Cytoplasmic Immunoglobulin Content in Multiple Myeloma
Bart Barlogie, Raymond Alexanian, Mark Pershouse, Leslie Smallwood, and Lon Smith
The University of Texas System Cancer Center, M. D. Anderson Hospital and Tumor Institute,
Texas Medical Center, Houston, Texas 77030
Bone marrow cells of 82 patients with multiple myeloma were
subjected to flow cytometric analysis of DNA and cytoplasmic
immunoglobulin (CIg) content using propidium iodide and
with nonsecretory myeloma, there was conformity in the im-
munoglobulin type derived from immunoelectrophoresis and
plasma cell CIg staining. One patient with nonsecretory my-
eloma exhibited monotypic CIg staining, while the second
showed no reaction. In eight patients with IgG lambda myeloma,
the same tumor cells contained both lambda and kappa light
chains, suggesting the productive rearrangement of both light
chain genes. 14 patients with previously unrecognized plasma
cells of low RNA content, all of whom were resistant to
chemotherapy, were identified by CIg staining. By revealing
previously unrecognized plasma cells with low RNA content,
CIg analysis identified more patients with treatment-refractory
assays. Except for two patients
The plasma cells of patients with multiple myeloma usually
contain an abnormal DNA and an increased RNA content,
permitting the quantitation of plasma cells by flow cytometry
(1-3). Higher response rates to both initial and salvage therapy
have been noted with increasing RNA content (4, 5), and a
shorter survival time has been found in patients with marked
bone marrow plasmacytosis (4). A typical feature of plasma
cells is the presence of cytoplasmic immunoglobulin (CIg).'
We have developed a flow cytometric assay for the concurrent
analysis of DNA and CIg content in order to quantify plasma
cells when studies ofDNA and RNA content could not provide
a clear distinction from normal cells. Whether CIg staining
might clarify the biology of plasma cells in patients with
nonsecretory disease or with other atypical clinical features
was also evaluated.
Studies were performed in 82 patients with advanced stages ofmultiple
myeloma including 23 without prior therapy. The bone marrow of all
Address reprint requests to Dr. Barlogie, Department of Hematology,
6723 Bertner-Box 55, Houston, TX 77030.
Receivedfor publication 28 November 1984 and in revisedform S
1. Abbreviations used in this paper: CIg, cytoplasmic immunoglobulin;
CV, coefficient of variation.
patients contained at least 7% plasma cells on routine morphological
examination. For flow cytometry studies, bone marrow aspirates were
separated by Hypaque-Ficoll gradient centrifugation (6). Interphase
cells were collected and washed once in phosphate-buffered saline. One
aliquot of cells was always stained with the metachromatic dye acridine
orange for concomitant DNA and RNA analysis using a mercury arc
flow cytometer (Phywe Co., Gottingen, Germany) (7). Routinely,
10,000 cells were analyzed and tumor cells gated on the basis of
abnormal DNA and RNA content (1, 4). In this manner, tumor cell
DNA and RNA index values were derived from the relation to normal
lymphocytes (specifically, the ratio of median fluorescence channel
numbers of tumor to normal diploid cells) (1). The proportion of cells
with abnormal DNA and/or RNA content in the entire sample was
A second bone marrow aliquot was processed for biparametric
DNA-CIg analysis with an EPICS V flow sorter (Coulter Electronics,
Inc., Hialeah, FL). Cells were fixed in 70% ice-cold ethanol for at least
24 h. Single cell suspensions were then exposed separately to anti-
light-chain and anti-heavy-chain reagents (fluorescein-conjugated F(ab')2
fragments; Capell Laboratories, Westchester, PA) at dilutions of 1:200
(0.1 mg/ml) and counterstained for DNA with propidium iodide (8).
In three instances of dual light chain reaction, different antisera were
evaluated in the same manner (Tago Inc., Burlingame, CA).
