Racial differences in transcriptomics and reactive oxygen species burden in myometrium and leiomyoma

Abstract

STUDY QUESTION

Are there differences in Mediator Complex Subunit 12 mutations (MED12) mutation, transcriptomics, and protein expression in uterine myometrium and leiomyomas of Black and White women?

SUMMARY ANSWER

RNA sequencing, tissue microarray, and immunohistochemistry data revealed that Black and White women have significant differences in their myometrium and leiomyoma profiles.

WHAT IS KNOWN ALREADY

Black women develop uterine leiomyoma earlier than White women, and are more likely to be anemic, have multiple tumors, undergo hysterectomy at an earlier age, have a higher uterine weight, and report very severe pelvic pain.

STUDY DESIGN, SIZE, DURATION

Uterine tissues were collected from premenopausal women undergoing hysterectomy or myomectomy at Northwestern University Prentice Women’s Hospital (Chicago, IL) from 2010 to 2021. Tissues were collected from a total of 309 women, including from 136 Black women, 135 White women, and 38 women from other racial groups. A total of 529 uterine leiomyomas (290 from Black women, 184 from White women, and 55 from women of other racial groups) were subjected to molecular analysis. Leiomyoma and matched myometrium from a total of 118 cases including 60 Black women and 58 White women, were used for tissue microarrays, along with 34 samples of myometrium without leiomyoma from White women.

PARTICIPANTS/MATERIALS, SETTING, METHODS

Tissues from the above patient cohorts were analyzed by tissue microarray, immunohistochemistry, RNA sequencing, and mutation analysis.

MAIN RESULTS AND THE ROLE OF CHANCE

The results indicated that leiomyoma from Black women have a higher rate of MED12 mutations (79.0%) than those from White women (68.5%) (*P ≤ 0.05). RNA-sequencing analysis in myometrium revealed differentially expressed genes (270 upregulated, 374 downregulated) dependent on race, wherein reactive oxygen species, hypoxia, and oxidative phosphorylation pathways were positively correlated with samples derived from Black patients. The levels of proteins associated with oxidative DNA damage and repair, 8-hydroxyguanosine (8-OHdG), 8-oxoguanine glycosylase (OGG1), heme oxygenase-1 (HO-1), and kelch-like ECH-associated protein 1 (KEAP1), were higher in leiomyoma and matched myometrium, particularly those from Black patients, compared to the control myometrium (with leiomyoma) (***P ≤ 0.001).

LARGE SCALE DATA

The datasets are available in the NCBI (The BioProject number: PRJNA859428).

LIMITATIONS, REASONS FOR CAUTION

Myometrium without leiomyoma derived from White patients was used as a control in the tissue microarray analysis, as myometrium without leiomyoma from Black patients was not accessible in large numbers. The RNA sequencing was performed on myometrium tissue with leiomyoma present from 10 White and 10 Black women. However, one sample from a Black woman yielded low-quality RNA-sequencing data and was excluded from further analysis.

WIDER IMPLICATIONS OF THE FINDINGS

Women with symptomatic leiomyomas have a considerable loss in their quality of life. This study provides information on underlying genetic and molecular defects that may be necessary for future therapeutics targeted at leiomyomas.

STUDY FUNDING/COMPETING INTEREST(S)

This work was supported by grants from NCI (R01CA254367) and NICHD (P01HD057877). The authors declare no conflict of interest.

Introduction

Uterine leiomyomas (LMs) are the most common, yet understudied, disease in women. LMs, commonly known as fibroids, are non-cancerous tumors of the uterus that arise from the myometrium with a prevalence of over 70% in women. Although most women with LMs are asymptomatic, ∼30% of them present with severe symptoms. Treatment options are limited, with surgery being the most common. As such, LMs are the leading cause of hysterectomy in the USA and have a profound impact on health care costs worldwide.

