CELLULAR IMMUNOLOGY 132, 102-114 (1991)

Unique Human Neutrophil Populations Are Defined by Monoclonal Antibody ED1 2F8ClO’

CHRISTOPHERC.BROWN,HARRY L.MALECH,ROBERTJ.JACOBSON,* CAROLYNF.SHRIMPTON,~PETERC.BEVERLY,I.

ANTHONYW.SEGAL,~-ANDJOHNLGALLIN~

Bacterial Diseases Section, Laboratory ~J’Clinical Investigation, National Institute ofAllergy and Infectious Diseases. National Institutes of Health, Bethesda. Maryland 20892; *Department of Medicine,

Georgetown University, Washington. D.C.; and the TDepartment qf!fMedicine, Rayne Institute, University, College London, London, England

Received April I I, 1990: accepted August 9, 1990

A mouse IgGl monoclonal antibody ED12F8CIO (CIO) binds a constant percentage ofperipheral blood neutrophils in the same individual when studied over time, defining a distinct subset of neutrophils in all normal individuals studied to date. Bone marrow studies confirm that the heterogeneity is present to the same degree at all stages of neutrophil development from the myelocyte to the mature neutrophil. Neither in vivo nor in vitro activation of neutrophils explains or significantly alters the relative percentages of CIO-positive and -negative neutrophils in the same individual. With both activation and exudation, however, expression of the CIO-defined epitope increases in intensity in the C 10 binding subpopulation. Studies of NBT reduction, phago- cytosis, adherence, light scattering characteristics, and monoclonal antibody surface binding have failed to demonstrate physical or functional differences between the ClO-defined populations. We examined Cl0 binding in patients with different defects of phagocyte function. In two patients with neutrophil-specific granule deficiency. less than I% of the neutrophils were found to be Cl0 positive, while neutrophils from a patient with idiopathic leukemoid reaction and recurrent in- fections demonstrated greater than 99% C IO binding. Although the present study does not delineate the physiologic significance of Cl0 binding heterogeneity, it firmly supports the concept of neu- trophil heterogeneity at the level of surface antigen expression. a 1991 Academx press. hc.

INTRODUCTION

The existence of neutrophil heterogeneity was first suggested by Florence Sabin in 1923 (1). Sabin observed that neutrophils have heterogenous motile responses to the same stimuli, observations which were later confirmed by Howard and by Harvarth (2, 3). Since Sabin, a variety of reports have appeared claiming that neutrophils were heterogenous (for review see Ref. (4)) with regard to rosette formation (57), density (S), membrane depolarization (9), protein synthesis (IO), alkaline phosphatase content (1 I), and oxidative metabolism (9, 12). Criticism of these studies has consistently

i Data were presented in part at the annual meeting of the American Federation for Clinical Research, San Diego, CA, May 1-4, 1987 (1).

* To whom correspondence and reprint requests should be addressed at National Institute of Allergy and Infectious Diseases, Building 10, Room I IC103, 9000 Rockville Pike, Bethesda, Maryland 20892.

102

0008-8749/9 1$3.00 Copyright 0 I99 I by Academic Press, Inc. All rights of reproduction in any form resewed

NEUTROPHIL HETEROGENEITY 103

focused on three major points. First, it is not always clear that the heterogeneity de- scribed is more than nonpurposeful biologic variation. Second, the level of maturity of individual cells could affect their function, making less mature neutrophils function at a lower level. Third, oxidative, chemotactic and other responses of neutrophils are thought by many to depend upon the state of activation and level of “priming.” Neutrophils in various states of activation could, therefore, function in a heterogenous fashion.

Recently, a number of monoclonal antibodies which label neutrophils in a heter- ogenous fashion have been developed ( 13- 15). These monoclonal antibodies clearly demonstrate that phenotypic heterogeneity exists among neutrophils with regard to surface epitope composition. It is unlikely that nonpurposeful biologic variation could account for this heterogeneity. State of activation and maturity are still concerns, however. To date, only one neutrophil-specific monoclonal antibody, 3 1 D8, has been reported to identify neutrophil subpopulations that differ in functional capacity (13). Biologic variation and state of activation do not appear to explain the heterogeneity seen with 31D8. Studies in neonates, trauma patients and normal volunteers given endotoxin suggest, however. a correlation between neutrophil immaturity and the subset of 3 1 D8 defined “dull” neutrophils ( 16- 18). In the present report we describe findings with an IgG 1 mouse monoclonal antibody, ED 12F8C 10 (C 10). We show that C 10 defines both Cl0 binding and C 10 nonbinding neutrophil populations in all normal individuals studied to date. Biologic variation and state of maturity do not appear to explain or influence C 1 O-defined populations. Similarly, although in vitro and in viva activation and exudation of neutrophils increases the intensity of Cl0 epitope expression on the binding population, the relative number of binding (Cl0 positive) and nonbinding (Cl0 negative) neutrophils does not change.

