CTHRC1 induces non-small cell lung cancer (NSCLC) invasion through upregulating MMP-7/MMP-9
The strong invasive and metastatic nature of non-small cell lung cancer (NSCLC) leads to poor prognosis. Collagen triple helix repeat containing 1 (CTHRC1) is involved in cell migration, motility and invasion. The object of this study is to investigate the involvement of CTHRC1 in NSCLC invasion and metastasis.
A proteomic analysis was performed to identify the different expression proteins between NSCLC and normal tissues. Cell lines stably express CTHRC1, MMP7, MMP9 were established. Invasion and migration were determined by scratch and transwell assays respectively. Clinical correlations of CTHRC1 in a cohort of 230 NSCLC patients were analysed.
CTHRC1 is overexpressed in NSCLC as measured by proteomic analysis. Additionally, CTHRC1 increases tumour cell migration and invasion in vitro. Furthermore, CTHRC1 expression is significantly correlated with matrix metalloproteinase (MMP)7 and MMP9 expression in sera and tumour tissues from NSCLC. The invasion ability mediated by CTHRC1 were mainly MMP7- and MMP9-dependent. MMP7 or MMP9 depletion significantly eradicated the pro-invasive effects mediated by CTHRC1 on NSCLC cells. Clinically, patients with high CTHRC1 expression had poor survival.
CTHRC1 serves as a pro-metastatic gene that contributes to NSCLC invasion and metastasis, which are mediated by upregulated MMP7 and MMP9 expression. Targeting CTHRC1 may be beneficial for inhibiting NSCLC metastasis.
KeywordsLung cancer CTHRC1 MMP7 MMP9 Invasion/metastasis
Circulating tumour cell
Collagen triple helix repeat containing 1
Matrix metalloproteinase 7
Matrix metalloproteinase 9
Non-small cell lung cancer
Lung cancer is one of the most common malignant tumours and remains the leading cause of cancer-related death in China and around the world . Among all the lung cancers, non-small cell lung cancer (NSCLC) is the most common and aggressive type, accounting for ~ 85% of cases [2, 3]. Surgical resection remains the preferred clinical treatment for NSCLC patients in the early stages of disease. Despite advances in radio- and chemotherapy and the development of new targeted therapies in the past few years, 5-year survival rate remains poor in NSCLC patients due to unresectable advanced or metastatic disease at diagnosis. The high mortality and low cure rates for NSCLC are largely attributed to the strong ability of lung cancer cells to invade surrounding tissue or metastasize to other remote sites [4, 5, 6]. Hence, understanding the molecular mechanisms underlying NSCLC invasion and metastasis is essential.
Tumor metastasis is a complex process involving cell adhesion and proteolytic degradation of the extracellular matrix (ECM) [7, 8]. Matrix metalloproteinases (MMPs) are characterized by their ability to degrade extracellular matrix (ECM) proteins and expose cryptic sites within the matrix molecules to facilitate tumour invasion and metastasis [9, 10, 11, 12]. A previous study has shown MMP7 in promoting ovarian cancer cell invasion . Additionally, mice deficient in MMP9 are resistant to tumour metastasis . MMP9 is highly involved in strengthening the invasion capability of NSCLC . Clinically, MMP7 and MMP9 expression correlates with poor prognosis of NSCLC [16, 17, 18]. To further understand MMP modulation mechanisms in NSCLC to search for new therapeutic targets is thus imperative.
Collagen triple helix repeat containing 1 (CTHRC1) was originally identified in balloon-injured rat arteries, and its overexpression in fibroblasts is associated with increased cell migration, motility and invasion . CTHRC1 is widely upregulated in several solid tumours, including melanoma and cancers of the gastrointestinal tract, breast, thyroid, liver and pancreas . Furthermore, recombinant CTHRC1 protein augments the migration and invasion capacities of primary gastrointestinal stromal tumours . According to Chen et al. , CTHRC1 promotes tumour invasion and predicts poor prognosis in hepatocellular carcinoma. In our previous study, CTHRC1 overexpression in NSCLC cells was associated with tumour aggressiveness ; however, how CTHRC1 is involved in tumour cell migration and metastasis has yet to be fully elucidated. In a recent study by Park et al. , CTHRC1 regulated pancreatic cancer migration and adhesion by inducing Src, MEK and Rac1 activation. We previously showed the ability of CTHRC1 to increase the invasive capability of epithelial ovarian cancer cells by provoking constitutive activation of Wnt/β-catenin signalling . However, whether CTHRC1 is involved in cancer cell invasion and metastasis has not been completely clarified. Additionally, the mechanisms employed by CTHRC1 to regulate MMPs remain uncharacterized by previous investigations.
