Cytometry (Communications in Clinical Cytometry) 42:371–378 (2000)

Flow Cytometric Method for Enumeration and

Classification of Reactive Immature
Granulocyte Populations
Hiroyuki Fujimoto,1* Takashi Sakata,1 Yukio Hamaguchi,1 Shuichi Shiga,2 Kaoru Tohyama,2
Satoshi Ichiyama,2 Fu-sheng Wang,3 and Berend Houwen3
1
Reagent Group, Sysmex Corporation, Kobe, Japan
2
Department of Central Clinical Laboratory, Kyoto University Hospital, Kyoto, Japan
3
Department of Pathology and Laboratory Medicine, Loma Linda University School of Medicine, Loma Linda, California

We developed a flow cytometric method for the enumeration and classification of nonmalignant immature
granulocytes (IG). In this study, IG are defined as most immature (IG stage 1: promyelocytes and myelocytes)
and as more mature (IG stage 2: metamyelocytes). Blood specimens from 46 patients with documented
infectious or inflammatory disease and known presence of IG (by routine manual microscopy) were analyzed.
For a reference manual differential count, we used a 400 white blood cell (WBC) differential and separated
granulocytes into promyelocytes and myelocytes combined, metamyelocytes, and included band cells in the
mature, segmented neutrophil population. The flow cytometric method is based on three-color staining of
whole, anticoagulated blood with CD45-PerCP, CD16-FITC, and CD11b-PE–labeled monoclonal antibodies
and a three-step gating procedure. The flow cytometric results were confirmed by cell sorting and
microscopic evaluation of the sorted cells. A total of 10,000 events, excluding debris, were recorded per
specimen and IG stage 1 (CD16-/CD11b-), IG stage 2 (CD16-/CD11bⴙ), and mature neutrophils (CD16ⴙ/
CD11bⴙ) were categorized. Regression and correlation between flow cytometric IG and the manual differ-
ential showed y ⴝ 1.34x ⴙ 0.95, r2 ⴝ 0.86 for IG stages 1 and 2 combined versus promyelocytes,
myelocytes, and metamyelocytes. For IG stage 1 versus microscopic counts of promyelocytes and myelo-
cytes, the results were y ⴝ 1.53x ⴙ 1.24, r2 ⴝ 0.76; for IG stage 2 versus manual metamyelocyte count,
y ⴝ 0.77x ⴙ 0.21, r2 ⴝ 0.58. Reproducibility of the flow cytometric method showed a coefficient of variation
(CV) of 6.8% for all IG combined compared with a CV of 50.2% for manual differential IG count (based on
a routine 100 WBC count). Samples were found stable at least 12 h at 25°C and at least 48 h at 4°C for flow
cytometry. After staining and lysing, the sample was stable for at least 120 min at room temperature. We
analyzed samples from patients with myelodysplastic and myeloproliferative disease separately. We found
that CD16- mature neutrophils falsely elevated the flow cytometric IG count. Similar results were obtained
in blood from patients treated with granulocyte-colony stimulating factor (G-CSF). Although this restricts the
use of the method somewhat, we believe that this flow cytometric method is useful for enumerating reactive
IG, as well as for evaluating automated methods for IG identification by hematology analyzers. Cytometry
(Comm. Clin. Cytometry) 42:371–378, 2000. © 2000 Wiley-Liss, Inc.

Key terms: immature granulocytes; flow cytometric method; enumeration; classification

Immature granulocytes (IG) can be found in the bone
marrow. With the exception of newborns, they are rarely
observed in the peripheral blood of normal, healthy indi-
viduals. Therefore, the presence of IG in a patient blood
sample indicates increased myeloid cell production,
which can be the result of infection (especially of bacte-
rial origin) or severe inflammatory disease. IG are also
frequently seen in myelodysplastic syndromes (MDS), my-
eloproliferative disease such as chronic myeloid leukemia
(CML), myelofibrosis, and in metastatic bone marrow in-
filtration by a malignancy. Accordingly, enumerating and
classifying IG can be helpful for detecting and sometimes
monitoring these diseases.
IG enumeration is conventionally done by means of
blood film morphology. A stained blood film is examined
by microscopy and the IG are classified and counted as
promyelocytes, myelocytes, and metamyelocytes. How-
*Correspondence to: Dr. Hiroyuki Fujimoto, Reagent Group, Sysmex
Corporation, 4-4-4 Takatsukadai, Nishi-ku, Kobe 651-2271, Japan.
E-mail: hiroyuki_fujimoto@sysmex.co.jp
Received 9 June 2000; Accepted 2 October 2000

© 2000 Wiley-Liss, Inc.

