MultiK: an automated tool to determine optimal cluster numbers in single-cell RNA sequencing data.

Single-cell RNA sequencing (scRNA-seq) provides new opportunities to characterize cell populations, typically accomplished through some type of clustering analysis. Estimation of the optimal cluster number (K) is a crucial step but often ignored. Our approach improves most current scRNA-seq cluster methods by providing an objective estimation of the number of groups using a multi-resolution perspective. MultiK is a tool for objective selection of insightful Ks and achieves high robustness through a consensus clustering approach. We demonstrate that MultiK identifies reproducible groups in scRNA-seq data, thus providing an objective means to estimating the number of possible groups or cell-type populations present.

Sensitivity to targeted therapy differs between HER2-amplified breast cancer cells harboring kinase and helical domain mutations in PIK3CA.

BACKGROUND: HER2-amplified breast cancer is a clinically defined subtype of breast cancer for which there are multiple viable targeted therapies. Resistance to these targeted therapies is a common problem, but the mechanisms by which resistance occurs remain incompletely defined. One mechanism that has been proposed is through mutation of genes in the PI3-kinase pathway. Intracellular signaling from the HER2 pathway can occur through PI3-kinase, and mutations of the encoding gene PIK3CA are known to be oncogenic. Mutations in PIK3CA co-occur with HER2-amplification in ~ 20% of cases within the HER2-amplified subtype. METHODS: We generated isogenic knockin mutants of each PIK3CA hotspot mutation in HER2-amplified breast cancer cells using adeno-associated virus-mediated gene targeting. Isogenic clones were analyzed using a combinatorial drug screen to determine differential responses to HER2-targeted therapy. Western blot analysis and immunofluorescence uncovered unique intracellular signaling dynamics in cells resistant to HER2-targeted therapy. Subsequent combinatorial drug screens were used to explore neuregulin-1-mediated resistance to HER2-targeted therapy. Finally, results from in vitro experiments were extrapolated to publicly available datasets. RESULTS: Treatment with HER2-targeted therapy reveals that mutations in the kinase domain (H1047R) but not the helical domain (E545K) increase resistance to lapatinib. Mechanistically, sustained AKT signaling drives lapatinib resistance in cells with the kinase domain mutation, as demonstrated by staining for the intracellular product of PI3-kinase, PIP3. This resistance can be overcome by co-treatment with an inhibitor to the downstream kinase AKT. Additionally, knockout of the PIP3 phosphatase, PTEN, phenocopies this result. We also show that neuregulin-1, a ligand for HER-family receptors, confers resistance to cells harboring either hotspot mutation and modulates response to combinatorial therapy. Finally, we show clinical evidence that the hotspot mutations have distinct expression profiles related to therapeutic resistance through analysis of TCGA and METABRIC data cohorts. CONCLUSION: Our results demonstrate unique intracellular signaling differences depending on which mutation in PIK3CA the cell harbors. Only mutations in the kinase domain fully activate the PI3-kinase signaling pathway and maintain downstream signaling in the presence of HER2 inhibition. Moreover, we show there is potentially clinical importance in understanding both the PIK3CA mutational status and levels of neuregulin-1 expression in patients with HER2-amplified breast cancer treated with targeted therapy and that these problems warrant further pre-clinical and clinical testing.

CRISPR-based knock-in mutagenesis of the pioneer transcription factor FOXA1: optimization of strategies for multi-allelic proteins in cancer cells.

