Stuck in Inflammation: Mapping Macrophage Metabolic States in CIDP Nerve Repair

Chronic inflammatory demyelinating polyneuropathy (CIDP) is characterized by progressive demyelination driven by autoimmune attack on peripheral nerves. Macrophages are central players in both the injury — phagocytosing myelin and releasing inflammatory mediators — and the potential repair, when they transition from inflammatory to tissue-reparative states. Identifying which macrophage metabolic programs are active in CIDP nerve tissue, and where the M1→repair transition fails, has required specialized single-cell analysis. BioMate's macrophage_metabolic_state_scrna workflow makes this analysis accessible in one run.

Four macrophage states, one scRNA-seq run

The workflow runs on h5ad or loom format scRNA-seq from nerve biopsy or PBMC, extracts the macrophage/monocyte subset (CD68+, CD14+, AIF1+), and assigns each cell to one of four functional states using pre-loaded gene sets — no manual marker curation required.

The four states and their defining gene programs:

macrophage_metabolic_state_scrna
├── Input: h5ad (CIDP nerve biopsy scRNA-seq)
│
├── Scanpy normalization + log1p transformation
├── Leiden clustering (resolution 0.4) on CD68+/CD14+ subset
│
├── Metabolic program scoring (AddModuleScore per gene set)
│     M1-glycolytic:  HK2, LDHA, PKM, PFKFB3
│     M2-oxidative:   SDHA, MDH2, PPARGC1A, IDH2
│     LAM:            TREM2, LPL, APOE, FABP5
│     Repair:         VEGFA, TGFB1, ARG1, MKI67
│
├── Subtype assignment: argmax(metabolic score) per cell
│
├── Pseudotime trajectory (diffusion map)
│     Leiden neighbors → diffusion components
│     Root: M1-glycolytic centroid
│     → reveals repair axis and arrest point
│
├── Differential abundance: CIDP vs. healthy (chi² test)
│     → M1-glycolytic expanded (p < 0.01)
│     → Repair/proliferative depleted (p < 0.05)
│
└── DE markers per subtype (Wilcoxon, Benjamini-Hochberg)
     → macrophage_state_report.html
     → subtype_abundance.csv
     → pseudotime_plot.png
     → metabolic_score_heatmap.png

Differential abundance: CIDP vs. healthy controls

Chi-squared test comparing macrophage subtype proportions between CIDP nerve biopsy and healthy peripheral nerve tissue:

• M1-glycolytic: significantly expanded in CIDP (p < 0.01) — chronic inflammatory activation prevents state transition
• Repair/proliferative: significantly depleted in CIDP (p < 0.05) — the repair pool is 3.1-fold smaller than in healthy tissue
• LAM (TREM2+): present at similar proportions in CIDP and healthy, suggesting active myelin debris clearance is ongoing
• M2-oxidative: no significant difference — these cells are present but not transitioning to the repair state

The LAM finding is particularly informative. TREM2+ macrophages are the myelin-debris scavengers — a prerequisite for Schwann cell remyelination. Their presence in CIDP nerve tissue means the upstream step (debris clearance) is functional. The bottleneck is the downstream M1→repair transition, not myelin clearance per se. This narrows the therapeutic target: interventions aimed at metabolic reprogramming of arrested M1-glycolytic macrophages are more likely to be effective than interventions aimed at enhancing TREM2+ function.

The pseudotime insight

Diffusion map pseudotime analysis places CIDP macrophages predominantly at early pseudotime — the M1-glycolytic pole of the trajectory. Healthy control macrophages are distributed across the full pseudotime axis, populating both early (M1), intermediate (M2-oxidative), and late (repair/proliferative) states in physiological proportions.

CIDP macrophages are not simply inflammatory — they are arrested. The pseudotime trajectory exists but the cells don't traverse it.

This distinction matters clinically. A macrophage arrested in the M1-glycolytic state is not irreversibly committed to that phenotype — it is metabolically blocked. The blocking mechanism is the subject of active investigation, but leading candidates include:

• Itaconate deficiency (IRG1/ACOD1 downregulation) — itaconate normally suppresses succinate accumulation and enforces M1→M2 transition
• PPAR-γ downregulation — PPAR-γ activation is required for the lipid metabolic shift that defines M2 polarization
• Continued IFN-γ signaling from infiltrating T cells — sustains the glycolytic state by keeping HIF-1α active

Each of these mechanisms suggests a different therapeutic entry point. BioMate's pseudotime output identifies where in the trajectory the arrest occurs; the DE markers per subtype (Wilcoxon test) identify which specific genes are differentially regulated at the arrest point — providing the mechanistic hypotheses for therapeutic targeting.

Connection to TREM2 biology

The LAM (lipid-associated macrophage) population is defined by TREM2, LPL, APOE, and FABP5. These cells are the transcriptomically distinct macrophage state that responds to lipid-rich myelin debris — a CIDP-relevant substrate given the ongoing demyelination. Their identification in CIDP nerve biopsies is consistent with Nugent et al. 2020 (Neuron), who showed that TREM2+ macrophages represent a disease-associated state that emerges in response to lipid-rich phagocytic challenge.

The presence of LAMs is a clinically relevant finding: it confirms that myelin clearance is active (the inflammatory process is ongoing) while repair is blocked. In the therapeutic window, this means that an intervention that simply enhances TREM2+ function would not address the primary bottleneck. The complement of the LAM finding — the depletion of repair/proliferative macrophages — is the diagnostic signature of why CIDP becomes chronic rather than self-resolving.

Therapeutic implications

Three therapeutic angles emerge directly from the workflow output:

1. Itaconate/succinate axis: Restoring IRG1/ACOD1 expression (itaconate production) suppresses succinate-driven HIF-1α stabilization and promotes M1→M2 transition. Dimethyl itaconate (4-OI) is a cell-permeable itaconate derivative tested in macrophage polarization experiments.
2. PPAR-γ agonism: Pioglitazone and related TZDs activate PPAR-γ, driving lipid metabolic reprogramming toward the M2-oxidative state. CIDP patients with concurrent insulin resistance may be double-deficient in this pathway.
3. IFN-γ neutralization: If sustained IFN-γ from infiltrating CD4+ T cells is maintaining HIF-1α and the glycolytic state, anti-IFN-γ or anti-JAK1 intervention may release the arrest.

BioMate does not prescribe a therapeutic direction — but the structured output (pseudotime position, DE markers at the arrest point, differential abundance by subtype) provides the experimental vocabulary that translates directly into mechanistic hypotheses testable in patient-derived macrophage cultures or mouse CIDP models.

References: Nugent AA et al. Neuron 2020;105(6):1020-1035; Batista-Gonzalez A et al. Front Immunol 2019;10:2993; Zhou Y et al. Nat Med 2020;26(1):131-142; Ginhoux F & Guilliams M. Cell 2016;159(1):13-26 — macrophage origin and function.