Short abstract
Background
This study analyzed the associations between single nucleotide polymorphisms (SNPs) in the mutY homolog gene (MUTYH) and the calcium release-activated calcium channel gene (ORAI1) with susceptibility to osteoarthritis in the Chinese Han population.
Methods
A total of 350 patients diagnosed with osteoarthritis from October 2013 to May 2016 were selected as the study group, together with 350 age- and gender-matched healthy controls. MUTYH SNP rs3219463 and ORAI1 SNPs rs712853, rs12313273, rs6486795, rs12320939, and rs7135617 were analyzed by Sanger sequencing. Serum MUTYH levels were measured by enzyme-linked immunosorbent assay. The relationship between SNPs in MUTYH and ORAI1 and osteoarthritis susceptibility was analyzed and compared with the level of serum MUTYH in the osteoarthritis and control groups.
Results
MUTYH rs3219463 G allele carriers (GG or GA genotypes) and ORAI1 rs7135617 T allele carriers had a higher risk of osteoarthritis than patients with other genotypes. The level of serum MUTYH in the study group was significantly higher than in the control group (22.05 ± 19.14 ng/mL vs. 14.15 ± 13.54 ng/mL).
Conclusions
MUTYH and ORAI1 SNPs are associated with osteoarthritis susceptibility in the Chinese Han population.
Keywords: Osteoarthritis, MUTYH, single nucleotide polymorphisms, ORAI1, Chinese Han, Sanger sequencing
Introduction
Osteoarthritis (OA) is a joint disease with various manifestations including articular cartilage fibrosis, ulcers, pain, stiffness, and attrition. The etiology of OA is still largely unknown, but various studies have indicated that its occurrence is associated with age, obesity, joint injury, and genetic factors.1–3
Human MutY homolog (MUTYH) is a unique DNA glycosylase that excises adenine bases that have mispaired with 8-oxo-G during DNA replication. Dysfunctional MUTYH variants may lead to G:C to A:T transversions in the genome.4 Previous studies revealed that MUTYH inactivation caused by genetic polymorphisms plays a crucial role in cancer-associated inflammatory responses.5,6
Calcium release-activated calcium channel protein 1 (ORAI1), located on the plasma membrane, is an essential component of store-operated calcium channels (SOCCs). The interaction between ORAI1 and stromal interaction molecular 1 regulates intracellular calcium ion homeostasis through activating SOCCs and mediating their calcium entry.7,8 Additionally, the haplotypes of ORAI1 single nucleotide polymorphisms (SNPs) rs12313273 and rs7135617 are associated with the risk of developing HLA-B27-positive ankylosing spondylitis (AS).9 Kung et al. previously showed that MUTYH SNPs rs3219463 and rs3219476 were associated with susceptibility to rheumatoid arthritis (RA) in Taiwanese patients.10 Based on these results, we speculated that polymorphisms of MUTYH and ORAI1 might affect the risk of OA in the Chinese Han population, but little is known of this potential association. Therefore, we carried out a case–control study to examine the association between MUTYH and ORAI1 polymorphisms and susceptibility to OA in the Chinese Han population. Specifically, we examined five tagged SNPs (tSNPs) of ORAI1 (rs712853, rs12313273, rs6486795, rs12320939, and rs7135617) with minor allele frequencies (MAFs) >10% from the HapMap Han database, and the MUTYH SNP rs3219463.
Materials and methods
Subject selection
Patients with OA diagnosed at the General Hospital of Ningxia Medical University (Ningxia, China) from October 2013 to May 2016 were included in the study. Patients were diagnosed with OA according to published guidelines.11 The severity of OA was assessed using the Kallgren–Lawrence grading system; other types of arthritis were excluded from the study. Age- and gender-matched healthy control subjects were also recruited. Informed consent was obtained from all subjects and the study was approved by the Ethical Committee of the General Hospital of Ningxia Medical University.
