Obesity-Induced Increase in Cystatin C Alleviates Tissue Inflammation

Obesity-induced inflammation is a crucial driver of insulin resistance and type 2 diabetes. In obesity, white adipose tissue (WAT) expansion is associated with local infiltration of proinflammatory cells, with macrophages being the quantitatively predominant cell type. The latter accounts for a large proportion of adipose tissue–derived proinflammatory proteins such as interleukin (IL)-1β, IL-6, tumor necrosis factor-α (TNF-α), or hypoxia-inducible factor-1α (HIF1α) (1), altering the secretory pattern of adipocytes. Consequently, the release of proinflammatory cytokines and free fatty acids (FFA) is increased, contributing to the vicious circle of obesity-associated chronic low-grade inflammation locally as well as systemically (e.g., in the liver and in skeletal muscle) (2,3). As a result of the central role of inflammation in the pathogenesis of type 2 diabetes, identification of anti-inflammatory targets to combat insulin resistance and its related comorbidities is warranted (4).

We previously found that removal of one kidney (uninephrectomy [UniNx]) in obese mice ameliorates obesity-induced adipose tissue and liver inflammation (5,6). Cystatin C (CysC) is a marker of the glomerular filtration rate (7) and is increased in patients with renal failure (8). Because CysC was previously reported to have immunomodulatory function, we hypothesized that increased circulating CysC levels contributed to the blunted inflammation in UniNx mice. CysC was found to be internalized into lipopolysaccharide (LPS)-stimulated monocytes and to downregulate the production of proinflammatory cytokines such as TNF-α or IL-1β (9). Moreover, an association between elevated circulating CysC levels and type 2 diabetes has been reported (10,11). However, the pathophysiological implication of such an association remains to be elucidated. Herein, we propose that elevated CysC levels in obesity counterbalance obesity-associated inflammation, thereby restoring glucose tolerance.

Serum concentrations of CysC, IL-10, and lipocalin-2, as well as mRNA expression in subcutaneous or visceral WAT, were determined in 63 individuals (age 19–82 years; BMI 25–75 kg/m²) who underwent laparoscopic sleeve gastrectomy, cholecystectomy, hernia repair, or Roux-en-Y gastric bypass surgery. In addition, CysC mRNA expression in subcutaneous WAT was analyzed in a previously described subgroup of individuals with insulin-sensitive (n = 30) or insulin-resistant obesity (n = 30) (12). All investigations were approved by the University of Leipzig Ethics Committee (Leipzig, Germany) (#159-12-21052012) and were performed in accordance with the Declaration of Helsinki. All participants provided witnessed written informed consent before study entrance. mRNA expression was measured by quantitative RT-PCR using the Assay-on-Demand gene expression kit and calculated relative to the expression of HPRT1 mRNA. For primer details see Supplementary Table 1.

CysC-deficient mice (13) were bred to C57BL/6J wild-type (WT) mice to eventually obtain homozygous CysC knockout (KO) and WT littermates.

Mice were housed in a pathogen-free environment with an alternating 12 h dark and light cycle. Male WT and CysC KO mice (6 weeks old) were fed ad libitum a regular chow (ProvimiKliba) or a HFD (59 kcal% fat, D12331 mod. Surwit; ssniff Spezialdiäten GmbH) for 20 weeks. All protocols conformed to the Swiss animal protection laws and were authorized by the Cantonal Veterinary Office in Zurich, Switzerland.

Glucose and insulin tolerance were assessed as previously described (14).

Hyperinsulinemic-euglycemic clamp studies were performed as previously described (15).

Energy expenditure, locomotor activity, and food intake were assessed in 24-h single-housed mice using the PhenoMaster system (TSE Systems).

FFA concentrations were determined using an enzymatic colorimetric assay (Wako Chemicals GmbH). Plasma CysC (BioVendor) and insulin (Crystal Chem) levels were measured using ELISA kits.

Distribution of adipocyte diameter was determined by the Multisizer 3 Coulter Counter (Beckman Coulter Life Sciences).

cDNA was amplified using TaqMan assays (Applied Biosystems) and normalized to 18s RNA using the 2^−∆∆CT method. For primer details see Supplementary Table 2.

