Article Figures and data Abstract Editor's evaluation Introduction Results Discussion Materials and methods Data availability References Decision letter Author response Article and author information Metrics Abstract Wolfram syndrome 1 (WS1) is a rare genetic disorder caused by mutations in the WFS1 gene leading to a wide spectrum of clinical dysfunctions, among which blindness, diabetes, and neurological deficits are the most prominent. WFS1 encodes for the endoplasmic reticulum (ER) resident transmembrane protein wolframin with multiple functions in ER processes. However, the WFS1-dependent etiopathology in retinal cells is unknown. Herein, we showed that Wfs1 mutant mice developed early retinal electrophysiological impairments followed by marked visual loss. Interestingly, axons and myelin disruption in the optic nerve preceded the degeneration of the retinal ganglion cell bodies in the retina. Transcriptomics at pre-degenerative stage revealed the STAT3-dependent activation of proinflammatory glial markers with reduction of the homeostatic and pro-survival factors glutamine synthetase and BDNF. Furthermore, label-free comparative proteomics identified a significant reduction of the monocarboxylate transport isoform 1 (MCT1) and its partner basigin that are highly enriched on retinal glia and myelin-forming oligodendrocytes in optic nerve together with wolframin. Loss of MCT1 caused a failure in lactate transfer from glial to neuronal cell bodies and axons leading to a chronic hypometabolic state. Thus, this bioenergetic impairment is occurring concurrently both within the axonal regions and cell bodies of the retinal ganglion cells, selectively endangering their survival while impacting less on other retinal cells. This metabolic dysfunction occurs months before the frank RGC degeneration suggesting an extended time-window for intervening with new therapeutic strategies focused on boosting retinal and optic nerve bioenergetics in WS1. Editor's evaluation The primary goal of this paper is to characterize retinal dysfunction and retinal ganglion cell degeneration in the Wfs1exon8del murine model of Wolfram Syndrome 1. The study provides fundamental insight into the timelines of degeneration as well as valuable transcriptomic and proteomic datasets. The methodologies performed are rigorous and the conclusions reached are generally well supported by the data. The results of this study are highly relevant for molecular mechanisms in Wolfram Syndrome 1 and are of potential interest to scientists interested in oligodendrocyte and neuron communication. https://doi.org/10.7554/eLife.81779.sa0 Decision letter Reviews on Sciety eLife's review process Introduction Wolfram syndrome 1 (WS1) is a rare and multisystemic genetic disease. In most cases the first clinical sign is the development of non-autoimmune diabetes mellitus (DM) during infancy followed by optic atrophy (OA) which leads to progressive visual loss in adolescence progressing relentlessly to blindness in adulthood (Wolfram and Wagener, 1938; Barrett et al., 1995). While DM symptoms are clinically treatable with antidiabetic therapies, interventions to arrest or delay the loss of sight are lacking (Urano, 2016). Although concurrent DM and OA are pathognomonic for WS1, patients might develop additional serious neurological illnesses, most commonly hearing loss, ataxia, epilepsy, and peripheral neuropathy (Minton et al., 2003; Chaussenot et al., 2011). Serious psychiatric disturbances have also been described in WS1 patients including psychosis, episodes of severe depression, and impulsive/aggressive behavior (Swift et al., 1990; Swift and Swift, 2005). OA is caused by the selective loss of retinal ganglion cells (RGCs) and their axons in the optic nerve with a pattern of degeneration that mainly interests the central part of the optic tract rather than the periphery and is accompanied by different degrees of demyelination (Zmyslowska et al., 2019; Barboni et al., 2022). Noteworthy, neuroimaging studies revealed diffuse alterations in the gray and white matters of the brain with significant loss of volume of the brainstem and microstructure abnormalities and signs of degeneration of myelin (Lugar et al., 2016; Samara et al., 