OUP user menu

Quantitative Assessment of Myeloid Nuclear Differentiation Antigen Distinguishes Myelodysplastic Syndrome From Normal Bone Marrow

Sara A. McClintock-Treep MD, Robert C. Briggs PhD, Keith E. Shults, Leanne A. Flye-Blakemore MS, Claudio A. Mosse MD, PhD, Madan H. Jagasia MD, Andrew A. Shinar MD, William D. Dupont PhD, Greg T. Stelzer PhD, David R. Head MD
DOI: http://dx.doi.org/10.1309/AJCP00SHTQCVUYRI 380-385 First published online: 1 March 2011


By using flow cytometry, we analyzed myeloid nuclear differentiation antigen (MNDA) expression in myeloid precursors in bone marrow from patients with myelodysplastic syndrome (MDS) and control samples from patients undergoing orthopedic procedures. The median percentage of MNDA-dim myeloid precursors in MDS cases was 67.4% (range, 0.7%–97.5%; interquartile range, 44.9%–82.7%) of myeloid cells, with bimodal MNDA expression in most MDS samples. Control samples demonstrated a median MNDA-dim percentage in myeloid precursors of 1.2% (range, 0.2%–13.7%; interquartile range, 0.6%–2.7%), with no bimodal pattern in most samples. The area under the receiver operating characteristic curve for MNDA-dim percentage in myeloid precursors was 0.96 (P = 9 × 10–7). Correlation of MNDA-dim levels with World Health Organization 2008 morphologic diagnoses was not significant (P = .21), but correlation with patient International Prognostic Scoring System scores suggested a trend (P = .07). Flow cytometric assessment of MNDA in myeloid precursors in bone marrow may be useful for the diagnosis of MDS.

Key Words:
  • Myeloid nuclear differentiation antigen
  • MNDA
  • Myelodysplastic syndrome
  • MDS
  • Diagnostic test

Myelodysplastic syndrome (MDS) is a set of clonal marrow failure disorders that are difficult to diagnose owing to the variability of standard diagnostic parameters, admixture of normal marrow in MDS samples, and the large differential diagnosis of MDS.13 Subclassification of MDS is important because the low-grade subtypes (refractory anemia [RA] with ringed sideroblasts, 5q– syndrome, and, to a lesser extent, RA) mean relatively stable disease with survival approaching that of age-matched peers, while high-grade subtypes (refractory cytopenia with multilineage dysplasia [RCMD] and RA with excess blasts [RAEB]) are progressive and usually fatal without intervention.1,3

Unfortunately, subclassification of MDS also remains difficult and imprecise. A single marker that reliably distinguishes MDS from non-MDS would greatly facilitate diagnosis and might aid MDS subclassification. Gene expression profiling has identified myeloid nuclear differentiation antigen (MNDA) as a highly down-regulated gene in MDS, in particular in high-grade cases.4,5 MNDA is a hematopoietic/lymphoid-cell specific nuclear protein that is uniformly expressed in normal differentiating myeloid precursors.6 MNDA belongs to the interferon-regulated IFN200 gene family, members of which contain an amino terminal pyrin domain that may regulate programmed cell death.7,8 MNDA expression may provide a protective advantage to cells by inhibiting programmed cell death.9 We previously reported decreased expression of MNDA protein in myeloid precursors in intermediate- to high-grade MDS using immunohistochemical and flow cytometric analysis.9 To extend this work, in the current study, we used quantitative flow cytometric analysis to evaluate MNDA expression in myeloid precursors in a series of MDS and normal control marrow samples.

Materials and Methods

Case Selection

MDS marrow samples were obtained from patients with MDS undergoing bone marrow examinations for clinical purposes who were receiving only supportive care. Control marrow samples were obtained intraoperatively from random, otherwise healthy patients with no known hematologic comorbidity undergoing elective orthopedic joint replacement procedures; samples were removed intraoperatively by trephine from the medullary cavity preparatory to placement of the prosthesis and processed in parallel with MDS samples. Marrow samples were collected under protocols approved by the institutional review board.

