TSG-6 released from intraperitoneally injected canine adipose tissue-derived mesenchymal stem cells ameliorate inflammatory bowel disease by inducing M2 macrophage switch in mice
- 688 Downloads
Inflammatory bowel disease (IBD) is an intractable autoimmune disorder that markedly deteriorates one’s quality of life. Mesenchymal stem cells (MSCs) alleviate inflammation by modulating inflammatory cytokines in inflamed tissues, and have been suggested as a promising alternative for IBD treatment in human and veterinary cases. Furthermore, tumor necrosis factor-α-induced gene/protein 6 (TSG-6) is a key factor influencing MSC immunomodulatory properties; however, the precise mechanism of TSG-6 release from canine MSCs in IBD remains unclear. This study aimed to assess the therapeutic effects of canine adipose tissue-derived (cAT)-MSC-produced TSG-6 in an IBD mouse model and to explore the mechanisms underlying the immunomodulatory properties.
Mice with dextran sulfate sodium-induced colitis were administered cAT-MSCs intraperitoneally; colon tissues were collected on day 10 for histopathological, quantitative real-time polymerase chain reaction, and immunofluorescence analyses.
cAT-MSC-secreted TSG-6 ameliorated IBD and regulated colonic expression of pro- and anti-inflammatory cytokines such as tumor necrosis factor-α, interleukin-6, and interleukin-10. To investigate the effect of cAT-MSC-secreted TSG-6 on activated macrophages in vitro, a transwell coculture system was used; TSG-6 released by cAT-MSCs induced a macrophage phenotypic switch from M1 to M2. The cAT-MSC-secreted TSG-6 increased M2 macrophages in the inflamed colon in vivo.
TSG-6 released from cAT-MSCs can alleviate dextran sulfate sodium-induced colitis by inducing a macrophage phenotypic switch to M2 in mice.
KeywordsTSG-6 Mesenchymal stem cells Inflammatory bowel disease Canine Cell therapy Immunomodulation M2 macrophage
Canine adipose tissue-derived mesenchymal stem cell
Canine peripheral blood mononuclear cell
Dextran sulfate sodium
Enzyme-linked immunosorbent assay
Glyceraldehyde 3-phosphate dehydrogenase
Institutional Animal Care and Use Committee
Inflammatory bowel disease
Inducible nitric oxide synthase
Mesenchymal stem cell
Quantitative reverse transcription polymerase chain reaction
Small interfering ribonucleic acid
Seoul National University
Tumor necrosis factor
Tumor necrosis factor-α-induced gene/protein-6
Inflammatory bowel disease (IBD) is an intractable autoimmune disease that leads to abdominal pain, diarrhea, fever, or other symptoms that may be caused by chronic inflammation of the digestive system. Depending on the pattern and site of inflammation, IBD can be categorized as either Crohn’s disease or ulcerative colitis . Although the exact underlying pathogenesis of IBD is unknown, it is thought to be associated with genetic and environmental factors and gut flora [2, 3]. In addition, the disease occurs naturally in dogs by a similar pathogenesis, and data from therapeutic trials for canine IBD may be excellent references for human IBD . Although IBD leads to a decreased quality of life in both humans and dogs, no effective treatments for IBD have been developed.
Macrophages are important immune cells related to inflammatory diseases and release inflammatory cytokines which are associated with acquired immune cells, such as lymphocytes . In inflamed tissues, macrophages can be classified into two subtypes: M1 and M2 macrophages [6, 7]. M1 macrophages induce inflammatory responses by secreting cytokines such as tumor necrosis factor (TNF)-α, interleukin (IL)-1β, and IL-6, whereas M2 macrophages exert anti-inflammatory responses by releasing anti-inflammatory cytokines such as IL-10 . According to several recent studies, M1 macrophages are present predominantly in inflamed tissues of inflammatory disease model animals; however, the percentage of M2 macrophages increased markedly when the model animals recovered.
