Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
Autosomal dominant polycystic kidney disease is a common disorder, primarily characterised by progressive renal enlargement and cyst formation, leading, in the majority of patients, to end stage renal disease. Cyst formation has also been observed in liver and pancreas. Cerebral artery aneurysm has also been reported.1
This disorder is genetically heterogeneous, but about 85% of cases are the result of a mutation in the PKD1 gene, located on the short arm of chromosome 16 (16p13.3). This gene spans about 50 kb of genomic DNA, is composed of 46 exons, and the ORF is 13 kb.2–5
The incidence of PKD1 in the population is about 1/1000 subjects, making it the most common genetic disorder in humans.6 However, routine clinical diagnosis by molecular analysis has not been extensively developed. This is because of two main characteristics of the gene. Firstly, the coding region is large and spans 46 exons. The second and main reason is that sequences homologous to 70% of the coding region are present as nearly identical transcribed copies located closer to the centromere on the short arm of chromosome 16.7,8 Some of these sequences are transcribed.2 This situation has led, until recently, to identification of mutations in only 30% of the non-repeated part of the coding region. Previously, authors have reported RT-PCR based mutation detection, but they only screened the PKD1 gene from exon 16 to exon 46. Recently, two papers have reported different screening strategies for mutation detection in the repeated region of PKD1.9–11 One involves a long range RT-PCR approach (PKD1 is expressed in lymphocytes) designed to amplify a 13 kb RT-PCR product.12 This can probably not be used as a routine screening test in a hospital setting. The second approach involves PTT tests and sequencing of genomic DNA.13 This approach is more practical and has met with some success, although the detection of mutations is mainly based on protein truncation tests, which allows the detection of premature protein translation.
The molecular diagnosis of this disorder is, however, important not only for patients but also for mutation based selection of patients for therapeutic trials in the future. We therefore designed primers from database sequences, spanning regions in which we detected nucleotide differences between PKD1 and its homologues. We used these primers to amplify RT-PCR products, ranging from 750 bp to 2.3 kb, a fairly normal range, and have shown that they are PKD1 specific. We used this test to screen for mutations in nine polycystic patients, with a clinical presentation of PKD1, with or without a family history. Surprisingly, we found 17 polymorphisms and six mutations. In conclusion, we have developed a strategy for mutation screening over the entire PKD1 coding region by a quick and reliable method.
MATERIAL AND METHODS
This study comprised nine patients with autosomal dominant polycystic kidney disease (ADPKD). Subjects gave informed consent and the hospital ethics committee approved the study. All patients, except patient 4, had a familial history of ADPKD. The patients under 30 years old all had at least two cysts in one kidney or one cyst in each kidney, those aged between 30 and 59 had at least two cysts in each kidney, while those aged 60 or older had at least four cysts in each kidney. The parents of patient 4 had no kidney cysts at an age greater than 60.
Eight patients (1, 2, 3, 4, 5, 6, 7, and 11) had end stage renal disease. Seven (1, 2, 3, 4, 5, 6, and 7) were undergoing haemodialysis treatment and patient 9 had received a transplant. Patient 8 had normal renal function. Families 1, 6, 7, and 8 had a history of cerebral aneurysms.
The polycystic kidney disease segregates over at least three generations in eight of the families. Patient 4, however, had no familial history and is a sporadic case. However, this patient fulfilled all the PKD1 criteria and his clinical presentation was identical to that of the familial cases. The presence of another nephropathy segregating in the families was excluded.
Seven of the families are of European origin, family 5 is from North Africa, and family 6 is from the Middle East. No linkage analysis was performed because some members of the families refused to contribute to the study.
RT-PCR analysis of PKD1
Using Blast2 software, we compared the sequence of PKD1 (accession number L33243) with the sequence of a BAC containing PKD1 homologues (accession number AC002039) and detected the regions with the lowest nucleotide identity. We used these data to design a set of nine primer pairs specific for the PKD1 transcript, which encompass the whole coding region. A summary of these primers is presented in table 1.
Lymphocytes were isolated using a Ficoll gradient and total RNA was extracted using the Rneasy Quiagen kit, according to the manufacturer's instructions (Quiagen SA). Total RNA was resuspended in Rnase free water. A total of 1-5 μg of total RNA was reverse transcribed using a RT-PCR kit (RT Life Technology GIBCO BRL), according to the manufacturer's instructions, and 3 μl of the RT products were amplified using a standard hot start PCR technique using the primers; the PCR temperature is shown in table 1. MgCl2 concentration in PCR reactions was 1.5 mmol and 35 cycles of amplification were performed.
