INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS AND DISCUSSION REFERENCES

Cystic fibrosis (CF) is caused by defects in the cystic fibrosis transmembrane conductance regulator (CFTR) gene, which encodes a cyclic adenosine monophosphate (cAMP)-regulated chloride channel (1). CF is characterized by chronic pulmonary disease, pancreatic exocrine insufficiency (PI), male infertility, and increased concentrations of electrolytes in sweat (1). However, there is wide variability in its mode of presentation, organ involvement, and severity (1). The CFTR gene is expressed in utero (2), and some of the clinical symptoms of CF may present already at birth. Almost all males with CF are infertile because of congenital bilateral absence of the vas deferens (CBAVD). Approximately 15% of CF patients are born with meconium ileus, and in some cases this may be evident in utero (1). PI may also present at birth or develop during the first weeks of life.

Genotype-phenotype studies have shown that genetic factors can affect the degree of disease expression in CF. There are mutations associated with PI and a severe form of the disease; other mutations are associated with pancreatic sufficiency (PS) and a milder form of the disease. Some mutations are associated with variable clinical presentation in all clinical parameters (5). It has been shown that different mutations in CF can be classified on the basis of defects in protein production and function (1, 5). Class V mutations affect the levels of messenger RNA (mRNA) transcripts and protein. This class of mutations includes mutations that affect correct splicing of the CFTR gene before mRNA transcription, by exon skipping or inclusion of cryptic exon(s). One example of a class V mutation, the 5T mutation, can lead to the skipping of exon 9 (6- 8). In males carrying the 5T mutation who have CBAVD but normal lung function, higher levels of normal CFTR transcripts were found in nasal epithelia than in epididymal samples, suggesting an inverse correlation between the level of normal CFTR transcripts and disease severity in these organs (6).

Class V includes another mutation, 3849 + 10 kb C T, which is generally associated with a milder CF phenotype, characterized by PS in more than 60% of affected patients, as well as by intermediate or normal sweat chloride levels, a later age of diagnosis, and occasionally male fertility (9, 10). Despite the tendency toward an overall milder phenotype, there is a marked variability in the severity of disease presentation among patients and among different organs of the same patient carrying the 3849 + 10 kb C T mutation. This mutation creates a partially active splice site in intron 19 of the CFTR gene, which can lead to the insertion of a new 84-bp "exon," harboring an in-frame stop codon, between exons 19 and 20 (9). The aim of the present study was to investigate whether there is variability in the level of correctly and aberrantly spliced CFTR mRNA among different fetal organs carrying the 3849 + 10 kb C T mutation, and whether this variability is associated with histopathologic expression of the disease.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS AND DISCUSSION REFERENCES

Fetal samples

After parental decision for elective termination of the pregnancy, tissues from an aborted fetus with the 3849 + 10 kb C T mutation in the CFTR gene were immediately received, dissected, and processed for examination. Histologic inspections and RNA analysis were performed on lung, trachea, colon, ileum, pancreas, and heart tissues of the 22-wk CF fetus. The tissues were obtained with approval of the Human Research Committee of Brigham and Women's Hospital (Boston, MA).

RNA Extraction and Single-Strand cDNA Synthesis

RNA was extracted from the fetal tissues through the guanidinium thiocyanate method and was purified by centrifugation through a CsCl cushion (11). Complementary DNA (cDNA) was synthesized according to a previously described method (6).

Nondifferential Polymerase Chain Reactions

CFTR cDNA was amplified through the polymerase chain reaction (PCR), using recombinant Taq DNA polymerase (Boehringer Mannheim). A nondifferential PCR reaction was designed in which the two reverse transcription (RT)-PCR products (from the correctly and aberrantly spliced transcripts) were of the same size (524 bp and 526 bp, respectively) (Figure 1A). The oligonucleotide primers used in the nondifferential PCR reactions permitted amplification of the cDNA region between exon 17b and either the 19/20 junction or the 19/84-bp junction, consisting of 17b 5'-GGACGGCAGCCTTACT TTGAA- 3' in both reactions, and p19/20 5'-AAGAGGCCCACCCTCTGGC-3' for the amplification of the normal transcript or p19/84 5'-GATGACAAGTCAACCTC TGGC-3' for the amplification of the aberrantly spliced transcript. The underlined nucleotides are common to both primers. Reactions in which RNA was not added were used as controls. The cDNA samples were heated at 94° C for 3 min and then subjected to 35 cycles of denaturation at 94° C for 1 min, 57° C for 30 s, and 65° C for 1 min, followed by final extension for 7 min at 65° C. Amplification of human -actin was done with rat primers having the following base sequences (which are homologous to the human sequences): 822(+) 5'-GAAACAACATACAATTCCATCATGAAGTGTGAC-3' and 996(-) 5'-AGGAGCGATAATCTTGATCTTCATGGTGCT-3'.

