Discussion
FCCTX comprises a group of HNPCC families with a higher risk of developing CRC and other associated cancers, but whose genetic basis is still unknown. Some of the genes that have been associated with FCCTX in recent studies include RPS20 (Nieminen et al., 2014),BRCA2 (Garre et al., 2015) and other fanconi anemia genes such asFAN1 or BRIP1 (Esteban-Jurado et al., 2016; Seguí et al., 2015; Smith et al., 2013), BMPR1A (Evans et al., 2018),SEMA4 (Schulz et al., 2014), OGG1 (Garre et al., 2011) andSETD6 (Martín-Morales et al., 2017). However most cases remain unexplained (Valle, 2017; Zetner & Bisgaard, 2017). With the aim of identifying new genes involved in the cancer predisposition of FCCTX, the whole exome was sequenced in 2 or 3 members of 13 FCCTX families. A thorough filtering and prioritization of the identified variants allowed a final selection of candidate variants for each family.
Here we describe the most promising candidate variant for family cc765, a novel frameshift mutation in the PTPRT gene: c.4090dup p.(Asp1364GlyfsTer24). PTPRT encodes a tyrosine phosphatase that has been proven to behave as a tumor suppressor that is involved in relevant pathways, such as the PI3K-Akt and the IL6-JAK-STAT3 pathway, through which it regulates the expression of genes involved in cell survival, apoptosis, cell proliferation, growth and migration (X. Zhang et al., 2007; Y. Zhao et al., 2017). In fact, PTPRT somatic inactivating alterations are frequently found in many tumors, including CRC, and have been reported to act as driver mutations that promote tumor development and progression (L.-E. Wang et al., 2013). Based on the relevance of the gene and the effect of the mutation, this variant was selected for further characterization.
PTPRT D1364Gfs*24 showed a compatible cosegregation with the disease within the family, given that it was carried by 3 members affected with different cancers while it was not present in two elderly cancer-free relatives aged 91 and 85. However, it was also absent in a relative diagnosed with CRC at the age of 85, but taking into account the elevated age of onset in this patient and the prevalence of CRC in the general population this could be perfectly explained as a phenocopy. The cancers of which the PTPRT mutation carriers had been diagnosed included two from the HNPCC spectrum (CRC and endometrial cancer) and a breast cancer, which although does not belong to the spectrum occurs with relative frequency within CRC families.
On the other hand, there was not a clear LOH of the wild-type allele in any of the tumors tested. Nonetheless, PTPRT promoter has been recently reported to be frequently hypermethylated in sporadic CRC and other tumors (Laczmanska et al., 2013; Peyser et al., 2016). Indeed, the promoter region of PTPRT was found to be hypermethylated in the two tumors of the carriers that were tested (breast and colon tumors), which could be a different mechanism of inactivation of the wild-type allele in the tumors from the carriers. In fact, an allele-specific expression assay showed that the expression of PTPRT ´s wild-type allele was significantly reduced in the colorectal tumor when compared to healthy colon from the same individual, while no decrease was observed in the expression of the mutant allele. This supports that the epigenetic silencing mainly affects the wild-type allele and could be considered a second hit involved in the inactivation of this tumor suppressor gene.
Regarding the effects of the variant on the protein, D1364Gfs*24 is a frameshift mutation that affects PTPRT’s second catalytic domain, known as D2. PTPRT D1364Gfs*24 results in the loss of the last 97 amino acids of the protein, including 36.2% of the D2 domain and a considerable amount of essential D2 residues. Actually, the majority (69.2%) of residues directly surrounding the substrate’s phospho-tyrosine are lost with this mutation (Lui et al., 2014). Although the first phosphatase domain (D1) is the one known to be responsible for the phosphatase activity of the protein per se , D2 is responsible for the regulation of this activity and has been proven to be essential for PTPRT’s activity (Y. Zhao et al., 2017). As a matter of fact, Wang et al. reported that just a missense mutation affecting D2’s residue 1368, which is lost in our mutant, was enough to decrease the enzyme’s activity by half (Z. Wang et al., 2004). The relevance of this second catalytic domain was also pointed out by Zhang et al., who showed how the deletion of the D2 domain had significant effects on the levels of phosphorylated PTPRT substrates and the expression of its downstream target genes (X. Zhang et al., 2007). Therefore, it can be predicted that this mutation will result in significant consequences for the activity of this phosphatase.
