Biliverdin Reductase-A integrates insulin signaling with mitochondrial metabolism through phosphorylation of GSK3β

Abstract

Brain insulin resistance links the failure of energy metabolism with cognitive decline in both type 2 Diabetes Mellitus (T2D) and Alzheimer's disease (AD), although the molecular changes preceding overt brain insulin resistance remain unexplored. Abnormal biliverdin reductase-A (BVR-A) levels were observed in both T2D and AD and were associated with insulin resistance. Here, we demonstrate that reduced BVR-A levels alter insulin signaling and mitochondrial bioenergetics in the brain. Loss of BVR-A leads to IRS1 hyper-activation but dysregulates Akt-GSK3β complex in response to insulin, hindering the accumulation of pGSK3β^S9 into the mitochondria. This event impairs oxidative phosphorylation and fosters the activation of the mitochondrial Unfolded Protein Response (UPRmt). Remarkably, we unveil that BVR-A is required to shuttle pGSK3β^S9 into the mitochondria. Our data sheds light on the intricate interplay between insulin signaling and mitochondrial metabolism in the brain unraveling potential targets for mitigating the development of brain insulin resistance and neurodegeneration.

Keywords: Biliverdin reductase-A; Brain insulin resistance; GSK3β; Mitochondrial metabolism; Mitochondrial unfolded protein response; Oxidative stress.

Graphical abstract

Fig. 1

Brain insulin signaling alterations are associated with an impaired mitochondrial metabolism in GK rats. Changes of BVR-A and of the insulin signaling pathway's proteins levels, i.e., IR, IRS1, Akt and GSK3β were evaluated in the hippocampus of Wistar (n = 6–9 independent samples) and GK (n = 7–8 independent samples) rats. (a) Representative Western blot images and densitometric evaluation of (b) BVR-A protein levels (p = 0.006, GK vs Wistar); (c) IR protein levels (p = 0.38, GK vs Wistar), and IR activation (evaluated as pIR^{Y1158/1162/1163}/IR ratio; p = 0.83, GK vs Wistar); (d) IRS1 protein levels (p = 0.51, GK vs Wistar), IRS1 inhibition (evaluated as S³⁰⁷/IRS1 ratio; p = 0.04, GK vs Wistar) and IRS1 activation (evaluated as Y⁶³²/IRS1 ratio; p = 0.35, GK vs Wistar); (e) IRS1 activation state index (evaluated as IRS1 Y⁶³²/S⁶³⁶ ratio; p = 0.03, GK vs Wistar); (f) Akt protein levels (p = 0.01, GK vs Wistar), and Akt activation (evaluated as S⁴⁷³/Akt ratio; p = 0.12, GK vs Wistar); (g) GSK3β protein levels (p = 0.03 GK vs Wistar) and GSK3β inhibition (evaluated as S⁹/GSK3β ratio; p = 0.01 GK vs Wistar). (h) Representative Western blot images and (i) densitometric evaluation of the Akt/GSK3β complex isolated through the immunoprecipitation assay from the hippocampus of Wistar (n = 3 independent samples) and GK (n = 4 independent samples) rats; p = 0.03 GK vs Wistar. Lanes description: lanes 1 to 3: Wistar; lanes 4 to 7: GK; Lane 8: empty; lane 9: beads alone (B); lane 10; beads plus the anti-GSK3β primary antibody, but no sample (B+A); lane 11: empty; lane 12: supernatant collected following beads magnetization; and lane 13: Input. (j) A significant association was found between the Akt/GSK3β complex, and the respective BVR-A protein levels evaluated in the same samples (Pearson r = 0.6; p = 0.02). (k–o) Representative traces showing oxygen consumption rates (OCR) in Wistar and GK rats. Arrows indicate the addition of substrates and/or inhibitors (1 mM NADH, 5 mM Succinate, 2.6 μM Antymicin, 2.13 mM Ascorbate, 0.8 mM N,N,N′,N′-Tetramethyl-p-phenylenediamine (TMPD) and 1 mM Sodium Cyanide (CN). Quantification of the OCR, normalized for proteins content (pmol O₂/s/mg proteins) after the addition of (m) NADH (Complex I, p = 0.03 GK vs Wistar); (n) Succinate (Complex I+II, p = 0.06 GK vs Wistar); (o) Antimycin to block Complex III, Ascorbate and TMPD (Complex IV, p = 0.2 GK vs Wistar). (p) Representative Western blot images and densitometric evaluation of (q) Complex I (subunit NDUFB8, p = 0,02 GK vs Wistar), (r) Complex II (subunit SDHB, p = 0.01 GK vs Wistar), (s) Complex III (subunit UQCRC2, p = 0.04 GK vs Wistar), (t) Complex IV (subunit MTCO1, p = 0.007 GK vs Wistar), and (u) Complex V (subunit ATP5A, p = 0.2 GK vs Wistar). The densitometric values are given as percentage of Wistar set as 100 %. Data are presented as means ± SEM. Statistical significance was determined using Student t-test analysis (*p < 0.05, **p < 0.01).