METHODS

Study? populution. Thirteen normal volunteers were used for injection studies, bone marrow aspirate, or blister exudate analysis, in accordance with NIH protocols 74-I- 99 (bone marrow aspiration), 76-l-349 (steroid administration), 77-I-185 (skin blister exudate), 80-I-96 (epinephrine administration), and 79-I-91 (endotoxin administra- tion). Thirty-seven additional subjects were normal, healthy volunteers while seventeen patients had various disorders of phagocyte function [specific granule deficiency (n = 2) preleukemia (n = 2), idiopathic leukemoid reaction with recurrent infections (n = l), chronic myelogenous leukemia (n = I), idiopathic neutropenia (n = l), chronic granulomatous disease of childhood (n = 5) Chediak-Higashi syndrome (n = I), uncharacterized monocyte dysfunction (n = l), and Hyperimmunoglobulin E-recurrent infection (Job’s) syndrome (n = l)].

In vivo studies. Three normal volunteers were given intravenous injections with 0.1 mg epinephrine (NIH protocol 80-I-96). Venous blood was collected before injection of epinephrine, immediately after and at 5 min and I hr following injection. Three normal volunteers received intravenous hydrocortisone, 200 mg, (NIH protocol 76- 1-349). Venous blood was collected before and at 1, 2, 4, and 6 hr after the injection. Two normal individuals were injected with 3 rig/kg and one volunteer received 1 ng/ kg Escherichia co/i RE-2 endotoxin (Federal reference standard; FDA BB-IND 1309, NIH protocol 79-1-91). Venous blood was collected before injection and at 1, 2, 4, and 6 hr after injection.

Blister exudate ana/wis. Three normal volunteers had blisters raised on one forearm

104 BROWN ET AL.

with an eight-well blister suction unit (Nemo Probe Inc., Bethesda, MD) as described previously (18, 19) (NIH protocol 77-I-185). Samples were obtained at the time of maximal neutrophil response (5 hr).

Cell preparation. Neutrophils were separated from heparinized venous blood and bone marrow by dextran sedimentation alone or by Hypaque-Ficoll gradient centrif- ugation followed by dextran sedimentation. The latter method resulted in >95% neu- trophils. In some studies, leukocytes from IOO-~1 samples of whole blood were stained, lysed, and fixed with Coulter Immunolyse (Coulter, Hialeah, FL). When multiple samples were obtained from a volunteer over time, processing of each individual blood sample was begun (neutrophil purification, antibody labeling, cell fixation) within 10 min of being drawn.

Production of antineutrophil antibody EDIZF8ClO. The monoclonal antibody Cl0 was prepared using previously described methods (22, 26, 28). Supernatants from fusion products were screened for binding to neutrophils and to the immunizing protein cytochrome b-245. Cl0 was found to bind a subset of neutrophils but not to cytochrome b-245. Limiting dilutions of cells from positive wells were used for subcloning. An- tibodies were obtained from supernatants of media from growing hybridoma cultures. The Ig subclass of the monoclonal antibody (IgGl) was determined by fluorescence- activated cell analysis using subclass-specific fluoresceinated goat anti-mouse mono- clonal antibodies (Coulter Clone, Hialeah, FL).

Antibody preparation and cell labeling. Monoclonal antibody ED 12FBC 10 was ob- tained either as culture medium supernatant from growing hybridoma cells or was purified by ammonium sulfate precipitation from hybridoma supernatants (13). Monoclonal antibody 3 1 D8 was obtained as culture medium supernatant or ascites. Control antibody was polyclonal mouse Ig anti-horse spleen ferritin (Jackson Im- munoresearch Laboratories, Avondale, PA). For indirect labeling we used fluorescein conjugated goat Fabz anti-mouse IgG (Cooper Biomedical, Malvern, PA) as the second antibody. For some studies we used ammonium sulfate-purified C 10, 3 1 D8 or control mouse antibody, and biotinylated them with Biotin-X-NHS (Calbiochem, La Jolla, CA) or conjugated them with fluorescein (Sigma, St. Louis, MO) as previously described (23). Blocking and double labeling experiments used biotinylated or fluorescein con- jugated MO1 (Coulter Immunology, Hialeah, FL), fluorescein-conjugated rabbit an- tihuman lactofenin (Jackson Immunoresearch, Avondale, PA), and 3G8 (bind human IgG Fc receptors; supplied by Howard Fleet). Freshly prepared live cells ( 1 06/ml) were labeled at 4°C or 30 min in phosphate-buffered saline which contained 1% bovine serum albumin (0.5 mg/ml human IgG), Cooper Biomedical, Malvern, PA) and either hybridoma media supernatant, control antibody, or directly conjugated antibodies. Labeled cells were then washed and immediately fixed with phosphate-buffered saline containing 1% paraformaldehyde. All samples were stored in the dark at 4°C. Control studies showed that fixation and storage in the dark after labeling did not alter the fluorescence patterns significantly over a period of several days before analysis.

