OUP user menu

Flow Cytometric Immunophenotyping and Minimal Residual Disease Analysis in Multiple Myeloma

Ritu Gupta MD, DNB, Archana Bhaskar MSc, Lalit Kumar MD, Atul Sharma DM, Paresh Jain MD
DOI: http://dx.doi.org/10.1309/AJCP1GYI7EHQYUYK 728-732 First published online: 1 November 2009


Presence of normal plasma cells (PCs), hemodilution of bone marrow aspirate, and changes in the immunophenotype are important considerations in minimal residual disease (MRD) assessment in multiple myeloma (MM). We evaluated 124 subjects—107 with MM, 11 with Hodgkin lymphoma, and 6 allogeneic stem cell transplantation donors—for the immunophenotype of neoplastic, reactive, and normal PCs respectively. Of the patients with MM, 36 were evaluated for MRD and 23 for a change in immunophenotype after chemotherapy. The immunophenotype of normal and reactive PCs was similar and differed from that of neoplastic PCs with respect to CD19, CD45, CD56, CD52, CD20, and CD117. At least 2 antigens were aberrantly expressed in all cases and 3 in 90.7% of MM cases. A change in the immunoprofile of PCs was observed in 18 (78%) of 23 cases. By using flow cytometry, we detected MRD in all samples, and a neoplastic PC index (percentage of neoplastic PCs/total bone marrow PCs) of less than 30 could differentiate immunofixation (IFx)− from IFx+ samples (complete and partial responders, respectively).

Key Words:
  • Multi-parametric flow cytometry
  • Immunophenotype
  • Minimal residual disease
  • Multiple myeloma

High-dose chemotherapy followed by autologous stem cell transplantation and newer chemotherapeutic agents such as thalidomide, lenalidomide, and bortezomib have considerably improved the response rates in multiple myeloma (MM).13 Sensitive and robust assays are, therefore, required to quantify minimal residual disease (MRD) after therapy to monitor the effects of these novel treatment strategies and to identify patients who are at risk of disease progression and may require additional treatment. Although allele-specific oligonucleotide real-time polymerase chain reaction is more sensitive than multiparametric flow cytometry (MFC) for MRD analysis, it is more labor-intensive and time-consuming and less readily available. The patchy nature of residual plasma cell (PC) infiltrate after therapy and hemodilution of bone marrow aspirates may undermine the quantification of PCs with respect to overall bone marrow cellularity. This may result in false-negative MRD flow cytometric results when compared with immunofixation (IFx) or allele-specific oligonucleotide real-time polymerase chain reaction.

Previous studies have demonstrated the use of CD38, CD138, and/or CD45 for identifying PCs in a milieu of hematopoietic cells.4,5 Furthermore, CD19 and CD56 have been used to differentiate between neoplastic and normal PCs.68 It has been suggested that additional markers may help in making this distinction in larger number of patients.4 The use of such additional markers for MRD analysis would depend on their specificity for neoplastic PCs, frequency of aberrant expression, and stability over time. CD20, CD52, and CD117 have been variably studied in patients with MM for identifying patients who may benefit from antigen-directed therapy.912

In this study, we compared the various gating strategies for identification of PCs and evaluated the frequency of aberrant antigen expression of CD19, CD56, CD45, CD52, CD20, and CD117 in 107 treated and untreated patients with MM. We repeated immunophenotypic evaluation in 23 patients to determine the stability of aberrant antigen expression after 4 to 6 months of non–antigen-directed therapy. Furthermore, in patients with 5% or fewer PCs on morphologic examination of bone marrow aspirates, we correlated serum and urine protein electrophoresis and IFx findings with flow cytometric MRD quantification based on aberrant PC immunophenotypes. Furthermore, we describe the neoplastic PC Index (neoplastic PCs as a percentage of total PCs) as a novel statistic that may be used to reliably differentiate complete responders (IFx–) from partial responders (IFx+).

