INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Cystic fibrosis (CF), the most common lethal autosomal recessive disease among Caucasians, is caused by mutations in the cystic fibrosis transmembrane conductance regulator (CFTR) gene. Defects in CFTR cause abnormal chloride and sodium transport across the apical membrane of epithelial cells in the airways, pancreas, intestine, and in the male genital system (1). In the respiratory system CFTR dysfunction predisposes to recurrent respiratory infections causing suppurative bronchiectasis which leads to respiratory failure and early death in most patients. At present, treatment of CF is mainly symptomatic, aimed at removing the retained secretions and treating chronic infection (1).

Since the identification of the CFTR gene, more than 800 CFTR mutations have been identified (Cystic Fibrosis Genetic Analysis Consortium, personal communication) including missense, small deletion or insertion, frameshift, and nonsense mutations (2). Nonsense mutations are mutations containing a premature termination signal causing the formation of truncated or unstable protein. Premature stop mutations account for approximately 5% of the total mutant alleles in CF patients (3, 4). However, in certain subpopulations the incidence of this class of mutation is much higher. In the Ashkenazi Jewish population the W1282X mutation is the most common CF-causing mutation and together with other nonsense mutations accounts for 64% of all CFTR alleles (5). CFTR nonsense mutations produce little or no CFTR chloride channels (8). The phenotype of patients carrying stop mutations is severe, similar to that of the F508 mutation (5, 11).

The aminoglycoside antibiotics, in addition to their antimicrobial activity, can increase the frequency of erroneous insertion of the nonsense codon, thereby permitting protein translation to continue to the normal end of the gene. This has been demonstrated in prokaryotic and eukaryotic cells (12). Recently, Howard and coworkers (9) demonstrated in HeLa cells transfected with plasmid vector carrying the CFTR nonsense mutations, that gentamicin induced a dose-dependent increase in expression of full-length CFTR. Subsequently, Bedwell and coworkers (10) have shown in a CF bronchial epithelial cell line carrying the CFTR W1282X premature stop mutation, that gentamicin was capable of restoring CFTR expression on the apical membrane.

The aim of this pilot study was to determine if gentamicin can, in vivo, activate mutant CFTR in CF patients carrying stop mutations. We used the nasal potential difference (PD) methodology to measure sodium and chloride transport before and after topical application of gentamicin drops to the nasal epithelium. This technique has been used extensively in studies assessing the response of patients with CF to gene transfer and to various novel pharmacological agents (15). Correction of the abnormal PD may suggest that gentamicin can correct the primary defect among patients carrying stop mutations.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patients and Treatment

Nine patients with CF carrying stop mutations (mean age 23 ± 11 yr, range 12 to 46 yr) were studied. Four patients were homozygous for the W1282X mutation, three were compound heterozygous W1282X/ G542X, one patient was compound heterozygous W1282X/3849 + 10kb C T, and one patient was compound heterozygous W1282X/ F508. The clinical data of the patients participating in the study are shown in Table 1. Diagnosis of CF was made in all patients by typical respiratory and/or gastrointestinal manifestations together with elevated sweat chloride levels. No patient received gentamicin for at least 4 wk before the study and no patient was using nasal drugs including nasal steroids. After measurement of the baseline nasal PD, all the patients received drops of Garamycin eye/ear drops (Schering-Plough Laboratory, Brussels, Belgium) containing gentamicin (as sulfate) 3 mg/ml, 2 drops three times daily to both nostrils for a total of 14 d. Patients or parents were instructed to instill the drops to both nostrils with the head tilted backwards and to remain still for several minutes afterwards. The patients underwent repeat PD measurements 14 d after starting gentamicin treatment. Nasal PD was also measured in 35 non-CF control subjects (mean age 26 ± 7 yr) and in 17 CF patients (mean age 20 ± 10 yr).