Fluorescence-activated cell analysis and sorting (FACS). Samples were analyzed with a fluorescence-activated cell sorter (FACS II, Becton-Dickinson, Sunnyvale, CA), an Epics 753 or Profile (both from Coulter, Hialeah, FL). The Epics 753 was used for sorting. Forward and 90 degree light scatter characteristics were used to distinguish neutrophils from monocytes and lymphocytes (gating verified by sorting on at least three occasions). This type of gating does not distinguish basophils or eosinophils from neutrophils. However, we have shown previously that 3 1 D8 does not label eosinophils

NEUTROPHIL HETEROGENEITY 105

or basophils and similar results were obtained for C 10 ( 13). Therefore, although eo- sinophils and basophils are present in small numbers, they contribute to the ClO- negative peak in all samples analyzed. Sorting studies (see below) confirmed that eo- sinophils were found only in the CIO-negative peak, and in less than 1% of the cells from 3 ID8 bright or dull peaks. For each sample analyzed 10,000 to 50,000 cells (gated to analyze only neutrophils) were counted. For some studies, the gating param- eters for forward light scatter or side light scatter were changed in order to examine portions of the neutrophil population with different light scatter characteristics. Re- gardless of the type of cell preparation and labeling used, we observed no significant differences in relative percentages of C I O-positive or -negative cells.

Data were displayed with the number of cells on the vertical axis and the logarithm of fluorescence intensity on the horizontal axis. Negative fluorescence was defined as all cells with fluorescent intensity below l-2% of the upper limit fluorescence of cells labeled with control antibody. Using a Model 2210-0.43.C graphic digitizer tablet (Numonics Corp., Montgomeryville, PA), version 3.1 Sigma-Scan software (Jandel Scientific, Corte Madera, CA) and a Compaq portable III computer (Compaq Com- puter Corp., Houston, TX). we derived values for percentage of C lo-negative neutro- phils in a given sample by dividing the measured integral of the area of the ClO- negative peak by the integral of the total area of the fluorescence analyzer-generated histogram of ClO-determined cell fluorescence. This value was subtracted from 100 to obtain the value for percentage of C lo-positive neutrophils within a given sample. In most cases C I O-positive and C 1 O-negative peaks were clearly distinguishable with minimal overlap. In these circumstances the integrals ofthe areas ofboth ClO-positive and CIO-negative peaks were easily measured, the upper limit of the negative peak coinciding with the point which corresponded with the upper l-2% of the negative control fluorescence peak. In the few instances when the C 1 O-positive and C lo-negative peaks overlapped. the outline of a symmetric CIO-negative curve was generated using the mode and shape of the ClO-negative peak and the point on the fluorescence axis (s axis) below which greater than 98% of negative control cell fluorescence was located. The integral of the area of this generated C 10 curve was measured and divided by the total integral of the area of cell fluorescence as described above, giving a value for the percentage of C lo-negative or nonbinding cells in the histogram. All measurements were made by one investigator (CB). On separate occasions the investigator determined the percentage of C lo-positive cells for five different histograms. Five separate deter- minations were made for each histogram. Repeat measurements of the same histogram always showed a standard deviation of less than 1.8% C 1 O-positive neutrophils, with 99% confidence limits within 4.4%).

To determine changes in fluorescence intensity of the ClO-positive peak under different conditions, we used the Sigma-Scan to measure the distance on the horizontal axis of a histogram between the mode of the ClO-negative peak and the mode of the C IO-positive peak. Under the different conditions tested, only the ClO-positive peak demonstrated changes in fluorescence intensity. The distance from the mode of the CIO-negative peak on the horizontal axis (measuring fluorescence) to the mode ofthe C 1 O-positive peak was measured using the Sigma Scan software.

