OUP user menu

Hydrolethalus Syndrome
Neuropathology of 21 Cases Confirmed by HYLS1 Gene Mutation Analysis

Anders Paetau MD, PhD, Heli Honkala MSc, Riitta Salonen MD, PhD, Jaakko Ignatius MD, PhD, Marjo Kestilä PhD, Riitta Herva MD, PhD
DOI: http://dx.doi.org/10.1097/NEN.0b013e318180ec2e 750-762 First published online: 1 August 2008


Hydrolethalus syndrome is a lethal malformation syndrome with a severe brain malformation, most often hydrocephaly and absent midline structures. Other frequent findings are micrognathia, polydactyly, and defective lobation of the lungs. Hydrolethalus syndrome is inherited in an autosomal recessive manner and is caused by a missense mutation in the HYLS1 gene. Here, we report the neuropathologic features of 21 genetically confirmed cases. Typically, 2 separated cerebral hemispheres could be identified, but they lacked midline and olfactory structures and were situated basally with a massive accumulation of cerebrospinal fluid. Temporal and occipital lobes were hypoplastic, and normally developed hippocampi were not found. Primitive thalami and basal ganglia were fused in the midline. A hypothalamic hamartoma was a frequent finding, and brainstem and cerebellum were hypoplastic. Three cases were hydranencephalic, and 1 was anencephalic. A midline "keyhole" defect in the skull base was a constant finding. Histologically, the cortex was dysplastic. This pattern of brain pathology, clearly belonging to the midline patterning defects, seems to be unique for the hydrolethalus syndrome and combines features of disturbed neurulation, prosencephalization, and migration. Despite variation in the clinicopathologic phenotype, all cases in the series carried the same homozygous missense mutation in HYLS1.

Key Words
  • Arhinencephaly
  • Callosal agenesis
  • Hydrocephalus
  • Hydrolethalus syndrome
  • Hypothalamic hamartoma
  • Midline patterning defects
  • Occipitoschisis


Hydrolethalus syndrome (HLS; MIM 236680) is a lethal malformation syndrome that was described in 1981 in Finland during a nationwide study on Meckel syndrome (MIM 249000) (1). Both of these syndromes are characterized by a severe central nervous system (CNS) malformation and polydactyly as the main findings, but children with HLS do not have the cystic dysplasia of the kidneys or fibrous changes of the liver that are the most constant findings of Meckel syndrome. Clinically, HLS is characterized by a large hydrocephaly observed prenatally and by a voluminous polyhydramnios; for a mnemonic of these main findings (i.e. hydrocephaly, polyhydramnios, and lethality), the syndrome was named hydrolethalus. Hydrolethalus syndrome is always lethal during the very first hours or days after birth, and in many cases it leads to premature stillbirth. Today, most of the pregnancies are terminated because the CNS malformation can be detected by early ultrasound scan. At ultrasound, the fetus is small for dates, but the head is large in proportion to the rest of the body. Typically, the midline of the brain is defective or not visible, and an abnormal fluid space can be seen in the midline. The neck is bulging, and the movements and posture of the legs may be abnormal, with club foot and medially deviating double big toes (2, 3).

Recently, a missense mutation in the HYLS1 gene located on chromosome 11q24.2 was found to cause HLS (4). HYLS1 consists of 6 exons spanning a genomic region of 17 kb, which has several alternative transcripts. Each of the transcripts, however, has the same translated region coded only by exon 6. The disease-causing mutation is a point mutation leading to an A to G transition in exon 6. This mutation results in an amino acid substitution of aspartic acid 211 to glycine of the polypeptide of 299 amino acids in a well-conserved region with a function that is so far unidentified (4). Hydrolethalus syndrome is one of the rare autosomal recessive conditions enriched in Finland that together form the Finnish disease heritage. The incidence in Finland is at least 1:20000 births (5). The carrier frequency of this mutation in the Finnish population is 1.1% in the Western part of the country and 2.5% in Central and Eastern Finland (4). To date, 64 cases carrying a homozygous exon 6 mutation have been found in Finland. No other mutations have hitherto been found, and no mutation-confirmed cases have been reported outside of Finland.

The clinical features of 55 HLS patients from 39 families have been published previously (5) but without detailed neuropathologic description and molecular genetic confirmation of the diagnosis. In this study, we present the spectrum of neuropathologic findings in the CNS in HLS based on 21 cases verified by analysis of the HYLS1 mutation.

Materials and Methods


This study was approved by the ethical committee of Joint Authority for the Hospital District of Helsinki and Uusimaa, Finland. Parental consent was obtained for the collection and study of the autopsy samples. None of the parents were consanguineous.