372 FUJIMOTO ET AL.
ever, this manual method is imprecise due to the small
number of cells counted, typically not more than 100
white blood cells (WBC) total; it is also labor intensive and
time-consuming. On the other hand, flow cytometric
methods have inherent advantages over manual micros-
copy. For example, a large number of cells can be ana-
lyzed and several parameters (i.e., forward scatter [FSC],
side scatter [SSC], and fluorescent labels) rather than mor-
phological appearance alone can be used.
The purpose of this study was to develop a flow cyto-
metric method for the classification and enumeration of
reactive IG. We used three-color staining by CD45, CD16,
and CD11b monoclonal antibodies. CD45, also called leu-
kocyte common antigen, is a transmembrane glycoprotein
that is present in high concentrations on virtually all
nucleated human blood cells in different isoforms. It has
an Mr ranging from 180 to 222 kDa (1). We could separate
WBC (CD45⫹cells) effectively from the debris produced
by the lysing procedure.
Granulocytes, including mature neutrophils, eosino-
phils, and IG, have higher SSC intensity because of their
granular cytoplasmic content (2) and lower expression of
CD45 than other leukocytes. Therefore, granulocytes
could be further identified from lymphocytes, monocytes,
and basophils by SSC and CD45 expression.
CD16 antigen is a low-affinity receptor for IgG
(Fc␥RIII). In neutrophil populations, it is anchored in the
cell membrane through glycosyl-phospatidylinositol
(GPI). Fc␥RIII has an Mr ranging from 50 to 65 kDa (3,4).
CD11b antigen (Mac-1␣) is the ␣-subunit of complement
receptor 3 (Mac-1; CD11b/CD18 heterodimer). Mac-1␣ is
a transmembrane glycoprotein molecule with an Mr of 17
kDa (5).
Granulocytes show differences in CD16 and CD11b
expression: both appear during maturation for promyelo-
cytes to neutrophils, but not at a synchronous rate. Thus,
for most IG, the promyelocytes are negative for both; for
most mature granulocytes, the neutrophils carry both
markers (6,7). Eosinophils also show negative expression
of CD16 (8) but can be excluded from the IG population
based on their very high level of CD45 expression (9).
Because CD11b is an earlier marker of IG maturation
than CD16, three types of neutrophilic cells including IG
can be expected: CD11b-/CD16-, CD11b⫹/CD16-, and
CD11b⫹/CD16⫹. Although this sequence in which CD11b
and CD16 appear in IG is well established for reactive IG,
this is not always the case for IG that are part of a
neoplastic or myeloproliferative process. In certain malig-
nant diseases (i.e., CML [10], MDS [11,12]) as well as in
individuals treated with granulocyte-colony stimulating
factor (G-CSF; 13), a low level or absence of CD16 expres-
sion is found on mature neutrophils. Therefore, when
using these markers, it may not be possible to enumerate
IG accurately in such circumstances. In this study, we
focused our analysis mainly on blood specimens from
patients in whom IG in the peripheral blood were the
result of an infectious or inflammatory process. We refer
to such IG as “reactive.” We performed a limited analysis
of IG in samples from patients with neoplastic disease and
from patients treated with G-CSF. We compared this flow
cytometric method with manual IG counts and evaluated
its reproducibility and the effects of sample storage.
MATERIALS AND METHODS
Blood Samples
All blood samples were anticoagulated with K2 or
K3EDTA. They were obtained with informed consent from
10 normal, healthy adult volunteers, and as the residual
material from patient samples drawn for clinical testing
purposes. The clinical diagnosis was verified to ensure
that all patient samples with reactive IG were obtained
from patients with an inflammatory or infectious disorder.
Specimens from patients with a malignant disease such as
CML, MDS, or from patients treated with hematopoietic
growth hormones such as G-CSF were analyzed sepa-
rately.
Sample Preparation
All blood samples were analyzed within 8 h from draw-
ing unless specified otherwise. As a first step, EDTA anti-
coagulated whole blood was diluted with phosphate-buff-
ered saline (PBS, pH 7.3) until the WBC concentration was
approximately 2,000 cells/␮L.
After this dilution, 100 ␮L of sample was incubated with
10 ␮L each of CD16 (clone: NKP15), fluorescein isothio-
cyanate (FITC), CD11b (clone: D12), phycoerythrin (PE),
CD45 (clone: 2D1), and peridinin chlorophyll protein
(PerCP) for 30 min in the dark at room temperature.
Mouse IgG1 FITC and mouse IgG2a PE were used as iso-
typic controls. All antibodies were from Becton Dickinson
(BD; San Jose, CA).
After incubation, the sample was mixed with 2 mL of
lysing reagent (8.26 g/L of NH4Cl, 1.0 g/L of KHCO3, and
0.037g/L of Na4EDTA in H2O) and incubated for 5 min at
room temperature. All flow cytometric measurements
were carried out within 60 min after sample preparation
unless specified otherwise.
Data Acquisition
Flow cytometric analysis was performed on a FACSCali-
bur (BD). FSC, SSC, and three-color fluorescence signals
(green, orange, and red fluorescence at 530, 585, and 650
nm, respectively) were determined for each cell and
stored in list mode data files. A total of 10,000 WBC events
were recorded for each specimen. All measurements were
performed under fixed instrument settings (i.e., photo-
multiplier tube [PMT] voltages, gains, and compensa-
tions). Before each measurement, quality control on the
instrument was performed with CaliBRITE3 beads (BD)
throughout the study.
Data Analysis
List mode data files were analyzed by a three-step gating
procedure using CellQuest software (BD). As a first step,
WBC were separated from debris by CD45-PerCP and SSC
dot plot (Fig. 1A). The WBC population in the dot plot was
gated in order to exclude debris and to count total WBC
events. In the same dot plot, four populations of WBC
(granulocytes, monocytes, lymphocytes, and basophils)
were identified and the granulocyte population (mature
 CLASSIFICATION OF IG 373

