|Home | About | Journals | Submit | Contact Us | Français|
Exceptional progress has been made during the past two decades in identifying genes causing inherited retinal diseases such as retinitis pigmentosa. An inescapable consequence is that the relationship between genes, mutations, and clinical findings has become very complex. Success in identifying the causes of inherited retinal diseases has many implications, including a better understanding of the biological basis of vision and insights into the processes involved in retinal pathology. From a clinical point of view, there are two important questions arising from these developments: where do we stand today in finding disease-causing mutations in affected individuals, and what are the implications of this information for clinical practice? This perspective addresses these questions specifically for retinitis pigmentosa, but the observations apply generally to other forms of inherited eye disease.
The goal of this perspective is to summarize the current state of the molecular diagnosis of retinitis pigmentosa (RP) and its relevance to clinical practice. The comments are limited largely to nonsyndromic, nonsystemic forms of RP, using autosomal dominant RP (adRP) as an example. It is important to recognize, though, that what is true for simple RP is true in general for most other forms of inherited retinal degeneration. There has been rapid progress in identifying genes and mutations causing all forms of retinal disease, including multifactorial diseases such as age-related macular degeneration. Of course, the specific genes are different and the clinical findings are distinct, but the implications for clinical practice are similar. For example, what is true for RP alone is also true for Usher syndrome, Bardet-Biedl syndrome, and familial macular degeneration. That is, many genes and mutations are also known for these diseases and have relevance to clinical practice. A list of genes causing RP and other retinopathies can be found at the RetNet Web site.1 A number of recent reviews address the biological bases of RP.2-5
Retinitis pigmentosa encompasses many different diseases with many distinct causes and diverse biological pathways but with overlapping symptoms and similar consequences.4 It is no more a single disease than is “fever of unknown origin.” If one word more than any other comes to mind in describing RP, it is complicated. There are dominant, recessive, and X-linked forms of inheritance in addition to rare mitochondrial and digenic forms. Retinitis pigmentosa may occur alone or as part of a more complex syndrome. Even simple RP is strikingly complicated. Each genetic type is caused by mutations in several or many different genes. For most genes, many different mutations with similar consequences are known, yet other mutations in the same gene may cause different diseases. Perhaps most surprisingly, the same mutation in different individuals may cause distinctly different symptoms, even among individuals within the same family.
Ironically, the great success in identifying genes and mutations causing RP during the past 2 decades has revealed the extent of the complexity but also offers hope of taming it—by defining RP at a molecular level rather than clinically. Still, in spite of the progress in genetics, a careful clinical description is and will be an essential prerequisite for molecular diagnosis. Further, the molecular description of RP is intrinsically complicated. Molecular diagnosis alone will therefore neither replace clinical testing nor fully resolve the complexity. Nonetheless, clinical testing coupled with molecular diagnosis of RP is a powerful combination of approaches for diagnosing patients and families and will eventually lead to treatment and prevention.
Retinitis pigmentosa is a class of diseases involving progressive degeneration of the retina, typically starting in the midperiphery and advancing toward the macula and fovea.6 Typical symptoms include night blindness followed by decreasing visual fields, leading to tunnel vision and eventually legal blindness or, in many cases, complete blindness. Clinical hallmarks are an abnormal fundus with bone-spicule deposits and attenuated retinal vessels; abnormal, diminished, or absent electroretinographic findings; and reduced visual fields. Symptoms typically start in the early teenage years, and severe visual impairment occurs by ages 40 to 50 years. However, there are early-onset forms of RP (the earliest is indistinguishable from Leber congenital amaurosis [LCA]) and other late-onset or even nonpenetrant forms. The underlying genetic cause is a useful predictor of severity in some cases, but the inverse is usually not true: the phenotype alone is not a good predictor of the gene or mutation.
In addition to simple forms of RP, there are syndromic forms involving multiple organs and pleiotropic effects as well as systemic forms wherein the retinal disease is secondary to a systemwide pathology (although the distinction is more historic than biological). The most frequent form of syndromic RP is Usher syndrome, which manifests as early-onset or congenital hearing impairment followed by development of RP by the early teenage years.7,8 The second most common syndromic form is Bardet-Biedl syndrome, which includes RP, polydactyly, obesity, renal abnormalities, and mental retardation.9 In addition, many other complex, pleiotropic conditions include RP as a component.1
Table 1 shows the overall prevalence of RP and the proportions of the most common genetic subtypes. Retinitis pigmentosa, broadly defined to include simple, syndromic, and systemic disease, has a worldwide prevalence of 1 case per 3000 persons to 1 case per 7000 persons.10 This is a relatively narrow range of estimates given the inherent difficulty of counting RP cases in large populations. In contrast, estimates of the fractions of the various genetic subtypes vary 10-fold between studies (summarized by Haim10). Part of the reason is that definitions and clinical criteria differ significantly between surveys. However, there are also substantial differences between populations in the prevalence of specific mutations and, hence, in the proportion of specific genetic types. Therefore, the proportions in Table 1 should be taken with a grain of salt.