The study material consisted of archival autopsy specimens of 21 HLS cases collected during the years 1981 to 2008 (Table 1). Many of the samples, particularly those from the induced abortions, were small and very fragmented and thus difficult to study in detail. In addition, the older samples in particular were available only as paraffin blocks selected at the time. Therefore, resampling of crucial areas sometimes could not be performed.

View this table:

All cases were confirmed by the mutation analysis that became possible after the identification of the HYLS1 gene (4). Fourteen cases were induced abortions, and their gestational age varied from 12 to 24 weeks. One of these was an anencephalic fetus, and 3 were hydranencephalic fetuses of which 2 were sibs. Seven cases were born after pregnancies that varied from 33 + 5 to 40 + 6 weeks of gestation. Three of these were stillborn, and 4 were born alive. These 4 cases survived from 1 minute to 6 days. The infant that survived 6 days was born at 33 + 5 weeks with mild manifestations. She was treated in the neonatal intensive care unit, and her HLS diagnosis was made only at autopsy. Two sibling pairs are included in the series (Cases 2a, 2b and 6a, 6b; Table 1).

Brain Specimens

The brain specimens were studied after fixation in neutral 10% formaldehyde. After macroscopic dissection, selected paraffin-embedded samples were sectioned at 6 to 10 μm and stained using the following methods: hematoxylin and eosin, Luxol fast blue-cresyl violet, and Holmes silver impregnation. Immunohistochemistry (IHC) was performed on selected blocks using monoclonal antibodies against the following antigens: glial fibrillary acidic protein (GFAP; M0761, 6F2; Dako, Carpinteria, CA; dilution, 1:300), neurofilament proteins (SMI-32 and SMI-311; Sternberger Monoclonals, Inc., Baltimore, MD; dilutions, 1:250 and 1:2500, respectively), nestin (MAB5326; Chemicon, Temecula, CA; dilution 1:200), microtubule-associated protein 2 (MAP-2; M-4403; Sigma, St. Louis, MO; dilution, 1:5000), calretinin (M7245, DAK-Calret 1; Dako; dilution, 1:50), epithelial membrane antigen (M0613, E29; Dako; dilution, 1:500), cell cycle-associated proliferation antigen Ki-67 (MIB-1, M7240; Dako; dilution, 1:50), and reelin (MAB5366; Chemicon; dilution, 1:2500). Pretreatment was performed with citrate buffer for GFAP and MAP-2 and with tris-EDTA for the other antibodies. The detection kit was Envision Advanced (Dako) for GFAP and MAP-2, and Envision (Dako) for the other antibodies except for reelin, where Vectastain Elite ABC-kit with 3-amino-9 ethylcarbazole (Vector Laboratories, Burlingame, CA) was used as the chromogen.

DNA Isolation

DNA was isolated from several tissues using standard methods and from paraffin-embedded tissues by E.Z.N.A. Tissue DNA Kit (Omega Bio-Tek, Lilburn, GA) according to the manufacturer's instructions. In addition, the octane DNA extraction method (6) (Arto Orpana, personal communication) was used. In this method, the tissue pieces were first obtained from the paraffin block with a needle. The paraffin was removed by adding 1 ml of octane (Fluka, Buchs, Switzerland), followed by adding 100 μl of methanol (centrifugation, 13,000 rpm for 2 minutes), discarding the supernatant, and adding 400 μl of proteinase K buffer containing 0.5% sodium dodecyl sulfate, 0.1 mol/L of NaCl, 50 mmol/L of Tris-HCl (pH 8.0), 20 mmol/L of EDTA, and 2 mg/ml proteinase K (Roche, Basel, Switzerland). The mixture was incubated at 55°C for approximately 48 hours, vortexing a couple of times during this time. DNA extraction was performed by heating the samples at 95°C for 20 minutes, followed by 10-second centrifugation. A mixture of 400 μl of phenol/chloroform/isoamyl alcohol (25:24:1) was added to the samples, subjected to vortex mixing, and centrifuged for 5 minutes at 11,000 rpm at 4°C. The upper phase was moved to clean tubes, 700 μl of −20°C absolute ethanol, and after that, 400 μl of 0.2 mol/L of sodium acetate (pH 7.5) was added. Tubes were mixed by flicking and turning until the DNA was precipitated. Tubes were incubated in −80°C for 30 minutes or in −20°C overnight. The samples were centrifuged for 30 minutes at 11,000 rpm, the supernatant was discarded, and the pellet was washed 2 times with −20°C 70% ethanol with 20-minute centrifugation at 11,000 rpm. The DNA was vacuum-dried and dissolved in 1 × Tris-EDTA buffer for 72 hours.

