Endogenous thrombopoietin promotes non‐small‐cell lung carcinoma cell proliferation and migration by regulating EGFR signalling

Abstract Thrombopoietin (TPO) is a haematopoietic cytokine mainly produced by the liver and kidneys, which stimulates the production and maturation of megakaryocytes. In the past decade, numerous studies have investigated the effects of TPO outside the haematopoietic system; however, the role of TPO in the progression of solid cancer, particularly lung cancer, has not been well studied. Exogenous TPO does not affect non‐small‐cell lung cancer (NSCLC) cells as these cells show no or extremely low TPO receptor expression; therefore, in this study, we focused on endogenous TPO produced by NSCLC cells. Immunohistochemical analysis of 150 paired NSCLC and adjacent normal tissues indicated that TPO was highly expressed in NSCLC tissues and correlated with clinicopathological parameters including differentiation, P‐TNM stage, lymph node metastasis and tumour size. Suppressing endogenous TPO by small interfering RNA inhibited the proliferation and migration of NSCLC cells. Moreover, TPO interacted with the EGFR protein and delayed ligand‐induced EGFR degradation, thus enhancing EGFR signalling. Notably, overexpressing TPO in EGF‐stimulated NSCLC cells facilitated cell proliferation and migration, whereas no obvious changes were observed without EGF stimulation. Our results suggest that endogenous TPO promotes tumorigenicity of NSCLC via regulating EGFR signalling and thus could be a therapeutic target for treating NSCLC.

Although other treatment options are available, including chemotherapy, radiotherapy, immunotherapy and molecularly targeted therapy, 3 lung cancer exhibits a very poor prognosis and nearly half of all patients die within one year of diagnosis, with a 5-year survival rate of 11%. 4 A better understanding of NSCLC initiation and development mechanisms is urgently needed, and studies are required to identify more oncogenes and suppressor genes as well as targetable gene alterations.
Thrombopoietin (TPO) is a haematopoietic cytokine that is mainly produced by the liver and kidneys and whose main function is to regulate megakaryocyte progenitor expansion and differentiation. 5 Scientists first successfully purified TPO in 1994; after it was cloned, thrombopoietic agents rapidly entered clinical development. PEG-rhMGDF and rh-TPO were the first-generation thrombopoietic drugs. Soon afterwards, TPO receptor agonists romiplostim and eltrombopag were produced as second-generation thrombopoietic drugs that had a better efficacy and rare side effects. These agents are mainly used to cure various types of primary or secondary thrombocytopenia, particularly thrombocytopenia caused by chemotherapy. Exogenous TPO acts by binding to the TPO receptor (C-MPL) and initiating various signal transduction pathways. [6][7][8] Most solid tumour tissues and cell lines, including NSCLC, do not express C-MPL or exhibit extremely low expression. [9][10][11] Consequently, exogenous TPO and TPO receptor agonists cannot affect the progress of these tumours, and thrombopoietic drugs were considered safe to cure thrombocytopenia in most cancer patients. In 1990s, a study reported TPO mRNA expression in some solid tumour cell lines 11 ; however, the function of endogenous TPO in solid tumours has generally been neglected. Our study focuses on the endogenous TPO produced by NSCLC cells and whether it affects the occurrence and development of lung cancer.
Epidermal growth factor receptor (EGFR) is a receptor tyrosine kinase belonging to the ErbB family that plays crucial roles in both normal physiological and pathological processes. 12 EGFR is activated upon binding to its ligands, after which a series of conformational changes occur in both the extracellular and intracellular domains. 13 This leads to the trans-autophosphorylation of tyrosine residues in the C-terminal regulatory domain and initiates intracellular signalling cascades, 13 including numerous signalling pathways such as the PI3K/ AKT, RAS/MAPK, STAT and JNK pathways. 14 This process can regulate cell apoptosis, proliferation, migration and differentiation, which have important effects on cancer phenotypes. 15 Increased EGFR expression has been found in multiple types of human cancers, including lung, breast, colon, oral and kidney cancers. 16 Thus, EGFR serves as a prognostic indicator in many types of cancer as its high expression is related to poor prognosis. EGFR mutations in lung cancer were first detected in 2004 and have subsequently been widely examined, resulting in the development of new therapeutic strategies for patients with NSCLC. EGFR tyrosine kinase inhibitors and monoclonal antibodies against EGFR have become critical for treating NSCLC as they prolong the survival of patients with advanced NSCLC. 17 However, EGFR-targeted therapy still has various limitations in clinical practice that necessitate further studies of EGFR and EGFR signalling.
Co-expression of EGFR and its ligands is commonly found in primary lung cancer. 18 EGF is the best characterized EGFR-specific ligand serving as agonist of EGFR signalling. 19 Its binding to the extracellular domain of EGFR initiates EGFR signalling which, once activated, is controlled by a negative feedback mechanism to avoid constant activation. 20,21 Endocytic trafficking, which involves internalization, endosomal sorting and lysosomal degradation, can attenuate EGFR signalling by removing and down-regulating activated receptors from the plasma membrane 22 and plays an important role in regulating the extent and duration of EGFR signalling. 23,24 Endocytic trafficking and degradation is a hot topic in EGFR signalling research because of its crucial role and high degree of complex- ity. An increasing number of studies have attempted to determine the behaviour of this complex process; however, the exact mechanism remains unclear.
In this study, we investigated the expression of TPO in NSCLC tissues and its clinicopathological relevance, as well as explored the biological role of endogenous TPO in NSCLC cells. We found that TPO interacts with EGFR and influences EGFR signalling. This haematopoietic cytokine may enhance the development and progression of NSCLC and may be a useful therapeutic target.

