Neutrophilic Differentiation Induced by Human Upper Airway Fibroblast-derived Granulocyte/Macrophage Colony-stimulating Factor (GM-CSF) Carlo Vancheri, Takayuki Ohtoshi, Gerard Cox, Antoni Xaubet, John S. Abrams, Jack Gauldie, Jerry Dolovich, Judah Denburg, and Manel Jordana Departments of Pathology, Pediatrics, and Medicine, McMaster University, Hamilton, Ontario, Canada, and DNAX Research Institute of Molecular and Cellular Biology Inc., Palo Alto, California
We have established primary lines of fibroblasts from nasal polyp (NP) tissues as well as from normal nasal (NN) mucosa and have examined the ability of these cells to release hormone-like peptide messenger molecules (cytokines). Our results show that human upper airway fibroblasts release granulocyte/macrophage colony-stimulating factor (GM-CSF), granulocyte-CSF (G-CSF), and macrophage-CSF (M-CSF) in vitro. We also show that fibroblasts derived from NP tissue express the gene for GM-CSF at a higher level, and release the GM-CSF product in greater amounts, than NN fibroblasts. In addition, we have examined the ability of these fibroblasts and their conditioned medium (CM) to induce differentiation of human hemopoietic progenitor cells. After 7 d, cultures of these cells in RPMI-lO% fetal bovine serum contained 5 ± 2.5% (mean ± SD) neutrophils. In contrast, culture of progenitor cells with fibroblasts resulted in significantly greater neutrophilic differentiation (18 ± 4 %). Culture in fibroblast-CM induced a similar degree of differentiation, and fibroblast-CM from NP fibroblasts elicited greater differentiation compared to CM from NN fibroblasts (17.5 ± 3 versus 12 ± 3%). The neutrophilic differentiation induced by fibroblast-CM can be fully inhibited by preincubating this CM with a monoclonal neutralizing antibody to human GM-CSF. Thus, our results demonstrate: (1) the ability of human upper airway fibroblasts to release GM-, G-, and M-CSF in vitro; (2) that fibroblasts derived from NP tissues express the gene and release the product GM-CSF at greater levels compared to NNfibroblasts; and (3) that fibroblast-derived GM-CSF causes neutrophilic differentiation of human hemopoietic progenitors.
Nasal polyps, a condition of unknown etiology occurring as often in patients with atopy as in the general population, are grape-like formations arising from the posterior nasal and sphenoid sinus mucosae (1). Nasal polyps are characterized by epithelial cell and varying degrees of fibroblast hyperplasia, as well as by the accumulation, in variable numbers, of inflammatory cells, including neutrophils, eosinophils, metachromatic cells, and lymphocytes (1-3). These inflammatory cells can release a variety of enzymes and mediators thought
(Received in original form August 28, 1989 and in revised form June 11, 1990) Address correspondence to: Manel Jordana, M.D., McMaster University, Department of Pathology, Room 4HI7-21, 1200 Main Street West, Hamilton, Ontario L8N 3Z5, Canada. Dr. Vancheri's present address is: Istituto di Malattie dell'Aparato Respiratorio e Tisiologia, Via Passo Gravina 187, 4500 Catania, Italy. Abbreviations: conditioned medium, CM; fetal bovine serum, FBS; granulocyte colony-stimulating factor, G-CSF; granulocyte/macrophage colonystimulating factor, GM-CSF; macrophage colony-stimulatingfactor, M-CSF; 2-mercaptoethanol, 2-ME; normal nasal, NN; nasal polyp, NP; nonadherent mononuclear cells, NAMe. Am. J. Respir. Cell Mol. BioI. Vol. 4. pp. 11-17, 1991
to be involved in the expression of symptoms associated with this condition. Thus, the mechanisms modulating the accumulation and/or activation of those cells in the nasal mucosa are central to the pathogenesis of nasal polyps. Traditionally, the role ascribed to the stromal cell population has been circumscribed to reparative processes. However, it is increasingly apparent that stromal cells such as fibroblasts are also capable of releasing a number of hormone-like messenger peptides (cytokines) and are thus effector cells themselves (4-6). Our laboratory has previously proposed the hypothesis of "microenvironmental control of airway inflammation" to indicate that key aspects of the inflammatory process such as proliferation and activation of mature inflammatory cells as well as differentiation of inflammatory cell progenitors might be modulated at the tissue level by signals originating, at least in part, from the tissue structural cells (7, 8). However, the precise interactions between cells and cytokines involved in this modulation remain to be clarified. We have examined the contribution of human upper airway fibroblasts, derived from nasal polyp (NP) and normal nasal (NN) tissue, to the differentiation of human hemopoietic progenitors. We show here that conditioned media
12
AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 4 1991
from NP fibroblasts induce neutrophilic differentiation of these progenitors to a significantly greater degree than that elicited by NN fibroblasts. We also show that these conditioned media contain granulocyte/macrophage colony-stimulating factor (GM-CSF), granulocyte-CSF (G-CSF), and macrophage-CSF (M-CSF), and that NP fibroblasts release, compared to NN fibroblasts, significantly greater amounts of GM-CSF. Finally, we show that the neutrophilic differentiation induced by fibroblast-conditioned medium can be fully abrogated by preincubating this conditioned medium with a neutralizing monoclonal rat anti-human GM-CSF antibody.
