Remember me
Clinical Research and Public HealthImmunologyPublic Health
Open Access | 
10.1172/jci.insight.199245
1The Jackson Laboratory for Genomic Medicine, Farmington, Connecticut, USA.
2Immunai, New York, New York, USA.
3UConn Center on Aging, UConn Health, Farmington, Connecticut, USA.
4Department of Microbiology,
5Global Health and Emerging Pathogens Institute,
6Marc and Jennifer Lipschultz Precision Immunology Institute,
7Icahn Genomics Institute,
8Department of Medicine, Division of Infectious Diseases,
9The Tisch Cancer Center, and
10Department of Pathology, Molecular and Cell-Based Medicine, Icahn School of Medicine at Mount Sinai, New York, New York, USA.
11Immunoledge LLC, Montclair, New Jersey, USA.
12Department of Genetics and Genome Sciences, University of Connecticut Health Center, Farmington, Connecticut, USA.
Address correspondence to: Duygu Ucar or Jacques Banchereau, 10 Discovery Dr., Farmington, Connecticut, 06032, USA. Phone: 207.288.6000; Email: duygu.ucar@jax.org (DU). Phone: 469.877.7416; Email: Jacques.banchereau@gmail.com (JB). Or to: George A. Kuchel, 263 Farmington Avenue, Farmington, Connecticut, 06032, USA. Phone: 860.679.8400; Email: kuchel@uchc.edu.
Authorship note: GE and YYY contributed equally to this work. GAK, JB, and DU were joint supervisors of this work.
Find articles by Eryilmaz, G. in: PubMed | Google Scholar
1The Jackson Laboratory for Genomic Medicine, Farmington, Connecticut, USA.
2Immunai, New York, New York, USA.
3UConn Center on Aging, UConn Health, Farmington, Connecticut, USA.
4Department of Microbiology,
5Global Health and Emerging Pathogens Institute,
6Marc and Jennifer Lipschultz Precision Immunology Institute,
7Icahn Genomics Institute,
8Department of Medicine, Division of Infectious Diseases,
9The Tisch Cancer Center, and
10Department of Pathology, Molecular and Cell-Based Medicine, Icahn School of Medicine at Mount Sinai, New York, New York, USA.
11Immunoledge LLC, Montclair, New Jersey, USA.
12Department of Genetics and Genome Sciences, University of Connecticut Health Center, Farmington, Connecticut, USA.
Address correspondence to: Duygu Ucar or Jacques Banchereau, 10 Discovery Dr., Farmington, Connecticut, 06032, USA. Phone: 207.288.6000; Email: duygu.ucar@jax.org (DU). Phone: 469.877.7416; Email: Jacques.banchereau@gmail.com (JB). Or to: George A. Kuchel, 263 Farmington Avenue, Farmington, Connecticut, 06032, USA. Phone: 860.679.8400; Email: kuchel@uchc.edu.
Authorship note: GE and YYY contributed equally to this work. GAK, JB, and DU were joint supervisors of this work.
Find articles by Yazici, Y. in: PubMed | Google Scholar
1The Jackson Laboratory for Genomic Medicine, Farmington, Connecticut, USA.
2Immunai, New York, New York, USA.
3UConn Center on Aging, UConn Health, Farmington, Connecticut, USA.
4Department of Microbiology,
5Global Health and Emerging Pathogens Institute,
6Marc and Jennifer Lipschultz Precision Immunology Institute,
7Icahn Genomics Institute,
8Department of Medicine, Division of Infectious Diseases,
9The Tisch Cancer Center, and
10Department of Pathology, Molecular and Cell-Based Medicine, Icahn School of Medicine at Mount Sinai, New York, New York, USA.
11Immunoledge LLC, Montclair, New Jersey, USA.
12Department of Genetics and Genome Sciences, University of Connecticut Health Center, Farmington, Connecticut, USA.
Address correspondence to: Duygu Ucar or Jacques Banchereau, 10 Discovery Dr., Farmington, Connecticut, 06032, USA. Phone: 207.288.6000; Email: duygu.ucar@jax.org (DU). Phone: 469.877.7416; Email: Jacques.banchereau@gmail.com (JB). Or to: George A. Kuchel, 263 Farmington Avenue, Farmington, Connecticut, 06032, USA. Phone: 860.679.8400; Email: kuchel@uchc.edu.
Authorship note: GE and YYY contributed equally to this work. GAK, JB, and DU were joint supervisors of this work.
Find articles by Marches, R. in: PubMed | Google Scholar
1The Jackson Laboratory for Genomic Medicine, Farmington, Connecticut, USA.
2Immunai, New York, New York, USA.
3UConn Center on Aging, UConn Health, Farmington, Connecticut, USA.
4Department of Microbiology,
5Global Health and Emerging Pathogens Institute,
6Marc and Jennifer Lipschultz Precision Immunology Institute,
7Icahn Genomics Institute,
8Department of Medicine, Division of Infectious Diseases,
9The Tisch Cancer Center, and
10Department of Pathology, Molecular and Cell-Based Medicine, Icahn School of Medicine at Mount Sinai, New York, New York, USA.
