Mammary glands undergo morphologic and biochemical changes during various physiologic stages of life, specifically during the transition from virgin to pregnancy, lactation, and involution (1–3). The complete cycle of structural and functional differentiation depends on the coordinated action of prolactin, insulin, adrenal corticoids, and ovarian hormones (4,5). Atypical ductal hyperplasia is an abnormal ductal epithelial cell proliferative condition that does not invade the periductal stroma (6). In women, atypical ductal hyperplasia, which may become more aggressive and ultimately fill the lumen of the duct, is considered to be a physiologic precursor to the development of ductal carcinoma in situ (DCIS). Although the histopathology of DCIS subtypes is well defined, there are few experimental models to evaluate the molecular mechanisms underlying DCIS formation or to evaluate cancer chemopreventive agents. One model is the mouse mammary gland organ culture (MMOC) (7,8), in which the entire cycle of mammary gland morphology and physiology can be simulated with appropriate hormonal supplementation of a chemically defined medium. We have shown previously that the MMOC model is useful for evaluating the underlying molecular mechanisms of lesion formation because mammary glands exposed to 7,12-dimethylbenz[a]anthracene (DMBA) develop hyperplastic mammary alveolar lesions in the presence of the nonovarian steroid hormones aldosterone and hydrocortisone (9). These lesions do not regress to the nonproliferating state comprised mainly of ducts and a few end-buds after the removal of growth-promoting hormones. Moreover, cells isolated from these lesions form adenocarcinomas when transplanted into syngeneic mice (10). In addition, the model is useful for testing the efficacy of chemopreventive agents to inhibit DMBA-induced lesions (11,12), although tamoxifen did not reduce lesion formation. Because of the proposed role of ovarian hormones in breast carcinogenesis, we evaluated the effects of estrogen and progesterone on DMBA-induced lesions in the MMOC model. Ductal lesions were induced by incubating mouse mammary glands in serum-free Waymouth MB752/1 medium (7,8), in which aldosterone and hydrocortisone were replaced with estradiol17 (0.001 g/mL) and progesterone (1 g/mL). The glands were treated with DMBA (2 g/mL) for 24 hours on the third day of culture. After 24 days, glands were fixed in 10% formalin, and histopathologic sections were evaluated for ductal lesions. For progesterone receptor analyses, mammary glands were incubated with growth-promoting hormones either alone or in the presence of estradiol-17 (1 nM) and/or tamoxifen (1 M) for 6 days. On day 6, the glands were fixed in 10% formalin, and histologic sections were prepared and immunostained with antibodies to the progesterone receptor (Neomarkers, Fremont, CA) (13). Sections were evaluated semiquantitatively for the expression of progesterone receptor according to the intensity of staining (14). The difference between means of percent incidence (of atypical ductal hyperplasia) and percent induction (of progesterone receptor expression) of control and treated groups was analyzed by use of Student’s t test for independent samples (SPSS® statistical software; version 6.1.3; Chicago, IL). All statistical tests were twosided. Histologically, ovarian hormonedependent lesions induced by DMBA in MMOC were predominantly of ductal rather than alveolar origin, similar to human breast hyperplastic and premalignant lesions. The ducts in control glands (i.e., not treated with DMBA) were largely lumina lined with one or two layers of epithelial cells (Fig. 1, A). The ducts in DMBA-treated glands were thickened and lined with five to six layers of hyperplastic cells (defined by a thickness of more than three cell layers) (Fig. 1, C). The lumina of some ducts were occluded completely by intraductal outgrowths, and the epithelial cells often formed alveolar and papillary structures (Fig. 1, E). Transverse sections through the lesions showed a combination of proliferating epithelial cells and areas of necrosis (Fig. 1, F). Close examination of the intraductal lesions showed that aggressive lesions (i.e., those completely occluded with ductal epithelial cells) were composed of atypical cells (variable in size and form), nuclei with intense chromatin staining, reduced intracellular spaces, and a reduced number of mitotic figures (Fig. 1, G). Light microscopic images of the sectioned glands were converted to digitized images by use of software written in MATLAB® (Source: MATLAB version 5.3 licensed to the Department of Electrical Engineering and Computer Science, University of Illinois, Chicago) to quantify the ductal lesions. The score for the area covered by epithelial cells was 0.05 for control glands (Fig. 1, B), 0.72 for hyperplastic lesions (Fig. 1, D), and 1.13 for aggressive ductal lesions (image not shown). Various degrees (ranging from greater than three layers of epithelial cells to completely occluded ducts) of intraductal atypical hyperplastic lesions were found in 387 of 486 fields from histologic sections of 61 DMBA-treated glands, which was equivalent to an incidence of 76.1% (95% confidence interval [CI] 69.3% to 82.9%). There were a few falsepositive lesions in four of 74 fields from sections of seven control glands, which was equivalent to an incidence of 5.4% (95% CI 2.8% to 8%). There was a
Abstract Transcriptomic screens of brain tumor (glioblastoma; GBM) parenchymal cells indicated tumor supporting traits of the GBM microenvironment, but largely excluded mitochondrially enriched gene-sets. Here, we show that a mitochondrial transcript of GBM parenchymal cells contributes to therapy resistance. We inspected the non-coding transcriptome of human GBM associated myeloid cells (GAM) and observed an upregulation of the mitochondrial ribosomal subunit MT-RNR2, which contains an open reading frame for the signaling peptide humanin. Immunohistology disclosed that humanin was preponderant in GAM. In vitro assays with a range of human stem-like GBM cells (GSCs) revealed that nanomolar concentrations of humanin can drive tumor cell expansion and chemoresistance to temozolomide. A series of genetic, pharmacological and Western blotting experiments showed that extracellular humanin increased GSC viability via stimulation of the GP130/IL6ST receptor and MAPK (Erk) activation. Humanin responsiveness was subject to inter-individual heterogeneity and predominated in GSC with high expression levels of the IL6ST subunit. Mechanistically, humanin promoted the ATR-dependent DNA-repair machinery in GSCs via induction of the DNA-clamp component HUS1. A slice culture model containing human microglia and GSCs confirmed these observations. Humanin is absent from the mouse brain. Exogenous delivery of humanin into orthotopic GBM mouse models or over-expression specifically of a secreted humanin variant recapitulated the in vitro findings. Blockade of the trimeric interleukin receptor with the brain permeant, FDA-approved drug bazedoxifene blocked humanin-mediated chemoresistance in vitro and in vivo. Overall, we identified a clinically applicable compound together with a predictive marker to prevent chemoresistance in a human-specific GBM model.
