Circulating CD5⁺ dendritic cells link gut microbiome function to immunotherapy response in cancer patients

Technical Summary: Circulating CD5⁺ Dendritic Cells Link Gut Microbiome Function to Immunotherapy Response in Cancer Patients

Problem Statement
While immune checkpoint inhibitors (ICIs) have revolutionized oncology, durable clinical benefit remains restricted to a minority of patients. Current biomarkers, such as PD-L1 expression and tumor mutational burden (TMB), focus on tumor-intrinsic features and local T-cell engagement, failing to capture the functional competence of the broader systemic immune apparatus or the upstream steps of the cancer-immunity cycle. Furthermore, while the gut microbiome is recognized as a modifiable determinant of ICI responsiveness, the field is hindered by inconsistent taxonomic associations and a lack of mechanistic understanding regarding how microbial metabolic output interfaces with host immunity. There is a critical need for biomarkers that reflect the coordinated operational capacity of the systemic immune system and its interaction with the microbiome.

Methodology
This prospective study enrolled 82 immunotherapy-naïve patients across five solid tumor types (melanoma, NSCLC, RCC, cSCC, and HNC). The study employed a multimodal, longitudinal profiling framework:

• Immune Profiling: Peripheral blood mononuclear cells (PBMCs) were analyzed using a custom 30-parameter spectral flow cytometry panel to resolve myeloid heterogeneity (including cDC1, cDC2, and CD5⁺ cDC2 subsets) and lymphoid populations. Serum was subjected to targeted proteomic profiling (Olink).
• Microbiome Profiling: Gut metagenomic sequencing (shotgun) and oral 16S rRNA sequencing were performed on baseline and 3-month post-treatment samples.
• Clinical Endpoints: Response was assessed at 3 and 6 months using RECIST v1.1. The 3-month disease control rate (CR/PR/SD) served as the primary endpoint for biomarker discovery due to attrition at later time points. Immune-related adverse events (irAEs) were graded per CTCAE v5.0.
• Statistical Analysis: The study utilized hierarchical gating strategies to define immune subsets within specific analytical layers (e.g., myeloid layer, HLA-DR⁺ myeloid layer). Associations were tested using Wilcoxon rank-sum tests, logistic regression, and false discovery rate (FDR) correction. Microbial functional potential was analyzed via metagenomic pathway reconstruction (MetaCyc) and species-resolved "per-cell" functional output modeling using DESeq2.

Key Results

• Peripheral Immune Biomarkers:

  • CD5⁺ cDC2 as a Predictor: Elevated baseline frequencies of CD5⁺ conventional dendritic cells type 2 (cDC2), specifically measured as a proportion of total cDC2 or HLA-DR⁺ myeloid cells, independently predicted early (3-month) clinical response. A 10% increase in the CD5⁺ cDC2/cDC2 ratio doubled the odds of response.
  • B Cell Activation: Attenuated B cell activation, defined by lower baseline expression of CD54 (ICAM-1) on B cells, was an independent negative predictor of response.
  • Long-term Benefit: While CD5⁺ cDC2 predicted early response, durable (6-month) benefit was associated with a broader expansion of the total DC pool (cDC1, cDC2, and CD5⁺ cDC2).
  • Longitudinal Dynamics: Responders exhibited a coordinated resolution of CD4⁺ T cell activation and a shift toward memory-like states. In contrast, non-responders displayed sustained CD4⁺ T cell activation and progressive B cell hyperactivity (increasing CD54). Non-responders also showed a post-treatment enrichment of immature, PD-1high myeloid cells, suggesting a failure to recruit functional effectors to the tumor.

• Microbiome Signatures:

  • Taxonomic Signatures: Baseline gut metagenomics identified an 11-species signature distinguishing responders from non-responders. Responders were enriched in Dorea formicigenerans, Faecalibacterium MIC7145, Coprococcus_A catus, Eisenbergiella sp900066775, and Eubacterium_I ramulus. Non-responders were enriched in Faecalicatena gnavus, Enorma massiliensis, Blautia_A sp000437735, Megasphaera sp900066485, Enterococcus faecalis, and Dorea sp000509125.
  • Functional Decoupling: A paradoxical decoupling was observed where taxonomic abundance did not strictly dictate functional output. For example, Faecalicatena gnavus was more abundant in non-responders but contributed higher functional activity in responders. Conversely, Faecalibacterium MIC7145 showed congruent taxonomic and functional enrichment in responders.
  • Microbiome-Immune Axis: Community-level microbial functional pathways, specifically peptidoglycan biosynthesis, UMP biosynthesis, and aromatic compound biosynthesis, positively correlated with the abundance of CD5⁺ cDC2s. No direct correlation was found between individual species abundance and immune parameters.
  • Oral Microbiome and Toxicity: While the oral microbiome did not predict response, a distinct signature dominated by periodontopathogens (Fusobacterium, Tannerella, Treponema, Filifactor) and higher baseline oral alpha diversity predicted severe irAEs.

Significance and Claims
The study identifies a baseline immunological configuration—characterized by elevated CD5⁺ cDC2 frequency and attenuated B cell activation—as a cornerstone of the antigen-presenting checkpoint required for effective ICI response. The authors claim that CD5⁺ cDC2s serve as a critical, developmentally distinct subset capable of priming anti-tumor T cells, and their peripheral abundance acts as a robust predictor of early clinical benefit.

Furthermore, the paper posits that the gut microbiome influences immunotherapy efficacy not merely through the presence of specific taxa, but through its collective metabolic output. The positive correlation between microbial biosynthetic pathways (peptidoglycan, UMP, aromatic compounds) and CD5⁺ cDC2 abundance suggests that the microbiome's functional capacity helps maintain a "priming-competent" immune state. The study concludes that the antigen-presenting step in the cancer-immunity cycle is a potential bottleneck, and that the gut microbiome's metabolic output may be exploitable to sustain the cellular substrate (CD5⁺ cDC2s) necessary for durable anti-tumor immunity. Finally, the identification of an oral pathobiont signature for severe toxicity highlights the need to consider multi-site microbial dynamics in managing immunotherapy safety.