Anserine and Carnosine Induce HSP70-Dependent H2S Formation in Endothelial Cells and Murine Kidney

Anserine and carnosine have nephroprotective actions; hydrogen sulfide (H2S) protects from ischemic tissue damage, and the underlying mechanisms are debated. In view of their common interaction with HSP70, we studied possible interactions of both dipeptides with H2S. H2S formation was measured in human proximal tubular epithelial cells (HK-2); three endothelial cell lines (HUVEC, HUAEC, MCEC); and in renal murine tissue of wild-type (WT), carnosinase-1 knockout (Cndp1-KO) and Hsp70-KO mice. Diabetes was induced by streptozocin. Incubation with carnosine increased H2S synthesis capacity in tubular cells, as well as with anserine in all three endothelial cell lines. H2S dose-dependently reduced anserine/carnosine degradation rate by serum and recombinant carnosinase-1 (CN1). Endothelial Hsp70-KO reduced H2S formation and abolished the stimulation by anserine and could be restored by Hsp70 transfection. In female Hsp70-KO mice, kidney H2S formation was halved. In Cndp1-KO mice, kidney anserine concentrations were several-fold and sex-specifically increased. Kidney H2S formation capacity was increased 2–3-fold in female mice and correlated with anserine and carnosine concentrations. In diabetic Cndp1-KO mice, renal anserine and carnosine concentrations as well as H2S formation capacity were markedly reduced compared to non-diabetic Cndp1-KO littermates. Anserine and carnosine induce H2S formation in a cell-type and Hsp70-specific manner within a positive feedback loop with CN1.


Introduction
The histidine-containing dipeptide carnosine is synthesized in muscle and kidney by carnosine synthase and is methylated to anserine via carnosine methyltransferase [1]. Both dipeptides are degraded by carnosinase 1 (CN1), a product of the Cndp1 gene [2]. Diabetic patients with Cndp1 gene variants, which are associated with a lower serum CN1 activity, have a lower risk of nephropathy [3]. Serum CN1 concentrations have been correlated with renal fibrosis, oxidative stress and tubular injury [4]. In rodent models of diabetes type 1 and 2, carnosine treatment reduced oxidative stress, carbonyl stress and advanced glycation end-product (AGE) formation, and improved glucose homeostasis, all of which are associated with less structural and functional renal damage [5,6]. In contrast to rodents, anserine and carnosine are rapidly metabolized by serum CN1 in humans; still, the first clinical trials yielded some positive results [7,8], possibly by carnosine delivery to the kidney via erythrocytes [9]. Carnosine has been approved as a nutritional supplement and is well tolerated [10]. The mode of action of anserine and carnosine, however, is not

Hsp70-Knockout MCEC
Hsp70-knockout was generated by transfection (Neon Transfection System, Invitrogen, Waltham, MA, USA) of 10 6 with a vector from Sigma-Aldrich, targeting the stress-inducible Hsp70 variant Hspa1a (Gene ID: 193740; targeting sequence of the gRNA: TGTGCTCA-GACCTGTTCCG). The vector contained the respective gRNA target sequences, the Cas9 endonuclease gene and a fluorescent reporter gene (GFP for Hspa1a), that was used for single cell isolation by FACS. Clones were cultured, and genome, mRNA and protein analysis were performed to confirm successful knockout of Hspa1a.

Maximal H 2 S Production Capacity
H 2 S was identified by detection of Ag 2 S according to Ahn et al. [31]. Plates were coated with AgNO 3 /Nafion/PVP solution and dried at room temperature (1 h) before incubation start. Cells were seeded on 96-well plates (2 × 10 4 per well for HK-2, 5 × 10 4 per well for HUAEC, HUVEC and MCEC) followed by a 22 h growth period and treated with L-homocysteine, anserine and carnosine for 48 h; L-homocysteine was renewed after 24 h. Incubation with a CBS inhibitor (O-(Carboxymethyl)-hydroxylamin-hemihydrochlorid; AOAA) and a CSE inhibitor (DL-proparglycine; PAG) for 48 h was used to demonstrate the involvement of those enzymes in H 2 S production capacity.
Murine kidney tissue (20 mg in protease-inhibited DPBS buffer) was homogenized and centrifuged at 4 • C and 10,000× g for 10 min. Supernatant was treated for 6 h with L-cysteine.
Absorbance was measured at 310 nm, and Ag 2 S production was normalized to protein concentration (DC™ Protein Assay; Bio-Rad Laboratories, Hercules, CA, USA).

