Fig 2: Gene dosage of microglial HK2 selectively modulates microglial gene expression in the 5xFAD mice.(A and B) Effect of HK2 gene dosage on gene expression respect to 5xFAD;HK2wt/wt control mice. Differences for 770 genes in Nanostring glia panel are expressed as volcano plot. Adjusted P values were obtained from Rosalind on ramp software. Significance set at adj-p < 0.05 and fold change >1.5, < −1.5 (n=7–8).(C) Heatmaps depicting the directed significance scores for genes annotated as markers of neurodegenerative phenotype (MGnD), DAM signature and complement system. The score magnitude of these gene sets was reduced only in the HK2 haploinsufficient AD mice.(D) Cell abundance scores reveal a microglial specific effect of HK2 partial loss. Cell type scores are based on the NanoString Cell Type Profiling Module *p< 0.05 (n=5, Unpaired t-test).(E) Schematic of the experimental procedure for the analysis of primary cultures of microglia. 4-OHT (4-hydroxytamoxifen).(F) qPCR analysis of selected microglial genes in primary cultures of HK2-deficient microglia. *p< 0.05, **p< 0.01, ***p< 0.001, ****p< 0.0001 (n=4–5 per group, One-way ANOVA followed by Tukey’s post-hoc test).

Fig 3: Gene dosage of microglial HK2 selectively modulates disease progression in the 5xFAD mice.(A) Mice that have a conditional deletion of one or two copies of HK2 gene in microglial cells were generated by crossing 5xFAD;CX3CR1-CreERT2 mice with mice harboring floxed HK2 alleles. At 2 months, five i.p. injections of TAM were administrated to induce recombination. Mice were assessed at 5 months old.(B and C) qPCR analysis of HK2 and HK1 expression in the cortex of 5xFAD mice harboring two (HK2wt/wt), one (HK2Fl/wt) or no copies (HK2Fl/Fl) of HK2 gene. **p< 0.01, *p< 0.05 (n=4–7 per group, One-way ANOVA followed by Tukey’s post-hoc test).(D–G) Immunofluorescent analysis of HK2 (red), microglia (Iba1, white) and Aβ (ThioS, green) in the cortical region of control and HK2 deficient mice. E. Quantification of the percentage of microglial area containing HK2 immunoreactivity shown in D, n=12 slices from 4 to 5 mice. F and G. Quantification of cortical area fraction of ThioS (plaque load) and microglia (Iba1+ cells). *p< 0.05, **p< 0.01, ****p< 0.0001 (n=4–5 per group, One-way ANOVA followed by Tukey’s post-hoc test).(H) Representative immunoblots of fibrillar and monomeric Aβ detected by MOAB2 antibody and Trem2 from soluble fraction of cortical lysates of 5xFAD;HK2wt/wt and 5xFAD;HK2Fl/wt mice treated with TAM. (I–K) Immunoreactive bands were quantified, normalized to β-actin and expressed as the fold change respect to 5xFAD;HK2wt/wt mice. *p< 0.05 and **p< 0.01 (n=5, Unpaired t-test).(L) Representative immunoblots of fibrillar and monomeric Aβ detected by MOAB2 antibody and Trem2 from soluble fraction of cortical lysates of 5xFAD;HK2wt/wt and 5xFAD;HK2Fl/Fl mice treated with TAM. (M–O) Immunoreactive bands were quantified, normalized to β-actin and expressed as the fold change respect to 5xFAD;HK2wt/wt mice. *p< 0.05 and **p< 0.01 (n=3–4 per group, Unpaired t-test).(S) Percentage of spontaneous alternation behavior in a Y-maze as an evaluation of spatial working memory. (n=6–14 per group, One-way ANOVA followed by Tukey’s post-hoc test)