11Immunoledge LLC, Montclair, New Jersey, USA.
12Department of Genetics and Genome Sciences, University of Connecticut Health Center, Farmington, Connecticut, USA.
Address correspondence to: Duygu Ucar or Jacques Banchereau, 10 Discovery Dr., Farmington, Connecticut, 06032, USA. Phone: 207.288.6000; Email: duygu.ucar@jax.org (DU). Phone: 469.877.7416; Email: Jacques.banchereau@gmail.com (JB). Or to: George A. Kuchel, 263 Farmington Avenue, Farmington, Connecticut, 06032, USA. Phone: 860.679.8400; Email: kuchel@uchc.edu.
Authorship note: GE and YYY contributed equally to this work. GAK, JB, and DU were joint supervisors of this work.
Find articles by Mimitou, E. in: PubMed | Google Scholar
1The Jackson Laboratory for Genomic Medicine, Farmington, Connecticut, USA.
2Immunai, New York, New York, USA.
3UConn Center on Aging, UConn Health, Farmington, Connecticut, USA.
4Department of Microbiology,
5Global Health and Emerging Pathogens Institute,
6Marc and Jennifer Lipschultz Precision Immunology Institute,
7Icahn Genomics Institute,
8Department of Medicine, Division of Infectious Diseases,
9The Tisch Cancer Center, and
10Department of Pathology, Molecular and Cell-Based Medicine, Icahn School of Medicine at Mount Sinai, New York, New York, USA.
11Immunoledge LLC, Montclair, New Jersey, USA.
12Department of Genetics and Genome Sciences, University of Connecticut Health Center, Farmington, Connecticut, USA.
Address correspondence to: Duygu Ucar or Jacques Banchereau, 10 Discovery Dr., Farmington, Connecticut, 06032, USA. Phone: 207.288.6000; Email: duygu.ucar@jax.org (DU). Phone: 469.877.7416; Email: Jacques.banchereau@gmail.com (JB). Or to: George A. Kuchel, 263 Farmington Avenue, Farmington, Connecticut, 06032, USA. Phone: 860.679.8400; Email: kuchel@uchc.edu.
Authorship note: GE and YYY contributed equally to this work. GAK, JB, and DU were joint supervisors of this work.
Find articles by Kenyon-Pesce, L. in: PubMed | Google Scholar
1The Jackson Laboratory for Genomic Medicine, Farmington, Connecticut, USA.
2Immunai, New York, New York, USA.
3UConn Center on Aging, UConn Health, Farmington, Connecticut, USA.
4Department of Microbiology,
5Global Health and Emerging Pathogens Institute,
6Marc and Jennifer Lipschultz Precision Immunology Institute,
7Icahn Genomics Institute,
8Department of Medicine, Division of Infectious Diseases,
9The Tisch Cancer Center, and
10Department of Pathology, Molecular and Cell-Based Medicine, Icahn School of Medicine at Mount Sinai, New York, New York, USA.
11Immunoledge LLC, Montclair, New Jersey, USA.
12Department of Genetics and Genome Sciences, University of Connecticut Health Center, Farmington, Connecticut, USA.
Address correspondence to: Duygu Ucar or Jacques Banchereau, 10 Discovery Dr., Farmington, Connecticut, 06032, USA. Phone: 207.288.6000; Email: duygu.ucar@jax.org (DU). Phone: 469.877.7416; Email: Jacques.banchereau@gmail.com (JB). Or to: George A. Kuchel, 263 Farmington Avenue, Farmington, Connecticut, 06032, USA. Phone: 860.679.8400; Email: kuchel@uchc.edu.
Authorship note: GE and YYY contributed equally to this work. GAK, JB, and DU were joint supervisors of this work.
Find articles by Handrejk, K. in: PubMed | Google Scholar
1The Jackson Laboratory for Genomic Medicine, Farmington, Connecticut, USA.
2Immunai, New York, New York, USA.
3UConn Center on Aging, UConn Health, Farmington, Connecticut, USA.
4Department of Microbiology,
5Global Health and Emerging Pathogens Institute,
6Marc and Jennifer Lipschultz Precision Immunology Institute,
7Icahn Genomics Institute,
8Department of Medicine, Division of Infectious Diseases,
9The Tisch Cancer Center, and
10Department of Pathology, Molecular and Cell-Based Medicine, Icahn School of Medicine at Mount Sinai, New York, New York, USA.
11Immunoledge LLC, Montclair, New Jersey, USA.
12Department of Genetics and Genome Sciences, University of Connecticut Health Center, Farmington, Connecticut, USA.
Address correspondence to: Duygu Ucar or Jacques Banchereau, 10 Discovery Dr., Farmington, Connecticut, 06032, USA. Phone: 207.288.6000; Email: duygu.ucar@jax.org (DU). Phone: 469.877.7416; Email: Jacques.banchereau@gmail.com (JB). Or to: George A. Kuchel, 263 Farmington Avenue, Farmington, Connecticut, 06032, USA. Phone: 860.679.8400; Email: kuchel@uchc.edu.