Progesterone and estrogen are thought to be key to the survival and growth of LM, whereas the initiators responsible for early tumor synthesis are currently unknown (Medikare et al., 2011). A previous study has demonstrated that cells derived from myometrium with (MyoF) and without (MyoN) fibroids exhibit differential responses to progesterone treatment in culture. Of importance, primary cells from MyoF present with a hyper-responsiveness to this treatment, thus indicating that progesterone could be a contributing factor in the mechanism behind the origin of LM (Omar et al., 2019). Risk factors for uterine LM include age (<40 years), early menarche (younger than 10 years), nulliparity, obesity, alcohol consumption, genetic factors (family history) and, of importance to this study, race (Ross et al., 1986; Chiaffarino et al., 1999; Ryan et al., 2005). Black women are at a higher risk of developing LM than White women and frequently experience much more severe symptoms. A study from Baird et al. (2003) demonstrated that amongst women with no previous diagnosis of LM, ∼65% of all White women had developed LM by age 50. Strikingly, almost 60% of African-American women had developed LM by age 35, with this figure jumping to nearly 80% by the age of 50 (Baird et al., 2003), supporting the concept that Black women develop LM earlier than White women. Another study found that of 409 Black women screened, 89% had LM, compared with an incidence of 59% in White women. Additionally, Black women were more likely to be anemic, have seven or more LM, undergo hysterectomy at an earlier age, have a higher uterine weight and report very severe pelvic pain, in comparison to White women (Kjerulff et al., 1996). One study concluded that even with a higher prevalence of risk factors within the Black population, it still did not account for the prominent excess in the rate of LM development (Marshall et al., 1997). In actuality, the occurrence of LM is likely to be underestimated, as most studies include only symptomatic women with clinical diagnoses.

MED12 is part of the human mediator complex, a genome-wide regulator of transcription, that interacts extensively with the RNA polymerase II enzyme, regulating its ability to express protein-coding genes. The mediator influences transcription by connecting gene-specific transcription factors to the RNA polymerase II initiation complex (Taatjes, 2011). MED12 mutations are the most common somatic alteration found in LM (Ordulu, 2016). Interestingly, MED12 has been shown to be mutated at different frequencies dependent on race; MED12 exon 2 mutations occur at a frequency of about 50% in LM from South African women (Mäkinen et al., 2011a), whereas high-throughput exome sequencing has revealed that MED12 mutations are found in ∼70% of all LM in the Finnish population, 67% of LM in the North American population (McGuire et al., 2012), and 54% of LM in the Han Chinese populace (Ye et al., 2015). Of interest, MED12 mutations have also been identified at high frequencies in fibroepithelial tumors (Lim et al., 2014; Yoshida et al., 2015). A recent meta-analysis containing data from 25 studies indicated that MED12 mutations were more common in Black (74.5%) than in White (65.8%) women (He et al., 2022). Given that MED12 mutations have been observed in women from numerous racial and ethnic backgrounds, it is thought that mutations in exon 2 of MED12 might be implicated in driving development of uterine LM (Yang et al., 2022). However, it remains unclear whether the prevalence of MED12 mutations differentially drives LM development in Black and White women.

Despite the racial disparities in LM prevalence having been identified numerous decades ago, not enough has been done to reveal the underlying causes for this. In this study, we demonstrate that the prevalence of MED12 mutations was higher in LM from Black women compared to White women. We show that LM exhibits distinct genomic differences in Black and White women and hypothesize that the increased burden of the oxidative stress pathway is important in the development of LM. Additionally, we demonstrate transcriptomic differences in matched LM myometrial tissue obtained from Black and White women. Finally, due to the enrichment of genes associated with the oxidative stress pathway, we explored levels of reactive oxygen species (ROS) associated markers, many of which showed racial differences in expression.

Materials and methods

Patient cohorts

Uterine tissues were collected from premenopausal women undergoing hysterectomy or myomectomy at Northwestern University Prentice Women’s Hospital (Chicago, IL) from 2010 to 2021. Informed consent was obtained from all patients following an Institutional Review Board (IRB)-approved protocol. Pathology reports concluded that all LM were found to be ‘usual type’; the most common primary tumor type. Patient age at surgery, race, uterine weight, number of tumors, tumor size, hormonal status, and hormonal treatment were documented and summarized in Table I. Tissues were collected from a total of 309 women in this study, including 136 Black women, 135 White women and 38 women from other racial groups. A total of 529 LM (290 from Black women, 184 from White women, and 55 from women of other racial groups) were subjected to analysis.

Table I.

Summary of cases for MED12 mutation analysis.