Fluor~sscence-uctivatcd sorting. Peripheral blood neutrophils or marrow leukocytes, labeled with ClO, were sorted live in Epics 753 within 3 hr of obtaining the sample from the volunteer. Gates for the sort of ClO-positive cells were set to include only cells with fluorescence intensity above the mode of the C lo-positive peak in order to

106 BROWN ET AL.

minimize any contribution of C IO-negative cells. Gates for the sort of ClO-negative cells were set to include cells below the mode of the C 1 O-negative peak. These limits were chosen to ensure minimal overlap of one C 1 O-defined neutrophil population to another. A minimum of one million cells was sorted from each gated region, spun down, resuspended in buffer to a concentration of 100,000 cells/ml, and distributed on slides with a cytospin (Shandon-Southern). The cytospins were stained with a modified Giemsa stain and 200 cell differential counts were performed by two separate observers. A X2 analysis was used to determine any significance of correlation between leukocyte type and Cl0 labeling pattern.

In vitro activation and aging of neutrophik For activation studies, purified neutro- phils were suspended at a concentration of 100,000 cells/ml in Hank’s buffer with magnesium and calcium. The suspensions included a variety of activators which had been previously added (phorbol myristate acetate (Sigma P8139), 10 and 20 rig/ml; f-met-leu-phe (Sigma F6632), 1 nM, 100 nM, and 10 yM; RE-2 E. coli endotoxin (Bureau of Biologics) 10 rig/ml; 10% zymosan (ICN 10 1237) activated serum (vol/ vol); granulocyte-macrophage colony stimulating factor (Genzyme RH-CSF) 300 PM, methylprednisolone (Sigma M0639), 1 mg/ml. These samples were incubated for 20 min in a shaking water bath at 37°C washed at 4°C and labeled with Cl0 as de- scribed above.

For in vitro aging of neutrophils, heparinized whole blood was collected and in- cubated at time intervals up to 72 hr under sterile conditions overnight at room tem- perature and at 37°C in Teflon-coated beakers. Neutrophils from these samples were counted, purified, and labeled with antibody at different time points.

In vitro functional studies (phagocytosis, NBT reduction and adherence to endothelial cell monolayers) of CIO-positive and -negative neutrophils. C 1 O-positive and -negative neutrophils were sorted as described above, and studied for their phagocytic and NBT- reducing capacity as previously described (24). Differences in ability of ClO-positive and -negative neutrophils to adhere to human umbilical cord endothelial monolayers were assessed as previously described (25).

RESULTS

Staining of normal peripheral blood with Cl 0. C 10 defines two distinct neutrophil populations (binding and nonbinding) in all normal individuals studied to date (n = 37). The single peak of background fluorescence achieved with a nonbinding control antibody approximates the same area as the nonbinding or CIO-negative peak. Al- though overlap between the ClO-positive and -negative peaks can occur, both peaks are recognizable as distinct and separate (Fig. 1). The mean and standard deviation of ClO-positive neutrophils in the population of normal, healthy volunteers (n = 37) was 58 + 19.8%, with a relatively uniform distribution of percentage of ClO-positive neutrophils among 37 normal individuals studied (Fig. 2). Ten individuals (7 normals and 3 patients with defects of phagocyte function) were studied at irregular intervals over a period from weeks to months. The standard deviation of percentage of C 1 O- positive circulating neutrophils sampled for each individual over time averaged 3.0%, range 0 to 5.1%. The variation range average for all 10 individuals studied was 3.0%.

Cl 0 binding and its relationship to cell maturity. To examine whether C 10 binding was dependent on the state of neutrophil maturity, we sorted C 10 positively and negatively labeled peripheral blood neutrophils obtained from normal volunteers (Table

NEUTROPHIL HETEROGENEITY 107

Cl0 Fluorescence I Log) -

FIG. 1. Cl0 determined peripheral blood labeling patterns in nine (A-l) randomly selected normal vol- unteers. The histogram peaks tilled in with hatchmarks represent the negative control antibody labeling pattern for each individual tested. Note that the negative control peak overlaps the CIO-negative peak and that CIO-positive and CIO-negative peaks are clearly delineated in each example. The calculated values for percentage of ClO-positive neutrophils in each example are as follows: A = 88%. B = 53’70, C = 30%. D = 61%. E = II?,. F = 54%. G = 29W, H = 498. I = 78’1,.

1). There was no significant correlation between morphological maturity of neutrophils (i.e., band forms) and pattern of Cl0 labeling. A similar sort and analysis of C 10 labeled bone marrow cells from a healthy normal donor was done as described under Methods. We noted no staining of erythrocytes, lymphocytes, or monocytes in any patient studied. There was no significant correlation between morphological maturity of neutrophils (band forms) and the ClO-positive or -negative labeling pattern. Cl0 labeling was seen as early as the metamyelocyte stage. In related studies the effect of aging of purified peripheral blood neutrophils in vilru on the relative percentages of C 1 O-positive and C 1 O-negative neutrophils was assessed using heparinized blood from normal volunteers. No change was noted in the C lo-labeling pattern of cells analyzed before or after aging for up to 72 hr at room temperature.

