A Method for the Detection of Multiple Surface Antigens on
Small Heterogeneous Cell Populations
ALAN C. HOMANS, M.D., EDWIN N. FORMAN, M.D., AND BARBARA E. BARKER, PH.D.
A technic using fluorescent immunospheres was developed to
identify simultaneously two surface antigens on individual lymphocytes while preserving Wright's stained cell morphology.
Small numbers (10,000-20,000) of cells were studied from peripheral blood or bone marrow samples from normal control
subjects and from patients with chronic lymphocytic leukemia
(CLL) and acute lymphoblastic leukemia (ALL). Samples were
studied for surface antigens using OKT-11, OK-Ia-1, Bl, and
J5. Comparison was made with results obtained from the same
patients by indirect immunofluorescence. Correlation between
results obtained with immunospheres and indirect immunofluorescence was excellent (r = 0.97). In addition, 35 cerebrospinal
fluid samples from children with ALL were tested using the immunosphere method alone. Results obtained with spinal fluid
lymphocytes agreed with previously reported results using similar
methodology. It is concluded that the use offluorescentmicrospheres provides a method for the combined evaluation of cell
morphology and surface antigens on small, heterogeneous cell
populations. (Key words: Microspheres; Immunofluorescence;
Cerebrospinal fluid) Am J Clin Pathol 1986; 86: 469^474
SEVERAL METHODS have been described using
monoclonal antibodies for the identification of specific
cell surface antigens on homogenous blood mononuclear
cell populations.' s n Difficulties may be encountered,
however, when applying conventional immunologic labeling technics to samples with quite small cell populations or cell populations with mixtures of monocytes and
other cell types.'' Technical limits exist in acquiring sufficient cell numbers for usual immunologic tests in some
pediatric patients, in patients with lymphopenia, or in
conditions such as early leukemic relapse where the cells
of interest may represent only a minority of the mononuclear cells present. In addition, usual immunologic
methods may not be easily adapted for use with other
body fluids such as cerebrospinal fluid (CSF), where small
(1-10 cells/mm 3 , 1-10 X 106 cells/L), cell concentrations
Received October 22, 1985; received revised manuscript and accepted
for publication February 14, 1986.
Supported by Rhode Island Foundation Grant No. 84037, the Julie
E. Trudeau Fund, and the Rhode Island Lions Children's Cancer Fund.
Address reprint requests to Dr. Homans: Department of Pediatric
Hematology/Oncology, Rhode Island Hospital, Providence, Rhode Island
02902.
Departments of Pediatrics and Pathology, Rhode Island
Hospital and Brown University, Providence, Rhode Island
are the rule. Finally, most technics for immunologic labeling of cell surface antigens do not provide detailed simultaneous information about cell morphology unless
additional separation and staining procedures are employed. 414
Several investigators have described the use of microscopic plastic spheres bound with monoclonal antibodies
(immunospheres) as a direct or indirect marker for single
specific cell surface antigens on blood or bone marrow
cells.2-6-8"1015 In addition, we have previously reported on
the use of immunospheres as a marker for indirect labeling
of CSF cells.7 This report describes the use of fluorescent
microscopic polystyrene spheres to identify simultaneously two cell surface antigens and cell morphology of
individual lymphocytes in blood, bone marrow, and CSF.
This technic makes it feasible to study surface antigens
on small (10 3 -10 4 cells), heterogeneous cell populations.
Excellent correlation was found between results obtained
with this technic and with visual indirect immunofluorescence.
