ALBERT

Description: Key Points The mechanism of bone marrow failure (BMF) in PNH is not known. Novel CD1d-restricted, GPI-specific T cells are present in PNH patients and might be responsible for BMF.

Description: PNH is a rare acquired clonal disorder of the hematopoietic stem cell, characterized by a somatic mutation that inactivates the X-linked PIGA gene: this in turn results in deficiency on the cell surface of all proteins anchored by the glycosylphosphatidylinositol (GPI) molecule. Two of these proteins,CD55 andCD59, are complement regulators and their deficiency is responsible for the susceptibility of red cells (RBCs) from the mutant clone to lysis by activated complement. Since oxidative damage is another well-known mechanism of hemolysis (as in G6PD deficient red cells), we have investigated whether this plays a role also in PNH. To this end, we have carried out experiments on RBCs from healthy donors and on PNH-like RBCs (obtained in vitro from the same donors through the use of anti-CD55 and anti-CD59 blocking moAb). After exposure to AB0-compatible serum (in which thecomplement alternative pathway was activated by mild acidification) all PNH-like (but not normal) RBCs were lysed. In parallel experiments in which complement was blocked by eculizumab (ECU) - a moAb that binds to the complement component C5 and controls intravascular hemolysis in PNH patients - we measured the levels of reactive oxygen species (ROS) by the dichlorofluorescin diacetate assay. We found no significant difference of ROS levels between normal RBCs and PNH-like RBCs. We next tested in a similar way G6PD-deficient RBCs, because these are known to be exquisitely sensitive to oxidative damage. We found that ROS levels were significantly higher in the G6PD deficient RBCs that have been made PNH-like (Fig. 1). Thus, complement activation on the surface of PNH-like RBCs results in the production of ROS that can be demonstrated when C5-blockade prevents complement-mediated lysis of RBCs. The notion that G6PD deficiency can interact with PNH was strongly corroborated by the clinical observation of a 40yo woman from Sardinia (Italy) with a 2 years history of pancytopenia, who then developed florid hemolytic PNH: she had anemia with normal granulocyte and platelet counts, dark urine, high reticulocytosis, LDH up to 5x upper normal level, 95% GPI-negative granulocytes. When the patient was started on ECU. LDH levels promptly returned to normal, PNH RBCs rose from 20% (before ECU) to 42%, but reticulocyte count (~250x109/L) and blood transfusion requirement remained high (10 units in the last year). 39% of the GPI-negative RBCs had bound C3 fragments The peripheral blood smear revealed marked macro-anisocytosis, poikilocytosis, spherocytes, and hemighosts: a picture consistent with oxidative damage as seen in G6PD deficient patients during a hemolytic attack. The RBC G6PD activity was about one-half of normal (5 IU/g Hb), and DNA analysis revealed heterozygosity for the G6PD Mediterranean (Med) mutation. By mRNA sequence analysis we found that the GPI-negative clone expressed only the G6PD Med allele, suggesting that the PIG-A mutation took place in a stem cell in which the normal G6PD gene was on the inactive X-chromosome (G6PD, like PIG-A, is on the X chromosome); therefore, all the patients' PNH RBCs were also all G6PD deficient. We have previously shown that the clinical expression of PNH can be influenced by inherited factors: specifically, a polymorphism of the complement receptor 1 (CR1) gene correlates with the blood transfusion requirement of patients on ECU (Rondelli et al, Haematologica 2014). However, the patient here reported was homozygous for the more favorable allele ofCR1. Instead, in keeping with our experimental data, the poor response to ECU seen in this patient results probably from a unique interaction, within the same population of RBCs, between the acquired PNH abnormality and her inherited G6PD deficiency This type of interaction is novel and it seems to have pharmacogenetic implications. Indeed, on its own G6PD deficiency affects mildly the clinical expression of PNH, because complement activation causes RBC lysis regardless; however, paradoxically, when the lysis of PNH RBCs is prevented by C5 blockade, complement activation results in oxidative damage, with which PNH G6PD deficient RBCs are unable to cope. Except for one case previously reported by Oni et al (Blood 1970), this is the first detailed study of PNH associated with G6PD deficiency. Since in some parts of the world the frequency of G6PD deficiency can be as high as 30% or more, we expect that more cases of this association will be discovered in the future. Figure 1 Figure 1. Disclosures No relevant conflicts of interest to declare.

