Variants in the 14q32 miRNA cluster are associated with osteosarcoma risk in the Spanish population

Abstract

Abstract Association studies in osteosarcoma risk found significant results in intergenic regions, suggesting that regions which do not codify for proteins could play an important role. The deregulation of microRNAs (miRNAs) has been already associated with osteosarcoma. Consequently, genetic variants affecting miRNA function could be associated with risk. This study aimed to evaluate the involvement of all genetic variants in pre-miRNAs described so far in relationship to the risk of osteosarcoma. We analyzed a total of 213 genetic variants in 206 pre-miRNAs in two cohorts of osteosarcoma patients (n = 100) and their corresponding controls (n = 256) from Spanish and Slovenian populations, using Goldengate Veracode technology (Illumina). Four polymorphisms in pre-miRNAs at 14q32 miRNA cluster were associated with osteosarcoma risk in the Spanish population (rs12894467, rs61992671, rs58834075 and rs12879262). Pathway enrichment analysis including target genes of these miRNAs pointed out the WNT signaling pathways overrepresented. Moreover, different single nucleotide polymorphism (SNP) effects between the two populations included were observed, suggesting the existence of population differences. In conclusion, 14q32 miRNA cluster seems to be a hotspot for osteosarcoma susceptibility in the Spanish population, but not in the Slovenian, which supports the idea of the existence of population differences in developing this disease.

Full text

1
SCIENTIFIC RePORts | (2018) 8:15414 | DOI:10.1038/s41598-018-33712-4
www.nature.com/scientificreports
Variants in the 14q32 miRNA
cluster are associated with
osteosarcoma risk in the Spanish
population
Idoia Martin-Guerrero1, Nerea Bilbao-Aldaiturriaga2, Angela Gutierrez-Camino2,
Borja Santos-Zorrozua2, Vita Dolžan
3, Ana Patiño-Garcia4 & Africa Garcia-Orad2,5
Association studies in osteosarcoma risk found signicant results in intergenic regions, suggesting that
regions which do not codify for proteins could play an important role. The deregulation of microRNAs
(miRNAs) has been already associated with osteosarcoma. Consequently, genetic variants aecting
miRNA function could be associated with risk. This study aimed to evaluate the involvement of all
genetic variants in pre-miRNAs described so far in relationship to the risk of osteosarcoma. We analyzed
a total of 213 genetic variants in 206 pre-miRNAs in two cohorts of osteosarcoma patients (n = 100)
and their corresponding controls (n = 256) from Spanish and Slovenian populations, using Goldengate
Veracode technology (Illumina). Four polymorphisms in pre-miRNAs at 14q32 miRNA cluster were
associated with osteosarcoma risk in the Spanish population (rs12894467, rs61992671, rs58834075 and
rs12879262). Pathway enrichment analysis including target genes of these miRNAs pointed out the
WNT signaling pathways overrepresented. Moreover, dierent single nucleotide polymorphism (SNP)
eects between the two populations included were observed, suggesting the existence of population
dierences. In conclusion, 14q32 miRNA cluster seems to be a hotspot for osteosarcoma susceptibility
in the Spanish population, but not in the Slovenian, which supports the idea of the existence of
population dierences in developing this disease.
Osteosarcoma is the most common primary malignant bone tumor, mainly occurring in children and adoles-
cents. e precise etiology of the disease remains partially unknown1, but genetic factors seem to play a key role
in its pathogenesis2,3. To date, several case-control studies have reported associations of common genetic vari-
ants with osteosarcoma risk3–5, but these studies were mainly focused on regions codifying for proteins, because
results are easily interpreted biologically. However, a genome wide association study (GWAS) in osteosarcoma
showed that 8 out of the 13 most signicant genetic variants were located in regions with no clear functional
consequence6, results that are more dicult to interpret. Similar results were found in other GWAS in dierent
cancer types, in which 44% of signicant signals were described to be located in intergenic regions7. All these data
together suggest that regions which do not codify for proteins could play an important role in the risk of cancer,
in general, and in osteosarcoma, in particular. One of the most studied non-coding RNAs are microRNAs (miR-
NAs), molecules of 20 nucleotides that regulate gene expression at the post-transcriptional level by binding to the
3′ untranslated region (UTR) of a target mRNA8, leading to its translation inhibition or degradation. rough this
mechanism, miRNAs can regulate more than half of human genes9. More than 600 miRNAs have been proposed
to be involved in osteogenesis regulation10, so it is reasonable to think that miRNAs deregulation can be linked
to osteosarcoma susceptibility. In fact, alterations of miR-34c aecting Notch signaling pathway were associated
with the pathogenesis of osteosarcoma11, and the deregulation of the 14q32 miRNA cluster was also linked to the
1Department of Genetics, Physical Anthropology and Animal Physiology, Faculty of Science and Technology,
University of the Basque Country, UPV/EHU, Leioa, Spain. 2Department of Genetics, Physical Anthropology and
Animal Physiology, Faculty of Medicine and Nursery, UPV/EHU, Leioa, Spain. 3Institute of Biochemistry, Faculty of
Medicine, Ljubljana, Slovenia. 4Laboratory of Pediatrics, University Clinic of Navarra, Pamplona, Spain. 5BioCruces
Health Research Institute, Barakaldo, Spain. Idoia Martin-Guerrero and Nerea Bilbao-Aldaiturriaga contributed
equally. Correspondence and requests for materials should be addressed to A.G.-O. (email: africa.garciaorad@ehu.