2019). The gene mutated in WS1 is WFS1 which encodes for a multi-transmembrane protein and resident in the endoplasmic reticulum (ER) named wolframin (Inoue et al., 1998; Strom et al., 1998). Mounting evidence suggests that WFS1 gene loss induces ER stress-mediated apoptosis of insulin producing pancreatic β-cells with disruption of cellular calcium homeostasis (Fonseca et al., 2005; Fonseca et al., 2010; Riggs et al., 2005; Yamada et al., 2006). In particular, wolframin is able to increase the Ca2+ uptake within the ER, partially through the regulation of the sarco/endoplasmic reticulum Ca2+-ATPase SERCA pump (Zatyka et al., 2015). As a consequence, wolframin loss results in the Ca2+ leakage from the ER with a simultaneous increase of its cytosolic concentration, activation of the Ca2+-dependent protease calpain which, in turn, activates programmed apoptosis (Lu et al., 2014; Takei et al., 2006). Moreover, recent findings have suggested that wolframin is required for the proper release of both insulin and neurotransmitters from the ER and their delivery to the cell periphery through assisted vesicle trafficking (Wang et al., 2021). Increasing reports have described dysfunctional ER-mitochondria communication caused by wolframin loss. In fact, WFS1 deficiency deranges the inositol 1,4,5-trisphosphate receptor (IP3R)-mediated release by destabilizing the NCS1 protein and, thereby, leading to reduced Ca2+ uptake by the mitochondria with consequent functional impairments (Schlecker et al., 2006; Cagalinec et al., 2016; Angebault et al., 2018). Moreover, a recent study unveiled an enriched localization of wolframin at the MAM sites and in WS1 patient fibroblasts, wolframin deficiency correlated with MAMs loss and, therefore, with defective Ca2+ transfer from the ER to the mitochondria and their related dysfunctions (La Morgia et al., 2020). Remarkably, pharmacological activation of the sigma-1 receptor (S1R), an ER-resident protein involved in calcium ion transfer, was shown to counteract the functional alterations of MAMs due to wolframin deficiency (Crouzier et al., 2022a). Altogether, these results highlight mitochondrial dysfunctions as a relevant pathogenetic event triggered by wolframin and leading to cell death. However, most of these findings were obtained using cell lines and patients’ fibroblasts, and, therefore whether these pathogenetic alterations have the same relevance in neural cells and represent the main and only cause of neurodegeneration remain to be further investigated. Different Wsf1 targeted mutant mouse and rat lines have been generated and showed to recapitulate the main cardinal pathological traits of the WS1 patients. Furthermore, zebrafish mutant lines for the two homologue genes wfs1a and wfs1b have been recently produced showing retinal abnormalities and behavioral visual dysfunctions associated with ER stress and impaired UPR response in both strains (Cairns et al., 2021; Crouzier et al., 2022b). Between the mouse models, the neurological and behavioral alterations of the Wfs1exon8del targeted mice have been extensively characterized showing diabetic glucose intolerance and anxious-like behaviors upon stressful environment exposure (Kato et al., 2008; Luuk et al., 2008). Visual impairment in these mice is less characterized with only a histopathological study that reported some thinning of the retinal tissue with lower retinal thickness/longitudinal diameter ratio in 4 months old Wfs1-deficient mice (Waszczykowska et al., 2020). More recently, a mutant rat line was obtained through Wfs1 exon 5 disruption showing the hallmarks of WS1, with diabetes, optic atrophy, and neurodegeneration (Plaas et al., 2017). The increase of ER stress levels was detected in mutant retina, brainstem and pancreas, coupled with a volumetric decrease in all these districts. These features made it very useful to test the efficacy of the pharmacological treatments (Plaas et al., 2017). Prolonged administration of the glucagon like peptide 1 receptor (GLP1-R) agonists have demonstrated beneficial effects in preventing diabetes, excitotoxicity, and ER stress in aged rats and delaying disease progression (Toots et al., 2018; Seppa et al., 2019). Very recently the synergistic