Materials and Methods
Patients and Cells
Peripheral blood mononuclear cells were obtained from
normal adult control patients and from peripheral blood
or diagnostic bone marrow specimens from patients with
chronic lymphocytic leukemia (CLL) and from bone
marrow specimens from patients with acute lymphoblastic
leukemia (ALL). Cells were separated by density gradient
centrifugation on Ficoll-Hypaque.3 The final cell concentration was adjusted to 2 X 105 cells/mL (2 X 108 cells/
L) in RPMI 1640 medium with 10% fetal calf serum
(FCS). CSF cells were obtained from pediatric patients
with ALL in bone marrow and central nervous system
(CNS) remission undergoing routine-surveillance lumbar
469
470
HOMANS, FORMAN, AND BARKER
A.J.C.P. • October 1986
•
FIG. 1. A (left). Photomicrograph of microsphere preparation of a bone marrow lymphoblast from a patient with ALL showing Wright's-stained
morphology under normal illumination (XlOO). B (center). Same cell examined under blue light (490 nm wavelength) showingfluorescenceof
attached green microspheres coated with la (X100). C (right). Same cell examined under green light (546 nm wavelength) showingfluorescenceof
attached red microspheres coated with J5 (XlOO).
punctures for intrathecal prophylaxis. CSF samples were
also obtained from five children with lymphoblasts noted
on cytocentrifuged specimens in the absence of peripheral
blood contamination. Aliquots of CSF (1-2 mL) were
centrifuged at 200 g for 10 minutes, and the supernatant
was used for routine chemistry determinations. This investigation was approved by our institutions's Committee
on Protection of Human Subjects.
and washed twice in phosphate-buffered saline (PBS) with
10% heat-inactivated agamma FCS to saturate the sphere's
remaining free protein binding sites. After washing, the
spheres were resuspended in 0.1 mL PSB with 10% inactivated FCS and 0.2% sodium azide and stored at 4 °C
or frozen at -20 °C. Spheres were used within 3 weeks
of preparation. Negative control spheres were prepared
by incubatingfluorescentmicrospheres with agamma FCS
or goat anti-mouse IgG but without monoclonal antibody.
Fluorescent Microsphere Preparation
A microsphere labeling technic was developed, which
is a modification of the method reported by Mirro and
Stass.910 Commercially available, affinity purified, monoclonal antibodies were used in the procedure. OKT-11
(Ortho Pharmaceutical, Raritan NJ) was used to identify
a broad range of T-lymphocytes.18 The antibodies Bl
(Coulter Clone, Hialeah, FL) and the anti-HLA Dr associated antibody OK-Ia-1 (Ortho) were used to identify
differentiation-associated antigens present on subsets of
B-lymphocytes.' The presence of the common acute lymphoblastic leukemia antigen (CALLA) was identified with
J5 (Coulter).13
Monoclonal antibody (10 ^g) was dissolved in 0.1 mL
distilled water and incubated with 10 nL of a 1.4% suspension of 0.7 n diameter fluorescent modified polystyrene
spheres (Covasphere MX, Ann Arbor, MI). Spheres of
two colors were used: green (with maximum emission at
a wavelength of 472 nm) and red (with a maximum emission at a wavelength of 577 nm). Spheres were first sonicated at 4 °C for five seconds to disperse them in suspension. After a 75-minute incubation at room temperature, the spheres were centrifuged at 8,700 g for 4 minutes
Cell Labeling
Aliquots of the peripheral blood mononuclear cell suspension, 0.1 mL (2 X 104 cells) or the CSF cell pellet were
studied. Cells withfirstincubated in a 12 X 75 mm plastic
tube with 0.1 mL of 40% heat inactivated human AB
serum in RPMI 1640 medium to decrease nonspecific
binding of monoclonal antibody to Fc-receptor-bearing
mononuclear cells. After a subsequent wash with medium
containing 10% inactivated FCS, two aliquots of dissimilar
colored antibody-coated microspheres were added to each
cell pellet. Combinations included OKT-11 (green
spheres) with Bl (red spheres) and OK-Ia-1 (green spheres)
with J 5 (red spheres). To assess the possibility of nonspecific binding of microspheres to cells, a negative control
was provided by incubating mononuclear cells from normal adults with the previously described negative control
spheres. For CSF cells, only OKT-11 (green spheres) and
J 5 (red spheres) were used as part of a continuing study
of CSF cell surface markers. The suspension of cells and
immunospheres was thoroughly mixed for 20 to 30 seconds, then centrifuged at 200 g for 10 minutes to bring
the cells and spheres into close proximity. After a one-
DETECTION OF MULTIPLE SURFACE ANTIGENS
Vol. 86 • No. 4
Li.
