This article is mentioned in:

Introduction

It is well established that a variety of protein kinases induce cell transformation by phosphorylating downstream protein substrates. For example, Aurora-A is a Serine/Threonine (Ser/Thr) kinase and a widely known oncoprotein associated with the development of numerous human cancers. Aurora-A transforms cells (1–3) by phosphorylating a wide range of downstream substrates (4,5). For example, Aurora-A phosphorylates hepatoma upregulated protein (HURP) at four serine residues (6,7). This process relays the cell transforming activity from Aurora-A to HURP (6,7).

In contrast to these findings, protein phosphatases, which remove phosphate groups from protein substrates result in the prevention or termination of kinase-induced cell proliferative signals and therefore are usually considered as tumor suppressors (8). However, the role of the Ser/Thr phosphatase protein phosphatase 1α (PP1α) in tumor progression remains undetermined. PP1α has been revealed to prevent oncogenic transformation in NIH3T3 fibroblasts (9). PP1α activity is detrimental to cell growth by causing activation of Rb-dependent G1 arrest in bone osteosarcoma cells (10,11) and induction of apoptosis in T and fibroblast cells (12–15). In contrast to these observations, PP1α is overexpressed in ascites hepatoma (16–18), prostate cancer (19), diffuse large B-cell lymphoma (20), and oral squamous cell carcinoma cell lines (21). These studies seemingly indicate a differential influence of PP1α on different cancer types. Moreover, the contribution of PP1α to lung cancer remains ambiguous. For example, PP1α was revealed to be involved in the RIF1-induced tumor growth of the lung cancer cell line H1299 (22), while overexpression of PP1α in H1299 cells restricted tumor formation in vivo (23). Furthermore, downregulation or inactivation of PP1α was revealed to be associated with poor prognosis (24) and radioresistance (25) of lung cancer cells.

Numerous signaling cascades are regulated from protein kinases or protein phosphatases during cancer formation. The MAPK pathway and the AKT cascade are frequently altered in human cancer types (26,27). The mitogen activated protein kinase (MAPK) superfamily consists of the ERK, p38 and Jun N-terminal kinase (JNK) proteins, and functions by transmitting upstream signals to its downstream effectors, thereby regulating several aspects of cancer development (26,28). ERK and p38 are engaged in oncogenic processes, whereas JNK demonstrates an oncosuppressive role. Furthermore, AKT is activated by PI3K/PDK1 and inactivated by PTEN. The activation of AKT leads to the activation and/or the inactivation of a variety of downstream effectors, such as cyclin D1, and GSK3β or c-Raf, respectively (27,29–31).

To explore the role of PP1α in lung cancer and determine whether PP1α contributes to cell proliferation or transformation according to each tissue type, it was examined by literature review whether the DNA region covering the PP1α locus was subjected to frequent alterations, such as gain or loss in various types of cancer. Certain studies demonstrated that PP1α was upregulated in lung cancer tissues, and that it was required for the cell transforming activity of the lung cancer cell line A549 by activating the AKT signaling pathway. However, PP1α inhibited 293T cell proliferation by regulating the JNK and AKT cascades. Collectively, the data indicated that PP1α contributes to lung cancer formation and that it exhibits differential growth-stimulating effects in different cell types via distinct mechanisms of action.