Mutation Analysis of the HYLS1 Gene

The region harboring the mutation in exon 6 of the HYLS1 gene was polymerase chain reaction amplified and sequenced using genomic DNA from affected individuals with primers 5′-AGAGAAGGAATGGGCTCTCC-3′ and 5′-ACCCCAGCGTAATTCCTTTC-3′ (Sigma-Aldrich, Haverhill, UK). Polymerase chain reactions were performed in 25 μl volume, containing 5 μl of 4 ng/μl DNA, 10× PCR buffer containing 15 mmol/L of MgCl2 (Applied Biosystems, Foster City, CA), 0.5 μl of 25 mmol/L of MgCl2 solution, 2.5 μl of 2 mmol/L of deoxynucleoside triphosphate, 0.5 μl of each primer at 20 μmol/L, and 0.2 μl of 5 U/μl of AmpliTaq Gold polymerase (Applied Biosystems). The reactions were subjected to an initial denaturation step of 94°C for 10 minutes, followed by 30 cycles of 94°C for 30 seconds, 64 to 60°C (64°C for the first 3 cycles, 62°C for the next 2 cycles, and 60°C for the rest of the cycles) for 30 seconds, 72°C for 30 seconds, and a final elongation step of 72°C for 7 minutes. The reaction products were analyzed on 1% agarose gel. Polymerase chain reaction products were purified with enzymatic treatment (ExoI/SAP; Applied Biosystems) and then sequenced on both strands using ABI Big Dye v.3.1 (Applied Biosystems) chemistry and ABI 3730xl sequencer (Applied Biosystems). Sequences were analyzed by GeneComposer version


General Autopsy and Genetic Findings

The central data for the individuals in the material and the major general pathologic findings are presented in Table 1. The sequencing analysis of the HYLS1 gene showed that all the 21 HLS cases studied were homozygous for the missense mutation changing aspartic acid 211 to glycine. The general appearance of the cases at autopsy was quite uniform: a large head, postaxial polydactyly in hands and feet and typically hallux duplex, and a keyhole-shaped defect in the skull base (Fig. 1; Table 1).


Phenotypic variation of hydrolethalus syndrome. (A) A severely affected stillborn fetus with hallux duplex and postaxial polydactyly in hands (Case 16). (B) A mildly affected female infant who lived for 6 days (Case 13). (C) The keyhole defect in the base of the skull of the fetus (B).

Macroscopic Neuropathologic Findings

In 17 of 21 cases, the cardinal finding was a unique type of hydrocephalus in which the lateral ventricles opened widely into an interhemispheric space filled with cerebrospinal fluid. This space was covered by a distended and torn arachnoid membrane (Figs. 2A-D). The cerebral hemispheres were found lying separated on the skull bottom and creating an "open-book" appearance with massive amount of cerebrospinal fluid above especially in the cases born in the second and third trimesters of pregnancy (Fig. 2D). The olfactory bulbs and tracts were invariably absent; we observed only 1 case with a small unilateral remnant of the olfactory tract (Case 13; Table 1). A complete interhemispheric fissure could be identified, indicating hemispheric cleavage; no cases with a true alobar or semilobar holoprosencephaly were observed in the series.


Macroscopic neuropathologic features of hydrolethalus syndrome. (A) A superior view of a fetal brain specimen (Case 2a; Table 1). The anteriorly fused cerebral hemispheres are otherwise widely separated with a prominent cleft in the thalamic region. Midline structures are missing. The rhombencephalic flexure is prominent with an abnormal mesencephalic quadrigeminal region. A small cerebellum (arrowhead) is seen with an open fourth ventricle below. (B) Case 4 fetal brain specimen showing the basal brain structures. An arrow points at the hypothalamic plate protrusion-hamartoma; there is a small cerebellum (arrowhead). This case also demonstrates a steep rhombencephalic flexure with hypoplasia of posterior fossa contents and occipitotemporal regions. (C) The base of the brain of Case 14 shows absence of olfactory bulbs and tracts, leptomeningeal heterotopia in the Sylvian fissures, and a large hypothalamic hamartoma that covers the brainstem (asterisk). (D) In Case 10, there are even more widely separated hemispheres without midline structures; only meningeal remnants cross the midline. This demonstrates the classical open-book appearance of HLS. There is a small cerebellum (arrowhead). (E) In this anterior frontal coronal section of Case 13, there are small proximal stumps of callosal fibers with a small longitudinal Probst bundle element (arrowhead). A lateral periventricular nodular heterotopia is indicated with an arrow. (F) A more posterior midthalamic coronal slice of Case 13. The cortical surface is quite smooth without normal gyration; midline commissures are missing, and fused thalami can be seen. The arrow points to a small periventricular nodular heterotopia.

