Every year, about 1.8 million new lung cancer cases are diagnosed, claiming 1.6 million deaths globally. Non-small cell lung cancer (NSCLC) is a major form of lung cancer, comprising approximately 84% of all lung cancers. Although many treatment options, including targeted therapy, are available, the survival rate is still very low because of ineffective treatment outcome and acquired drug resistance. In Hong Kong, the survival rate of NSCLC is about 15% only (Ho, 2011). Although cigarette smoking has predominant association with lung cancer, non-smokers constitute a substantial minority of lung cancer patients. In the USA, 17,000–26,000 annual deaths of non-smokers from lung cancer account for the seventh leading cause of cancer mortality (Rivera and Wakelee, 2016). With the exposure to passive smoke and cooking fumes, non-smoking women represent an under-studied subset of NSCLC patients (Shen et al., 1998).

The current therapeutic approaches mainly target the molecular mechanisms of tumor cell growth caused by the aberrant activity of epidermal growth factor receptor (EGFR) (Camp et al., 2005). EGFR tyrosine kinase inhibitors (TKIs) are small molecules designed to bind to the tyrosine kinase domain of EGFR so that its phosphorylation and the subsequent receptor activation and signal transduction can be inhibited. However, the cancer progression remains because of the acquired resistance to EGFR TKIs (Yu et al., 2013). To date, no effective therapy is available for NSCLC patients with such resistance. Multiple molecular mechanisms of resistance have been explored, such as the activation of alternative receptor tyrosine kinases (RTKs) that bypass the EGFR signaling pathway, and the secondary mutations in EGFR T790M (Camp et al., 2005; Zhang et al., 2012). The resistance to anti-EGFR therapy may be caused by the activation of other RTKs, such as hepatocyte growth factor receptor (c-MET), Ron (a protein tyrosine kinase related to c-MET), platelet-derived growth factor receptor (PDGFR), and insulin-like growth factor receptor-1 (IGF-1R), which share the same downstream signaling cascade with EGFR pathway (Camp et al., 2005). The other activated RTKs are able to substitute the function of EGFR by activating overlapping signal transduction. The downstream molecule phosphoinositide-3-kinase (PI3K) shared by both EGFR and IGF-1R is associated with the EGFR TKIs resistance (Huang and Fu, 2015).

For testing the normality of data, Shapiro–Wilk test was performed to SELDI-TOF-MS intensity data of both non-cancer and adenocarcinoma groups. Normality cannot be assumed in non-cancer group (p < 0.0001) and adenocarcinoma group (p = 0.001). So, the difference between two groups was examined using non-parametric test, independent samples Mann–Whitney U test. It was found that the intensities of m/z = 17230.8335 Da were significantly different between two groups (p = 0.002, median in non-cancer = 0.1898, median in adenocarcinoma = 0.3075, see Supplementary Materials). This implies that the protein expression of SIAH2 is significantly upregulated in adenocarcinoma.

Search Tool for Recurring Instances of Neighbouring Genes predicted the interactions among PI3K, CBL, UBE2D2, and SIAH2 (Figure 1). All the interactions were supported by experimental results and co-mentioning in literature. The CBL-UBE2D2 interaction was further supported by the pathways curated from third-party databases, such as Kyoto Encyclopedia of Genes and Genomes (KEGG). Datasets relevant to these interactions are shown in Table 1. Although there was no evidence showing that PIK3R2 directly interact with SIAH2, TargetScan predicted that miR-30-5p family concurrently targets them. The involved members of miR-30-5p family include miR-30a-5p, miR-30b-5p, miR-30c-5p, miR-30d-5p, and miR-30e-5p. The type of pairing site between PIK3R2 and miR-30-5p family members is 8mer and the target site is conserved among mammals. The type of pairing site between SIAH2 and miR-30-5p family members is 7mer-A1 and the target site is conserved among a number of vertebrates. It is interesting to note that miR-30-5p family also targets UBE2D2.

