Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
X linked retinitis pigmentosa (XLRP) is a heterogeneous disease causing a severe form of retinal degeneration. Patients typically present with night blindness and constricted visual fields within the first two decades of life owing to peripheral photoreceptor degeneration. As the disease progresses, impairment of central vision occurs resulting in loss of visual acuity and complete functional blindness often by the age of 40-50 years.1 The gene that causes one form of this disease, RP2, has recently been positionally cloned and has been shown to account for between 15-20% of XLRP.2 The RP2 gene consists of five exons encoding a polypeptide of 350 amino acids and is ubiquitously expressed. There are currently few functional data available about this protein. One functional clue is a similarity to cofactor C with the predicted RP2 amino acid sequence having 30.4% identity over 151 amino acids.2 Cofactor C was initially thought to play a role in the folding of β-tubulin,3,4 suggesting that RP2 could also be involved in tubulin biogenesis. A recent study has shown, however, that RP2 undergoes N-terminal acyl modification and thus has a predominantly plasma membrane localisation in cultured cells.5 For this reason it seems unlikely that RP2 functions exclusively in tubulin folding and its precise function and any specific role in the retina are at present unknown. Mutation screening in XLRP patients has identified over 20 different pathogenic mutations in the RP2 gene, including missense, frameshift, insertion, and deletion changes.2,6–11 There are also three different identified nonsense mutations causing a premature stop within the first two exons of the gene.2,6–8
In this study, we have investigated a potential drug mediated therapy to restore RP2 function in patients with nonsense mutations, in particular the opal nonsense mutation converting CGA (arginine 120) to TGA (termination codon) at nucleotide position 358 in exon 2, which is present in a large pedigree from Moorfields Eye Hospital.7 The aminoglycoside antibiotics, such as gentamicin, have been shown to suppress premature stop codons both in transcription/translation reactions, cultured cells, and whole animals.12–20 This interesting and potentially clinically beneficial phenomenon is believed to be caused by aminoglycoside antibiotics interacting with ribosomes during translation, reducing the usual stringency of codon-anticodon pairing.12,13 This sometimes results in the insertion of alternative amino acids at the site of the internal stop codon of the mutated gene, thus permitting the ribosomes to continue reading through to the end of the gene and make the full length polypeptide. It has been suggested, therefore, that aminoglycoside antibiotics could be used to treat almost any genetic disease caused by nonsense mutations. As there are several large pedigrees of subjects with this particular type of mutation in RP2, we have investigated the therapeutic potential of aminoglycosides for the RP2 mutation Arg120stop as the first therapy for X linked retinitis pigmentosa. This type of treatment would have several potential advantages over a traditional gene replacement approach. There is already a large amount of pharmacological data available on the use of aminoglycoside antibiotics and the approach requires little information on the function or expression profile of RP2, so this potential therapy could be taken into the clinic relatively quickly.
In this study, we investigated whether treatment with gentamicin could restore full length protein production in an in vitro model. As RP2 appears to be ubiquitously expressed,2,5 we used human lymphoblastoid cells from affected males, with the Arg120stop mutation, as a model to assess the effect of gentamicin on RP2 protein expression. These cells harbour the actual patient mutation in context, not engineered mutations constructed for the experiments which may not have had the same significance for the treatment of our patient group. Demonstration of effective stop codon read through in this clinically relevant mutation would have promised a new potential therapy for the clinic; however, our results suggest that aminoglycoside therapy is not a practical treatment at this stage for the Arg120stop nonsense mutation in RP2.
Human lymphoblastoid cell lines corresponding to a large Moorfields pedigree with the Arg120stop nonsense mutation in RP27 and randomly selected control males were obtained from the European Collection of Cell Cultures (ECACC, Salisbury, UK). Cells were maintained in suspension culture in RPMI 1640 Glutamax-I (Life Technologies, Paisley, UK) supplemented with 10% fetal bovine serum (Sigma, Poole, UK) with media changes every two to three days. For aminoglycoside antibiotic treatment, cells were cultured for 12 hours, three days, and 10 days in the presence of 0-500 μg/ml gentamicin (Sigma), with media changes, where appropriate, every three days.