Western blots were obtained as previously described (14). For antibody details see Supplementary Table 3.

The stromal-vascular fraction of epididymal WAT (eWAT) was collected as described (16). For staining details see Supplementary Table 4. Measurements were performed with a flow cytometer (Beckman Coulter), and data were analyzed with FlowJo software. Relative frequency of single, live lymphocytes of the respective subpopulation was assessed.

An osmotic minipump (Model 1004; ALZET) was implanted into the intraperitoneal cavity of anesthetized 12-week-old mice. Minipumps were filled with LPS from Escherichia coli (055:B5; Sigma-Aldrich) to infuse 300 µg/kg · day for 4 weeks.

Data were analyzed by Spearman correlation, unpaired two-tailed Student t test, or by one- or two-way ANOVA, followed by the Bonferroni post hoc test.

All data generated or analyzed during this study are included here and in the Supplementary Material or are available from the corresponding author upon reasonable request.

First, we explored the effect of HFD feeding on circulating CysC levels in WT mice. As depicted in Fig. 1A, 20 weeks of the HFD increased plasma CysC concentrations in systemic and portal (Supplementary Fig. 1A) blood compared with age-matched chow-fed mice independently of kidney function since plasma creatinine levels were similar among chow- and HFD-fed WT mice (Supplementary Fig. 1B). Moreover, CysC plasma levels were significantly increased in HFD-fed UniNx compared with sham-operated mice (Fig. 1B). Of note, UniNx had a protective effect on HFD-induced adipose tissue inflammation and hepatic insulin resistance (5). To explore a potential role of increased CysC on adipose tissue function, we analyzed CysC KO mice (13). As expected, CysC was absent in the circulation of CysC KO mice (Fig. 1C). CysC KO mice exhibited lower body weight under the chow diet, whereas it was no longer different in HFD-fed mice (Fig. 1D). Furthermore, the mass of inguinal and mesenteric fat depots was significantly increased in HFD-fed KO mice (Fig. 1E). Such increased fat accumulation might be due to a pronounced positive energy balance since we observed diminished locomotor activity and decreased energy expenditure in HFD-fed CysC KO mice that may have outweighed lower food intake in these mice (Supplementary Fig. 2). Of note, whereas energy expenditure was similarly decreased in chow-fed CysC KO mice, their locomotor activity and food intake were not different when compared with WT mice (Supplementary Fig. 3).

In inguinal WAT (iWAT) of HFD-fed mice, levels of the adipocyte differentiation markers peroxisome proliferator–activated receptor-γ (PPARγ) and CEBPβ were decreased (Fig. 1F and G). Consistently, adipocyte size was enlarged in HFD-fed CysC KO mice (Fig. 1H). In contrast, PPARγ and CEBPβ protein levels were similar in iWAT of chow-fed WT and CysC KO mice (Supplementary Fig. 4A). Moreover, recombinant CysC treatment did not affect differentiation of 3T3-L1 adipocytes (Supplementary Fig. 4B), indicating that CysC has no direct effect on the latter. Obesity-associated adipose tissue dysfunction is accompanied by impaired angiogenesis (17). In accordance, levels of the proangiogenic protein vascular endothelial growth factor (VEGF) were decreased in iWAT of HFD-fed CysC KO mice (Supplementary Fig. 5A). Similarly, VEGF protein levels were decreased in skeletal muscle (Supplementary Fig. 5B). Taken together, lack of CysC aggravates HFD-induced adiposity and reduces adipocyte differentiation markers, consistent with adipose tissue dysfunction in these mice.