eus)
Received: 5 July 2018
Accepted: 26 September 2018
Published: xx xx xxxx
OPEN
Content courtesy of Springer Nature, terms of use apply. Rights reserved

www.nature.com/scientificreports/
2
SCIENTIFIC RePORts | (2018) 8:15414 | DOI:10.1038/s41598-018-33712-4
progression and prognosis of osteosarcoma12. Genetic variations in miRNAs can alter their function aecting
their gene targets. ese variants can modify the miRNA expression levels if they are located in the pre-miRNA or
the mRNA-miRNA binding if they are located in the seed region. Consequently, genetic variations in pre-miRNAs
aecting their function could be involved in the risk of cancer. Several works have already described polymor-
phisms in miRNAs associated with the susceptibility to dierent types of cancer13,14. Despite all these evidences,
few studies have analyzed the involvement of miRNA single nucleotide polymorphism (SNPs) in the risk of
osteosarcoma so far. Although only a low number of SNPs were analyzed, signicant results were found with two
variations belonging to miR-34 family15,16 and with one located in mir-124a17.
Considering that the number of annotated miRNAs has increased substantially up to 2500 miRNAs approx-
imately18, the aim of this study was to evaluate the contribution in the risk of osteosarcoma of variants in
pre-miRNAs. With that objective, all variants in pre-miRNAs with a minor allele frequency (MAF) higher than
1% were analyzed in a representative group of osteosarcoma patients from two populations.
Materials and Methods
Patients. e study population included 100 patients (<34 years) diagnosed of osteosarcoma at the Oncology
Unit of the Department of Pediatrics of the University Clinic of Navarra (n = 74) between 1985 and 2003 and
University Children’s Hospital of Liubliana (n = 26) between 1990 and 2008. Both patient cohorts were residents
in Spain or Slovenia at the moment of diagnosis and had West European ancestry. Moreover, 256 healthy individ-
uals of European origin with no previous history of cancer (n = 160 and n = 96 from Spain and Slovenia, respec-
tively) were added (Table1). Informed consent was obtained from all patients or their parents before sample
collection. e study was approved by the Spanish Ethics Committees for Clinical Research of Euskadi (CEIC-E)
(CEISH/102R/2011/GARCIA-ORAD CARLES 67/02/12) and the University of Navarra (105/2009), and by the
Slovenian Ethics Committee for Research in Medicine (bilateral project BI- ES/04-05-016) and was carried out
according to the Declaration of Helsinki.
Selection of polymorphisms in miRNAs. We selected all the pre-miRNAs including SNPs with a
MAF > 0.01 in European/Caucasian populations described in the databases until May 2014. Since, on the one
hand, osteosarcoma is a polygenic disease in which associated genes are not totally dened, and, on the other
hand, a single miRNA can regulate several transcripts which are not completely known nowadays, we decided to
analyze all polymorphic miRNAs to date. MAF > 0.01 was selected because this frequency was required to detect
signicant dierences in our sample size.
e SNP selection was performed using miRNA SNiPer (www.integratomics-time.com/miRNA-SNiPer/),
NCBI and literature review. Finally, a total of 213 SNPs in 206 pre-miRNAs were included.
Genotyping. Peripheral blood samples were obtained as the source of DNA from Spanish patients and all
healthy controls, while in Slovenian osteosarcoma patients DNA was extracted from the areas of formalin xed
paran embedded (FFPE) material veried by an experienced pathologist to be representative of normal tis-
sue. Most FFPE samples were osteogenic (>96%) from histological point of view, and all of them were primary
malignancy. Genomic DNA was extracted using standard procedures19. DNA was quantied using PicoGreen
(Invitrogen Corp., Carlsbad, CA). For each sample, 400 ng of DNA were genotyped using the GoldenGate
Genotyping Assay with Veracode technology according to the published Illumina protocol. Data were ana-
lyzed with Genome Studio soware for genotype clustering and calling. As quality control, duplicate samples
and CEPH trios (Coriell Cell Repository, Camden, NJ) were genotyped across the plates, following the Illumina
recommendations.
Statistical analysis. e association between genetic polymorphisms and the risk of osteosarcoma was
evaluated by the χ2 or Fisher’s exact test. e eect sizes of the associations were estimated by the OR’s from
univariate logistic regression. e most signicant test among codominant, dominant, recessive and additive was
used to determine the statistical signicance of each SNP. e results were adjusted for multiple comparisons by
the False Discovery Rate (FDR)20. In all cases the signicance level was set at 5%. Analyses were performed by
Tot a l Controls Cases
Participants (n) 356 256 100
Population (n;%)
Spain 234 160 (68.37) 74 (31.62)
Slovenia 122 96 (78.68) 26 (21.31)
Age (mean; sd)
Spain 69.01 (17.5) 14.5 (4.7)
Slovenia 46 (9.3) 19.5 (8.6)
Sex (f/m)a
Spain 111/120 81/79 30/41
Slovenia 51/71 35/58 13/13
Table 1. Study population. Abbreviations: n, number of individuals; sd, standard deviation; f, female; m, male.
aSex was not available for all patients.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

www.nature.com/scientificreports/
3
SCIENTIFIC RePORts | (2018) 8:15414 | DOI:10.1038/s41598-018-33712-4
using R v2.11 soware. χ2 test was used to search for any deviation of Hardy-Weinberg equilibrium (HWE) in
controls. ose SNPs that showed deviations from HWE in control population were removed from the analyses.