M
Q
I—I
X
H
I00-,
O
a.
W5
LLI
o
A. *
80-|
Table 1. Results of Immunosphere Testing of CSF
Samples from Children without Evidence
of CNS Leukemia (n = 30)
70
60-
>
«V
90
50-
/
A
Antibodies Tested
40-
• 0KT-1I
30-
o J5
2010
Bl
r = 0.97
471
WBC/mm 3
(Neubauer
Chamber)*
%CALLA
Positive
Cells
% OK.T 11
Positive
Cells
1.3
0.64
0.97
(0-4)
72.9
9.4
(50-89)
Mean
S.D.
Range
(0-5)
• Multiply by 10J for cells/L.
CSF = cerebrospinal fluid; CNS = central nervous system; WBC»white blood cells;
CALLA = common acute leukemia antigen.
1
1 1 1 1 1 1 1 1 1
10 20 30 40 50 60 70 80 90 100
% CELLS POSITIVE WITH IMMUNOSPHERES
FlG. 2. Correlation between results obtained with fluorescent microspheres and indirect immunofluorescence (IDIF). Each point represents
data for a specific surface antigen from a patient measured for antigen
positivity with each method. N = 37 samples, 11 patients. Correlation
coefficient of r = .97 for pooled data.
hour incubation at 4 °C, the cells and spheres were resuspended with a glass pasteur pipette and underlayered
with 2 mL of FCS. The suspension of spheres and cells
bound with spheres was then centrifuged through the FCS
at 200 g for 10 minutes, and all but 0.2 mL of the cellfree supernatant was discarded along with the excess, unbound spheres suspended on top of the FCS. The remaining cells were resuspended, placed in a cytocentrifuge
cup (Cytospin® II, Shandon Elliot, Sewickley PA), and
centrifuged at 750 g for 5 minutes onto glass slides. The
slide preparations were subsequently Wright's-stained and
examined under oil immersion at X100 using fluorescent
microscopy equipment (Olympus, model no. BH2-RFL).
Slides were examined under white light to characterize
cell morphology, under blue light (wavelength 490 nm)
to examine cells bound with green spheres, and under
green light (wavelength 546 nm) to examine cells bound
with red spheres. Two hundred lymphocytes per slide were
counted for each peripheral blood sample, and all lymphocytes (range, 40-200) were counted for each CSF
sample. As has previously been reported,2,8'9 cells binding
three or more spheres of the same color were considered
positive for the antigen identified by the antibody bound
to that color of sphere. Cells were not counted if clusters
of spheres were bound to the cell at a single point or if
the cells appeared greatly damaged or nonviable. In practice, usually many more than three spheres were bound
to positive cells (Fig. 1). Monocytes and granulocytes can
be recognized by cell morphology and by their ability to
phagocytize the antibody-bound spheres. Because atypical
monocytes may occasionally be difficult to differentiate
from lymphoblasts on CSF cytocentrifuge preparations,
the visible phagocytosis of microspheres is useful, and we
preferred the use of viable, nonfixed cells. Results are expressed as the percentage of antigen positive lymphocytes
of the total number of lymphocytes counted. The immunosphere assay was performed without knowledge of
results obtained by indirect immunofluorescence testing.
Indirect Immunofluorescence Testing
Aliquots of peripheral blood mononuclear cells obtained from the same patients described earlier were also
tested for surface antigens using visual indirect immunofluorescence (IDIF). Mononuclear cells were isolated
on Ficoll-Hypaque, washed three times in PBS, and adjusted to a final concentration of 20 X 106 cells/mL (20
X 109 cells/L) in PBS. Aliquots of cells (0.5 X 106 cells)
were incubated with 5 fiL of murine monoclonal antibody
(Ortho or Coulter, as described earlier) for 30 minutes at
4 °C. After washing in PBS with 10% FCS, the cells were
Table 2. Results of Surface Antigen Testing of CSF
Lymphocytes from Children with Cytologic
Evidence of CNS Leukemia (n = 5)
Patient
no.
