• multiple myeloma;
  • CD56;
  • plasmablast;
  • flow cytometry;
  • prognosis


  1. Top of page
  2. Abstract
  3. Patients and methods
  4. Results
  5. Discussion
  6. References

Summary. We analysed CD56 expression in 70 patients with multiple myeloma (MM) to determine its clinicopathological and prognostic significance. Fifty-five (79%) patients were CD56+. CD56 patients (n = 15) had higher β2 microglobulin levels and a higher incidence of extramedullary disease, Bence Jones protein, renal insufficiency and thrombocytopenia than CD56+ patients. Their myelomas more frequently had a plasmablastic morphology. Overall survival was significantly lower in CD56 than CD56+ patients (22 vs 63 months, P = 0·0002). We conclude that CD56 MM is a discrete entity associated with more aggressive disease. The higher incidence of plasmablastic cases suggested that CD56 MM may develop from a less mature plasma cell than CD56+ MM.

CD56, a neural cell adhesion molecule, was first reported by Van Camp et al (1990) to be frequently expressed on myeloma cells. In addition, these authors reported that patients with CD56 multiple myeloma (MM) predominantly had aggressive disease. A lack of CD56 expression also has been associated with a higher frequency of extramedullary involvement (Pellat-Deceunynck et al, 1998), whereas, using cryopreserved samples from MM patients, Mathew et al (1995) reported that lack of CD56 expression was not related with a poor prognosis factor. However immunostaining of cryopreserved samples is sometimes difficult to evaluate. Therefore, we analysed the CD56 expression on fresh bone marrow samples using three-colour flow cytometry and determined the clinicopathological and prognostic significance of this molecule.

Patients and methods

  1. Top of page
  2. Abstract
  3. Patients and methods
  4. Results
  5. Discussion
  6. References

Patients.  The expression of CD56 on myeloma cells was studied in the bone marrow (BM) of 70 patients with MM (56 cases at diagnosis and 14 cases in relapse). The clinical and laboratory characteristics were obtained from the patients at diagnosis (Table I). The diagnostic criteria for MM were those of the South-west Oncology Group (Durie, 1986). Patients with monoclonal gammopathy of undetermined significance (MGUS) and smoldering MM were excluded. The median age of the patients was 68 years (range 41–91), with 34 men and 36 women. The monoclonal component was: IgG in 41% of cases, IgA in 31%, IgD in 1% and Bence Jones protein (BJP) in 21% of patients. In three cases, no monoclonal protein was found. The ratio of κ/λ was 1·2. The patients were distributed according to Durie–Salmon stage (Durie & Salmon, 1975) as follows: stage IA, 10%; IIA, 26%; IIB, 3%; IIIA, 44%; and IIIB, 17%. Sixty-four of these patients were treated with either melphalan–prednisone (n = 21) or alternating cycles of VMMD (vincristine 1·2 mg/m2 i.v. d 1, ranimustine 40 mg/m2 i.v. d 1, dexamethasone 40 mg/body i.v. d 1–4, 9–12, 17–20 and melphalan 8 mg/m2 p.o. d 1–6) with (n = 24) or without (n = 19), interferon alpha (IFN-α).

Table I.  CD56 status and clinico-biological disease characteristics.
CharacteristicCD56+ MM (n = 55)CD56 MM (n = 15)P-value
  • *

    P-value was calculated using the Mann–Whitney test and the chi-squared test.

Age > 65 years (median)61% (66 ± 9)80% (72 ± 10)0·1774
Sex (male)26/55 (47%)8/15 (53%)0·6794
 I–II22/55 (40%)5/15 (33%)0·6406
 III33/55 (60%)10/15 (67%) 
Presence of bone disease82%60%0·0766
Presence of extramedullary disease7%47%0·0002
Haemoglobin < 8·5 g/dl36% (9·5 ± 2·1)60% (9·2 ± 2·1)0·1020
WBC > 5·0 × 109/l43% (5·4 ± 2·3)53% (6·7 ± 4·6)0·4627
Platelets < 150 × 109/l7% (22·5 ± 8·8)47% (18·3 ± 9·4)0·0002
LDH > 400 IU/l27% (355 ± 101)33% (375 ± 130)0·6475
Calcium > 2·76 mmol/l49% (10·4 ± 1·2)43% (10·1 ± 1·5)0·6790
CRP > 10 mg/l18% (0·8 ± 2·1)13% (0·3 ± 0·5)0·6610
β2-microglobulin > 60 mg/l35% (5·2 ± 2·9)70% (7·2 ± 3·7)0·0420
Creatinine > 176·8 µmol/l11% (1·2 ± 0·6)40% (2·8 ± 2·6)0·0085
Proteinuria > 1 g/24 h39% (1·0 ± 1·5)50% (1·8 ± 2·0)0·5976
Presence of urine light chains59%100%0·0009
Monoclonal protein  0·0177*
 Bence Jones15%47% 
Treatment  0·2285*
PB morphology7/55 (13%)10/15 (67%)< 0·0001
CR ratio10/46 (22%)5/13 (38%)0·2254

