Postmortem Diagnosis of the Proteus Syndrome by Next Generation Sequencing of Affected Brain Tissue

Abstract

We report a case of a somatic overgrowth syndrome diagnosed at forensic autopsy with the aid of next generation sequencing as Proteus syndrome. Somatic overgrowth syndromes result from spontaneous somatic mutations that arise early in development and display a mosaic pattern of expression in patient tissues. Due to the temporal and anatomic heterogeneity of these syndromes, phenotypes vary widely, resulting in clinical overlap. Furthermore, the variable ratio of mutated to nonmutated cells in patient tissue can result in low-level mutations that could be missed using Sanger sequencing. Due to these factors, recent literature points to next generation sequencing (NGS) as an adjunct to diagnosis of these rare entities. A male in his fourth decade of life presented to our forensic autopsy service with physical features suggestive of a somatic overgrowth syndrome. Due to the paucity of clinical information accompanying the individual, a definitive diagnosis based on physical characteristics, alone, was not possible. Next generation sequencing of affected formalin-fixed and paraffin-embedded brain tissue confirmed the presence of the variant in AKT1 (c.49G>A, p.Glu17Lys, in 14.13% of reads) found in Proteus syndrome. To our knowledge, this is the first report of the mosaic variant of AKT1 detected in brain tissue and the first reported case of a postmortem diagnosis of Proteus syndrome with the aid of NGS. We conclude that NGS can be used as an adjunctive method to support a specific diagnosis among the somatic overgrowth syndromes postmortem in the absence of sufficient clinical history.

Introduction

Somatic overgrowth syndromes (SOSs) result from spontaneous somatic mutations that arise early in development. This results in a mosaic pattern of expression that could affect any organ system. The progressive overgrowth of affected tissues may not become apparent until months or years after birth. Due to the temporal and anatomic heterogeneity of these syndromes, phenotypes vary widely, resulting in clinical overlap. Furthermore, the variable ratio of mutated to nonmutated cells in patient tissue can result in low-level mutations that could be missed using Sanger sequencing. Due to these factors, recent literature points to next generation sequencing (NGS) as an adjunct to diagnosis of these rare entities (1 -3). We describe herein a case of an adult male with features of SOS first recognized at the time of postmortem neuropathological examination. Whole exome NGS of affected brain tissue confirmed the specific etiology to be Proteus syndrome (PS).

Case Report

The decedent was a male in his fourth decade of life with a history of developmental and cognitive delay, seizures, scoliosis, asthma, spinal fusion, and right eye blindness. He had collapsed suddenly and experienced a seizure just prior to death. A full autopsy was performed and the brain was retained for later neuropathologic examination. The postmortem examination revealed a focally hyperostotic skull with right-sided bony protuberances that projected from the outer and inner surfaces resulting in an overall thickened appearance of the bone ( Image 1A ). The neuropathologic examination showed depressions in the brain surface subjacent the overlying hyperostosis. The brain was heavy (1620 g) and exhibited right hemimegalencephaly with an abnormal gyral pattern with both widened gyri and small undulations resembling polymicrogyria ( Image 1B ). The cut surface of the enlarged cerebral hemisphere showed an irregularly thickened cortical ribbon with intermittent blurring of the gray–white junction, as well as gray matter heterotopia. Abnormalities of the right brain also included a 0.1 cm vascular malformation within the temporal lobe and cystic degeneration of the caudate nucleus ( Image 1B and D ). The vermis was atrophic. Microscopic examination of the grossly affected hemisphere showed indistinct cortical layering beyond the recognizable molecular layer. In addition to the lack of normal cortical architecture, neurons were haphazardly arranged and dysmorphic. Dysmorphic features included enlarged cell size and coarse, irregularly dispersed deposits of Nissl substance. No balloon cells were identified ( Image 1C ). These findings prompted a review of photos taken at the time of initial autopsy. Postmortem external examination revealed dysmorphic facies (long face, downslanting palpebral fissures, low nasal bridge, open mouth at rest), bony protuberances of the right face and head, and asymmetrical upper extremities (right larger than left). Gross examination of the internal organs was only notable for increased pulmonary edema and slightly granular kidneys with dilated pelvis. Microscopic examination of the lungs, kidneys, liver, and heart yielded nonspecific findings, including patchy increased pulmonary edema, patchy chronic inflammatory infiltrates around the airways, and rare globally sclerotic glomeruli with mild thickening of the renal arterioles. Vitreous electrolytes were within expected ranges, and comprehensive toxicology testing on peripheral blood and urine was only positive for naloxone in the blood. Taken together, the phenotype was a good fit for PS; however, due to the paucity of clinical information available at forensic autopsy, the decedent did not meet all of the clinical criteria required to definitively render that diagnosis. For these reasons combined with the known overlap in clinical presentation between SOSs, we implemented whole exome NGS to further categorize the overgrowth syndrome.

Image 1.

A, Photograph of hyperostotic portion of skull overlying the hemimegalencephaly. At the right-hand side of the image is a thinner, less affected portion of skull for comparison. B, Gross photograph of the brain demonstrating the abnormal gyral pattern and overall enlargement of the right hemisphere. C, Photomicrograph of a hematoxylin and eosin-stained section of cortex from the enlarged right cerebral hemisphere. There is architectural disarray and individual enlarged neurons with abnormal distribution of Nissl substance. (×200 original magnification). D, Gross photograph of a representative coronal slice of the brain at the level of the basal ganglia, demonstrating enlargement of the right hemisphere, heterotopic gray matter, and cystic degeneration of the caudate nucleus.