In our present study, we detected CTHRC1 overexpression in NSCLC tissues by performing a proteomic analysis and further confirmed the results through western blotting and in IHC assays at the tissue and cell levels. Furthermore, IHC analysis revealed a close relationship between CTHRC1 overexpression and lymph node metastasis, clinical stage and overall survival. Moreover, we demonstrated that CTHRC1 promoted tumour invasion by regulating MMP7 and MMP9 expression, which is mediated by the AP-1/c-Jun and NF-κB pathways, respectively. Additionally, serum concentration of CTHRC1 correlates with metastasis, clinical stage and circulating tumour cell (CTC) number and functions as an important prognostic factor for NSCLC patients. In summary, our findings represent an important step forward in understanding the role of CTHRC1 in NSCLC metastasis.
Patient and tissue information
Sera and primary tumour tissues were collected from a total of 230 cases of clinically and immunohistologically verified NSCLC (obtained from 2006 to 2011) identified in the pathology archives of the Affiliated First Hospital, Sun Yat-sen University, and the Central Hospital of Wuhan. NSCLC was verified by performing haematoxylin and eosin (HE) staining and immunohistochemistry as shown in Additional file 1: Figure S1A. Patients’ clinical characteristics are listed in Additional file 1: Table S1.
Proteomic analysis was performed. NSCLC tissues (n = 20) and adjacent non-tumour tissues (n = 20) were used to extract proteins for analysis. Tumour tissues were homogenized in lysis buffer via sonication on ice. A 2-D CleanQ2-Up Kit (Amersham Biosciences, UK) and a 2-D Quant Kit (GE Healthcare, London, UK) were used for protein purification and concentration qualification, respectively, according to the manufacturers’ instructions. Two-dimensional gel electrophoresis was performed using an immobilized pH gradient (IPG) strip (24 cm, pH 3–10 NL; GE Healthcare), in which proteins were separated according to their isoelectric point (pI) and molecular weight. Visualized stained proteins were selected using an Ettan Spot Handling Workstation (GE Healthcare), and spots of interest were digested with trypsin. Peptide mass mapping was performed via matrix-assisted laser desorption time-of-flight mass spectrometry (MALDI-TOF MS) using an ABI Voyager DE-STR mass spectrometer. The MASCOT Database (http://www.matrixscience.com/search_form_select.html) was employed to identify the original proteins. The search criteria were as follows: Homo sapiens, trypsin cleavage, and no constraints on either the molecular weight or the isoelectric point of the protein.
After electrophoresis, Membranes were incubated with primary antibodies in 5% milk/TBST at 4 °C overnight. The membranes were then washed with TBST and subsequently incubated with HRP-conjugated anti-rabbit IgG at room temperature for 60 min. Membranes were washed with TBST, and signals were detected via enhanced chemiluminescence (ECL). Cell nucleoprotein was extracted using EpiQuik Nuclear Extraction Kit (Epigentek, Farmingdale, NY) according to the manufacturer’s instructions.
Reverse dot hybridization
All assays were performed as previously described  and in strict accordance with the instructions provided with the kit (Roche, USA). Specific probes targeting MMP genes were designed and listed in Additional file 1: Table S2. siRNA sequences and RT-PCR primers are listed in Additional file 1: Tables S3 and S4 respectively.
Luciferase reporter gene assay
NCI-H1975 cells were seeded in 24-well plates in triplicate and allowed to adhere for 24 h. Luciferase reporter plasmids (200 ng) containing different fragments of the MMP7 or MM9 promoters were transfected into cells using Lipofectamine 2000 with 1 ng of pRL-SV40 Renilla luciferase as an internal control. Cell extracts were prepared 24 h after transfection, and luciferase signals were measured using the Dual-Luciferase Reporter Assay System (Promega, USA) according to the manufacturer’s instructions.
