Introduction
Uterine fibroids (UFs) are non-malignant pelvic tumours of the uterine smooth muscle that affect 70% of women in their pre-menopausal years (Igboeli et al. 2019). Globally, UF results in infertility, recurrent pregnancy loss, heavy menstrual bleeding, dyspareunia and frequent urination (Navarro et al. 2021). Approximately 65% of women with UF are asymptomatic and late presentation of the condition sometimes results in early pregnancy failure and complications such as placenta abruption and preterm labour (Chill et al. 2019). Despite the associated clinical outcomes, long-term non-invasive treatment options are not available for fibroids resolution. The condition therefore generates a massive healthcare burden in terms of surgeries, mostly hysterectomy (complete surgical removal of the uterus) and myomectomy (removal of tumors) and prolonged hospital stays (Stewart et al. 2016). Recurrent fibroids that require recurrent myomectomy are seen in 20% of UF cases (Kuznetsova et al. 2022). UFs are the cause of over 40 – 60% of all gynecological surgeries in African countries like Ghana, Cameroon and Nigeria (Opare-Addo et al. 2014, Egbe and Egbe 2018). Additionally, Ghana reports fibroids as one of the top 10 causes of hospitalisation over the past five years (Ofori-Dankwa et al. 2019).
UFs affect a wide cross-section of the population; however, African women show a higher prevalence, early onset and rapidly growing number of tumors compared to other races (Al-Hendy et al. 2017). Genetic predispositions have been implicated in UF, especially in developed countries. Genetic variation, such as single nucleotide polymorphisms that may be associated with fibroids susceptibility, has been identified (Cha et al. 2011). However, the pathogenesis of UF is not understood due to the heterogeneity of fibroids. There is also a paucity of information on the genetic basis of the increased incidence of fibroids; previous studies focused more on Caucasians and a few African Americans but none on native Africans (Eltoukhi et al. 2014). African Americans’ genome is approximately 80% African due to admixture (Bryc et al. 2015). Therefore, identified UF variants in such populations may be attributable to admixture and may have no contributing effects on the increased risk of UF in the African population.
Studies have identified environmental risk factors for UF in an effort to understand its etiology. In Ghana, studies using retrospective data identified age, obesity, parity and history of reproductive tract infections (RTI) as putative risk factors for UF (Opare-Addo et al. 2014). There is a need for a more comprehensive study to gain insights into the prevalent environmental factors that predispose Ghanaian women to UF. Recent evidence in developed countries also suggests altered microbiota as a risk factor (Mao et al. 2022). Such alterations from the direct effect of microbes or the cross-talk of their metabolites can alter the immune system, host gene expression and metabolic signalling (Sobstyl et al. 2022). There is a high burden of RTI in developing countries (Torrone et al. 2018). In addition, heterologous bacteria can translocate from one organ to colonise other tissues or organs and influence human health (Wang et al. 2021). Considering the aforementioned, there is a need to investigate the uterine microbiota as a risk factor for UF in this population.
Overall, the genetic and other host-related basis of UF pathogenesis remain underexplored in high-risk populations such as Africa. Therefore, a comprehensive assessment of factors that predispose Ghanaian women to UF development is critical. A population-based study integrating environmental and genetic factors will provide knowledge on how these factors influence genetic predisposition and UF development. The outcome of this study will provide information for candidate biomarkers for UF prevention, diagnosis and treatment and ultimately improve management strategies.
Research Questions
This study will answer the following questions:
What are the non-genetic risk factors predisposing Ghanaian women to UF?
How does the alteration of the uterine microbiota contribute to UF tumorigenesis?
What are the unique genetic variants in the Ghanaian population associated with UF?
Study aims and Objectives
Aim 1: To investigate the influence of epidemiological factors and the uterine microbial community on fibroid tumorigenesis amongst Ghanaian women:
Identify the association of demographic factors and clinical manifestations of UF amongst Ghanaian women;
Determine the uterine microbial profile associated with UF;
Analyse the role of patient-derived microbiome on fibroid growth in-vitro.
Aim 2: To profile molecular markers associated with fibroid tumorigenesis:
Characterise and identify variants associated with fibroid tumorigenesis;
Functionally analyse identified SNPs in-vivo;
Validate the unique disease markers in extended cohort.
