Immune checkpoint protein VSIG4 as a biomarker of aging in murine adipose tissue

Abstract Adipose tissue is recognized as a major source of systemic inflammation with age, driving age‐related tissue dysfunction and pathogenesis. Macrophages (Mφ) are central to these changes yet adipose tissue Mφ (ATMs) from aged mice remain poorly characterized. To identify biomarkers underlying changes in aged adipose tissue, we performed an unbiased RNA‐seq analysis of ATMs from young (8‐week‐old) and healthy aged (80‐week‐old) mice. One of the genes identified, V‐set immunoglobulin‐domain‐containing 4 (VSIG4/CRIg), encodes a Mφ‐associated complement receptor and B7 family‐related immune checkpoint protein. Here, we demonstrate that Vsig4 expression is highly upregulated with age in perigonadal white adipose tissue (gWAT) in two mouse strains (inbred C57BL/6J and outbred NIH Swiss) independent of gender. The accumulation of VSIG4 was mainly attributed to a fourfold increase in the proportion of VSIG4+ ATMs (13%–52%). In a longitudinal study, VSIG4 expression in gWAT showed a strong correlation with age within a cohort of male and female mice and correlated strongly with physiological frailty index (PFI, a multi‐parameter assessment of health) in male mice. Our results indicate that VSIG4 is a novel biomarker of aged murine ATMs. VSIG4 expression was also found to be elevated in other aging tissues (e.g., thymus) and was strongly induced in tumor‐adjacent stroma in cases of spontaneous and xenograft lung cancer models. VSIG4 expression was recently associated with cancer and several inflammatory diseases with diagnostic and prognostic potential in both mice and humans. Further investigation is required to determine whether VSIG4‐positive Mφ contribute to immunosenescence and/or systemic age‐related deficits.


| INTRODUC TI ON
Mφ are sentinels of the immune system, detecting and responding to various stressors (e.g., infection, tissue damage, cell transformation) in order to maintain tissue homeostasis (Wynn, Chawla, & Pollard, 2013). Mφ play pivotal roles in clearing pathogens and cellular debris, modulating both innate and adaptive branches of immunity, and orchestrating tissue remodeling, inflammatory response, and subsequent resolution. Aging is associated with a general decline of proper immune function. Age-related changes in Mφ are associated with delayed wound healing, cancer progression, and the development of age-associated inflammatory diseases (e.g., atherosclerosis, rheumatoid arthritis, and peripheral neuropathy; Becker et al., 2017;Mahbub, Deburghgraeve, & Kovacs, 2012;Oishi & Manabe, 2016;Rowe et al., 2016;Thevaranjan et al., 2017;Yuan et al., 2018).
Since aged ATMs were recently shown to possess fundamental markers of aging associated with frailty (e.g., Il6, p16 Ink4a , senescence-associated β-galactosidase activity) (Hall et al., 2016Lumeng et al., 2011), we performed whole transcriptome analysis as an unbiased approach to identify age-related differential gene expression in perigonadal ATMs. Genes exhibiting increased Mφassociated expression with age were chosen for confirmation as candidate biomarkers of aged adipose tissue and were subsequently validated for their differential gene expression in mice of two genetic backgrounds (C57BL/6J and NIH Swiss). One of the genes identified, VSIG4 (CRIg/Z39Ig), was of interest as a Mφassociated protein with roles spanning both innate and adaptive immunity. As a tissue-resident Mφ marker, VSIG4 is a multifunctional cell surface protein with potent immunosuppressive and anti-inflammatory roles [as recently reviewed (Small, Al-Baghdadi, Quach, Hii, & Ferrante, 2016)]. Notably, induction of VSIG4 expression is associated with several inflammatory diseases and cancers in both mice and humans. VSIG4 functions as an immune checkpoint regulator, suppressing T-lymphocyte function and promoting cancer development and progression (Bianchi-Frias et al., 2019;Liao et al., 2014). In this study, we demonstrate that increased VSIG4 expression by ATMs correlates with age and physiological frailty in mice, suggesting that accumulation of VSIG4 may reflect relevant phenotypic changes that underlie age-related pathologies. Thus, we present VSIG4 as a novel biomarker of aged murine ATMs. Further studies are required to elucidate whether elevated levels of VSIG4 in adipose and other tissues promote age-related deficits, such as immunosenescence, dysregulation of tissue homeostasis, and/or tumor establishment.