Precise genome engineering of living cells has been revolutionized by the introduction of the highly specific and easily programmable properties of the clustered regularly interspaced short palindromic repeats (CRISPR) technology. This has greatly accelerated research into human health and has facilitated the discovery of novel therapeutics. CRISPR-Cas9 is most widely employed for its ability to inactivate or knockout specific genes, but can be also used to introduce subtle site-specific substitutions of DNA sequences that can lead to changes in the amino acid composition of proteins. Despite the proven success of CRISPR-based knock-in strategies of genes in typical diploid cells (i.e., cells containing two sets of chromosomes), precise editing of cancer cells, that typically have unstable genomes and multiple copies of chromosomes, is more challenging and not adequately addressed in the literature. Herein, we detail our methodology for replacing endogenous proteins with intended knock-in mutants in polyploid cancer cells and discuss our experimental design, screening strategy, and facile allele frequency estimation methodology. As proof of principle, we performed genome editing of specific amino acids within the pioneer transcription factor FOXA1, a critical component of estrogen and androgen receptor signaling, in MCF-7 breast cancer cells. We confirm mutant FOXA1 protein expression and intended amino acid substitutions via western blotting and mass spectrometry. In addition, we show that mutant allele frequency estimation is easily achieved by topoisomerase-based cloning combined with allele-specific PCR, which we later confirmed by next-generation RNA-sequencing. Typically, there are 4 - 5 copies (alleles) of FOXA1 in breast cancer cells, making the editing of this protein inherently challenging. As a result, most studies that focus on FOXA1 mutants rely on ectopic overexpression of FOXA1 from a plasmid. Therefore, we provide an optimized methodology for replacing endogenous wild-type FOXA1 with precise knock-in mutants to enable the systematic analysis of its molecular mechanisms within the appropriate physiological context.

FGFR4 regulates tumor subtype differentiation in luminal breast cancer and metastatic disease.

Mechanisms driving tumor progression from less aggressive subtypes to more aggressive states represent key targets for therapy. We identified a subset of luminal A primary breast tumors that give rise to HER2-enriched (HER2E) subtype metastases, but remain clinically HER2 negative (cHER2-). By testing the unique genetic and transcriptomic features of these cases, we developed the hypothesis that FGFR4 likely participates in this subtype switching. To evaluate this, we developed 2 FGFR4 genomic signatures using a patient-derived xenograft (PDX) model treated with an FGFR4 inhibitor, which inhibited PDX growth in vivo. Bulk tumor gene expression analysis and single-cell RNA sequencing demonstrated that the inhibition of FGFR4 signaling caused molecular switching. In the Molecular Taxonomy of Breast Cancer International Consortium (METABRIC) breast cancer cohort, FGFR4-induced and FGFR4-repressed signatures each predicted overall survival. Additionally, the FGFR4-induced signature was an independent prognostic factor beyond subtype and stage. Supervised analysis of 77 primary tumors with paired metastases revealed that the FGFR4-induced signature was significantly higher in luminal/ER+ tumor metastases compared with their primaries. Finally, multivariate analysis demonstrated that the FGFR4-induced signature also predicted site-specific metastasis for lung, liver, and brain, but not for bone or lymph nodes. These data identify a link between FGFR4-regulated genes and metastasis, suggesting treatment options for FGFR4-positive patients, whose high expression is not caused by mutation or amplification.

B Cells and T Follicular Helper Cells Mediate Response to Checkpoint Inhibitors in High Mutation Burden Mouse Models of Breast Cancer.

This study identifies mechanisms mediating responses to immune checkpoint inhibitors using mouse models of triple-negative breast cancer. By creating new mammary tumor models, we find that tumor mutation burden and specific immune cells are associated with response. Further, we developed a rich resource of single-cell RNA-seq and bulk mRNA-seq data of immunotherapy-treated and non-treated tumors from sensitive and resistant murine models. Using this, we uncover that immune checkpoint therapy induces T follicular helper cell activation of B cells to facilitate the anti-tumor response in these models. We also show that B cell activation of T cells and the generation of antibody are key to immunotherapy response and propose a new biomarker for immune checkpoint therapy. In total, this work presents resources of new preclinical models of breast cancer with large mRNA-seq and single-cell RNA-seq datasets annotated for sensitivity to therapy and uncovers new components of response to immune checkpoint inhibitors.

Ki-67 is required for maintenance of cancer stem cells but not cell proliferation.

