Fig 2: Proposed mechanism for LAM response regulation by BHLHE40/41.Left) In wild-type mouse microglia or human iPSC-derived microglia (iMGLs) and THP-1 macrophages (MACs) treated with scrambled siRNA (and thus with normal levels of BHLHE40/41), when LXR:RXR and MiT/TFE family TFs stimulate the expression of cholesterol clearance and lysosomal processing genes that contribute to the LAM transcriptional program, they also stimulate the expression of BHLHE40 and BHLHE41. In turn, BHLHE40/41 oppose the activity of LXR:RXR and MiT/TFE family TFs by binding to the promoters of their target genes and leading to decreased expression of cholesterol clearance and lysosomal processing genes, in a negative feedback loop. Right) In mouse microglia or human iMGLs and MACs with complete loss or reduced levels of BHLHE40/41, this negative feedback loop is interrupted or weakened resulting in increased expression of cholesterol clearance and lysosomal processing genes to partially recapitulate the LAM transcriptional and cellular response, including increased cholesterol efflux to APOE and storage in LDs, as well as increased lysosomal mass, acidification and proteolysis.

Fig 3: Knockout of BHLHE40/41 partially recapitulates the LAM transcriptional response in human iPSC-derived microglia (iMGLs).A) Rank-rank hypergeometric overlap (RRHO) heatmaps visualizing significant overlaps in gene expression changes between each pair of BHLHE40/41 knockout (KO) iMGL and human LAM (Jaitin et al. [3], Dataset S6, FDR Adj.P-value < 0.05) transcriptional signatures. Adj.P-value in color temperature scale represents Benjamini-Hochberg corrected P-value of hypergeometric overlap test. B) Venn diagrams of most significant overlaps between genes up-regulated in both BHLHE40/41 KO iMGL and human LAM transcriptional signatures, corresponding to the warmest pixel in the bottom-left quadrant of the respective heatmap (left) and Venn diagrams of most significant overlaps between genes down-regulated in both BHLHE40/41 KO iMGL and LAM transcriptional signatures, corresponding to the warmest pixel in the upper-right quadrant of the respective heatmap (right). P-values are calculated using the hypergeometric overlap test restricted to the universe of 517 human LAM genes (Jaitin et al. [3], Dataset S6, FDR Adj.P-value < 0.05). C) Pathways (“diseases and biological functions” category) found by IPA to be significantly enriched for RRHO overlapping genes. Red bars represent pathways predicted by IPA to be more active (i.e., with positive Z-scores), blue bars represent pathways predicted by IPA to be less active (i.e., with negative Z-scores), grey bars represent pathways with non-attributed Z-scores. Adj.P-value on the x-axis represents Bonferroni-Holm corrected P-value of pathway enrichment test. 40KO = BHLHE40 KO iMGLs, 41KO = BHLHE41 KO iMGLs, DKO = BHLHE40 and BHLHE41 double KO iMGLs, all compared to iMGLs derived from the parental iPSC line.

Fig 4: Knockout of BHLHE40/41 increases expression of lipid and lysosomal clearance genes, cholesterol efflux and lipid droplet content, lysosomal mass and degradative capacity in human iPSC-derived microglia (iMGLs).A) Expression of lipid and lysosomal clearance genes measured by RT-qPCR, N=5/group. log2(fold change) (log2FC) is calculated with WT iMGLs as reference. Separate plots for each gene in Supplementary Figure 4. B) Intracellular APOE normalized to Actin measured by western blot, N=3/group. C) ABCA1 normalized to Actin measured by western blot, N=3/group. D) Secreted APOE normalized to total protein measured by ELISA, N=5/group. E) Percentage of cholesterol efflux, N=5/group. F) LDs content (BODIPY gMFI) measured by flow cytometry in BODIPY-positive cells (top), representative flow-cytometry histograms (bottom), N=6/group. G) Lysosomal mass (LysoTracker gMFI) measured by flow cytometry in LysoTracker-positive cells (top), representative flow-cytometry histograms (bottom), N=5/group. H) Lysosomal acidification (LysoSensor gMFI) measured by flow cytometry in LysoSensor-positive cells (top), representative flow-cytometry histograms (bottom), N=5/group. I) Lysosomal proteolysis (DQ-BSA gMFI) measured by flow cytometry in DQ-BSA-positive cells (top) representative flow-cytometry histograms (bottom), N=5/group. J) DQ-BSA red fluorescent signal (total integrated density) measured over time using the Incucyte S3 live imaging system, N=3/group. N represents the number of independent iMGLs differentiations. Flow cytometry gates were drawn based on FMO (fluorescence minus one) controls (F) or cells treated with Bafilomycin A (G–I) to prevent fusion of autophagosomes with lysosomes and thus obtain LysoSensor, LysoTracker and DQ-BSA negative signals. Differences of means between groups were tested using one-way ANOVA with repeated measures followed by Dunnett’s post-hoc test. * P-value < 0.05, ** P-value < 0.01, *** P-value < 0.001. Data plotted as mean ± SEM. Detailed statistics are shown in Supplementary File 1. 40KO = BHLHE40 KO iMGLs, 41KO = BHLHE41 KO iMGLs, DKO = BHLHE40 and BHLHE41 double KO iMGLs, WT = iMGLs derived from the parental iPSC line.

Fig 5: Knockdown of BHLHE40/41 increases expression of lipid clearance genes, cholesterol efflux and lipid droplet content in human THP1 macrophages (MACs).A) Expression of lipid and lysosomal clearance genes measured by RT-qPCR, N=7/group. log2(fold change) (log2FC) is calculated with SCR MACs as reference. B) Secreted APOE normalized to total protein measured by ELISA, N=6/group. C) Percentage of cholesterol efflux, N=6/group. D) LD content (BODIPY gMFI) measured by flow cytometry in BODIPY-positive cells, N=6/group. N represents the number of independent siRNA transfections. Flow cytometry gates were drawn based on FMO (fluorescence minus one) controls. Differences of means between groups were tested using one-way ANOVA with repeated measures followed by Dunnett’s post-hoc test. * P-value < 0.05, ** P-value < 0.01, **** P-value < 0.0001. Data plotted as mean ± SEM. Detailed statistics are shown in Supplementary File 1. 40KD = MACs treated with BHLHE40 siRNA, 41KD = MACs treated with BHLHE41 siRNA, DKD = MACs treated with BHLHE40 and BHLHE41 siRNA, SCR = MACs treated with scrambled siRNA.