In vitro and ex vivo suppression by aminoglycosides of PCDH15 nonsense mutations underlying type 1 Usher syndrome
- First Online:
- Cite this article as:
- Rebibo-Sabbah, A., Nudelman, I., Ahmed, Z.M. et al. Hum Genet (2007) 122: 373. doi:10.1007/s00439-007-0410-7
- 195 Views
Type 1 Usher syndrome (USH1) is a recessively inherited condition, characterized by profound prelingual deafness, vestibular areflexia, and prepubertal onset of retinitis pigmentosa (RP). While the auditory component of USH1 can be treated by cochlear implants, to date there is no effective treatment for RP. USH1 can be caused by mutations in each of at least six genes. While truncating mutations of these genes cause USH1, some missense mutations of the same genes cause nonsyndromic deafness. These observations suggest that partial or low level activity of the encoded proteins may be sufficient for normal retinal function, although not for normal hearing. In individuals with USH1 due to nonsense mutations, interventions enabling partial translation of a full-length functional protein may delay the onset and/or progression of RP. One such possible therapeutic approach is suppression of nonsense mutations by small molecules such as aminoglycosides. We decided to test this approach as a potential therapy for RP in USH1 patients due to nonsense mutations. We initially focused on nonsense mutations of the PCDH15 gene, underlying USH1F. Here, we show suppression of several PCDH15 nonsense mutations, both in vitro and ex vivo. Suppression was achieved both by commercial aminoglycosides and by NB30, a new aminoglycoside-derivative developed by us. NB30 has reduced cytotoxicity in comparison to commercial aminoglycosides, and thus may be more efficiently used for therapeutic purposes. The research described here has important implications for the development of targeted interventions that are effective for patients with USH1 caused by various nonsense mutations.
Usher syndrome (USH) is the most common cause of combined sensorineural deafness and blindness, characterized by bilateral sensorineural deafness and progressive loss of vision due to retinitis pigmentosa (RP). The prevalence of USH varies from 1/16,000 to 1/50,000 in different populations. The majority of USH cases can be classified into one of three clinical subtypes, the most severe of which is Usher type 1 (USH1), characterized by profound prelingual hearing loss, vestibular areflexia, and prepubertal onset of RP (Petit 2001).
Six USH1 loci (USH1B-USH1G) have been mapped, and the causative genes have been identified for five of them (MYO7A, USH1C, CDH23, PCDH15, and SANS) (Kremer et al. 2006). Interestingly, while truncating mutations, including nonsense mutations, in each of these genes cause USH1, certain missense mutations in some of the very same genes cause nonsyndromic deafness, which is not associated with RP. Specifically, certain mutations of MYO7A underlie USH1B, while others cause nonsyndromic dominant deafness DFNA11, and possibly nonsyndromic recessive deafness DFNB2 (Astuto et al. 2002; Liu et al. 1997; Luijendijk et al. 2004; Weil et al. 1995); Certain mutations of the USH1C gene underlie USH1C, while others cause nonsyndromic deafness DFNB18 (Ahmed et al. 2002; Ouyang et al. 2002); Certain mutations of CDH23 underlie USH1D, while others cause nonsyndromic deafness DFNB12 (Bork et al. 2001); USH1F and nonsyndromic deafness DFNB23 are caused by allelic mutations of PCDH15 (Ahmed et al. 2001, 2003). These observations suggest that partial or low level activity of the proteins encoded by these genes may be sufficient for normal retinal function although not for normal hearing.
In individuals who have USH1, deafness is congenital and profound, but the loss of vision due to RP is delayed in the onset, initially presenting as loss of night vision, and later in life as a gradual restriction of the visual field, eventually leading to blindness. While the auditory deficit can be successfully treated with cochlear implants (Brownstein et al. 2004; Pennings et al. 2006), to date there is no effective treatment for the ophthalmic component of USH1. Interventions to enable at least some translation of full-length protein may delay the onset and/or progression of RP in individuals with USH1 due to nonsense mutations. One such possible therapeutic approach is suppression of nonsense mutations by aminoglycosides.