Ki-67 expression is correlated with cell proliferation and is a prognostic marker for various cancers; however, its function is unknown. Here we demonstrate that genetic disruption of Ki-67 in human epithelial breast and colon cancer cells depletes the cancer stem cell niche. Ki-67 null cells had a proliferative disadvantage compared to wildtype controls in colony formation assays and displayed increased sensitivity to various chemotherapies. Ki-67 null cancer cells showed decreased and delayed tumor formation in xenograft assays, which was associated with a reduction in cancer stem cell markers. Immunohistochemical analyses of human breast cancers revealed that Ki-67 expression is maintained at equivalent or greater levels in metastatic sites of disease compared to matched primary tumors, suggesting that maintenance of Ki-67 expression is associated with metastatic/clonogenic potential. These results elucidate Ki-67's role in maintaining the cancer stem cell niche, which has potential diagnostic and therapeutic implications for human malignancies.

TMSB4Y is a candidate tumor suppressor on the Y chromosome and is deleted in male breast cancer.

Male breast cancer comprises less than 1% of breast cancer diagnoses. Although estrogen exposure has been causally linked to the development of female breast cancers, the etiology of male breast cancer is unclear. Here, we show via fluorescence in situ hybridization (FISH) and droplet digital PCR (ddPCR) that the Y chromosome was clonally lost at a frequency of ~16% (5/31) in two independent cohorts of male breast cancer patients. We also show somatic loss of the Y chromosome gene TMSB4Y in a male breast tumor, confirming prior reports of loss at this locus in male breast cancers. To further understand the function of TMSB4Y, we created inducible cell lines of TMSB4Y in the female human breast epithelial cell line MCF-10A. Expression of TMSB4Y resulted in aberrant cellular morphology and reduced cell proliferation, with a corresponding reduction in the fraction of metaphase cells. We further show that TMSB4Y interacts directly with ß-actin, the main component of the actin cytoskeleton and a cell cycle modulator. Taken together, our results suggest that clonal loss of the Y chromosome may contribute to male breast carcinogenesis, and that the TMSB4Y gene has tumor suppressor properties.

MACROD2 overexpression mediates estrogen independent growth and tamoxifen resistance in breast cancers.

Tamoxifen is effective for treating estrogen receptor-alpha (ER) positive breast cancers. However, few molecular mediators of tamoxifen resistance have been elucidated. Here we describe a previously unidentified gene, MACROD2 that confers tamoxifen resistance and estrogen independent growth. We found MACROD2 is amplified and overexpressed in metastatic tamoxifen-resistant tumors. Transgene overexpression of MACROD2 in breast cancer cell lines results in tamoxifen resistance, whereas RNAi-mediated gene knock down reverses this phenotype. MACROD2 overexpression also leads to estrogen independent growth in xenograft assays. Mechanistically, MACROD2 increases p300 binding to estrogen response elements in a subset of ER regulated genes. Primary breast cancers and matched metastases demonstrate MACROD2 expression can change with disease evolution, and increased expression and amplification of MACROD2 in primary tumors is associated with worse overall survival. These studies establish MACROD2 as a key mediator of estrogen independent growth and tamoxifen resistance, as well as a potential novel target for diagnostics and therapy.

Single copies of mutant KRAS and mutant PIK3CA cooperate in immortalized human epithelial cells to induce tumor formation.

The selective pressures leading to cancers with mutations in both KRAS and PIK3CA are unclear. Here, we show that somatic cell knockin of both KRAS G12V and oncogenic PIK3CA mutations in human breast epithelial cells results in cooperative activation of the phosphoinositide 3-kinase (PI3K) and mitogen-activated protein kinase (MAPK) pathways in vitro, and leads to tumor formation in immunocompromised mice. Xenografts from double-knockin cells retain single copies of mutant KRAS and PIK3CA, suggesting that tumor formation does not require increased copy number of either oncogene, and these results were also observed in human colorectal cancer specimens. Mechanistically, the cooperativity between mutant KRAS and PIK3CA is mediated in part by Ras/p110α binding, as inactivating point mutations within the Ras-binding domain of PIK3CA significantly abates pathway signaling. In addition, Pdk1 activation of the downstream effector p90RSK is also increased by the combined presence of mutant KRAS and PIK3CA. These results provide new insights into mutant KRAS function and its role in carcinogenesis.

Androgen receptor as a targeted therapy for breast cancer.

