Confirmation study of PTEN mutations among individuals with autism or developmental delays/mental retardation and macrocephaly

Abstract

There is a strong genetic component to autism spectrum disorders (ASD), but due to significant genetic heterogeneity, individual genetic abnormalities contribute a small percentage to the overall total. Previous studies have demonstrated PTEN mutations in a sizable proportion of individuals with ASD or mental retardation/developmental delays (MR/DD) and macrocephaly that do not have features of Cowden or Bannayan–Riley–Ruvalcaba syndrome. This study was performed to confirm our previous results. We reviewed the charts of individuals who had PTEN clinical sequencing performed at our institution from January 2008 to July 2009. There were 93 subjects tested from our institution during that period. PTEN mutations were found in 2/39 (5.1%) ASD patients and 2/51 (3.9%) MR/DD patients. Three additional patients without mutations had no diagnostic information. Multiple relatives of individuals with a PTEN mutation had macrocephaly, MR, or early onset cancer (breast, renal, and prostate). Of those relatives tested, all had the familial PTEN mutation. None of the affected relatives had previously been diagnosed with Cowden or Bannayan–Riley–Ruvalcaba syndrome. We noted in our previous study several adult relatives without any findings who carried a mutation. Combined with data from our previous cohort, we have found PTEN mutations in 7/99 (7.1%) of individuals with ASD and 8/100 (8.0%) of individuals with MR/DD, all of whom had macrocephaly. We recommend testing for mutations in PTEN for individuals with ASD or MR/DD and macrocephaly. If mutations are found, other family members should be offered testing and the adults offered cancer screening if they have a PTEN mutation.

It is well established that autism spectrum disorders (ASD) have a strong genetic component. Familial aggregation and twin studies demonstrate high sibling recurrence risks and high heritability [reviewed in Abrahams & Geschwind, 2008; Muhle, Trentacoste, & Rapin, 2004]. Many studies have now shown that the cause is heterogeneous, with specific genetic abnormalities individually accounting for only small percentages of the overall total [O'Roak & State, 2008]. Clinical genetic testing in individual patients will yield a specific cause in perhaps 15% [Herman et al., 2007b; Schaefer, Mendelsohn, & Professional Practice and Guidelines, 2008], most frequently copy number variants such as duplication of 15q11-q13 and deletion of 16p11.2 [Cook & Scherer, 2008], fragile X syndrome [Hatton et al., 2006], Rett syndrome [Amir et al., 1999], and tuberous sclerosis [Baker, Piven, & Sato, 1998]. More recently, individuals with ASD and macrocephaly have been found to have mutations in the PTEN gene [Butler et al., 2005; Buxbaum et al., 2007; Goffin, Hoefsloot, Bosgoed, Swillen, & Fryns, 2001; Herman et al., 2007a; Parisi et al., 2001; Varga, Pastore, Prior, Herman, & McBride, 2009; Zori, Marsh, Graham, Marliss, & Eng, 1998].

The phosphatase and tensin homologue deleted on chromosome ten (PTEN–MIM 601728) is a tumor suppressor gene that functions as a dual-specificity phosphatase active in numerous pathways involved with cellular growth [Maehama & Dixon, 1999; Maehama, Taylor, & Dixon, 2001]. Its primary function is as a lipid phosphatase that downregulates the phosphoinositol 3-kinase/AKT pathway. Heterozygous germline mutations in PTEN have been identified in Cowden syndrome (CS–MIM 158350), Bannayan–Riley–Ruvalcaba syndrome (BRRS–MIM 153480), and Proteus and Proteus-like syndromes, now collectively referred to as PTEN hamartomas syndrome [Orloff & Eng, 2008]. Individuals with CS typically develop multiple benign hamartomas and have an increased risk of certain cancers, particularly of the breast, uterus, and thyroid. BRRS is characterized by macrocephaly, hamartomas (including lipomas, hemangiomas, or intestinal polyps), penile freckling in males, and developmental delays (DD). Approximately 80% of individuals who meet diagnostic criteria for CS and 60% who meet clinical criteria for BRRS have detectable PTEN mutations [Zhou et al., 2003].

Autism is not a criterion for either CS or BRRS, but there have been several reports of individuals with ASD occurring in families with CS or BRRS [Goffin et al., 2001; Parisi et al., 2001; Zori et al., 1998]. In addition, macrocephaly is an overlapping feature found in CS, BRRS, and over 20% of individuals with an ASD [Lainhart et al., 2006]. This suggests PTEN could be involved in ASD.