I!?&Y c?f’h~drocorti.sone. epinephrine, endotoxin, and exudation on Cl0 labeling of neutrophils (Table 2). Following intravenous epinephrine there was no change in pe-

FIG. 2. Distribution of percentage of CIO-positive neutrophils in 37 normal individuals.

% Cl0 POSITIVE PMN

108 BROWN ET AL

TABLE 1

DIFFERENTIAL WHITE BLOOD COUNT FOR C IO ANTIBODY-POSITIVE AND -NEGATIVE CELLS IN BONE MARROW AND PERIPHERAL BLOODY

Percent positive

PMN Band Pro/Meta/Myelo Eos Erythroid precursor

Blood c-IO(f) ClO(-)

Marrow clo(+) CIO(-)

83 17 - 1 - 79 II - II -

58 39 <I 2 <l 30 24 17 9 20

0 Values obtained are from cell preparations sorted by flow cytometry and then fixed and stained for morphologic determination. Data are means oftwo samples from blood and one sample from bone marrow.

ripheral blood percentage of neutrophils positive for Cl0 staining and no difference in the fluorescence intensity of the ClO-positive cells. Similarly, there was no change in the percentage of neutrophils positive for Cl0 staining or in the intensity of Cl0 labeling of neutrophils from three individuals 6 hr after intravenous injection of 200 mg hydrocortisone, a time when poststeroid injection leukocytosis was maximal. In three subjects challenged with endotoxin, no change in percentage of ClO-positive neutrophils was seen. Two of three patients showed an increase in fluorescence of ClO-positive neutrophils. Furthermore, there was no difference in the relative per- centage of C IO-positive and -negative neutrophils between exudate and simultaneously

TABLE 2

CHANGES IN PERCENTAGE CIO-POSITIVE PMN AFTER INTRAVENOUS INFUSION WITH HYDROCORTISONE, EPINEPHRINE, OR ENDOTOXIN

CIO-Positive PMN

% Positive* Fluorescence intensity*

Agent” Subject Preinfusion Postinfusion Ratio (post/preinfusion)

Hydrocortisone I 31 32 1.46 2 90 93 1.01 3 23 21 1.23

Epinephrine 4 55 57 0.98 5 69 63 I .07 6 71 71 0.88

Endotoxin 7 59 66 1.82 8 75 81 1.42 9 64 62 1.14

” Subjects 1-3 received 200 mg intravenous hydrocottisone, subjects 4-6 received 0. I mg epinephrine, subject 7 received 1 mg/kg .%cherichiu co/i RE-2 endotoxin, subjects 8 and 9 received 3 mg/kg E. culi RE- 2 endotoxin.

* P > 0.05 for each group, paired sample I test, postinfusion data obtained during maximum neutrophil response

NEUTROPHIL HETEROGENEITY 109

FIG. 3. Differences in percentage of CIO-positive neutrophil distribution and intensity of Cl0 binding in three individuals. Peripheral blood and simultaneously collected exudative neutrophils (after 5 hr exudation) were studied for C IO binding by flow cytometry. In the right panel. the fluorescence intensity (log scale) of the CIO-positive peak is compared in three individuals in simultaneously collected peripheral blood and exudate neutrophils.

collected peripheral blood cells. However, in exudate neutrophils, the mean fluorescence intensity of the ClO-positive cells was significantly greater (P < 0.05) than simulta- neously collected peripheral blood cells (Fig. 3).

c&t ofin vitro activation gj’neutrophils on Cl0 labeling. Hypaque-Ficoll purified neutrophils were treated in vitro with concentrations of f-met-leu-phe, phorbol myristate acetate, endotoxin, zymosan activated serum, granulocyte-macrophage colony stim- lating factor (GM-CSF). or prednisolone known to modify neutrophil function. The different types of treatment protocols resulted in no significant change in the relative percentages of ClO-positive and -negative neutrophils (data not shown). However, after treatment for 30 min with phorbol myristate acetate (20 ng/mI) there was an increase in the mean fluorescence of the C 1 O-positive cells while no significant change was noted in the percentage of ClO-positive cells (Fig. 4). A similar increase in C 10 fluorescence was seen in cells activated with zymosan and f-met-leu-phe (data not shown).