Bioinformatic analyses. miRNAs secondary structures prediction. e RNAfold web tool (http://rna.tbi.
univie.ac.at) was used to calculate the minimum free energy secondary structures and to predict the most stable
secondary structures of the miRNAs showing signicant SNPs.
miRNAs expression analyses. Expression levels of miRNAs were analyzed in a series of 18 osteosarcoma cell
lines (<34 years) and 4 normal bone samples, using data publicly available in GEO database under the acces-
sion GSE2842321. T-tests were performed using the GEO2R web tool, applying a Benjamini and Hochberg FDR
adjusted p-value cut-o of 0.05.
Gene targets selection and pathways analysis. MiRWalk22 (http://zmf.umm.uni-heidelberg.de/apps/zmf/miR-
walk2/) database was used to select miRNA targets. Only targets predicted by at least 8 dierent algorithms
provided by miRWalk were selected. Enriched pathway analyses of putative target genes were determined with
ConsensusPath database (CPdB) (http://consensuspathdb.org/)23 using the over-representation analysis module.
Gene list were analyzed against the default collection of KEGG24, Reactome25 and BioCarta (http://cgap.nci.nih.
gov/Pathways/BioCarta_Pathways) pathway databases. A conservative p-value cuto (0.001) was used.
Results
Genotyping results. Genotyping analyses were performed in 100 patients diagnosed of osteosarcoma (74
Spanish and 26 Slovenian) and 256 cancer-free controls (160 and 96, respectively). Successful genotyping was
obtained in 350 of 356 DNA samples (98.3%). Finally, a total of 140 SNPs were included in the association anal-
yses, aer eliminating SNPs with genotyping failures (<80%), monomorphic in the studied populations, or with
deviations from HWE in controls (TableS1).
Genotype association study. We found 23 SNPs signicantly associated with osteosarcoma risk; 14 SNPs
in 14 miRNAs in the Spanish population and 9 SNPs in 8 miRNAs in the Slovenian. When the two populations
were analyzed together, 11 SNPs at 11 miRNAs were signicant.
In the Spanish population, 4 out of 14 signicant SNPs were located at 14q32 region (Fig.1). Among them,
rs12894467 at miR-300 showed the most signicant association value under the log-additive model (CC vs CT vs
TT). e frequency of TT genotype was found to be 2.5 times higher in patients than in controls (OR = 2.01, 95%
CI: 1.32–3.06; P = 0.001). With regard to the other three signicant SNPs at 14q32 region, we found an increase in
the risk of osteosarcoma for the genotypes AG + AA for rs61992671, CT for rs58834075 and CG for rs12879262
located at miR-412, miR-656 and miR-4309, respectively (OR = 2.21, OR = 4.98 and OR = 1.99). Other 10 SNPs
showed statistically signicant results (P < 0.05), 6 located in pre-miRNAs, 2 in mature miRNAs and 1 in the seed
region (Table2). Aer FDR correction, no SNP remained signicant.
Figure 1. 14q32 miRNA cluster. (A) Diagram of the 14q32 miRNA cluster, including miRNAs analyzed in
our study (in bold), miRNAs with signicant SNPs (highlighted with an asterisk), and miRNAs described in
the literature to be downregulated. (B) Secondary structures of the 14q32 miRNAs showing signicant SNPs,
predicted by RNAfold web tool.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

www.nature.com/scientificreports/
4
SCIENTIFIC RePORts | (2018) 8:15414 | DOI:10.1038/s41598-018-33712-4
In the Slovenian population, 9 SNPs were signicant. Among them, rs35613341 at miR-5189 showed the
most signicant association. e genotype CG for rs35613341 showed a protective eect (OR = 0.07, 95% CI:
0.01–0.59; under codominant model), association that remained signicant aer FDR correction. Another geno-
type in the same miRNA (AG + AA for rs56292801) also showed protective eect (OR = 0.25; 95% CI:0.08–0.80).
Other 7 SNPs displayed signicant results (P < 0.05), 4 located in pre-miRNAs and 3 in the seed region (Table3).
None of the miRNAs signicant in the Spanish population were signicant in the Slovenian.