| Identification of Vsig4 as a biomarker of aged ATMs
To identify Mφ-specific biomarkers that accumulate with age in  (Table S1). In contrast, analysis of several internal reference genes (B2m,Sap130,Sdha,Tbp,Tubb5) was similar across analyzed samples (Table S1).
Mφ-associated genes exhibiting age-related upregulation were then selected based on the following criteria: (a) expression level (≥100 normalized counts in the Mφ-enriched SVF from aged mice), (b) age-associated upregulation (≥3-fold difference in Mφ-enriched SVFs between young and aged mice), and (c) Mφ-associated gene expression (≥2-fold enrichment in the Mφ-enriched versus Mφdepleted SVFs from aged mice) (Figure 1a). A total of 23 candidate genes were identified (Figure 1b), 20 of which were selected for confirmation studies. To evaluate the list of candidate genes, we confirmed age-dependent differential gene expression using a custom-built NanoString Technologies nCounter ® codeset platform. Analysis of the Mφ-enriched SVF samples used for RNA-seq confirmed that 17 genes (85%) exhibited a ≥3-fold age-dependent increase in expression (Table S2). Differential gene expression for the remaining 3 genes (Csmd1, Tacr3, and Tmem132e) could not be accurately assessed due to low-level expression. The majority of confirmed genes encoded secreted cytokines and chemokines previously associated with inflammation and/or aging (Ccl2,Ccl8,Cxcl1,Cxcl2,Cxcl5,Cxcl13,Il6,Inhba,Lif,Saa3) (Song et al., 2012;Yousefzadeh, Schafer, et al., 2018;Hudgins et al., 2018;Greene & Loeser, 2015;Loria et al., 1998;Xu, Palmer, et al., 2015;Knight, 2001;Den Hartigh et al., 2014artigh et al., 2014, although not all of these genes have been linked to age-related processes in adipose tissue and/or Mφ. In addition, we confirmed an age-related increase of Mφ-associated genes encoding surface receptors with roles in phagocytosis, including collagen-like scavenger receptor Marco, several C-type lectin receptors (Cd209d, Cd209f, and Cd209g), and complement receptor and immune checkpoint protein, Vsig4.
Analysis of several independent pools of Mφ-enriched SVF revealed the strongest and most consistently upregulated genes in aged male mice to be tissue-resident Mφ marker Vsig4, cytokine Saa3, and chemokines Ccl8 (Mcp2), Cxcl5, and Cxcl13 ( Figure S1A).
Of these genes, Vsig4 was of significant interest as a gene not previously associated with aging, encoding a multifaceted cell surface receptor with roles in both innate and adaptive immunity (Small et al., 2016). Vsig4 was undetectable in ATM-enriched SVF from young mice but showed a substantial signal in aged mice (8.0-fold increase above background signal; p = .008; Figure 2a). F I G U R E 1 Schematic of data acquisition and identification of genes upregulated in Mφ from old mice. (a) The stromal vascular fraction (SVF) was isolated via enzymatic digestion of perigonadal white adipose tissue (gWAT) pooled from young (Yng; n = 12) or chronologically aged (Old; n = 6) mice. Mφ were then segregated into Mφ-enriched and resultant Mφ-depleted fractions (see Section 4). Total RNA was subsequently extracted and analyzed via whole transcriptome sequencing. Following data processing and quantitation of expression, genes were filtered using several criteria, including (i) expression level, (ii) upregulation in aged Mφ, (iii) and Mφ-associated gene expression (i.e., genes enriched in Mφ-enriched fraction). Confirmation of age-dependent differential gene expression in adipose tissue Mφ was performed via NanoString analysis, using the same RNA extracts submitted for RNA-seq analysis (Table S2). (b) Differential gene expression of analyzed RNA-seq data, with genes ranked by the extent of upregulation in the ATM-enriched SVF between age groups (blue bars). For each gene, the extent of enrichment between the ATM-enriched and ATM-depleted SVF from aged mice is depicted (red bars)