Authorship note: GE and YYY contributed equally to this work. GAK, JB, and DU were joint supervisors of this work.
Find articles by Jangra, S. in: PubMed | Google Scholar
1The Jackson Laboratory for Genomic Medicine, Farmington, Connecticut, USA.
2Immunai, New York, New York, USA.
3UConn Center on Aging, UConn Health, Farmington, Connecticut, USA.
4Department of Microbiology,
5Global Health and Emerging Pathogens Institute,
6Marc and Jennifer Lipschultz Precision Immunology Institute,
7Icahn Genomics Institute,
8Department of Medicine, Division of Infectious Diseases,
9The Tisch Cancer Center, and
10Department of Pathology, Molecular and Cell-Based Medicine, Icahn School of Medicine at Mount Sinai, New York, New York, USA.
11Immunoledge LLC, Montclair, New Jersey, USA.
12Department of Genetics and Genome Sciences, University of Connecticut Health Center, Farmington, Connecticut, USA.
Address correspondence to: Duygu Ucar or Jacques Banchereau, 10 Discovery Dr., Farmington, Connecticut, 06032, USA. Phone: 207.288.6000; Email: duygu.ucar@jax.org (DU). Phone: 469.877.7416; Email: Jacques.banchereau@gmail.com (JB). Or to: George A. Kuchel, 263 Farmington Avenue, Farmington, Connecticut, 06032, USA. Phone: 860.679.8400; Email: kuchel@uchc.edu.
Authorship note: GE and YYY contributed equally to this work. GAK, JB, and DU were joint supervisors of this work.
Find articles by
Schotsaert, M.
in:
PubMed
|
Google Scholar
|
1The Jackson Laboratory for Genomic Medicine, Farmington, Connecticut, USA.
2Immunai, New York, New York, USA.
3UConn Center on Aging, UConn Health, Farmington, Connecticut, USA.
4Department of Microbiology,
5Global Health and Emerging Pathogens Institute,
6Marc and Jennifer Lipschultz Precision Immunology Institute,
7Icahn Genomics Institute,
8Department of Medicine, Division of Infectious Diseases,
9The Tisch Cancer Center, and
10Department of Pathology, Molecular and Cell-Based Medicine, Icahn School of Medicine at Mount Sinai, New York, New York, USA.
11Immunoledge LLC, Montclair, New Jersey, USA.
12Department of Genetics and Genome Sciences, University of Connecticut Health Center, Farmington, Connecticut, USA.
Address correspondence to: Duygu Ucar or Jacques Banchereau, 10 Discovery Dr., Farmington, Connecticut, 06032, USA. Phone: 207.288.6000; Email: duygu.ucar@jax.org (DU). Phone: 469.877.7416; Email: Jacques.banchereau@gmail.com (JB). Or to: George A. Kuchel, 263 Farmington Avenue, Farmington, Connecticut, 06032, USA. Phone: 860.679.8400; Email: kuchel@uchc.edu.
Authorship note: GE and YYY contributed equally to this work. GAK, JB, and DU were joint supervisors of this work.
Find articles by
García-Sastre, A.
in:
PubMed
|
Google Scholar
|
1The Jackson Laboratory for Genomic Medicine, Farmington, Connecticut, USA.
2Immunai, New York, New York, USA.
3UConn Center on Aging, UConn Health, Farmington, Connecticut, USA.
4Department of Microbiology,
5Global Health and Emerging Pathogens Institute,
6Marc and Jennifer Lipschultz Precision Immunology Institute,
7Icahn Genomics Institute,
8Department of Medicine, Division of Infectious Diseases,
9The Tisch Cancer Center, and
10Department of Pathology, Molecular and Cell-Based Medicine, Icahn School of Medicine at Mount Sinai, New York, New York, USA.
11Immunoledge LLC, Montclair, New Jersey, USA.
12Department of Genetics and Genome Sciences, University of Connecticut Health Center, Farmington, Connecticut, USA.
Address correspondence to: Duygu Ucar or Jacques Banchereau, 10 Discovery Dr., Farmington, Connecticut, 06032, USA. Phone: 207.288.6000; Email: duygu.ucar@jax.org (DU). Phone: 469.877.7416; Email: Jacques.banchereau@gmail.com (JB). Or to: George A. Kuchel, 263 Farmington Avenue, Farmington, Connecticut, 06032, USA. Phone: 860.679.8400; Email: kuchel@uchc.edu.
Authorship note: GE and YYY contributed equally to this work. GAK, JB, and DU were joint supervisors of this work.
Find articles by Kuchel, G. in: PubMed | Google Scholar
1The Jackson Laboratory for Genomic Medicine, Farmington, Connecticut, USA.
2Immunai, New York, New York, USA.
3UConn Center on Aging, UConn Health, Farmington, Connecticut, USA.
4Department of Microbiology,
5Global Health and Emerging Pathogens Institute,
6Marc and Jennifer Lipschultz Precision Immunology Institute,
7Icahn Genomics Institute,
8Department of Medicine, Division of Infectious Diseases,
9The Tisch Cancer Center, and
10Department of Pathology, Molecular and Cell-Based Medicine, Icahn School of Medicine at Mount Sinai, New York, New York, USA.
