Genetic variation in CFTR and modifier loci may modulate Cystic Fibrosis disease severity

Abstract

In patients with cystic fibrosis (CF), genetic variants within and outside the CFTR locus contribute to the variability of the disease severity. CFTR transcription is tightly regulated by cis-regulatory elements (CREs) that control the three-dimensional structure of the locus, chromatin accessibility and transcription factor recruitment. Variants within these CREs may contribute to the pathophysiology and to the phenotypic heterogeneity by altering CFTR transcript abundance. In addition to the CREs, variants outside the CFTR locus, namely “modifiers genes”, may also be associated with the clinical variability. This review addresses variants at the CFTR locus itself and CFTR CREs, together with the outcomes of the latest modifier gene studies with respect to the different CF phenotypes.

Background

Cystic fibrosis (CF) is the most common severe autosomal recessive genetic disease in Caucasians caused by variants in the gene encoding the cystic fibrosis transmembrane conductance regulator (CFTR), a chloride channel expressed in epithelial cells throughout the body (1). To date, over 2,000 variants in the CFTR gene have been identified world-wide. The international CF genetics research community has contributed these data to the CFTR2 database (http://www.cftr2.org), which provides information on different CFTR variants and their disease liability by a phenotype-driven approach (2). Available data include genomic and phenotypic analyses, together with molecular studies (CFTR RNA and protein expression) and sweat chloride measurements. Compiling information from more than 88,000 patients with CF, together with in vitro studies, 412 variants were annotated and 346 assigned as “CF causing”, 37 as “of varying consequences”, 21 as “non-CF causing”, and 8 “of unknown significance” (2). Missense variants are the most prevalent (40%), compared to frameshift (16%), splicing (12%) and non-sense variants (8%) (3, 4). The most common variant (70% of patients) is a three base deletion (c.1521_1523delCTT) that removes a phenylalanine residue at position 508 (F508del). Annotating these CFTR variants will benefit both clinical diagnosis and personalized therapies (5).

CF affects many organs, including the pancreas, the liver, the intestine, and most critically the lungs. Although CF is recognized as a single gene disorder, disease severity varies between mutation types and between patients with the same CFTR variant (6–9). This variability complicates treatment options and the efficacy of therapeutics. The only strong CFTR genotype-phenotype association is with pancreatic exocrine function, which is either deficient (PI for pancreatic insufficiency), or normal (PS for pancreatic sufficiency). Generally, patients carrying two CF-causing variants have a classical form of CF associated with PI, whereas others have a milder form of disease associated with PS (4). Besides environmental factors, additional genetic variants, both within and outside the CFTR locus, may also contribute to this phenotypic variability (6–8, 10). CFTR transcription is tightly regulated by cis-regulatory elements (CREs) that control the three-dimensional structure of the locus, chromatin accessibility and transcription factor recruitment (11). Variants within these CREs may contribute to pathophysiology by altering CFTR transcript abundance, as shown for those in the gene promoter (10). Utilization of unique CREs in tissue-specific or cell-typespecific contexts may be involved in the different disease manifestations and severities between affected organ systems. In addition to the CREs, variants outside the CFTR locus, in other “modifiers genes”, are also associated with the clinical heterogeneity in CF (6–9).

This review, based in part on a symposium session presented at the 16th ECFS Basic Science Conference, Dubrovnik, Croatia, 27 to 30 March, 2019, describes the results of investigating variants at the CFTR locus itself and CFTR CREs, together with the outcomes of the latest modifier gene studies with respect to the different CF phenotypes.

VARIANTS AT THE CFTR LOCUS AND CFTR REGULATION

Variants at the CFTR locus

Extensive research has focused on mutational diversity of the CFTR gene and the relationship with disease severity. Current therapeutic options are limited to treating patients with relatively few (albeit common) known variants in the coding region, whereas variants elsewhere in the locus (Figure 1) also contribute to the disease-associated alleles. Large deletions and genomic rearrangements constitute a significant proportion of rare CFTR alleles with variable disease phenotype (12, 13). In a screen for rare CF-causing mutations that are currently unidentified, several variants in 460 kb encompassing the CFTR gene mapped to known CFTR CREs (10). These include variants at the gene promoter, enhancer elements at the locus, and structural elements that may result in dysregulated CFTR transcription. Many mechanisms could account for how variants in CREs disrupt CFTR transcription, though these have yet to be proven experimentally. Single base changes may introduce transcription factor binding sites such as the c.−8G>C CFTR promoter variant, which creates a binding motif for zinc finger protein 300 (ZNF300), a known transcriptional repressor (14). Similarly, variants that destroy the binding site for architectural proteins may change locus organization and hence gene expression.