In the global analysis, a total of 11 signicant SNPs were detected. Nine of them had been already found sig-
nicant in the Spanish or in the Slovenian populations. Among them, 3 SNPs showed more signicant and 5 less
signicant P values in the total population than those found in each population separately. e other 3 out of 11
signicant associations detected were new (TableS2). From the total of signicant SNPs observed in the Spanish
SNP miRNA Localization Genotype N (%)controls
N = 160 N (%)cases
N = 69 OR (CI 95%) P
1 rs12894467 mir-300 14q32
pre-miRNA
CC 74 (46.2) 19 (27.5)
2.01 (1.32–3.06) 0.001 (add)CT 71 (44.4) 34 (49.3)
TT 15 (9.4) 16 (23.2)
2 rs356125 mir-2278 9q22
pre-miRNA
GG 140 (87.5) 68 (98.6) 1
0.002 (codom)AG 20 (12.5) 1 (1.4) 0.1 (0.01–0.78)
AA 0 (0 0 (0) 0.00 (0.00)
3 rs77639117 mir-576 4q25
pre-miRNA
AA 156 (97.5) 60 (87.0) 1
0.003 (codom)AT 4 (2.5) 9 (13.0) 5.85 (1.74–19.71)
TT 0 (0) 0 (0) 0.00 (0.00)
4 rs7247237 mir-3188 19p13
pre-miRNA
CC 72 (45.3) 23 (34.3) 1
0.004 (rec)CT 77 (48.4) 31 (46.3) 3.59 (1.49–8.66)
TT 10 (6.3) 13 (19.4)
5 rs60871950 mir-4467 7q22.1
miRNA
GG 35 (22.0) 27 (39.1) 1
0.009 (dom)AG 83 (52.2) 23 (33.3) 0.44 (0.24–0.81)
AA 41 (25.8) 19 (27.5)
6 rs10505168 mir-2053 8q23.3
pre-miRNA
AA 78 (49.1) 25 (36.2) 1
0.009 (codom)AG 65 (40.9) 42 (60.9) 2.02 (1.11–3.65)
GG 16 (10.1) 2 (2.9) 0.39 (0.08–1.81)
7 rs61992671 mir-412 14q32
miRNA
GG 57 (35.6) 13 (20.0) 1
0.018 (dom)AG 66 (41.2) 35 (53.8) 2.21 (1.11–4.41)
AA 37 (23.1) 17 (26.2)
8 rs58834075 mir-656 14q32
pre-miRNA
CC 157 (98.1) 63 (91.3) 1
0.021 (codom)CT 3 (1.9) 6 (8.7) 4.98 (1.21–20.55)
TT 0 (0) 0 (0)
9 rs10406069 mir-5196 19q13
pre-miRNA
GG 110 (69.6) 43 (62.3) 1
0.021 (codom)AG 39 (24.7) 26 (37.7) 1.71 (0.93–3.13)
AA 9 (5.7) 0 (0.0) 0.00 (0.00)
10 rs12879262 mir-4309 14q32
pre-miRNA
GG 111 (69.8) 39 (56.5) 1
0.022 (codom)CG 43 (27.0) 30 (43.5) 1.99 (1.10–3.59)
CC 5 (3.1) 0 (0.0) 0.00 (0.00)
11 rs702742 mir-378h 5q33
pre-miRNA
AA 117 (73.1) 58 (86.6) 1
0.022 (dom)AG 41 (25.6) 8 (11.9) 0.42 (0.19–0.93)
GG 2 (1.2) 1 (1.5)
12 rs10422347 mir-4745 19p13
miRNA
CC 138 (87.3) 51 (75.0) 1
0.025 (dom)CT 19 (12.0) 17 (25.0) 2.30 (1.12–4.73)
TT 1 (0.6) 0 (0.0)
13 rs2289030 mir-492 12q22
pre-miRNA
CC 130 (81.2) 60 (87.0) 1
0.026 (rec)CG 30 (18.8) 6 (8.7) 0.00
GG 0 (0.0) 3 (4.3)
14 rs35770269 mir-449c 5q11
seed
AA 61 (38.4) 35 (50.7)
0.64 (0.42–0.98) 0.038 (add)AT 71 (44.7) 28 (40.6)
TT 27 (17.0) 6 (8.7)
Table 2. Polymorphisms in miRNAs associated with osteosarcoma risk in the Spanish population.
Abbreviations: OR, Odd Ratio; CI, Condence Interval; add, additive; codom, codominant; dom, dominant;
rec, recessive.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

www.nature.com/scientificreports/
5
SCIENTIFIC RePORts | (2018) 8:15414 | DOI:10.1038/s41598-018-33712-4
(n = 14) or in the Slovenian population (n = 9), 14 did not show signicant results when both population were
analyzed together.
miRNAs secondary structures. We analyzed in silico the energy change (|ΔΔG|) and the secondary struc-
tures of the miRNAs with signicant SNPs. In the Spanish population, 4/14 miRNAs showed drastic energy
changes (>2.0 Kcal/mol) and 7 showed altered secondary structure (Fig.S1). With regard to the SNPs at 14q32
region, rs61992671 in miR-412 and rs58834075 in miR-656 induced positive energy changes which turned the
miRNA hairpins from a stable to an unstable status. In the Slovenian population, 2/9 miRNAs showed energy
changes >2.0 Kcal/mol and 3 displayed secondary structure changes (Fig.S2). In the global analysis, 2 of the 3
new detected miRNAs showed energy changes >2.0 Kcal/mol and all of them showed changes in the secondary
structure (Fig.S3).
miRNA expression. We studied the expression levels of miRNAs of interest in osteosarcoma cell lines using
the public database Gene Expression Omnibus (GEO). Out of 22 miRNAs with signicant SNPs, 5 miRNAs were
represented in the GSE28423 database (miR-300, miR-412, miR-492, miR-576 and miR-656). From them, mir-
300 was found signicantly down-regulated in osteosarcoma cell lines group (logFC = −1.545; adj-p = 0.006).
e rest of miRNAs showed no signicant results (p > 0.05).
Pathway analysis. We performed a pathway enrichment analysis with miRNAs of 14q32 region that mod-
ied the secondary structure, miR-412 and miR-656, using miRWalk database and ConsensusPathDB web tool.
MiR-300 (the most signicant SNP) was also included in pathway enrichment analysis although no remark-
able results were observed (data not shown). For miR-412, we found two pathways over-represented, being
both WNT signaling predicted by KEGG and Biocarta (TableS3). Regarding miR-656, only Ca2+ pathway was
over-represented, with 7/55 genes targeted by this miRNA (TableS4). Of these 7 genes, 5 overlapped with WNT
signaling pathway. When both miRNAs were analyzed together, 5 pathways were over-represented, being WNT
signaling pathway the most signicant (p = 0.000177) (TableS5), with 16/143 genes targeted by miR-412 and
miR-656 (TableS6).