FIG. 1. Flow cytometric analysis of a pateint sample containing IG and a sample from a normal, healthy volunteer after staining with CD45-PerCP,
CD16-FITC, and CD11b-PE. In the first step, granulocytes (IG ⫹ neutrophils ⫹ eosinophils) were identified (A). In the second step, neutrophils and IG
were separated from eosinophils (B). In the third step, (C), IG stage 1 (CD16⫺/CD11b⫺) and stage 2 (CD16⫺/CD11b⫹) were identified as well as mature
neutrophils (CD16⫹/CD11b⫹). No IG were identified in any samples from normal, healthy adult individuals.

neutrophils, eosinophils, IG) could be gated (Fig. 1A).
Using this granulocyte gate, eosinophils were excluded
from IG and neutrophils by using a CD16-FITC and CD45-
PerCP combination (Fig. 1B). The neutrophils and IG were
identified by a CD16-FITC and CD11b-PE combination. As
a result, populations of IG stage 1, IG stage 2, and mature
neutrophils were classified (Fig. 1C).
The quadrant gate setting for CD16 and CD11b was
defined by 10 normal healthy adult specimens using pos-
itive staining and isotype controls. The channel number
for mean fluorescence intensity (MFI) for CD16 on neu-
trophils and IG was recorded as the positive control; the
MFI for isotype staining on the same specimen was re-
corded as the negative control. The boundary for CD16-
positive cells was determined as the center point between
the positive and the negative control MFI (Fig. 2). A similar
procedure was repeated for CD11b (Fig. 2). This gate
setting is maintained throughout the study. The percent-
age of all IG, IG stage 1, and IG stage 2 was calculated as
a proportion of all WBC by the following equations: IG %
⫽ ([IG stage 1 events ⫹ IG stage 2 events])/(total WBC
events) ⫻ 100, IG stage 1% ⫽ (IG stage 1 events)/(total
WBC events) ⫻ 100, and IG stage 2% ⫽ (IG stage 2
events)/(total WBC events) ⫻ 100.
Manual Microscopic IG Count
Manual 400 WBC differential counts (the 400 cell refer-
ence method) were performed as described by NCCLS
H20-A (14). Blood films were prepared by a manual-wedge
technique and stained with May-Grünwald-Giemsa. Two
medical technologists each performed a 200 WBC differ-
ential count on separate blood films and counted IG as
promyelocytes, myelocytes, and metamyelocytes,
whereas band cells were included in the mature, seg-
mented neutrophil population.
Reproducibility
Each of four patient samples was analyzed 10 times by
flow cytometry and by manual microscopic IG counting.
For this reproducibility study, we did not use the refer-
ence 400 WBC. Instead, we used a 100 WBC differential
because this is routine, clinical practice. Samples for the
flow cytometric method were individually stained and
lysed, whereas 10 separate blood films prepared from
each sample were available for manual microscopy.
Sample Stability
Long-term stability prior to staining was tested under
storage conditions at 4° and 25°C. Four patient samples
were stored and analyzed at 8, 24, and 48 h after blood
collection. Sample stability for flow cytometry was tested
at 30, 60, and 120 min after completion of the staining
procedure.
374 FUJIMOTO ET AL.
Cell Sorting
The flow cytometric results were confirmed by cell
sorting and subsequent microscopy of the sorted cells.