Description: Abstract 4408 PNH is a hemolytic disorder resulting from dysregulation of complement on the rbc, and is associated with immune mediated marrow failure and marked hypercoagulability. Thrombosis can be prevented with anticoagulation or the complement inhibitor eculizumab, and it often involves the hepatic, portal, and splenic veins. This can result in an increase in portal pressure and complications that we term the “Thrombosis-Splenomegaly-Thrombocytopenia” syndrome (TST). TST is frequently associated with abdominal pain, and thrombocytopenia may be simultaneously exacerbated by marrow failure. Particularly, thrombocytopenia can complicate efforts at anticoagulation, leaving the patient exposed to the risk of further events. In some patients thromboses can reversed with tPA, but others will not be candidates because they have presented late after their thromboses. Ablating splenic function would be desirable: however, with a high risk of perioperative thrombosis, surgical splenectomy may be hazardous, particularly in patients who have already had thrombosis, and will confer a risk of sepsis. Here we report on 4 patients with PNH and late-presenting TST. We referred these 4 patients for selective splenic artery embolization (SSAE), which involves cannulating branches of the splenic artery– beyond the hilum– and introducing gelfoam and microcoils. We planned a multi-session stepwise approach: we started with the inferior branches of the splenic artery, to decrease the risk of pleural effusions. We planned to infarct no more than 1/3 of the spleen at any one time and to allow weeks to months for recovery. We routinely administered vaccinations and discontinued anticoagulation temporarily, and patients were given prophylactic antibiotics, fluids, analgesics, and antipyretics, and they were observed in the hospital with a back-up surgical team. Prior to the procedure, the median platelet count was 17 and the median spleen size was 22 cm. Patients 1–4 were treated with 3,2,1, and 3 procedures respectively, which resulted in a significant reduction in spleen volume in all 4 patients. The post procedure platelet counts were respectively 123, 12, 44, and 90, which represented a significant increase for all patients except patient 2, who remained thrombocytopenic; since there was evidence of bone marrow failure, she underwent a successful unrelated SCT. Patients 1,2, and 4 all had had abdominal pain before the procedure, which very significantly improved after recovery from the procedure, which we attribute to decreased venous return from the spleen into the portal circulation. Patients 1–3 were treated before eculizumab was available and patient 1, 3, and 4 are now on it. Patient 4 is a special case in that she had been on eculizumab for several months prior to the SSAE, but had had only a partial reduction in the red cell transfusion requirement; C3d deposition on rbcs was documented by flow cytometry, and it was thought that extravascular hemolysis was limiting her response to eculizumab, as has been described. Of note, her hemoglobin began to rise after the embolization procedure, concurrent with the increase in the platelet count, and a significant further reduction in the red cell transfusion requirement, suggesting that partially reducing splenic function markedly reduced C3-mediated extravascular hemolysis. Of the 9 procedures performed on these 4 patients, only one was complicated by a clinically significant left sided pleural effusion, which was drained by thorascope. All patients are doing well between 3.5 and 11 years after their procedures. We conclude that SSAE, when performed by an experienced interventional radiologist, is relatively safe in patients with PNH and TST, it produces sustained correction of hypersplenism without the risk associated with surgery or the asplenic state, and can be of benefit to patients on eculizumab who have hypersplenism. Pt Age at diagnosis Lowest plt count pre-SSAE Spleen size (cm) pre-SSAE Abdominal pain pre-SSAE Bone marrow Plt count post SSAE Spleen size post SSAE Abdominal pain after recovery from SSAE Complications of SSAE Follow up (yrs) 1 30 42 21 + Hypercellular Megakaryocyte aggregates 123 12 No Abd pain and fever 10 2 18 19 19 + Hypocellular Megas reduced 12 12 No Pain, pleural effusion SCT 3 27 14 36 – Erythroid hyperplasia Megas adequate 44 NA No Abd pain 11 4 26