Nonsyndromic, nonsystemic RP encompasses 65% of all cases, or about 65 000 people in the United States. Of the total number of nonsyndromic, nonsystemic cases, roughly 30% are adRP, 20% are autosomal recessive RP, 15% are X-linked RP, and 5% are early-onset forms of RP that are typically diagnosed as recessive LCA. The remaining cases, at least 30%, are isolated or simplex cases. The simplex cases are likely to include many individuals with recessive mutations, but dominant-acting de novo mutations are also found in these individuals.11,12
In the past few decades, rapid progress has been made in finding genes and mutations causing inherited retinal diseases. The Figure shows the progress in gene identification since 1980.1 Genes and the underlying mutations within these genes have been identified by a number of methods. Many genes were first localized to a chromosomal site by linkage mapping in families or, more recently, by homozygosity mapping.13,14 Once mapped, the underlying gene can be found by various targeted sequencing strategies. Other disease genes were identified by sequencing candidate genes in selected patient populations. A retinal gene may be a disease candidate because of its functional properties, because it is similar to a gene known to cause retinal disease, or because it is the cause of retinal disease in an animal model.
To date, 181 genes causing inherited retinal diseases have been mapped to a specific chromosomal site, and 129 of these have been identified at a sequence level. Also, at least 5 additional genes are known to contribute to the lifetime risk of multifactorial diseases such as age-related macular degeneration.15-22
What is true for retinal disease genes in general is especially true for RP. Currently, mutations in 17 different genes are known to cause adRP, mutations in 25 genes cause recessive RP, mutations in 13 genes cause recessive LCA, mutations in 2 genes cause dominant LCA, and mutations in 6 genes cause X-linked RP.1 Table 2 lists the genes that are currently known to cause nonsyndromic, nonsystemic RP. However, a simple listing of genes in each category is misleading because many genes can cause more than 1 form of disease. For example, although rhodopsin mutations usually cause dominant RP, other rare rhodopsin mutations cause recessive RP. Mutations in NRL can also be either dominant or recessive acting. Further, mutations in some genes, such as RDS, can cause dominant RP, dominant macular degeneration, or other distinct forms of retinopathy. Therefore, Table 2 also lists the alternate phenotypes that can arise for mutations in RP genes and lists some genes in more than 1 section.
In total, mutations in 53 genes are known to cause nonsyndromic, nonsystemic RP or LCA (counting each gene once only, even if it causes more than 1 type of retinopathy). As stunning as this number may be, a question more important than the number of genes is the total fraction of patients in whom disease-causing mutations can be detected. In other words, how close are we to knowing all of the RP genes?
One way to answer this question is to summarize the fraction of mutations detected in each gene based on surveys of appropriate patient populations. Table 3 is a compilation of the percentage of patients with detectable mutations in each major RP gene as reported in representative surveys. A gene is “major” if it accounts for at least 1% of cases. In summary, with a number of simplifying assumptions, it is now possible to detect disease-causing mutations in 56% of patients with adRP, roughly 30% of patients with recessive RP, more than 70% of patients with recessive LCA, and nearly 90% of patients with X-linked RP. This is a remarkable achievement given that the first gene known to cause RP, the rhodopsin gene, was described only 17 years ago.40,41
The percentages in Table 3 come with several caveats. First, many of the numbers are “soft” because disease definitions are not consistent between reports, sample sizes may be small, different segments of the gene may have been screened, and the definition of a mutation differs significantly from study to study. In fact, very few published percentages include confidence intervals, which are usually large. Further, most of these studies are of Americans of European origin and Europeans. Other ethnic and geographic groups have different fractions of disease-causing mutations.42,43 Finally, these are the fractions of mutations detected in carefully designed studies with optimal methods; screening in practice may be less efficient.
But caveats aside, across all of the categories of inherited retinopathy, careful screening of known disease genes leads to detection of pathogenic mutations in 25% to 90% of patients, an extraordinary accomplishment. At the same time, however, linkage studies and other evidence show that that there are more, perhaps many more, RP genes to be found.