Breast cancer occurs at a high frequency in women and, given this fact, a primary focus of breast cancer research has been the study of estrogen receptor α (ER) signaling. However, androgens are known to play a role in normal breast physiology and therefore androgen receptor (AR) signaling is becoming increasingly recognized as an important contributor towards breast carcinogenesis. Moreover, the high frequency of AR expression in breast cancer makes it an attractive therapeutic target, but the ability to exploit AR for therapy has been difficult. Here we review the historical use of androgen/anti-androgen therapies in breast cancer, the challenges of accurately modeling nuclear hormone receptor signaling in vitro, and the presence and prognostic significance of AR in breast cancer.

Omics and therapy - a basis for precision medicine.

A founding premise of the human genome project was that knowledge of the spectrum of abnormalities that comprise cancers and other human diseases would lead to improved disease management by identifying molecular abnormalities that could guide disease detection and diagnosis, suggest new therapeutic strategies and be developed as markers to predict response to therapy. This project led to elucidation of a reference normal human genome sequence and normal polymorphisms therein against which sequences from diseased tissues can be compared to enable identification of causal abnormalities. It also stimulated development of an array of computational tools for genomic analysis and catalyzed public and private sector development of revolutionary tools for genome analysis that transformed analysis of whole genomes from an enterprise that required international teams and hundreds of millions of dollars to a process that can be carried out in core facilities for only a few thousand dollars per sample. Indeed, the $1000 genome is nearly upon us. Applications of these technologies to human cancers in international cancer genome projects are now revealing the spectra of abnormalities that comprise thousands of individual cancers. Analyses of these data are leading to the promised improvements in disease management. We review several aspects of cancer genomics with emphasis on aspects that are relevant to improving cancer therapy.

The growth response to androgen receptor signaling in ERα-negative human breast cells is dependent on p21 and mediated by MAPK activation.

INTRODUCTION: Although a high frequency of androgen receptor (AR) expression in human breast cancers has been described, exploiting this knowledge for therapy has been challenging. This is in part because androgens can either inhibit or stimulate cell proliferation in pre-clinical models of breast cancer. In addition, many breast cancers co-express other steroid hormone receptors that can affect AR signaling, further obfuscating the effects of androgens on breast cancer cells. METHODS: To create better-defined models of AR signaling in human breast epithelial cells, we took estrogen receptor (ER)-α-negative and progesterone receptor (PR)-negative human breast epithelial cell lines, both cancerous and non-cancerous, and engineered them to express AR, thus allowing the unambiguous study of AR signaling. We cloned a full-length cDNA of human AR, and expressed this transgene in MCF-10A non-tumorigenic human breast epithelial cells and MDA-MB-231 human breast-cancer cells. We characterized the responses to AR ligand binding using various assays, and used isogenic MCF-10A p21 knock-out cell lines expressing AR to demonstrate the requirement for p21 in mediating the proliferative responses to AR signaling in human breast epithelial cells. RESULTS: We found that hyperactivation of the mitogen-activated protein kinase (MAPK) pathway from both AR and epidermal growth factor receptor (EGFR) signaling resulted in a growth-inhibitory response, whereas MAPK signaling from either AR or EGFR activation resulted in cellular proliferation. Additionally, p21 gene knock-out studies confirmed that AR signaling/activation of the MAPK pathway is dependent on p21. CONCLUSIONS: These studies present a new model for the analysis of AR signaling in human breast epithelial cells lacking ERα/PR expression, providing an experimental system without the potential confounding effects of ERα/PR crosstalk. Using this system, we provide a mechanistic explanation for previous observations ascribing a dual role for AR signaling in human breast cancer cells. As previous reports have shown that approximately 40% of breast cancers can lack p21 expression, our data also identify potential new caveats for exploiting AR as a target for breast cancer therapy.

Mutation of a single allele of the cancer susceptibility gene BRCA1 leads to genomic instability in human breast epithelial cells.