1
2
3
4
5t
WBC/mm 3
(Neubauer
Chamber)*
% CALLA
Positive
Lymphoid
Cells
% OKT11
Positive
Lymphoid
Cells
% Blasts
Seen on
Cytocentrifuge
0
297
40
1
15
13
99
58
10
0
67
1
23
57
69
15
100
89
10
2
* Multiply by 10J for cells/L.
t Patient with prior history of both CNS leukemia and viral CNS infection.
CSF = cerebrospinal fluid; CNS = central nervous system; WBC - white blood cells;
CALLA - common acute leukemia antigen.
472
HOMANS, FORMAN, AND BARKER
A.J.C.P. -October 1986
FIG. 3. A (upper). Photomicrograph of Wright's-stained microsphere
preparation of 2 CSF cells from patient 18 showing a lymphocyte (left)
and a lymphoblast (right) (XI00). B (center). Same cells as A viewed
under blue light (490 nm wavelength) showingfluorescenceof OKT-11
coated green spheres attached to lymphocyte only (X100). C (lower).
Same cells viewed under green light (546 nm wavelength) showing fluorescence of J5 coated red spheres attached to lymphoblast only (X100).
•^B^
•
subsequently incubated with fluoresceinated goat antimouse IgG (Tago Corp., Burlingame, CA). After a 30minute incubation at 4 °C, the cells were again washed
and resuspended in buffered glycerol. Coverslip preparations were examined forfluorescenceat X100 under oil
immersion on a Zeiss® microscope. Results are expressed
as the percent antigen-positive cells of 200 lymphocytes
counted. Correlation was then determined between fluorescent microsphere assay results and results obtained
with the conventional indirect immunofluorescence. Due
to the small number of cells present in CSF specimens,
they were assayed with the immunosphere method only.
Results
Figure 1 (a-c) includes photomicrographs of an individual bone marrow cell from a patient with ALL. These
photographs illustrate the ability of this method to identify
two cell surface antigens (la and CALLA in this case) on
an individual Wright's-stained cell. Peripheral blood or
bone marrow samples from 11 patients were studied.
Samples were obtained from three normal control subjects, four patients with CLL, and four patients with ALL.
By studying samples from patients with ALL and CLL,
and from persons without disease, cells were obtained with
a broad range of expression of T- and B-cell antigens.
With the immunosphere method, values obtained for
OKT-11 positivity ranged from 0-84%, for OK-Ia positivity 60-100%, for Bl positivity 5-95%, and for CALLA
positivity 0-100%. Ten trials were performed with negative control spheres. A mean of 0.45% positivity (range,
0-1%) was noted on negative control slides, indicating an
extremely low background binding of cells with microspheres. No negative control cell was bound with greater
than three spheres.
Figure 2 illustrates the correlation between results obtained with fluorescent microsphere assay and conventional indirect immunofluorescence. The correlation coefficient for the two sets of results (obtained with IDIF and
microspheres) for all antibodies tested was 97%. Correlation for the two methods for individual antigens are as
follows: OK-Ia (n = 7) r = 0.94, Bl (n = 8) r = 0.97,
CALLA (n = 10) r = 0.99, and OKT-11 (n = 13) r
= 0.99. The slope of the line described by the pooled data,
1.06, calculated by the method of least squares, verifies
DETECTION OF MULTIPLE SURFACE ANTIGENS
Vol. 86 • No. 4
that the two methods yield quite similar results. Minor
differences between results obtained with the two labeling
methods were not clinically significant.
