An adapted protocol to derive microglia from stem cells and its application in the study of CSF1R-related disorders

Abstract

Background: Induced pluripotent stem cell-derived microglia (iMGL) represent an excellent tool in studying microglial function in health and disease. Yet, since differentiation and survival of iMGL are highly reliant on colony-stimulating factor 1 receptor (CSF1R) signaling, it is difficult to use iMGL to study microglial dysfunction associated with pathogenic defects in CSF1R.

Methods: Serial modifications to an existing iMGL protocol were made, including but not limited to changes in growth factor combination to drive microglial differentiation, until successful derivation of microglia-like cells from an adult-onset leukoencephalopathy with axonal spheroids and pigmented glia (ALSP) patient carrying a c.2350G > A (p.V784M) CSF1R variant. Using healthy control lines, the quality of the new iMGL protocol was validated through cell yield assessment, measurement of microglia marker expression, transcriptomic comparison to primary microglia, and evaluation of inflammatory and phagocytic activities. Similarly, molecular and functional characterization of the ALSP patient-derived iMGL was carried out in comparison to healthy control iMGL.

Results: The newly devised protocol allowed the generation of iMGL with enhanced transcriptomic similarity to cultured primary human microglia and with higher scavenging and inflammatory competence at ~ threefold greater yield compared to the original protocol. Using this protocol, decreased CSF1R autophosphorylation and cell surface expression was observed in iMGL derived from the ALSP patient compared to those derived from healthy controls. Additionally, ALSP patient-derived iMGL presented a migratory defect accompanying a temporal reduction in purinergic receptor P2Y12 (P2RY12) expression, a heightened capacity to internalize myelin, as well as heightened inflammatory response to Pam₃CSK₄. Poor P2RY12 expression was confirmed to be a consequence of CSF1R haploinsufficiency, as this feature was also observed following CSF1R knockdown or inhibition in mature control iMGL, and in CSF1R^WT/KO and CSF1R^WT/E633K iMGL compared to their respective isogenic controls.

Conclusions: We optimized a pre-existing iMGL protocol, generating a powerful tool to study microglial involvement in human neurological diseases. Using the optimized protocol, we have generated for the first time iMGL from an ALSP patient carrying a pathogenic CSF1R variant, with preliminary characterization pointing toward functional alterations in migratory, phagocytic and inflammatory activities.

Keywords: ALSP; CD68; CSF1R; Inflammation; Microglia; Migration; P2RY12; Phagocytosis; iPSC.

Fig. 1

Characterization of iMGL 2.9. A Schematic of the 2.0 and 2.9 protocols. B-H iMGL 2.0 and 2.9 differentiation were carried out side-by-side from the same healthy control iPSC lines. B Viable iMGL cell yield assessed by trypan blue exclusion assay. Cell numbers per well of a 6-well plate are presented. A t-test was performed. n = 13 differentiation batches from 6 iPSC lines. *** p < 0.001. Connecting lines show side-by-side experiments. C Crystal violet assay. A Mann–Whitney test was performed. n = 4 lines, * p < 0.05, O.D. = optical density. D Phase contrast images of iMGL 2.0, iMGL 2.9 and primary human microglia. Scale bar = 150 $\mu m$. E Representative images of IBA1 and PU.1 immunostaining. Blue = Hoechst 33342, green = IBA1 or PU.1, scale bar = 200 μm. F Quantification of IBA1- and PU.1-immunopositivity. n = 4 lines. G PCA plot of RNA-sequencing data. H Heatmap showing key microglia marker expression assessed by qRT-PCR. n = 4 donors for primary microglia, 6 lines for iMGL 2.0/2.9 and iPSCs. I Flow cytometry assessment of cell surface microglia marker expression in iMGL 2.0 and 2.9. Data from THP-1 cells are shown as negative controls. Mann–Whitney tests were performed. n = 6 lines, * p < 0.05. MFI = median fluorescence intensity

Fig. 2

Scavenging abilities of iMGL 2.9. iMGL 2.0 and 2.9 were differentiated side-by-side from the same healthy control iPSC lines. A Expression of genes encoding phagocytosis/scavenger receptors or positive regulators of phagocytosis in iMGL 2.9 compared to 2.0. Red bars represent genes that were significantly different (adjusted p < 0.05) between iMGL 2.0 and 2.9 whereas black bars represent genes that were not significantly different. n = 4 lines. B-D iMGL 2.0, iMGL 2.9 and primary human microglia were treated side-by-side with vehicle or pHrodo™ Green-labelled myelin, α-synuclein preformed fibrils (α-syn PFF), opsonized red blood cells (IgG-RBC) or E. coli for three hours and then counterstained with Hoechst 33342. B Representative fluorescence images. Scale bar = 50 $\mu m$. C-D Quantification of mean green fluorescence intensity (MFI) per cell. Mann–Whitney tests were performed in C. n = 4 lines, * p < 0.05 in C. Average values from n = 2 lines (iMGL) or donors (primary microglia) are presented in D

