Histone H3F3A and HIST1H3B K27M mutations define two subgroups of diffuse intrinsic pontine gliomas with different prognosis and phenotypes
- 7.3k Downloads
Diffuse intrinsic pontine glioma (DIPG) is the most severe paediatric solid tumour, with no significant therapeutic progress made in the past 50 years. Recent studies suggest that diffuse midline glioma, H3-K27M mutant, may comprise more than one biological entity. The aim of the study was to determine the clinical and biological variables that most impact their prognosis. Ninety-one patients with classically defined DIPG underwent a systematic stereotactic biopsy and were included in this observational retrospective study. Histone H3 genes mutations were assessed by immunochemistry and direct sequencing, whilst global gene expression profiling and chromosomal imbalances were determined by microarrays. A full description of the MRI findings at diagnosis and at relapse was integrated with the molecular profiling data and clinical outcome. All DIPG but one were found to harbour either a somatic H3-K27M mutation and/or loss of H3K27 trimethylation. We also discovered a novel K27M mutation in HIST2H3C, and a lysine-to-isoleucine substitution (K27I) in H3F3A, also creating a loss of trimethylation. Patients with tumours harbouring a K27M mutation in H3.3 (H3F3A) did not respond clinically to radiotherapy as well, relapsed significantly earlier and exhibited more metastatic recurrences than those in H3.1 (HIST1H3B/C). H3.3-K27M-mutated DIPG have a proneural/oligodendroglial phenotype and a pro-metastatic gene expression signature with PDGFRA activation, while H3.1-K27M-mutated tumours exhibit a mesenchymal/astrocytic phenotype and a pro-angiogenic/hypoxic signature supported by expression profiling and radiological findings. H3K27 alterations appear as the founding event in DIPG and the mutations in the two main histone H3 variants drive two distinct oncogenic programmes with potential specific therapeutic targets.
KeywordsApparent Diffusion Coefficient Diffuse Intrinsic Pontine Glioma MYCN Amplification K27M Mutation Distribute Diffusion Coefficient
High-grade gliomas are the most common paediatric malignant brain neoplasms and among them diffuse intrinsic pontine glioma (DIPG) is a leading cause of death from solid tumours in children, with no improvement in outcome in decades. The median survival below 1 year does not properly describe the various outcomes encountered clinically, from rapid growth over a few months to more prolonged survival up to 2 years in some cases . Histological grading plays no part in diagnosis nor prognostication . Surgery is not offered due to the infiltrative nature of the neoplasm in a delicate brain structure, and biopsy was abandoned in favour of a clinical and radiological diagnosis only . Imaging parameters, when limited to morphological alterations, have not been associated with survival either .
Paediatric high-grade gliomas (pHGG) were considered to mimic their adult counterparts until recent genomic studies unravelled significant differences between tumours arising at different ages [2, 10, 41, 42, 58]. A unique epigenetic reprogramming has recently been suggested in pHGG by the discovery of recurrent mutations in genes encoding histone H3 variants, never described in any other type of cancer of any location or age [50, 59]. These hotspot mutations were used to define distinct epigenetic and biological subgroups of HGG specifically seen in children . Supratentorial HGG have recurrent mutations in H3F3A encoding histone H3.3, with G34R/V restricted to hemispheric tumours, and K27M to those occurring in the midline . DIPG exclusively harbours K27M mutations, but in histone H3.1 as well as H3.3 genes [8, 20, 56, 59, 60].
K27M mutations in both genes substitute a key lysine residue on the histone H3 tail for a methionine, and have been shown to exert biochemical inhibition of the Polycomb Repressor Complex 2 (PRC2) resulting in a global loss of trimethylation of lysine 27 on all histones H3 molecules either wild types or mutated . However, a closer examination of the epigenome in mutated cells identified subtler deregulation with focal gains and maintenance of H3K27me3, for which the mechanism remains elusive [3, 11]. Although differing only by five amino acids, these two histone H3 proteins are distinct in terms of expression during the cell cycle, chaperones for incorporation in the nucleosomes, localisation on the genome and presumably physiological functions [23, 54]. Consequently, we sought to elucidate if K27M mutations in the distinct histone H3 variants (i.e. HIST1H3B and H3F3A) were associated with a specific biology and performed comprehensive histological, radiological, transcriptome and CGH array (aCGH) analyses on an extended cohort of DIPG cases biopsied at diagnosis selected with stringent clinical, radiological and histological integrated criteria.
