Early immune responses have long-term associations with clinical, virologic, and immunologic outcomes in patients with COVID-19

Zicheng Hu, Kattria van der Ploeg, Saborni Chakraborty, Prabhu S. Arunachalam, Diego Martinez Mori, Karen B. Jacobson, Hector Bonilla, Julie Parsonnet, Jason R. Andrews, Haley Hedlin, Lauren de la Parte, Kathleen D. Press, Maureen Ty, Gene S. Tan, Catherine A. Blish, Saki Takahashi, Isabel Rodríguez-Barraquer, Bryan Greenhouse, Atul J. Butte, Upinder Singh, Bali Pulendran, Taia T. Wang, Prasanna Jagannathan

Abstract

The great majority of SARS-CoV-2 infections are mild and uncomplicated, but some individuals with initially mild COVID-19 progressively develop more severe symptoms. Furthermore, mild to moderate infections are an important contributor to ongoing transmission. There remains a critical need to identify host immune biomarkers predictive of clinical and virologic outcomes in SARS-CoV-2-infected patients. Leveraging longitudinal samples and data from a clinical trial of Peginterferon Lambda for treatment of SARS-CoV-2 infected outpatients, we used host proteomics and transcriptomics to characterize the trajectory of the immune response in COVID-19 patients within the first 2 weeks of symptom onset. We define early immune signatures, including plasma levels of RIG-I and the CCR2 ligands (MCP1, MCP2 and MCP3), associated with control of oropharyngeal viral load, the degree of symptom severity, and immune memory (including SARS-CoV-2-specific T cell responses and spike (S) protein-binding IgG levels). We found that individuals receiving BNT162b2 (Pfizer–BioNTech) vaccine had similar early immune trajectories to those observed in this natural infection cohort, including the induction of both inflammatory cytokines (e.g. MCP1) and negative immune regulators (e.g. TWEAK). Finally, we demonstrate that machine learning models using 8-10 plasma protein markers measured early within the course of infection are able to accurately predict symptom severity, T cell memory, and the antibody response post-infection.

Introduction

While extensive research has focused on hospitalized COVID-19 patients, the vast majority of infections are mild or moderate. These cases are highly heterogeneous: some patients recover rapidly, while others progress to severe disease or fail to develop robust long-term immunity. Understanding the "opening move" of the immune system is critical to predicting these divergent outcomes. This study identifies specific early molecular signals—biomarkers measured within days of symptom onset—that dictate clinical and immunologic trajectories over a seven-month period.

Study Design and Methodology

The researchers employed a multi-omics approach using longitudinal samples from a Phase 2 clinical trial of Peginterferon Lambda involving 108 outpatients with mild-to-moderate COVID-19. The study design integrated whole-blood RNA sequencing and Olink multiplex proteomic panels (184 plasma proteins) at enrollment (Day 0) and Day 5.

Figure 1: Study design and longitudinal sampling timeline for the Lambda COVID-19 trial, tracking patients from enrollment through seven months.

To analyze the data, the team used quadratic regression to map immune activation waves (Interferons → NK cells → T cells → B cells) and xCell for immune cell type estimation. Machine learning, specifically Random Forest models with leave-one-out cross-validation, was used to identify the most predictive biomarkers for long-term outcomes.

Characterizing the Early Immune Trajectory

A primary finding of the study was that the time since symptom onset is the dominant driver of variation in both the transcriptome and proteome. Principal component analysis (PCA) revealed that PC1 for RNA-seq and PC2 for Olink data were strongly correlated with the timing of the infection and the development of the antibody response.

Figure 2: Analysis of RNA-seq and Olink proteomic data, showing the relationship between molecular principal components, time since symptom onset, and S protein-binding IgG levels.

Key Insights and Predictive Power

  • The RIG-I "Smoke Signal": RIG-I, a viral RNA sensor typically found inside cells, was detected in the plasma. Its levels correlated with cell death (apoptosis), suggesting it "leaks" into the blood as the virus kills cells. High early RIG-I predicted lower viral loads but also higher symptom severity and stronger long-term antibody responses.
  • The CCR2 Axis: Ligands for the CCR2 receptor (MCP1, MCP2, MCP3) were identified as core predictors of the downstream adaptive immune response, helping recruit immune cells to the infection site.
  • Vaccine vs. Infection: The first dose of the BNT162b2 vaccine mirrors the early protein trajectory of natural infection, including inflammatory cytokines and regulators like TWEAK. The second dose, however, triggers a "cleaner" adaptive response without the inflammatory neutrophil signature.
  • Predictive Accuracy: Machine learning models using just 8–10 plasma proteins achieved high Area Under the Curve (AUC) scores for predicting outcomes:
  • Symptom Severity: 0.84 (0.96 in an independent validation set).
  • T-cell Memory (Day 28): 0.77.
  • Antibody Levels (Month 7): 0.75.

Conclusion and Future Directions

This study demonstrates that a small set of early biomarkers can accurately predict a patient's clinical and immunologic path months in advance. While these correlations are robust, future work is needed to determine if these pathways are causative through perturbation studies. Validating the T-cell and antibody memory models in larger, independent cohorts remains a priority for translating these findings into clinical tools.

Summary updated

Abstract

Abstract The great majority of SARS-CoV-2 infections are mild and uncomplicated, but some individuals with initially mild COVID-19 progressively develop more severe symptoms. Furthermore, mild to moderate infections are an important contributor to ongoing transmission. There remains a critical need to identify host immune biomarkers predictive of clinical and virologic outcomes in SARS-CoV-2-infected patients. Leveraging longitudinal samples and data from a clinical trial of Peginterferon Lambda for treatment of SARS-CoV-2 infected outpatients, we used host proteomics and transcriptomics to characterize the trajectory of the immune response in COVID-19 patients within the first 2 weeks of symptom onset. We define early immune signatures, including plasma levels of RIG-I and the CCR2 ligands (MCP1, MCP2 and MCP3), associated with control of oropharyngeal viral load, the degree of symptom severity, and immune memory (including SARS-CoV-2-specific T cell responses and spike (S) protein-binding IgG levels). We found that individuals receiving BNT162b2 (Pfizer–BioNTech) vaccine had similar early immune trajectories to those observed in this natural infection cohort, including the induction of both inflammatory cytokines (e.g. MCP1) and negative immune regulators (e.g. TWEAK). Finally, we demonstrate that machine learning models using 8-10 plasma protein markers measured early within the course of infection are able to accurately predict symptom severity, T cell memory, and the antibody response post-infection.
BibTeX
@inproceedings{hu2022early,
  author = {Zicheng Hu and Kattria van der Ploeg and Saborni Chakraborty and Prabhu S. Arunachalam and Diego Martinez Mori and Karen B. Jacobson and Hector Bonilla and Julie Parsonnet and Jason R. Andrews and Haley Hedlin and Lauren de la Parte and Kathleen D. Press and Maureen Ty and Gene S. Tan and Catherine A. Blish and Saki Takahashi and Isabel Rodríguez-Barraquer and Bryan Greenhouse and Atul J. Butte and Upinder Singh and Bali Pulendran and Taia T. Wang and Prasanna Jagannathan},
  title = {{Early immune responses have long-term associations with clinical, virologic, and immunologic outcomes in patients with COVID-19}},
  booktitle = {Research Square},
  year = {2022},
  doi = {10.21203/rs.3.rs-847082/v1},
}
Figure