Background

Alzheimer’s disease (AD) is a progressive, neurodegenerative disease characterized by a wide spectrum of pathological hallmarks including amyloid-beta (Aβ) plaques, intraneuronal neurofibrillary tangles, atrophy of synapses, and dendritic arbors, with associated cognitive decline [1]. While histological studies have characterized the morphology and process complexity of glial cells in a wide-range of neurodegenerative diseases, specific genomic and proteomic data regarding how microglia and astrocytes are modulated by disease, and vice versa, remain unknown. The lack of data regarding microglial reactivity in neurodegeneration is somewhat due to controversy regarding the utility of specific AD mouse models as windows into or prototype for human disease, and a paucity of studies using panoramic unbiased molecular profiling approaches to better describe local pathological changes occurring within the Aβ niche.

Numerous studies using mouse AD models have identified and pre-clinically validated mechanisms targeted at either neuro-inflammatory or innate immune pathways [2,3,4,5]. However, the predictive value of these data is predicated on the translational relevance of the selected mouse model. Hence, clear understanding of how molecular pathways, particularly those related to the neuroimmune response, change with respect to disease pathology in these models is a key to accurate data analysis and interpretation. RNAseq is well suited to map** out transcriptome changes across the entire genome and has significant advantages to low-throughout platforms and microarrays which are biased to known transcripts and generally remain limited to comparative inferences. One goal of the current study was to complete longitudinal evaluation of two different APP mouse models using RNAseq in brain regions demonstrating heavy plaque burden to evaluate their overlap with gene expression signatures previously identified in human AD brain samples.

Microglia are the brain’s resident immune cells that maintain central nervous system (CNS) homeostasis, constantly survey their environment, and react to injury by initiating an inflammatory reaction [6]. They express a diverse set of pattern recognition receptors capable of sensing these damage-associated signals and pathogens in the extracellular milieu and respond to neuronal injury and tissue pathogens by rapidly extending processes toward and chemotaxing into the pathological niche. Not surprisingly, in the AD brain, microglia cluster and surround Aβ plaques, a phenomenon also noted in many mouse models [7,8,9]. Despite the similar observations between mouse and humans, it remains to be seen how the plaque niche in AD mouse models may be dysregulated with regard to expression of human AD signatures, inflammatory versus innate immune pathways, and what transcriptome-wide profiling insights can be gained about biological changes occurring in the plaque-associated tissue. Notably, drugs targeting inflammation in AD, such as COX inhibitors and nonsteroidal anti-inflammatory drugs (NSAIDs), have failed in the clinic underscoring the need for validating translational relevance of mouse models with respect to neuroinflammation such that novel insights [10,11,12,13,14], therapeutic hypotheses, and targets are more likely to translate to clinical success.

Large-scale AD GWAS studies have provided genetic hypotheses for neuroinflammation in AD [15,16,17,18,19,20,21,22,3a). Conversely, the majority of the genes upregulated in the plaque versus non-plaque comparison were also upregulated in 6–10-month-old TgCRND8 animals compare to WT (Fig. 4b). Several genes (12,130) were detected in both experiments, including 342 genes that were upregulated in TgCRND8 vs WT cortex (p < 0.001 at any age) and 67 genes upregulated in plaque vs non-plaque (p < 0.001), of which 46 were overlap** between both signature sets (p < 2e-56) (Fig. 4c). We confirmed the findings from our RNAseq data using a customized NanoString nChip showing that key plaque-associated genes were upregulated (Trem2, Tyrobp, Cd68, Clec7a, Tspo, Itgfax) (Additional file 1: Figure S1). Data were analyzed for expression of a set of markers associated with resident microglia or peripheral macrophages as outlined by Hickman et al. to characterize the phenotype of cells surrounding the plaque [36]. Results show that 7 out of the top 25 most abundant genes in resident microglia were significantly upregulated in the plaque niche compared to non-plaque samples (Additional file 8: Figure S4). Conversely, only 1 gene associated with peripheral macrophage expression was significantly upregulated in the plaque niche, Complement C4-b (C4b) (Additional file 8: Figure S4). This phenotypic analysis was more pronounced in whole cortex samples resulting in 13 out of the top 25 microglial-specific genes were significantly upregulated in TgCRND8 mice compared to WT compared to only 1 gene in the peripheral macrophage signature (Additional file 8: Figure S4).