Aminoglycosides can reduce the fidelity of translation of mRNA into protein, predominantly by inhibiting ribosomal proofreading, thus leading to read-through of nonsense mutations and generation of full-length protein. The amino acid that is inserted at the position of the stop codon is not necessarily the amino acid present in the wild type (wt) protein. However, the resulting protein may still be fully or partially functional. Suppression of premature stop codons by aminoglycoside antibiotics was described in bacteria, yeast and mammalian cells (Burke and Mogg 1985; Gorini and Kataja 1964; Singh et al. 1979). The degree of suppression varied significantly, and depended on several factors, including the nature of the stop mutation and the neighboring DNA sequence context (Keeling and Bedwell 2002; Manuvakhova et al. 2000).
Suppression of nonsense mutations by aminoglycosides has been tested as a potential therapy for several genetic diseases, including cystic fibrosis (Wilschanski et al. 2003), Duchenne muscular dystrophy (Politano et al. 2003), Hurler syndrome (Keeling et al. 2001), ataxia telangiectasia (Lai et al. 2004), and spinal muscular atrophy (Mattis et al. 2006). As a first step toward using this approach as a potential therapy for RP in USH1 patients due to nonsense mutations, we chose to focus on nonsense mutations of the PCDH15 gene. Here, we show suppression by aminoglycosides of several PCDH15 nonsense mutations, both in vitro and ex vivo.
Materials and methods
Commercial aminoglycosides used in this study were purchased from sigma. NB30 was synthesized and characterized as previously reported (Nudelman et al. 2006) (compound 3). Both commercial aminoglycosides and the synthetic NB30 were in their sulfate salt forms.
In vitro transcription/translation assay
Expression constructs were generated by insertion of a 1 kb fragment of PCDH15 cDNA, harboring either the wt allele or the p.R245X mutation, into the ClaI restriction site of the pCS2 + MT vector (Roth et al. 1991). COS-7 cells were grown in six well plates on coverslips (for immunofluorescence purposes) or in 60 mm plates (for Western blot analysis), in DMEM medium containing 10% fetal bovine serum (FBS), 1% penicillin/streptomycin and 1% glutamine (Biological Industries) at 37°C and 5% CO2. Twenty-four hours later, cells were transfected with either the wt or mutant PCDH15-myc expression construct, using the jetPEI transfection reagent (polyplus-transfection). Five hours later, medium was changed to fresh medium without streptomycin, and cells were incubated with the tested compounds for 48 h (1 or 2 mg/ml of commercial aminoglycosides; 1, 2 or 4 mg/ml of NB30).
Following incubation with the tested compounds, cells were fixed with 4% formaldehyde for 10 min at room temperature (RT). After rinses in 1XPBS, membrane permeabilisation in 0.1% Triton X-100 was performed for 5 min at RT, followed by a 30 min blocking in 5% FBS at RT. After rinses in 1XPBS, cells were incubated over night at 4°C with a monoclonal antibody directed against the myc-tag epitope at a dilution of 1:800 in 1% FBS solution. The following day, after rinses in 1XPBS, cells were incubated with Cy3-conjugated AffiniPure Donkey anti-Mouse IgG (Jackson ImmunoResearch) at a dilution of 1:100 in a 2% FBS, 2% BSA, and 1% Tween 20 solution for 1 h at RT. Coverslips were mounted onto slides using Vectashield with DAPI (Vector Laboratories). Images were taken with the Image-Pro Plus software connected to an inverted fluorescent microscope (Zeiss Axiovert 135).
Western blot analysis
Following incubation with the tested compounds, cells were lysed with M-PER mammalian protein extraction reagent (Pierce). Protein samples (40 μgr each) were subjected to denaturing polyacrylamide gel electrophoresis (4–20% Precise protein gels, Pierce). Proteins were then transferred to a nitrocellulose membrane (GE Healthcare), which was incubated with a monoclonal antibody directed against the myc-tag epitope (diluted 1:2,000), followed by a peroxidase-conjugated AffiniPure goat-anti-mouse IgG secondary antibody (diluted 1:1,000) (Jackson ImmunoResearch Laboratories). Signals were visualized by chemiluminescence using the Amersham ECL Western blotting analysis System (GE Healthcare). To control for the amount of loaded protein, the membranes were subsequently stripped and re-probed with a monoclonal antibody against α-tubulin (Sigma) at a dilution of 1:5,000. To control for transfection efficiency, DNA was extracted from each culture and subjected to PCR amplification with a primer-pair specific to the PCDH15-myc construct (F: GCCACCGTGAATGAGCTCACTCCAG and R: GTGAGGTCGCCCAAGCTCTCC), under non-saturating conditions (25 cycles of amplification). PCR products were subjected to electrophoresis on a 2% agarose gel and relative band intensities were quantified. Quantification of both Western blot signals and PCR products was performed with ImageMaster (Bio-Rad) and TotalLab (Phoretix International) software.