Materials and methods

Chemicals, antibodies, plasmids and shRNAs

Fetal bovine serum (FBS), Dulbecco's modified Eagle's medium (DMEM), penicillin, streptomycin, and Lipofectamine™ were purchased from GIBCO-BRL; Thermo Fisher Scientific, Inc. The catalog numbers and suppliers for the antibodies used in the present study are as follows: PP1α (cat. no. sc-271762), GFP (cat. no. sc-9996), tubulin (cat. no. sc-5286), actin (cat. no. sc-8432) and NF-κB (cat. no. sc-8008; all from Santa Cruz Biotechnology, Inc.); cyclin E1 (HPA018169; Sigma-Aldrich; Merck KGaA); GAPDH (cat. no. sc-32233; Santa Cruz Biotechnology, Inc.); phospho-AKT pathway sampler kit, (AKT, AKT p308, AKT p473, phospho-c-Raf, phospho-GSK3β, phospho-PTEN, phospho-PDK1; cat. no. 9916), MAPK family sampler kit, (ERK, p38, JNK; cat. no. 9926), and phospho-MAPK family antibody sampler kit, (phospho-ERK, phospho-p38, phospho-JNK; cat. no. 9910; all from Cell Signaling Technology, Inc.); GSK3β (cat. no. sc-53931), c-Raf (cat. no. sc-52827) and PTEN (cat. no. sc-7974; all from Santa Cruz Biotechnology, Inc.). The antibodies against PP1α, tubulin, and actin, were purchased from Santa Cruz Biotechnology, Inc., and the antibodies against AKT or MAPK pathways were purchased from Cell Signaling Technology, Inc. The p-HURP antibodies were generated by immunizing rabbits with synthesized peptides containing phospho-Ser as follows: p627, VKLFSGLSVSSEGP; p725, CLSSERMSLPLLA; p757, EGMELNSSITSQDV; p830, EHARH ISFGGNLI. All the antisera were further purified by antigenic peptide-bound column. EGFP-PP1α was obtained from Dr Monique Beullens (51). The shRNAs used in the study were obtained from the National RNAi Core Facility at Academia Sinica in Taiwan and the targeting sequence for LacZ and PP1α was CGCGATCGTAATCACCCGAGT and TGAGTGCAAGAGACGCTACAA respectively.

Tissue procurement

The study protocol in terms of the collection of the biopsies of patients was approved by the Ethics Committee at Taichung Veterans General Hospital. No patient had previously received any neoadjuvant treatment such as chemotherapy before surgery. Patients (31) were recruited from 2004/4/7 to 2005/12/28 in the study with males accounting for 71%, and a mean age of 65 ranging from 41 to 87. All patients provided informed consent and signed the consent form individually. The study samples were obtained after surgery from a non-necrotic area of the tumor and from adjacent non-tumorous tissue from neighboring sites outside the tumor. Both tumor and adjacent non-tumor tissues were confirmed by pathologists. The tissue samples were placed immediately in cryovials, frozen in liquid nitrogen, and stored at −80°C until analysis by western blotting.

Cell culture

The cell lines used in this study were purchased from the American Type Culture Collection. The culture medium for 293T cells was Dulbecco/Vogt Modified Eagle's Minimal Essential Medium with 5% fetal bovine serum, and for A549 cells it was Roswell Park Memorial Institute (RPMI)-1640 medium with 10% fetal bovine serum, 1% nonessential amino acids, and 1% sodium pyruvate. Moreover, 2 mM glutamine, 100 U/ml penicillin, and 100 g/ml streptomycin were added in all media. All cell culture-related reagents were purchased from Invitrogen; Gibco; Life Technologies; Thermo Fisher Scientific, Inc. Cells were maintained in a humidified incubator at 37°C in the presence of 5% CO2.

Single cell proliferation assay

The assay was based on our previous study (22). Briefly, cells were seeded in 24-well plates and transfected with desired constructs tagged with EGFP. The following day, the cells were replated with low density to 10-cm dishes, to prevent cell contact. The cells were allowed to proliferate for 1–4 days, and the percentage of cells with proliferation judged by formation of ‘mini-colonies’ with cell numbers ≥2 was counted. If cells could not proliferate, they remained single in the 10-cm dish after 36-h culture.

Colony formation assay

Cells (2×103) were seeded in 10-cm dishes and cultured for 10 days. The cell foci were fixed with methanol and stained with 5% Giemsa solution. The focus number was then counted.

PolyHEMA-based anchorage-independent growth

PolyHEMA (2.5 mg/ml) was added to 6-cm dishes and heated on a hotplate until most of the solution evaporated. The coated dishes were sterilized under UV irradiation before use. Cells (5×105) were then seeded onto the dishes and allowed attachment and proliferation for 2 days. The cells were trypsinized and counted with a hemocytometer.

Migration wound healing assays

Cells with full confluence were incubated with serum-free medium for overnight. The cells were then scraped with a 20-µl gel-loading tip to create an empty space which allows for cell migration. Serum containing the medium was then added to cells and the migrated cells were photographed at 0, 3, 6 and 9 h with an CCD camera (Olympus, Model DP71) attached to an inverted microscope (Olympus Optical Co., Ltd., Tokyo, Japan). The cells with migration were counted.

