EHMT1 mosaicism in apparently unaffected parents is associated with autism spectrum disorder and neurocognitive dysfunction

Kleefstra syndrome (KS; OMIM #610253) is caused by haploinsufficiency of the euchromatin histone methyltransferase 1 (EHMT1) gene located at chromosome 9q34.3. Almost all cases of KS are due to de novo microdeletions or intragenic loss-of-function mutations in this gene [1, 2]. So far, no inherited cases have been reported, except for a few cases on the basis of parental mosaicism. Mosaicism for EHMT1 defects appears to be rare. At present, only four cases have been reported: three with a 9q34.3 microdeletion [3, 4] and one with a splice-site mutation in EHMT1 [5]. These carriers of a genetic EHMT1 mosaicism were reported to function normally in daily life. However, the focus of these reports was on the genetic and somatic features of the subjects. No data was presented on the possible psychopathology, which is the most pronounced comorbidity in patients with KS caused by a germline EHMT1 defect [6–9].

The full KS phenotype is characterized by the core triad of intellectual disability (ID), (childhood) hypotonia, and distinct facial features. Additional clinical features include specific behavioral characteristics, heart and urogenital defects, epilepsy, and overweight [1, 2, 10]. Recently, we studied the specific profile of behavioral characteristics in KS, which led us to conclude that KS patients are extremely vulnerable to developing severe psychiatric disorders. In our cohort, the prevalence of autism spectrum disorder (ASD) is nearly 100% [6]. In addition, we found high prevalence of mood disorders and psychotic disorders. Moreover, patients above the age of 18 years in our cohort showed severe regression with a loss of at least 30–50% of their initial functioning [6, 8, 11]. Interestingly, recent studies have shown that de novo postzygotic mutations (PZM) in important developmental genes contribute to ASD [12, 13]. Based on this evidence, we set out to investigate the presence of psychopathology in the parents with an EHMT1 mosaicism.

In this study, we examined three subjects with EHMT1 mosaicism for the presence of psychopathology. All three subjects are seemingly unaffected parents of children diagnosed with KS caused by a germline EHMT1 defect. We hypothesized that individuals with EHMT1 mosaicism are vulnerable to psychopathology which may have significant clinical consequences.

Subject 1 is a 44-year-old married male. He is a father of two children; the youngest son was diagnosed with KS. After finishing a higher professional education, he started working as a data engineer. He does not report social or communicative problems himself; meaningful social contacts are limited, even with family members and colleagues. He had a normal development. His medical history includes an inguinal hernia and hydrocele in childhood, both of which were surgically corrected. The family history does not mention any medical or developmental problems. At physical examination, no typical facial characteristics were observed.

Subject 2 is a 43-year-old female, mother of two children, and her youngest daughter was diagnosed with KS. The somatic characteristics of subject 2 and her youngest daughter have been previously described, represented as family 2 [3]. At the age of 13 years, the subject emigrated from Morocco to the Netherlands. She experienced learning difficulties and did not finish high school. Besides running the household, she started working as a care assistant for elderly people. She was not able to take care of her own financial administration, because of learning difficulties. Her psychiatric history reported a MDD during pregnancy for which fluoxetin was prescribed. Further medical history mentioned three spontaneous abortions, psoriasis, and nephritis. Additionally, she experiences sleep disturbances comprising apneas and frequent awakenings during the night.

Her social network was limited to her family. She had no friends and experienced problems in initiating and maintaining contacts and conversations. Small talk was difficult for her. At the time of examination, she was divorced and not able to work due to mental health problems, consisting of problems in social reciprocity, reactive depressive complaints, and overstimulation and burnout.

At the time of our investigations, she experienced a second major depressive episode with psychotic features (imperative hallucinations and delusions). The imperative hallucinations had led to a suicide attempt. Her general practitioner was informed after our examination, and intensive outreaching care was suggested based on our observation. Physical examination of this subject at the age of 43 years showed obesity and minor dysmorphic features as previously described, including a hypoplastic midface, small upslanting palpebral fissures, depressed nasal root, and anteverted nares.

Subject 3 is a 39-year-old male with three children: the second child was diagnosed with KS. There is no further family history of somatic or neuropsychiatric disorders. After finishing primary school, he struggled to finish high school (practical education level) because of learning disabilities. He went through different jobs, but at the time of the current evaluation, he was not able to work due to his impaired somatic and mental health. He lives an isolated life, together with his wife and children. Despite his social isolation, the patient did not experience any problems in social behavior.

His birth was complicated by a forceps delivery. The remaining medical history revealed atrial fibrillation, knee surgery, tonsillectomy, tympanostomy, and mild auditory and visual impairment. Pharmacotherapy started after the atrial fibrillation was diagnosed and comprised the following: metoprolol, acenocoumarol, simvastatin, and perindopril. Physical examination at the age of 39 years showed obesity, an occipito-frontal circumference of 59 cm (nearly + 1 SD), and minor dysmorphic features including a hypoplastic midface and synophrys.

The genetic, psychopathological, and neurocognitive characteristics are presented in Table 1.

