Splice variants of the extracellular region of RON receptor tyrosine kinase in lung cancer cell lines identified by PCR and sequencing

RON is a member of the Met family of receptor tyrosine kinases (RTKs) and ample evidence suggests its deregulated expression and functioning in a number of cancers. RON is overexpressed in cancers involving organs such as skin, colon, breast, lung and kidney [1‐8] and its overexpression has been shown to be associated with tumor aggressiveness and metastasis [9]. Knockdown of RON expression in different cancer cell lines using siRNA/shRNA suppressed tumorigenic properties [10‐12]. Validation of overexpressed RON as therapeutic target, however, has been hampered by the production of several of its isoforms by tumors. Several alternatively spliced transcripts and their protein products have been described for RON in various cancer cell lines as well as in solid and pleural tumors [13]. Differential localization of RON isoforms has been confirmed by their presence in the tumor microenvironment, cytoplasm and nucleus [14‐16]. Recently, Western blotting studies indicated the presence of several RON isoforms in each of several lung cancer cell lines, while many of these same cell lines did not contain wild type RON [15].

The alternatively spliced transcripts of RON, which are formed by skipping of one or more exons, and their isoform products, show considerable sequence similarity with each other and to the wild type, despite exhibiting diverse functionalities in cancers. The isoforms include both hyperactive [2, 8, 17] and N-terminally truncated dominant negative variants [13, 18‐20]. Notably, ectopic expression of some of the RON splice variants in NIH3T3 cells lead to tumor formation in vivo [13].

Recently, we reported the presence of several, frequently occurring novel transcript variants involving exons coding for the intracellular region of RON in lung cancer cell lines [21, 22]; the current study was designed to identify the splicing variants that affect the extracellular region of RON using lung cancer cell lines.

Several alternatively spliced forms of RON transcripts and their protein products have previously been reported. However, PCR amplification and sequencing of the entire coding sequence (4200 bps, consisting of 20 exons) might have failed to provide a complete picture of all the alternative splicing events due to the formation of multiple products, which are expected to be similar in sequence except for one or a few missing exons. Hence, we determined the splicing alterations involving exon(s) skipping by amplification of two short, overlapping sections of RON cDNA, which together code for the complete extracellular portion of wild type RON protein. Results are summarized in Table 1.

Sequencing from 5′ end of 1187 bps RON cDNA PCR products of cell lines indicated the presence of an overlapping sequence starting at nucleotide 1231 (immediately following exon 1) of RON reference sequence (Fig. 1A). The new sequence that appeared following exon 1 exactly matched with that of exon 4 of reference sequence suggesting the skipping of exons 2 + 3 in some of the RON transcripts. The exons 2 + 3 skipping splicing variant was found to be present together with the wild type RON transcript in cell lines H249, H69 and H526.

Additionally, sequencing from 5′ end of 1187 bps PCR products also showed the appearance of two overlapping sequences starting at nucleotide 1231 of RON reference sequence (Fig. 1B). Of the two overlapping sequences, one matched with that of exon 3 while the other matched with that of exon 4, suggesting the presence of two unique splicing alterations; one formed by skipping of exon 2 and the other formed by skipping of exons 2 + 3. H82 was the only cell line that contained both these splice variants along with the wild type transcript.

Sequencing the 1187 bps PCR products from the 5′ end also showed the appearance of an overlapping sequence starting at nucleotide 1720 of RON reference sequence (Fig. 1C). The newly appearing overlapping sequence matched with that exon 7 suggesting the presence of a splice variation involving skipping of exons 5 + 6, together with the wild type RON transcript. The exons 5 + 6 skipping splice variant was detected only in cell line H524.

In some cell lines, sequencing the 1187 bps PCR products from the 3′ end showed the appearance of an overlapping sequence starting at nucleotide 2046 of RON reference sequence (Fig. 1D). The overlapping sequence matched with that of exon 5 suggesting the presence of a splice variant formed by skipping of exon 6. RON splice variant formed by skipping of exon 6, which was present along with wild type RON transcript, was found in cell lines H358, H661, SW900, H522, H526 and H146.

Sequencing the 789 bps PCR products of lung cancer cell lines from 3′ end showed an abrupt change of sequence from nucleotide 2441 of RON reference sequence, and the new sequence matched with the 5′ sequence of exon 7 (Fig. 1E). This implied the presence of a splice variant resulting from the combined loss of exons 8 and 9 due to direct splicing of exon 7 with 10. Further, in addition to the direct splicing of exon 7 with 10, a codon CAG was also retained from the intron between exons 9 and 10. RON splice variant formed by skipping of exons 8–9, together with its wild type sequence, was found in cell lines H358, H661, H146, H524, A549, H1437, H2170, SKLU1, H249, SKMES, H69 and H522. However, in cell line SW900, normally spliced wild type RON variant was not found in cell line SW900, while exons 8 + 9 skipping variant was detected.

The specific exons involved/skipped in alternative splicing are shown schematically in Fig. 2, in the context of full length structure of RON protein (showing constituent domains) and coding sequence of RON transcript (showing the constituent coding exons). Also, a supplementary table (Additional file 1: Table S1) is presented showing all the splicing variations of RON identified so far in our lab (including this study and previous two reports [21, 22]) using lung cancer cell lines.