HEK-293, COS-7 or MDCK cells were grown in 96-well plates (5,000 cells/well) in DMEM medium containing 10% FBS, 1% penicillin/streptomycin and 1% glutamine (Biological Industries) at 37°C and 5% CO2 over night. The following day, medium was changed to medium without streptomycin and different concentrations of the tested compounds were added. Forty-eight hours later, a cell proliferation assay (XTT based colorimetric assay, Biological Industries) was performed, using the 5-h incubation protocol, according to the manufacturer’s instructions. OD was read using an Elisa plate reader. Cell viability was calculated as the ratio between the number of living cells in cultures grown in the presence of the tested compounds, versus cultures grown without compound. The concentration of half-maximal lethal dose for cells (LD50) was obtained from fitting concentration-response curves to the data of at least two independent experiments, using GraFit 5 software (Leatherbarrow 2001).
Suppression of PCDH15 nonsense mutations by commercial aminoglycosides in vitro
Eight USH1-causing nonsense mutations of PCDH15 have been reported to date. Among them are p.R3X, which was identified as the cause for USH1 in two families, from Pakistan and India (Ahmed et al. 2001; Alagramam et al. 2001); p.R643X, which was identified in USH1 families from Pakistan and France (Ahmed et al. 2003; Roux et al. 2006); p.R929X, which was identified in an USH1 patient of European descent (Ouyang et al. 2005); and p.R245X, which we identified as a major cause of USH1 in Ashkenazi Jews (Ben-Yosef et al. 2003).
Our first step was to test in vitro suppression of each of these mutations by aminoglycosides, using a series of reporter constructs harboring each of the four nonsense mutations (the pDB constructs; Fig. 1a). In vitro transcription/translation of the wt constructs resulted in the production of a 35-kDa polypeptide, while a 25-kDa polypeptide was produced from the mutant constructs. Addition of aminoglycosides to the transcription/translation reaction of the mutant constructs resulted in the synthesis of a 35-kDa polypeptide in addition to the 25-kDa polypeptide, indicating partial read-through of the nonsense mutation (Fig. 1a–c). A faint 35-kDa polypeptide could be occasionally observed among reaction products of the mutant constructs without the addition of aminoglycosides. This product presumably represents spontaneous read-through of the nonsense mutations and comprised up to 6% of total reaction product. Similar findings have been previously reported when the same experimental system was used (Keeling and Bedwell 2002; Manuvakhova et al. 2000).
Maximum in vitro suppression levels achieved for PCDH15 mutations
Supp. level (%)a
Conc. (μg/ml) b
Supp. level (%)a
Conc. (μg/ml) b
Supp. level (%)a
Conc. (μg/ml) b
Supp. level (%)a
Conc. (μg/ml) b
91 ± 0.5
50 ± 3
49 ± 6
21 ± 9
46 ± 0.1
11 ± 1.5
9 ± 3
6 ± 0.3
57 ± 6
47 ± 8
34 ± 6
13.5 ± 1
62 ± 3
29 ± 3
47 ± 3
13 ± 4
All four nonsense mutations tested generate a UGA stop codon (Fig. 1a). However, there were marked differences between suppression rates obtained for each of the mutations (Fig. 1; Table 1). It had been shown that aminoglycoside-mediated suppression of nonsense mutations depends on several factors, including the nature of the nucleotide located immediately following the stop codon. The highest rate of read-through of the UGA stop codon is obtained when a C is located at this position (Manuvakhova et al. 2000). Indeed, our experimental results show that highest suppression rates were obtained for p.R3X (UGAC) (Table 1). The nucleotide located immediately following the UGA stop codon for both p.R245X and p.R929X is A. Yet, p.R929X was more efficiently suppressed by all compounds tested (Table 1). These results further demonstrate that the sequence context beyond the tetranucleotide termination signal can also influence the level of read-through induced by various aminoglycosides (Manuvakhova et al. 2000).
