Deletion of 17q12 was first reported in MRKH syndrome by Cheroki et al. in 2008 (81) and to this date, 21 deletions have been reported ( Table 1 ) (30, 71–73, 75, 76, 81–86). Ledig et al. reported a 17q12 deletion in one MRKH syndrome patient (85), whose mother and sister were later reported with the same deletion and other uterovaginal anomalies (124). Deletions of 17q12 are typically 1.4 Mb in size. They are highly penetrant with variable expressivity and cause a multisystem disorder which may include kidney disease, neurocognitive impairment, endocrinological disease including HNF1B-related maturity-onset diabetes of the young (MODY), genital malformations, liver disease, and other manifestations (80). Two candidate genes for MRKH syndrome, LHX1 and HNF1B, are located at this locus, both of them being involved in MD development.

Lhx1, formerly referred to as Lim1, encoding the transcription factor LIM homeobox 1, is involved in normal kidney and MD development (125, 126). Lhx1-null female mice have normal ovaries but lack their reproductive tract, which results from a disruption of MD elongation and epithelium formation (125–127). Besides deletions, six single nucleotide variants in LHX1 ( Table 2 ) (72, 85, 119, 120) have been reported including one missense variant with functional evidence of decreased transcriptional activity (120). Still, LHX1 mutational analysis of larger cohorts did not report any variants, suggesting that sequence variants of LHX1 are no major cause of MRKH syndrome (30, 128).

HNF1B has long been a candidate gene for MRKH syndrome since Lindner et al. already in 1999 reported a Norwegian family with MODY type 5 and progressive parenchymal kidney disease following autosomal dominant inheritance caused by a 75 bp in-frame deletion in exon 2 of HNF1B. Notably, two of four female variant carriers also had uterovaginal agenesis, supporting MRKH syndrome as part of the HNF1B disease spectrum (51). Subsequent studies have reported HNF1B variants associated with various uterovaginal malformations (52, 121, 129, 130), including a ~59 kb whole-gene deletion of HNF1B found by WES analysis of a 9-year-old girl presenting with precocious puberty. After the genetic result, reverse phenotyping by imaging confirmed uterus agenesis (121). Recently, Thomson et al. (83) investigated the function of Hnf1b following conditional ablation in mice, which resulted in a hypoplastic uterus and renal anomalies (including renal agenesis) similar to an MRKH syndrome type II phenotype. They performed single-cell RNA sequencing of the Hnf1b-ablated embryonic uterine tissue and found dysregulated gene expression involved in cell proliferation, patterning, and differentiation. This supports HNF1B as a causative gene in MRKH syndrome independent of LHX1 involvement.

In 2011, Nik-Zainal et al. (71) reported four cases with recurrent deletions of 16p11.2 and suggested TBX6, previously implicated in paraxial mesoderm development (131), as a candidate gene. To date, a large number of cases have been identified with both deletions (30, 71–79) and TBX6 sequence variants (72, 75, 103, 105, 111, 112). The 16p11.2/TBX6 locus is also associated with congenital scoliosis (132). The genetics hereof is complex and does not follow typical Mendelian inheritance, requiring one TBX6-null allele and a particular hypomorphic trans allele, as described in the compound inheritance gene dosage model (132, 133).

Ma et al. reported 16 rare TBX6 variants enriched in a large MRKH syndrome patient cohort compared to controls. They performed various functional analyses of 13 missense variants and found evidence for loss-of-function in 7 variants, which together do support the role of TBX6 variants. However, in contrast to null alleles associated with scoliosis, no second risk alleles were reported in MRKH syndrome (105). As of now, no clear biological mechanism for monoallelic TBX6 variants causing MRKH syndrome has been established, which challenges interpretation and warrants further studies.

Both deletions (60, 71, 81, 85, 87–90) and duplications (81, 85, 91, 92) of 22q11 have been reported in MRKH syndrome patients ( Table 1 ). The CNVs vary in size and location with no single common overlapping region, and therefore no particular candidate genes at 22q11 have been identified. Of note, 22q11 duplication has also been associated with other uterovaginal malformations (124). 22q11 deletions are often associated with DiGeorge and velocardiofacial syndromes. Features hereof are partly thought to be caused by TBX1 haploinsufficiency, however, some of the deletions reported with MRKH syndrome do not include this particular gene. Overall, the evidence for 22q11 imbalances causing MRKH syndrome remains low.

