Identification of Breast Cancer Immune Subtypes by Analyzing Bulk Tumor and Single Cell Transcriptomes

Background: The histological and molecular classification of breast cancer (BC) is being used in the clinical management of this disease. However, subtyping of BC based on the tumor immune microenvironment (TIME) remains insufficiently explored, although such investigation may provide new insights into intratumor heterogeneity in BC and potential clinical implications for BC immunotherapy.Methods: Based on the enrichment scores of 28 immune cell types, we performed clustering analysis of transcriptomic data to identify immune-specific subtypes of BC using six different datasets, including five bulk tumor datasets and one single-cell dataset. We further analyzed the molecular and clinical features of these subtypes.Results: Consistently in the six datasets, we identified three BC subtypes: BC-ImH, BC-ImM, and BC-ImL, which had high, medium, and low immune signature scores, respectively. BC-ImH displayed a significantly better survival prognosis than BC-ImL. Triple-negative BC (TNBC) and human epidermal growth factor receptor-2-positive (HER2+) BC were likely to have the highest proportion in BC-ImH and the lowest proportion in BC-ImL. In contrast, hormone receptor-positive (HR+) BC had the highest proportion in BC-ImL and the lowest proportion in BC-ImH. Furthermore, BC-ImH had the highest tumor mutation burden (TMB) and predicted neoantigens, while BC-ImL had the highest somatic copy number alteration (SCNA) scores. It is consistent with that TMB and SCNA correlate positively and negatively with anti-tumor immune response, respectively. TP53 had the highest mutation rate in BC-ImH and the lowest mutation rate in BC-ImL, supporting that TP53 mutations promote anti-tumor immune response in BC. In contrast, PIK3CA displayed the highest mutation rate in BC-ImM, while GATA3 had the highest mutation rate in BC-ImL. Besides immune pathways, many oncogenic pathways were upregulated in BC-ImH, including ErbB, MAPK, VEGF, and Wnt signaling pathways; the activities of these pathways correlated positively with immune signature scores in BC.Conclusions: The tumors with the strong immune response (“hot” tumors) have better clinical outcomes than the tumors with the weak immune response (“cold” tumors) in BC. TNBC and HER2+ BC are more immunogenic, while HR + BC is less immunogenic. Certain HER2+ or HR + BC patients could be propitious to immunotherapy in addition to TNBC.

Follicular Lymphoma Microenvironment Signatures Define Patients Subsets Obtaining Long Term Clinical Benefit after Single-Agent First-Line Anti-CD20 Immunotherapy

