Dysfunction of Lipid Raft Microdomains and Defective Apoptotic Signal Transduction in PNH.

Mutation of PIG-A gene in the hematopoietic stem cell is the primary pathogenic lesion in paroxysmal nocturnal hemoglobinuria (PNH). However, the key element to the understanding of the disease evolution is the ability of the defective hematopoietic clone to outgrow their normal counterparts. It is believed that this growth advantage operates only in the pro-apoptotic milieu that exists in the context of immune-mediated bone marrow failure. The lack of glycosylphosphatidyl inositol-anchored proteins (GPI-AP) on the cell surface of affected cells is likely responsible for the ability of the PNH clone to evade immune attack and expand. GPI-AP are known to be associated with lipid raft microdomains (LRM). LRM contain a large number of receptors and regulatory proteins, whose function depend upon the presence of GPI-AP. Consequently, we hypothesized that GPI-AP deficiency can result in altered LRM which can either blunt proaptoptotic or inhibitory signals or possibly lead to exaggeration of prosurvival stimuli. We utilized paired GPI-AP-deficient (CD−) and wild type (CD+) EBV-transformed B cells derived from a PNH patient, as well as similarly paired lines derived from K562. LRM isolation was performed using sucrose density gradient centrifugation. PNH cells clearly displayed an altered protein repertoire in the LRM, based on analysis by 2D gel electrophoresis and SDS-PAGE, some of which is likely due to the loss of GPI-AP. Immunoblotting revealed lack of the GPI-AP CD55, while Lyn kinase, a known raft-associated protein was present in both PNH and WT LRM. However, an altered LRM function in PNH cells was illustrated via phosphotyrosine immunoblotting: rafts from PNH cell showed a distinct phosphorylation pattern from that observed in WT cells, suggesting distinct raft-dependent signaling in these paired cell lines. In subsequent experiments we have studied the functional consequences of an altered LRM structure in PNH. Two models were established. I) Based on the association of Fas with LRM we stipulated that it may not adequately transduce apoptotic signals in GPI-AP deficient rafts. While Fas was present in LRM of both CD+ and CD− cells as shown by Western blot, in GPI-AP deficient CD− cells, the apoptotic response to Fas agonist CH11 was attenuated in comparison to CD+ cells (23% vs. 60% of apoptotic cells). II) In the second model, we measured differential response of PNH cells to TNFa signaling, shown to be associated with LRM. As a readout, we used TNFa-induced p38 MAPK phosphorylation, which was shown to suppress hematopoiesis; PNH as compared to WT cells showed markedly reduced phosphorylation of p38 after TNF treatment, and exhibited increased basal phosphorylation of NFkB and AKT. These results suggest decreased inhibitory and increased survival signaling in PNH cells treated with TNF. Importantly, we observe increased activity of PDK1, the PI3K-dependent enzyme upstream of AKT, in LRM of PNH cells versus WT cells. Thus, our data show that lipid rafts in PNH cells orchestrate a skewed pro-survival signaling response that may play an important role in their ability to persist in the immune-mediated pro-apoptotic milieu that is a hallmark of bone marrow failure.