NSNP miRNA Localization Genotype N (%) controls
N = 96 N(%) cases
N = 25 OR (CI 95%) P
1 rs35613341 mir-5189 16q24 pre-miRNA
CC 49 (51.0) 16 (69.6) 1
0.0001* (codom)CG 41 (42.7) 1 (4.3) 0.07 (0.01–0.59)
GG 6 (6.2) 6 (26.1) 3.06 (0.86–10.85)
2 rs4674470 mir-4268 2q35 pre-miRNA
TT 50 (52.1) 17 (85.0) 1
0.002 (codom)CT 40 (41.7) 1 (5.0) 0.07 (0.01 0.58)
CC 6 (6.2) 2 (10.0) 0.98 (0.18 5.33)
3 rs2070960 mir-3620 1q42 seed
CC 75 (78.9) 21 (87.5) 1
0.008 (codom)CT 20 (21.1) 1 (4.2) 0.18 (0.02–1.41)
TT 0 (0.0) 2 (8.3) 0
4 rs56292801 mir-5189 16q24 pre-miRNA
GG 51 (53.1) 18 (81.8) 1
0.010 (dom)AG 41 (42.7) 3 (13.6) 0.25 (0.08–0.80)
AA 4 (4.2) 1 (4.5)
5 rs2273626 mir-4707 14q11 seed
AA 31 (32.3) 2 (8.7) 1
0.013 (dom)AC 45 (46.9) 15 (65.2) 5.01 (1.10–22.72)
CC 20 (20.8) 6 (26.1)
6 rs6726779 mir-4431 2p16 pre-miRNA
TT 34 (35.8) 12 (63.2) 1
0.027 (codom)CT 51 (53.7) 4 (21.1) 0.22 (0.07–0.75)
CC 10 (10.5) 3 (15.8) 0.85 (0.20–3.62)
7 rs9877402 mir-5680 8q22 pre-miRNA
AA 88 (93.6) 17 (85.0) 1
0.030 (rec)AG 6 (6.4) 1 (5.0) 0
GG 0 (0.0) 2 (10.0)
8 rs243080 mir-4432 2p16 pre-miRNA
CC 30 (31.6) 12 (57.1) 1
0.030 (dom)CT 49 (51.6) 5 (23.8) 0.35 (0.13–0.91)
TT 16 (16.8) 4 (19.0)
9 rs3746444 mir-499a 20q11 seed
TT 64 (66.7) 18 (75.0) 1
0.044 (rec)CT 30 (31.2) 3 (12.5) 6.71 (1.06–42.73)
CC 2 (2.1) 3 (12.5)
Table 3. Polymorphisms in miRNAs associated with osteosarcoma risk in the Slovenian population.
*Signicant aer FDR correction. Abbreviations: OR, Odd Ratio; CI, Condence Interval; add, additive;
codom, codominant; dom, dominant; rec, recessive.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

www.nature.com/scientificreports/
6
SCIENTIFIC RePORts | (2018) 8:15414 | DOI:10.1038/s41598-018-33712-4
Discussion
In the Spanish population, the most interesting result was that 4 genetic variants in miRNAs belonging to the
14q32 miRNA cluster were statistically associated with the risk of osteosarcoma. From these, rs12894467 T allele
at miR-300 showed the most signicant result, conferring a 2.01-fold increased risk. is polymorphism was also
found signicant when Spanish and Slovenian populations were analyzed together, what means that it showed
the same trend in both cohorts (although it was not signicant in the Slovenian sample individually). e other
3 signicant SNPs of the cluster in the Spanish population (rs61992671, rs58834075 and rs12879262 at miR-412,
miR-656 and miR-4309, respectively) were also associated with an increased risk of osteosarcoma. Interestingly,
miRNAs of this cluster were found to be under-expressed in osteosarcoma in previous studies26,27. is miRNAs
downregulation was correlated with MYC overexpression, that it is known to be related to the development of
osteosarcoma27. e miRNAs down-expression was conrmed for mir-300 in a series of osteosarcoma cell lines
using GEO dataset GSE28423. Moreover, the block of 14q32 miRNAs was shown to increase the tumorigenic
potential in osteoblasts, suggesting that they could work as tumor suppressors. Consequently, the loss of function
of these miRNAs could be considered as a causative factor in osteosarcomagenesis27. Supporting this idea, the bio-
informatical analysis predicted that the SNPs in miR-412 and miR-656 decreased the stability of the miRNA hair-
pins, which has been suggested that may reduce the product of the mature miRNA28. is reduction in miRNA
levels could increase the expression of their target genes. Interestingly, pathway analyses pointed out the WNT
pathway as the most over-represented pathway, which is known to play an important role in osteoblastogenesis29.
Other authors have also pointed out the involvement of WNT pathway in the development of osteosarcoma30,31.
Dysregulation of Wnt signaling pathway allows β-catenin to accumulate and translocate into the nucleus, where
it activates downstream oncogenes including MYC32. Considering these previous studies, we can hypothesize
that variations in the pre-miRNAs miR-300, miR-412, miR-656 or miR-4309 could lead to their downregulation,
altering the Wnt pathway which ultimately would lead to the overexpression of MYC. All these results would
support the hypothesis that this region is a hotspot for the development of osteosarcoma. In fact, recent studies in
early-onset osteosarcoma have shown that inherited imprinting defects in14q32 region aects gene and miRNA
expression in this area, which could be associated with the pathobiology of osteosarcoma33.