The maximal dosage used for each aminoglycoside was limited by its toxic effect on overall translation, leading to marked reduction in total product amount. This observation indicates that USH1-causing nonsense mutations of PCDH15 are sensitive to suppression by aminoglycoside antibiotics, including some that are clinically relevant. However, this effect may be of limited efficacy, due to the known toxicity of aminoglycosides.
Suppression of the p.R245X nonsense mutation by commercial aminoglycosides ex vivo
Characterization of NB30, a new aminoglycoside derivative
A major limitation in the use of commercially available aminoglycoside antibiotics in humans is their marked toxicity in mammals, which is mainly manifested as nephrotoxicity and ototoxicity (Forge and Schacht 2000; Mingeot-Leclercq and Tulkens 1999). In an effort to address this problem, we have been designing and synthesizing new aminoglycoside-derivatives, and searching for new compounds with improved stop codon read-through activity and lower toxicity in mammalian cells. We have recently synthesized and characterized NB30, a new paromomycin derivative. NB30 induced efficient read-through of the p.R3X mutation in vitro, and had reduced in vitro translation inhibition activity in comparison to commercial aminoglycosides (Nudelman et al. 2006; compound 3). We have thus tested its ability to suppress other PCDH15 nonsense mutations in vitro. NB30 was able to partially suppress all of the tested mutations although suppression rates were lower than the ones achieved by commercial aminoglycosides (Table 1).
We next tested the ability of NB30 to suppress the p.R245X mutation ex vivo, using the PCDH15-myc construct described above. As seen for commercial aminoglycosides, addition of NB30 (1–4 mg/ml) to the culture media of cells transfected with the p.R245X-mutant construct lead to partial read-through of the mutation, as indicated by positive staining of transfected cells with the anti-myc tag antibody (Fig. 2h). Quantitative Western blot analysis revealed that the read-through level obtained with 2 mg/ml of NB30 was approximately 5%. Upon addition of 4 mg/ml of NB30 read-through level increased to over 6% (Fig. 2b).
LD50 values (mg/ml) calculated for different aminoglycosides in kidney-derived cell lines
0.70 ± 0.08
0.49 ± 0.06
0.51 ± 0.07
0.77 ± 0.14
0.90 ± 0.10
1.00 ± 0.06
4.82 ± 0.92
7.03 ± 1.85
7.53 ± 1.21
Usher syndrome is the most common cause of combined sensorineural deafness and blindness, affecting individuals of various origins. USH1 is the most severe form of this genetic condition. A therapy which will delay the onset and/or progression of vision loss in individuals with USH1 will have a remarkable impact on their quality of life. One such possible therapy is suppression of nonsense mutations by small molecules such as aminoglycosides. Clinical trials testing this approach on human subjects have thus far produced encouraging results. In one clinical trial, three of the four DMD patients harboring nonsense mutations in the dystrophin gene showed re-expression of dystrophin in muscle fibers (Politano et al. 2003). Moreover, clinical trials of the effect of gentamicin on nasal potential difference measurements in CF patients carrying stop mutations in the CFTR gene demonstrated significant repolarization of the nasal epithelium, representing at least some restoration of chloride transport (Clancy et al. 2001; Wilschanski et al. 2003). We decided to test this approach as a potential therapy for RP in USH1 patients due to nonsense mutations, initially focusing on nonsense mutations of the PCDH15 gene.
The nature of the tested gene, PCDH15, generates some limitations in terms of experimental methodology. PCDH15 encodes for protocadherin 15, a member of the cadherin superfamily of calcium-dependent adhesion proteins, which plays an important role in the development and maintenance of neurosensory cells in both retina and cochlea (Ahmed et al. 2006; Kremer et al. 2006). The lack of a biochemical assay for measuring protocadherin 15 activity makes it hard to evaluate the level of functional protein produced due to aminoglycoside-mediated suppression. The assays we used, both in vitro and ex vivo, quantify the suppressive effect based on the amount of full-length protein generated. Both assays do not test the actual activity of the full-length protein. The data presented here clearly demonstrate that partial read-through of PCDH15 nonsense mutations can be induced by commercial aminoglycoside antibiotics and their derivatives. The effect on protocadherin 15 activity will have to be tested in vivo, by physiological measures of retinal function in an animal model.