Cell sorting was carried out on FACStar plus (BD). Sorted
cells were suspended in 1.0% bovine serum albumin (BSA;
Nakarai Tesque, Kyoto, Japan) in PBS before making cy-
tospin preparations (Cytospin, Shandon, Pittsburgh, PA)
and subsequent staining with May-Grünwald-Giemsa.
Statistical Analysis
For overall statistical analysis and method comparison,
we followed established and published methods (15,16).
For comparison between flow cytometry and manual mi-
croscopic results (the 400 WBC reference method), we
used simple regression analysis and a coefficient of deter-
mination (r2). Imprecision (reproducibility) was measured
by calculating the coefficient of variation (CV) of multiple
measurements. A paired t-test was applied to determine
whether imprecision results for flow cytometry and man-
ual microscopy were statistically different.
RESULTS
Typical results for flow cytometric analysis of a patient
sample with IG and a normal sample without IG are
shown in Figure 1. In the patient sample, three cell cate-
gories can be observed in the CD16/CD11b dot plot; IG
stage 1 (CD16-/CD11b-), IG stage 2 (CD16-/CD11b⫹), and
mature neutrophils (CD16⫹/CD11b⫹). In samples from
normal, healthy adult individuals, no cells negative for
CD16 or CD11b were detected.
FIG. 2. Control data analysis and gate set-
tings for enumeration of IG. Negative and
positive controls were used for setting the
gate in the CD16/CD11b dot plot. Bound-
aries of CD16⫺/⫹ and CD11b⫺/⫹ were set in
the center point of MFI of negative and pos-
itive controls.
Cell Sorting
The cell populations identified by flow cytometry were
confirmed by cell sorting and subsequent morphological
examination. Microscopy of the sorted cells showed that
the most immature IG were mainly found in the region of
IG stage 1, metamyelocytes in the region of IG stage 2,
whereas band form and segmented neutrophils were
found in the mature, segmented neutrophil population
(Fig. 3).
Comparison Between the Flow Cytometric and
Microscopic IG
Comparison between flow cytometry and manual, mi-
croscopic 400 WBC reference counts on 46 samples
showed the following results: for all IG combined for both
methods, y ⫽ 1.34x ⫹ 0.95, r2 ⫽ 0.86; for IG stage 1
versus microscopic promyelocytes ⫹ myelocytes, y ⫽
1.53x ⫹ 1.24, r2 ⫽ 0.76; for IG stage 2 versus microscopic
metamyelocytes, y ⫽ 0.77x ⫹ 0.21, r2 ⫽ 0.58. All results
for the manual microscopy count were based on a 400
WBC differential count (Fig. 4).
Reproducibility
Results are shown in Table 1. These results were based
on each of four patient samples analyzed independently
10 times by either flow cytometry or by manual micros-
copy (100 WBC differential count). The differences be-
tween the flow cytometric method and manual micros-
copy were statistically significant (P ⬍ 0.01).
 CLASSIFICATION OF IG 375

FIG. 3. Microscopy of sorted cells stained with May-Grünwald-Giemsa. The most IG (promyelocytes and myelocytes) were mainly found in the region
labeled IG stage 1 (A). The more mature IG, metamyelocytes, were mainly found in the region labeled IG stage 2 (B). Band forms and mature neutrophils
were found in the region labeld mature neutrophil (C).