Genomic DNA extraction and MUTYH and ORAI1 genotyping
A total of 10 mL of venous blood was collected from each subject. Genomic DNA was extracted from 2 mL blood using the QIAamp DNA Blood Mini Kit (Qiagen, Valencia, CA, USA). MUTYH and ORAI1 genotyping was performed by Sanger sequencing.12 Sequencing primers were synthesized as previously described.10,12
Quantitative determination of MUTYH in the serum
The remaining 8 mL of blood was used as a serum sample and diluted 200-fold in dilution buffer (5% skim milk, w/v). MUTYH levels were then quantified by enzyme-linked immunosorbent assay (ELISA) (USCN Life™ Science & Technology, Wuhan, China). Recombinant MUTYH protein (PeproTech Inc., Rocky Hill, NJ, USA) was used as the standard. The amount of MUTYH in the serum was determined according to the standard curve and dilution ratio.
Statistical analysis
IBM SPSS Statistics, version 20.0 (IBM Corp., Armonk, NY, USA) was used for statistical analysis. Statistical differences in genotypes and allelic frequencies between cases and controls were assessed using the chi-squared test. The association of each SNP with OA was assessed using multivariate logistic regression analysis. The linkage disequilibrium (LD) map of the five tSNPs in ORAI1 was constructed using Haploview software (version 4.2; http://www.broad.mit.edu/mpg/haploview/). The odds ratios (ORs) of genetic loci were calculated from genotype and allele frequencies with a 95% confidence interval (CI). Results were considered statistically significant when P values were less than 0.05.
Results
Subject characteristics
This study included a total of 350 patients with OA and 350 healthy controls. Subject characteristics are shown in Table 1. There was no significant difference in basal characteristics between the patients and controls.
Table 1.
Basal characteristics of patients with osteoarthritis and controls.
| Characteristics | Patients with OA (n = 350) |
Healthy controls (n = 350) |
t/χ2value | P value |
|---|---|---|---|---|
| Age (years) | 63.80 ± 11.52 | 64.10 ± 12.24 | 0.334 | 0.739 |
| Gender: male, n (%) | 50 (14.29%) | 55 (15.71%) | 0.280 | 0.597 |
| Course (years) | 6.10 ± 2.14 | – | ||
| Education: n (%) | 0.625 | 0.731 | ||
| Primary school or below | 75 (21.43%) | 71 (20.29%) | ||
| Middle school | 212 (60.57%) | 208 (59.43%) | ||
| Graduate and above | 63 (18.00%) | 71 (20.29%) | ||
| Obesity: n (%) | 189 (54.00%) | 191 (54.57%) | 0.023 | 0.879 |
| Heavy physical labor: n (%) | 98 (28.00%) | 120 (34.29%) | 3.224 | 0.073 |
| Smoking history: n (%) | 91 (26.00%) | 114 (32.57%) | 3.649 | 0.056 |
| Drinking history: n (%) | 134 (38.29%) | 158 (45.14%) | 3.384 | 0.066 |
Association between ORAI1 rs7135617, MUTYH rs3219463, and the susceptibility of OA
We selected five tSNPs of ORAI1 (rs712853, rs12313273, rs6486795, rs12320939, and rs7135617) with MAFs >10% from the HapMap Han database. Genotype and allele frequencies were consistent with the Hardy–Weinberg equilibrium (HWE). A comparison of the difference in genotype and allele frequencies of SNPs between patients and controls showed that individuals carrying the rs7135617 TT genotype had a significantly higher risk of developing OA than individuals with other genotypes (OR = 1.332, 95%CI = 1.069–1.661, P =0.01) (Table 2).
Table 2.
Genotype and allele frequencies of ORAI1 and MUTYH and the risk of OA.