| Age-related upregulation of Vsig4 in adipose tissue is independent of gender and genetic background
An ideal biomarker is one that is reliably associated with a specific condition and can be assessed following minimal tissue processing. Therefore, we assessed the expression of Vsig4 and other candidate genes in total SVF (without Mφ enrichment) and whole adipose tissue to determine whether these samples were suitable for the detection of differential expression between young and old mice via the NanoString platform. Further, we examined whether age-dependent increases were influenced by gender and/or genetic background. Perigonadal fat from young male C57BL/6J mice exhibited elevated levels of Vsig4 expression compared to young female mice (2.0-fold difference) (Figure 2b,c). Aged male mice showed a greater increase in Vsig4 expression compared to female littermates (6.7-fold and 3.2-fold increase, respectively). Analysis of intact gWAT from both inbred C57BL/6J and outbred NIH Swiss mice also demonstrated an age-dependent increase in Vsig4 expression (3.1-fold and 4.3-fold, respectively; Figure 2d,e). Notably, as the only gene to show a ≥3-fold increase across all samples, Vsig4 was validated as our strongest biomarker candidate ( Figure S1B). Gene expression of other Mφ-specific markers (Adgre1, Csf1r, Mrc1) was analyzed in parallel, revealing minor changes relative to Vsig4; in the SVF and intact gWAT from C57BL/6J mice, a ≤1.5-fold age-related increase was observed in classical Mφ markers, whereas gWAT from NIH Swiss mice exhibited up to a twofold decrease ( Figure S2A-D).
These data indicate that, unlike other Mφ-specific markers, Vsig4 is strongly upregulated with age. In summary, our efforts to validate age-related gene expression in gWAT revealed a strong, consistent increase of Vsig4 expression in two strains of chronologically aged mice, C57BL/6J and NIH Swiss, independent of gender.

| Age-dependent increase in the proportion of VSIG4-positive ATMs in gWAT
We next evaluated age-related changes in VSIG4 protein expression in gWAT via immunoblot analysis. We found that gWAT from males expressed more VSIG4 compared to females and that both genders  Figure S4). Concomitantly, we observed a dramatic reduction in the age-associated anti-VSIG4 immunoreactive band via immunoblot ( Figure 2g). Other bands of relatively higher and lower molecular weight were unchanged, indicating their non-specific immunoreactivity. These data indicate that VSIG4 expression of ATMs was elevated in aged mice consistent with Vsig4 transcript expression, with male mice exhibiting substantially higher levels in gWAT than female mice from the same cohort.
To determine whether the increased expression of VSIG4 in adipose tissue with age was site-specific, we characterized several adipose tissue depots via immunoblot. VSIG4 was expressed in both gWAT and iWAT from young mice and was substantially increased with age in both male and female C57BL/6J and male NIH Swiss mice ( Figure 2h; Figure S5A). However, VSIG4 expression in perirenal white adipose tissue (rWAT) was highly variable with no apparent age-related trend, and in brown adipose tissue (BAT) isolated from the interscapular depot, little to no VSIG4 expression was observed ( Figure S5A). Expression analysis was also expanded to a panel of several tissues, revealing that, while most tissues either lack substantial levels of VSIG4 expression or do not show an appreciable increase with age, the heart, thymus, and peritoneal cavity exhibit a strong accumulation of VSIG4 in aged mice ( Figure S5B-H). In the thymus, immunofluorescence analysis revealed an extensive increase of VSIG4-positive Mφ in the remnants of the cortex following thymic involution ( Figure S5H). Overall, our analysis reveals a substantial tissue-specific accumulation of VSIG4 expression with age.
We next characterized age-related expression of VSIG4 in gWAT via immunofluorescent staining. In both young and aged mice, VSIG4 staining co-localized with staining for tissue-resident Mφ marker CD206 (Figure 2i) mice was analyzed via flow cytometry (see Figure S6A,B for gating scheme). Notably, the anti-VSIG4 antibody used for flow cytometry showed strict co-localization with the different VSIG4 antibody used for immunofluorescent staining ( Figure S7). The percentage F I G U R E 2 Age-related increase of Vsig4 gene expression in murine adipose tissue. (a-e) NanoString gene expression analysis of Vsig4 in gWAT samples from young and old mice. (a) Vsig4 mRNA expression analysis of Mφ-enriched SVF from several independent pools of 3-12 mice each, either young (8-week-old; n = 2 pools) or aged (78-to 88-week-old; n = 3 pools), adjusted to gene expression of three pan-Mφ genes Adgre1, Mrc1, and Csf1r. Data presented as mean ± SD; nd = not detectable. **p < .01, compared to young mice. Vsig4 mRNA expression analysis of total SVF of gWAT from different-aged cohorts of male (b) or female (c) mice, and of intact gWAT from male C57BL/6J (d) and NIH Swiss (e) mice, adjusted to gene expression of five internal reference genes B2m, Sap130, Sdha, Tbp, and Tubb5. Data presented as a single measurement of pooled samples (n = 3-8 mice/group) ± error propagation from SD of normalized internal reference genes. (f-h) Immunoblot analysis of age-dependent VSIG4 expression from intact adipose tissues of (f) gWAT from three individual male and female C57BL/6J mice; (g) gWAT from old male C57BL/6J mice (102-week-old) 1 week after a single administration of PBS, Encapsome (Encap; liposomal PBS), or Mφ depletion reagent Clodrosome (Clod; liposomal clodronate); (h) pooled samples (n = 3 mice) of gWAT and inguinal white adipose tissue (iWAT) collected from young (Y; 8-to 18-week-old) and old (O; 90-to 96-week-old) C57BL/6J mice. GAPDH or β-actin were used as immunoblot loading controls, as indicated. *Non-specific α-VSIG4 immunoreactive bands, as indicated in panel (g). (i) Representative fluorescence microphotographs of intact gWAT from young (14-week-old) and old (84-week-old) mice depicting immunofluorescent staining of Mφ-related markers VSIG4 (red) and CD206 (green), and a merged overlay with DAPI nuclear counterstain (blue). Scale bar = 50-μm. (j-n) Representative experiment of flow cytometric analysis of single-cell suspensions of the SVF isolated from three individual young (19-week-old) and old (120-week-old) male C57BL/6J mice. (j) Pseudocolor plots from flow cytometric analysis of VSIG4 expression in live-stained, viable ATMs (CD45 + CD11b + F4/80 + ). All gates were defined using age-matched fluorescence minus one (FMO) controls (see gating scheme in Figure S6A,B). Quantitation of (k) the proportion of VSIG4 + Mφ and (l) the median fluorescent intensity (MFI) of αVSIG4-APC stained Mφ minus MFI from negative control staining, that is, FMO(VSIG4-APC). (m) SVF total cell yield per mouse.
Despite higher yields of SVF cells isolated from aged gWAT, the total number of ATMs was found to be similar between young and aged mice (Figure 2m,n). Together, these data indicate that the increased abundance of VSIG4 in aged gWAT is attributed to both an increased proportion of tissue-resident Mφ expressing VSIG4 and an age-dependent increase in VSIG4 expression within the VSIG4 + Mφ population.