Fig. 3

TLR4 signaling response of iMGL 2.9 following LPS treatment. iMGL 2.0 and 2.9 were differentiated side-by-side from the same healthy control iPSC lines. A IL-6, TNF and IL-10 concentrations in supernatants from iMGL 2.0, iMGL 2.9 and primary human microglia treated with vehicle or LPS (100 ng/mL) for 24 h. Mann–Whitney tests were performed. n = 5 lines for iMGL, n = 6 donors of primary microglia, * p < 0.05, ** p < 0.01, *** p < 0.001. B Representative Western blotting and quantification of NF-κB in iMGL 2.9 treated with vehicle or LPS (100 ng/mL) for 1.5 h. A Mann–Whitney test was performed. n = 4 lines, * p < 0.05. C Flow cytometry assessment of cell surface CD14 and TLR4 expression in iMGL 2.0 and 2.9. Mann–Whitney tests were performed. n = 6 lines, * p < 0.05. D Heatmap showing expression of genes encoding pattern recognition receptor and their co-receptors. A two-way ANOVA was performed. n = 4 lines, *** p < 0.001 vs iMGL 2.0. E Representative Western blotting and quantification of inflammasome components in iMGL 2.9 treated with vehicle or LPS (100 ng/mL) for 24 h. Mann–Whitney tests were performed. n = 5 lines, * p < 0.05. F IL-1β and TNF concentrations in cell supernatants of human primary microglia and iMGL 2.9 treated with vehicle or LPS (100 ng/mL) for 24 h, followed by ATP (5 mM) or not for 30 min. Kruskal–Wallis tests followed by Dunn’s multiple comparison tests were performed. n = 4 donors of primary microglia and n = 5 lines for iMGL, * p < 0.05, ** p < 0.01, ns = non-significant. G Heatmap showing the expression of genes encoding NLRP3 inflammasome components and their substrates. n = 4 donors of primary microglia, 4 lines for iMGL and iPSCs

Fig. 4

Pedigree of the proband’s family, MRI images, and genotyping data. A Pedigree of the proband’s family (prepared via https://cegat.com/). Shading indicates carriers of the c.2350G > A (p.V784M) CSF1R variant with ALSP diagnosis. Symbol with a dot indicates a carrier without clinical manifestations of ALSP. B-E Brain MRI of subject II.3. B Sagittal FLAIR T2-weighted MR image showing thinning and hyperintense signal of the anterior portion of the body of the corpus callosum (red circle). C-D Axial T2-weighted MR images showing the presence of bilateral multifocal and confluent lesions in the frontal lobes (red circles) and multifocal lesions in the left fontal and parietal lobe with areas of restricted diffusion in the diffusion weighted image (E, red circles). F–H Brain MRI of a healthy control, matched by sex and age to subject II.3. F Sagittal T1-weighted image showing normal size and signal of the brain parenchyma, specifically of the corpus callosum. G-H Axial T2-weighted images showing no atrophy, normal ventricular size and normal signal at the level of the cerebral white matter. (I) Chromatogram of PBMC-derived DNA showing heterozygous c.2350G > A CSF1R variant (shown here as position 190)

Fig. 5

Derivation of microglia-like cells from the ALSP-CSF1R patient with a c.2350G > A (p.V784M) CSF1R variant. A Phase contrast images of ALSP-CSF1R iPSC line differentiated into iMGL using the 2.0 or 2.9 protocol. Scale bar = 150 $\mu m$. B Viable cell yield per well of a 6-well plate assessed by trypan blue exclusion assay following 2.0 and 2.9 differentiation of the ALSP-CSF1R iPSCs into iMGL. n = 7 differentiation batches. C Viable cell yield per well of a 6-well plate assessed by trypan blue exclusion assay following side-by-side differentiation of healthy control lines (n = 14 differentiation batches from 6 lines) and the ALSP-CSF1R line (n = 14 differentiation batches) using the 2.9 protocol. A t-test was performed, * p < 0.05. (D) Quantification of IBA1- and PU.1-immunopositivity. n = 4 healthy control lines and 4 batches of a single ALSP-CSF1R line, differentiated side-by-side using the 2.9 protocol. E qRT-PCR assessment of microglia marker expression on day 0, 14 and 28 of microglial differentiation. n = 4 healthy control lines and 4 batches of a single ALSP-CSF1R line, differentiated side-by-side using the 2.9 protocol. Two-way ANOVA were performed, followed by Sidak’s post hoc tests. *** p < 0.001. F Flow cytometry assessment of P2RY12, CX3CR1 and MerTK cell surface expression. T-tests were performed. n = 4 healthy control lines and 4 batches of a single ALSP-CSF1R line, differentiated side-by-side using the 2.9 protocol. ** p < 0.01. G Flow cytometry assessment of CSF1R cell surface expression. A t-test was performed. n = 4 healthy control lines and 4 batches of a single ALSP-CSF1R line, differentiated side-by-side using the 2.9 protocol. * p < 0.05. H Western blot assessment of CSF1R and its tyrosine 723-phosphorylated form, and GAPDH. A t-test was performed. n = 5 healthy control lines and 5 batches of a single ALSP-CSF1R line, differentiated side-by-side using the 2.9 protocol. ** p < 0.01