PTPRT is known to dephosphorylate two main target proteins, STAT3 and paxillin (X. Zhang et al., 2007; Y. Zhao et al., 2017, 2010), both of which are well-known oncogenes (Bromberg et al., 1999). These two proteins are activated upon phosphorylation by different protein tyrosine kinases, so the PTPRT-mediated removal of the phosphate group results in their inactivation. This results, in turn, in the inhibition of their downstream pathways, through a decreased phosphorylation of paxillin’s target proteins (Y. Zhao et al., 2017, 2010) and a decreased expression of STAT3’s target genes (X. Zhang et al., 2007).BCL-XL and SOCS3 are two of those STAT3’s target genes, which have been proven to show increased expression upon PTPRT depletion, and even upon deletion of the D2 domain (X. Zhang et al., 2007). Consistent with the hypothesis that this germline PTPRTvariant may be involved in the cancers of this FCCTX family, the tumors from the two carriers tested presented a significantly increased expression of BCL-XL , which is an oncogenic driver in CRC (Scherr et al., 2016). This was observed in both the colon and breast tumors when compared to healthy tissue controls.
In contrast, SOCS3 expression was decreased in both tumors. Nonetheless, SOCS proteins are negative feedback regulators of the JAK-STAT signaling pathway (Inagaki-Ohara, Kondo, Ito, & Yoshimura, 2013; Jiang et al., 2017), and it has been proven that SOCS3 is usually downregulated in CRC, even when IL-6 and STAT3 are upregulated (Chu et al., 2017). This is thought to occur through different mechanisms, such as the hypermethylation of SOCS3 gene promoters, allowing inflammatory cytokines IL-6 to activate STAT3’s signaling pathway while inhibiting the expression of SOCS3 (Chu et al., 2017), with the purpose of inactivating its negative feedback. This negative regulation of SOCS3 expression, mediated by the activation of IL-6/STAT3, leads to imbalance and sustained activation of STAT3 signaling pathway (Chu et al., 2017). The same study by Chu et al. also showed that SOCS3 plays an important role inhibiting tumor development and that reduced SOCS3 ’s expression affects tumorigenesis and CRC progression, promoting growth and metastasis (Chu et al., 2017). For all the things mentioned, both results are compatible with a pathogenic role of this PTPRT variant.
As previously discussed, PTPRT ’s association with cancer has been well studied, but this is the first time that PTPRT is linked to hereditary cancer. Interestingly, a somatic mutation affecting the same codon (COSV62009665) is associated with CRC according to COSMIC (Tate et al., 2019). Even though the role of this protein in tumor development is undeniable, more studies are needed to confirm its involvement as a cancer susceptibility gene. Studying this gene in a larger cohort would help with this task. As a matter of fact, the screening of PTPRTin an independent familial/early-onset CRC cohort identified one additional rare missense germline variant located inside the D1 domain in an early-onset CRC patient. However, the functional effect of this variant has not been determined.
Last but not least, it should be pointed out that even though thisPTPRT mutation had the highest potential for the explanation of the increased cancer risk, another three candidate variants were prioritized for this family: two missense variants in MAP3K6 andABTB1 , and a splice region variant in INVS . MAP3K6 is involved in the regulation of VEGF expression and has also been reported to act as a tumor suppressor (Gaston et al., 2014). In addition, germline mutations in this gene have been associated with familial gastric cancer (Gaston et al., 2014). ABTB1 is a mediator of the PTEN signaling pathway reported to suppress the growth of cancer cells by the inhibition of the cell cycle (Unoki & Nakamura, 2001). Finally, INVS acts as a molecular switch between the different Wnt signaling pathways, inhibiting the canonical Wnt pathway (Simons et al., 2005), and homozygous INVS mutations have been associated with juvenile nephronophthisis (Bellavia et al., 2010). Although PTPRTD1364Gfs*24 is the best candidate variant for this family and the results presented in this report support its causality, we cannot rule out the possibility that these candidate variants – or even other genetic or environmental factors – may be contributing, independently or together, to the increased cancer susceptibility of the family or modifying the effect of this PTPRT mutation.
Taken together, the results here presented point to a probable causal role of the germline variant PTPRT c.4090dup p.(Asp1364GlyfsTer24) in the cancer susceptibility of the carrier family. For that reason, we propose PTPRT as a novel cancer predisposition gene. However, more research is necessary to confirm the causality, penetrance, conferred risk and preferred cancer location. The screening of this gene in additional familial colorectal cancer cohorts – or even in other high-risk families – will help us clarify its role in cancer predisposition. Although PTPRT’s role in cancer initiation and progression has been well stablished, this is the first time that aPTPRT germline variant is linked with cancer susceptibility and hereditary cancer, which highlights the relevance of this work.