Growth‐associated protein 43 promotes thyroid cancer cell lines progression via epithelial‐mesenchymal transition

Abstract Thyroid cancer is maintaining at a high incidence level and its carcinogenesis is mainly affected by a complex gene interaction. By analysis of the next‐generation resequencing of paired papillary thyroid cancer (PTC) and adjacent thyroid tissues, we found that Growth Associated Protein 43 (GAP43), a phosphoprotein activated by protein kinase C, might be novel markers associated with PTC. However, its function in thyroid carcinoma has been poorly understood. We discovered that GAP43 was significantly overexpressed in thyroid carcinoma and these results were consistent with that in The Cancer Genome Atlas (TCGA) cohort. In addition, some clinicopathological features of GAP43 in TCGA database showed that up‐regulated GAP43 is significantly connected to lymph node metastasis (P < 0.001) and tumour size (P = 0.038). In vitro experiments, loss of function experiments was performed to investigate GAP43 in PTC cell lines (TPC‐1 and BCPAP). The results proved that GAP43 knockdown in PTC cell significantly decreased the function of cell proliferation, colony formation, migration, and invasion and induced cell apoptosis. Furthermore, we also indicated that GAP43 could modulate the expression of epithelial‐mesenchymal transition‐related proteins, which could influence invasion and migration. Put those results together, GAP43 is a gene which was associated with PTC and might be a potential therapeutic target.


| INTRODUC TI ON
Thyroid cancer (TC) is maintaining at a high incidence level and it becomes one of the most common cancers in worldwide. 1,2 The US National Cancer Institute anticipated 53 990 new cases besides 2110 death numbers of patients because of TC in the USA in 2018. 3,4 Besides, the papillary thyroid cancer (PTC) is the commonest subtype of TC (accounts for 80%-90%). 5 Although the incidence of PTC is high, it is generally considered that PTC is relatively curable and has a good prognosis. 6,7 After a surgical operation or radio-iodinated therapy, patients with PTC show satisfactory prognosis besides a general 10-year survival rate of 90%. 7 However, clinical course of PTC usually follows indolent and carries excellent prognosis, it is still highly metastatic and relapses after routine treatment. 5 Lymph node metastasis (LNM) is the main factor of locoregional recurrence and a distinct risk factor of mortality. [8][9][10] LNM has been shown to occur in approximately 20% of all PTC patients, and regional recurrence occurred in 10% of patients undergoing total thyroidectomy. 11 So, it is vital for us to facilitate the in-depth investigation of the mechanism of PTC and help doctors to provide an appropriate treatment approach.
The incidence and development of PTC are primarily affected by genomic alternation, involving stimulation of oncogene and silencing of a tumour suppressor gene. Accumulating reports have been published about molecular mechanisms of PTC over nearly a period of 20 years. B-type Raf kinase (BRAF) V600E, a famous gene mutation, promotes PTC tumorigenesis and progression by abnormally stimulating the mitogen-activated pathway kinase pathway. 12 Besides, some notable mutations such as RAS mutation, 13 TERT mutation, 14 PTEN mutation, 15 PIK3CA mutation 16,17 and TP53 mutation 18 also show a significant part in thyroid carcinoma. Numerous researchers have made remarkable progress in TC gene research, but many characters of PTC are yet relatively unspecified. Hence, examining for new potential markers and clarifying molecular mechanisms in the improvement of thyroid carcinoma is still essential.
In our previous study, we have conducted next-generation of 19 pairs of PTC tissues and adjacent normal thyroid tissue. 19 After evaluating all the information, we identified that growth-associated protein 43 (GAP43), a gene encoding an axonal membrane protein, as an oncogene promotes PTC tumorigenesis and tumour progression. 19 As for the normal function of this protein, it is normally expressed in neuronal somata, axons and growth cones during pre-and early postnatal development. 20,21 Consistently, it also plays a fundamental function in neural growth, axonal regeneration and stabilisation of synaptic function. 22,23 A study indicated that GAP43 genetic variants and its gene network interaction may associate with the susceptibility to Hirschsprung disease. 24 A considerable report of the mechanisms indicated that GAP43 may promote non-small lung cancer (NSLC) cell migration by activating Rac1 and mediating F-actin cytoskeleton polymerisation. 25 Emerging shreds of evidence have manifested that GAP43 has an important function in many aspects; however, whether the GAP43 gene also plays a critical role in PTC remains unclear.
To evaluate the mRNA expression profiles of PTC, 19 paired PTC samples and adjacent non-carcinomatous samples were conducted to whole transcriptome sequencing. We firstly discovered GAP43 was significantly overexpressed in PTC tissue. Hence, we collected 50 pairs of tumour samples and adjacent normal tissues to validate the expression of GAP43 by using quantitative reverse transcriptionpolymerase chain reaction (qRT-PCR). Moreover, we also analysed the relationship of GAP43 expression and clinical features in PTC.
Then, we performed a series of cell line experiments and Western blot by siRNA to explore the role of GAP43 in TC. We established that loss of function about GAP43 effectively inhibits the proliferation and invasion of two PTC cell lines (TPC-1, BCPAP) and promotes apoptosis in vitro. Down-regulation of GAP43 in those cell lines decreased the N-cadherin and vimentin and increased E-cadherin.
All in all, we firstly proved that GAP43 has a potential role in PTC cell progression and could be a novel target in PTC patients.

