Role of Kallikrein 7 in Body Weight and Fat Mass Regulation

Increased plasma and adipose tissue protease activity is observed in patients with type 2 diabetes and obesity. It has been proposed that specific proteases contribute to the link between obesity, adipose tissue inflammation and metabolic diseases. We have recently shown that ablation of the serine protease kallikrein-related peptidase 7 (Klk7) specifically in adipose tissue preserves systemic insulin sensitivity and protects mice from obesity-related AT inflammation. Here, we investigated whether whole body Klk7 knockout (Klk7−/−) mice develop a phenotype distinct from that caused by reduced Klk7 expression in adipose tissue. Compared to littermate controls, Klk7−/− mice gain less body weight and fat mass both under chow and high fat diet (HFD) feeding, are hyper-responsive to exogenous insulin and exhibit preserved adipose tissue function due to adipocyte hyperplasia and lower inflammation. Klk7−/− mice exhibit increased adipose tissue thermogenesis, which is not related to altered thyroid function. These data strengthen our recently proposed role of Klk7 in the regulation of body weight, energy metabolism, and obesity-associated adipose tissue dysfunction. The protective effects of Klk7 deficiency in obesity are likely linked to a significant limitation of adipocyte hypertrophy. In conclusion, our data indicate potential application of specific KLK7 inhibitors to regulate KLK7 activity in the development of obesity and counteract obesity-associated inflammation and metabolic diseases.


Introduction
Obesity increases the risk to develop type 2 diabetes, cardiovascular and fatty liver diseases, likely in the context of promoting an underlying pro-inflammatory state [1]. antibodies (CST#7074). Chemiluminescence was detected using the G:BOX Chemi XX9 documentation system with GeneTools analysis software (Syngene, Cambridge, UK).

Proteomics of AT Depots
Formalin-fixed, paraffin embedded specimens of iWAT and eWAT depots of control or ATKlk7 −/− mice were processed [27] and liquid chromatography tandem mass spectrometry (LC-MS/MS) was performed as described [28]. Each group consisted of four biological replicates. Data analysis was performed with MaxQuant 1.6.12, using tryptic specificity with up to two missed cleavages, fixed cysteine carbamidomethylation, no variable modifications, and a mouse reference proteome sequence file downloaded from EMBL-EBI on 4th February 2020. Statistical analysis of quantitative proteomic data was performed using linear models for microarray analysis (limma) [29,30]. Proteins with an FDR-adjusted p-value < 0.05 and a quantitative alteration of >50% were considered to be significantly affected (Supplementary Tables S2-S5). Proteins were defined as exclusively expressed, whenever they were detected in N ≤ 1 samples of group A and N ≥ 3 samples of group B for the respective comparisons. For a broad over-representation analysis, proteins with an unadjusted p-value < 0.01 irrespective of fold change were considered, together with proteins exclusively detected in either control or ATKLK7 −/− mice. Over-representation analysis was performed using the R topGO package (Version 2.42.0) [31] and gene ontology (GO) annotation. A broad gene list enrichment analysis for pathways was performed using Enrichr (www.amppharm.mssm.edu/Enrichr) [32,33].

Statistical Analyses
Statistical analyses were performed using GraphPad Prism 7.0. Data are presented as means ± SEM. Differences between groups were analyzed as indicated in the figure legends. Sample sizes for every figure or table are presented in Supplementary Table S6.

Generation of Klk7 −/− Mice
Constitutive Klk7 knockout was achieved by deleting exons two to five of the Klk7 gene as previously described [16]. Heterozygous genetic deletion generates a Cre PCR product of a 100 bp as well as 276 bp (control) and 402 bp (floxed) instead of a 276 bp band found in control animals. Homozygous Klk7 −/− animals showed a 100 bp Cre PCR product (Supplementary Figure S1A). Klk7 −/− mice are fertile and were obtained with the expected Mendelian frequency. Using PCR analysis of genomic DNA, we confirmed Klk7 gene disruption (Supplementary Figure S1B).  Table 1). Lower body weight of Klk7 −/− mice was accompanied by a~40% significant reduction of total fat mass under HFD ( Figure 1C,D). In chow-fed mice, relative eWAT fat mass was significantly reduced, and HFD leads to both, a significant reduction in eWAT and iWAT fat mass in Klk7 −/− vs. control mice ( Figure 1E,F). Lean body mass was not altered in Klk7 −/− mice (Supplementary Figure S2). Similarly, liver weight and liver glycogen content were not different ( Figure 1E and Supplementary Figure S3). iWAT fat mass compared to obese controls. Data are presented as means ± SEM; significance tested by two-way ANOVA with correction for multiple testing (Sidak; A,B) or multiple t-tests correction for multiple testing (Holm-Sidak; C-F); * p < 0.05; ** p < 0.01; *** p < 0.001. for Klk7 −/− mice beginning at an age of 10 weeks and independent of the diet. (E,F) Determination of relative organ weights show that Klk7 −/− mice under chow diet exhibit decreased eWAT fat mass compared to chow-fed control mice, while HFD-fed Klk7 −/− mice have both reduced eWAT and iWAT fat mass compared to obese controls. Data are presented as means ± SEM; significance was tested by two-way ANOVA with correction for multiple testing (Sidak; A,B) or multiple t-tests with correction for multiple testing (Holm-Sidak; C-F); * p < 0.05; ** p < 0.01; *** p < 0.001.