Chromatin immunoprecipitation (ChIP)
ChIP was performed using a Chromatin Immunoprecipitation Kit (Upstate) according to the manufacturer’s instructions. Briefly, NCI-H1975 cells were treated with 1% formaldehyde to cross-link proteins to DNA in a 100-mm culture dish. Sonication was applied to the cell debris to shear DNA into 300–1000-bp fragments. Equal amounts of chromatin supernatants, containing an antibody against CTHRC1 (1 μg) or an equal amount of control IgG, were incubated overnight at 4 °C with shaking. PCR was performed after the reverse cross-linking of protein/DNA complexes to release DNA.
Paraffin-embedded tissues were sectioned (4 μm) and incubated with anti-CTHRC-1 (Abcam, Cambridge, UK), anti-MMP7 (Abcam, Cambridge, UK) and anti-MMP9 (Abcam, Cambridge, UK) primary antibodies at 4 °C overnight. After washing with PBS, sections were then incubated with an HRP-conjugated goat anti-rabbit secondary antibody for 1 h at room temperature. Peroxidase was visualized with 3,3′-diaminobenzidine, and haematoxylin was used as a counterstain.
CTC enrichment using the NanoVelcro system
CTC were detected by NanoVelcro system as we previously described . Blood specimens were collected in EDTA tubes. Blood samples were processed within 24 h. NH4Cl was added to whole blood at a ratio of 10:1 v/v and incubated for 20 min at room temperature to lyse the red blood cells. Samples were centrifuged at 200 g for 5 min, and the supernatants were removed. Cell pellets were re-suspended. Immunocytochemistry was applied to visualize cells captured on the SiNW substrate. The microchannels were loaded with 100 μl of fluorophore-labelled antibody solution (20 μl/1 ml of the initial concentration) and incubated at 4 °C overnight. CTCs were identified based on positive staining for cytokeratin (PE) and negative staining for CD45 (FITC). An experienced pathologist characterized the phenotypes and morphologies of tumour cells.
Enzyme-linked immunosorbent assay (ELISA)
Sera were collected from NSCLC patients or healthy controls. The concentrations of CTHRC1 in the sera were measured by ELISA as previous described .
All above experiments were performed at least three times. Statistical analysis was carried out using SPSS software (version 16.0; SPSS, Chicago, IL, USA). The χ2 test was applied to analyse the relationships between CTHRC1, MMP7, and MMP9 expression and clinicopathologic parameters. An unpaired, two-tailed Student’s t-test was used to determine the between- group significance. Bivariate correlation analysis was calculated as Spearman’s rank correlation coefficient. Survival curves were plotted using the Kaplan-Meier method and compared with the log-rank test. ROC curve analysis was carried out to determine the CTHRC1 cut-off points for metastasis and recurrence status. P values < 0.05 were considered significant.
CTHRC1 overexpression in NSCLC tissues correlates with clinical metastasis status in NSCLC
CTHRC1 overexpression is reportedly associated with tumour invasion and metastasis [20, 29]. To investigate whether NSCLC also exhibits strong CTHRC1 expression compared to ANTs, we first measured CTHRC1 expression in twenty paired NSCLC and ANT samples by performing western blotting. Compared to individual corresponding ANTs, CTHRC1 expression was significantly higher in NSCLC tissues (Fig. 1c), which was further confirmed at the RNA level by RT-PCR (Additional file 1: Figure S1C). Consistent with the western blot and RT-PCR results, IHC data further verified the upregulation of CTHRC1 in primary NSCLC tissues (Fig. 1d). Next, we investigated the correlation between CTHRC1 expression and NSCLC metastasis. Pathologically verified NSCLC tumour tissues and ANTs were collected. According to the IHC results, CTHRC1 was expressed at very low levels in normal lung tissue. In contrast, CTHRC1 expression was very high in 55.2% of primary NSCLC tissues, and furthermore, CTHRC1 expression in NSCLC tissues was associated with tumour metastasis (Additional file 1: Table S1). Additionally, compared to early stages, advanced stages characterized by localized invasion or distant metastasis had significantly higher levels of CTHRC1 (Fig. 1e).