This work investigated the roles of PIK3R2, CBL, UBE2D2, SIAH2, and miR-30-5p family that may be related to the resistance to anti-EGFR therapy. Among these factors, SIAH2 has not been studied sufficiently and its role in cancer treatment remains unclear. In the comparison between plasma samples of independent groups, we showed the significant upregulation of SIAH2 protein in NSCLC. Moreno et al. (2015) obtained similar result but the SIAH2 expression in lung cancer was compared with that in adjacent normal tissues. Sun et al. (2016) found that SIAH2 was overexpressed in human breast cancer tissues. Our finding suggests that SIAH2 may be a novel theranostic target in the circulation of NSCLC patients.

Protein–protein interaction database search revealed the experimental proofs using the reconstituted complex assays that the E2 ubiquitin-conjugating enzyme, UBE2D2, interacts with both CBL and SIAH2, E3 ubiquitin ligases. CBL ubiquitylates active RTKs for degradation and attenuates directly the receptor signaling, including that mediated by EGFR (Joazeiro et al., 1999). The potential theranostic role of CBL has been shown in Klapper et al. (2000) that some antibodies may enhance directing HER-2 to CBL-regulated proteolysis. Similar to CBL, SIAH2 ubiquitylates protein for degradation. However, there is no evidence at the moment showing that the ubiquitination of EGFR is related to SIAH2. On the other hand, it was shown in Szargel et al. (2009) that SIAH2 is responsible for the ubiquitination of alpha-synuclein and synphilin-1 in Lewy bodies of PD patients. The only association with EGFR pathway is that both SIAH2 and MAPK are sensitive to hypoxia and MAPK, a downstream protein of EGFR can activate SIAH2 by phosphorylation (Nakayama et al., 2009).

TargetScan predicted that mRNAs of PIK3R2, UBE2D2, and SIAH2 are concurrently targeted by miRNA-30-5p family. These three RNA transcripts sharing the same conserved MREs are called competing endogenous RNA (ceRNA). CeRNAs interact indirectly by competing for their common targeting miRNAs in cells. The up-regulation of a ceRNA could indirectly promote the expression of the other ceRNAs (Marques et al., 2011). The molecular mechanism, through which the ceRNAs are co-regulated, paves the undiscovered cascades in the signaling pathway. Such co-regulation can be detected by the co-expression of ceRNAs. It was shown that PTEN and its putative ceRNAs are co-expressed in prostate cancer and glioblastoma (Tay et al., 2011). The ceRNA crosstalk is susceptible to the miRNA:target concentration ratio and, in other words, mediated by the targeting miRNAs (Bosson et al., 2014).

Among the miRNA-30-5p family members, microRNA-30a expression was found negatively associated with the malignant potential of various NSCLC cell lines, positively with the expression of E-cadherin, which is an epithelial marker, and negatively with the expression of N-cadherin, which is a mesenchymal marker. In other words, the epithelial-to-mesenchymal transition could be inhibited by miR-30a (Kumarswamy et al., 2012). Based on the above evidence, interaction of PIK3R2-SIAH2 ceRNAs and its mediator, miRNA-30-5p family, can be considered as a potential treatment target and theranostic panel in NSCLC, particularly adenocarcinoma. The treatment effect can be examined by intravenous injection of Lipofectamine reagents of miRNA-30-5p inhibitor/mimic in xenograft model and clinical trials.

The study was exempt from ethics approval because only the archived plasma samples were analyzed. No prospective study was performed.

LC wrote this paper, provided the research idea and performed the analysis of this work. FW advised about the association of microRNA-30 family with non-small cell lung cancer and the bioinformatics analysis. LW advised about the bioinformatics analysis. FM advised about the functional role of microRNA-30 family in non-small cell lung cancer. SW advised about the molecular and genomics perspective for theranostic markers in non-small cell lung cancer. JA advised about the clinical perspective for theranostic markers in non-small cell lung cancer. SY advised about the cellular self-homeostasis under stress, particularly hypoxic response. WC provided the proteomics data and advised about the clinical diagnostic need and the bioinformatics analysis.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.