Body Weight
Energy expenditure was assessed using indirect calorimetry ( Figure 2). While there was no change in oxygen consumption ( Figure 2A) and total energy expenditure (EE, Figure 2B) in chow-fed animals, Klk7 −/− mice under HFD showed significantly higher oxygen consumption ( Figure 2D) and EE during the light phase compared to controls (p < 0.05; Figure 2E). Locomotor activity was not affected by Klk7 −/− . Interestingly, Klk7 −/− exhibited a preference for carbohydrate oxidation reflected by significantly higher respiratory exchange ratios (RER) compared to controls ( Figure 2C,F). As observed in ATKlk7 −/− mice [16], food intake was higher in Klk7 −/− mice, both under chow and HFD ( Figure 2G,H).

Consequences of Klk7 −/− on Adipose Tissue
We performed adipocyte size distribution measurements and AT histology analyses. Expression of genes involved in thermogenesis (Ucp1, Cidea, Dio2, Elovl3, Fgf21, Pgc1a, Prdm16) and WAT browning (Tmem26, Tbx1) were analyzed in BAT, iWAT and eWAT. In HFD-fed mice, adipocytes were significantly smaller in eWAT ( Figure 3A,B) and iWAT ( Figure 3C,D) of Klk7 −/− mice compared to controls. There were no differences in the expression of thermogenic genes ( Figure 3E). Macrophage infiltration into eWAT of HFDfed Klk7 −/− mice was diminished, with more M2 and significantly less M1 polarized AT macrophages in the AT samples analyzed (Supplementary Figure S4). In Klk7 −/− mice, HFD-induced whitening of BAT was reduced ( Figure 3F) in parallel to increased expression of Dio2 and Prdm16 ( Figure 3G). In chow-fed Klk7 −/− mice, gene and protein expression related to BAT activity, mitochondrial activity and WAT browning were significantly higher compared to controls ( Figure 3I,J).

Consequences of Klk7 −/− on Adipose Tissue
We performed adipocyte size distribution measurements and AT histology analyses. Expression of genes involved in thermogenesis (Ucp1, Cidea, Dio2, Elovl3, Fgf21, Pgc1a, Prdm16) and WAT browning (Tmem26, Tbx1) were analyzed in BAT, iWAT and eWAT. In HFD-fed mice, adipocytes were significantly smaller in eWAT ( Figure 3A,B) and iWAT ( Figure 3C,D) of Klk7 −/− mice compared to controls. There were no differences in the expression of thermogenic genes ( Figure 3E). Macrophage infiltration into eWAT of HFDfed Klk7 −/− mice was diminished, with more M2 and significantly less M1 polarized AT macrophages in the AT samples analyzed (Supplementary Figure S4). In Klk7 −/− mice, HFD-induced whitening of BAT was reduced ( Figure 3F) in parallel to increased expression of Dio2 and Prdm16 ( Figure 3G). In chow-fed Klk7 −/− mice, gene and protein expression related to BAT activity, mitochondrial activity and WAT browning were significantly higher compared to controls ( Figure 3I,J).  Data are presented as means ± SEM; significance was tested by t-tests (D) or multiple t-tests with correction for multiple testing; * p < 0.05, ** p < 0.01, *** p < 0.001.