Type		Black	White	Other	P-value
No. cases (n)		136	135	38	Black vs White	Three groups
Age (years)	Mean ± SEM	42.19 ± 0.459	44.98 ± 0.471	45.27 ± 0.841	<0.0001	<0.0001
Tumor size (cm)	Mean ± SEM	9.498 ± 0.447	9.030 ± 0.526	8.797 ± 0.986	0.2599	0.1901
Uterine weight [g]	Mean ± SEM	1050 ± 82.37	849.4 ± 83.09	830.8 ± 158.7	0.0081	0.0063
No. of tumors	<5	52	82	26	0.0009	<0.0001
	≥5	72	47	7
	NA	12	6	5
Endometrial phase	PE	54	61	18	0.2356	0.3108
	SE	34	30	5
	IA/AT	14	25	6
MED12 mutation		229/290	126/184	41/55	0.0103	0.0372

SEM, standard error; NA, not available; PE, proliferative endometrium; SE, secretory endometrium; IA/AT, inactive or atrophic endometrium.

Tissue microarray

A total of 152 samples were included for tissue microarray (TMA), including 60 samples from Black women and 58 from White women of myometrium with LM, as well as 34 samples from White women of myometrium without LM. After slide review, the Formalin-Fixed Paraffin-Embedded tissue of LM and matched myometrium were selected. Tissue cores of 1.5 mm in diameter were taken to create TMAs. TMAs were sectioned at 4 μm. The first and last slides of each TMA were stained with Hematoxylin and Eosin (H&E) for quality assurance to confirm the correct tissue types.

Immunohistochemistry

Immunohistochemical staining was performed on a Ventana Nexus automated system (Ventana Nexus, USA) as described previously (Bertsch et al., 2014). Antibody information is summarized in Supplementary Table SI, including antibodies for 8-hydroxyguanosine (8-OHdG), 8-oxoguanine glycosylase (OGG1), heme oxygenase-1 (HO-1), kelch-like ECH-associated protein 1 (KEAP1), superoxide dismutase 2 (SOD2), SOD^K122Ac, SOD^K68Ac, estrogen receptor (ER), and progesterone receptor (PR). Immunostaining was scored semi-quantitatively by percentage and intensity by two pathologists. Intensity was scored as negative (0), weak (1+), moderate (2+), or strong (3+), and the percentage of positive tumor cells was scored from 0% to 100%. Selected immunostaining markers were imaged and further quantified by H-scores in Image J (quantitative analysis) (Varghese et al., 2014).

RNA sequencing

Total RNA was isolated using the Qiagen miRNeasy Mini Kit according to the manufacturer’s instructions (Qiagen, USA, cat. # 217004). RNA samples were quantified using a Qubit 2.0 Fluorometer (Life Technologies, USA) and RNA integrity was checked using Agilent TapeStation 4200 (Agilent Technologies, USA). RNA integrity numbers ≥9 were used for library preparation using the NEBNext Ultra II RNA Library Prep Kit for Illumina following the manufacturer’s instructions (NEB, USA). Briefly, mRNAs enriched using Oligo (dT) beads were fragmented for 15 min at 94°C. First-strand and second-strand complementary DNA (cDNA) was then synthesized. cDNA fragments were end repaired and adenylated at the 3′ ends, and universal adapters were ligated to cDNA fragments, followed by index addition and library enrichment by limited-cycle PCR. The sequencing libraries were validated on the Agilent TapeStation (Agilent Technologies, USA) and quantified using a Qubit 2.0 Fluorometer (Invitrogen, USA) as well as by quantitative PCR (KAPA Biosystems, Germany). RNA sequencing was performed by Genewiz (Genewiz, USA) on the Illumina HiSeq4000 and sequenced using a 2 × 150 bp Paired-End (PE) configuration.

RNA-sequencing processing and analysis

The quality of DNA reads, in FASTQ format, was evaluated using FastQC (Babraham Bioinformatics, UK). Adapters were trimmed and reads of poor quality or those aligned to ribosomal RNA (rRNA) sequences were filtered out. The clean reads were mapped to the Homo sapiens GRCh38 reference genome available on ENSEMBL using the STAR aligner v.2.5.2b, and feature Counts from Subread package v.1.5.2 was applied to calculate unique gene hit counts. After extraction, the gene hit counts table was used for downstream differential expression analysis using the DESeq2 package in R. Differentially expressed genes were identified using threshold adjusted P-value <0.05 and absolute log2 fold change >1 (myometrium from Black versus from White women). To perform clustering analyses on a group of samples, a union of all genes and their expression RPKM values within that group was generated to build a read count matrix for the group of interest. Various unsupervised and other machine learning techniques were applied to the composite read count matrix of interest. The packages ggplot2 and Pheatmap in R were used to build various heatmaps and volcano plots. ‘Factoextra and FactoMineR’ packages were used for principal component analysis (PCA) to reveal similarity between samples based on the distance matrix. The ‘fgsea’ and ‘enrichplot’ packages were run for Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis. Functional analysis was performed by Gene Set Enrichment Analysis (GSEA) using the Molecular Signatures Database, MSigDB (http://www.broad.mit.edu/gsea/msigdb/index.jsp). RNA sequence data have been submitted to NCBI (The BioProject number: PRJNA859428).