Figure 1. CFTR locus and regulatory elements.

Integrative genomics viewer (IGV) image of CFTR and adjacent genes. DNase I Hypersensitive Sites are named based on their location relative to the CFTR translational start site (upstream), corresponding intron, or translational end site (downstream). TAD boundaries at −80.1kb and +48.9kb are noted in dark red. CFTR promoter is shown in light green.

Regulation of CFTR

CFTR regulatory mechanisms differ between tissues and cell types at organ sites affected by CF (15). This diversity is evident from observations on long-range chromatin looping interactions and the location of open chromatin sites, which mark active CREs, at the locus (16). Cell-type-selective regulatory elements tightly control CFTR expression by recruiting specific transcription factors and establishing the three-dimensional structure of the expressed locus. CFTR lies within a single topologically associated domain (TAD) maintained by the architectural proteins CCCTC-binding factor (CTCF) and cohesin complex (17). Although the CFTR TAD is invariant in all cells, different CREs within it are recruited in cell-type-selective 3D interactions with the promoter. Open chromatin/CRE sites were shown by several protocols to be different in airway and intestinal epithelial cell populations (18–20). CREs within intron 1 (185 + 10kb) (K) and intron 11 (1811 + 0.8kb) were found to control CFTR expression in the intestine (21). Among activating transcription factors at these two sites are forkhead box protein A1/A2 (FOXA1/A2), hepatocyte nuclear factor 1 (HNF1) and caudal type homeobox 2 (CDX2) (22). In contrast, CREs at open chromatin peaks located 44kb and −35kb upstream of the CFTR transcriptional start site are airway-selective (23, 24).

The −35 kb site appears to be involved in immune regulation of CFTR, recruiting interferon regulatory factor 1 (IRF1) and nuclear factor Y (NF-Y). The −44 kb site may respond to oxidative stress in airway epithelial cells. In addition to activating factors CFTR expression in airway epithelial cells is regulated by a number of transcription factors acting as repressors such as Ets homologous factor (EHF) and Krüppel-like factor 5 (KLF5) (25). Determining the interactions of CREs at the locus and their associated transcription factors will enhance understanding of the CFTR regulatory landscape in CF-relevant tissues. A number of approaches are being taken to upregulate CFTR transcription for therapeutic benefit, including activation by CRISPRa (26) and steric inhibition of 5’ UTR regulatory mechanisms by antisense oligonucleotides (27).

Cell-Type Specificity of Expression

The abundance of CFTR transcripts varies between tissue types and in different cells within the same tissue (28). In the small intestine and the colon, the majority of CFTR mRNA is found in the mucosal epithelium with a decreasing gradient along the crypt to tip axis (29). Also, a rare epithelial cell type that expresses high levels of CFTR was observed in several studies. CFTR expression is very high in most ductal epithelial cells in the pancreas (30–32). As lung disease is the primary cause of mortality in CF patients, there is an ongoing effort to characterize CFTR-expressing cells in the airway epithelium. CFTR protein is reported in ciliated cells in the epithelium though levels of expression are apparently lower than in intestinal epithelial cells (33). Recent single cell RNA-seq sequencing data from mouse tracheal epithelial cells and primary human bronchial epithelial cells also identified a rare high CFTR-expressing cell type in the lung, named the pulmonary ionocytes, which may have a similar role to the high CFTR-cells in the intestinal epithelium (34, 35). Further work is underway to resolve the cellularity of CFTR expression in the human lung. Identifying the unique combinations of CREs and activating factors utilized by different CFTR expressing epithelial cells will advance understanding of the cell-specific disease mechanisms.

CYSTIC FIBROSIS MODIFIER GENES

For more than a decade, modifier genes in CF were sought using candidate gene approaches, and early studies were hampered by small sample size and poor study design, yielding mixed results (7). One of the major limitations of the candidate gene approach is that it does not identify genetic locations other than those suspected to influence the disease. That is, it does not detect modifying genes or pathways beyond those involved in our limited understanding of the disease (6). Thus, it became important to move forward by studying the whole genome. Although whole genome sequencing was still out of range in large part through costs and lack of resources, it became feasible to pursue genome-wide association studies (GWAS). A GWAS can generate information on hundreds of thousands of single nucleotide polymorphisms (SNPs) chosen to capture linkage disequilibrium (LD) structure without bias imposed by pre-existing models. Moreover, the use of imputations made it possible to explore millions of polymorphisms based on the initial GWAS results. Thus, the GWAS data identified novel genes, regulatory loci, and pathways not previously considered. However, the main disadvantage in testing so many variants is that there are statistical penalties that increase as the number of comparisons rises, and, thus, power is a major limitation (36). To overcome this constraint, collaborations were set up in North-America and France to create an international GWAS consortium. Data from these collaborative studied revealed several novel regions of the genome that are associated with numerous CF phenotypes (Figure 2).