To give a more detailed perspective, what follows is a look at the genes and mutations causing just 1 form of retinal disease, adRP. However, many of the conclusions from the study of adRP are broadly applicable to other inherited retinal diseases. Therefore, this section ends with observations that apply generally to all forms of RP.
In a recent survey, we tested a panel of affected individuals from 200 families with adRP for mutations in most of the known dominant RP genes (Table 2).23 To be included in the study, a family had to have a diagnosis of adRP by a knowledgeable clinical specialist and either 3 affected generations with affected females or 2 affected generations with male-to-male transmission. The latter requirement was to reduce the likelihood of including families with X-linked RP. This possibility arises because some mutations in the X-linked gene RPGR affect female carriers; thus, the disease in these families can be misinterpreted as adRP.44-46
The cohort of patients with adRP was screened (largely by DNA sequencing) for mutations in the protein-coding regions and intronexon junctions of all adRP genes or gene regions causing at least 1% of cases. Open reading frame 15 (ORF15), the “hot spot” for dominant-acting mutations in RPGR, was also tested in families without male-to-male transmission. Determining whether a novel, rare variant is pathogenic can be challenging.47 We used several computational and genetic tools for this purpose.23 Generally, once a definite disease-causing mutation was identified in a family, other genes were not tested further in these individuals.
We found definite or probable mutations in 53.5% of the families with adRP. In subsequent studies, we tested several of the remaining families for linkage to genetic markers within or close to the known adRP genes and to RPGR.24 The logic here was to uncover mutations that might have been missed by sequencing or to locate genes that have been mapped but not identified yet. In 1 large family, we found linkage to the PRPF31 gene, even though careful resequencing failed to disclose a DNA change. Further testing revealed that affected members of the family have a complex deletion and insertion in PRPF31. This rearrangement was not detected earlier because only the nondeleted, homologous chromosome was sequenced; that is, the deletion is “invisible” to sequencing.
We then tested the remaining families for deletions in PRPF31 using multiplex ligation-dependent probe amplification (MLPA).48,49 Surprisingly, we found 4 large deletions, including 2 that encompass genes adjacent to PRPF31.24 This brings the fraction of detected mutations to 56% (Table 3).
These studies have a number of implications that go beyond just adRP. First, 14 different, common mutations account for up to 30% of the families with adRP in this survey; that is, each of these mutations accounts for at least 1% of the cases.23 Thus, screening for this handful of mutations alone will resolve at least 30% of the cases. Common mutations are found in other RP genes, and numerous inexpensive, high-throughput techniques exist for detecting these variants.50,51
Second, another 20% of mutations were novel and could only be detected by sequencing entire genes. Further, each novel mutation requires careful evaluation of pathogenicity. As a consequence, the main bottleneck in genetic testing of patients with RP is the need to screen and analyze many genes by expensive, time-consuming methods. Fortunately, promising high-throughput resequencing techniques, such as microarray gene chips, may relieve this bottleneck.52 Nonetheless, interpretation of novel, rare variants will still require professional evaluation.47
Third, some families thought to have adRP actually have digenic or X-linked mutations. Digenic RP is the result of 1 mutation in RDS and a second in ROM1.53 Different individual mutations in RDS and ROM1 can cause adRP, but each of the digenic mutations alone is not pathogenic. Digenic and polygenic inheritance is true of other forms of retinal disease, such as Bardet-Biedl syndrome, which can be “triallelic.”9 Another misleading mode of inheritance among families diagnosed with “adRP” is X-linked inheritance of RPGR mutations with significant disease in carrier females.44-46 Both of these phenomena are important reminders that the molecular diagnosis can radically change genetic counseling.
Fourth, at least 2.5% of adRP mutations are genomic rearrangements or deletions in PRPF31 that are not detectable by conventional screening methods.24 Whether there are disease-causing deletions in other adRP genes or in recessive or X-linked genes is an active area of research. This is likely, though, because deletions are a common cause of other inherited and acquired diseases.54-56 For example, large deletions cause up to 17% of familial breast cancer.57
The existence of disease-causing deletions has significant implications for molecular testing of patients with RP. For one, routine testing methods may miss deletions (eg, sequencing does not detect the breast cancer deletions). For another, deletions may explain reported anomalies in the frequency and segregation of RP mutations. If so, here again, the molecular diagnosis will affect counseling. Finally, this finding suggests that there may be other subtle mutations in known RP genes that are missed by standard methods.