11Immunoledge LLC, Montclair, New Jersey, USA.
12Department of Genetics and Genome Sciences, University of Connecticut Health Center, Farmington, Connecticut, USA.
Address correspondence to: Duygu Ucar or Jacques Banchereau, 10 Discovery Dr., Farmington, Connecticut, 06032, USA. Phone: 207.288.6000; Email: duygu.ucar@jax.org (DU). Phone: 469.877.7416; Email: Jacques.banchereau@gmail.com (JB). Or to: George A. Kuchel, 263 Farmington Avenue, Farmington, Connecticut, 06032, USA. Phone: 860.679.8400; Email: kuchel@uchc.edu.
Authorship note: GE and YYY contributed equally to this work. GAK, JB, and DU were joint supervisors of this work.
Find articles by Banchereau, J. in: PubMed | Google Scholar
1The Jackson Laboratory for Genomic Medicine, Farmington, Connecticut, USA.
2Immunai, New York, New York, USA.
3UConn Center on Aging, UConn Health, Farmington, Connecticut, USA.
4Department of Microbiology,
5Global Health and Emerging Pathogens Institute,
6Marc and Jennifer Lipschultz Precision Immunology Institute,
7Icahn Genomics Institute,
8Department of Medicine, Division of Infectious Diseases,
9The Tisch Cancer Center, and
10Department of Pathology, Molecular and Cell-Based Medicine, Icahn School of Medicine at Mount Sinai, New York, New York, USA.
11Immunoledge LLC, Montclair, New Jersey, USA.
12Department of Genetics and Genome Sciences, University of Connecticut Health Center, Farmington, Connecticut, USA.
Address correspondence to: Duygu Ucar or Jacques Banchereau, 10 Discovery Dr., Farmington, Connecticut, 06032, USA. Phone: 207.288.6000; Email: duygu.ucar@jax.org (DU). Phone: 469.877.7416; Email: Jacques.banchereau@gmail.com (JB). Or to: George A. Kuchel, 263 Farmington Avenue, Farmington, Connecticut, 06032, USA. Phone: 860.679.8400; Email: kuchel@uchc.edu.
Authorship note: GE and YYY contributed equally to this work. GAK, JB, and DU were joint supervisors of this work.
Find articles by Ucar, D. in: PubMed | Google Scholar
Authorship note: GE and YYY contributed equally to this work. GAK, JB, and DU were joint supervisors of this work.
Published February 23, 2026 - More info
Published in Volume 11, Issue 4 on February 23, 2026
Abstract
BACKGROUND. The SARS-CoV-2 pandemic provided a rare opportunity to study how human immune responses develop to a novel viral antigen delivered through different vaccine platforms. However, to date, no study has directly compared immune responses to all 3 FDA-approved COVID-19 vaccines at single-cell multiomic resolution.
METHODS. We longitudinally profiled SARS-CoV-2–naive adults (n = 31) vaccinated with BNT162b2, mRNA-1273, or Ad26.COV2.S, integrating plasma cytokines, antibody titers, and single-cell multiomic data (DOGMA-Seq).
RESULTS. We discovered a distinct, transient IFN program termed ISG-dim, which emerged specifically 1–2 days after the first mRNA dose in approximately 10% of myeloid cells. This state was characterized by ISGF3 complex activation and its target genes (e.g., MX1, MX2, DDX58), with transcriptional and epigenetic profiles distinct from the robust IFN program observed after mRNA boosting or a single Ad26.COV2.S dose (ISG-high). In vitro stimulation of human monocytes showed that IFN-α alone recapitulates ISG-dim, whereas both IFN-α and IFN-γ are required for ISG-high.
CONCLUSION. These findings define dose-dependent IFN programming in human myeloid cells and highlight mechanistic differences between priming and boosting, with implications for optimizing vaccine platform choice, dose scheduling, and formulation.
FUNDING. NIH grants AI142086, U19 AI135972, U01 AI165452, U01 AI165452, R01 AI160706, and P30 AG067988.
IntroductionThe SARS-CoV-2 pandemic posed unprecedented global health challenges but also created a rare scientific opportunity to study how human immune responses develop to a novel viral antigen introduced through different vaccine platforms. Among these, the novel and highly effective mRNA vaccines BNT162b2 (Pfizer-BioNTech) and mRNA-1273 (Moderna) utilized lipid nanoparticles to deliver nucleoside-modified RNA encoding the SARS-CoV-2 spike protein (1–4). An alternative platform, Ad26.COV2.S (Johnson & Johnson), delivered spike-encoding DNA via a nonreplicating adenoviral vector (5). Although all 3 vaccines target the same antigen, their distinct formulations and platforms provide a unique framework to dissect human immune priming and boosting. Despite this opportunity, to our knowledge, no prior study has performed a head-to-head, single-cell multiomic comparison of all 3 FDA-approved COVID-19 vaccines in SARS-CoV-2–naive individuals.