| VSIG4 protein levels elevated with age but not diet-induced obesity
VSIG4 expression is associated with several inflammatory diseases that are linked to aging. Both aging and diet-induced obesity

F I G U R E 4
Correlation of VSIG4 expression in gWAT with chronological age and health throughout life. (a-j) A longitudinal study of a single cohort of chronologically aged male and female NIH Swiss littermates, where physiological frailty index (PFI; a multi-parameter assessment of health) was analyzed throughout the life span of the mice. At indicated ages, a final PFI was performed and gWAT was collected. Immunoblot analysis of VSIG4 and Mφ marker F4/80 in gWAT from individual male (a) and female (b) mice (n = 3). β-actin was used as a loading control. *Non-specific α-VSIG4 immunoreactive bands. Densitometric analysis of immunoblot images for male (c) and female (d) mice, evaluating the expression level of VSIG4 and F4/80 with respect to β-actin for each mouse. Data represent mean ± SD of values normalized to the youngest group for each gender. ns, not significant (p > .05), *p < .05, **p < .01, compared to young mice (26-or 30-week-old mice for males and females, respectively). (e-j) Correlation of VSIG4 immunoblot densitometric analysis (normalized to β-actin), chronological age and PFI in male (e-g) and female ( (DIO) provoke inflammation in adipose tissue, albeit with differing etiology (Garg et al., 2014;Lumeng et al., 2011). We examined whether VSIG4 expression was upregulated in response to changes in adipose tissue elicited by obesity. Expression of VSIG4 was analyzed in gWAT and iWAT from obese, Western-type highfat diet-fed mice exhibiting increased intraabdominal fat and signs of impaired glucose tolerance ( Figure S8A-E). Immunoblot analysis revealed no differences in VSIG4 expression in gWAT or iWAT from DIO mice compared to normal diet-fed, age-matched controls (Figure 2o-q). Analysis of gWAT tissue via immunofluorescent staining indicates that, unlike aging, DIO is associated with an extensive infiltration of Mφ into adipose tissues, consistent with literature data Li et al., 2017; Figure   S8F). However, the proportion of VSIG4 + Mφ was decreased in DIO mice, suggesting that infiltrating Mφ are largely negative for VSIG4 expression. Robust Vsig4 upregulation was associated with aging but not DIO, reflecting intrinsic differences between these two conditions.