Biallelic inactivation of cancer susceptibility gene BRCA1 leads to breast and ovarian carcinogenesis. Paradoxically, BRCA1 deficiency in mice results in early embryonic lethality, and similarly, lack of BRCA1 in human cells is thought to result in cellular lethality in view of BRCA1's essential function. To survive homozygous BRCA1 inactivation during tumorigenesis, precancerous cells must accumulate additional genetic alterations, such as p53 mutations, but this requirement for an extra genetic "hit" contradicts the two-hit theory for the accelerated carcinogenesis associated with familial cancer syndromes. Here, we show that heterozygous BRCA1 inactivation results in genomic instability in nontumorigenic human breast epithelial cells in vitro and in vivo. Using somatic cell gene targeting, we demonstrated that a heterozygous BRCA1 185delAG mutation confers impaired homology-mediated DNA repair and hypersensitivity to genotoxic stress. Heterozygous mutant BRCA1 cell clones also showed a higher degree of gene copy number loss and loss of heterozygosity in SNP array analyses. In BRCA1 heterozygous clones and nontumorigenic breast epithelial tissues from BRCA mutation carriers, FISH revealed elevated genomic instability when compared with their respective controls. Thus, BRCA1 haploinsufficiency may accelerate hereditary breast carcinogenesis by facilitating additional genetic alterations.

PIK3CA mutations and EGFR overexpression predict for lithium sensitivity in human breast epithelial cells.

A high frequency of somatic mutations has been found in breast cancers within the gene encoding the catalytic p110α subunit of PI3K, PIK3CA. Using isogenic human breast epithelial cells, we have previously demonstrated that oncogenic PIK3CA "hotspot" mutations predict for response to the toxic effects of lithium. However, other somatic genetic alterations occur within this pathway in breast cancers, and it is possible that these changes may also predict for lithium sensitivity. We overexpressed the epidermal growth factor receptor (EGFR) into the non-tumorigenic human breast epithelial cell line MCF-10A, and compared these cells to isogenic cell lines previously created via somatic cell gene targeting to model Pten loss, PIK3CA mutations, and the invariant AKT1 mutation, E17K. EGFR overexpressing clones were capable of cellular proliferation in the absence of EGF and were sensitive to lithium similar to the results previously seen with cells harboring PIK3CA mutations. In contrast, AKT1 E17K cells and PTEN -/- cells displayed resistance or partial sensitivity to lithium, respectively. Western blot analysis demonstrated that lithium sensitivity correlated with significant decreases in both PI3K and MAPK signaling that were observed only in EGFR overexpressing and mutant PIK3CA cell lines. These studies demonstrate that EGFR overexpression and PIK3CA mutations are predictors of response to lithium, whereas Pten loss and AKT1 E17K mutations do not predict for lithium sensitivity. Our findings may have important implications for the use of these genetic lesions in breast cancer patients as predictive markers of response to emerging PI3K pathway inhibitors.

Knockin of mutant PIK3CA activates multiple oncogenic pathways.

The phosphatidylinositol 3-kinase subunit PIK3CA is frequently mutated in human cancers. Here we used gene targeting to "knock in" PIK3CA mutations into human breast epithelial cells to identify new therapeutic targets associated with oncogenic PIK3CA. Mutant PIK3CA knockin cells were capable of epidermal growth factor and mTOR-independent cell proliferation that was associated with AKT, ERK, and GSK3beta phosphorylation. Paradoxically, the GSK3beta inhibitors lithium chloride and SB216763 selectively decreased the proliferation of human breast and colorectal cancer cell lines with oncogenic PIK3CA mutations and led to a decrease in the GSK3beta target gene CYCLIN D1. Oral treatment with lithium preferentially inhibited the growth of nude mouse xenografts of HCT-116 colon cancer cells with mutant PIK3CA compared with isogenic HCT-116 knockout cells containing only wild-type PIK3CA. Our findings suggest GSK3beta is an important effector of mutant PIK3CA, and that lithium, an FDA-approved therapy for bipolar disorders, has selective antineoplastic properties against cancers that harbor these mutations.

The multiple myeloma associated MMSET gene contributes to cellular adhesion, clonogenic growth, and tumorigenicity.