The upper midline commissural structures, corpus callosum, septum pellucidum, and fornices were absent. A short and small frontal bundle of Probst could be identified in only 1 case (Case 13; Table 1; Fig. 2E). Basally, the thalami and small basal ganglia were widely fused (Fig. 2F). None of the cases had an anterior commissure. The optic nerves and chiasm were hypoplastic in most cases. The detailed ophthalmic pathology has been previously reported (7).

In several cases, there was a polypoid hypothalamic hamartoma (Table 2). These were of considerable size in some of the cases born during the last trimester (Fig. 2C). In the fetal cases, the hamartoma was difficult to observe, and in many instances, only a bulging of the region of the hypothalamic plate was identified (Fig. 2B; Table 2). Developed and identifiable mamillary bodies could not be seen in any case. The pituitary gland varied from absent to normal. The sella was sampled in only a few cases.

View this table:

The gyration of the brain was often grossly abnormal in the frontoparietal regions, and the temporal and occipital regions were severely hypoplastic. In most cases, the gyration seemed coarse, microgyric, or nodular (Table 2). Normally developed hippocampi or amygdala could not be identified. A narrow cleft that probably corresponded to a hypoplastic third ventricle was seen at the junction of thalami and rostral mesencephalon. It seemed to communicate with a slit-like aqueduct in some cases.

Features of rhombic roof dysgenesis were seen as absent tectal structures; corpora quadrigemina could not be identified. The brainstem and cerebellum were hypoplastic; the mesencephalon was the most normal-appearing region of the brainstem. Most cases had a slit-like aqueduct and narrow crura cerebri. At the level of the pons, the tegmentum was only mildly atrophic, whereas the basis pontis was severely narrowed. The medulla oblongata was also small, with a slit-like fourth ventricle and severely narrowed pyramids. The inferior olives could not be identified. The cerebellum was invariably small, and it was difficult to identify nucleus dentatus in any of the cases.

Occipitoschisis (i.e. a cleft in the base of the skull in the midline of the occipital bone) was a constant finding. The cleft extended from the foramen magnum to form a keyhole-shaped opening in the base of the skull (Fig. 1C); the meninges or cerebellum sometimes bulged through the defect in the neck beneath the intact skin.

In 4 cases, the cardinal macroscopic finding was anencephaly/hydranencephaly instead of open-book hydrocephalus (Table 1; Cases 6a, 6b, 9, and 12). Two fetuses with hydranencephaly were sibs (Cases 6a and 6b). Both had basal occipitoschisis, and the latter had absence of the dorsal arch of the first and second cervical vertebrae and the pituitary. One fetus with anencephaly (Case 12) had gliovascular tissue on the base of the cranium. In addition, this fetus had palatoschisis, micrognathia, hypertelorism, broad nose, postaxial polydactyly in hands, and preaxial polydactyly in feet with hallux duplex.

Microscopic and Immunohistochemical Neuropathologic Findings

The torn meningeal membrane covering the upper open midline in the cases consisted of both epithelial membrane antigen-positive arachnoidal cells and an underlying thin gliovascular membrane (Figs. 3A, B). The leptomeningeal heterotopic layer, which was prominent at the base of the brain, contained gliovascular tissue and MAP-2-positive (probably neuronal) elements (Figs. 3C, D).


Histologic neuropathologic features of hydrolethalus syndrome of surface structures and small dysplastic hippocampus demonstrated in Case 13. (A) Epithelial membrane antigen-positive immunostaining in meningeal remnant surface from the superior midline. (B) The same area is shown with glial fibrillary acidic protein (GFAP) immunostain demonstrating a probably heterotopic glial inner layer in the thin meningeal membrane. (C) The surface of the base of the brain shows a large leptomeningeal heterotopia between the ba and brain tissue surface (asterisk), demonstrating here glial elements with GFAP immunostain. (D) The same heterotopic area also includes neuronal elements that are immunopositive for microtubule-associated protein 2 (MAP-2). (E) A medial section through the hypoplastic temporal lobe displays a miniature dentate fascia (arrow). (F) At slightly higher magnification, the small hippocampal structure is seen in this section immunostained for SMI-32 neurofilament. There is an asterisk in the terminal folium. Paraffin sections, epithelial membrane antigen immunohistochemical staining (IHC), 100× (A); GFAP IHC, 100× (B); GFAP IHC, 40× (C); MAP-2 IHC, 40× (D); hematoxylin and eosin, 20× (E), and SMI-32 IHC 40× (F). ba, basilar artery.

The hypothalamic hamartomas mostly consisted of neuronal and some glial cells, also including some smaller, probably immature, cells and neuropil (Fig. 4A). By IHC, sparse neurofilament-positive mature neuronal cells were identified (Fig. 4B), as well as GFAP-positive astroglial cells (Fig. 4C). Only low to moderate proliferative activity was demonstrated by MIB-1 immunostaining (Fig. 4D).