effects of GLP1-R agonist and the brain-derived neurotrophic factor mimetic 7,8-DHF co-treatment have shown neuroprotective capacity on rats’ visual pathway rescuing visual activity (Seppa et al., 2021). Interestingly, chronic ER stress and autophagy alterations caused by Wfs1 loss results in the acceleration of Tau protein aggregation and relative pathology in mouse hippocampus suggesting a novel mechanism by which wolframin deficiency can lead to neurodegeneration (Delpech et al., 2021; Chen et al., 2022). Thus, rodent models of Wfs1 deficiency offers pivotal systems where to dissect and further validate the pathogenetic basis of this disease. Herein, we employed the Wfs1exon8del mutant mice to better characterize the progressive impairments of visual physiology and activity over time and identify novel molecular and biochemical alterations by genome-wide transcriptomics and proteomics on isolated retinal tissues. These targeted analyses reveal neuroinflammation and bioenergetic substrate loss as plausible causes for the selective vulnerability of RGCs and their degeneration in WS1 mice. Results Progressive visual impairments in Wfs1 mutant mice To determine RGCs functionality and visual signal conduction to the brain in Wfs1 mutant mice, we exploited whole field photopic electroretinogram (pERG), which measures action potentials produced by the retina when it is stimulated by light of adequate intensity and it is the composite of electrical activity from the photoreceptors, inner retina, and RGCs. In photopic conditions, the negative response (PhNR) wave represents the RGC spiking activity originating from light-adapted cones and allows the detection of RGC-specific alterations (Gotoh et al., 2004; Kinoshita et al., 2016). Whilst pERG measurements were comparable between wild-type and Wfs1 mutant mice at 2 months of age (WT: 68±5; Wfs1-KO: 72±6; p<0.05), a prolonged implicit time, which refers to the interval between the stimulation and the peak of the negative response, was detected in 4 months old Wfs1-deficient animals (WT: 76±5; Wfs1-KO: 82±6; p<0.05) (Figure 1A). This early indication of RGC abnormalities became increasingly more pronounced in 8 (WT: 73 ± 6; Wfs1-KO: 83 ± 9; p < 0.05) and 12 (WT: 76 ± 5; Wfs1-KO: 85 ± 5; p < 0.0001) months old mutant mice, with the latter group that showed, beside a delayed response , also a significant PhNR amplitude decrease (WT: 11±2; Wfs1-KO: 8±2, p<0.0001) at the latest time period (Figure 1A). To further deepen our neurophysiological evaluation, we sought to investigate light-induced electric conduction from the retina to the brain visual cortex by flash visual evoked potential (fVEP) recordings. These measurements represent a direct readout of the optic nerve functional activity by providing two important parameters: the latency, which measures the propagation speed of the visual stimulus along the nerve fibers until its central detection; the amplitude, that represents the potential difference generated after administration of the visual stimulus, allowing quantification of the power of the electrical signal produced in the optic nerve. VEP trace typically consists of three waves, a first positive wave (P1), followed by both a negative (N1) and a second positive wave (P2). We measured the latency of N1 and the peak-to-peak amplitude of the N1-P2 complex to assess the optic nerve function (Figure 1A). While 2 and 4 months old mutant mice did not display detectable VEP alterations, 8-month-old mutant mice displayed increased latencies (WT: 46±4; Wfs1-KO: 51±2, p<0.001), indicative of slower transmissions along the visual pathway (Figure 1A). However, both VEP parameters were significantly affected as mouse age progressed to 12 months (VEP amplitude WT: 40±14; Wfs1-KO: 28±2; p<0.001; VEP latencies WT: 47±2; Wfs1-KO: 52±3, p<0.001) (Figure 1A). We then asked if these neurophysiological deficits correlate with manifested loss of visual acuity as recorded by the optomotor reflex response (OMR) based on tracking the unconditioned compensatory head movements of the animal while trying to stabilize the image of the moving environment (rotating vertical stripes). Wfs1 mutant mice showed a progressive reduction in visual acuity with