Immune responses, particularly adaptive immune responses following the second dose of mRNA vaccines, have been well characterized, including the robust expansion of CD4+ and CD8+ T cells, memory B cells, and the production of SARS-CoV-2–neutralizing antibodies (6–15). In contrast, early innate responses, especially in human myeloid cells after the first dose, remain poorly understood. Although some studies noted IFN-stimulated gene (ISG) induction (13, 16–18), the magnitude, timing, and cell-type specificity of this response have not been resolved. Our study fills this critical gap by providing a single-cell, multiomic map of early innate priming in humans after mRNA vaccination.
Here, we leveraged the unique conditions of the early days of the pandemic to study how immune responses develop in SARS-CoV-2–naive healthy adults (n = 31) vaccinated with either mRNA (BNT162b2) (3), mRNA-1273 (4), or the adenoviral vector (Ad26.COV2.S) (5) COVID-19 vaccines. Using single-cell multiomics alongside cytokine and antibody profiling, we mapped the temporal dynamics of innate responses at single-cell resolution. We uncovered a distinct myeloid priming program induced by type I IFNs after the first mRNA dose, and a broader boosting response involving both type I and type II IFNs after the second mRNA dose or Ad26.COV2.S. By defining dose-specific and vaccine platform–specific IFN programs in human myeloid cells, our study provides a mechanistic framework for optimizing vaccine platform choice, scheduling, and formulation, with implications for improving protection in vulnerable populations.
ResultsHead-to-head comparison of 3 COVID-19 vaccines in SARS-CoV-2–naive adults. Thirty-one healthy adults with no prior history of SARS-CoV-2 infection or vaccination were enrolled between spring and summer 2021 at the UConn Health Center. Of these, 28 participants received mRNA vaccines, including both BNT162b2 (n = 22) and mRNA-1273 (n = 6), and 3 received the adenoviral vector vaccine Ad26.COV2.S (Figure 1A and Supplemental Table 1; supplemental material available online with this article; https://doi.org/10.1172/jci.insight.199245DS1). Age distributions were comparable across groups (Figure 1A and Supplemental Table 1).
Figure 1A longitudinal systems immunology study of immune responses to COVID-19 vaccines. (A) Study design. Blood samples were collected longitudinally from individuals (n = 31) vaccinated with either 2 doses of mRNA vaccines (Pfizer-BioNTech or Moderna) or 1 dose of the adenovirus-based Johnson & Johnson vaccine. Antibody titers were measured before and 3–4 weeks after each vaccination. A subset of donors (6 Pfizer, 6 Moderna, 3 Johnson & Johnson) was selected for in-depth profiling, including serum cytokine analysis and DOGMA-Seq. Age distributions among vaccine groups were compared using the Kruskal-Wallis test. (B) IgG titers against SARS-CoV-2 spike protein measured by ELISA before and after vaccination. Antibody titers further increased after the mRNA booster dose. (C) Serum levels of CXCL10 and IFN-γ at key time points were quantified using ELLA. Both cytokines increased on day 1 (D1) after the adenoviral vaccine and following the mRNA booster dose. (D and E) UMAP projections of DOGMA-Seq data for transcriptome (D, scRNA-Seq) and chromatin accessibility (E, scATAC-Seq) modalities. Canonical lineage markers (surface protein and gene expression) annotated major immune cell populations. For B and C, statistical significance was assessed using 2-tailed paired t tests: *P < 0.05, ****P < 0.0001.
mRNA vaccines were administered in 2 doses, whereas Ad26.COV2.S was administered as a single dose, according to CDC recommendations. Blood samples were collected at key time points: baseline (prevaccination), day 1 (innate responses), day 7 (early adaptive responses), and days 25–35 (antibody responses) after each vaccine dose (Figure 1A and Supplemental Table 1), totaling 7 time points for mRNA vaccines and 4 time points for Ad26.COV2.S. The isolated PBMCs were used for genomic profiling, and plasma samples were used for cytokine and antibody titer profiling.
Anti–SARS-CoV-2 spike protein IgG titers were measured by ELISA at baseline and 3–4 weeks after each vaccination. All vaccine groups showed a significant increase in antibody titers after the first dose (P < 0.0001), followed by an additional increase after the second mRNA dose (Figure 1B). After 2 doses, individuals vaccinated with mRNA vaccines had significantly higher titers (P = 0.0369) than those receiving a single dose of Ad26.COV2.S (Supplemental Figure 1A). Despite the dose difference between BNT162b2 (30 μg) and mRNA-1273 (100 μg), antibody titers were comparable between the 2 mRNA groups at both time points (Supplemental Figure 1A). No correlation was observed between antibody titer levels and participant age (Supplemental Figure 1B).
