Identifying Pathogenic FOXP3 Variants in Patients Presenting with NDM Provides a Window of Opportunity to Monitor and Treat

We report 65 males with 39 different pathogenic or likely pathogenic FOXP3 variants, 16 of which are novel and were classified according to ACGS best practice guidelines (ESM Table 1) [15]. Where a maternal sample was available, 50/55 (91%) patients had inherited the variant from their unaffected mother. In the remaining 5 cases, the variant had arisen de novo.

The presenting feature was diabetes in 50/65 (77%) patients. Of those, 47/50 (95%) presented with NDM and 3/50 (5%) were diagnosed aged between 6 months and 1 year. The remaining 15 presented with enteropathy and/or other features (ESM Table 1, ESM Fig. 1). Of the 50 who presented with diabetes, 20 (40%) are known to have developed enteropathy (i.e., have classical IPEX) during follow-up (median follow up 48 weeks, [IQR 17.2–329 weeks]). The median time from diabetes onset to enteropathy was 23.5 weeks [IQR 7.8–126.8 weeks]. No individuals were on immunosuppressive therapy at the time of sampling; individuals with diabetes were treated with insulin.

TSDR/CD4 Measurement Can Distinguish Between NDM Patients with an IPEX-Causing FOXP3 Variant, Healthy Controls, and Individuals with Other Tregopathies

In 41 genetically confirmed IPEX patients where there was sufficient DNA, we next measured the % demethylated FOXP3 TSDR to assess if it could aid in rapid diagnosis. The median duration of disease at sampling was 11.2 weeks (IQR 3.6–83.1). We found that these individuals had slightly reduced demethylation at the % demethylated CD4 and increased % demethylated Foxp3 TSDR (both as a percentage of demethylated GAPDH from whole blood) compared to healthy controls (ESM Fig. 2). Levels of % demethylated FOXP3 TSDR as a proportion of % demethylated CD4 (%TSDR/CD4) were therefore higher (median 13.6% [IQR 10.5–22.3]) compared with male controls (n = 29; median TSDR/CD4 8.5%, [IQR 7.7–10.4], p < 0.0001) or males with other IPEX-like Tregopathies (n = 13, median TSDR/CD4 8.7% [IQR 7.4–12.7], p = 0.01) (Table 1, Fig. 1; ESM Table 1, Table 3).

Table 1 Cohort characteristics. W, weeks; IQR, interquartile range; N/A, not applicableFig. 1

a %TSDR/CD4 in controls (n = 29, median 8.5%, IQR 7.7–10.4%) compared with patients with IPEX (n = 41, median 13.6%, IQR 10.5–22.3, p < 0.0001) and patients with overlapping IPEX-like phenotypes due to other monogenic Tregopathies (n = 13, median 8.7%, IQR 7.4–12.7, p = 1.0; blue dots represent patients with homozygous pathogenic LRBA variants, red dots individuals with homozygous pathogenic IL2RA variants, and yellow dots individuals with gain-of-function STAT3 variants [see ESM Table 3 for detailed information]). Horizontal lines represent the median and interquartile range. b Receiver-operator characteristic curve for TSDR/CD4 based on data from 41 IPEX patients and 29 controls. The area under the curve (AUC) was 0.81

We conducted receiver-operator characteristic analysis on our control and IPEX cohort to assess the discriminatory ability of the %TSDR/CD4 and identify the optimal threshold for discrimination. The overall area under the curve was 0.81 [95% CI 0.71–0.91] and the optimal cut-off based on 95% specificity was 13.4%, which had 51% sensitivity to identify IPEX patients (Fig. 1b). This supports that %TSDR/CD4 is a useful tool to aid diagnosis in individuals with a FOXP3 variant of uncertain significance. For 11 out of 16 patients with novel Foxp3 mutation variants, the percentage of TSDR/CD4 was available. Six of these patients (55%) also showed an increased %TSDR/CD4 (value greater than or equal to 13.4%) (Fig. 2a).

