We used a combination of direct sequencing and DHPLC to search for mutations in 8 patients with DMD and a carrier sister. Likely causative mutations were found in 6 of 8 males and confirmed in the carrier sister of one. In our study, one likely causative mutation (patient 03 del exn21) was found just by PCR analysis and 5 likely causative mutations were first detected by direct sequencing but each was subsequently detected by DHPLC. Our experience indicates that the multiple different conditions of DHPLC would have detected all of these likely causative mutations and all polymorphisms. Direct sequencing of all the fragments is still more costly than DHPLC and our results indicate that initial analysis using the far cheaper DHPLC should precede sequencing thus reducing cost of the analysis. Causative mutations were not found in two of the eight patients. Due to the enormous size of the dystrophin gene (2.4 million base pairs), finding 100% of mutations is improbable using these fragments because we cannot examine all the sequences and situations that might affect expression. It is possible, but unlikely, that one of the currently unknown alterations that we found in patients 05 and 06 will eventually be proven a causative mutation rather than a polymorphism.
Two of the patients in which we found disease-causative mutations have affected brothers, three in addition to patient 01, have potential carrier sisters, and all six have mothers who could also be carriers. We have shown that, 1) using the DHPLC system, it is now not only feasible but actually a simple process to determine if any of these relatives have the same point mutation as their brother/son and, 2) that carrier testing and pre-natal diagnostic testing is now available to any of these patients who wish to use it.
As larger cohorts of patients are tested, a more accurate estimate of the percentage of undetectable causative mutations will emerge and present a clear new challenge. We estimate based on our small sample, and other unpublished data, that such cases will comprise a small percentage (less than 8 percent) of total DMD cases. Possible explanations for such cases include duplications which we may have missed, mutations in unknown enhancers or translation modifiers hidden in exons or introns, mutations that create novel splice sites and changes in the coding region which might be pathogenic but which, due to our lack of knowledge, are thought to be polymorphic. For example, changes to an amino acid that is essential for some protein/protein interaction (potentially transportation), or is modified/processed on the protein level but which we currently assume is just a polymorphic change. Other possible mechanisms include mosaicism, in which DNA extracted from blood lymphocytes has different sequence than DNA extracted from muscle cells, and cryptic chromosomal rearrangements. These will require dedicated efforts to resolve either individually case by case or to develop new, more comprehensive, routine tests, including RNA and protein analysis. Fortunately, there are other methods for detection of point mutations that can be compared to, or used in addition to, the method presented here, including PTT[15
], and DOVAM-S[19
]. More information on DGGE analysis of the DMD can be found in the Leiden Muscular Dystrophy web site http://www.dmd.nl/
Clinical laboratories planning to begin testing for point as well as large mutations must clearly evaluate possible technologies in at least three areas: effectiveness, convenience and cost.
DHPLC followed by sequencing improves the effectiveness of mutation detection from 65%, using the existing multiplexed PCR technology that detects large mutations only, to approximately 92% by including detection of approximately 75% of point mutations as well. Clinical laboratories that are planning to screen cohorts of patients using the technique presented here will produce three important outcomes. The first will be a more accurate measure of the effectiveness of DHPLC screening for sequence variation followed by direct sequencing. The second will be improvements in the conditions for mutation detection using DHPLC. As new mutations are discovered, they could be entered into the database along with any suggested improvements to the DHPLC conditions for a given fragment/alteration. The third will be a collection of DNA, RNA and tissue from patients for whom all 86 fragments were sequenced with no likely disease causing mutations detected. This will prove extremely useful for further investigations into the more subtle causes of dystrophin-deficient muscular dystrophy. Ultimately, this will provide procedures for the detection of mutations in dystrophin-absent patients that will be more comprehensive.
The convenience of DHPLC followed by sequencing is readily apparent. It requires neither radioisotopes nor ethidium bromide gels when combined with a core sequencing facility or capillary sequencer (See Additional file B
We found that the cost of reagents for DHPLC screening followed by direct sequencing was only moderately higher than the cost for the existing multiplexed PCR test. The reagent cost of the existing multiplexed PCR diagnostic is estimated at $25.00 per patient. Increasing the percentage of mutations detected in patients from the current 65% to approximately 92% by including point mutations would come at the moderate increase in reagent costs of approximately $57.75 per patient plus a moderate increase in other costs such as consumables and technician time. Although the initial investment in a DHPLC system is not minimal at approximately $80,000, the cost can be amortized over many patients and the DHPLC system can be used in the molecular diagnosis of many other diseases in addition to DMD. The same, of course, is true for the purchase of an automated sequencer. Details of the reagent costs per patient calculations are attached as Additional file A
. Briefly, we assumed that likely disease-causative mutations would be found, on average, within approximately 43 fragments. We calculated the cost of reagents per 50μl PCR and the cost of reagents to run a sample on the DHPLC system and multiplied by the number of samples required to screen a patient. We estimated that there will be four fragments per patient that require sequencing. We calculated the reagent costs to amplify, purify and sequence these four strands per patient in both directions. We then combined the cost of the existing multiplex test for 100% of patients with the cost of DHPLC screening followed by direct sequencing for 35% of patients to arrive at an average cost per patient and the increase over the average cost per patient for the existing multiplexed PCR test alone.