June 26th, 2009
Re: Ms1731504354252538: " A novel study of Copy Number Variations in Hirschsprung disease
using Multiple Ligation-dependent Probe Amplification (MLPA) technique."
Dear Editor
Thank you for your mail of June 18th containing the referees´ comments and inviting us to send
a revised version of our manuscript. We have amended the manuscript following your
recommendations, and the comments of the referees were addressed as follows:
Reviewer 1
“The authors describe results of an MLPA study of copy number variations in Hirshsprung
disease. This is a straight-forward paper and discussion illustrating that CNVs in 4 commonly
associated genes appear not to involve CNVs. My concern here about the conclusions from the
authors is the sample source. The authors indicate the patients have the disorder, but they
never indicate if these patients have a known mutation that is causative of the disorder or if they
are idiopathic cases. If the cases are all idiopathic, then this should be indicated. If, however,
the cases all have mutations in a known Hirshsprung gene, then these are not the cases to
evaluate for CNVs. This needs to be clarified”.
We completely agree with the referee that a clarification is needed justifying the inclusion of our
patients in the present study. In this sense, it is fundamental to state that Hirschsprung disease
(HSCR) stands as a model for genetic dissection of complex diseases. In this model, a major
gene, RET, is involved in most if not all cases of isolated HSCR, in conjunction with other
autosomal susceptibility loci under a additive/multiplicative model. As commented in the text, to
date it is known that HSCR susceptibility alleles in RET can harbor either heterozygous coding
sequence mutations or, more frequently, a polymorphism within intron 1, leading to a
hypomorphic RET allele (see paragraph 2). It is conceivable that such RET mutations, together
with other events occurring in RET or other genes (such as CNVs), might give rise to the
phenotype, which could explain the non-mendelian features of the disease. Our HSCR cases
had been previously screened for both RET coding point mutations (Ruiz-Ferrer et al., 2006 and
unpublished data), and RET “enhancer mutation” within the intron 1 (Fernandez et al., 2005). In
addition, as explained in the text, the presence of variants in the remaining 3 genes (GDNF,
EDN3, ZFHX1B) that might affect the MLPA procedure had been also analyzed by dHPLC
(unpublished data). Other HSCR genes had been also evaluated. However, the presence of
such mutational events in some of our patients does not exclude the possibility that putative
CNVs in RET or other HSCR genes could also be playing a major or a modifier role in the
pathogenesis of the disease. This is the reason why we have analyzed our complete series in
this study, and not only the “idiopathic” cases without mutations in HSCR genes. With the aim to
clarify this point, we have included a paragraph in the “Patient and Control subjects” section
explaining the mutational status of our HSCR patients and justifying why they have been all
analyzed for CNVs. Moreover, a little paragraph emphasizing the complex, multifactorial, nonmendelian nature of Hirschsprung disease has been added within the Background section.
2. Regarding the patient samples indicated with the mutations near the MLPA probes, it seems
these patients should have been screened previously for mutations in these genes, or they are
not idiopathic cases. In fact, it seems possible the 2 mutaitons detected could actually be
causative. For the first patient, the "mutation" is indicated P953L, but DNA, not protein should
be written in the text--or both. The second patient has a 2 bp deletion, which would result in
a frameshift.
The referee is right when supposing that the 2 mutations affecting the hybridization process
within the MLPA technique (RET P953L and EDN3 c.572delAA) could actually be causative of
the disease. However, in both cases, the mutations had been inherited from one of the
unaffected parents, which suggest that additional mutations are required to produce the
phenotype in the patients. This is in fact one of the typical features of complex diseases and, as
previously commented, justifies the inclusion of both patients in this study. Regarding the
nomenclature for the mutations, we have followed the recommendation of annotating them both
at the protein and DNA levels.
3. Page 4, lines 5-6. Sentence is not clear.
The original sentence was “Nevertheless, additional studies are required to deep in their real
involvement in the disease”, and has been substituted by “Nevertheless, additional studies are
required to dig in their real involvement in the disease”.
4. Would recommend a primary English speaker review the details of the text for clarity.
As suggested, the whole manuscript has been reviewed and edited by a native English speaker.
Reviewer 2
1. This is a poorly written and generally weak paper. The English is very weak (see examples
below).
The whole manuscript has been reviewed and edited by a native English speaker in order to
improve its quality.
2. While 2 examples were found of decreased dosage, the authors subsequently demonstrated
that these represented sequence variations that affected probe annealing, rather than true copy
number changes. The authors do not discuss why these sequence variants were not detected
during their dHPLC screening prior to the use of MLPA.
It is evident that there must have been a misunderstanding regarding this point. The 2 variants,
(as well as many others not affecting the probes annealing) had been previously detected by
dHPLC and characterized by direct sequencing. The fact is that we conducted the MLPA
analysis on our complete series of patients without knowing a priori the exact probes binding
regions. Only after observing the decreased dosages for RET exon 17 and EDN3 exon 4, a
detailed inspection of the corresponding probe binding sites demonstrated that they were
affected by the presence of the previously characterized point mutations. In order to clarify this
point we have performed some modifications in the text. For instance, we have substituted the
sentence “Once we identified the patients with the altered profiles we sought to confirm if they
presented any sequence variant that could be affecting our results”, by “We therefore
investigated the exact probe binding sites for the 2 altered exons and sought to confirm if they
were affected by the presence of any sequence variant previously detected in the
corresponding patients.”