Western Blotting
Cell samples were lysed in RIPA buffer (radio-immunoprecipitation assay buffer: 150 mM NaCl, 0.1% Triton X-100, 0.5% sodium deoxycholate, 0.1% SDS and 50 mM Tris-HCl; pH 8.0) and protease inhibitor (cOmplete tablets, Mini EASYpack, Roche Diagnostics, Mannheim, Germany) and separated by SDS-PAGE in 8% polyacrylamide gels. Samples were transferred to a nitrocellulose membrane by semi-dry blot. The membrane was then blocked with 5% milk (1 h at room temperature) and incubated with anti-Hsp70-antibody (HSP70 Polyclonal antibody, ProteinTech, Rosemont, IL, USA, 1:10,000 in 5% milk; 1.5 h at room temperature). After washing with Tris-buffered saline with Tween 20 (TBS-T), the membrane was incubated with a secondary HRP-conjugated antibody (1:1000 in 5% milk) for 1 h at room temperature. Protein expression of the target protein was normalized to β-Actin expression of the representative sample.

Carnosinase Activity
CN1 activity was assayed according to the method described by Teufel et al. (2003) [2]. The reaction was initiated by addition of carnosine to human serum carnosinase or recombinant enzyme (rCN1; R&D Systems, Minneapolis, MN, USA) at pH of 7. The reaction was stopped after defined periods by adding 1% trichloracetic acid (final concentration in the test 0.3%). Liberated histidine was derivatized by adding o-phtaldialdehyde (OPA), and fluorescence was read using a plate reader (MicroTek International Inc., Hsinchu, Taiwan).

Statistical Analysis
Data were obtained from at least three independent experiments and are given as mean ± standard deviation (SD). Statistical analysis was performed with GraphPad prism 9 using analysis of variance (ANOVA) with Tukey's test. p-values of <0.05 were considered significant.

Dipeptide-Induced H 2 S Formation in Tubular Epithelial and Endothelial Cells
Homocysteine dose-dependently increased H 2 S formation in HK-2 cells (Supplementary Figure S1, Supplementary Table S1). At a homocysteine concentration of 5 mM, the H 2 S formation rate in HK-2, HUVEC, HUAEC and MCEC varied between 15 and 73 pmol H 2 S/mg protein ( Figure 1, Supplementary Table S2). The addition of anserine dosedependently increased H 2 S formation in HK-2 cells ( Figure 1A, Supplementary Table S3); H 2 S formation from cysteine was below the detection limit in HK-2. Addition of 1 mM anserine increased H 2 S formation in all cell types, and the addition of 1 mM carnosine increased H 2 S formation in HK-2 cells only. Combined anserine and carnosine exposure had no additive effect beyond the anserine-mediated effects ( Figure 1).
To study the interaction of H 2 S on anserine and carnosine availability, CN1 activity was measured at increasing concentrations of the H 2 S donor sodium disulfide (Na 2 S). Na 2 S dose-dependently reduced recombinant CN1 and human serum CN1 carnosine and anserine degradation activity ( Figure 2, Supplementary Table S4).

Anserine-Induced H 2 S Formation Depends on HSP70
Since anserine, carnosine and H 2 S have previously been described to modulate Hsp70 expression, we studied H 2 S formation capacity in MCEC with Hsp70-KO.

Anserine-Induced H2S Formation Depends on HSP70
Since anserine, carnosine and H2S have previously been described to modulate Hsp70 expression, we studied H2S formation capacity in MCEC with Hsp70-KO. H2S

Anserine-Induced H2S Formation Depends on HSP70
Since anserine, carnosine and H2S have previously been described to modulate Hsp70 expression, we studied H2S formation capacity in MCEC with Hsp70-KO. H2S

H2S Formation in Murine Kidney Tissue
We then studied ex vivo H2S formation capacity in kidney tissue of WT mice. Addition of homocysteine and cysteine at equimolar concentrations to kidney tissue homogenate both increased H2S formation, and homocysteine to a smaller extent (Supplementary Figure S4A, Supplementary Table S6). H2S formation was not influenced by five hours of fasting (Supplementary Figure S4B, Supplementary Table S6) and was markedly reduced by inhibition of both CBS and CSE activity, indicating that both enzymes are involved in

