PATIENT AND METHODS

Patient

The patient, 42 years old, had moderate mental retardation with hearing loss and mild renal impairment. Several dysmorphic features were observed (fig 1), including a long face with hypoplastic malae and retrognathia, a low nuchal hairline, downward slanting palpebral fissures, simply formed ears, a large nose with a high nasal bridge, full lips and a high arched palate, adenoma sebaceum on the chin and in the nasolabial region, and a webbed neck. He had finger-like thumbs, clinodactyly of the second finger of the right hand, and a subungual fibroma on the third finger. There was cutaneous syndactyly of the second and third fingers of the left hand. He had a wide space between the first and second toe of his feet and subungual fibroma on the second and third toe of the right foot and on the third toe of the left foot. X rays showed a triphalangeal thumb on the right hand but not on the left hand. The second and third cervical vertebrae were fused. On his back were two shagreen patches. Further investigation, following the tuberous sclerosis protocol including MRI of the brain and echocardiography, was refused. A CT scan of the abdomen was, however, performed at the age of 44 years, showing enlarged cystic kidneys and multiple cysts in the liver, as seen in ADPKD. In the following 14 months plasma creatinine levels increased from 320 to 500 μmol/l indicating decline of renal function. The patient suffered from hypertension and progressive anaemic symptoms. Haemodialysis was started but this treatment was not well tolerated, and the patient died after three months. Necropsy was not performed. Family history was negative for tuberous sclerosis, polycystic kidney disease, and dysmorphic features. A diagnosis of tuberous sclerosis with an acrofacial dysostosis-like phenotype and polycystic kidneys was made.

Cytogenetic analysis

Chromosome analysis on cultured peripheral blood lymphocytes was performed according to standard G banding procedures.

For fluorescence in situ hybridisation (FISH), metaphase preparations from EBV transformed harvested cells were prepared as described by Landegent et al.⁷ FISH analysis was performed as described previously⁸ with all cosmids and PAC clones. Fine mapping of the deletion breakpoints was done by fibre-FISH using PAC clone 64.12C and 77.3D according to the protocol described by Giles et al.⁹

Clones

Chromosome 16 cosmids encompassed the TSC2 gene (LADS4, ZDS5¹⁰), the PKD1 gene (ZDS5, REP59, 2H2^11,12), and cosmids located more proximally (cos 3, provided by Dr S Reeders, Yale, USA, 218+218.10¹³) or more distally (UW3+5, cos2B,¹⁴ cos40¹⁵) from the region. Chromosome 16 PACs were 109.8C, 77.3D, 91.8B, 97.10G, 1.8F, 96.4B, and 64.12C.¹⁶

Database analysis

Blast searches of the databases of the National Center for Biotechnology Information (NCBI, http://www.ncbi.nlm.nih.gov) were performed using the DNA sequences of ESTs or predicted transcripts and to obtain the sequences of PAC clones. The NIX program at www.hgmp.mrc.ac.uk, which is a tool for identifying unknown nucleic acid sequences, was used to analyse the sequences of the PAC clones 1.8F and 96.4. Gene predictions are now also available using the map viewer on the NCBI website. (Accession numbers: PAC clone 47-2H, AC005363; PAC clone 109.8C, AC005606; PAC clone 109.9G, AC005600.1; PAC clone 1.8F, AC00711; PAC clone 96.4B,: AC005212.)

RESULTS

The patient is a man with an acrofacial dysostosis-like phenotype, tuberous sclerosis, and cystic kidneys; the latter were not initially noticed. Cytogenetic analysis showed a normal male karyotype. Whether a chromosomal aberration involving the TSC1 gene on chromosome 9q34 had occurred was initially investigated, as Zori et al² described a patient with Nager anomaly and an apparently balanced translocation, 46,X,t(X;9)(p22.1;q32). DNA analysis using highly polymorphic microsatellite markers flanking the TSC1 gene on chromosome 9 did not show a deletion (data not shown). After renal impairment and polycystic kidneys had been detected, FISH analysis with cosmids mapped to chromosome 16p13.3 was performed, showing that the cosmids encompassing the TSC2 gene (LADS4, ZDS5¹⁰) and the PKD1 gene (ZDS5, REP59, 2H2¹¹) were deleted on one copy of chromosome 16. We defined the deletion boundaries using a set of overlapping P1 clones (fig 2)¹⁶ and observed a deletion of (large parts of) P1 64.12C, 96.4B, 1.8F, 97.10G, 91.8B, and 77.3D. We did, however, detect signals more proximally on chromosome 16p for the P1 clones 97.10G and 91.8B as well as for the cosmids REP59 and 2H2, as these clones contain (part of) a 40 kb repeated region within the PKD1 gene and recognise several copies of the homologous sequence located on chromosome 16p13.1.¹¹

