The lab’s research works to identify novel drug targets of triple-negative breast cancer via integrative analysis of cellular phenotype assays and multi-omics molecular characterization. 

 p53 mutations and triple-negative breast cancer 
The main goal of this ongoing project is to enable the use of molecularly targeted therapies in triple-negative breast cancer (TNBC) as either adjuvants or alternatives, in order to improve clinical outcomes and reduce the toxicities associated with current cytotoxic regimens. 

TNBC, a virulent subset of the triple negative breast cancers, is the most aggressive and lethal subtype. It disproportionately affects younger populations and is often missed by mammography. Management of TNBC is particularly challenging, as most TNBC tumors lack inhibitor-targetable proteins, limiting therapeutic options to radio-chemotherapies with their consequent toxicities. Furthermore, TNBC is very heterogeneous with a wide variation in responses to anti-cancer therapies. 

Clearly, different molecular pathways dominate in different patients’ tumors, and we need a better mechanistic understanding of which combinations of mutated genes and pathways drive TNBC, especially toward its more aggressive behaviors. Mutation in TP53 (or p53) is common to TNBC, occurring in >80% of these cancers (Fig. 1), and there are strong arguments suggesting that p53 behaves as both a tumor suppressor and oncogene in these cancers. In addition to the common loss of tumor suppressing function, more than 130 different types of mostly non-synonymous point mutations may lead to neomorphic gain-of-function activities of the p53, which likely contribute to phenotypic heterogeneity of breast cancer. 

Fig. 1. Integrated analysis of somatic mutations and gene expression profeiles of breast cancer showed prevalent p53 mutations in triple-negative subtype. *Data source:TCGA

However, despite the evidence for p53 neomorphic activity contributing to both oncogenesis and heterogeneity, different p53 mutants have rarely been compared and little is known about how they differentially impact cancer phenotypes.

Characterization of cellular and molecular impacts of different p53 mutants
Jin Park’s lab generated MCF10A cell lines  that stably express each of the 10 most frequently occurring TP53 missense mutations (Fig. 2) in breast cancer and developed functional assays to examine their effects on the hallmark phenotypes of cancer, including serum-free survival, resistance to apoptosis, resistance to anoikis, migration, invasion, and the loss of polarity. The results showed that the mutants R248W, R273C, and Y220C were the most aggressive and showed increased cell migration and invasion, and anoikis/apoptosis resistance, while the least aggressive mutants were G245S and Y234C (Fig. 3). 

Fig. 2. Somatic mutations profile of TP53 in breast cancer (cBIO Portal) and 10 most prevalent p53 mutations in breast cancer of our research focus (shown in red). 

Fig. 3. Phenotypic heterogeneity of mammary epithelial cells expressing 10 different p53 mutant proteins. As an example, morphological changes (A) and image-based quantification of cell polarity (B) in 3D mammosphere are shown. Phenotypic measurements were normalized to those of the parental MCF10A cells, and the log2 ratios are shown as a heat map (C).

In addition, the lab also performed RNA-Seq and integrated pathway analysis on these mutant p53-expressing cell lines to understand better the neo-morphic activities of different missense mutations. Researchers applied PLS (partial least squares) regression, a multivariate statistical regression method, that builds robust linear models on continuous response variables and particularly suitable for high dimensional data with a small sample size. 

With PLS, the lab can build models on individual pathways to test whether the expression levels of pathway genes can explain the continuous trend of phenotypes by measuring the coefficient of correlation of the models. When the regression results were projected onto the PLS components, for the focal adhesion and adherens junction pathways, the most invasive p53 mutants (R273C, Y220C and R248W) were clustered together while the least invasive mutants H179R and Y234C clustered on the other end (Fig. 4A). To highlight the genes that contribute most to the phenotypic differences, researchers overlaid the pathway diagram with the correlation between the expression levels of each gene and the phenotypic scores (Fig. 4B) and summarized the PLS-based analysis results as a heat map (Fig. 4C).

Fig. 4. Phenotypic heterogeneity of mammary expressing 10 different p53 mutant proteins. As an example, mophological changes (A) and image-based quantification of cell polarity (B) in 3D mammosphere are shown. Phenotypic measurements were normalized to those of the parental MCF10A cells, and the log2 ratios are shown as a heat map (C). 

As p53 is a transcription factor that regulates hundreds of genes upon DNA damage and other stress responses, mutations in p53 proteins likely affect the DNA binding properties, such as affinity and motif preference, and consequently expression of potentially different sets of target genes. Thus, to profile transcriptional targets of mutant p53 that contribute to invasion, the lab performed ChIP-Seq analyses on 2 invasive (R273C and Y220C) and 2 noninvasive (R273H and Y234C) cell lines, along with the WT-OE cells expressing wild type p53 protein as a reference and the parental MCF10A cells as a negative control. 

The ChIP-seq results showed that, in general, more peaks were identified in p53 WT-OE cells compared to the p53 mutant expressing cell lines, which had similar numbers of peaks to the negative control MCF10A cells, suggesting significantly reduced DNA binding of the p53 mutants. Of all the p53 mutant cell lines, only R273H showed an increased DNA binding. Only 6.7% and 4.7% peaks for R273C and R273H, respectively, were identified within promoter regions, which were much lower than 17.3% in Y220C, 21.9% in Y234C, and 18.3% promoter peaks in p53 WT-OE cells. Thus, mutations at R273 to C and H apparently separate general DNA binding from site-specific recognition within promoters, especially R273H which binds DNA but not specifically.

Principal investigators: Josh LaBaer and Jin Park

Funding: Breast Cancer Research Foundation Investigator Grant (PI: Josh LaBaer)