Klk7 −/− Mice Exhibit Lower Body Temperature and Altered Circulating T3
Thyroid hormones are major physiological regulators of metabolism and AT function. Surprisingly, body temperature of chow-fed Klk7 −/− was significantly lower compared to control mice ( Figure 4A). As multiple kallikrein proteases including KLK7 are expressed in thyroid tissue [37], we investigated potential effects of Klk7 deficiency on thyroid hormone generation, circulating thyroid stimulating hormone (TSH), total T4 (TT4), and free T3 (fT3) concentrations. While TSH and TT4 remained unchanged, fT3 serum levels were in tendency lower in Klk7 −/− compared to control mice ( Figure 4B-D). In the liver, expression of thyroid hormone responsive genes Pdk4 and thyroxine-binding globulin (Tbg) were decreased, while deiodinase 1 (Dio1) expression was not altered ( Figure 4E). Expression of thyroglobulin solubilizing and degrading enzymes were evaluated in thyroid lysates of Klk7 −/− mice, but protein amounts of aspartic cathepsin D as well as the cysteine proteases cathepsins B and L were not different from controls ( Figure 4F and Supplementary Figure S5). In addition, thyroid protein analyses using gel-electrophoresis or immunostaining did not indicate genotype-related abnormalities ( Figure 4G,H). Finally, thyroid histology and phenotyping revealed that thyroid follicle size and distribution is similar in control and Klk7 −/− mice ( Figure 4I).
In contrast to chow-fed animals, fasted and fed blood glucose levels were not significantly different between KLK7 −/− and control mice in response to HFD ( Figure 5D). Despite the leaner phenotype of HFD-fed KLK7 −/− mice, glucose tolerance was not different between the genotypes ( Figure 5E). In contrast to KLK7 −/− mice, HFD-fed controls almost lost responsiveness to exogenous insulin ( Figure 5F). The leptin to adiponectin ratio and MCP-1 levels were significantly lower in both, chow-and HFD-fed KLK7 −/− mice compared to controls ( Figure 6A, Table 1).
Interestingly, in chow-fed mice glucose infusion rate (GIR) during hyperinsulinemiceuglycemic clamp studies was not different between KLK7 −/− and controls ( Figure 6B). Hepatic glucose production (HGP) and insulin-induced suppression of HGP were indistinguishable between KLK7 −/− and control animals ( Figure 6C). At the organ level, we measured significantly higher uptake of glucose into the heart of KLK7 −/− mice ( Figure 6D). Fully differentiated primary adipocytes from eWAT and iWAT of KLK7 −/− mice showed significantly lower insulin-induced AKT activation in Western blot analyses ( Figure 6E).  and revealed no significant changes. (G) Thyroid proteins were separated by reducing and non-reducing SDS-PAGE and analyzed after silver-staining, revealing similar banding patterns for both genotypes. Western blot analysis of thyroglobulin expression in thyroid tissue of control and Klk7 −/− mice did not show differences, as did (H) immunostaining of thyroglobulin in tissue sections. (I) Immunostaining of collagen IV and staining of cytoplasm with CellMask™ Orange (CMO) and nuclei using Draq5™ of thyroid tissue from control and Klk7 −/− mice (left panels). Follicle size and distribution is similar, with larger follicles in the periphery (middle panels) and smaller follicles in the thyroid lobes' center (right panels). Follicle area, lumen area, epithelial extensions, and cell counts per follicle area (right panels, respectively) appear comparable, again supporting unaltered thyroglobulin synthesis, deposition and solubilization in Klk7 −/− mice. Data are presented as means ± SEM; significance was tested by Student's t-test (A-D), two-way ANOVA with correction for multiple testing (Tukey; (E)); * p < 0.05, ** p < 0.01.
Interestingly, in chow-fed mice glucose infusion rate (GIR) during hyperinsulinemiceuglycemic clamp studies was not different between KLK7 −/− and controls ( Figure 6B). Hepatic glucose production (HGP) and insulin-induced suppression of HGP were indistinguishable between KLK7 −/− and control animals ( Figure 6C). At the organ level, we measured significantly higher uptake of glucose into the heart of KLK7 −/− mice ( Figure 6D). Fully differentiated primary adipocytes from eWAT and iWAT of KLK7 −/− mice showed significantly lower insulin-induced AKT activation in Western blot analyses ( Figure 6E).

(C) Hepatic glucose production (HGP) during clamp (as % from basal HGP) is not affected by the genotype (n = 6/3). (D)
Tissue glucose uptake (dpm/g/mL) is significantly higher in hearts of Klk7 −/− mice compared to controls (n = 6/3). (E) Insulin-induced AKT phosphorylation analyzed by Western blot in differentiated primary adipocytes from eWAT (left) and iWAT (right) from three animals per genotype. Densitometric quantification of bands is presented next to the blots. Phosphorylation of AKT by insulin is significantly dampened in all primary adipocytes from Klk7 −/− mice. Data are presented as means ± SEM; significance was tested by Student's t-test (A-C), uncorrected multiple t-tests (D) or two-way ANOVA with correction for multiple testing (Tukey; (E)); * p < 0.05, ** p < 0.01, *** p < 0.001.