| VSIG4 expression in gWAT correlates with healthspan
VSIG4 expression was found to be highly upregulated in aged mice; however, it is unclear why VSIG4 expression changes and whether elevated VSIG4 expression reflects deleterious changes that occur with age. To evaluate the relationship between VSIG4 expression and health, we conducted a cross-sectional study of a single cohort of male and female NIH Swiss mice where physiological frailty index (PFI), a diverse multi-parameter assessment of animal health (Antoch et al., 2017), was measured at multiple ages. In parallel, gWAT was collected for assessing VSIG4 expression via immunoblot analysis.
These samples revealed an age-dependent increase in VSIG4 expression, with densitometric analysis revealing at least a ninefold increase in both male and female mice up to ages 132 and 108 weeks, respectively (Figure 4a-d). Notably, gender-specific differences were observed in the kinetics of acquired VSIG4 expression; while male mice exhibited a consistent increase in VSIG4 expression starting at 78 weeks old, increased expression of VSIG4 in females was observed as early as 44 weeks old and plateaued at 82 weeks old.
By comparison, little to no change was observed in the expression of F4/80 between aged cohorts. Significant heterogeneity exists in the expression of VSIG4 in gWAT among middle-aged male mice.
We sought to determine whether this variation was reflective of differences in health between mice, as assessed by PFI. In male and female mice, PFI showed a strong correlation with age in the subset of mice analyzed for VSIG4 expression (r = .82 and .70, respectively; p ≤ .01; Figure 4e,h). In male mice, VSIG4 expression in gWAT was strongly correlated with both age and PFI (r = .84 and .82, respectively; p ≤ .002; Figure 4f,g). In female mice, VSIG4 expression was strongly correlated with age (r = .80; p = .002) but lacked a statistically significant correlation with PFI (r = .42; p = .17; Figure 4i,j).
Aging is associated with the accumulation of a heterogeneous population of cells expressing elevated levels of CDK inhibitor protein p16 Ink4a , the clearance of which, via genetic means or chemical intervention (e.g., senolytics), has demonstrated an improvement of health and/or extension of median life span across multiple studies (Baker et al., 2011;Xu, Palmer, et al., 2015;Xu et al., 2018;Yousefzadeh, Zhu, et al., 2018). To determine whether VSIG4 expression was associated with the presence of p16 Ink4a -positive cells in adipose tissue, gWAT was analyzed from young (18-week-old) and aged (96-week-old) female INK-ATTAC mice, a strain that enables the detection of cells with elevated p16 Ink4a promoter activity (Yousefzadeh, Zhu, et al., 2018). Mice of each age group were acutely treated with vehicle or the senolytic flavonoid fisetin (100 mg/kg for two consecutive days by oral gavage) recently shown to deplete p16 Ink4a -positive cells in vivo and extend murine healthspan (Yousefzadeh, Zhu, et al., 2018;Zhu et al., 2017). Analysis of the proportion of p16 Ink4a cells (via CyTOF) and VSIG4 expression re-