Flow cytometry.  Heparinized bone marrow was obtained from patients for the analysis of CD phenotypes. Cells were incubated with fluorescein isothiocyanate (FITC)-labelled anti-CD38 or R-phycoerythrin (PE)-labelled anti-CD19 (Coulter Immunotech, Miami, FL, USA) and anti-CD56 (Pharmingen, San Diego, CA, USA) monoclonal antibody. These cells were incubated at 4°C for 30 min and then washed three times with phosphate-buffered saline (PBS) containing bovine serum albumin (BSA). After washing, 20 000 events were counted using a flow cytometer (FACS Calibur, Becton Dickinson, San Jose, CA, USA). Cells were gated by their forward- and side-scatter characteristics and strong CD38 expression (CD38++). The amount of antigen expression was measured indirectly using the mean fluorescence intensity of bound antibody as reported (Rawstron et al, 1999). An antigen was defined as positive when the mean fluorescence intensity was 10 times greater than that of the CD3-negative control.

Morphological classification.  BM smears from the patients were prepared with Wright–Giemsa stain. Myeloma cells was classified as plasmablastic (PB) type or non-PB (including mature, intermediate and immature) type according to the criteria of Greipp et al (1985). To avoid bias or dominant influence by one investigator, two reviewers sorted 200 cells independently, without knowledge of the patient's course, and the myeloma subtype was defined as plasmablastic (2% or more plasmablasts among the myeloma cell population) only if both reviewers agreed.

Statistical analysis.  The relationship between CD56 expression and prognosis was analysed in 70 MM patients. Overall survival (OS) was measured from the diagnosis of MM up to death. OS was determined according to the Kaplan–Meier method. Statistical analysis was carried out using the chi-squared test, Mann–Whitney U-test and the log rank test using the statview 4·5 J program (Abacus Concepts, Berkeley, CA, USA). The influence of each variable on survival was analysed first univariately by the Cox proportional hazards model. Statistically significant factors (P < 0·05) were further tested in the multivariate analysis with stepwise forward selection.


  1. Top of page
  2. Abstract
  3. Patients and methods
  4. Results
  5. Discussion
  6. References

Of 70 patients evaluated for CD56 expression, 55 (79%) were positive and 15 (21%) were negative. Only two cases (both CD56) were positive for CD19. The clinical features of the CD56+ and CD56 groups are summarized in Table I. No significant differences were found between the two groups in terms of age, sex, bone lesion extension or Durie–Salmon stage. In addition, several prognostic indicators such as the serum level of calcium, lactate dehydrogenase (LDH), C-reactive protein (CRP), haemoglobin level, white blood cell (WBC) counts and 24 h excretion of proteinuria were similar in both groups of patients, whereas the serum β2 microglobulin level was significantly higher and the platelet counts were significantly lower in the CD56 group. In addition, CD56+ cases had a significantly higher prevalence of BJP-type myeloma, BJP in urine, renal insufficiency and extramedullary involvement than the CD56 group. The CD56 cases had a plasmablastic morphology more frequently than the CD56+ cases.

The median survival of the 70 patients was 59 months (range 4–122 months). Fifteen of 59 (25%) patients, who received chemotherapy and were evaluated for more than 6 months, achieved complete remission (CR), defined by disappearance of myeloma cells and M-protein. The median duration of CR was 24 months. The CR rate was similar in both groups. The CD56 group had a significantly shorter median survival time than the CD56+ group (22 vs 63 months, P = 0·0002) (Fig 1).


Figure 1. Overall survival of CD56 positive and negative patients. Median survival was 63 vs 22 months respectively (P = 0·0002).

Download figure to PowerPoint

Six of 15 analysed variables, which included β2 microglobulin level, haemoglobin level, platelet counts, presence of extramedullary disease, plasmablastic morphology and CD56 expression, were found to be significantly associated with survival in univariate analysis. In multivariate analysis, haemoglobin level (P = 0·03) and presence of extramedullary disease (P = < 0·0001) were found to have independent prognostic importance for survival.