Statistical analysis

Statistical analysis of RNA sequencing was performed using related packages in R. A two-way ANOVA (paired t test) was constructed to identify genes that were differentially expressed in all LM compared with the matched myometrium samples. Ordinary one-way ANOVA, Brown–Forsythe and Welch ANOVA, or Kruskal–Wallis test was performed for multiple comparisons depending on the distribution and variances of the data on differential gene expression in non-matched tissue samples. Other statistical analysis was performed using GraphPad Prism version 8.0 (GraphPad Software). All data represent the mean ± SEM of a minimum of three independent experiments and data were considered statistically significant if the P-value was <0.05.

Results

MED12 mutations are associated with racial differences in leiomyoma

The frequency of mutations within the MED12 gene is remarkably high in uterine LMs, indicting a strong link between these mutations and the incidence of LM. We hypothesized that racial differences in uterine LM incidence and severity would be linked to MED12 mutation, and so to evaluate this association, we collected a large cohort of 529 LM from 136 Black women (290 LM tumors), 135 White women (184 LM tumors), and 38 women of other racial groups (55 LM tumors) to perform MED12 mutation analysis (Fig. 1A). Overall, MED12 mutations were highly enriched at c.130-131 (54% in White and 57% in Black women) in exon 2. Deletions accounted for 10.6% of all mutations in MED12, followed by point mutations at other sites (9.75%) in the exon (Fig. 1B and C). Overall, the number of MED12 mutations was significantly higher in LM from Black women (79.0%) than in LM from White women (68.5%) (Fig. 1B–D). Given the noticeable disparity between race and MED12 mutations, we performed a comparative analysis of all mutation types to further examine whether a specific form of mutation (point mutations at c.130-131, other point mutations and insertion/deletions) was preferentially expressed in Black women with LM. The results indicated that all mutation types were present at significantly higher numbers in LM from Black women than in LM from White women (Chi square 9.089, df 3.000, P < 0.05) (Fig. 1E). As expected, mutations in MED12 accounted for a large portion of the incidence of LM (74.9%) in these cases. In addition to this, point mutations at c.130-131 accounted for the incidence of LM in 55.0% of the cases (Fig. 1E). Further analysis revealed that 56.3% of missense mutations were G>A, 24.6% were G>T, and 19.2% were G>C in LM from Black women which was significantly different from the proportions in LM from White women (P < 0.05, Fig. 1F). The data here demonstrate that MED12 mutations are more prevalent in the tumors of Black women.

Figure 1.

Mutation analysis in leiomyoma (LM) based on race. (A) A schematic diagram illustrating the number of patients (cases) and leiomyomas (tumors) collected from Black women, White women and women from other racial groups, and the number of tumors. (B, C) MED12 mutation patterns and their distribution in 529 leiomyomas from white (B) and black (C) women. The actual mutation distribution in MED12 exon 2 and the intron 1-exon 2 boundary are illustrated along with a sequence map. Each dot represents one mutation. Blue lines under the sequence indicate the actual deletion site and extension. (D) Percentage of MED12 mutations (MED12 Mut, light blue) and non-MED12 mutations (MED12 NM, light purple) in LM from Black, White and other racial groups. The number of tumors (n) is shown below the graph. (E) Differences in MED12 mutation types by race (Black: blue dots; White; pink dots). Mutation types are: NM, non-MED12 mutation; SV, structural variants, insertions/deletions; Other mut: all point mutations other than c.130-131; c.130-131 mut: mutation at codon 130 and 131 guanine. (F) Percentage of c.130-131 missense mutations G>A (purple), G>C (light purple), and G>T (blue) by race.

