Differentially expressed microRNAs in bone marrow mesenchymal stem cell-derived microvesicles in young and older rats and their effect on tumor growth factor-β1-mediated epithelial-mesenchymal transition in HK2 cells
- 2.3k Downloads
The prevalence of renal fibrosis is higher in older than in younger individuals. Through paracrine activity, bone marrow mesenchymal stem cell-derived microvesicles (BM-MSC-MVs) influence the process of renal fibrosis. Differences in microRNA (miRNA) expression of BM-MSC-MVs that correlate with the age of the subjects and the correlation between miRNA expression and the process of renal fibrosis have not been established. The present study aimed to analyze differences in miRNA expression of BM-MSC-MVs between young or older rats and its influence on tumor growth factor-beta 1 (TGF-β1)-mediated epithelial-mesenchymal transition (EMT) of HK2 cells to explore the causes of renal fibrosis in aged tissues.
miRCURY LNA Array (version 18.0) was used to identify differentially expressed miRNAs in BM-MSC-MVs of 3- and 24-month-old Fisher344 rats. Reverse transcription-polymerase chain reaction was used to verify miRNA levels in BM-MSC-MVs and in the serum of rats. A TGF-β1-mediated EMT model was used to study the effects of BM-MSC-MVs and differentially expressed miRNAs on EMT.
BM-MSCs from older rats showed more severe aging phenotypes compared with those of young rats. In addition, the growth rate and cell migration of BM-MSCs derived from older rats were significantly reduced. In secreted BM-MSC-MVs, the expression of miR-344a, miR-133b-3p, miR-294, miR-423-3p, and miR-872-3p was significantly downregulated in older rats than in younger rats (P < 0.05), and the serum level of these miRNAs exhibited the same patterns. Intervention using BM-MSC-MVs resulted in the weakening of TGF-β1-mediated EMT in the aged rats. MiR-344a, miR-133b-3p, and miR-294 affected TGF-β1-mediated EMT in HK2 cells. Among these, miR-133b-3p and miR-294 significantly inhibited TGF-β1-mediated EMT in HK2 cells (P < 0.05).
In older rats, the inhibitory effect of BM-MSC-MVs on TGF-β1-mediated HK2 cell EMT was weaker than that observed in younger rats. In addition, miR-133b-3p and miR-294, which were downregulated in BM-MSC-MVs of older rats, remarkably inhibited TGF-β1-mediated EMT in HK2 cells, suggesting that these may play a role in the fibrosis of aging renal tissues.
KeywordsChronic Kidney Disease miRNA Expression Renal Fibrosis Beyotime Institute Renal Interstitial Fibrosis
Alpha-smooth muscle actin
Bone marrow mesenchymal stem cell
Bone marrow mesenchymal stem cell-derived microvesicle
Chronic kidney disease
Dulbecco’s modified Eagle’s medium/F12
Fetal bovine serum
Gene Expression Omnibus
Reverse transcription-polymerase chain reaction
Transforming growth factor-beta 1
Chronic kidney disease (CKD) affects older people at a higher rate compared with younger subjects. According to systematic reviews and analyses, the incidence of CKD has significantly increased with the improvement on the life span of the general population . Based on Chinese epidemiological surveys on CKD, the average age of patients with stage 3 CKD or higher is 63.6 years old. With an increase of 10 years in age, the risk of epidermal growth factor receptor (eGFR) is less than 60 ml/min per 1.73 m2, which represents the renal glomerular filtration rate, increasing by 1.74-fold . According to the US National Health and Nutrition Examination Survey, the incidences of phase 1 or 2 CKD were 2 %–3 % among people within the age of 20 to 39 years and 9 %–10 % among people above 70 years old, whereas those of phase 3 or 4 CKD were 0.2 %–0.7 % among people with the age range of 20 to 39 years old and 27.8 %–37.8 % among people above 70 years old . Tubulointerstitial fibrosis plays an important role in the course of CKD and is a typical characteristic of aged kidney tissues [2, 3]. Tumor growth factor-beta 1 (TGF-β1) is an important growth factor that induces epithelial-to-mesenchymal transition (EMT) and promotes tubulointerstitial fibrosis .
