Fig 2: Increased Fcγ‐receptors expression and enhanced Syk phosphorylation in TREM2‐deficient BMDM Representative histograms for Fcγ‐receptors‐PE (FcγR‐PE) expression levels as used for quantification. Stacked histograms for log PE fluorescence intensity of wt and Trem2 ko BMDM are shown for the respective FcγR (I, IIB/III, and IV).Relative quantification of cell surface levels of FcγR molecules. Absolute number of cell surface FcγR‐PE molecules was determined by the BD QuantiBRITE® method (see methods section for details) and normalized to expression levels of the respective wt control. (n = 4, ± SEM, t‐test, two‐tailed; wt vs. ko: FcγRI‐PE P = 0.0008, FcγRIIB/III‐PE P = 0.0033, FcγRIV‐PE P = 0.7001).mRNA levels of FcγR are increased in Trem2 ko BMDM. Fold changes of the respective FcγR (I, IIB, III, IV) mRNA levels in Trem2 ko BMDM were determined by quantitative real‐time PCR. (n = 6, ± SEM; one‐sample t‐test, two‐tailed; wt vs. ko: FcγRI P = 0.0118, FcγRIIB P = 0.0006, FcγRIII P = 0.0004, FcγRIV P = 0.0284)Phosphorylated Syk (P‐Syk) and total Syk (T‐Syk) levels were determined by Western blotting in lysates from wt and Trem2 ko BMDM after 1 h treatment with Aβ alone, together with antibody 2D8 (Aβ‐2D8), or an isotype control (Aβ‐IC). Actin was used as a loading control.Quantification of P‐Syk normalized to T‐Syk. (n = 4, ± SEM; 2 way ANOVA, interaction P = 0.0490, genotype P = 0.0898, treatment P < 0.0001. Bonferroni‐corrected pair‐wise post hoc tests, **P = 0.0075 vs. wt). Source data are available online for this figure.

Fig 3: Compensation of reduced amyloid plaque clearance by elevated antibody dose Cryosections from unfixed brain of 6‐month‐old APP/PS1 mice were pre‐incubated with increasing concentrations of mAb11 (0.001, 0.01, 0.1, 1, 5 μg/ml). BMDM from wt or Trem2 ko mice were added for 24 h. Sections were stained with methoxy‐X04. Scale bar, 500 μm.Methoxy‐X04 signals were quantified from the entire sagittal section. (n = 5, ± SEM; two‐way ANOVA, interaction P = 0.0082, genotype P < 0.0001, treatment P < 0.0001. Fisher's LSD post hoc comparisons; * show statistics between wt and ko under the same experimental condition. # in black shows wt compares to no‐antibody stimulation; # in gray shows ko compares to no‐antibody stimulation; wt vs. ko for the following conditions: no antibody P = 0.0053, mAb11 0.001 μg/ml P = 0.0003, mAb11 0.01 μg/ml P < 0.0001, mAb11 0.1 μg/ml P = 0.0001, mAb11 1 μg/ml P = 0.0007, mAb11 5 μg/ml P = 0.0011; following conditions compare to wt/no antibody: wt/mAb11 0.01 μg/ml P = 0.0166, wt/mAb11 0.1 μg/ml P = 0.0002, wt/mAb11 1 μg/ml P < 0.0001, wt/mAb11 5 μg/ml P < 0.0001; following conditions compare to ko/no antibody: ko/mAb11 0.1 μg/ml P = 0.0099, ko/mAb11 1 μg/ml P < 0.0001, ko/mAb11 5 μg/ml P < 0.0001.