Fig. 2
figure 2

Localization of TREM2 and Cd33 around amyloid-beta (Aβ) plaques in the aged TgCRND8 mouse. Brain slices were prepared from 35-week-old wild-type and TgCRND8 mice and processed for amyloid-beta immunohistochemistry (6E10 labeling, magenta) combined with TREM2 (panel a) or CD33 (panel b) in situ hybridization via RNAScope (brown). Visualization of TREM2 and Cd33 confirms their associated expression pattern with Aβ pathology, supporting the rationale for laser capture microdissection of Thio-S-labeled plaques for transcriptome-wide RNA sequencing. Representative images show section sampling for LCM and RNA seq (c, d). Images show Iba1 (brown) and 6E10 (magenta) immunostainings. Scale bars, 50 and 100 μm

Full size image
Fig. 3
figure 3

Plaque-niche gene expression differences in 6-month-old TgCRND8. a Student’s t test p value distributions for gene expression differences plaque and normal samples in TgCRND8 cortex at 6 months of age. Gray line indicates expected false discovery rate (FDR) given multiple test comparisons. b Heatmap showing log10 ratio values from each sample (y-axis) for each gene (x-axis) with t test p value < 0.001 between plaque and normal tissue in TgCRND8 cortex. Samples are ordered manually by genotype as indicated. Genes are ordered by agglomerative clustering. c Heatmaps showing log10 ratio values from each sample (y-axis) for each gene (x-axis) within the indicated gene sets. Samples are ordered manually by genotype as indicated. Genes are ordered by agglomerative clustering within each set

Full size image
Fig. 4
figure 4

Overlap of TgCRND8 progression signature and plaque-niche signature. a Heatmap showing log10 ratio values from each LCM sample (y-axis) for each gene (x-axis) detected in both the LCM and whole cortex studies and with t test p < 0.001 between TgCRND8 and WT cortex at one or more ages. Samples are ordered manually by genotype as indicated. Genes are ordered by agglomerative clustering. b Heatmap showing log10 ratio values from each whole cortex sample (y-axis) for each gene (x-axis) detected in both the whole cortex and LCM studies and with t test p < 0.001 between plaque and normal tissue in TgCRND8 cortex. Samples are ordered manually by genotype as indicated. Genes are ordered by agglomerative clustering. c Venn diagram depicting number of genes in common between signatures in a and b, and the associated hypergeometric p value given 12,130 genes detected in both experiments

Full size image

To further characterize the microglia phenotype observed in the TgCRND8 plaque niche, we compared the gene signatures identified here to those of the microglial subtype, disease-associated microglia (DAM), recently shown to surround plaques in another mouse AD model expressing five familiar AD mutations (5XFAD) [37]. Indeed, genes reported to be expressed higher in the DAM population (> 1.5 fold, p < 1e-5) were strongly enriched among the upregulated gene signatures in both the TgCRND8 whole cortex and the LCM-dissected TgCRND8 plaque tissue signatures (p = 4e-25 and p = 8e-13, respectively) (Fig. 5a–d; Additional file 4: Table S2, Additional file 6: Table S3, and Additional file 7: Table S4). Across all three gene sets, eight genes were commonly expressed: Ccl6, Cd9, Ctsz, Gusb, Lyr2, Npc2, Trem2, and Tyrobp. In contrast, the DAM marker genes were not significantly represented among the Tg2576-related signatures (p = 0.3) (Fig. 5a, c). In additional analyses, expression of markers of M1 and M2 microglia as outlined in [38, 39] were also analyzed for differential expression in plaque vs. non-plaque samples as well as in whole cortex samples from TgCRND8 compared to WT controls. Results show a significant increase in expression of only one marker of M1 activation, CD16a, in the plaque niche compared to non-plaque controls (3.06-fold; p < 0.0001). No markers of M2 activation were significantly upregulated in plaque samples compared to non-plaque. Whole cortex TgCRND8 displayed significant (p < 0.001) expression of four markers of M1 activation, CXCL10 (34.52-fold; p < 0.0001), CCR5 (1.51-fold; p < 0.0001), CD86 (2.25-fold; p < 0.0001), and CD16a (2.62-fold; p < 0.0001) in TgCNRD8 compared to WT, and only two markers of M2 activation/repair, Clec7a (72.33-fold; p < 0.0001), and TGFβ (2.23-fold; p < 0.0001).

Fig. 5
figure 5

Overlap of TgCRND8 progression signature and plaque-niche signature. a, b Heatmap showing log10 ratio values from each whole cortex sample (a) or each LCM sample (b) for genes reported to be expressed higher (> 1.5 fold, p < 1e-5) in the DAM population [37]. Samples are ordered manually by genotype and age as indicated. Genes are ordered by agglomerative clustering. c Venn diagram depicting number of genes in common between the Tg576/TgCRND8 whole cortex signatures (p < 0.001, any age) and the genes expressed higher (> 1.5 fold, p < 1e-5) in the DAM cells. d Venn diagram depicting number of genes in common between the TgCRND8 LCM plaque signature (p < 0.001) and the genes expressed higher (> 1.5 fold, p < 1e-5) in the DAM population

Full size image

Discussion

Many studies have demonstrated a strong association between inflammation, innate immunity, and amyloid plaques in both human AD and preclinical models [5c). Thus, our data suggests that the TgCRND8 model is a more relevant model to understand polarization states and plaque niche.