| Cell lines and cell culture
A549, H1299, H460, SK-MES-1 and H292 cells were purchased from the Cell Bank of the China Academy of Sciences (Shanghai).
HBE cells were obtained from the American Type Culture Collection (ATCC). All cells were cultured according to the instructions of the ATCC/CTCC. All media were purchased from Gibco, and FBS was purchased from Clark.

| Plasmid transfection and small interfering RNA treatment
Transfection was performed with Lipofectamine 3000 reagent For TPO knockdown, TPO-specific small interfering RNA (siRNA) and negative control siRNA were purchased from RiboBio. The TPO-siRNA sequence was GGATACACGAACTCTTGAA.

| Cell proliferation assay and colony formation assay
Assays were performed as described previously. 27 For the Cell Counting Kit-8 (CCK-8) assay, 3000 cells were added to each well of a 96-well plate containing 100 μL medium. Absorbance was quantified at 450 nm every 24 h for 5 days to generate a growth curve. For the colony formation assay, 1000 cells were added to each 60-mm cell culture dish containing 4 mL medium and incubated in a 5% CO 2 incubator at 37°C for 10-14 days. Images were acquired with a bio-imaging system (DNR). All experiments were repeated independently at least three times.

| Cell migration analysis
Assays were performed as described previously. 25 Cells were col- Cells were stained after incubation for 20 h. All experiments were repeated independently at least three times.

| Evaluation of TPO in cell medium by ELISA
The cell lines were cultured in RPMI 1640 medium containing 10% FBS. After 48 h, the medium of each cell line was collected for TPO detection. TPO was evaluated using an ELISA kit (KeyGen Biotech, Nanjing, China) according to the manufacturer's instructions. The minimum detection range was 62 pg/mL. All experiments were repeated independently at least three times.

