Peripheral sTREM2-Related Inflammatory Activity Alterations in Early-Stage Alzheimer's Disease

Abstract

Alzheimer's disease (AD) has been linked to multiple immune system-related genetic variants. Triggering receptor expressed on myeloid cells 2 (TREM2) genetic variants are risk factors for AD and other neurodegenerative diseases. In addition, soluble TREM2 (sTREM2) isoform is elevated in cerebrospinal fluid in the early stages of AD and is associated with slower cognitive decline in a disease stage-dependent manner. Multiple studies have reported an altered peripheral immune response in AD. However, less is known about the relationship between peripheral sTREM2 and an altered peripheral immune response in AD. The objective of this study was to explore the relationship between human plasma sTREM2 and inflammatory activity in AD. The hypothesis of this exploratory study was that sTREM2-related inflammatory activity differs by AD stage. We observed different patterns of inflammatory activity across AD stages that implicate early-stage alterations in peripheral sTREM2-related inflammatory activity in AD. Notably, fractalkine showed a significant relationship with sTREM2 across different analyses in the control groups that was lost in later AD-related stages with high levels in mild cognitive impairment. Although multiple other inflammatory factors either differed significantly between groups or were significantly correlated with sTREM2 within specific groups, three inflammatory factors (fibroblast growth factor-2, GM-CSF, and IL-1β) are notable because they exhibited both lower levels in AD, compared with mild cognitive impairment, and a change in the relationship with sTREM2. This evidence provides important support to the hypothesis that sTREM2-related inflammatory activity alterations are AD stage specific and provides critical information for therapeutic strategies focused on the immune response.

Figure 1:. Plasma sTREM2 and inflammatory factors by disease status.

Soluble TREM2 was significantly lower in AD, compared to MCI (A). Two immunoregulatory inflammatory factors, were significantly lower in AD, compared to MCI (B). Four anti-inflammatory factors exhibited significantly different levels by disease status (C). Twelve pro-inflammatory factors exhibited significantly different levels by disease status (D). Levels were compared between disease groups by Mann-Whitney tests. Here asterisks denote significance upon outlier removal. Outliers were removed per disease group for each inflammatory factor using the ROUT method. See Table II for p-values with and without outliers. P-value notations: *p<0.05; **p<0.01; ***p<0.001.

Figure 2:. Inflammatory factors by sTREM2 tertile status.

sTREM2 Tertile 1 showed a range in sTREM2 MFI of 16.75–28.25, and Tertile 3 had a range of 43.0–108.0 MFI (A). The immunoregulatory/ pleiotropic factors that were significantly higher in the high sTREM2 tertile included: fractalkine, IL-6, IL-7, IL-9, ad TGF-α (B). The anti-inflammatory factors that were significantly higher in the high sTREM2 tertile included: FGF-2, G-CSF, IL-4, IL-5, IL-10, and IL-13 (C). The pro-inflammatory factors that were significantly higher in the high sTREM2 tertile included: EGF, GM-CSF, GRO, IFNα2, IFNγ, IL-1α, IL-1β, IL-2, IL-3, IL-8, IL-12p40, IL-12p70, IL-15, IL-17A, IP-10, MCP-3, MIP-1α, sCD40L, TNFα, TNFβ, and VEGF (D). Levels were compared between sTREM2 Tertile 1 (low) and sTREM2 Tertile 1 (high) by Mann-Whitney tests after removal of outliers. Table 3 shows numbers of outliers removed per disease group for each cytokine using the ROUT method as well as significance between groups. *p<0.05; **p<0.01; ***p<0.001; ****p<0.0001.

Figure 3:. Inflammatory factor correlation matrices by sTREM2 tertile.

sTREM2 and each cytokine were correlated to each other in (A) sTREM2 Tertile 1 (n=67) and (B) sTREM2 Tertile 3 (n=67). Data used is without outlier removal (see Supplemental Table I for p-values with and without outlier removal). Steiger tests to compare matrices indicate significant differences between low and high sTREM2 tertiles. Correlation directions are shown by a color gradient; positive correlations are shown in blue and negative correlations are in red. Significant correlations are noted by displaying asterisks for p values. *p<0.05; **p<0.01; ***p<0.001.

Figure 4:. Number of individuals in each sTREM2 tertile.

CN, MCI, and AD participants were separated into tertiles based on low (Tertile 1) and high (Tertile 3) levels of plasma sTREM2. There were significantly different numbers of CN vs MCI (A), and MCI vs AD (B), but not CN vs AD (C) by Fisher’s exact test.

Figure 5.. ATN distribution within disease group:

Participants in each disease category with CSF AD-related biomarker data were classified into ATN groups.

Figure 6:. Plasma sTREM2 and inflammatory factors by ATN status. Levels of sTREM2 were not different between ATN categories (A).