undetectable differences at 2 and 4 months of age while a significant loss was recorded at 8 months which markedly worsened 4 months later (4 months - WT: 0.40±0.5; Wfs1-KO 0.36±0.8 c/deg, 8 months - WT: 0.39±0.4; Wfs1-KO: 0.30±0.3 c/deg, p<0.05, 12 months - WT: 0.38±0.6; Wfs1-KO: 0.19±0.5 c/deg, p<0.01; Figure 1B). These results imply that the initial visual impairment in Wfs1 mutant mice is represented by RCG alterations in the retinal tissue followed by a delay of the visual stimulus along the nerve fibers suggesting a progressive optic nerve damage. Subsequently, RGC functions are extensively altered causing a drastic loss of visual acuity at 12 months of age. Figure 1 with 1 supplement see all Download asset Open asset Wfs1 mutant mice show progressive and severe impairment of visual activity. (A) Representative pERG and fVEP waveforms at 4, 8, and 12 months of age in wild-type (WT) and Wfs1 mutant (KO) animals. PhNR amplitude and implicit time are quantified by ERG measurements (n=15, *p<0.05; **p<0.01; ***p<0.001, Student t-test). Data are presented as mean ± SEM. N1 latency and N1-P2 amplitude quantifications are calculated from the fVEP traces (n=15, *p<0.05; **p<0.01; ***p<0.001, Welch’s t-test for N1-P2 amplitude and Student t-test for N1 latency). Data are presented as mean ± SEM. (B) Quantification of visual acuity measuring the opto-motor reflex expressed as cycles per degree in 2, 4, 8, and 12 months old wild-type (WT) and Wfs1 mutant (KO) mice (n=8, *p<0.05; **p<0.01; **p<0.001; ****p<0.0001). Data are presented as mean ± SEM. Optic nerve degeneration precedes RGC loss in Wfs1 mutant mice No deep characterization of the Wfs1 mutant visual structures has been yet reported and the data available refers exclusively to eye morphology in relatively young animals (Waszczykowska et al., 2020). At first, we evaluated the gross retinal architecture in 8 months old mice by optical coherence tomography (OCT) and fluorescein angiography (FA), in terms of both morphology and neovascularization. FA assessment was necessary to exclude confounding effects due to diabetic retinopathy (DM), which represents one of the major eye complications of DM but whose causative mechanisms are highly different from those in WS1 (Hilson et al., 2009). The collected images did not reveal any signs of retinal vasculature dysfunction in either group of animals. Indeed, OCT-angiography did not highlight any significant difference in the eye fundi or perfusion and vessel length density and tortuosity (Figure 1—figure supplement 1). Moreover, we also assessed the thickness of retinal nerve fiber layer (RNFL), and no significant differences were disclosed by structural OCT (Figure 1—figure supplement 1). Since mutant elderly mice often displayed corneal haze, with consequent poor quality images, we managed to sample only three 12 months old clear-lensed eyes that still did not show big abnormalities in RNFL thickness at OCT segmentation (Figure 1—figure supplement 1). To investigate the cellular dysfunctions underlying the neurophysiological recordings, we performed transmission electron microscopy analyses (TEM) of transversal optic nerve sections of Wfs1 mutant and littermate control mice at the same timepoints (8 and 12 months). Post-hoc TEM imaging analysis enabled us to estimate average RGC axonal densities (mean axon count/area), average myelin area/field also quantifying the empty space (i.e. uncovered by axons) and g-ratio (axonal area divided by axonal area including that of myelin sheets). These elaborations revealed that 8 months old Wfs1 mutant optic nerves presented incipient myelin thinning (g-ratio - WT: 0.52±0.18; Wfs1-KO: 0.54±0.14, p<0.05) with reduced myelin area and a higher proportion of empty space, but without axonal count discrepancies (Figure 2A). TEM imaging in elderly mutant animals (12 months of age) unveiled signs of more severe myelin degeneration (g-ratio - WT: 0.42±0.18; Wfs1-KO: 0.50±0,18, p<0.05) with pathological features and degenerating fibers (Figure 2B). In particular, we observed a significant reduction of axonal density and area occupied by myelin together with evident signs of axonal damage with increased g-ratio (WT: 0.41±0.04; Wfs1-KO: 0.51±0.003, p<0.05; Figure 