Bone marrow mesenchymal stem cells (BM-MSCs) are a pluripotent cell population that can differentiate into various cell types such as fat, bone, muscle, and skin cells as well as show significant therapeutic effects against many diseases [5, 6, 7]. It has been recently found that BM-MSCs change accordingly with the age of an individual [8, 9, 10]. With aging, BM-MSCs decrease in number [11, 12], and the functional factors they secrete may also change [12, 13, 14]. Microvesicles (MVs) are small-particle exosomes that are secreted by BM-MSCs and contain a number of bioactive substances such as proteins, lipids, mRNA, and microRNA (miRNA). These also have good therapeutic effects against acute and chronic diseases [15, 16, 17]. miRNAs are important genetic regulatory factors of BM-MSC-MVs that identify target mRNAs through base pairing. Moreover, based on the degree of complementarity, miRNAs can guide silencing complexes to degrade target mRNAs or downregulate the expression of mRNAs. Previous studies have shown that miRNAs are closely linked to the occurrence of diseases and the aging process [18, 19, 20]. However, information on differences in miRNA expression in BM-MSC-MVs between young and elder individuals and whether these differences can further promote renal fibrosis and aging of kidney is unknown. The present study aimed to analyze differences in miRNA expression of BM-MSC-MVs between young or elder rats as well as the effect of differentially expressed miRNAs on TGF-β1-mediated EMT of HK2 cells in relation to the pathogenesis of renal fibrosis in aged tissues.
Cell culture and identification of BM-MSCs
Animal welfare and experimental procedures were all carried out in accordance with the Guide for the Care and Use of Laboratory Animals (Ministry of Science and Technology of China, 2006) and were approved by the animal ethics committee of the General Hospital of the People’s Liberation Army. BM-MSCs were extracted from 3- and 24-month-old Fisher344 rats . The isolated primary BM-MSCs were cultured in complete medium (Cyagen Bioscience Inc., Santa Clara, CA, USA). After the cells were passaged to P2, BM-MSCs were identified by flow cytometry by using cell surface markers, CD44, CD105, CD146, and CD45 (BioLegend Inc., San Diego, CA, USA).
Proliferation, migration, and differentiation of BM-MSCs
BM-MSC proliferation assay
Both P2 BM-MSCs from young and elder rats were counted after trypsinization by using 0.25 % trypsin (Corning, Corning, NY, USA), and then approximately 10,000 cells were resuspended in 100 μl of media and seeded onto each well of a 96-well plate. After culturing for 4 days, each well was supplemented with 10 μl of a cck-8 solution (Dojindo Inst. Biotech, Shanghai, China). The plate was cultured for 3 h in an incubator. The absorbance at a wavelength of 450 nm was measured by using a microplate reader. Each group had five replicate experimental sets.
BM-MSC migration assay
Polycarbonate membrane transwell inserts (8 μm) were used (Corning). Before cell inoculation, upper transwell chambers were equilibrated for 30 min by using serum-free RPMI 1640 medium (Corning). Then in the uncoated upper chambers, 1 × 105 P2 BM-MSCs, which had been resuspended in serum-free RPMI 1640 medium, were inoculated. In the lower chambers, normal medium containing 10 % serum (Cyagen Bioscience Inc.) was added. After incubating for 24 h, cells that had not migrated across the polycarbonate membranes were gently wiped off. Then the cells on the underside of the membranes were fixed for 30 min with 4 % formaldehyde and stained with crystal violet (Beyotime Institute of Biotechnology, Jiangsu, China). Cells that migrated across the polycarbonate membrane were imaged and counted by using an inverted microscope.
BM-MSC osteogenic differentiation assay
When the cells were 60 %–70 % confluent, the complete medium was replaced with osteogenic and induction medium (Cyagen Bioscience Inc.). After incubation for 3–4 weeks, bone nodules were stained with Alizarin Red S.
BM-MSC adipocyte differentiation assay
When the cells were 70 %–80 % confluent, the complete medium was replaced with an adipogenic induction medium (Cyagen Bioscience Inc.). After incubation for 2 weeks, lipid droplets were stained by using Oil Red O.