View larger version (10K): [in this window] [in a new window]   View larger version (9K): [in this window] [in a new window]   Figure 1.   Analysis of correctly and aberrantly spliced CFTR transcripts with nondifferential RT-PCR. (A) Correctly and aberrantly spliced transcripts were amplified with a common primer in exon 17b and primers either in the junction between exons 19/20 (for the amplification of correctly spliced transcripts) or in the junction between exons 19/"84 bp" (for the amplification of aberrantly spliced transcripts). The RT-PCR products were of 524 bp or 526 bp, respectively. (B) Calibration of semiquantitative, nondifferential RT-PCR conditions. RT-PCR was performed on a series of 1:3 dilutions of RNA from a control sample in each RT-PCR experiment. Only results from reactions in the linear phase were included. In the present example, the linear range was 90-830 ng RNA.

Differential PCR

A differential PCR reaction was designed in which the RT-PCR products from the correctly and aberrantly spliced transcripts were of 402 bp and 486 bp, respectively. The oligonucleotide primers used in the differential PCR reactions were C1-1D and X21A (12). The cDNA samples were heated at 94° C for 3 min and then subjected to 35 cycles of PCR at 94° C for 1 min, 50° C for 30 s, and 65° C for 1 min, followed by final extension for 7 min at 65° C.

Semiquantitative RT-PCR

Semiquantitative PCR conditions were established by RT of 2.5 µg of RNA from the lung of the embryo. PCR was then performed on each of six serial tertiary dilutions of the initial RT product (see the previous discussion). The relative intensities of the RT-PCR products for each dilution were measured with a Phosphorimager (Molecular Dynamics, Sunnyvale, CA). The range for which the results reflected the initial relative amounts of normal and aberrantly spliced transcripts was determined. Only results in the linear quantitative range were considered in any of the PCR systems (Figure 1B).

Hybridization to RT-PCR Products

Electrophoresis was done on 40 µl of each RT-PCR reaction on 1% agarose gels, with subsequent blotting. Hybridizations were done with exon 19-specific oligonucleotide primers that identified both the normally and aberrantly spliced transcripts; C1-1D (12) for the nondifferential RT-PCR and F2R (13) for the differential RT-PCR. The filters were hybridized at 37° C in buffer containing 50% formamide. The blots were washed twice in 5× standard saline citrate (SSC) at room temperature for 15 min each, followed by two washes in 2× SSC at 42° C (for the nondifferential RT-PCR) and at 54° C (for the differential RT-PCR) for 15 min each. The relative intensities of the PCR products were measured with a Phosphorimager.

	RESULTS AND DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS AND DISCUSSION REFERENCES

Routine second trimester ultrasonography revealed an echogenic bowel in the fetus under study. CFTR mutation analysis of the parents revealed that the mother carried the 3849 + 10 kb C T mutation and the father carried the F508 mutation. The F508 allele leads to nonfunctional CFTR protein and is associated with severe CF (1). Analysis of amniotic fetal cells showed that the fetus was compound heterozygous for the F508 and the 3849 + 10 kb C T mutations. The parents elected to terminate the pregnancy. Gross inspection analysis of organs usually affected in CF showed that the distal small bowel (ileum) had markedly dilated, thin-walled segments alternating with regions of narrow diameter and thick walls. The lumen was filled with abnormally thick, sticky, dark green meconium (Figure 2). The colon was also filled with meconium. The lungs, trachea, spleen, heart, liver, kidneys, adrenal glands, and stomach appeared normal. Histologic analysis showed that the ileum was dilated and filled with proteinaceous material. The external diameter of the small bowel was abnormally large and the villi were flattened, suggesting distension resulting from an obstruction. The colon, lung, liver, heart, spleen, kidney, adrenal, stomach, and gall bladder had normal architecture when examined light-microscopically.

View larger version (126K): [in this window] [in a new window]   Figure 2.   Sections through the ileum of 22-wk fetuses. Left: normal fetus, right: the CF fetus in the present study.