The immunoregulatory/pleiotropic factor, fractalkine, was significantly higher in the A+T−N−, compared to A−T−N− and A+T+N. IL-7 was significantly higher in the A+T−N− group, compared to A−T−N− and A+T+N− (B). The anti-inflammatory factors, FGF-2 and IL-5 were significantly higher in A+T+N−, compared to A+T+N+ (C). The pro-inflammatory factors GRO,, IFNγ, IL-1β, IL-2, IL-3, IL-8, IL-17A and MCP-3 are lower in A+T+N+, compared to A+T+N−, while GM-CSF is lower in A+T+N−, compared to A+T−N−, and MCP-3 is higher in A+T−N−, compared to A−T−N− (D). Levels were compared between A−T−N− vs. A+T−N−, A+T−N− vs A+T+N−, and A+T+N− vs A+T+N+ by Mann-Whitney tests. Asterisks denote significance upon outlier removal. Outliers were removed per ATN group for each inflammatory factor using the ROUT method. See Table V for p-values with and without outliers. P-value notations: *p<0.05; **p<0.01; ***p<0.001.

Figure 7:. Inflammatory factor correlation matrices by disease and ATN status.

Plasma sTREM2 and each inflammatory factor were correlated to each other within disease groups (CN: n=88, MCI: n=37, AD: n=75 (A) and ATN groups (A−T−N−: n=31, A+T−N−: n=17, A+T+N− n=37, A+T+T+: n=39) (B). Spearman correlations for these correlation matrices were performed without outlier removal (see Supplemental Table III for p-values with and without outliers removed). Steiger tests to compare matrices were performed and significant differences were identified for CN verses MCI (p<0.0001) and CN verses AD (p<0.0001). Correlation directions are shown by a color gradient; positive correlations are shown in blue and negative correlations are red. Significant correlations are noted by displaying asterisks for p values: *p<0.05; **p<0.01; ***p<0.001.

Figure 8.. Inflammatory factor networks by disease and ATN status.

Upon stratification by clinical diagnosis, a large sTREM2 node was observed in CN with connections to multiple inflammatory factors, while there were fewer connections in MCI and AD as exhibited by the modularity, average degree and number of communities (A). Upon stratification by ATN, a large sTREM2 hub was observed in A−T−N−, which was absent in A+T−N−. The differences between ATN categories is also exhibited by differences in the modularity, average degree and number of communities. A large sTREM2 node was observed in A+T+N− and A+T+N+ where interconnectedness with sTREM2 was different than other ATN categories (B). Inclusion of inflammatory factors was based on Spearman r>0.8 and p<0.05. Inclusion of sTREM2 was based on p<0.05 without regard to Spearman r values. Node size and thickness of lines are related to levels of Spearman rho. Modularity is an indicator of the complexity of the networks. Average degree is an indicator of the how many of the nodes are connecting with other nodes. Communities are highlighted as one color and is a function of stronger and more frequent connections.

Figure 9.. Summary of altered inflammatory factor levels and altered inflammatory factor relationships with sTREM2 in AD.

After outlier removal some inflammatory factors exhibited an early increase in the MCI stage (IFNy, IL-17A, IP10) while several were decreased later in the AD stage including sTREM2 (black line). Other inflammatory factors exhibited an early decrease (IL-1RA, MIP-1B) (grey line) while two inflammatory factors were significantly lower in the later AD stage compared to controls (Flt-3L, IL-2) or also in MCI compared to AD (IL-2) (grey line) (A). After outlier removal several inflammatory factors were uniquely significantly correlated with sTREM2 in the control group. Inflammatory factors significantly correlated with sTREM2 in controls, none in MCI, and five in AD (IL-10, EGF, VEGF, IL-3, IL-1B) and two were unique to AD (IL-3, IL-1B) (B). After outlier removal a few inflammatory factors were significantly increased in A+T−N−,compared to A−T−N− (Fractalkine, IL-7, MCP-3) and/or exhibited a decrease in the A+T+N− stage, compared to A+T−N− (Fractalkine, IL-7, GM-CSF) while several other inflammatory factors were significantly decreased later in the A+T+N+, compared to A+T+N− stage (C). After outlier removal individuals without AD-pathology (A−T−N−) uniquely exhibited significant correlations between sTREM2 and three inflammatory factors; MDC, IL-5, IP-10, while EGF was shared with A+T+N−. IL-4 and IL-12P70 were unique to the A+T−N− stage and TGF-α and IL-10 were unique to the A+T+N− stage while IL-1β overlapped with A+T+N+ and six were unique to the A+T+N+ group; IL-7, FGF2, GM-CSF, GRO, IFNα2 and IL-3 (D). Multiple inflammatory factors were either positive (+) or negative (−) for significant differences in levels or the relationship with sTREM2 by disease or ATN status. Only FGF-2, GM-CSF and IL-1β were positive for all four categories (E).