Glucose metabolism was assessed next in chow- and HFD-fed CysC KO and WT mice. Whereas CysC KO mice were slightly more glucose tolerant under the chow diet, glucose tolerance was further deteriorated in HFD-fed CysC KO compared with WT mice (Fig. 2A–C). Accordingly, the negative impact of the HFD on glucose tolerance was significantly more pronounced in CysC KO mice (Fig. 2D). Insulin tolerance was increased in chow-fed but similar in HFD-fed CysC KO mice compared with WT mice (Supplementary Fig. 6A–D). Circulating insulin levels were similar in both genotypes under both diets (Supplementary Fig. 6E). Hyperinsulinemic-euglycemic clamp studies revealed a similar glucose infusion rate reflecting similar whole-body insulin sensitivity in HFD-fed CysC KO and WT mice (Fig. 2E). In contrast, endogenous glucose production was significantly higher in CysC KO mice (Fig. 2F), indicating impaired hepatic insulin sensitivity. Glucose uptake into oxidative soleus muscle was trend-wise increased in CysC KO mice (Fig. 2G). Possibly, increased glucose uptake into oxidative skeletal muscle counteracted the blunted insulin response in the liver, resulting in an overall similar glucose infusion rate (Fig. 2E). Albeit not significantly different, glucose uptake into iWAT was almost 50% lower in CysC KO (Fig. 2G). In addition, insulin’s capacity to reduce circulating FFA during clamps was significantly blunted in CysC KO mice (Fig. 2H), suggesting impaired WAT insulin sensitivity. Taken together, CysC KO mice are more susceptible to HFD-induced deterioration of glucose tolerance.

CysC may have anti-inflammatory properties (9). Consistently, mRNA expression of proinflammatory proteins was increased in both iWAT and eWAT of HFD-fed CysC KO mice (Fig. 3A and B). In line, infiltration with proinflammatory macrophages was enhanced in eWAT of HFD-fed CysC KO compared with WT mice as assessed by flow cytometry (Fig. 3C and Supplementary Fig. 7). Similar to WAT, expression of proinflammatory cytokines was increased in the liver and skeletal muscle of HFD-fed CysC KO mice (Fig. 3D and E). In contrast, mRNA expression of these cytokines was not different between both genotypes under the chow diet (Supplementary Fig. 8).

To explore whether chow-fed CysC KO mice were more susceptible to exogenous proinflammatory stimuli, LPS, which was demonstrated to promote adipose tissue inflammation in lean mice (18), was chronically administrated to chow-fed CysC KO and WT mice using minipumps. We hypothesized that an isolated increase in circulating LPS levels would aggravate adipose tissue inflammation in chow-fed mice lacking the protective effect of CysC. Indeed, adipose tissue inflammation was pronounced in lean CysC KO mice, as demonstrated by increased mRNA expression of Il6 (Fig. 3F). In line, phosphorylation of IL-6 downstream targets, such as STAT3 and extracellular signal–regulated kinase (ERK) 1/2, were increased in CysC KO mice (Supplementary Fig. 9).

CysC can be internalized into LPS-stimulated macrophages, thereby downregulating the expression and secretion of proinflammatory cytokines (9,19). CysC treatment of LPS-stimulated RAW cells decreased mRNA expression of Hif1α (Fig. 3G), which has been previously linked to adipose tissue inflammation (20). In parallel, the release of TNF-α and MCP1 was reduced (Supplementary Fig. 10). To test potential anti-inflammatory properties of CysC in obesity-associated adipose tissue inflammation, eWAT explants of HFD-fed mice were incubated with or without recombinant CysC. Importantly, CysC treatment significantly reduced Hif1α mRNA levels in eWAT explants (Fig. 3H). In addition, Vegfa mRNA expression was upregulated in response to CysC treatment (Supplementary Fig. 11).

Taken together, CysC modulates inflammation in insulin-sensitive tissues in mice.

Consistent to the observation in rodents (Fig. 3A), circulating CysC levels correlated negatively with HIF1α (Fig. 4A) and TNFα (Fig. 4B) mRNA expression in subcutaneous WAT of human subjects. Similarly, HIF1α and IL6 mRNA expression in visceral WAT associated negatively with serum CysC concentration (Fig. 4C and D). Moreover, the latter correlated negatively with the proinflammatory protein lipocalin-2 (Fig. 4E) but positively with the anti-inflammatory cytokine IL-10 (Fig. 4F). In further support of anti-inflammatory properties of CysC, healthy (i.e., insulin-sensitive) obese subjects had higher mRNA expression of the CysC gene CST3 in subcutaneous WAT than insulin-resistant obese patients (Fig. 4G).

