Weakly activated core neuroinflammation pathways were identified as a central signaling mechanism contributing to the chronic neurodegeneration in Alzheimer’s disease

Fuhai Li, Abdallah M. Eteleeb, William Buchser, Christopher Sohn, Guoqiao Wang, Chengjie Xiong, Philip Payne, Eric McDade, Celeste M. Karch, Oscar Harari, Carlos CruchagaFrontiers in Aging Neuroscience 2022 · DOI

Abstract

Neuro-inflammation signaling is a recognized hallmark of Alzheimer's disease (AD), yet the core signaling pathways driving this process remain poorly understood. This study shifts the focus from strongly activated molecular signals to "weakly activated" ones, hypothesizing that chronic neurodegeneration is driven by low-level, persistent inflammation rather than acute "firestorms." By analyzing large-scale genomics datasets from the Mayo Clinic and RosMap, researchers identified a core "X-signaling" network of nine pathways—including MAPK, NF-kappa B, and TNF—that link chronic inflammation and hypoxia to apoptosis. These findings suggest that targeting these subtle signaling hubs could offer new avenues for AD prevention and treatment.

Introduction: The "Slow Burn" of Alzheimer’s

For decades, Alzheimer’s research has centered on amyloid-β plaques and tau tangles. While genetic risk factors like APOE4 are well-documented, the intermediate signaling mechanisms that translate these biomarkers into neuronal death have remained elusive. Most genomic studies prioritize "strongly activated" genes with high fold-changes. However, AD is a chronic condition spanning decades. The authors argue that if inflammation were strongly activated, it would likely manifest as an acute, rapid disease (similar to sepsis or severe COVID-19). To understand the slow progression of AD, we must instead look for the "weak" signals that drive cumulative damage over time.

Methodology: Listening to the "Whispers"

The researchers analyzed RNA-seq data from two major independent cohorts: the Mayo Clinic (77 control, 81 AD) and RosMap (97 control, 260 AD). To capture the "whispers" of chronic inflammation, the study employed a novel approach: relaxing the standard fold-change thresholds.

Instead of the traditional 2.0x or 3.0x cutoff, the team included genes with a fold-change as low as 1.1 (a 10% increase). This allowed for the identification of a much broader set of genes that are consistently, albeit subtly, up-regulated in AD brains across both datasets.

Table 1: Number of Differentially Expressed Genes (DEGs) identified at various fold-change and p-value thresholds. Note the significant increase in common DEGs when the threshold is lowered to 1.1.

Following differential expression analysis, the team used KEGG pathway enrichment and network inference (Dijkstra’s algorithm) to map these genes to biological circuits and determine their proximity to cell death machinery.

The "X-Core" and Viral Mimicry

The analysis revealed seven categories of signaling pathways implicated in AD, many of which are typically associated with viral infections (e.g., Epstein-Barr or HIV-1). This suggests the AD brain exists in a state of "pseudo-infection."

Central to these findings is the identification of the "X-core" signaling network, a group of nine proteins: MAPK, Rap1, NF-kappa B, HIF-1, PI3K-Akt, Wnt, TGF-beta, Hippo, and TNF. This network acts as a hub, responding to low-level inflammation and hypoxia (oxygen deficiency) to trigger chronic neurodegeneration. Specifically, the up-regulation of TNF receptors (TNFRSF1A, 10A, and 10B) was found to be a direct trigger for "executioner" enzymes (Caspase 6 and 7) that initiate apoptosis.

Results: Enhanced Diagnostic Separation

A key validation of this "weak signal" approach is the improved statistical separation between AD and healthy control samples. When using the entire set of protein-coding genes, the two groups are difficult to distinguish. However, when focusing specifically on the 417 weakly up-regulated genes identified through this method, the groups become highly distinct.

Figure 1: Boxplots demonstrating that the 417 identified genes provide significantly better separation between Control and AD groups compared to using all protein-coding genes across both datasets.

In the Mayo dataset, this refined gene set improved the statistical significance of the separation from a p-value of 0.037 to $9.5 \times 10^{-9}$. Network analysis further showed that genes in the X-core and virus-related pathways are significantly "closer" in the signaling chain to apoptosis genes than random genes are.

Conclusion and Future Directions

This study provides a compelling roadmap for understanding AD as a chronic inflammatory process. By identifying the X-core signaling network, researchers have highlighted potential therapeutic targets that sit at the intersection of environmental stress and neuronal death.

However, limitations remain. The use of bulk RNA-seq data means these signals are an average across all brain cell types; future single-cell studies are needed to pinpoint whether these signals originate in neurons, microglia, or astrocytes. Furthermore, drugging "weakly" activated pathways requires extreme precision to avoid disrupting essential immune functions. Nevertheless, shifting the focus to these subtle signals opens a new frontier in the fight against Alzheimer’s.

Summary updated

Abstract

The potential core neuroinflammation and associated signaling targets and pathways were identified using integrative network analysis on two large-scale transcriptomics datasets of AD.
BibTeX
@inproceedings{li2022weakly,
  author = {Fuhai Li and Abdallah M. Eteleeb and William Buchser and Christopher Sohn and Guoqiao Wang and Chengjie Xiong and Philip Payne and Eric McDade and Celeste M. Karch and Oscar Harari and Carlos Cruchaga},
  title = {{Weakly activated core neuroinflammation pathways were identified as a central signaling mechanism contributing to the chronic neurodegeneration in Alzheimer’s disease}},
  booktitle = {Frontiers in Aging Neuroscience},
  year = {2022},
  doi = {10.3389/fnagi.2022.935279},
}
Figure