Druggable genetic targets in endometrial cancer

Abstract

Background:

FBXW7 is frequently somatically mutated in grade 3 endometrioid endometrial cancers (G3EECs) and serous endometrial cancers (SECs), high-risk cancers associated with poor prognosis. CRISPR-edited cell lines identified the proteomic and phosphoproteomic effects of FBXW7 mutation in 3 high-risk endometrial cancers (ECs), including altered protein levels of L1CAM and TGM2. This result is important because L1CAM and TGM2 are druggable proteins that could represent new therapeutic targets.

Methods:

We used cBioPortal for Cancer Genomics to analyze data in The Cancer Genome Atlas (TCGA). We used the UCSC Xena Browser to analyze gene expression. For differential gene expression analysis, the gene ontology molecular function 2018 version was used. The analysis was focused on determined genes.

Results:

FBXW7 mutations affect gene expression of L1CAM but are unrelated to TGM2 gene expression. L1CAM gene expression is significantly related to survival. Patients with lower L1CAM gene expression have better survival. FBXW7 mutations are unrelated to survival. TGM2 gene expression is unrelated to FBXW7 mutations. TGM2 gene expression is unrelated to survival, all tumor grades or grade 3 alone.

Conclusion:

We agree with Urick et al. that L1CAM may be a promising druggable target in endometrial carcinoma. The lack of relationship of TGM2 expression with FBXW7 mutations and endometrial cancer survival suggests that TGM2 might not be of as much value as a druggable target, compared to L1CAM. However, the fact that a certain alteration is not prognostic for cancer survival does not necessarily mean that the alteration will not be targetable. More data, such as inhibition of each gene by calculating drug targetability, may be required to support this conclusion.

Identification of pathogenic mutations within uterine and endometrial carcinomas may facilitate diagnosis, clinical care and development of novel treatments [1]. For example, L1CAM levels in curettage or plasma samples measured before surgery predict lymph node metastases and provide further information on patient prognosis [2]. Novel plasma biomarkers with good accuracy in diagnosing endometrial carcinoma include TGM2 [3]. In uterine leiomyosarcoma the vast majority of patients (81%) carry at least one mutation in either TP53, RB1, ATRX or PTEN [4].

FBXW7 is frequently somatically mutated in 6%–18% grade 3 endometrioid endometrial cancers (G3EECs) and 15%–29% serous endometrial cancers (SECs), high-risk cancers associated with poor prognosis. Urick et al. used CRISPR-edited cell lines to identify the proteomic and phosphoproteomic effects of FBXW7 mutation in 3 high-risk endometrial cancers (ECs), including altered protein levels of L1CAM and TGM2. They found decreased levels of L1CAM and TGM2 proteins after reversion of endogenous FBXW7 mutations in 2 cell lines derived from grade 2 endometrioid endometrial cancers (G2EECs). This result is important because L1CAM and TGM2 are druggable proteins that could represent new therapeutic targets [5].

To further investigate the relationship of FBXW7, L1CAM and TGM2 to endometrial cancer, we consulted the Cancer Genome Atlas (TCGA).

Methods

TCGA is a project, begun in 2005, to catalog genetic mutations responsible for cancer, employing genome sequencing and bioinformatics. We used cBioPortal for Cancer Genomics to analyze data in TCGA. cBioPortal provides visualization, analysis and download of large-scale cancer genomics data sets [6]. We used the GISTIC (Genomic Identification of Significant Targets in Cancer) module in cBioPortal to identify regions of the genome that are significantly amplified or deleted across a set of samples. Gene expression is quantitated as Fragments Per Kilobase of transcript per Million mapped reads upper quartile (fpkm-uq), which is an RNA-Seq-based expression normalization method [7].

