Immunodynamics of Paroxysmal Nocturnal Hemoglobinuria

In the context of immune-mediated attack in aplastic anemia (AA), PIGA mutations produce a clonally-restricted escape phenotype in an attempt to restore hematopoiesis. Along with PIGA lesions, HLA class I/II mutations may represent an alternative way to render cytotoxic T cell recognition less effective. Microdeletions of 6p21, uniparental disomy [UPD] of 6p and mutations of HLA genes have been previously identified in AA patients. One could stipulate that paroxysmal nocturnal hemoglobinuria (PNH) may be an alternative mode of clonal evolution as opposed to HLA mutation acquisition or instead be a redundant pathway of immune escape co-existing with such lesions. Here, we took advantage of a large cohort (n=92) of patients with hemolytic PNH (granulocytes clone size >20% and LDH>x2.5 ULN) to elucidate the immunodynamics of PIGA/HLA clone(s) (Fig1A).

We first used a deep targeted sequencing panel covering HLA classical loci and applied an in-house newly developed pipeline on 62 patients with PNH. Overall, 35% of cases had HLA alterations, which were found at a similar rate in patients studied at onset (33%) and during the course of the disease (32%). These alterations were classified as allelic losses in 15% of cases while majority of mutations were missense (57%) followed by 5' or 3' UTR (20%), nonsense (15%) and frameshifts (8%) with a median variant allelic frequency (VAF) of 5% (range, 2-95). Of note, the canonical exon 1 HLA B14:02 mutation (c.19C>T, p.R7X) ^1,2was found in 2 patients who both had de novo PNH with a granulocyte clone size of 60% at disease onset (Fig1B).

To clarify whether the PNH phenotype conveys an absolute resistance from immune attack, one should determine whether in these patients HLA mutations arise indeed within non-PNH cells. Consequently, we also performed HLA sequencing on 9 patients whose samples (n=18) were previously sorted by flow cytometry to enrich for CD59+ and CD59- fractions. HLA mutations were identified in both CD59+ and CD59- fractions at similar rates (44% each) and VAF (5% vs 3%, p=0.3) while no "canonical" exon 1 mutation was found.

When looking at associations with disease-specific characteristics, we did not identify any significant differences in terms of burden of HLA events and disease ontogeny (primary vs secondary to AA cases) as well as baseline demographics characteristics (gender, age at presentation), dominant erythrocytes type (II vs III), and thrombotic events. No correlations with specific types of PIGA mutations nor their burden were found in HLA-mutated vs non-mutated groups as also confirmed by the absence of relationship with PNH granulocyte clone size (62% vs 72%, p=0.5). Of note, in HLA mutants the VAFs of HLA lesions were significantly lower than those of PIGA (5% vs 18%, p<0.001) underlying their possible less powerful role in driving immune escape within the context of AA overshooting reactions. To validate these results, we have designed a highly quantitative flow cytometric method for detection of HLA expression (A, B, C for class I; DQ, DR, DP for class II) on CD59+ vs CD59- fractions of CD34+ and CD15+ cells obtained from PNH patients. Preliminary results in 7 patients showed a general downregulation of HLA expression within PNH clones (delta HLA MFI of CD59+ vs CD59- fractions), consistent with our ability to detect HLA mutations also in PNH cells (Fig1C). This preferential down-modulation on PNH cells suggests a possible selection mechanism operating agonistically and in a non-mutually exclusive fashion with the PNH-intrinsic privilege. Moreover, these findings may also represent a byproduct of antigen processing machinery impairment due to proteasomal accumulation of proteins, which would have been GPI-linked and exported to the membrane.

In sum, our study indicates that PNH phenotype conveyed by PIGA mutations operates independently from HLA mutations as means of immune escape. In addition, based on these results we stipulate that HLA down-modulation via various mechanisms is operative in the context of the PNH phenotype and the immune privilege of PIGA clones is not absolutely protective. Patients with PNH may have elements of immune suppression. In fact, their responses in the context of reduced-intensity transplant do eliminate hosts PNH clones following transplantation. ³

Figure 1

View largeDownload slide

Figure 1

View largeDownload slide

Close modal