Polymorphisms of mitochondrial DNA control region are associated to endometriosis

Abstract

Purpose

Polymorphisms in the control region of mitochondrial DNA (mtDNA) can affect generation of reactive oxygen species and impact in the pathogenesis of endometriosis. This study investigated the association of mtDNA polymorphisms with endometriosis.

Methods

Patients were divided in two groups: endometriosis (n = 90) and control (n = 92). Inclusion criteria were as follows: women between 18 and 50 years, with histological diagnosis and surgical staging of endometriosis (endometriosis group) or undergoing gynecological surgery for tubal ligation, leiomyoma, or ovarian cysts, with no evidence of endometriosis (control group). DNA extraction was performed from peripheral blood. Sanger sequencing of mtDNA control region was performed, and polymorphisms were determined comparing the sequences obtained with the Cambridge Reference Sequence.

Results

The frequency of polymorphisms T16217C (14.4 and 5.4% of endometriosis and control group, respectively; p = 0.049) and G499A (13.3 vs. 4.3%; p = 0.038) was higher in the endometriosis group, while T146C (32.6 vs. 18.9%; p = 0.042) and 573.2C (5.6 vs. 29.3%; p < 0.001) were lower. No difference was observed in haplogroups between groups.

Conclusion

mtDNA polymorphisms T16217C and G499A were associated with endometriosis, while T416C and 573.2C were shown to be associated with an absence of disease.

Electronic supplementary material

The online version of this article (10.1007/s10815-017-1082-4) contains supplementary material, which is available to authorized users.

Introduction

Endometriosis is characterized by the presence of an endometrium outside of the uterine cavity. A recent cohort study of 58,427 women of reproductive age estimated the prevalence of laparoscopically confirmed endometriosis to be 6% overall and up to 16% among infertile women [1]. While the mechanisms underlying the pathogenesis of this disease are not well-understood, the fact that it occurs more frequently in individuals with an affected twin or family member is suggestive of a genetic component [2]. Furthermore, previous studies have suggested that endometriosis may be associated with an imbalance between the oxidative and anti-oxidative activities of the inflammatory system, which results in the increased production of reactive oxygen species (ROS) and thereby induces oxidative stress [3].

Mitochondria are responsible for intracellular oxygen metabolism, including the production of ATP via oxidative phosphorylation (OXPHOS); thus, they are also a major source of ROS production. Human mitochondrial DNA (mtDNA) comprises a 16,569-bp circular duplex molecule that encodes 37 genes, all of which are required for OXPHOS [4, 5]. It also contains a “displacement loop” (D-loop) non-coding control region extending from position 16024–576, which includes three hypervariable regions, HV1 (16024–16383), HV2 (73–372), and HV3 (438–574) [6]. The D-loop has been shown to be more frequently mutated than any other mtDNA regions. Since it regulates mtDNA transcription and replication, such polymorphisms often lead to the altered replication and/or transcription of mitochondrial genes and thereby affect overall mitochondrial function and/or the intracellular generation of ROS, causing various disease phenotypes [5].

Alterations to mtDNA have been previously identified in the peritoneal fluid and lesions of patients with endometriosis [7, 8]. Furthermore, the mtDNA control region polymorphisms 189G [4], 310C [4], and 16189C [4, 9] and the M5 [10] and N haplogroups [4] have been reported to be associated with the disease in Korean and Indian populations. Given that previous research has shown endometriosis-associated polymorphisms to be population-specific and that the Brazilian population comprises an admixture of a multiple ethnicities [11], the objective of the present study was to identify endometriosis-associated mtDNA control region polymorphisms in a cohort of Brazilian women.

Methods

Sample size

No previous studies investigating the association between mtDNA and endometriosis in a Brazilian population have been reported. Thus, based on previous results that showed T16189C was the only polymorphism in the control region to be associated to endometriosis in both Indian and Korean population [4, 9], we assumed this polymorphism to occur at a prevalence of 17% in the general population and 41% among endometriosis patients. By considering an alpha value of 0.05 (using a two-tailed test) to be acceptable, the required sample size was estimated to be 87 women with endometriosis and 87 control subjects, to achieve 95% study power [12].