Mesenchymal stem cells (MSCs) have been suggested as a promising tool for treating various inflammatory diseases, including pancreatitis, peritonitis, rheumatoid arthritis, and atopic dermatitis, as well as IBD [9, 10, 11, 12, 13]. In addition, the mechanisms underlying the therapeutic effects have been investigated and several studies have shown that MSCs exert anti-inflammatory effects through secretory factors . TNF-α-induced gene/protein 6 (TSG-6) is a well-known secretory factor responsible for immunomodulation, and several recent studies have shown that it plays important roles in reducing inflammatory responses in lung injury, corneal injury, skin wound, peritonitis, pancreatitis, and IBD [15, 16, 17, 18, 19, 20]. Moreover, MSCs derived from dogs, cats, or horses have been shown to have immunomodulatory effects on activated immune cells and cell-based therapy using MSCs is a potential treatment for intractable inflammatory diseases in veterinary medicine [21, 22, 23].
In canine medicine, however, few studies have characterized secretory factors from canine MSCs, and cross-talk mechanisms between canine MSCs and immune cells are not well understood. Evaluating the efficacy of canine MSCs and accumulating therapeutic results from canine medicine would be useful in human medicine as well as veterinary medicine, particularly for intractable inflammatory diseases such as IBD. Therefore, in this study, we assessed the anti-inflammatory effects and mechanisms of canine adipose tissue-derived (cAT)-MSCs in a dextran sulfate sodium (DSS)-induced colitis model.
Isolation and characterization of cAT-MSCs
Canine adipose tissues were collected using a protocol approved by the Institutional Animal Care and Use Committee (IACUC) of Seoul National University (SNU; protocol no. SNU-170724-6). MSCs were isolated from the tissues and cultured as described in Additional file 1: Supplementary materials. Cells were characterized for the expression of several stem cell markers by flow cytometry before they were used in the experiments (Additional file 2: Figure S1). Additionally, the differentiation ability of the cells was confirmed (Additional file 3: Figure S2), and isolated cAT-MSCs at passage 3–4 were used in the following experiments.
Small interfering RNA (siRNA) transfection of cAT-MSCs
When cAT-MSCs reached approximately 70% confluence they were transfected with TSG-6 siRNA or control siRNA (sc-39,819 and sc-37,007, respectively; Santa Cruz Biotechnology, Santa Cruz, CA, USA) for 24 h using Lipofectamine RNAiMAX (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s instructions. TSG-6 knockdown was confirmed by quantitative reverse transcription polymerase chain reaction (qRT-PCR) (Additional file 4: Figure S3). cAT-MSCs transfected with siRNA were used for further experiments immediately after the transfection protocol was completed.
Male C57BL/6 J mice aged 6 weeks were purchased from Nara Biotech (Seoul, Korea) and housed under controlled temperature, humidity, and light cycle conditions. Only male mice were used in this study to simplify interpretation of the results by avoiding the effects of gender-related differences. All experimental procedures involving animals were approved by the IACUC of SNU (protocol no. SNU-170804-2), and the protocols were performed in accordance with approved guidelines (n = 4 for the naive group, n = 6 for the other groups). Colitis was induced by ad libitum administration of 3% DSS (36–50 kDa; MP Biomedical, Solon, OH, USA) in the drinking water from day 0 to day 7, whereas mice receiving normal drinking water were used as the naive group. On day 1 the following experiments were performed: 2 × 106 cAT-MSCs transfected with TSG-6 siRNA in 200 μL phosphate-buffered saline (PBS); 2 × 106 cAT-MSCs transfected with scrambled siRNA control in 200 μL PBS; 2 × 106 control cAT-MSCs in 200 μL PBS; or the identical volume of PBS was injected intraperitoneally into the colitis mice. The body weight of each mouse was assessed every 24 h. The mice were sacrificed on day 10 and colon tissues were collected for further processing.
Evaluating colitis severity
The disease activity index was determined by scoring the body weight loss (grades 0–4: 0, none; 1, < 5% loss of the initial body weight; 2, 5–10% loss of the initial body weight; 3, 10–20% loss of the initial body weight; 4, > 20% loss of the initial body weight), stool consistency (grades 0–2: 0, none; 1, mild to moderate diarrhea; 2, severe diarrhea), rectal bleeding (grades 0–2: 0, none; 1, mild to moderate bleeding; 2, severe bleeding), and general activity (grades 0–2: 0, normal; 1, mildly to moderately depressed; 2, severely depressed).