| Co-immunoprecipitation assay
Assays were performed as described previously. 28

| Growth factors
Recombinant human EGF and recombinant human TPO were purchased from PeproTech (AF-100-15-100) and R&D Systems (288-TP-005/CF), respectively. F I G U R E 1 TPO is highly expressed in NSCLC tissues A, TPO expression was negative in (a) paired normal bronchial and (b) alveolar epithelial cells but was positive in NSCLC tissues: (c) highly differentiated adenocarcinoma; (d) poorly differentiated adenocarcinoma; (e) highly differentiated squamous carcinoma; and (f) poorly differentiated squamous carcinoma; (g) normal liver tissue; and (h) normal kidney tissue. Magnification, ×200. B, Western blot analysis indicated that TPO was highly expressed in fresh non-small-cell lung cancerous tissues (C) compared to corresponding non-cancerous tissues (N). Relative quantification of protein expression was analysed by ImageJ software. *P < 0.05; **P < 0.01

| TPO is highly expressed in NSCLC tissues and has significant clinical relevance
We  Table 1, TPO expression was also positively correlated with clinicopathological parameters of NSCLC patients, including differentiation (P = 0.015), P-TNM stage (P < 0.01), lymph node metastasis (P < 0.01) and tumour size (P < 0.01). We also stained 6 tissue samples of normal liver and kidney with the same antibody as positive controls ( Figure 1A).
Furthermore, we detected TPO expression in 10 paired fresh NSCLC and corresponding non-cancerous tissues by Western blotting, finding that TPO was highly expressed in NSCLC specimens compared to the surrounding normal tissue ( Figure 1B). cells compared to that in HBE cells but was weakly expressed in H460 cells (Figure 2A,B). We also detected whether the secreted TPO exists in the medium of these NSCLC cell lines and HBE cells.

| TPO expression and subcellular localization in
ELISA results revealed that there was no detectable TPO secreted from NSCLC or HBE cells ( Figure 2C). Immunofluorescence analysis of A549, H1299, SK-MES-1 and H292 cells showed that TPO was localized in both the cytoplasm and nucleus ( Figure 2D). As above,

F I G U R E 3 TPO suppression inhibits NSCLC cell proliferation and migration. A, B, CCK-8 and colony formation assays demonstrated
that A549 and H1299 cell proliferation was down-regulated when TPO was suppressed. C, Transwell assays showed that TPO knockdown inhibited A549 and H1299 cell migration. Magnification, ×200. D, Cyclin E1, cyclin E2, CDK2, c-Myc, P27, RhoA and RhoC protein levels were detected by Western blotting when TPO was down-regulated in A549 and H1299 cells. Relative quantification of protein expression was analysed by ImageJ software ( Figure S2). E, AKT, P-AKT (Ser473), mTOR, P-mTOR (Ser2448), EGFR and P-EGFR (Tyr1068) protein levels were detected by Western blotting when TPO was down-regulated in A549 and H1299 cells. Relative quantification of protein expression was analysed by ImageJ software ( Figure S2). *P < 0.05; **P < 0.01. Data are presented as the mean ± SD of three independent experiments

| TPO suppression inhibits NSCLC cell proliferation and migration
As TPO expression was correlated with NSCLC progression, we analysed whether TPO affects the biological functions of NSCLC cells. We used short interfering RNA (siRNA) to suppress TPO expression in A549 and H1299 cells and used Cell Counting Kit-8 (CCK-8) assay and colony formation assay to examine the growth and colony formation abilities of these cells. The results showed that down-regulating TPO suppressed the proliferation of these two cell lines ( Figure 3A and B). We also detected the protein levels of molecules involved in proliferation; cyclin E1, cyclin E2, CDK2 and c-Myc were down-regulated and P27 was up-regulated when TPO was suppressed ( Figure 3D). Transwell migration assays showed that suppressing TPO inhibited the migration of A549 and H1299 cells compared to control cells ( Figure 3C). Consistently, the expression of the migration-related proteins RhoA and RhoC was down-regulated when TPO was knocked down ( Figure 3D). As we found that normal bronchial epithelial HBE cells express detectable levels of TPO, we also suppressed TPO expression in HBE cells by siRNA; however, CCK-8, colony formation and Transwell assays showed no significant differences. These results suggest that the effect of TPO on proliferation and migration is not generalized but cancer-specific ( Figure   S1D-F).