3. Apart from the poor English, my major criticism is that this paper reveals little that is novel,
either in terms of results or methodology. The authors used a commercial kit, developed by
MRC-Holland. Had they designed their own assay and used this publication to demonstrate it’s
benefits over an off the shelf kit, I might have been more favorable towards this article.
In response to this comment we have to argue that although the results are essentially negative,
they lead us to discard CNVs in RET and in the other 3 genes as a common cause of
Hirschsprung disease, and this constitutes per se a novel although modest finding. The referee
is right when stating that the use of a commercial kit is not a novel approach in methodology.
However, it does not make sense to design a new assay when a perfectly optimized kit is
available for the Scientific Community and is absolutely appropriate to reach our initial
objectives.
4. There is clearly a fine line between routine diagnostics and research and my view is that the
contents of this paper fall more on the side of diagnostics. It should be routine, these days, for
any condition that results from alterations of known genes to look for copy number variations as
well as sequence changes in those genes. As such, it is only novel methods or variations on
known methods that should be worthy of publication.
Against what is considered by this referee, this paper is focused in investigation on
Hirschsprung disease, the paradigm of complex traits in which the contribution of several
mutational events seems to be necessary to produce the phenotype. It is not our purpose to
present the MLPA method as a diagnostic method for the disease, but to investigate in the
presence of CNVs with an appropriate methodological approach. In other words our main
interest when planning this study was centred in the potential results (finding CNVs associated
to the disease), and not in the implementation of the method for diagnosis.
Reviewer 3
1. The Authors should explain more clearly why they chose those genes. In fact, if we exclude
Ret gene, the others are involved only in a minority of the patients. In particular, ZFHX1B is
specific for the Mowat-Wilson syndrome, which is extremely rare.
We agree with this referee that with the exception of RET, the remaining genes here evaluated
have a minor impact in Hirschsprung. The reason why those 3 genes were included in the study
is simply because the MLPA commercial kit used for RET testing, also included probes for their
detection. With the aim to clarify this matter, we have modified the last paragraph of the
Background section. The sentence “... in the present study we have analyzed the presence of
CNVs for 4 “HSCR genes” in our series of patients...” has been substituted by “...in the present
study we have sought to analyze the presence of CNVs for RET, the “major HSCR gene” in our
series of patients, using the Multiple Ligation-dependent Probe Amplification (MLPA) technique,
which is optimal to detect alterations in the genes dosages [7]. In addition, another 3 genes
responsible for a minority of the cases, (GDNF, EDN3, ZFHX1B) were simultaneously analyzed
since the MLPA kit employed also contained probes for their detection.
2. Basing on previous considerations, the Authors should indicate how many of their patients
have a syndromic HSCR and how many have a isolated form of the disease. This is of utmost
importance when discussing the results, particularly if certains syndromes such as Down,
Ondine, and Waardenburg-Shah are present in their series of patients. In fact those syndromes
are well known to be determined by mutations of other genes but those screened by the
Authors. Were patients with those syndromes excluded or not?
Following this recommendation, we have specifically stated in the Method section that all our
cases were isolated forms, without any additional clinical feature suggestive of a syndrome.
3. As stated by the Authors, the study did not include non-coding Ret regions and a number of
important syndromic HSCR genes. Basing on this consideration, I would strongly appreciate
further studies on this regard. In particular, it would be of great interest to screen those patients
with clinical features suggestive for a well defined syndrome who turned out not to have any
known mutation. An example could be to apply MLPA or CGH-array to those patients with
HSCR associated congenital central hypoventilation syndrome but no PHOX2B mutation that
could carry CNV on the PHOX2B gene. This would increase the diagnostic power of genetic
screening of these syndromes.
This is a particularly interesting observation and will be taken into account for future studies. In
fact we have already planned to design our own MLPA assay to analyze other different HSCR
genes/loci in the context of both isolated and syndromic forms. In this way, the final conclusion
of our paper, in accordance with the opinion of this referee, is that further studies are warranted
to completely discard the CNVs as the molecular cause of some forms of Hirschsprung disease.
We hope you will find the changes appropriate and that our manuscript may now be suitable for
publication in BMC Medical Genetics.
Thank you very much for your time and consideration.
Yours sincerely,
Salud Borrego, MD, PhD
Unidad de Gestión Clínica de Genética, Reproducción y Medicina Fetal
Hospitales Universitarios Virgen del Rocío,
Avenida Manuel Siurot s/n, 41013 Sevilla, Spain.
Voice:+34 955 012778, Fax:+34 955 013473
E-mail: salud.borrego.sspa@juntadeandalucia.es