Colon tissues were fixed in 10% formaldehyde for 48 h, embedded in paraffin, and cut into 4-μm sections. The sections were stained with hematoxylin and eosin. A total of 20 fields per group was selected randomly and histological examinations were performed in a blinded manner. The severity of symptoms was calculated by scoring the extent of bowel wall thickening (grades 0–3: 0, none; 1, mucosa; 2, mucosa and submucosa; 3, transmural), damage to the crypt (grades 0–3: 0, none; 1, loss of goblet cells; 2, only surface epithelium intact; 3, loss of entire crypt and epithelium), and infiltration of inflammatory cells (grades 0–2: 0, none; 1, mild to moderate; 2, severe).
Enzyme-linked immunosorbent assay (ELISA)
Total proteins were extracted from the colon tissue using PRO-PREP Protein Extraction Solution (Intron Biotechnology, Seongnam, Korea) according to the manufacturer’s instructions and stored at −80 °C until use. The concentrations of TNF-α, IL-6, and IL-10 were measured using a commercial ELISA kit (all from eBiosciences, San Diego, CA, USA) according to the manufacturer’s instructions.
Obtaining canine peripheral blood mononuclear cell (cPBMC)-derived macrophages
Canine macrophages were obtained from the peripheral blood as previously described . Briefly, the blood of healthy canine donors was obtained from the SNU Veterinary Medical Teaching Hospital and PBMCs were isolated using Ficoll-Paque PLUS (GE Healthcare Life Sciences, Little Chalfont, UK). cPBMCs were resuspended in Roswell Park Memorial Institute (RPMI)-1640 medium (PAN Biotech, Aidenbach, Germany) containing 20% fetal bovine serum (FBS; PAN Biotech) and 20% macrophage colony-stimulating factor medium obtained from the supernatant of L929 immortalized cells. The cPBMCs were plated at 2 × 106 cells/well in 24-well plates and incubated at 37 °C in a humidified atmosphere of 5% CO2. After 24 h, the wells were washed to remove nonadherent cells. The remaining adherent cells were incubated with fresh medium for 5 days for differentiation into macrophages.
Coculture of cPBMC-derived macrophages with cAT-MSCs
After cPBMC-derived macrophages were stimulated with 200 ng/ml lipopolysaccharide (LPS; Sigma-Aldrich, St. Louis, MO, USA) for 24 h, the LPS-stimulated macrophages were plated at a density of 2 × 105 cells per well in 24-well plates. Subsequently, 2 × 104 cAT-MSCs, control siRNA-cAT-MSCs, or TSG-6 siRNA-cAT-MSCs were seeded onto 0.4-μm pore-sized Transwell inserts (SPL Life Science, Pocheon, Korea) and incubated for 48 h. The macrophages were harvested for further experiments.
RNA extraction, cDNA synthesis, and qRT-PCR
Total RNA was extracted from homogenized colon tissue or cPBMC-derived macrophages using the Easy-BLUE Total RNA Extraction kit (Intron Biotechnology) according to the manufacturer’s instructions. cDNA was synthesized using LaboPass M-MuLV Reverse Transcriptase (Cosmo Genetech, Seoul, Korea) and the samples were analyzed using 10 μL AMPIGENE qPCR Green Mix Hi-ROX with SYBR Green dye (Enzo Life Sciences, Farmingdale, NY, USA) and 400 nM forward and reverse primers (Cosmo Genetech). Expression levels of the target genes were normalized to that of glyceraldehyde 3-phosphate dehydrogenase (GAPDH). Primer sequences used in the present study are listed in Additional file 5: Table S1.
Flow cytometric analysis
Flow cytometry was conducted using a FACSAria II system (BD Biosciences, Franklin Lakes, NJ, USA) and analyzed using FlowJo software (Tree Star, Ashland, OR, USA). To characterize MSCs derived from canine adipose tissues, the cells were harvested and resuspended in PBS. Subsequently, the cells were stained with fluorescein isothiocyanate (FITC)-, phycoerythrin (PE)-, or allophycocyanin (APC)-conjugated antibodies against the following proteins: CD29-FITC, CD34-PE, and CD73-PE (BD Biosciences); and CD44-FITC, CD45-FITC, and CD90-APC (eBiosciences). To evaluate M2 macrophage polarization, PBMC-derived macrophages cocultured with cAT-MSCs were detached and resuspended in PBS. Next, the macrophages were stained with PE-conjugated CD11b (Abcam, Cambridge, UK) and FITC-conjugated CD206 (Santa Cruz Biotechnology).
