Gene therapy for type 1 diabetes. New approaches.

Type 1 or insulin-dependent diabetes mellitus is a metabolic disease that results from the destruction of pancreatic beta-cells. Most patients receive subcutaneous insulin injections to reduce blood glucose levels. However, strict glucose control by multiple insulin injections is associated with an increased risk of hypoglycemia and weight gain, while a less strict glucose control is insufficient to prevent chronic complications such as nephropathy, neuropathy and retinopathy. Gene therapy holds a tremendous therapeutic potential to improve glycemic control by restoring endogenous insulin production. This review focuses on recent advances directed at producing insulin in an ectopic tissue as well as inducing pancreatic beta-cell neogenesis. The strategies include constitutive and promoter-regulated insulin expression in the liver; increasing hepatic glucose oxidation; insulin production from intestinal cells; and islet cell neogenesis in liver and pancreas.

Challenges for gene therapy of type 1 diabetes.

Type 1 or insulin-dependent diabetes mellitus is caused by autoimmune attack and selective destruction of the pancreatic beta cells. Despite the development of various insulin replacement therapies, insulin injection still remains the mainstay treatment for type 1 diabetes. However, exogenous insulin administration cannot achieve the same degree of glycemic control as provided by endogenous insulin produced from the pancreatic beta cells. Insulin gene transfer is being developed to improve the quality of glycemic control by restoring endogenous insulin production in type 1 diabetes. Nevertheless, attempts to achieve adequately regulated insulin production are stymied by the lack of appropriate surrogate cells that are able to detect blood glucose variations and release insulin in a glucose-dependent manner. Although limited success has been made to control insulin gene expression in ectopic cells using hormone/glucose-regulated expression systems, these transcriptionally regulated systems are relatively slow in the "on-" and "off"-kinetics of insulin production, raising a serious safety concern for clinical application. In this article, we will review recent advances made to address this concern and highlight the importance of insulin gene transfer to cell types that possess an intrinsic ability to kinetically mimic the pancreatic beta cells in terms of glucose-responsive insulin secretion.

Hepatic insulin expression improves glycemic control in type 1 diabetic rats.

Low levels of hepatic insulin production have been shown to prevent lethal ketoacidosis associated with type 1 diabetes. To assess the beneficial effects of sustained hepatic production of insulin on glycemic control in type 1 diabetes, we have employed the adenovirus-mediated gene delivery system to transfer an engineered rat preproinsulin gene to the livers of streptozotocin-induced diabetic nude rats. Hepatic insulin production resulted in the reduction of blood glucose in treated diabetic rats, the degree of blood glucose reduction correlated with both the vector dose and the level of hepatic insulin expression. At moderate vector doses, 0.3-0.7 ng/ml of plasma insulin was produced in treated diabetic animals, resulting in significant reduction of nonfasting hyperglycemia and improvement in glucose tolerance. Furthermore, these animals maintained euglycemia after 12-h fast. At higher vector doses, greater than 1 ng/ml of plasma insulin was produced, completely reversing nonfasting hyperglycemia in treated rats. However, all of the treated animals developed severe hypoglycemia upon fasting. This study has defined the maximal tolerable level of hepatic insulin production that is sufficient to reduce the degree and ameliorate the adverse effects of nonfasting hyperglycemia without risk of fasting hypoglycemia in type 1 diabetic rats.

Toxicity associated with repeated administration of first-generation adenovirus vectors does not occur with a helper-dependent vector.

BACKGROUND: Certain gene therapy protocols may require multiple administrations of vectors to achieve therapeutic benefit to the patient. This may be especially relevant for vectors such as adenoviral vectors that do not integrate into the host chromosome. Because immunocompetent animal models used for gene transfer studies develop neutralizing antibodies to adenoviral vectors after a single administration, little is known about how repeat administrations of vectors might affect transgene expression and vector toxicity. MATERIALS AND METHODS: We used mice deficient in the membrane spanning region of immunoglobulin (IgM), which do not develop antibodies, to evaluate the effect of repeated intravenous administration of first-generation and helper-dependent adenoviral vectors expressing human alpha 1-antitrypsin (hAAT). The duration and levels of transgene expression were evaluated after repeated administration of vectors. Toxicity was assessed by measuring the level of liver enzymes in the serum and the degrees of hepatocyte hypertrophy and proliferation. RESULTS: We found that previous administration of first-generation adenoviral vectors can alter the response to subsequent doses. These alterations included an increase in transgene expression early (within 1 and 3 days), followed by a rapid drop in expression by day 7. In addition, previous administrations of first-generation vectors led to an increase in toxicity of subsequent doses, as indicated by a rise in liver enzymes and an increase in hepatocyte proliferation. In contrast to first-generation vectors, use of the helper-dependent adenovirus vector, Ad-STK109, which contained no viral coding regions, did not lead to increased toxicity after multiple administrations. CONCLUSIONS: We conclude that the response of the host to adenoviral vectors can be altered after repeated administration, compared with the response after the initial vector dose. In addition, these experiments provide further evidence for the relative safety of helper-dependent adenoviral vectors for gene therapy, compared with first-generation vectors.