Morphologic Review With Confirmation of Diagnosis

For MDS cases, Wright-Giemsa–stained peripheral and bone marrow aspirate smears, H&E-stained particle preparation and biopsy sections, and iron stains were reviewed with CBC, clinical flow cytometry, and cytogenetic results for confirmation of diagnosis (S.A.M.-T. and D.R.H.). Morphologic review was blinded to MNDA results. Diagnosis, subclassification, and International Prognostic Scoring System (IPSS) risk categories were assigned by using published criteria.1,10

Flow Cytometric Methods

Following collection, all samples were diluted 50% with RPMI and kept at room temperature until analysis. Flow cytometric methods have been published previously.9 Briefly, analyses were performed using an FC500 flow cytometer (Beckman Coulter, Fullerton, CA) using a standard quality control program.11 Cells were stained with CD45-phycoerythrin and CD34–Texas red (Beckman Coulter), washed twice in phosphate-buffered saline + 2% fetal calf serum, and then permeabilized with PermiFlow (InVirion, Oak Brook, IL). A rat monoclonal antibody specific for MNDA (clone 3C1) was conjugated to Alexa 488 using a protein labeling kit (Invitrogen, Carlsbad, CA) and used in these studies at a concentration proven to be specific for granulocyte-monocyte precursors. (This antibody is currently available at Cell Signaling Technology, Danvers, MA, and Trillium Diagnostics, Bangor, ME.) Analysis was performed using Winlist 5.0 software (Verity Software, Topsham, ME) with dynamic data exchange (DDE) links to Excel (Microsoft, Redmond, WA) using modifications of published methods. Analyses were blinded to clinical results. Gating to identify cell subsets is shown in Figure 1A. Myeloid precursors were identified as high side scatter/intermediate CD45/CD34– cells. Mean fluorescence intensity for MNDA expression in the lymphocyte population (defined as cells with low side scatter/high CD45/CD34– cells) was used as an internal biologic control above which the levels of MNDA in the granulocyte-monocyte progenitors were determined to be MNDA-diminished or MNDA-normal.

Statistical Considerations

The distribution of MNDA expression levels in MDS cases compared with controls was analyzed by using the Wilcoxon rank sum test. The sensitivity and specificity of different MNDA cut points for diagnosing MDS were assessed by receiver operating characteristic (ROC) curves.12 The nonparametric trend test of Cuzick13 was used to assess how MNDA expression levels varied with morphologic subclassification and IPSS categories in patients with MDS. All analyses were performed using the Stata statistical software package (Stata Statistical Software, release 11, StataCorp, College Station, TX).

Figure 1

A, CD45 vs side scatter display of cell populations in a control sample. Distributions are based on CD45 and side scatter, as in routine clinical sample analysis by flow cytometry. Blasts were confirmed with CD34 positivity (not shown). BD, Myeloid nuclear differentiation antigen (MNDA) expression intensity vs number of events in myeloids for representative control (B) and myelodysplastic syndrome (MDS; C and D) samples. The control sample (B) shows a homogeneous peak of normal MNDA expression, with few cells showing diminished expression. By contrast, the MDS samples show a homogeneous peak of normal expression and a distinct population with diminished MNDA expression (C, a small MNDA-dim population; D, a large MNDA-dim population). Lymphs, lymphocytes; Monos, monocytes; NRBCs, nucleated RBCs.


Patient Information

Marrow samples were analyzed from 20 patients with MDS. Patient characteristics are summarized in Table 1. There were 12 male and 8 female patients, with a median age of 58 years (range, 5–82 years). Marrow blast percentages varied from 0% to 13%. Cytogenetic abnormalities were typical for patients with MDS. Case distribution based on study diagnosis using World Health Organization (WHO) criteria1 was as follows: 12 patients with RCMD, 3 with RAEB-1, 4 with RAEB-2, and 1 therapy-related MDS. Of note, 1 patient with a WHO study diagnosis of RCMD (case 20) had marginal multilineage dysplasia with ringed sideroblasts and a clinical diagnosis of RA with ringed sideroblasts. The IPSS distribution was as follows: 3 low, 7 intermediate-1, 7 intermediate-2, and 3 high.10 Of the patients, 11 had previously received only supportive care for MDS. In the remaining patients, previous therapeutic interventions included lenalidomide, darbepoetin alfa, and cytotoxic chemotherapy; these patients were receiving only supportive care at the time of this study.

View this table:
Table 1

For the study, 20 marrow samples were analyzed from 19 normal control subjects Table 2, including simultaneous bilateral samples from 1 patient. There were 8 men and 11 women, with median age of 61.9 years (range, 45–91 years). Control subjects had no history of antecedent hematopoietic disease.

View this table:
Table 2

Samples were processed a mean time of 3.2 hours after collection, with no difference in elapsed time between control and MDS samples. In all, 30,000 to 50,000 events were processed per sample, at least half being myeloid events.

