Coupling germline genetics to interferon-alpha responses in polycythemia vera treatment

Pegylated interferon alfa-2 (PEG-IFN) induces high response rates in many patients with polycythemia vera (PV), while some patients do not respond well. In order to identify specific genetic predictors, a new study by Jaeger and colleagues in Blood investigated specific germline diplotypes of the interferon lambda 4 (IFNL4) gene locus and their correlation with PEG-IFN treatment outcome.¹

Methods

Samples from 122 patients treated with PEG-IFN in the PROUD-PV and CONTINUATION-PV trials were used for the primary analysis, while samples from 27 patients treated in the PEGINVERA trial (NCT01193699) were used to replicate results.^1,2 The samples were analyzed with Affymetrix SNP 6.0 microarrays at 12, 18, 24, 30, and 36 months of follow-up. IFNL4 variants were sequenced and phased with SHAPEIT.³

Results

Using a genome-wide association study analysis tagging single nucleotide polymorphism (SNP) and correlating allelic differences with hematologic response (HR) or molecular response (MR) in PEG-IFN-treated patients did not reveal any significant genome-wide associations. However, in a hypothesis-driven approach investigating the role of the IFNL4 gene locus variants, the authors demonstrated a significant genomic correlation between non-functional IFNL4 variants and improved MR (not HR).

The IFNL4 gene is a polymorphic pseudogene, which translates into the functional IFNL4 protein only in some humans. The gene is polymorphic in humans, and the TT allele for the variants rs368234815 and rs8099917 produces a frameshift in the coding region leading to a non-coding mRNA. Similarly, for the rs12979860 variant, a CC allele leads to lack of IFNL4 expression. In contrast, the rs117648444 variant can come with a G allele leading to the expression of an impaired IFNL4-S70 protein, while the non-G allele produces the fully functional IFNL4-P70 protein. Interestingly, non-coding IFNL4 SNPs (rs12979860, rs8099917) have previously shown to be associated with response to PEG-IFN and viral clearance during hepatitis C virus infection.

In accordance with results for hepatitis C virus treatment, the investigated PV population showed a significant genotypic association of the non-coding alleles of rs368234815 (OR, 6.10; 95% CI, 1.49–36.64; p = 0.00065), rs8099917 (OR, 4.60; 95% CI, 1.39–16.69; p = 0.0071), and rs12979860 (OR, 6.10; 95% CI, 1.49–36.64; p = 0.0065) for MR at 36 months follow-up.

Based on that finding, the authors investigated diplotypes of these variants in patients, which comprises three functional classes: no IFNL4, nonfunctional IFNL4-S70, and active IFNL4-P70. As was expected, MR (but not HR) was strongly associated with the absence of IFNL4 (89.3%) versus a completely functional IFNL4-P70 (43.3%) protein. Interestingly, the impaired IFNL4-S70 induced similar responses to no IFNL4 (Table 1). 

Table 1. Influence of IFNL4 diplotypes on the molecular response in PEG-IFN treatment of PV¹

CI, confidence interval; IFNL4, interferon lambda 4; n, number of patients; NR, non-responders; N (no IFNL4); OR, odds ratio; P, IFNL4-P70; R, responders; S, IFNL4-S70.
Variant/ diplotype	Diploid status	n	NR, n	R, n	R, %	OR, 95% CI	P value
Contribution of IFNL4-S70 and no-IFNL4 compared with IFNL4-P70 alone
IFNL4-P70 alone	P/N and P/P	26	15	11	42.3	Reference
IFNL4-S70 and S70/P70	N/S and P/S	15	3	12	80.0	5.23, 1.06–35.87	0.0254
no IFNL4	N/N	28	3	25	89.3	10.80, 2.39–69.97	0.000391
Contribution of IFNL4-S70 compared with no-IFNL4
no IFNL4	N/N	28	3	25	89.3	Reference
IFNL4-S70 and S70/P70	N/S and P/S	15	3	12	80.0	0.49, 0.06–4.20	0.6474

These results were confirmed in 27 patients treated in the PEGINVERA trial, where a consistent trend for a better MR after PEG-IFN treatment was observed in patients lacking functional IFNL4 at 24 months follow-up (p = 0.068).

Conclusions

In patients with PV treated with PEG-IFN, lack of functional IFNL4 expression status confers a significantly better MR indicating deeper responses which the potential for cure. In contrast, no differences were observed for HR. These data may have implications for optimizing patient management during PEG-IFN treatment, warranting further confirmation in bigger patient cohorts.