Administration of helper-dependent adenoviral vectors and sequential delivery of different vector serotype for long-term liver-directed gene transfer in baboons.

The efficiency of first-generation adenoviral vectors as gene delivery tools is often limited by the short duration of transgene expression, which can be related to immune responses and to toxic effects of viral proteins. In addition, readministration is usually ineffective unless the animals are immunocompromised or a different adenovirus serotype is used. Recently, adenoviral vectors devoid of all viral coding sequences (helper-dependent or gutless vectors) have been developed to avoid expression of viral proteins. In mice, liver-directed gene transfer with AdSTK109, a helper-dependent adenoviral (Ad) vector containing the human alpha(1)-antitrypsin (hAAT) gene, resulted in sustained expression for longer than 10 months with negligible toxicity to the liver. In the present report, we have examined the duration of expression of AdSTK109 in the liver of baboons and compared it to first-generation vectors expressing hAAT. Transgene expression was limited to approximately 3-5 months with the first-generation vectors. In contrast, administration of AdSTK109 resulted in transgene expression for longer than a year in two of three baboons. We have also investigated the feasibility of circumventing the humoral response to the virus by sequential administration of vectors of different serotypes. We found that the ineffectiveness of readministration due to the humoral response to an Ad5 first-generation vector was overcome by use of an Ad2-based vector expressing hAAT. These data suggest that long-term expression of transgenes should be possible by combining the reduced immunogenicity and toxicity of helper-dependent vectors with sequential delivery of vectors of different serotypes.

Use of a liver-specific promoter reduces immune response to the transgene in adenoviral vectors.

Previous studies using adenoviral (Ad) vectors expressing human alpha1-antitrypsin (hAAT) under the control of ubiquitous promoters (RSV, mPGK) elicited the production of antibodies to hAAT in some mouse strains (C3H/HeJ and BALB/c) but not in others (C57BL/6J). In contrast, when a helper-dependent Ad vector (AdSTK109) with all viral coding sequences deleted and expressing hAAT from human genomic DNA with the endogenous promoter was used, C3H/HeJ mice failed to develop antibodies and demonstrated long-term expression. These results suggested that promoter choice and/or properties of the vector itself might influence the host immune response to the transgene product. Direct comparison of first-generation vectors expressing the hAAT cDNA from a ubiquitous mouse PGK promoter rather than from a liver-specific mouse albumin promoter demonstrated that an antibody response to hAAT occurred with the mPGK promoter but not with the albumin promoter in C3H/HeJ mice. As expected, neither vector elicits an antibody response in C57BL/6J mice. Coinjection of the two first-generation vectors containing the mPGK and albumin promoter in C3H/HeJ mice induced an antibody response with resulting loss of detectable hAAT from the sera of the injected mice in 3-4 weeks. From these data, we conclude that under certain conditions, the choice of promoter with its associated liver-specific expression can modulate the host immune response to the transgene independent of viral backbone.

Toxicological comparison of E2a-deleted and first-generation adenoviral vectors expressing alpha1-antitrypsin after systemic delivery.