Materials and Methods Cells HL-60 cells were obtained from the American Type Culture Collection (Rockville, MD). Experiments were carried out with cells between the thirtieth and fiftieth passage. Routinely, these cells were subcloned twice weekly to a density of 2.5 x 105 cells/ml with RPMI 1640 (Rosswell Park Memorial Institute) supplemented with 10% heat-inactivated fetal bovine serum (FBS) (GIBCO, Grand Island, NY), 1% penicillin/streptomycin (GIBCO), and 25 mmol/ liter Hepes buffer (Boehringer Mannheim Canada Ltd., Dorval, Quebec, Canada). Cells were kept in a humidified atmosphere at 5 % CO2 and 37° C. Primary lines of human fibroblasts were established in our laboratory using an outgrowth from an explant method as previously described (9). After removal of the explants, fibroblasts were expanded and frozen in liquid nitrogen for future use. In the experiments described here, fibroblasts were always used at a passage earlier than the tenth. Tissue Source Characterization We established nine primary fibroblast lines from NP tissue removed at the time of polypectomy. Four patients had tested positive to one or more skin tests performed in the week prior to the surgical procedure. Four patients tested negative to a battery of skin tests, including 12 antigens, and one patient refused to be skin-tested. Control nasal tissue refers to nasal mucosa tissue from five patients who did not have a history of nasal allergy, had negative skin tests, and did not have nasal polyps. Tissue was obtained at the time of surgery for mechanical obstruction or cosmetic correction. These studies conformed to guidelines for human experimentation and were approved by the Chedoke-McMaster Hospital Ethics Committee. In Vitro Proliferative Characteristics of Nasal Fibroblasts Fibroblasts from confluent dishes were trypsinized and counted. A total of 2 x 105 fibroblasts were then delivered in 4 ml of RPMI-supplemented medium into each well of 6-well plates (Falcon; Becton Dickinson, Oxnard, CA). The cells were incubated for various time intervals at 37° C and 5% CO,. At the end of the incubation period, that is, at days 1, 3, 8, and 11, the cells were trypsinized and counted with a hemocytometer. Two wells were prepared for each data point, and two counts were always performed per well. Cell viability as assessed by trypan blue exclusion was 95 % or greater in all experiments.