Successive flow cytometric analyses were performed on 10,000
cells, each stained for both light chains, and, in a subset of patients,
stained also for two heavy chains (typically, IgG and IgA). Except for
eight instances of dual light chain reaction, marked differences in
staining intensity were observed between pairs of samples stained for
light and heavy chains (see Fig.
DNA-abnormal stemlines, bright CIg fluorescence was usually limited
to aneuploid cells, while residual diploid cells showed the same dim
fluorescence observed after staining with the opposite light chain
reagent. The proportion of tumor cells was readily determined by
applying an electronic gating procedure to the brightly stained cells
with an abnormal DNA content (Fig. 1). To obtain a measure of CIg
content for each patient, a CIg index was computed from the ratio of
median CIg fluorescence intensities of aneuploid and diploidGIcells
of the same sample.
For the remaining DNA-diploid samples, both the percentage of
plasma cells and their CIg index were determined in comparison with
the nonspecific staining pattern of cells reacted with the opposite light
chain antiserum. Specifically, kappa and lambda distribution curves
were superimposed electronically to identify the lower level of specific
CIg fluorescence. The CIg index was then computed from the ratio of
median intensities of specific and nonspecific (opposite light chain)
fluorescence. The degree ofheterogeneity ofCIg content among plasma
cells from the same patient was expressed as a coefficient of variation
(CV) ofthe monotypic CIg distribution. The CV (percent) was computed
from the ratio of the standard deviation and the mean of CIg
fluorescence channel numbers times 100.
There were eight patients with dual cytoplasmic light chain expres-
sion. In the six aneuploid cases, all DNA-abnormal cells demonstrated
both kappa and lambda staining. This finding and fluorescence micro-
scopic evaluation (after double-staining with fluorescein- and rhodamine-
conjugated antisera) indicated that the same plasma cells reacted with
both light chain antisera. As in the majority of cases with monotypic
CIg staining, the CIg index was computed in relation to dimly stained
normal diploid cells. In the two cases of diploid myeloma with dual
light chain staining, assessment of the degree of marrow plasmacytosis
1). In the "80% of patients with
Cytoplasmic Immunoglobulin Content in Multiple Myeloma
J. Clin. Invest.
© The American Society for Clinical Investigation, Inc.
Volume 76, August 1985, 765-769
and of the CIg index was performed in comparison with normal bone
marrow samples separately stained with kappa and lambda light chain
DNA-CIg analysis of bone marrow cells from a patient with
IgG kappa myeloma. A bright staining reaction was observed
in hyperdiploid cells using anti-gamma and anti-kappa sera,
but not with anti-alpha and anti-lambda reagents. A second
example illustrates a patient with concurrent multiple myeloma
and acute myelogenous leukemia (Fig. 2). Compared with
normal marrow, two discrete DNA stemlines (diploid and
hyperdiploid) with markedly increased RNA content were
identified. Using light chain antisera, only the hyperdiploid
DNA stemline contained monoclonal CIg of kappa type, and
hence, represented the myeloma tumor population; the diploid
DNA stemline showed similar weak staining reactions with
both kappa and lambda reagents and represented the leukemic
immunoelectrophoretic results. Concordance of light chain
phenotypes was observed in 80 of 82 patients, including 18
patients on whom additional heavy chain analyses were con-
ducted. Two patients lacked a monoclonal protein by immu-
noelectrophoresis, but showed marrow involvement micro-
scopically (25 and 42%) and by DNA-RNA flow cytometry
(18 and 69%). One patient had kappa staining within the
cytoplasm ("low secretor"), while the second showed no de-
tectable anti-light-chain reaction ("low producer"). A third
1 shows a typical example of bivariate flow cytometric
I summarizes the relationship between CIg and
Figure. 1. Myeloma phenotyping by CIg flow cytometry. Bone mar-
row cells were stained with anti-heavy and anti-light chain reagents
and counterstained for DNA with propidium iodide. Note the posi-
tive reaction of hyperdiploidGI cells (window 1) only with anti-
gamma and anti-kappa and not with anti-alpha and anti-lambda
antisera, confirming the presence of IgG kappa monoclonal protein.