Klk7 −/− Mice Have Depot-Specific Protein Expression Signatures in AT
As we observed major alterations in AT of Klk7 −/− mice, we aimed to unravel changes in protein expression that may contribute to improved AT function. Therefore, we performed label-free quantitative proteome analyses in eWAT and iWAT depots of adiposespecific Klk7 −/− knockout mice (ATKlk7 −/− ) and control mice fed either chow or HFD. First, we compared expression of proteins in eWAT and iWAT between the HFD and chow diet group in controls only. Proteins over-and exclusively expressed in HFD versus chow fed mice were considered as markers of obesity. As shown in Supplementary Table S2, pathways related to PPAR signaling, fatty acid, glucose and cholesterol metabolism were increased in HFD induced obesity. We then focused on the impact of Klk7 deletion in AT on proteome composition in ATKlk7 −/− mice. Significantly affected proteins were only identified in iWAT (23 under chow and 59 under HFD conditions) ( Figure 7A,B). The number of proteins exclusively detected in AT samples was higher in iWAT (52 under  creased in HFD induced obesity. We then focused on the impact of Klk7 deletion in AT on proteome composition in ATKlk7 −/− mice. Significantly affected proteins were only identified in iWAT (23 under chow and 59 under HFD conditions) ( Figure 7A,B). The number of proteins exclusively detected in AT samples was higher in iWAT (52 under chow and 66 under HFD) than in eWAT (8 under chow and 21 under HFD). These data indicate a more prominent role of Klk7 in subcutaneous AT. There was little overlap between identified proteins in chow and HFD conditions (Supplementary Tables S3 and S4).   Table S4). This was accompanied by a number of significantly downregulated markers, thereof multiple tight junction proteins (MYLPF, MYH2, ACTN3, NEB). Pathway analysis revealed enrichment of proteins involved in ER processing of proteins, the phagosome, in amino and nucleotide sugar metabolism as well as PPAR signaling and sucrose metabolism to be affected by loss of Klk7 in AT. GO term analysis indicated upregulation of proteins involved in the regulation of immune responses as well as intracellular signal transduction and cell adhesion ( Figure 7C). Repressed pathways were related to cGMP-PKG as well as calcium signaling (Supplementary Table S5) and proteins related to GO terms of muscle contraction, ATP metabolism and monocarboxylic acid metabolism were significantly repressed in iWAT of chow-fed ATKlk7 −/− mice compared to controls ( Figure 7C).
In iWAT of HFD-fed animals, COL3A1, CD163, NCK1, HFE, DNAJC3 and AGT were downregulated ATKlk7 −/− mice compared to controls. KEGG pathway analysis revealed a highly significant induction of metabolic proteins in iWAT of obese ATKlk7 −/− mice that are involved in fatty acid elongation and degradation (Supplementary Table S5). GO terms enrichment revealed significant reduction of proteins involved in extracellular structure organization and regulation of transport ( Figure 7D). We compared our data to a dataset comprising potential proteolytic substrates of KLK7 identified by degradome analyses [38]. Overall, putative KLK7 substrates (~20% of differentially expressed proteins) were only observed in iWAT of HFD fed animals, independent of the genotype (Supplementary  Tables S3 and S4). Our analyses corroborate a more pronounced impact of diet over KLK7 genotype on AT proteome. In subcutaneous AT, KLK7 seems to play an important role in the regulation of immune response and metabolism.