Preparation of cell extracts and western blot analysis

To prepare the cell-free extracts, cells were lysed in extraction buffer (20 mM PIPES, pH 7.2, 100 mM NaCl, 1 mM EDTA, 0.1% CHAPS, 10% sucrose, 1 mM PMSF, 1 mM DTT, 1 mM Na3VO4, and 10 µg/ml each of leupeptin, aprotinin, chymostatin, and pepstatin). After incubation at 4°C for 30 min, cellular debris was removed by centrifugation at 15,856 × g for 90 min in an Eppendorf centrifuge. Protein concentrations were determined using the Bradford assay (Bio-Rad Laboratories, Inc.). The resulting samples were heated at 95°C for 10 min and loaded onto a 10% SDS-polyacrylamide electrophoresis gel (SDS-PAGE) and then transferred onto a polyvinylidene difluoride membrane (PVDF; EMD Millipore). The PVDF membrane was then blocked with 5% bovine serum albumin (BSA)/PBST (0.1% Tween-20 in PBS). Various antibodies were incubated with the membranes at 4°C for overnight. The membranes were washed with PBST at room temperature for 30 min and repeated for 3 times. Secondary antibodies against mouse (cat. no. sc-2005) or rabbit (cat. no. sc-2004) IgG conjugated with horseradish peroxidase (Santa Cruz Biotechnology, Inc.) were added for 1 h at room temperature followed by washing with PBST for 3×30 min. NBT and BCIP substrates (Zymed Laboratories, Inc.) were added to develop the membrane. The protein level was assessed using the software (Gel-Pro analyzer 4.5), and the signal intensity of the protein of interest was normalized against that of actin, tubulin, or GAPDH. When comparing the changes of protein between the control and experimental groups, a relative level of each protein of interest was used, where [protein of interest/actin]experimental was normalized with [protein of interest/actin]control.

Immunoprecipitation

One milligram of cell extracts with protease inhibitors were incubated with protein A/G beads in 500 µl immunoprecipitation washing buffer (50 mM HEPES, pH 7.6, 2 mM MgCl2, 50 mM NaCl, 5 mM EGTA, 0.1% Triton X-100, 40 mM glycerolphosphate) at 4°C, for 1 h, to preabsorb unwanted proteins. Antibodies in the amount of 1 µg were then added to the cell extracts for 4 h at 4°C. The cell extracts were then incubated with protein A/G-beads for 1 h, followed by 6 washes with PBS for 3 h at 4°C. The resulting samples were heated at 95°C for 10 min and applied to SDS-PAGE-based electrophoresis and western blotting.

Statistical analysis

All data were collected from three independent experiments and presented as the mean ± SD. Statistical significance determined by independent samples Student's t-tests were represented as P<0.05, P<0.01, and P<0.001.

Results

PP1α is upregulated in lung cancer tissues

In order to examine the potential involvement of PP1α in cancer formation, we initially discovered that the DNA region covering the PP1α locus, i.e. 11q13.2, was subjected to DNA alteration in various types of human cancer (Table I). The majority of these alterations were DNA amplifications. For example, in the digestive system the following cancer types were present: Esophageal squamous cell carcinoma, hepatocellular carcinoma and colorectal cancer, whereas in the immune system, myeloid leukemia and plasmablastic lymphoma were noted. Moreover, PP1α protein levels were increased in lung cancer. At least, a 2-fold increase in PP1α protein levels was observed in 15 out of 31 lung cancer pairs of cancerous and adjacent normal tissues (Fig. 1). These findings indicated that deregulation of PP1α occurred concomitantly with cancer formation.

Figure 1. PP1α is upregulated in lung cancer tissues. Protein expression of PP1α in a collection of 31 samples of lung cancer tissue. (A) Biopsies from paired lung tumor (T) and adjacent normal tissues (N) were subjected to western blotting. (B) The quantification of PP1α protein levels in the collected paired lung tissues is displayed . The numbers on the (A) N and T or (B) x-axis represent the codes of the patients. The intensity of PP1α or β-actin was quantified by densitometry. The levels of the PP1α protein in each paired tissue were normalized against those of β-actin. The ratio of normalized PP1α levels in the tumor tissues to those observed in the normal tissues was calculated and plotted in the lower image. PP1α, protein phosphatase 1α.

Table I. The DNA alteration covering PP1α locus 11q13.2 in various cancer tissues.

Table I.

The DNA alteration covering PP1α locus 11q13.2 in various cancer tissues.