2B). These results provide evidence that myelin alterations and axonal thinning are the first pathological signs followed by manifested myelin degeneration and axonal loss in Wfs1 mutant optic nerves. This pathological progression is perfectly matching the timeline of the visual functional deficits previously uncovered by electrophysiological recordings and the optomotor analysis. To assess whether optic axonal damage was occurring with simultaneous alterations in RGC cell bodies within the retinal tissue, we performed a detailed count of RGC number by immunocytochemical staining using the pan-RGCs marker RBMPS (Figure 2C). Unbiased stereology counting did not reveal significant RGC number changes in 8 months old Wfs1 mutant mice (Figure 2C). However, 4 months later the loss of the RGC population became evident and statistically relevant (28% ± 6% loss in Wfs1-KO vs WT retinas) indicating a progressive loss of RGC somata over time (Figure 2C). These results illustrate that axonal damage and myelin degeneration anticipate the frank loss of RGC cell bodies in the retinas, even if within a relative delayed time frame a severe loss of RGCs can be also detectable in the Wfs1-deficient retinas. Figure 2 Download asset Open asset Progressive demyelination in the Wfs1 mutant optic nerves. (A) Representative EM images of optic nerve cross-sections of wild-type (WT) and Wfs1 mutant mice at 8 months of age showing increased space among the axons (red asterisks). Quantifications of the g-ratio (n=193, *p<0.05, unpaired t-test), total number of axons (n=34), myelin (n=29) and empty areas (n=29) normalized by total area (n=29) (**p<0.01; ***p<0.001, two-way ANOVA with Bonferroni’s post-hoc test). Data are presented as mean ± SEM. (B) Representative EM images of optic nerve cross-sections of wild-type (WT) and Wfs1 mutant mice at 12 months of age showing dramatic expansion of space among the axons (red asterisks). Quantifications of the g-ratio (n=192, *p<0.05, unpaired t-test), total number of axons (n=29), myelin (n=27) and empty areas (n=27) normalized by total area (n=27) (***p<0.001, two-way ANOVA with Bonferroni’s post-hoc test). Data are presented as mean ± SEM. (C) Representative immunohistochemical images of retinal cross-sections of wild-type (WT) and Wfs1 mutant (KO) animals at 8 and 12 months of age with RBPMS and DAPI staining in purple and blue, respectively. Relative quantification of RBPMS positive cells over total cell number in the GCL at 8 months (n=9, p=0.39, two-tails unpaired t-test) and at 12 months (n=8, **p<0.01, two-tails unpaired t-test). Scale bar: 50 µm. Transcriptomic analysis reveals aberrant gliosis which precedes neuronal loss in Wfs1 mutant retinas To unveil specific molecular alterations underlying the pathological deficits caused by the Wfs1 gene loss, we performed transcriptome analysis of isolated mutant and control mouse retinas. To investigate prodromic events that anticipate the morphological changes, we performed RNA-seq analysis on tissues from 8 months old mice. RNAs were extracted from whole retinas of three different mice per group (WT vs Wfs1-KO) and subjected to global gene expression analysis. t-SNE analysis of the RNA-seq datasets confirmed the expected closer similarity and correlation between all mutant retinas in comparison to the analogous wild-type samples (Figure 3A). Computational analysis identified 396 genes differentially expressed (DEGs) in mutant retinas compared to the control counterparts, with a similar proportion of either up- or downregulated genes (Figure 3B), suggesting not profound transcriptome alterations but rather selective changes in mutant samples at this stage. Interestingly, gene enrichment analysis (Gene Ontology, GO) revealed many deregulated genes related to pathways like cell death, UPR response and protein folding (Figure 3C). Intriguingly, all these pathways were down-regulated in the mutant retinas, in line with the general assumption from literature that wolframin loss impairs the cellular stress-coping capacities. On the contrary, mutant retinas upregulated key genes of neuroinflammatory pathways (Figure 3C) as those associated with reactive glial cells like, Gfap, Lcn2, Edn2, Alpk1, and C4 (Figure 3D; Liu