The Alzheimer's disease gene SORL1 regulates lysosome function in human microglia

Abstract

The SORL1 gene encodes the sortilin related receptor protein SORLA, a sorting receptor that regulates endo-lysosomal trafficking of various substrates. Loss of function variants in SORL1 are causative for Alzheimer's disease (AD) and decreased expression of SORLA has been repeatedly observed in human AD brains. SORL1 is highly expressed in the central nervous system, including in microglia, the tissue resident immune cells of the brain. Loss of SORLA leads to enlarged lysosomes in hiPSC-derived microglia like cells (hMGLs). However, how SORLA deficiency contributes to lysosomal dysfunction in microglia and how this contributes to AD pathogenesis is not known. In this study, we show that loss of SORLA results in decreased lysosomal degradation and lysosomal enzyme activity due to altered trafficking of lysosomal enzymes in hMGLs. Phagocytic uptake of fibrillar amyloid beta 1-42 and synaptosomes is increased in SORLA deficient hMGLs, but due to reduced lysosomal degradation, these substrates aberrantly accumulate in lysosomes. An alternative mechanism of lysosome clearance, lysosomal exocytosis, is also impaired in SORL1 deficient microglia, which may contribute to an altered immune response. Overall, these data suggest that SORLA has an important role in proper trafficking of lysosomal hydrolases in hMGLs, which is critical for microglial function. This further substantiates the microglial endo-lysosomal network as a potential novel pathway through which SORL1 may increase AD risk and contribute to development of AD. Additionally, our findings may inform development of novel lysosome and microglia associated drug targets for AD.

Keywords: Alzheimer’s disease; SORL1; hiPSC-derived microglia; lysosomes; phagocytosis.

Figure 1.. Loss of SORL1 results in decreased lysosomal degradation in hMGLs.

(A) Schematic illustrating experimental design using DQ-BSA, a fluorogenic protease substrate that fluoresces only upon lysosomal degradation by proteases. (B) A pulse chase experiment performed with WT and SORL1 KO hMGLs and Mean fluorescence intensity (MFI) measured using flow cytometry. SORL1 KO hMGLs show decreased MFI of DQ-BSA at all timepoints indicating decreased lysosomal degradation. The lysosomotropic agent chloroquine that inhibits lysosomal degradation was used as control. Chloroquine treatment shows decreased lysosomal degradation of DQ-BSA in both WT and SORL1 KO hMGLs. (C) Flow cytometry showing increased MFI of BSA-594 in SORL1 KO hMGLs relative to WT hMGLs suggestive of increased uptake of BSA in hMGLs with loss of SORL1. (D-F) Immunocytochemistry of hMGLs treated with DQ-BSA using the experimental design in (A), cells fixed for 2h and labeled with the microglia marker, CX3CR1 showing decreased lysosomal degradation in SORL1 KO hMGLs as compared to WT hMGLs, validating results obtained from the experiments using flow cytometry. Similar to the flow cytometry experiment, chloroquine was used as a control and showed decreased MFI in both WT and SORL1 KO hMGLs indicating reduced lysosomal degradation Quantification of fluorescence intensity of DQ-BSA using Image J software after immunocytochemistry with microglia marker, CX3CR1 and treatment with DQ-BSA. (E-F) I. 2 isogenic clones per genotype (WT and SORL1 KO) and 9 independent replicates (3 differentiations and 3 technical replicates per differentiation) per clone per genotype (N=18 independent replicates) were used for these experiments. Data represented as mean ± SD and analyzed using parametric two-tailed unpaired t test and 2-way ANOVA with Tukeys multiple comparison test. Significance while comparing WT to SORL1 KO hMGLs was defined and depicted as a value of *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001, ns=not significant and while comparing different treatments of each genotype was defined and depicted as a value of #p < 0.05, ##p < 0.01, ###p < 0.001, and ###p < 0.0001, ns=not significant.

Figure 2.. Loss of SORL1 results in decreased lysosomal enzyme activity in hMGLs.