RT-PCR products, synthesised as previously described, were diluted to 1/1000 and subjected to a nested PCR, using internal primers derived from the PKD1 sequence, in order to generate fragments of about 200-300 bp (table 2), except for PCR products corresponding to fragments 4 and 8, which were directly sequenced. A total of 40 μl of PCR product was mixed with 40 μl of 0.1 mol/l NaOH and transferred to ice. This mixture was deposited at the top of a 30% acrylamide gel and subjected to electrophoresis (16 hours at 10 mA at 18°C). If necessary, a second analysis was performed at 4°C. Following migration, the gel was stained with ethidium bromide and fragments were visualised by UV. Fragments (PCR products) showing an abnormal migration compared to standard DNA from unaffected subjects were sequenced on both strands.
Sequences were analysed and compared to the PKD1 gene sequence (accession number L33243) and homologous sequences (accession numbers AC002039, AC009065,8 AC012171,8 NT_024838, and NT_010543), using the basic Blast software (http://www.ncbi.nlm.nih.gov/blast/).
Designing an RT-PCR assay specific for the PKD1 transcript
Using database sequences (see Methods), we compared the sequence of PKD1 to sequences of PKD1 homologous transcripts.7 The comparison of these sequences was used to design a set of nine oligonucleotide primer pairs (table 1), for which at least the 3` end is specific for PKD1. The distribution of these primers is shown in fig 1 and covers the entire coding sequence.
We then extracted RNA from normal lymphocytes and reverse transcribed it using a standard procedure. We amplified the cDNA using the primers described above. The amplified products were directly sequenced and compared to PKD1 and related sequences.7,8 Sequencing of DNA fragments obtained by our RT-PCR assay clearly corresponds to unique sequences, corresponding to PKD1 only and not to any of the related sequences (an example is shown in fig 2A). Moreover, we used enzymes that specifically digest the PKD1 sequence and not the related sequences, and showed that each amplified fragment is specific for PKD1. An example is presented in fig 2B.
These data clearly indicate that the RT-PCR assay described above amplifies the PKD1 transcript only and not transcripts from related sequences.
Mutation screening of PKD1 using a specific RT-PCR assay
In order to test the ability of our RT-PCR assay to identify mutations in ADPKD patients, we selected nine unrelated patients who presented the classical symptoms of the disorder. RNA was extracted from fresh blood lymphocytes and RT-PCR was performed. Before screening for mutations, we ruled out the possibility that the transcript corresponding to the mutant allele was unstable and degraded in the majority of our patients, as this would have rendered our assay ineffective for PKD1 mutation screening. In order to do this, we used known PKD1 polymorphisms and digested the RT-PCR product with the appropriate restriction enzymes. An example is given in fig 3. We can see that two alleles can be detected in the RNA of ADPKD patients. Moreover, we found polymorphisms in all tested patients (see below) except one (patient 1), indicating that the two alleles have been amplified in the RT-PCR reaction.
As a next step, we screened for mutations in RT-PCR products, using a standard SSCP approach, followed by sequencing (fig 4A, B). Surprisingly, we found 17 differences from the published sequence (table 3). These differences clearly represent polymorphisms in the repeated region, since the variation in the sequence is present in several patients, as well as in unaffected subjects, some of which have already been published. It can be seen in table 2 that 13 mutations involve the base combination of C/T (or G/A), while only four affect other base combinations (75%/25%). Moreover, of the missense mutations, seven of 17 polymorphisms (six already reported and one new one) correspond to a substitution present in at least one of the PKD1 homologues (40%).
In addition, we detected six mutations which were each present in only one patient and were not detected in unaffected subjects. One is a 4 bp insertion (patient 6), which has not been previously described (2986ins4). This mutation introduces a stop codon 300 bp downstream. The second is a splice mutation (patient 1) that causes partial deletion (74 bp) of exon 14 and introduces a stop codon 400 bp downstream, which has already been described.12 This splicing abnormality is caused by a mutation in the splice acceptor site of exon 14 and the use of a cryptic splice site in exon 14. The third is a splice site mutation that causes exon 44 to be skipped, thus deleting 45 amino acids coding for the tenth transmembrane domain.13 The sequence of the corresponding region of this patient's genomic DNA showed a T>C transition at the intron 44 donor site (fig 5). This mutation was detected in patient 4, a sporadic case of PKD with no previous familial history. The fourth mutation (patient 5) is also a splicing mutation that leads to intron 45 being unspliced, adding 30 extra amino acids to the protein. Sequencing exons 45 and 46 and intron 45 only showed a T>C transition in the intron (51385, IVS45-14, fig 6). These three splicing variants were detected by RT-PCR products of an abnormal size (for example, see fig 5). Abnormal fragments were sequenced, as well as the corresponding genomic region. The fifth mutation (patient 9) corresponds to a 3 bp deletion (9412-9415), encoding a valine (3138) which has been strongly conserved throughout evolution (human, mouse, Fugu). This valine is part of a PLAT domain, which is involved in signal transduction.21 This mutation segregates with polycystic kidney disease in this family. The sixth mutation (patient 7) is a missense mutation (Q987H) that has not been previously identified. The absence of these mutations in 100 unrelated wild type chromosomes has been verified. This mutation is likely to be pathogenic as it changes a negatively charged amino acid (glutamic acid) into a positively charged one (histidine). Moreover, the mutated glutamic acid is conserved in PKD1 from human to Fugu.