| ORAI1 | Genotype | Patients with OA (n = 350) |
Healthy controls (n = 350) |
OR (95% CI) |
P
value |
Adjusted P value |
|---|---|---|---|---|---|---|
| rs712853 | ||||||
| CC | 34 | 37 | 0.875 (0.507–1.509) | 0.609 | 0.704 | |
| CT | 150 | 155 | 0.921 (0.665–1.275) | 0.607 | 0.663 | |
| TT | 166 | 158 | ref | |||
| Allele | ||||||
| C | 218 | 229 | 0.93 (0.738–1.172) | 0.528 | 0.566 | |
| T | 482 | 471 | ref | |||
| rs12313273 | ||||||
| CC | 27 | 30 | 0.890 (0.490–1.614) | 0.683 | 0.791 | |
| CT | 142 | 141 | 0.894 (0.487–1.640) | 0.699 | 0.809 | |
| TT | 181 | 179 | ref | |||
| Allele | ||||||
| C | 196 | 202 | 0.959 (0.755–1.218) | 0.722 | 0.767 | |
| T | 504 | 498 | ref | |||
| rs6486795 | ||||||
| CC | 52 | 47 | 1.090 (0.670–1.775) | 0.714 | 0.804 | |
| CT | 163 | 170 | 0.945 (0.676–1320) | 0.729 | 0.791 | |
| TT | 135 | 133 | ref | |||
| Allele | ||||||
| C | 267 | 264 | 1.018 (0.816–1.271) | 0.869 | 0.912 | |
| T | 433 | 436 | ref | |||
| rs12320939 | ||||||
| TT | 85 | 82 | 1.003 (0.644–1.562) | 0.990 | 1 | |
| TG | 173 | 179 | 0.935 (0.643–1.360) | 0.713 | 0.782 | |
| GG | 92 | 89 | ref | |||
| Allele | ||||||
| T | 344 | 342 | 1.011 (0.816–1.254) | 0.915 | 0.957 | |
| G | 356 | 358 | ref | |||
| rs7135617 | ||||||
| TT | 79 | 54 | 1.728 (1.112–2.686) | 0.010 | 0.014 | |
| TG | 144 | 146 | 1.165 (0.826–1.643) | 0.364 | 0.411 | |
| GG | 127 | 150 | ref | |||
| Allele | ||||||
| T | 302 | 254 | 1.332 (1.069–1.661) | 0.009 | 0.010 | |
| G | 398 | 446 | ref | |||
| MUTYH | ||||||
| rs3219463 | ||||||
| AA | 51 | 45 | 1.792 (1.091–2.945) | 0.014 | 0.020 | |
| AG | 194 | 139 | 2.207 (1.569–3.104) | 0.000 | 0.000 | |
| GG | 105 | 166 | ref | |||
| Allele | ||||||
| A | 297 | 230 | 1.506 (1.204–1.83) | 0.000 | 0.000 | |
| G | 403 | 470 | ref |
ref, reference.
The genotype and allele frequency distribution of MUTYH SNP rs3219463 is shown in Table 2. The frequencies were consistent with the HWE. Individuals with the rs3219463 GA or GG genotype had a significantly higher risk of developing OA than individuals with other genotypes (OR =1.506, 95%CI = 1.204–1.830, P = 0.000).
Haplotype analysis of ORAI1 SNPs in OA susceptibility
To further confirm the existence of a correlation between ORAI1 haplotypes and OA, we constructed an LD map (Figure 1) and estimated the differences of haplotype frequencies between OA patients and controls.
Figure 1.
Linkage disequilibrium map of five ORAI1 tSNPs.
As shown in Table 3, the haplotype analysis revealed that there was no significant correlation in rs12313273/rs7135617 or rs7135617/rs6486795 pairwise allelic comparisons.
Table 3.
Haplotype frequency of the ORAI1 gene in patients with OA and normal controls.
| Haplotype | Patients with OA (n=350) |
Healthy controls (n=350) |
OR (95% CI) | P value |
|---|---|---|---|---|
| rs12313273/rs7135617 | ||||
| T/T | 147 (42.00%) | 143 (40.86%) | 1.093 (0.748–1.597) | 0.698 |
| T/G | 106 (30.29%) | 106 (30.29%) | 1.063 (0.707–1.598) | 0.834 |
| C/G | 95 (27.14%) | 101 (28.86%) | ref | |
| rs7135617/rs6486795 | ||||
| T/T | 144 (41.14%) | 142 (40.57%) | 1.006 (0.710–1.427) | 0.971 |
| G/T | 69 (19.71%) | 75 (21.43%) | 0.913 (0.596–1.400) | 0.661 |
| G/C | 132 (37.71%) | 131 (37.43%) | ref |
Haplotype frequencies <1% were excluded. ref, reference.