Materials and methods
Patients and tumour samples
Patients were selected based on classical clinical and radiological criteria, i.e. short clinical history (<3 months of symptom duration) and the presence of a pontine tumour infiltrating at least 50 % of the pons (suppl. Table S1; suppl. Fig S1) . All patients underwent systematic stereotactic or surgical biopsy at Necker Hospital (Paris, France). Diagnosis of glial infiltrative non-pilocytic neoplasm was histologically confirmed in all patients and we obtained snap-frozen tumour material from 91 children. A smear of each biopsy was performed before freezing and the presence of tumour cells was assessed before their use for the genomic analyses. Informed consent for the translational research programme was obtained from the parents or guardian according to the IRB approved protocol (number DC-2009-955 for tumour banking). Site of the biopsy (routinely the junction between the pons and the cerebellar peduncle where there was an hypersignal on FLAIR sequences) was checked on the post-biopsy imaging .
Biopsies from patients with pHGG in a non-brainstem location (n = 93) were obtained during the same period at Necker hospital. Their histone H3.1/H3.3 mutation status was determined by Sanger sequencing.
Histone H3 genes were analysed by direct sequencing of PCR-amplified products from tumour DNA using primers listed in suppl. Table 2.
Morphological sequences of the MRI were reviewed independently by three clinicians (neurosurgeon, neuroradiologist and oncologist). The following criteria were scored on the diagnostic MRI: location, contrast enhancement, ring contrast enhancement, large area of necrosis, cysts, presence of stripes in the infiltrated brain on T2/FLAIR sequences and tumour size. In case of discrepancy, definitive scoring was obtained by consensus. Response to radiotherapy was judged by the clinical improvement only in patients with stable or decreasing doses of steroids to avoid misinterpretation of the radiology due to pseudoprogression . Disease extension was registered throughout the follow-up (i.e. local only, loco-regional or metastatic). Patients without an MRI in the last 2 months were excluded, considering that data were incomplete.
For the diffusion maps acquired at diagnosis, regions of interest (ROI) were drawn over the T2 hyperintensity corresponding to the tumour in each slice, creating a volume of interest (VOI) carefully avoiding necrotic areas, by an experienced neuroradiologist blind to the clinical data. These VOIs were transferred to the co-registered diffusion maps and every voxel value was individually registered. Histograms were created from the apparent diffusion coefficient (ADC) and distributed diffusion coefficient (DDC) voxel data .
The performance in terms of prognostication of the recently published “DIPG survival model” was evaluated in comparison to other biological stratification. This score is based on the assessment of the following parameters: age at diagnosis, interval between first symptoms and diagnosis, presence of a ring enhancement and use of adjuvant chemotherapy in addition to radiotherapy .
Histology, immunohistochemistry (IHC) and FISH analyses
Tumours were histologically classified according to WHO 2007 criteria whenever possible. Emphasis was put on the presence of an oligodendroglial component in the tumour cells (morphology, negativity of the tumour cells for vimentin, positivity for OLIG2), presence of interstitial oedema (semi-quantitative) and the presence or absence of necrosis.
A systematic panel of IHC markers was routinely performed: OLIG2, vimentin, GFAP, p53 (DO-7), PTEN, EGFR and MIB-1 as previously described . Additional stainings were developed to detect the loss of nuclear expression of the trimethylation mark at position K27 of the histone 3 tail (1:1000, polyclonal rabbit antibody, Diagenode, Belgium), the nuclear expression of the K27M form of histone H3 (1:1000, polyclonal rabbit antibody, Millipore, CA) and loss of ATRX nuclear expression (1:200, polyclonal rabbit antibody, Sigma-Aldrich, MO).