  1. Top of page
  2. Abstract
  3. Patients and methods
  4. Results
  5. Discussion
  6. References

Van Camp et al (1990) demonstrated that eight of 12 cases of CD56 MM had progressive or refractory disease. Pellat-Deceunynck et al (1998) reported that CD56 MM presented a leukaemic phase more frequently than CD56+ MM. Rawstron et al (1999) reported that the expression of CD56 showed an inverse correlation with the degree of marrow infiltration and peripheral blood plasma cell numbers. However, in these studies, clinicopathological information and the prognostic significance of CD56 expression were not reported. Mathew et al (1995) analysed cryopreserved samples from 68 patients and reported that lack of CD56 expression did not appear to be a poor prognostic factor. However, our data, acquired by cytometric analysis, show that the OS of CD56 MM was significantly shorter than that of CD56+ MM. This discrepancy may be explained by differences in the methods used for analysis. While Mathew et al (1995) analysed cryopreserved samples, we used fresh bone marrow cells for immunophenotypic study. We used the sequential gating strategy proposed by Harada et al (1993) which could clearly distinguish myeloma cells from other haematopoietic cells and detected CD56 expression on myeloma cells even in the BM samples containing few myeloma cells. Thus, our methodology may have resulted in increased sensitivity of detection of CD56 expression on myeloma cells.

Van Camp et al (1990) reported a higher prevalence of BJP myeloma cases in the CD56 group than in the CD56+ group. We found, in addition, a higher incidence of urine BJP and renal insufficiency in the CD56 group; all 15 patients with CD56 MM were positive for BJP in urine.

We demonstrated that myeloma cells in the CD56 group were more likely to have a PB morphology according to Greipp's classification (Greipp et al, 1985). Patients with PB features also tended to have more advanced and aggressive disease (Greipp et al, 1985, 1998). In our study, OS in PB was significantly shorter than that of non-PB patients (19 vs 63 months, P = 0·0001). Greipp et al (1998) hypothesized that the aggressive behaviour of plasmablasts in PB MM may, in part, reflect an increased interleukin 6 responsiveness associated with mutation of ras, which should be analysed in CD56 MM.

We could not find any prognostic values of CD56 expression in the multivariable Cox analysis. This may be due to the strong correlation between CD56 expression and the presence of extramedullary disease (P = 0·0002, R = −0·444, by Spearman rank correlation) or the limited number of cases analysed. The results of the currently ongoing, prospective randomized study by the Japan Adult Leukaemia Study Group may provide further information on this.


  1. Top of page
  2. Abstract
  3. Patients and methods
  4. Results
  5. Discussion
  6. References
  • Durie, B.G.M. (1986) Staging and kinetics of multiple myeloma. Seminars in Oncology, 13, 300309.
  • Durie, B.G.M. & Salmon, S.E. (1975) A clinical staging system for multiple myeloma. Correlation of measured myeloma cell mass with presenting clinical features, response to treatment, and survival. Cancer, 36, 842854.
  • Greipp, P.R., Raymond, N.M., Kyle, R.A. & O'Fallon, W.M. (1985) Multiple myeloma: significance of plasmablastic subtype in morphological classification. Blood, 65, 305310.
  • Greipp, P.R., Leong, T., Bennet, J.M., Gaillard, J.P., Klein, B., Stewart, J.A., Oken, M.M., Kay, N.E., Van Ness, B. & Kyle, R.A. (1998) Plasmablastic morphology – an independent prognostic factor with clinical and laboratory correlates: eastern cooperative oncology group (ECOG) myeloma trial E9486 report by the ECOG myeloma laboratory group. Blood, 91, 25012507.
  • Harada, H., Kawano, M.M., Huang, N., Harada, Y., Iwato, K., Tanabe, O., Tanaka, H., Sakai, A., Asaoku, H. & Kuramoto, A. (1993) Phenotypic difference of normal plasma cells from mature myeloma cells. Blood, 81, 26582663.
  • Mathew, P., Ahmann, G.J., Witzig, T.E., Roche, P.C., Kyle, R.A. & Greipp, P.R. (1995) Clinicopathological correlates of CD56 expression in multiple myeloma: a unique entity? British Journal of Haematology, 90, 459461.
  • Pellat-Deceunynck, C., Barille, S., Jego, G., Puthier, D., Robillard, N., Pineau, D., Rapp, M.-J., Harrousseau, J.-L., Amiot, M. & Bataille, R. (1998) The absence of CD56 (NCAM) on malignant plasma cells is a hallmark of plasma cell leukemia and of a special subset of multiple myeloma. Leukemia, 12, 19771982.
  • Rawstron, A., Barrans, S., Blythe, D., Davies, F., English, A., Pratt, G., Child, A., Morgan, G. & Jack, A. (1999) Distribution of myeloma plasma cells in peripheral blood and bone marrow correlates with CD56 expression. British Journal of Haematology, 104, 138143.
  • Van Camp, B., Durie, B.G., Spier, C., De Waele, M., Van Riet, I., Vela, E., Frutiger, Y., Richter, L. & Grogan, T.M. (1990) Plasma cells in multiple myeloma express a natural killer cell-associated antigen: Cd56 (NKH-1; Leu-19). Blood, 76, 377382.