Altered DNA region	Cancer types	DNA alteration	(Refs.)
llql3.2	esophageal squamous cell carcinoma	amplification	(32)
llql3.2	esophageal squamous cell carcinoma	amplification	(33)
llql3.2	esophageal squamous cell carcinoma	amplification	(34)
llql3.2	intrahepatic cholangiocarcinoma	amplification	(43)
llql3.2	hepatocellular carcinoma	amplification	(44)
llql3.2	colorectal cancer	amplification	(49)
llql3.2	myeloid leukemia	amplification	(37)
llql3.2	plasmablastic lymphoma	amplification	(46)
llql3.2	myeloid leukemia	amplification	(47)
llql3.2	ovanan cancer	amplification	(35)
llql3.2	ovanan cancer	deletion	(36)
llql3.2	prostate cancer	amplification	(38)
llql3.2	prostate cancer	amplification	(39)
llql3.2	breast cancer	amplification	(40)
llql3.2	breast carcinoma	deletion	(41)
llql3.2	invasive lobular carcinoma	amplification	(45)
llql3.2	urothelial carcinoma	amplification	(42)
llql3.2	peritoneal mesothelioma	deletion	(48)
llql3.2	skin cancer	amplification	(50)

[i] PP1α, protein phosphatase 1α.

PP1α is required for the transformation of the A549 lung cancer cell line

To explore the role of PP1α in lung cancer development, an ectopic overexpression model of PP1α was initially performed in A549 cells, and it was further demonstrated that PP1α could promote cell proliferation (Fig. 2A). Secondly, PP1α expression was knocked down in A549 cells and the PP1α-depleted cells exhibited a considerably slower rate under normal or low serum conditions (Fig. 2B). Low growth was also observed in the presence of a cellular environment that would not allow strong attachment of the cells (Fig. 2C) and in the presence of a low number of cells (Fig. 2D). Moreover, A549 cells harboring a PP1α shRNA sequence exhibited reduced migratory activity (Fig. 3). All these observations revealed that PP1α was a potential oncoprotein in lung cancer cells.

Figure 2. PP1α is required for the cell transforming activity of A549 cells. (A) PP1α stimulated cell proliferation in A549 cells. The cell proliferative activity of A549 cells transfected with EGFP or EGFP-PP1α was assessed (see Materials and methods). The % of proliferation was counted and plotted. (B) Knockdown of PP1α impaired cell proliferation under normal or low serum conditions. A549 cells harboring LacZ shRNA (shLacZ) or PP1α shRNA (shPP1α) were examined for the relative knockdown efficiency (left) and analyzed for cell proliferative activity by counting the number of cells using a hemocytometer in 10% (middle) or 0.5% (right) serum. (C) Knockdown of PP1α impaired anchorage-independent growth. A549 cells harboring shLacZ or shPP1α were analyzed for cell proliferative activity on polyHEMA-coated plates. (D) Knockdown of PP1α reduced colony formation activity. A549 cells harboring shLacZ or shPP1α were analyzed using the colony formation assay. All data were collected from three independent experiments with error bars representing the SD. *P<0.05, **P<0.01 and ***P<0.001 represented significant differences as determined using Student's t-test, respectively. PP1α, protein phosphatase 1α.

Figure 3. PP1α is required for cell migration. (A) A549 cells harboring shLuc or shPP1α were analyzed for cell migration by wound healing assay for 0–9 h. (B) The cells that migrated into the scraped wound were counted and plotted. The data were collected from three independent experiments with error bars representing the SD. *P<0.05 and **P<0.01 represented significant differences as determined using Student's t-test, respectively. PP1α, protein phosphatase 1α.

PP1α activates the PDK1/AKT pathway

To explore the underlying mechanisms by which PP1α causes cell transformation, the expression levels of the proteins involved in MAPK and/or the AKT signaling cascades were investigated (Fig. 4). Although PP1α did not affect the MAPK pathway, the protein phosphatase regulated the AKT cascade. It was demonstrated that knockdown of PP1α downregulated the levels of PDK1 and AKT p473, whereas it upregulated the levels of p-Gsk3β and p-c-Raf.

