Background
Biallelic mutations of the C16orf57 (OMIM*613276) gene underlie Poikiloderma with Neutropenia (PN; OMIM#604173), an inherited genodermatosis characterized by early onset poikiloderma, pachyonychia, palmo-plantar hyperkeratosis, skeletal defects and non-cyclic neutropenia. This condition results in recurrent infections during infancy and childhood, primarily of a pulmonary nature, and contributes to the postnatal growth delay in weight and height of the patients.
The syndrome was first described by Clericuzio in Navajo Indians [
1] and subsequently in patients of Caucasian ancestry from different geographic areas [
2‐
5]. Following discovery of the causative gene [
6] and molecular evidence for distinct genetic control between PN and Rothmund-Thomson syndrome (RTS; OMIM#268400) [
7], 31 PN patients have been tested and found to bear 17 different mutations in the responsible
C16orf57 gene, 84% of the patients were found in the homozygous state [
6,
8‐
12] and only six were compound heterozygous [
6,
11,
13]. Interestingly this cohort includes patients previously diagnosed as affected with Dyskeratosis Congenita (DC; OMIM#224230) and with Rothmund-Thomson syndrome illustrating significant phenotypic overlap among these entities [
10,
13]. All the identified mutations, six nonsense, six frameshifts and five splicing mutations including the apparent missense change, c.502A>G [
6], lead to loss-of-function. Despite the limited mutational repertoire a few recurrent mutations have been identified, including c.496delA, common among those of Athabaskan ancestry [
11], c.531delA in Turkish families [
10,
12] and c.179delC in patients of North African origin [
8,
12], consistent with founder mutations restricted to different geographic areas.
Genotyping further patients could better define the spectrum, the type and the geographical distribution of the C16orf57 sequence changes.
The
C16orf57 gene is phylogenetically conserved and ubiquitously expressed, suggesting a housekeeping function [
10].
Whenever cDNA analysis has been performed, mutant alleles have been detected [
6,
11] hinting they might be translatable. The function of the unidentified C16orf57 protein remains obscure, so we focused on bioinformatic tools to predict the possible structure and ancestry of C16orf57 and to gauge the functional consequences of the different mutations on the folded protein chain, a start point to address genotype-phenotype correlations in the patients.
In this study we characterize six PN patients and define at the DNA and cDNA level the underlying
C16orf57 mutations. The mutations all present in the homozygous state include a novel early truncating mutation, a novel IVS2 splice-site mutation and an IVS6 splice-site mutation already described but in the heterozygous condition [
11].
We also present the computational analysis of the C16orf57 protein chain, showing that two conserved H-X-S/T-X tetrapeptide motifs (where X is a hydrophobic residue) likely mark the active site of a two-fold pseudosymmetric structure related to the 2H phosphoesterase superfamily [
14,
15]. Based on this analysis we predict that C16orf57 belongs to this enzyme class and the effect of all
C16orf57 mutations reported so far cause disruptions of both protein fold and catalytic site that hold the critical function of the protein.
Methods
Patients
Six patients, three males and three females, referred to us by clinical geneticists and dermatologists, were enrolled in this study. Patients and their parents provided appropriate informed consent.
DNA extraction and mutation analysis
Genomic DNA was isolated from buccal swabs of patient #26 with NucleoSpin Tissue (Macherey-Nagel, Bethlehem, PA 18020, USA) and from whole peripheral blood of patients #11, #16, #17a, #21, #25 and their relatives using the Wizard Genomic DNA Purification Kit (Promega Corporation, Madison, WI, USA).
About 100 ng of DNA were amplified using GoTaq polymerase (Promega) with previously published primers and conditions [
6]. Amplicons were sequenced using Big Dye Terminator v.1.1 Cycle Sequencing Kit according to the manufacture's protocol on the ABI PRISM 3130 sequencer (Applied Biosystems, Foster City, CA, USA). Electropherograms were analyzed with ChromasPro software 1.42 (Technelysium Pty Ltd, Tewantin QLD, Australia) using the wild type sequence of
C16orf57 gene [ENSG00000103005] as reference.
