graphic

Follicular lymphomas (FL) are spawned by elusive cancer precursor cells (CPC) that evade current therapies and thereby contribute to disease relapse. Schroers-Martin, Soo, Brisou, Scherer, and colleagues used sensitive liquid biopsy techniques to define the somatic genetic lesions of these CPCs to interrogate serial prediagnostic blood, saliva, and tissue samples collected from two large screening studies from patients who ultimately developed FL. In so doing, key recurrent somatic mutations in FL CPCs were identified with lesions in CREBBP's KAT domain as specific founder mutations of premalignant B-cell precursors detectable years before FL diagnosis.

See article, p. 1310.

graphic

The human pancreas is notoriously difficult to study due to rapid postmortem degradation and lack of indication to obtain tissue in the absence of disease. Utilizing a brain-dead organ donation pipeline to obtain disease-free pancreata with no warm ischemia time, Carpenter, Elhossiny, Kadiyala, and colleagues identified pancreatic cancer precursor lesions, known as PanINs, in the majority of organs. Integrating single-cell RNA sequencing and spatial transcriptomics revealed a unique microenvironment in sporadic PanIN lesions that was distinct from that of tumors. Conversely, PanIN epithelial cells from healthy pancreata were remarkably transcriptionally similar to cancer cells, suggesting early activation of neoplastic pathways in tumorigenesis.

See article, p. 1324.

graphic

Barrett's esophagus (BE) and gastric intestinal metaplasia (GIM) are believed to be precancerous lesions of gastric and esophageal adenocarcinoma, respectively. While these cancers are in many respects molecularly similar, a debate exists about the origin and progression of precancerous lesions. Using single-cell RNA sequencing, Nowicki-Osuch and colleagues demonstrated molecular similarity at the precancerous BE and GIM stages. Individual cells of both lesions were characterized by cellular mosaics defined as a simultaneous expression of gastric and intestinal transcriptional profiles that already had the hallmarks of cancer. Furthermore, a comparison of BE and GIM microenvironments showed that the fibroblasts of precancerous lesions display activation profiles of cancer-associated fibroblasts.

See article, p. 1346.

graphic

The proclivity of cancer to adapt and evade environmental pressures through classical evolutionary principles makes it difficult to effectively treat. Additionally, an understanding of these evolutionary processes can inform future treatment strategies. Spain, Coulton, Lobon, Rowan, Schnidrig, and colleagues utilized research autopsy to overcome the limitations of sampling tumors during life as well as multiomic profiling of metastatic melanoma to provide high-resolution insight into the many evolutionary pathways that can ultimately lead to patient death, including widespread copy-number alterations, extrachromosomal DNA, and mutational processes. Also featured are analyses of dormancy in brain metastases and polyclonal seeding at the level of whole-genome doubling.

See article, p. 1364.

graphic

Predicting response to antineoplastic drugs continues to be a challenge, and complementary precision medicine paradigms could improve the prioritization of personalized treatments. Mundi, Dela Cruz, and colleagues used patient-derived xenograft models obtained from tumors refractory or intolerant to standard-of-care treatment to assess the efficacy of 35 drugs predicted by OncoTarget and OncoTreat, two transcriptome-based tools that identify high-affinity inhibitors of master regulator (MR) proteins and drugs that invert transcriptional activity of hyperconnected MR modules, respectively. Both platforms demonstrated ability to predict antitumor drug activity in vivo, with validation in a case study supporting their further clinical development.

See article, p. 1386.

graphic

The BCL2 inhibitor venetoclax (VEN) in combination with azacitidine (5-AZA) is transforming acute myeloid leukemia (AML) therapy. The heterogenous composition of AML cells at various differentiation stages is a major problem for response prediction. Waclawiczek, Leppä, Renders, and colleagues combined transcriptomic, functional, and clinical data to show that immature leukemic stem cells (LSC) are the primary targets of 5-AZA/VEN and their elimination determines therapy outcome. A flow cytometry–based “Mediators of apoptosis combinatorial score” (MAC-Score) that combines BCL2 family expression ratios in LSCs can predict initial response and event-free survival. In summary, combinatorial levels of BCL2 family members in LSCs are a key denominator of response, and MAC scoring reliably predicts response to 5-AZA/VEN.

See article, p. 1408.

graphic

Pancreatic cancer is characterized by an undruggable genome, an immunosuppressive microenvironment, and stromal inflammation, rendering it highly resistant to chemotherapy and immunotherapy. Using KRASTP53 genomic cooperativity as a model for high-risk biology, Bianchi and colleagues identify Cxcl1 as a key cancer cell–intrinsic factor that not only governs the recruitment and suppressive behavior of tumor-associated neutrophils, but also spatially constrains effector T cells from their antitumor functions. Moreover, Cxcl1–CXCR2 engagement promotes neutrophil-derived TNF signaling, which further elicits stromal inflammation and T-cell dysfunction via transmembrane TNF–TNFR2 interactions. Targeting TNFR2 signaling improves sensitivity to cytotoxic chemotherapy, offering a novel strategy to improve outcomes in this deadly disease.

See article, p. 1428.

graphic

The metastatic and primary tumor microenvironments are significantly different. Faget and colleagues found that limiting immune stromal p38MAPKα signaling restrained metastatic breast progression by shifting macrophages away from an immunosuppressive phenotype and increasing CD4+ Th1 T-cell antitumor responses. Furthermore, Tnfsf4 (OX40L) was found to be poorly expressed in the bone metastatic microenvironment, and combining p38α inhibition with agonist anti-OX40 synergized to decrease metastatic growth. When this approach was combined with cytotoxic T-cell engagement, it led to clearance of metastatic disease and long-term immunologic memory. Finally, a signature in patients was identified that might predict who will respond to this approach to limit metastatic progression.

See article, p. 1454.

graphic

Cancer-associated IDH mutants produce (R)-2HG, an oncometabolite that seemingly indiscriminately inhibits 2-oxoglutarate–dependent enzymes. To identify functionally relevant targets of (R)-2HG, Gunn and colleagues used positive-selection CRISPR–Cas9 screens and identified three H3K4 histone lysine demethylases, KDM5A, KDM5C, and KDM5D, as tumor suppressor enzymes whose inhibition by (R)-2HG contributes to mutant IDH–mediated transformation. These findings provide novel molecular insights into the biology of IDH-mutant cancers, including acute myeloid leukemia and glioma.

See article, p. 1478.