Background and Study Impact
BACKGROUND
Uterine Fibroids (UF)
Uterine fibroids, also called leiomyoma, are benign tumour growths that arise from the proliferation and differentiation of single cells. UF growth can be influenced by genetic and non-genetic factors (Fig. 1). Histologically, UF-associated markers are smooth muscle fibres with increased extracellular matrix (ECM) contents, mainly fibronectin, proteoglycans and collagen (Mas et al. 2014). Fibroids can grow as single or multiple tumours in the uterine wall (myometrium), with sizes ranging from pea to watermelon. They occupy different layers of the uterus and are subtyped as intramural, submucosal and subserosal. Depending on the location and lesion size, symptoms range from pelvic pain, heavy bleeding, anaemia and frequent urination to miscarriages, preterm birth and infertility (Bulun 2013). In Africa, the majority of women appear with huge fibroids and seek medical attention late due to inadequate access to proper healthcare services and high costs (Igboeli et al. 2019). In Ghana, there is a paucity of information on the burden of UF; nonetheless, a significantly high level of UF cases has been reported in a tertiary hospital (Ofori-Dankwa et al. 2019). Age, obesity and a history of sexually transmitted infections have been associated with UF development in the country (Opare-Addo et al. 2014).
Aetiology and Risk factors
Characteristic risk factors associated with UF pathogenesis include African descent, family history and early menarche (Fig. 2). In spite of differences in racial and ethnic background, age is consistently linked to the incidence of UF. Women between 40 and 60 years old are at a 4–11 times higher risk of developing UF than women between 20 and 30 and those above the age of 60 (Morhason-Bello and Adebamowo 2022). An increase in oestrogen and progesterone regulating genes hormone markers also results in the proliferation of UF cells (Lora et al. 2012). A low level of vitamin D is also an UF risk factor and the black race has a higher risk of deficiency (Ciebiera et al. 2021).
Obesity is an established risk factor for UF and it is characterised by the secretion of adipokines and chronic inflammation. Inflammatory cytokines could lead to increased levels of reactive oxygen species, promote the proliferation of UF cells and enhance ECM deposition.
Altered Microbiome and Epigenetic regulation
Microbiome (microbiota and its environment) is known to play a role in human health and disease. The microbiota in the reproductive tract of women is important in the regulation of reproduction, a major physiological process in mammals (Heil et al. 2019). The uterine layer has always been considered sterile; however, sequencing technology revealed it is colonised by a unique microbiota that, when altered, is associated with various disease states (Peric et al. 2019). Recently, bacterial communities have been seen to ascend to the uterine cavity from the vagina through the cervix, resulting in microbiome imbalance (Wang et al. 2021). Such colonisation, as evidenced by the co-occurrence of multiple organisms, the outgrowth of beneficial species and the release of virulence factors, may alter host gene expression and induce epigenetic change. These changes in the UF genome activate signalling pathways that play an essential role in UF pathologic processes (Mas et al. 2014). There is no study on the uterine microbiota and its role in UF development. Shotgun metagenomics can give insights on how microbial communities influence disease processes by interacting with the host.
Genetic Alterations of UF
Genetic factors are major players in UF development. Findings via genome-wide linkage analysis and SNP arrays in non-African populations identified a proportion of genetic variants associated with UF (Cha et al. 2011). The most closely associated variant with a high risk of UF development is mutations in the mediator complex subunit 12 (MED12). Mutations in this gene arise under the influence of infection or mechanotransduction and result in dysfunction of the metabolic pathway associated with ECM formation and myoblast proliferation (Baranov et al. 2019). Others include rearrangement of the high mobility group AT-hook gene (HMGA1), collagen type IV alpha 5 and 6 (COL4A5 and COL4A6) deletion and biallelic inactivation of fumarate hydratase – FH - that result in autosomal dominant syndrome (Baranov et al. 2019). Some variants in non-Africans were not recurrent in African-Americans, indicating fibroid heterogeneity and involvement of different pathways in UF tumorigenesis from different regions (Mehine et al. 2016). High-throughput sequencing technology provides nearly full coverage of variations in the genome and can contribute to unknown aetiology. There are little or no genetic studies on UF pathogenesis in Africa. Hence, a population-based study to identify and experimentally validate population-specific UF-associated variants would be an optimal approach. The poor pathophysiology of UF has resulted in limited identification of drug targets that can prevent UF development in high-risk groups.