| DISCUSS ION
The pathogenesis of age-related diseases in mice and humans is thought to be driven in large part by deleterious changes in adipose tissue over time that give rise to systemic tissue dysfunction through its continuous secretion of pro-inflammatory cytokines. As sentinels of the immune system and mediators of inflammatory response, Mφ play a central role in maintaining tissue homeostasis. Phenotypic changes in ATMs likely reflect processes underlying and/or contributing to increases in physiological deficits with age, yet these cells remain largely uncharacterized. Here, we performed an unbiased transcriptome analysis of an ATM-enriched population of SVF cells to elucidate age-related differential gene expression.
Analysis of age-associated gene upregulation of ATM-enriched SVF revealed inflammatory cytokines and chemokines previously associated with aged ATMs as drivers of "inflammaging": Il6, Ccl2, Cxcl1, and Cxcl2 (Garg et al., 2014;Lumeng et al., 2011). However, consistent with other studies, the analysis of these inflammatory markers in total SVF did not reveal increased gene expression in aged mice (Liu, Shen, Ueno, Patel, & Kraemer, 2011;Starr et al., 2013;Wu et al., 2007). Most genes identified in our study are not exclusively expressed by Mφ, and age-related differences in ATMs can be diluted within SVF or in whole adipose tissue by other expressing cell types. Moreover, expansion of the B-and T-lymphocyte population in adipose tissues with age further reduces the proportion of Mφ, whose numbers remain largely unchanged (Bapat, Myoung Suh, et al., 2015). A notable caveat of this study is that the ATM enrichment procedure developed relies on their phagocytic ability that may enrich for a subpopulation of highly phagocytic ATMs and/or elicit phagocytosis-related changes in gene expression (A-Gonzalez et al., 2017). While differential age-dependent responses of ATMs to various stimuli are likely relevant to changes in adipose with age (Starr et al., 2009), in this study we focused on identifying stable Mφspecific aging biomarkers that could be assessed with minimal tissue processing. Importantly, we confirmed that our strongest marker,  Vogt et al., 2006). In humans, VSIG4 is highly expressed in Mφ in the liver, lungs, adrenal glands, and placenta (Helmy et al., 2006).
Elevated expression of VSIG4 occurs in the context of autoimmune disease and cancer in both mice and humans [e.g., foam cell Mφ (Lee et al., 2006), synovial Mφ in arthritic joints (Lee et al., 2006), Mφ in tumor microenvironment (Liao et al., 2014;Sturtz et al., 2014;Xu, Jiang, et al., 2015)], and has been proposed as a diagnostic and prognostic indicator of several diseases, including cancer, preeclampsia, and rheumatoid arthritis (Roh et al., 2017;Textoris et al., 2013;Xu, Jiang, et al., 2015;Zheng et al., 2016). A survey of several adipose tissue depots revealed that VSIG4 was consistently upregulated in gWAT and iWAT, but not perirenal WAT. VSIG4 was absent from brown adipose tissue and did not appear with age. Adipocytes from adipose tissue depots are highly heterogeneous, expressing unique developmental gene signatures and exhibiting many functional and metabolic differences (Schoettl, Fischer, & Ussar, 2018 (Li et al., 2017). However, this study did not ascertain whether VSIG4 was upregulated in wild-type mice in response to obesity-induced dysfunction. Our data indicate that VSIG4 expression was unchanged in both gWAT and iWAT following several months of feeding on a Western-like high-calorie diet. These data are consistent with findings that Mφ accumulating in the fat of obese mice do not acquire Vsig4 expression (Lumeng et al., 2007). Thus, while the presence of VSIG4 appears to play a role in promoting adipose tissue homeostasis during the onset of diet-induced obesity, our data suggest that age, but not diet-induced obesity, induces a significant accumulation of VSIG4-positive ATMs in gWAT.
Analysis of VSIG4 expression across tissues revealed that age-related upregulation was not restricted to adipose tissues, as a strong increase was also observed in the heart and thymus. In aged mice, VSIG4-positive Mφ were found in remnants of the thymic cortex following involution, a process of age-related thymic atrophy, conversion to adipose tissue, and immunosenescence (Thomas, Wang, & Su, 2020). The potential role of VSIG4 in the suppression of conventional T-lymphocyte proliferation and accumulation of Tregs associated with thymic involution warrants investigation.