The 3849 + 10 kb C T mutation is generally associated with a mild CF phenotype (9, 10), including a high rate of PS, absence of meconium ileus, and a better nutritional status. This is the first report of severe intrauterine gastrointestinal involvement associated with the 3849 + 10 kb C T mutation, presenting as meconium ileus with severe histopathologic abnormalities of the ileum. Thus, the nature of gastrointestinal disease might vary among patients carrying the 3849 + 10 kb C T mutation.

The 3849 + 10 kb C T allele was inherited by the fetus from his Ashkenazi Jewish mother. We have previously shown that all 3849 + 10 kb C T alleles of Jewish origin carry the same extended haplotype and do not carry an additional CFTR mutation (14). Thus, correctly spliced RNA transcribed from the 3849 + 10 kb C T allele carried by Jewish individuals is normal CFTR mRNA that is expected to produce a functional CFTR protein. The studied fetus was compound heterozygous, with 3849 + 10 kb C T/F508, and normal transcripts producing functional CFTR protein could only have been those that were transcribed from the 3849 + 10 kb C T allele and correctly spliced. The F508 allele produces correctly spliced RNA, although it leads to a nonfunctional protein. If we assume that the 3849 + 10 kb C T and F508 alleles are transcribed at equal rates, then the amount of normal CFTR mRNA expected to yield a functional protein (i.e., the correctly spliced mRNA from the 3849 + 10 kb C T allele) is equal to 50% of the total CFTR mRNA minus the aberrantly spliced RNA transcribed from the 3849 + 10 kb C T allele. However, we could not exclude the possibility that F508 and the 3849 + 10 kb C T mutation are transcribed at different efficiencies with different promoters. In addition, the relative stabilities of the F508 and the 3849 + 10 kb C T transcripts were measured in the embryo lung sample by RT- PCR, and were found to be similar (data not shown).

The level of normal RNA transcribed from the 3849 + 10 kb C T allele (as a percent of the total CFTR mRNA) was analyzed through semiquantitative, nondifferential RT-PCR (Figure 1). To verify the specificity of the primers (p19/84 and p19/20 for the aberrantly and correctly spliced RNAs, respectively), we amplified samples from three individuals with CF who do not carry the 3849 + 10 kb C T mutation. Following hybridization with a specific probe for exon 19, C1-1D, RT- PCR products were detected only in reactions run with the p19/20 primer (data not shown). An alternative system, differential RT-PCR, was used to confirm that the partial differences between the p19/20 and p19/84 primers did not introduce a bias. In this reaction the PCR products differed only by 84 bp (20%), and therefore no preferential amplification was expected. The levels of aberrantly spliced RNA (as a percent of total CFTR mRNA) transcribed from the 3849 + 10 kb C T allele in four individuals compound heterozygous for the 3849 + 10 kb C T mutation and another CF mutation were 0%, 9%, 22%, and 46% in the nondifferential RT-PCR system, as compared with 0%, 8.5%, 23%, and 48%, respectively, with the differential RT-PCR system. Thus, our nondifferential RT-PCR system is a suitable system for analyzing the levels of correctly and aberrantly spliced RNA transcribed from the 3849 + 10 kb C T allele. Six different organs of the fetus were analyzed: lung, trachea, pancreas, ileum, colon, and heart (for each organ the analysis was repeated at least twice). The analysis revealed considerable variability among the different tissues in the levels of correctly and aberrantly spliced CFTR mRNA (Figure 3). As expected, neither normal nor aberrantly spliced CFTR transcripts could be detected in the heart. A control PCR reaction for the -actin gene indicated that the heart sample had RNA in both PCR reactions (data not shown). Among the other organs expected to transcribe the CFTR gene, the levels of normal CFTR mRNA (as a percent of the total) were variable, ranging from 1% to 26% (Table 1 and Figure 4). These results show for the first time that there is variability in the level of correctly spliced CFTR mRNA transcribed from the 3849 + 10 kb C T mutation among different organs of the same individual, and that this variability already exists in utero. It is important to note that our analysis was done on organs of a 22-wk fetus. Alternative splicing is a dynamically regulated mechanism than might change during the developmental period and might therefore lead to changes in the relative levels of aberrantly spliced transcripts. Furthermore, our results show that variability in the level of correctly spliced CFTR mRNA includes the gastrointestinal system, in addition to the variability between the respiratory and genital systems, as previously identified (6).