Extraction and identification of MVs
Cells were starved overnight in serum-free medium supplemented with 0.5 % bovine serum albumin (Sigma-Aldrich, St. Louis, MO, USA) prior to collection of MVs. Cell debris was removed from the supernatant after centrifugation at 2000 × g for 20 min. The supernatant was collected and centrifuged at 100,000 × g at 4 °C for 1 h by using a high-speed refrigerated centrifuge (CP100WX; Hitachi, Tokyo, Japan). The pellet was resuspended in serum-free RPMI 1640 medium, which was again centrifuged at 100,000 × g at 4 °C for 1 h. The final pellet was considered to consist of MVs . The collected MVs were treated with phosphate-buffered saline (PBS) and fluorescein isothiocyanate (FITC) or phycoerythrin (PE)/CY7-labeled anti-CD44, CD29, and alpha 4-integrin antibodies (BioLegend Inc.). Murine IgG labeled with FITC or PE/CY7 (BioLegend Inc.) was used as a negative control. Flow cytometry was used to identify cell surface markers.
A senescence-associated beta-galactosidase (SA-β-Gal) staining kit was used (Beyotime Institute of Biotechnology). P2 BM-MSCs were seeded onto a six-well plate. When cells reached 70 % confluency, the medium was aspirated and the cells were washed twice with PBS. The cells were subsequently fixed with 4 % formaldehyde for 30 min and then stained for 16 h at 37 °C with an SA-β-Gal staining reagent. Positively stained (blue) cells were counted by using an inverted microscope, and the positive rates between the young and elder groups were compared.
Microarray analysis of miRNAs in BM-MSC-MVs
The technology of miRCURY LNA Array (version 18.0) (Exiqon, Vedbaek, Denmark) was adopted. RNA was extracted and purified from BM-MSC-MVs of both young and elder groups. With an Exiqon miRCURY Hy3/Hy5 Power Labeling Kit, miRNAs were fluorescently labeled and then hybridized in an miRCURY LNA Array Station (version 18.0). A GenePix 4000B microarray reader was used to measure chip fluorescence intensity. Then the fluorescence intensity was converted to raw numeric data by using GenePix pro version 6.0. Triplicates were set up for both young and elder groups. Signals with fluorescence intensities of 30 or above were selected for group analysis. The raw signals were normalized with the median fluorescence intensity of each chip. With the normalized signals, different expression levels of miRNAs between the young and elder groups were calculated. The statistical significance of the differences in miRNA expression between both groups was calculated by using the t test. miRNAs with 1.5-fold or above expression difference and P values of less than 0.05 were selected and defined as those showing significant differential expression. The microarray data were deposited in the National Center for Biotechnology Information Gene Expression Omnibus (GEO) public repository and are accessible under GEO Series accession number GSE72198.
Verification of mRNA differential expression in BM-MSC-MVs and sera
Blood from five 3-month-old and five 24-month-old Fisher344 rats was collected from the orbital sinus. After centrifugation at 3000 × g for 15 min, 100 μl of serum was collected from the supernatant. Based on methods earlier described, BM-MSCs were extracted and cultured, and the derived MVs were collected from the 10 rats. Total RNAs in sera and MVs were extracted by using an Exiqon miRCURY RNA Isolation Kit. The corresponding cDNAs were synthesized by using SYBR PrimeScript miRNA reverse transcription-polymerase chain reaction (RT-PCR) Kit (Takara, Tokyo, Japan). Lastly, an ABI-Prism 7500 Sequence Detection System (Applied Biosystems, Waltham, MA, USA) and SYBR PrimeScript miRNA RT-PCR Kit (Takara) were used to detect the expression level of specific miRNAs (miR-344a, miR-294, miR-872-3p, miR-133b-3p, and miR-423-3p) and the loading control, miR-191.