The TgCRND8 transcriptional profiles align with previous findings of robust transcriptional responses in recently discovered genetic links between microglia, innate immunity, and AD. Numerous GWAS and meta-analyses in various AD cohorts’ studies have identified novel microglial-associated receptors TREM2 and CD33 for AD [16, 18, 29]. This has naturally led to investigations of their roles in AD in preclinical models. In the current study, regional assessment of the Aβ plaque niche revealed TREM2 and its signaling adaptor TYROBP (which encodes the DAP12 protein) are significantly (p < 0.001) upregulated in the Aβ plaque niche compared to non-plaque areas. Both markers show a 3-fold increase in expression near the plaque compared to non-plaque samples. An expanded view beyond these hub genes confirms that several ITIM/ITAM-associated transcripts are also enriched around plaques. In TgCRND8 cortical homogenates, this innate immune panel was shown to be upregulated at 6 and 10 months of age, as well. It is important to highlight, however, that the ITIM/ITAM-associated module is not invariably associated with Aβ plaque pathology. In the Tg2576 model, we did not observe significant induction of this AD-implicated innate immune signature. Therefore, when juxtaposed with robust progressive signature induction in the TgCNRD8 mice, it is clear that not all mouse APP models are equivalent in this regard.

Exploring more closely specific ITIM/ITAM-associated changes, we identified novel dysregulated pathway nodes as potential modulatory factors, including genes known to be enriched in microglia such as receptor and non-receptor tyrosine kinases, as well as associated phosphatases predicted to inhibit phagocytic signal transduction. For instance, it is widely known that DAP12 signals via SYK induction and its downstream signaling may be negatively regulated by various phosphatases such as PTPN6 and INPP5D. Interestingly, not only do we observe robust SYK upregulation around amyloid plaques (> 3-fold) but associated protein tyrosine phosphatases PTPN6 and INPP5D are also enriched (upregulated 6.0-fold and 2.7-fold respectively). Both PTPN6 and INPP5D are involved in the signaling via inhibitory receptors found in innate immune cells, such as CD33 for example. Moreover, in a meta-analysis of major AD GWAS, the rs35349669 locus on chromosome 2 encoding INPP5D was recently identified as a novel AD risk-associated locus [60,61,62]. Within the plaque niche, we also observed that several inhibitory ITIM-containing receptors in addition to CD33 are also enriched, such as Slamf9 (7.8-fold, p = 0.011), FCGR1 (2.2-fold, p = 0.002), and LAIR1 (2.0-fold, p = 0.001). Likewise, we identified several additional ITAM-associated receptors such as FCGR1 (2.2-fold, p = 0.002), FCGR3 (3.1-fold, p = 1.4e-06), and CSF1R (1.5-fold, p = 0.0013) are significantly increased in the plaque niche compared to non-plaque areas.

Functional studies have begun to identify a role for the TREM2/DAP12 pathway in promoting cell survival and phagocytic capacity [63]. TREM2 deficiency reportedly impairs the phagocytosis of apoptotic neurons in microglia [64], possibly through apolipoprotein E and lipid-sensing mechanisms [65,66,67,68]. Similarly, in primary microglia isolated from TREM2 knockout mice, impaired phagocytic capacity of microbeads as well as Escherichia coli bacteria imply that TREM2 plays an important regulatory role in innate immune activity and possibly host defense mechanism [69]. Similar functional insights have been reported for CD33. A study by Gricuic et al. observed increased expression of CD33 positive microglia in AD brains, a significant positive correlation between Aβ pathology and CD33 expression as well as an increase in microglial uptake of Aβ in vitro after CD33 inactivation [32]. Interestingly, clinical data show an increase in PiB imaging in patients with the risk allele for CD33 [16], further demonstrating a relationship with amyloid pathology and potential to track the progression of disease-related PET-imaging biomarkers. These findings serve as the foundation for the therapeutic hypothesis that blockade of CD33 function could be a compelling alternative to accelerate Aβ clearance [32]. Thus, when considering collectively the functional roles for TREM2 and CD33 in microglia and AD, it appears that converging and opposing innate immune regulation may exist and dictate risk status for human AD.

Conclusions

The current study highlights the importance of careful consideration and selection of appropriate APP mouse models specifically for exploring neuroimmune modulation, in particular ITIM/ITAM-associated mechanisms, such as TREM2 and CD33. Combined with the insights from human genetics, these findings provide broad evidence for the presence of a complex relationship between activating and inhibitory mechanisms which collectively may determine innate immune status around the Aβ plaque niche. Future studies utilizing single-cell profiling of microglia in AD mouse models to further characterize the co-expression of immune activating (e.g., ITAM) and suppressing (e.g., ITIM) mechanisms and will help to define how the underlying regulation of mechanisms controlling this microglial innate immune rheostat are shaped by neuropathology. Such studies will also help clarify whether distinct heterogeneous sub-populations of microglia exist around amyloid plaques, or whether competing inhibitory and activating nodes are commonly present within individual plaque-associated microglia. The observations provide tissue-level insights into plaque-associated transcriptional signatures to encourage more fine characterization of cell-specific pathway changes across such broad panels of markers.