Diagnosis and sample collection

Participants were prospectively recruited at the Gynecology and Obstetrics Department of Hospital das Clinicas HCFMUSP, Faculdade de Medicina, Universidade de Sao Paulo, a tertiary referral center for endometriosis, and all laboratory-based analyses were performed at the Department of Legal Medicine, Ethics, and Occupational Health of the same institution.

Female patients (aged 18–50 years) were divided into endometriosis (n = 90) and control (n = 92) groups according to whether they had been histologically diagnosed with and surgically staged for endometriosis (endometriosis group) or had instead undergone benign gynecological surgery and shown no evidence of endometriosis (control group). Exclusion criteria comprised the diagnosis of adenomyosis, and diseases previously shown to be associated with mtDNA polymorphism, including renal failure, hepatocellular carcinoma, and coronaropathy [5, 13].

All individuals gave informed consent for participation in the study (approved by the Institutional Review Board CAPPesq, protocol 534714/14). They provided details of their medical history and underwent a transvaginal ultrasound scan (TVUS) during screening. Surgery indications for the control group comprised tubal ligation, uterine fibroids, and/or benign adnexal masses. For patients in the endometriosis group, surgery was performed to address pelvic pain and/or infertility. During laparoscopy, patients in the endometriosis group were also classified as being stage I–IV according to the American Society for Reproductive Medicine (ASRM) [14] and as exhibiting either (1) superficial endometriosis (peritoneal lesions only), (2) deep infiltrative endometriosis (characterized by the presence of any lesion deeper than 5 mm), or (3) ovarian endometriosis (characterized by the presence of ovarian endometrioma, with no deep lesions). For each group, 95 women were eligible and peripheral blood samples (4 ml) were collected in EDTA vacutainers and stored at 4 °C for a maximum of 24 h until DNA extraction.

Genotyping and analyses of the mtDNA control region

Patient DNA was extracted from the peripheral blood samples. DNA samples were stored at − 20 °C, and an aliquot was transferred to be diluted and quantified prior to use. The mitochondrial control sub-regions HV1, HV2, and HV3 were then amplified via polymerase chain reaction (PCR) using the L15879 (5-AATGGGCCTGTCCTTGTAGT-3) and H727 (50-AGGGTGAACTCACTGGAACG-3) primers (Invitrogen Life Technologies). The utilized PCR mix comprised 50 ng of genomic DNA, 2.5 pmol of each primer, 1.25 mM of dNTP, and 2 units of Taq polymerase (Thermo Fisher Scientific). A Mastercycler thermocycler (Eppendorf) was used to achieve cycling conditions comprising 95 °C for 1 min (initial denaturation), followed by 36 cycles of 95 °C for 1 min (denaturation), 60 °C for 1 min (annealing), and 72 °C for 1 min (extension) and a final cycle of 72 °C for 7 min (final extension).

The resulting PCR product was purified using Exonuclease I/Shrimp Alkaline Phosphatase (Thermo Fisher Scientific) according to manufacturer’s instructions and sequenced using a BigDye Terminator Cycle Sequencing Kit (Applied Biosystems), according to the manufacturer’s instructions. Capillary electrophoresis was then performed using an ABI3130 sequencer (Applied Biosystems). The resulting sequence was compared with the revised Cambridge Reference Sequence (rCRS) [6], and haplotypes were defined using BioEdit software (http://www.mbio.ncsu.edu/BioEdit/BioEdit.html) and the Haplosite program (http://www.haplosite.com/haplosearch/). Identified differences between sequences and the rCRS were noted using the recommended nomenclature [15]. mtDNA haplogroups were classified using the Haplogrep 2 program (v2.1.0, https://haplogrep.uibk.ac.at) and confirmed using Phylotree software (http://www.phylotree.org/tree/main.htm) [16].