Discussion
We tested the hypothesis that ablation of Klk7 at the whole body level may cause a metabolically "healthier" phenotype that is distinct from that observed in ATKlk7 −/− mice.
A key finding of the study is that whole-body deletion of Klk7 resembles the phenotype of AT-specific KO in many aspects. However, beyond the expected growth curve differences from ATKlk7 −/− mice [16], Klk7 −/− mice exhibit a leaner phenotype with aging. Body weight differences between Klk7 −/− and control mice became significant under chow fed conditions at an age of 30 weeks. Both under chow and HFD, body fat mass was significantly lower in Klk7 −/− compared to control mice, whereas lean body mass or organ weights were not affected by Klk7 ablation.
In contrast to ATKlk7 −/− , Klk7 −/− mice exhibit a generalized fat mass reduction suggesting an additional systemic role of Klk7 reduction beyond AT-specific effects on adipogenesis or expandability. Counter-intuitively to lower fat mass and preserved lean mass, Klk7 −/− mice had higher food intake than controls. This may represent an (not entirely sufficient) attempt to compensate for higher energy expenditure of Klk7 −/− compared to control mice as found in metabolic chamber experiments. In addition, lower fat mass was accompanied by lower circulating leptin, which may have contributed to higher food intake of Klk7 −/− mice. However, based on our data, we are not able to distinguish potential causality chains whether effects of Klk7 ablation primarily cause activation of energy metabolism or affect central regulation of food intake. Lower leptin levels, a higher RER, absence of adipocyte hypertrophy and increased thermogenic gene expression may additionally suggest AT browning as another mechanism explaining the leaner phenotype and higher energy expenditure of Klk7 −/− mice. As result of lower fat mass and/or healthier AT, we find circulating adipokine differences between the genotypes with lower levels of systemic inflammatory markers (e.g., MCP1) and higher circulating adiponectin. This phenotype was preserved under HFD conditions and mostly reflects the phenotype of ATKlk7 −/− mice [16].
Thyroid hormones are major physiological regulators of metabolism [39] and partly of AT function [40]. In rats, kallikrein-related peptidases are regulated by thyroid hormones [41]. Although the overall thyroid state remained unaffected in Klk7 −/− mice as indicated by unchanged serum TT4 and TSH levels, fT3 levels were lower in comparison to control animals. Our data suggest that T4-to-T3 conversion rates or T4-inactivation are affected by loss of Klk7. However, expression of the outer-and inner-ring deiodinating enzyme Dio1 in the liver was not affected by Klk7 deficiency, arguing against a liver-mediated change in TH metabolism. To further investigate the role of thyroid hormones on the phenotype of Klk7 −/− mice, we performed detailed biochemical and morphometric analyses. The results indicated unchanged protein amounts of aspartic and cysteine cathepsins, which are the main thyroglobulin processing and degrading enzymes acting in the follicle lumen for solubilization from cross-linked multimers (cathepsins B and L) and in endo-lysosomal degradation of the TH precursor (cathepsins B, D, and L) [34,42,43]. Because thyroglobulin degradation states did not differ between the genotypes, we conclude that thyroglobulin biosynthesis, cross-linking for storage in the follicle lumen and utilization for TH liberation are not altered in Klk7 −/− mice. However, we cannot exclude that decreased fT3 concentrations in the circulation feed back onto thyroid functionality. Hence, a thorough analysis was performed using semi-automated Cell Profiler-based histological examination of thyroid tissue sections [34][35][36], indicating that KLK7 is dispensable for proper thyroid development and function. This is an interesting finding, because the thyroid protease web is characterized by a high degree of redundancy when cysteine peptidases are considered [42,43]. Obviously, serine proteases such as KLK7 are not redundantly co-regulated with thyroglobulin-processing enzymes, and therefore likely play a different role in thyroid physiology.
Based on the phenotype similarities between ATKlk7 −/− and Klk7 −/− mice, we postulate that AT development and metabolism plays an important role in determining the phenotype of Klk7 −/− mice. Compared to iWAT of control mice, we found significant differences in expression of proteins previously connected to AT development, function and expansion, such as COL3A1, CD163, NCK1, HFE, DNAJC3 and AGT [44][45][46][47][48][49]. Thus, the proteome signature of iWAT from ATKlk7 −/− mice reflects a maintained normal AT function under HFD conditions. In addition, regulators of immune response were higher expressed, whereas proteins promoting macrophage-infiltration during the progression of obesity-associated inflammation, such as CD5L [50], were reduced. Together, AT signatures distinctly associated with reduced Klk7 expression may contribute to lower macrophage infiltration into AT [16].
The (AT-specific) consequences of Klk7 disruption may be explained by KLK7-related processes in AT including a reduced activation of pro-inflammatory adipokines such as chemerin. Interestingly, chemerin was only detected in iWAT of Klk7 −/− mice with diet induced obesity. We have previously shown that KLK7 activates human chemerin by cleaving the preform into the active chemerin156 [51]. The effects of chemerin on adipocyte insulin sensitivity and glucose uptake are disputed [52,53] and may not alone explain impaired glucose uptake observed in primary adipocytes of Klk7 −/− mice. However, Klk7 −/− mice did not exhibit lower circulating chemerin, suggesting that modulation of chemerin activation by KLK7 does not represent the main mechanism explaining the leaner phenotype. Yet, chemerin is also involved in adipogenesis [54], and this may be in accordance with an increase of iWAT mass [16]. A higher expression of cathepsin L protease in iWAT of Klk7 −/− mice may further contribute to the expansion of this depot, as loss this protease has been shown to affect adipocyte differentiation and lipid accumulation [7].
Both chow and HFD-fed Klk7 −/− mice are more responsive to exogenous insulin. In hyperinsulinemic-euglycemic clamps, we could exclude that higher sensitivity to exogenous insulin was caused by higher insulin sensitivity of Klk7 −/− compared to control mice. Differences in the response to intraperitoneal versus intravenous insulin application could be related to local effects of KLK7 action on insulin degradation. Insulin is cleaved by KLK7 in vitro [55], and inhibition of KLK7 activity is associated with prolonged insulin circulation and action in vivo [15]. One limitation of our study is that we are not able to provide formal proof of the concept that KLK7 modulates insulin action in vivo. Circulating, putatively KLK7-derived insulin fragments or differences in insulin degradation dynamics between Klk7 −/− and control mice would have supported our hypothesis that KLK7-mediated insulin degradation occurs in vivo and contributes to prolonged insulin action. Despite the better response to insulin of Klk7 −/− mice, endogenous insulin does not seem to be sufficient to maintain glucose homeostasis, particularly after a glucose challenge. Under chow fat conditions and despite lower fat mass, Klk7 −/− mice are characterized by higher fasted, fed and GTT glucose levels. These differences were not confirmed under HFD. Because we consider AT as the critical organ of KLK7 action with regard to glucose homeostasis, we further investigated AT insulin sensitivity. Indeed, we find that primary adipocytes from Klk7 −/− mice display lower insulin-induced AKT-phosphorylation compared to adipocytes obtained from controls and may therefore be considered insulin resistant. In contrast, glucose uptake into AT during the steady state of the clamp was not significantly different between Klk7 −/− and control mice. Lower expression levels of AT SERCAs (sarcoplasmic/endoplasmic reticulum calcium ATPases SERCA3 and SERCA1), voltage-dependent anion-selective channel 1 (VDAC1) and ADP/ATP translocase 2 (SLC25A5) detected in our proteome analysis point towards impaired calcium homeostasis and increased cytosolic calcium levels. While we did not analyze hepatic proteome changes, similar changes in hepatic calcium levels could dysregulate liver glucose production via increased gluconeogenesis, thereby contributing to the observed hyperglycemia in chow-fed AT-Klk7 −/− mice.

