Functional analysis of the p.[Arg74Trp;Val201Met;Asp1270Asn]/p.Phe508del CFTR mutation genotype in human native colon

Abstract Background The impact of complex alleles on CFTR processing and function has yet not been investigated in native human tissue. Methods Intestinal current measurements (ICM) followed by CFTR immunoblot were performed on rectal biopsies taken from two siblings who are compound heterozygous for the CFTR mutations p.Phe508del and the complex allele p.[Arg74Trp;Val201Met;Asp1270Asn]. Results Normal and subnormal chloride secretory responses in the ICM were associated with normal and fourfold reduced amounts of the mature glycoform band C CFTR, respectively, consistent with the unequal clinical phenotype of the siblings. Conclusion The combined use of bioassay and protein analysis is particularly meaningful to resolve the CFTR phenotype of “indeterminate” borderline CFTR genotypes on a case‐to‐case basis.


| INTRODUCTION
Cystic fibrosis (CF, OMIM 219700) is caused by mutations in the CFTR (GenBank M28668.1) (Elborn, 2016;Ratjen et al., 2015). The CF Mutation Database (http:// www.genet.sickkids.on.ca) currently lists more than 2,000 mutations in the CFTR, but only a small number of them are clearly defined as CF-causing based on epidemiological data and functional studies (http://www.cftr2.org) (Sosnay et al., 2013). Genotype-phenotype correlations become even more complicated in case of complex alleles, that is, two or more CFTR mutations in cis on the same allele. The first reported case was the revertant mutation p.Arg553Gln which, when carried in cis together with the major mutation p.Phe508del, was associated in the index case with borderline chloride levels in sweat test (Dörk et al., 1991) and which, when expressed in heterologous cells, could partially correct the processing and gating defect caused by the p.Phe508del mutation (Teem et al., 1993). On the other hand, the combination of mutations may be additive on phenotype (Terlizzi et al., 2017) so that the processing and/or function of CFTR become more and more impaired. For example, when expressed in heterologous systems, p.Arg74Trp CFTR behaved like a neutral polymorphism, the double mutant p.[Arg-74Trp; Asp1270Asn] exhibited a reduced cAMP-mediated anion transport and the triple mutant p.[Arg74Trp;Val-201Met;Asp1270Asn] was moreover impaired in the posttranslational processing and trafficking of CFTR (Fanen et al., 1999;Terlizzi et al., 2017). The corresponding clinical phenotype, however, is less clear. Individuals with p. Phe508del and p.[Arg74Trp;Val201Met;Asp1270Asn] in trans have been reported to be healthy (Brugnon et al., 2008) or to suffer from a CFTR-related disorder (CFTR-RD) or CF (Claustres et al., 2004).
Here, we report on the yet uncharacterized phenotype of this complex CFTR allele in patient's tissue in order to unravel of how the CFTR mutation genotype translates into basic defect and clinical phenotype in vivo.

| Ethical compliance
The study (no. 2771) was approved by the Ethics Committee of Hannover Medical School.

| Intestinal current measurements (ICM)
The electrogenic transport of ions across the intestinal epithelium was measured as short circuit current (I SC ) by ICM following the Standard Operating Procedure (SOP), version 2.7, of the ECFS Diagnostic Network Working Group.
The luminal and basolateral compartments were filled with a HCO − 3 containing buffer of the following composition: 128 mM NaCl, 4.7 mM KCl, 20.2 mM NaHCO 3 , 10 mM HEPES, 0.3 mM Na 2 HPO 4 , 1 mM MgCl 2 , 1.3 mM CaCl 2 , 10 mM d-glucose. The solution was kept at 37°C and gassed continuously with a mixture of 95% O 2 /5% CO 2 , which maintained the pH at 7.4. Experiments were performed under short circuit conditions, and I SC was recorded continuously throughout the experiment.
To determine CFTR Cl − channel function, rectal tissues were equilibrated in Ussing chambers for 40 min in the presence of amiloride (10 µM, luminal) to block electrogenic Na + absorption and indomethacin (10 µM, basolateral) to inhibit prostaglandin E2 synthesis and endogenous cAMP formation. Previous studies demonstrated that endogenous CFTR activity is largely inhibited under these experimental conditions (Bronsveld et al., 2000). To assess CFTR-mediated | 3 of 7 Schucht et al.
Cl − transport, we next measured lumen-positive (Cl − secretory) I SC responses induced by cAMP-dependent stimulation with 3-isobutyl-1-methylxanthine (IBMX, 100 μM) and forskolin (1 μM) added to the basolateral compartment. In normal human colon, CFTR-mediated Clsecretion (lumenpositive I SC responses) is augmented by cholinergic co-activation, which leads to an increase in intracellular Ca 2+ and stimulation of basolateral Ca 2+ -dependent K + channels that increase the electrical driving force for luminal Cl − secretion via CFTR (Bronsveld et al., 2000;Roth et al., 2011). In CF colon, cholinergic co-activation results in an initial inverse lumen-negative I SC response reflecting luminal K + secretion, whereas the lumen-positive Cl − secretory response is absent or reduced depending on the severity of mutant CFTR malfunction. To increase the driving force for CFTR-mediated Cltransport, rectal tissues were therefore activated with carbachol (CCH; 100 μM, basolateral) in the presence of IBMX and forskolin, and CCH-induced lumen-negative (K + secretory) and lumen-positive (Cl − secretory) I SC responses were determined. Thereafter, all non-CFTR chloride channels were inhibited with 0.2 mM 4,4′-Diisothiocyano-2,2′-stilbenedisulfonic acid (DIDS) followed the addition of 0.5 mM histamine to evoke again a chloride secretory response (Bronsveld et al., 2000).
Measurements were performed in four rectal mucosa biopsy specimens. After ICM, the biopsies were frozen and stored at −80°C until use.