(A-C) Enzyme activity assays based on fluorometric detection show reduced enzyme activity of lysosomal enzymes (A) HEXB, (B) Cathepsin B and (C) Cathepsin D in SORL1 KO hMGLs as compared to WT hMGLs. Both WT and SORL1 KO hMGLs show reduced enzyme activity of all the enzymes tested when treated with the lysosome function inhibitor, chloroquine, which was used as a control. (D-I) Western blotting demonstrates loss of SORLA protein expression (D) but no change in total protein levels of immature and mature forms of HEXB (E-F), Cathepsin B (G-H) and Cathepsin D (I-J) in SORL1 KO hMGLs relative to WT hMGLs. For the enzyme activity assays, 2 isogenic clones per genotype (WT and SORL1 KO) and 6 independent replicates (2 differentiations and 3 technical replicates per differentiation) per clone per genotype (N=12 independent replicates) were used and for the western blot experiments, 2 isogenic clones per genotype (WT and SORL1 KO) and 2 independent replicates (1 differentiation and 2 technical replicates) per clone per genotype (N=4 independent replicates) were utilized. Data represented as mean ± SD and analyzed using parametric two-tailed unpaired t test and 2-way ANOVA with Tukeys multiple comparison test. Significance while comparing WT to SORL1 KO hMGLs was defined and depicted as a value of *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001, ns=not significant and while comparing different treatments of each genotype was defined and depicted as a value of #p < 0.05, ##p < 0.01, ###p < 0.001, and ###p < 0.0001, ns=not significant.

Figure 3.. Loss of SORL1 results in altered trafficking of lysosomal enzymes in hMGLs

Immunocytochemistry and colocalization analysis used to measure colocalization of lysosomal enzymes HEXB, Cathepsin B and Cathepsin D with TGN and lysosomes using TGN38 and LAMP1 antibodies respectively. As compared to WT hMGLs, SORL1 KO hMGLs show increased colocalization of all the enzymes with TGN38 (Figure 3A,C,E) and decreased colocalization of all enzymes with LAMP1 (Figure 3B,D,F) suggestive of altered trafficking of lysosomal enzymes from TGN to lysosomes. Colocalization was measured using JACOP plugin in Image J software and presented as Manders correlation coefficient. Scale bar=10μm. 2 isogenic clones per genotype (WT and SORL1 KO) and 10 images per clone per genotype (N=20 independent replicates) were used for these experiments. Each image comprised of at least 5–10 cells and hence a total of 50–100 cells per clone per genotype were analyzed for colocalization analysis. Data represented as mean ± SD and analyzed using parametric two-tailed unpaired t test. Significance while comparing WT to SORL1 KO hMGLs was defined and depicted as a value of *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001, ns=not significant.

Figure 4.. Loss of SORL1 results in reduced protein levels and dysfunctional trafficking of Cation-independent Mannose-6-phosphate receptor (CIMPR), from the TGN to the lysosomes in hMGLs

(A-B) Immunocytochemistry and colocalization analysis was used to measure colocalization of CIMPR with TGN and late endosomes and lysosomes using TGN38 and LAMP1 antibodies respectively. As compared to WT hMGLs, SORL1 KO hMGLs show no change in colocalization of CIMPR with TGN38 (A) and decreased colocalization of CIMPR with LAMP1 (B). (C-D) Western blot and quantification showing decreased protein levels of CIMPR in SORL1 KO hMGLs as compared to WT hMGLs. Colocalization was measured using JACOP plugin in Image J software and presented as Manders correlation coefficient. Scale bar=10μm. 2 isogenic clones per genotype (WT and SORL1 KO) and 10 images per clone per genotype (N=20 independent replicates; 1 differentiation) were used for these experiments. Each image comprised of at least 5–10 cells and hence a total of 50–100 cells per clone per genotype were analyzed for colocalization analysis. Data represented as mean ± SD and analyzed using parametric two-tailed unpaired t test. Significance while comparing WT to SORL1 KO hMGLs was defined and depicted as a value of *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001, ns=not significant.