Tumor microenvironments can promote stem cell maintenance, tumor growth, and therapeutic resistance, findings linked by the tumor-initiating cell hypothesis. Standard of care for glioblastoma (GBM) includes temozolomide chemotherapy, which is not curative, due, in part, to residual therapy-resistant brain tumor-initiating cells (BTICs). Temozolomide efficacy may be increased by targeting carbonic anhydrase 9 (CA9), a hypoxia-responsive gene important for maintaining the altered pH gradient of tumor cells. Using patient-derived GBM xenograft cells, we explored whether CA9 and CA12 inhibitor SLC-0111 could decrease GBM growth in combination with temozolomide or influence percentages of BTICs after chemotherapy. In multiple GBMs, SLC-0111 used concurrently with temozolomide reduced cell growth and induced cell cycle arrest via DNA damage in vitro. In addition, this treatment shifted tumor metabolism to a suppressed bioenergetic state in vivo. SLC-0111 also inhibited the enrichment of BTICs after temozolomide treatment determined via CD133 expression and neurosphere formation capacity. GBM xenografts treated with SLC-0111 in combination with temozolomide regressed significantly, and this effect was greater than that of temozolomide or SLC-0111 alone. We determined that SLC-0111 improves the efficacy of temozolomide to extend survival of GBM-bearing mice and should be explored as a treatment strategy in combination with current standard of care.
Gliomas are recalcitrant brain tumors. Anti-glioma immunity and immunopathogenic responses are critical contributors for better survival of isocitrate dehydrogenase-mutant (IDHmut) over wild-type (IDHwt) gliomas. Despite this correlative pattern of immunity and survival, an unbiased understanding of cell-type specific transcriptomic and epigenomic states of glioma-derived myeloid cells beyond immunosuppressive paradigms remains elusive.
Methods
To this end, we performed single-cell RNA-sequencing (scRNA-seq) on 140,000 tumor-associated immune cells from eighteen IDH mutation classified primary (N=8) and recurrent (N=10) human gliomas and three non-glioma brains (NGBs). We performed unsupervised clustering and cell annotation based on overlapping canonical lineage and signal dependent transcription factors. Gene ontology and gene set enrichment analyses were performed to define functional states of glioma associated myeloid cells.
Results
Our analyses revealed twelve myeloid cell types across glioma subgroups. We noted abundant microglial cells in IDHmut than IDHwt gliomas. Concomitantly, continnum of microglia and macrophage phenotypes were observed in IDHwt glioma, which exhibit more severe tumor pathologies. Strikingly, we identified a hybrid microglia/macrophage cell subset with enriched interferon module inferred from our gene ontology analyses. These hybrid phagocytes were significantly increased in recurrent IDHwt gliomas. As tissue macrophages exhibit multifaceted polarization in response to microenvironmental cues, we clarify the existence of microglia/macrophage functional states beyond M1/M2 paradigms exemplified by the presence of palmitic-, oleic- acid, and glucocorticoid responsive polarized states. Specifically, certain microglia and monocyte-derived subpopulations were associated with antigen presentation gene modules, akin to cross-presenting dendritic cells. Furthermore, immune related gene ontology analysis identified enriched antigen presentation and phagocytosis gene modules in distinct microglia-like clusters. Importantly, the phagocytic immunomodulator; Triggering Receptor Expressed on Myeloid Cells 2 (TREM2) was upregulated in these microglia-like cells. Contrary to tumor promoting role of TREM2 myeloid cells in non-brain cancers, we identify TREM2 axis as a regulator of antigen presentation. Additionally, single cell- Assay for Transposase-Accessible Chromatin using sequencing (sc-ATAC-seq) on ~40,000 tumor-associated microglia revealed genes associated with IFN-gamma/IL-12/IL-10 pathway were negatively regulated in IDHmut/IDHwt microglia compared to homeostatic microglia in NGBs.
Conclusions
In summary, our study sculpts transcriptional and epigenomic details and re-defines glioma-specific immune contexture for downstream immunogenomics applications. We specifically reveal interferon and TREM2 nodes on microglia-like phagocytic cells as clinically tractable anti-glioma immunotherapy target.
'%3e%3cpath fill='%23012169' d='M0 0v30h60V0z'/%3e%3cpath stroke='%23fff' stroke-width='6' d='m0 0 60 30m0-30L0 30'/%3e%3cpath stroke='%23C8102E' stroke-width='4' d='m0 0 60 30m0-30L0 30' clip-path='url(%23b)'/%3e%3cpath stroke='%23fff' stroke-width='10' d='M30 0v30M0 15h60'/%3e%3cpath stroke='%23C8102E' stroke-width='6' d='M30 0v30M0 15h60'/%3e%3c/g%3e%3c/svg%3e)