Transfections and dual luciferase reporter assay
Chinese hamster ovary cells were maintained in Dulbecco's modified Eagle's medium/F12 supplemented with 10% fetal bovine serum. Transfections were performed using lipofectamine reagents (Life Technologies). The reporter plasmid p2luc constructs were a gift from Professor John Atkins (University of Utah). Briefly, 1 × 105 Chinese hamster ovary cells were plated on to 12 well plates and grown for 48 hours. Cells were transfected with 300 ng of plasmid DNA for 15 hours in serum free media. Fresh media containing serum and varying levels of gentamicin were added for 24 hours before being assayed. Cells were lysed and luciferase activity was determined using the dual luciferase reporter assay (Promega, Southampton, UK) according to the manufacturer's protocols. Stop codon readthrough was calculated by comparing the ratio of firefly to renilla luciferase activity in cells transfected with p2luc stop codon constructs, relative to the ratio of luciferase activity in cells transfected with p2luc control constructs.19,21
Preparation of cell lysates
Lymphoblastoid cells were Dounce homogenised on ice in 20 mmol/l Tris-HCL, pH 7.5, 500 mmol/l NaCl, 12.5 mmol/l KCl, 1 mmol/l EDTA, 1 mmol/l dithiothreitol containing a protease inhibitor cocktail (Sigma). The concentration of protein in the homogenates was determined using the Bio-Rad (Hemel Hempstead, UK) DC assay, following manufacturer's protocols.
The cell lysates were prepared for electrophoresis by the addition of sample buffer (100 mmol/l Tris-HCl, pH 6.8, 2% glycerol, 0.4% SDS, 1% 2-β mercaptoethanol, 0.01% bromophenol blue, final concentration) and heating to 96°C for five minutes. A total of 50 μg or 100 μg of total protein were loaded on a 12% SDS-polyacrylamide gel and after electrophoresis were electroblotted onto nitrocellulose membrane (Schleicher and Schuell, Dassel, Germany). Non-specific binding sites on the nitrocellulose were blocked by incubation with 5% marvel, 1 × PBS, 0.05% Tween 20 overnight at 4°C. The blots were hybridised with sheep polyclonal antisera S974, raised against full length recombinant His tagged RP2, and affinity purified5 at a titre of 1:1000 for one hour at room temperature. Anti-sheep polyclonal secondary antibody conjugated to peroxidase (Sigma) was used at a titre of 1:2000. Immunoreactive bands were visualised using enhanced chemiluminescence (Amersham Pharmacia, Little Chalfont, UK). Standards containing known amounts of the recombinant His tagged RP2 were included to ensure sufficient minimum detection levels of protein. The immunoreactive bands were quantified and compared using imaging software (Kodak Digital Science™ 1D image analysis software, Eastman Kodak Company, Rochester, NY).
RNA was extracted from patient and control lymphoblastoid cells using the RNeasy kit from Qiagen (Crawley, UK) following the manufacturer's protocols. RT-PCR was performed using the reverse transcription system from Promega (Southampton, UK) to obtain the first strand cDNA using oligo(dT) primers as per the manufacturer's protocol. PCR was performed using exonic primers RP2-5 (GCGGATCCGCCATGGGCTGCTTCTTCTCCAAGAG) and RP2-1-199 (GGATCCTCTAGATCAAACATAGTCCTGAACCACAGC). The following conditions were used: 94°C for five minutes, then 40 cycles of 94°C for 30 seconds, 60°C for 30 seconds, 72°C for one minute, then a final extension of 72°C for five minutes.
Confocal microscopy of cells
Lymphoblastoid cells from RP2 patients and male controls were washed with PBS and trypsinised to reduce clumping of the cells. After further PBS washes, the cells were fixed in 3.7% formaldehyde followed by detergent permeabilisation in 0.05% Triton X-100. Non-specific binding sites were blocked by incubation with 3% bovine serum albumin, 10% normal rabbit serum, and PBS for one hour at room temperature. Affinity purified S974 was used at a titre of 1:250 for one hour at 20°C. Cy™3 conjugated anti-sheep secondary antibody (Jackson ImmunoResearch, West Grove, USA) was used at a titre of 1:150 for one hour at 20°C. The stained lymphoblastoid cells were fixed to glass slides by cytospin centrifugation and analysed using a Zeiss laser scanning confocal microscope.
