SRSF6 expression in colon and colorectal cancer has been studied extensively. Most studies demonstrated that SRSF6 was overexpressed in cancer tissues. For example, Cohen-Eliav et al. reported that SRSF6 mRNA was overexpressed in colon cancer, and its gene was amplified in some colon cancer patients (37%) (Cohen-Eliav et al., 2013). Wan et al. showed that SRSF6 is significantly overexpressed in colorectal cancer patients from the TCGA database and their cohort in transcriptional level. Furthermore, SRSF6 overexpression is significantly associated with poor overall survival (Wan et al., 2019). Park et al. showed that SRSF6 protein was overexpressed in a cohort of colorectal patients by Western blot (Park et al., 2016), which was further confirmed in another study by Western blot (Si et al., 2021). However, Lin et al. showed that SRSF6 protein was downregulated in colorectal cancer tissues in eight patients compared with adjacent normal tissues by Western blot (Lin et al., 2017). More studies in larger cohorts may be required to evaluate SRSF6 expression in colorectal cancer.

SRSF6 is significantly overexpressed in a set of subtypes of skin cancer, including basal-cell carcinoma, squamous-cell carcinoma, and malignant melanoma (Jensen et al., 2014). SRSF6 gene is also reported to be amplified in some lung cancer patients (12%) (Cohen-Eliav et al., 2013), and overexpressed (Cohen-Eliav et al., 2013; Kim et al., 2016).

In ovarian cancer, Mayer et al. showed that cancer tissues expressed significantly higher SRSF6 than normal tissues in a small cohort (Mayer et al., 2015). However, another study showed that the expression level of SRSF6 is not significantly superior to normal controls in a small cohort of patients with ovarian cancer. Interestingly, patients with metastasis showed significant higher SRSF6 expression (Iborra et al., 2013). More studies in larger cohorts are required to determine the association between SRSF6 expression and ovarian cancer. In addition, Li et al. showed that pancreatic cancer tissues expressed less SRSF6 than adjacent normal tissues, suggesting that SRSF6 may be not overexpressed in some cancers (Li et al., 2020).

So far, only a few studies reported the relationship between SRSF6 expression and cancer patient prognosis. Besides colorectal cancer, SRSF6 expression is also associated with poor prognosis in T-cell acute lymphoblastic leukemia (T-ALL) (Zhou et al., 2020b). More pieces of evidence are required to determine the association between SRSF6 expression and cancer patient prognoses.

In summary, SRSF6 may be overexpressed in most cancers. However, the association between SRSF6 expression and disease prognosis remains largely unclear. Therefore, to understand the value of SRSF6 expression in cancer diagnosis and prognosis, more studies are required to investigate the expression and clinical significance of SRSF6 in cancers.

The overexpression of some SR proteins, such as SRSF1 (Karni et al., 2007; Anczukow et al., 2012) or SRSF3 (Jia et al., 2010) can induce cell transformation. It is also true for SRSF6. Similar to SRSF1 and SRSF3, SRSF6 overexpression also transforms mouse embryonic fibroblast (Jensen et al., 2014). In non-transformed mouse or human lung epithelial cells, SRSF6 overexpression enables these cells to form colonies in soft agar and tumors in nude mice (Cohen-Eliav et al., 2013). In addition, in non-transformed mammary epithelial cells, SRSF6 overexpression induced significantly larger and dysmorphic acini morphology and increased the amount of proliferating acini in a short period of time, thereby indicating the potent transformation capability of SRSF6 (Park et al., 2019). Notably, most breast cancers originate from the mammary acini, which is the basic anatomical unit of the mammary gland. SRSF6 regulates the alternative splicing of genes associated with cell-cycle regulation, apoptosis, and cell adhesion, specifically increases exon10 inclusion of ARAP2, which is essential for cellular focal adhesion formation (Park et al., 2019).

Although SRSF6 overexpression can transform cells, it cannot be ruled out that the depletion of SRSF6 may also be involved in tumorigenesis in some tissues. Splicing factor SRSF3 is an example of this. SRSF3 overexpression in mouse embryonic fibroblast induced significantly tumor formation (Jia et al., 2010). However, specific knockout of SRSF3 in mouse hepatocytes impaired hepatocyte maturation and metabolism and induced spontaneous hepatocellular carcinoma (Sen et al., 2015). SRSF3 is required for protecting mice from tetrachloride-induced fibrosis and carcinogenesis in liver (Sen et al., 2015). Later, in human hepatocellular carcinoma cells, SRSF3 was found to be dephosphorylated and inactivated (Chen et al., 2021a). Therefore, the precise function of SRSF6 in a specific type of tissue or cell should be evaluated individually. In general, SRSF6 overexpression may be positively associated with cancers.