Materials and Methods


A total of 124 patients, including 107 with MM and 11 with Hodgkin lymphoma (reactive plasmacytosis), and 6 allogeneic stem cell transplantation (ASCT) donors were evaluated in this study. Of the 107 patients with MM, 67 were evaluated at the time of diagnosis and 40 had been partially treated. In all cases, the bone marrow aspirate was obtained after obtaining informed consent from participants per the guidelines of the ethical committee of the institute.

Immunophenotyping Studies

Immunophenotyping studies were carried out on the bone marrow aspirates using pretitrated volumes of the following monoclonal antibodies: CD19 fluorescein isothiocyanate (FITC), CD20 FITC, CD45 FITC, CD117 phycoerythrin (PE), CD56 PE, CD38 peridinin chlorophyll protein–cyanine 5.5, CD138 allophycocyanin (BD Biosciences, San Jose, CA), and CD52 PE, κ FITC, and λ PE (Serotec, Kidlington, England). The panel used was as follows: (1) CD19/CD56/CD38/CD138; (2) CD45/CD52/CD38/CD138; (3) CD20/CD117/CD38/CD138, and (4) κ/λ/CD38/CD138. Staining was done using standard whole blood lysis technique. For assessing surface antigens, an aliquot of cells containing 1 × 106 cells was labeled with pretitrated volumes of preconjugated monoclonal antibodies, and to study cytoplasmic κ and λ light chain expression, fixation and permeabilization before staining were carried out per the manufacturer’s recommendation (Serotec). The cells were then washed with phosphatebuffered saline and suspended in 1% paraformaldehyde. Acquisition was done on a flow cytometer (BD FACSCalibur and BD FACSCanto, BD Biosciences) equipped with facility for at least 4-color immunophenotyping, and at least 105 events were acquired in each tube. Analyses were carried out using FCS Express V3 (De Novo Software, Los Angeles, CA). Negative limits were set using autofluorescence alone.

We compared various gating strategies using 1 antibody (CD38 or CD138), 2 antibodies (CD38 and CD138; CD38 and CD45; or CD138 and CD45), or all 3 antibodies (CD38, CD138, and CD45) for distinguishing PCs from other hematopoietic cells Image 1 . We found that CD38 and CD138 could reliably detect PCs in 98.1% of the cases, so these were used for gating at analysis Table 1 .

The immunophenotypes of PCs in patients with MM, patients with reactive plasmacytosis, and ASCT donors were compared to identify immunologic criteria that may help in differentiating reactive from neoplastic plasmacytosis. An aberrant immunophenotype was defined as expression of antigens not normally expressed or lack of expression of antigens normally expressed by the PCs of ASCT donors. In a patient with MM, expression of an antigen was considered positive when at least 10% of the PCs expressed it at the time of diagnosis.

To evaluate the clinical usefulness of the immunoprofile of PCs for the detection of MRD, immunophenotyping studies were done in 36 patients with MM receiving chemotherapy with 5% or fewer PCs in bone marrow imprints or aspirate smears and adequate bone marrow aspirate. Adequacy of the bone marrow aspirate for MRD evaluation by flow cytometry was determined by documenting the presence of myeloid and erythroid precursors on morphologic evaluation of Jenner-Giemsa–stained smears prepared from the same sample that was submitted for flow cytometry. A PC was defined as neoplastic when it displayed at least 2 aberrant antigens. A complete response in MM was defined as 5% or fewer PCs in the bone marrow and a negative IFx test result; IFx results were compared with the results of immunophenotyping studies in all patients when evaluating for MRD.13

To evaluate if the immunophenotype of PCs changes after therapy, a repeated evaluation was carried out in 23 patients with MM after 4 to 6 months of chemotherapy and compared with the baseline immunophenotypes of PCs seen in the diagnostic samples. Up-regulation of an antigen was defined as expression of an antigen not expressed in the diagnostic sample or a 3-fold or greater increase in PCs with expression of a particular antigen. Similarly, down-regulation of an antigen was defined as loss of expression of an antigen expressed in the diagnostic sample or a 3-fold or greater decrease in PCs with expression of a particular antigen.