Second-generation adenoviral vectors, mutated in E2a, have been proposed to decrease host immune responses against transduced cells, reduce toxicity, and increase duration of expression as compared with first-generation vectors deleted only in E1. To test these hypotheses further, we have developed an E2a-deleted adenoviral vector expressing human alpha1-antitrypsin (hAAT). Toxicity of first-generation and E2a-deleted vectors, as determined by hematological indices, liver function tests, and histological analyses, was evaluated in C3H mice for 21 days after vector administration at increasing doses starting at 1 x 10(12) particles/kg. Both vectors induced dose-dependent abnormalities including transient thrombocytopenia, elevated ALT levels in serum, and increased hepatocyte proliferation followed by inflammation and then hypertrophy. Differences in the ratio of particles to plaque-forming units among vector preparations led to differences in hAAT expression at similar particle doses. There were no differences in toxicity between the two vectors when measured at matching levels of hAAT expression. However, the E2a-deleted vector was demonstrated to have slightly reduced hepatocyte toxicity at an intermediate particle dose. This suggests that hepatocyte toxicity is related primarily to viral entry and expression, rather than to the presence of noninfectious particles, and implies that vectors with complete elimination of viral gene expression, such as vectors with all viral coding sequences deleted, are likely to have substantial advantages in terms of safety and toxicity.

Genomic DNA transfer with a high-capacity adenovirus vector results in improved in vivo gene expression and decreased toxicity.

Many applications for human gene therapy would be facilitated by high levels and long duration of physiologic gene expression. Adenoviral vectors are frequently used for gene transfer because of their high cellular transduction efficiency in vitro and in vivo. Expression of viral proteins and the low capacity for foreign DNA limits the clinical application of first- and second-generation adenoviral vectors. Adenoviral vectors with all viral coding sequences deleted offer the prospect of decreased host immune responses to viral proteins, decreased cellular toxicity of viral proteins and increased capacity to accommodate large regulatory DNA regions. Currently most vectors used in vivo for preclinical and clinical studies express cDNAs under the control of heterologous eukaryotic or viral promoters. Using an adenoviral vector with all viral coding sequences deleted and containing the complete human alpha1-antitrypsin (PI) locus, we observed tissue-specific transcriptional regulation in cell culture and in vivo; intravenous injection in mice resulted in high levels of very stable expression for more than ten months and decreased acute and chronic toxicity. These results indicate significant advantages of regulated gene expression using genomic DNA for gene transfer and of adenoviral gene transfer vectors devoid of all viral coding sequences.

High doses of a helper-dependent adenoviral vector yield supraphysiological levels of alpha1-antitrypsin with negligible toxicity.

Optimal gene therapy for many disorders will require efficient transfer to cells in vivo, high-level and long-term expression, and tissue-specific regulation, all in the absence of significant toxicity or inflammatory responses. While recombinant adenoviral vectors are efficient for gene transfer to hepatocytes, their usefulness is limited by short duration of expression related, at least in part, to immune responses to viral proteins and by a low capacity for foreign DNA. A number of systems have been developed for producing adenoviral vectors devoid of all viral coding sequences. Using AdSTK109, a vector lacking all viral coding sequences and carrying the complete human alpha1-antitrypsin (hAAT) genomic DNA locus, we have demonstrated sustained expression for longer than 10 months in mice. Utilizing high doses of this vector for hepatic gene transfer in mice, we find that supraphysiological levels of hAAT can be achieved without hepatotoxicity.

Immune responses to reporter proteins and high viral dose limit duration of expression with adenoviral vectors: comparison of E2a wild type and E2a deleted vectors.

Experiments designed to evaluate the effect of deletion of E2a on duration of expression using adenoviral vectors led to a series of observations regarding host responses to adenoviral vectors and reporter proteins. In studies using human alpha1-antitrypsin (hAAT) as a reporter gene, we found that the duration of expression is very brief for C3H/J and CBA/J mice but is prolonged for C57BL/6J mice, that disappearance of hAAT from the blood is correlated with the appearance of antibodies, and that immunization against hAAT can prevent appearance of the protein in the blood after administration of an adenoviral vector. Deletion of E2a in hAAT vectors did not prolong expession in C3H/J or CBA/J mice and did not shorten duration of expression in C57BL/6J mice. Using similar vectors expressing Escherichia coli beta-galactosidase (beta-Gal) in immunocompetent mice, short duration of expression with a beta-Gal reporter was remarkably different from the long expression with an identical vector expressing hAAT in C57BL/6J. In the case of vectors expressing hAAT, adenoviral sequences persisted in the liver, and inflammatory responses were minimal compared to vectors expressing beta-Gal, where adenoviral sequences disappeared from the liver concomitant with a prominent inflammatory response. The duration of expression of beta-Gal in hepatocytes was increased in transgenic mice expressing the reporter in keratinocytes, indicating that host immune responses to the reporter can limit duration of expression. Dosage studies indicated that persistence of expression of hAAT can be markedly decreased by administration of high doses of vector in a manner consistent with a nonimmune-mediated toxicity following injection. These experiments indicate that host responses to reporter genes rather than host responses to adenoviral proteins can be the primary determinant of duration of expression under many experimental conditions.