The current study shows that elevated circulating CysC levels in obesity reduce (adipose) tissue inflammation and, thus, have a beneficial impact on glucose tolerance. Circulating CysC is increased not only in kidney failure but also in obese humans and mice (21,22). Consistently, we found increased circulating CysC levels in UniNx as well as in HFD-fed mice. Of note, WAT from obese humans revealed higher expression and secretion of CysC, suggesting a direct contribution of WAT to the obesity-associated increase in circulating CysC (23). Similarly, murine WAT expresses significant amounts of CysC (24), supporting the notion that WAT is an important source of circulating CysC levels.

In support of the suggested anti-inflammatory properties of CysC, its circulating levels correlated negatively with inflammatory markers in human subcutaneous and visceral WAT, and additionally, CysC treatment reduced inflammatory markers in murine fat explants. Moreover, lack of CysC promotes an aggravation of HFD-induced WAT, liver, and skeletal muscle inflammation in mice as determined by mRNA expression of proinflammatory cytokines and accumulation of proinflammatory macrophages in WAT. A putative role of obesity-associated macrophage accumulation as a major cytokine source and, hence, as driver of WAT inflammation has been suggested before (1). CysC may inhibit the secretion of proinflammatory cytokines by LPS-stimulated macrophages (9). Consistently, we found herein that LPS-mediated WAT inflammation was aggravated in lean CysC KO mice, suggesting a direct immunomodulatory effect of CysC on WAT inflammation. In the latter experiment, chronically LPS-infused CysC KO mice displayed not only pronounced Il6 upregulation but also increased activation of its downstream targets STAT3 and ERK 1/2 compared with WT mice. Of note, a regulatory effect of CysC on LPS-stimulated ERK 1/2 phosphorylation has been previously reported in human monocytes (9).

Obesity-associated WAT inflammation may inhibit adipogenesis, resulting in adipocyte hypertrophy (17,20). In fact, inflammatory mediators such as IL-6 and TNF-α, which are both elevated in WAT of HFD-fed KO mice, may negatively affect adipocyte differentiation and/or induce dedifferentiation of white adipocytes (25,26). In support of this notion, elevated Il6 and Tnfα expression observed in HFD-fed CysC-deficient mice was associated with diminished protein levels of the adipocyte differentiation factors PPARγ and CEBPβ as well as hypertrophied adipocytes, further supporting our hypothesis that CysC elicits anti-inflammatory properties and may thus counteract WAT inflammation. Moreover, we observed decreased levels of the proangiogenic protein VEGF-A, which is important for healthy WAT expansion and vascularization (17), in WAT of HFD-fed CysC KO mice. Consistently, CysC has been reported to increase VEGF (27). These findings further support a beneficial role of CysC in the adipose tissue of obese.

In conclusion, increased CysC in obesity may counterbalance obesity-associated inflammation and, thus, protects from additional obesity-associated deterioration of glucose homeostasis. Besides elevated inflammation, increased HFD-induced body weight gain and/or impaired vascularization may contribute to impaired glucose tolerance in HFD-fed CysC KO mice.

Funding. This work was supported by grants from the Wolfermann-Nägeli Foundation, the Children’s Research Centre, University Children’s Hospital Zurich, the Universität Zürich Forschungskredit “Candoc” (FK-18-027 all to M.A.D.), and from the Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung (Swiss National Science Foundation) (310030-160129 and 310030-179344 to D.K.).

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. M.A.D. designed and performed experiments, analyzed data, and wrote the manuscript. S.W. designed and performed experiments. T.D.C., F.C.L., T.R.J.A., M.Bo., A.A.M., and S.V. performed experiments. J.P.S. designed experiments. M.Bl. provided human samples and designed experiments. D.K. designed experiments, analyzed data, and wrote the manuscript. All authors reviewed and commented on the manuscript. D.K. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Prior Presentation. Parts of this study were presented as a poster presentation at the 80th Scientific Sessions of the American Diabetes Association, 12–16 June 2020.