Sample Stability the percentage of stage 2 decreased within 12 h, whereas

Long-term stability of unstained samples showed negli- the percentage of IG stage 1 increased (Fig. 5). The net
gible effect up to 48 h when stored at 4°C. When samples result of these changes is that the total IG percentage
were stored at room temperature (25°C), we noticed that remained stable for at least 48 h. Stained samples were
stable for at least 120 min at room temperature (Fig. 6).

Samples From Patients With Hematological Disease and/or

FIG. 4. Correlation between the manual, microscopic reference IG
count (400 WBC counted) and the flow cytometry results (10,000 events
analyzed).
Who Received G-CSF Treatment
We evaluated a small number of patients with malignant
disease. We observed problems in samples from patients
with a diagnosis of MDS and from patients receiving G-CSF
treatment. In each of these cases, the MFI of CD16-FITC
for mature neutrophils was significantly lower than ob-
served in patients with reactive IG. As a result, correlation
between flow cytometric and manual microscopic IG
count was poor (Fig. 7).
376 FUJIMOTO ET AL.

Table 1
Reproducibility Results for 10 Repeat Analyses on Four Patient Samples
With Total IG Counts Ranging From 1 to 4/100 WBC
Mean CV%
Flow cytometry Manual microscopy
All IG combined 6.8 (range 5.1–8.8) 50.2 (range 39.7–65.7)
IG stage 1 7.9 (range 5.9–10.0) 64.9 (range 51.1–86.4)
IG stage 2 13.1 (range 7.0–21.1) 123.9 (range 76.6–161.0)

DISCUSSION expression increases through myeloid maturation into

Cellular membrane antigen expression and its detection
by monoclonal antibodies have been widely used for cel-
lular identification by flow cytometry. Some cellular anti-
gens, such as common leukocyte antigen (CD45), have
broad specificity. Others, such as CD3 for T cells or CD34
for hematopoietic progenitor cells (17), are highly specific
cell markers. Myeloid cells typically do not have specific
single markers, but can usually be recognized by using
combinations such as CD13 and CD14 that have been
recommended for monocytes (18,19). Similarly, a single,
specific marker expressed only on IG has not yet been
discovered. Therefore, we developed a strategy for detec-
tion of IG by using markers that relate to myeloid matu-
ration.
Several antigens have been studied in relation to gran-
ulocyte maturation: CD66b, CD66c, CD16, and CD11b.
For all of these, antigen expression varies according to the
degree of maturation and can either increase or decrease.
For example, CD66b (NCA-95) expression increases from
the promyelocyte to the metamyelocyte stage, but then
decreases in band and segmented neutrophil stages. Some-
what similarly, expression of CD66c (NCA-90) decreases
from the immature promyelocyte stage to the more ma-
ture band and segmented stage (20).
Expression of CD16 and CD11b increases during my-
eloid maturation from the myeloblast stage to the mature
segmented neutrophil, but at independent rates. Bone
marrow studies by Terstappen et al. (6) using the same
monoclonal antibody clones as in our study showed, sim-
ilarly to our observations, that immature myeloid cells do
not, or only minimally express, CD11b. Although CD11b
metamyelocytes, CD16 expression does not start until the
cell reaches the postmetamyelocyte maturation level (Fig.
2). We correlated our method of measuring CD11b and
CD16 expression with the morphology of sorted cells. We
also used direct correlation between flow cytometry data
and morphology from blood films. The difference be-
tween the morphology system that was used by Terstap-
pen et al. (6) and our group is illustrated in Table 2.
Briefly, Terstappen et al. (6) used a seven-step morpho-
logical description from promyelocyte to segmented neu-
trophil as opposed to the five-step system used by our
group. The overall outcome in correlation between the
two morphology systems and flow cytometry is virtually
the same. However, it should be realized that flow cyto-
metric analysis of maturation markers such as CD11b and
CD16 more or less reflects the biological continuum of
myeloid differentiation, whereas the morphological de-
scription follows a stepwise discrimination between cell
types differing in maturation. Overlap between morpho-
logically classified cells and flow cytometric categories
becomes inevitable.
The results of cell-sorting experiments showed that IG
could be separated from mature neutrophils by the use of
CD16-FITC and that these IG could be classified into more
and less mature by the use of CD11b-PE. Correlation
between flow cytometric IG combined and manual refer-
ence IG (promyelocytes ⫹ myelocytes ⫹ metamyelo-
cytes) was good (r2 ⫽ 0.86). However, we observed that
there was a tendency for IG counts determined by flow
cytometry to be higher than by microscopy (slope ⫽
1.34). We believe that this discrepancy is due to prepara-
tion techniques for blood films. Large cells, such as mono-
cytes, located at the feathered and lateral edges in blood
films prepared by the wedge method, result in significant
losses (18,21). The phenomenon is well documented for
monocytes. Because IG are also large cells, it is likely that