Figure 4. Effects of PP1α on the MAPK or the AKT pathway. A549 cells harboring shLuc or shPP1α were analyzed for the expression of the proteins involved in the MAPK or AKT pathways. The relative expression of each affected component was measured, calculated and plotted on the right of the Figure. The data were collected from three independent experiments with error bars representing the SD. *P<0.05 and **P<0.01 represented significant differences as determined using Student's t-test, respectively. PP1α, protein phosphatase 1α.

PP1α inhibits cell proliferation of 293T cells

In addition to the amplification of the chromosomal region nearby the PP1α locus 11q13.2, the data further indicated that 11q13.2 was deleted in certain cancer types of sex hormone-dependent tissues, such as ovarian and breast cancer, and in other cancer types, such as peritoneal mesothelioma (Table I). To examine whether PP1α exerts a negative contribution to cell transformation, PP1α was overexpressed in 293T cells and it was demonstrated that the cell proliferative rate was largely impaired (Fig. 5).

Figure 5. PP1α inhibits cell growth in 293T cells. 293T cells were transfected with EGFP or EGFP-PP1α and were assayed for cell proliferative activity for 1–4 days by the cell proliferation assay. The data were collected from three independent experiments with error bars representing the SD. and the symbols *P<0.05 and **P<0.01 represented significant differences as determined using Student's t-test, respectively. PP1α, protein phosphatase 1α.

PP1α regulates the dephosphorylation of the Aurora-A downstream substrate HURP

To explore the mechanism by which PP1α retards cell proliferation, we expanded our analysis on the finding that PP1α acted against Aurora-A by dephosphorylating its substrates. The previous studies conducted by our group established that Aurora-A could phosphorylate HURP and promote cell transformation (6,7). Based on these findings the potential involvement of HURP was investigated in the PP1α-dependent inhibition of cell proliferation. It was initially revealed that PP1α could interact with HURP (Fig. 6A). HURP is phosphorylated by Aurora-A during mitosis (7). Therefore, it was assessed whether the interaction of Aurora-A with PP1α could reduce the phosphorylation levels of HURP. The data demonstrated that overexpression of PP1α decreased the levels of the four p-HURP isoforms catalyzed by Aurora-A in untreated cells or nocodazole-induced mitosis (Fig. 6B), whereas knockdown of PP1α exhibited the opposite effects (Fig. 6C). These isoforms were derived based on the type of the amino acid residue that was phosphorylated each time. It has been reported that the Aurora-A-induced phosphorylation of HURP stimulates cell proliferation by downregulating p-JNK (6). Therefore, the expression levels of JNK were investigated and it was demonstrated that PP1α significantly increased the levels of p-JNK in 293T cells (Fig. 7A). Notably, PP1α did not affect p-JNK levels in A549 cells (Fig. 4), while it reduced the levels of AKT p473 in 293T cells (Fig. 7B), revealing a differential action of PP1α in the two cell lines.

Figure 6. PP1α interacts with HURP and controls the dephosphorylation of HURP. (A) PP1α interacts with HURP. Immunoprecipitation using an antibody against PP1α was conducted in 293T cells transfected with EGFP or EGFP-HURP. sup, supernatant; ppt, pellet. (B) Overexpression of PP1α reduced the levels of p-HURP. 293T cells transfected with EGFP or EGFP-PP1α were treated with or without nocodazole and they were subsequently analyzed for the expression levels of the four p-HURP isoforms at S627, S725, S757 and S830 residues. The levels of each p-HURP isoform and HURP in each treatment were measured, the ratio of p-HURP/HURP was calculated, and [p-HURP/HURP]EGFP-PP1α/[p-HURP/HURP]EGFP was finally obtained and plotted. Since HURP may act at the mitotic phase, cells were treated with nocodazole in order to examine p-HURP levels during mitotic cell division. (C) Knockdown of PP1α increased the levels of p-HURP. 293T cells harboring shLuc or shPP1α were treated with or without nocodazole and subsequently analyzed for the levels of various p-HURPs. The levels of each p-HURP isoform and HURP in each treatment were measured, the ratio of p-HURP/HURP was calculated, and [p-HURP/HURP]EGFP-PP1α/[p-HURP/HURP]EGFP was finally obtained and plotted. All data were collected from three independent experiments with error bars representing the SD. *P<0.05, **P<0.01 and ***P<0.001 represented significant differences as determined using Student's t-test, respectively. PP1α, protein phosphatase 1α; HURP, hepatoma upregulated protein.

