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

Chiara Lanzillotta, Antonella Tramutola, Simona Lanzillotta, Viviana Greco, Sara Pagnotta, Caterina Sanchini, Silvia Di Angelantonio, Elena Forte, Serena Rinaldo, Alessio Paone, Francesca Cutruzzolà, Flavia Agata Cimini, Ilaria Barchetta, Maria Gisella Cavallo, Andrea Urbani, D. Allan Butterfield, Fabio Di Domenico, Bindu D. Paul, Marzia Perluigi, Joao M.N. DuarteEugenio Barone

Research output: Contribution to journalArticlepeer-review

3 Scopus citations

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 pGSK3βS9 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 pGSK3βS9 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.

Original languageEnglish
Article number103221
JournalRedox Biology
Volume73
DOIs
StatePublished - Jul 2024

Bibliographical note

Publisher Copyright:
© 2024 The Authors

Funding

To emphasize the role for BVR-A in the observed alterations, we performed similar analyses in the hippocampus of a knock-out mouse model for BVR-A (BVR-A\u2212/\u2212). We observed a significant reduction of IR activation along with a consistent increase of IRS1 inhibition, demonstrated by elevated S307/IRS1 ratio (Supplementary Figs. 2b and c). These findings support the development of brain insulin resistance in the hippocampus of BVR-A\u2212/\u2212 mice. Downstream in the pathway we observed that Akt activation (S473/Akt) did not show significant changes, while a consistent reduction of GSK3\u03B2 inactivation was observed in BVR-A\u2212/\u2212 when compared to WT mice (Fig. 3b and c). These observations strengthen the idea that loss of BVR-A impairs the Akt-mediated inhibition of GSK3\u03B2 in agreement with data collected in GK rats and with previous studies [24]. The reduction of GSK3\u03B2S9 phosphorylation parallels the impairment of mitochondrial activity, as highlighted by the significant reduction in Complex I (Fig. 3f) and Complex I+II (Fig. 3g) activities reported in BVR-A\u2212/\u2212 when compared to WT mice. No significant changes were observed for mitochondrial OXPHOS complexes levels (Supplementary Figs. 2d\u2013i). An increase of oxidative stress levels in BVR-A\u2212/\u2212 mice was also observed, as demonstrated by the accumulation of PC and 3-NT modifications (Supplementary Fig. 2j). We also measured the aforementioned UPRmt markers in this model. We found that proteins belonging to the canonical axis of the UPRmt were mainly upregulated. These include Hsp60, Chop, and the master regulator of UPRmt, Grp75, which showed a slight increase close to the significance (Supplementary Fig. 3 c,d and f). These results confirm the role for BVR-A in the observed alterations found on insulin signaling and mitochondrial metabolism in GK rats. We also acknowledge few differences between the two models likely since the complete loss of BVR-A in BVR-A\u2212/\u2212 mice might accelerate and worsen the development of brain insulin resistance (increased IRS1 inhibition) and of mitochondrial impairment, resulting in the accumulation of oxidatively damaged proteins (PC and 3-NT). To further explore whether loss of BVR-A also has a role in the alterations of intracellular signalling pathways regulating cell energy metabolism and neuronal plasticity mechanisms found altered in GK rats, hippocampal samples were tested by label-free quantitative proteomics. We identified 989 proteins and isoforms with significant differences in expression (Supplementary Table 2). The Volcano plot highlights that proteins of the UPRmt canonical axis are among the most significantly upregulated in BVR-A\u2212/\u2212 mice (Hsp60, Hsp90, Sod2, Grp75, Lonp1) while proteins associated with cell energy metabolism (ATPase inhibitor F1, ATP synthase, Pip5k1a, Pkm, Ldhb and Mdh2) were downregulated (Fig. 3i). Remarkably, the IPA analysis displays similar altered pathways as those observed in GK rats (Fig. 3j). In addition, the comparison between the GK vs Wistar rats and BVR-A\u2212/\u2212 vs WT mice datasets reveals the overlapped alteration of 375 proteins, which account for the 27.7 % of the total amount of identified proteins (Fig. 3k). The comparative