Figure 2. Cystic fibrosis modifier genes.

Schematic representation of genes identified as associated with CF complications variability and/or treatment efficacy. Ets Homologous Factor (EHF), Apaf-1 interacting protein (APIP), mucin 4 (MUC4), mucin 20 (MUC20), solute carrier family 9 member A3 (SLC9A3), class II major histocompatibility complex (HLA Class II), angiotensin II receptor type 2 (AGTR2), solute carrier family 6 member 14 (SLC6A14), nongastric form of the H⁺/K⁺-ATPase (ATP12A), serine protease 1 (PRSS1), solute carrier family 26 member 9 (SLC26A9), transcription factor 7-like 2 (TCF7L2), ATP binding cassette subfamily A member 4 (ABCA4), glutathione S-transferase (GSTP1), and a mannose-binding lectin (MBL) haplotype, serpin family A member 1 (SERPINA1).

Lung function modifiers

Lung disease remains the major cause of morbidity and mortality in CF. CF sibling and twin studies suggested that more than 50% of the variation in lung function could be explained by modifier genes (37). To evaluate these modifiers, a quantitative lung phenotype is computed for each patient (38). In addition, this enables direct comparison of the lung function of CF patients irrespective of age and gender, thus reducing the need to stratify individuals and thereby increasing statistical power and avoiding distortions. By applying genome-wide approaches (linkage and association analyses), North-American teams initially showed in 3,444 PI-CF patients that the intergenic region between the Ets Homologous Factor (EHF) and Apaf-1 interacting protein (APIP) loci (chr11p13) was associated with lung disease severity (39). The study was further extended with additional patients from North America and France. The meta-analysis of 6,365 CF patients confirmed the previous genome-wide association with the EHF/APIP locus, and identified four new regions that contain genes with high biological relevance for the pathophysiology of CF lung disease: regions on chr3q29 (mucins 4 and 20: MUC4/MUC20), chr5p15.3 (solute carrier family 9 member A3: SLC9A3), chr6p21.3 (class II major histocompatibility complex: HLA Class II) and chrXq22-q23 (angiotensin II receptor type 2/solute carrier family 6 member 14: AGTR2/SLC6A14) (Figure 2) (40). Of note the mechanisms connecting high P-value SNPs in intergenic regions with the expression or function of nearby genes, has proven to be a major challenge.

Genes associated with meconium ileus susceptibility

Around 15% of the patients with CF are born with meconium ileus (MI), an intestinal obstruction at birth which affects only PI patients. Similarly to the approach applied for lung disease modifiers, the first GWAS in 3,763 North-American patients showed 2 loci near solute carrier genes, SLC6A14 at chXq23–24 and solute carrier family 26 member 9 (SLC26A9) at ch1q32.1, associated with MI susceptibility (41). Recently, the meta-analysis of >6,500 CF patients from the International CF Gene Modifier Consortium confirmed these associations and identified two new loci comprising the nongastric form of the H⁺/K⁺ATPase (ATP12A) and serine protease 1 (PRSS1) genes (Figure 2) (42). Further transcriptomics analysis of nasal epithelial cells from CF patients and pancreatic gene expression data from GTEx, showed that lung disease severity is associated with variations in SLC6A14 gene expression in nasal epithelia, and meconium ileus susceptibility with variations in expression of SLC6A14, SLC26A9 and ATP12A in the pancreas.