Plasma levels of 6 cytokines were quantified at baseline, day 1, and day 7 after each vaccination using the ELLA platform (Bio-Techne) (Figure 1C, Supplemental Figure 1C, and Supplemental Table 2). The first dose of mRNA vaccines modestly elevated CXCL10 levels at day 1 (P = 0.14 for Moderna, P = 0.013 for Pfizer), and the second dose induced stronger induction of both CXCL10 and IFN-γ (CXCL10: P = 0.019 for Moderna, P = 0.00036 for Pfizer; IFN-γ: P = 0.019 for Moderna, P = 0.0125 for Pfizer, Figure 1C). In contrast, both IFN-γ and CXCL10 levels increased to levels comparable to those induced by mRNA vaccines at day 1 after Ad26.COV2.S. Furthermore, we observed a correlation between IFN-γ levels after a booster vaccine (day 1, vaccine dose 2) and fold change of IgG titers (day 42 vs. baseline), suggesting a potential link between IFN responses and humoral immunity (Supplemental Figure 1D). No significant changes were observed for IL-8, TNF-α, CCL2, or IL-10 levels after any vaccine (Supplemental Figure 1C).
Together, these data showed robust antibody responses and limited cytokine induction after the first mRNA dose, followed by enhanced humoral and innate cytokine responses after the second dose. No significant differences were observed between the 2 mRNA platforms despite significant differences in their doses.
Longitudinal profiling of PBMCs with a single-cell multiome assay. To investigate the cellular dynamics underlying vaccine-induced immune responses, we performed longitudinal profiling of PBMCs from 15 participants who received mRNA-1273 (n = 6), BNT162b2 (n = 6), or Ad26.COV2.S (n = 3) (Figure 1A) using DNA-Oriented Genome and Transcriptome Mapping by Analysis sequencing (DOGMA-Seq). DOGMA-Seq is a multimodal single-cell assay that enables the simultaneous measurement of surface protein expression, gene expression, and chromatin accessibility from the same cells (19). DOGMA-Seq libraries were generated at baseline, day 1, and day 7 for Ad26.COV2.S; at baseline and day 1 after the first mRNA vaccine dose; and at baseline, day 1, and day 7 after the second mRNA dose (Figure 1A). A total of 69 samples were multiplexed using Cell Hashing (20) and sequenced in batches of 6–8.
After quality control and filtering, we obtained 354,958 high-quality cells for single-cell RNA-Seq (scRNA-Seq) and 216,320 for single-cell assay for transposase-accessible chromatin sequencing (scATAC-Seq) (Figure 1, D and E, Supplemental Figure 1E, Supplemental Figure 2, Supplemental Figure 3, A and B, and Supplemental Table 3). We recovered approximately 5,000 high-quality cells per sample (donor and time point). Surface protein expression and canonical marker genes identified major immune cell populations including monocytes, NK cells (CD335), B cells (CD19), CD4+ T cells (CD4), CD8+ T cells (CD8), γδ T cells (TCR-Vδ2), and mucosal-associated invariant T (MAIT) cells (TCR-Vα7.2) (Figure 1D and Supplemental Figure 1E). Cell-type annotations from the RNA-Seq modality were transferred to the ATAC-Seq modality (Figure 1E). This dataset enabled a high-resolution, longitudinal analysis of vaccine-induced immune responses, with transcriptional and epigenetic programs captured from the same cells providing direct insights into coordinated regulatory states at single-cell resolution for each vaccine.
Strong transcriptional IFN responses in myeloid cells after the second mRNA and Ad26.COV2.S vaccines. Clustering of single-cell transcriptomic profiles from the myeloid compartment (n = 71,184 cells) revealed 3 major subsets: CD14+ monocytes, CD16+ monocytes, and DCs (Figure 2A and Supplemental Figure 4A). We obtained a median of 628 CD14+ monocytes, 112 CD16+ monocytes, and 40 DCs per sample (donor and time point) (Supplemental Figure 4B). One day after Ad26.COV2.S vaccination, the frequency of CD14+ monocytes increased substantially, comprising up to 50% of total PBMCs on average, a 2.7-fold expansion relative to baseline (Supplemental Figure 4C). No significant changes in cell composition were observed for CD16+ monocytes and DCs. In contrast, mRNA vaccination elicited a more modest but statistically significant increase in CD14+ monocytes after both the first dose (P = 0.002, fold change = 1.25) and the second dose (P = 0.0034, fold change = 1.72) (Supplemental Figure 4C).
Figure 2mRNA vaccines induce a robust IFN response after the booster dose. (A) UMAP representation of the myeloid compartment from DOGMA-Seq, annotated using canonical markers to identify CD14+ monocytes, CD16+ monocytes, and DCs. IFN-stimulated gene (ISG) subsets were defined based on the expression of ISG markers in each subset. (B) Kernel density estimation plots of ISG scores for each vaccine and time point show that Johnson & Johnson elicits a strong IFN response 1 day after vaccination; mRNA vaccines induce a robust response after the booster. (C) Differential gene expression analysis in CD14+ monocytes comparing postvaccination time points (D1V1, D1V2) to baseline. Thresholds for significance: log2FC > 0.5 and P < 0.05. ISGs are highlighted in green. Venn diagram shows the overlap of differentially expressed genes (DEGs) from D1 (Johnson & Johnson), D1V2 (Moderna), and D1V2 (Pfizer). A χ2 test was conducted across the 3 groups to evaluate whether the overlap of DEGs was statistically significant, indicating a shared transcriptional response. (D) Heatmap showing the top 25 IFN-related genes upregulated 1 day after the booster (D1V2) for mRNA vaccine recipients. (E) Type-I IFN expression score calculated from manually curated list (n = 43) shows stronger type I IFN response after Johnson & Johnson vaccination compared with mRNA responses. Statistical significance between time points was assessed using a 2-tailed paired t test; comparisons between adenovirus vaccine and mRNA vaccine (Moderna and Pfizer combined) groups were performed using the unpaired t test: NS, nonsignificant, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.
