C for later analysis. Samples were thawed in MAS buffer (70 mM sucrose, 220 mM mannitol, 5 mM KH₂PO₄, 5 mM MgCl₂, 1 mM EGTA, 2 mM HEPES pH 7.4), finely minced with scissors, and then homogenized with a glass Dounce homogenizer. The resulting homogenate was spun at 1000 g for 10 min at 4 °C. The supernatant was collected and immediately used for protein quantification by BCA assay (Thermo Fisher Scientific, 23225). Each well of the Seahorse microplate was loaded with 8 µg (liver samples) or 10 µg (gastrocnemius muscle samples) of homogenate protein. Each biological replicate is comprised of three technical replicates. Samples from all tissues were treated separately with NADH (a complex I substrate, 1 mM) or Succinate (a complex II substrate, 5 mM) in the presence of rotenone (a complex I inhibitor, 2 µM), then with the inhibitors rotenone (2 µM) and Antimycin A (4 µM), followed by TMPD (0.45 mM) and Ascorbate (1 mM) to activate complex IV, and finally treated with Azide (40 mM) to assess non-mitochondrial respiration. All respiration data were normalized to mitochondrial content, quantified using MitoTracker Deep Red (MTDR, Thermo Fisher, M22426) as described (Sarver et al., 2024). Briefly, lysates were incubated with MTDR (1 µM) for 10 min at 37 °C, then centrifuged at 2000 × g for 5 min at 4 °C. The supernatant was carefully removed and replaced with 1 x MAS solution and fluorescence was read with excitation and emission wavelengths of 625 nm and 670 nm, respectively. To minimize non-specific background signal contribution, control wells were loaded with MTDR and 1 x MAS and subtracted from all sample values.

Total RNA was isolated from tissues using Trizol reagent (Thermo Fisher Scientific) according to the manufacturer’s instructions. Library preparation and bulk RNA sequencing of liver, skeletal muscle (gastrocnemius), gonadal white adipose tissue (gWAT), and inguinal white adipose tissue (iWAT) of HFD-fed Ctrp10-KO female mice and WT controls were performed by Novogene (Sacramento, California, USA) on an Illumina platform (NovaSeq 6000) and pair-end reads were generated. Sample size: 6 WT and 6 KO for each tissue. All raw sequencing files are available from the NCBI Sequence Read Archive (SRA) accession PRJNA971939.

Transcript features were assembled from raw fastq files and aligned to the current version of mouse transcriptome (Mus musculus.GRCm39.cdna) using kallisto -aln (Bray et al., 2016). Version-specific Ensembl transcript IDs were linked to gene symbols using biomart. Estimated counts were log normalized and filtered for a limit of sum >5 across all samples. Logistic regressions comparing WT vs KO samples across tissues were performed using limma (Ritchie et al., 2015). Differential expression results were visualized using available R packages in CRAN: ggplot2, ggVennDiagram and pheatmap. Gene set enrichment analyses of the DEGs were performed using Enrichr (Kuleshov et al., 2016). Scripts for analyses and visualization are available at the github link below (Data availability section).

All the datasets and scripts to perform analyses are available at the github linked here. Male and Female human data were obtained from Genotype-Tissue Expression (GTEx; GTEx Consortium, 2013) and filtered for sufficient comparison of inter-tissue transcript correlation as described (Velez et al., 2022; Sarver et al., 2023). CTRP10/C1QL2 and other human orthologues for mouse DEGs were identified by intersecting mouse gene symbols with known human orthologues from the vertebrate homology resource at Mouse Genome Informatics (MGI; The Alliance of Genome Resources Consortium, 2020). Co-correlation between all human orthologues DEGs were calculated in either self-reported male or female subjects in GTEx using the bicorAndPvalue() function in Weighted Genetic Coexpression Network Analysis (WGCNA) package (Langfelder and Horvath, 2008). To compare sex-differences of regression coefficients, Wilcoxon t-tests were compared between coefficients using the R package ggpubr. To investigate the degree of conservation of CTRP-engaged pathways, we mapped the DEGs identified from Ctrp10 KO versus WT mice to their human orthologs, including human CTRP10/C1QL2, in the GTEx database for transcriptional correlations. Individuals were stratified by sex to examine sex-specific gene connectivity, consisting of 210 males and 100 females to compare gene expression across tissues. Gene-connectivity analyses were performed based on population correlation significances summarized by cumulative -log10(pvalues) as previously described (Velez et al., 2022; Koplev et al., 2022; Massa et al., 2023; Zhou et al., 2024). Additional details on the filtering criteria and normalization methods, as well as the human DEGs file, GTEx data and subject characteristics, can be found in the GitHub repository.

Total RNA was isolated from tissues using Trizol reagent (Thermo Fisher Scientific). Purified RNA was reverse transcribed using an iScript cDNA Synthesis Kit (Bio-Rad). Real-time quantitative PCR analysis was performed on a CFX Connect Real-Time System (Bio-Rad) using iTaq Universal SYBR Green Supermix (Bio-Rad) according to manufacturer’s instructions. Data were normalized to the stable housekeeping gene β-actin or 36B4 (encoding the acidic ribosomal phosphoprotein P0) and expressed as relative mRNA levels using the ∆∆Ct method (Schmittgen and Livak, 2008). Real-time qPCR primers used to assess Ctrp10 expression across mouse tissues were: Ctrp10 forward, 5’-cggcttcatgacacttcctga-3’ and reverse, 5’-agcagggatgtgtcttttcca-3’. qPCR primers used to confirm the absence of Ctrp10 in KO mice were: forward (qPCR-m10-F2), 5’-CACGTACCACATTCTCATGCG-3’ and reverse (qPCR-m10-R1), 5’-TCGTAATTCTGGTCCGCGTC-3’.

Sample size is indicated in figure and/or figure legend. All results are expressed as mean ± SEM. Statistical analysis was performed with Prism 9 software (GraphPad Software, San Diego, CA). Data were analyzed with two-tailed Student’s t-tests, one-way ANOVA or two-way ANOVA (with Sidak’s post hoc tests). Two-way ANOVA was used for body weight over time, fasting-refeeding response, and all tolerance tests. p<0.05 was considered statistically significant.