Paraffin-embedded colon tissue sections were cut at a thickness of 4 μm. Sections were deparaffinized in xylene and rehydrated sequentially in 100%, 95%, and 80% ethanol solutions, and antigen retrieval was carried out using 10 mM citrate buffer (Sigma-Aldrich). After the sections were washed, they were blocked with blocking buffer containing 5% bovine serum albumin (BSA) and 0.3% Triton X-100 (both from Sigma-Aldrich) for 1 h. The sections were incubated overnight at 4 °C with antibodies against F4/80 (1:250) or FITC-conjugated CD206 (1:250; both from Santa Cruz Biotechnology). After three washes, the slides incubated with F4/80 antibody were incubated with PE-conjugated secondary antibody (1:500; Santa Cruz Biotechnology) for 1 h at 20 °C in the dark. The colon sections stained with antibody against either F4/80 or CD206 were washed three times and mounted in Vectashield mounting medium containing 4′,6-diamidino-2-phenylindole (DAPI; Vector Laboratories, Burlingame, CA, USA). The slides were visualized with a confocal laser scanning microscope (LSM710; Carl Zeiss, Jena, Germany), and immunoreactive cells were counted in 20 random fields per group.
Annexin-V and propidium iodide (PI) staining
Colon tissue slides were stained with FITC-conjugated annexin-V and PI using the Annexin V-FITC apoptosis detection kit plus (Enzo Life Sciences) according to the manufacturer’s instructions. The slides were observed using an EVOS FL microscope (Life Technologies, Carlsbad, CA, USA). Apoptotic cells identified as FITC-positive cells were counted in 20 random fields per group.
Generation of the GAPDH standard curve
Standard curves for evaluating the migratory ability of intraperitoneally injected canine MSCs were generated by administering serial dilutions of cAT-MSCs to mouse organs as described previously . Briefly, 2 × 102, 2 × 103, 2 × 104, or 2 × 105 cAT-MSCs were added to whole mouse organs prior to homogenization. Total RNA was extracted from the samples using the Easy-BLUE Total RNA Extraction kit (Intron Biotechnology), and cDNA was synthesized (LaboPass M-MuLV Reverse Transcriptase; Cosmo Genetech) using 1 μg of RNA. Next, qRT-PCR using canine-specific mitochondrial cytochrome b primers (forward primer, 5′-CCT TAC TAG GAG TAT GCT TG-3′; reverse primer, 5′-TGG GTG ACT GAT GAA AAA G-3′) was performed to generate the standard curves. The curves were corrected by performing parallel qRT-PCR with primers for universal eukaryotic 18S ribosomal RNA (forward primer, 5′-GCT ACT ACC GAT TGG ATG GTT TAG-3′; reverse primer, 5′-CTA CGG AAA CCT TGT TAC GAC TTT-3′).
Data are shown as the mean ± standard deviation. Mean values among different groups were compared by one-way analysis of variance using the GraphPad Prism v.6.01 software (GraphPad, Inc., La Jolla, CA, USA). A P value < 0.05 was considered statistically significant.
Intraperitoneally administered cAT-MSC-secreted TSG-6 plays a crucial role in ameliorating IBD
cAT-MSC-secreted TSG-6 reduced inflammatory response and apoptosis in the colon
In addition, we examined apoptosis that could be induced by TNF-α in colon sections. Annexin-V-positive cells were increased markedly in the colons of DSS-induced mice (Fig. 2b). Administration of cAT-MSCs led to a significant decrease in annexin-V-positive cells in the colon sections compared with the PBS-treated groups, which was abolished by knockdown of TSG-6 with siRNA transfection (Fig. 2b).
