Objective
We sought to comprehensively evaluate the association of laminin gamma-1 ( LAMC1 ) and advance pelvic organ prolapse.
Study Design
We conducted a candidate gene association of patients (n = 239) with stages III-IV prolapse and controls (n = 197) with stages 0-I prolapse. We used a linkage disequilibrium (LD)–tagged approach to identify single-nucleotide polymorphisms (SNPs) in LAMC1 and focused on non-Hispanic white women to minimize population stratification. Additive and dominant multivariable logistic regression models were used to test for association between individual SNPs and advanced prolapse.
Results
Fourteen SNPs representing 99% coverage of LAMC1 were genotyped. There was no association between SNP rs10911193 and advanced prolapse ( P = .34). However, there was a trend toward significance for SNPs rs1413390 ( P = .11), rs20563 ( P = .11), and rs20558 ( P = .12).
Conclusion
Although we found that the previously reported LAMC1 SNP rs10911193 was not associated with nonfamilial prolapse, our results support further investigation of this candidate gene in the pathophysiology of prolapse.
Pelvic organ prolapse is a major public health issue for women because 40% of postmenopausal women are affected. Symptomatic prolapse results in a marked reduction in quality of life because of pelvic discomfort and bladder and bowel dysfunction. Although prolapse can be managed surgically, the recurrence rate is as high as 30%, and the annual health care costs of surgery alone total more than a billion dollars in the United States. Given the public health impact of prolapse, it is critical to understand the underlying pathophysiology of this disease to develop effective prevention strategies and specific treatment recommendations for high-risk women.
Currently, the underlying mechanisms of prolapse are poorly understood; however, a growing body of evidence supports a genetic predisposition. Epidemiologic studies suggest that a family history of prolapse is a significant risk factor for disease. A pedigree analysis of familial prolapse found that inheritance occurs in an autosomal dominant fashion with incomplete penetrance, and a linkage analysis of affected sibling pairs identified a predisposition gene for pelvic floor disorders on chromosome 9q21.
Hypotheses regarding the pathophysiology of prolapse include abnormal synthesis or degradation of collagen and elastin; thus, genetic mutations in extracellular matrix (ECM) pathways may result in the genesis and progression of prolapse. Nikolova et al identified a single-nucleotide polymorphism (SNP) (rs10911193) in the LAMC1 gene encoding laminin, an ECM component in a study of familial prolapse. However, when studying nonfamilial prolapse in a case-control association study of 265 Caucasians and 146 African Americans, Chen et al reported that this laminin mutation and 2 additional functional SNPs in LAMC1 (rs20563 and rs20558) were not associated with advanced prolapse. One limitation of the study by Chen et al may have been the smaller sample size.
Given the significant findings of LAMC1 in familial prolapse by Nikolova et al, we felt that further investigation of LAMC1 as a potential candidate gene for advanced pelvic organ prolapse was warranted in a larger cohort of unrelated individuals. In this study, we sought to perform a comprehensive analysis of the entire LAMC1 gene, using a linkage disequilibrium (LD)-tagged SNP approach, in a larger case-control association study of non-Hispanic white women.
Materials and Methods
Study population
After institutional review board approval, we recruited subjects from the Division of Urogynecology at Duke University Medical Center for this case-control association study from February 2009 to December 2010. Patients were defined as nonpregnant women with stages III-IV prolapse based on the Pelvic Organ Prolapse Quantification (POP-Q) examination. Controls were women with stages 0-I prolapse who had no prior history of prolapse surgery. Women with a prior hysterectomy for nonprolapse indications were eligible to be controls.
We excluded women with a gynecological malignancy or inherited or acquired connective tissue diseases, including rheumatoid arthritis, systemic lupus erythematosus, scleroderma, polymyositis, dermatomyositis, or Marfan or Ehlers-Danlos syndromes from both groups. Sociodemographic data, past medical and obstetric history, and physical examination information were collected, which included age, race, ethnicity, parity, body mass index (BMI), and POP-Q points.
Although we recruited subjects of all ages with advanced prolapse, one approach commonly used in genetic epidemiology studies is to preferentially enroll younger subjects with more advanced disease to select for a high-risk group with a presumably higher genetic load. In this study, we preferentially recruited younger women with advanced prolapse for our patients, and for our control group, we preferentially recruited older women. Because the prevalence of prolapse increases with age, older controls help to ensure that these individuals are not at risk for developing prolapse. By selecting these more extreme cases, we attempted to enhance our ability to detect genetic differences between our 2 groups.