Development of a complementing cell line and a system for construction of adenovirus vectors with E1 and E2a deleted.

Although adenovirus vectors offer many advantages, it would be desirable to develop vectors with improved expression and decreased toxicity. Toward this objective, an adenovirus vector system with deletion of both the El and E2a regions was developed. A 5.9-kb fragment of the adenovirus type 5 (Ad5) genome containing the E2a gene and its early and late promoters was transfected into 293 cells. A complementing cell line, designated 293-C2, expressed the E2a mRNA and protein and was found to complement the defect in Ad5 viruses with temperature-sensitive or deletion mutations in E2a. A deletion of 1.3 kb removing codons 40 to 471 of the 529 amino acids of E2a was introduced into plasmids for preparation of viruses and vectors. An Ad5 virus with disruption of the El gene and deletion of E2a grew on 293-C2 cells but not on 293 cells. Vectors with E1 and E2a deleted expressing Escherichia coli beta-galactosidase or human alpha1-antitrypsin were prepared and expressed the reporter genes after intravenous injection into mice. This vector system retains sequences in common between the complementing cell line and the vectors, including 3.4 kb upstream and 1.1 kb downstream of the deletion. These vectors have potential advantages of increased capacity for insertion of transgene sequences, elimination of expression of E2a, and possibly reduction in expression of other viral proteins. Although the titers of the vectors with deleted are about 10- to 30-fold below those of vectors with E2a wild-type regions, the former vectors are suitable for detailed studies with animals to evaluate the effects on host immune responses, on duration of expression, and on safety.

CFTR haplotypic variability for normal and mutant genes in cystic fibrosis families from southern France.

In order to contribute to a better understanding of the dispersion of cystic fibrosis (CF) mutations in the South of France, seven diallelic and three multiallelic markers [three upstream of the cystic fibrosis transmembrane conductance regulator (CFTR) gene (XV-2c, KM 19 and J44) and seven intragenic polymorphism (IVS6A, IVS8CA, M470V, T854T, IVS17BTA, IVS17BCA and TUB18)] were analyzed for 143 delta F508 chromosomes, 100 CF chromosomes carrying 85 non-delta F508 and 15 unknown mutations, and 198 normal CFTR alleles. The study provides haplotypic data for 39 different CF mutations, which should be useful in diagnosis by haplotypic analysis and detection of the associated mutations. A major haplotype [2-1-2-7-16-2-1-(30/31)-13-1] was found in normal chromosomes, which should be the most ancient in the Caucasoid population. The most frequent haplotypes in normal chromosomes were associated with 16 different non-delta F508 mutations, suggesting that there was no preferential haplotype on which these mutations arose. Several mutations were each associated with more than one haplotype, as the result of slippage at one or two of the three microsatellites (delta F508, G542X, N1303K, G85E, E585X, K710X and 2184delA) or recombination (1717-1G-->A, R334W, L206W, R1162X and Y122X). Haplotypes for the most common CFTR mutations (delta F508, G542X, N1303K) revealed that a large number of alleles were generated by slippage at the microsatellite loci, suggesting that they are the most ancient CF mutations. Other mutations were associated with haplotypes that were different either at several diallelic sites (R334W) or at both diallelic and microsatellite markers (R1162X and R1158X), which is more suggestive of recurrence. Twenty recombinations were detected among the CF mutant alleles analyzed, 75% of them occurring in the second half of the CFTR gene. The higher mutational heterogeneity and the haplotypic variability reported in this small population from the Mediterranean area are consistent with an earlier appearance of CFTR mutations in southern Europe than in central and northern Europe, and an earlier origin and expansion of this population.

Haplotype analysis of 94 cystic fibrosis mutations with seven polymorphic CFTR DNA markers.