Results of surface antigen testing for CSF cells in children without CNS leukemia are shown in Table 1. Of the
CSF lymphocytes from children with ALL in CNS remission, 73% (range, 50-89%; SD 9.4%) were positive for
OKT-11, whereas only 0.6% (range, 0-3%; SD 0.97%) of
the cells were positive for CALLA. Results for CALLA
positivity of remission CSF did not differ significantly from
that of the negative controls. These results agree with our
previously reported results obtained with the use of a related methodology using nonfluorescent immunospheres
as indirect markers for single cell surface antigens.7 For
patients 1-4, who had evidence of CNS leukemia on cytocentrifuge, CALLA was detected with J5 antibody on
13-99% of the CSF lymphoid cells, including most of the
CSF blast forms (Table 2). Positive OKT-11 cells were
found on 1-67% of the cells in these three patients and
accounted for most of the small lymphocytes with normal
morphology found in these samples. The decreased binding of OKT-11-coated microspheres indicates specific
binding of CALLA and provides an internal control
against the possibility of nonspecific binding of microspheres by lymphoblasts. Figure 3 shows photomicrographs of CSF cells from patient 18, illustrating the ability
of this method to discriminate individual positive OKT11 lymphocytes in mixed CSF cell populations with
CALLA positive lymphoblasts. Patient 5 had a prior history both of CNS leukemia and a viral CNS infection.
The CSF of this patient showed 2% lymphoblasts in the
presence of other reactive mononuclear cells, yet normal
numbers of positive OKT-11 cells and no positive CALLA
CSF cells were noted. The CSF results in this patient illustrate the usefulness of this methodology in differentiating infectious from neoplastic causes of CSF pleocytosis.
Discussion
Synthetic microspheres may be bound with a variety
of antibodies and used for immunologic characterization
of cell populations visually, by electron microscopy, or
with the aid of flow cytometry.2,91015,16 Immunologic
characterization of surface antigens by flow cytometry has
the definite advantage of being able rapidly to study large
numbers of cells; however, it requires specialized equipment and trained personnel. There is only a single report
of the use of flow cytometry (examining aneuploidy) in
the study of the small heterogeneous cell populations
found in routine CSF samples.12 Only limited information
about cell morphology is available from flow cytometry
473
unless additional cell separation procedures (with loss of
cell yield) and subsequent cell staining are carried out.4,1719
In addition, cells expressing autofluorescence may render
results obtained with flow cytometry less accurate.14 Immunofluorescence testing, either with visual methods or
with flow cytometry, does not allow the simultaneous examination of detailed morphology of those individual cells
expressing a specified antigen. The study of CSF cell immunophenotypes is confounded by all of these factors
because quite small, heterogeneous cell populations are
routinely present. Because of the low concentration of
cells (about 103/mL [106 cells/L]) in CSF and other body
fluids, the separation of mononuclear cells on Ficoll-Hypaque is not feasible.
The technic described in this report has the advantage
of allowing the simultaneous study of two cell surface
antigens with conventional Wright's-stained morphology
of individual cells. This permits the immunologic characterization of quite small cell populations (2 X 104 cells
or less) and allows the determination of surface antigen
profiles of morphologically identified lymphocytes and
lymphoblasts without prior mononuclear cell separation.
Investigators have also shown that other cytochemical
stains may be performed on cells previously labeled with
immunospheres.8 If additional procedures are performed,
fluorescent microspheres may be used to provide quantitative information about the degree of cell surface antigen expression.6
The results obtained with the methods outlined earlier
correlated well with the results obtained with conventional
indirect immunofluorescence methods. Our results with
spinal fluid cells demonstrate that this technic is well
adapted for the study of individual lymphoblasts in body
fluids other than blood or bone marrow where small, heterogeneous cell populations are present. The study of
larger numbers of CSF samples will be required to confirm
the clinical usefulness of this method, but we have found
microsphere labeling technics useful for confirming CNS
leukemia in the absence of CSF pleocytosis, diagnosing
primary CNS lymphoma, and differentiating infectious
from neoplastic causes of meningitis (as was noted in patient 5). We are currently using this method in a larger
longitudinal study of the CSF cell populations of pediatric
patients with ALL.