Variable-sized deletions at 1q21.1 have been associated with MRKH syndrome (30, 82, 85, 93). Duplications of 1q21.1 have also been identified in an MRKH syndrome patient (81) and one case with uterus didelphys (82). The overlapping region has been determined to be GRCh38: chr1:145,779,056-146,019,795 (82) with RBM8A as the proposed candidate gene in which sequence variants/polymorphisms also have been identified associated with MRKH syndrome (112).

Deletions of this region may also cause thrombocytopenia–absent radius (TAR) syndrome in compound heterozygosity with certain non-coding polymorphisms on the trans allele (134). Notably, TAR syndrome has previously been reported in a case with MRKH syndrome (135, 136). The possible causal role of 1q21.1 deletions/RBM8A gene variants in MRKH syndrome is, however, still unclear warranting further studies to establish causality.

Pathogenic variants in PAX8 are an established monogenic cause of congenital hypothyroidism due to thyroid dysgenesis (CH, OMIM #218700), first described in 1998 by Macchia et al. (137). In an experimental study, Pax8-deficient female mice were reported with infertility independently of thyroid replacement therapy, and a possible role in MD development was suggested (138).

Years later, Ma et al. reported on the investigation of a 12-year-old girl with global developmental delay, MD agenesis, and hypothyroidism, found to carry a large 10.79 Mb deletion of 2q13q14.2 (94). A partially overlapping 2q12.1q14.1 deletion was reported in another case with CH, atrial septal defect, intellectual disability, and MD agenesis with anterior displacement of the anus (95). The two deletions share an overlapping region of 4.3 Mb spanning PAX8 as the proposed candidate gene for MRKH syndrome.

Most recently, Chen et al. (75) reported on a mutational burden analysis of 19 candidate genes based on WES data from 442 cases and 941 controls. Among cases, they found enrichment for predicted loss-of-function variants in PAX8. When including results from a replication cohort (n=150), a CH cohort (n=5), and missense variants, a total of 11 PAX8 variants were found associated with MRKH syndrome. In three cases with available parental DNA, paternal inheritance was confirmed, showing a sex-limited expressivity of infertility. Functional analysis of the five missense variants by luciferase reporter assay found evidence for loss-of-function in two variants. Finally, reverse-phenotyping of five female CH cases, revealed uterovaginal aplasia in one individual, emphasizing the pleiotropy of PAX8. This confirms MRKH syndrome as a part of the PAX8 disease spectrum in females, which the authors refer to as CH-MRKHS (75).

In 2017, three unrelated studies identified GREB1L variants as a new autosomal dominant cause of congenital anomalies of the kidney and urinary tract (CAKUT, OMIM #617805) (109, 110, 139). Some of the female cases, mainly fetuses affected by bilateral renal agenesis, were also described with uterovaginal malformations supporting a possible link to MD development (109, 110).

In 2019, Herlin et al. (101) reported on the investigation of a three-generational family with four cases of renal agenesis, including two adult female cousins with MRKH syndrome type II with unilateral renal agenesis. Whole-exome sequencing analysis in this family identified a segregating missense variant in GREB1L, supporting GREB1L variants as a novel monogenic cause of MRKH syndrome associated with incomplete penetrance and sex-limited expressivity and a phenotype mirroring the early descriptions of hereditary renal/urogenital adysplasia (41–43, 140, 141). Jacquinet et al. (102) identified GREB1L variants in 5 of 63 (7.9%) sporadic MRKH syndrome patients and segregating variants in four multiplex families presenting renal and uterovaginal malformations. Buchert et al. reported a stop-gain variant in a monozygotic twin-pair discordant for MRKH syndrome with the other twin having unilateral renal agenesis (98). Most recently, Jolly et al. performed an unbiased rare variant enrichment analysis based on WES data from a large American-European cohort (n=148), identifying GREB1L as the only gene approaching exome-wide significance based on seven detected variants. From a replication cohort of 442 Han Chinese cases, additional variants were found, including in six cases with type I MRKH syndrome. Of note, besides kidney and uterovaginal malformations, GREB1L variants have also been associated with inner ear malformations and deafness as well as complex congenital heart disease (142, 143).