Modifiers of cystic fibrosis related diabetes (CFRD)

CF-related diabetes (CFRD) is one of the most important complications of CF. While the age at onset varies considerably, CFRD prevalence increases with age, affecting more than 90% by the 5^th decade in PI-CF patients. Onset of CFRD is recognized to be associated with a more rapid decline in lung function and a reduced survival (43). A pilot candidate-gene study found that the gene encoding the transcription factor 7-like 2 (TCF7L2), which is known to be associated with type II diabetes prevalence in the general population, was also associated with the risk of developing CFRD (44). In particular, the variant rs7903146 increased the risk of diabetes threefold, and decreased the age of onset by 7 years. A further GWAS in 3,059 North-American patients including 644 patients with diabetes confirmed the association of CFRD with TCF7L2, and identified a new association with SLC26A9 (Figure 2) (45). Interestingly, as the SLC26A9 gene is also associated with the risk of onset of MI at birth (see above) this suggests a pleiotropic action of this gene (41). Finally, a recent study provides new data suggestive of an impact of SLC26A9 promoter variants on its expression (46).

Cystic fibrosis liver disease (CFLD) modifiers

Focal biliary cirrhosis is the most clinically relevant form of CF liver disease (CFLD) in PI-CF patients, since extension of the initially focal fibrogenic process may lead to multilobular biliary cirrhosis with subsequent portal hypertension and related complications (47, 48). Currently, multilobular cirrhosis ranks as the third leading cause of death in patients with CF after respiratory failure and transplantation-related complications (49). To date, only candidate-gene approaches have been performed to look for liver modifiers (50). Some genes have been shown to be associated to an increased frequency of liver disease, such as ATP binding cassette subfamily A member 4 (ABCA4), glutathione S-transferase (GSTP1), and a mannose-binding lectin (MBL) haplotype (51, 52); and other to a more severe liver disease such as serpin family A member 1 (SERPINA1, previously known as the α1antitrypsin gene) (Figure 2) (53). The French CF Modifier Gene Study recently reported the incidence of CFLD and severe CFLD in patients with CF and showed that the risk of CFLD increases with age, with a frequency up to 32% by age 25, and is associated with several risk factors, including male gender, CFTR F508del homozygosity, and a history of meconium ileus at birth (54). In a cohort of unprecedented size for such study (n = 3,328), we confirmed that the incidence of CFLD increased more rapidly in patients carrying the SERPINA1 Z allele, with up to 47% of the Z allele carriers developing liver disease before the age of 25 compared to only 30% for non-carrier patients (55).

Pharmacogenomics

In recent years, considerable efforts have led to the development of therapies that target the CFTR protein. Significant clinical benefits of ivacaftor, such as gain of lung function and reduced number of pulmonary exacerbations, were initially observed in patients older than 12 years and carrying at least one Gly551Asp (G551D) CFTR variant (56). Subsequently, ivacaftor was approved for other CFTR-gating variants and for patients older than 2 years (57–60). Since the airway response to ivacaftor was variable between individuals, it was hypothesized that modifier genes could play a role. Though genome-wide analyses are as yet incomplete, candidate studies targeted SLC26A9, a gene that may have a pleiotropic role in CF as discussed above (41, 42, 45). A pilot study in Canadians showed that one variant of this gene was indeed associated with the lung-response to ivacaftor, explaining 28% of the variability (Figure 2) (61). We also found that this gene was associated with the variability of the lung response in French (62). Specifically, focusing on the variant rs7512462 of the SLC26A9 gene in the patients carrying at least one G551D CFTR mutation, we observed that the CC genotype was associated with a lesser response, with a decrease in FEV_1pp of −7.7% compared to the patients carrying the ancestral allele T. Pharmacogenetics testing in CF patients is a growing current and future challenge that may be predictive for efficacy and for adverse drugs effects.

Summary and Future directions

Recent advances in research into the genetics of CF provide some insights on the phenotypic heterogeneity of the disease, potential targets to predict disease severity and personalize therapeutic care, and new therapeutic avenues for CF patients. Studying larger cohorts and using cutting-edge protocols of functional genomics may allow the identification of new CFTR variants and modifier genes that could better explain the heterogeneity of the disease. In addition, exploring the mechanisms of CFTR transcriptional regulation and evaluating the impact of CFTR regulatory variants and modifier genes in CF is crucial to better understand the disease and improve patient care. Current challenges are to develop clinical diagnostic tools based on identified CFTR variants and modifier genes as well as new specific and personalized therapies to improve survival and quality of life of all the patients with CF.

Highlights.

• Genetic variants contribute to the clinical heterogeneity in cystic fibrosis

• Genetic variants are located within and outside the CFTR locus

• Variants within cis-regulatory elements may modulate CFTR transcript abundance

• Modifier genes variants are associated with the variability of the disease severity

• Variants impact lung/intestinal/hepatic/pancreatic diseases and therapies’ efficacy