Nasal PD Measurements

Transepithelial nasal PD was determined by standard criteria as described by Knowles and coworkers (15). The PD between a fluid-filled exploring bridge on the nasal mucosa and a reference bridge inserted into the subcutaneous space of the forearm was measured. The reference bridge (21-gauge needle) filled with Ringer's solution in 4% agar was inserted into the subcutaneous space of the forearm. A specially constructed exploring bridge (PE-50 tubing) was marked at 0.5-cm intervals and perfused at a constant rate (0.2 ml/min) with Ringer's solution. Both bridges were linked by calomel electrodes to a high-impedance, low-resistance buffer amplifier. The output of the buffer amplifier was connected to a digital multimeter and a strip chart recorder. Initially, PD measurement of the skin was determined by touching the exploring catheter to the palm and recording the PD values on at least two occasions. A PD of 60 to 70 mV was obtained. Using direct vision with an otoscope, the exploring catheter was advanced through the inferior meatus of both nostrils and PD was recorded at various sites. The maximal PD from both nostrils was measured and the nostril with the highest maximal PD was used in the subsequent perfusion studies. After consistent basal PD measurements were obtained, the effects of amiloride superfusion through a second tube overriding the exploring catheter were evaluated. The exploring catheter (constructed of double-barreled PE-50 tubing) was repositioned under the inferior turbinate at the site of maximal PD. Baseline measurements of potential difference were recorded for at least 1 min during perfusion with Ringer's lactate at 0.2 ml/min, and amiloride hydrochloride (10⁴ M/L) dissolved in Ringer's solution was perfused at 5 ml/min (35 to 36° C) for 3 min via the perfusion catheter. To study nasal chloride permeability and cyclic adenosine monophosphate (cAMP) activation of chloride permeability, a large chloride chemical gradient was generated across the apical membrane by superfusion of the nasal mucosa with a chloride-free solution (Ringer's solution with gluconate substituted for chloride) for 3 min and then a chloride-free solution containing 10⁵ M/L isoproterenol was perfused at a rate of 5 ml/min. The change in voltage response to isoproterenol over 3 min served as an index of cAMP activation of the epithelial chloride permeability. The resultant change in PD was recorded and expressed as both an absolute change, and as a percentage change from the baseline maximal PD value. The residual PD (absolute voltage remaining) was also noted.

The Human Ethics Committee of the Israeli Ministry of Health approved this study and informed written consent was obtained from the patients or parents.

Statistical Analysis

The two-tailed Mann-Whitney U test was used to compare continuous data allowing for non-normal distribution and small group sizes. Data are presented as group means (± SD) unless specified otherwise.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Ten treatment trials were performed in the nine patients who participated in the study. One patient (Patient 1) underwent two treatment trials with a 1-mo interval in between. All patients were highly motivated to participate in the study and declared full compliance.

Before gentamicin application the nasal PD tracing was typical for CF in all the studied patients; basal PD was elevated, the depolarizing response to amiloride superfusion was exaggerated, and after superfusion with chloride-free/isoproterenol solution there was no increase in PD (Figure 1B, Table 2). Fourteen days after gentamicin treatment the most prominent and significant effect of gentamicin application was restoration of chloride transport causing repolarization (Tables 2 and 3, Figure 2). After perfusion with chloride-free/isoproterenol solution, nasal PD was significantly increased toward the typical response in our non-CF control group (Figures 1C and 2). Two patients had a third PD measurement 1 wk after discontinuation of gentamicin application, and PD values returned to their previously abnormal levels (data not shown).

View larger version (14K): [in this window] [in a new window]   Figure 1.   Representative nasal potential difference measurement recordings: (A) a subject without CF; (B) a study patient before gentamicin therapy; (C ) the same patient 2 wk after gentamicin therapy. The arrows indicate change in perfusate.

View larger version (12K): [in this window] [in a new window]   Figure 2.   The change in PD after perfusion of chloride-free and isoproterenol solutions in non-CF controls, CF patients, study patients prior to gentamicin (Patient 1 performed the trial twice and the mean value is shown), and study patients after gentamicin treatment.