We have analyzed 416 normal and 467 chromosomes carrying 94 different cystic fibrosis (CF) mutations with polymorphic genetic markers J44, IVS6aGATT, IVS8CA, T854, IVS17BTA, IVS17BCA, and TUB20. The number of mutations found with each haplotype is proportional to its frequency among normal chromosomes, suggesting that there is no preferential haplotype in which mutations arise and thus excluding possible selection for specific haplotypes. While many common mutations in the worldwide CF population showed absence of haplotype variation, indicating their recent origins, some mutations were associated with more than one haplotype. The most common CF mutations, delta F508, G542X, and N1303K, showed the highest number of slippage events at microsatellites, suggesting that they are the most ancient CF mutations. Recurrence was probably the case for 9 CF mutations (R117H, H199Y, R347YH, R347P, L558S, 2184insA, 3272-26A-->G, R1162X, and 3849 + 10kbC-->T). This analysis of 94 CF mutations should facilitate mutation screening and provides useful data for studies on population genetics of CF.

Analysis of microsatellites by direct blotting electrophoresis and chemiluminescence detection.

We describe a fast and reliable method for the nonradioactive analysis of microsatellites. For three dinucleotide repeats within the cystic fibrosis transmembrane conductance regulator (CFTR) gene, the separation of polymerase chain reaction (PCR) products generated with biotinylated primers on a direct blotting electrophoresis system and subsequent chemiluminescence detection is shown. In direct blotting electrophoresis, the separation of DNA fragments depended linearly on size. The reproducible resolution allowed reliable assignment of allele lengths to a given signal. The nonradioactive detection protocol was advantageous compared to radioactive methods: samples could be analyzed within one day due to the fast signal development by 3-(4-methoxyspiro[1,2-dioxetane-3,2'-(5'- chloro)tricyclo[3.3.1.1.3,7]decan]-4-yl)phenylphosphate disodium salt (CSPD). Variation of exposure times enabled differentiation between major bands and byproducts of comparable intensity that are due to the slippage of the Taq polymerase during PCR amplification.

Complete detection of mutations in cystic fibrosis patients of Native American origin.

An increased incidence of cystic fibrosis (CF) has been reported in some populations of Native Americans of the Southwest such as the Pueblo, which is a genetic isolate. As the most common mutation found in Caucasians (delta F508) was absent and only one chromosome carried the G542X mutation, we decided to analyze the entire coding sequence of the CFTR gene in eight Pueblo CF patients. We have identified four different mutations: G542X, R1162X, 3849+10kbC-->T, and D648V that account for these 16 haplotypes. The R1162X was found on 11 chromosomes. Using intragenic microsatellites, we have compared the haplotypes of those chromosomes to those of Italian origin where the R1162X mutation was initially reported. These haplotypes turned out to be identical, suggesting a common origin and an admixture with Italian or Spanish settlers, followed by typical founder effect. In contrast the 3849+10kbC-->T mutation, which was found on three chromosomes, is associated with different haplotypes than those on chromosomes carrying the same mutation in Caucasians. A novel mutation, D648V, observed on one chromosome has not been found outside the Pueblo population.

Independent origins of cystic fibrosis mutations R334W, R347P, R1162X, and 3849 + 10kbC-->T provide evidence of mutation recurrence in the CFTR gene.

Microsatellite analysis of chromosomes carrying particular cystic fibrosis mutations has shown different haplotypes in four cases: R334W, R347P, R1162X, and 3849 + 10kbC-->T. To investigate the possibility of recurrence of these mutations, analysis of intra- and extragenic markers flanking these mutations has been performed. Recurrence is the most plausible explanation, as it becomes necessary to postulate either double recombinations or single recombinations in conjunction with slippage at one or more microsatellite loci, to explain the combination of mutations and microsatellites if the mutations arose only once. Also in support of recurrence, mutations R334W, R347P, R1162X, and 3849 + 10kbC-->T involve CpG dinucleotides, which are known to have an increased mutation rate. Although only 15.7% of point mutations in the coding sequence of CFTR have occurred at CpG dinucleotides, approximately half of these CpG sites have mutated at least once. Specific nucleotide positions of the coding region of CFTR, distinct from CpG sequences, also seem to have a higher mutation rate, and so it is possible that the mutations observed are recurrent. G-->A transitions are the most common change found in those positions involved in more than one mutational event in CFTR.

The origin of the major cystic fibrosis mutation (delta F508) in European populations.

delta F508 is the most frequent cystic fibrosis (CF) mutation and accounts for approximately 70% of CF chromosomes worldwide. Three highly polymorphic microsatellite markers have been used to study the origin and evolution of delta F508 chromosomes in Europe. Haplotype data demonstrate that delta F508 occurred more than 52,000 years ago, in a population genetically distinct from any present European group, and spread throughout Europe in chronologically distinct expansions, which are responsible for the different frequencies of delta F508 in Europe.