To avoid issues of confounding by race, also known as population stratification, we focused on non-Hispanic white women in this study. By focusing on non-Hispanic white women, our goal was to further ensure that our cases and controls were derived from the same source population. In other words, we matched on race and ethnicity to minimize the risk of differences in the genetic background of our 2 cohorts. It is critically important to consider racial and ethnic groups in genetic epidemiologic studies because different racial/ethnic groups can have markedly different population frequencies of genetic variants as well as different linkage disequilibrium structures.
Genotyping and SNP selection
Genomic deoxyribonucleic acid was extracted from frozen stored peripheral venous whole-blood samples. To enable a more comprehensive evaluation of the LAMC1 gene, we utilized an LD tagged SNP approach. A tagged SNP approach is an efficient methodology for comprehensive evaluation of a gene because most SNPs are in LD (ie, genetically linked and thus correlated) with other SNPs. That is, within each gene, there are groups of SNPs that are highly correlated with each other. Representative SNPs from each group (ie, tagged SNPs) can be assessed, which represents a detailed evaluation of the gene without having to genotype every SNP. Thus, a tagged SNP approach is comprehensive as well as efficient, in terms of time and cost.
We utilized the Tagger subsection of Haploview 4.2 software (Broad Institute, Cambridge, MA) to identify tagged SNPs in LAMC1 as well as in the 5 kb upstream and downstream of LAMC1 and used a correlation (r ) of greater than 0.8 and SNPs with a minor allele frequency (MAF) of greater than 0.05. We also included known and putative functional coding SNPs with an MAF greater than 0.05 from any given LD bin.
Genotyping was performed using the 7900 HT TaqMan system, which incorporates a standard PCR-based, dual-fluorescence, allele discrimination assay with a dual-laser scanner. Assays by Design and Assays on Demand were purchased through Applied Biosystems (Applied Biosystems, Foster City, CA). Quality control (QC) samples, composed of 12 reference controls, were included in each quadrant of the plate. All SNPs examined were successfully genotyped for 98% or greater of the individuals in the study. Error rate estimates for SNPs meeting QC benchmarks were 0.2%.
Statistical analysis
Sociodemographic data and physical examination findings for cases and controls were compared using a χ 2 test for discrete variables and a Student t test for continuous variables. A 2-sided alpha of 0.05 was considered significant. For all SNPs, deviation from Hardy-Weinberg equilibrium (HWE) was tested using a χ 2 test. Genotype clusters and genotyping calling for SNPs that were significantly out of HWE were re-reviewed and genotyping accuracy confirmed. To test for association of single SNPs with pelvic organ prolapse, we constructed logistic regression models and evaluated both unadjusted and adjusted models, using both additive and dominant models of genetic risk. The multivariable models adjusted for age, vaginal parity (dichotomous variable regarding any prior vaginal deliveries), and BMI.
A full-tagged SNP approach was conducted in an initial cohort of 396 non-Hispanic white women. Because 3 SNPs showed a nominal P value of .10, we expanded the final cohort to 436 women to increase the power. Given these independent effects of several SNPs, we also performed a haplotype analysis of all 2-way SNP combinations using the tool set of PLINK to evaluate the possible association of a genetic background of more than 1 SNP for pelvic organ prolapse risk.
A power calculation was conducted for a nonfamilial case-control association using the QUANTO program (Gauderman and Morrison, http://hydra.usc.edu/gxe . Accessed Oct. 24, 2008.). A range of allele frequencies of 0.05-0.5 and varying effect sizes as well as dominant and additive models were considered. We recognize that these calculations are based on model parameters that we cannot know with certainty; however, we can derive a sense of the expected power for the proposed sample sizes. Given a sample size of 150 cases and 150 controls, we have greater than 80% power to detect an effect size of approximately 2.0 and higher for most allele frequencies and genetic models. We chose an effect size of 2.0 because we believed that this is a clinically meaningful difference to detect and represented a balance between clinical/biological significance and feasibility. All statistical analyses were conducted with SAS version 9.2 (SAS Institute, Cary, NC) unless otherwise noted.
Results
We enrolled a total of 436 women, of whom 239 were cases with advanced prolapse and 197 were controls with none to minimal prolapse. Cases were slightly younger (64.8 ± 10.3 vs 69.0 ± 10.2 years, P < .001) and more likely to have had 1 or more vaginal deliveries (96.6% vs 82.2%, P < .001) when compared with controls ( Table 1 ) . Cases and controls were similar in the Charlson comorbidity index, smoking status, and BMI (27.8 ± 5.1 vs 28.5 ± 6.3 kg/m 2 , P = .18). Of controls, 57.4% had stage 0 prolapse and 42.6% had stage I prolapse. For cases, a majority had stage III prolapse (94.6%) and 5.4% had stage IV prolapse.