Another interesting result in the Spanish population was found for rs35770269, located in the seed region
of miR-449c. In this case, the T allele was observed to decrease the risk of osteosarcoma (OR = 0.64). is allele
was proposed to alter the secondary structure of the miRNA (in silico), so the T allele could have a double action
in the miRNA, one aecting its levels and another, the miRNA-mRNA binding. Of note, miR-449c is part of the
highly conserved miR-449 cluster belonging to the miR-34 family34, a key regulator of tumor suppression35. SNPs
in the miR-34 family had already been found involved in the risk of osteosarcoma: rs4938723 C and rs72631823
A were associated with a reduction of miR-34b and miR-34a, respectively15,16. In addition, the underexpression
of miR-34a was shown to downregulate the suppression of the proto-oncogene C-MET, promoting osteosarcoma
cell proliferation and migration16. Since miRNAs belonging to the same family usually share target genes, we can
hypothesize that rs35770269 could aect the binding of miR-449c to MET.
e other 9 signicant miRNA variants detected in the Spanish population also showed a putative eect on
target genes with known involvement in osteosarcoma. For instance, rs77639117 T allele could increase the risk
of osteosarcoma through upregulating miR-576, which in turn might downregulate RB1, a tumor suppressor gene
inactivated in 35% of osteosarcoma patients1. e genotype rs2289030 GG could alter miR-492, aecting its target
PTEN. is gene was previously shown to be downregulated in osteosarcoma cells36–38. Rs6087195 could alter
the expression levels of miR-4467, which consequently could alter the expression of its putative target gene SF1,
involved in DNA reparation function39. In this case, the miRNA dysfunction could be explained by a modication
of the pre-miRNA secondary structure and a drastic energy change (3.9 Kcal/mol), which has been suggested to
aect the stability of the miRNA28.
In the Slovenian population, rs35613341 and rs56292801 (both located at miR-5189) showed the most
remarkable results. In this case, the signicant association was caused by a decrease of the percentage of hete-
rozygotes and an increase of the percentage of homozygotes. is fact suggests the presence of a deletion in this
region in which a copy number variation (CNV) (according to the database of Genomic Variations) has been
described. To the best of our knowledge, this is the rst time that this CNV is associated with osteosarcoma
risk. Another interesting nding was observed for rs3746444 located in the seed region of the pre-miR-499. e
GG genotype was associated with increased risk of osteosarcoma. Similar results were observed in two previous
meta-analyses studying the involvement of this polymorphism in cancer susceptibility in Caucasians (although
not signicant)40,41.
When both populations were analyzed together, a total of 6 SNPs increased the signicance of association
with respect to the individual analyses. ese results indicate that all these SNPs showed the same trend in both
populations, so they could be considered as disease markers. Among them, rs2910164 at miR-146a was previously
associated with diverse types of cancer42,43. is SNP was also analyzed in relation to the risk of osteosarcoma
in Chinese, showing the same trend as in our population (but it was not signicant)16. When a meta-analysis
including the three populations (Chinese, Spanish and Slovenian) was performed, a signicant association was
found under the dominant model (P = 0.003). e CG + CC rs2910164 genotype showed an OR = 0.57 (95% CI:
0.39–0.83) (Fig.S4). However, 5 SNPs decreased their signicance level, what means that opposite results were
detected in the two populations. is suggests that these SNPs are population specic, which indicates remarkable
population dierences in factors contributing to osteosarcoma risk.
is study has some limitations that might be addressed, such as the limited sample size. Nevertheless, consid-
ering the scarcity of the disease, we think that the number of patients included in the present study was enough
to obtain valid results. Another possible weakness of the study was the relatively high failure rate in genotyping
technique. However, this high chance of failure was accepted from the beginning, because despite the predicted
problem with the technique, no other design option to amplify these polymorphisms was possible.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

www.nature.com/scientificreports/
7
SCIENTIFIC RePORts | (2018) 8:15414 | DOI:10.1038/s41598-018-33712-4
In conclusion, the most important ndings of the present study indicated that SNPs located at the 14q32
miRNA cluster can be involved in the susceptibility of osteosarcoma in the Spanish population, conrming the
interest of this region in the disease. Our results also conrm the existence of population dierences in the risk of
developing osteosarcoma. To our knowledge, this is the rst study analyzing in depth so many SNPs at miRNAs
in relation with the risk of osteosarcoma, which opens a promising approach to search for new susceptibility
markers in this disease. New large-scale studies including functional analyses will help to validate our ndings.
Ethics approval and consent to participate. All procedures performed in studies involving human par-
ticipants were in accordance with the ethical standards of the institutional and/or national research committee
and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.
References
1. osenberg, A. E. et al. Conventional osteosarcoma. In: Fletcher, C. D. M., Bridge, J. A., Hogendoorn, P. C. W., Mertens, F., eds WHO
Classication of Tumours of So Tissue and Bone. 4th edn. Lyon, France: IAC Press, 282–8 (2013).
2. Hameed, M. & Dorfman, H. Primary malignant bone tumors–recent developments. Semin Diagn Pathol 28(1), 86–101 (2011).
3. Gianferante, D. M., Mirabello, L. & Savage, S. A. Germline and somatic genetics of osteosarcoma - connecting aetiology, biology and
therapy. Nat ev Endocrinol 13(8), 480–491 (2017).
4. Mirabello, L. et al. A comprehensive candidate gene approach identies genetic variation associated with osteosarcoma. BMC Cancer
11, 209 (2011).
5. Broadhead, M. L., Clar, J. C., Myers, D. E., Dass, C. . & Choong, P. F. e molecular pathogenesis of osteosarcoma: a review.
Sarcoma 2011, 959248 (2011).
6. Savage, S. A. et al. Genome-wide association study identies two susceptibility loci for osteosarcoma. Nat Genet 45(7), 799–803
(2013).