The P1 clones 64.12C, 96.4B, and 77.3D gave weaker than expected signals, indicating that part of these clones were deleted. Using fibre-FISH with clones 64.12C and 77.3D, the deletion breakpoints were precisely mapped. Fig 2 shows a fusion of the signals of these two P1s, which on fibres of the normal chromosome 16 are 200 kb apart. As several of these clones have been sequenced by the Center for Human Genome Studies (DOE Joint Genome Institute, Los Alamos National Laboratory), we could improve the physical map of this chromosomal region. However, a few small gaps still exist in the sequence and only a draft sequence is available for clone 1.8F. We concluded that the deletion ranges from the middle of P1 clone 64.12C to the middle of P1 clone 77.3D and estimated the size to be approximately 280 kb.

Sequence analysis

The acrofacial dysostosis-like phenotype suggests that besides the PKD1 and TSC2 genes, one or more additional genes are contributing to the patient's phenotype. From centromere to telomere, genes encoding the following proteins are deleted. The ABC3 transporter (L48758, L48760, L75924, L75925) is involved in transport of molecules into or out of cells and across subcellular membranes.¹⁷ The human S phase prevalent DNA/RNA binding protein RNPS1 (accession No L75926 and L75927) is suggested to be involved in splicing or in processing of precursor RNAs into mature mRNA.¹⁸ Dodecenoyl-coenzyme A delta isomerase (DCI) is a mitochondrial 3,2-trans-enoyl-coenzyme A isomerase suggested to have a role in oxidation of unsaturated fatty acids.¹⁹ The deoxyribonuclease I-like2 (DNASE1L2) gene²⁰ is related to deoxyribonuclease I, which plays a role in the breakdown of nucleic acids in the gastrointestinal tract. An additional role for deoxyribonuclease I in apoptosis has also been proposed.²¹ The exact function of DNAS1L2, however, has still to be elucidated. The adenovirus E1A regulated transcription factor E4F1, the human homologue of the transcription factor ϕAP3 (accession No L48762, L48763), is a negative regulator of genes encoding proteins responsible for the inhibition of the cell cycle.²² The human homologue of the rat RAB26 ras related GTPase (L48770, L48771), which is a GTP binding protein involved in vesicular transport,²³ is the final deletion.

The following genes are located distally from the PKD1 and TSC2 genes. The endonuclease III-like 1 gene (NTLH1) located in the P1 clone 77.3D but not in 109.8C, encodes an enzyme with DNA glycosylase and DNA lyase activity and could be involved in DNA repair.²⁴NTLH1 is located 5` to 5`, “head to head”, with the TSC2 gene with only 63 bp in between. Furthermore, the gene is located in a 3` to 3`, “tail to tail”, orientation with the SLC9A3R2 gene.²⁵ This gene encodes the regulatory factor 2 of the solute carrier family 9 (sodium/hydrogen exchanger), but has recently been shown to be a regulator for phospholipase-β3.²⁶ The distal deletion breakpoint is not exactly known. One or both of the genes NTLH1 and SLC9A3R2 may be deleted or disrupted in the patient.

Exon trapping has previously identified four of the genes located in the 280 kb deletion interval, ABCA3, RNPS1, E4F1, and RAB26. The authors identified one additional transcript, “I”, located between E4F1 and RAB26, with unknown function.²⁷

The SazD transducin gene,²⁸ the human ERV1 gene,²⁹ and the gene encoding the ribosomal protein-like 3 (RPL3L)³⁰ are located distally from the deletion interval.¹⁶ The human somatostatin receptor 5 gene³¹ also maps outside the interval, on cosmid 349E16, although in the Puffer fish Fugu, the gene is located within 10 kb proximal of the PKD1 gene.³¹

