Severity of Biallelic ABCA4 Dysfunction Is Associated with Age of Symptom Onset in Stargardt Disease

Abstract

Purpose : To test the hypothesis that age of symptom onset in Stargardt Disease (STGD) is predicted by the combined biallelic severity of the variants in ABCA4.

Methods : We have developed an ABCA4 mutation severity score that estimates the ability of the combined product of the two alleles to perform ABCA4’s physiological function in the retinal cell, based on experimental characterization of disease variants. This protein functionality score, F, considers both alleles’ abilities to express proteins that correctly embed in the cell membrane and function as efficient transporters. We express F as the fraction of the fully functional genotype consisting of two healthy alleles, and thus F ranges from 0% (two non-expressing alleles) to 100% (normal). We applied F scores to the 27 STGD patients seen in our department who had a documented age of symptom onset (SO), defined as first failed vision exam or first verbal complaint, in their medical records. We supplemented these data with that of 207 individuals in the literature for whom biallelic diagnosis and SO were reported. In our full sample (N=234), SO ranged from 3–72 (median 10) years, and F ranged from 0%–89% (median 18%). We then developed a model to predict SO (years) from F, assuming that the relation of SO to F followed the common logistic form:

F = (SO - D)ⁿ/(kⁿ + (SO - D)ⁿ) for SO ≥ D,

where D is the typical SO in a patient with F=0%, k+D is the predicted SO in a patient with F=50%, and n is related to the slope of the curve. To produce estimates of SO by F, we isolated SO in the above equation,

SO = ((F*kⁿ)/(1 - F))^{1/n + D},

and optimized the values of k, n, and D by fitting to our full sample.

Results : For our logistic model, r² was 0.36. The predicted SO for a person with F=0% was 9 (i.e., D=9), with F=25% was 15, with F=50% was 19 (i.e., k=10), and with F=75% was 27; n was 1.9.

Conclusions : We found a positive, logistic association between the appearance of STGD symptoms and ABCA4 functionality. Although, at this point, we can only confirm that our approach captures a source of the phenotype/genotype correlation in STGD, we hope that more personalized characterizations of STGD patients will lead to improved prognostication and therefore to timely and effective intervention.

This abstract was presented at the 2023 ARVO Annual Meeting, held in New Orleans, LA, April 23-27, 2023.