Racial disparity in the transcriptome profile of myometrium with LM

Given that the progression of LM in Black women is more aggressive, it was hypothesized that the gene expression profile in myometrium with LM (MyoF) would be different in comparison to White women. As such, we assessed the transcriptome of myometrium derived from 10 White and 10 Black women. One sample from a Black woman yielded low quality of RNA-sequencing data and was excluded in further analysis. RNA-sequencing analysis on these patient samples revealed 270 upregulated and 374 downregulated transcripts that varied dependent on race alone, as summarized in the heatmap (Fig. 2A, Supplementary Table SII). Additionally, a volcano plot of the data showed differentially expressed transcripts that were significantly higher (red) or lower (blue) in the MyoF from Black compared to White women (Fig. 2B). GO pathway analysis of the differentially expressed genes indicated that the top-ranking pathways included genes associated with cell adhesion, oxidation–reduction process, organ development, inflammation, and others (Fig. 2C). PCA analysis of differentially regulated genes further highlighted the differences in White and Black derived patient samples (Fig. 2D). Gene enrichment plot pathway analysis showed that genes associated with ROS, hypoxia, and oxidative phosphorylation pathways were significantly altered in myometrium from Black compared to White patients (Fig. 2E), amongst others such as p53, phosphoinositide 3-kinase (PI3K), estrogen response pathways (Supplementary Fig. S1). This data were expanded upon with the use of KEGG enrichment analysis, the process of mapping genes to molecular interactions and related networks (Supplementary Fig. S2). Interactions with cytokines, Toll-like receptor pathway, chemokine signaling, and others support increased inflammatory and signaling pathways. Altogether, these analyses demonstrate transcriptomic differences in MyoF of White and Black women converge around increased oxidative stress, signaling, and inflammatory pathways.

Figure 2.

Evaluation of differentially expressed genes in myometrium of Black and White patients with leiomyoma. (A) RNA-sequencing heatmap of myometrium tissue present for leiomyomas (MyoF) in 10 White and 9 Black women, comparing global transcriptional change across the groups (left: Black, right: White). (B) Volcano plot showing (270 upregulated and 374 downregulated genes) significantly differentially expressed genes in MyoF (with >5 fibroids) between Black and White women of comparable age, tumor number, and hormone status. (C) Gene Ontology (GO) pathway analysis revealed that the top-ranked pathways were related to cell adhesion, oxidation-reduction process, organ development, inflammation, and others. P-value is ranked from low (blue) to high (red). (D) Principal component analysis (PCA) showing distinct gene expression in 10 White versus 9 Black patients (each dot represents one tissue sample). (E) Gene enrichment plot pathway analysis from GSEA revealing significantly changed pathways in Black (red, positively correlated) versus White (blue, negatively correlated) patients (n = 19) for reactive oxygen species, hypoxia, and oxidative phosphorylation.

Racial disparity in ROS burden and oxidative stress response

As the ROS pathway was enriched by both GO pathway and gene enrichment plot analysis, we investigated oxidative stress pathway-related markers to further evaluate the racial disparity observed at the transcript level, and potentially uncover hints as to an underlying mechanism. Uterine LM and myometrium from 60 Black and 58 White women, as well as 34 control myometrial samples (each from a uterus without LM, defined as MyoN) were collected and a tissue microarray was made. The LM samples from White and Black women were taken from those with similar sized tumors, uterine weights, and number of tumors (Table II). Immunohistochemical staining for 8-OHdG, an oxidized nucleoside of DNA that serves as a biomarker of oxidative stress, was done on both the tumors and matched myometrium in the TMA (Fig. 3A). The quantitative immunoscores (H-Score) demonstrated significantly higher 8-OHdG levels in both LM and matched MyoF from Black women (P < 0.001) with an overall 1.5-fold difference (Fig. 3B).

Table II.

Summary of cases for tissue microarray analysis.

Type		Black	White	MyoN	P-value
No. cases (n)		60	58	34	Black vs White
Age (years)	Mean ± SEM	42.52 ± 0.577	45.21 ± 0.496	42.15 ± 1.075	0.0006
Tumor size (cm)	Mean ± SEM	8.32 ± 0.630	7.27 ± 0.590		0.1344
Uterine weight [g]	Mean ± SEM	890.2 ± 102.7	642.9 ± 103.1		0.0067
No. of tumors	<5	23	33		0.0435
	≥5	37	25
Endometrial phase	PE	26	30		0.0985
	SE	23	12
	IA/AT	11	16

SEM, standard error; NA, not available; PE, proliferative endometrium; SE, secretory endometrium; IA/AT, inactive or atrophic endometrium.