FIG. 5. Long-term stability of four samples was measured up to 48 h FIG. 6. Short-term stability of four stained samples was measured up
at 4°C and 25°C. Closed diamond, A; closed square, B; asterisk, C; to 120 min at room temperature. Closed diamond, A; closed square, B;
closed triangle, D. asterisk, C; closed triangle, D.
 CLASSIFICATION OF IG
FIG. 7. Patient sample with MDS and G-
CSF treatment.
a lower morphological IG count occurred for the same
reason.
It is obvious from the distribution of cells in the scat-
terplot (Fig. 2) that there are more than the three popu-
lations mentioned so far: promyelocytes and myelocytes;
metamyelocytes; and mature neutrophils. Between meta-
myelocytes and the mature neutrophils, a transitional cell
population exists. Most likely, these cells represent “left
shifted” or less mature neutrophils, also called “band” or
“stab” cells. These cells have been very difficult to identify
consistently by morphological criteria (22). Therefore, we
have abstained from correlating this transitional group
with a morphological substrate. Instead, we grouped
these cells with the mature neutrophil population.
Hübl et al. (9) described the use of CD16 in combina-
tion with CD45, CD14, and SSC for the detection of left
shift. They compared results with manual microscopy and
flagging results from two hematology analyzers (9). Their
results showed a marginal improvement in efficiency of
CD16- neutrophils over the neutrophil count alone (81 ⫾
3% versus 70 ⫾ 4% respectively at 95% specificity). Their
definition of left shift consisted of the morphological pres-
ence of any of the following cell types exceeding the
99.85 percentile of a reference population (or if more
than one count exceeded the 97.5 percentile): band cells,
metamyelocytes, myelocytes, and promyelocytes. Because
the band cells predominated over all other cell types
combined, it is not surprising that a morphology bias as
well as the use of a single antibody for myeloid maturation
resulted in a disappointing performance. Our objective
was to establish flow cytometric criteria to define IG in
the peripheral blood, not to define left shift. Because IG as
a group can be reasonably well defined by morphology,
and probably because we used a double-antibody label for
Table 2
Correlation Between Flow Cytometric Analysis and Morphology
Terstappen et al. (6)
CD11b CD16 Predominant morphological feature
⫺ ⫺ Promyelocytes Early myelocytes
⫹ ⫺ Early myelocytes Myelocytes
⫹⫹ ⫺ Myelocytes Metamyelocytes
⫹⫹ ⫹ Bands Metamyelocytes
⫹⫹ ⫹⫹ Segmented Presegmented
neutrophils neutrophils
a
This population is assumed by us to contain transitional or band neutrophils (see text).
377
myeloid differentiation, the correlation between flow cy-
tometry and morphology for all IG combined is stronger in
our study than found by Hübl et al. (r2 ⫽ 0.86 or r ⫽ 0.927
as opposed to r ⫽ 0.541, respectively). The use of a CD16-
myeloid population alone would include a significant
number of transitional myeloid cells, the left shifted cells,
possibly causing the large number of outliers as men-
tioned in their study.
Reproducibility was far better than could be achieved
by the routine IG count (mean CV for the flow cytometric
IG was 6.8% versus 50.2% for the manual microscopic IG).
This is in major part due to the higher number of cells
counted by flow cytometry (10,000 versus 100 by manual
microscopy), thereby reducing Poisson error. As stated
above, we used a 100 cell differential and not a 400 cell
reference count for the manual method, because this is
routine, clinical practice. It is likely that inconsistent mor-
phological identification also plays a role. This is because
the CV for the manual count of IG is far greater than for
other easily recognizable, but generally low-frequency,
cell types such as eosinophils.
Stability tests showed that there was no effect on the
