Mitochondrial cyclophilin D ablation is associated with the activation of Akt/p70S6K pathway in the mouse kidney

The mitochondrial matrix protein cyclophilin D (CypD) is an essential component of the mitochondrial permeability transition pore (MPTP). Here we characterized the effects of CypD ablation on bioenergetics in the kidney. CypD loss triggers a metabolic shift in Ppif−/− male and female mouse kidneys towards glycolysis and Krebs cycle activity. The shift is accompanied by increased glucose consumption and a transcriptional upregulation of effectors of glucose metabolism in the kidney. These included activation of Akt, AMPK (only in males) and p70S6K kinases. Gender specific differences between the Ppif−/− male and female mouse kidneys were observed including activation of pro-surviving ERK1/2 kinase and inhibited expression of pro-apoptotic and pro-fibrotic JNK and TGFβ1 proteins in Ppif−/− females. They also showed the highest expression of phosphorylated-ERK1/2 and Akt S473 proteins of all four investigated animal groups. Furthermore, Ppif−/− females showed higher lactate concentrations and ATP/ADP-ratios in the kidney than males. These metabolic and transcriptional modifications could provide an additional level of protection to Ppif−/− females. In summary, loss of mitochondrial CypD results in a shift in bioenergetics and in activation of glucose-metabolism regulating Akt/AMPK/p70S6 kinase pathways that is expected to affect the capability of Ppif−/− mice kidneys to react to stimuli and injury.

Western blot analysis. For Western blots, tissue extracts were loaded onto Biorad Bis-Tris HCl Criterion gels (various percentages). Proteins were separated using a Biorad Criterion electrophoresis system operating at 120 V and then transferred from the gel to an Immobillon-P membrane (200 mA, Millipore, Billerica, MA). Membranes were incubated with the primary antibody at 4 °C overnight, after blocking with 5% milk/2% BSA in PBS-Tween buffer. Antibodies used in this study included (the majority were raised in a rabbit and to a smaller amount in a mouse): PTEN (phosphatase and tensin homolog); protein phosphatase 2 A (PP2A); AKT unmodified and phosphorylated at Ser473 and Thr308; p70S6K (unmodified and phosphorylated at T389); p38 and p42/44 MAPK (ERK1/2) unmodified and phosphorylated at Thr180/Tyr182 and Thr202/Tyr204, respectively; JNK (unmodified and phosphorylated at Thr183/Tyr185); PAK2 (p21-activated kinase) unmodified and phosphorylated at Ser20; p70 S6 kinase unmodified and phosphorylated at Thr389 and Thr421/Ser424; β-actin (source of all antibodies above Cell Signaling Technology, Danvers, MA); TGFβ1 and TGFβ3 (transforming growth factor, Santa Cruz Biotechnology, Santa Cruz, CA). After membranes were washed three times, the secondary antibody (goat-anti rabbit and horse anti-mouse antibodies conjugated to horseradish peroxidase, Cell Signaling Technology) was added. Membranes were subsequently treated with Pierce SuperSignal West Pico or Femto Solution (Pierce) following the manufacturer's instructions. A UVP system (BioImaging Systems, Upland, CA) was used to detect the horseradish peroxidase reaction on the membrane. ImageJ software (NIH, Bethesda, MD) was used for quantitative densitometric analysis of select gel band intensities. Densitometry data were normalized to β-actin.
Quantification of high-energy phosphate metabolites in kidney tissues. To assess the effect of CypD ablation on energy metabolism, we determined high-energy phosphate metabolite concentrations in snap-frozen kidneys using a previously described assay 23 . An average of 100 mg kidney tissue was homogenized in a mortar grinder over liquid nitrogen and extracted with 6 mL ice-cold PCA (12%) as described previously 23 . The samples were centrifuged, the liquid phase removed and neutralized to a pH of 7.0-7.3 using KOH. To separate from perchlorate salts, the neutralized samples were centrifuged again, and the supernatant lyophilized overnight. Lyophilisates were then re-dissolved in 0.5 ml water and adjusted to pH 6.5. An Agilent series 1100 HPLC (Agilent Technologies, Palo Alto, CA) coupled to an API4000 triple quadrupole mass spectrometer (AB Sciex, Concord, ON) equipped with an electrospray ionization (ESI) source was employed for quantitation of nucleotide mono-, di-, triphosphates, FAD(H 2 ) and NAD(H). Energy charge was calculated as [ATP + (0.5 × ADP)]/ (ATP + ADP + AMP).