VSIG4-mediated suppression of cytotoxic T lymphocytes and inflammatory responses (via administration of a VSIG4-Ig fusion protein) exerts a therapeutic benefit in experimental models
of autoimmune disease (Chen, Muckersie, Luo, Forrester, & Xu, 2010;Jung et al., 2015;Katschke et al., 2007). However, VSIG4 expression within the tumor microenvironment is associated with increased tumor growth and poor prognostic outcomes for several cancers (Bianchi-Frias et al., 2018;Byun et al., 2017;Roh et al., 2017;Xu, Jiang, et al., 2015). VSIG4 may play a direct role in cancer progression, as Vsig4 deficiency was shown to suppress the growth of Lewis lung carcinoma (LLC1) isografts (Liao et al., 2014). Here, we observed that VSIG4 expression was upregulated in lungs of aged mice bearing spontaneous tumors and in an A549 lung carcinoma xenograft model. In both cases, expression of VSIG4 was strongly induced in the adjacent normal tissue.
Similarly, VSIG4 was previously identified as a marker highly upregulated by invasive breast cancer near tumor-adjacent adipose tissue compared to a distant site from the same breast (Sturtz et al., 2014). VSIG4 is also locally upregulated within the microenvironments of inflammatory diseases, such as arthritis (Lee et al., 2006;Zheng et al., 2014). The regulation of VSIG4 expression is poorly understood. Elucidation of the factors responsible for the local induction of VSIG4 within a subset of Mφ associated with inflammatory lesions and cancer may provide insight into the mechanisms that drive VSIG4 upregulation within specific tissues with age.
The accumulation of senescent cells and their bioactive secretions (i.e., SASP) is implicated as local and systemic pathogenic agents, with adipose tissue a major site of senescent cell burden with age (Stout et al., 2017). In addition, the proportion of Mφ expressing p16 Ink4a and β-galactosidase (two classical markers of senescence that may indicate Mφ activation rather than bona fide senescence) increases in adipose tissue with age and likely contributes to inflammaging and factors driving age-related changes in adipose tissue (Hall et al., 2016Yousefzadeh, Zhu, et al., 2018). Here, we demonstrate that the proportion of cells with activated p16 Ink4a promoter expression in INK-ATTAC mice was not associated with the level of VSIG4 detected in whole gWAT from female mice. This may reflect gender-specific differences in aging; for NIH Swiss mice, males showed a more robust induction of VSIG4 than female mice that was strongly correlated with physiological frailty. However, We instead chose to utilize the senotherapeutic agent fisetin in our study due to its senolytic activity against a wide range of cell types (including mesenchymal cells, endothelial cells, T lymphocytes, and NK cells) but reported inability to remove p16 Ink4a -positive Mφ (Yousefzadeh, Zhu, et al., 2018). In this way, regulation of VSIG4 expression in Mφ could be assessed without the confounding influence of Mφ depletion. In addition, fisetin directly exerts immunosuppressive effects against Mφ and other cell types that may also contribute to its therapeutic benefit (Kim, Kim, Kim, & Cho, 2015;Liu et al., 2010;Paul, Majhi, Mitra, & Ray Banerjee, 2019;Zheng, Ock, Kwon, & Suk, 2008). Acute treatment of mice with fisetin was unable to influence the short-term expression of VSIG4, indicating that neither the senolytic nor immunomodulatory effects of fisetin were able to abruptly or substantially modulate VSIG4.
Notwithstanding, VSIG4 levels may require additional time to respond to senescent cell clearance or could be impacted by other senolytics that target different senescent cell types than fisetin.
Whether VSIG4 levels are a suitable pharmacodynamic marker of improved tissue homeostasis following preventive and/or interventional therapies against senescent cells, other mechanisms of aging, or inflammatory diseases is an exciting prospect that warrants investigation.
Investigation of age-related changes in ATMs reveals an increase in several M2-related markers associated with an anti-inflammatory phenotype, and yet, exhibits increases in other markers associated with M1 polarization and exaggerated pro-inflammatory responses (van Beek et al., 2019;Zeyda et al., 2007). Despite several mechanisms through which VSIG4 exerts anti-inflammatory effects, including both intrinsic and extrinsic signaling, the extent to which VSIG4 expression exerts anti-inflammatory effects in aged mice is unknown. For instance, recently a transmembrane protein, TRIM72, was shown to directly bind to and antagonize VSIG4 signaling, resulting in an elevated inflammatory response following bacterial clearance (Nagre et al., 2018). In addition, depending on the underlying mechanisms, VSIG4 expression can promote pathogenic inflammation and disease progression. For example, VSIG4 was found to exacerbate an experimental model of colitis in which NLRP3 inflammasome activation is associated with limiting disease progression (Huang et al., 2019).
In this study, we have identified VSIG4, a multifaceted checkpoint protein of recent interest in fields of inflammatory disease and cancer, as a novel biomarker of aging Mφ in tissues associated with age-related systemic inflammation and immunosenescence. In male mice, VSIG4 accumulation was strongly correlated with increased F I G U R E 5 Schematic of a proposed model of VSIG4-driven increase in physiological frailty with age. (a) Our research indicates that VSIG4 expression accumulates throughout various tissues with age, notably white adipose tissue depots. In addition, VSIG4 expression is strongly associated with site of cancer growth and several inflammatory diseases. (b) VSIG4 expression promotes immune tolerance through its immunosuppressive activities, including (i) intrinsic suppression of Mφ responses following VSIG4 activation, (ii) efficient clearance of complement-opsonized pathogens and debris, (iii) promotion of Treg differentiation and reinforcement of Foxp3 + Treg phenotype, (iv) suppression of conventional CD4+ and cytotoxic CD8+ T-lymphocyte activation, cytokine response, and proliferation, and (v) suppression of T cell-mediated isotype switching in B lymphocytes. We demonstrated that VSIG4 expression in gWAT strongly correlates with physiological frailty in male mice. This schematic depicts scenarios in which VSIG4 activities may promote age-related tissue dysfunction. VSIG4-mediated suppression of conventional CD4+ and cytotoxic CD8+ T-lymphocyte response and expansion and a shift toward CD4 + Foxp4 + Treg phenotypes may help to drive immunosenescence of T lymphocytes. Further, the suppression of both innate and adaptive immunity and promotion of immune tolerance by VSIG4 encourage a pro-tumorigenic microenvironment and may prevent efficient clearance of damaged and/or senescent cells, resulting in unresolved inflammation and tissue dysfunction physiological frailty, suggesting that this novel biomarker reflects a fundamental process underlying natural organismal aging. Based on the current understanding of VSIG4 function, we present here a hypothetical model detailing how VSIG4 upregulation may drive age-related changes ( Figure 5). However, future studies are required to elucidate the mechanism of VSIG4 upregulation and to evaluate whether VSIG4-expressing Mφ in adipose and/or other tissues (e.g., thymus) directly contribute to life span or age-related deficits.