PDGFRA gene copy number was assessed by fluorescent in situ hybridization (FISH) using PDGFRA/CEN4 Dual Color Probe (Abnova, Tapei, Taiwan) on interphase nuclei. In brief, four-micron sections of tumour were mounted on SuperFrost Plus slides (Erie Scientific CA., Portsmouth, NH) and the probed area determined in accordance with haematoxylin and eosin-stained section. The sections were deparaffinised in xylene, rehydrated through an ethanol series, air-dried and incubated in pre-treatment solution (1 M NaSCN-Tris) at 80 °C for 25 min. Slides were then treated with a 0.01 % pepsin solution (Sigma-Aldrich, Saint Louis, USA) at 37 °C for 8 min. After dehydration, 10 µl of probe mixture was applied to each sample, slides were coverslipped and co-denatured at 75 °C for 5 min and hybridized at 37 °C for 48 h using thermobrite system (Leica Biosystems, Richmond, IL). A post-hybridization wash was performed in 2 × SSC at 73 °C for 2 min. Preparations were dehydrated and counterstained with 4,6-diamidino-phenyl-indole (DAPI). Signals were scored in at least 100 non-overlapping interphase nuclei. PDGFRA gene amplification was considered as positive in (A) specimens that have ≥40 % of cells displaying ≥4 copies of the PDGFRA signal, (B) specimens that display PDGFRA gene amplification, according to one of the following criteria: (a) a PDGFRA to CEN4 ratio ≥2 over all scored nuclei and calculated using the sum of PDGFRA divided by the sum of CEN4 when mean CEN4 per cell is ≥2 copies; (b) the presence of gene cluster (≥4 spots) in ≥10 % of tumour cells; (c) at least 15 copies of the PDGFRA signals in ≥10 % of tumour cells. Results were recorded using a DM600 imaging fluorescence microscope (Leica Biosystems, Richmond, IL) and digital imaging software (Cytovision, v7.4).
Genomic and statistical analysis
Gene expression (GE) profiling and comparative genomic hybridization on array (aCGH) were conducted for patients with enough material available of required quality on an Agilent platform. For all statistical analysis, the level of significance was 5 %.
Gene expression microarrays analysis
The data analysed is the result of the gathering of DIPG samples belonging to three different cohorts of young patients with high-grade glioma. In such experimental design, a well-known undesired bias is the batch effect, which is purely technical. To correct for this effect, we replicated some samples in at least two of three batches and performed the ComBat analysis  as recommended in Chen et al. . A normal brainstem sample from commercial source, hybridized on chips in the three batches, was used as a reference to normalize the data for the intensity bias as recommended in Do et al. . The normalization of fluorescence intensities is performed in two steps, inspired by Bolstad et al. . The first one is a loess normalization using the normal brainstem samples as a common baseline array for all arrays in the cohort. The second step is a quantile normalization.
Differential analysis was then performed using a moderated t test, comparing mean of log2(intensities) in both H3.1- and H3.3-mutated DIPG samples, implemented in the limma R package. The null hypothesis H0 for each gene is that the mean of log2(intensities) is the same in both H3.1- and H3.3-mutated DIPG samples. The alternative hypothesis H1 for each gene is that they are different. The obtained p values were adjusted for multiple testing using the Benjamini–Hochberg procedure. The a priori defined level of significance was 5 % after correction for multiple testing.
Comparative genomic hybridization array analysis
Log2(ratio) between raw signals of the reference DNA and DIPG DNA was first normalized for the dye and intensity effects and also for the local GC % content, using loess procedure, implemented in the limma R package . Data were then centred according to normalized log2(ratio) distribution values by an in-house script using the EM (expectation maximization) approach and aberrations status calling was automatically performed. Normalized centralized values were then segmented using the Circular Binary Segmentation (CBS) algorithm , implemented in the DNAcopy R package. The normal copy number interval [log2(ratio) = 0 which means 2 DNA copies as in the DNA reference] was calculated for each sample by multiplying with a define factor the median absolute deviation (MAD) of the normalized data for each single sample.
Survival curves comparisons
Survival functions were estimated with the Kaplan–Meier method and all survival function estimate comparisons were performed using a log-rank test. The null hypothesis H0 was that the two considered survival function estimates were the same. The alternative hypothesis H1 was that they were different.