Description: Abstract 4239 Paroxysmal Nocturnal Hemoglobinuria (PNH) is associated with clonal expansion of stem cells with an acquired somatic PIG-A mutation. The PIG-A gene is essential for the biosynthesis of glycosylphosphatidylinositol (GPI), and mature blood cells derived from the PNH stem cell clone exhibit a loss of all proteins that require this structure for attachment to the cell surface, notably the complement inhibitors CD55 and CD59. The loss of these proteins on the surface of red cells is responsible for hemolysis, and thrombosis may be the consequence of the loss of these proteins from the surface of platelets: both hemolysis and thrombosis can be attenuated by anti-complement therapy. Thrombosis can occur, however, despite anticomplement therapy and despite anticoagulation and has been historically the most important determinant of death in patients with PNH. Some patients will present with thrombosis and some patients will not be candidates for anticoagulation or anticomplement therapy: therefore treatment of thrombosis remains an important part of the management of PNH patients. Thrombolysis with tissue plasminogen activator (tPA) in PNH has been reported in small series or case reports, generally with encouraging outcomes. Here we report what we believe to be the largest series on the outcome of the use of tPA. Of 38 patients with PNH who had at least one thrombotic event, 13 were thought to have had a thrombus sufficiently recent to be amenable to fibrinolysis; of these, 4 patients were regarded as ineligible on account of active hemorrhage or high risk of hemorrhage. Of the 9 eligible patients who received tPA, all of whom had potentially life-threatening thromboses, 3 also required tPA on subsequent hospitalizations, and the results of a total of 15 hospitalizations during which tPA infusions were given are reported here. tPA was given in the ICU by systemic infusion through a peripheral vein at a dose of 1 mg/kg delivered over 24 hours, with anticoagulants withheld temporarily during this time. Response was monitored by follow-up imaging, and most patients required several 24 hour infusions. Platelets were given for thrombocytopenia and FFP was given to reverse oral anticoagulation or when low circulating plasminogen was documented. On all 15 occasions a radiologically documented response was obtained, including reversal of thrombosis in hepatic veins, portal veins, the IVC, cerebral dural venous sinuses, and an intrahepatic portocaval shunt. Among the 15 courses of tPA, serious hemorrhagic complications developed in 3 cases. At last follow-up visit, of the 9 patients treated, 3 have expired, one patient (who has been non-compliant with post treatment anticoagulation and anticomplement therapy) was in good clinical condition despite extensive residual occlusions, and 5 others were in good to excellent condition in terms of clinical and radiological outcome. The only patient in whom tPA may have contributed to a fatal outcome also had complications of ‘heparin induced thrombocytopenia with thrombosis’ (HITT), which we diagnosed in a milder form in 3 additional patients. The other two fatalities were associated with bowel edema (probably due to progressive small vessel thrombosis) in one case, and a progressive concurrent myeloproliferative disorder associated with a JAK2 mutation in the other case. On the other hand, we feel tPA must be credited as having been immediately life-saving in 2 patients who had been moribund with Budd-Chiari syndrome, and in one who had impending renal failure associated with an IVC thrombosis. Given the high incidence of HITT, we favor the use of direct thrombin inhibitors or fondaparinux rather than heparin products in patients with PNH. Given the high mortality and morbidity associated with thrombosis in PNH patients, and given the excellent radiographic responses, we conclude that, in spite of the risk of hemorrhage, thrombolysis is strongly indicated to reverse intra-abdominal and intracranial thromboses. Disclosures: No relevant conflicts of interest to declare.

Description: Key Points Increase in HDAC binding is required for HDAC inhibitors to enhance gene transcription. G6PD deficiency in erythroid precursors can be restored by HDAC inhibitor-mediated increased transcription of the variant gene.

Description: Abstract 4237 PNH is a chronic, life-threatening, acquired disease associated with deficiency of GPI-anchored complement inhibitory proteins on blood cells. The resulting defective regulation of terminal complement activation is responsible for hemolysis and can lead to thromboembolism (TE), chronic kidney disease (CKD) and pulmonary hypertension. The risk of TE is high, with an observed 6.24 venous TE events per 100 patient years, or approximately 62-fold higher compared to the general population: in fact, TE accounts for 40–67% of PNH related deaths. The effectiveness of anticoagulation (AC) in PNH patients (pts) is uncertain, as AC treated PNH may still experience TE. The terminal complement inhibitor eculizumab reduces intravascular hemolysis rapidly and significantly; it also leads to a reduction in TE events, pulmonary hypertension and improvements in CKD and quality of life. Here, we report on prolonged treatment of PNH patients with eculizumab for safety and sustained patient outcomes. Methods: All pts (N=195) in the PNH eculizumab clinical trials (Pilot (N=11), TRIUMPH (N=87) and SHEPHERD (N=97)) and subsequent Extension studies were assessed for long term safety and efficacy. Median age was 40 yrs, 54% female, 29% had a history of aplastic anemia and 1.5% with history of myelodysplasia. TE was reported in 32% (63/195) of pts prior to eculizumab treatment. There was high adherence to long term treatment; 90% (175/195) of pts completed the parent and extension trials. Results: The median eculizumab treatment duration was 29 mo (1 -66; IQR:23-32m); with a total eculizumab exposure of 474.1 patient-years. Intravascular hemolysis was rapidly reduced in 100% of pts after eculizumab treatment. LDH was reduced from a median of 2,133 U/L (∼10x ULN) at baseline to 310 U/L at 1 month of treatment (P