Here we describe a PKD1 mutation screening strategy based on RT-PCR amplification of the transcript, using specific primers. In order specifically to amplify the PKD1 transcript, and not any of the homologous sequences, nine sets of primers specific to the PKD1 sequences were designed on the basis of comparisons with PKD1 homologues present in the databases. In order to check the specificity of this assay, we sequenced RT-PCR products. The sequence of these fragments corresponds to a unique sequence, which is the sequence of PKD1 and not one of its six homologues. We cleaved the RT-PCR product with selected restriction enzymes, allowing the discrimination of PKD1 from related sequences and confirming that the sequence is definitely the PKD1 sequence.
Using RT-PCR, we amplified the PKD1 transcript from nine unrelated patients and checked that the two alleles could be amplified and analysed, using a published polymorphism.22 This is extremely important, as mutations can often lead to unstable transcripts that cannot be detected. Three patients out of nine show the two polymorphic alleles, indicating that the transcripts from the two alleles can be amplified. Moreover, we have detected polymorphisms by SSCP in all patients except one. This indicates that we can screen both copies, even if one is more abundant than the other (it is well known that PCR techniques normalise the quantity of transcripts when more than 35 rounds of amplification are used). We therefore think it probable that our approach can be applied to most samples from PKD1 patients.
We found six mutations that are likely to be the cause of the disorder: a 4 bp insertion, three splice mutations, a 3 bp deletion, and one missense point mutation. Only one of the splice mutations has been previously identified. Moreover, using our approach we also detected a mutation in a sporadic case. Unfortunately, we have no access to the DNA of the parents, so we cannot know if this sporadic mutation is de novo or if the case is sporadic because of the small size of the family (a situation becoming more frequent in Europe with the decrease in family size). One of the splicing mutations (IVS45-14T>C) is interesting as it did not affect a consensus splice site, but affected the sequence of the intron. The only explanation for this (we have never found this splice variant in the either normal population or other patients) is that this mutation affects the secondary structure of the mRNA, necessary for the normal formation of the lariat in the first steps of splicing. These mutations did not correspond to the sequence of PKD homologues, ruling out the possibility of some non-specificity of our test.
Therefore, six mutations were detected in the nine patients. In spite of the small size of the tested population, this rate of success (67%) is higher than other published screening methods. This success is partially the result of the large number of spliced mutations detected in the assay (3), which can accurately be detected only by RT-PCR. However, this should be confirmed in a larger population. The failure to find mutations in 30% of patients may be because: (1) some of the patients were not PKD1 patients; (2) SSCP screening does not have a 100% success rate and often gives false negative results; or (3) there was an error in the sequence of abnormal SSCP fragments. In order to increase our rate of success, we plan to use more sensitive and reliable approaches, such as mutation screening using the DHPLC technique.
In addition, we have detected numerous variations that appear to be polymorphisms, as they are present in both patients and unaffected subjects. These represent a high rate of nucleotide substitution. The most frequent base substitution in humans is C to T, which is also the case in the present study.
One explanation for this high rate of base substitution could be, as has already been proposed,23,24 that gene conversion occurs between PKD1 and related sequences, introducing mutations into PKD1. However, more than 60% of the amino acid or nucleotide changes detected in PKD1 are not present in PKD1 related transcribed sequences. Moreover, we may note that no putative mutations observed in our patients match the PKD related sequences. The gene conversion hypothesis cannot, therefore, explain the majority of sequence variants in the PKD1 transcript. However, this observation supports our data, indicating that our technique specifically amplifies PKD1 and not the related sequences. These findings also confirm the frequency of polymorphisms of PKD1 previously described.13,15 This surprisingly high rate of mutations/polymorphisms has so far not been elucidated. However, a high rate of mutations is consistent with the two hit strategy of cyst formation, as this theory implies a high rate of somatic mutations in order to have the chance to form a cyst.
In conclusion, we have developed a quick and reliable RT-PCR approach that specifically amplifies the whole PKD1 transcript, allowing the screening of all types of mutations over the entire PKD1 gene.
We thank M Mitchell for correction of the English in the manuscript (and sometimes rewriting). SB is supported by an ADEREM fellowship, Société Française de Néphrologie fellowship, and poste d'accueil INSERM. This work was supported by an INSERM/MERCK grant.