Fifth, there are definitely additional, unknown adRP genes. We failed to detect mutations in 40% of the families we tested. Some, but not all, of the remaining mutations may be deletions or subtle changes in known genes that have not been detected to date.58 Linkage mapping continues to locate new adRP genes—most recently RP31 and RP33.59,60 Likewise, new recessive and X-linked genes are reported regularly.1
It is impossible to predict whether there are several or many more RP genes that have yet to be discovered. Completion of the Human Genome Project, new high-throughput screening methods, and development of powerful bioinformatic approaches have dramatically reduced the time it will take to find new genes. In spite of these technical advances, the need for thorough, knowledgeable, innovative clinical characterization of patients and families has never been greater.
What does the current state of RP genetics say of the future? A reasonable hope is that within 5 years, molecular testing of newly diagnosed patients with RP will be a routine part of clinical practice and will uncover the underlying disease-causing mutation (or mutations) in at least 90% of cases. For this hope to come true, 4 conditions must be met:
Before addressing these necessary conditions, it is worth asking why finding the underlying disease-causing mutation should matter to the patient or the clinician. After all, RP is currently an untreatable condition, so wouldn't the molecular information be of no use?
There are several compelling reasons why molecular testing is important for clinical care. For one, identifying the underlying mutation(s) can establish the diagnosis, which may be problematic otherwise. This is particularly important for childhood retinopathies wherein the molecular diagnosis may portend distinctly different clinical outcomes.33,61 Also, knowing the genetic cause is essential for family counseling and for predicting recurrence risk and prognosis. In addition, each new mutation that is found contributes to a better understanding of ocular biology. Finally and of the most importance, the era of gene-specific and mutation-specific treatments for inherited retinal diseases is quickly approaching.62,63 Knowing the underlying genetic cause will be essential for enrolling patients in clinical trials, a few of which have begun already or will begin shortly.64,65 It is a safe prediction that in the near future, there will be many more treatment and prevention strategies based on knowledge of the underlying mutation(s) in affected individuals and families.
Then, how close are we to routine molecular diagnosis of RP? Identification of new RP genes is proceeding swiftly. An educated guess (at best) is that most of the major genes, at least in Americans of European origin and Europeans, will be found within 5 to 10 years. Whether current screening methods, such as sequencing or microarray testing, can detect all or even most of the mutations in known genes is debatable. Not all of the gene regions that could harbor mutations are tested routinely. For example, large intervening sequences and noncoding regulatory regions are usually ignored. Also, current methods do not detect large deletions or rearrangements. Venturing another educated guess, though, existing methods and methods under development will be able to detect most mutations, ie, more than 90%, within 5 years.
Currently, the greatest road-block to molecular diagnosis of RP is the availability of genetic testing. Large commercial interests have not yet entered the field, primarily because there are so many genes to test and so many inherent complications. However, methods for rapid, inexpensive detection of known RP mutations exist today and will be routinely available soon.51,66 Also, targeted screening of genes and gene regions that are frequent causes of inherited retinal diseases is being offered on a fee-for-service basis by a few institutions in the United States and Europe (see the GeneTests Web site for further information67). In addition, the National Eye Institute has recently developed a program, eyeGENE™,68 to facilitate genetic testing of inherited eye diseases. Finally, it is reasonable to expect that new high-throughput sequencing methods will make genetic testing of all diseases affordable and efficient within 10 years.69
In our opinion, the major impediment to routine molecular diagnosis of RP is not technical or commercial but rather informational. No aspect of understanding, interpreting, and explaining the molecular causes of RP is routine. Skilled, informed clinical diagnosis must precede testing. Even if genetic testing is standardized, interpretation of novel variants will require sophisticated analysis. Understanding the results of genetic testing will be challenging, especially if novel findings such as polygenic inheritance are involved. Making sense of this to patients and families in a helpful and supportive way will require good counseling skills. Finally, when gene-specific and mutation-specific treatments become available, which is inevitable, even greater levels of knowledge and understanding will be demanded.
None of this is unique to RP: molecular diagnostics will enrich all aspects of medical care in future years. What is unusual, though, is the extent of the current knowledge of the molecular causes of inherited retinal diseases and the recognition of the underlying complexity. Thus, clinical ophthalmology has the unique opportunity to prepare for the near future by enhancing training in genetics, incorporating genetic counseling at all levels of care, and developing specialized centers for the diagnosis and treatment of inherited eye diseases. Another reasonable prediction is that the ophthalmology profession will lead the way for other branches of medicine.
Funding/Support: This study was supported by grants from the Foundation Fighting Blindness, the William Stamps Farish Foundation, the Gustavus and Louise Pfeiffer Research Foundation, and the Hermann Eye Fund as well as by grants EY007142 and EY014170 from the National Institutes of Health.
Financial Disclosure: None reported.