Description: Paroxysmal nocturnal hemoglobinuria (PNH) is characterized by complement-mediated intravascular hemolysis because of the lack from erythrocyte surface of the complement regulators CD55 and CD59, with subsequent uncontrolled continuous spontaneous activation of the complement alternative pathway (CAP), and at times of the complement classic pathway. Here we investigate in an in vitro model the effect on PNH erythrocytes of a novel therapeutic strategy for membrane-targeted delivery of a CAP inhibitor. TT30 is a 65 kDa recombinant human fusion protein consisting of the iC3b/C3d-binding region of complement receptor 2 (CR2) and the inhibitory domain of the CAP regulator factor H (fH). TT30 completely inhibits in a dose-dependent manner hemolysis of PNH erythrocytes in a modified extended acidified serum assay, and also prevents C3 fragment deposition on surviving PNH erythrocytes. The efficacy of TT30 derives from its direct binding to PNH erythrocytes; if binding to the erythrocytes is disrupted, only partial inhibition of hemolysis is mediated by TT30 in solution, which is similar to that produced by the fH moiety of TT30 alone, or by intact human fH. TT30 is a membrane-targeted selective CAP inhibitor that may prevent both intravascular and C3-mediated extravascular hemolysis of PNH erythrocytes and warrants consideration for the treatment of PNH patients.

Description: Abstract 983 Eculizumab, a monoclonal antibody that binds to the complement component C5, controls intravascular hemolysis in PNH patients and causes significant clinical improvement. However, the extent to which patients respond is variable in terms of hemoglobin levels, and some patients remain red blood cell (RBC) transfusion-dependent in spite of regular eculizumab treatment. We have hypothesized that genetic polymorphisms of complement-related genes might contribute to determining this variability of response to eculizumab. In order to test this hypothesis, we have analyzed polymorphic alleles of the Complement Receptor 1 (CR1) gene and of the Complement Component 3 (C3) gene. We have genotyped by standard methods (1) the HindIII restriction length polymorphism (RFLP) in the CR1 intron 27 (rs11118133 A〉T) whose co-dominant alleles, H (common) and L (rare), are associated with high and low levels of CR1 expression on RBC, respectively; (2) the single nucleotide polymorphism rs2230199 C〉G of the C3 gene, responsible for the allelic electrophoretic variants, S (slow, common allele) and F (fast, rare allele), that influence the complement alternative pathway activity. We have studied 72 patients with hemolytic PNH treated with eculizumab for at least 6 months. Patients with clear evidences of bone marrow failure (platelets ≤ 50,000/μL, neutrophils ≤1000/μL) have not been included in this study because this condition may affect the clinical response independently of the control of hemolysis. In order for our criterion for response to be stringent and objective we used blood transfusion requirement: patients who received no RBC transfusion have been defined as Responders; patients who received any RBC transfusion have been defined as Poor Responders. In this series of 72 hemolytic PNH patients on eculizumab the allelic frequency of the CR1 rare allele L was 0.31, and that of the C3 rare allele F was 0.19. We found no correlation between response and the C3 genotype (Chi square for trend: p=0.939). On the other hand the proportion of Poor Responders is significantly higher in patients who were heterozygotes (H/L) for the CR1 polymorphism, and even higher in those who were homozygous for the rare allele (L/L) of the CR1 HindIII RFLP polymorphism (Chi square for trend: p=0.016): see figure 1. We do not know yet the precise mechanism whereby the CR1 genotype affects the response to eculizumab. CR1 enhances the decay of C3 and C5 convertase by binding to C3b and C4b and it plays a role in the clearance of immune complex and in phagocytosis. Thus, the variable levels of CR1 expression on RBC membrane associated with the HindIII RFLP polymorphism may affect the regulation of complement alternative pathway. Moreover, it is well established that in almost all PNH patients on eculizumab a significant fraction of GPI-negative RBCs are opsonized by C3 and therefore may undergo extravascular hemolysis. Thus, a possible explanation of our findings is that the lower levels of CR1 expression on RBCs associated with H/L and L/L genotypes are responsible for a more intense extravascular hemolysis in PNH patients on eculizumab. In conclusion, our data show that the CR1 HindIII polymorphism may predict the poor response to eculizumab in PNH patients and provide additional insight in the mechanisms responsible for the variable clinical effectiveness of complement blockade. Disclosures: Risitano: Alexion: Membership on an entity's Board of Directors or advisory committees, Research Funding. Peffault de Latour:Alexion: Consultancy. Iori:Alexion: invited speaker Other. Socie:Alexion: Consultancy. Luzzatto:Alexion: Consultancy. Notaro:Alexion: Consultancy.