Functional Studies of Missense TREM2 Mutations in Human Stem Cell-Derived Microglia

Abstract

The derivation of microglia from human stem cells provides systems for understanding microglial biology and enables functional studies of disease-causing mutations. We describe a robust method for the derivation of human microglia from stem cells, which are phenotypically and functionally comparable with primary microglia. We used stem cell-derived microglia to study the consequences of missense mutations in the microglial-expressed protein triggering receptor expressed on myeloid cells 2 (TREM2), which are causal for frontotemporal dementia-like syndrome and Nasu-Hakola disease. We find that mutant TREM2 accumulates in its immature form, does not undergo typical proteolysis, and is not trafficked to the plasma membrane. However, in the absence of plasma membrane TREM2, microglia differentiate normally, respond to stimulation with lipopolysaccharide, and are phagocytically competent. These data indicate that dementia-associated TREM2 mutations have subtle effects on microglia biology, consistent with the adult onset of disease in individuals with these mutations.

Keywords: Nasu-Hakola disease; TREM2; dementia; frontotemporal dementia; iPSC-microglia; microglia; neuroinflammation.

Figure 1

An Efficient Protocol for the Generation of Microglia from Pluripotent Stem Cells (A) PSCs are differentiated to microglia via embryoid bodies and PMPs. PMPs are produced continuously in culture and are terminally differentiated into microglia when required. (B) A high proportion of stem cell-derived microglia express the microglial/macrophage markers Iba1, CD45, and TREM2. Scale bars represent 100 μm, except PSC and embryoid bodies (1 mm). n = 5–6 biological replicates. Error bars represent SDs.

Figure 2

The Transcriptome of Human Stem Cell-Derived Microglia Is Similar to Primary Microglia (A) At the whole-transcriptome level, iPSC-derived primitive macrophage precursors (iPMPs) and iPSC-derived microglia (iMG) cluster with primary microglia cultured in vitro (In vitro MG (2)). (B) When compared using a subset of genes enriched in murine microglia over other CNS myeloid cells (Bennett et al., 2016), iMG and iPMPs again cluster with primary microglia cultured in vitro (In vitro fetal MG [fMG], In vitro MG (1), and In vitro MG (2)), and additionally with an alternative iMG method (Alternate iMG; Abud et al., 2017) and monocyte-derived macrophages (MDM). (C) FPKM (fragments per kilobase million) counts of “microglia signature” genes indicate comparable levels of expression in iMG compared with in vitro and ex vivo primary microglia. Monocytes, CD14+/CD16− monocytes (Abud et al., 2017); MDM, monocyte-derived macrophages (Zhang et al., 2015); iPMPs, iPSC-derived primitive macrophage precursors (this study); iMG, iPSC-derived microglia (this study); iMG (alternate), iPSC-microglia derived using an alternative method (Abud et al., 2017); In vitro fMG, in vitro fetal microglia (Abud et al., 2017); In vitro MG (1), in vitro microglia (Abud et al., 2017); In vitro MG (2), in vitro microglia (Gosselin et al., 2017); Ex vivo MG/Mac, primary sorted CNS CD45+ microglia/macrophages (Zhang et al., 2016); Ex vivo MG, primary sorted microglia (Gosselin et al., 2017). For iPMPs and iMG, n = 3 independent differentiations of 2 genetic backgrounds (n = 5 iPMPs, n = 6 iMG). In (A) and (B), genetic backgrounds of iPMPs and iMG are distinguished by square and triangle symbols. For (C), iMG FPKM values were averaged across differentiations to give values for each genetic background (n = 2). Error bars represent SDs.

Figure 3

Stem Cell-Derived Microglia Are Functionally Similar to Primary Microglia (A) Microglia efficiently phagocytose pHrodo-E. coli, in a process sensitive to cytochalasin D. (B) Upon exposure to 100 ng/mL LPS, microglia secrete pro-inflammatory cytokines; an effect augmented by interferon γ (IFN-γ). (C) Microglia migrate into preformed cortical organoids. Upon migration, microglia tessellate throughout organoids and assume a pronounced ramified morphology, which is demonstrated by live 2-photon imaging of organoid/microglia co-cultures. Scale bars represent 200 μm (A) and 100 μm (C), while scale grid markings at high magnification represent 12.4 μm. In (B), n = 3 biological replicates; ^∗ p < 0.05, ^∗∗ p < 0.01 treatment versus control; Dunnett's post hoc test. Error bars represent SDs.

Figure 4

The TREM2 Receptor Is Mislocalized and Aberrantly Processed in TREM2 Mutant Microglia (A) Staining for the TREM2 receptor with an N-terminally directed primary antibody reveals expression of TREM2 protein in microglia from TREM2 wild-type and mutant microglia when cells are permeabilized. Upon omission of permeabilization, TREM2 receptor surface staining is only detected in TREM2 wild-type and heterozygous mutant microglia, and is absent in microglia derived from homozygous mutant backgrounds. (B) Probing of whole-cell lysates reveals expression of immature and mature forms of the full-length (F/L) TREM2 receptor in wild-type microglia. Mutations in TREM2 cause a gene-dosage-dependent accumulation of immature TREM2, and a reduction in mature forms of TREM2. (C) Probing with a C-terminally directed antibody reveals a weak band corresponding to the CTF of TREM2 in wild-type microglia, indicating efficient turnover of the CTF. Overnight treatment with DAPT (10 μM) results in accumulation of the TREM2-CTF in TREM2 wild-type microglia, a barely detectable accumulation in heterozygous mutant microglia and no detectable accumulation in homozygous mutant microglia. Scale bar represents 50 μm.

Figure 5

Microglia Harboring TREM2 Mutations Respond Appropriately to LPS Challenge (A) Upon exposure to LPS, microglia from TREM2 wild-type and mutant backgrounds release pro-inflammatory cytokines in a dose-dependent manner. (B) Further experiments in two additional clones from patient T66M/T66M confirm a similar response to LPS challenge across all clones of this genotype. Results in (A) are an average of all data from two independently performed treatments, n = 2 biological replicates per genotype and error bars represent SDs. Error bars in (B) represent SE of 2–3 technical well replicates, and x-axis values are LPS concentration in pg/mL.

Figure 6

Microglia from TREM2 Mutant Backgrounds Are Phagocytically Competent (A and B) Homozygous TREM2 mutant microglia phagocytose pHrodo-E. coli with a similar efficiency and capacity to microglia from a wild-type background. In contrast, cytochalasin D significantly attenuates pHrodo-E. coli uptake. (C and D) Similarly, after serum starvation, TREM2 mutant microglia phagocytose E. coli as efficiently as TREM2 wild-type microglia. (E–G) TREM2 mutant microglia internalize AcLDL almost as efficiently as TREM2 wild-type microglia. Scale bar represents 200 μm. All data from two (C, D, and E–G) or three (A and B) independent experiments were averaged to produce values for each genotype. n = 3 (TREM2 wild-type-untreated; TREM2 wild-type cytochalasin D-treated) or 2 (TREM2 heterozygous; TREM2 homozygous) biological replicates. ^∗p < 0.05, ^∗∗p < 0.01 versus untreated TREM2 wild-type microglia, Dunnett's post hoc test. n.s., not significant. Error bars represent SDs.