Multivariate survival analysis
The multivariate survival analysis was conducted on H3.1- or H3.3-mutated patients only. The other patients (wild type or H3.2 mutated) were excluded from this analysis. The pool of initial covariates to include in the Cox model was: (1) H3-variant mutation, (2) presence of metastasis, (3) MRI contrast enhancement, (4) treatment type and (5) DIPG Risk score. We first checked for multicollinearity. H3-variant mutation and presence of metastasis are correlated (cor = 0.32) as well as MRI contrast enhancement and the DIPG Risk score (cor = 0.56). We started the analysis with 4 different pools of covariates to separate the correlated pairs of variables and then performed a backward stepwise variable selection using the pec R package. At the end of the procedure, two models retained no variables while the other retained two (‘H3-variant mutation’ and ‘DIPG Risk score’) and three variables (‘H3-variant mutation’, ‘treatment type’ and ‘MRI contrast enhancement’), respectively. For both models, the coefficient for ‘H3-variant mutation’ was 1.4109 and 1.6389, respectively, while the other coefficients were much lower (DIPG Risk score coef = 0.0873, radiotherapy coef = 0.1101, Tarceva coef = −0.4963, Temodal coef = 0.5126, MRI contrast enhancement coef = 0.6573). The hazard ratios were 3.98 and 5.00, respectively, and the confidence intervals were (0.25; 0.92) and (0.67; 1.93), respectively. The log-likelihood p values were 3.39e−5 and 7.69e−5, respectively. The H3-variant mutation is then the most important variable in the multivariate Cox model.
Distribution analyses (diffusion data and age at diagnosis)
To test if both ADC and DDC values were drawn from the same distribution for H3.1- and H3.3-mutated patients, a Mann–Whitney test was performed in each case. The null hypothesis H0 is that ADC (or DDC, respectively) values are drawn from the same distribution law for both H3.1- and H3.3-mutated patients. The alternative hypothesis H1 is that they follow two different distribution laws.
The same methodology was adopted to compare distributions of age at diagnosis between H3.1- and H3.3-mutated patients.
To compare sex ratios in H3.1- and H3.3-mutated patients, we performed a Chi-squared proportion test as none of the theoretical headcounts was below 5. The null hypothesis H0 was that the male–female proportions are the same in the two considered groups. The alternative hypothesis H1 was that they are different.
For the response to radiotherapy comparison, there was one theoretical headcount below 5, so we performed a Fisher exact test instead of the Chi-squared proportion test. The null hypothesis H0 was that the proportions of good and bad response to radiotherapy were the same in the two considered groups.
Loss of H3K27me3 and histone H3 mutation as hallmarks of DIPG
All but one H3-K27M mutation found by sequencing could also be accurately detected by IHC, including a novel mutation a gene encoding the H3.2 variant, HIST2H3C, not previously described (Fig. 1c, suppl. Fig S2b, c). However, a second novel mutation, a lysine-to-isoleucine substitution (K27I) in H3F3A resulting from two nucleotide changes within the same codon (c.[83A>T; 84G>T]) failed to be detected by the H3-K27M staining but presented a loss H3K27me3 immunoexpression (suppl. Fig S2b, c). In addition, we observed qualitative differences in the labelling intensity of the H3-K27M in the nuclei of the tumour cells between H3.3/H3.2- (strong staining) and H3.1-mutated alleles (weak staining, cross-reactivity corresponding to the important similarity among the two variants already shown ) (Fig. 1 a, b, suppl. Fig S2d).
We performed further Sanger sequencing of histones HIST1H3B, H3F3A and for wild-type cases. we subsequently examined HIST1H3C and HIST2H3C in an extended cohort of 183 pHGG from diverse anatomical regions. We identified H3.3-K27M mutations in midline tumours whereas H3-G34R/V mutations were restricted to the cerebral hemispheres. Conversely, H3.1- and H3.2-K27M, as well as H3.3 K27I, were only found in pontine tumours (Fig. 1d).
Landscape of genomic alterations in H3.3- and H3.1-mutated DIPG
Given the discrepancies in H3.1 and H3.3 functions and genomic localization, we next explored the differences between H3.3- and H3.1-mutated tumours using integrated DNA copy number and gene expression analysis.