CTHRC1 promotes NSCLC cell migration and invasion
CTHRC1 regulates MMP7 and MMP9 expression in vitro and in vivo
We next examined the CTHRC1-mediated regulation of MMP7 and MMP9 in in vitro experiments. Both MMP7 and MMP9 were upregulated when CTHRC1 was overexpressed in NSCLC cells. In contrast, CTHRC1 knockdown in NCI-H1975 and NCI-H2122 cells downregulated MMP7 and MMP9 expression at the protein level, as measured by western blotting (Fig. 3e-h).
NSCLC invasion and migration mediated by CTHRC1 are MMP7- and MMP9-dependent
CTHRC1 enhances MMP7 promoter activity through AP-1/c-Jun pathway
According to the PCR results obtained from the ChIP assay, the physical interaction site between CTHRC1 and the MMP7 gene may be located in region 2 (nucleotides − 176 to + 50) of the MMP7 promoter (Fig. 5e, f). Because CTHRC1 itself does not contain DNA-binding sites, thus, there may be other transcription factors that cooperate with CTHRC1 to enhance promoter activity. Next, the MMP7 promoter region was screened for transcriptional binding sites using prediction tools. A potential binding site in region 2, specifically nucleotides − 176 to + 50 within the MMP7 promoter, was identified as an activator protein (AP)-1–binding element (ABE), as indicated in Fig. 5e. Furthermore, silencing AP-1/c-Jun using siRNA significantly inhibited the binding efficiency of CTHRC1 to the MMP7 promoter (Fig. 5f), implying that AP-1 served as a “bridge protein” between CTHRC1 and MMP7 promoter. Moreover, according to our western blot and RT-PCR results, MMP7 expression was significantly increased in NCI-H1975-CTHRC1 cells compared to NCI-H1975 cells. However, the increased expression of MMP7 in NCI-H1975-CTHRC1 cells was abolished when c-Jun siRNA was introduced (Fig. 5g, h), and MMP7 concentrations in the culture supernatants, as measured by ELISA, significantly decreased when c-Jun was depleted (Additional file 1: Figure S7A, C). Taken together, these findings confirm the involvement of the AP-1 pathway in the CTHRC1-mediated regulation of MMP7.
CTHRC1 enhances MMP9 promoter activity through the NF-κB and AP-1 pathways
Increased expression of CTHRC1 was correlated with CTC and predicts progression and poor prognosis of NSCLC
Given the role of CTHRC1 in the tumor invasion and metastasis, we made further efforts to identify the clinical significance of CTHRC1 for prognosis prediction in NSCLC patients. Patients with high CTHRC1 expression had significant lower 5-year survival rate (2.4%) than those with low CTHRC1 expression (51.5%) (Fig. 7e). Moreover, early stage (stages I and II) and late stage (stages III and IV) patients with high CTHRC1 expression had shorter survival durations than those with low CTHRC1 expression (Fig. 7f-h). In the early and advanced stages, 5-year survival rates for low CTHRC1 expression were 55.1% and 21.4%, respectively. All the advanced stage patients with high CTHRC1 expression died within 5 years.
In a univariate analysis, CTHRC1 levels were correlated with overall survival (Additional file 1: Table S6). Further, a multivariate analysis performed using the COX proportional hazard regression model indicated CTHRC1 overexpression was an independent prognosis-related marker for NSCLC (Additional file 1: Table S7). We then evaluated how predictive prognostic power was related to the combined expression of CTHRC1, MMP7 and MMP9 in NSCLC patients. Compared to patients with a low expression pattern, patients with a high expression pattern for CTHRC1, MMP7 and MMP9 appeared to have poorer survival (Fig. 7i, j).
Cancer invasion and metastasis are among the biological hallmarks acquired during the multistep development of human tumours. Underlying these hallmarks is genome instability, which leads to the genetic diversity that eventually expedites their acquisition [30, 31]. In this report, CTHRC1 upregulation was associated with lymphatic metastasis, distant metastasis, and MMP7 and MMP9 overexpression in NSCLC patients, indicating CTHRC1 plays a critical role in promoting cancer invasiveness. Consistent with the above findings, we identified MMP7 and MMP9 as substrates of CTHRC1 and revealed an unknown function of CTHRC1 in promoting strong cell invasion in an MMP7- and MMP9-dependent manner through the transcriptional upregulation of MMP7 and MMP9 via interactions with their corresponding promoters.