The first report directly linking PTEN and autism was published by Butler et al. [2005], who investigated a small cohort of individuals with ASD and macrocephaly for mutations in PTEN, identifying a change in 3/18 individuals. Our group subsequently performed a similar study, identifying PTEN mutations in 2/16 individuals with macrocephaly and ASD [Herman et al., 2007a]. Two cohort studies extended these results. Using individuals with ASD and macrocephaly drawn from the Paris Autism Research International Sibpair (PARIS) study, the Autism Genetic Research Exchange (AGRE), and separately recruited patients, Buxbaum et al. found PTEN mutations in 1/88 subjects tested [Buxbaum et al., 2007]. Our group studied 114 subjects with isolated macrocephaly, ASD or mental retardation/developmental delays (MR/DD), and found PTEN mutations in 5/60 (8.3%) ASD subjects (including the 2 previously identified), 6/49 (12.2%) with MR/DD and 0/5 with isolated microcephaly [Varga et al., 2009]. All individuals with a PTEN mutation had macrocephaly. These combined data suggest mutations in PTEN among individuals with ASD and macrocephaly may be relatively common, but there is a discrepancy in the frequency between the cohort studies.

To address this discrepancy, we sought to replicate our initial findings. We repeated the study at our institution with the same study designed used previously. Briefly, we retrospectively reviewed all PTEN clinical sequencing tests that had been performed through our institution from January 1, 2008 to June 30, 2009. Medical records were assessed for diagnosis, and for those individuals found to have a mutation, the medical history and family history (where available) were abstracted from the chart. The diagnosis of ASD was made by DSM-IV criteria, with confirmation performed by ADOS in approximately 20% of individuals. Institutional Review Board approval was obtained for the study.

There were 93 subjects tested from our institution during that period. A total of 39 subjects had a diagnosis of ASD, 51 had MR/DD, and 3 had no diagnostic information. Missense PTEN mutations were found in 2/39 (5.1%) ASD patients and 2/51 (3.9%) MR/DD patients (Table I). All subjects with a mutation had macrocephaly, defined as an occipitofrontal circumference over 2.0 standard deviations above the mean, standardized by age and sex (Z score>2.0).

Table I. Subject Characteristics in Those with a PTEN Mutation

Family	Diagnosis	Age	Sex	OFC (Z score)	Base change	Amino acid change	Protein domain	Inheritance
The catalytic core of PTEN is amino acids 124–130. ASD, autism spectrum disorder; DD, developmental delay; MR, mental retardation; N.D., not determined; OFC, occipito-frontal circumference; Z score, age and sex standardized score in standard deveiations from the mean.
1	DD	8 mo	F	+5.38	c.821 G>T	p.W274L	C2	Maternal
2	ASD	4.5 yo	M	+3.29	c.518 G>A	p.R173H	Phosphatase	Paternal
3	MR	9.3 yo	F	+6.06	c.401 T>C	p.M134T	Phosphatase	Maternal
4	ASD	2.5 yo	F	+3.63	c.369 C>G	p.H123Q	Phosphatase	N.D.

Detailed medical history was available on all four subjects with a mutation, while pedigree information was complete in only three families (Fig. 1). Individual 1 was an 8-month-old girl with DD and macrocephaly. Exam showed only macrocephaly. Multiple family members have macrocephaly and DD, including a maternal uncle with freckling of the penis consistent with a diagnosis of BRRS. The affected mother, who shares the mutation with her daughter, is the only other tested individual thus far. Individual 2 was a 41/2-year-old boy with autism (ADOS-II score of 22; Leiter-R IQ score of 97). Macrocephaly was the only abnormal feature on exam. Chromosome analysis, fragile X testing, and array CGH testing were normal. Several family members have macrocephaly or cancer, but thus far only the father, who has the same mutation as his son, has been tested in this family. Individual 3 was a 9-year-old girl who presented to clinic with mild MR, obesity, conductive hearing loss, affective disorder, and behavior problems (oppositional and anger issues) requiring treatment with respiridone and clonidine. Findings included macrocephaly and obesity only. Chromosome analysis, fragile X testing, and array CGH testing were normal. Multiple family members were affected with MR/DD, macrocephaly and/or cancer, with all tested affected individuals possessing the same PTEN mutation. Individual 4 was seen at 30 months of age by a neurologist for DD (walked at 17 months, first words at 18 months) and autism by DSM-IV criteria. Examination revealed only macrocephaly. Investigation included cranial MRI, chromosome analysis, fragile X testing, and array CGH testing, all normal. Pedigree information was incomplete on this family.

Figure 1. Pedigrees of three families with PTEN mutations. Families are named by the number key from Table I, followed by their specific mutation. The proband is indicated by an arrow and the letter P to the lower left of the symbol. Individuals affected with autism, learning disability (LD), developmental delay (DD), mental retardation (MR), macrocephaly (Macroceph), or cancer (Ca) are indicated with filled symbols. Status of PTEN testing is shown to the upper left of each symbol as positive for familial mutation (POS), negative for familial mutation (NEG), or not tested/unknown (UNK).