| TPO overexpression does not affect the biological functions of NSCLC cells
We predicted that TPO overexpression may promote the biological functions of NSCLC cells; however, no positive results were obtained initially. The results of the CCK-8, colony formation and Transwell assays for A549 and H1299 cells transfected with the TPO plasmid showed no significant differences compared to the control groups (data not shown).

| TPO interacts with EGFR and influences EGFR/ PI3K/AKT/mTOR signalling
We next evaluated the changes in some critical signalling pathways related to cell proliferation and migration and found that P-AKT (Ser473) and P-mTOR (Ser2448) protein levels were significantly down-regulated when TPO was suppressed ( Figure 3E). Meanwhile, no visible changes were observed in P-ERK, P-MEK, P-P65, P-JNK or active β-catenin levels. Mass spectrometric analysis enabled us to identify EGFR as one among the many TPO-interacting proteins ( Figure 4A). As EGFR is a key regulator of PI3K/AKT/mTOR signalling, 30 we verified the interaction between transfected and endogenous TPO and EGFR in A549 and H1299 cells by co-immunoprecipitation assays ( Figure 4B,C) and detected the change in P-EGFR (Tyr1068) and EGFR levels when TPO was suppressed ( Figure 3E). Interestingly, the P-EGFR level changed significantly with TPO expression but the EGFR level showed no notable change.
Immunofluorescence analysis also showed TPO and EGFR to be colocalized in the cytoplasm ( Figure 4D). These data suggest that

| TPO influences EGFR signalling by delaying ligand-induced EGFR degradation
To investigate the mechanism via which TPO regulates EGFR signalling, we first assessed EGFR mRNA levels by RT-PCR when TPO was knocked down. However, in contrast, EGFR mRNA level was upregulated compared to that in the control group ( Figure 5A) represented the degradation rate of the receptor and its signalling.
We found that EGF-induced EGFR and P-EGFR degradation was delayed and impaired in TPO-overexpressed cells ( Figure 5D,F), whereas the opposite effects were observed when TPO was suppressed ( Figure 5C,E). Based on the above-mentioned results, we thought that the transcription change of EGFR is a compensation response due to the change in the EGFR degradation rate. Our study also demonstrated that TPO overexpression becomes functional in EGF-stimulated NSCLC cells, which is discussed in detail below.
Hence, we also analysed the EGFR mRNA level in EGF-stimulated A549 and H1299 cells, and found it to be down-regulated when TPO was overexpressed ( Figure 5B). Thus, our findings indicate that TPO is a new regulator of EGFR endocytic trafficking and degradation, which enhances EGFR stability and increases the duration and intensity of EGFR signalling.

| TPO overexpression facilitates the proliferation and migration of EGF-stimulated NSCLC cells
When A549 and H1299 cells were transfected with TPO plasmids, no obvious changes in biological function were observed.
The results shown above suggested that TPO affects the biological functions of NSCLC cells by increasing EGFR stability, delaying its down-regulation and enhancing its signalling. As A549 and H1299 cells contain wild-type EGFR, TPO overexpression may not promote their proliferation and migration as EGFR signalling and endocytic trafficking activity are relatively low. Thus, we stimulated A549 and H1299 cells with 100 ng/mL EGF, with TPO overexpression enhancing the proliferation and migration of these NSCLC cells ( Figure 6A-C). We also detected proliferation-,

