Genetic and Phenotypic Variability in Siblings With Friedreich Ataxia

Abstract

Objectives

Friedreich ataxia (FRDA) is an autosomal recessive disorder caused by FXN gene mutations involving GAA trinucleotide repeat expansions. This study explores phenotypic heterogeneity between siblings, focusing on differences in age at onset (AAO) and shorter GAA repeat (GAA1) length to improve understanding of disease variability.

Methods

We analyzed AAO and genotype of siblings with FRDA. Linear regression examined AAO differences and genetic predictors (GAA1 length and SIRT6 S46N single nucleotide polymorphism), while logistic regression assessed discordant clinical manifestations and GAA1 heterogeneity.

Results

This study included 150 siblings with FRDA from 70 families. GAA1 and AAO differences between siblings were not significant, although discordance between siblings in the diagnosis of hypertrophic cardiomyopathy and scoliosis was noted in approximately a quarter of the families. Differences in GAA1 length predicted a modest amount of AAO heterogeneity (R² = 0.075). The S46N polymorphism in SIRT6 did not predict the differences in AAO.

Discussion

Genetic and phenotypic variability between paired siblings with FRDA was moderate to small, with GAA1 differences explaining some of the variance in AAO. Other factors (genetic or environmental) or data collection bias may explain the remaining variance. These findings highlight the complexity of FRDA and reiterate the role of GAA1 length in disease severity.

Introduction

Friedreich ataxia (FRDA) is an autosomal recessive disorder caused by mutations in the FXN gene. FRDA affects 1 in 50,000–100,000 individuals, most commonly with an age at onset (AAO) of 7–15 years, although this varies.^1-3 Approximately 96% of patients with FRDA carry a biallelic expansion of GAA trinucleotide repeats in intron 1 of FXN. The remaining 4% carry a GAA expansion on one allele and a point mutation or deletion on the other.^1-6 Both GAA expansions and other mutations reduce levels of functional frataxin protein, leading to mitochondrial dysfunction and the clinical manifestations of FRDA, including ataxia, sensory loss, scoliosis, hypertrophic cardiomyopathy (HCMP), optic atrophy (OA), and elevated risk of diabetes mellitus (DM).^7,8

The length of the shorter GAA repeat (GAA1) correlates with frataxin expression and FRDA disease severity. However, the heterogeneity in GAA1 only explains 50% of the variability in AAO, the most important predictor of disease progression.^5,8,9 Previous studies have noted intrafamilial phenotypic variability among siblings with similar GAA1 lengths, suggesting the influence of other factors on disease expression.^10,11

Recently, a S46N single nucleotide polymorphism (SNP) in the Sirtuin6 (SIRT6) gene was identified as a modifier of disease severity in FRDA, suggesting that other unidentified modifiers may exist.⁸ Still, no other modifiers have yet been identified in genomic studies.¹² To explore an alternative to whole-exome sequencing for identifying genetic modifiers, this study focuses on phenotypic differences between paired siblings with FRDA, based on the assumption that siblings share a similar genetic background. This allows us to investigate the influence of genetic differences and potential modifiers.

Methods

Patient Cohort

Data were collected from the Friedreich Ataxia Outcome Measures Study (FACOMS), which has enrolled patients with genetically confirmed FRDA since 2003.¹³ At the Children's Hospital of Philadelphia site, 167 individuals with at least 1 affected sibling were identified. After excluding 17 patients with missing sibling data, 150 individuals from 70 families remained.

Siblings Comparisons

Within each family, AAO and GAA1 length were compared between siblings. In families with more than 2 siblings (n = 6), comparisons were made with the oldest sibling. The difference in AAO was calculated by subtracting the younger sibling's AAO from the older sibling's, with negative values indicating an earlier AAO for younger siblings. Similar analyses were performed for GAA1 length. The rate of disease progression was assessed using mFARS slopes for siblings with at least 4 years of follow-up (n = 89). To compare progression within families, the percentage difference in slopes was calculated relative to the oldest sibling.