The role of first-line single agent Rituximab immunotherapy in follicular lymphoma (FL) is still a matter of debate. Although a subset of patients (pts) may obtain long term benefit with upfront immunotherapy, first-line therapy with standard chemoimmunotherapy offers better results in terms of progression free survival (PFS) compared to single agent Rituximab. On the other hand, there is no clear demonstration of sizeable long term benefit with first-line Rituximab compared to an initial wait and see approach, as most FL pts will ultimately be exposed to chemotherapy at some point in their disease course. Here we show that the efficacy of front line anti CD20 immunotherapy in FL could be dependent on microenvironmental factors, and that specific immune signatures could define FL pts subsets obtaining maximal benefit from upfront anti CD20 immunotherapy. First, we retrospectively analyzed the outcome of our single center cohort of 81 FL pts treated with first-line single agent Rituximab therapy with (n=53) or without (n=28) maintenance. The vast majority of pts considered in this analysis were treated in the context of several clinical trials exploring the efficacy of upfront single agent originator (SAKK 35/98, 35/03, 35/10: n= 68) or biosimilar (JASMINE and REFLECTIONS B328/06 n= 9) Rituximab, which were ran at our center from 2000 to 2018. Four patients received off-label single agent Rituximab. Median age was 55 years (y), 52% (42) of patients were female, 72% (58) stage III-IV, 76% (62) were Follicular international prognostic index (FLIPI) score 0-2, 26% (21) were bulky, 52% (42) were high tumor burden according to the GELF criteria. After a median follow-up of 9.5 y, the overall survival (OS) and PFS rates were 82% and 35% respectively. With a long follow-up, these data are in line with previous findings indicating that a sizeable fraction of FL pts derive long term benefit from upfront single agent Rituximab maintaining a continuous complete remission. Paraffin embedded tissue from initial diagnosis was available in 39 pts. In order to dissect determinants of Rituximab immunotherapy efficacy in this homogeneous chemo-naive population, we analyzed FL diagnostic biopsies (n=39 FL + 5 healthy controls) with targeted gene expression profiling (T-GEP) on the NanoString platform, using the PanCancer Immune Profiling panel, which includes 730 genes belonging to the most relevant immunologic checkpoints and pathways. A 20-gene signature including genes involved in chemokine-cytokine signaling, T-regulatory cells, natural-killer cell activity and interleukin-17 signaling was significantly associated with the achievement of a complete response after induction +/- maintenance treatment. A simple 6-gene immune signature (hereafter ImSig) was found to be significantly associated with PFS, with IL22RA2, CCL22, TNFRSF4, IL17RB, CCL19 overexpression and CD209 downregulation being associated with worse outcome. By applying the maxstat package 10-y PFS was 65% for ImSig low pts vs 6% for ImSig high pts (p<0.0001). In multivariate analysis only the 6-gene ImSig and Rituximab maintenance retained independent prognostic value (p<0.001 and 0.002 respectively). As opposite, GELF criteria (present/absent = 27/12) and FLIPI score (0-2/3-5 = 23/16) were not associated with PFS. The 6-gene immune signature was validated in silico in 2 independent publicly available cohorts of FL pts treated with upfront chemoimmunotherapy: a cohort of 137 pts (Silva et al 2019, Affimetryx platform) and a cohort of 50 pts (Bararia et al. 2020), the latter analyzed on the NanoString Platform with the same T-GEP panel (PanCancer Immune Profiling) used in our discovery cohort. The 6-gene signature was confirmed to be a powerful outcome predictor in these 2 chemoimmunotherapy-treated cohorts, and notably the 10-y PFS rates of ImSig low vs high patients mirrored the results observed with single agent immunotherapy in the discovery cohort. The results here reported indicate that pts with a favorable immune signature could derive maximal benefit from a first-line chemo-free treatment approach with single agent Rituximab, achieving and maintaining complete remission in the long term, irrespective of the tumor burden and other clinical variables. Thus, profiling of FL microenvironment with T-GEP could provide a useful tool for selecting patients who may be suitable for a chemo-free upfront treatment with anti CD20 immunotherapy. Disclosures Derenzini: TG-THERAPEUTICS: Research Funding; ASTRA-ZENECA: Consultancy, Other: ADVISORY-BOARD; ADC-THERAPEUTICS: Research Funding; TAKEDA: Research Funding; BEIGENE: Other: ADVISORY BOARD. Pileri: CELGENE: Other: ADVISORY BOARD; NANOSTRING: Other: ADVISORY BOARD; ROCHE: Other: ADVISORY-BOARD. Tarella: ADC-THERAPEUTICS: Other: ADVISORY BOARD; Abbvie: Other: ADVISORY BOARD. OffLabel Disclosure: First line single agent Rituximab in Follicular Lymphoma

Daily caloric restriction limits tumor growth more effectively than caloric cycling regardless of dietary composition

Cancer incidence increases with age and is a leading cause of death. Caloric restriction (CR) confers benefits on health and survival and delays cancer. However, due to CR’s stringency, dietary alternatives offering the same cancer protection have become increasingly attractive. Short cycles of a plant-based diet designed to mimic fasting (FMD) are protective against tumorigenesis without the chronic restriction of calories. Yet, it is unclear whether the fasting time, level of dietary restriction, or nutrient composition is the primary driver behind cancer protection. Using a breast cancer model in mice, we compare the potency of daily CR to that of periodic caloric cycling on FMD or an isocaloric standard laboratory chow against primary tumor growth and metastatic burden. Here, we report that daily CR provides greater protection against tumor growth and metastasis to the lung, which may be in part due to the unique immune signature observed with daily CR.

A novel gene expression signature-based on B-cell proportion to predict prognosis of patients with lung adenocarcinoma

Background This study aimed to develop a reliable immune signature based on B-cell proportion to predict the prognosis and benefit of immunotherapy in LUAD. Methods The proportion of immune cells in the TCGA-LUAD dataset was estimated using MCP-counter. The Least Absolute Shrinkage and Selector Operation was used to identify a prognostic signature and validated in an independent cohort. We used quantitative reverse transcription-polymerase chain reaction (qRT-PCR) data and formalin-fixed paraffin-embedded (FFPE) specimens immunohistochemistry to illustrate the correlation between prognostic signature and leukocyte migration. Results We found that the relative abundance of B lineage positively correlated with overall survival. Then, we identified a 13-gene risk-score prognostic signature based on B lineage abundance in the testing cohort and validated it in a cohort from the GEO dataset. This model remained strongly predictive of prognoses across clinical subgroups. Further analysis revealed that patients with a low-risk score were characterized by B-cell activation and leukocyte migration, which was also confirmed in FFPE specimens by qRT-PCR and immunohistochemistry. Finally, this immune signature was an independent prognostic factor in the composite nomogram of clinical characteristics. Conclusions In conclusion, the 13-gene immune signature based on B-cell proportion may serve as a powerful prognostic tool in LUAD.