MNDA Values for MDS Samples vs Healthy Control Samples

MNDA results, expressed as the percentage of myeloid progenitors with MNDA-diminished expression relative to normal myeloid levels, are summarized for patients with MDS in Table 1 and for control subjects in Table 2. Typical MDS and control patient results for MNDA expression in myeloid progenitors are shown in (Figure 1A) Figure 1B, Figure 1C, and Figure 1D. Most patients with MDS showed bimodal populations of MNDA-diminished and MNDA-normal myeloid cells, with mean fluorescent intensity for the 2 populations separated by 1 to 1.5 logs of a 4-log display. The median level of MNDA-dim myeloid precursors for MDS samples, expressed as a percentage of myeloid cells, was 67.4% (range, 0.7%–97.5%; interquartile range, 44.9%–82.7%) of myeloid cells Figure 2A. Comparison of MDS results with the results for normal control samples was striking. The analogous median level of MNDA-dim myeloid precursors for normal control samples was 1.2% (range, 0.2%–13.7%; interquartile range, 0.6%–2.7%; Figure 2A). The area under the ROC curve (concordance index12) for this comparison was 0.96, which suggests that MNDA expression levels can be used to discriminate between normal cases and patients with MDS with great sensitivity and specificity Figure 2B. Classifying cases with at least 8.4% MNDA-dim myeloid progenitor cells as MDS yields a sensitivity of 95% and a specificity of 94.7%. The distribution of the percentage of MNDA-diminished cells differed between control subjects and patients with MDS with overwhelming statistical significance (P = 9 × 10–7).

Only 2 samples in the MDS data set overlapped the MNDA values in the control population. One (case 18) had a bimodal MNDA distribution pattern, with a distinct MNDA-diminished population (8.4%). The second (case 20) had normal MNDA expression (MNDA-diminished level, 0.7%) and no bimodal distribution. Of note, this patient lacked increased blasts, had normal cytogenetics, and had isolated persistent anemia with ringed sideroblasts and a previous diagnosis of low-grade RA with ringed sideroblasts; because of modest trilineage dysplasia, the study review WHO 2008 diagnosis was RCMD.1 All other cases seemed to be high-grade MDS (RCMD, RAEB, or therapy-related MDS). Excluding the case with ringed sideroblasts, the MDS-diminished range for MDS would be 8.4%–97.5%. A single pediatric case of MDS (case 7) was distinctly abnormal in MNDA expression (MNDA-diminished, 73.8%), similar to other patients with MDS. One control subject (case 15) had a bimodal MNDA distribution with an MNDA-diminished population of 13.7%, overlapping the 2 low MDS cases. This control subject was 83 years old, with no pertinent hematologic medical history and no clinical evidence of MDS; clinical follow-up was limited to 5 months, during which time no hematopoietic abnormalities developed. Whether a case such as this represents incipient MDS, a predisposition to develop MDS, or simply a normal variant remains to be explored. In summary, despite the small number of overlapping cases, the MNDA-diminished data sets show a remarkable divergence between samples from patients with MDS and control subjects.

Figure 2

Distribution and interquartile ranges (A) and receiver operating characteristic (ROC) curve display (B) of myeloid nuclear differentiation antigen (MNDA) results in myeloids for control vs myelodysplastic syndrome (MDS) samples. A, In myeloid progenitors in patients with MDS, the median percentage of MNDA-dim cells was 67.4% (range, 0.7%–97.5%; interquartile range, 44.9%–82.7%). The analogous median percentage of MNDA-dim cells in control patients was 1.2% (range, 0.2%–13.7%; interquartile range, 0.6%–2.7%). B, The area under the ROC curve was 0.96 (P = 9 × 10–7), indicating remarkable sensitivity and specificity for use of MNDA expression to detect MDS, with almost complete discrimination between MDS cases and controls.

MNDA Expression vs WHO MDS Subtype, Marrow Blast Percentage, and IPSS Score

Cases of RCMD tended to lower MNDA-diminished levels (median, 52.2%; range, 8.4%–97.5%, excluding case 20) than cases of RAEB-1 (median, 72.1%; range, 37.5%–72.1%) or RAEB-2 (median, 80.7%; range, 62.7%–83.8%). However, these differences were not significant (P = .21), with small numbers in each group. In patients with MDS, MNDA expression did not correlate with marrow blast percentage (r2 = 0.12; P = .15). Comparison of MNDA-dim levels with patient IPSS scores suggested a trend (P = .07), again with small numbers of patients in each group.


