Uterine adenomyosis is an oligoclonal disorder associated with KRAS mutations

Abstract

Uterine adenomyosis is a benign disorder that often co-occurs with endometriosis and/or leiomyoma, and impairs quality of life. The genomic features of adenomyosis are unknown. Here we apply next-generation sequencing to adenomyosis (70 individuals and 192 multi-regional samples), as well as co-occurring leiomyoma and endometriosis, and find recurring KRAS mutations in 26/70 (37.1%) of adenomyosis cases. Multi-regional sequencing reveals oligoclonality in adenomyosis, with some mutations also detected in normal endometrium and/or co-occurring endometriosis. KRAS mutations are more frequent in cases of adenomyosis with co-occurring endometriosis, low progesterone receptor (PR) expression, or progestin (dienogest; DNG) pretreatment. DNG's anti-proliferative effect is diminished via epigenetic silencing of PR in immortalized cells with mutant KRAS. Our genomic analyses suggest that adenomyotic lesions frequently contain KRAS mutations that may reduce DNG efficacy, and that adenomyosis and endometriosis may share molecular etiology, explaining their co-occurrence. These findings could lead to genetically guided therapy and/or relapse risk assessment after uterine-sparing surgery.

Conflict of interest statement

The authors declare no competing interests.

Figures

Fig. 1. Number and allele frequency of single-nucleotide variations (SNVs) identified in adenomyosis and co-occurring lesions.

a Number of somatic mutations in samples of adenomyosis (n = 51) and co-occurring cases of endometriosis (n = 7), leiomyoma (n = 8), and ovarian cancer (n = 5) presented as Tukey’s Box-and-Whisker plots. b Variant allele frequency (VAF) identified in the adenomyosis and co-occurring endometriosis, leiomyoma, and ovarian cancer samples in a presented as Tukey’s Box-and-Whisker plots. c–e Recurrence of the indicated gene mutations in adenomyosis (c), co-occurring endometriosis (d), and co-occurring leiomyoma (e). Top boxes: each box represents an individual SNV that was identified by whole exome sequencing (WES). SNVs were validated (grey) or not (white) by targeted deep sequencing (TDS). Middle boxes: samples were subjected to WES (purple) or TDS (yellow). Adenomyosis patients with co-occurring endometriosis (blue), leiomyoma (orange), and/or ovarian cancer (purple), and low progesterone receptor (PR) expression (black) are indicated above the patient number. Bottom boxes: each box represents an individual mutation of KRAS (red), PIK3CA (turquoise), PPP2R1A (blue), or MED12 (brown). f VAF values for KRAS mutations in adenomyosis lesions (blue) or peripheral blood (red) as determined by TDS. g Schematic representation of somatic mutations in KRAS.

Fig. 2. Somatic mutations in the epithelial component of adenomyosis lesions.

a, b HE staining of epithelial cells from adenomyosis Patient #6 (a) before and (b) after isolation by laser capture microdissection (LCM). Scale bars, 100 μm. c–e Quantification of VAF as determined by TDS of the indicated genes in peripheral blood (Blood), frozen adenomyosis lesions (Bulk-A), LCM-acquired epithelial component of adenomyosis (LCM-A), or LCM-acquired adjacent muscle cells (LCM-ADJ) from the indicated patients.

Fig. 3. Oligoclonality in adenomyosis tissues as revealed by multi-regional sampling.

a Left panel: coronal plane (top) and sagittal plane (bottom) schemes indicating the sites of multi-regional adenomyosis samples (1–6) and adjacent histologically normal myometrium (NM) that were acquired from Patient #28 and subjected to mutational profiling. R, right; L, left. Right panel: VAFs of mutations in the indicated genes in the indicated multi-regional samples profiled for each individual adenomyosis lesion, NM, or peripheral blood (B) sample from Patient #28. VAF values are shown using a color scale. b, c Top panels: sagittal plane schemes indicating the sites of multi-regional adenomyosis samples plus NM and leiomyoma (LM) samples that were acquired from Patient #29 (b) or Patient #5 (c) and subjected to mutational profiling. Bottom panels: analysis of VAFs for the samples in the top panels as for a. For a–c, raw VAF values (%) for each sample are shown in Supplementary Data 14–16.

Fig. 4. Multi-regional sampling reveals shared mutations in adenomyosis, endometriosis, and adjacent normal endometrium.