RP2 expression in control and patient lymphoblastoid cells
The expression of RP2 was investigated in lymphoblastoid cells from an unaffected control male and compared with RP2 expression in human SH-SY5Y neuroblastoma cells and human retina by western blotting using antisera S974, as previously characterised5 (fig 1). The predicted molecular weight of the 350 amino acid RP2 is 39.6 kDa. A band of approximately this size was observed in all of the samples. The expression levels of RP2 in the lymphoblastoid cells were much higher than in human retina but lower than in the SH-SY5Y cells; these differences agree with previously estimated RP2 expression levels in human retina and SH-SY5Y cells.5 Interestingly, sera S974 appeared to cross react with another protein of 68 kDa that was present only in the human lymphoblastoid cells and not in retina, SH-SY5Y cells, or any other tissues examined. The reactivity of S974 with this 68 kDa protein and RP2 could be removed by blocking S974 with 100 μg of recombinant RP2 protein during the antibody incubations (data not shown), suggesting that this lymphoblastoid protein may share epitopes with RP2.
RP2 mRNA expression in the lymphoblastoid cells from males affected with the Arg120stop mutation and cells from control males was also investigated. A full length RP2 transcript was detected in both the male patient and the control male lymphoblastoid cells (data not shown). The expression levels of RP2 mRNA, however, were greatly reduced in the male patient cells compared to the male control cells, suggesting that RP2 containing the Arg120stop mutation may undergo nonsense mediated decay.22
RP2 protein expression in males affected with the Arg120stop mutation and control males was investigated in lymphoblastoid cells (fig 2A). The expression levels of RP2 protein were compared between the cells from five male RP2 patients and five control males. A band of the predicted size for RP2 was observed consistently in all of the control cells, but no band of the expected size for full length RP2 or the predicted 119 amino acid truncated protein product was detectable in any of the patient cells. We have recently shown that sera S974 is immunoreactive with the first 15 amino acids of RP2 (data not shown), indicating that any truncated protein produced is rapidly degraded. The immunoreactive band corresponding to the 68 kDa protein was observed in both the patient and the control lymphoblastoid cells, suggesting that, although the protein may share epitopes with RP2, it does not correspond to a post-translational modification of RP2 and is not a product of the RP2 gene (fig 2B). Although the basis for the cross reactivity between S974 and this 68 kDa protein remains to be resolved, it did provide a useful internal loading and transfer control for the western blot assays.
Determining the minimum detection levels of RP2
It was necessary to determine the minimum detection level of the RP2 protein before analysing the effects of aminoglycoside therapy on RP2 expression. From comparison to recombinant RP2 standards, we estimate that RP2 protein represents approximately 0.085% of the total protein in lymphoblastoid cells. This figure is higher than in other human tissues where RP2 protein has been estimated to represent approximately 0.01% of the total protein.5 A standard curve of recombinant His tagged RP2 protein of known concentration was used for immunoblotting (data not shown). Using our methodology, we could easily detect below 0.5 ng of recombinant RP2, or approximately 0.4% of the total RP2 expressed in control lymphoblastoid cells. This detection limit was confirmed by mixing RP2 patient lymphoblastoid cell lysates and control lymphoblastoid cell lysates in different fixed ratios and loading the same amount of total protein into each well. This confirmed that the western assay could detect at least 0.5% of control RP2 levels (fig 2B).