Conclusions
Whole body Klk7 deficiency significantly decreases body weight gain, both age-related and diet-induced. These data confirm and strengthen our recently demonstrated role of Klk7 in the regulation of body weight, energy metabolism, and obesity-associated adipocyte dysfunction using the AT-specific knockout [16]. Yet, in the constitutional knockout, the protective effects of Klk7 deficiency against obesity-associated adipose dysfunction and subsequently insulin resistance under HFD can be mainly attributed to lower fat mass (both, of eWAT and iWAT). This is a striking difference to the conditional ATKlk7 −/− , where the protective effects were observed despite no significant differences in HFD-induced weight gain and obesity.
Together with rodent and human studies providing evidence that elevated plasma protease activity is associated with obesity and type 2 diabetes [4][5][6], this argues for the contribution of specific proteases to obesity and related diseases. Our data further suggest KLK7 as a treatment target for these diseases. Over the last years, various small molecule or peptide-based inhibitors of KLK7 have been described [56][57][58][59][60][61][62][63][64][65][66]. Application of these molecules to Klk7 −/− mice may further improve our understanding of KLK7-related or -dependent mechanisms in AT function and obesity development.

Institutional Review Board Statement:
The study was conducted according to the guidelines of the Declaration of Helsinki, and all animal experiments were approved by the local authorities of the state of Saxony (Germany), as recommended by the responsible local animal ethics review board (Landesdirektion Leipzig, Germany, TVV23/12, TVV16/15, T02/13 and T08/16).
Data Availability Statement: All data generated or analyzed during this study are included in the published article and its online Supplementary files. The resources (e.g., Klk7 −/− and ATKlk7 −/− mice) generated and analyzed during the current study are available from the corresponding author upon reasonable request.

Acknowledgments:
We would like to thank Eva Böge, Viola Döbel, Frauke Beermann, Claudia Gebhardt, Sabine Päschke and Daniela Kern for technical assistance.