| CFTR immunoblot analysis
Biopsies of the elder sibling were sampled and processed in 2012 following the procedure described previously (van Barneveld et al., 2010). Biopsies were homogenized in the presence of 10 mM iodoacetamide, 20 mM PMSF, 20 mg/ ml pepstatin and antipain, 100 µg/ml leupeptin and antiprotonin, and 500 µg/ml soybean trypsin-inhibitor in Tris buffer (20 mM Tris/HCl, 150 mM NaCl, pH 8). The lysis started by incubation with 0.03% SDS for 60 min, followed by 1% (v/v) Triton X-100% and 0.5% (w/v) sodium deoxycholate for 2 hr. After centrifugation (16,000 g, 4°C, 20 min), the supernatant was incubated with the specific pre-immune serum and protein A-sepharose for 60 min. Immunoprecipitation (IP) was carried out with the in-house polyclonal anti-CFTR antibodies R40, R66, and R16 in the presence of protein A-and protein G-agarose. Pellets were washed several times (van Barneveld et al., 2010). CFTR-immunoreactive bands were detected on 5% SDS-PAGE separated PVDF-membranes with mABs 570 and 596 (1:500) in 0.2% I-Block (Tropix, Applied Biosystems) in 0.05% Tween-TBS (T-TBS) and pre-adsorbed anti-mouse IgG-HRP from donkey (1:300,000, Abcam) in 0.2% I-block in T-TBS incubating with ECL Advance (GE Healthcare) for 20 s. The CFTR-immunoreactive signal was quantified by densitometry of Hyperfilms ECL (GE Healthcare) exposed to ECL Advance-covered immunoblot for 1, 5, 10, 30, and 55 s. As markers for calibration of the CFTR B-and C-bands in rectal biopsies, the B-and C-bands of T84 immunoprecipitates separated on the same gel and the protein marker Precision plus Protein Standards All Blue (10-250 kDa, Bio-Rad) were used. The amount of protein was determined by the Bradford assay. Biopsies of the younger sibling were sampled and processed in 2017. Frozen biopsies were lysed in 50 µl buffer (50 mM Tris, pH 6.8; 10% glycerol; 0.1 M DTT; 10 −4 diluted protease inhibitor cocktail [SRE 0055, Sigma]; 2% SDS) supplemented with 0.5 µl PMSF and 0.5 µl 1:20-diluted Omnicleave endonuclease (Epicentre) for 10 min at room temperature and thereafter for 30 min at 37°C. After removal of insoluble debris by centrifugation, a mixture of 15 µl supernatant/15 µl glycerol was separated at 4°C by 6% SDS-PAGE with 1.5 V/cm for 17 hr and then 9 V/cm for 5 hr (Kälin, Claass, Sommer, Puchelle, & Tümmler, 1999) in a Bio-Rad Mini-PROTEAN Tetra Cell. These electrophoresis conditions were chosen to achieve a high resolution of proteins of 100 kDa and larger such as CFTR, albeit all proteins below 70 kDa will be eluted into the lower buffer chamber by iontophoresis. Electrotransfer of within-gel remaining proteins onto Amersham Protran Supported 0.45 NC membranes was performed for 18 hr at 44 mA and 0°C (Kälin et al., 1999). CFTR immunoreactive bands were detected on the blot by sequential incubation with first anti-CFTR mAbs 217, 570, 596, 660 (1:1,600 dilution, 4°C, overnight), then secondary goat anti-mouse IgG (Abcam) (1 hr, room temperature), and finally SuperSignal West Femto Maximum Sensitivity Substrate (Thermo) according to the instructions of the manufacturers.