Diagnosis and subclassification of MDS remain difficult, hampered by the absence of definitive testing for these purposes, the standard tools (dysplastic morphologic features, cytopenias, cytogenetic abnormalities, and blast percentage) being neither sufficiently sensitive nor specific to accurately diagnose and subclassify all cases.13 Flow cytometry is rapid, efficient, and commonly available in clinical laboratories. While flow cytometric demonstration of aberrant differentiation and antigen expression has been proposed as an additional tool for diagnosis of MDS,1416 this use requires large panels of antibodies and analytic experience and has not yet entered routine diagnostic practice. Microarray expression analysis and flow cytometric and immunohistochemical analysis have previously demonstrated down-regulation of MNDA message and protein levels in MDS samples.4,5

We investigated flow cytometric quantitation of MNDA antigen expression in myeloid precursors as a potential diagnostic tool for MDS, with very encouraging results, finding marked down-regulation of MNDA expression in myeloid cells in MDS compared with frequency-matched controls, with only minimal overlap in the 2 sets of data. The ROC curve for our results, with near complete separation of MDS and control samples, suggests this single test has the potential to be as effective as the combined set of tests in current use for the diagnosis of MDS.

Our cases of MDS consisted of high-grade, potentially progressive MDS (RCMD, RAEB, therapy-related MDS), with 1 exception, a case of WHO RCMD, based on modest multilineage dysplasia, but with ringed sideroblasts. Excluding this case, most MDS samples had 2 populations of myeloid cells, 1 expressing MNDA normally and 1 with diminished MNDA expression, suggesting there is mixed normal and myelodysplastic hematopoiesis in patients with MDS. This admixture may be one source of difficulty in the diagnosis and subclassification of MDS and, if not taken into account, could compromise investigative studies of MDS samples. This observation also provides a potential means of refining such testing, ie, by restricting analysis to the demonstrably abnormal MNDA-diminished population by using cell sorting or electronic segregation of data.

Our study suggests multiple areas for future investigation. Validation of the diagnostic usefulness of this testing will require analysis of additional MDS samples, normal control samples, and cases constituting the differential diagnosis of MDS. The 1 patient in our MDS set with ringed sideroblasts, uniquely among the MDS set, had MNDA expression identical to that in normal control samples. While requiring validation with analysis of more cases, this observation raises the possibility that MNDA analysis may provide additional information for discriminating subtypes of MDS. Pediatric MDS, although uncommon, has a bad prognosis similar to that of high-grade MDS in adults; of interest, the single pediatric MDS case in our series had bimodal MNDA expression similar to adult patients. The trend to correlation of reduced MNDA expression with IPSS results is interesting but requires validation with a larger set of patients. Although there was increasing median loss of MNDA expression in progressively higher WHO MDS grades (median MNDA-dim levels for RAEB-2 > RAEB-1 > RCMD), these results were not statistically significant, nor, by extension, were results of MNDA loss vs blast percentages (on which this WHO grading is based).

Approximately half of acute myeloid leukemia (AML) cases are related to MDS in pathogenesis.3,17 This subset of AML occurs predominantly in elderly people and has a very poor prognosis without stem cell transplantation. Recognition of this poor-risk AML subset is often difficult because cytogenetics and history are frequently uninformative, leaving imprecise distinction by default based on patient age. We hypothesize that assessment of MNDA expression in background hematopoiesis or possibly in leukemic blasts in AML may facilitate this distinction, again requiring further investigation.

By using quantitative flow cytometry, we demonstrated decreased levels of MNDA antigen expression in marrow samples from patients with MDS, with a typical bimodal pattern of normal and dim MNDA expression. These results diverge markedly from frequency-matched control samples and suggest that such testing may provide a new objective parameter for the diagnosis of MDS.


Upon completion of this activity you will be able to:

  • list 3 types of myelodysplastic syndrome (MDS) that are low-grade and relatively stable in clinical outcomes, compared to controls, and to list at least 2 types that are at high risk for progression to leukemia.

  • describe the biologic origin and role of the protein myeloid nuclear differentiation antigen (MNDA).

  • describe a potential application of use of MNDA in the diagnosis of MDS.

The ASCP is accredited by the Accreditation Council for Continuing Medical Education to provide continuing medical education for physicians. The ASCP designates this educational activity for a maximum of 1 AMA PRA Category 1 Credit™ per article. This activity qualifies as an American Board of Pathology Maintenance of Certification Part II Self-Assessment Module.

The authors of this article and the planning committee members and staff have no relevant financial relationships with commercial interests to disclose.

Questions appear on p 477. Exam is located at www.ascp.org/ajcpcme.


We thank Jean McClure for assistance with figure preparation for this article.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
View Abstract