Subcellular localisation of RP2
The subcellular localisation of RP2 in lymphoblastoid cells from control males and RP2 Arg120stop patients was investigated by immunofluorescent staining and confocal microscopy. RP2 protein localisation in the lymphoblastoid cells from control males was predominantly on the plasma membrane of the cells (fig 3A). The intensity of plasma membrane staining was variable within each individual cell and also between cells. This observation is in agreement with recently published data on the subcellular localisation of RP2 in other cultured cell types.5 In addition to the plasma membrane staining, a diffuse stain was also observed throughout the cells but at a much lower intensity. The plasma membrane staining was not seen in lymphoblastoid cells from patients (fig 3B). The staining pattern in the cells from the patients was much more diffuse and did not appear to have a specific subcellular localisation. A similar staining pattern was seen in lymphoblastoid cells from the control males and RP2 patients probed with the Cy™3 conjugated anti-sheep antibody alone, without a previous S974 primary antibody incubation (fig 3C, D). Cross reactivity with human tissues appears to be a problem in many commercially available anti-sheep antibodies. Therefore, the diffuse intracellular staining pattern may in some part correspond to cross reactivity between the human lymphoblastoid cells and the secondary antibody and the reactivity between S974 and the 68 kDa protein.
Gentamicin treatment of cells
To determine whether gentamicin could lead to suppression of the premature stop codon in RP2 patients, lymphoblastoid cells from male patients with the Arg120stop mutation were cultured in the presence of various gentamicin doses ranging from 0-500 μg/ml for 12 hours, three days, and 10 days. At doses above 500 μg/ml, significant cytotoxicity was observed within the 12 hour period. Even with sensitive, calibrated western blots with detection levels down to below 0.4% of control RP2 levels, we could not detect any increase in full length or truncated RP2 protein expression in these cells after treatment with gentamicin at all of the tested doses and time points (fig 4).
The efficacy of the gentamicin used in this study was confirmed using a dual luciferase reporter system, which uses firefly and renilla luciferase coding sequences separated by stop codons or control codons in different sequence contexts.19,21 We examined the effect of gentamicin on two constructs encoding the opal stop codon UGA. We compared the intervening sequence that contained the sequence UGAG, the stop codon context in the RP2 Arg120stop mutation, with the intervening sequence UGAC, the opal stop codon context found to achieve the greatest levels of aminoglycoside induced readthrough.19 Using the same methodology used in these earlier studies, we achieved gentamicin induced stop codon readthrough of between 1.5% and 5% for the sequence UGAC, similar to readthrough levels achieved in previous studies, using gentamicin at concentrations of up to 500 μg/ml. However, the base immediately after the stop codon had a profound effect on the efficiency of gentamicin induced translation. By substituting the cytosine base for a guanine base, as found in the Arg120stop mutation, readthrough levels were reduced by over 75%.
Aminoglycoside antibiotics have been shown to suppress nonsense mutations both in vitro and in vivo.12–20 The primary aim of this study was to investigate whether this phenomenon could be used as a potential treatment for RP2 patients with the Arg120stop nonsense mutation.7 We have shown that, at present, aminoglycoside antibiotic therapy does not appear to be a viable therapy for RP2 patients with this mutation. No RP2 protein could be detected in cells from affected males treated with a wide range of doses and treatment time points of the aminoglycoside antibiotic, gentamicin.
Studies on the cystic fibrosis transmembrane conductance regulator (CFTR) suggested that full length CFTR protein production in transfected HeLa cells and bronchial epithelial cells expressing an opal (UGA) nonsense mutation could be restored to levels of 20-35% of that in wild type cells after treatment with aminoglycoside antibiotics at doses of 100-200 μg/ml.14,15 Furthermore, the CFTR produced appeared to be functional by investigating cyclic AMP-activated chloride channel activity in the cells. Similar observations were made in an in vivo model for Duchenne muscular dystrophy, the mdx mouse,16 where, before aminoglycoside antibiotic treatment, muscle expressed no full length dystrophin protein. After treatment with gentamicin, however, dystrophin levels were detectable at up to 10-20% of that in muscle of control mice expressing the wild type dystrophin protein, and the mice displayed a significant degree of protection against contractile induced damage, suggesting some functional recovery. Recent studies, however, have brought into question the accuracy of these high levels of protein restoration and suggest that the true level of readthrough may be much lower, possibly less than 5%, while highlighting the importance of the context of the stop mutation.18,19 This lower level of protein restoration may still have clinical applications. A study of premature stop codons in the Hurler syndrome gene, IDUA, has suggested that restoration of activity to 3%, mediated by readthrough, may be sufficient to mediate some cellular recovery.20 In our study, however, we observed no restoration of full length RP2 protein production to less than 0.5% by western blotting in our opal mutation cell lines on gentamicin treatment at the doses and times that have been shown to be effective in these paradigms. As the function of RP2 is unknown at present, it was not possible to investigate the restoration of any function in the cells or to know what levels of RP2 expression would be necessary for some function to be restored. Once the function of RP2 has been elucidated, a more sensitive assay may be used to test the effectiveness of aminoglycoside antibiotic treatment for the Arg120stop nonsense mutation.