MUTYH expression levels in the serum
Serum levels of MUTYH protein from 350 OA patients and 350 controls were detected by ELISA. The mean MUTYH level of OA patients was significantly higher than that of healthy controls (Figure 2, P < 0.005). Individuals carrying mutant rs3219463 (GA+GG) genotypes also had significantly higher serum MUTYH levels than those with the wild-type genotype (AA) (Figure 2, P < 0.05).
Figure 2.
MUTYH expression levels in the serum of OA patients and controls. P < 0.05 represents statistical significance.
Discussion
OA is a chronic disease that can lead to the total degeneration of articular cartilage structure and function. The development of OA may also result in changes to other soft tissue parts of the body.13–15 The most common symptoms include joint pain, swelling, stiffness, and functional limitations.16 OA can be induced by a number of predisposing causes, including both occupational and non-occupational factors. Coggon et al.17 showed that individuals who kneeled or squatted for prolonged periods had a higher risk of developing knee OA, which was also associated with obesity.18 Moreover, Valdes et al.19 analyzed the genotypes of 36 SNPs in 17 candidate genes, and found that multiple genetic variants could predict the risk of knee OA. Thus, genetic factors appear to be critical for the development of OA.
OA is traditionally thought to be associated with degeneration and trauma. However, increasing evidence suggests that it is not a single disease caused by aging or biological stress, but rather the accumulation of multiple factors such as metabolic disorders and inflammatory senescence. Lipid metabolism and body fluids are also involved in the initiation and progression of disease. Because the systemic inflammatory response associated with RA is believed to play an important role in the development of AS, it appears that OA, RA, and AS may share common features regarding the inflammatory response, while the incidence of OA and RA may be related to calcium signaling. Moreover, MUYTH and ORAI1 SNPs have been significantly associated with the risk of chronic RA.6,12 However, few studies have investigated the role of MUYTH and ORAI1 SNPs in the susceptibility to OA in the Chinese Han population.
Here, we performed a case–control association study of MUTYH SNP rs3219463 and five ORAI1 tSNPs (rs712853, rs12313273, rs6486795, rs12320939, and rs7135617). Gender was previously shown to be one of the key factors affecting the risk of OA,20 so the 350 controls were age- and gender-matched to the 350 OA patients.
Other than the association between MUTYH SNPs and RA risk, MUTYH was also chosen as a target gene in this study because of the reported correlation between its polymorphisms and cancer-linked inflammatory responses.21,22 Indeed, tumor inflammatory responses including tumor necrosis factor-α and interleukin-1β are important factors during OA development.23,24
The Yen group previously reported that only the MUTYH SNP rs3219463 was relevant to the development of RA, so we focused on the association between rs3219463 and OA. As shown in Table 2, individuals with the rs3219463 G allele (GA or GG genotypes) had a higher risk of developing OA than those with other genotypes. We also found that individuals with GA or GG rs3219463 genotypes had higher serum MUTYH levels than wild-type individuals with the AA genotype. Although it is not obvious why MUTYH accumulates in the serum, we propose that this might reflect the intracellular function of MUTYH.25 We believe that SNP rs3219463 affects the expression level of MUTYH, but more in vitro studies will be needed to confirm this hypothesis.
The ORAI1-mediated calcium signaling pathway is associated with a range of diseases. ORAI1 mutations may lead to disruption of the store-operated Ca2+ influx and the suppression of immune responses.26 A number of ORAI1 SNPs have been reported to be associated with human disease, including the correlation of rs7135617 with AS,9 while ORAI1 polymorphisms were shown to influence calcium signal transduction by affecting protein splicing.27 We found that individuals with the rs7135617 TT genotype had a higher risk of developing OA. This may be because rs7135617 polymorphisms affect calcium signal transduction, causing Ca2+ signaling abnormalities during OA development.28 Wen et al.29 revealed differences in the characteristics of Ca2+ signals in different layers of normal cartilage tissue, which are regulated by extracellular Ca2+ concentrations. Moreover, the Ca2+ signal in OA chondrocytes differs from that in normal cells. Therefore, abnormalities of OA cartilage tissue may be linked to Ca2+ signal abnormalities, which could reflect differences in the expression of genes such as MUTYH and ORAI1. In our study, ORAl1 rs7135617 T allele carriers and MUTYH rs3219463 G allele carriers had abnormal cartilage tissue Ca2+ signal responses, which would likely result in the dysfunction of OA cartilage cell voltage-regulated calcium channels and calcium-release activated channels. Additionally, Berenbaum et al.30 showed that inflammatory cytokines play an important role in the pathogenesis of OA.