The effect of 14 d of gentamicin treatment on sodium transport was less prominent. Although the basal PD was reduced, it did not reach statistical significance (Table 2). Nonetheless, in two patients (Patients 5 and 6) the basal PD was reduced to normal levels (in Patient 5 the change was from 70 mV to 20 mV, and in Patient 6 the change was from 40 mV to 18 mV). There was no significant change in the response to amiloride perfusion in any of the patients.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This pilot study in patients with CF carrying stop mutations demonstrates a significant correction of the basic electrophysiological abnormalities in CF using topical application of gentamicin drops to the nose. In most patients the main effect of gentamicin was activation of transmembrane chloride transport that approached the normal range in seven of the nine patients studied. It is likely that this effect is a result of the ability of gentamicin to increase the frequency of erroneous insertion of nonsense codons, thereby permitting the translation of CFTR alleles carrying missense mutations to continue reading to the end of the gene. It is interesting to note that recent studies that assessed new therapies aiming to modulate mutated CFTR revealed that the measurable effect was only on chloride conductance (16). The clinical significance of this effect on electrolyte conductance warrants further investigation.

CFTR nonsense mutations produce little or no CFTR chloride channels (8). It is possible that the effect on chloride conductance after gentamicin may reflect a relatively small number of active CFTR channels or via activation of outwardly rectified chloride channels (ORCC) frequently observed in the respiratory epithelia. The ORCCs are present in the epithelia of CF airways but are inactive, and the cAMP-dependent regulation of this channel is restored when functional CFTR is expressed (23). However, in the in vitro studies performed by Bedwell and coworkers (10), it was clearly shown that the effect of gentamicin was primarily due to the restoration and expression of a sufficient amount of full-length functional CFTR protein at the apical membrane, which also restored the cAMP-dependent chloride conductance through the ORCC. Furthermore, the correction of sodium conductance observed in two of our patients may indicate that gentamicin caused expression of active CFTR in the nasal epithelial cells.

The in vitro studies showed that after treatment with gentamicin, full-length CFTR was observed by immunoprecipitation with response increasing with doses from 0.1 to 0.4 mg/ml (9). The standard gentamicin drops that we used in vivo were approximately 10 times this concentration. It is difficult to extrapolate directly from the in vitro studies the maximal effective dose of topical application because the drug is diluted by nasal surface liquid and needs to cross the extracellular barriers before it reaches the respiratory epithelium.

Quantification studies have shown that after aminoglycoside incubation, the amount of full-length CFTR produced is as much as 25% (in the R553X mutation) to 35% (in the G542X mutation) of that observed in cells transfected with a wild-type CFTR complementary DNA (cDNA) (9, 10). This increase in functional CFTR might be above the threshold that is required for normal respiratory epithelial cell function. Activation by gentamicin of functional CFTR in nasal epithelial cells may reach this threshold leading to normalization of cell membrane function in our patients. One patient (Patient 8) who carried the W1282X/3849 + 10kbC T genotype showed a marked response which led to normalization of the nasal PD. The 3849 + 10kbC T mutation is a splicing mutation in intron 18 which creates a partially active splice site in intron 19 of the CFTR gene. This can lead to the insertion of a new 84 bp "exon," harboring an in-frame stop codon between exon 19 and 20 (26). Thus, two types of messenger RNA (mRNA) are produced; normally spliced mRNA leading to a normal protein, and aberrantly spliced mRNA which contains a stop mutation. This patient is, therefore, a carrier of two stop mutations but is expected to have some normally spliced RNA resulting in functional CFTR. Studies in patients with CF carrying CFTR mutations affecting splicing showed that respiratory disease severity was correlated with the levels of aberrantly spliced CFTR mRNA (27). The pulmonary function in these patients correlated with the amount of correctly spliced RNA, and there was a threshold level of normally spliced RNA that was associated with normal lung function. The low level of FEV₁ (44% predicted) that this patient (Patient 8) had, suggests that the level of normally spliced RNA was below the threshold. The marked PD response to gentamicin may indicate that partial correction of the stop mutation on both alleles together with the small amount of residual functioning CFTR, increased her functional CFTR to a level that significantly improved her nasal epithelial cell function.

