Paroxysmal nocturnal hemoglobinuria (PNH) is characterized by somatic mutation of the PIGA gene, resulting in a clonal disorder of hematopoietic stem cells (HSC) that lack glycosylphosphatidyl inositol-anchored proteins (GPI-AP) on the cell surface. According to the extrinsic theory, PNH stem cells enjoy a selective growth advantage in the context of a cellular immune attack eliciting proinflammatory cytokines, such as seen during the course of aplastic anemia (AA). Since GPI-AP associate with lipid rafts in the plasma membrane and we have previously shown that lipid rafts exist as heterogeneous microdomains on the cell surface, we proposed that GPI-AP deficiency in PNH cells may result in altered raft-dependent signaling pathways to confer a potential growth advantage on PNH cells.

The p38 MAPK pathway has been shown to mediate the suppressive effects of proinflammatory cytokines on HSC. When we stimulated GPI-AP deficient K562 leukemic cells (PNH cells) with 0.2 to 20 ng/ml TNFα , there was markedly reduced p38 phosphorylation compared to K562 cells with intact GPI-AP (WT cells), determined by immunoblot analysis. While TNF receptor 1 has been shown to associate with lipid rafts, not all TNFR1 signaling is raft-dependent. To determine if all or only a subset of TNF signaling was affected by lack of GPI-AP, NFκ B p65(S536) phosphorylation was also examined. We observed increased levels of phospho-NFκ B in unstimulated PNH cells that was further induced upon TNFα stimulation. When methyl-β-cyclodextrin (MCD) was used to disrupt lipid rafts, there was a differential effect on PNH versus WT cells. PNH cells displayed increased phospho-NFκ B after MCD treatment, while WT cells increased phospho-p38 slightly, with no change in phospho-NFκ B. This finding suggests that lipid rafts in PNH cells normally sequester signaling components negatively regulating NFκ B, and disruption of rafts allows for potentiated NFκ B signaling. This imbalance in signaling pathway activation is manifest after culturing WT and PNH cells in the presence of MCD. Following a 5 hour exposure to MCD, a notable decrease in the population of WT cells (from 53% untreated to 33% after MCD) correlates with an increase in PNH cells (from 43% untreated to 59% after MCD), determined by flow cytometric analysis of CD55 and CD59 surface co-expression. When WT and PNH cells were cultured overnight with MCD, WT cells showed extensive apoptosis, from 90% viable untreated, to 8% viable after MCD. However, in agreement with the signaling analysis, PNH cells appeared morphologically identical to their untreated controls, with 83% viable in the MCD-treated group and 90% in the untreated control.

These observations in paired wild type and GPI-AP cell lines were confirmed in primary cell cultures. Upon culture of monocytes from PNH patients in the presence of TNFα , GPI-AP-deficient monocytes selectively expanded while normal monocytes decreased numbers. Thus, in comparison to normal cells, the relative increase of NFκ B to p38 activation in PNH cells after TNFα exposure may play a role in their selective survival/proliferation advantage the context of an immune attack.

In summary, we propose that altered lipid raft-dependent signaling in GPI-AP deficient cells may cue different responses to proinflammatory cytokines than normal cells with intact GPI-AP. Hence, an increased proportion of cells with a PNH phenotype, having survived the hostile bone marrow environment, will contribute a greater share to reconstitution of mature hematopoietic lineages.

Author notes

Corresponding author

Sign in via your Institution