Fig. 2

a %TSDR/CD4 values in patients with either previously reported (“known”) or novel variants. The dotted line is the cut off of 13.4%. b %TSDR/CD4 values in individuals with either a missense or protein truncating variant (“PTV”)

To assess whether FOXP3 genotype and the %TSDR/CD4 were related, we compared values in patients with missense variants (n = 31, median 13.1% [IQR 10.6–20.0%]) and protein truncating variants (n = 8, median 12.5% [IQR 9.3–23.1]). These values were similar, although more variable, in individuals with missense variants regardless of whether they were in the forkhead box domain (which is where the majority of pathogenic variants cluster) or not (Fig. 2b).

Case Study: Use of %TSDR/CD4 in Determining Pathogenicity of a Novel FOXP3 Variant

We found that determining the %TSDR/CD4 was useful in interpreting a case without classic IPEX features who had a novel missense variant in the alternatively spliced exon two of FOXP3, p.(Pro75Leu) (individual not included in main analysis; Fig. 3, ESM Table 4).

Fig. 3

Partial pedigree of a family with novel FOXP3 variant in the alternatively spliced exon two (p.Pro75Leu). Arrow denotes proband. Frequencies refer to %TSDR/CD4

During our study, a male patient from Palestine was referred for genetic testing after developing isolated diabetes aged 6 months (insulin treated) and presenting with severe diabetic ketoacidosis aged 5 years resulting from lack of access to insulin which was subsequently fatal (patient III.1). Sequencing of FOXP3 revealed this patient had a novel missense variant in the alternatively spliced exon two, p.(Pro75Leu), which was originally classified as a VUS by ACMG guidelines [15]. We subsequently measured %TSDR/CD4 and found them to be more than double the mean value in controls (23.5% vs. 10.1%), providing evidence to suggest the variant was causal. On further discussion with the clinician, we established that there was a family history of early onset diabetes or immunodysregulatory features affecting his brother, male cousin, and an uncle. No family member had enteropathy, the most common and severe feature of IPEX syndrome. We were then able to confirm co-segregation of the variant, establish that %TSDR/CD4 values were raised in the affected (30.3%, 23.5%, 28.1%, and 15.2%) but not unaffected (10.21, 10.14, and 9.5%) male relatives, and establish pathogenicity using ACMG guidelines. The two younger affected relatives of the proband have now be referred for hematopoietic stem cell transplantation (HSCT).

%TSDR/CD4 May Offer Insight into IPEX Severity in Patients Who Presented with Diabetes

To assess the prognostic potential of the %TSDR/CD4, we looked at clinical features developed during follow-up by patients who had initially presented with diabetes. We categorized additional features according to the system/organ affected (see ESM Table 5 for categories). Using the cut off of 13.4%, we assessed whether the number of affected systems differed between patients who presented with diabetes and had %TSDR/CD4 values above or below the threshold (Fig. 4). Those with a higher %TSDR/CD4 were more likely to have \(\ge\) 2 systems affected (\(\ge\) 13.4% 17/19 [89%] vs. < 13.4% 7/16 [44%], p = 0.009). The %TSDR/CD4 may therefore offer insight into IPEX severity, though we were unable to determine if this reflects underlying pathophysiology or is a consequence of severity. The median age at latest follow-up did not differ between the two groups (ESM Fig. 3) but was relatively short (< 13.4%: n = 15, median 26.1 weeks, IQR 13–886 weeks vs \(\ge\) 13.4%: n = 18, median 43.5 weeks, IQR 19.8–156.3 weeks, p = 0.6), and some patients may have subsequently developed additional features.

Fig. 4

Proportion of IPEX patients with 1, 2, 3, 4, 5, or 6 systems affected in those with a TSDR/CD4 value of < 13.4% (n = 16) and \(\ge\) 13.4% (n = 19). Of those with a %TSDR/CD4 \(\ge\) 13.4%, 17 (89%) had \(\ge\) 2 systems affected by the end of follow-up compared with 7 (44%) in those with a %TSDR/CD4 < 13.4%. p = 0.009

We do not suggest that TSDR/CD4 should be used to guide clinical practice as this data are not longitudinal and a high proportion of patients with TSDR/CD4 below 13.4% have ≥ 2 systems affected. While this tool appears to add understanding to clinicians for some patients, the lack of full longitudinal studies means caution is required in interpreting the data. In a disease as rare as IPEX, potentially valuable data should be monitored but, due to the lack of population studies, interpreted cautiously.