The CIg index as a measure of plasma cell CIg content was com-
puted in relation to nonspecific fluorescence (window 2).
Normal Bone Marrow
AML and Myoloma
Zs 72n'o64444 1S.S66046
Figure 2. Comparison ofDNA-RNA and DNA-CIg analysis in a
patient with concurrent myeloma and acute myeloid leukemia
(AML). Upper panels represent DNA-RNA histograms; lower panels
show DNA-CIg histograms. Compared with normal bone marrow
with a prevalence of diploid cells with low RNA content (upper left),
both diploid and hyperdiploid DNA stemlines with markedly in-
creased RNA content were found in a patient with both myeloma
and AML (upper right). DNA-CIg analyses with anti-lambda and
anti-kappa reagents revealed monoclonal kappa expression in the
hyperdiploid DNA stemline, whereas diploid cells showed nonspecific
staining. Hence, the hyperdiploid clone represented Myeloma, and
the diploid cells with high RNA content represented AML cells (see
reference 7). Compared with the example in Fig. 1, nonspecific
fluorescence (lower left) is shifted toward higher channel numbers due
to instrument setting. The CIg index of specific kappa fluorescence is
relatively low (lower right).
Table I. Immunoglobulin Typing by Immunoelectrophoresis
and CIg Analysis using Flow Cytometry
§ Subset of 18 patients in whom heavy chain analyses were also con-
Patient with lambda light chain in cytoplasm and urine; however,
gamma heavy chain was only detected in cytoplasm and not in
Barlogie, Alexanian, Pershouse, Smallwood, and Smith
patient with a positive reaction of cells for lambda light chain
and with Bence Jones proteinuria also exhibited cytoplasmic
IgG in the absence of a monoclonal serum peak.
Fig. 3 illustrates the correlations between the proportions
of tumor cells identified by DNA-RNA or DNA-CIg flow
cytometry and those enumerated by morphology for the 77
patients in whom all three assays were available. Both cyto-
metric assays showed similar degrees of plasmacytosis in the
48 patients with DNA aneuploidy, where abnormal cells were
< 0.001) (Fig. 3 A). Among the 29 patients with diploid
myeloma, there were 8 patients with a DNA-RNA pattern
indistinguishable from normal marrow but a monoclonal light
chain reaction was evident in 5-69% of cells (median, 26%).
These cells corresponded to plasma cells on microscopic
examination (Fig. 3 B). One patient with nonsecretory myeloma
showed 60-70% plasma cells with high RNA content but
lacked a monotypic CIg pattern ("low-producing myeloma").
There were also 6 patients with aneuploid myeloma, in whom
a second diploid DNA stemline with 5-20% plasma cells was
uncovered by monotypic CIg staining (four kappa and two
lambda) (Fig. 4). ThUs, the combined analysis of DNA-CIg
and DNA-RNA permitted the identification of previously
unrecognized cell populations with low RNA content in -20%
of our patients.
Eight patients with a discrete DNA-RNA abnormality had
plasma cells which expressed both kappa and lambda light
lambda in all instances. An example of such double-staining
is illustrated in Fig. 5 with three different DNA stemlines in
the diploid, the low degree hyperdiploid, and the tetraploid
(R = 0.83,
range, all of which expressed kappa, lambda, and IgG immu-
noglobulin (the latter not shown). Fluorescence microscopy
revealed concurrent kappa and lambda reactions in the same
plasma cells. The dual light chain reaction was confirmed in
three patients using a different source of anti-light-chain
There was marked variation in the CIg staining per plasma
cell in individual patients, which was expressed as a CV
ranging from 30 to 80% (median, 50%). To compare the CIg
content per cell among different patients, a CIg index was
defined from the ratio of the median fluorescence intensities
of specific vs. nonspecific anti-light-chain reactions. Among
23 patients studied at diagnosis, the CIg index ranged from 2
to 27, with a median value of 10. There was no correlation
between the RNA and the CIg index, but the CIg index
decreased with increasing marrow tumor infiltrate (R = -0.25,
P < 0.01).