GREB1L is considered to be involved in retinoic acid signaling, although its protein remains poorly characterized (109). Greb1l-knockdown in zebrafish causes abnormal pronephros morphogenesis (110, 139), and injection of wild-type human mRNA has been shown to rescue the phenotype (110). Knock-in mutagenesis of one missense variant in mice has also been shown to cause renal agenesis (139). Homozygous knock-out of Greb1l in mice has been shown to cause absence of the kidneys, Wolffian ducts, and Müllerian ducts (109). However, GREB1L variants reported in humans are monoallelic and predominantly missense variants ( Table 2 ), and the pathogenic mechanism of how these missense variants cause MRKH syndrome is still unknown requiring further functional analysis.

Taken together, the current evidence with 30 reported variants ( Table 2 ) including disease-segregating variants in extended pedigrees, epidemiological evidence of rare variant enrichment in larger cohorts, and functional evidence from knock-out mice, suggest GREB1L as a major causative gene in MRKH syndrome.

Gervasini et al. (62) first reported on the association of partial SHOX duplications in MRKH syndrome in 2010. They investigated 30 cases and 53 controls and identified 5 cases with SHOX duplications including a sib-pair with MRKH syndrome. No duplications were identified among controls. Guerrier and Morcel (64) reported similar findings with three SHOX duplications and one duplication downstream of the gene. In contrast, Sandbacka et al. found no duplications among 101 Finnish cases questioning the role of SHOX in MD development (63). More recently, Mikhael et al. reported two missense variants from an investigation of 72 candidate genes in 111 cases (103). The functional role of these duplications (including their insertion sites) and missense variants is unknown and causality has not been established.

Wnt9b encodes a protein involved in MD development and Wnt9b knock-down in mice causes uterovaginal and renal agenesis. Wnt9b is expressed in the Wolffian duct epithelium providing signals guiding MD elongation (144). WNT9B has therefore been considered a candidate gene in MRKH syndrome and has been investigated in several studies identifying a total of 9 sequence variants in MRKH syndrome type I, of these 7 missense variants ( Table 2 ) (98, 103, 113, 114). One patient has been reported with two variants (114). Other studies found no variants in MRKH syndrome patients (145, 146). WNT9B variants have also been reported in other uterovaginal malformations (114, 146). The functional role of these variants remains to be ascertained.

As previously described, WNT4 was the first gene identified as a monogenic cause of MD agenesis in females (OMIM #158330) (115). Since its discovery, a total of seven missense variants have been reported (103, 115–118). A missense variant of WNT4 has also been reported with another uterovaginal anomaly with renal agenesis (Herlyn-Werner-Wunderlich syndrome) (147).

Importantly, MD agenesis caused by WNT4 variants is associated with clinical and biochemical hyperandrogenism, representing a phenotype distinct from MRKH syndrome in general as described by Biason-Lauber et al. (116), and is often considered a separate entity. This is also supported by several investigations reporting no variants in larger MRKH syndrome and MD anomaly cohorts (30, 148, 149).

Lrp10 encodes low-density lipoprotein receptor-related protein 10, which has been proposed as a negative regulator of Wnt/β-catenin signaling (150), a pathway involved in MD development (151). In 2015, Rall et al. reported on a SNP-array analysis of CNVs in discordant twin pairs. In one affected twin, a 585 kb duplication at 14q11.2 spanning LPR10 was identified, not present in the other twin, and this gene was suggested as a candidate gene for MRKH syndrome. Two subsequent studies have reported a total of five missense variants in LRP10, hereof two variants in the same patient (74, 103). Functional evidence for these variants is lacking and causality has not been established.

The development of the field shows the importance of detailed family history in the genetic assessment, as it may include important information to suggest an underlying genetic trait or perhaps provide information on the genetic nature of the family’s trait. The early descriptions of hereditary urogenital adysplasia families, reported long before molecular genetic testing became available, is a good example hereof suggesting an autosomal dominant trait with sex-limited expressivity (41–43, 140, 141). In recent years, GREB1L variants have been identified as a major cause of MRKH syndrome, particularly in families with a hereditary urogenital adysplasia-like presentations (101, 102).