E.ypression of‘CR3, IgFcR, lactofkrrin or 31D8 on Cl O-positive and -negative PMN. To determine whether C 1 O-positive and -negative neutrophils differed in expression of known surface components such as CR3. the IgG Fc receptor, and lactoferrin, we activated neutrophils for 30 min at 37°C with 10 rig/ml phorbol myristic acetate and then double labeled with C 10 antibody and either MO 1, anti-lactoferrin, 3 lD8, or 3G8 (anti-human IgG Fc receptor). No differences in the labeling patterns of anti- MO I, anti-lactoferrin. 3 I D8 or 3G8 were noted between C IO-positive and -negative

.

? J

FIG;. 4. Effect of incubating neutrophils for 30 min with buffer (inactivated) or PMA (20 rig/ml) (activated) on Cl0 binding. Left panel shows percentage of CIO-positive neutrophils and right panel the fluorescence

intensity. P values are paired sample f test.

110 BROWN ET AL.

neutrophils in a basal or activated state. An example of anti-MO1 staining before (upper panel) and after (lower panel) activation is presented in Fig. 5. Note that the increase in mean fluorescence of anti-MO 1 is equivalent in both ClO-positive and C 1 O-negative populations.

In related experiments incubation of neutrophils with unlabeled antibodies against CR3 (MOl, OKM-I), FcR (3G8, 11 A, 1 lB, 1 IC), antilactoferrin, or 3 lD8 did not prevent directly labeled C 10 from binding to neutrophils while in control studies 1 O- fold excess unlabeled Cl0 resulted in complete inhibition of binding with directly labeled C 10. Incubation of purified neutrophils with Cl0 resulted in no change in viability, NBT reduction, or superoxide production, compared to untreated neutrophils.

Characteristics of CIO-positive and -negative neutrophils. Following incubation of Hypaque-Ficoll purified neutrophils with endothelial cell monolayers, the relative percentages of C 1 O-positive and -negative neutrophils in the adhered and nonadhered populations was not significantly different (7% adherent). Similarly, when endotoxin- stimulated neutrophils were adhered to endothelial cell monolayers, the percentage of adherent cells increased to 56% but the relative percentage of CIO-positive cells in the adherent and nonadherent populations did not change.

Forward light scatter (indicator of cell size) and right angle light scatter (indicator of internal light scattering characteristics, or granularity) were evaluated. No differences in forward or right angle light scatter characteristics between ClO-positive and ClO- negative labeling neutrophils were seen (Fig. 6).

In two studies the ability of purified (cell sorter) positive and negative neutrophils to reduce NBT indicated that 96 k 2% of sorted ClO-negative neutrophils reduced NBT following activation with PMA ( 100 pg/ml). In addition, 90 t 4% of C IO-positive neutrophils phagocytized opsonized Candida albicans, with an average of 3.1 candida/ cell, whereas 92 k 5% of sorted ClO-negative neutrophils phagocytized opsonized C. albicans, with an average of 3.0 Candidalcell.

Anti- MO1 Fluorescence ( Log k-

FIG. 5. Increase in ClO-positive fluorescence which occurs after activation for 30 min with phorbol my&ate acetate (20 rig/ml) (bottom panel) compared with neutrophils incubated in buffer (upper panel). Note the increased intensity of anti-MO1 binding in activated cells without change in percentage of ClO-positive binding of cells.

NEUTROPHIL HETEROGENEITY 111

Forward Scatter -

9oD Light Scatter -

FIG. 6. Right angle and forward angle laser determined light scatter in CIO-positive and Cl20-negative neutrophils from the peripheral blood of a normal volunteer. The horizontal line in the middle of each panel denotes the separation between C IO and C I O-negative neutrophils.

Cl0 e,xpression in patients with phagocyte dqf2ct.r (Table 3, Fig. 7). C 10 binding to neutrophils from patients with chronic granulomatous disease, Chediak-Higashi, hy- perimmunoglobulin E-recurrent infection (Job’s) syndrome, and uncharacterized monocyte dysfunction was normal. In two patients with inherited deficiency of neu-

TABLE 3

PERCENTAGE OF C IO NELJTROPHILS IN PATIENTS WITH DEFECTS OF PHAGOCYTE FUNCTION

Phagocyte defect a % Cl0 Neutrophils*

None (normal) (37) Specific granule deficiency (2) Preleukemia (2) Idiopathic leukemoid reaction (1) Myelogenous leukemia (I) Acquired immune deficiency (2) Idiopathic neutrophenia (1) Chronic Granulomatous disease of childhood (5) Chediak-Higashi syndrome (I) Uncharacterized monocyte dysfunction (I) Hyperimmunologobulin E-recurrent infection

Job’s Syndrome (1)

58 i 19.8 0. 0 55,58 100 55 58.68 55 53, 60, 62. 65. 65 52 58

34

a Number of different patients studied is in parentheses. b Each value represents the mean of at least two different studies in an individual