We also examined the clinical implications ofCIg analysis.
Among the 23 previously untreated patients, neither CIg index
nor CIg dispersion (CV) correlated with response to chemo-
therapy. By CIg analysis, we had identified 14 patients with a
low RNA index (less than four times the RNA content of
normal lymphocytes) in diploid plasma cells, either as their
sole tumor cell population (8 patients) or in addition to an
aneuploid DNA stemline (6 patients). Five individuals in
each category were evaluated after 1-3 mo of resistance to
initial chemotherapy (primary unresponsive disease); the re-
maining patients were studied at diagnosis, and none responded
to standard chemotherapy. Thus, CIg analysis unmasked diploid
plasma cells with low RNA content, which was associated with
complete resistance to preceeding or subsequent chemotherapy.
* ~*: A-
% Abnormal Cells (DNA-GIg)
Figure 3. Correlations between the proportion of tumor cells in the
bone marrow identified by flow cytometry of DNA-RNA and DNA-
CIg as well as by microscopy. (A) Close concordance between DNA-
RNA and DNA-GIg cytometric assays in aneuploid myeloma (v).
Among the 29 diploid cases (., o), there were 8 with undetectable
plasma cells on DNA-RNA histograms (A) but with monotypic CIg
staining (o), consistent with monoclonal plasma cells (B). *, o, o;
% Abnormal Cells (DNA-Cg)
Diploid, n =29, r =0.40, P < 0.01. &; Aneuploid,n =48,r = 0.83,
P < 0.01. (B) Correlation between marrow plasmacytosis by micros-
copy and by DNA-CIg cytometry. One patientlacked amonotypic
CIg reaction, but had a high RNA content withtypical plasmacell
morphology (o). o,
Aneuploid, n =48, r =0.62, P < 0.01.
., o; Diploid, n =29,r =0.75,P < 0.01. A,
Cytoplasmic Immunoglobulin Content in Multiple Myeloma
32 _A "'
Figure 4. Biclonal myeloma DNA stemlines identified by DNA-CIg
flow cytometry. (A) DNA-RNA analysis demonstrates only one cell
population with increased RNA and DNA content, as typically found
in myeloma (4). (B) Monoclonal light chain reaction with anti-kappa
serum in bob diploid and hyperdiploid DNA stemlines. (C) Nonspe-
cific staining with anti-lambda serum.
Plasma cells represent the last phase of B lymphocyte differ-
entiation and are readily identified by the presence ofmonotypic
immunoglobulin in the cytoplasm (CIg) and, more recently,
by monoclonal antibodies reacting specifically with the surface
membrane of plasma cells (9, 10). While flow cytometry has
been used to probe the lymphocyte differentiation pathway,
this method has not been applied to assess the biologic and
clinical relevance of CIg in patients with multiple myeloma
(1 1). Using a direct immunofluorescence technique, there was
excellent agreement between immunoelectrophoresis and CIg
flow cytometry in immunoglobulin phenotyping. In conjunction
with morphology and DNA-RNA flow cytometry, DNA-CIg
analysis helped define special categories of"low-secretory"and
There were eight instances of concurrent kappa and lambda
reaction in the same aneuploid cells. This observation must
be interpreted with caution, as all eight patients produced
lambda light chains. The dual staining reaction, however, was
confirmed in three patients using a different reagent; in one of
those examined further, we only eliminated the lambda, but
not the kappa staining of the doubly stained cells by preincu-
bation with free lambda light chains. Thus, our findings seem
consistent with those rare reports of coexpression of kappa
Figure 5. Dual light chain expression in myeloma. (A) DNA-RNA
analysis reveals three discrete populations with diploid (1), low degree
hyperdiploid (2), and near tetraploid DNA stemlines (3), with typi-
cally increased RNA content in populations (2) and (3). (B), (C)
DNA-CIg analysis demonstrates positive reactions in all three DNA
stemlines with both kappa (B) and lambda (C) reagents. (D), (E) On
microscopic inspection of doubly stained cells (kappa-fluorescein iso-
thiocyanate, D; lambda-rhodamine isothiocyanate; E), the same
plasma cells react with both light chain reagents.