The immunophenotype of PCs in ASCT donors was similar to the immunophenotype in reactive plasmacytosis, ie, CD19+, CD45+, CD56–, CD52–, CD20–, CD117–, and polyclonal with respect to κ and λ light chains (κ/λ ratio ranging from 1.1 to 2.1). The PCs were thus assigned to have an aberrant phenotype when they expressed either CD56 or CD52 or CD20 or CD117 or lacked expression of CD19 or CD45. The antigens most commonly found to be aberrantly expressed in patients with MM, in decreasing order of frequency, were CD19 (96.3%), CD45 (84.1%), CD56 (82.2%), CD117 (50.5%), CD52 (32.7%), and CD20 (10.3%) Table 2 . Expression of at least 2 aberrant antigens was seen in all 67 cases of untreated MM, at least 3 aberrant antigens were seen in 61 cases (91%), and more than 3 aberrant antigens were seen in 45 cases (67%).

Image 1

Gating strategies for the identification and enumeration of plasma cells. The boxed events represent gated plasma cells. APC, allophycocyanin; Cy5.5, cyanine 5.5; FITC, fluorescein isothiocyanate; PerCP, peridinin chlorophyll protein.

View this table:
Table 1
View this table:
Table 2

Of 36 patients with MM receiving chemotherapy who were evaluated for residual disease, neoplastic PCs were identified in all and ranged from 0.01% to 5.23% of all acquired events. Immunofixation studies revealed an M band in 21 (58%) of 36 samples, and the neoplastic PCs ranged from 0.08% to 5.23% of all acquired events in these samples. In the 15 IFx− samples (42%), neoplastic PCs were identifiable in all and ranged from 0.01% to 0.12% of all the acquired events. The neoplastic PCs were then enumerated as a percentage of total PCs, ie, the neoplastic PC index that ranged from 1.7 to 26.7 in IFx− and 36.1 to 99.4 in IFx+ samples. Thus, a neoplastic PC index value of less than 30 could reliably discriminate between IFx− and IFx+ cases (complete and partial responders, respectively) Figure 1 .

A change in the immunophenotype of PCs was seen in 18 (78%) of 23 patients after therapy, with changes in the expression of CD56, CD117, CD52, and CD20 in 9 (39%), 10 (44%), 11 (48%), and 1 (4%) of 23 patients, respectively. Stable expression of CD56, CD117, CD52, and CD20 was seen in 14 (61%), 13 (57%), 12 (52%), and 22 (96%) of 23 patients, respectively Table 3 . Simultaneous up-regulation of 2 or more antigens was seen in 4 cases, 2 with up-regulation of CD56 and CD117, 1 with up-regulation of CD52 and CD117, and 1 with up-regulation of CD52, CD56, and CD117. Simultaneous down-regulation of 2 or more antigens was seen in 2 cases, 1 with down-regulation of CD52 and CD56 and 1 with down-regulation of CD52, CD56, and CD117; the former case also expressed CD117 and the latter case CD20, which was useful in identifying neoplastic PCs in these cases.


PCs are characterized by a CD38+++, CD138+, CD45–/low phenotype and have been identified using variable gating strategies. The combined use of CD38, CD138, and CD45 in a single tube is ideal for gating PCs.4 Since neoplastic PCs may express dim to bright CD45, an initial gate based on CD45 expression may result in their exclusion from analysis. In our study using 4-color flow cytometry, a combination of CD38 and CD138 was most useful compared with other combinations in delineating PCs in the majority of samples (98.1%; Table 1).

After chemotherapy or stem cell transplantation, monoclonal neoplastic PCs are admixed with normal polyclonal PCs and can be identified by light chain restriction alone only when the monoclonal cells constitute at least 30% of the total cells.4 Since at least 2 markers are required for gating PCs, it is not possible to combine the markers for light chain restriction and aberrant phenotype in a single tube in 4-color flow cytometry assays. In 6-color immunophenotyping assays, all PCs with aberrant phenotypes have been shown to exhibit monoclonality.14,15 Thus, in an MFC-based MRD assay for MM when the total number of fluorescent detectors is limited to 4, assessment based on detection of PCs with aberrant phenotypes rather than light chain restriction may be preferable for enumerating low numbers of neoplastic PCs in a background of normal polyclonal PCs.