RNA extraction and RT-PCR analysis
EBV-transformed lymphoblastoid cell lines (LCLs) were established for patients #21, #17a and their parents. LCLs were cultured in RPMI 1640 medium with 2 mM L-glutamine (EuroClone, Milano, Italy) supplemented with 20% foetal bovine serum (Lonza, Walkersville, MD, USA) and 1% Penicillin, Streptomycin and Ampicillin in 37°C humidified incubator with 5% CO2.
Total RNA was extracted from LCLs using TRI-Reagent RNA Isolation reagent (Sigma-Aldrich, Saint-Louis, MI, USA) and from whole blood of patient #26 using PAXgene Blood RNA Kit (PreAnalitix, Hombrechtikon, Swiss).
500 ng of total RNA was reverse-transcribed into cDNA using High Capacity cDNA Reverse Transcription Kit (Applied Biosystems) with random hexamers. PCR amplification of the
C16orf57 gene transcripts was performed for patients and positive controls using GoTaq polymerase (Promega). Specific primers and conditions for
C16orf57 transcripts are listed in additional file
1, Table S1. Nucleotide sequences were compared to the major
C16orf57 transcript reference sequence [ENST00000219281].
Computational analysis
An evolutionary and structural profile of C16orf57 was constructed by iterative PsiBLAST searches of Genbank [
16] followed by PsiPRED secondary structure prediction of the resulting multiple sequence alignment [
17]. This alignment was also used by HHrep [
18] to look for internally repeated segments of sequence and predicted structure, and by HHpred to locate potentially related folds from the PDB by sensitive HMM-HMM comparison [
19]. Comparative modeling of the C16orf57 three-dimensional structure from fold recognition-derived templates was performed by MODELLER [
20] and I-TASSER [
21]. Surface patterns of conservation and variability in the resulting model were derived by ConSurf [
22]. Structure manipulation and electrostatic potential surface visualization of C16orf57 and related structures were done with Pymol
http://www.pymol.org.
Discussion
Since the discovery of the causative gene, Poikiloderma with Neutropenia syndrome can now be confirmed by molecular diagnostic testing for mutations in the
C16orf57 gene and thereby differentiated from the phenotypically similar clinical entities Rothmund-Thomson syndrome and Dyskeratosis Congenita. Based on the molecular test, patients previously diagnosed on clinical grounds as having RTS or DC have been successfully reclassified [
13].
We have identified biallelic mutations of the
C16orf57 gene in six patients including #26, previously described as atypical Rothmund-Thomson [
24] and #21, reported as affected with Poikiloderma associated with Osteopetrosis [
23]. All the patients herein molecularly confirmed as PN display the consensus PN clinical signs, i.e. early-onset poikiloderma involving the extremities and extending to the face, pachyonychia, palmo-plantar hyperkeratosis and non-cyclic neutropenia (Table
1). It is worth mentioning that our patient #17a shows poikiloderma also on the ear helix, as previously reported in other patients [
3,
9].