H 2 S Formation in Murine Kidney Tissue
We then studied ex vivo H 2 S formation capacity in kidney tissue of WT mice. Addition of homocysteine and cysteine at equimolar concentrations to kidney tissue homogenate both increased H 2 S formation, and homocysteine to a smaller extent (Supplementary Figure S4A, Supplementary Table S6). H 2 S formation was not influenced by five hours of fasting (Supplementary Figure S4B, Supplementary Table S6) and was markedly reduced by inhibition of both CBS and CSE activity, indicating that both enzymes are involved in kidney H 2 S formation (Supplementary Figure S5, Supplementary Table S7). To demonstrate the role of renal HSP70 on H 2 S formation, kidney tissues of mice with a global Hsp70-KO were studied in 23-to 28-week-old animals. Ex vivo kidney H 2 S formation capacity was 50% lower in female Hsp70-KO kidneys compared to WT mice ( Figure 3B). In male Hsp70-KO mice, H 2 S formation was within the range of WT littermates. Kidney anserine concentrations were similar in all four groups ( Figure 3C, Supplementary Table S8).

Kidney Anserine Abundance and H 2 S Formation
Incubation of kidney tissue with anserine and carnosine for six hours did not increase H 2 S formation (Supplementary Figure S6; Supplementary Table S9). We then studied the effect of a persistent increase of endogenous kidney anserine and carnosine concentrations on H 2 S formation in global Cndp1-KO mice. Male Cndp1-KO mice had 3-fold higher renal anserine concentrations, and female Cndp1-KO mice had 5-to 12-fold higher concentrations than respective WT controls, depending on age (Table 1). Renal carnosine concentrations were below the detection limit in WT and markedly increased in Cndp1-KO mice. H 2 S formation was 2-to 3-fold higher in female Cndp1-KO mice compared to their respective WT controls, but not different between male Cndp1-KO and WT animals. H 2 S formation and anserine/carnosine concentrations were correlated in 23-to 25-and 47-to 51-week-old mice (Figure 4).
We then studied the impact of diabetes on kidney anserine and carnosine concentrations and H 2 S formation. Kidney anserine and carnosine concentrations were similar in type 1 diabetic and non-diabetic WT mice, and carnosine concentrations were below detection level. In diabetic Cndp1-KO mice, kidney anserine and carnosine concentrations, however, were lower compared to non-diabetic Cndp1-KO littermates and in diabetic mice not related with H 2 S formation (Table 1, Figure 5). In female diabetic Cndp1-KO mice, kidney H 2 S formation was reduced compared to female non-diabetic Cndp1-KO mice and similar to diabetic WT mice. In male diabetic Cndp1-KO mice, kidney H 2 S formation was even reduced compared to the male diabetic WT littermates. We then studied the impact of diabetes on kidney anserine and carnosine concentrations and H2S formation. Kidney anserine and carnosine concentrations were similar in type 1 diabetic and non-diabetic WT mice, and carnosine concentrations were below detection level. In diabetic Cndp1-KO mice, kidney anserine and carnosine concentrations, however, were lower compared to non-diabetic Cndp1-KO littermates and in diabetic mice not related with H2S formation (Table 1, Figure 5). In female diabetic Cndp1-KO mice, kidney H2S formation was reduced compared to female non-diabetic Cndp1-KO mice and similar to diabetic WT mice. In male diabetic Cndp1-KO mice, kidney H2S formation was even reduced compared to the male diabetic WT littermates.

Discussion
Anserine and carnosine are experimentally well-established compounds mitigating diabetic nephropathy, and the first clinical trials yielded beneficial effects [7,[33][34][35]. Similarly, H2S improves outcome in acute and chronic kidney impairment, experimental sepsis, hemorrhagic shock and following ischemia/reperfusion [20]. The first clinical trials investigating the effects of the H2S donor Na2S2O3 in myocardial infarct in humans are