| Samples collection
As a validated cohort, we gathered 50 fresh PTC and paired normal tissues from patients through surgery at the First Affiliated Hospital of Wenzhou Medical University between 2014 and 2018. All those patients did not acquire chemotherapy or radiotherapy for pre-treatment. The samples were instantly achieved at the time of initial surgery and were frozen in liquid nitrogen instantly after lesion resection and then stored at −80 Celsius before RNA extraction. All tumour tissues were histologically reviewed by two pathologists, and the cases were retrospectively reviewed by two senior pathologists to confirm the histological diagnosis. Informed consent for the scientific use of the biological material was obtained from each patient. All patient-de-

| The Cancer Genome Atlas (TCGA) database
A total of 502 PTC patients with complete clinicopathological characteristics (such as age, sex, LNM, tumor size, clinical stage (ACJJ7), and histological type) were collected for further analysis. All data were downloaded from the TCGA data portal (https ://tcga-data.nci. nih.gov/docs/publi catio ns/tcga/).

| RNA isolation and real-time reverse transcription/polymerase chain reaction (qPCR)
The total RNA from the tissues was isolated using TRIzol reagent (Invitrogen; Thermo Fisher Scientific, Inc), according to the manufacturer's protocol (Life Technologies, Carlsbad, CA), cDNA was prepared using a ReverTra Ace qPCR RT kit (Toyobo, Osaka, Japan).
To evaluate the RNA quality and quantity, we used the A260/A280 ratio as the control value. cDNA samples were all stored at −80°C.

| Colony formation assay
The two transfected or control groups were planted into six-well plates at

| Migration and invasion assays
In the migration assays, we used

| Protein extraction and Western blot analysis
The protein of transfected cells lysates was lysed in cell lysis buffer

| GAP43 overexpression in PTC
To investigate the function of GAP43 in PTC, we conducted 19 pairs of primary PTC tissues and adjacent normal tissues via whole transcriptome sequencing. After analysing this data, we identified GAP43 as up-regulated in most PTCs (Table 1). Furthermore, we detected the mRNA expression level of GAP43 in 50 patients via RT-qPCR to certify our sequencing results. In tumour tissues, expression level of GAP43 was significantly higher comparing with adjacent non-tumour tissue ( Figure 1A). This result is consistent with the data revealed in the TCGA cohort ( Figure 1B). TA B L E 1 The expression of growth associated protein-43 gene in 19 cases of thyroid papillary carcinoma was higher than that in normal tissue by whole transcriptome sequencing F I G U R E 1 The mRNA expression of GAP43 both in local validated cohort and TCGA cohort. A, GAP43 expression was significantly up-regulated in PTC tissues compared with the adjacent non-cancerous thyroid tissues using qRT-PCR (****P < 0.0001). B, Data including 502 PTC samples and 58 adjacent non-cancerous thyroid samples, GAP43 expression was also significantly down-regulated in PTC in TCGA cohort (****P < 0.0001). Abbreviations: GAP43, growth-associated protein 43; PTC, papillary thyroid carcinoma; qRT-PCR, quantitative reverse transcription-polymerase chain reaction; TCGA, The Cancer Genome Atlas