Multiple myeloma (MM) is an incurable hematologic malignancy characterized by recurrent chromosomal translocations. Patients with t(4;14)(p16;q32) are the worst prognostic subgroup in MM, although the basis for this poor prognosis is unknown. The t(4;14) is unusual in that it involves 2 potential target genes: fibroblast growth factor receptor 3 (FGFR3) and multiple myeloma SET domain (MMSET). MMSET is universally overexpressed in t(4;14) MM, whereas FGFR3 expression is lost in one-third of cases. Nonetheless, the role of MMSET in t(4;14) MM has remained unclear. Here we demonstrate a role for MMSET in t(4;14) MM cells. Down-regulation of MMSET expression in MM cell lines by RNA interference and by selective disruption of the translocated MMSET allele using gene targeting dramatically reduced colony formation in methylcellulose but had only modest effects in liquid culture. In addition, MMSET knockdown led to cell-cycle arrest of adherent MM cells and reduced the ability of MM cells to adhere to extracellular matrix. Finally, MMSET knockdown and knockout reduced tumor formation by MM xenografts. These results provide the first direct evidence that MMSET plays a significant role in t(4;14) MM and suggest that therapies targeting this gene could impact this particular subset of poor-prognosis patients.

Tamoxifen-stimulated growth of breast cancer due to p21 loss.

Tamoxifen is widely used for the treatment of hormonally responsive breast cancers. However, some resistant breast cancers develop a growth proliferative response to this drug, as evidenced by tumor regression upon its withdrawal. To elucidate the molecular mediators of this paradox, tissue samples from a patient with tamoxifen-stimulated breast cancer were analyzed. These studies revealed that loss of the cyclin-dependent kinase inhibitor p21 was associated with a tamoxifen growth-inducing phenotype. Immortalized human breast epithelial cells with somatic deletion of the p21 gene were then generated and displayed a growth proliferative response to tamoxifen, whereas p21 wild-type cells demonstrated growth inhibition upon tamoxifen exposure. Mutational and biochemical analyses revealed that loss of p21's cyclin-dependent kinase inhibitory property results in hyperphosphorylation of estrogen receptor-alpha, with subsequent increased gene expression of estrogen receptor-regulated genes. These data reveal a previously uncharacterized molecular mechanism of tamoxifen resistance and have potential clinical implications for the management of tamoxifen-resistant breast cancers.

A PCR-based high-throughput screen with multiround sample pooling: application to somatic cell gene targeting.

Here, we describe a method of systematic PCR screening with multiround sample pooling for the isolation of rare PCR-positive samples. As an example, we have applied this protocol to the recovery of gene-targeted clones in human somatic cells comprising only 0.02-0.17% of cells transduced with targeting vectors. Initially, cells infected with targeting vectors are seeded and grown in fourteen 96-well tissue culture plates. Samples are then collected from these plates and subjected to two rounds of pooling to yield twelve 'superpools' used for an initial PCR. After identifying PCR-positive samples, de-pooling is carried out with successive rounds of PCR screening, using samples of decreasing complexity. Single-cell cloning is subsequently performed to isolate gene-targeted clones. The entire protocol can be completed in 4-8 weeks depending on the proliferative capacity of the cell line.

Knock-in of mutant K-ras in nontumorigenic human epithelial cells as a new model for studying K-ras mediated transformation.

The oncogenic function of mutant ras in mammalian cells has been extensively investigated using multiple human and animal models. These systems include overexpression of exogenous mutant ras transgenes, conditionally expressed knock-in mouse models, and somatic cell knockout of mutant and wild-type ras genes in human cancer cell lines. However, phenotypic discrepancies between knock-in mice and transgenic mutant ras overexpression prompted us to evaluate the consequences of targeted knock-in of an oncogenic K-ras mutation in the nontumorigenic human breast epithelial cell line MCF-10A and hTERT-immortalized human mammary epithelial cells. Our results show several significant differences between mutant K-ras knock-in cells versus their transgene counterparts, including limited phosphorylation of the downstream molecules extracellular signal-regulated kinase and AKT, minor proliferative capacity in the absence of an exogenous growth factor, and the inability to form colonies in semisolid medium. Analysis of 16 cancer cell lines carrying mutant K-ras genes indicated that 50% of cancer cells harbor nonoverexpressed heterozygous K-ras mutations similar to the expression seen in our knock-in cell lines. Thus, this system serves as a new model for elucidating the oncogenic contribution of mutant K-ras as expressed in a large fraction of human cancer cells.