Myeloid Nuclear Differentiation Antigen (MNDA) Expression: A Useful Marker in the Diagnosis of Myelodysplastic Syndromes By Flow Cytometry

Background: Diagnostic workup for suspected myelodysplastic syndromes (MDS) is increasingly done with the aid of multiparamter flow cytometry (MFC) detecting aberrant antigen expression. Myeloid nuclear differentiation antigen (MNDA) is a hematopoietic protein expressed strongly in mature myeloid cells, but only weakly in myeloid progenitor cells. Recently, reduced MNDA expression was detected by MFC in monocytes (M), granulocytes (G) and myeloid progenitor cells (MP) in patients (pts) with MDS and MNDA as additional marker for MFC was proposed to improve diagnostic capabilities of MFC.

Aim: To verify differences in MNDA expression in M, G and MP of pts with MDS vs. those without MDS and to test its value as additional MFC parameter in a cohort with suspected MDS.

Patients and methods: We analysed bone marrow from 131 pts (median age 74, range 17-93 years) with suspected MDS by cytomorphology (CM), standard 10-color-MFC and cytogenetics (CG) in parallel. For detection of MNDA expression we applied a readily available five color intracellular staining assay using monoclonal antibodies against MNDA, CD45, CD64, CD15 and myeloperoxidase (MDS-Quant, Trillium Diagnostics, Bangor, ME). Different gating strategies to best define M [SSC/CD45 plot (A) vs CD16/CD15 plot (B)], G (A vs B) and MP (A) were used. MNDA expression in pts diagnosed with no MDS (n=35), possible MDS (n=40; based on morphologic dysplastic changes insufficient for diagnosis of MDS) and MDS (n=56) by CM was compared. We also correlated results to findings in standard MDS-MFC (n=128) and CG (n=120).

Results: Based on CM diagnosis, pts with MDS showed higher percentages of G and M with weak expression of MNDA (%dimG and %dimM) than no-MDS pts irrespective of the gating strategy [mean±SD, %dimG: 16±17 vs 6±7, p<0.001 (A) and 3±3 vs 1±2, p=0.011 (B); %dimM: 22±21 vs 15±12, p=0.054 (A) and 14±25 vs 4±4, p=0.004 (B)]. Diagnostically challenging cases with "possible MDS" by CM also displayed significantly higher %dimG and %dimM than no-MDS cases [%dimG 13±14 vs 6±7, p=0.006 (A), %dimM: 23±20 vs 15±12, p=0.027 (A) and 17±24 vs 4±4, p=0.002 (B)]. Differences between MDS and possible MDS existed only in higher %dimGra in MDS gated in plot B (2.7±3.3 vs 1.6±1.9, p=0.036). Conversely, MDS pts and pts with possible MDS had higher percentages of MP with high MNDA expression (%hiMP) than no-MDS pts (17±16 vs 9±8, p=0.005 and 14±13 vs 9±8, p=0.041, respectively). MNDA expression levels measured by mean fluorescence intensity (MFI) were diminished in M of pts with MDS vs no-MDS [26±12 vs 33±9, p=0.006 (A) and 30±14 vs 39±10, p=0.002 (B), respectively] and in pts with possible MDS vs no-MDS pts [26±12 vs 33±9, p=0.004 (A) and 30±2 vs 39±10, p=0.005, respectively]. Looking at the low level MNDA MP (dimMP), higher MFIs were found in MDS compared to no-MDS (1±0.8 vs 0.7±0.4, p=0.002) and in possible MDS vs no-MDS (1.0±0.7 vs 0.7±0.4, p=0.011). Comparing results of MNDA MFC and standard MFC we also found higher MFI in dimMP (1.0±0.8 vs 0.7±0.3, p=0.039) in pts diagnosed MDS by MFC (n=63) vs those without signs of MDS (n=14). In the inexplicit cases diagnosed MDS possible by MFC (aberrant antigen expressions not sufficient for MDS diagnosis, n=51) higher percentages of dimG [16±18 vs 9±10, p=0.015 (A) and 3±3 vs 1±2, p=0.005 (B)] and dimM [16±25 vs 7±13, p=0.02 (B)] were seen as compared to no-MDS, in line with CM results. Moreover, a lower MFI in M [26±12 vs 31±11, p=0.024 (A) and 29±14 vs 36±13, p=0.02 (B)] and a higher MFI in dimMP (1.0±0.8 vs 0.9±0.6, p=0.021) could be detected. Considering CG, cases with an aberrant karyotype (n=41) had higher values for %dimM (27±23 vs 18±16, p=0.035) as compared to those with a normal karyotype (n=79). Including the most significant markers [%dimG>12 (A), %dimM>13 (B), %hiMP>15 (A), 1 point each], we created a new MNDA score with a score of ≥2 indicating MDS. Of 14 pts with a score of ≥2, 13 pts were concordantly diagnosed MDS using CM for validation; 1 pt had no MDS by CM but was classified MDS by standard MDS-MFC.

Conclusions: Reduced MNDA levels and higher percentages of M and G with low MNDA expression as well as a higher percentage of MP with high MNDA expression could be confirmed in pts with MDS. Applying our newly defined MNDA score for MFC, identification of a subset of pts with MDS was possible at high specificity. Further analyses will have to evaluate this MNDA score incorporated into standard MDS-MFC panels to improve MFC-based diagnostic approaches for MDS.