7. Cheetham, S. W., Gruhl, F., Mattic, J. S. & Dinger, M. E. Long noncoding NAs and the genetics of cancer. Br J Cancer 108(12),
2419–2425 (2013).
8. yan, B. M., obles, A. I. & Harris, C. C. Genetic variation in microNA networs: the implications for cancer research. Nat ev
Cancer 10(6), 389–402 (2010).
9. uov, J. L., Wilentzi, ., Jae, I., Vinther, J. & Shomron, N. Pharmaco-mi: lining microNAs and drug eects. Brief Bioinform
15(4), 648–659 (2014).
10. van Wijnen, A. J. et al. MicroNA functions in osteogenesis and dysfunctions in osteoporosis. Curr Osteoporos ep 11(2), 72–82
(2013).
11. Bae, Y. et al. miNA-34c regulates Notch signaling during bone development. Hum Mol Genet 21(13), 2991–3000 (2012).
12. elly, A. D. et al. MicroNA paran-based studies in osteosarcoma reveal reproducible independent prognostic proles at 14q32.
Genome Med 5(1), 2 (2013).
13. Xia, L. et al. Prognostic role of common microNA polymorphisms in cancers: evidence from a meta-analysis. PLoS One 9(10),
e106799 (2014).
14. Srivastava, . & Srivastava, A. Comprehensive review of genetic association studies and meta-analyses on miNA polymorphisms
and cancer ris. PLoS One 7(11), e50966 (2012).
15. Tian, Q. et al. A causal role for circulating mi-34b in osteosarcoma. Eur J Surg Oncol 40(1), 67–72 (2014).
16. Lv, H., Pei, J., Liu, H., Wang, H. & Liu, J. A polymorphism site in the pre-mi-34a coding region reduces mi-34a expression and
promotes osteosarcoma cell proliferation and migration. Mol Med e p 10(6), 2912–2916 (2014).
17. Shi, Z. W., Wang, J. L., Zhao, N., Guan, Y. & He, W. Single nucleotide polymorphism of hsa-mi-124a aects ris and prognosis of
osteosarcoma. Cancer Biomar 17(2), 249–57 (2016).
18. ozomara, A. & Griths-Jones, S. miBase: annotating high condence microNAs using deep sequencing data. Nucleic Acids es
42(Database issue), D68–73 (2014).
19. Sambroo, J. & ussell, D. Molecular cloning: a laboratory manual. ird edition ed. New Yor: Cold Spring Harbor (2001).
20. Benjamini, Y. & Hochberg, Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. Journal of
the oyal Statistical Society, Series B (Methodological) 57(1), 289–300 (1995).
21. Namløs, H. M. et al. Modulation of the osteosarcoma expression phenotype by microNAs. PLoS One 7(10), e48086 (2012).
22. Dweep, H. & Gretz, N. miWal2.0: a comprehensive atlas of microNA-target interactions. Nat Methods 12(8), 697 (2015).
23. amburov, A., Stelzl, U., Lehrach, H. & Herwig, . e ConsensusPathDB interaction database: 2013 update. Nucleic Acids es
41(Database issue), D793–800 (2013).
24. anehisa, M., Furumichi, M., Tanabe, M., Sato, Y. & Morishima, . EGG: new perspectives on genomes, pathways, diseases and
drugs. Nucleic Acids es 45(D1), D353–D361 (2017).
25. Fabregat, A. et al . e eactome pathway nowledgebase. Nucleic Acids es 44(D1), D481–7 (2018).
26. Maire, G. et al. Analysis of miNA-gene expression-genomic proles reveals complex mechanisms of microNA deregulation in
osteosarcoma. Cancer Genet 204(3), 138–146 (2011).
27. ayanithy, V. et al. Perturbation of 14q32 miNAs-cMYC gene networ in osteosarcoma. Bone 50(1), 171–181 (2012).
28. Gong, J. et al. Genome-wide identication of SNPs in microNA genes and the SNP eects on microNA target binding and
biogenesis. Hum Mutat 33(1), 254–63 (2012).
29. Angulo, P. et al. Natural compounds targeting major cell signaling pathways: a novel paradigm for osteosarcoma therapy. J Hematol
Oncol 10(1), 10 (2017).
30. Chen, C. et al. Aberrant activation of Wnt/β-catenin signaling drives proliferation of bone sarcoma cells. Oncotarget 10(6(19)),
17570–83 (2015).
31. Tian, J., He, H. & Lei, G. Wnt/β-catenin pathway in bone cancers. Tumour Biol 35(10), 9439–45 (2014).
32. Zou, Y., Yang, J. & Jiang, D. esveratrol inhibits canonical Wnt signaling in human MG-63 osteosarcoma cells. Mol Me d ep 12(5),
7221–6 (2015).
33. Shu, J. et al. Imprinting defects at human 14q32 locus alters gene expression and is associated with the pathobiology of osteosarcoma.
Oncotarget 7(16), 21298–314 (2016).
34. Yang, X. et al. mi-449a and mi-449b are direct transcriptional targets of E2F1 and negatively regulate pb-E2F1 activity through
a feedbac loop by targeting CD6 and CDC25A. Genes Dev 23(20), 2388–2393 (2009).
35. Misso, G. et al. Mi-34: a new weapon against cancer? Mol er Nucleic Acids 3, e194 (2014).
36. Tian, Z. et al. Upregulation of micro-ribonucleic acid-128 cooperating with downregulation of PTEN confers metastatic potential
and unfavorable prognosis in patients with primary osteosarcoma. Onco Targets er 7, 1601–1608 (2014).