Complete sequencing of mitochondrial control sub-regions HV1, HV2, and HV3 was successfully completed in 90 patients in endometriosis group and 92 patients in control group, which were included in the analysis.

Statistical analysis

A Student’s t test was used to compare the mean values of continuous variables. A chi-square coefficient (χ2), likelihood ratio, and Fisher’s test were used to categorize variables. All statistical analyses were performed using SPSS software (v10), assuming a confidence interval (CI) of 95% and a significance level of 5%.

Results

The present study assessed 90 and 92 Brazilian women with and without endometriosis, respectively. The mean patient age in the endometriosis group was 39.2 ± 6.2 years and similarly in the control group was 40.1 ± 9.8 years. Likewise, patients in each group showed a similar incidence of comorbidities, smoking status, and family history of endometriosis (Table 1). Patients in the endometriosis group exhibited a higher rate of infertility (25.8% compared to 5.5%, respectively; p < 0.001) and pain symptoms than those in the control group (Table 1).

Table 1.

Demographic characteristics of included participants

Characteristic	Group n (%)		p
	Control (n = 92)	Endometriosis (n = 90)
BMI (kg/m2)	27.4 ± 5.1	26.7 ± 4.6	0.302a
Skin color
White	36 (39.6)	44 (49.4)	0.605b
Brown	42 (46.2)	35 (39.3)
Black	12 (13.2)	9 (10.1)
Yellow	1 (1.1)	1 (1.1)
No comorbidities	59 (64.1)	57 (63.3)	0.911c
Smokers	9 (9.9)	5 (5.6)	0.285c
Family history of endometriosis	3 (3.3)	8 (9)	0.111c
Infertility	5 (5.5)	23 (25.8)	< 0.001 c
Dysmenorrhea (EVA ≥ 7)	38 (41.8)	56 (62.2)	0.006 c
Deep dyspareunia	28 (30.8)	48 (53.3)	0.002 c
Acyclic pelvic pain	21 (23.1)	51 (56.7)	< 0.001 c
Cyclic bowel symptoms	0 (0)	21 (23.3)	< 0.001 c
Cyclic dysuria	2 (2.2)	9 (10)	0.028 c

Data expressed as mean ± standard deviation. BMI body mass index

^aStudent’s t test

^bLikelihood test

^cChi-square test

Of the patients in the endometriosis group, 77.8% (n = 70) had deep infiltrative lesions, 18.9% (n = 17) exhibited only peritoneal lesions, and 3.3% (n = 3) had ovarian endometriosis (without deep lesions). According to ASRM staging [14], 20.0% (n = 18) and 80.0% (n = 72) of these patients were diagnosed with stage I–II and III–IV endometriosis, respectively. Of the patients in the control group, 32.6% (n = 30) underwent tubal ligation, while 47.8% (n = 44), 17.4% (n = 16), and 2.2% (n = 2) were diagnosed with leiomyoma, benign ovarian cysts, and hydrosalpinx, respectively.

The HV1, HV2, and HV3 mitochondrial D-loop regions were successfully genotyped in all groups (i.e., both those with and without endometriosis). The distribution of mitochondrial haplogroups was found to be similar between the endometriosis and control groups (Table 2).

Table 2.

Distribution of mitochondrial macrohaplogroups in endometriosis and control groups

Macrohaplogroup	Endometriosis (n = 90)	Control (n = 92)	p
African	36 (40)	45 (48.9)	0.263
Amerindian	30 (33.3)	28 (30.4)
European	19 (21.1)	18 (19.6)
Asiatic	5 (5.6)	1 (1.1)

Likelihood test

We identified 205 polymorphisms in the mtDNA control region, comprising 120 located in HVR1 (58.5%), 35 in HVR2 (17.0%), 20 in HVR3 (9.8%), and 30 (14.6%) in sequences between hypervariable regions (Supplemental Table 1). We observed one major rearrangement, consisting of a 15-bp duplication between positions 16032–16118 (TCTCTGTTCTTTCAT), in a control-group patient.