We further conducted genome-wide aCGH analysis to determine the subgroup specificity of DNA copy number alterations. We observed a more frequent gain of chromosomes 1q (83 vs. 44 %; p value = 0.035) and 2 (75 vs. 16 %; p value = 0.0008) in the H3.1 subgroup, whereas loss of 17p13.1 (8 vs. 48 %; p value = 0.0272) was preferentially found in H3.3 tumours (Fig. 2). These data were significantly correlated with gene expression modulations of the genes located in these regions (data not shown). As the latter locus corresponds to TP53, previously shown mutated at least in 40 % of DIPG [8, 20, 24, 60], we evaluated expression by IHC and found exclusive p53 accumulation in H3.3-K27M samples (Fig. 2, S3b; p value = 0.0001). Gene set enrichment analysis (GSEA) of our expression profiling data accordingly showed downregulation of TP53 targets (suppl. Fig S3b). Likewise, gain or amplification of PDGFRA locus (4q12) was only seen in H3.3 tumours and the overexpression of genes upregulated in PDGFRA-amplified pHGG  was only observed in the H3.3-K27M subgroup (suppl. Fig S3c, d). ATRX expression was also evaluated by IHC, and ATRX loss was only found in 24 % of H3.3-mutated tumours (suppl. Fig S3e, f). Finally, ACVR1 mutations were exclusively found in H3.1-K27M tumours (Fig. 2, suppl. Fig S3g) as previously described [8, 20, 56, 60].
K27M mutations in H3.3 and H3.1 mutations drive two distinct oncogenic programmes
HIST1H3B K27M mutation is associated with a less aggressive behaviour in DIPG
Mutations in the N-terminal tail of histone H3 genes have been recently found in midline pHGG and at particular high frequency in DIPG. Our results support that substitution of the lysine at position 27 of genes encoding H3, which can be detected by IHC through its consequences, i.e. the global loss of trimethylation on this residue, is the driving event in DIPG oncogenesis. In addition to K27M mutations in genes encoding H3.3 (H3F3A) and H3.1 (HIST1H3B and HIST1H3C) found in the majority of samples as previously reported, we also identified two novel lysine 27 substitutions. The first was a novel K27M mutation in a gene encoding histone H3.2 (HIST2H3C) implying that whichever histone H3 variant is targeted by the mutation (H3.1, H3.2 or H3.3), it may lead to the development of a DIPG. A further argument for the importance of alterations of these residues is the discovery of a second novel mutation, two base changes being required to produce a K27I lysine-to-isoleucine substitution in H3.3 (H3F3A). Although reported here in a DIPG sample for the first time, systematic in vitro modification of the lysine 27 predicted K27I as the only substitution other than K27M to result in a repressive effect on H3K27me3 . In summary, our data underline the pivotal importance in this disease of H3K27 mutations and loss of H3K27me3, which is consequently characteristic of DIPG and likely represents the first genomic event leading to transformation in these tumours. DIPG thus appears as a homogeneous tumour group defined as a glial neoplasm with a stereotypical epigenetic drive consecutive to these histone H3 mutations.
All H3K27 mutations described in DIPG appear to have the same epigenomic consequences on the PRC2 complex in general [11, 34] despite the distinct functions and genomic distribution of the different variants. Critically, it is clear that the type of histone targeted by K27 alterations greatly influences the survival length of patients. H3.1-mutated tumours were found to respond better to the treatment, mainly radiotherapy, have a less aggressive course and metastasize less frequently. As the significant effect on OS length was confirmed by the re-analysis of an independent series , assessment of the type of histone mutation could be used as a prognostic stratification factor in future prospective studies. Moreover, in multivariate analysis considering also the effect of treatment, the type of histone H3 mutated was a better predictor for survival length than the DIPG clinico-radiological risk score  or the presence of an ACVR1 mutation identified in a subset of DIPG with a less aggressive course [56, 60]. The latter results concerning ACVR1 mutation can be explained by the segregation of this alteration with H3.1-K27M mutation rather than the reverse. Indeed, no difference was observed among the H3.1 tumour subgroup between ACVR1 WT- and ACVR1-mutated samples (Fig S7e). Also, the high frequency of metastatic progression in H3F3-mutated DIPG, if confirmed in an independent series, could suggest a different therapeutic approach in the subset of patients with tumours prone to dissemination.
Our GE profiling data suggest that the two main histone H3 mutations drive distinct oncogenic programmes, an observation further strengthened by the integration of histological and MR-imaging data. Notably, the less aggressive H3.1 tumours appeared to be more hypoxic and oedematous, as well as exhibiting higher expression of neo-angiogenic markers than their H3.3 counterpart. A bias towards oligodendroglial and astroglial differentiation was observed for H3.3 and H3.1 tumours, respectively. Also, the dramatic metastatic evolution of the disease—observed almost exclusively in H3.3-K27M samples, could be partially explained by the identification of deregulated genes encoding ion-gated channels and adhesion molecules known to play a role in this invasion process [15, 52]. Despite their different progression rate, we did not find discrepancies in MIB-1 proliferative index between the two groups.