37. Shen, L., Chen, X. D. & Zhang, Y. H. MicroNA-128 promotes proliferation in osteosarcoma cells by downregulating PTEN.
Tumour Biol 35(3), 2069–2074 (2014).
38. Gao, Y., Luo, L. H., Li, S. & Yang, C. mi-17 inhibitor suppressed osteosarcoma tumor growth and metastasis via increasing PTEN
expression. Biochem Biophys es Commun 444(2), 230–234 (2014).
Content courtesy of Springer Nature, terms of use apply. Rights reserved

www.nature.com/scientificreports/
8
SCIENTIFIC RePORts | (2018) 8:15414 | DOI:10.1038/s41598-018-33712-4
39. Fairman-Williams, M. E., Guenther, U. P. & Janowsy, E. SF1 and SF2 helicases: family matters. Curr Opin Struct Biol 20(3),
313–324 (2010).
40. Qiu, M. T. et al. Hsa-mi-499 rs3746444 polymorphism contributes to cancer ris: a meta-analysis of 12 studies. PLoS One 7(12),
e50887 (2012).
41. Fan, C., Chen, C. & Wu, D. e association between common genetic variant of microNA-499 and cancer susceptibility: a meta-
analysis. Mol Biol ep 40(4), 3389–3394 (2013).
42. Peng, S. et al. Association of microNA-196a-2 gene polymorphism with gastric cancer ris in a Chinese population. Dig Dis Sci
55(8), 2288–2293 (2010).
43. Xu, Z., Zhang, L., Cao, H. & Bai, B. Mi-146a rs2910164 G/C polymorphism and gastric cancer susceptibility: a meta-analysis. BMC
Med Genet 15, 117 (2014).
Acknowledgements
Special thanks to Slovenian Osteosarcoma Study Group for their collaboration in sample collection. The
“Slovenian Osteosarcoma Study Group” is conformed by Katja Goričar from the Institute of Biochemistry, Faculty
of Medicine of Ljubljana, Viljem Kovač from the Pharmacogenetics Laboratory, Institute of Biochemistry, Faculty
of Medicine of University of Ljubljana, Janez Jazbec from the Institute of Oncology Ljubljana, Janez Lamovec
from the Oncology and Hematology Unit, University Children’s Hospital, University Medical Centre of Ljubljana
and Prof. Vita Dolžan included in the authorship of this article. e authors would like to thank Leire Iparraguirre
for her technical assistance with gures. is study was funded by the Basque Government (IT661-13, IT989-16),
UPV/EHU (UFI11/35).
Author Contributions
A.G.O. and I.M.G. conceived and planned the experiments and N.B.A. performed the computations and B.S.Z.
performed the analytic calculations for miRNA expression analysis. I.M.G. and N.B.A. veried the analytical
methods. I.M.G. and N.B.A. wrote the manuscript with support from A.G.C. and A.G.O.V.D. and A.P.G.
helped supervise the project. All authors provided critical feedback and helped shape the research, analysis and
manuscript. All authors discussed the results and contributed to the nal manuscript.
Additional Information
Supplementary information accompanies this paper at https://doi.org/10.1038/s41598-018-33712-4.
Competing Interests: e authors declare no competing interests.
Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional aliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International
License, which permits use, sharing, adaptation, distribution and reproduction in any medium or
format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Cre-
ative Commons license, and indicate if changes were made. e images or other third party material in this
article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the
material. If material is not included in the article’s Creative Commons license and your intended use is not per-
mitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the
copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
© e Author(s) 2018
Content courtesy of Springer Nature, terms of use apply. Rights reserved

1.
2.
3.
4.
5.
6.
Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers and authorised users (“Users”), for small-
scale personal, non-commercial use provided that all copyright, trade and service marks and other proprietary notices are maintained. By
accessing, sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of use (“Terms”). For these
purposes, Springer Nature considers academic use (by researchers and students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and conditions, a relevant site licence or a personal
subscription. These Terms will prevail over any conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription
(to the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of the Creative Commons license used will
apply.
We collect and use personal data to provide access to the Springer Nature journal content. We may also use these personal data internally within
ResearchGate and Springer Nature and as agreed share it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not
otherwise disclose your personal data outside the ResearchGate or the Springer Nature group of companies unless we have your permission as
detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial use, it is important to note that Users may
not:
use such content for the purpose of providing other users with access on a regular or large scale basis or as a means to circumvent access
control;
use such content where to do so would be considered a criminal or statutory offence in any jurisdiction, or gives rise to civil liability, or is
otherwise unlawful;
falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association unless explicitly agreed to by Springer Nature in
writing;
use bots or other automated methods to access the content or redirect messages
override any security feature or exclusionary protocol; or
share the content in order to create substitute for Springer Nature products or services or a systematic database of Springer Nature journal
content.
In line with the restriction against commercial use, Springer Nature does not permit the creation of a product or service that creates revenue,
royalties, rent or income from our content or its inclusion as part of a paid for service or for other commercial gain. Springer Nature journal
content cannot be used for inter-library loans and librarians may not upload Springer Nature journal content on a large scale into their, or any
other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not obligated to publish any information or
content on this website and may remove it or features or functionality at our sole discretion, at any time with or without notice. Springer Nature
may revoke this licence to you at any time and remove access to any copies of the Springer Nature journal content which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or guarantees to Users, either express or implied
with respect to the Springer nature journal content and all parties disclaim and waive any implied warranties or warranties imposed by law,
including merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published by Springer Nature that may be licensed
from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a regular basis or in any other manner not
expressly permitted by these Terms, please contact Springer Nature at
onlineservice@springernature.com