Statistical Analyses

Data analysis was conducted using STATA version 18.5 (College Station, TX). Descriptive statistics summarized demographics and clinical characteristics. Older siblings were excluded from GAA1 and AAO analyses to prevent skewing because their differences are inherently 0. Logistic regression examined discordant clinical manifestations (HCMP, DM, scoliosis, and OA) in relation to GAA1 differences among siblings. Linear regression analyzed AAO differences and predictors (GAA1 variability, S46N SNP in SIRT6), with statistical significance set at p < 0.05. SIRT6 S46N SNP status, available for 82 individuals with an allele frequency of 13.9%, was obtained from a previous study⁸; individuals with missing data were excluded from analyses involving SIRT6 polymorphisms.

Standard Protocol Approvals, Registrations, and Patient Consents

Written informed consent was obtained from all patients or their surrogates at enrollment and renewed annually. The FACOMS was approved by the local institutional review board and registered with ClinicalTrials.gov (NCT03090789).

Data Availability

The FACOMS, along with relevant study information, is part of the Friedreich Ataxia Integrated Clinical Database, accessible upon request at the Critical Path Institute's Data Collaboration Center.

Results

Demographics and Follow-Up Time

Among 150 siblings with FRDA, 47 experienced early-onset (0–7 years, 31%), 70 typical-onset (8–14 years, 47%), 22 intermediate-onset (15–24 years, 15%), and 11 late-onset (>24 years, 7%) FRDA (Table 1). Overall, 8.7% of siblings with FRDA were compound heterozygotes (carrying a point mutation or deletion), slightly more common in late-onset groups. The median follow-up time was 5.5 years (IQR 2–10). During follow-up, 77% were diagnosed with scoliosis (116 siblings, 48 concordant pairs), 51% with HCMP (76 siblings, 30 concordant pairs), 6.7% with DM (10 siblings, 2 concordant pairs), and 5.3% exhibited OA (8 siblings, 1 concordant pair).

Table 1.

Demographics and Follow-Up Characteristics by Onset Group

Onset group	0–7 y (early)	8–14 y (typical)	15–24 y (intermediate)	>24 y (late)	Overall
N (% of overall)	47 (31)	70 (47)	22 (15)	11 (7)	150
Male sex (%)	45	50	45	64	49
Age at onset (AAO), y	5 [4–6]	10 [9–12]	17 [15–18]	30 [28–38]	10 [7–14]
GAA1a	766 [666–858]	700 [600–850]	533 [366–599]	325 [199–325]	690 [546–800]
GAA2	1,000 [887.5–1,162.5]	950 [893–1,100]	875 [833–1,025]	1,025 [966–1,025]	966 [866–1,100]
Point mutations (%)	6 (12.8)	4 (5.7)	1 (4.5)	2 (18.2)	13 (8.7)
Age, y	20 [14–24]	24.5 [18–29]	32 [29–40]	65 [54–72]	25 [18–33]
Age at last visit, y	16 [12–21]	21 [16–25]	29.5 [26–38]	61 [52–65]	22 [16–31]
HCMP (%)	26 (55)	44 (63)	5 (23)	1 (9)	76 (51)
DM (%)	5 (10.6)	4 (5.7)	1 (4.5)	0 (0)	10 (6.7)
Scoliosis (%)	37 (79)	61 (87)	14 (64)	4 (36)	116 (77)
Optic atrophy (%)	5 (10.6)	2 (2.9)	1 (4.6)	0 (0)	8 (5.3)
Ambulatory at enrollment (%)	37 (79)	60 (86)	21 (95)	6 (55)	124 (83)
Ambulatory at last visit (%)	24 (51)	40 (57)	15 (68)	5 (45)	84 (56)
Follow-up, y	5 [0–10]	6 [3–10]	5.5 [4–14]	4 [0–10]	5.5 [2–10]
Patients without follow-up (%)	11 (23)	9 (13)	2 (9)	4 (36)	26 (17)

Data are reported as median [IQR].

^aExcluding point mutations (n = 13) and participants with missing repeat length information (n = 6).

