Characterization of Erythrocytes in PNH: Can We Predict Response to Complement Inhibition Based on Different Erythrocyte Markers?

Paroxysmal nocturnal hemoglobinuria (PNH) is a rare disease characterized by complement-mediated hemolysis, bone marrow failure and thrombosis. Despite the promising results with C5 complement inhibitors (CIs), approximately one third of C5 CI treated patients (pts) have hemolysis, >50% suffer from fatigue, and 10-30% require transfusions. Novel therapeutic options are available: a C3 inhibitor, a Factor D and a Factor B inhibitor are approved along with C5 inhibitors. Red blood cells (RBCs) in PNH have a central role in the pathophysiology and clinical presentation of the disease. Nonetheless, there are not many studies focusing on their biological characterization and their potential to serve as biomarkers for the prediction of treatment response.

Aim:Therapeutic decisions should aim at the optimal CI treatment, hence the question is: which pt. will benefit most from a specific CI? We wanted to examine various RBCs parameters in untreated and treated pts. These results will be compared with clinical parameters (Hgb, LDH, percentage of reticulocytes, history of thrombosis, and others) to unravel potential links between RBC characteristics and clinical markers.

Methods: Ten patients with PNH (8 on treatment with CI, 2 treatment-naïve pts., 3 men, 7 women) and 10 healthy age-matched individuals were analyzed. All samples were examined for spontaneous, osmotic, mechanical, and oxidative hemolysis, the oxidation status of Hb and the overall oxidative burden (reactive oxygen species; ROS) before and after oxidation using phenylhydrazine (PHZ), diamide, and tert-butyl hydroperoxide (tBHP), and the activity of the proteasome machinery. Moreover, the percentage of mitochondria-positive RBCs was assessed. P value < 0.05 was used to determine statistical significance.

Results: Patients with PNH showed elevated hemolysis (0.067 ± 0.033 vs 0.005±0.003 %, PNH vs healthy individuals, p<0.001) and their RBCs exhibited increased susceptibility to lysis after exposure to osmotic, mechanical, and oxidative stress (e.g., mean corpuscular fragility: 0.485±0.023 vs 0.401±0.036 % [NaCl], PNH vs healthy individuals, p<10^-5; PHZ-induced lysis: 6.73±3.67 vs 2.50±1.23 mg Hb/dL, PNH vs healthy individuals, p=0.005). While there was only a trend for increased ROS accumulation with and without external stimulation (e.g., intrinsic ROS: 2293±593 vs 1756±657 MFI/mg of protein, PNH vs healthy individuals, p=0.080), the oxidation of Hb was higher, as indicated by the levels of metHb and the presence of hemichromes on the membrane (5.7±1.5 vs 4.1±1.4 nM/mg, PNH vs healthy individuals, p=0.026). Interestingly, while cytosolic proteasomal activity was similar between the two groups, there was greater membrane-related activity in the patients’ RBCs (e.g., caspase-like activity: 111954±31862 vs 70001±22667 MFI/mg of protein, PNH vs healthy individuals, p=0.0042). Additionally, a non-negligible proportion of PNH mature RBCs also retained their mitochondria. As expected, reticulocytosis was present in PNH samples, and it presented positive correlations with the osmotic fragility of the cells. Additionally, ROS accumulation was proportional to the levels of bilirubin (R=0.663, p<0.05). While the low number of sub-categorized samples doesn’t allow drawing solid conclusions at the moment, it is notable that while treated patients were characterized by improved hematological indices and reduced hemolysis markers (e.g., LDH: 357±152 vs 2152±128, treated vs untreated, p=0.005; hemolysis: 0.054±0.028 vs 0.094±0.014 %, treated vs untreated, p=0.062), their RBCs exhibited susceptibility to mechanical lysis (1.079±0.364 vs 0.656±0.143 %, treated vs untreated, p=0.049).

Summary/Conclusion: Our preliminary results show evidence that RBCs in PNH, even in treated patients, differ significantly from the controls in terms of hemolysis, oxidative insults, and proteostasis, while the use of complement inhibitors leads to improvement in the clinical manifestation of hemolysis. The fact that mechanically-induced hemolysis is higher in treated patients, needs to be further elucidated along with the mitochondria-positive RBCs. The analysis is ongoing on a larger cohort of patients with omics data. A multivariate analysis is ongoing to enable determine the best treatment choice for each patient and potentially introduce a “biomarker” for disease activity-treatment benefit.