Kinesin Family Member 20A (KIF20A) Is Specifically down Regulated in 5q- CD34+ and CD61+ Cells and Uniparental Disomy Is Not a Feature of 5q- Syndrome.

The specificity of the clinicomorphological characteristics in 5q- syndrome has led to a search for putative tumour suppressor genes localised to the critical deleted region (CDR) of 5q-. To date, microarray expression studies in 5q- syndrome patients have been carried out on CD34+ cells, neutrophils or total bone marrow (BM) cells. We have used a Affymetrix microarray approach to investigate allelic imbalance (AI) and the transcription profile of CD61+ and CD34+ cells from patients (n=17) presenting with 5q- of which 10 have an “isolated 5q-“. the rest having additional cytogenetic abnormalities. BM CD34+ and CD61+ cells from normals (n=10) were controls in expression profiling. DNA or total RNA from CD61+ or CD34+ cells and constitutional DNA was prepared and hybridised to the 500K mapping or U113 plus 2.0 arrays. GC-RMA normalized expression data was analysed using non-parametric testing with Genespring GX software to generate lists of differentially regulated genes. DChip was used to analyse SNP data. From a list of changing genes 48/720 occur in both the CD61+ and CD34+ cell profiles. 16/48 of these genes map to 5q of which 10/16 map to 5q31. Of these genes KIF20A showed the most marked change (5 and 3.8 fold down) in CD34+ and CD61+ cells respectively. KIF20A is a microtubule-associated protein that is upregulated in mitosis and localizes to the spindle midzone during anaphase. Knock out of KIF20A protein results in binucleate nuclei indicating its involvement in both cytokinesis and nucleokinesis. Haploinsufficiency of this gene, may account for mononuclear megakaryocytes that are so characteristic of 5q- syndrome. Interestingly other members of the kinesin family expressed in 5q- CD34+ cells are down regulated in 5q- CD61+ cells (KIF21B, KIF2, KIF3A and KIF1B) indicating that in CD34+ cells, a compensatory mechanism for KIF20A haploinsufficiency may exist which is absent in megakaryocytes. The 5q- deletion (q14.2–q33.3) was detected in 6/10 CD61+ and 11/17 CD34+ cell populations by SNP analysis. The disparity reflects the admixture of normal and MDS CD61+ and CD34+ cells. Three patients presented with additional cytogenetic abnormalities, two had −7 or del 7(q22–q36), both being detected with the SNP array. The third patient had trisomy 8 that was not detected. SNP microarray identified 5q31.1–5q33.2 as the minimal deleted segment, which includes previously described CDR. Paired analysis of CD61+ and CD34+ with constitutional DNA revealed only 2/17 patients had additional abnormalities in CD34+ cells. One with a del12(q21.1–22) (21Mb). The 2^nd had −13(q12.3–34;83Mb),+ 13(q11–12.3; 12.5Mb) and +11(q22.3-qterm; 30Mb) These changes were not detected by cytogenetic analysis. 2/6 patients shared 13Mb of UPD on chromosome 3(p12.2–q11.2) and chromosome 9(q33.1–q34.1) in CD34+ cells as well as constitutive DNA. We conclude that unlike AML, 5q- Syndrome does not have significant rate of UPD. Only a small number of genes are under expressed in both CD61+ and CD34+ cells and amongst these KIF20A is the most significant change and may be responsible for some of the clinicomorphological features of 5q- abnormality.