Our study has a number of limitations. First, we were unable to collect clinical biochemical data from OA patients, such as ORAI1 expression levels, so we only focused on the association between SNPs and the risk of OA. Second, we did not perform an in vitro analysis of the mechanisms affecting OA-associated risk factors. Finally, the small sample size of this study meant that no distinction could be made among individuals of different nationalities in China. Therefore, further study is required to confirm whether these SNPs are affected by different ethnic groups or sites of OA.
In conclusion, MUTYH SNP rs3219463 and ORAI1 SNP rs7135617 appear to be associated with susceptibility to OA in the Chinese Han population. The SNPs may affect protein expression and function, which in turn affect the Ca2+ signaling pathway, resulting in disease development.
Declaration of conflicting interest
The authors declare that there is no conflict of interest.
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
References
- 1.Cross M, Smith E, Hoy D, et al. The global burden of hip and knee osteoarthritis: estimates from the Global Burden of Disease 2010 study. Ann Rheum Dis 2014; 73: 1323–1330. [DOI] [PubMed] [Google Scholar]
- 2.DeGroot J, Verzijl N, Wenting-Van Wijk MJG, et al. Accumulation of advanced glycation end products as a molecular mechanism for aging as a risk factor in osteoarthritis. Arthritis Rheum 2004; 50: 1207–1215. [DOI] [PubMed] [Google Scholar]
- 3.Sharma L, Lou C, Cahue S, et al. The mechanism of the effect of obesity in knee osteoarthritis: the mediating role of malalignment. Arthritis Rheum 2000; 43: 568–575. [DOI] [PubMed] [Google Scholar]
- 4.Banda DM, Nunez NN, Burnside MA, et al. Repair of 8-oxoG:A mismatches by the MUTYH glycosylase: Mechanism, metals and medicine. Free Radic Biol Med 2017; 107: 202–215. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Tsai-Wu JJ, Su HT, Wu YL, et al. Nuclear localization of the human mutY homologue hMYH. J Cell Biochem 2000; 77: 666–677. [PubMed] [Google Scholar]
- 6.Chen SY, Chen HH, Huang YC, et al. Polymorphism and protein expression of MUTYH gene for risk of rheumatoid arthritis. BMC Musculoskelet Disord 2017; 18: 69. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Wei-Lapierre L, Carrell EM, Boncompagni S, et al. Orai1-dependent calcium entry promotes skeletal muscle growth and limits fatigue. Nat Commun 2013; 4: 2805. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Zhang W, Qi Z, Wang Y. BTP2, a store-operated calcium channel inhibitor, attenuates lung ischemia-reperfusion injury in rats. Inflammation 2017; 40: 778–787. [DOI] [PubMed] [Google Scholar]
- 9.Wei JC, Yen JH, Juo SH, et al. Association of ORAI1 haplotypes with the risk of HLA-B27 positive ankylosing spondylitis. PLoS One 2011; 6: e20426. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Kung YJ, Tsai KS, Huang CM, et al. MUTYH gene polymorphisms as risk factors for rheumatoid arthritis. Biomed Res Int 2015; 2015: 893796. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Decker JL, Weiss TE, Howell DS, et al. American rheumatism association. Arthritis Rheum 1973; 16: 113–138. [Google Scholar]
- 12.Yen JH, Chang CM, Hsu YW, et al. A polymorphism of ORAI1 rs7135617, is associated with susceptibility to rheumatoid arthritis. Mediators Inflamm 2014; 2014: 834831. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Westra J, Ashbeck EL, Guermazi A, et al. Burden of MRI-detected pathological features and future trajectories of pain and function in participants with radiographic knee osteoarthritis. Osteoarthr Cartil 2017; 25: S372–S373. [Google Scholar]
