Disease Progression-Dependent Effects of TREM2 Deficiency in a Mouse Model of Alzheimer's Disease

Abstract

Neuroinflammation is an important contributor to Alzheimer's disease (AD) pathogenesis, as underscored by the recent identification of immune-related genetic risk factors for AD, including coding variants in the gene TREM2 (triggering receptor expressed on myeloid cells 2). Understanding TREM2 function promises to provide important insights into how neuroinflammation contributes to AD pathology. However, studies so far have produced seemingly conflicting results, with reports that amyloid pathology can be both decreased and increased in TREM2-deficient AD mouse models. In this study, we unify these previous findings by demonstrating that TREM2 deficiency ameliorates amyloid pathology early, but exacerbates it late in disease progression in the APPPS1-21 mouse model of AD. We also demonstrate that TREM2 deficiency decreases plaque-associated myeloid cell accumulation by reducing cell proliferation, specifically late in pathology. In addition, TREM2 deficiency reduces myeloid cell internalization of amyloid throughout pathology, but decreases inflammation-related gene transcript levels selectively late in disease progression. Together, these results suggest that TREM2 plays distinct functional roles at different stages in AD pathology.

Significance statement: Alzheimer's disease (AD) is a devastating neurodegenerative disorder and there are currently no effective treatments that modify disease progression. However, the recent identification of genetic risk factors for AD promises to provide new insight into AD biology and possible new therapeutic targets. Among these risk factors, variants in the gene TREM2 (triggering receptor expressed on myeloid cells 2) confer greatly elevated risk for developing the disease. We demonstrate that TREM2 deficiency has opposing effects on AD-related pathologies at early and late stages of disease progression, unifying previous work in the field. In addition, we examine how TREM2 affects the function of the brain immune cell populations in which it is expressed throughout disease progression to understand possible mechanisms underlying its differential impacts on pathology.

Keywords: AD; genetic risk factor; inflammation; microglia; myeloid cell; neuroinflammation.

Figure 1.

TREM2 deficiency reduces amyloid pathology early but exacerbates it late in disease progression. a, b, The 6E10-immunoreactive area was assessed in the cortex of APPPS1;Trem2^+/+ and APPPS1;Trem2^−/− mice at 2 months of age (a) and 8 months of age (b). c, d, Thioflavin S+ plaque number per square millimeter was assessed in APPPS1;Trem2^+/+ and APPPS1;Trem2^−/− mice at 2 months of age (c) and 8 months of age (d). e, There was a significant reduction in 6E10 immunoreactive area in the cortex of 2-month-old TREM2-deficient mice (WT, n = 3 M/2 F; KO, n = 3 M/2 F) and a significant increase in 6E10 immunoreactive area in the cortex of 8-month-old TREM2-deficient mice compared with controls (WT, n = 4 M/2 F; KO, n = 4 M/3 F). f, There was a trend toward a reduction in the density of ThioS+ plaque number in the cortex of 2-month-old APPPS1;Trem2^−/− cortex compared with APPPS1;Trem2^+/+ controls (WT, n = 4 M/2 F; KO, n = 3 M/3 F), but no significant difference in ThioS+ plaque density in the cortex of 8-month-old TREM2-deficient mice (WT, n = 4 M/2 F; KO, n = 5 M/3 F). g, Analysis of 6E10+ plaque size revealed no significant differences in 2-month-old TREM2-deficient mice (WT, n = 3M/2F; KO, n = 3M/3F), but a significant increase in 8-month-old TREM2-deficient mice compared with controls (WT, n = 4 M/2 F; KO, n = 4 M/2 F). h, Transcript levels of human APP were assessed in 2-month-old mice (WT, n = 4 M/2 F; KO, n = 3 M/3 F) and 8-month-old mice (WT, n = 4 M/2 F; KO, n = 3 M/3 F). i, Western blots were used to examine the protein levels of human APP using 6E10 in cortical lysates from 2-month-old mice (WT, n = 3 M/2 F; KO, n = 3M/2F) and 8-month-old mice (WT, n = 3 M/2 F; KO, n = 3 M/2 F) mice. j, APP protein levels were normalized to actin and the fold change expressed to the WT for each age. ns, Not significant; *p < 0.05; **p < 0.01.

Figure 2.

TREM2 deficiency reduces myeloid cell number. a, Analysis of myeloid cell markers by qPCR revealed no significant changes in expression of the myeloid cell markers PU.1, Tmem119, or CD45 in cortical lysates from 2-month-old APPPS1;Trem2^−/− mice compared with APPPS1;Trem2^+/+ controls (WT, n = 3 M/3 F; KO, n = 3 M/3 F). However, there was a trend toward reduction in expression levels of PU.1 and Tmem119 in 8-month-old TREM2-deficient mice and a significant reduction in CD45 expression (WT, n = 2 M/2 F; KO, n = 3 M/4 F). b–d, Immunohistochemistry for CD45 at 2 months of age (b; WT, n = 3 M/2 F; KO, n = 3 M/4 F) and 8 months of age (c; WT, n = 4M/2F; KO, n = 4M/2F) revealed a significant decrease in the density of cells expressing high levels of CD45 in the cortex of APPPS1;Trem2^−/− mice compared with APPPS1;Trem2^+/+ controls (d). ns, Not significant; **p < 0.01; ***p < 0.001.