Barlogie, Alexanian, Pershouse, Smallwood, and Smith
and lambda light chains by the same tumor cells (12-14). Download full-text
Interestingly the dual light chain expression was only observed
in lambda-secreting myeloma. This fits the concept of a
hierarchy of light chain activation, where kappa gene rear-
rangement precedes that of lambda (15, 16). Among 18 B cell
samples examined from patients with chronic lymphocytic
leukemia and long term Epstein-Barr virus-transformed normal
human B cell lines, lambda constant region genes remained
in the germ line configuration in all eight samples producing
kappa light chains (16). In contrast, 10 lymphocyte samples
producing lambda light chains showed deletion of both kappa
constant region alleles in 9 instances and productive rearrange-
ment of 1 kappa allele in 1 B cell line. The latter case indicates
the possibility of productive rearrangement of both light chain
genes in lambda-secreting cells. Our CIg flow cytometric data
suggest that this event may occur in as many as 10% of
patients with myeloma.
There was close conformity in the degree of bone marrow
involvement by tumor cells using morphologic and flow cy-
tometric assays. The latter have the advantage of objectivity
and reproducibility. DNA-CIg analysis identified some patients
with apparently normal bone marrow on DNA-RNA studies,
whose plasma cells were now recognized with a very low RNA
content. This assay also revealed diploid plasma cells with a
low RNA content in addition to aneuploid cells previously
demonstrated by DNA-RNA analysis. Both of these observa-
tions are of major importance because of the established
resistance to chemotherapy of myeloma patients either with a
low RNA index or multiple DNA stemlines (4, 5, 17).
As we had noted with RNA content, the CIg content per
plasma cell varied markedly within and among patients. In
previously untreated patients, increasing bone marrow plas-
macytosis was associated with a decrease in CIg index and
RNA index (18), which might reflect less differentiated and
perhaps more aggressive features of myeloma. The resistance
to chemotherapy of low RNA index myeloma has been
established (4, 5, 17) and may be related in part to a greater
degree of RNA heterogeneity with increasing tumor burden
(18, 19). Similar inferences cannot yet be made for CIg index
and CIg dispersion in view of the small number of patients
evaluated at diagnosis.
In summary, DNA-CIg flow cytometry aids in several
aspects of myeloma research and clinical management. We
were able to recognize patients with "low-secretory" and "low-
producing" disease. The biological significance ofCIg dispersion
and the coexpression by the same tumor cell of both light
chains requires further study. The importance of objective and
quantitative assessment of the degree of marrow tumor infil-
tration has been demonstrated (4, 17). DNA-CIg analysis is
superior to DNA-RNA cytometry for assessing marrow plas-
macytosis because monoclonal CIg is more specific for abnormal
plasma cells. In conjunction with DNA-RNA flow cytometry,
CIg analysis permitted the identification of more patients with
a poor prognosis, whose plasma cells either had a low RNA
content, biclonal DNA abnormalities, or both.
This work was supported by National Cancer InstitutegrantsCA16672,
CA28771, and CA37 161.
1. Barlogie, B., J. Latreille, D. Swartzendruber, L. Smallwood, A.
Maddox, A. M. Raber, B. Drewinko, and R. Alexanian. 1982. Quan-
titative cytology in myeloma research. In Clinics in Haematology.
W. R. Schmidt, editor. W. B. Saunders Co. 19-45.
2. Bunn, P., S. Krasnow, R. Makuch, M. Schlam, and G. Schechter.
1982. Flow cytometric analysis of DNA content of bone marrow cells
in patients with plasma cell myeloma. Clinical implications. Blood.