- 14.Dequeker J, Mokassa L, Aerssens J, et al. Bone density and local growth factors in generalized osteoarthritis. Microsc Res Tech 1997; 37: 358–371. [DOI] [PubMed] [Google Scholar]
- 15.Glyn-Jones S, Palmer AJR, Agricola R, et al. Osteoarthritis. Lancet 2015; 386: 376–387. [DOI] [PubMed] [Google Scholar]
- 16.Chang NJ, Shie MY, Lee KW, et al. Can early rehabilitation prevent posttraumatic osteoarthritis in the patellofemoral joint after anterior cruciate ligament rupture? Understanding the pathological features. Int J Mol Sci 2017; 18: pii: E829. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Coggon D, Croft P, Kellingray S, et al. Occupational physical activities and osteoarthritis of the knee. Arthritis Rheum 2000; 43: 1443–1449. [DOI] [PubMed] [Google Scholar]
- 18.Yucesoy B, Charles LE, Baker B, et al. Occupational and genetic risk factors for osteoarthritis: a review. Work 2015; 50: 261–273. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Valdes AM, Doherty M, Spector TD. The additive effect of individual genes in predicting risk of knee osteoarthritis. Ann Rheum Dis 2008; 67: 124–127. [DOI] [PubMed] [Google Scholar]
- 20.Gosvig KK, Jacobsen S, Sonne-Holm S, et al. Prevalence of malformations of the hip joint and their relationship to sex, groin pain, and risk of osteoarthritis: a population-based survey. J Bone Joint Surg Am 2010; 92: 1162–1169. [DOI] [PubMed] [Google Scholar]
- 21.Yamaguchi S, Ogata H, Katsumata D, et al. MUTYH-associated colorectal cancer and adenomatous polyposis. Surg Today 2014; 44: 593–600. [DOI] [PubMed] [Google Scholar]
- 22.Mazzei F, Viel A, Bignami M. Role of MUTYH in human cancer. Mutat Res 2013; 743–744: 33–43. [DOI] [PubMed] [Google Scholar]
- 23.Maksymowych WP, Russell AS, Chiu P, et al. Targeting tumour necrosis factor alleviates signs and symptoms of inflammatory osteoarthritis of the knee. Arthritis Res Ther 2012; 14: R206. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Ye W, Zhou J, Markova DZ, et al. Xylosyltransferase-1 Expression Is Refractory to Inhibition by the Inflammatory Cytokines Tumor Necrosis Factor α and IL-1β in Nucleus Pulposus Cells: Novel Regulation by AP-1, Sp1, and Sp3. Am J Pathol 2015; 185: 485–495. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Hayashi H, Tominaga YS, Mckenna AE, Nakabeppu Y, Matsumoto Y. Replication-associated repair of adenine:8-oxoguanine mispairs by MYH. Curr Biol 2002; 12: 335–339. [DOI] [PubMed] [Google Scholar]
- 26.Feske S, Gwack Y, Prakriya M, et al. A mutation in Orai1 causes immune deficiency by abrogating CRAC channel function. Nature 2006; 441: 179. [DOI] [PubMed] [Google Scholar]
- 27.Liu HX, Chew SL, Cartegni L, et al. Exonic splicing enhancer motif recognized by human SC35 under splicing conditions. Mol Cell Biol 2000; 20: 1063. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.He Z, Leong DJ, Zhuo Z, et al. Strain-induced mechanotransduction through primary cilia, extracellular ATP, purinergic calcium signaling, and ERK1/2 transactivates CITED2 and downregulates MMP-1 and MMP-13 gene expression in chondrocytes. Osteoarthritis Cartilage 2016; 24: 892–901. [DOI] [PubMed] [Google Scholar]
- 29.Wen JB, Gong XY, Yang L, et al. Changes of spontaneous Ca2 + signal in chondrocytes of osteoarthritis knee joint in situ. Zhonghua Knee Arthritis: Electronic Edition 2016; 10: 319–325. [Google Scholar]
- 30.Berenbaum F, Eymard F, Houard X. Osteoarthritis, inflammation and obesity. Curr Opin Rheumatol 2013; 25: 114–118. [DOI] [PubMed] [Google Scholar]