Quantitative Autofluorescence and ABCA4 Disease

The manuscript by Müller et al.,¹ describing a study in which a small group of 26 carriers of pathogenic alleles of the Stargardt gene, ABCA4, were imaged by quantitative autofluorescence (qAF), was of particular interest to us, as we had very recently published a similar study on a cohort of 75 individuals.² Although the results of Müller and colleagues¹ were very similar to ours, their interpretation of the results was often inconsistent with the established knowledge in the field. 

First, the title: “Monoallelic ABCA4 Mutations Appear Insufficient to Cause Retinopathy: A Quantitative Autofluorescence Study,” is confusing. Because all diseases caused by mutations in the ABCA4 gene are recessive (i.e., require mutant ABCA4 alleles on both chromosomes for an expression of a phenotype), the carriers of ABCA4 alleles are, by genetic definition, asymptomatic. The argument of Müller et al.¹ in the introduction, that “there has been an ongoing debate on whether or not monoallelic ABCA4 mutations might also cause disease at the benign end of the spectrum of ABCA4-related retinopathy,” is incorrect, because there has been no such debate. What has been debated is whether ABCA4 carriers are more susceptible to AMD. Regardless the status of that debate, it has never been suggested that carrying one ABCA4 pathogenic allele alone will result in a disease phenotype. Age-related macular degeneration is a complex trait that, by definition, suggests that a combination of many alleles in many genes, together with environmental factors, cause the late-onset disease. It is therefore clear that one mutation in one gene cannot cause AMD; if this were correct, then we would be dealing with a case of late-onset dominant Mendelian trait with reduced penetrance. This is exactly the case with the TIMP3 gene, in which dominant (i.e., single) mutations cause Sorsby macular dystrophy³ and also, as most recently shown,⁴ cause the disease at an advanced age, thereby confusing the phenotype with AMD. “Causing a disease” and “increasing disease susceptibility” are two different terms describing very different disease mechanisms. 

Second, the authors equate Stargardt disease with one of its constituent phenotypes, the increased lipofuscin accumulation,^5–7 suggesting that no increased lipofuscin (i.e., not-elevated qAF) means “no disease.” However, it was shown 16 years ago by Fishman and colleagues⁸ and confirmed by many, including us,⁶ that a substantial fraction of STGD1 patients do not present with increased lipofuscin. This is particularly true for patients harboring the p.Gly1961Glu mutation,^6,8 and this trend is clearly apparent in the study by Müller and colleagues,¹ where three-fourths of STGD1 patients with the p.Gly1961Glu variant exhibited only marginally increased qAF values. It is noteworthy that limited or no increase in qAF was noted in five more patients in their study; therefore, 8 (∼30%) of 26 STGD1 patients in the Müller et al.¹ study had a limited increase in lipofuscin as measured by qAF.¹ These numbers are comparable to those in the studies by Burke et al.⁶ (∼20%) and Fishman et al.,⁸ in which approximately 30% of patients presented with the absence of dark choroid on fluorescein angiography. It is also worth stressing that qAF values change with the disease stage so that very advanced patients with large areas of atrophy have lower qAF values, probably due to the absence of RPE (and photoreceptor) cells.⁶ 

Third, Müller et al.¹ repeat the A2E measurements done by Travis and colleagues in 1999 and 2001 in Abca4^+/− and Abca4^−/− mice^9,10 and confirm that heterozygous animals accumulate approximately 50% of the A2E the knock-outs do. The difference between humans and mice can be caused by a variety of reasons; however, the two methods used, qAF of the macular area in patients and HPLC measurements in whole eye cups in mice, are not directly comparable. Coincidentally, it has been demonstrated that although the A2E accumulation in the Abca4^−/+ animals is 50% of that in Abca4^−/− mice if measured by HPLC in whole eyecups, the qAF measurements in the same animals showed only an approximately 20% increase compared with wild-type mice.¹¹ 

Fourth, the authors try to explain the increased qAF in one of the carriers by “another, mild ABCA4 allele,” rendering that individual not a heterozygous carrier but rather one an affected by STGD1 due to compound heterozygosity. However, the allele under question, p.Asn1868Ile, is a well-known common variant with the allele frequency of 7% in the general population of European descent (ExAC database; http://exac.broadinstitute.org/). Thus, it cannot be the cause of an increased AF in that individual. 

In summary, the Müller et al. study,¹ although reaching valid results, has some issues with interpreting and/or presenting those. Even the main conclusion of the study: “the findings may have implications for therapeutic approaches such as gene replacement therapy,” is confusing. How would the findings that ABCA4 carriers do not accumulate increased lipofuscin^1,2,6 have any implications for the application of gene supplementation (i.e., addition of a functional gene, not replacement, which is not yet available) therapy? The authors may want to explain their conclusion better, because the gene supplementation therapy, if successful, remains currently the best option for all recessive diseases, including STGD1.