F I G U R E 4 TPO interacts with the EGFR protein.
A, Mass spectrometric analysis predicted that TPO could interact with EGFR. B, Interaction between transfected TPO and EGFR was verified by co-immunoprecipitation assays in A549 and H1299 cells. A549 and H1299 cells were collected 48 h after transfection with pCMV6-Myc-DDK-TPO plasmids. Cell lysates were immunoprecipitated with anti-Myc antibodies (# 2276; Cell Signaling Technology) or control IgG and examined by anti-EGFR antibody (#4267; Cell Signaling Technology) and anti-Myc antibodies (# 2278; Cell Signaling Technology). C, Interaction between endogenous TPO and EGFR was verified by coimmunoprecipitation assays in A549 and H1299 cells. Cell lysates were immunoprecipitated with anti-TPO antibody (sc-374045; Santa Cruz Biotechnology) or control IgG and examined by anti-EGFR antibody (#4267; Cell Signaling Technology) and anti-TPO antibody (ab196026; Abcam). D, Immunofluorescence staining in A549 and H1299 cells showed that endogenous TPO and EGFR were colocalized in the cytoplasm. Magnification, ×400
These data further indicate that TPO influences the biological function of NSCLC cells by regulating EGFR signalling.  Figure S2). E, AKT, P-AKT (Ser473), mTOR, P-mTOR (Ser2448), EGFR and P-EGFR (Tyr1068) protein levels were detected by Western blotting when TPO was overexpressed in EGF-stimulated A549 and H1299 cells. Relative quantification of protein expression was analysed by ImageJ software ( Figure S2). *P < 0.05; **P < 0.01. Data are presented as the mean ± SD of three independent experiments

| D ISCUSS I ON
As a haematopoietic cytokine, TPO is mainly expressed in hepatic parenchymal cells and renal tubular cells. The lack of TPO receptor (C-MPL) is implicated in most solid tumour cells including NSCLC cells, and hence, these are not affected by exogenous TPO that acts by binding to this receptor. [9][10][11] We verified this conclusion using exogenous rh-TPO to stimulate A549 and H1299 cells.
CCK-8 and Transwell results showed that rh-TPO does not affect the proliferation or migration of these two cells ( Figure S1B,C).
Our study indicated that TPO is highly expressed in NSCLC tissues and cell lines but does not act like a typical cytokine. During our research, we found that TPO is well expressed but is not secreted in many NSCLC cells, which is different from the TPO expression in hepatic or renal cells. This is an interesting phenomenon, and In view of our current findings, the influence of TPO on EGFR signalling activity could be attributed to the regulation of EGFR degradation rate. However, there is an interesting phenomenon: When we knocked down TPO or overexpressed TPO in EGFstimulated NSCLC cells, the EGFR level did not change notably, whereas the P-EGFR level was altered significantly. This suggests that a negative feedback mechanism exists as EGFR mRNA level changes in a manner opposite to that of TPO and P-EGFR levels.
When the EGFR degradation rate is changed continuously, NSCLC cells tend to compensate by modifying the EGFR transcription process.
EGFR-targeted therapy has become a first-line treatment as it considerably improves the clinical outcome of NSCLC patients with sensitizing EGFR mutations. Mutant EGFR is constitutively activated and initiates ligand-independent signalling, resulting in high EGFR signalling and enhanced endocytic trafficking in NSCLC cells. 33 As our results showed that TPO promotes NSCLC progression by regulating EGFR degradation, we hypothesize that TPO could be a possible therapeutic target for treating NSCLC. In particular, the combination of EGFR-targeted and TPO-targeted therapy may have greater clinical efficacy. TPO-targeted therapy is not a new concept; besides haematopoietic system diseases, it has been studied and developed to assist in treating some solid tumours such as ovarian tumour and breast tumour. 34,35 However, these current studies targeted the exogenous TPO produced in the liver or kidneys, aiming to counter the paraneoplastic thrombocytosis, which may lead to platelet-dependent cancer progression. For NSCLC treatment, it will be quite different as anti-TPO treatment is supposed to specifically target the endogenous TPO in NSCLC cells.
Our study encountered some limitations. We were unable to analyse the prognostic information for the specimens used for im- the ratio of these two pathways differs. However, we were unable to take this uncertainty into account during our study. Determining the precise mechanism and exact timing of every step of EGFR endocytic trafficking has always been the main challenge in EGFR signalling research.

ACK N OWLED G EM ENTS
We thank the Department of Pathology in the Basic Medical Sciences College of China Medical University.