We show here that the H3-K27M-mutated DIPG group recognized by previous studies [7, 8, 31] could be further divided into two subgroups, H3.1-K27M and H3.3-K27M. There is significant overlap between the H3.3-K27M subgroup and the PDGFRA-driven DIPG shown previously and between the H3.1-K27M subgroup and the tumours with a mesenchymal gene expression signature [40, 46]. The wild-type DIPG (with respect to histone H3 mutation) described by other groups [8, 31], is also found in our series, but we did not find differences in survival between WT and H3-K27-mutated groups in this population of strictly defined DIPG (suppl. Fig S7f). Importantly, all but one of the silent tumours we describe had a loss of the trimethylation mark at K27 of the histone H3. As this is an indication of the functional impairment of the PRC2 complex, alternative mechanisms have to be explored to confirm these tumours could be grouped together with the histone H3-27 M-mutated ones. We did not find a MYCN amplification, nor an IDH1/2 mutation in the only sample without trimethylation loss at K27 of the histone H3 (case P188; suppl. Fig S2). MYCN-amplified cases may correspond to a PNET-like variant and are in fact relatively rare (less than ten cases reported so far in the whole literature with MYCN amplification).
The differences between the H3.3/H3.1 subgroups may be a result of distinct cells of origin for these tumours, or reflect a mechanistic consequence of the type of histone mutated. Interestingly, we identified a differential expression between these subgroups of numerous homeobox genes, known to be pivotal for neural stem cell identity and tissue patterning. This coupled with the earlier age of onset of the H3.1-mutated tumours as well as their restricted pontine location (as opposed to H3.3 mutations also found in HGG of other midline regions) could imply that distinct progenitors are targeted by the two mutations. Equally, homeobox clusters are known to be regulated by the PRC2 complex  and the epigenetic consequences of the PRC2 complex inhibition might have regional differences depending on the histone type. Secondary genetic alterations also co-segregate between the two subgroups. PDGFRA amplification/mutation, a known inducer of oligodendrocytic differentiation [6, 19, 48], is seen together only with histone H3.3 mutations, whilst ACVR1 mutations that activate the BMP pathway driving astrogenesis are only concomitantly observed with mutant H3.1. Still, the gene expression signature between H3.1 and H3.3 subgroups may not be due to the H3 mutations themselves but rather the accompanying alterations (PDGFRA vs. ACVR1) or differential modification of the microenvironment by tumour cells. However, the astrocytic phenotype was observed in the H3.1-mutated tumours independently of the ACVR1 mutation and future work in vitro should help unravel this matter.
In conclusion, biopsies performed at diagnosis allowed us to confirm a DIPG in almost every case by the identification of a specific histone H3K27 alteration. The type of histone H3 mutated could also predict the outcome of DIPG patients more efficiently than clinical and radiological characteristics of the tumours. Moreover, two distinct oncogenic pathways associated with the mutated histone-driven subgroups have been identified in DIPG and different therapeutic approaches may be developed to target these specific alterations and phenotypic changes. Conversely, mutually exclusive alterations observed in the two DIPG subgroups may influence the response to a given targeted agent. Crucially, our results justify the re-introduction of stereotactic biopsies at diagnosis in the management of DIPG for better treatment stratification decisions.
Tumor and Blood Bank of Necker-Enfants Malades Hospital, Paris (Pr. Thierry Molina and Caroline Werl). CRB-ADN, Imagine, Université Paris Descartes.
DC acknowledges funding from Programme PICRI 2011 “Partenariats Institution Citoyens pour le Recherche et l’Innovation”, CP acknowledges funding from Programme “Brainomics” de l’Agence Nationale pour la Recherche and JG and MAD acknowledge support from Charities “Etoile de Martin”, “Défi de Fortunée”, “Lemos Family and Friends”. SP acknowledges support from the Association pour la Recherche en Neurochirurgie Pédiatrique. KRT, AM, and CJ acknowledge NHS funding to the National Institute of Health Research Biomedical Research Centres and support from Abbie’s Army.
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.