analysis between the Canonical Pathways identified by IPA in both mice and rats, supports the primary common dysfunction of mitochondrial metabolism (e.g., oxidative phosphorylation, sirtuin signaling pathway and TCA cycle), proteostasis network (chaperone-mediated autophagy signaling pathway, sirtuins signaling pathway and Nrf-2-mediated antioxidant response) and neuronal plasticity mechanisms (synaptogenesis, SNARE signaling, Rho/Rac-mediated pathways, LTP and actin dynamics) (Fig. 3l). Furthermore, the analysis of the Downstream Effects related to observed altered pathways reports the common alteration of several processes associated with neuronal development, growth, morphogenesis and plasticity (Fig. 3m). Accordingly, BVR-A\u2212/\u2212 mice exhibited impaired behavioral functions (Supplementary Fig. 3j-m).To examine the relevance of this mechanism in humans, we evaluated the activation of UPRmt in peripheral blood mononuclear cells (PBMCs) isolated from healthy (Ctr) and T2D donors, focusing on the disposition of BVR-A and Grp75. In our initial analysis, we did not observe any statistically significant differences between Ctr and T2D groups (Fig. 8b and c). However, upon closer examination of Grp75 levels, a noteworthy pattern emerged, revealing the presence of two distinct subgroups. One subgroup exhibited higher Grp75 levels, while the other subgroup is characterized by lower Grp75 levels (Fig. 8c). Hence, we categorized our samples based on high and low Grp75 levels to gain additional insights from these two distinct subgroups. We found that individuals with higher Grp75 levels show reduced BVR-A levels compared to Ctr subjects, along with a nearly significant increase of Hsp60 but no changes of Atf5 in PBMCs (Fig. 8 d-g). Furthermore, pGSK3\u03B2S9 levels were significantly reduced in both subgroups, despite no changes observed for total GSK3\u03B2 protein levels (Fig. 8h and i). The multivariable regression model showed that, among patients with T2D, high levels of UPRmt proteins were associated with the concurrent use of multiple antidiabetic medications compared to those with low UPRmt protein levels [median (range) antidiabetic agents ongoing: high UPRmt: 2.5 (1\u20133) vs low UPRmt: 1 (1\u20132), p = 0.04]. This association was independent of gender, age, HbA1c values, BMI, and duration of diabetes (Table 5). This finding suggests that, to achieve similar glycemic control, these patients require targeting of multiple pathophysiological pathways. It could imply a more severe diabetic condition leading to greater cellular stress despite comparable disease duration. PCA analysis reveals distinct clusters for each of these 3 groups (Fig. 8j). Considering that T2D is a major risk factor for the development of AD, and that dysfunctional glucose metabolism, mitochondrial defects and insulin resistance are also key pathological hallmarks of AD brain, we performed the same analyses in post-mortem brain samples from MCI, AD and age-matched controls, to unravel whether AD development parallel alterations in UPRmt. We found that BVR-A protein levels are significantly reduced in MCI with respect to Ctr, while BVR-A levels are comparable to those of Ctr in AD subjects (Fig. 8l). However, an increase of 3-nitrotyrosine (3-NT) levels on BVR-A in both MCI and AD post-mortem brain was observed (Supplementary Fig. 7b), suggesting a consistent impairment despite unaltered protein levels in AD and in agreement with our previous work [55]. Alterations of BVR-A are associated with a significant increase of Atf5 both in MCI and AD brain, while no changes were observed for Grp75 and Hsp60 (Fig. 8m-o). Regarding GSK3\u03B2 we observed a nearly significant decrease of pGSK3\u03B2S9 in MCI brain with respect to age-matched controls (Fig. 8p). Considering the average age of the subjects (\u223C90 yrs), these results describe a very late phase in the progression of the pathology supporting the failure of UPRmt activation in MCI and AD. PCA analysis performed by taking into consideration BVR-A, GSK3\u03B2, UPRmt markers, age, sex and MMSE scores, shows that only AD group can be identified by a distinct