One heteroplasmic point mutation (73G/A) was observed in two individuals from the endometriosis and two individuals from the control group (p > 0.999). One heteroplasmy (573.2C) was more frequently observed in the control group (5.6%) than in the endometriosis group (29.3%; p < 0.001; Table 3).

Table 3.

Polymorphisms of control region of mtDNA in control (CT) and endometriosis (EDT) groups

SNP	CT (n = 92)	EDT (n = 90)	p	ASRM stage		p
				I–II (n = 18)	III–IV (n = 72)
T146C	30 (32.6)	17 (18.9)	0.042	4 (22.2)	13 (18.0)	0.45
G185A	0 (0)	6 (6.7)	0.013	1 (5.5)	5 (6.9)	0.65
A189G	11 (12)	12 (13.3)	0.826	2 (11.1)	10 (13.8)	0.55
T236C	0 (0)	5 (5.6)	0.028	0 (0)	5 (6.9)	0.30
310.1C	44 (47.8)	48 (53.3)	0.463	10 (55.6)	38 (52.8)	0.52
310.2C	5 (5.4)	10 (11.1)	0.187	3 (16.6)	7 (9.7)	0.31
G499A	4 (4.3)	12 (13.3)	0.038	3 (16.7)	9 (12.5)	0.44
573.2C	27 (29, 3)	5 (5, 6)	< 0.001	3 (16.7)	2 (2.7)	0.052
T16,189C	34 (37.0)	39 (43.3)	0.450	9 (50.0)	30 (41.7)	0.35
T16,217C	5 (5.4)	13 (14.4)	0.049	3 (16.7)	10 (13.8)	0.50

Fisher’s exact test; SNP single nucleotide polymorphism

The frequencies of the previously identified, endometriosis-associated polymorphisms T16189C, A189G, and 310.1C were found to be similar in the two groups. We identified four polymorphisms that occurred at a higher frequency in the endometriosis compared to the control group, comprising T16217C (14.4 vs. 5.4%; p = 0.049, respectively), G499A (13.3 vs. 4.3%; p = 0.038), T236C (5.6 vs. 0%; p = 0.028), and G185A (6.7 vs. 0%; p = 0.013). Conversely, the T146C (18.9 vs. 32.6%; p = 0.042) polymorphism was more frequently observed in the control than in the endometriosis group. No significant difference was observed between the presence of polymorphism and I–II and III–IV stage of endometriosis (Table 3).

Of the 182 identified haplotypes, 180 were each observed in only one patient, while two were found to be common to more than one individual. No significant differences in the incidence of these haplotypes were observed between the two patient groups.

Discussion

Endometriosis is a benign condition associated with oxidative stress and inflammation. The mtDNA control region is essential for the replication and translation of mitochondrial genes, all of which mediate oxygen metabolism and the production of ROS [17]; thus, mutations in the mtDNA region may interfere with overall mitochondria function and contribute to the pathogenesis of various diseases [7].

Among the analyzed cohort, approximately 80% of patients comprised women with ASRM [14] stage III–IV endometriosis, and almost 25% were affected by infertility. These patients also frequently exhibited pain symptoms, such that 62% were affected by dysmenorrhea, 53% by deep dyspareunia, and 56% suffered from acyclic chronic pelvic pain. These results are consistent with those reported previously, which estimated 35% of women with deep endometriosis to be affected by infertility and 62% to exhibit dysmenorrhea [18, 19].