Histologic features of the hypothalamic hamartoma in hydrolethalus syndrome Case 14. (A) With hematoxylin and eosin stain, the hamartoma consists of slightly nodular areas of small, round, probably partly immature, neuronal, and possibly some oligodendroglial cells, capillaries, astroglial cells, and occasional larger pyramidal neuronal cells. (B) A large pyramidal neuron is shown with neurofilament SMI-311 immunostaining; many small round cells also show some staining of the scant perinuclear cytoplasm. (C) A nodule consisting of small neuronal cells is surrounded (on the right) by glial fibrillary acidic protein (GFAP)-positive astroglial cells. (D) Only few nuclei are immunolabeled with the Ki-67/MIB-1 cell proliferation marker. Mitotic figures are not observed. Paraffin sections: hematoxylin and eosin, 100× (A); neurofilament SMI-311 immunohistochemistry (IHC), 400× (B); GFAP IHC, 100× (C); and MIB-1, 400× (D).

Histologically, the cerebral cortex was immature, with features of focal, mainly unlayered, polymicrogyria and also lissencephaly type 2-like areas, disorganized neuroblastic rosettes (Figs. 5A-F), and leptomeningeal heterotopia. Most cases had areas of both polymicrogyria and lissencephaly type 2-like areas and disorganized areas that were difficult to classify. Ectopic neuroblastic rosettes were frequently observed under the cortical region especially in the germinal matrix-rich areas in the fused basal midline. These rosettes displayed a high proliferation index by MIB-1 IHC. The proliferative activity was accentuated near the lumina of the rosettes (Fig. 5F). Calretinin-positive neuronal clusters were prominent in superficial cortical areas; these cells also largely expressed reelin by IHC (Figs. 5G, H). Both cortically and in the germinal matrix-rich areas at fused thalami, occasional fragmented, probably apoptotic, nuclei could be seen. This was, however, a minor feature, and we did not perform any terminal deoxynucleotidyl transferase-mediated dUTP-biotin end labeling of fragmented DNA assays in these cases.


Histologic neuropathologic features of dysplastic cortex in hydrolethalus syndrome (HLS). (A) Microgyric nodular architecture with abundant gliovascular strands (arrows) can be seen in the lateral parietal cortex of HLS Case 2b. (B) Microtubule-associated protein 2 (MAP-2) immunostaining demonstrates haphazard irregular geographic areas that somewhat resemble lissencephaly type 2 in the cortex of HLS Case 13. On the top, there is a heterotopic glioneuronal wart-like eruption (asterisk). (C) The same sample as in (A) with glial fibrillary acidic protein (GFAP) immunostaining shows irregular bundles of probable radial glia (arrow). (D) The same area as in (C) with nestin immunostaining shows the same appearance of irregular and disrupted bundles of radial glial structures (arrow). (E) A disorganized cortex with many primitive neuroepithelial rosettes under the outer zone of the cortical plate is seen in the frontolateral cortex of Case 2a. (F) The same rosette-rich region seen in MIB-1 staining shows very high proliferative activity in the inner luminal zones of the rosettes. (G) There are many calretinin (CR)-positive neurons (arrow) in Layer 1 of Case 13 with CR immunostaining. (H) Several of the CR-positive superficial neuronal cells are also immunopositive for reelin. Paraffin sections: hematoxylin and eosin 40× (A), MAP-2 immunohistochemistry (IHC), 20× (B); GFAP IHC, 40× (C); nestin IHC, 40× (D); hematoxylin and eosin, 100× (E); MIB-1 IHC, 100× (F); CR IHC, 200× (G); and reelin IHC, 200× (H).

The fused thalami were somewhat sparsely cellular with respect to mature neuronal components, but the germinal matrix was prominent in the midline. The germinal matrix zone higher in the separated hemispheres was scant. The lateral borders of the fused thalami continued without a sharp border in immature areas of glioneuronal tissue. In some cases, there were areas that vaguely resembled caudoputaminal areas. The inner walls of the separated hemispheres were covered by a much flattened and focally patchy ependymal layer. This ependymal layer was mostly detectable in the base and lower parts of the hemispheres; in the superior parts of the widely separated hemispheres, it was often totally flattened or missing and without any subependymal matrix layer. Fragments of normal-looking choroids plexus were seen in many cases at the basal corners of the open interhemispheric space and also often in the region of the fourth ventricle.

A normal visual cortex could not be identified in the hypoplastic occipitotemporal areas. Only in Case 13 (Table 1), one of the milder and near-term cases, did we observe a miniature dentate fascia in the medial temporal lobe (Figs. 3E, F). A developed identifiable amygdala was not identified.