Whole Exome NGS and Analysis

DNA extraction was performed on formalin-fixed paraffin-embedded tissue from the hemimegalencephalic portion of the brain containing both gray and white matter. Briefly, the overall flow was as follows: DNA (200 ng) was sheared to an average size of 350 bp using a Covaris, LLC S220 sonicator. Sheared DNA was exome enriched followed by NGS. Enriched libraries were prepared as described in the Agilent SureSelect XT Human All Exon V6 Target Enrichment System for Illumina Multiplexed Sequencing manual. Next generation sequencing was performed following the manufacturer’s protocol (Illumina). Paired-end (2X125 cycles) sequencing was performed on the Illumina HiSeq2500 utilizing version 4 reagents and software; 15 GB of data were obtained to ensure an overall average of 100× coverage. Upon completion of the sequencing run, basecalls were converted to Fastq files using the Illumina Bcl2 software. Secondary analysis was performed according to the genome analysis tool kit best practices (4), annotation in Golden Helix’s VarSeq v2 (Golden Helix, Inc). Variant pathogenicity was following published guidelines (5). Annotations were provided from many sources including RefSeq 105 Interim release (6), dbSNP 155 (7), gnomAD Exomes 2.1.1 (8), and ClinVar (9). Variant evaluation was limited to AKT, AKT2, AKT3, CDKN1C, GNAQ, GNAS, MTOR, PIK3CA, PIK3R2, PTEN, TSC1, and TSC2.

Results

Next generation sequencing of the whole exome followed by targeted analysis of 12 genes associated with overgrowth syndromes revealed a previously described variant, AKT1 (c.49G>A, p.Glu17Lys), in 14.13% of reads confirming the diagnosis of PS.

Discussion

Proteus syndrome is exquisitely rare, occurring at an estimated incidence of less than 1 per 1 million individuals (10). Proteus syndrome likely first surfaced in the medical literature with the description of Joseph Merrick in 1885. Interestingly, PS was only first described as a discrete entity nearly a century later (11,12). Proteus syndrome is progressive in course and highly variable in presentation. Overgrowth can affect cells from any tissue type and in any combination with the exception of the global involvement of any one organ system or individual patient. It is postulated that expression of the causative mutation in every cell would be incompatible with life. Since the initial descriptions of PS, clinicians have relied on diagnostic criteria based solely on clinical history and patient phenotype in order to render the diagnosis. In 2011, Lindhurst and colleagues first identified a mosaic activating variant of AKT1, believed to cause PS, that was present in 90% of individuals studied (13). This finding confirmed a long-standing hypothesis proposed decades earlier by Happle regarding the mosaic nature of SOSs (14). Since its initial discovery, a number of reports have reported detection of the same and alternative AKT1 variants in association with the clinical picture of PS (15 -21). Moreover, AKT has been targeted for the purpose of treating PS in humans (22,23). Of note, variants can arise in neoplasms of individuals without a syndrome (15), so clinical correlation is required in determining whether an individual has a syndrome or a nonsyndromic neoplasm. Due to the complex nature of diagnosing PS, revised clinical criteria have been proposed to incorporate the presence of a mosaic AKT1 variant (24).

Of note, hemimegalencephaly is a brain malformation that is seen in both syndromic and nonsyndromic clinical presentations. To date, several genes have been implicated in this phenotype, including AKT1, AKT3, PTEN, PIK3CA, and MTOR (25). Lack of normal cortical lamination and the presence of dysmorphic neurons have previously been documented in hemimegalencephaly.

The decedent had no known prior diagnosis of SOS. The physical features suggestive of SOS primarily involved the right side of the body: enlarged right upper extremity, dysmorphic facies, hyperostotic right skull, and right hemimegalencephaly. Although the physical features, alone, were sufficient to diagnose an overgrowth syndrome, it was the detection of the AKT1 variant in affected brain tissue that confirmed the etiology. The low variant allele frequency, 14.13%, reflects the mosaic nature of the variant.

Given the nature of a single case report, limitations are eminent. The targeted approach to gene analysis identified a pathogenic mutation; however, additional gene variants were not fully excluded. The NGS results were not confirmed using a second, alternative method of genetic testing. However, the American College of Medical Genetics guidance suggests that such confirmation studies are not necessitated in certain instances when the quality of data is adequate (26). Additionally, due to the mosaic nature of the overgrowth syndromes, interrogation of other affected and grossly unaffected tissues for the purpose of comparing the variant allele frequencies would have been useful. Unfortunately, these avenues of testing were not possible due to limited resources.

Conclusions

To our knowledge, this case represents the first detection of the AKT1 variant linked to PS in formalin-fixed paraffin-embedded brain tissue. In addition, this is the first report of a primary diagnosis of PS postmortem with the aid of NGS. We conclude that NGS may be used as an adjunctive method to support a specific diagnosis among the SOSs postmortem in the absence of sufficient clinical history. Furthermore, targeted sequencing for AKT1 or a panel of overgrowth-related genes may be more cost effective than whole exome sequencing for confirmation of an already suspected diagnosis.

AUTHORS

Tiffany G. Baker, MD, PhD, Department of Pathology and Laboratory Medicine, Medical University of South Carolina

Roles: A, C, D, E, 1

William B. Glen, PhD, Department of Pathology and Laboratory Medicine, Medical University of South Carolina

Roles: B, C, D, E, 6

Robert C. Wilson, PhD, Department of Pathology and Laboratory Medicine, Medical University of South Carolina

Roles: B, C, D, E, 6

Nicholas I. Batalis, MD, Department of Pathology and Laboratory Medicine, Medical University of South Carolina

Roles: A, C, D, E, 4

Daynna J. Wolff, PhD, Department of Pathology and Laboratory Medicine, Medical University of South Carolina

Roles: A, C, D, E, 4

Cynthia T. Welsh, MD, Department of Pathology and Laboratory Medicine, Medical University of South Carolina

Roles: A, C, D, E 4