Gottardo, Raphaël

Abstract Spatial transcriptomics links gene expression to tissue architecture, providing a mechanistic view of cellular organization. Yet existing datasets cover few donors and miss the complexity of human disease. Experimental costs remain prohibitive, and large-scale profiling is impractically slow for population-level studies. Accurate computational methods are urgently needed. Predicting gene expression from standard histology, however, remains an open problem, as current approaches transfer poorly to unseen cohorts and diseases. Here, we present Phoenix, a (latent) flow matching generative model that infers pan-cancer spatially resolved single-cell gene expression with high accuracy. Phoenix analyzes treatment response in silico: Applied to 763 head and neck cancer patients, it identified three new spatial biomarkers that we validated across two cancers (breast cancer, n = 84; ovarian cancer, n = 157) and treatment regimens (platinum, trastuzumab). Phoenix generalizes beyond carcinomas: In a large sarcoma cohort (802 tissue microarray cores), it accurately predicted cell-type-specific signatures in held-out samples and captured chemotherapy-induced immune remodeling. Phoenix also extends across species: In a mouse model, it accurately predicted the expression of pancreatic cancer lineage markers and the mutant mKras^G12D allele in silico. In total, we evaluated Phoenix on over 10,000 patients. Our results establish virtual spatial transcriptomics as a scalable framework for studying tissue organization, therapeutic response, and disease mechanisms.

AbstractPurpose: Immunotherapy as a treatment for cancer holds the promise of complete and durable tumor remission, yet the immunosuppressive environment created by many tumors, advanced patient age, and previous treatments with cytotoxic agents may limit the approach. The activity of motolimod (VTX-2337), a potent and selective Toll-like receptor 8 (TLR8) agonist, was therefore assessed in the context of advanced, late-stage cancer patients.Experimental Design: The repertoire of mediators induced from human peripheral blood mononuclear cells in response to motolimod was characterized. Translational studies in cynomolgus monkeys elucidated the activity of motolimod on an intact immune system, identified biomarkers of TLR8 activation, and defined the relationship between the pharmacokinetic and pharmacodynamic (PK/PD) response. The PK/PD relationship for motolimod in cancer patients was assessed, compared with preclinical findings, and contrasted with activity in healthy volunteers.Results: In late-stage cancer patients, plasma levels of multiple biomarkers, including IL6, G-CSF, MCP-1, and MIP1-β, increased with increasing motolimod dose. The magnitude and breadth of the biomarker response closely aligned with the response seen in preclinical studies, demonstrating that advanced cancer patients remained responsive to TLR8 activation. In addition, the PK/PD response in cancer patients closely aligned with the activity of motolimod seen in healthy volunteers.Conclusions: Late-stage cancer patients are highly sensitive to TLR8 activation by motolimod. Tumor burden, advanced age, and prior treatment history with cytotoxic agents did not moderate or modify the response predicted by nonclinical studies and confirmed in healthy volunteers. Clin Cancer Res; 21(24); 5445–52. ©2015 AACR.