Figure 7. Ectopic PP1α expression causes an increase in the levels of p-JNK and a decrease in the levels of pAKT (residue 473), respectively. (A) The ectopic expression of PP1α affected the levels of p-HURP that in turn regulated the downstream factors of the pathway. A549 cells transfected with EGFP or EGF-PP1α were analyzed for the expression of the proteins regulated by p-HURP. The relative expression levels of the affected proteins were measured, calculated and plotted on the right part of the Figure. (B) PP1α decreased the levels of AKT p473. 293T cells transfected with EGFP or EGFP-PP1α were analyzed for the expression of proteins involved in the AKT cascade. All data were collected from three independent experiments with error bars representing the SD. **P<0.01 represented a significant difference as determined using Student's t-test. PP1α, protein phosphatase 1α; HURP, hepatoma upregulated protein.

Discussion

The chromosomal region covering PP1α is subjected to DNA alterations in human cancers including amplification and deletion, which reflects the distinct effect of PP1α on the development of different cancer types. The present study revealed that knockdown of PP1α attenuated the cell transforming activity and inactivated the AKT signaling cascade. This effect included the decrease in the p-AKT levels at the T473 residue, whereas the MAPK/JNK pathway was not affected in A549 cells. In contrast to these findings, overexpression of PP1α inhibited cell proliferation, while enhancing the dephosphorylation of HURP. Concomitantly, it inhibited p-JNK levels induced by p-HURP and it increased the levels of p-AKT at the T473 residue in 293T cells. Therefore, PP1α upregulated p-AKT levels and positively regulated cell transformation in A549 cells. PP1α downregulation of p-AKT levels increased the expression of p-JNK and inhibited 293T cell proliferation. The data indicated that PP1α exerted differential growth regulatory effects on A549 and 293T cells via distinct or even opposite mechanisms. PP1α has been revealed to activate the MAPK pathway in androgen-responsive cell lines, such as PC3 cells (52). In the present study, it was reported that PP1α inactivated the MAPK cascade by increasing the levels of p-JNK in 293T cells, where the androgen receptor was not expressed in these cell line models (53–55). Therefore, the distinct effect of PP1α on the MAPK pathway may depend on the androgen levels.

Furthermore, the present study demonstrated the essential role of PP1α in the maintenance of the cell transforming activity of lung cancer A549 cells. Initially, the overexpression of ectopic PP1α promoted cell proliferation. Moreover, the silencing of PP1α impaired cell migration and blocked cell proliferation under normal and low serum conditions, and in the presence of a low number of cells and a cell culture environment that did not allow strong attachment. In addition, PP1α was revealed to be overexpressed in 15 out of 31 lung cancer tissues. Collectively, the data supported the oncogenic role of PP1α in lung cancer formation, which is in line with previous studies highlighting that downregulation or inactivation of PP1α is associated with poor prognosis (24) and radioresistance (25) of lung cancer cells. However, negative correlation of PP1α with lung cancer has also been reported (23).

Acknowledgements

We would like to thank Ms. Mei-Chun Liu from the Instrument Center and Ms. Jen Miao at Taichung Veterans General Hospital and Ms. Li-Wen Lee from the Department of Surgery at Taichung Veterans General Hospital for their technical support.

Funding

The present research was supported by the Taichung Veterans General Hospital-National Chi Nan University Joint Research Program (TCVGH-NCNU1077902 and TCVGH-NCNU1067902), Taichung Veterans General Hospital (TCVGH-1083207C and TCVGH-1083208D), the Ministry of Science and Technology (MOST 105-2320-B-260-001-MY3), and the China Medical University and Hospital grant (DMR-107-116, 108-119) awarded to SCC.

Availability of data and materials

The datasets used and/or analyzed during the present study are available from the corresponding author on reasonable request.

Ethics approval and consent to participate

The Institutional Review Board and Human Ethics Committee of Taichung Veterans General Hospital (Taichung, Taiwan) approved the study. All patients signed informed consent for participation in the present study.

Authors' contributions

KCC, SCC, JYH and CTRY designed the study, collected the literature, analyzed and interpreted the data. CTRY prepared the manuscript and arranged the manpower. JMMC, RYC and YRJH performed the experiments. All authors approved the final manuscript and agree to be accountable for all aspects of the research in ensuring that the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Patient consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

References