Figure 5.. Loss of SORL1 causes increased lysosomal accumulation of fibrillar Aβ 1–42 (fAβ) and synaptosomes in hMGLs

(A-D) Immunocytochemistry and colocalization analysis with and without treatment with chloroquine, showing increased colocalization of fAβ (A-B) and synaptosomes (C-D) with lysosome marker, LAMP1, in SORL1 KO hMGLs as compared to WT hMGLs suggestive of increased lysosomal accumulation. Chloroquine treated WT and SORL1 KO hMGLs show increased lysosomal accumulation of both substrates as compared to their untreated counterparts. Colocalization analysis performed using JACOP plugin of Image J software and data is presented as Manders correlation coefficient. Scale bar=10μm. 2 isogenic clones per genotype (WT and SORL1 KO) and 10 images per clone per genotype (N=20 independent replicates; 1 differentiation) were used for these experiments. Each image comprised of at least 5–10 cells and hence a total of 50–100 cells per clone per genotype were analyzed for colocalization analysis. Data represented as mean ± SD and analyzed using parametric two-tailed unpaired t test and 2-way ANOVA with Tukeys multiple comparison test. Significance while comparing WT to SORL1 KO hMGLs was defined and depicted as a value of *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001, ns=not significant and while comparing different treatments of each genotype was defined and depicted as a value of #p < 0.05, ##p < 0.01, ###p < 0.001, and ###p < 0.0001, ns=not significant

Figure 6.. Loss of SORL1 causes reduced lysosomal exocytosis in hMGLs

(A-B) Surface localization of LAMP1 measured by immunocytochemistry and flow cytometry as a readout of LE A) Immunocytochemistry with antibody specific for the N terminal luminal epitope of LAMP1 and microglia marker, CX3CR1 demonstrating decreased surface localization of LAMP1 in SORL1 KO hMGLs as compared to WT hMGLs. B) Quantification by flow cytometry showing decreased MFI of LAMP1 antibody in SORL1 KO hMGLs relative to WT hMGLs suggestive of decreased LE with loss of SORL1 in hMGLs. C) Western blot and quantification shows no change in protein expression of LAMP1 in SORL1KO hMGLs relative to WT hMGLs. (D-F) Extracellular enzyme activity of lysosomal enzymes measured as a readout of LE. Enzyme activity assays demonstrate decreased lysosomal enzyme activity of D) HEXB, E) Cathepsin B and F) Cathepsin D in SORL1 KO hMGLs as compared to WT hMGLs suggestive of decreased LE with loss of SORL1 in hMGLs. (G-L) Cytokine secretion of IL-6 measured by ELISA shows increased extracellular secretion of IL-6 upon calcimycin treatment in both WT and SORL1 KO hMGLs without increase in intracellular protein expression but SORL1 KO hMGLs show decreased secretion of IL-6 in response to calcimycin as compared to WT hMGLs (G, J). SORL1 KO hMGLs show reduced extracellular secretion and intracellular protein levels of IL-6 in response to pro-inflammatory stimuli LPS and IFNγ relative to WT hMGLs (H, I K, L). 2 isogenic clones per genotype (WT and SORL1 KO) and 6 independent replicates (2 differentiations and 3 technical replicates) per clone per genotype (N=12 independent replicates) were used for these experiments. Data represented as mean ± SD and analyzed using parametric two-tailed unpaired t test and 2-way ANOVA with Tukeys multiple comparison test. Significance while comparing WT to SORL1 KO hMGLs was defined and depicted as a value of *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001, ns=not significant and while comparing different treatments of each genotype was defined and depicted as a value of #p < 0.05, ##p < 0.01, ###p < 0.001, and ###p < 0.0001, ns=not significant.

Figure 7.. Working model.

We propose the following working model that underpins the abnormal lysosomal phenotypes observed in SORL1 KO hMGLs. Loss of SORL1 (orange circle) leads to decreased lysosomal degradation and decreased lysosomal enzyme activity caused by impaired trafficking of CIMPR and lysosomal enzymes from the TGN. Loss of SORL1 also causes enhanced phagocytic uptake due to increased cell surface localization of phagocytic receptors, but due to the decreased degradative capacity of the lysosomes, there is lysosomal accumulation of substrates including fibrillar Aβ 1–42 and synaptosomes. Alternate pathways of lysosomal clearance such as lysosomal exocytosis are also impaired, contributing to a blunted neuroimmune response. Altogether, loss of SORL1 causes increased lysosomal stress in human microglia.