Fig. 6

Functional phenotype of iMGL derived from the ALSP-CSF1R patient with a c.2350G > A (p.V784M) CSF1R variant. All iMGL were generated using the 2.9 protocol. Quantification (A) and images (B) of iMGL migratory activity toward ADP assessed by Boyden chamber assay. Cells were concomitantly treated or not with PSB0739. Kruskal–Wallis tests followed by Dunn’s multiple comparison tests were performed. n = 6 healthy control lines and 6 batches of a single ALSP patient line, differentiated side-by-side. White = Hoechst 33342, scale bar = 75 μm in (B). Quantification of green fluorescence intensity per cell (C) and representative images (D) of iMGL exposed to vehicle or pHrodo.™ Green-labelled myelin, opsonized red blood cells (IgG-RBC) or E. coli for three hours and then counterstained with Hoescht 33342. T-tests were performed. n = 3 healthy control lines and 3 batches of a single ALSP patient line, differentiated side-by-side, * p < 0.05. Scale bar = 250 $\mu m$ in (D). Quantification of mean fluorescence intensities (MFI; E) and representative images (F) of LAMP1, CD68 and IBA1 immunostaining of iMGL. T-tests were performed. n = 3 healthy control lines and 3 batches of a single ALSP patient line, differentiated side-by-side, ** p < 0.01, *** p < 0.001. Scale bar = 100 $\mu m$ in (F). G Cytokine secretion assessed in cell supernatants following a 24-h treatment with vehicle, LPS (100 ng/mL) or Pam₃CSK₄ (100 ng/mL). A two-way ANOVA was performed, followed by Tukey’s post hoc test. n = 4 healthy control lines and 4 batches of a single ALSP patient line, differentiated side-by-side, *** p < 0.001, ns = non-significant. H qRT-PCR performed after three hours of Pam₃CSK₄ (100 ng/mL) vs. vehicle treatment. T-tests were performed. n = 3 healthy control lines and 3 batches of a single ALSP patient line, differentiated side-by-side, * p < 0.05

Fig. 7

Effect of CSF1R loss of function on microglial phenotype. A-C CSF1R^WT/WT and CSF1R^WT/KO iPSCs were differentiated into iMGL side-by-side. A Viable cell yield per well of a 6-well plate assessed by trypan blue exclusion assay following 2.0 and 2.9 differentiation. A Kruskal–Wallis test was performed, followed by Dunn’s post hoc test. n = 4 differentiation batches. B Phase contrast images of iMGL 2.9. Red triangles show presence of round or dysmorphic floating cells. Scale bar = 150 $\mu m$. C Flow cytometry assessment of cell surface marker expression. Kruskal–Wallis tests were performed, followed by Dunn’s post hoc test. n = 4 differentiation batches using the 2.9 protocol, * p < 0.05, ** p < 0.01 vs. unstained control. MFI = median fluorescence intensity, a.u. = arbitrary unit. D-G CSF1R^WT/REV and CSF1R^WT/E633K iPSCs were differentiated into iMGL side-by-side. D Viable cell yield per well of a 6-well plate assessed by trypan blue exclusion assay following 2.0 and 2.9 differentiation. A Kruskal–Wallis test was performed, followed by Dunn’s post hoc test. n = 4 differentiation batches. E Phase contrast images of iMGL 2.9. Scale bar = 150 $\mu m$. F qRT-PCR assessment of microglia markers. T-tests were performed. n = 5 differentiation batches using the 2.9 protocol, * p < 0.05 vs CSF1R^WT/REV iMGL. G Flow cytometry assessment of cell surface marker expression. Mann–Whitney tests were performed. n = 5 differentiation batches using the 2.9 protocol * p < 0.05, ** p < 0.01. MFI = median fluorescence intensity. H-I Healthy control, mature iMGL were generated following the 2.9 protocol. H Microglia marker expression assessed by qRT-PCR in iMGL three days after siCON or siCSF1R transfection. T-tests were performed. n = 3 lines, * p < 0.05 vs siCON. I Microglia marker expression assessed by qRT-PCR in iMGL following a 3-day treatment with PLX3397 (1 μM, every other day). T-tests were performed. n = 4 lines, * p < 0.05 vs vehicle treatment. J-L eGFP and either WT or V784M CSF1R were stably co-expressed in ALSP-CSF1R iHPCs using lentiviruses and cells were differentiated into iMGL following the 2.9 protocol. J Merged phase contrast and green fluorescence images of ALSP-CSF1R iMGL. Scale bar = 150 μm. (K) Viable cell yield per well of a 6-well plate assessed by trypan blue exclusion assay. A t-test was performed. n = 4 differentiation batches. L Flow cytometry assessment of eGFP signal and microglia marker expression. MFI = median fluorescence intensity. n = 1 differentiation batch