The data in this study have shown a clear difference in RP2 subcellular localisation between the RP2 patient and control lymphoblastoid cells and this could have been used to determine if any full length protein produced was correctly targeted. However, owing to the background cross reactivity of the secondary antibody and the potential problems caused by the cross reacting 68 kDa protein, immunoblotting represented a more sensitive detection method for RP2. By immunoblotting we were able to detect full length protein down to levels of less than 0.4% of wild type protein.
Although the potential use of aminoglycoside therapy for nonsense mutations is very attractive, our observations that full length protein production was not restored in cells with the Arg120stop nonsense mutation show that there may be severe limitations to their application. A major consideration is that mRNAs with premature stop codons may be subject to their own quality control via nonsense mediated decay (NMD).22 We have observed that although the patient cell lines do express the RP2 mRNA, it appears to be present at much lower levels in the patient cells than in controls. This level of RNA surveillance reduces the amount of mutant mRNA available for translation and readthrough assistance from the aminoglycoside. Furthermore, the type of nonsense mutation and composition of the RNA sequence flanking the stop codon may also have a major effect on the efficiency of aminoglycoside mediated readthrough.12,13,18,19,23,24 Although the opal premature stop mutation, UGA, which was investigated in our study, appears to show greatest translational readthrough, the nucleotide in the position immediately downstream from the stop codon appears to be a major modifier (in the order C > U > A > G).18,19 The presence of a G immediately after the stop codon, as in the Arg120stop patients, reduced readthrough levels by up to 75%. On the basis of the type of mutation and its immediate context, it could be predicted that readthrough of up to 3% could occur in our model.19 However, the effects of other context characteristics in the mRNA on readthrough are unknown, such that the combination of mutation context and the lower levels of RP2 mRNA caused by NMD result in levels of readthrough below the sensitivity of our western assay.
It should also be noted than in the in vivo model of mdx, not all animals with the same nonsense mutation responded to the gentamicin treatment.16 This raises the possibility that even if the “context” of the mutation is conducive for the aminoglycoside treatment to be effective, other factors may modulate the beneficial effect of the drug. As variation within different people may be important, we investigated the effect of gentamicin on the Arg120stop mutation covering all of the previously tested doses and time points in lymphoblastoid cells from five male RP2 patients. It should be noted that doses of gentamicin higher than 500 μg/ml were not used in this study as there was evident toxicity to the lymphoblastoid cells at 500 μg/ml and higher doses. It is possible, however, that the lymphoblastoid cells have cell line specific differences in RP2 expression, NMD, and aminoglycoside mediated readthrough distinct from retina. Nevertheless, it must be taken into consideration that these patient derived cell lines possess the stop mutation within the actual context of the RP2 gene and within the context of the patients' genetic background. Therefore, they represent a much better in vitro model system in which to study these compounds' therapeutic potential than cell free transcription/translation or transfection of standard cell lines.
Even if aminoglycoside therapy had proven to be a potential treatment for RP2 patients, it would still have been important to consider their possible toxicity. Barton-Davies et al16 have suggested that administration of gentamicin below the maximum recommended human dosage (for antibiotic use) could prove effective in restoring protein function. There are, however, no data available on the consequences of the long term use of this drug, particularly at the high doses that may be required for effective readthrough therapy. Retinal toxicity of gentamicin should also be given careful consideration before any treatment of patients with ocular disease. The toxicity of gentamicin in the retina is well documented, the majority of cases being reported after its prophylactic use in vitrectomy or routine ocular surgery, at doses considered to be safe.25 There is also evidence of gentamicin toxicity in primates after intravitreous injection, leading to damage within the inner retinal layers.26 Aminoglycoside antibiotics cause full length polypeptides to be made as they interfere with the usually stringent codon-anticodon pairing during translation, causing alternative amino acids to be inserted in the place of the premature stop codon. This means that as well as inserting the correct amino acid to produce functional full length protein, it may also introduce erroneous insertions of other incorrect amino acids leading to the possible production of aberrant proteins with unknown toxicity. Such proteins may not fold correctly and form aggregates, or may exhibit gain of function effects.