In the brainstem, substantia nigra neurons were seen in most cases as well as an aqueduct that was often slit-like. The basis pontis was also histologically severely atrophic, including in the corticospinal tracts (Fig. 6A). Narrowing of the tegmental area was much less marked. In the medulla oblongata, there was a deep slit-like fourth ventricle and mostly identifiable hypoglossal nuclei. Normal inferior olives could not be found, but there were small dysplastic areas of probable olivary neurons lateral to extremely flattened and small pyramids (Fig. 6B).


Brainstem, cerebellum, and spinal cord in hydrolethalus syndrome (HLS). (A) A markedly narrowed basis pontis with minute corticospinal tract (asterisk) and a larger tg above are seen in Case 19. (B) A minute dysplastic inferior olive can be seen as only some neuronal groups (asterisk), whereas the totally narrowed pyramid is seen medially below this asterisk in the right lower corner of the medulla oblongata of Case 13. (C) All layers are narrowed in the atrophic cerebellar cortex of HLS Case 13. (D) A patchy distribution of small Purkinje cells (arrow) can be seen in the case in (C) with SMI-32 neurofilament staining. (E) A dysplastic dentate nucleus with irregular neuronal strands (arrow) and intermingled matrix-like small cells is seen in the cerebellar hilus from Case 19. (F) The cervical spinal cord from Case 17 shows a polypoid bulge (asterisk) from the posterior part of the cord and a slit-like central canal (arrow). (G) The lumbar spinal cord from Case 9 shows a duplicated central canal (arrow); the posterior columns behind the arrow seem disorganized. (H) The ventral horn of Case 9 shows normal-appearing motor neurons with an SMI-311 neurofilament immunostain (arrows). Paraffin sections: Luxol fast blue-cresyl violet (LFB), 20× (A); hematoxylin and eosin (H&E), 40× (B); H&E, 20× (C); SMI-32 neurofilament immunohistochemistry (IHC), 100× (D); LFB, 200× (E); H&E, 20× (F); H&E, 20× (G); and SMI-311 neurofilament IHC, 100× (H). tg indicates tegmentum.

There was narrowing of all layers in the cerebellar cortex; the molecular layer was diminished; there were unevenly distributed, small Purkinje cells, and there was a sparse internal granular layer (Figs. 6C, D). When it was identified (which was only in the older near-term cases), the dentate nucleus was dysplastic (Fig. 6E). This was probably due to the small cerebellar area in the fetal cases, also making parasagittal sampling to include the dentate nucleus difficult.

Histology of the spinal cord was studied in 4 cases (Cases 2b, 9, 15, and 17). In the cervical portions, the central canal was slit-like, and the posterior columns were hypoplastic. One spinal cord was bifid in the ventral part, 1 had a dorsal hamartomatous bulge in the cervical portion (Fig. 6F), and 2 spinal cords had double central canal in the lumbar region (Fig. 6G). Motor columns seemed to be normally populated (Fig. 6H).


Neuropathology of HLS-Developmental Interpretation

This is the first study to describe the neuropathologic findings of HLS in a clinically defined series of 21 cases that have been verified by the HYLS1 gene mutation analysis. The typical cases display an open-book type of hydrocephalus with widely separated hemispheres devoid of midline structures. Olfactory aplasia, fused thalami, hypothalamic hamartoma, hypoplastic posterior fossa, and a distinct keyhole occipital skull base defect are additional typical features. The cerebral cortex is immature and dysplastic. There seems to be a unique pattern of brain pathology which, when combined with the other findings of this syndrome, enables a more specific differential diagnosis to morphologically resembling syndromes.

The severity of the CNS abnormality varies in HLS. In one end of the spectrum, the patients may exhibit a substantial mass of malformed brain, or they may present with hydranencephaly or anencephaly. The pituitary may show findings from absence to normal, and the hypothalamic hamartoma may vary from none to massive and bulging. It should be pointed out, however, that the hamartoma is difficult to identify in early fetuses.

Absent midline structures of the brain, most of the craniofacial anomalies, the congenital heart defect, abnormalities of the respiratory tract, bifid uterus, and other genital anomalies can be regarded as defects of the midline developmental field (8), and the neuropathologic findings indicate that HLS belongs to the midline patterning defects (9). The pattern of arhinencephaly, agenesis of corpus callosum, and widely separated cerebral hemispheres could be interpreted as a special form of lobar holoprosencephaly (9). In our opinion, however, because there was a completely developed interhemispheric fissure in all cases of HLS and at the rostral and superior ends of the hemispheres, HLS is separate from cases of true holoprosencephaly (10).