flow cytometric results when samples were stored for at
least 48 h at 4°C after blood collection, which is far better
than for manual microscopy, and for at least 120 min after
completion of the staining procedure. When stored at
room temperature (25°C), we found a change in IG stage
1 and 2 populations. Because the percentage of all IG does
not change, we assume that IG stage 2 shifted to stage 1
due to a decrease in CD11b expression. This is supported
by a decrease in MFI for CD11b in mature neutrophils
when stored under similar conditions (data not shown).
Therefore, for analysis of IG staging, we recommended
storing blood specimens at 4°C rather than at room tem-
This paper
CD11b CD16 Predominant morphological feature
⫺ ⫺ Promyelocytes Myelocytes
⫺ ⫺ Promyelocytes Myelocytes
⫹ ⫺ Metamyelocytes
⫹ Dim Bandsa
⫹ ⫹ Segmented neutrophils Bands
378 FUJIMOTO ET AL.
perature. Storage conditions should not influence the mor-
phological count of all IG regardless of stage of matura-
tion.
We showed the same pattern for CD11b and CD16
distribution as shown in Figure 2 in two patients with
acute promyelocytic leukemia. We were able to monitor
induction of myeloid differentiation by all-trans retinoic
acid (ATRA) chemotherapy in both patients (results not
shown). However, results from tests on blood samples
from patients with CML, MDS, and from patients treated
with G-CSF showed that there is a limitation for this flow
cytometric method. Abnormal expression of CD16 or
CD11b has been known to exist in certain disorders: a
decreased expression of CD16 and elevated expression of
CD11b has been observed in the mature neutrophils from
patients with MDS (11,12,23). Our own data showed a
flow cytometric IG (CD16- neutrophils) count significantly
higher than by manual microscopy (flow cytometry 25.5%
versus manual IG 10.0%). These results support the pres-
ence of CD16- or CD16dim mature neutrophils in MDS
patients. A decreased CD16 expression in mature, seg-
mented neutrophils has also been reported in patients
treated with G-CSF (13). Our data support this.
While acknowledging the limitations of our method,
our data also indicate that IG can be reliably identified by
the use of CD16 and CD11b antibodies in a flow cytomet-
ric setting. The method has a high degree of precision and
is accurate in patients with nonmalignant disease and who
have not received G-CSF. In addition, the method is not
influenced by distribution effects as with wedge blood
film preparations. The method would thus be appropriate
for monitoring reactive IG in patients with infectious
diseases. Additionally, the method could be of use in the
evaluation of automated methods for IG enumeration by
hematology analyzers, provided samples analyzed are
taken from such patients only. Although this restricts the
method’s use somewhat, it would still be of significant
utility. This is because in many cases of an infectious
process leading to the presence of IG in the peripheral
blood, such presence is often limited to a few cells, mak-
ing method comparison by manual microscopy very diffi-
cult. The availability of a flow cytometric IG count with
much greater precision, albeit not suitable for all occa-
sions, would be of great assistance to evaluate IG counting
performance by hematology analyzers.
LITERATURE CITED
1. Schwinzer R. Cluster report CD45/CD45R. In: Knapp W, et al.,
editors. Leukocyte typing IV: white cell differentiation antigens. Ox-
ford, England: Oxford University Press; 1989. p 628 – 634.
2. Robinson JP. Study of cell function: oxidative metabolism. In: Robin-
son JP, et al., editors. Current protocols in cytometry. New York: John
Wiley & Sons; 1997. p 9.7.1-9.7.14.
3. Tomasello E, Revello V, Poggi A. CD16 workshop panel report. In:
Kishimoto T, et al., editors. Leukocyte typing VI. White cell differen-
tiation antigens. New York: Garland Publishing; 1997. p 269 –271.