Quantification of Krebs cycle, and purine degradation metabolites in kidney tissues and urine. Citrate, succinate, glucose, α-ketoglutarate and lactate were quantified in PCA kidney tissue extracts (vide supra) and 1:20-diluted mouse urine samples (diluted with 40 nM NaOH aqueous buffer). The internal standard solution was added at a 25:1 ratio. It was prepared in 40 nM NaOH aqueous buffer and contained 200 µM d6-glucose and d4-succinate.
Scientific RepoRts | 7: 10540 | DOI:10.1038/s41598-017-10076-9 Kidney histology. For hematoxylin and eosin (H&E) staining, kidney tissue samples were fixed in 10% buffered formaldehyde and embedded in paraffin, incubated for 5 minutes in Harris hematoxylin solution and for 60 seconds in eosin solution. Sections were washed with plain water, differentiated in 1% hydrochloric acid (HCl) + 50% ethanol, and stain intensity was optimized in ammonia water. Finally, sections were rinsed in 70% ethyl alcohol and dehydrated in xylene solution. Semi quantitative scoring system. The slides were scanned using an Aperio Scanscope with Spectrum Software (v. 10.0.1346.1805). The images were then analyzed using Aperio Imagescope software (v. 9.1.19.1574; APERIO Technologies, Vista, CA). Inflammatory infiltration of the kidneys was assessed by measurement of the density of nuclei using color saturation. In addition, 25 high-powered fields were examined in the tubulo-interstitium of the cortex and outer medulla of each section. The kidneys were examined for inflammation, epithelial necrosis, loss of brush border, and tubular dilatation. Twenty five glomeruli were examined for histologic changes and for vascular congestion. Their sizes as well as the average area were determined for each slide. To determine the number of glomeruli, the number of glomeruli within a region of the cortex was counted. The results are reported as the number of glomeruli corrected for the area analyzed.
Statistical analysis. All numerical data are presented as mean ± standard deviation. One-way ANOVA with Tukey's post hoc test was used to determine differences among groups (either between the WT and Ppif−/− animals within the same gender and between genders). The level of significance was set at p < 0.05 for all tests (SPSS, version 24.0, IBM/SPSS, Armonk, NY).

Results
Kidney histology. Comparison of kidney histologies revealed no visible differences (Fig. 1A), with the only exception that the size (area) of glomeruli were slightly enlarged in Ppif−/− mice as compared to their wild-type counterparts (in females and males, respectively, Fig. 1B). At the same time, the number of glomeruli remained unchanged (Fig. 1C).

Glucose metabolism.
Metabolite changes in the kidney. Kidneys of male Ppif−/− animals showed a significantly higher energy charge as compared to their wild-type counterparts ( Fig. 2A). This augmented energy production was accompanied by increased glucose uptake and increased Krebs cycle activity, as suggested by the significantly higher levels of glucose, citrate and succinate in the kidney tissues of Ppif−/− males (Table 1). Interestingly, female Ppif−/− animals showed only significant increase in succinate when compared to WT females (Table 1). Their expression of glucose transporter Glut1, however, was the highest among all investigated animal groups (Fig. 2B), and so was the concentrations of the glycolysis product lactate (Table 1). Ppif−/− mice of both genders had higher kidney lactate concentrations as compared to their WT counterparts (Table 1).
In terms of gender differences, higher intracellular kidney glucose and citrate concentrations were detected in Ppif−/− males as compared to corresponding females (Table 1).
Metabolites excretion in the urine. Only urinary citrate concentrations were higher in Ppif−/− males as compared to WT (Table 2). Similarly, in Ppif−/− females, the urinary concentrations of Krebs cycle and glycolysis intermediates and products did not differ as compared to WT females ( Table 2).
Glucose Metabolism Regulation: Akt/mTOR Pathway. Activity of Akt kinase that serves as a glucose uptake and metabolism regulator was measured through its phosphorylation at two sites: S473 and T308. A significant increase in the phospho-Akt S473 to Akt ratio was observed in Ppif−/− animals (Fig. 3A), whereas no change in the activation of Akt at the T308 site was noted (Fig. 3B). Although no change in the phospho-Akt T308 to Akt ratio was observed, expression of its downstream target, phospho-p70S6 kinase (expressed as phospho-p70S6K/ p70S6K-ratio) was significantly higher in Ppif−/− animals (Fig. 3C). The expression of the energy regulator AMPK was significantly higher in kidneys of malePpif−/− animals (Fig. 3D). Interestingly, these increases of the phosphorylated AktS473, p70S6K and AMPK proteins, were not accompanied by changes in the expression of PPP2A phosphatases (Fig. 3E). Other than noted in heart or lung tissues 24 , but in alignment with the expression in aortic tissue 24 , naïve PTEN did not change in the kidneys of WT versus Ppif−/− animals (Fig. 3E).