It is difficult to evaluate the pathogenesis of the open midline in HLS. The development of the dorsal neural tube and also the dorsal forebrain patterning seem to involve bone morphogenetic proteins and Wnt proteins, and the sonic hedgehog (SHH) signaling pathway which participates in ventral patterning of the midbrain and the forebrain (9). It remains to be determined how HYLS1 may be involved in these pathways. Because we do not consider HLS to be a real holoprosencephaly, it leaves us, in addition to other defects, with agenesis of the corpus callosum. The pathogenesis of callosal agenesis also remains uncertain in many known syndromes (9). Hydrolethalus syndrome also includes features of rhombic roof dysgenesis, considering the mesencephalic region, where no regular corpora quadrigemina could be identified.

The cortical dysplasia in HLS includes a totally immature cortical plate containing numerous neuroepithelial rosettes in early cases and mainly unlayered polymicrogyria and also some areas reminiscent of lissencephaly type 2 (11, 12) in older cases. The numerous ectopic cortical and subcortical neuroepithelial rosettes, with accentuated proliferative activity in the lumina, represent an interesting feature. This could mean that an aberrant ongoing signal for the development of the dorsal neural tube is present in HLS especially in disorganized and immature regions. In our opinion, these rosettes could represent aberrant miniature "neural tubes." We previously observed a similar feature but in a milder form in study of Meckel syndrome (13). The clustering of calretinin-positive cells in Layer 1 seen also in near-term HLS cases is also notable. Because most of these cells also expressed reelin, they can be considered as Cajal-Retzius cells. The persistence of Cajal-Retzius cells in cases of polymicrogyria have been reported earlier (14), and it is probably a common phenomenon in migration disorders.

Differential Diagnosis

Pallister-Hall syndrome (PHS; MIM 146510) caused by nonsense and splicing mutations of GLI3 gene at chromosome 7p13 (15-17) has the most resemblance to HLS. Very recently, somatic mutations in GLI3 were also reported in hypothalamic hamartomas (18, 19). The hypothalamic hamartoblastoma is characteristic to PHS and may also be seen in HLS (16, 20-22). In the HLS cases, we consider the hypothalamic hamartomas to represent the similar phenomenon to that in PHS, but at the time when the name hamartoblastoma was chosen, IHC and proliferation studies were not used. In HLS, the hamartomas did not invade deeper brain base structures, but they could be seen as polypoid protrusions and tumor-like masses later in pregnancy. By IHC, they contained many cells that showed both neuronal and glial differentiation, some oligodendroglia-like cells, and some less differentiated, more immature elements. The proliferative activity assessed by MIB-1 IHC was only low to moderate, and mitotic figures were not seen. In a recent review of hypothalamic hamartomas (23) and in a textbook (10), PHS is also included in the discussion of this type of lesion. Moreover, the most recent World Health Organization classification of CNS does not include the term hamartoblastoma (24). We therefore designate the lesions in HLS as "hypothalamic hamartomas." In addition to hamartoma/hamartoblastoma, other shared features between HLS and PHS are polydactyly, micrognathia, occasional cleft/lip palate, abnormal lobation of lungs, and heart defects. Pallister-Hall syndrome patients, however, do not seem to exhibit the open-book brain anomaly typical of HLS, and PHS patients also typically show renal abnormalities and imperforate anus that are not seen in HLS.

The acrocallosal syndrome also shares some features with HLS (25-28) and has been reported to display a GLI3 mutation (29). The so-called holoprosencephaly-diencephalic hamartoblastoma association is apparently heterogeneous, and some patients diagnosed with this condition may share a common pathophysiologic pathway with the pathways in HLS (30).

In addition, several other syndromes share CNS midline malformative features, including the oral-facial-digital syndrome (OFD) Type IV (Mohr-Majewski or Baraitser-Burn syndrome; MIM 258860) (31); 1 case of this syndrome has been associated with occipitoschisis (32). In our series, however, basal occipitoschisis is a constant and almost pathognomonic finding in HLS and should be specifically identified. It is easily seen on x-ray, which is an informative and practical examination for cases of suspected HLS (33). The OFD Type VI (Varadi-Papp syndrome; MIM 277170) also shares features with HLS. A common finding is bifid hallux, and OFD Type VI patients also show brain malformations. On neuropathologic examination, however, they demonstrate cerebellar midline gap and absent vermis but, unlike in HLS, do not exhibit gross cerebral abnormalities (34). Muenke et al (35) have described a case with overlapping features of PHS, HLS, and OFD Type VI, demonstrating the difficulty in diagnosing these phenotypes. The invariable lethality in HLS, however, is one important feature that differentiates it from many cases of the previously discussed entities. A concept of multiplex phenotype, the cerebro-acro-visceral early lethality multiplex syndrome, has been suggested for cases sharing several of the features in common in these syndromes (36).