Figure 3.

Immunohistochemistry analysis of oxidized guanine (8OHdG) in leiomyoma (LM) and matched myometrium in Black and White women. (A) Immunostaining for 8-OHdG, a marker for oxidative stress, in LM and matched myometrium with LM (MyoF) from 60 Black women (left) and 58 White women (mid panel), and myometrium without LM (MyoN) from 34 White women (right) in a high-density tissue microarray (TMA). (B) Plots illustrating the quantification of 8-OhdG via H-scores. Each plot indicates scores for: myometrium without LM (MyoN, light blue), myometrium with LM from White women (MyoF-W, blue), LM from White women (LM-W, dark blue), myometrium with LM from Black women (MyoF-B, light red), and LM from Black women (LM-B, dark red). ***P < 0.001, ^#Significance compared with MyoN. Scale bars = 1.5 mm.

To further evaluate the mechanism underlying the greater ROS burden and DNA damage response seen in uterine tissue from Black women, we selected an additional eight markers (Fig. 4, Supplementary Fig. S3) for examination by immunohistochemistry: OGG1 (a base excision repair enzyme for 8-OHdG), ROS response markers and metabolites (HO-1, KEAP1, SOD, SOD^K122Ac, SOD^K68Ac), and steroid hormone receptors (ER, PR). Immunostaining for OGG1 was significantly higher in both the myoF and LM from Black women compared to those of White women as well as MyoN, which showed the lowest levels of OGG1 (Fig. 4A). HO-1 and KEAP1 (Fig. 4A) and SOD^K122Ac, SOD^K68Ac (Supplementary Fig. S4) showed similar levels of staining between the Black and White tissues but were significantly higher than in MyoN. Total SOD, ER, and PR showed similar staining intensities for all samples (Supplementary Fig. S4). The unsupervised cluster analysis by heatmap allowed for the visualization of differences between the MyoF and LM from Black and White women (Fig. 4B). Correlation matrix of all the data in Fig. 4C, highlighted the association between the markers assessed with the strongest correlations highlighted in the dark blue circles. Specifically, three clusters of strong correlation were found among ER/PR, 8-OHdG/HO1/OGG1, and SOD/KEAP1, indicative of gene expression corresponding to their functions (Fig. 4C). The strong correlations between 8-OHdG and OGG1 (R = 0.66) and HO-1 (R = 0.62) in individual tumors and myometrium are summarized in Fig. 4D. The results here show the interaction of ROS-associated genes which are maintained in tumors from Black versus White women. Interestingly, both myometrium and LM from Black women aggregated proximately in PCA analysis based on these nine markers (Fig. 4E).

Figure 4.

Immunohistochemistry analysis of reactive oxygen species (ROS) and related gene products in leiomyoma (LM) and matched myometrium in Black and White women. (A) Plots illustrating the quantification of selected markers (OGG1, HO-1, and KEAP1) via H-scores. Each plot indicates scores for: myometrium without LM (MyoN, light blue), myometrium with LM from White women (MyoF-W, blue), LM from White women (LM-W, dark blue), myometrium with LM from Black women (MyoF-B, light red), and LM from Black women (LM-B, dark red). *P < 0.05; **P < 0.01; ***P < 0.001, ^#Significance compared with MyoN. (B) Heatmap illustrating the overall expression levels of nine selected biomarkers of ROS and target genes in MyoN, MyoF-W, LM-W, MyoF-B, and LM-B. (C) Correlation analysis of nine selected markers in a correlation matrix. The size and intensity of the blue circle indicates the level of correlation. (D) Correlation analysis of biomarkers 8-OHdG, OGG1, and HO-1. Each dot represents one tissue sample, and five different colors (right top corner) indicate the racial group and tissue type. (E) Principal component analysis (PCA) showing distinct gene expression. Each dot = one tumor.

Discussion

Uterine LM play a key role in socioeconomic burden associated with medical care and should be considered a major public health issue. Women with symptomatic LMs have a considerable loss in quality of life, particularly within racial minorities or those in lower income brackets (Marsh et al., 2018). Black women tend to have earlier disease onset, with more numerous and larger tumors and more severe symptoms of LM than White women (Kjerulff et al., 1996). Our results indicated that LM from Black women have a higher rate of MED12 mutations (79.0%) than that from White women (68.5%) (Fig. 1B and C). We have shown that White and Black women have considerable differences in their LM profiles. As such, this may be one contributing factor, amongst several, which collectively lead to the pathogenesis of LM, potentially shedding light on why LM are so prevalent in Black women. In this study, we have illuminated genetic differences between White and Black patients with LM.

