RESEARCH

How Tfh cells shape antibody immunity in people.

We study human T follicular helper cells and the germinal centers they organize, across vaccination, infection and transplantation. Our work follows four connected themes.

01GERMINAL CENTERS · INFLAMMATION

Germinal centers and inflammatory signaling

How the inflammation present at the start of an immune response shapes the Tfh–B cell partnership and the memory it leaves behind.

Tfh cells guide B cells through selection in the germinal center, which determines the quality and durability of antibody responses. The inflammatory signals around a response, from infection or from aging, can change how this process unfolds for months or years afterward.

We showed that SARS-CoV-2 infection leaves an interferon-linked imprint on memory CD4 T cells that persists through later vaccination. We now ask how acute and age-associated inflammation alters Tfh differentiation and germinal center output, including in people living with long COVID.

Long COVIDAgingInterferon signalingLymph node sampling
UMAP of spike-specific CD4 T cell clusters and density plots comparing vaccine-primed and infection-primed individuals
Spike-specific memory CD4 T cells: clusters (top) and cell density in vaccine-primed versus infection-primed people (bottom). Gray-Gaillard et al., Science Immunology 2024.
RELATED PAPERS
SCIENCE IMMUNOLOGY · 2024SARS-CoV-2 inflammation durably imprints memory CD4 T cellsIMMUNITY · 2026Lymphatic constraint of germinal centers optimizes protective antibody responsesISCIENCE · 2023mRNA COVID-19 vaccine elicits potent adaptive immune response without the acute inflammation of SARS-CoV-2 infection
02XENOTRANSPLANTATION

Tfh contributions to xenograft rejection

What human T cells do when they encounter a pig organ, and how Tfh help could drive the antibodies that reject it.

Pig-to-human transplantation could ease the organ shortage, but antibody-mediated rejection remains a central barrier, and those antibodies depend on help from Tfh cells in germinal centers.

As part of NYU Langone’s pig-to-human decedent xenotransplant program, we follow human T cells after kidney xenotransplantation: which clones recognize pig antigens, whether they differentiate into Tfh and drive germinal center and anti-pig antibody responses, and how immunosuppression changes these responses. We also work to define what xenoreactive T cell receptors recognize.

Decedent modelTCR repertoireGerminal centersImmunosuppression
03VACCINES · PREDICTION

Cellular predictors of vaccine responses

Reading early cellular signals to know who a vaccine will protect, and why it fails for others.

Vaccines do not work equally well for everyone. Responses are weaker in older adults, in people receiving immunotherapy or immunosuppression, and in a subset of otherwise healthy adults. We use vaccination as a controlled immune challenge in people, sampling blood and draining lymph nodes to find the cellular features that separate durable responses from poor ones.

Our hepatitis B vaccine studies link poor responses to Tfh dysfunction in the lymph node. Related work spans influenza, mpox and SARS-CoV-2 vaccines, building on our earlier finding that circulating Tfh are early, sensitive readouts of vaccine responses.

Hepatitis BInfluenzaMpoxSystems immunology
H&E and immunofluorescence images of germinal centers in a lymph node core biopsy stained for IgD, PD-1 and Ki67
Germinal centers in a lymph node core biopsy after hepatitis B vaccination, stained for IgD, PD-1 and Ki67. Chen et al., bioRxiv 2026.
RELATED PAPERS
BIORXIV · 2026Tfh dysfunction is associated with poor responses to HBV vaccinationNATURE IMMUNOLOGY · 2022PD-1 directed immunotherapy alters Tfh and humoral immune responses to seasonal influenza vaccineSCIENCE IMMUNOLOGY · 2017Successive annual influenza vaccination induces a recurrent oligoclonotypic memory response in circulating T follicular helper cells
04SPATIAL ANALYSIS · SOFTWARE

Algorithm development for spatial immunology

Open-source tools for measuring how immune cells organize inside human tissue, and whether that organization predicts outcomes.

Multiplex imaging and spatial transcriptomics now record where every cell sits in a lymph node. Turning those maps into biology requires statistics built for human samples, where differences between people are large and each tissue is its own experiment. We develop R packages that test cell–cell organization across patients rather than within a single image.

ESImap (Effect Size Interaction mapping) scores whether two cell types sit next to each other more often than chance in each sample, then compares groups with effect sizes across biological replicates. Using it, we found that contacts between activated Tregs and exhausted CD8 T cells in melanoma sentinel lymph nodes are associated with later recurrence. RippleR, developed by Yee Hoon Foong, analyzes Xenium and Giotto data in concentric rings around tumor regions to follow how cell composition and gene expression change with distance.

Spatial statisticsMultiplex imagingXeniumOpen-source R
HOW WE WORK

From clinical samples to single cells

Human cohorts
Vaccine and infection studies with longitudinal blood and lymph node sampling.
Single-cell multi-omics
Transcriptome, surface protein and paired TCR/BCR sequencing from the same cells.
Repertoire tracking
Following individual T and B cell clones across time and tissues.
Spatial biology
Mapping cell neighborhoods inside human lymphoid tissue.
Supported by NIAID, the NIH RECOVER Initiative and the Hevolution Foundation.