SRSF6 can promote cancer cell metastasis. SRSF6 overexpression increased migration and invasion in breast cancer cells (Park et al., 2019) and induced epithelial–mesenchymal transition (EMT) in colorectal cancer cells (Kong et al., 2016). Knockdown or using specific inhibitor of SRSF6 significantly inhibited colorectal cancer cell migration and invasion in vitro and metastasis in vivo (Wan et al., 2019). In principle, SRSF6 silence increased exon 23 inclusion of ZO-1 gene, which is an important cell adhesion molecule. ZO-1 isoform with exon 23, not isoform without exon 23, showed significant inhibitory role in cell motility (Wan et al., 2019). In skin cancer, SRSF6 promoted exon 10-15 inclusion of Tnc (extracellular-matrix protein tenascin C) gene, which can promote cell migration through its isoform with 10–15 exons (Jensen et al., 2014).

Immunosuppression helps cancer cells to escape from the immune system and progress. PBMCs in breast cancer patients contains several immune suppressive cells, such as myeloid-derived suppressor cells (MDSCs). Interestingly, peripheral mononuclear cells (PBMCs) from patients with metastasis showed dramatic increase in SRSF6 RNA than those without metastasis (Moradpoor et al., 2020). So far, little is known about the function of SRSF6 in cancer immunosuppression.

SRSF6 regulates the expression of a set of immune-associated genes. For example, IL-1b, the first altered signal was induced by SRSF6 (Jensen et al., 2014) upon wounding (Morasso and Tomic-Canic, 2005). CD44, a cell surface adhesion molecule, mediates T-cell homing (de la Hera et al., 1989), as well as tumor metastasis (Chen et al., 2018). The knockdown of SRSF6 increased CD44 alternative exon v7 and v10 inclusion in U2OS cells (Filippov et al., 2007). However, SRSF6 overexpression decreased exon v6 inclusion in breast cancer cells (Loh et al., 2016). CD45 is a transmembrane tyrosine phosphatase expressed by all leucocytes (Charbonneau et al., 1988) and required for TCR-mediated T cell activation (Pingel and Thomas, 1989). During T cell activation, the expression level of SRSF6 significantly increases, thereby promoting CD45 exon 4 inclusion (Lemaire et al., 1999). SRSF6 may regulate cancer-associated immunosuppression via these genes. In addition, Lu et al. found that the genome-wide modification of pre-mRNA alternative splicing induced neoantigens and elicited anti-tumor immunity (Lu et al., 2021), which showed a new way to enhance cancer immunotherapy. The suppression of SRSF6 expression or function may also induce neoantigens for immunotherapy because SRSF6 controls a number of target genes.

SRSF6 overexpression can increase the resistance of immortal mouse lung epithelial cells to cis-platinum treatment (Cohen-Eliav et al., 2013). In gastric cancer, SRSF6 is required for the resistance of gastric cancer cells to oxaliplatin and 5-FU. In principle, SRSF6 promotes phosphatidylinositol-binding clathrin assembly protein (PICALM) exon 14 inclusion to produce a full-length PICALM protein, which is required for the autophagy-induced resistance of gastric cancer cells to oxaliplatin and 5-FU (Zhang et al., 2021). By contrast, PICALM protein without exon 14 sensitizes cancer cells to chemotherapy (Zhang et al., 2021). PICALM participates autophagic precursor formation (Moreau et al., 2014) and can form a PICALM-MLLT10 fusion gene in leukemia, which is often associated with poor outcome (Savage et al., 2010).

SRSF6 controls the alternative splicing of many target genes. Sometimes, SRSF6 may play an anti-tumorigenesis role. Angiogenesis is an example of this. VEGF is a key regulator of angiogenesis in cancers. The alternative splice acceptor site usage in exon 8 (a terminal exon) produce two isoform families of VEGF. Isoforms that use proximal acceptor site encode pro-angiogenic VEGFxxx proteins. On the contrary, isoforms that use distal acceptor site encode anti-angiogenic VEGFxxxb proteins (Bates et al., 2002). SRSF6 binds to a 35-nucleotide motif in exon 8 and promotes the usage of distal acceptor site and VEGF165b expression (Nowak et al., 2008), which was further confirmed in systemic sclerosis patients (Manetti et al., 2011). Therefore, in terms of angiogenesis, increased SRSF6 may inhibit angiogenesis and have an adverse effect on tumor development.

However, VEGF165b isoform expression actually decreased in colorectal cancer (Diaz et al., 2008), which had SRSF6 overexpression. Other splicing factors may function against SRSF6 to suppress VEGF165b expression in cancers. Nowak et al. showed that oncogene SRSF1 can inhibit VEGF165b expression by relatively suppressing distal acceptor site usage (Nowak et al., 2008).