| GAP43 was associated with the clinicopathological features of PTC
To facilitate in-depth investigation of GAP43 in PTC, the relationship of GAP43 and the clinicopathological features was further analysed. In TCGA cohort, we divided 502 patients into high GAP43 expression (n = 251) and low GAP43 expression (n = 251) groups according to the median value. Results showed that histological type (P < 0.001), tumour size (P = 0.038) and LNM (P < 0.001) were significantly related to high GAP43 expression (Table 2). Conversely, no significant associations were found between GAP43 expression and gender, age, distant metastasis, disease stage (AJCC7) and multi-focality. Similarly, in our validated cohort, high GAP43 expression was correlated with the tumour size (P = 0.021) and LNM (P = 0.041), but disease stage (AJCC7) was insignificant (P = 0.774), as shown in Table 3.
Overall, those results have shown that GAP43 might serve as a oncogene in PTC.

| High GAP43 expression intensifies the risk of LNM in PTC patients
Through logistic regression analysis, we assessed the relationship be- TA B L E 2 The association between growth associated protein-43 expression and clinicopathologic features in The Cancer Genome Atlas cohort

| GAP43 knockdown restrains TPC-1 and BCPAP cell proliferation and colony formation
We also assessed the GAP43 expression level both in several PTC cell lines and normal thyroid cell lines (HTORI-3) by using RT-qPCR and WB. The results showed that GAP43 gene was commonly highexpressed in TPC-1, KTC-1 and BCPAP (Figure 2A, 2B). Given the GAP43 gene is commonly overexpressed in PTC, we hypothesised GAP43 plays an oncogenic function in TC cell progression. Next, we selected TPC-1 and BCPAP as relatively high-expression cell lines to explore the function of GAP43. We down-regulated the GAP43 expression level via effective small interfering RNA (siRNA-1, siRNA-2).
Obviously, Si-RNA could knock off the expression of GAP43 more than half ( Figure 2C, 2D). Next, down-regulation of GAP43 significantly inhibited TPC-1 and BCPAP cell line proliferation as assessed by CCK-8 and colony formation assays ( Figure 2E, 2F).

| GAP43 gene knockdown induces the cell apoptosis of PTC cell lines in vitro
We presumed that GAP43 also plays a role in the cycle of cell, so

| GAP43 modulates epithelial-mesenchymal transition in thyroid carcinoma cell lines
Accumulating studies reported that epithelial-mesenchymal transition (EMT) is a critical process especial during cancer cell metastasis. [26][27][28] As shown in Figure 5A-D, the expression of vimentin and N-cadherin was deceased and expression of E-cadherin was enhanced in si-GAP43 cell lines. Those results suggested that GAP43 may promote thyroid cell metastasis by changing EMT.

| D ISCUSS I ON
Thyroid cancer, as one of the quickly growing malignant tumour, is expected to surpass colorectal cancer and became the fourth most commonly diagnosed cancer in 2030. 29 Accumulating evidence has proved that GAP43 has a crucial function in the process of neural cell growth and development, but the underlying molecular mechanisms of GAP43 in thyroid carcinoma are still poorly understood.
Next-generation sequencing has been widely used in the study of the gene expression variations and cancer molecular mechanism.
In our previous study, 19  In a summary, the present study investigated the association between GAP43 and clinicopathological features in two PTC cohorts (TCGA and local). Additionally, we showed that GAP43 knockdown impairs TC cell proliferation, colony formation, migration, invasion and induces cell apoptosis via EMT pathway. These interesting findings provide a potential molecular marker of PTC for diagnosis and therapy.

CO N FLI C T O F I NTE R E S T
Above authors declare that they have no conflict of interest.

AUTH O R CO NTR I B UTI O N S
Chen Zheng did the main experiments and wrote the manuscript.
Rui-da Quan and Cheng-Yong Wu help gathered and analyzed the raw data. Adheesh Bhandari helped to revise the article. Xiaohua Zhang conceptualized and designed the study and provided supervision.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data sets supporting the conclusions of this study are included in this article and its additional images. Raw data are available on the main electronic data storage system of First Affiliated Hospital of Wenzhou Medical University and access can be provided upon request to the authors.