| Characteristic | Controls (n = 197) | Cases (n = 239) | P value |
|---|---|---|---|
| Age, y | 69.0 ± 10.2 | 64.8 ± 10.3 | < .001 |
| Vaginal parity | 162 (82.2) | 230 (96.6) | < .001 |
| Charlson comorbidity index | 0.48 ± 0.90 | 0.42 ± 0.83 | .51 |
| Current smoker | 8 (4.1) | 16 (6.8) | .26 |
| Body mass index, kg/m 2 | 28.5 ± 6.3 | 27.8 ± 5.1 | .18 |
| Prolapse stage | |||
| 0 | 113 (57.4) | 0 (0.0) | |
| I | 84 (42.6) | 0 (0.0) | |
| II | 0 (0.0) | 0 (0.0) | |
| III | 0 (0.0) | 226 (94.6) | |
| IV | 0 (0.0) | 13 (5.4) |
We genotyped 15 SNPs based on our LD tagging selection. The Figure is a schematic of the LAMC1 gene structure, which indicates exons, introns, and sites of translation initiation and termination. The genotyped SNPs are marked on the schematic with descriptions of the different types of SNPs, whether nonsynonymous coding, synonymous coding, or intronic. All SNPs were in Hardy-Weinberg equilibrium except rs17477864 ( P < .001). This SNP was regenotyped and remained out of HWE and thus was not included in the final analysis. At an r of 0.80, the remaining 14 SNPs provided 99% coverage of the gene.
In this cohort of non-Hispanic white women, there was no association between the previously reported LAMC1 SNP rs10911193 with advanced pelvic organ prolapse, based on both the dominant and additive adjusted models (odds ratio [OR], 1.29; 95% confidence interval [CI], 0.77–2.17; P = .34) and OR, 1.23; 95% CI, 0.77–1.97; P = .39, respectively) ( Table 2 ) . However, there was a trend toward significance for the following SNPs in the dominant model: rs1413390, an intronic SNP (adjusted OR, 1.45; 95% CI, 0.92–2.29; P = .11), rs20558; a nonsynonymous coding SNP (adjusted OR, 1.44; 95% CI, 0.91–2.26; P = .12); and rs20563, a nonsynonymous coding SNP (adjusted OR, 1.44; 95% CI, 0.92–2.28; P = .11) ( Table 2 ). These 3 SNPs were in perfect linkage disequilibrium with each other (r =1.0) ( Figure ). No other SNPs in LAMC1 were associated with prolapse in either the dominant or additive adjusted models.
| SNP | Minor allele (frequency) | Dominant adjusted model | Additive adjusted model | ||
|---|---|---|---|---|---|
| OR (95% CI) | P value | OR (95% CI) | P value | ||
| rs10911193 | T (0.11) | 1.29 (0.77–2.17) | .34 | 1.23 (0.77–1.97) | .39 |
| rs1413390 | G (0.45) | 1.45 (0.92–2.29) | .11 | 1.22 (0.90–1.65) | .21 |
| rs20558 | T (0.45) | 1.44 (0.91–2.26) | .12 | 1.19 (0.88–1.61) | .25 |
| rs20563 | A (0.46) | 1.44 (0.92–2.28) | .11 | 1.20 (0.89–1.63) | .23 |
| rs10911206 | A (0.48) | 1.37 (0.85–2.20) | .19 | 1.16 (0.85–1.57) | .35 |
| rs2296291 | C (0.45) | 1.36 (0.85–2.16) | .20 | 1.14 (0.84–1.56) | .39 |
| rs12041030 | G (0.11) | 1.36 (0.80–2.29) | .25 | 1.28 (0.80–2.07) | .30 |
| rs12739316 | T (0.41) | 0.79 (0.51–1.23) | .30 | 0.82 (0.60–1.11) | .20 |
| rs3768617 | T (0.42) | 0.83 (0.53–1.31) | .42 | 0.84 (0.61–1.16) | .29 |
| rs2483675 | A (0.22) | 1.15 (0.76–1.74) | .51 | 1.17 (0.82–1.68) | .38 |
| rs10911211 | T (0.40) | 0.89 (0.57–1.39) | .62 | 0.90 (0.66–1.24) | .52 |
| rs41475048 | T (0.25) | 1.04 (0.69–1.56) | .86 | 1.08 (0.77–1.50) | .66 |
| rs1058177 | C (0.09) | 1.01 (0.58–1.76) | .97 | 1.01 (0.58–1.76) | .97 |
| rs12073936 | G (0.08) | 1.00 (0.56–1.78) | .99 | 0.96 (0.55–1.68) | .88 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