Intraperitoneally infused cAT-MSCs did not migrate to the inflamed colon
TSG-6 produced by cAT-MSCs induced phenotypic switching from M1 to M2 macrophages in vitro
cAT-MSC-secreted TSG-6 increased M2 macrophages in the inflamed colon
Numerous studies have shown that MSC administration may be an important treatment option for IBD [26, 27, 28, 29, 30, 31]. Although they are not fully understood, the mechanisms underlying the anti-inflammatory effects of MSC have been described previously [32, 33, 34]. It is well known that the anti-inflammatory ability and mechanisms of MSCs vary depending on their source . Moreover, inflammatory cytokines released by activated immune cells are important triggers for MSCs to exert immunomodulatory properties, indicating that the disease-specific inflammatory microenvironment is crucial for the therapeutic effects of administered MSCs . In this study, we first demonstrated the anti-inflammatory effects of cAT-MSCs and determined the underlying mechanisms in a DSS-induced colitis mouse model.
Recent studies reported that MSCs reduce inflammation through soluble factors such as indoleamine 2,3-dioxygenase, transforming growth factor-β, prostaglandin E2 (PGE2), hepatocyte growth factor, and TSG-6 [12, 32, 33, 37, 38]. Among these, TSG-6 has been shown to be pivotal for the immunomodulatory effects of MSCs in several inflammatory disease models such as corneal inflammation, wound injury, acute lung injury, peritonitis, and pancreatitis [10, 16, 17, 20, 39]. We also previously demonstrated that MSCs derived from human adipose tissue exert therapeutic effects against DSS-induced colitis by secreting TSG-6 . Here, we showed that TSG-6 secreted from intraperitoneally infused cAT-MSCs may ameliorate the symptoms of DSS-induced colitis and that weight loss and disease activity indices were reduced. In addition, by evaluating the length of the colon and assessing the histologic scores of the colon tissue sections, we demonstrated the therapeutic effects of cAT-MSC-secreted TSG-6 against DSS-induced colitis mice. Moreover, TSG-6 released from cAT-MSCs played an important role in modulating inflammatory cytokines such as TNF-α, IL-6, and IL-10 in the colon; reduced TNF-α secretion led to a significant decrease in annexin V-positive apoptotic cells in colon tissue sections.
In the present study, we administered cAT-MSCs into an immunocompetent mouse model of IBD. We focused on alterations in mouse immune cells after treatment with canine MSCs. Moreover, MSCs are immunoprivileged, partly because of the low expression of major histocompatibility complex class II molecules . Similar approaches involving the injection of xenogeneic MSCs into immunocompetent mouse models have been used by several groups, and cross-species-induced immunological responses have not been reported [17, 21, 39, 41]. Additionally, in this study, none of the mice injected with cAT-MSCs showed any side effects or died until sacrifice. In addition, Wang et al. reported that intraperitoneal infusion of MSCs showed better amelioration of DSS-induced colitis compared with local anal injection, suggesting that systemic immunomodulation rather than reducing local inflammation is required for effective therapy . Therefore, we determined that the optimal conditions were 2 × 106 cAT-MSCs infused intraperitoneally.
We also evaluated the distribution of intraperitoneally administered cAT-MSCs using qRT-PCR, which is known to have relatively higher sensitivity and specificity compared to fluorescence-mediated cell tracking. Two hours after cAT-MSC infusion less than 1% of the infused cells were detected in the heart, lung, liver, spleen, kidney, and colon tissues. At 1 and 3 days after cAT-MSC administration, the percentages of cells detected in the tissues were less than 0.5%. Furthermore, infused cAT-MSCs were not detected in colon tissues, despite inflammatory responses observed at days 1 and 3. Similar results were obtained when cAT-MSCs transiently transfected with control or TSG-6 siRNA were administered. These results are consistent with those of our previous study, indicating that intraperitoneally injected cAT-MSCs formed aggregates in the peritoneal cavity and alleviated DSS-induced colitis at sites distant from the colon through soluble factors, such as TSG-6 .
Considering that inflammatory cytokines modulated in the inflamed colon treated with cAT-MSCs were principally derived from macrophages, we carried out in vitro experiments to evaluate whether TSG-6 secreted from cAT-MSCs could switch the macrophage phenotype from M1 to M2. LPS-stimulated cPBMC-derived macrophages exhibit a conventional M1 type pattern and were cocultured with cAT-MSCs transfected with siRNA or with naive cAT-MSCs in a transwell system. CD206-expressing M2 macrophages were increased markedly in the cAT-MSC group but were inhibited in the siTSG6-cAT-MSC group. Moreover, TSG-6 secreted from cAT-MSCs contributed to the decrease in the expression levels of M1 markers such as iNOS and IL-6.