4. Edberg JC, Kimberly RP. Modulation of Fc gamma and complement
receptor function by the glycosyl-phosphatidylinositol-anchored form
of Fc gamma RIII. J Immunol 1994;152:5826 –5835.
5. Hogg N. CD11b workshop panel report. In: Kishimoto T, et al.,
editors. Leukocyte typing VI. White cell differentiation antigens. New
York: Garland Publishing; 1997. p 345–347.
6. Terstappen LW, Safford M, Loken MR. Flow cytometric analysis of
human bone marrow. III. Neutrophil maturation. Leukemia 1990;4:
657– 663.
7. Kansas GS, Muirhead MJ, Dailey MO. Expression of the CD11/CD18,
leukocyte adhesion molecule 1, and CD44 adhesion molecules during
normal myeloid and erythroid differentiation in humans. Blood 1990;
76:2483–2492.
8. Gopinath R, Nutman TB. Identification of eosinophils in lysed whole
blood using side scatter and CD16 negativity. Cytometry 1997;30:
313–316.
9. Hübl W, Andert S, Thum G, Ortner S, Bayer PM. Value of neutrophil
CD16 expression for detection of left shift and acute-phase response.
Am J Clin Pathol 1997;107:187–196.
10. Kabutomori O, Iwatani Y, Koh T, Yanagihara T. CD16 antigen density
on neutrophils in chronic myeloproliferative disorders. Am J Clin
Pathol 1997;107:661– 664.
11. Bowen KL, Davis BH. Abnormal patterns of expression of CD16
(FcRgIII) and CD11b (CRIII) antigens by developing neutrophils in
the bone marrow of patients with myelodysplastic syndrome. Lab
Hematol 1997;3:292–298.
12. Felzmann T, Gisslinger H, Krieger O, Majdic O, Ludwig H, Koller U.
Immunophenotypic characterization of myelomonocytic cells in pa-
tients with myelodysplastic syndrome. Br J Haematol 1993;84:428 –
435.
13. Felzmann T, Gisslinger H, Krieger O, Majdic O, Ludwig H, Koller U.
Altered surface marker expression and function of G-CSF-induced
neutrophils from test subjects and patients under chemotherapy. Br J
Haematol 1994;87:31–38.
14. NCCLS Approved Standard H20-A. Reference leukocyte differential
count (proportional) and evaluation of instrumental methods. Vil-
lanova, PA: National Committee for Clinical Laboratory Standards;
1992.
15. ICSH Expert Panel on Cytometry. Guidelines for the evaluation of
blood cell analyzers including those used for differential leucocyte
and reticulocyte counting and cell marker applications. Clin Lab
Haematol 1994;16:157–174.
16. Weber N, Mast BJ, Houwen B. Performance evaluation of hematology
analyzers: an outline for clinical laboratory testing. Sysmex J Int
1995;5:103–112.
17. Gratama JW, Orfao A, Barnett D, Brando B, Huber A, Janossy G,
Johnsen HE, Keeney M, Marti GE, Preijers F, Rothe G, Serke S,
Sutherland DR, Van der Schoot CE, Schmitz G, Papa S. Flow cytomet-
ric enumeration of CD34⫹ hematopoietic stem and progenitor cells.
European Working Group on Clinical Cell Analysis. Cytometry 1998;
34:128 –142.
18. Lebeck LK, Mast BJ, Houwen B. Flow cytometric white blood cell
differentials: a proposed alternate reference method. Sysmex J Int
1993;3:61– 69.
19. Hübl W, Hauptlorenz S, Tlustos L, Jilch R, Fischer M, Bayer PM.
Precision and accuracy of monocyte counting. Comparison of two
hematology analyzers, the manual differential and flow cytometry.
Am J Clin Pathol 1995;103:167–170.
20. Boccuni P, Pardo C, Alfinito F, Vacca C, Rotoli B, Vecchio L. CD66
family workshop: CD66 family expression during myeloid expres-
sion. In: Leukocyte typing VI. White cell differentiation antigens.
Garland Publishing; 1997. p 1003–1005.
21. Koepke JA, Dotson MA, Shifman MA. A critical evaluation of the
manual/visual differential leukocyte counting method. Blood Cells
1985;11:173–186.
22. Cornbleet PJ, Novak RW. Classifying segmented and band neutro-
phils. CAP Today 1994;8:37– 41.
23. Ohsaka A, Saionji K, Igari J, Watanabe N, Iwabuchi K, Nagaoka I.
Altered surface expression of effector cell molecules on neutrophils
in myelodysplastic syndromes. Br J Haematol 1997;98:108 –113.