112 BROWN ET AL.

Fluorescence ( Log 1 -

FIG. 7. Cl0 labeling patterns of a normal volunteer, compared with those of specific granule deficient and idiopathic leukemoid reaction patients. (A) A fluorescence histogram of neutrophils from a normal volunteer incubated with a fluorochrome-labeled negative control antibody. The C IO labeling pattern of the same normal volunteer is presented in B. (C) The negative control example for the Cl0 pattern (<I% CIO- positive) seen in a patient with specific granule deficiency (D). (E) The negative control example for the Cl0 labeling pattern (F) of the patient with idiopathic leukemoid reaction. The vertical line represents the negative control determined cutoff for negative fluorescence.

trophil-specific granules, C 10 bound to less than 1% of peripheral blood neutrophils on each patient’s neutrophils. In contrast, in a patient with idiopathic leukemoid reaction studied three times, greater than 99% of neutrophils were Cl0 positive on each occasion. The very high percentage of ClO-positive cells in the patient with idiopathic leukemoid reaction was not a result of clonal expansion of a single stem cell population since the patient’s neutrophils proved heterogenous for G6PD with both A and B G6PD isoenzymes found in erythrocytes and leukocytes (data not shown). Studies in a patient with chronic myelogenous leukemia, with clonal expansion of a single stem cell population, revealed a normal percentage of C 1 O-positive and -negative neutrophils.

DISCUSSION

The meaning of neutrophil heterogeneity is not clear (4). In order to demonstrate that truly unique neutrophil populations exist, one must establish that the populations are constant and do not exist simply as a function of immaturity, senescence, state of activation, or random, nonpurposeful biologic variation. Most importantly, one should be able to demonstrate unique physical and functional characteristics for such neu- trophil populations.

In this report we describe an IgG 1 mouse monoclonal antibody (C 10) that is specific for neutrophils, identifies two distinct (binding and nonbinding) neutrophil populations in 37 normal individuals studied to date, and binds to an epitope whose surface expres- sion is enhanced when the neutrophil is activated in vitro or in vivo.

The heterogeneity defined by this monoclonal antibody is unique compared to that of previous reports of neutrophil heterogeneity. For example, the heterogeneity is distinct from that dependent on differences in Fc receptors (5, 7), lactoferrin content

NEUTROPHIL HETEROGENEITY 113

(29). NBT reduction (30). membrane depolarization (9) chemotaxis (l-3) light scat- tering characteristics. size, age (1) adherence (3 I), phagocytosis (6), and distribution in the various blood pools (4). Further characterization of ClO-positive neutrophil populations revealed that the relative number of C lo-positive and -negative cells persists regardless of maturity of the cells, unlike the monoclonal antibody 31D8, where a correlation can exist between 3 1 D8 “dull” neutrophils and bands ( 13, 18). In contrast to 3 1 D8, C 10 demonstrates little or no overlap in the fluorescence intensity of C 1 O- positive and -negative cells.

Ball et al. ( 14) reported a monoclonal antibody, which was prepared from myeloblasts from a patient with acute myelogenous leukemia, that recognized normal monocytes as well as a subpopulation of neutrophils. Marked heterogeneity in the density of the antigen among cells expressing the antigen was also observed, which is different from the C 10 antigen. To our knowledge the observation by Ball et al. has not been extended to study of bone marrow elements or pathologic states.

Clement et u/. ( 15) raised two monoclonal antibodies that bound specifically to 57%) and 5 1 Yj of neutrophils with a high degree of overlap in the binding among neutrophils, much greater than seen with Cl0 antibody used in the current report (Fig. I). The antibodies, called lB5 and 4Dl. bound to different antigenic determinants on the neutrophils. IB5 was a y2a,K antibody, and 4D 1 was a y 1 ,K antibody. Antibody 4D 1 bound to a 59-kDa protein: the binding site for the lB5 was not identified, although it was not the 59-kDa protein. The antibodies did not compete with binding to each other. Functional studies of the different neutrophil populations were not reported.

We found no functional differences between C 1 O-positive and -negative neutrophils either in riro (epinephrine, hydrocortisone endotoxin challenge and exudation) or in ri/ro. Two of three patients treated with endotoxin demonstrated an increase in CD 1 O- positive fluorescence. possibly due to complement-mediated activation by endotoxin. We therefore studied Cl0 distribution in a variety of patients selected for defects in neutrophil function. It is provocative that three patients exhibited abnormal Cl0 expression. One of these patients had a leukemoid reaction with >99% CIO-positive neutrophils. This patient had heterogenous G6PD expression indicating the presence of a polyclonal neutrophil population. The other two patients, with abnormal specific granule maturation, had no C 10 expression. Although this suggests that the C 10 epitope may reside in specific granules. we do not know the cellular location since cellular fractionation studies have not been performed and proteins from other intracellular compartments. such as the gelatinase compartment, are also known to be upregulated by activation (32). Each of the three patients with abnormal Cl0 epitope expression has abnormal neutrophil maturation. The possible association of C 10 epitope expres- sion with early neutrophil maturation is intriguing and warrants further investigation.