Another possible treatment for diseases caused by premature stop mutations could be the use of suppressor tRNA gene therapy,27 as opposed to conventional gene replacement therapy. This is enabled by engineering mutant tRNAs that can read premature stop codons as sense codons and hence restore full length polypeptide production.28 This may represent a more focused approach to the treatment of diseases caused by nonsense mutations, but may be unsuitable for the treatment of RP2 at present. As the function and localisation of the protein are at present unknown, it would be difficult to direct this treatment towards the necessary specific target cell, and similar problems of erroneous protein production would need to be considered as a possible side effect. However, as more information becomes available about the function of RP2, this may be a viable therapy, as there are a large number of patients with nonsense mutations in RP2.
An extremely useful finding of this study is that it should be possible to use immunoblotting or immunocytochemistry as a diagnostic test for mutations in RP2. As the majority of RP2 mutations are protein truncating, most patient mutations could be detected by using immunoblotting for the presence of full length protein. Alternatively, immunocytochemistry could be used to detect mutations that also affect correct protein targeting.5 Immunoblotting and/or immunocytochemistry with a suitable antibody, such as the sheep sera S974, as a primary screen would be relatively inexpensive and less time consuming than sequencing the whole of the RP2 and/or the phenotypically indistinguishable RP3 gene for every potential XLRP patient. The RP2 protein appears to be ubiquitously expressed2,5 and we were able to show that protein truncations can be detected in lymphoblastoid cells. Therefore, a diagnostic test could be carried out using lymphocytes from blood samples that would usually be taken for DNA analysis. Choroideremia, another X linked retinal dystrophy, already has a protein based diagnostic test29 and, as more genes are cloned, many other diseases may be diagnosed using this type of method.
The clinical potential of using aminoglycoside antibiotics to suppress premature stop mutations has been heightened by recent results showing their effectiveness both in vitro and in vivo. The purpose of this study was to determine whether aminoglycoside therapy could be a viable treatment for X linked retinitis pigmentosa patients with the Arg120stop nonsense mutation in RP2.
The expression of RP2 mRNA and protein in Arg120stop patient lymphoblastoid cells were compared to control cells from unaffected males. Lymphoblastoid cells from male patients with this mutation were treated with a range of gentamicin doses in an in vitro system and expression levels of RP2 protein were compared to those in control lymphoblastoid cells from males unaffected by RP2 mutations using immunoblotting. Differences in expression levels of RP2 were determined by densitometry. RP2 mRNA was detectable in both patients and controls, although expression levels were reduced in the patient cells. The RP2 protein was only detectable in control lymphoblastoid cells and not in patient cells by both western blotting and immunocytochemistry. When cells containing the premature stop mutation were treated with gentamicin, under a wide range of conditions, no induced expression of full length RP2 protein could be detected down to 0.4% of control levels.
Aminoglycoside antibiotic therapy does not appear to be a viable treatment for RP2 patients with the Arg120stop nonsense mutation at the present time. The use of immunoblotting or immunocytochemistry of peripheral blood cells could, however, be a useful tool for the rapid diagnosis of new patients with protein truncating or targeting defect mutations in RP2.
This study has shown that, unfortunately, aminoglycoside therapy does not appear to be a viable treatment for Arg120stop RP2 patients at this time. We propose, however, that immunoblotting or immunocytochemistry for RP2 may be a potentially useful diagnostic tool. As more becomes known about the function of RP2, and more specifically its function in the retina, more potential therapies may be investigated.
We thank Professor John Atkins for providing the p2luc vector constructs. This work was supported by The Wellcome Trust and Fight For Sight. CG is a Fight For Sight Prize Student.