Sagittal plane scheme (top) and VAF analysis (bottom) of sites of multi-regional adenomyosis (1–6) plus normal endometrium (NE), normal myometrium (NM), leiomyoma (LM), and blood (B) samples from Patient #8. VAF values are depicted as in Fig. 3. b, c Coronal plane scheme (top) and VAF analyses (bottom) of multi-regional adenomyosis plus NE and/or NM samples from the indicated patients. d–f Coronal plane scheme (top), sagittal plane scheme (middle), and VAF analyses of multi-regional adenomyosis samples plus endometriosis samples (EN) and/or NM samples from the indicated patients. For a–f, raw VAF values (%) for each sample are shown in Supplementary Data 17–19. g–i Quantification of VAF values as determined by TDS of the indicated genes in frozen (Bulk) and LCM tissues from the indicated patients. Bulk tissues included peripheral blood (B), adenomyotic lesions (A), and endometriotic lesions (EN). LCM tissues included epithelial component of adenomyosis (A), endometriosis (EN), and normal adjacent tissues (ADJ). For g–i, raw VAF values (%) for each sample are shown in Supplementary Data 21.

Fig. 5. Increased mutation frequency and VAF of KRAS-mutated clones in the histologically normal endometrium adjacent to adenomyotic or endometriotic lesions.

a, b Histologically normal myometrium (NM; negative control) and endometrium (NE) were macro-dissected from the indicated three groups of patients. a Occurrence of the indicated gene mutations in NE of patients with adenomyosis (patient ID column: orange), patients with neither adenomyosis nor endometriosis (patient ID column: yellow), and patients with endometriosis but without co-occurring adenomyosis (patient ID column: purple). Bottom boxes: each box represents an individual mutation of KRAS (red), PIK3CA (turquoise), or PPP2R1A (blue). VAF values are shown using individual color scales. b Quantification of VAF values as determined by TDS of the indicated mutations in the macro-dissected NM (negative control) or NE from the patients in the group A (adenomyosis; orange in a), group Non-A/E (patients with neither adenomyosis nor endometriosis; yellow in a), and group E (patients with endometriosis but without co-occurring adenomyosis; purple in a). Source data are provided as a Source Data file. Data are values for individual patients plus the mean. *p < 0.05 by Welch’s t-test.

Fig. 6. Mutant KRAS expression abrogates the progestin-induced anti-proliferative effect of DNG through downregulation of PR.

a Immunoblotting to detect ectopically expressed, HA-tagged wild-type (WT) KRAS and PIK3CA, or KRAS or PIK3CA bearing the indicated amino acid substitutions, in immortalized human uterine epithelial cells. Vinculin, loading control. Viability (b) and proliferation (c) of WT immortalized uterine endometrial epithelial cells or those overexpressing the indicated mutations of KRAS or PIK3CA. b Cells were exposed to 1 μM DNG or vehicle (DMSO) for 72 h. Results are the percentage of viable cells relative to vehicle control cells. Data are the mean + SD of six independent experiments, each with three technical replicates per group. *p < 0.05 by Welch’s t-test. c Cells were exposed to 1 μM DNG or vehicle (DMSO) for 48 h. Results are the percentage of BrdU+ cells relative to vehicle control cells. Data are the mean + SD of four independent experiments per group without technical replicates. *p < 0.05 by Welch’s t-test. d Quantitative RT–PCR determination of mRNA levels of PR-A/B in the immortalized cells in a. Values were normalized to GAPDH. Data are the mean fold change + SD relative to levels in control cells stably infected with empty vector. Six independent experiments, each with three technical replicates per group, were conducted. *p < 0.05 by Welch’s t-test. Source data are provided as a Source Data file.

Fig. 7. Downregulation of progesterone receptors (PRs) in KRAS-mutated adenomyosis as assessed by immunohistochemical (IHC) analyses.

Representative photomicrographs of (upper) IHC of PR and (lower) HE staining of adenomyotic lesions from the indicated patients. Arrowheads, epithelial component. Scale bars, 100 µm.

Fig. 8. Mutant KRAS expression is linked to epigenetic silencing of PR.

a Schematic diagram of the PR-A/B promoters and exon 1. Transcription start sites (arrows), CpG islands, and regions of bisulfite sequencing are indicated. b Frequency of methylated CpG in PR-B (top) and PR-A (bottom) as determined by bisulfite sequencing of DNA from KRAS-WT (n = 20) and KRAS-mutated (n = 11) adenomyosis samples. Data are the mean + SD. *p < 0.05 by Welch’s t-test.