Cancer is considered a kind of aberrant wound healing process (Dvorak 1986; Sundaram et al., 2018). For example, skin cancer shares similar gene-expression profile with wounded normal skin (Schafer and Werner, 2008). Notably, SRSF6 overexpression dramatically upregulated the expression of a set of genes involved in wound-healing by 13–154 fold, including keratin 6, keratin 16, IL-1b, Cxcl2, and Ccl3 in a transgenic mouse model (Jensen et al., 2014). Normal skin may only upregulate SRSF6 expression for several days after injury in contrast to the continuous SRSF6 overexpression in cancer (Jensen et al., 2014). This study revealed that SRSF6 exerted an important role in wound healing process, as well as in cancer when continuously overexpressed.

An important regulatory mechanism of SRSF6 expression is the inclusion of its alternative exon 3, which is conserved across multiple species (Lareau and Brenner, 2015). Transcripts including this poison exon, which contains a pre-mature stop codon, are mostly degraded by nonsense-mediated decay (NMD) and may also encode a truncated SRSF6 protein. The inclusion of exon 3 is positively regulated by Nova1 (Lin et al., 2016), SRSF4 and SRSF5 (Leclair et al., 2020), and leads to reduce full-length SRSF6 protein level. SRSF6 can also promote the inclusion of exon 3 and autoregulate its own expression, which is also a conserved autoregulatory mechanism in SR family (Leclair et al., 2020). This mechanism exists not only in carcinoma cells, but also in leukemia cells (Leclair et al., 2020; Zhou et al., 2020b). Cancer cells prefer to impair autoregulatory mechanism and produce transcripts without exon 3 and then increase SRSF6 protein level (Zhou et al., 2020b). This phenomenon also exists in other SR proteins. For example, SRSF3 poison exon 4 inclusion is significantly downregulated in oral cancer (Guo et al., 2015).

Competing endogenous RNA (ceRNA) regulation is another important regulatory mechanism of SRSF6 expression. CeRNAs can attenuate the inhibition of SRSF6 expression mediated by microRNAs. SRSF6 is targeted by miR-660. circFBXL5 functions as a ceRNA to sponge miR-660 and upregulate SRSF6 expression in breast cancer (Zhou et al., 2020a). Another example is that lncRNA ZNF561-AS1 can sponge miR-26a-3p and miR-128-5p to upregulate SRSF6 expression in colorectal cancer (Si et al., 2021).

Ubiquitination is an important regulatory mechanism of protein stability mediated by the ubiquitin-proteasome system (Hershko, 1983). Several splicing factors have been found to be ubiquitinated and then degraded by proteasome (Du et al., 2021). LncRNA can directly regulate SRSF6 protein stability or function. For example, lncRNA CRNDE binds to SRSF6 protein and causes its ubiquitination and degradation by proteasome in gastric cancer cell (Zhang et al., 2021). In T-ALL, SRSF6 expression was enhanced by the increased USP7, an ubiquitinase, which can deubiquitinate and stabilize SRSF6 protein (Zhou et al., 2020b).

The oncogenic function of SRSF6 may be neutralized by its antagonistic protein. For example, by using a genetic screen of randomly overexpressing genes, Fernando et al. discovered that the brat (brain tumor protein) gene of Drosophila, a tumor suppressor and post-transcriptional repressor of myc, can overcome the effects caused by SRSF6 overexpression (Fernando et al., 2015). The human homolog of brat is TRIM3 gene, which also suppresses tumorigenesis by ensuring asymmetric cell division of neural cells, attenuating stem-like characteristics of glioblastoma cells, and suppressing c-Myc expression (Chen et al., 2014).

SRSF6 expression is also regulated by a list of key tumor-related genes. For example, DNA damage can induce SRSF6 expression in colorectal cancer cells lacking p53, not in cells with p53 expression, indicating that p53 may downregulate SRSF6 in cancer cells (Filippov et al., 2007). On the other hand, SRSF6 may also regulate alternative splicing of p53 pre-mRNA, which often mis-spliced due to mutations in SRSF6 binding motifs in Li–Fraumeni and Li–Fraumeni-Like syndrome patients, two hereditary cancer predisposition syndromes commonly with somatic mutation in p53 (Kouidou et al., 2009). Estrogen indirectly inhibits SRSF6 expression in breast cancer cells. Estrogen receptor-positive (ER⁺) breast tumors had decreased abundance of SRSF6 compared with ER⁻ tumors (Lal et al., 2013). In colorectal cancer, lncRNA LINC01133 can block SRSF6 function in metastasis by interacting with SRSF6 protein. However, TGF-β can induce LINC01133 downregulation, and then allow SRSF6 to promote tumorigenesis in colorectal cancer (Kong et al., 2016). Pnn is a desmosome associated protein, and is overexpressed and associated with poor prognosis in cancers (Mini et al., 2019). Silence of Pnn significantly reduced SRSF6 expression (Chiu and Ouyang, 2006).