Therapeutics and Diagnostics
Currently, UF treatment options and preventative strategies are limited. Available invasive surgical procedures require lengthy recovery time, they are associated with complications and are not suitable for women who desire fertility. The non-invasive options lessen UF-associated symptoms; however, they present side effects and require a perpetual regimen. Additionally, there is a risk of tumour size increase or fibroid regrowth following cessation of treatment (Aninye and Laitner 2021). The rate of UF recurrence following myomectomy after the initial removal of all existing fibroids is high. Prevention strategies, such as early diagnosis and prophylaxis treatment for women at high risk, are necessary.
Diagnosis of UF is done by imaging procedures (pelvic and transvaginal ultrasound, MRI) that are expensive and not readily available in low-resource areas. However, cell free DNA (CfDNA) detected in body fluids (plasma and serum) has recently emerged as a diagnostic and prognostic marker for benign and malignant disorders (Ganesamoorthy et al. 2022). Available serum biomarkers that can accurately diagnose UF will improve early-stage diagnosis, aid routine screening, prevent misdiagnosis and may predict women's risk of recurrent UF.
Study Impact
Uterine fibroids result in high morbidity in Africa and oftentimes mortality from treatment-related complications, especially blood loss during myomectomy (Igboeli et al. 2019). Frequent exposure and translocation of pathogens due to factors such as RTI could alter the uterine-wall microbiota and initiate pathways that can provoke the development of tumour. Previous studies on leiomyoma etiology focused on hormones and genetic predisposition, but none considered the potential role of the altered uterine microbiota. The mechanism behind the increase fibroids incidence in the African population remains understudied. There are little or no biological studies on UF genetic predispositions in the African population. To date, only hormonal treatment exists as a non-invasive treatment option due to the limited knowledge of disease mechanisms. An integrated approach that considers trait-associated genetic variants and environmental risk factors can provide information for biomarkers for prophylaxis, diagnostic tools and precise treatment. This will increase UF management efficiency and reduce morbidity and the associated socio-economic burden. Overall, this study will contribute to the understanding of the etiology of UF tumorigenesis in a high-risk population. Consequently, we will overcome the public health problem.
Implementation
Study workflow
To achieve the study objectives, the following workflow (Fig. 3) will be adopted.
Study Design
This will be a case-control study using clinically-confirmed UF patients and non-affected individuals from selected regions in Ghana (Fig. 4) based on cluster sampling (Xiong et al. 2004). Epidemiological mapping will be generated online and on-site.
Ethical Clearance
Study approval will be obtained from all sample sites and Ghana Health Service Ethics Review Committee. The study participants will be counselled and signed informed consent will be obtained before enrolment into the study. Ethical approval will also be obtained from the Institutional Animal Care and Use Committee, Noguchi for animal studies.
Participation will be voluntary and study participants will be anonymised using unique study identification codes and all information treated with strict confidentiality.
Sample size determination
The sample size will be determined using the Cochran formula, where d is precision level, Z is critical value and P is sample proportion. In this study, a proportion of 70% will be used because probability is 70% in pre-menopausal women.
Sample calculation will be done as shown in:
n = Z2P (1-P) /d2
n = (1.96)2 (0.70) (1-0.70) / 0.052
n = 322.4 ~ 350
Therefore, the sampling distribution will be as shown (Fig. 5)
Sample Collection
Women aged 20 - 60 years with or without UF in the selected sites enrolled into the study will be provided with a questionnaire. Those within the reproductive age with confirmed UFs scheduled for hysterectomy or myomectomy in selected hospitals will form the cases, while those admitted for hysterectomy for any other indications apart from UF will form the control. Those who have not attained menarche and unwilling to participate will be excluded. Fibroid tissue and normal adjacent myometrium tissue, blood and cervico-vaginal swabs will be obtained from cases, while blood and cervico-vaginal swabs obtained from controls (Fig. 5). Tumour number and sizes will be recorded and vitamin D level measured in the serum.