Craniofacial dysmorphisms have been reported only in a few PN patients [
5,
32]; facial features such as a saddle nose and midfacial hypoplasia are displayed by four out of our six patients, suggesting they are quite common and hence should be recorded during PN clinical evaluation. Dental defects are common in our cases, in agreement with the literature [
4,
13,
26,
33]. Hypogonadism was identified in male patients #17a and #26 and previously reported [
3,
34]. In contrast to
RECQL4-positive RTS, where bony changes constitute a major diagnostic sign [
7], skeletal involvement has only seldom been described in PN patients [
3,
13,
26]. Indeed, apart from our patients #17a and the infant #25, all herein described PN patients display overt skeletal signs including zygodactyly between the second and third digit (#26), multiple bone fractures (#11), intermediate osteopetrosis (#21) or X-ray detectable skeletal findings (#16) (Table
1). A relationship may be envisaged between zygodactyly and the swan neck hand hyperflexibility noticed in a few described patients [
3,
5,
33]. More generally delayed bone maturation has been recorded [
32] in the girl subsequently confirmed to carry two distinct
C16orf57 mutations [
6], and diffuse osteosclerosis has been underlined by Porter [
26] in the patient found to harbour
C16orf57 mutations [
10]. This patient is one of three cases reported as RTS [
10] with defects in the bone marrow, as reported for other clinically diagnosed RTS cases [
32‐
34]. We believe these cases, previously classified as affected by RTS, more correctly represent cases of PN as confirmed by molecular testing of some of them [
6,
10]. All our patients display non-cyclic neutropenia, the hallmark of PN, and this is a fundamental sign differentiating PN from both RTS or DC, which should be searched also in suspected cases where it remains silent [
13]. Neutropenia leads to recurrent infections, which are widely documented in the paediatric histories of PN cases and confers upon them a high risk to develop myelodysplasia from the second decade of life, as attested in at least ten
C16orf57- positive patients [
5,
10] and our patient #17a. When considering bone marrow hypocellularity, increased myeloid precursors and delayed neutrophils maturation, a higher percentage of
C16orf57-positive cases displays this feature [
4,
5,
10,
11] highlighting the sensitivity of the myeloid lineage to
C16orf57 mutations. Evolution to acute myeloid leukemia has been reported in a few PN patients [
10,
26] and in other untested cases [
25,
33‐
35]. This evidence features PN as a cancer predisposing syndrome affecting the myeloid compartment and connects PN to RTS, a syndrome with an increased risk for osteosarcoma and skin cancer [
7], and DC which predisposes to a wide variety of haematological and solid tumors [
36]. Molecular characterization is thus compulsory for assigning appropriate oncological surveillance.
As depicted in Figure
3 the recurrence of the c.531delA, c.496delA and c.179delC mutations delineates three clusters according to the geographic origin of the patients. A common ancestor is the likely hypothesis to explain the recurrence of these mutations restricted to specific ethnic groups, considering the very low frequency of the PN syndrome and the prevalence of patients with homozygous mutations (31 out of 37). Genetic analysis and reconstruction of ancient genetic links through haplotype segregation analysis could confirm this assumption.
Expansion of the
C16orf57 mutational repertoire and validation of the observed geographical distribution may allow assessment of clinical variability of PN phenotype in distinct founder mutation cohorts, as it has been described for another autosomal recessive developmental disorder [
37].
Lack of information of the function of C16orf57 protein makes it difficult to establish a link between mutations and the onset and evolution of syndromic presentation.
Bioinformatic analysis can provide a preliminary tool to predict the severity of specific mutations.
The structure-based inference that human C16orf57 is a member of the 2H phosphoesterase superfamily, despite very little sequence identity, rests on the recognition of a common protein fold by sensitive algorithms that weave together evolutionary information, in the form of sequence patterns that are conserved across C16orf57 orthologs, and accurate secondary structure predictions to comprehensively scan structural databases. The most prominent pattern in C16orf57 is a two-fold repeated sequence and structural segment with signature H-X-T/S-X tetramotifs; these conserved features are precisely mirrored in the fold-recognition-derived matches with bilobal 2H phosphoesterase folds that rely on the symmetrically poised His residues for catalytic activity. The 2H phosphoesterase superfamily is a diverse grouping of enzymes with a common core architecture, a basic hydrolytic focus for cyclic phosphates, and some functional variability in both substrates and reactions [
14,
15]. This degree of functional diversification is observed in other structurally-assembled superfamilies of enzymes [
38]. Of the various biological tasks performed by 2H active site enzymes, we argue (by closer structural resemblance) that C16orf57 is perhaps an RNA ligase though the actual targets or substrates of this activity remain unknown.
All the mutations reported in our patients are in the homozygous state, which facilitates transcript analysis and understanding the effect of the mutations.