Discussion
Anserine and carnosine are experimentally well-established compounds mitigating diabetic nephropathy, and the first clinical trials yielded beneficial effects [7,[33][34][35]. Similarly, H 2 S improves outcome in acute and chronic kidney impairment, experimental sepsis, hemorrhagic shock and following ischemia/reperfusion [20]. The first clinical trials investigating the effects of the H 2 S donor Na 2 S 2 O 3 in myocardial infarct in humans are ongoing [27].
Both anserine/carnosine and H 2 S act via multiple mechanisms, and activation of HSP70 is a prominent common pathway [14,15,19,36]. We now demonstrate the cell type-specific increase in H 2 S synthesis capacity by anserine and carnosine, mediated by HSP70 activation (Figure 6). Ex vivo studies of kidney tissue reconfirm a correlation of H 2 S synthesis with tissue anserine and carnosine concentrations and demonstrate the mechanistic role of HSP70 in kidney H 2 S synthesis.  There is growing evidence that histidine-containing dipeptides exert m tive functions in various disease states by interfering with specific pathomecha Carnosine acts as a carbonyl scavenger [12,37], an ion-chelating agent [10], as tensin-Converting Enzyme (ACE) inhibitor [38,39] and multifunctional [40,41]. In HK-2 cells, anserine has a higher antioxidative capacity than carn These activities are mediated via hormetic processes involving Nrf2, Sirt-1, T the glutathione system [12,17,41] and modulation of the nitric oxide formati There is growing evidence that histidine-containing dipeptides exert major protective functions in various disease states by interfering with specific pathomechanisms [17]. Carnosine acts as a carbonyl scavenger [12,37], an ion-chelating agent [10], as an Angiotensin-Converting Enzyme (ACE) inhibitor [38,39] and multifunctional antioxidant [40,41]. In HK-2 cells, anserine has a higher antioxidative capacity than carnosine [15]. These activities are mediated via hormetic processes involving Nrf2, Sirt-1, Trx, Hsp70, the glutathione system [12,17,41] and modulation of the nitric oxide formation and metabolism [42,43]. Anserine and carnosine increase HSP70 expression depending on the cell type. Carnosine increases HSP70 in podocytes [14], independent of glucose concentrations, but not in HK-2 cells. Anserine increases HSP70 in HK-2 cells in the presence of oxidative and glucose stress [14,15], and incubation with carnosine and anserine are well tolerated by cells [13]. Since H 2 S increases myocardial and cerebral HSP70 [19,34], we investigated a putative interaction of anserine and carnosine with H 2 S synthesis in vitro and demonstrated that carnosine dose-dependently increases H 2 S synthesis capacity in HK-2 cells, and anserine increases this capacity in capillary, aortic and umbilical vein endothelial cells, i.e., in endothelial cells derived from capillaries and large arterial and venous vasculature. The induction of H 2 S synthesis in endothelial cells by anserine was entirely HSP70-dependent. Hsp70-KO in capillary endothelial cells abolished the stimulatory effect of anserine on H 2 S formation, and transfection of Hsp70 into Hsp70-KO cells restored it. In line with this, in Hsp70-KO mice, kidney H 2 S synthesis was reduced by 50%. In contrast, the carnosineinduced upregulation of H 2 S synthesis in HK-2 cells should be HSP70-independent, since carnosine does not activate Hsp70 in HK-2 cells [15].
We then studied the impact of H 2 S within the highly regulated metabolism of anserine and carnosine via CN1. Carnosine is provided by nutrition and tissue carnosine synthase (Carns1), but the impact of the latter is uncertain. In Carns1-KO mice, brain and muscle are carnosine-deficient, and stores can be replenished by oral intake [44], but the Carns1-KO had no impact on kidney and brain markers of carbonyl and oxidative stress in healthy and diabetic mice [45]. In contrast, anserine and carnosine are rapidly degraded by CN1, and kidney tissue CN1 activity in healthy and in type 2 diabetic mice is correlated with kidney anserine and carnosine concentrations and the tissue carbonyl and oxidative stress level [46]. These findings suggest a gatekeeping role and protective function of CN1 for kidney histidine-containing dipeptide concentration. In line with this, diabetic patients, who are homozygous carriers of the Cndp1 gene variant CTG5, have significantly lower serum CN1 concentrations and activity, and its concentrations independently predict eGFR [47]. Previous enzyme kinetic studies and molecular dynamic simulations revealed inhibition of carnosine degrading CN1 activity by competitive inhibition with anserine [48], and by thiol-containing compounds due to allosteric interactions [49]; the latter interaction might be the underlying mechanism of the CN1 inhibition by H 2 S. In the same direction, we now demonstrate dose-dependent CN1 inhibition by H 2 S donor Na 2 S, suggesting interaction of anserine and of H 2 S in a positive feedback loop, i.e., mutual reinforcement of both protective mechanisms.