TGF-β1 stimulation and miRNA transfection
HK2 cells (American Type Culture Collection, Manassas, VA, USA) were cultured in 10 % fetal bovine serum (FBS) Dulbecco’s modified Eagle’s medium/F12 (DMEM/F12) medium (Corning). When the cells reached 50 % confluency, serum-free DMEM/F12 medium was used to synchronize the cells for 18 h. The HK2 cells were incubated with recombinant human TGF-β1 (PeproTech Inc., Rocky Hill, NJ, USA) at a concentration of 8 ng/ml for 48 h to induce fibrosis with young MV (105 young MSCs released overnight) or old MV (105 old MSCs released overnight) treatment. For the miRNA transfection groups, miR-344a, miR-294, miR-872-3p, miR-133b-3p, and miRNA mimic control (GeneCopoeia Inc., Rockville, MD, USA) were transfected into the synchronized HK2 cells in accordance with the instructions of the jetPRIME® transfection reagent (Polyplus Transfection Inc., USA). Approximately 24 h after transfection, the medium was replaced with 10 % FBS DMEM/F12 medium (Corning, Shanghai, China) containing 8 ng/ml of TGF-β1 (PeproTech Inc.) and then incubated for 48 h.
Western blot analysis
Rabbit monoclonal anti-E-cadherin (1:800 dilution; Biogot Biotechnology Co. Ltd., Nanjing, China)
Rabbit monoclonal anti-alpha-smooth muscle actin (anti-α-SMA) (1:500 dilution; Abcam, Cambridge, MA, USA)
Anti-β-actin (1:5,000 dilution; Beyotime Institute of Biotechnology).
After excess primary antibodies were washed off, the membranes were incubated with secondary antibodies (Beyotime Institute of Biotechnology) diluted in 1:1000 for 2 h at room temperature. The target band was detected chemiluminescently by using an enhanced chemiluminescence (ECL) Western Blotting kit (Applygen Technologies Inc., Beijing, China). While β-actin was used as internal control, the relative expression levels of E-cadherin and α-SMA were calculated in each experimental group.
HK2 cell suspensions were seeded onto autoclaved glass coverslips placed in six-well plates (105 cells per well). After overnight incubation at 37 °C in 5 % CO2, cells had adhered to the coverslips. After synchronization and transfection based on previous methods, 10 % FBS DMEM/F12 medium (Corning) with or without 8 ng/ml TGF-β1 (PeproTech Inc.) was used, and the cells were further incubated for 48 h. After fixing in 4 % paraformaldehyde (Beyotime Institute of Biotechnology) at room temperature for 15 min, the cells were permeabilized in 0.2 % Triton X-100 (Sigma-Aldrich) for 5 min and then blocked in 1× casein (Sigma-Aldrich) at room temperature for 30 min. Rabbit monoclonal anti-E-cadherin (1:100, Biogot Biotechnology Co. Ltd.) and anti-α-SMA (1:50; Abcam) primary antibodies were diluted and added to the coverslips and then incubated overnight at 4 °C. After washing with PBS, fluorescent anti-rabbit conjugated with CY3 or FITC secondary antibodies (1:400, Jackson ImmunoResearch Laboratories, Bar Harbor, ME, USA) were added and incubated for 2 h at room temperature. After washing with PBS, the coverslips were removed from the six-well plates and placed on a glass slide, and mounting medium containing 4ʹ,6-diamidino-2-phenylindole (DAPI) (ZSGB-BIO, Beijing, China) was added. Random fields were chosen under a fluorescence microscope (Olympus America Inc., Center Valley, PA, USA), and the expressions of E-cadherin and α-SMA in HK2 cells were observed.
Results were expressed as the mean ± standard deviation by using SPSS 17.0 software (IBM Corporation, Armonk, NY, USA). Differences between experimental groups were analyzed by using one-way analysis of variance. P values of less than 0.05 were considered statistically significant.
Identification and senescence staining of BM-MSCs
Aged BM-MSCs show decreased proliferation and migration
BM-MSCs of older rats showed significantly slower growth rate in vitro than that of younger rats (Additional file 3a). In the transwell migration assay, compared with BM-MSCs derived from younger rats, the migration capacity of BM-MSCs from older rats also remarkably decreased, and this was likely due to aging (Additional file 3b).
Change in miRNA expression profiles in BM-MSC-MVs caused by aging
Expression of miRNAs in blood
Compared with younger rats, the serum levels of miR-344a, miR-133b-3p, miR-294, and miR-872-3p in aged rats decreased, and those of miR-133b-3p, miR-294, and miR-872-3p showed a significant decrease (P < 0.05) (Fig. 2d).