Among Indian population, one study has observed an association of the mtDNA control region polymorphisms A189G (41% endometriosis group versus 17%; p < 0.05) and 310.1C (33.5% versus 22.7%; p < 0.05) with endometriosis stages III–IV [4]. Controversially, we observed in our samples a similar distribution between A189G (12% endometriosis group versus 13.3% control group; p = 0.82) and 310.1C (47.8 versus 53.3%; p = 0.46). Also, polymorphism T16189C has been associated with endometriosis in two studies [4, 9] in Korean and Indian populations; however, no such association was found in the present study. The frequency of polymorphisms and mitochondrial haplogroups is specific of each geographic location. Therefore, polymorphisms in homogeneous populations such as Indian and Korean may not have external validation, especially when evaluated in a high-admixture population such as the Brazilian [20].

In our study, the T16217C polymorphism occurred at a higher frequency in the endometriosis compared to control group (14.4 vs. 5.4%; p = 0.049, respectively). In 2013, a study of a Korean population [9] showed this polymorphism to occur similarly in both groups (1.9 versus 0.7%; p > 0.05). Other diseases related to oxidative stress, such as diabetes mellitus, have also been associated with the presence of T16217C [21]. The present study demonstrates, for the first time, that the polymorphism G499A (13.3 vs. 4.3%; p = 0.03, respectively) also is associated with endometriosis. G499A has not yet been shown to contribute to human disease [22].

In the present cohort, polymorphism T146C appeared to be associated with the absence of endometriosis (18.9% control group vs. 32.6% endometriosis group; p = 0.042). These results are different with those of Govatati et al., who compared 152 Indian women with endometriosis stages III–IV and 150 women without the disease who identified T146C to occur at the same frequency between endometriosis and control, respectively (15.8 vs. 22.7%; p > 0.05). The polymorphism T146C has previously been described as an independent risk factor for the progression of non-Hodgkin lymphoma and hepatocellular carcinoma [13, 23].

A significant difference was also observed in the frequency of the heteroplasmic polymorphism 573.2C between the endometriosis and control groups (5.6 vs. 29.3%; p < 0.001, respectively). Heteroplasmic mutations are frequently observed in homopolymeric mtDNA sequences; however, their biological relevance is not fully understood. The length of heteroplasmy can vary, and the accumulation of heteroplasmic polymorphisms among the tissues of a given individual has been shown to mediate the severity and/or the phenotype of induced human diseases [24].

Mitochondrial ancestry is known to be determined by population migration patterns. The Brazilian population represents a highly admixed genetic population, due to multiple immigration events from African, South-American, and European populations. In the present study, the distribution of mitochondrial haplogroups was found to be similar between the endometriosis and control groups and is consistent with those reported by previous studies of healthy individuals in São Paulo, Brazil [11, 25–27]. The differences for the polymorphisms between the two groups are not related to a particular haplogroup, since their distribution is similar in both groups.

Despite G185A (6.7% endometriosis group vs. 0% control group; p = 0.01, respectively) and T236C (5.6 vs. 0%; p = 0.028, respectively) appeared to be associated with the presence of the endometriosis in our sample, the small frequency observed of these polymorphisms can impact on external validation of the results observed.

Limitations of this study include the small number of individuals analyzed in early stages (I–II) of the disease (80% of women included were stages III–IV) which impaired the subgroup analysis. Increasing the number of analyzed individuals in different stages of the disease could better clarify the role of these polymorphisms in the etiology of endometriosis.

On the other hand, the strength is that we evaluated the potential association of mtDNA control-region polymorphisms with endometriosis in a Brazilian population for the first time. Since our population is highly heterogeneous due to mixture of three populations (European, African, and Amerindians), it has unique characteristics. Understanding the pathogenesis of the disease and determining serum biomarkers, such as the mtDNA polymorphisms, may identify risk patients of developing endometriosis and advanced stages of the disease, allowing the primary prevention of endometriosis and the prevention of disease progression, respectively.

Conclusion

The present study showed the mitochondrial DNA control region polymorphisms T16217C and G499A to be positively associated and 573.2C and T146C to be negatively associated with endometriosis in a sample of Brazilian population. No difference was observed in haplogroups between groups.

Compliance with ethical standards

Conflicts of interest

The authors declare that they have no conflicts of interest.