Fig. 2

UPRmt is activated as an adaptive response to face alterations of insulin signaling-mitochondria axis in the brain. (a) Densitometric evaluation of markers, i.e., protein carbonyls (PC) (p = 0.15, GK vs Wistar), 3-nitrotyrosine (3-NT) (p = 0.03, GK vs Wistar) and proteins-bound 4-hydroxyl-2-nonenals (HNE) (p = 0.0051, GK vs Wistar) evaluated in the hippocampus of Wistar (n = 8 independent samples) and GK (n = 8 independent samples) rats. (b) Densitometric evaluation of 4-HNE (p = 0.03, GK vs Wistar) and 3-NT (p = 0.02, GK vs Wistar) evaluated in hippocampal mitochondrial extracts isolated from Wistar (n = 4 independent samples) and GK (n = 4 independent samples) rats. Changes of the Unfolded Protein Response (UPRmt) proteins [Atf5, Grp75, Hsp60, Atf4, Chop, Sirt3, Pgc1α and Tfam] and of the antioxidant enzymes [Catalase, Gpx and Sod-2] were evaluated in the hippocampus of Wistar (n = 8–9 independent samples) and GK (n = 7–8 independent samples) rats. (c) Representative Western blot images and densitometric evaluation of (d) Atf5 (p = 0.01, GK vs Wistar), (e) Grp75 (p = 0.002, GK vs Wistar), (f) Hsp60 (p = 0.04, GK vs Wistar), (g) Atf4 (p = 0.04, GK vs Wistar), (h) Chop (p = 0.04, GK vs Wistar), (i) Sirt3 (p = 0.03, GK vs Wistar), (j) Pgc1α (p = 0.01, GK vs Wistar) and (k) Tfam (p = 0.01, GK vs Wistar). (l) Representative Western blot images and densitometric evaluation of (m) Catalase (p = 0.03, GK vs Wistar) (n) Gpx (p = 0.33, GK vs Wistar) and (o) Sod-2 (p = 0.07, GK vs Wistar). Values are given as percentage of Wistar set as 100 %. Data are presented as means ± SEM. Statistical significance was determined using Student t-test analysis (*p < 0.05, **p < 0.01) (p) Wistar and GK hippocampal samples were tested by label-free quantitative proteomics approach and analyzed through the Ingenuity Pathway Analysis (IPA) software (QIAGEN IPA). Mass spec identified 3461 proteins and isoforms listed in Supplementary material. Volcano plot indicating differentially expressed proteins in Wistar and GK rats. Proteins showing a fold-change significantly (p < 0.05) reduced or elevated in GK vs Wistar rats are shown in blue. Highlighted in yellow proteins relevant for the current study. (q) Bubble diagram of Canonical Pathways across the entire dataset of proteins. The colors indicate the z-score (see legend at top right), while the size of the bubbles increases with the number of overlapping proteins. The IPA analysis display a negative regulation of pathways associated with Oxidative phosphorylation, synaptogenesis signaling pathway and chaperone mediated autophagy. A positive regulation of pathways related to mitochondrial dysfunctions and sirtuin signaling and axonal guidance signalling was observed. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Fig. 3