View larger version (64K): [in this window] [in a new window]   Figure 3.   Examples of different levels of correctly and aberrantly spliced CFTR mRNA in different organs of the studied fetus. Correctly (upper panel ) and aberrantly (with the 84-bp exon, lower panel ) spliced RNA, from the same filter, transcribed in different organs of the fetus. Products of RT-PCR were hybridized to C1-1D. Correctly spliced RT-PCR products were transcribed from both CFTR alleles, whereas aberrantly spliced products were transcribed from the 3849 + 10 kb C T allele only.

View larger version (16K): [in this window] [in a new window]   Figure 4.   Levels (in percent) of normal and mutant CFTR transcripts in different organs of the fetus. Normal CFTR mRNA (solid bars) refers to correctly spliced RNA transcribed from the 3849 + 10 kb C T allele, whereas mutant mRNA (open bars) refers to aberrantly spliced RNA transcribed from the 3849 + 10 kb C T allele and RNA transcribed from the other CF allele (F508).

We further investigated the relevance of this variability to disease expression. Our RNA analysis showed that only 1% of the total CFTR mRNA in the ileum was normal. The other 99% of CFTR transcripts were defective owing either to the F508 mutation or to the additional cryptic exon in aberrantly spliced 3849 + 10 kb C T transcripts. Thus, the severe phenotype of ileal disease, which is a consequence of primary intestinal involvement by CF (15, 16), is associated with low levels of normal CFTR mRNA in utero.

All CF patients born with meconium ileus develop PI. Thus, it is reasonable to assume that the fetus that we studied would have developed PI. Unfortunately, no histopathologic analysis of the pancreas was available to us, and as a result, the significance of the very low levels (2%) of normal CFTR mRNA in the pancreas is not clear, although it might indicate that the pancreas of the fetus was affected. The colon had a relatively higher level of normal CFTR mRNA (19%), and histopathologic analysis showed that the colon was normal.

In CF, respiratory disease usually does not present during the neonatal period (1). Thus, normal histopathologic findings were expected for the fetal respiratory tissues (lung and trachea). Barker and colleagues (17) measured the transepithelial potential difference in monolayer preparations derived from the lung acini of the same fetus used in our study. Their analysis revealed cAMP-sensitive Cl secretory responses similar to those seen in normal fetal lung, suggesting that the CFTR function in the lung of this fetus was normal. In accordance with these results, higher levels of normal CFTR mRNA were identified in the respiratory tissues of the fetus (17% in the trachea and 26% in the lung).

The CF-affected organs analyzed in our study are usually not available for study, either from patients or from fetuses. Therefore, although we investigated only one fetus with CF, the results are important for understanding the molecular basis for disease of differing severity in different organs of the same patient carrying the 3849 + 10 kb C T mutation. Overall, the levels of normal CFTR mRNA detected in the different fetal organs appeared to correlate with the histopathologic observations on these organs (Table 1). Those organs (colon, trachea, lung) that had no histopathologic abnormalities had higher levels of normal CFTR mRNA, whereas the severely affected ileum had a very low level of normal CFTR mRNA. This is in accordance with the inverse correlation found between disease severity in cells from the respiratory and the male genital systems and the level of RNA in these systems (6). However, in the present study, we show that the low level of normal CFTR mRNA correlates with severe gastrointestinal disease already in utero.

Alternative splicing is a complex regulatory mechanism involving cis- and trans-acting factors. The variable levels of correctly spliced RNA transcribed from the 3849 + 10 kb C T allele in different organs of the same fetus are obviously not associated with variability in cis elements. Therefore, the variability might be associated with trans-acting factors involved in regulation of the alternative splicing mechanism. Variability among different tissues in the relative amounts of alternative splicing factors (18, 19) and/or tissue-specific alternative splicing factors probably contributes to the different patterns of alternative splicing found among different organs of the same individual. Other nongenetic factors, such as viral infections or environmental factors, might also be involved. We cannot, however, exclude the possibility that the different levels of correctly and aberrantly spliced CFTR transcripts among the different organs in cases of CF result from differences in the rates of degradation of CFTR mRNA.

In summary, alternative splicing regulation might contribute to the variability in organ involvement and expression of CF and a broad range of other inherited human diseases caused by splicing mutations. Further understanding of the mechanisms regulating alternative splicing will contribute to potential treatment for patients carrying such splicing mutations.