Fig 4: TREM2 deficiency reduces efficacy of antibody‐stimulated Aβ uptake by phagocytic cells Schematic of the mouse Trem2 locus and the TREM2 protein. Sequence alignment of wild‐type N9 (N9 wt) and TREM2 mutant N9 (N9 mu) surrounding the gRNA target site. The gRNA sequence is in cyan, and protospacer‐adjacent motif (PAM) is marked with a line. The single nucleotide insertion is labeled in red. Schematic representation of wild‐type TREM2 (NP_112544.1) and CRISPR/Cas9‐modified TREM2 (N9 mu). TM, transmembrane domain; SP, signal peptide.Western blot analysis of lysates and media from wt and mutant N9 cells (N9 wt /mu) using the antibody anti‐murine TREM2 (clone 5F4), which is raised against the murine TREM2 extracellular domain. sTREM2, soluble TREM2. *indicate unspecific bands. Calnexin was used as a loading control.Phagocytosis of 1 μM HiLyte™ Fluor 488 Aβ1‐42 (fAβ42) by N9 wt and N9 mu in the presence or absence of antibody 2D8 or the non‐binding antibody 6687. Cytochalasin D (CytoD, 10 mM) was used as control to verify phagocytic uptake. (n = 4, ± SEM; two‐way ANOVA, interaction P = 0.61, genotype P < 0.0001, treatment P = 0.0001; post hoc tests wt vs. mu for the following conditions: fAβ42 P = 0.0043, fAβ42‐2D8 P = 0.0436).Western blot of BMDM derived from wt and Trem2 knockout (ko) animals using antibody 5F4. *indicate unspecific bands.Phagocytosis of fAβ42 by BMDM from wt and Trem2 ko animals in the presence or absence of 2D8, or the non‐binding control antibody 6687. (n = 3, ± SEM; two‐way ANOVA, interaction P = 0.0005, genotype P < 0.0001, treatment P < 0.0001; post hoc tests wt vs. ko for the following conditions: fAβ42 P = 0.0021, fAβ42‐2D8 1 μg/ml P < 0.0001, fAβ42‐2D8 5 μg/ml P < 0.0001, fAβ42‐2D8 10 μg/ml P < 0.0001, fAβ42/6687 10 μg/ml P = 0.0007).Quantification of relative fAβ42 uptake to lowest antibody concentration used (n = 3, ± SEM).Phagocytosis of fAβ42 by BMDM from wt and Trem2 ko animals in the presence or absence of mAb11, or an isotype control antibody (IC). (n = 4, ± SEM; two‐way ANOVA, interaction P = 0.0223, genotype P < 0.0001, treatment P < 0.0001; post hoc tests wt vs. ko for the following conditions: fAβ42‐mAb11 1 μg/ml P = 0.0391, fAβ42‐mAb11 5 μg/ml P = 0.0069, fAβ42‐mAb11 10 μg/ml P < 0.0001, fAβ42‐mAb11 20 μg/ml P = 0.0001, fAβ42‐mAb11 50 μg/ml P < 0.0001).Quantification of relative fAβ42 uptake to lowest antibody concentration used (n = 4, ± SEM).Recombinant mouse sTREM2 does not rescue fAβ42 uptake in Trem2‐deficient BMDM. Increasing amounts of sTREM2 were added to the media of wt or Trem2 ko BMDM in the presence or absence of mAb11 (10 μg/ml) (n = 4, ± SEM).Western blot of primary microglia from wt or Trem2 ko animals using antibody 5F4. *indicate unspecific bands.Phagocytosis of fAβ42 by primary microglia from wt and Trem2 ko animals in the presence or absence of mAb11, or an isotype control antibody (IC). (n = 5, ± SEM; two‐way ANOVA, interaction P = 0.4797, genotype P < 0.0001, treatment P < 0.0001; post hoc tests wt vs. ko for the following conditions: fAβ42‐mAb11 5 μg/ml P = 0.0449, fAβ42‐mAb11 10 μg/ml P = 0.0370, fAβ42‐mAb11 20 μg/ml P = 0.0299, fAβ42‐mAb11 50 μg/ml P = 0.0120).Quantification of relative fAβ42 uptake to lowest antibody concentration used (n = 5, ± SEM).Data information: (C, E, G, K) Quantification of internalized fAβ42 was normalized to wt without antibody. Bonferroni‐corrected pair‐wise post hoc tests were used.Source data are available online for this figure.