Next, we evaluated M2 macrophages in colon tissue sections of DSS-induced colitis mice treated with cAT-MSCs to analyze the macrophage polarization ability of TSG-6 released by cAT-MSCs in vivo. Recent studies suggested that MSCs have the capacity to induce phenotypic alterations in macrophages in acute kidney injury, spinal cord injury, and skin wound animal models [43, 44, 45]. Consistent with these reports, the expression levels of M2 markers in the colon tissue were increased in the siCTL-cAT-MSC-treated group and naive cAT-MSC-treated group compared with the PBS-treated group. However, the siTSG6-cAT-MSC-treated group showed no significant changes in the expression levels of these markers compared with the PBS group. Taken together, we demonstrated that TSG-6 secreted by cAT-MSCs plays an essential role in switching the phenotype of macrophages from M1 to M2 in the inflamed colon.
We could not rule out the possibility that other factors secreted from cAT-MSCs contributed to M2 macrophage polarization in colitis mice. Interestingly, other groups recently showed that PGE2 released by human MSCs induced macrophage phenotypic alteration [46, 47]. Further experiments on other factors secreted from cAT-MSCs, such as PGE2, are required to verify their effects on macrophages in IBD models. However, our findings suggest that TSG-6 released by cAT-MSCs plays an important role in switching the macrophage phenotype from M1 to M2 in vitro and in DSS-induced colitis mice.
In conclusion, we demonstrated that TSG-6 secreted by cAT-MSCs ameliorated DSS-induced colitis in mice by inducing macrophages to switch to the M2 phenotype. In addition, our findings reveal a possible mechanism underlying the TSG6-induced therapeutic effects in various inflammatory disease models previously published. These results may help in the development of cell-based therapies for treating IBD.
We are very grateful to the Research Institute for Veterinary Science and the BK21 PLUS Program for Creative Veterinary Science Research.
This study was supported by the Research Institute for Veterinary Science and the BK21 PLUS Program for Creative Veterinary Science Research.
Availability of data and materials
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.
W-JS designed the study, performed the experiments, analyzed the data, and wrote the manuscript; QL contributed to the planning, and collected and analyzed the data; M-OR and J-OA collected and analyzed the data; DHB analyzed the data and edited the manuscript; YCJ collected and analyzed the data; H-YY designed and supervised the study. All authors read and approved the final manuscript.
Canine adipose tissue was obtained from a 4-year-old beagle dog and the owner provided informed, written consent for its use in research. The protocol was performed in accordance with the approved guidelines by the Institutional Animal Care and Use Committee (IACUC) of Seoul National University.
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.
- 24.Viana KF, Aguiar-Soares RD, Roatt BM, Resende LA, da Silveira-Lemos D, Correa-Oliveira R, Martins-Filho OA, Moura SL, Zanini MS, Araujo MS, Reis AB, Giunchetti RC. Analysis using canine peripheral blood for establishing in vitro conditions for monocyte differentiation into macrophages for Leishmania chagasi infection and T-cell subset purification. Vet Parasitol. 2013;198:62–71.CrossRefPubMedGoogle Scholar
- 32.Gonzalo-Gil E, Perez-Lorenzo MJ, Galindo M, Diaz de la Guardia R, Lopez-Millan B, Bueno C, Menendez P, et al. Human embryonic stem cell-derived mesenchymal stromal cells ameliorate collagen-induced arthritis by inducing host-derived indoleamine 2,3 dioxygenase. Arthritis Res Ther. 2016;18:77.CrossRefPubMedPubMedCentralGoogle Scholar
- 39.Roddy GW, Oh JY, Lee RH, Bartosh TJ, Ylostalo J, Coble K, et al. Action at a distance: systemically administered adult stem/progenitor cells (MSCs) reduce inflammatory damage to the cornea without engraftment and primarily by secretion of TNF-alpha stimulated gene/protein 6. Stem Cells. 2011;29:1572–9.CrossRefPubMedGoogle 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.