Preparation of Conditioned Medium (CM) Fibroblast-CM was obtained from 2 X 106 fibroblasts seeded in 35-mm ¢ wells of 6-well plates (Falcon; Becton Dickinson) with 4 ml of RPMI-supplemented medium (5 x 105/ml). In most experiments, fibroblast-CM was used immediately upon harvesting. In some instances, the CM was aliquoted and frozen at -20° C for future use. Experimental Cultures Fibroblasts, 1.5 X 106 cells, were seeded in 100-mm tissue culture dishes (Corning, Corning, NY) and allowed to adhere overnight. The next day, the medium was washed off and 5 x 105 HL-60 cells in 10 ml of supplemented RPMI were added. After 3 d, 5 ml of fresh medium was added to each dish. In another series of experiments, HL-60 cells were cultured with fibroblast-CM obtained as described above. In this protocol, cells were recovered after 3 d of culture, spun down at 1,000 rpm, and resuspended in either fresh CM or supplemented RPMI. Unless otherwise indicated, fibroblast-CM was used at 50% concentration. Following the same protocol, HL-60 cells were in some instances cultured with various concentrations of recombinant human cytokines. In addition, in some specific experiments in which we examined the effect of a rat anti-human GM-CSF and a mouse anti-human G-CSF antibody, the fibroblast-CM was preincubated with the antibody for 2 h at 37° C before being added to the HL-60 cells. All experiments were terminated at day 7 of culture. At that time, HL-60 cells were recovered from each dish after at least two washes, spun down at 1,000 rpm for 10 min, and resuspended in 2 ml of medium. The cells were counted with a hemocytometer, and viability was assessed by using 0.1% trypan blue exclusion (GIBCO). Morphology Cytocentrifuge preparations were made using a ShandonSouthern Cytospin (Johns Scientific, Toronto, Ontario, Canada). In all experiments, smears were stained with the DiffQuik'" modification of the May-Giemsa technique (American Scientific Products, McGraw Park, IL). Assay of Colony-forming Units Mononuclear cells were isolated from heparinized peripheral blood using Ficoll-Hypaque density sedimentation (specific gravity, l(177) and washed twice with McCoy's 5A medium (GIBCO), supplemented with 15% FBS, 1% penicillin-streptomycin, and 5 X 10-5 2-mercaptoethanol (2-ME). Nonadherent mononuclear cells (NAMC) were obtained by removal of adherent cells after incubation in plastic tissue culture flasks (Falcon ~024; Becton Dickinson) for 2 h at 37° C. A total of 106 NAMC per 35-mm culture dish (Falcon 3001) were cultured in 0.9% methylcellulose containing Iscoves modified Dulbecco's medium (GIBCO) supplemented with 20% FBS and 5 x 10-5 M of 2-ME with or without fibroblast-CM. Granulocyte colonies of ~ 40 cells were counted under an inverted microscope (Olympus, Tokyo,Japan) at day 14. Reagents A rat anti-human monoclonal neutralizing anti-GM-CSF antibody was generously provided by Dr. John S. Abrams
Vancheri, Ohtoshi, Cox et al.: Neutrophilic Differentiation by Fibroblast-derived GM-CSF
(DNAX Research Institute, Palo Alto, CA). In a colony assay, 1:100 dilution of antibody is capable of neutralizing up to 5 Dlml recombinant GM-CSF (10 pg/ml). A mouse antihuman monoclonal neutralizing anti-G-CSF antibody was kindly donated by Dr. Bruce W. Altrock (Amgen Corp., Thousand Oaks, CA); the neutralizing activity of this antibody in a CFV-GM bone marrow assay is 100,000 neutralizing Uzrnl. Immunoassay of GM-, G-, and M-CSF in Fibroblast-CM The content of human GM-CSF in fibroblast-CM was detected by means of a specific immunoassay performed by Dr. John S. Abrams. The limit of sensitivity of this assay is 30 pg/rnl. Human G-CSF content was quantitated in a monospecific sandwich immunoassay by Dr. Bruce W. Altrock, and the limit of sensitivity of this assay is 0.2 ng/ml (10). Human M-CSF content was assessed using a radioimmunoassay by Dr. Peter Ralph (Cetus Corp., Emeryville, CA); the limit of sensitivity of this assay is 0.8 ng/ml (11). RNA Preparation and Analysis Fibroblast RNA was isolated following essentially the same procedure as previously described (12). Briefly, fibroblasts were lysed with 5 M guanidium isothiocyanate and total RNA was recovered by ultracentrifugation through a 5.7 M CsCI gradient, resuspended in diethylpyrocarbonate-treated water, and then denatured with formaldehyde. The RNA was then transferred onto a nitrocellulose blot and prehybridized at 42° C for4t06hin4x SSC (IX SSC = 0.15 mM NaCl/15 rnM Na citrate, pH 7.0), 2x Denhardt's, 50 mM Na phosphate, pH 7.0,20%formamide, 10% dextran sulphate, 1 mM Na pyrophosphate, 50 ttg/ml ATP, and 50 ttg/ml sonicated salmon sperm DNA (13). GM-CSF mRNA was detected using a human cDNA probe kindly provided by Dr. G. Wong (14). The probe was labeled to 109 cpm/ug using hexanucleotide primers and 32P-dCTP (15). A total of 107 counts per minute of probe were added to 10 ml of prehybridization solution, and the blot was hybridized at 42° C overnight. Afterwards, the blot was washed in 0.2 X SSC/OJ% NaDodS04 , dried, and exposed for 48 h to Kodak X-Omat X-ray film with intensifying screens. Statistical Analysis Statistical comparisons were made using a nonpaired Student's t test.