Impact of the HYLS1 Gene Mutation

HYLS1 is a novel gene that encodes a protein the function of which is currently unknown. Because 1 amino acid change causes a severe lethal malformation syndrome, however, HYLS1 must have a critical role in fetal development. Our previous in silico analyses of the protein show that the site in which the mutated amino acid is located is highly conserved in different orthologs from Caenorhabditis elegans to human; the isoelectric point of the protein is changed in the mutated form, and a protease cleavage site is lost in the mutated region (4). We have also studied the effect of the mutation on the protein with the PolyPhen (polymorphism phenotyping) program that predicts the D211G change to be probably damaging for the protein. All of these studies support the significance of the mutation site for the structure and function of the protein.

The expression profile of the Hyls1 gene in the mouse was demonstrated by in situ studies (4) and correlates with the neuropathologic findings of HLS. The mouse embryos showed strong expression in the CNS, including the telencephalon, the midbrain, the medulla, developing cortex, choroid plexus, and the ganglionic eminence. In addition, brain sections from a 3-month-old adult mouse showed expression in the hippocampal region. Thus, there are significant similarities between the expression pattern of HYLS1 and the neuropathologic abnormalities in HLS cases, including missing midline structures, dysplastic cortical regions with features of disturbed migration, and missing or extremely hypoplastic hippocampi.

Because radial glia cells are required for normal migration of neurons during neocortical development (37), the disorganized and disrupted radial glial fibers demonstrated in HLS by GFAP and nestin IHC most likely cause the severe defect in neuronal migration during CNS development of HLS patients. Interestingly, some genes that participate in cell cycle regulation and cell migration (e.g. cyclin D1) can also be a part of cancer-causing actions when the expression level of the gene is abnormal (38, 39). This might at least partly explain the hamartoma, and the findings of neuroepithelial rosettes in the brain (i.e. the neuroepithelial rosette structures in the HLS brains) suggest a severe disturbance in the early stage of neuronal development.

The main findings affecting the midline structures of the developing fetus might suggest a role in the SHH signaling pathway for HYLS1. Castori et al (30) also propose the SHH pathway as possible for HYLS1 because of the overlapping findings in HLS and holoprosencephaly-diencephalic hamartoblastoma. However, the question remains in which part of the pathway HYLS1 would be situated because defects in any of several steps in the SHH pathway may lead to similar clinical phenotypes presumably because of the functionally equal effects on downstream target genes (40).

As previously noted, the pathogenesis of agenesis of the corpus callosum is still uncertain (9). Thus, it is currently difficult to speculate how the agenesis of corpus callosum and, moreover, the other midline defects in HLS occur and whether it is a primary or a secondary effect of HYLS1 malfunction. The absence of the corpus callosum in HLS cases might, however, suggest a severe defect in axon guidance because the interhemispheric axonal projections in the brain are conducted across this part. This would further give HYLS1 a role as a part of the axon guidance machinery.

The present material of mutation-confirmed HLS cases shows that there is significant variation in the spectrum of the CNS malformations. This is also true for the overall phenotypic variation in HLS, although HLS in Finland is caused by the same missense mutation in all cases. It should be considered that the phenotypic variability might result from secondary effects such as unknown genetic or environmental factors during fetal development.


The pattern of neuropathology of HLS is unique, and both keyhole defect and hypothalamic hamartoma are rare key findings, the presence of which could be a reason for the HYLS1 gene mutation study. The neuropathologic pattern seen in HLS has a quite constant combination of findings that clearly connects this syndrome to the midline patterning defects. It displays features of multiple stage defects: in dorsal neural tube development, forebrain patterning, and migration. Although the precise function of the HYLS1 protein and the pathway it belongs to is currently unknown, the severe effects of the mutation suggest an important role of HYLS1 in fetal development. It is hoped that in the future, further studies of HYLS1 function and possible animal models would offer novel findings and shed light on the neuropathology of this syndrome and give us essential information regarding the signal routes and molecular actions that participate in important steps of fetal development.


The authors thank Olli Tynninen, MD, Department of Pathology, University of Helsinki, Helsinki, Finland, for valuable help with the digital images; Arto Orpana, PhD, HUSLAB Laboratory of Molecular Genetics, Helsinki, Finland, for helpful technical advice; and Ritva Timonen, Department of Molecular Medicine, National Public Health Institute, Helsinki, Finland, for excellent technical assistance.


  • This study was supported by Grant Nos. 211124 and 118468 from the Academy of Finland (MK), Helsinki Biomedical Graduate School (HH), Oulu University Hospital EVO grants (RH), and the Department of Medical Genetics Väestöliitto (RS) is funded by Finland's Slot Machine Association (RAY).


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
View Abstract