The most common causes of cancer-associated mortality are the occurrence of local invasion or distant metastasis, rather than the presence of the primary tumours themselves. Among clinically diagnosed NSCLC patients, almost half have confirmed distant metastases . As a candidate tumour marker in this study, CTHRC1 was identified in NSCLC samples by performing quantitative assessments involving 2D-PAGE gels and mass spectrometry in comparison with adjacent non-tumour tissues. Furthermore, elevated preoperative serum CTHRC1 levels were associated with tumour metastasis. According to previous studies [20, 22, 29], and our own . CTHRC1 promotes cancer progression and activates relevant signalling molecules, which urged us to determine whether CTHRC1 plays a similar role in determining the aggressiveness of NSCLC. Thus, we evaluated CTHRC1 functions in lung cancer cell migration and invasion. Consistent with the published effects of CTHRC1 on the invasive phenotype of NSCLC cells, CTHRC1 knockdown greatly decreased cell invasion and inhibited cell migration, whereas endogenous CTHRC1 overexpression significantly increased invasive ability. Additionally, based on the IHC analysis of 230 clinical NSCLC specimens, CTHRC1 overexpression was significantly correlated with clinical stage, indicating CTHRC1 upregulation may facilitate NSCLC metastasis.
Tumour metastasis is initiated by a CTC sub-group that has been observed in patient blood. CTCs are used as a marker to predict disease progression in metastatic patients, including breast, colorectal and prostate cancers. They are involved in metastatic spread to distant organs, leading to the formation of secondary sites of lung cancer . Here, increased plasma levels of CTHRC1 were positively correlated with the presence of CTCs, suggesting plasma CTHRC1 attracts CTCs into circulation. However, the molecular mechanisms underlying tumour metastasis mediated by CTHRC1 require further investigation.
Proteolytic degradation of the stromal ECM is well known for its contribution to malignant invasion and metastasis [34, 35]. MMPs, as a family of zinc-dependent endopeptidases, are involved in degrading ECM and facilitating tumour invasion [11, 36, 37]. As shown in the present study, significantly higher expression of CTHRC1, MMP-7 and MMP-9 was observed in a cohort of NSCLC sera and surgically resected tumour tissues. CTHRC1 was positively correlated with MMP7 and MMP9 expression at both the protein and mRNA levels. MMP regulation, generally by hormones and cytokines, occurs primarily at the transcription level [38, 39]. MMPs are also regulated via the initiation of pro-MMP cleavage and proteolytic activity inhibition by specific inhibitors [40, 41, 42]. Our observations here indicate MMP7 and MMP9 may serve as the potential target genes of CTHRC1 in NSCLC, which may explain why CTHRC1 enhances NSCLC progression in vivo. However, it remains unclear whether the involvement of MMP7 and MMP9 in NSCLC progression is mechanistically regulated by CTHRC1.
CTHRC1, generally recognized as a secreted protein , was firstly confirmed in the nuclear localization of NSCLC cells. This builds the foundation for us to deeply explore the regulation mechanisms of CTHRC1 on MMP7 and MMP9. MMP7 and MMP9 expression is generally regulated via promoter binding sites for multiple transcription activators, such as AP-1 and β-catenin/TCF4 for MMP7 [43, 44] and AP-1, AP-2, NF-κB, and SP-1 for MMP9 [45, 46]. Based on previous findings from a luciferase-based ChIP assay, CTHRC1 binds to the MMP7 promoter (− 120 to + 50 bp) and MMP9 promoter (− 102 to + 31 bp and − 810 to − 510 bp). However, CTHRC1 does not contain DNA-binding sites .
Thus, CTHRC1 may cooperate with other transcription factors to bind to the promoters of downstream genes and initiate to their transcriptional activation. Sequence analysis revealed a potential binding site for AP-1 in the MMP7 promoter between nucleotide − 176 to + 50 bp. Two binding sites for NF-κB and AP-1 are also present in the MMP9 promoter between nucleotides − 164 to − 3 bp and − 690 to − 506 bp, respectively. Furthermore, AP-1/c-Jun depletion decreased the binding efficiency of CTHRC1 to the MMP7 promoter. Simultaneously, the knockdown of both NF-κB and AP-1/c-Jun suppressed CTHRC1 binding to the MMP9 promoter. In agreement with our hypothesis, c-Jun depletion downregulated MMP7 expression, and knockdown of either NF-κB or c-Jun decreased MMP9 expression. However, whether the invasion mediated by CTHRC1 depends on MMP7 and MMP9 remains unclear.