Analysis of the CFTR gene confirms the high genetic heterogeneity of the Spanish population: 43 mutations account for only 78% of CF chromosomes.

We have analysed 972 unrelated Spanish cystic fibrosis patients for 70 known mutations. Analysis was performed on exons 1, 2, 3, 4, 5, 6a, 6b, 7, 10, 11, 12, 13, 14a, 14b, 15, 16, 17b, 18, 19, 20 and 21 of the cystic fibrosis transmembrane regulator gene using single strand conformation polymorphism analysis and denaturing gradient gel electrophoresis. The major mutation delta F508 accounts for 50.6% of CF chromosomes, whereas another 42 mutations account for 27.6% of CF chromosomes, with 21.8% of Spanish CF chromosomes remaining uncharacterized. At present, we have identified 36 mutations that have frequency of less than 1% and that are spread over 15 different exons. This indicates that, in the Spanish population, with the exception of delta F508 (50.6%) and G542X (8%), the mutations are not concentrated in a few exons of the gene nor are there any predominating mutations. This high degree of genetic heterogeneity is mainly a result of the different ethnic groups that have populated Spain and of the maintenance of separated population sets (Basques, Arab-Andalusian, Mediterranean, Canarian and Gallician). The high proportion of CF chromosomes still unidentified (21.8%) together with association analysis with intragenic markers suggest that at least 100 different mutations causing CF are present in our population.

Cystic fibrosis in a low-incidence population: two major mutations in Finland.

The incidence of cystic fibrosis (CF) in Finland, 1:25,000 newborn, is one of the lowest in Caucasian populations. The delta F508 mutation accounts for 18/40 (45%) of CF chromosomes in Finland. Other mutations were therefore sought among the remaining 55%. Twelve out of 40 chromosomes (30%) were found to carry 394delTT, whereas G542X and 3732delA were each detected in one chromosome. Eight mutations remained unidentified using a testing panel for 26 mutations. Mutation 394delTT was associated exclusively with haplotype 23-36-13. Five unknown mutations were associated with different haplotypes for microsatellite markers, whereas three shared the same haplotype. Most delta F508 mutations and all unidentified mutations originated from regions of old and dense settlement in the coastal regions, whereas 394delTT was geographically clustered and enriched in a rural location, consistent with a local founder effect. The remote location of Finland and her population history give a plausible explanation for the rarity of CF in Finland.

Microsatellite haplotypes for cystic fibrosis: mutation frameworks and evolutionary tracers.

Highly informative intragenic microsatellite markers within the cystic fibrosis (CF) transmembrane conductance regulator (CFTR) gene allow the analysis of associations between specific mutations and haplotypes. We have analysed 440 Spanish CF families carrying 22 different CF mutations and have established haplotypes in 1,036 chromosomes for microsatellites IVS8CA, IVS17BTA and IVS17BCA. No new alleles were detected at the three CFTR microsatellites, in more than 3,000 meiosis analysed (estimated mutation rate of less than 3.3 x 10(-4)). The evolution of 16 haplotypes associated with the most common CF mutation, delta F508, and the low mutation rate at these microsatellite loci suggest that delta F508 originated within the 23-31-13 haplotype at least 53,000 years ago, very early in the history of the European population. The number of haplotype changes seen for two other common mutations, G542X (haplotype 23-33-13) and N1303K (23-31-13), suggests that they originated at least 35,000 years ago. Microsatellite allele variability associated with delta F508, G542X and N1303K demonstrates that slippage and mispairing is the main mechanism generating microsatellite alleles. In spite of the haplotype variability detected for these 3 common mutations, the association between haplotype and mutations is very strong. Mutations 1609delCA, 3667del4, delta I507 and G551D are all associated with haplotype 16-7-17, which has a frequency of 14.5% in normal chromosomes. 5 haplotypes bearing specific CF mutations were not found in normal chromosomes. Haplotype 16-46-13 is strongly associated with CF mutations E92K and 3601-111G-->C. About 23% of CF chromosomes with unknown mutations show significant linkage disequilibrium for microsatellite haplotypes.(ABSTRACT TRUNCATED AT 250 WORDS)