Family histories should not only include incidents of MRKH syndrome but also associated extragenital malformations and other uterovaginal malformations in family members. The subtle nature of many renal and uterovaginal malformations is, however, important to consider, as these may not yet have been revealed in asymptomatic relatives, requiring radiological imaging. The relevance of other uterovaginal malformations in the genetic assessment of MRKH syndrome patients is highlighted by the many genes reported in both phenotypes, including 17q12 deletion (85, 124), EMX2 (75, 152), GREB1L (102, 109, 110), HNF1B (52, 129, 130), HOXA10 (58, 75, 153), SPECC1L (74, 154, 155), WNT4 (115, 147), and WNT9B (114). This could suggest a spectrum of uterovaginal anomalies with shared causes.

Exome and genome sequencing have proven useful in the discovery of new genes during recent years, including GREB1L (99, 101, 102) and PAX8 (75). With the continuously developing field of genomic sequencing technologies and multi-omics analyses, there is reason to believe that additional genetic causes remain to be identified. Recently, the first complete telomere-to-telomere genome reference was published, which will support even better detection of genetic variation in future genetic studies (156). Also, our understanding of the role of non-coding parts in the genome such as topologically associated domains and long non-coding RNAs is increasing (157, 158). Genome sequencing today is typically based on short-read sequencing due to its cost-effectiveness and accuracy. Long-read sequencing has increased in use and provides some advantages in terms of de novo assembly, read-mapping, detection of structural variants, phasing of variants, and transcript isoform identification (159). Another technology for improved analysis of structural variation is optical genome mapping, which has been applied by Brakta et al. (73). All along the technical developments, it will still be relevant to revisit sequencing data of unsolved cases, as new knowledge emerges supporting variant interpretation. Sequenced MRKH syndrome cohorts will increase in size, as will available genomic data resources, which will improve rare variant discovery from association analyses.

It is also relevant to look for somatic variation or differential gene expression in affected tissues as previously performed on surgically-removed uterine remnants (35, 98, 106–108). This may also be helpful in functional analyses of candidate variants of unknown significance to provide evidence for pathogenicity. However, it should be considered that organogenesis in the embryo is an accurately orchestrated process with precise spatiotemporal control of gene expression (160). Therefore, gene expression in adult uterine remnants may not be representative of the gene expression occurring during embryonic development, which in that case would require mouse modeling. As the causal effect of many gene variants remains unknown, functional analyses will be important for further progress.

Finally, it will be important to investigate candidate variants in extended pedigrees to understand the mode of inheritance and disease segregation, also considering more complex mechanisms in developmental genomics as proposed in the compound inheritance gene dosage model (161).

Identifying genetic causes of MRKH syndrome is of immense importance including both academic pursuits towards improving our understanding of genetic factors underlying human uterovaginal development as well as establishing the necessary evidence to provide informed care and counseling for the individual patient and their families in clinical settings. Rare disease patients often undergo lengthy diagnostic odysseys from the first suspicion until the time of (genetic) diagnosis. Although the diagnostics in MRKH syndrome primarily involve finding the gynecological cause for primary amenorrhea, questions of “why did this happen?” remain and feelings of guilt and responsibility are prevalent among patients and their parents (11, 12). Identifying the underlying cause of malformation syndromes may reduce the burden for patients and families, even if it does not lead to changes in treatment (162).

Genetic counseling and management following the identification of a gene variant depends on the ‘clinical actionability’ of the variant, which requires both a valid gene-condition association and sufficient evidence for pathogenicity (163). As evidence for genetic causes in MRKH syndrome continues to increase, it may contribute to better disease classification in the future based on molecular causes instead of the current clinical distinction of type I and II. It can also be expected to guide future patient management such as examinations for associated diseases and anomalies specific to the particular gene. Clinically actionable variants may lead to genetic testing of family members at risk, and variant carriers may then be referred to relevant specialists for further examination. Finally, genetic counseling may address the recurrence risk if patients/couples with an identified genetic cause wish to pursue genetic parenthood either through gestational surrogacy or UTx. The possibility of recurrence in offspring was demonstrated in a recent report, describing the first case of mother-to-daughter inheritance following surrogacy (22). As UTx approaches clinical implementation (24) and the evidence for genetic causes increases, questions about offspring recurrence risk and reproductive choices will likely become more prevalent and so will requests for preimplantation genetic testing as part of the in vitro fertilization preceding the UTx procedure.

However, in the current state of knowledge, many reported variants associated with MRKH syndrome are still to be considered as variants of uncertain significance, which warrant cautious interpretations and counseling in clinical care.