A major limitation in the use of commercially available aminoglycoside antibiotics in humans is their marked toxicity in mammals. The aminoglycoside geneticin (G-418) shows the best termination suppression activity in vitro (Manuvakhova et al. 2000 and the current report); however, its use as a therapeutic agent is not possible since it is lethal even at very low concentrations (Chernikov et al. 2003). Aminoglycosides such as gentamicin and paromomycin, which are clinically used, have adverse side effects, mainly nephrotoxicity and ototoxicity (Forge and Schacht 2000; Mingeot–Leclercq and Tulkens 1999). The origin of this toxicity is still controversial, and probably results from a combination of different mechanisms, including formation of free radicals, disturbance of membrane functions by interaction with phospholipids, inhibition of phospholipases and interference with the translation mechanism (Forge and Schacht 2000). Ototoxicity and nephrotoxicity are the dose-limiting factors in the use of different aminoglycosides, and the use of subtoxic doses in clinical trials probably limits their potential benefits.
The situation described above creates a need to develop new aminoglycoside-derivatives and other small molecules, which maintain their read-through activity, while having reduced toxicity. Several attempts have been made to address this issue, both by us and by other groups worldwide (Mattis et al. 2006; Nudelman et al. 2006; Welch et al. 2007). We have previously designed and synthesized a series of new paromomycin-derivatives, and tested their read-through ability and in vitro toxicity. One of these compounds, NB30, demonstrated efficient read-through activity and reduced in vitro toxicity, in comparison to paromomycin and gentamicin (Nudelman et al. 2006). In the present work, we further characterized this compound, and tested its ability to suppress additional nonsense mutations. Interestingly, NB30-mediated suppression in vitro was lower than suppression achieved by commercial aminoglycosides, even when a higher compound dosage was used (Table 1). However, in cultured cells the effect of NB30 was similar to the effect of gentamicin (Fig. 2b). This finding demonstrates the two major relative advantages of NB30 over commercial aminoglycosides: improved cellular permeability and reduced toxicity.
All natural and semisynthetic aminoglycosides share a similar structure, consisting of several sugar rings that are linked via glycosidic bonds. As a consequence of their polar structure, these drugs inefficiently cross biological membranes. NB30 was generated by selectively removing one sugar ring from the paromomycin structure. It is based on rings I and II of paromomycin, with a 5-amino ribose as a third sugar ring (Nudelman et al. 2006). This more compact structure may lead to improved cellular permeability of NB30, and thus a higher intracellular concentration. In addition, due to its reduced toxicity, NB30 can be used at a higher dosage. For example, in our ex vivo experiments, the maximal dosage of paromomycin and gentamicin used was 2 mg/ml, since higher concentrations of both compounds were extremely toxic to the cells. However, NB30 was used at a dosage of up to 4 mg/ml, which lead to an increase in suppression level. On the basis of our experimental results, the read-through inducing activity of NB30 is higher than paromomycin, and similar to gentamicin. Due to its reduced toxicity, the use of NB30 for suppression of nonsense mutations may be more beneficial, since it is expected to be accompanied by less negative side effects. These observations support our previous findings, and place NB30 as a potential new therapeutic agent for RP and other genetic conditions caused by nonsense mutations. This potential needs to be further evaluated, in terms of in vivo toxicity, and in vivo suppressive activity.
In the present work, we chose to initially test aminoglycoside-mediated suppression of nonsense mutations of one USH1-related gene, PCDH15. However, the research described here will have important implications for the development of targeted therapeutic strategies that are effective for individuals with nonsense mutations of other USH1-related genes, and possibly for individuals with nonsyndromic RP and other genetic conditions as well.
We thank David Bedwell for pDB650, and Ami Aronheim for pCS2 + MT and the anti-myc tag antibody. We thank Sara Selig for help with immunofluorescence, Ido Stahl for technical assistance, and Tom Friedman and Sara Selig for critical reading of this manuscript. This work was supported by research grants from the German–Israeli Foundation for Scientific Research & Development and from the Hedson Fund for Medical Research to T. Ben-Yosef, NIDCD/NIH (1 ZO1 DC000035-09 and 1 ZO1 DC000039-09) to T. Friedman and the Mizutani Foundation for Glycoscience to T. Baasov and T. Ben-Yosef.