Within each myeloid subset, we detected ISG+ cells characterized by the high expression of ISGs, including critical antiviral and signaling molecules MX1, GBP2, WARS, and STAT1 (Figure 2A and Supplemental Figure 4A). These ISG+ populations significantly expanded at day 1 after Ad26.COV2.S and after the mRNA second dose (Supplemental Figure 4D), as confirmed by cumulative ISG scores calculated at the single-cell level using previously published gene set modules (n = 96 genes) (21) (Figure 2B). Notably, this IFN response was stronger after Ad26.COV2.S compared with the mRNA second dose (P = 0.004).
To further quantify transcriptional changes upon vaccination responses, we performed differential expression analysis by comparing postvaccination time points to baseline in CD14+ and CD16+ monocytes and DCs (Figure 2C and Supplemental Table 4). The highest number of differentially expressed genes was detected at day 1 after Ad26.COV2.S and at day 1 after the second mRNA vaccination in CD14+ monocytes (Figure 2C). Across all platforms, 389 genes were commonly induced (χ2 test, P = 3.66 × 10–19) (Figure 2C), and these were enriched in IFN-associated pathways (Supplemental Figure 4E and Supplemental Table 4), suggesting a shared core IFN response in CD14+ monocytes upon mRNA booster vaccination and after Ad26.COV2.S.
Although relatively few differentially expressed genes were detected after the first mRNA vaccination in CD14+ monocytes, several ISGs, including OAS3, IFI44, and MX2, were upregulated (Figure 2D). Analysis of curated gene sets (Supplemental Table 5) revealed that type I IFN response genes (n = 43) were upregulated after the first mRNA vaccine, whereas type II IFN response genes (n = 53) were not (Figure 2E and Supplemental Figure 4F). In contrast, both type I and type II IFN response genes were activated after the second mRNA vaccine and after Ad26.COV2.S vaccination. Notably, Ad26.COV2.S induced a significantly higher type I IFN response than the mRNA booster, whereas the type II IFN responses were comparable.
Together, these results show that mRNA vaccination elicits a modest type I IFN response after the first dose in myeloid cells, followed by a robust, broader IFN program after the second dose. The transcriptional IFN response to Ad26.COV2.S resembled that of the mRNA booster but was markedly stronger in magnitude, driven by the stronger type I IFN responses.
The first mRNA dose induces a distinct type I IFN response state in myeloid cells. We next focused on myeloid cell responses to the first mRNA vaccination by performing a detailed analysis of CD14+ monocytes. This analysis revealed 3 distinct IFN-associated transcriptional states: ISG-low, ISG-dim, and ISG-high (Figure 3A and Supplemental Figure 5A). ISG-dim cells selectively expressed a subset of canonical type I IFN response genes, including antiviral effectors MX1 and MX2; ISGs IFI44, IFI44L, and ISG15; and the pattern recognition receptor (PRR) DDX58/RIG-I (Figure 3, B and C, and Supplemental Table 6). In contrast, ISG-high cells displayed a broader IFN signature, including the upregulation of IFN-γ–induced guanylate-binding proteins (GBPs), the tRNA synthetase WARS, and immunoproteasome components such as PSMB9 (LMP2) (22–24) (Figure 3C). Notably, the ISG-dim state expanded after the first mRNA dose; approximately 12% of CD14+ monocytes were in this state after BNT162b2 or mRNA-1273 vaccination, compared with approximately 5% at baseline (~2.36-fold increase) (Figure 3D).
Figure 3Distinct ISG states associated with primary and booster mRNA vaccination in CD14+ monocytes and DCs. (A) UMAPs highlight ISG states (ISG-low, ISG-dim, ISG-high) in CD14+ monocytes and demonstrate expression of marker gene. (B) Box plots show expression of MX1 and GBP1. (C) Heatmap showing the expression of genes differentially expressed between 3 ISG states in Moderna (M) and Pfizer (P) vaccine recipients. (D) The proportion of ISG-dim and ISG-high cells within CD14+ monocytes across mRNA vaccines and time points. Note that the ISG-dim population expands specifically after the first mRNA vaccination. (E) UMAP highlights ISG states (ISG-low, ISG-dim, ISG-high) in cDC2s upon reanalysis of publicly available data. (F) The proportion of ISG-dim and ISG-high cells within cDC2s across time points. (G) The heatmap displays the expression levels of marker genes for ISG states in cDC2s. For D, statistical comparisons between time points were performed using the 1-sided Wilcoxon test: ***P < 0.001.