Aberrant expression of RON in tumors involves frequent alternative splicing of transcripts leading to the formation of several isoform products which differ in their functionalities. However, high similarity among the structures of transcript variants or the protein isoforms poses problems in therapeutic target discovery and validation. In this study, we screened lung cancer cell lines for splice variants between exons 1 and 10 of RON transcripts through partial cDNA amplification by PCR and sequencing. Results revealed the presence of splice variants lacking exons 2, 2–3, 5–6, 6 and 8–9 in lung cancer cell lines. Of these, skipping of exons 2, 2–3 and 5–6 merely involved loss of 63, 106 and 109 amino acids, respectively, but did not cause frameshift; however, splice variants formed by skipping of exons 6 and 8–9 resulted in frame-shift together with the appearance of premature termination codons.

Alternative pre-mRNA splicing that eliminates exon 2 resulting in in-frame deletion of 63 amino acids in the extracellular domain of the RON β chain caused the production of a 165 kda variant that localized in the intracellular compartment [23]. Although it lacked tyrosine phosphorylation activity, the RONΔ165E2 variant was shown to phosphorylate and activate phosphatase and tensin homolog (PTEN). In the human colorectal cancer cell line HCT116, an in-frame deletion of 106 amino acids in the extracellular domain of RON β-chain caused by the deletion of exon 2 and exon 3 was shown to result in the production of a 160 kDa RON isoform, RONΔ160(E2E3), that localized intracellularly [24]. This isoform had no tyrosine phosphorylation ability, but it could constitutively activate Akt activity in transfected HEK293 epithelial cells. Splicing changes affecting the extracellular region have been shown to form dominant negative isoforms of RON, either secreted or intracellular, in cancers. Variant RONΔ85 is a soluble/secreted protein derived from a differentially spliced mRNA transcript arising out of retention of 49 nucleotide intron sequence between exons 5 and 6 [25]. RONΔ90 is another secreted RON isoform, coded by a transcript lacking exon 6 and as a result is N-terminally truncated [18]. Both these isoforms were detected in conditioned media in in vitro cell culture studies and were shown to be capable of binding both MSP and RON and block MSP/RON signaling. RON160 is a previously reported splice variant generated by deletion of 109 amino acids encoded by exons 5 and 6 [13, 20]. RON160 showed cell-transforming activities in cell focus formation and mediated anchorage-independent growth and RON160-mediated EMT was shown to be associated with increased motile/invasive activity [26]. In our study, splicing involving loss of exons 5–6 was only identified in 1 of 23 lung cancer cell lines indicating that this is not a common occurrence in lung cancer.

The splice variants formed by skipping of exon 6 and exons 8–9 result in reading frame-shift and were found in several cell lines examined in this study. RON transcripts with combined deletion of exons 8 and 9 have not been reported previously. Deletion of exons 8 + 9 result in the appearance of a premature termination codon; this may lead to the absence of the transmembrane sequence, which is coded by exon 12 of wild type RON; consequently, the N-terminally truncated peptide with most of the extracellular sequence of wild type RON may be secreted out. However, the transcript variants caused by loss of exons without concurrent frame-shift may lead to constitutively active or dominant negative activities. We speculate that the secreted RON isoforms may also bind and dimerize wild type RON located on the surface of tumor cells or on cells such as macrophages of the tumor microenvironment. MSP-RON signaling in macrophages suppresses inflammation by downregulating iNOS and COX-2 expression and increasing the production of the anti-inflammatory cytokine IL-10 [27‐29]. Hence, the soluble RON isoforms circulating in blood may serve to increase inflammation in distant organs, an important hallmark of cancer, and thereby facilitating metastasis.

Results from the present and the two earlier studies (performed in our lab) [21, 22] on RON splicing changes involving skipping of exons using lung cancer cell lines has revealed that incidence of splice variants is common in lung cancer. RON transcripts in H82 cell line showed splice variations at 7 locations suggesting their possible occurrence in various combinations in full length transcripts. H249, SW900, H526, H69, H524, A549 and H358 were the other cell lines that contained 4 or more alternative splicing events. Also, a majority of the exons of RON that were analyzed in these studies were involved in alternative splicing. Furthermore, in agreement with the identification of multiple alternatively spliced RON transcripts, our previous work showed the presence of multiple RON isoforms of different sizes in several lung cancer cell lines as identified by Western blotting [15].

This study was mainly focused on evaluating the incidence of splicing changes which may interfere with the qualitative and quantitative outcomes of various methods currently employed to study the role of RON signaling in cancer. Amplification and sequencing of smaller regions of the coding cDNA sequence has proved powerful enough to reveal the presence of several alternatively spliced mRNA transcripts in lung cancer that may have been missed by other methods. The results of this and the other two previous studies [21, 22] strongly suggest that the presence of various alternatively spliced mRNA transcripts and their isoform products must be taken into consideration while correlating expression of wild type RON to its functions in lung cancer.

Alternative splicing is a major general mechanism of genetic alteration used by cells to increase the repertoire and the functions of proteins and cancer cells stand to gain significantly by using this mechanism. Even though more than 90% of known transcripts are understood to undergo alternative splicing, this fact, which can interfere with both qualitative and quantitative methods, is not given due consideration in basic research or during target validation studies. Identification of several alternatively spliced variants together with the structural data from this study shed light on the frequent and ubiquitous nature of alternative splicing in RON in lung cancer and suggests that isoform specific quantification and functional determination may be an important prerequisite for understanding the deregulated RON signaling in cancers.