Simple statistics were calculated to identify patterns of mutual exclusivity or co-occurrence. For each pair of query genes (e.g., FBXW7 and L1CAM), an odds ratio (OR) is calculated (Eq. (1)) that indicates the likelihood that the events (e.g., mutations) in the two genes are mutually exclusive or co-occurrent across the selected cases,

$$\text{OR}=\left(\text{A}*\text{D}\right)/\left(\text{B}*\text{C}\right)$$ (1)

where A = number of cases altered in both genes; B = number of cases altered in FBXW7 but not L1CAM; C = number of cases altered in L1CAM but not FBXW7; and D = number of cases altered in neither gene. Each pair was then assigned to one of three categories indicative of a tendency toward mutual exclusivity, of a tendency toward co-occurrence, or of no association. To determine whether the identified relationship is significant for a gene pair, Fisher’s exact test was performed [6].

We used the UCSC Xena Browser to analyze gene expression [8]. (https://xenabrowser.net). UCSC Xena allows exploration of functional genomic data sets for correlations between genomic or phenotypic variables. For differential gene expression analysis (Fig. 6), the gene ontology molecular function 2018 version was used [9,10].

Fig. 6.

Differential gene expression analysis of the GDC TCGA endometrioid endometrial cancer data set demonstrated that phosphorylated proteins exhibited significantly different levels in grade 2 (G2) versus grade 3 (G3) cancers, with significant down-regulation in grade 2.

Results

We analyzed data from 522 patients in the GDC TCGA endometrioid endometrial cancer data set. Mean age was 64 ± 11. Distribution of tumors by grade is shown in Fig. 1. 57% of patients had grade 3 tumors.

Fig. 1.

Distribution of tumors by grade in the GDC TCGA endometrioid endometrial cancer data set. Grade 3 (G3) was most common.

An Oncoprint diagram shows mutations in FBXW7, L1CAM, and TGM2 (Fig. 2). A plot of FBXW7 mutations versus mRNA expression is displayed in Fig. 3. Mutations in FBX7, L1CAM, and TGM2 tended to co-occur (Table 1).

Fig. 2.

Oncoprint diagram showing mutations in FBXW7, L1CAM, TGM2, and TP53.

Fig. 3.

Plot of FBXW7 mutations versus mRNA expression; putative copy number alterations are from GISTIC.

Table 1.

Mutations in FBX7, L1CAM, and TGM2 all tended to co-occur. P values are Bonferroni adjusted. Q value is derived from the Benjamini-Hochberg false discovery rate correction procedure for multiple comparisons.

A	B	Neither	A Not B	B Not A	Both	Log2 Odds Ratio	p-Value	q-Value	Tendency
L1CAM	TGM2	215	14	5	6	>3	<0.001	<0.001	Co-occurrence
FBXW7	L1CAM	191	29	11	9	2.43	0.001	0.002	Co-occurrence
FBXW7	TGM2	196	33	6	5	2.307	0.017	0.017	Co-occurrence

FBXW7 mutations (Fig. 4) affect gene expression of L1CAM but are unrelated to TGM2 gene expression. L1CAM gene expression is significantly related to survival. Patients with lower L1CAM gene expression had better survival (p = 0.00000032, log rank test, Fig. 5). FBXW7 mutations are unrelated to survival (p = 0.11, log rank test). TGM2 gene expression is unrelated to FBXW7 mutations (Fig. 4). TGM2 gene expression is unrelated to survival, all tumor grades (p = 0.5, log rank test) or grade 3 alone (p = 0.4, log rank test). Survival analyses and the reported p values were not adjusted for the tumor grades.

Fig. 4.

FBXW7 mutations (column B) affect gene expression of L1CAM (column C) but are unrelated to TGM2 gene expression (column D). Each row contains data from a single sample. Horizontal bars in Column A indicate number of rows with 50 samples. Row order is determined by sorting the rows by their column values. Column B value is used to sort the rows. In case of a tie, the next columns to the right are used to break the tie. (xenabrowser.net).

Fig. 5.

Survival of endometrial cancer patients is significantly related to L1CAM gene expression. Patients with lower L1CAM gene expression (blue) had better survival (p = 0.00000032, log rank test).