25
Fibroblast-induced Neutrophilic Differentiation HL-60 cell cultures in RPMI containing 10% FBS showed a relatively small amount of spontaneous neutrophilic differentiation (5 ± 2.5%). In contrast, a statistically significant greater number of neutrophils was found (18 ± 4 %) when the HL-60 cells were cocultured with NP-derived fibroblast monolayers (Figure 1). A similar result was obtained by culturing HL-60 cells with NP fibroblast-CM, indicating that a soluble factor released by the fibroblasts was responsible for the effect. As shown in Figure 2, CM derived from different cultures of NN fibroblasts induced neutrophilic differentiation above the level of spontaneous HL60 differentiation but significantly less than NP fibroblast-
T
20
-ET Z ~
15
10 5
CONTROL 9) ( n
=
NP (n
= 3)
Figure 1. Fibroblast-induced neutrophilic differentiation of HL-60 cells. The empty bar depicts the neutrophilic differentiation (percentage of neutrophils in the differential count) of HL-60 cells cultured in RPMI-lO% FBS and harvested at day 7. Results represent mean ± SD of nine experiments. The solid bar depicts the neutrophilic differentiation of HL-60 cells cocultured with a monolayer of nasal polyp (NP) fibroblasts. Results represent mean ± SD of three experiments in which three different primary fibroblast lines were used.
CM (12 ± 3% versus 17.5 ± 3%, respectively). The effect offibroblast-CM on neutrophilic differentiation was dose dependent (Figure 3). We carried out additional experiments in which the effect of fibroblast-CM, specifically NP-CM, on HL-60 cells was evaluated at day 10 of culture as opposed to day 7. We found that there were no significant differences between these two time points in regard to the extent of neutrophilic differentiation (data not shown). Finally, Table 1 shows the results of a detailed morphologic characterization of HL-60 cells cultured with CM from NP fibroblasts. The data, which are combined from four experiments using four different NP-
25
20
-ET
15
Z
~
Results
13
10
S.D.
5
0 CONTROL ( n
= 9)
NN (n
= 5)
NP (n
= 4)
Figure 2. Fibroblast-conditioned medium (CM)-induced neutrophilic differentiation of HL-60 cells. The empty bar shows, as in Figure 1, the neutrophilic differentiation of HL-60 cells cultured in RPMI-lO% FBS. The solid bars depict the neutrophilic differentiation ofHL-60 cells cultured with fibroblast-CM from normal nasal (NN) or nasal polyp (NP) fibroblasts and harvested at day 7. Results represent mean ± SD of five and four experiments, respectively. In each experiment, a different primary fibroblast line was used to generate the CM.
14
AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 4 1991
TABLE 2
20
Cytokine content offibroblast-conditioned medium (Fib-CM) 15
Fib-CM
GM-CSF*
G-CSFt
M-CSF+
(pg/ml)
(ng/mll
(ng/mll
-GZ
10
Normal nasal tissue
194 (n
Nasal polyp tissue
5
663 (n
5
10
25
50
±
69
= 5)
± 258 = 12)
1.3
± 0.9
(n
= 5)
4.8 (n
± 0.8 = 2)
0.6
± 0.3
5.0
± 0.0
(n
= 5)
(n
= 2)
* Mean ± SEM; sensitivity of the assay is 30 pg/m!. t Mean ± SEM; sensitivity of the assay is 0.2 ng/ml. t Mean ± SEM; sensitivity of the assay is 0.8 ng/m!.
90
Fib-CM Concentration ("10)
Figure 3. Dose response of fibroblast-conditioned medium (CM)induced neutrophilic differentiation of HL-60 cells. The empty bar represents differentiation of cells cultured in RPMI alone. Results depict mean + SD of three experiments using CM from different nasal polyp fibroblast cultures.