Loss of BVR-A links alterations of brain insulin signaling and mitochondrial dysfunctions. Changes of Akt and GSK3β activation were evaluated in the hippocampus of WT (n = 4 independent samples) and BVR-A^−/− (n = 4 independent samples) mice. (a) Representative Western blot images and densitometric evaluation of (b) Akt protein levels (p = 0.03, WT vs BVR-A^−/−) and Akt activation (evaluated as S⁴⁷³/Akt ratio; p = 0.09, WT vs BVR-A^−/−), (c) GSK3β protein levels (p = 0.65, WT vs BVR-A^−/−) and GSK3β inhibition (evaluated as S⁹/GSK3β ratio; p = 0.01, WT vs BVR-A^−/−). (d and e) Representative traces showing oxygen consumption rates (OCR) in WT and BVR-A^−/− mice samples normalized for proteins content. Arrows indicate the addition of substrates and/or inhibitors (1 mM NADH, 5 mM Succinate, 2,6uM Antymicin, 2,13 mM Ascorbate, 0,8 mM N,N,N′,N′-Tetramethyl-p-phenylenediamine (TMPD) and 1 mM Sodium Cyanide (CN). Quantification of OCR normalized for protein content (pmol O₂/s/mg proteins) after the addition of (f) NADH (Complex I, p = 0.01, WT vs BVR-A^−/−), (g) Succinate (Complex I+II, p = 0.03, WT vs BVR-A^−/−), (h) Antimycin to block Complex III, Ascorbate, and TMPD (Complex IV, p = 0.2, WT vs BVR-A^−/−). Values are given as percentage of WT set as 100 %. Data are presented as means ± SEM. Statistical significance was determined using Student t-test (*p < 0.05). (i) WT and BVR-A^−/− hippocampal samples were tested by label-free quantitative proteomics approach and analyzed through the Ingenuity Pathway Analysis (IPA) software (QIAGEN). Mass spec analysis identified 989 proteins and isoforms listed in Supplementary material. (i) Volcano plot indicating differentially expressed proteins in WT and BVR-A^−/− mice. Proteins showing a fold-change significantly (p < 0.05) reduced or elevated in WT and BVR-A^−/− mice are shown in green. Highlighted in yellow proteins relevant for the current study. (j) Bubble diagram of Canonical Pathways across the entire dataset of proteins. The colors indicate the z-score (see legend at top right), while the size of the bubbles increases with the number of overlapping proteins. The IPA analysis displays a negative regulation of pathways associated with oxidative phosphorylation, glycolysis and chaperone mediated autophagy. A positive regulation of pathways related to mitochondrial dysfunctions, sirtuin signaling, EIF-2 signalling, NRF2- mediated response, mTOR, p70S6K and PI3K/Akt signalling was observed. (k) Venn diagram illustrating the outcomes of a comparative analysis conducted on label-free proteomic datasets obtained from GK vs Wistar rats and BVR-A^−/− vs WT mice. The analysis highlights a shared set of 375 proteins exhibiting a consistent pattern of alteration across both groups. This subset constitutes 27.7 % of the total identified proteins. (i) Comparative analysis between the Canonical pathways identified by IPA in rats (violet) and mice (blue) groups of analysis. Data show the common alteration of mitochondria, sirtuin signaling, synaptogenesis, among others. (m) Biological outcomes identified by IPA analysis in Wistar/GK rats (violet) and WT/BVR-A^−/− (blue) groups. Data show the common alteration of long-term potentiation of hippocampus, assembly of axon initial segments and morphology of neurons, among others. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Fig. 4