The metalloproteinases MMP7 and MMP9 are overexpressed in NSCLC and other types of cancers and are strongly associated with poor prognosis [16, 47, 48]. An inhibitor of MMPs, including MMP7 and MMP9, was shown to suppress tumour metastasis . Additionally, the application of MMP7 to colon cancer-bearing nude mice enhanced tumour metastasis  and MMP9-deficient mice were protected against tumour metastasis . Consistent with these data, the enhanced tumour cell migration and invasion mediated by CTHRC1 overexpression was eradicated when either MMP7 or MMP9 was knocked down. Based on these data, CTHRC1-mediated invasion is MMP7- and MMP9-dependent.
In conclusion, our current systematic study identified CTHRC1 as a invasion-driving gene that promotes NSCLC progression by activating c-Jun/MMP7, c-Jun/MMP9 and NF-κB/MMP9 signalling; thus, therapeutic targeting of CTHRC1 may be a promising strategy to enhance the therapeutic effects of anticancer drugs against NSCLC. Additionally, CTHRC1 may serve as a sensitive predictor of the low 5-year overall survival in NSCLC and as an effective biomarker for evaluating the poor clinicopathological characteristics of NSCLC.
CTHRC1 promotes NSCLC invasion by upregulating MMP7 and MMP9. Targeting CTHRC1 may be beneficial for inhibiting NSCLC progression.
This work was supported by grants from National Natural Science Foundation of China (30900650, 81372501, 81572260, 81773299, 81172232, 81570008, 81172564 and 31430030), Guangdong Natural Science Foundation (2011B031800025, S2012010008378, S2012010008270, S2013010015327, 2013B021800126, 20090171120070, 9451008901002146, 2013B021800126, 2014A030313052 and 2013B021800259), Guangdong Science and Technology Planning Program (2014 J4100132, 2015A020214010 and 2016A020215055) and Guangzhou Science and Technology Planning Program (201704020094, 2015ykzd07 and 16ykjc08). The funding body had no role in the design of the study and collection, analysis, and interpretation of data and in writing the manuscript.
Availability of data and materials
The datasets used and/or analyzed are available from the corresponding author on reasonable request.
ZK, WH, HZ, YW, YZ, XZ designed the study. YW, YL, YC, NJ, WJ, HW, ZW performed experiments. SL, YC, YS did the data analysis. HZ, WH, ZK, LW wrote the manuscript. All authors read and approved the final manuscript.
Ethics approval and consent to participate
This study was approved by the Institutional Review Board of the First Affiliated Hospital, Sun Yat-sen University, for the present retrieval of cancer specimens and associated clinical data from our institute (Reference number: 20150228). Consent forms were obtained prior to this study from every included patient.
All animal procedures were approved by the ethical committee of the First Affiliated Hospital, Sun Yat-sen University and performed in accordance with the guidelines provided by the National Institute of Health Guide for Care and Use of Animals.
Consent for publication
The authors declare that they have no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
- 21.Ma MZ, Zhuang C, Yang XM, Zhang ZZ, Ma H, Zhang WM, You H, Qin W, Gu J, Yang S, et al. CTHRC1 acts as a prognostic factor and promotes invasiveness of gastrointestinal stromal tumors by activating Wnt/PCP-rho signaling. Neoplasia. 2014;16(3):265–78. 278 e261-213CrossRefPubMedPubMedCentralGoogle Scholar
- 27.Sun Y, Chen YS, Li SH, Lei YY, Xu D, Jiang N, Zhang Y, Cao J Ke ZF. NanoVelcro-captured CTC number concomitant with enhanced serum levels of MMP7 and MMP9 enables accurate prediction of metastasis and poor prognosis in patients with lung adenocarcinoma. Int J Nanomedicine. 2017;12:6399–412.Google Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.