CM, illustrate that the majority of the induced cells (66.2 %) were either band or segmented, that is, mature neutrophils. Immunoassay of GM-, G-, and M-CSF in Fibroblast-CM Table 2 shows the human CSF content in CM from a number of individual primary lines of fibroblasts from different tissue sources. Fibroblast-CM contained detectable amounts of G- and M-CSF, but no significant differences related to tissue source (NN- versus NP-derived fibroblasts) were seen. Fibroblast-CM also contained GM-CSF, and the amounts present in conditioned media from NP fibroblasts were significantly greater than those present in CM from NN fibroblasts (663 ± 255 versus 194 ± 70 pg/ml [mean ± SEM)). Proliferative Characteristics of Nasal Fibroblasts Table 3 shows the growth characteristics of three primary lines of NN fibroblasts and of three primary lines of NP fibroblasts. It is apparent that the number of cells in culture of either type of fibroblasts is.very similar at days 1, 3, and 8. The data also show that the cell number of NN fibroblasts reaches a plateau, in contrast to NP fibroblasts, between days 8 and 11; in fact, the difference in cell number between both types of fibroblasts is statistically significant at day 11.
TABLE 1
Effect of Specific Neutralizing Antibodies on the Neutrophilic Differentiation Induced by Fibroblast-CM Figure 4 shows the effect of a neutralizing rat anti-human monoclonal neutralizing antibody on the neutrophilic differentiation induced by fibroblast-CM. The dose range of the antibody used for these experiments was similar to that which fully abrogates the effect of fibroblast-CM on eosinophil survival in vitro (16). It can be seen that this anti-GlvlCSF antibody inhibits the neutrophilic differentiation induced by fibroblast-CM in a dose-dependent fashion with full inhibition at a 1:50 dilution. Under the same experimental conditions, an anti-G-CSF antibody at concentrations up to 1:100 did not inhibit fibroblast-CM-induced neutrophilic differentiation of HL-60 cells. Effect of Fibroblast-CM on Colony Growth Table 4 shows the effect of a number of fibroblast-CM on colony growth in a methylcellulose assay system. The positive control in these experiments was CM from the Mo human T lymphocyte cell line, which contains GM-CSF (17). Mo-CM at 5 % concentration induced the emergence of about 24 colonies of which 4.5 were GM colonies. CM from five different primary lines of NP fibroblasts, used at a concentration of 15 %, induced an average of 32 colonies of which more than 12 were GM colonies. Rat monoclonal antiGM-CSF antibody at a concentration of 1:50 (vol/vol) inhibited colony formation induced by fibroblast-CM (Table 5). RNA Analysis Figure 5 illustrates that a low-intensity signal is seen in lanes where total RNA obtained from cultures of NN fibroblasts was loaded. The intensity of the signal is stronger in those
Fibroblast-induced neutrophilic differentiation of HL-60 cells Stage
Immature neutrophils Myelocytes Metamyelocytes Mature neutrophils Total
%
Morphology
Band Segmented
± 0.5* ± 1.2 7.0 ± 2.3 5.5 ± 1.0 18.8 ± 3.1 1.1 5.2
TABLE 3
Relative %
(6.1 (27.7 (36.3 (29.9
Proliferative characteristics of nasal fibroblasts in vitro
± 3.0)t ± 2.6) ± 7.0) ± 7~4)
(100%)
* Results are expressed as mean ± SD of four experiments in which four different nasal polyp-conditioned media were used. t The numbers in parentheses represent the proportion of cells at each stage to the total number of neutrophils.
Days in Culture Day I
NN-Fib* (n = 3) 2.1 NP-Fibt (n = 3) 2.1
* NN-Fib
± 0.2t ± 0.3
Day 3
6.8 5.5
± 2.6 ± 1.3
Day 8
14.1 11.6
± 3.9 ± 4.6
= normal nasal fibroblasts. t NP-Fib = nasal polyp fibroblasts. t Cell number (X 10'). Results expressed as mean ± SD.
Day II
13.8 18.0
± 2.9 ± 4.7
Vancheri, Ohtoshi, Cox et al.: Neutrophilic Differentiation by Fibroblast-derived GM-CSF
15
TABLE 5
15
Effect of anti-GM-CSF antibody on colony growth induced by fibroblast-conditioned medium (Fib-CM)*
S.D.
GM Colonies Total Colonies
10
Negative control! (n = 2) Positive control! (n = 2) Fib-CM§ (n = 3) Fib-CM§ + anti-GM-CSF 1:50 (n
-ez at 5
* Results expressed
=
1.0 2.5 7.5 3) 0.5
± ± ± ±
0.0 0.5 2.1 0.3
3.5 12.5 22.3 0.8
± ± ± ±
0.5 1.5 4.7 0.2
as mean ± SEM.
t Medium alone. t CM at a 5 % concentration from the Mo human T lymphocyte cell line, CONTROL FIb-eM
which contains GM-CSF. § Fib-CM was used at a 15% concentration.