| CASE REPORT
The currently 8-year and 5-year old male index cases are two siblings of German maternal and Moroccan paternal descent who are compound heterozygous for the CFTR mutations p.Phe508del on the maternal allele and p.[Arg74Trp;p. Val201Met;p.Asp1270Asn] on the paternal allele. The elder boy suffered from recurrent episodes of obstructive bronchitis and had recurrent detection of Staphylococcus aureus and of Haemophilus influenzae in respiratory specimens. The younger boy is healthier. He experienced fewer episodes of airway infections. Throat swabs were repeatedly positive for H. influenzae, but never for S. aureus. Spirometry is normal for age in both siblings. Multiple-breath nitrogen washout tests (Poncin, Singer, Aubriot, & Lebecque, 2017) yielded slightly elevated lung clearance indices of 8.0 and 7.8 for the elder and younger boy, respectively.
The basic defect was investigated with the CFTR biomarkers sweat chloride concentration in Gibson-Cooke pilocarpine sweat tests and chloride secretory responses in intestinal current measurements (ICM; Figure 1) (De Boeck et al., 2011) followed by immunoblot analysis of CFTR protein ( Figure 2) (van Barneveld et al., 2010).
Chloride levels in sweat tests were in the lower intermediate range between 40 and 45 mmol/L in both siblings ( Table 1). ICM of rectal biopsies taken from both siblings yielded normal chloride secretory responses to forskolin/IBMX (Table 1). Since p.Phe508del CFTR responses are within the range of a few percent of wild-type (van Barneveld et al., 2010), the cAMP-linked chloride secretion can be attributed to the CFTR triple mutant. In contrast, the transient pulses of chloride secretion evoked in the biopsies by basolateral Ca 2+ -dependent K + efflux induced by carbachol or histamine (Bronsveld et al., 2000) were donor-dependent either in the normal or in the CF range (Figure 1). The carbachol-and the histamine-mediated responses of the biopsies of the younger sibling were in the normal range, but the responses of the specimens of the elder sibling were in the CF range below the lowermost value ever measured with the same protocol in non-CF probands ( Table 1). The profile of a normal response to forskolin/IBMX and low responses to carbachol and histamine seen in the elder sibling's biopsies is typical for individuals with very mild CF (Stanke et al., 2008) and can be attributed to the drug concentrations selected by the current "SOP" for ICM. Even minute amounts of residual CFTR activity will generate a response to forskolin/IBMX, but the response will saturate at about 20% of wild-type CFTR activity. Conversely, the chloride secretory responses to carbachol and histamine are approximately proportional to the CFTR activity of the biopsy. In conclusion, the elder sibling exhibited the phenotype of a very mild CF in the ICM, whereas the younger sibling showed a normal response typical for a healthy non-CF individual.  Next, we determined the CFTR protein levels in the biopsies (Figure 2). The immunoblot detected fourfold lower amounts of complex glycosylated CFTR in immunoprecipitates of rectal biopsies collected from the elder sibling than in specimens taken from wild-type controls (Figure 2a). In contrast, the intensity of immunoreactive CFTR C-and B-bands was indistinguishable between samples from the younger sibling and non-CF controls ( Figure  2b). Thus, the low and normal responses of the intestinal epithelium to carbachol and histamine observed in the ICM corresponded to low and normal levels of the mature CFTR glycoform, respectively.

| DISCUSSION
The diagnosis of CF is typically based on clinical features, a pathological sweat test and the detection of two diseasecausing mutations in the CFTR (Elborn, 2016). The scenario becomes more complex in case of CFTR sequence variants that confer substantial residual activity (Bombieri et al., 2011;De Boeck et al., 2011;Sosnay et al., 2017). Subject-to-subject variation in phenotype may be so pronounced that the cftr.2 database classifies these sequence variants as "indeterminate" and recommends that "clinical criteria alone should be used whether a person with this variant has CF" (Sosnay et al., 2013). If isolated, all three mutations of the complex allele studied here, p.Arg74Trp, p.Val201Met; p.Asp1270Asn, fall into the category "indeterminate". Hence, to resolve the impact of the complex allele on CFTR function, we combined the CFTR bioassay ICM with immunochemical CFTR protein analysis. In agreement with the unequal clinical presentation, wildtype CFTR activity corresponded with wild-type levels of CFTR protein in the healthier sibling and reduced CFTR activity corresponded with reduced levels of the mature CFTR glycoform in the unhealthier sibling. Immunoblot analysis of CFTR glycoforms in patients' specimens is technically challenging and has yet been applied to only a few common CFTR genotypes causing typical full-blown CF (Kälin et al., 1999;Kartner, Augustinas, Jensen, Naismith, & Riordan, 1992;van Barneveld et al., 2010). However, as shown in this case report, the combined use of bioassay and protein analysis is particularly meaningful to resolve the individual CFTR phenotype of "indeterminate" borderline CFTR genotypes on a case-to-case basis.