Assessment of demographic factors and clinical manifestations of UF in Ghanaian women
This objective will assess age, parity, body mass index (BMI), family history and UF-related basic clinical features from enrolled participants using the survey questionnaire (Suppl. material 1). Data will be grouped and associations with UF will be determined accordingly.
Determining uterine microbial profile from UF cases and controls
This objective will assess differences in microbial type and abundance. Samples will be grouped into those with or without RTI.
Sample preparation
Swab and tumour tissue homogenate (PDM) will be suspended in 2 ml of phosphate buffer solution (PBS) and centrifuged (5 min, 12000 rpm). Nucleic acids will be extracted from pellets using the microbiome DNA/RNA extraction kit.
Shotgun metagenomics sequencing and data analysis
Metagenomics sequencing will be carried out as described previously (Quince et al. 2017). Briefly, libraries will be prepared using Nextera XT library kit then sequenced using NextSeq platform (Illumina). Bacteria taxonomic classification will be conducted from sequenced data after a quality check using Kraken2 and paraKraken for differential viral particles. Data will be analysed using the workflow below (Fig. 6). Differences in taxa abundance will be considered statistically significant if adjusted p-values are < 0.05.
Role of patient-derived microbiome on fibroid tumorigenesis
The effect of microbiome on uterine fibroid cells (UtLm) and primary uterine smooth muscle cells (HUtSMC) will be measured after infecting the cells with PDM.
Cell Culture
This will be carried out as previously described (Rabiei et al. 2019). Cells will be cultured in Dulbecco Modified Eagle’s Medium (DMEM) under the following conditions: 37°C, 5% CO2, and 95% humidity. At 80% confluence, cells will be detached using trypsin and then centrifuged. Pellets will be reconstituted in DMEM and 100 ul will be transferred into a 24-well plate at a density of 5x104 cells per well. Patient-derived microbiome (PDM) will be added to the cells using the various cell-culture conditions A-E (Fig. 7) and then incubated and monitored for cell viability.
Cell Viability assay
Viability media solution will be added after discarding the culture medium for viability determination. After 4 h incubation, the culture media will be discarded and solubilisation solution will be added into each well and optical density of viable cells will be measured (550-630 nm) using a microplate reader.
Immunohistochemical analysis
UF-associated markers including alpha smooth muscle actin (ACTA2), desmin, COL1 and COL4 (COL1-A, COL-94) will be measured. Proliferation, angiogenesis, apoptotic and hormone markers will be measured using primary antibody (anti-Ki67, anti-VEGF, anti-Bcl-2, anti-ER, anti-PR), respectively. Briefly, cells will be blocked in 3% bovine serum albumin for 1 h, followed by overnight incubation in the primary antibody. Secondary antibody conjugated to substrate will be added for enzymatic detection.
Profiling molecular markers associated with UF
This objective aims to identify UF-associated genetic variants amongst Ghanaian women using group 1 clinical samples (Fig. 5).
Sample Preparation
DNA from tissue samples will be extracted using AllPrep Universal kit, while total RNA will be extracted using Qiazol reagent following the manufacturer’s instructions. Blood samples (5 ml) will be centrifuged (10 min, 3000 rpm) to separate blood cells and plasma. Plasma aliquots will be re-centrifuged (10 min, 13000 rpm) and used for Plasma cfDNA extraction using the Circulating Cell free Nucleic acid kit.
Genomics profiling of UF tissue and controls using whole exome sequencing
Whole exome sequencing will be carried out as described in McGuire et al. (2012). DNA samples from tissue and blood will be subjected to in-solution exome enrichment using SureSelect Human All Exon Kit. Samples will be sent for sequencing (Illumina platform) following exome capture. Data will be aligned to the human genome GRCh37, quality control and downstream variant analysis will be performed. Focus will be on variants present in the fibroids, but absent in adjacent normal myometrium.
Variant calling and annotation
Variant filtering and quality control will be carried out using NextGENe software. Somatic single nucleotide variants (SNV) will be detected from sequenced data using VarScan2. Already-existing SNV associated with UF in literature will be searched for and used as control. Copy number aberrations (CNA) will be detected using IchorCNA and CNA from matched germline samples will serve as control. Structural variant (SV) calling will be observed using Manta.