None of the mutations found here has been found in 500 control alleles screened in our lab, nor have they been seen before in CS or BRRS. The R173H was present in a subject from our previous study who presented at 9 months with developmental delay, macrocephaly (Z score+4.4) and a family history of large head size and MR [Varga et al., 2009]. All four mutated positions are highly conserved, and the observed changes are predicted by POLYPHEN [Sunyaev et al., 2001] to be probably damaging and by SIFT [Ng & Henikoff, 2003] to be an intolerant change. Different substitutions with functional effects have been reported for positions 123 and 134, but the precise substitutions observed here have not been reported previously. The R173H change has been shown to have decreased phosphatase activity in vitro [Han et al., 2000]. At least two affected individuals with a PTEN mutation are noted in 3/4 families, and the R173H mutation occurs in two unrelated families, providing additional evidence the identified mutations are pathogenic.

Combining the results from this study and our first cohort, we have found PTEN mutations in 7/99 (7.1%) of individuals with ASD and 8/100 (8.0%) of individuals with MR/DD, all of whom had macrocephaly. Comparison of the study cohorts did not demonstrate a significant difference in the rate of PTEN mutations (2×2 contingency table, Fisher's exact test, ASD P=0.70; MR/DD P=0.16; Combined P=0.11). Only one subject met criteria for BRRS, and none had CS. Two relatives of the subjects could meet criteria for BRRS, but none of the family members carried a previous diagnosis of CS. A history of cancer (breast, uterine, renal, and prostate) could be found in 4/15 families, usually at a younger age. Among the seven parents who had a PTEN mutation, two were clinically normal, two had only macrocephaly, two had macrocephaly and MR, and one had macrocephaly, MR, and renal cell carcinoma (Table II).

Table II. Combined Cohort Data Showing Source of Subject Mutation, with Parental and Familial Phenotype Information, by Subject Diagnosis

Phenotype		Inheritance
Subject	Parent phenotype	Paternal	Maternal	De Novo
Combined data from current study and [Varga et al., 2009]. The parent from whom the PTEN mutation was inherited is listed, along with presence of specific features or diagnosis in the parent with the mutation. The source of the mutation has not been determined for 1 individual with ASD and 3 individuals with MR/DD. A family history of cancer was present for 2 families of individuals with ASD (both paternal inheritance) and 3 families of individuals with MR/DD (all maternal inheritance). ASD, autism spectrum disorder; BRRS, Banyan–Riley–Ruvalcaba syndrome; Ca, Cancer; DD, Developmental delay; FH, Family history; Macro, Macrocephaly; MR, mental retardation.
ASD
	Normal	2	0	0
	Affected	1 (Macro)	0	0
	Total	3	0	3
MR/DD
	Normal	0	0	0
	Affected	1 (BRRS)	3 (Macro-3; DD-2; Ca-1)	0
	Total	1	3	1

These findings carry broad implications. Upon identifying a child with ASD or MR/DD with a PTEN mutation, both parents (even if normal) should be considered for testing for that mutation as they may be at higher risk for developing specific cancers early in life and they may be at risk for having additional children with a similar problem. Testing of siblings may be indicated if a parent is found to have the mutation as well. If both parents test negative for the mutation, there is the possibility of germline mosaicism, and other sibs should be considered for testing if clinically warranted. The cancer risk is not known for individuals with a PTEN mutation who do not meet the criteria for CS. We are currently recommending cancer surveillance for our mutation positive patients over the age of 18 years, using guidelines for CS published by the National Comprehensive Cancer Network [Daly, 2009].

There do not appear to be any genotype–phenotype correlations of the observed PTEN mutations in these two cohorts. Several mutations have been seen previously in families with CS or BRRS, while others are private mutations. The R173H mutation has been found in an individual with ASD and in an individual with DD. Most of the mutations found in our ASD and MR/DD subjects involve the catalytic domain, similar to those with CS and BRRS [Orloff & Eng, 2008]. Further defining the relationship, if any exists, will require detailed analysis of existing CS cohorts for the presence of ASD and MR/DD, and conversely, cohorts of individuals with ASD will need to be studied for the presence of PTEN mutations and CS or BRRS among the subjects and their relatives. This will also help determine if ASD should be added as a criterion for CS and BRRS, possibly in an age-dependent manner.

Our study is limited by its retrospective collection of data. It is possible that similar findings over our two studies represent a systematic bias in subject collection. We believe this is unlikely for several reasons. Our institution is the only children's hospital for our area, and is the major center for autism treatment and investigation. We have also developed a standard protocol to investigate the genetic causes of autism in all individuals seen in the autism center and by developmental pediatricians, neurologists, and geneticists. This previously published protocol [Herman et al., 2007b] specifies PTEN gene testing for all individuals with ASD and macrocephaly. A prospective study of ASD subjects will be required to confirm our results.

In summary, we have replicated our previous results of PTEN mutations among individuals with ASD or MR/DD and macrocephaly. The yield of PTEN mutations in this group is estimated to be between 5–10%. We recommend this test be performed in individuals with ASD or MR/DD and macrocephaly. If a mutation is found, parents, even if clinically normal, should be considered for testing. At this time prudence suggests all individuals over 18 years should be offered cancer surveillance as currently recommended for CS. Future studies will need to address the presence of ASD in CS families, and prospective screening studies of individuals with ASD and macrocephaly should be performed to obtain a better estimate of the rate of PTEN mutations in this group.