Reduced BVR-A protein levels impairs mitochondrial bioenergetic in response to insulin. Changes of BVR-A and of the insulin signaling pathway's proteins levels, i.e., IRS1, Akt and GSK3β were evaluated in Ctr and siRNA-treated cells stimulated with 100 nM insulin at different time points 15’, 30’, 60’ and 120’ minutes (n = 5–6 independent cultures/group). (a) Representative Western blot images and densitometric evaluation of (b) IRS1 activation evaluated as IRS1^Y632 (*p = 0.03, Ctr vs 15’ Ins; *p = 0.0012, Ctr vs 30’ Ins; *p = 0.000016, Ctr vs 120’ Ins; °p = 0.01, 30’Ins vs 30’ siRNA+Ins) (c) IRS1 inhibition evaluated as IRS1^S307 (°p = 0.03, 60’ Ins vs 60'siRNA+Ins; °p = 0.0078, 120’ Ins vs 120'siRNA+Ins) (d) Akt activation (evaluated as S⁴⁷³/Akt ratio; *p = 0.02, Ctr vs 15’ Ins; *p = 0.003, Ctr vs 30’ Ins; *p = 0.01, Ctr vs 60’ Ins; *p = 0.01, Ctr vs 120’ Ins; °p = 0.01, 30’ Ins vs 30'siRNA+Ins; °p = 0.03, 120’ Ins vs 120'siRNA+Ins) (e) GSK3β inhibition (evaluated as S⁹/GSK3β ratio; *p = 0.007, Ctr vs 15’ Ins; *p = 0.003, Ctr vs 30’ Ins; *p = 0.01, Ctr vs 120’ Ins; °p = 0.001, 30’ Ins vs 30’ siRNA+Ins; °p = 0.001, 120’ Ins vs 120'siRNA+Ins). Values are given as percentage of Ctr 0’ set as 100 %. Data are presented as means ± SEM (One-way ANOVA with Fisher's LSD test). (f–i) Bioenergetic profile evaluated by seahorse in Ctr and siRNA-treated SHSY-5Y cells stimulated with 100 nM Insulin for 30’ and 120’ minutes. (f) ECAR measured in Ctr cells (Ctr Baseline vs 30’Ins Baseline ***p = 1.76E-06; Ctr Baseline vs 120’Ins Baseline °°°p = 1.82E-05; Ctr Stressed vs 30’Ins Stressed ***p = 2.03E-05; Ctr Stressed vs 120’Ins Stressed °°°p = 6.18E-06) (g) OCR measured in Ctr cells (Ctr Baseline vs 30’Ins Baseline *p = 0.022; Ctr Baseline vs 120’Ins Baseline °°p = 0.0040; Ctr Stressed vs 30’Ins Stressed *p = 0,018; Ctr Stressed vs 120’Ins Stressed °°p = 0.0051). (h) ECAR measured in siRNA-treated cells (siRNA Baseline vs siRNA 30’Ins Baseline **p = 0.0057; siRNA Baseline vs siRNA 120’Ins Baseline p = 0.087; siRNA Stressed vs siRNA 30’Ins Stressed **p = 0.00057; siRNA Stressed vs siRNA 120’Ins Stressed p = 0.075). (i) OCR measured in siRNA-treated cells (siRNA Baseline vs siRNA 30’Ins Baseline OCR p = 0.34; siRNA Baseline vs siRNA 120’Ins Baseline p = 0.97; siRNA Stressed vs siRNA 30’Ins Stressed p = 0.34; siRNA Stressed vs siRNA 120’Ins Stressed p = 0.84). Arrows indicate the addition of inhibitors: Olygomycin, FCCP and Rotenone. Values are given as means ± SEM of 2 independent experiment with 8 replicates each (one-way ANOVA followed by Holm-Sidak post-hoc comparison test). Mitochondrial superoxide production was detected with MitoSOX. (j) Representative images of MitoSOX in Ctr and siRNA-treated cells stimulated with 100 nM for 30’ and 120’. MitoSOX in red and DAPI for nuclei visualization in blue. (k). Immunofluorescence quantification of MitoSOX red was measured by ImageJ software (p = 0.0039, 120’ Ins vs 120'siRNA+Ins, p = 0.001, 30'siRNA+Ins vs 120’ siRNA+Ins). Values are given as percentage of Ctr 0’ set as 100 %. Data are presented as means ± SEM (one-way ANOVA with Fisher's LSD test). (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Fig. 5

