In the last few years, several studies have evaluated the clinical utility of flow cytometry immunophenotyping in myelodysplastic syndromes (MDS). We developed a flow cytometry approach for the evaluation of marrow dysplasia in MDS using discriminant analysis to generate erythroid and myeloid classification functions (

Hematol J
2004
; vol
5
S2: pag
S91
). However, a major problem in the immunophenotypic analysis of MDS is evaluation of erythroid dysplasia because of the limited availability of specific antibodies (Ab). In order to define new tools, we prospectively analyzed the expression of mitochondrial ferritin (MtF) and H ferritin subunit (HF) in bone marrow cells from 69 consecutive patients with MDS and 40 pathologic controls. Six-parameter, 4-color flow cytometry was performed with an EPIX XL flow cytometer (Beckmann Coulter, Fullerton, CA). Erythroid bone marrow cells were analyzed for the expression of HF using a monoclonal mouse Ab (rH02), expression of MtF using a polyclonal mouse Ab, and expression of Glycophorin A, CD71 and CD105 using monoclonal Abs (Instrumentation Laboratory, Milan, Italy). Data on HF and MtF expression obtained by flow cytometry were compared with those obtained on bone marrow smears by an immunocytochemical method (LSAB kit, Dakopatts, Glostrup, Denmark). Discriminant function analysis was performed to identify erythroid dysplasia, with the aim of distinguishing MDS without ringed sideroblasts (RS), MDS with RS, and controls. A significant correlation was found between immunocytochemically and cytometrically detected HF expression (r=.46, P=.016). Overall, the expression of HF was higher in MDS than in controls (P<.001), and was positively correlated with the degree of morphological erythroid dysplasia (r=.56, P<.001). A negative correlation was noticed between HF and CD71 expression (r=−.53, P<.001). Despite its mitochondrial localization, we were able to detect MtF by flow cytometry. Positive correlations were found between flow cytometry detection of MtF and both Prussian blue staining (r=.81, P<.001) and immunocytochemical data (r=.65, P<.001). MtF expression was significantly higher in MDS with RS than in MDS without RS or in controls (P<.001). Finally, compared with controls, MDS patients had a higher percentage of erythroblasts (P<.001), lower expression of CD71 (P<.001) and a higher expression of CD105 (P<.001). A classification function based on HF, CD71 and CD105 expression was defined with the aim of distinguishing MDS from controls:

\[Y=\ 12.44573{\ast}(HF_{X-mean})\ {-}\ 2.7698{\ast}(CD71_{X-mean})\ +\ 13.33752{\ast}(CD105_{X-mean})\ {-}\ 4.38426.\]

A patient will be classified as having erythroid dysplasia if the Y value obtained is greater than zero, or as not having dysplasia if the value is lower than zero. This function was integrated with the immunophenotypic evaluation of MtF (Xmean) in order to differentiate between MDS with and without RS. This function allowed us to classify 96% of MDS patients correctly with no false positive cases. Two MDS patients were incorrectly assigned to the RS subgroup by cytometric evaluation of MtF. Based on these findings, we conclude that flow cytometry analysis of MtF and HF expression is reliable and reproducible. Analysis of the expression of these antigens, together with CD71 and CD105, provide an accurate evaluation of erythroid marrow dysplasia and could be usefully included in the work-up of patients with MDS.

Topics:
dysplasia, flow cytometry, myelodysplastic syndrome, heart failure, transferrin receptors, antibodies, antigens, false-positive results, ferric hexacyanoferrate, ferritin

Author notes

Corresponding author

2005, The American Society of Hematology
2004
Sign in via your Institution