Figure 3.

TREM2 deficiency reduces plaque-associated myeloid cells. a, b, Immunohistochemistry for Iba1 and 6E10 was performed to assess accumulation of myeloid cells around plaques in the cortex of 2-month-old (a) and 8-month-old (b) APPPS1;Trem2^−/− and APPPS1;Trem2^+/+ mice. c–e, Quantification of confocal images from 2-month-old (c) and 8-month-old mice (d) revealed significant reductions in the number of Iba1+ cells around plaques in TREM2-deficient mice at both 2 months of age (WT, n = 3 M/2 F; KO, n = 3 M/3 F) and 8 months of age (WT, n = 2 M/1 F; KO, n = 2 M/1 F; e). f–h, Quantification of CD45-immunoreactive area around Congo Red+ plaques in 2-month-old (f) and 8-month-old (g) APPPS1;Trem2^+/+ and APPPS1;Trem2^−/− mice revealed significant reductions in the CD45^hi cell accumulation around plaques in TREM2-deficient mice at both 2 (WT, n = 3M/2F; KO, n = 3M/3F) and 8 (WT, n = 4M/2F; KO, n = 4M/2F) month time points (h). **p < 0.01; ***p < 0.001.

Figure 4.

TREM2 deficiency reduces myeloid cell proliferation late in disease progression. Mice were injected with 10 mg/kg BrdU intraperitoneally every 24 h for 72 h and killed 24 h after the last injection. a, b, BrdU immunohistochemistry was used to identify proliferating cells in 2-month-old (a) and 8-month-old (b) APPPS1;Trem2^−/− and APPPS1;Trem2^+/+ mice. c, Number of BrdU+ Iba1+ double-positive cells was quantified in the cortex of 2-month-old (WT, n = 1 M/2 F; KO, n = 1 M/1 F) and 8-month old (WT, n = 3 M/1 F; KO, n = 5 M/3 F) APPPS1;Trem2^−/− and APPPS1;Trem2^+/+ mice. There were no significant differences in the numbers of proliferating myeloid cells in 2-month-old TREM2-deficient mice, but there were significant decreases in 8-month-old APPPS1;Trem2^−/− mice compared with APPPS1;Trem2^+/+ controls. d, Brain myeloid cells were isolated from 6- to 9-month-old APPPS1 mice and cells were analyzed by flow cytometry. e, Cells were gated on CD11b and divided into CD45^lo and CD45^hi cells. f, When the percentage of BrdU+ cells was quantified within that CD45^hi population, there were very few proliferating cells (0.43%). g, However, there was a substantial population (7.53%) of proliferating CD45^lo cells. ***p < 0.001.

Figure 5.

TREM2 deficiency reduces amyloid internalization within myeloid cells. a, b, Confocal slices through amyloid plaques were acquired 1 μm apart for up to 30 slices in 2-month-old (a) and 8-month-old (b) APPPS1;Trem2^−/− and APPPS1;Trem2^+/+ mice. The total volume of colocalization between Iba1 and 6E10 was measured and normalized to plaque volume. c, This normalized volume was reduced significantly in TREM2-deficient mice at 2 months of age (WT, n = 2 M/1 F; KO, n = 2 M/1 F) and 8 months of age (WT, n = 1 M/2 F; KO, n = 3M). *p < 0.05.

Figure 6.

TREM2 deficiency reduces astrocytosis and inflammation. a, b, Immunohistochemistry for GFAP and 6E10 was performed to assess astrocytosis in 2-month-old (a) and 8-month-old (b) APPPS1;Trem2^−/− and APPPS1;Trem2^+/+ mice. c, Quantification of the GFAP-immunoreactive area around plaques revealed significant reductions in TREM2-deficient mice at 2 months of age (WT, n = 3 M/2 F; KO, n = 3 M/3 F) and 8 months of age (WT, n = 4 M/2 F; KO, n = 4 M/3 F). d–f, S100β-immunoreactive area around 6E10+ plaques was also examined at 2 months of age (d) and 8 months of age (e) and revealed no significant differences between genotypes at either time point (f; 2 M WT, n = 3 M/ 2F; 2 M KO, n = 3 M/3 F; 8 M WT, n = 4 M/2 F; 8 M KO, n = 3 M/2 F). g, Expression of inflammation-related transcripts was quantified using qPCR. There were no significant changes in expression of these genes in TREM2-deficient mice at 2 months of age (WT, n = 3 M/3 F; KO, n = 3 M/3 F). However, there was a significant reduction in the pro-inflammatory genes IL1β and TNFα and a significant increase in the anti-inflammatory marker Fizz1 in TREM2-deficient mice at 8 months of age (WT, n = 2 M/2 F; KO, n = 3 M/4 F). *p < 0.05; **p < 0.01; ***p < 0.001.