Small molecules or metal ion are applied to inhibit SRSF6 function. Wan et al. found that stable knockdown of SRSF6 significantly decreased the xenograft tumor growth and lung metastasis in nude mice. Moreover, they predicted the 3D structure of SRSF6 protein RRM2 domain by simulation software and virtually screened possible chemicals that were able to bind the binding pockets of RRM2 domain (Wan et al., 2019). They discovered that indacaterol, a β2-agonist for the treatment of chronic obstructive pulmonary disease (COPD) (Yum et al., 2017), could significantly inhibit SRSF6 function, colorectal cancer cell proliferation, and tumor formation in a mouse colorectal model (Wan et al., 2019).These findings implied that RRM2 domains of other SR proteins may be promising targets of developing specific inhibitors because RRM2 domain is not conserved in SR proteins.

The hyperphosphorylation of SR proteins inhibited their splicing activity (Prasad et al., 1999). Hyperphosphorylation of SRSF6 also reduces its splicing activity and stability (Lai et al., 2003). Zinc ion markedly and specifically phosphorylated SRSF6 and induced its hyperphosphorylation and loss of RNA-binding ability (Hara et al., 2013), thereby suggesting that zinc ion may be applied to suppress SRSF6 function as an inhibitor.

Small interfering RNAs (siRNAs) are synthetic double-stranded RNA and can interfere with its target gene expression. Anti-SRSF6 siRNAs can efficiently silence SRSF6 expression and may inhibit cancer cell proliferation and migration. However, the clinical application of siRNA is still facing many obstacles, such as off-target effects and in vivo delivery (Rautela et al., 2021). Therefore, the anti-tumor efficiency of anti-SRSF6 siRNA should be tested and improved in vivo.

Decoy RNA oligonucleotide is another type of synthetic small RNA, which specifically binds to RNA binding proteins and blocks their function by steric hindrance (Denichenko et al., 2019). This strategy has been successfully used to target SR protein SRSF1 and three other splicing factors, namely, PTBP1 and RBFOX1/2. The splicing function of PTBP1 was interfered by decoys. Breast cancer cells treated with PTBP1 decoy showed significantly retarded cell proliferation and reduced soft agar colony formation (Denichenko et al., 2019). SRSF6 recognizes specific motifs in RNA, especially purine-rich motifs (Nagel et al., 1998). For example, SRSF6 binds a consensus motif sequence of UGGAG in ZO-1 exon 23 (Wan et al., 2019), a sequence of UGCAGGA in Tnc exon 12 (Jensen et al., 2014), and a sequence of AGTAGA in HIV-1 pre-mRNA (Erkelenz et al., 2015). Alvelos et al. identified thousands of SRSF6 binding motifs in human pancreatic β-cells by integrating individual-nucleotide resolution UV-cross-linking, immunoprecipitation (iCLIP) and RNA sequencing. Importantly, they found that SRSF6 preferred to bind a purine-rich consensus motif that contains GAA triplets, and more contiguous GAA triplets were associated with stronger binding (Alvelos et al., 2021). These studies paved the way to design efficient decoys to block SRSF6 binding and correct aberrant alternative splicing in cancers.

Another promising approach is to take advantage of SRSF6 autoregulation mechanism by increasing the inclusion of its exon 3 with antisense oligonucleotides (ASOs) and then relatively decreasing the short isoform without exon 3, which encodes full-length oncogenic SRSF6 protein. This strategy has been successfully applied in decreasing overexpressed SRSF3 and inhibiting cell proliferation in oral cancer cells (Guo et al., 2018). Theoretically, this strategy depends on the identification of exonic splicing suppressors (ESSs) or intronic splicing suppressors (ISSs), which are responsible for the exclusion of SRSF6 exon 3. Then specific ASOs can be designed according to ESS or ISS sequence. These ASOs can bind to these motifs and block the interaction with regulatory factors and release the suppressive effects on exon 3 inclusion. The off-target effects of anti-splicing suppressor ASOs may be much less than siRNAs because splicing suppressors are hardly conserved in genome. However, similar to siRNAs, in vivo delivery is also the major obstacle for clinical application of ASOs (Gheibi-Hayat and Jamialahmadi, 2020).