Causal regulators of human T-cell and macrophage activation from perturbation screens, and a replicated coupling of T-cell exhaustion markers with macrophage M1 activation in patients under checkpoint blockade

Engineering anti-tumour immunity requires knowing which genes control T-cell and macrophage programs, and how the two compartments relate to each other in patients. We built ImmunoTwin, an analysis framework that addresses both questions with public data and with every test, baseline and criterion fixed before the result was computed. From a genome-scale CRISPR-interference Perturb-seq in primary human CD4+ T cells (22 million cells, four donors) we derive a map of regulators of the cytotoxic effector program that is calibrated, reverses direction under CRISPR activation (rho = -0.53), and tracks interferon-gamma production among the hits of independent sorting-based screens (rho = 0.69). From a Perturb-seq in primary human macrophages we recover the TLR4-NF-kappaB pathway as regulator of the lipopolysaccharide response (pathway-level P = 0.0004) and find that transcriptional effects track TNF protein measured in independent donors (rho = 0.58). In both cell types, however, a generic measure of perturbation strength recovers the hits of functional screens better than the program-specific effect (area under the ROC curve 0.96 versus 0.73 in T cells; 0.87 versus 0.72 in macrophages). In 90 patients from seven cohorts with tumour sampled before and during checkpoint blockade, T-cell and macrophage programs shifted together, most strongly T-cell exhaustion markers and the macrophage M1 program (r = 0.43). Fixed as a hypothesis and tested once in 86 further patients from six cohorts, this association replicated (r = 0.51, permutation P = 0.0001) and was specific against expression-matched random gene sets (P = 0.004). The baseline state of one compartment did not improve the forecast of the other, so the criterion we had set for calling the framework a patient-level digital twin was not met. ImmunoTwin provides two validated perturbation maps, a replicated cross-compartment association in patients, and a caution on how such maps are benchmarked.

Authors

Publication Details

Journal
bioRxiv (Cold Spring Harbor Laboratory)
Published
2026-11-19
DOI
https://doi.org/10.64898/2026.11.18.767605
Primary Topic
Cancer Immunotherapy and Biomarkers
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

Causal regulators of human T-cell and macrophage activation from perturbation screens, and a replicated coupling of T-cell exhaustion markers with macrophage M1 activation in patients under checkpoint blockade

Ivan Merelli A, Francesca Anna Cupaioli
bioRxiv (Cold Spring Harbor Laboratory)
Cancer Immunotherapy and Biomarkers
preprint

Causal regulators of human T-cell and macrophage activation from perturbation screens, and a replicated coupling of T-cell exhaustion markers with macrophage M1 activation in patients under checkpoint blockade

Ivan Merelli A, Francesca Anna Cupaioli
preprint en

Abstract

Engineering anti-tumour immunity requires knowing which genes control T-cell and macrophage programs, and how the two compartments relate to each other in patients. We built ImmunoTwin, an analysis framework that addresses both questions with public data and with every test, baseline and criterion fixed before the result was computed. From a genome-scale CRISPR-interference Perturb-seq in primary human CD4+ T cells (22 million cells, four donors) we derive a map of regulators of the cytotoxic effector program that is calibrated, reverses direction under CRISPR activation (rho = -0.53), and tracks interferon-gamma production among the hits of independent sorting-based screens (rho = 0.69). From a Perturb-seq in primary human macrophages we recover the TLR4-NF-kappaB pathway as regulator of the lipopolysaccharide response (pathway-level P = 0.0004) and find that transcriptional effects track TNF protein measured in independent donors (rho = 0.58). In both cell types, however, a generic measure of perturbation strength recovers the hits of functional screens better than the program-specific effect (area under the ROC curve 0.96 versus 0.73 in T cells; 0.87 versus 0.72 in macrophages). In 90 patients from seven cohorts with tumour sampled before and during checkpoint blockade, T-cell and macrophage programs shifted together, most strongly T-cell exhaustion markers and the macrophage M1 program (r = 0.43). Fixed as a hypothesis and tested once in 86 further patients from six cohorts, this association replicated (r = 0.51, permutation P = 0.0001) and was specific against expression-matched random gene sets (P = 0.004). The baseline state of one compartment did not improve the forecast of the other, so the criterion we had set for calling the framework a patient-level digital twin was not met. ImmunoTwin provides two validated perturbation maps, a replicated cross-compartment association in patients, and a caution on how such maps are benchmarked.

bioRxiv (Cold Spring Harbor Laboratory)
Cancer Immunotherapy and Biomarkers
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon