Is Nature Truly Healing Itself? Spontaneous Remissions and Clonal Replacement in Paroxysmal Nocturnal Hemoglobinuria

Paroxysmal nocturnal hemoglobinuria (PNH) is considered to be curable only through the means of allogeneic HSCT. One of the many fascinating and scientifically instructive aspects of the pathogenesis of this disease is the rare possibility of its spontaneous remission with disappearance of PNH clone and abatement of clinical symptoms, which has always captivated the research community. Due to the orphan nature of the condition, no clinical predictors have been identified so far as harbingers of this phenomenon. In a classical scenario, exhaustion of PNH clone may be associated with reappearance of aplastic anemia (AA), in which PNH clone reflects a semi-maladaptive attempt of recovery. Consequently, one could stipulate that the retraction of PNH clone(s) would have to be associated with a compensatory re-expansion of normal hematopoiesis should normal counts be maintained. The recent insights into the AA/PNH pathobiology shed light onto molecular underpinnings of polyclonal vs oligoclonal hematopoiesis and their dynamics. Here, through application of NGS we attempted to better discern the mechanism of PNH spontaneous remission taking advantage of our internal cohort of PNH patients.

Among 92 patients with a diagnosis of hemolytic PNH (M:F ratio 0.88, median age 38 years, range 9-84) 41% were primary PNH (pPNH) while 59% were secondary to AA (sPNH). Overall, patients were clinically followed-up for a median time of 68 months (2-339). Median granulocyte clone size was 73% (22-99) with the majority of cases being classified as having a type III dominant red blood cells (RBCs) clone (80%) while 20% type II. Within this cohort, a total of 3 patients underwent spontaneous remission.

UPN1 was a 69-year-old male diagnosed with pPNH at the age of 46 after an episode of deep vein thrombosis. He had been managed with prednisone, transfusions and anti-coagulation because of recurrent thrombotic episodes. Once available, he was started on eculizumab and later continued on ravulizumab. His initial flow cytometry study revealed the presence of a type III RBCs clone of 40% and a granulocyte clone of 89%. After 9 years of anti-complement therapy, the patient's clone started a slow decrease and the most recent study revealed a granulocyte clone of 0.02%. Molecular analysis performed at the time of eculizumab start showed a co-dominant mutational configuration by variant allelic frequency (VAF) with PIGA deletion (p.94_95del; VAF 29%) and a BCOR nonsense (p.Y1446X; VAF 27%). No HLA class I/II mutations were found in two longitudinal samples collected 1 year before and after eculizumab start. However, at the last sequencing performed after the complete disappearance of the PNH clone, the patient developed ASXL1 (p.E635Rfs*, VAF 26%) and ZRSR2 (p.E120Gfs*, VAF 42%) mutations along with retraction of the previous PIGA lesion. No decrease in blood counts was noted.

UPN2 was a 58-year-old male initially diagnosed with severe AA at the age of 48 and treated with ATG + CsA. At that time, he had a co-existing PNH granulocyte clone of 28%. After 1 year from IST his PNH clone dropped to 1% and since then has been consistently below 1%. Patient has never received anti-complement therapy. At the time of PNH clone retraction, no HLA class I/II or myeloid driver mutations were found and PIGA mutations were not detectable. However, longitudinal molecular studies performed after disappearance of PNH clone revealed the acquisition of ASXL1 p.Q512X mutation at an initial VAF of 23%, which doubled (45%) at last follow-up 5 years later while normal counts persisted.

UPN3 was instead a 59-year-old lady diagnosed with pPNH at the age of 30. She had a granulocyte clone as high as 43% with a type II RBCs clone of 17% and a typical PIGA splice site c.981+1G>A mutation (VAF 15%). She was initially treated with transfusions and steroids and her course was complicated by a cerebral venous sinus thrombosis. Patient was eventually given eculizumab and her PNH clone started decreasing until it vanished (last 0.04%) after 8 years. Analysis of samples prior to and after PNH disappearance did not show any HLA class I/II nor myeloid driver gene mutations with absence of PIGA alterations at last sequencing.

PNH spontaneous remissions are rare events. In addition to be replaced by polyclonal hematopoiesis, PIGA clones may be swept by CHIP lesions in myeloid genes (e.g. ASXL1) characterized by improved fitness advantage in a process of Darwinian selection.

Figure 1

View largeDownload slide

Figure 1

View largeDownload slide

Close modal