Reduced BVR-A protein levels trigger UPRmt activation in SHSY-5Y in response to insulin Changes of BVR-A and Unfolded Protein Response (UPRmt) proteins [Atf5, Grp75, Hsp60, Atf4, Chop] were evaluated in Ctr and siRNA-treated cells stimulated with 100 nM insulin for 30’ and 120’. Where indicated cells were pre-treated with 10 mM lithium chloride (LiCl) for 24 h before insulin administration (3–5 independent cultures/group) (a) Representative Western blot and densitometric evaluation of (b) Atf5 (p = 0.03, 120’Ins vs 120'siRNA+Ins) (c) Grp75 (p = 0.01, 30’Ins vs 30'siRNA+Ins; p = 0.04, 120’ Ins vs 120'siRNA+Ins; p = 0.0049, 30'siRNA+Ins vs 30'siRNA+Ins+ Li; p = 0.0040, 120'siRNA+Ins vs 120'siRNA+Ins+Li) (d) Hsp60 (p = 0.04, 30’ Ins vs 30'siRNA+Ins) (e) Atf4 (p = 0.03, 120’Ins vs 120'siRNA+Ins) (f) Chop. All densitometric values are given as percentage of Controls cells set as 100 %. Data are presented as means ± SEM (One-way ANOVA with Fischer LSD post-hoc test). (g,h) Bioenergetic profile evaluated by seahorse in siRNA-treated cells pretreated with 10 mM LiCl and stimulated with Insulin for 30’ and 120’ (g) ECAR and (h) OCR measured in siRNA-treated cells+Li. Arrows indicate the addition of inhibitors: Olygomycin, FCCP and Rotenone. Values are given as means ± SEM of 2 independent experiment with 8 replicates each (one-way ANOVA followed by Holm-Sidak post-hoc comparison test).

Fig. 6

BVR-A is a shuttle for GSK3β into the mitochondria in response to insulin Changes of BVR-A and GSK3β were evaluated in Ctr and siRNA-treated cells stimulated with 100 nM insulin for 30’ and 120’. Where indicated cells were pre-treated with 10 mM lithium chloride (LiCl) for 24 h before insulin administration (3–5 independent cultures/group) (a) Representative confocal immunofluorescence images of BVR-A (green), Mitotracker (cyan) and GSK3β^S9 (magenta) (scale bar: 10 μm, DAPI for nuclei visualization in grey) in Ctr and siRNA-treated cells stimulated with 100 nM insulin for 30’ and 120’. Where indicated cells were pre-treated with 10 mM lithium chloride (LiCl) for 24 h before insulin administration. (b) Quantification of GSK3β^S9 and Mitotracker signals colocalization at single cell level in Ctr and siRNA-treated SHSY-5Y cells upon the treatments described above (p = 2E-07, Ctr vs 30’Ins; p = 8E-07, Ctr vs 120’Ins; p = 7E-12, 30’Ins vs 30'siRNA+Ins; p = 3E-12, 120’Ins vs 120'siRNA+Ins; p = 0.00007, 30’Ins vs 30'siRNA+Ins+Li; p = 0.0012, 120’Ins vs 120'siRNA+Ins+Li; p = 0.00001 Ctr vs Ctr+Li). For the analysis relative to siRNA-treated cells only silenced cells were took into consideration (c) Quantitative analysis showing the colocalization of BVR-A and Mitotracker signals at single cell level in Ctr cells upon the different treatments (p = 2E-06, Ctr vs 30’Ins; p = 8E-08, Ctr vs 120’Ins; p = 0.008 Ctr vs Li). Values are given as mean ± SEM (n = 6–11 cells from each condition from 3 independent experiments; One-way ANOVA with Sidak's multiple comparison test). (d) Representative Western blot images and (e) densitometric evaluation of the BVR-A/GSK3β complex isolated through the immunoprecipitation assay from SHSY-5Y mitochondrial extracts following stimulation with 100 nM insulin for 30’. Validation of the purity of the subcellular fractions was determined by examining Complex 1. To achieve a sufficient amount of mitochondrial proteins for the IP analyses extracts were pooled, resulting in a sample size of n = 2 for each condition. IP lanes description: lanes 1 and 2: SHSY-5Y stimulated with 100 nM insulin; lanes 3 and 4: CTR; lane 5: empty; lane 6: beads alone, lane 7: beads plus the anti-BVR-A primary antibody but no sample (B+A); lane 8: empty; lanes 9 to 12: Supernatant collected following beads magnetization; lane 13 and 14: empty; lane 15: Recombinant Human BVR-A protein used as positive control. Changes of BVR-A and GSK3β^S9 levels evaluated in the hippocampal mitochondrial extract from Wistar (n = 3 independent samples) and GK (n = 4 independent samples) rats. (f) Representative Western blot images and densitometric evaluation of (g) GSK3β^S9 (p = 0.03, GK vs Wistar) and (h) BVR-A (p = 0.04, GK vs Wistar). Values are given as percentage of Wistar set as 100 %. Data are presented as means ± SEM (Student t-test). Statistical significance was determined using Student t-test analysis (*p < 0.05, **p < 0.01). (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Fig. 7