To demonstrate the impact of the interaction of anserine and carnosine with H 2 S synthesis capacity in vivo, we studied Cndp1-KO mice. These mice exhibit a kidney-selective, age-and gender0dependent 2-to 9-fold increase in kidney tissue anserine and carnosine concentrations, and kidney function is unaltered [32]. The increased kidney anserine and carnosine concentrations could be reconfirmed, underlying molecular mechanisms of sexspecific differences, such as the influence of estrogens, have not yet been studied in detail. In these mice, kidney anserine and carnosine concentrations correlated with the kidney tissue H 2 S synthesis capacity, underpinning the significance of the mechanistic interactions demonstrated in vitro in the in vivo setting. Experimental studies demonstrated the upregulation of vascular endothelial cell H 2 S synthesis by estrogen [50]. H 2 S has been previously recognized as a toxic gas, but has emerged as an important gaseous signaling molecule, and administration of the H 2 S donor thiosulfate (2 × 15 g) is well tolerated by humans [51]. The action of H 2 S involves a variety of molecular mechanisms, such as activation of PI3K/Akt/eNOS pathway, suppressing ferroptosis or the antioxidant effect mediated by Nrf2 signaling [52][53][54]. In line with the in vitro findings of carnosine inducing H 2 S synthesis only in proximal tubular epithelial cells, but anserine in all cell lines studied, correlations of kidney H 2 S synthesis capacity were higher with tissue anserine than carnosine concentrations. Of note, short-term incubation of kidney tissue with anserine did not increase H 2 S synthesis capacity, suggesting slower anserine-induced actions than observed in vitro, but unspecific alterations in the ex vivo tissue homogenate devoid of blood supply cannot be excluded.
The impact of diabetes mellitus was demonstrated in type 1 diabetic Cndp1-KO mice. While kidney anserine and carnosine concentrations were in a similar range in diabetic and non-diabetic WT mice, concentrations of both histidine-containing dipeptides were 3-fold lower in the kidney of diabetic versus non-diabetic Cndp1-KO mice. In the diabetic mice, H 2 S formation rate was not correlated with the low kidney anserine and carnosine. Thus, the type 1 diabetes mellitus Cndp1-KO model does not allow for firm conclusions on their interaction. In type 2 diabetes db/db mice of similar age, kidney anserine and carnosine concentrations were also reduced [46]. Plasma and kidney H 2 S levels have repeatedly been reported to be low in diabetes mellitus [20], except for one study with double transgenic Balb/c mice sacrificed at the age of 13.5 weeks that had increased kidney H 2 S [55], a finding which could be reproduced here in male, but not in female STZ mice with a similar duration of diabetes. In female STZ Cndp1-KO mice, H 2 S formation was lower compared to female non-diabetic KO littermates, presumably due to the 3-fold lower kidney anserine and carnosine concentrations.
Our findings do not exclude beneficial effects of pharmacological doses of anserine and carnosine in diabetes via induction of H 2 S. Numerous studies in rodents demonstrated kidney protection in diabetic mice by carnosine administration [5][6][7]56,57]. Only two studies reported on the administration of anserine [15,58], even though nephroprotective effects of carnosine may at least in part be exerted by methylation to anserine [59]. We recently demonstrated several-fold increased kidney anserine concentrations after oral carnosine supplementation in healthy and db/db mice for four weeks [60]. None of these studies evaluated the interaction of therapeutic anserine or carnosine supplementation with H 2 S metabolism in diabetic mice, but in view of our finding, it deserves validation in animal models and in clinical trials. Likewise, our findings are of interest with regard to the potential therapeutic impact of combined treatment such as in patients with ischemic heart disease treated with the H 2 S donor thiosulfate. Anserine may not only exert direct beneficial effects, but also increase H 2 S availability.
A limitation of our study is the measurement of H 2 S synthesis capacity, based on an endpoint quantification with Ag 2 S, providing a high specificity and reproducibility with reasonable sensitivity. This, however, does not provide information on actual cell and tissue H 2 S concentrations. Current methods to quantify H 2 S concentrations comprise colorimetric assays, gas chromatography, fluorescence probes and electrochemical technics, and the advantages and limitations have recently been summarized [61]. Future studies combining different methods may further increase insights and validity of our findings. In the same direction, we demonstrated the synthesis activity of CBS and CSE in murine kidney tissue, and the functional role of HSP70, but not the specific mode of action of HSP70 on enzyme expression and activity.

Conclusions
In conclusion, we demonstrate a novel mechanism of action of the histidine-containing dipeptides anserine and carnosine, i.e., the induction of H 2 S synthesis in proximal tubular epithelial cells and capillary, venous and aortic endothelial cells, which in endothelial cells is exerted via HSP70. The in vivo relevance of these interactions is demonstrated in Hsp70and Cndp1-KO mice, but awaits clarification in diabetes mellitus and ischemic heart disease. The positive feedback interaction of H 2 S, inhibiting anserine and carnosine degradation by CN1 together with the H 2 S synthesis-inducing effect of both dipeptides, should enhance the efficacy of therapeutic interventions with anserine, carnosine and with H 2 S donors.