Fig 4: HK2 as a molecular hub between metabolism and immune response in AD.(A) qPCR analysis of HK2 expression during disease progression in the cortex of male and female 5xFAD mice. Dashed line represents gene expression in Bl/6 control mice. Post hoc analysis revealed significant genotype-related increases (###p< 0.001, ####p< 0.0001) and a significant increase in HK2 expression in the cortex of female, compared with male 5xFAD mice at 6 and 8 months (*p< 0.05, ***p< 0.001, n=5–9 per group, One-way ANOVA followed by Tukey’s post-hoc test).(B–C) Western blot and quantification of HK1 and HK2 in the cortex of 8-mo old 5xFAD mice. Dashed line represents gene expression in Bl/6 control mice. Densitometric analysis of hexokinases isoforms were normalized to GAPDH levels (n=4, genotype difference: #p< 0.05, ####p<0.0001 and sex difference: **p<0.01. Unpaired t test).(D) Confocal images of microglial HK2 expression in the cortex of 8-month-old BL/6 (upper panels) and 5xFAD mice (lower panels). Microglia (Iba1+, red), HK2 (green) and deposited amyloid (6E10, purple). Scale bar: 25 μm.(E) Quantification of the integrated density of HK2 in Iba1+ cells. After binarization of Iba1 immunoreactive signal, the mean fluorescence intensity of HK2 was calculated and multiplied by the area of Iba1. *p<0.05 (n=8–9 per group, Unpaired t test)(F) Immunofluorescent staining of HK2 (white), Iba1 (red) and ThioS (green) of brain samples of AD patients.(G) Quantifications of mean fluorescent intensity of microglial HK2 in close contact with Aβ (in radius <15μm form Aβ) and away from Aβ (>15μm). *p<0.05 (n=8–9 per group, Unpaired t test).(H) qPCR analysis of HK2 expression in female vs male post-mortem samples of non-demented (ND) and AD patients normalized to the expression of Iba1+ (n=3–4).(I) HK2 expression obtained from a human transcriptomic dataset of prefrontal cortex tissue (PFC) of 157 nondemented controls and 310 AD patients (GSE33000). Data segregated by sex reveals a significant difference in the up-regulated levels of HK2 between woman and men.(J) AMP-AD consortium co-expression network analysis of RNA-seq data from AD cases and controls. HK2 node, is highlighted in yellow and co-expressed genes associated to microglial activation or identity are highlighted in pink. The network analysis performed by the AMP-AD consortium, uses an ensemble methodology to identify genes that show similar co-expression across individuals.(K) Gene ontology of HK2 co-expression network. The list of specific microglial genes co-expressed with HK2 in hAD cases were used to evaluate the top 15 biological process associated with its expression.

Fig 5: HK2 inhibition reduces ATP production and its gene dosage induce divergent effects on mitochondrial function.(A) In vitro evaluation of HK activity, ATP production and cellular metabolism in primary cultures of microglia haploinsufficient for HK2. (*p< 0.05, n=4–6 per group, Unpaired t-test) cultures(B) In vitro evaluation of HK activity, ATP production and cellular metabolism in primary cultures of microglia treated with different concentrations of HK2 inhibitor LND. (**p< 0.01, n=3 per group, Unpaired t-test and **p< 0.01, ****p<0.0001, n=9 per group, One-way ANOVA, tukey posttest, respectively)(D and E) Quantification of the mean fluorescence intensity (MFI) of mitotracker DR in primary microglia described in C. **p< 0.01, ***p<0.001, ****p<0.0001, n=9 per group, One-way ANOVA, tukey posttest, respectively)(F) Flow cytometry analysis of mitochondrial potential in CD11b+, CD45+ cells treated with LND. Left, representative histogram of the fluorescence intensity of mitotracker deep red (DR) in microglia from 5xFAD mice treated with vehicle or LND. Right, quantification of the geometric mean fluorescence intensity (gMFI) of mitotracker DR. (n=3 per group, Unpaired t-test)(G) Flow cytometry analysis of mitochondrial potential in CD11b+, CD45+ cells of 5xFAD HK2 haploinsufficient mice as described in F. (n=3 per group, Unpaired t-test).(H) Colocalization analysis of HK2 and VDAC in the 5xFAD mice treated with LND. Representative confocal microscopy images showing HK2 (green) VDAC (red) and it colocalization (white) in resting and activated microglia from mice treated with vehicle or LND. Pearson quantification shown that activated microglia display decreased levels of mitochondrial association and LND treatment fails to recover this effect. (***p< 0.001, ****p<0.0001, n=4, >50 cells per group, One-way ANOVA, tukey posttest)