Molecular annotation
SIFT and polyphen2 will be used to identify SNPs that will have an impact on protein function. These non-synonymous SNPs will further be validated in-vivo. To identify disease-causing genes commonly mutated in fibroids, variants will be compared across samples. These genes will be analysed for their biological functions using DAVID and ReactomePA to identify pathways impacted due to functional alteration of the genes.
RT-qPCR
The relative expression of identified disease-causing genes and existing UF driver genes in literature will be measured using RT-qPCR from normal and diseased tissues. The mRNA levels will be normalised using the endogenous gene, glyceraldehyde 3-phosphate dehydrogenase (GAPDH). The relative fold change in gene expression will be computed using the 2-Δ∆Ct formula.
Functional analysis of identified marker <italic>in-vivo</italic>
This objective sought to identify if the non-synonymous SNPs are disease-causing mutations using experimental xenograft model and CRISPR-Cas9 technology.
Design of ssDNA/sgRNA/Cas9 expression vectors
Single-stranded donor oligonucleotides (ssDNA) carrying desired mutations will be generated via site-directed mutagenesis and used as DNA repair templates. Cas9 mRNA, single guide RNAs (sgRNAs) with 20 bp of homology flanking either sides of the target integration sites and ssDNA will be synthesised from Biogene. The promoter driving expression of gene’s sgRNA will also be cloned into the backbone plasmid as previously described (Luo et al. 2018). Mutation will be confirmed using PCR and sequencing. The plasmids will be co-transfected in HutSMc and UtLm. Scrambled vector and non-transfected cells will serve as control.
Xenograft transplantation
To mimic human UF microenvironment, animals bearing UF xenograft will be used as previously described (Wang et al. 2021). Briefly, 4-5 weeks old germ-free female SCID mice will be purchased and grouped into two (six each), test and control group. The mice will be anaesthetised using chloroform inhalation. Cells (HutSMc, UtLm and mutants (ΔUtLm, ΔHutSMc)) mixed with matrigel will be grafted beneath the kidney capsule following subcutaneous implantation of oestradiol. They will be monitored daily; developed lesions and body weight will be measured twice per week using an electronic scale and vernier caliper.
Morphological and histological evaluation
Mice will be euthanised after 8 weeks by CO2 inhalation and tumours will be dissected and evaluated.
Genotyping the unique marker associated with UF
This will be done to further validate SNPs in the remaining subset of clinical samples, Group 2 (Fig. 5) and also confirm if SNPs are in circulation for blood-based diagnosis.
SNP genotyping
SNPs that show a strong association with the UF disease phenotype will be genotyped in CfDNA and tissue DNA. Specific primers will be designed for each polymorphism and samples will be amplified using the restriction fragment length polymorphism (PCR) approach and targeted genotyping.
Bioinformatics and Statistical Analyses
Bioinformatics analysis will be performed using packages anchored in R (version 4.1.3) as detailed in the Methods section. For statistical analysis, data will be subjected to normality test using the Shapiro-Wilk test. Association between assessed variables and UF will be determined using the Chi-square test. Measurements between two groups or amongst more than two groups will be compared using the t-test and ANOVA, respectively. Disease marker association will be determined using PLINKv.1.9. All tests will be two-tailed, significance level will be set at 5% and p-value less than 0.05 will be interpreted as significant. The crude and adjusted odds ratio and CI of 95% will be noted.
Expected outcomes
The epidemiological study will identify the prevalence of UF and provide knowledge on how factors (age, obesity, vitamin D levels, parity and family history) contribute to an increased risk of UF amongst Ghanaian women. The microbiome study will give insights into microbial groups that are risk alleles for UF and the role of the uterine microbiota in modulating the physical health of women with regards to fibroid development. This can influence the use of host gene expression and microbial profile as active surveillance biomarkers.
The genetic studies will identify unique UF-associated genetic variants in this population. In addition, essential genes and relevant molecular pathways that are activated or dysregulated in fibroid tissue compared to normal tissue will be identified. These can be harnessed for the design of preventative and fertility-friendly treatments for women of reproductive age. A functional study will provide insights on the function of SNPs and their availability in blood. Cell free DNA biomarker SNP array chips can be developed for UF diagnosis as well as predicting the risk of developing UF.