Loss of BVR-A impairs insulin signaling activation and GSK3β translocation into the mitochondria in the brain in response to insulin To confirm the activation of the insulin signaling in the brain Akt activation was evaluated in the hippocampus of WT and BVR-A^−/− mice treated with intranasal insulin for 30’ (n = 3 independent samples/group). (a) Representative Western blot of Akt^S473, Akt and densitometric evaluation of (b) Akt activation (evaluated as S⁴⁷³/Akt ratio; p = 0.0096, WT vs WT+Ins; p = 0.0007, WT+Ins vs BVR-A−/− + Ins; p = 0.03, WT vs BVR-A^−/−) (c) Representative Western blot images and (d) densitometric evaluation of the Akt/GSK3β complex isolated through the immunoprecipitation assay from the hippocampus from WT and BVR-A^−/− mice treated with intranasal insulin for 30’ (n = 5 independent samples/group). Lanes description: lanes 1 to 3: WT; lanes 4 to 6: WT stimulated with Intranasal Insulin; lanes 7 to 9: BVR-A^−/− stimulated with Intranasal insulin; lane 10: empty; lane 11: beads alone, lane 12: beads plus the anti-GSK3β primary antibody but no sample (B+A); lane 13: empty; lane 14: Supernatant collected following beads magnetization; and lane 14: Input. (p = 0.0045, WT vs WT+Ins; p = 0.0098, WT+Ins vs BVR-A^−/− +Ins). Evaluation of BVR-A, GSK3β and proteins of UPRmt i.e., Grp75, Hsp60 and Atf5 following subcellular fractionation of hippocampal samples collected from WT and BVR-A^−/− mice treated with intranasal insulin for 30’ (n = 3 independent samples/group). (e) Representative Western blot images and densitometric evaluation of (f) BVR-A (p = 0.03, WT vs WT+Ins), (g) GSK3β inhibition (evaluated as S⁹/GSK3β, p = 0.000038 WT vs WT+Ins; p = 0.000008, WT+Ins vs BVR-A^−/− + Ins; p = 0.02, WT vs BVR-A^−/− + Ins), (h) Grp75 (p = 0.0098, WT vs BVR-A^−/− + Ins; p = 0.04, WT+ Ins vs BVR-A^−/−), (i) Hsp60 (p = 0.02, WT+Ins vs BVR-A^−/− + Ins), (j) Atf5 (p = 0.01, WT vs BVR-A^−/− + Ins; p = 0.06, WT+Ins vs BVR-A^−/−). Purity of the subcellular fractions was determined by examining Complex I in the mitochondrial fraction and Polimerase II in the nuclear fraction by Western blot analysis. The densitometric values are given as percentage of WT set as 100 %. Data are presented as means ± SEM. One-way ANOVA with Fisher's LSD test. (k) Representative Western blot images and (l) densitometric evaluation of the BVR-A/GSK3β complex isolated through the immunoprecipitation assay from hippocampal mitochondrial extracts from WT mice treated with intranasal insulin for 30’. To achieve a sufficient amount of mitochondrial proteins for the IP analyses, hippocampal-derived extracts were pooled, resulting in a sample size of n = 2 for each condition. Lanes description: lanes 1 and 2: WT; lanes 3 and 4: WT stimulated with Intranasal Insulin; lane 5: empty; lane 6: beads alone, lane 7: beads plus the anti-GSK3β primary antibody but no sample (B+A); lane 8: empty; lanes 9 to 12: Supernatant collected following beads magnetization; lane 13: Input and lane 14: BVR-A−/− mice.