et al., 2022; Linnerbauer et al., 2022). We, next, performed immunocytochemical analysis confirming the aberrant activation of the ER stress effectors ATF6 (WT: 0.5±0.3; Wfs1-KO: 1.5±0.2) and CHOP (WT: 0.2±0.01; Wfs1-KO: 0.4±0.02) in mutant retinal samples (Figure 4A). Moreover, GFAP (8 months WT: 0.5±0.2; Wfs1-KO: 1.4±0.3; 12 months WT: 0.2±0.02; Wfs1-KO: 52±2) and Vimentin (VIM) (8 months WT: 20±2; Wfs1-KO: 38±5; 12 months WT: 2.2±0.5; Wfs1-KO: 4.6±0.8) staining was found strongly upregulated in mutant retinal Müller glia at both 8 and 12 months of age (Figure 4B). Furthermore, the glia-specific glutamine synthetase (GS) enzyme was found downregulated both at mRNA and protein levels in Wfs1-deficient retinas (Figures 3C and 4D). GS generates glutamine by assembling glutamate and ammonia and, therefore, having a critical role in preventing glutamate-dependent excitotoxicity and ammonia toxicity (Rose et al., 2013). Thus, GS reduction in Wfs1-deficient retinal glial cells might favor a toxic environment detrimental for neuronal survival. Next, the genes encoding for the key neurotrophic molecule BDNF, and the survival factors Nr4a1 and Nr4a3 were also reduced in Wfs1 mutant retinas (Figure 3D), further altering neuronal homeostasis and stress resilience. In fact, BDNF is a well-known neurotrophic factor expressed by RGCs and Müller glia in adulthood retina, where it exerts its autocrine or paracrine effects. Multiple lines of evidence suggest that BDNF promotes survival of RGCs and can ameliorates RGC death after traumatic optic nerve injury (Bennett et al., 1999). It is thus not surprising that its downregulation occurs before ganglion cell death. Finally, a number of microglial pro-inflammatory genes such as Il-18, Gpr37, Cybb, Lyz2, Lglas3, and Mpp12 were upregulated, suggesting that Wfs1 mutant microglia activate at least some reactive-specific pathways (Figure 3D and E). However, levels of interferon-inducible genes and NF-κB signaling components were unchanged suggesting the lack of a generalized inflammatory response, but rather a selective activation of specific neuro-inflammatory pathways. To further strengthen our results, RT-qPCR analysis were performed on independent biological samples and substantially confirmed the transcriptome data for the key selected genes (Figure 3E), further highlighting a selective pro-inflammatory profile in Wfs1-deficient retinas. On this line, we noted that the dysregulated Lcn2 and Vim genes in Wsf1 mutant samples have been shown to be direct targets of the STAT3 signaling (Herrmann et al., 2008). Intriguingly, previous studies reported that STAT3 can be stimulated by the ER stress-dependent PERK and IRE1α pathways in glial and cancer cells (Meares et al., 2014; Chen and Zhang, 2017). Thus, we profiled by western blotting both total and active (pY705) forms of STAT3 in control and Wfs1 mutant retinal lysates at 8 months of age. Interestingly, the active STAT3 variant was highly enhanced in mutant samples although with some variability among the samples (Figure 5A). Thus, we profiled by candidate gene expression other STAT3 regulated genes such as Fgf2, Il6, Vegfa, C1s, and C1rl finding them consistently upregulated in Wfs1 mutant retinas (Figure 5B). Altogether, these findings uncovered a prominent pro-inflammatory molecular program associated with reduction of survival and homeostatic factors which is promoted at least in part by the ER stress-dependent STAT3 activation. Figure 3 Download asset Open asset Transcriptome analysis of Wfs1 mutant and control retinas. (A) Whole-transcriptome analysis using the t-Distributed Stochastic Neighbor Embedding (t-SNE) analysis with a view of the sample distribution along the first two dimensions. (B) Significantly up- (red) and down- (blue) regulated genes among the transcriptomic profiles of the samples, shown as highlighted dots in the MA plots. Number of differentially expressed genes (DEGs) (up = 218; down = 178). (C) GO-term categories relative to transcriptional analysis, as in RNA-seq dataset calculated in the list of genes downregulated in Wfs1 mutant retinas. (D) Heatmap showing genes normalized count RPKM associated with inflammatory pathways. (E) Boxplot depicting the distribution