The RNA-sequencing results generated from MyoF of Black and White patients have provided a valuable insight into differentially regulated genes and pathways (Fig. 2A and B). Analyses revealed differentially expressed genes dependent on ethnicity, wherein ROS, hypoxia, and oxidative phosphorylation pathways were positively correlated with samples derived from Black patients. GSEA showed statistically significant differences between the two groups, in this case, based on race. Gene enrichment was positively correlated with numerous pathways in only the Black patients (Fig. 2E, Supplementary Fig. S1), implying that these biological processes could be responsible for the differences experienced by Black and White women with uterine LMs; this is an important step to elucidating a mechanism behind LM development.

Compelling evidence for the development of LM through MED12 mutations exist. Firstly, the mutation positions affect evolutionarily conserved and functionally important residues, for example, disrupting mediator association with Cyclin C-CDK8/CDK19 and the loss of mediator-associated CDK activity (Turunen et al., 2014). Secondly, Mäkinen et al. (2011b) demonstrated that monoallelic expression of mutant MED12 was observed in MED12 positive LMs, suggesting a requirement for functionally altered MED12. Recently, Takao et al. (2022) demonstrated that gain-of-function MED12 mutations can also induce partial LM cell properties in human uterine smooth muscle cells. LMs typically contain various extracellular matrix proteins, such as collagens, fibronectins, and laminins (Commandeur et al., 2015). The gain-of-function MED12 mutations generated by Takao et al. (2022) were shown to influence collagen production. Thirdly, over the last few decades, there have been numerous rodent models designed to capitulate LM development. There have also been several novel models of LMs proposed lately for a variety of animal species (Piccini et al., 2022). Patient-derived MED12 mutant stem cells are tumorigenic in mouse xenograft models (Ono et al., 2012). Similarly, in 2015, Mittal et al. (2015) generated a mouse model that conditionally expressed a Med12 missense variant (c.131G>A) in the uterus and were able to demonstrate that this alone led to the development of uterine LM.

The TMA immunostaining data of oxidative stress markers indicated a striking disparity between White and Black women, which was previously unknown (Fig. 3). Expression of 8-OHdG and the DNA repair enzyme, OGG-1, was increased and strongly correlated in myoF and LM of Black women, suggesting a high oxidative burden in the uterus of Black women resulting in an increase in base-excision repair activity and the ROS metabolic pathway (Fig. 4). Oxidative stress occurs when there is an imbalance of the production and elimination of ROS, which has damaging consequences on cells. Manganese superoxide dismutase (MnSOD) is an essential mitochondrial antioxidant enzyme that detoxifies reactive metabolites. We have previously demonstrated that LMs in general have increased levels of acetylated MnSOD at K122, which decreases its detoxification enzyme activity (Vidimar et al., 2016). Additionally, it has been demonstrated previously that genomic oxidized guanine, a marker of oxidative stress, was found at a significantly higher level in the MyoF that had multiple LM compared to MyoN (Li et al., 2022). It could be argued that the higher ROS burden experienced by Black women could influence MnSOD functionality, hence the disease is more aggressive. However, immunohistochemistry (IHC) staining revealed no difference in MnSOD^K122Ac levels in the tumors between Black and White patients, suggesting other sources of increased oxidative stress burden. Studies have suggested that there are racial disparities in oxidative stress (Morris et al., 2012; Szanton et al., 2012) which are consistent with disparities in chronic exposure to psychosocial and behavioral stressors. Higher levels of oxidative stress occur in Black people, in comparison to the White population (Feairheller et al., 2011; Morris et al., 2012). Ethnic and racial minorities experience disparities in morbidity and mortality from cancers and it is thought that their capacity to stimulate a response to oxidative stress will influence their susceptibility to cancers (Zhang et al., 2019). In particular, ROS production is higher in Black women than in men, with no observed gender differences in White individuals (Gardner et al., 2015). Oxidative stress is hypothesized to be a precursor to many health conditions (Price et al., 2013), in particular those that show racial disparities, e.g. cardiovascular disease and diabetes mellitus (Perrin et al., 2007; Van Der Lee et al., 2007). On average, despite the lower visceral fat tissue in Black than White people, women from the Black population still have similar or higher concentrations of inflammatory biomarkers, such as interleukin-6 (IL-6) and C-reactive protein (Carroll et al., 2009). Reports support that oxidative stress is influenced by lifestyle factors that include diet, physical activity, and psychosocial variables. We have shown in this study that there is a higher burden of oxidative stress activity in the myometrium and LM from Black women, which could have damaging effects on DNA and cell function.

