Tumor development is a dynamic evolutionary process, where the survival of the fittest translates into the persistence of cancer cell clones most adaptable to their surrounding environment. Although the tumor microenvironment (TME) is composed of non-cancerous cells, these cells are far from "normal" as their behavior often diverges dramatically from their counterparts in healthy tissues. In fact, many immune cell types within the TME become dysfunctional or, worse, co-opted into supporting tumor growth and progression.

For instance, we previously demonstrated that neutrophils recruited to the TME in non-small cell lung cancer (NSCLC) adopt pro-tumorigenic functions (1). These neutrophils, commonly referred to as N2, tumor-associated neutrophils (TANs), or myeloid-derived suppressor cells (MDSCs), are reprogrammed by cancer cells. One key mechanism of this reprogramming is intercellular communication via ligand-receptor signaling. In NSCLC, we showed that cancer-cell-derived CXCL5 and CXCL6 recruit pro-tumorigenic neutrophils through the CXCR2 receptor (1). In ovarian cancer, we identified IL-4 secretion by tumor cells as a driver of macrophage polarization toward an M2-like, tumor-promoting phenotype. These macrophages then contribute to CD8+ T cell exhaustion and immune evasion (2). In another case, we found that ovarian cancer cells downregulate CCL7 secretion, impeding the infiltration of immune cells, especially cytotoxic T cells, into the tumor (2). In lung cancer, TANs not only support tumor growth but also induce transdifferentiation of cancer cells into more aggressive forms (1).

These cancer-cell-driven alterations in the TME not only enhance tumor progression but also promote therapy resistance. For example, IL-4–mediated macrophage reprogramming leads to resistance against anti-PD-1 immune checkpoint inhibitors (2). Similarly, CXCL5/6-driven neutrophil recruitment facilitates adeno-to-squamous transdifferentiation, enabling lung adenocarcinomas to escape dependence on oncogenic drivers like EGFR or KRAS and resist targeted therapies (1).

Critically, tumor evolution is not a linear process. Each cancer cell finds its own adaptive means due to both cell-intrinsic variations such as mutational and epigenetic/translational differences as well as cell-extrinsic factors in the TME. This gives rise to intratumoral heterogeneity (ITH), a fundamental challenge in cancer therapy. ITH underlies why tumors that initially respond to treatments, such as EGFR inhibitors in lung adenocarcinoma, inevitably relapse. Resistant clones may emerge by rewiring signaling pathways, undergoing lineage plasticity, or adopting other survival strategies. In other cases, resistance is not acquired but intrinsic, as seen in ovarian tumors where IL-4–driven immune suppression results in de novo resistance to anti–PD-1 therapy (2).

Our research seeks to unravel the complex interplay between ITH and the TME, two hallmarks of cancer biology that drive aggressiveness and treatment failure. We aim to dissect the intercellular communication networks that shape tumor heterogeneity and immunosuppressive microenvironments. By understanding how immune pressure sculpts ITH and how distinct TMEs enable immune evasion, we hope to identify novel vulnerabilities and inform the development of more effective, durable cancer therapies.

We use mouse models to investigate intratumoral heterogeneity (ITH) and the tumor microenvironment (TME) in solid tumors, focusing primarily on lung and ovarian cancers, two of the most common and clinically challenging cancer types (1-3). However, accurately modeling ITH in mice and interrogating the TME in heterogeneous tumors presents significant technical challenges.

To address these limitations, we leverage Perturb-map, an advanced spatial functional genomics platform developed by the Brown Lab at Mount Sinai (4). Perturb-map allows us to spatially map cancer cells and their microenvironment in intact tissue sections at single-cell resolution, while simultaneously performing CRISPR-based perturbations. This powerful combination of spatial profiling and genome engineering enables us to introduce targeted genetic alterations, such as gene knockouts or expression modulation, and directly assess both cell-intrinsic and microenvironmental (extrinsic) consequences.

Using Perturb-map, we discovered that cancer cell-derived cytokines are key drivers of clonal TMEs, revealing a direct link between cancer cell heterogeneity and microenvironmental diversity. This approach has allowed us to elucidate how distinct immune niches influence tumor evolution and promote resistance to immunotherapy (2).

We continue to harness this innovative technology to dissect the dynamic interplay between ITH and the TME, two central features of cancer biology, with the goal of uncovering mechanisms of treatment resistance and identifying new therapeutic opportunities.