In regards to gender differences, kidneys of Ppif−/− females contained higher levels of phospho-Akt S473 as compared to Ppif−/− males (Fig. 3A). In addition, both WT and Ppif−/− females expressed higher levels of phospho-Akt T308 as well as phospho-p70S6 kinase when compared to their respective male counterparts (Fig. 3B,D).
Kidneys of Ppif−/− females contained significantly higher levels of phospho-p42/44 MAPK, but lower levels of phospho-JNK MAPK as compared to Ppif−/− males (Fig. 4A,C). TGF-β signaling. As an interaction between high glucose concentration and the synthesis of TGFβ in renal cells 25,26 had been reported 24,25 , we examined the expression of this growth factor in kidneys of Ppif−/− animals. Expression of TGFβ1 and TGFβ3, two proteins within the TGFβ-family associated with the development of renal fibrosis 27,28 , remained unchanged between the Ppif−/− and WT animals (Fig. 5A). However, Ppif−/− females showed significantly lower expression of TGFβ1 as compared to the corresponding males (Fig. 5A). This observation seems interesting when considering the often slower rate of progression of renal disease in women 22 .
In addition to TGFβ, we examined the expression and activation status of the PAK2 kinase that are also involved in the development and progression of renal fibrosis. No change in the expression of unmodified or phosphorylated PAK2 protein was observed (Fig. 5B).

Discussion
As shown in the present study, ablation of CypD is associated with reorganization of energy pathways in the kidney towards a state of higher glucose utilization.
The kidney contributes to glucose homeostasis through processes of gluconeogenesis, glucose filtration, glucose reabsorption, and glucose consumption. Krebs cycle intermediates are important substrates for renal   as measured in the kidneys from female and male Ppif−/− (KO) and wild-type (WT) mice. Data is presented as average ± SD (n = 5-6; in µM/ mg protein). One-way ANOVA significances are presented below. Tukey post hoc significance levels are as follows: *p < 0.05, **p < 0.01, ***p < 0.001 for WT versus Ppif−/− animals (male and female), and # p < 0.05 and ## p < 0.01 for male versus female (Ppif−/− and WT, respectively). Abbreviations: α-KGD: α-ketoglutarate.  Table 2. Comparison of metabolite as measured in the urine from female and male Ppif−/− (KO) and wildtype (WT) mice. Data is presented as average ± SD (n = 7-12 in µM per mM creatinine). One-way ANOVA significances are presented below. Tukey post hoc significance levels are as follows: *p < 0.05 for WT versus Ppif−/− animals (male and female), and ## p < 0.01 for male versus female (Ppif−/− and WT, respectively). Abbreviations: α-KGD: α-ketoglutarate. metabolism as they account for 10-15% of oxidative metabolism in the kidney 29 . Interestingly, in vitro studies have shown that renal cells can be salvaged from hypoxia-reoxygenation-induced mitochondrial injury by the supplementation of Krebs cycle intermediates which mediate ATP production and prevent an energy deficit [30][31][32] . Furthermore, while the renal cortex uses fatty acid oxidation as the main source of energy, the renal medulla is mainly producing and taking up lactate 33 . Morphologically, kidneys lacking CypD did not look different from those containing CypD with the only exception that CypD ablation led to enlargement of glomeruli. Metabolically, kidneys from CypD knockout mice displayed higher glucose consumption, Krebs cycle and glycolytical activity and ATP production than kidneys from WT mice. In Ppif−/− males, increased ATP levels were accompanied by a higher energy charge, whereas in Ppif−/− females concentration of AMP also increased resulting in an unchanged energy charge. Consistent with our data, previous metabolic profiling studies suggested a switch away from fatty acid oxidation towards glycolytic and Krebs cycle metabolism in hearts, brains and livers of Ppif−/− animals 12,13,34 . In the mouse heart, genetic deletion of CypD correlated with elevated levels of mitochondrial Ca 2+ and led to alteration in branched chain amino acid metabolism, pyruvate metabolism and the Krebs cycle 13 . However, CypD deletion  did not reduce the activity of proteins involved in cellular respiration 13,35 . Yet, in primary hepatocytes and embryonic fibroblasts, CypD loss triggered a global compensatory shift towards glycolysis, with increased glucose consumption and higher ATP production 34 . The same metabolic switch was seen in our study.