FIb-eM + anti GM-CSF ab
1:500
1:200
1:50
Figure 4. Effect of a neutralizing rat anti-human GM-CSF antibody on the neutrophilic differentiation of HL-60 cells induced by fibroblast-conditioned medium (CM). The empty bar shows the neutrophilic differentiation of HL-60 cells cultured in RPMI-lO% FBS. The solid bar depicts the neutrophilic differentiation induced by fibroblast-CM. The hatched bars illustrate the neutrophilic differentiation induced by this fibroblast-CM when preincubated with the anti-GM-CSF antibody.
lanes where RNA recovered from NP fibroblasts was loaded. In all instances, the RNA was obtained from the same number of cultured fibroblasts, plated at the same cell density, and incubated for 48 h.
Discussion There is increasing evidence that the tissue microenvironment represents a compartment where inflammatory and immune cell function is modulated. One component of this activity involves the release of soluble signals (cytokines) by the structural cells. These cytokines can interact with both immature and mature inflammatory cells and thereby modulate their differentiation, proliferation, and activation, thus participating in the regulation of the inflammatory response (18, 19). In regard to the effector capability of fibroblasts, Zucali and associates have shown that supernatants from recombinant human (rh) interleukin-l (IL-l)-stimulated human lung fibroblasts contain GM-CSF activity (20). Kaushansky and colleagues have shown that quiescent dermal fibroblasts constitutively release low amounts of G-, GM-, and M-CSF in culture and that rh-IL-l markedly enhance the transcription and release of both G- and GM-CSF (4). A similar observation has been made by Seelentag and cowork-
TABLE 4
Effect ofconditioned medium from nasal polyp fibroblast (Fib-CM) on colony growth* GM Colonies
Total Colonies
1.2 ± 0.2 4.5 ± 1.6 4.3 ± 0.5
3.2 ± 0.8 23.8 ± 7.6 15.2 ± 4.0
Negative control! (n = 3) Positive controlt (n = 3) Fib-CM§ (n = 5)
* Results
expressed as mean ± SEM.
t Medium alone. t CM at a 5% concentration from the Mo human T lymphocyte cell line,
which contains GM-CSF. § Fib-CM was used at a 5 % concentration.
ers; they have also shown that IL-l and tumor necrosis factor (TNF) have an additive stimulatory effect on G- and GMCSF production by skin fibroblasts (21). In addition to the colony-stimulating factors, fibroblasts express and release the gene product interleukin-6 (IL-6) (5), which has now recognized hemopoietic functions (22, 23) as well as a molecule, now referred to as interleukin-8 (IL-8) (24), which is chemotactic for neutrophils (6) and lymphocytes (24). With regard to upper airway inflammation, we have previously shown that a potential mechanism for inflammatory cell accumulation is the ingress and local differentiation of inflammatory cell progenitors (25-27). In the studies we report here, we sought to investigate whether human upper airway fibroblasts release cytokines that could stimulate the differentiation of hemopoietic progenitors and whether such activities were related to disease expression. To address this question, we examined the effect of fibroblasts and fibro-
5
2.5
JJRNA
NP NP NP NN NN NC Figure 5. Detection of GM-CSF-specific mRNA (slot-blot) in human upper airway fibroblasts. The first three lanes show the level ofGM-CSF expression of tot a! RNA extracted from three different primary lines of nasal polyp (NP) fibroblasts. Lanes 4 and 5 were loaded with RNA from two different lines of normal nasal (NN) fibroblasts. NC represents the negative control, which was calf liver.