DISCUSSION

Tuberous sclerosis (TSC) is a genetically heterogeneous disorder with genes on chromosome 9 (TSC1) and chromosome 16 (TSC2). In the patient presented in this paper, the diagnosis of TSC was based on adenoma sebaceum, nail fibroma, and shagreen patches. Along with TSC, he had dysmorphic features, triphalangeal thumbs, and hearing loss, consistent with a diagnosis of acrofacial dysostosis. A patient with acrofacial dysostosis and a balanced translocation 46,X,t(X;9)(p22.1;q32) inherited from a mosaic carrier mother was described by Zori et al,² suggesting the location of a gene for acrofacial dysostosis on chromosome 9 or on the X chromosome. The fact that the TSC1 gene is located on chromosome 9q34 prompted us to analyse this chromosomal region, but no chromosomal aberration was found. As the patient also had polycystic kidneys, it was a logical step to analyse the short arm of chromosome 16, uncovering a microdeletion of approximately 280 kb. Since dysmorphic facial features are not manifestations of either TSC or ADPKD, we suggest that the region also harbours an acrofacial dysostosis gene. Database analysis showed at least six other genes in the deletion interval.

The acrofacial dysostoses are a heterogeneous group of disorders with craniofacial anomalies and defects in limb development.¹ For a large number of malformation disorders, disruption or dosage effects of specific transcription factors or transcription enhancers/repressors are known to cause the disease. On this basis the ZFP-37 gene, a putative transcription factor belonging to the GLI-Kruppel gene family, was proposed as a candidate gene for the Nager acrofacial dysostosis syndrome.³² Interestingly, E4F1, deleted in our patient, is also a transcription factor related to the GLI-Kruppel family of zinc finger proteins.²² There have been reports of other deletions of patients with no dysmorphic phenotype,^6,33 which partly overlap with the deletion described here. The most proximal extending deletion extends 80 to 100 kb proximal from exon 1 of the PKD1 gene,⁶ which is probably distal to E4F1. Recently, a patient with an unbalanced translocation t(8;16)(q24.3;p13.3) with TSC, ADPKD, and hypomelanosis of Ito was described.³⁴ Downward slanting palpebral fissures, mild malar hypoplasia, and mild retrognathia were present in this patient, but acrodysostosis was not reported. In this patient, however, the most telomeric part of 16p is deleted, including the PKD1, TSC2, and α-globin genes, and the 8q24.3-qter region is duplicated. The chromosome 16 breakpoint is located in P1 clone 1.8F but the position in relation to E4F2 is not known.

The distal deletion breakpoint is located in the middle of P1 77.3D. Two genes are located very close to the breakpoint, the endonuclease III-like 1 gene and the gene encoding the regulatory factor 2 of the solute carrier family 9 (sodium/hydrogen exchanger).²⁵ As we did not observe reduced FISH signals for P1 109.8C, the breakpoint is located proximally to or at the beginning of this clone. The proximal breakpoint disrupts the ATP binding cassette transporter ABC3. The substrate of the transporter is not known but the highest mRNA expression has been seen in lung tissue.¹⁷

In general, deletions spanning TSC2 and PKD1 have been implicated in a severe and infantile form of polycystic kidney disease in TSC.³⁵ Progression of renal cystic disease is apparently accelerated when, in addition to inactivation of the PKD1 gene, TSC2 function is lost. A patient with neonatal presentation of polycystic kidney disease with paternally inherited ADPKD and maternally inherited TSC supports this observation.³⁶

A few other TSC2-PKD1 deletion patients with a milder phenotype have been described.³⁴ Sampson et al⁶ observed somatic mosaicism in four of 22 patients with contiguous deletions of TSC2 and PKD1. The relatively mild cystic phenotype found in our patient could be the result of somatic mosaicism. Alternatively, genes in the region counteracting the effect of the PKD1/TSC2 genes or modifying genes located on other chromosomes could explain the milder phenotype of the patient presented here. FISH analysis did not show any mosaicism in the patient presented in this paper, but this was only performed on EBV transformed lymphoblastoid cells. Unfortunately, the patient died from renal failure and inadequate compliance with haemodialysis. For this reason, a fresh blood or tissue sample to test for mosaicism could not be obtained. We therefore cannot definitively conclude whether the mild phenotype was caused by mosaicism or by other genes influencing the phenotype.

The PKD1-TSC2 region on chromosome 16p13.3 is particularly gene rich. We propose that haploinsufficiency of one of these genes can cause acrofacial dysostosis or that the deletion has exposed a mutation in a gene on the non-deleted chromosome. The transcription factor E4F1 seems to be an interesting candidate gene.