| Animals
Male and female C57BL/6J mice were obtained from Jackson Laboratory (Bar Harbor, ME) at 6-8 weeks of age, and Cr:NIH(S) mice (NIH Swiss) were obtained from Charles River (Wilmington, MA) at 6 weeks of age and were allowed to age at the Roswell Park CyTOF analysis, as previously described (Yousefzadeh, Zhu, et al., 2018). A portion of each whole adipose tissue sample was freshly frozen and stored for subsequent immunoblot analysis at RPCCC.

| Physiological Frailty Index
The following parameters were measured for the assessment of animal health at various ages, and Physiological Frailty Index (PFI) was calculated, as previously described (Antoch et al., 2017): body mass index, grip strength, non-invasive hemodynamic parameters (systolic, diastolic and mean blood pressure, heart rate, tail blood flow, and tail blood volume), whole blood cell counts, plasma concentration of cytokines (CXCL1, CRP) and triglycerides, and blood concen-

| Isolation of gWAT stromal vascular fraction
Perigonadal white adipose tissue was collected from mice into pre-weighed conical vials containing PBS on ice. Isolation of SVF cells from gWAT for flow cytometric analysis was performed using Collagenase II (Sigma) digestion of minced adipose tissue, as described (Cho, Morris, & Lumeng, 2014

| Immunoblotting
Whole adipose tissue protein extracts were obtained via homogenization in RIPA lysis buffer (Sigma) supplemented with Halt phosphatase (Pierce) and protease inhibitors (Sigma). Protein concentration was assessed via Pierce BCA assay kit. Immunoblots were performed as previously described (Frescas et al., 2017). Briefly, pro- Tris-buffered saline containing 0.1% Tween-20 prior to incubation with SuperSignal West Dura chemiluminescent peroxidase substrate (Thermo Scientific) and were exposed using FluorChem E System (Protein Simple). Band densitometry analysis of immunoblots was performed with ImageJ software (NIH).

| Immunofluorescent staining
Immunofluorescence staining of whole perigonadal adipose tissue was performed as previously described . Briefly, adipose tissues were fixed 4 hr in 4% paraformaldehyde in PBS at 4°C, washed in PBS overnight, incubated with blocking solution (PBS with 5% normal rat serum and 0.25% Triton X-100) 1 hr at room temperature, and stained with a cocktail of rat monoclonal antibod-

| Statistical analyses
Unless otherwise noted, data were analyzed by unpaired Student's two-tailed t test and are expressed as mean ± standard deviation. All statistical analyses were performed using GraphPad Prism 5.

CO N FLI C T O F I NTE R E S T
The authors have declared no conflict of interests.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data supporting the findings of this study are openly available in the NCBI GEO database (reference number GSE154832) and in Dryad at https://doi.org/10.5061/dryad.hdr7s qvfp (Hall et al., 2020) O RCI D Andrei V. Gudkov https://orcid.org/0000-0003-2548-0154