Fig. 8

UPRmt activation in T2D and AD Changes of BVR-A, GSK3β inhibition and of the mitochondrial unfolded protein response (UPRmt) markers, i.e., Atf5, Grp75, Hsp60 evaluated in PBMC isolated from healthy (Ctr = 9) and T2D donors (n = 18). (a) Representative Western blot images of BVR-A, Grp75, Hsp60, Atf5, GSK3β and GSK3β^S9 and densitometric evaluation of (b) BVR-A and (c) Grp75 (d–j) Densitometric values of BVR-A, Grp75, Hsp60, Atf5, GSK3β and GSK3β^S9 are reported following categorizing our samples based on high and low Grp75 levels in T2D subjects. (d) BVR-A (p = 0.0002, Ctr vs High; p = 0,000019, High vs Low); (e) Grp75 (p = 0.0003, Ctr vs High; p = 0.0000011, High vs Low); (f) Hsp60 (p = 0.08, Ctr vs High; p = 0.0020 High vs Low); (g) Atf5; (h) GSK3β^s9 (p = 0.06, Ctr vs High; p = 0.02 Ctr vs Low); (i) GSK3β. Data are presented as means ± SEM. Statistical analysis has been performed on each experiment by using one-way ANOVA followed by Tukey's multiple comparisons test. (j) Principal component analysis (PCA) performed on the results collected in human samples from Ctr and T2D donors. Changes of BVR-A, GSK3β inhibition and of the mitochondrial unfolded protein response (UPRmt) markers, i.e., Atf5, Grp75, Hsp60 evaluated in the inferior parietal lobule (IPL) of Control (Ctr) (n = 8), amnestic mild cognitive impairment (MCI, n = 6) and Alzheimer's disease (AD, n = 8) subjects. (k) Representative western blot images and densitometric evaluation of (l) BVR-A protein levels (p = 0.01, Ctr vs MCI; p = 0.005, MCI vs AD); (m) Atf5 (p = 0.02, Ctr vs MCI; p = 0.0001, Ctr vs AD; p = 0.058 MCI vs AD); (n) Grp75; (o) Hsp60; (p) GSK3β^S9 (p = 0.07, Ctr vs MCI) and (q) GSK3β. Data are presented as means ± SEM (One-way ANOVA with Fisher's LSD test). (r) Principal component analysis (PCA) performed on the results collected in post-mortem brain samples from Ctr, amnestic MCI and AD subjects.

Fig. 9

Schematic representation of the proposed molecular mechanisms through which BVR-A links insulin signaling activation with mitochondrial metabolism. Under Physiological Conditions (Left Panel): insulin binding to the insulin receptor (IR) initiates the phosphorylation of the insulin receptor substrate (IRS1) and BVR-A at Y residues. Phosphorylated BVR-A acts as a serine/threonine/tyrosine (S/T/Y) kinase, targeting inhibitory sites (S) on IRS1 to prevent IRS1 hyper-activation in response to insulin. Downstream from IRS1, Akt activation takes place. At this level, BVR-A serves as a scaffold protein, facilitating the physical interaction between Akt and GSK3β, leading to GSK3β phosphorylation at S9. Then, the BVR-A/pGSK3βS9 complex translocates into the mitochondria, promoting oxidative phosphorylation (OXPHOS) and ATP production to meet cellular metabolic demands. Under pathological conditions (right panel): reduced insulin sensitivity, leads to reduced BVR-A protein levels with the aim to increase IRS1 activation (by reducing the BVR-A-mediated inhibitory effect). However, this event creates a dual challenge downstream from IRS1. Loss of BVR-A compromises Akt-mediated inhibition of GSK3β, hindering the translocation of pGSK3βS9 into the mitochondria. This disruption results in mitochondrial stress and reduced ATP production. The ensuing cellular stress activates the mitochondrial unfolded protein response (UPRmt), prompting the translocation of transcription factors such as ATF5, ATF4, and CHOP into the nucleus. These factors promote the transcription of genes involved in the cellular stress response, including Grp75, HSP60, Lonp1, Cat, and Sod2, among others. Upregulation of these proteins aims to restore mitochondrial function, enhance ATP production, reduce oxidative stress-induced damage to proteins and lipids, and support synaptic plasticity mechanisms. Arrows: activation; lines:inhibition. The scheme has been created with BioRender.com.