Indeed with the exception of the Algerian patient #11, transcripts have been tested, detected and sequenced in all other patients (Figure
2). With regards to the c.693+1G>T mutation in patient #25, transcript analysis has been reported [
11] suggesting exon 6 skipping according to the size of the aberrant band; indeed a misspliced 84 base shorter transcript was found associated with the c.683_693+1del12 which affects the same IVS6 donor splice site [
6]. The general emerging feature is that the aberrant transcripts are detectable even for early truncating mutations, such as those of patients #21 and #26, pointing out they are relatively stable and translatable.
Conclusions
It is known that Dyskeratosis Congenita, Rothmund-Thomson and Poikiloderma with Neutropenia have many overlapping features. Starting from the clinical presentation it is very difficult to assess the correct diagnosis, and in fact most of the molecularly confirmed PN patients have a long history of wrong diagnosis, as DC or RTS. The availability of C16orf57 molecular testing allows the correct diagnosis which is compulsory for retargeting syndrome-specific oncosurveillance.
As the present study shows, all the 19 C16orf57 mutations linked to disease involve the predicted enzymatic domain of C16orf57 protein. By destroying the native fold, all mutations should cause a drastic loss of enzymatic activity. Future studies of protein presence and activity in PN patients could confirm whether and how the aberrant transcripts, which have been always detected whenever assayed, may be translated.
The
C16orf57 gene is ubiquitously expressed [
6,
10], but not all tissues are equally affected by the lack of correctly functioning C16orf57 protein during development and throughout life.
The onset of the poikiloderma, nail dystrophy and teeth malformations at early infancy reflects the perturbed morphogenesis of skin and cutaneous annexes, while neutropenia results from impaired homeostasis of the highly C16orf57 expressing myeloid cells [
6,
10]. The life long risk of myelodysplastic syndrome features the increased tendency to apoptosis and leukemic transformation of C16orf57-defective myeloid progenitor cells [
39].
However, whether different mutations impact differently on the clinical phenotype and on the risk of myelodysplastic syndrome awaits further clinical and molecular characterization of PN patients, along with delineation of patients' clinical expressivity in distinct geographical areas and dissection of the biological function of the C16orf57 protein.
Acknowledgements
We thank the patients and their relatives for intensive cooperation in the study; Galliera Genetic Bank-Network of Telethon Genetic Biobank-project GTB07001- for providing us lymphoblastoid cell lines from the affected siblings of the PN families and DNA of patient #21; we thank Dr. Guillemette Frémont (Departement of Dermatology, Hôpital Saint-Louis, Paris, France) for referring to us patient #11.
This work was supported by Associazione Italiana per la Ricerca sul Cancro (grant 2008-2009/4217 to L.L.), CARIPLO N.O.B.E.L. (project 2007-2009 to L.L), Nando Peretti Foundation (grant 2007-2011 to L.V.) and the "Dote di Ricerca": FSE, Regione Lombardia granted to Elisa Adele Colombo, Dipartimento di Medicina Chirurgia e Odontoiatria, Università degli Studi di Milano for the research project "Functional characterization of the new C16orf57 gene".
L.V. and L.L.: equal contribution.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
EAC performed molecular analysis and interpreted the predicted effects of mutations, drafted the manuscript and approved the final version. JFB performed and interpreted computational analysis, drafted and revised the article, approved the final manuscript. GN performed and interpreted molecular analysis, mined the literature and approved the final manuscript. CG supported the molecular work and software access. NE and DY contributed with clinical data and biological samples of patients#16 and #17a, provided helpful feedback and approved the final manuscript. NA and UC contributed with clinical data and sample of patient #26 and approved the final manuscript. AT, ADF and ML contributed with clinical data and sample of patient #21 and approved the final manuscript. SKS and ACY contributed with clinical data and sample of patient #25, revised and approved the manuscript. LV contributed to recruit patients, designed the geographical analysis, drafted, revised and approved the manuscript. LL designed and coordinated the study; drafted, revised and approved the manuscript.