Sibling comparisons: The differences in AAO were analyzed for 80 sibling pairs, while the differences in GAA1 length were examined for 74 pairs, depending on data availability. The median difference in AAO was 0 years (IQR −2 to 2) (mean = 0, SD = 6.06, 95% CI −1.35 to 1.35), with a median absolute difference of 2 years (IQR 1–5) (mean = 3.65, SD = 4.82). The AAO difference within families was not significant (p = 1.0) (Figure, A and B). The median difference in GAA1 length within families was 0 repeats (IQR −100 to 60) (mean = −9.34, SD = 143.66) (Figure, C). The median absolute GAA1 difference was 67 repeats (IQR 9–134) (mean = 97.45, SD = 105.36). Among the 70 families, 23% had discordant scoliosis (16 discordant pairs), 23% HCMP (16 discordant pairs), 9% DM (6 discordant pairs), and 9% OA (6 discordant pairs). To address disease progression discrepancies, the percentage difference in mFARS slopes compared with the oldest sibling was calculated (IQR −37.9% to 30.1%; mean = 11.6%, median = −3.7%, SD = 133).

Figure. Comparisons of Age at Onset and GAA1 in Siblings With FRDA.

(A) Difference in AAO of younger siblings compared with their older siblings. (B) AAO of younger siblings compared with their older siblings is similar. (C) Difference in GAA1 length. (D) Association of GAA1 length difference on AAO difference in siblings. AAO = age at onset; FRDA = Friedreich ataxia.

Regression Analysis

Regression analysis of AAO differences within families as a function of GAA1 differences revealed a modest association, with GAA1 length predicting a small portion of variability in AAO (p = 0.019, R² = 0.075, 95% CI −0.021 to 0.002) (Figure, D). Adding SIRT6 gene S46N SNP status increased the model's significance (p = 0.0072, R² = 0.1189), but SIRT6 SNP status did not predict AAO heterogeneity (p = 0.294, 95% CI −0.745 to 2.43). By contrast, GAA1 differences were the significant driver of the model (p = 0.004, 95% CI −0.017 to −0.0035). A logistic regression model did not find GAA1 differences as a significant predictor of clinical heterogeneity (eTable 1). Regression analysis of mFARS slope differences among siblings revealed a significant association with GAA1 differences, explaining a small portion of the variability (p = 0.0311, R² = 0.1084).

Discussion

This study demonstrates that phenotypic features of FRDA are generally concordant between affected siblings, with differences in GAA1 length partially explaining heterogeneity. Younger siblings typically had similar AAO and GAA1 lengths as their older counterparts. GAA1 differences accounted for 7.5% of intrafamilial AAO discrepancy, while a nonsynonymous SNP in SIRT6 had no predictive value. GAA1 length differences did not predict discordance in clinical manifestations among siblings, but they accounted for approximately 10.8% of variability in neurologic disease progression.

Several factors may explain the discordance in AAO between siblings. GAA lengths measured in blood cells might differ in affected tissues, influencing AAO. Other unidentified genetic modifiers could also play a role, necessitating a larger sample size for investigation. In addition, while environmental influences on AAO are possible, they remain unknown. Though not observed here, AAO can be a biased outcome measure, typically leading to earlier identification of younger siblings. Notably, one sibling pair exhibited a significant AAO discrepancy despite identical GAA1 lengths, suggesting self-reporting bias.

In conclusion, this study highlights genetic and phenotypic variability in FRDA among siblings, particularly the significant role of GAA1 length in AAO. The modest explanatory power of our models underscores the complexity of FRDA and suggests the presence of additional genetic or environmental modifiers. Future research should involve larger cohorts or discordant sibling pairs and include other measures of disease severity to better understand the factors driving variability in FRDA.

Author Contributions

K. Eshaghi: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; study concept or design; analysis or interpretation of data. P.H. Rao: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data. M.M. Shen: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data. D.R. Lynch: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; study concept or design; analysis or interpretation of data.

Study Funding

The authors report no targeted funding.

Disclosure

The authors report no relevant disclosures. Go to Neurology.org/NG for full disclosures.