Large CD45 Protein Tyrosine Phosphatase (PTP) Isoforms Predominate in Myelodysplastic Erythroid Precursors and Attenuate the Erythropoietin (EPO): STAT5 Signal.

Ineffective erythropoiesis in MDS is associated with impaired EPO-Receptor activation of STAT5 [Hoefsloot et al. Blood1997; 89:1690]. However, identification of regulatory molecules that impede the EPO-R signal remains elusive. CD45 is a transmembrane PTP expressed in hematopoietic cells that co-localizes with the EPO-R and attenuates signaling by dephosphorylating tyrosines on Jak-2 and Lyn to abolish kinase activity. Expression within the erythroid lineage is limited to CD71Hi erythroblasts and is down-regulated with differentiation. CD45 is comprised of several isoforms generated by alternate splicing. In lymphoid cells, dimerization of the smallest isoform, CD45RO, which predominates in normal erythroids, suppresses PTP activity, whereas larger isoforms such as CD45RA or -RB, favor monomeric distribution and primed PTP activity. We recently found that lenalidomide augments EPO-R:STAT5 activation in non-del5q MDS CD71Hi cells by inhibiting the CD45 PTP, implying inherent CD45 PTP over-activity. To investigate CD45 regulation of the EPO-R:STAT5 signal, we characterized CD45 isoform distribution in CD71Hi erythroids from normal [N=7] and MDS marrow specimens [N=9], and evaluated the relationship to EPO-R:STAT-5 activation and PTP activity. Flow quantitation of EPO-induced STAT-5 phosphorylation (PHN) confirmed that PHN capacity is significantly impaired in lower risk, non-del5q MDS compared to normal erythroid precursors (median [SD] log10 geometric pSTAT5 MFI EPO:isotype ratio, 0.435 [0.21] MDS vs. 1.284 [0.06] Normal; P<0.0001). Comparison of CD45 isoform distribution in showed that CD45-RO predominates in normal CD71Hi cells (mean [SD] RA:RO ratio, 0.33[0.23); RB:RO, 0.04[0.03)), whereas CD45-RA and -RB are preferentially expressed in MDS (RA:RO 1.9[0.54]; RB:RO 0.28[0.12]; P<0.0001). Regression analysis of the relationship between CD45 isoform distribution and STAT-5 PHN after EPO stimulation showed that cellular RA and RB fraction was highly correlated (r=0.927, P<0.001) and PHN capacity inversely correlated with CD45-RA or -RB proportion (P<0.0001). In contrast, the magnitude of lenalidomide enhancement of EPO:STAT5 PHN directly correlated with CD45RA and -RB isoform fraction (P=0.0008), and was significantly higher in MDS compared to normal erythroids (median log10 geometric MFI [SD] pSTAT5 EPO+CC-5013:EPO, 0.989 [0.25] MDS vs. 0.0195 [0.07] Normal; P=0.0005). Genomic DNA genotyping for CD45 polymorphisms by allele-specific PCR & sequencing showed no skewing in SNP distribution (exon 6, A138G; exon 4, C77G & C59A) among MDS patients [N=101] compared to normal controls [N=44]. These data provide the first evidence that CD45 PTP over-activity impairs the EPO-R signal in MDS. Our investigations show that the magnitude of EPO-R:pSTAT5 signal reduction directly correlates with the proportion of larger CD45 isoforms in erythroid precursors and is restored by CD45 PTP inhibition. These findings suggest that erythroid CD45 isoform profile may identify patients potentially responsive to rhu-EPO or selective PTP inhibitors.