CFTR function, in the in vitro studies, was regained within 18 h (10). We measured nasal PD after 14 d of therapy. We did not perform daily measurements because this may cause trauma affecting the electrophysiological properties of nasal mucosa. However, in one patient who agreed to continue the trial for another week, a further improvement in nasal PD measurements was noticed (data not shown). Thus, it may be that the effect of gentamicin on the nasal epithelia increases with time.

In a recent study Barton-Davis and colleagues (30) have shown that gentamicin restored dystrophin function to skeletal muscle in mice suffering from Duchenne muscular dystrophy possessing a premature stop codon at the dystrophin gene. Thus, the effect of aminoglycoside antibiotics to suppress stop mutations may not be exclusive to the CFTR gene. However, it is possible that in our study gentamicin improved the integrity of the nasal epithelium thus allowing more chloride transport. Therefore, further work in a placebo-controlled blinded study is essential both in patients carrying stop mutations and in other genotypes not potentially responsive to gentamicin. The examination of the effect of different variables such as time, dose, and the number of treatments per day on the response to gentamicin is also necessary as is the measurement of mRNA in the nasal epithelium. We also have not determined how long the effect lasts, though 1 wk after termination of gentamicin, in the two patients studied, nasal PD returned to the abnormal values typical of CF.

Two of the patients did not show changes in nasal PD in response to gentamicin application. This could be a result of poor patient compliance with the technique of nasal drops application, or low sensitivity of nasal cells to gentamicin. These two patients (Patients 2 and 3) were sisters whose parents were first cousins. Defects in other genes regulating factors involved in the process by which gentamicin increases the frequency of erroneous insertion of nonsense codons may explain the lack of response in these two sisters. Because the effect of aminoglycosides on the activation of CFTR stop mutations is dose-dependent (9, 10), it is possible that higher doses of gentamicin may be more effective.

It was noted that at baseline measurements, there was relatively higher cAMP-mediated chloride conductance in patients carrying stop mutations than the other patients with CF (in Figure 2). It is possible that there are small differences in chloride conductance among patients with different classes of mutations. Similar observations have been noted previously (31). The molecular mechanism by which gentamicin restores CFTR function is not fully understood. The mechanism of translation termination is highly conserved among most organisms, and is almost always signaled by an amber (UAG), ochre (UAA), or opal (UGA) termination codon. The nucleotide sequence surrounding the termination codon plays an important role in determining the efficiency of translation termination (32). Naturally occurring termination codons lie within a context that promotes efficient translation termination. In contrast, premature translation termination codons introduced by mutation would not be subject to such selective pressure, and may not be able to promote efficient translation termination. Aminoglycoside antibiotics can reduce the fidelity of translation (35), predominantly by inhibition of the ribosomal "proofreading." Proofreading operates to exclude poorly matched amino acyl-transfer RNA (tRNA) from incorporation into the polypeptide chain (36). In this way aminoglycosides increase the frequency of erroneous insertions at the nonsense codon and permit translation to continue to the end of the gene. This has been shown in eukaryotic cells (37) including human cells (38).

Gentamicin has been used for many years to treat patients with CF. It is regarded as safe when given by inhalation. The results of our study suggest that gentamicin may have a potential role in the therapy of patients with CF carrying stop mutations by correcting the primary defect. The next step would therefore be to assess if aerosolized gentamicin promotes the production of full-length CFTR through the suppression of premature stop mutations in lung epithelia. If successful, this approach may provide a therapeutic tool for intracellular modification of mutated CFTR to produce functional CFTR in a specific genotypic subgroup of patients with CF. This approach may also prove beneficial in the treatment of other genetic diseases caused by stop mutations.