Inhibition of TGF-β1-mediated EMT by BM-MSC-MV is weakened in older rats
Effects of miRNA downregulated in elder rat BM-MSC-MVs on TGF-β1-mediated EMT
HK2 cells were initially transfected with miR-344a, miR-294, miR-872 -3p, miR-133b-3p, and miRNA mimic control and then incubated in a medium containing 8 ng/ml TGF-β1 for 48 h. Compared with non-transfected HK2 cells, the expression level of α-SMA significantly decreased in HK2 cells transfected with miR-344a, miR-294, and miR-133b-3p (P < 0.05), and the expression level of E-cadherin significantly increased in HK2 cells transfected with miR-294 and miR-133b-3p (P < 0.05). Meanwhile, HK2 cells transfected with miR-872-3p exhibited no significant changes (Fig. 3c, d). We also validated these results by immunofluorescence staining, which showed that HK2 cells without TGF-β1 stimulation displayed high expression of E-cadherin and low expression of α-SMA. After TGF-β1 stimulation, HK2 cells presented lower expression of E-cadherin and higher expression of α-SMA. Moreover, miR-344a, miR-294, and miR-133b-3p weakened the upregulation of α-SMA and inhibited the downregulation of E-cadherin in HK2 cells induced by TGF-β1 (Figs. 4e-i and 5e-i).
Renal interstitial fibrosis is a characteristic pathology of aging kidneys and the basis of CKD progression. EMT of renal tubules is the leading cause of renal interstitial fibrosis. Several studies have shown that TGF-β1 enhances EMT through a variety of mechanisms [21, 22, 23]. In addition, various studies have demonstrated that BM-MSCs participate in renal tissue repair. Morigi et al.  showed that BM-MSCs can improve kidney functions, and BM-MSCs are the primary cell population involved in renal tubular and functional repair. Other recent investigations have also shown that BM-MSCs can reduce renal damage through paracrine activity [25, 26, 27].
MVs are small circular diaphragms (diameter of 100 nm to 1 μm) shed from the cell surface or released from cellular compartments. MVs contain proteins, enzymes, mRNA, and miRNA and may participate in tissue damage repair through paracrine activity [15, 17, 27]. Camussi et al.  revealed that both BM-MSCs and BM-MSC-MVs promote morphological restoration and renal function maintenance in animals with renal ischemia-reperfusion or unilateral nephrectomy. MVs secreted by BM-MSCs can enter the damaged kidney cells and play a role in kidney repair. Cantaluppi et al.  also observed that MVs in rats possessed a repair function for renal ischemia-reperfusion injury. However, pre-treatment with an efficient miRNA blocker resulted in a reduction in repair function, suggesting that MVs can transfer miRNAs to repair damaged kidneys.
In eukaryotes, miRNAs play an extremely important role in organ development and maturation as well as the pathogenesis and progression of disease . Previous studies have shown that in fibrotic renal tissues the expressions of both miR-200 family and E-cadherin are significantly downregulated and that exogenous miR-200 could reduce renal fibrosis [30, 31]. Pre-transfection of the miR-200 family can reverse TGF-β1-mediated EMT in tubular epithelial cells [30, 32, 33]. Through the regulation of Tgfbr2 and Gsk3-beta, miR-294/302 can inhibit TGF-beta and GSK3 pathways. On the other hand, miR-294/302 can indirectly inhibit Tgfbr2 as well as EMT of embryonic stem cells . Liu et al. discovered that, as a downstream gene of fibrotic TGF-β/Smad3 pathway, miR-133 can negatively regulate TGF-β/Smad3 . Along with physiological aging, BM-MSCs also age, and the cellular molecules these secrete significantly decrease or even disrupt secretion. Expression profiles of miRNAs can also significantly change with age, and its functions against diseases can decrease or become ineffective [36, 37].
In the present study, we found that the number of BM-MSCs derived from 24-month-old rats was significantly lower, displayed a senescent phenotype, and had a markedly slower in vitro proliferation rate compared with that observed in 3-month-old rats. Furthermore, these aged BM-MSCs showed irregular growth with increased cellular spacing, and their migration ability was significantly weaker. We also observed significant differences in miRNA expression of BM-MSC-MVs between younger and older rats. Compared with younger rats, five miRNAs in BM-MSC-MVs—miR-344a, miR-133b-3p, miR-294, miR-423-3p, and miR-872-3p—were significantly reduced in older rats. Additionally, compared with MVs from younger rats, the inhibitory effects of MVs from older rats on TGF-β1-induced EMTs were significantly reduced. We also observed that miR-133b-3p and miR-294, which were downregulated in older rats, imparted significant inhibitory effects on TGF-β1-induced EMT. Our data therefore suggested that the downregulation of miR-133b-3p and miR-294 in older rats may be an important cause of suppression of TGF-β1-induced EMT in aged rats.