pattern (Fig. 8r). The explanation could be that, given an equal dysfunction in the activation of the UPRmt, most likely attributable to the advanced age of the study subjects, AD neuropathology becomes a distinguishing factor.Indeed, studies in literature suggest that UPRmt-response is weakened in T2D and is responsible for mitochondrial and cellular dysfunctions as well as increased oxidative stress [66]. In a model of high fat diet-induced brain insulin resistance, a disrupted mitochondrial stress response led to mitochondrial dysfunction, excessive autophagy, and increased weight gain [61]. Short-term intranasal insulin application restored the expression of Atf4, Chop, Hsp60, Hsp10, ClpP, and Lonp1, suggesting that insulin signaling regulates mitochondrial stress response and ensures proper mitochondrial function [61]. Additionally, T2D mice exhibit a reduction of the Hsp60 mRNA in the hypothalamus that is sufficient to induce hypothalamic insulin resistance [ 67\u201369]. Furthermore, Hsp60 mRNA are decreased in post-mortem brain samples from T2D patients [67]. Our data collected on MCI and AD post-mortem brain samples further support the weaking of UPRmt with the progression of the pathology [70], whereby persistently elevated Atf5 levels parallel no changes on the downstream targets and mostly important are associated with clinically relevant neuropathology. Nevertheless, it is also important to underly that the abnormal activation of UPRmt leads to cells undergoing continuous mitochondrial recovery [51]. This process may elevate the risk of accumulating misfolded proteins with aging, playing a direct role in age-related deterioration [51]. Therefore, the precise regulation of UPRmt becomes crucial [51]. Identifying these early alterations in the brain might be valuable for better characterizing the impact of T2D on the brain and for identifying subjects at a higher risk of developing neurodegenerative diseases.This work has been supported by Fondi Ateneo grants funded by Sapienza University (#RM120172A3160B53 and #RG11916B87F55459) to EB; by the Alzheimer's Association grant (#2019-AARG-643091) to EB; by Banca d'Italia funded grant (#1130944/22) to EB; by the Swedish Research Council (#2019-01130) and the Albert P\u00E5hlssons stiftelse to JMND. JMND acknowledges support from The Knut and Alice Wallenberg foundation and the Lund University Diabetes Centre, which is funded by the Swedish Research Council (EXODIAB 2009-1039) and the Swedish Foundation for Strategic Research (IRC15-0067). This work has been supported by Fondi Ateneo grants funded by Sapienza University (#RM120172A3160B53 and #RG11916B87F55459) to EB; by the Alzheimer\u2019s Association grant (#2019-AARG-643091) to EB; by Banca d\u2019Italia funded grant (#1130944/22) to EB; by the Swedish Research Council (#2019-01130) and the Albert P\u00E5hlssons stiftelse to JMND. JMND acknowledges support from The Knut and Alice Wallenberg foundation and the Lund University Diabetes Centre, which is funded by the Swedish Research Council (EXODIAB 2009-1039) and the Swedish Foundation for Strategic Research (IRC15-0067).

FundersFunder number
UPRmt
Knut och Alice Wallenbergs Stiftelse
Università degli Studi di Roma Unitelma Sapienza
Bundesministerium des Innern, für Bau und Heimat
Ministry of Communications and Information, Singapore
Direktör Albert Påhlssons Stiftelse
Fondi Ateneo
Alzheimer's Association2019-AARG-643091, 1130944/22
Alzheimer's Association
UPRmt activation in MCILonp1, 67–69
Stiftelsen för Strategisk ForskningIRC15-0067
Stiftelsen för Strategisk Forskning
Lund University Diabetes CentreEXODIAB 2009-1039
Vetenskapsrådet2019-01130
Vetenskapsrådet

    Keywords

    • Biliverdin reductase-A
    • Brain insulin resistance
    • GSK3β
    • Mitochondrial metabolism
    • Mitochondrial unfolded protein response
    • Oxidative stress

    ASJC Scopus subject areas

    • Organic Chemistry

    Fingerprint

    Dive into the research topics of 'Biliverdin Reductase-A integrates insulin signaling with mitochondrial metabolism through phosphorylation of GSK3β'. Together they form a unique fingerprint.

    Cite this