Several studies have shown that residual disease above a level of 0.01% is clinically relevant in MM.7,16,17 The total tumor cell burden may, however, appear spuriously low in a hemodiluted bone marrow aspirate and is one of the major confounding factors in residual disease monitoring in MM. When we analyzed the neoplastic PCs as a percentage of all acquired events, we could not find a correlation between residual disease levels and protein electrophoresis results in our samples. Assuming that in a hemodiluted bone marrow aspirate the neoplastic and the normal PCs would be proportionately reduced and normal PCs would outnumber the neoplastic clone after successful therapy, we analyzed the neoplastic PCs as percentage of total PCs, ie, the neoplastic PC index, and found that a cutoff value of 30 could differentiate IFx+ from IFx− samples (partial vs complete responders, respectively; Figure 1). The presence and persistence of normal PCs after therapy has been known to be predictive of disease progression, and patients with 30% or more normal PCs have been reported to have significantly longer progression-free survival compared with patients who had fewer than 30% normal PCs.7,18 Determination of the neoplastic PC index using MFC provides a rapid means to differentiate partial responders from complete responders.

Figure 1

Neoplastic plasma cells as a percentage of total bone marrow plasma cells, ie, neoplastic plasma cell index in immunofixation (IFx)+ (n = 21) and IFx− (n = 15) samples. The mean ± SD (range) values were as follows: IFx+, 67.5 ± 17.5 (36.1–99.4); IFx–, 9.3 ± 7.0 (1.7–26.7).

View this table:
Table 3

Information on changes in the immunophenotype of neoplastic cells after therapy is important when such markers are used as targets for MFC-based MRD detection or antigen-directed therapy.19 We observed a change in the immunophenotype of PCs during treatment in the majority of patients (78%) when both up-regulation and down-regulation of antigen expression were taken into account. The down-regulation of aberrantly expressed antigens may influence MRD detection, particularly when more than 1 antigen is simultaneously down-regulated. Since there is a frequent change in the immunophenotype of neoplastic PCs (including up-regulation and down-regulation) in MM, MRD evaluation using a wide panel of antibodies would be desirable, and the use of only the markers that were aberrantly expressed at diagnosis may not be useful.

Currently, antigen-targeted therapy is being investigated in MM.912 Alemtuzumab (anti-CD52) and imatinib mesylate (inhibitor of c-kit) have been shown to have antitumor activity in in vivo experiments in mice and in vitro experiments, respectively.10,11 Response to anti-CD20 (rituximab) has been observed in a few patients.9,12 We observed up-regulation of CD52, CD117, and CD20 in a few patients during therapy (Table 3). It may, therefore, be worthwhile to explore the clinical usefulness of antigen-targeted therapy initially in patients who show up-regulation of a particular antigen during conventional chemotherapy.

Aberrant antigen expression at diagnosis and changes in the immunophenotype after therapy are frequent in MM, and, thus, a wide panel of monoclonal antibodies needs to be used for MRD detection by MFC. Residual disease monitoring based on the neoplastic PC index as determined by MFC can differentiate complete responders from partial responders, but its clinical usefulness in stratification of patients with myeloma with different risks of disease progression needs to be evaluated in future prospective studies.


We gratefully acknowledge support of Subrata Sinha, MD, PhD, and Neera Nath, PhD, Department of Biochemistry, who made available the flow cytometric laboratory facility to carry out this work. We are indebted to Mrinali Hakim for help and excellent technical assistance.


  • * Dr Jain is currently a Scientific Advisor with BD Biosciences, India.

  • Supported by grant SR/FT/L-29/2005 from the Department of Science & Technology, Government of India (to Dr Gupta) under the SERC-Fast track scheme.


  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.
  18. 18.
  19. 19.
View Abstract