In recent years, studies have shown that changes in body environment affect aging of organs. Through the establishment of Siamese animal models such as young and older rats, we observed that learning and memorization were enhanced in older rats but that these were weakened in younger rats. With injection of serum from older rats, younger rats presented brain aging phenotypes. Conversely, injection of serum from younger rats resulted in the alleviation of the brain aging phenotype in older rats . In the young-elderly rat model, cardiac hypertrophy of aged hearts was significantly less than that of the elder-elder rat model, and cardiac function significantly improved . In addition, in the young-elderly rat model, aged muscles, liver, neural stem/progenitor cells, and ovarian follicles were activated, and the regeneration capacity of tissues and organs was enhanced [40, 41, 42]. The miRNAs contained in BM-MSC-MVs can be released into blood circulation, where they are converted into circulating miRNAs that possess the ability to change the internal environment, and ultimately impart systemic effects. Our current experimental results showed that miRNAs that were downregulated in MVs from older rats also exhibited a comparatively low level in serum, suggesting that by secreting MVs and changing miRNA expression, BM-MSCs changed the body environment during the aging process. In older rats, the level of secreted miR-133b-3p and miR-294 decreased. Their levels in serum were also low and this may be an important factor in causing renal EMT and renal aging.
In the present study, we confirmed that, in older rats, the inhibitory effects of BM-MSC-MVs on TGF-β1-induced EMT had weakened, and this might be related to the decrease in the levels of miR-133b-3p and miR-294. However, the target genes, regulatory mechanisms, and signaling pathways of miR-133b-3p and miR-294 currently remain unclear and should be investigated in future studies.
The present study observed that the inhibitory effect of BM-MSC-MVs on TGF-β1-mediated HK2 cell EMT was weaker in older rats than in younger rats. Moreover, miR-133b-3p and miR-294, which were downregulated in BM-MSC-MVs and serum of older rats, could remarkably inhibit TGF-β1-mediated EMT in HK2 cells. These findings suggest that the downregulated miR-133b-3p and miR-294 in older rats play important roles in causing renal EMT and renal aging.
This work was supported by the grants from the Major State Basic Research Development Program of China (2013CB530800), National Natural Science Foundation of China (81270819), National Key Technology R&D Program (2011BAI10B00). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
- 28.Cantaluppi V, Gatti S, Medica D, Figliolini F, Bruno S, Deregibus MC, et al. Microvesicles derived from endothelial progenitor cells protect the kidney from ischemia–reperfusion injury by microRNA-dependent reprogramming of resident renal cells. Kidney Int. 2012;82:412–27.CrossRefPubMedGoogle Scholar
- 34.Guo WT, Wang XW, Yan YL, Li YP, Yin X, Zhang Q, et al. Suppression of epithelial-mesenchymal transition and apoptotic pathways by miR-294/302 family synergistically blocks let-7-induced silencing of self-renewal in embryonic stem cells. Cell Death Differ. 2014;12:1038–50.Google Scholar
- 35.Duan LJ, Qi J, Kong XJ, Huang T, Qian XQ, Xu D, et al. MiR-133 modulates TGF-β1-induced bladder smooth muscle cell hypertrophic and fibrotic response: implication for a role of microRNA in bladder wall remodeling caused by bladder outlet obstruction. Cell Signal. 2015;27:215–27.CrossRefPubMedGoogle Scholar
- 37.Pandey AC, Semon JA, Kaushal D, O'Sullivan RP, Glowacki J, Gimble JM, et al. MicroRNA profiling reveals age-dependent differential expression of nuclear factor κB and mitogen-activated protein kinase in adipose and bone marrow-derived human mesenchymal stem cells. Stem Cell Res Ther. 2011;2:49–60.PubMedCentralCrossRefPubMedGoogle 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.