Acknowledgments. The authors acknowledge the technical assistance
of Dr. E. Mazur, J. White, A. Schultz, M. O'Connell, C. Lefoley, and P.
Lawrence, and thank Drs. R. Trueworthy and J. Truman for CSF samples
from patients in cases 2, 3, and 5.
References
1. Anderson KC, Bates MP, Slaughenhoupt BL, Pinkus GS, Schlossman
SF, Nadler LM: Expression of human B cell-associated antigens
474
HOMANS, FORMAN, AND BARKER
on leukemias and lymphomas: A model of human B cell differentiation. Blood 1984; 63:1424-1433
2. Bernard J, Ternynck T, Zagury D: A general method for the cytochemical and ultrastructural studies of human lymphocyte subsets
defined by monoclonal antibodies. Immunol Let 1982; 4:65-73
3. Boyum A: Isolation of mononuclear cells and granulocytes from
human blood. Scand J Clin Lab Invest 1968; 97(suppl):77-81
4. Fishleder AJ, Tubbs RR, Savage RA, et al: Immunophenotypic
characterization of acute leukemia by immunocytology. Am J
Clin Pathol 1984;81:611-617
5. Foon KA, SchroffRW, Gale RP. Surface markers on leukemia and
lymphoma cells: recent advances. Blood 1982; 60:1-19
6. Higgins TJ, O'Neill HC, Parish CR: A sensitive and quantitative
fluorescence assay for cell surface antigens. J Immunol Methods
1981;47:275-287
7. Homans AC, Forman EN, Barker BE: Use of monoclonal antibodies
to identify cerebrospinalfluidlymphoblasts in children with acute
lymphoblastic leukemia. Blood 1985; 66:1321-1325.
8. Sjoberg O, Inganas M: Detection of Fc receptor-bearing lymphocytes
by using IgG-coated latex particles. Scand J Immunol 1979; 9:
547-552
9. Mirro J, Stass SA: Fluorescent microsphere detection of surface antigens and simultaneous cytochemistries in individual hematopoietic cells. Am J Clin Pathol 1985; 83:7-11
10. Mirro J, Schwartz JF, Civin CI: Simultaneous analysis of cell surface
antigens and cell morphology using monoclonal antibodies con-
11.
12.
13.
14.
15.
16.
17.
18.
19.
A.J.C.P. • October 1986
jugated tofluorescentmicrospheres. J Immunol Methods 1981;
47:39-48
Molday RS, Dreyer WJ, Rembaum A, Yen SPS: New immunolatex
spheres: Visual markers of antigens on lymphocytes for scanning
electron microscopy. 1975; 64:75-88
Prued'homme JL, Gugliemi P, Labaume S: Lymphocyte markers
in human leukemias and lymphomas: methodologic remarks.
Semin Hematol 1984; 21:296-301
Redner A, AndreefFM, Miller D, Steinherz P, Melamed M: Recognition of central nervous system leukemia by flow cytometry.
Cytometry 1984; 614-618
Ritz J, Pesando JM, Notis-McConarty J, Lazarus H, Schlossman S:
A Monoclonal antibody to human acute lymphoblastic leukaemia
antigen. Nature 1980; 283:583-585
Shapiro HM: Multistation multiparameter flow cytometry: A critical
review and rationale. Cytometry 1983; 3:227-243
Steinkamp JA, Wilson JS, Saunders GC, Stewart CC: Phagocytosis:
Flow cytometric quantitation with fluorescent microspheres. Science 1982;215:64-66
Traganos F: Flow cytometry: Principles and applications. Cancer
Invest 1984;2:149-163
Van Wauwe J, Goossens J, Decock W, Kung P, Goldstein GL:
Suppression of human T-cell mitogenesis and E-rosette formation
by the monoclonal antibody OKT 11A. Immunology 1981; 44:
865-871
Wheeless LL, Patten SF: Slit-scan cytofluorometry. Acta Cytol 1973;
17:333-339