Identifying novel ovarian tumor biomarkers through mining of the transcriptome of circulating immune cells: A proof-of-concept study.

OBJECTIVE: Treatment of high-grade serous ovarian cancer (HGSOC) will benefit from early detection of cancer. Here, we provide proof-of-concept data supporting the hypothesis that circulating immune cells, because of their early recognition of tumors and the tumor microenvironment, can be considered for biomarker discovery. METHODS: Longitudinal blood samples from C57BL/6 mice bearing syngeneic ovarian tumors and peripheral blood mononuclear cells (PBMC) from healthy postmenopausal women and newly diagnosed for HGSOC patients were subjected to RNASeq. The results from human immune cells were validated using Affymetrix microarrays. Differentially expressed transcripts in immune cells from tumor-bearing mice and HGSOC patients were compared to matching controls. RESULTS: A total of 1282 transcripts (798 and 484, up- and downregulated, respectively) were differentially expressed in the tumor-bearing mice as compared with controls. Top 100 genes showing longitudinal changes in gene expression 2, 4, 7, and 18 days after tumor implantation were identified. Analysis of the PBMC from healthy post-menopausal women and HGSOC patients identified 4382 differentially expressed genes and 519 of these were validated through Affymetrix microarray analysis. A total of 384 genes, including IL-1R2, CH3L1, Infitm1, FP42, CXC42, Hdc, Spib, and Sema6b, were differentially expressed in the human and mouse datasets. CONCLUSION: The PBMC transcriptome shows longitudinal changes in response to the progressing tumor. Several potential biomarker transcripts were identified in HGSOC patients and mouse models. Monitoring their expression in individual PBMC subsets can serve as additional discriminator for the diagnosis of HGSOC.

Different Human Immune Lineage Compositions Are Generated in Non-Conditioned NBSGW Mice Depending on HSPC Source.

NBSGW mice are highly immunodeficient and carry a hypomorphic mutation in the c-kit gene, providing a host environment that supports robust human hematopoietic expansion without pre-conditioning. These mice thus provide a model to investigate human hematopoietic engraftment in the absence of conditioning-associated damage. We compared transplantation of human CD34+ HSPCs purified from three different sources: umbilical cord blood, adult bone marrow, and adult G-CSF mobilized peripheral blood. HSPCs from mobilized peripheral blood were significantly more efficient (as a function of starting HSPC dose) than either cord blood or bone marrow HSPCs at generating high levels of human chimerism in the murine blood and bone marrow by 12 weeks post-transplantation. While T cells do not develop in this model due to thymic atrophy, all three HSPC sources generated a human compartment that included B lymphocytic, myeloid, and granulocytic lineages. However, the proportions of these lineages varied significantly according to HSPC source. Mobilized blood HSPCs produced a strikingly higher proportion of granulocyte lineage cells (~35% as compared to ~5%), whereas bone marrow HSPC output was dominated by B lymphocytic cells, and cord blood HSPC output was enriched for myeloid lineages. Following transplantation, all three HSPC sources showed a shift in the CD34+ subset towards CD45RA+ progenitors along with a complete loss of the CD45RA-CD49f+ long-term HSC subpopulation, suggesting this model promotes mainly short-term HSC activity. Mice transplanted with cord blood HSPCs maintained a diversified human immune compartment for at least 36 weeks after the primary transplant, although mice given adult bone marrow HSPCs had lost diversity and contained only myeloid cells by this time point. Finally, to assess the impact of non-HSPCs on transplantation outcome, we also tested mice transplanted with total or T cell-depleted adult bone marrow mononuclear cells. Total bone marrow mononuclear cell transplants produced significantly lower human chimerism compared to purified HSPCs, and T-depletion rescued B cell levels but not other lineages. Together these results reveal marked differences in engraftment efficiency and lineage commitment according to HSPC source and suggest that T cells and other non-HSPC populations affect lineage output even in the absence of conditioning-associated inflammation.

The role of RORγt at maternal-fetal interface during murine pregnancy.

PROBLEM: Innate lymphoid cells (ILCs, including NK cells) and their subsets are the most frequent lymphocytes at the maternal-fetal interface (decidua). Recent recognition of extensive ILC subset diversity at mucosal sites and the possible role they might play at different stages of pregnancy poses questions about their composition and lineage stability. Namely, RORγt-dependent ILC3s have been recognized as a key cellular mediator of tissue organization in the gut and secondary lymphoid organs, prompting examination of their distribution and role in decidua during pregnancy. METHOD OF STUDY: We employed highly polychromatic flow cytometry with conventional and machine learning-aided analysis to map ILC subsets and dissected the role of canonical transcription factor RORγt using fate-mapping animals and RORγt-/- animals. RESULTS: We demonstrate a comprehensive immunome map of ILCs/NKs, revealing a dynamic interface even in the absence of antigenic or allogeneic challenge. Strikingly, we demonstrate plasticity of RORγt expression in decidual ILCs with across gestation. However, gross reproductive efficiency is not affected in RORγt-/- animals. CONCLUSION: These results indicated that RORγt+ ILCs are highly plastic at the maternal-fetal interface, but dispensable for normal pregnancy, revealing a novel mechanism of transcriptional immunoregulation in pregnancy.

Decidual-Placental Immune Landscape During Syngeneic Murine Pregnancy.

Adaptive immune system, principally governed by the T cells-dendritic cells (DCs) nexus, is an essential mediator of gestational fetal tolerance and protection against infection. However, the exact composition and dynamics of DCs and T cell subsets in gestational tissues are not well understood. These are controlled in human physiology by a complex interplay of alloantigen distribution and presentation, cellular/humoral active and passive tolerance, hormones/chemokines/angiogenic factors and their gradients, systemic and local microbial communities. Reductive discrimination of these factors in physiology and pathology of model systems and humans requires simplification of the model and increased resolution of interrogative technologies. As a baseline, we have studied the gestational tissue dynamics in the syngeneic C57BL/6 mice, as the simplest immunological environment, and focused on validating the approach to increased data density and computational analysis pipeline afforded by highly polychromatic flow cytometry and machine learning interpretation. We mapped DC and T cell subsets, and comprehensively examined their maternal (decidual)-fetal (placental) interface dynamics. Both frequency and composition of decidual DCs changed across gestation, with a dramatic increase in myeloid DCs in early pregnancy, and exclusion of plasmacytoid DCs. CD4+ T cells, in contrast, were lower at all gestational ages and an unusual CD4-CD8-TCRαß+group was prominent at mid-pregnancy. Dimensionality reduction with machine learning-aided clustering revealed that CD4-CD8- T cells were phenotypically different from CD4+ and CD8+ T cells. Additionally, divergence between maternal decidual and fetal placental compartment was prominent, with absence of DCs from the placenta, but not decidua or embryo. These results provide a novel framework and a syngeneic baseline on which the specific role of alloantigen/tolerance, polymicrobial environment, and models of pregnancy pathology can be precisely modeled and analyzed.