REFERENCES

I. Sabin. F.. Bfrll .lo/lr~s lfoph. Ilos/>. 34. 217. 1923. 2. Howard, T. H. E.. Bloocl59, 946. 1983. 3. Harvarth. L.. Slrr~,. Irnmfrn. Rev. 2, 145. 1983. 4. Gallin. J. 1.. Blood 63, 977. 1984. 5. Wong. L.. and Wilson. J. D.. J Iwwurnol. .21&xk 7, 69. 1975. 6. Klempner. M. S.. and Gallin. J. 1.. Blood51, 659. 1978. 7. Fanger. M. W., Shen. L.. and Pug. J.. I+w. Null. kucl. Ser. c/S.4 77, 3640. 1980 8. Pember. S. 0.. Barnes. K. C.. and Brandt. S. J.. Blood 61, 1105. 1983. 9. Seligmann. B., Chused. T. M.. and Gallin. J. 1.. J. Clir~. Invw. 68, 1125. 1981.

114 BROWN ET AL.

10. Granelli-Piperno. A.. Vassali, J. D., and Reich, E., Medicine 146, 1693, 1977. 11. Fehr, J., and Grossman, H.. Amer. J. Hematol. 7, 369, 1979. 12. Bass, D. A., Olbrantz, P., and Szejda, P.. J. Immunol. 136, 860, 1986. 13. Seligmann, B., Malech, H. L., and Melnick, D. A,, J. Immunol. 135, 2647, 1985. 14. Ball, E. D., Graziano, R. F., Shen. L., and Fanger. M. W., Proc. Nat/. Acad. Sci. USA 79, 5374, 1984. 15. Clement, L. T., Lehmeryer, J. E., and Gartland. G. L., Blood 61, 326, 1983. 16. Krause, P. J., Malech, H. L.. Kristie, J., Kosciol. C. M., Herson, V. C., Eisenfeld, V., Pastuszak, W. T.,

Kraus, A., and Seligmann, B.. Blood 68, 200. 1986. 17. Krause. P. J., Maderazo, E. G., Bannon. P., Kosciol, C. M., and Malech, H. L., J. Lab. C/in. Med. 112,

208, 1988. 18. Brown, C. C., Malech. H. L., and Gallin, J. I., Ce//. Immunol. 123, 294, 1989. 19. Zimmerli. W., Seligmann, B., and Gallin, J. I., J. Clin. Invesf. 77, 925, 1986. 20. Zimmerli, W., and Gallin, J. I., J. Immunol. Me/hods 96, 11. 1987. 2 1. Harper, A., Dunne, M. J.. and Segal, A. W., Biochem. J. 219, 5 19, 1984. 22. Kohler, G., and Milstein, C., Eur. J. Immunol. 6, 51 1, 1976. 23. Baines, G., and Brakeslee, D., J. Immunol. Methods 13, 305, 1976. 24. Metcalf, J. A., Callin, J. I.. Nauseef, W. M., and Root, R. K., “Laboratory Manual of Neutrophil

Function.” pp. 87-96. Raven Press, New York, 1986. 25. Rotrosen, D.. and Gallin, J. I., J. CeN Biol. 103, 2379, 1986. 26. Gallin, J. I.. Jacobson, R. J., Seligmann. B. E.. Metcalf, J. A., McKay, J. H., Sacher, R. A., and Malech,

H. L., Blood 68, 343, 1986. 27. Dale, D. C., Fauci, A. S., Guerry, D. P., and Wolff. S. M., J. C/in. Invest. 56, 808, 1975. 28. Hammerling. G. L., Hammerling, U., and Kearney, J. F.. “Monoclonal Antibodies and T-Cell Hybri-

domas.” pp. 563-587. Elsevier/North-Holland. New York, 198 1. 29. Broxmeyer, H. E.. Ralph P.. Bognachi. J., Kinkade, P. W., and Desousa, M., J. Immunol. 125, 903,

1980. 30. Repine, J. E., Rasmussen, R., and Whiter, J. G., Amer. J. C/in. Palho/. 71, 582, 1979. 31. Klempner, M. S., Ann. Intern. Med. 92, 529, 1980. 32. Falloon. J.. and Gallin, J. I.. J. Allergy Clin. Immunol. 77, 653, 1986.