Urick et al. reported that phosphorylated proteins exhibited significantly different levels (p < 0.05) in CRISPR-edited FBXW7-mutant HEC-50B cells in comparison with matched FBXW7 nonmutant parental cells [5]. Differential gene expression analysis of the GDC TCGA endometrioid endometrial cancer data set demonstrated phosphorylated proteins at significantly different levels in grade 2 versus grade 3 cancers, with significant downregulation in grade 2 (Fig. 6).

Discussion

FBXW7 is a critical tumor suppressor. FBXW7 is located on chromosome 4q32, a genomic region deleted in more than 30% of all human cancers, and is frequently mutated [11].

L1CAM is a transmembrane protein member of the L1 protein family, encoded by the L1CAM gene. This protein, 200–220 kDa, is a neuronal cell adhesion molecule with a prominent role in cell migration, adhesion, neurite outgrowth, myelination, and neuronal differentiation.

L1CAM plays a key role in treatment-resistant cancers [12]. L1CAM overexpression in all types of cancer cells has been associated with poor prognosis, tumor progression and metastasis. Blocking L1CAM in mouse models holds promise as a cancer treatment [13]. As noted above, endometrial cancer patients with lower L1CAM gene expression had better survival.

Transglutaminase 2 (TGM2) affects cell growth and survival through the anti-apoptosis signaling pathway. TGM2 has been proposed as a novel marker for colorectal cancer survival and a therapeutic target [14]. TGM2 expression is aberrantly upregulated in multiple cancer cell types, particularly those that resist chemotherapy and radiation therapy and those isolated from metastatic sites. TGM2 may be a factor in inflammation-induced progression of breast cancer [15]. However, we found TGM2 expression to be unrelated to endometrial cancer survival.

One explanation is possible for the discrepancy between the cooccurrence of the 3 mutations (FBXW7, L1CAM and TGM2) shown in Table 1, but the lack of association of TGM2 with the other genes despite the mentioned co-occurrence. According to the Human Protein Atlas, TGM2 is expressed at a relatively low level in uterine cancer, when compared to other cancers such as thyroid cancer, pancreatic cancer, renal cancer, and prostate cancer [16].

Cancer gene alterations usually do not occur at random. Alterations of certain cancer genes tend to co-occur, indicating that they may work in tandem to drive tumor formation and development [6]. This may be the case with the co-occurring alterations in Table 1 of FBXW7, L1CAM, and TGM2. But some alterations may occur as passenger/ non-driver mutations and have no functional consequences. Not all alterations have significance for tumor biology.

The expression of phosphorylated proteins, e.g., MAPK (p-MAPK), is increased in actively proliferating tissues, such as normal proliferative-phase endometrium and endometrial carcinoma. In endometrial carcinoma, the MAPK signaling pathway appears to play a major role in tumor cell proliferation and survival [17]. Urick et al. noted the effect of phosphorylated proteins in their cell lines [5], and we have now documented the effect of phosphorylation in the GDC TCGA endometrioid endometrial cancer data set (Fig. 6).

Integrating clinic pathologic and molecular factors improves the risk assessment of patients with early-stage endometrial carcinoma. A weakness in our study is that we were not able to classify the TCGA endometrioid endometrial carcinomas, included in our analysis, as per the endometrial cancer molecular classifier that is currently used in clinical practice [18]. Given the difference in the biology within the endometrioid endometrial cancers, it would be worthwhile to have additional data on TP53 mutation, MMR genes, POLE gene alterations, and whether the genes that we analyzed in the current study were preferentially altered in a certain subgroup.

We agree with Urick et al. that L1CAM may be a promising druggable target in endometrial carcinoma. The lack of relationship of TGM2 expression with FBXW7 mutations and endometrial cancer survival suggests that TGM2 might not be of as much value as a druggable target, compared to L1CAM [19]. However, the fact that a certain alteration is not prognostic for cancer survival does not necessarily mean that the alteration will not be targetable. More data, such as inhibition of each gene by calculating drug targetability, may be required to support this conclusion.