of gene expression levels, confirming their trends in RNAseq. (n=3 *p<0.05; **p<0.01; ***p<0.001, Student t-test). Figure 4 Download asset Open asset Representative immunohistochemical analysis in wild-type and Wfs1 mutant retinal tissues at 8 and 12 months of age. (A) Representative images and relative quantification of ATF6 and CHOP immunofluorescence signal in wild-type (WT) and Wfs1 mutant (KO) retinas at 8 months of age. On the right, quantification of the signal (n=5, ATF6; n=12, CHOP, **p<0.01, Student t-test). Data are presented as mean ± SEM. Scale bar: 100 µm. (B) Representative images of GFAP immunofluorescence in 8- and 12-month-old retinas. On the right, quantification of the signal (n=5, 8 months; n=13, 12 months, **p<0.01, Student t-test). Data are presented as mean ± SEM. Scale bar: 100 µm. (C) Representative images of Vimentin (VIM) immunofluorescence in 8- and 12-month-old retinas. On the right, quantification of the signal intensity of VIM in wild-type (WT) and Wfs1 mutant (KO) retinas. (n=5, *p<0,05, **p<0.01, Student t-test). Data are presented as mean ± SEM. Scale bar: 100 µm. (D) Representative images and relative quantification of the glutamine synthetase (GS) immunofluorescence signal in wild-type (WT) and Wfs1 mutant (KO) retinas at 8 and 12 months. Data are presented as mean ± SEM. (n=3, **p<0.01, Student t-test). Scale bar: 100 µm. Figure 5 Download asset Open asset Impaired signal transducer and activator of transcription 3 (STAT3) signaling in Wfs1 mutant mice. (A) Representative immunoblotting of pSTAT3 and total STAT3 endogenous protein levels in adult retinae from WT and Wfs1 KO mice at 12 months of age. The ratio of p-STAT3 versus total signal transducer and activator of transcription 3 (STAT3) is shown in the graph after normalization over the respective protein load (actin). Quantification of pSTAT3/STAT3 protein signal intensity was performed in 12 months mice retinae (eyes from n=3 animals, *p<0.05; **p<0.01, Student t-test). Data are presented as mean ± SEM. (B) Boxplot depicting the distribution of expression levels of STAT3 regulated genes (p<0.05; **p<0.01; ***p<0.001, Student t-test). Figure 5—source data 1 Original western blot images used to make Figure 5A. https://cdn.elifesciences.org/articles/81779/elife-81779-fig5-data1-v2.pdf Download elife-81779-fig5-data1-v2.pdf Proteomics analysis of Wfs1 mutant retinas Given that relative few genes showed altered expression levels in Wfs1 mutant retinas, we postulated that additional key changes can be determined only at protein level. In fact, wolframin regulates the function of several proteins that control ER physiology and ER-mitochondrial interactions through post-transcriptional mechanisms. Thus, we collected eyes from WT and Wfs1 mutant mice at 8 months and isolated the retinas with their optic nerves to carry out label-free LC-MS comparative proteomics (Figure 6A). Three biological replicates for each of the two genotypes were analysed in triplicate and aligned through Multidimensional Algorithm Protein Mapping (Comunian et al., 2011). Overall, 2046 proteins were identified (Supplementary file 1) and the virtual 2D map showed the ability of the shotgun approach to identify proteins in a wide range of molecular weight and isoelectrical point (Figure 6—figure supplement 1). In this study, a label-free quantification was applied, considering the total number of peptides (evaluated as Peptide Spectrum Matches or Spectral Count) for each protein. The quantitative analysis, combined with statistical evaluation, led to the identification of 21 differentially expressed proteins between the two genotypes with 9 upregulated and 12 downregulated in the Wfs1-KO compared to the WT condition (Figure 6B and Supplementary file 1). Notably, the most upregulated protein upon Wfs1 gene loss was GFAP, confirming increased gliosis described earlier in the mutant retinas (Figure 6B). Conversely, among the most downregulated molecules in mutant samples, we identified MCT1 (Slc16a1) and basigin. MCT1 belongs to the family of monocarboxylate transporters (MCTs) that promotes the transport across cell membranes and the shuttling between cells of energy metabolites as lactate, pyruvate or ketone bodies (Felmle