In this report, we describe the consequences of EHMT1 mosaicism on psychopathology in three adult subjects. The three seemingly unaffected parents included in this study had been referred to the clinical genetics department only because of the diagnostic trajectory of each parent’s respective child. The discovery of EHMT1 mosaicism in these subjects was unexpected [3–5]. The degree of mosaicism that was measured in different tissues varied from 15 to 40% (subject 1) to 80% (subject 2). It is difficult to predict the clinical consequences of mosaicism, as the patterns and distribution of abnormal cells can vary widely between different tissues, depending on the timing of the mutation events [20]. Despite this predicted variation, all three subjects had an ASD, a (prior) MDD, and other associated neurocognitive dysfunctions. Some features that can be seen in patients with a germline EHMT1 mutation had a variable occurrence in the mosaic individuals. The results are indicative of a vulnerability to develop severe psychiatric disorders that carriers of an EHMT1 mosaic mutation possess. This correlates with the high prevalence of psychopathology, which has been reported for individuals with the full KS phenotype due to a germline EHMT1 mutation [6, 8, 21]. Based on the ADOS test, all mosaic subjects met the criteria for ASD and all had a (prior) MDD. This is further corroborated by our findings on the CANTAB. All subjects showed weaknesses in performance on the IED, which is suggestive of cognitive inflexibility. This might be related to the set-shifting problems often seen in subjects with ASD, although findings on executive function deficits in adults with ASD were inconsistent [22]. These deficiencies on the IED are on the continuum compared with a significant high dropout on the IED in our previous research on subjects with the full KS phenotype (unpublished data). Subject 3 had no other deviating scores compared to his norm group. Subject 1 showed performance problems on the PRM, but we could not clearly link this to his clinical profile. Subject 2 had low scores on all tasks, most likely due to (a combination of) her learning and mood problems.

Psychopathology has a direct impact on adaptive functioning. The adaptive behavior measured with the VABS implies a developmental age above 12 years, consistent with adult functioning, for all of our subjects. Nevertheless, all three of them struggle to fulfill their role in society, most likely as a direct consequence of their neurodevelopmental problems. They became socially isolated and experienced difficulties in finding the right type of education and employment that would suit their strengths and weaknesses. Therefore, the clinical relevance for diagnosing these subjects with EHMT1 mosaicism and consequent neuropsychiatric assessment is significant. Optimal treatment of the psychopathology may improve daily functioning and positively influence the well-being of the individuals treated and their families, including the child with KS, as well as their participation in society. Based on our findings, personalized psychiatric evaluation and treatment should be considered for subjects diagnosed with EHMT1 mosaicism. The diagnostic procedure should include a special attention to ASD and mood disorders. Furthermore, these cases illustrate that presence of mosaic deletions could potentially be missed, since the genetic defects were only discovered during the diagnostic phase of the children with the full KS phenotype. None of the subjects had been referred to a medical specialist for their own psychiatric problems prior to our studies. These subjects tend to have a more subtle presentation, but this can nevertheless cause impaired (social) functioning. Therefore, our study also demonstrates that genetic analysis should be considered in patients with a psychiatric disorder and other characteristics of KS without the classic appearance.

Despite the similarities in psychopathology between our subjects, the consequences of mosaicism in general are widespread and unique for each individual [23]. The phenotypic expression of somatic mosaicism has been investigated and reported in several studies including mosaic trisomy 21 (Down syndrome) and mosaic fragile X syndrome [24, 25]. These studies show severe phenotypic expression, as well as somatic mosaicism without clinical consequences. The expression of the behavioral phenotype is also variable. In mosaic trisomy 21, for example, the social adaptive skills were significantly better in the subjects with a mosaicism compared to non-mosaic subjects with Down syndrome. However, the intellectual performance did not relate to the level of mosaicism [24]. This necessitates interpretative caution, especially since we assessed the psychopathology of only three subjects with an EHMT1 mosaicism and each had different levels of mosaic expression. Interestingly, in recent studies, it was shown that rare postzygotic mutations in important developmental genes contribute significantly to ASD [12, 13]. Our present results obtained for three ASD subjects with such a rare mosaic defect in EHMT1 are in line with these reports. EHMT1 is an important developmental gene that is extremely intolerant to loss-of-function mutations (http://exac.broadinstitute.org) [26]. Further clinical studies are warranted to provide more insight into the wide spectrum of the EHMT1 mosaicism, although the extremely low prevalence of such mosaic defects hampers the careful collection of symptoms in larger cohorts. For future research, we recommend to include more specific measures on IQ, ASD, and mood disorders in subjects with an EHMT1 mosaicism. These measures should be in addition to a (semi)structured clinical observation as people with neurodevelopmental disorders, like ASD, are more likely to have deviating scores on questionnaires [27, 28]. Thus, clinical interviews and observations should always be performed to gain a more comprehensive and objective insight into the psychopathology.

Subjects with EHMT1 mosaicism seem vulnerable to developing psychiatric disorders, especially ASD and mood disorders. However, since the behavioral outcome of mosaicism is unpredictable, these results should be interpreted with caution. Based on our findings in this small but unique cohort, we recommend screening subjects with an EHMT1 mosaicism for psychopathology. In our experience, multidisciplinary care, combining expertise in psychiatry, clinical genetics, and clinical neuropsychology, is of additional value in these complex cases where a broad perspective is required. Ideally, a careful diagnostic assessment of psychopathology, cognition, and physical health should be performed to provide an integrative diagnosis. Deficits in daily functioning and behavioral problems may well be a result of an underlying genetic or otherwise somatic cause, which is often not clearly present. Clinical considerations should therefore include the possibility of a genetic mosaicism in parents who have one or more children with a neurodevelopmental disorder of genetic origin and a typical combination of complaints, as illustrated in this study.