We identified cell surface markers enriched in the ISG-dim and ISG-high states: CD169 (SIGLEC1) and CD64 (FCGR1A), respectively (Supplemental Figure 5B). Next, we performed flow cytometry on CD14+ monocytes from 6 mRNA-vaccinated donors to validate these transcriptional subsets at the protein level. Frequency of CD14+CD169+ cells, reflecting type I IFN responses, increased significantly after the first vaccination (P = 0.036) and reached its highest levels after the second dose (P = 0.014). In contrast, CD14+CD64+ cells, reflecting type II IFN responses, increased significantly only after the booster vaccination (P = 0.0007) (Supplemental Figure 5C), consistent with the transcriptomic patterns observed at corresponding time points (Supplemental Figure 5D). Notably, the ISG-dim transcriptional signature remained restricted to myeloid populations.
We reanalyzed cellular indexing of transcriptomes and epitopes by sequencing (CITE-Seq) data from DC-enriched PBMCs of 6 BNT162b2-vaccinated individuals (16) and confirmed the presence of all 3 ISG states (ISG-low, ISG-dim, ISG-high) in CD14+ monocytes in this independent cohort (Supplemental Table 3). Furthermore, we detected the same subsets in CD16+ monocytes and cDC2s (Figure 3E and Supplemental Figure 6A). In all 3 cell types, the ISG-dim state expanded on day 1 after the first vaccine, peaked at day 2, and declined by day 7 (Figure 3F and Supplemental Figure 6B). In contrast, the ISG-high state emerged on day 1 after the second vaccine (Figure 3F and Supplemental Figure 6B). Marker gene expression in this cohort mirrored our findings: ISG-dim cells highly expressed MX1, MX2, and DDX58, whereas ISG-high cells highly expressed a broader panel of ISGs, including IFN-γ–induced genes (Figure 3G and Supplemental Figure 6C).
Taken together, priming induced an early program (ISG-dim state), characterized by type I IFN response genes, while boosting elicited a robust response (ISG-high state), marked by broader activation of both type I and type II IFN response genes.
Priming induces epigenetic activation of ISGF3 complex in CD14+ monocytes. Transcription factors (TFs) such as IFN regulatory factors (IRFs) and STATs are key regulators of IFN responses (25, 26). To determine whether the ISG-dim and ISG-high states exhibit distinct TF activity, we first examined the ISGF3 complex — the primary effector of type I IFN signaling (27, 28). The ISGF3 components STAT1, STAT2, and IRF9 were upregulated in both ISG-dim and ISG-high CD14+ monocytes (Figure 4, A and B). In contrast, IRF1 and IRF8, critical to type II IFN–driven programs (26, 29), were only induced in the ISG-high state (Figure 4, A and B).
Figure 4ISGF3 transcription factor complex activity increases in ISG-dim subset. (A) Heatmap displaying expression of key IFN response regulatory transcription factors across ISG subsets. (B) Average expression of the members of the ISGF3 complex (IRF9, STAT1, and STAT2) and IRF1. (C) Average chromatin accessibility at ISGF3 member binding sites across 3 ISG subsets using validated binding sites from a published study. (D) We calculated z scores of ISGF3 target genes (n = 17) based on the average expression of these genes in individual donors across ISG-defined states. (E) Genome browser views of MX1 and DDX58 loci across ISG states, highlighting predicted and validated (ChIP-Seq) binding sites for IRF9, STAT1, and STAT2. (F) Pearson’s correlation between promoter accessibility and gene expression for MX1.
To explore the epigenetic characteristics of these ISG states (Figure 3A), we analyzed chromatin accessibility in CD14+ monocytes. Using previously published ChIP-Seq data from human monocytes for STAT1, STAT2, and IRF9 (27), we identified validated ISGF3 binding sites (27) and assessed their accessibility in our ATAC-Seq data. ISGF3 binding sites were significantly more accessible in both ISG-dim and ISG-high monocytes compared with ISG-low monocytes (Figure 4C and Supplemental Figure 7A). Only 17 genes had binding sites at their promoters for all 3 ISGF3 TFs (Figure 4D and Supplemental 7B). These genes included MX1, OAS1, OAS3, and DDX58 (RIG-I), markers of the ISG-dim state. The promoters of these genes exhibited increased chromatin accessibility in ISG-dim cells and harbored ISGF3 binding motifs (Figure 4, E and F, and Supplemental Figure 7B), further supporting the notion that they are direct ISGF3 targets.
ChromVAR analysis of IRF and STAT motifs, together with longitudinal transcriptomic data, demonstrated that the increased binding accessibility and expression observed for these TFs after the first mRNA vaccine dose was transient (Supplemental Figure 8A); no significant changes were detected at the second baseline prior to booster vaccination. Similarly, the changes observed upon booster vaccination (i.e., the ISG-high state) had resolved by day 7 after vaccination.
These analyses indicate that the transcriptional and epigenetic activation of the ISGF3 complex is a defining feature of the ISG-dim state induced by the priming by mRNA vaccines.
Exposing monocytes to type I and type II IFNs recapitulates ISG-dim and ISG-high states, respectively. We hypothesized that type I and type II
Comments (0)