3. Barlogie, B., M. Raber, J. Schumann, T. Johnson, B. Drewinko,
D. Swartzendruber, W. Gohde, M. Andreeff, and E. Freireich. 1983.
Perspectives in cancer research: flow cytometry in clinical cancer
research. Cancer Res. 43:3982-3997.
4. Barlogie, B., R. Alexanian, E. Gehan, L. Smallwood, T. Smith,
and B. Drewinko. 1983. Marrow cytometry and prognosis in myeloma.
J. Clin. Invest. 72:853-861.
B., L. Smith, and R. Alexanian.
treatment of advanced multiple myeloma refractory to alkylating
agents. N. Engl. J. Med. 310:1353-1356.
6. Boyum, A. 1968. Isolation of mononuclear cells and granulocytes
from human blood. Scand. J. Clin. Lab. Invest. 21:77-89.
7. Barlogie, B., J. Latreille, E. J. Freireich, C. T. Fu, D. Mellard,
M. Meistrich, and M. Andreeff. 1980. Characterization of hematologic
malignancies. Blood Cells. 6:719-744.
8. Crissman, H. A., P. F. Mullaney, and J. A. Steinkamp. 1975.
Methods and applications of flow systems for analysis and sorting of
mammalian cells. Methods Cell Biol. 9:179-246.
9. Anderson, K. C., E. K. Park, M. P. Bates, R. C. F. Leonard, R.
Hardy, S. F. Schlossman, and L. M. Nadler. 1983. Antigens on human
plasma cells identified by monoclonal antibodies. J. Immunol. 130:
10. Ruiz-Arguelles, G. J., J. A. Katzmann, P. R. Greipp, N. J.
Gonchoroff, J. P. Garton, and R. A. Kyle. 1984. Multiple myeloma:
circulating lymphocytes that express plasma cell antigens. Blood. 64:
11. Zeile, G. 1980. Intracytoplasmic immunofluorescence in multiple
myeloma. Cytometry. 1:37-41.
12. Bouvet, J. P., D. Buffe, R. Oriol, and P. Liacopoulos. 1974.
Two myeloma globulins, IgGl-K and IgGI-a, from a single patient
(Im). Immunology. 27:1095-1 101.
13. Hopper, J. E. 1977. Comparative studies on monotypic IgM
and IgG from an individual patient. IV. Immunofluorescent evidence
for a common clonal synthesis. Blood. 50:203-211.
14. Choi, Y., and M. Wong. 1981. Double light chain production
by leukemic cells of common clonal origin: a case report with a review
of pertinent literature. Am. J. Hematol. 11:93-98.
15. Korsmeyer, S. J., P. A. Hieter, J. V. Ravetch, D. G. Poplack,
T. A. Waldmann, and P. Leder. 1981. Developmental hierarchy of
immunoglobulin gene rearrangements in human leukemic pre-B-cells.
Proc. Natl. Acad. Sci. USA. 78:7096-7100.
16. Hieter, P. A., S. J. Korsmeyer, T. A. Waldmann, and P. Leder.
1981. Human immunoglobulin kappa light-chain genes are deleted or
rearranged in lambda-producing B cells. Nature (Lond.). 290:368-372.
17. Barlogie, B., R. Alexanian, L. Smith, D. Dixon, L. Smallwood,
and K. Delasalle. 1985. Prognostic implications of tumor cell DNA
and RNA content in multiple myeloma. Blood. In press.
18. Barlogie, B., and R. Alexanian. Cellular aspects of myeloma:
biologic and clinical implications. In Multiple Myeloma and Other
lishers. In press.
19. Goldie, J., and A. Coldman. 1979. A mathematical model for
relating the drug sensitivities of tumors to their spontaneous mutation
rate. Cancer Treat Rep. 63:1717-1733.
I. W. Delamore, editor. Churchill Livingstone Pub-
Cytoplasmic Immunoglobulin Content in Multiple Myeloma