In our study, we found a modestly higher, but statistically significant, increase in the rate of MED12 mutations in LM of Black women. The positive correlation of MED12 mutations with LM tumor numbers in the uterus has been reported previously (Bertsch et al., 2014), but this is the first study to report a racial disparity in MED12 mutation frequency (Fig. 1). The high rate of MED12 mutations in Black women raises questions of the underlying risk factors. We found high 8-OHdG (Fig. 3) levels and a high MED12 mutation rate (Fig. 1) in LM from Black women, which may in part, alongside other contributing causes, explain the observed racial disparities in LM frequency and severity. It has been hypothesized that the myometrium of Black women may be more susceptible to the development of LM due to underlying factors. As such Paul et al. (2022) examined the transcriptomic and DNA methylation profiles in the MyoF of Black versus White women. The results revealed several notable findings; MyoF samples clustered by race in both transcriptome and methylation profiles, whereas LM samples only clustered by race in their methylation profile. In addition to these, there was a greater number of differentially expressed genes in MyoF in comparison to LM (Paul et al., 2022). This demonstrated that a progenitor for the observed racial disparity could begin in the myometrium prior to development of LM themselves. In a separate study, we demonstrated that administrated oxidative stress in myometrial cells in vitro could significantly induce MED12 misrepair at c.130-c.131 (Li et al., 2022). Our findings suggested that there was a possible underlying mechanism which may explain the high rate of MED12 mutations and LM burden in women of reproductive age, particularly in Black women. There are, however, other contributing factors that may also influence the prevalence of LM in Black women; an expanded Myo stromal cell population has been shown to correlate with parity and LM number and could lead to an increase in tumor-initiating cells (Fernung et al., 2018).

The data in this article provide clear evidence of a disparity between Black and White women who have developed LM. Despite this, there are some potential limitations of the study. IHC expression in the high-density TMA relied on the use of MyoN only from White women as a control. This was due to the difficulty in obtaining a sufficient sample number of MyoN tissues from Black women. Furthermore, although mutation rates MED12 are prevalent in LM, its role in the development of LM remains to be demonstrated. A previously published study has shed some light on the possible origin of MED12 mutations in the myometrium prior to LM development; we demonstrated that ROS inducers cause oxidation of DNA, and that 8-OHdG can eventually be misrepaired and mutated as demonstrated using replacement of G with oxidized G at c.130 of the MED12 gene. Importantly, the myometrial cells were treated with paraquat (an inducer of ROS) and assessed via deep sequencing, and the results indicated that the chronic treatment of paraquat resulted in the development of MED12 mutations (Li et al., 2022). Therefore, it is very likely that the MED12 mutations occur in the myometrium of Black women prior to the development of LM. Our studies warrant further investigation by researchers in the field on the molecular changes that promote the development of LM and differences observed between White and Black women.

Conclusion

There are significant differences in the MyoF and LM profiles between Black and White women: differing transcriptomic profiles, responses to oxidative stress, and incidences of MED12 mutations. LM is a major health concern, given the frequency of occurrence and the harmful symptoms experienced especially in Black women. As such, it is of great importance that the underlying mechanisms are elucidated, particularly with regards to the observed racial disparities. Future directions should include studies to uncover the mechanisms of LM development and evaluate the role of MED12 mutation on the incidence of LM between races.

Data availability

The data underlying this article are available in NCBI and can be accessed at https://www.ncbi.nlm.nih.gov/bioproject/PRJNA859428.

Authors’ roles

Y.L., J.J.K., and J.-J.W. conceived experiments. Y.L. and Y.F. performed the experiments and data analysis. R.P.M. wrote the article. R.P.M., J.J.K., and J.-J.W. revised the article. All authors reviewed the article.

Funding

The authors have nothing to disclose. This study was supported by NCI (R01CA254367) and NICHD (P01HD057877).

Conflict of interest

The authors declare no conflict of interest.