We also observed an activation of AMPK, a crucial cellular energy sensor, in male CypD-deleted mouse kidneys. In females, the degree of activation was not as pronounced and did not reach statistical significance, as was the case with the energy charge. Once activated by impaired energy metabolism, AMPK promotes ATP production by increasing the activity or expression of proteins involved in catabolism while conserving ATP by switching off biosynthetic pathways 36 . This could mean that the CypD-deleted mouse kidney is reacting to an accumulation of mitochondrial Ca 2+ through a lower cell energy demand and redistribution of metabolic pathways, such as a reduction in fatty acid β-oxidation 13,34,37 , rather than stimulation of mitochondrial respiration 38,39 . The question if this state of metabolic alteration is ultimately causing mitochondrial stress while decreasing the kidney's metabolic reserves and rendering it more sensitive to decompensation even under subtle stress conditions, such as described for Ppif−/− heart 12 , remains to be investigated.
The alteration of kidney metabolism was associated with alterations in cellular signaling pathways. An upregulation of kinase cascades associated with insulin signaling including phosphorylated AMPK, Akt and mTOR was observed in kidneys of Ppif−/− mice. Previous studies had shown that Ppif−/− animals experience the expansion of insulin-producing β-cells and mild hyperinsulinemia 34 . Increased secretion of insulin could be the driving force behind the increase in Akt/mTOR and glycolytic activity, since the kidney is known to respond to insulin signaling 40,41 . Specifically, while no change in the phosphorylation of Akt at T308 occurred, kidneys of Ppif−/− mice had higher levels of Akt phosphorylated at S473. Downstream, these two sites have different targets: while Akt phosphorylated at S473 is a regulating member of the glucose and lipid metabolism 42,43 , Akt phosphorylated at T308 primarily leads to the activation of mTORC1, p70S6K, and protein synthesis 44,45 . Therefore, the observed activation of AktS473 could be responsible for the observed glucose metabolism shift and increase in glycolysis 42,43 . However, notwithstanding a lack of alteration in AktT308, an activation of p70S6 kinase in the kidneys of Ppif−/− mice was seen. This kinase has been shown to protect against I/R injury and to promote cell survival and protein production 46 .
The activation of Akt is tied to MAPK proteins including p38, p42/44 (ERK1/2) and JNK 47,48 . These MAPKs are regulated by stress or injury not only in the kidney but also other organs. ERK promotes cell survival while JNK and p38 lead to cell death. Some studies postulate that it may be the balance between the two (ERK versus JNK) that determines the fate of the cell 49 . Interestingly, while we did not observe a change in either of these MAPKs in kidneys of Ppif−/− males, Ppif−/− females expressed higher levels of phosphorylated ERK1/2 and lower level of activated JNK kinases as compared to their wild type counterparts. Ppif−/− females had the highest expression of phosphorylated ERK1/2 and Akt S473 proteins among all four investigated animal groups. Furthermore, expression of the glucose transporter Glut1 was the highest in the female Ppif−/− kidney, which metabolically also showed the highest lactate concentration. Based on these gender differences, it may be speculated that the CypD ablation is augmenting the protective effects of estrogen in the female kidney that is known to block inflammatory and apoptotic activities [50][51][52][53] . Increased expressions of phospho-AKT S473 and phospho-p70S6K and pAMPK (in males only) were observed in Ppif−/− mice. Additionally, Ppif−/− females (presented on the right side of the image) showed a reduced expression of TGFβ1 as well as higher expression of phosphorylated ERK1/2 and p38 proteins. ↑: increase compared to WT controls, ↓: decrease compared to WT controls. Please note that a direct link between the phosphorylation of Akt and phosphorylation of MAP kinases p38 and ERK1/2 is not fully confirmed in the here used mouse model. Please also note that, since our experiments were performed in the whole kidney, the exact contribution of each subcellular fraction to the observed changes requires further investigation.
With the close interaction between energy demand, supply and activity of protein kinases including Akt, AMPK and mTORC1/mTORC2 44 , the observed switch in energy metabolism following CypD ablation is not surprising. The reorganization of this central controlling hub that regulates cellular functions may also be responsible for the protective effects CypD ablation has during I/R injury or stroke 5, 12, 54-56 . In conclusion, it is reasonable to expect that the reorganization of cell signaling pathways and resulting metabolic changes following the loss of CypD as found in the present study (Fig. 6) leads to changes in the response of kidneys of Ppif−/− mice to stimuli and injury and that such responses may differ between genders.