16
AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 4 1991
blast-CM from both NP and NN tissues on the differentiation of the human HL-60 myeloid leukemia cell line, which is capable of differentiating into mature cells of different lineages, depending on the inducing agent (28-30). Thus, this assay system is useful to study biochemical and molecular events associated with differentiation of progenitor cells into functionally mature inflammatory cells. HL-60 cells differentiated toward neutrophils when cocultured with a monolayer of human upper airway fibroblasts (Figure I). Fibroblast-CM induced a similar degree of neutrophilic differentiation of HL-60 cells, indicating that a soluble factor present in the CM was responsible for the effect, and the differentiating activity induced by CM from NP-derived fibroblasts was significantly greater than that induced by CM from fibroblasts obtained from NN tissues (Figure 2). This difference appeared permanent in that it was maintained in fibroblasts that had gone through a number of passages in vitro. The neutrophilic differentiation induced by fibroblast-CM was not specific for the HL-60 target because, as shown in Table 3, this CM induced the emergence of GM colonies from a preparation of nonadherent peripheral blood mononuclear cells containing hemopoietic progenitors. Since the type of differentiation seen suggested the presence of colony-stimulating factors in the CM, we then determined the content of such factors in the CM using specific radioimmunoassays. As shown in Table 2, G-, GM-, and M-CSF were detected in fibroblast-CM from both polyp and control tissues, and the CM from NP fibroblasts contained, compared to that from fibroblasts derived from NN tissues, significantly greater amounts ofGM-CSF Table 3 shows that the cell number of NN and NP fibroblasts was very similar during the time at which the CM was collected (48 h). Thus, the differences in GM-CSF content in CM that we sec reflect differences in either GM-CSF output per cell or in the proportion of cells releasing GM-CSF at a given time, but not differences in the total number of cells in the culture. We also examined NP and NN fibroblasts for the presence of GMCSF-specific mRNA and, as seen in Figure 5, NP fibroblasts express the GM-CSF signal at a higher level than NN fibroblasts. In the light of these findings, we next examined the effect of a neutralizing monoclonal rat anti-human GM-CSF antibody with the fibroblast-CM and, as shown in Figure 4, this antibody completely inhibited the neutrophilic differentiation induced by NP fibroblast-CM. G-CSF has been shown to induce the emergence of neutrophil colonies in semisolid assay systems (31). However, in our experimental system, an anti-human G-CSF antibody was not capable of abrogating the neutrophilic differentiation induced by fibroblast-CM. Together, these results indicate that fibroblast-derived GMCSF is fully capable of mediating the neutrophilic differentiation of HL-60 cells induced by fibroblast-CM. From these findings, we conclude that human upper airway fibroblasts express the signal and release the gene product GM-CSF and that fibroblast-derived GM-CSF induces neutrophilic differentiation of human hemopoietic progenitors. Our data also show that fibroblasts derived from a disease condition such as nasal polyps elicit, in comparison with NN fibroblasts, an enhanced neutrophilic differentiation, release increased amounts of GM-CSF and express the mRNA GM-CSF at a higher level. Since the method of isolation and culture were identical and these differences were
maintained over a number of passages, the findings suggest that NP fibroblasts might be upregulated (activated) in vivo and that this could define a distinct fibroblast phenotype associated with disease. However, final proof depends upon studies examining the expression and content of GM-CSF directly in whole tissue, or the proportion of cells expressing GM-CSF in tissues from various sources. The observations we have documented here further support the concept that fibroblasts are effector cells themselves. In particular, the enhanced release of GM-CSF by fibroblasts derived from NP tissue has implications beyond neutrophilic differentiation of hemopoietic progenitors. We have recently shown that lung fibroblast-derived GM-CSF supports in vitro human eosinophil survival (16). In addition, GM-CSF is capable of stimulating proliferation of mature cells such as macrophages (32), as well as activation (33) and chemotaxis (34), including the release ofIL-l (35), of neutrophils. The observation that fibroblasts, as well as human upper airway epithelial cells (36), and endothelial cells (37) are capable of releasing cytokines, among them GM-CSF, involved in multiple aspects of the inflammatory process, lends much support to the notion that structural cells contribute to an inductive microenvironment that could significantly modulate the inflammatory response. Furthermore, the observations in NP fibroblasts would suggest that disregulations of structural cells' effector function might be an important determinant in chronic airway inflammation. Acknowledgments: We are much indebted to Drs. Bruce Altrock and Peter Ralph for their generous supply of reagents and for examining the content of specific cytokines in the supernatants. We are also indebted to Drs. D. Hitch and P. Lapp from the ENT Department who provided us with tissues for study. The excellent technical help of Mr. Jerry Schulman and Mrs. Karen Howie is gratefully acknowledged. This work has been funded by the Medical Research Council (Canada) and the Council for Tobacco Research Inc. (USA). Dr. Cox is a Research Fellow of the Canadian Lung Association. Dr. Xaubet is supported by grants from Hospital Clinic i Provincial de Barcelona, Spain: FISS (88/2331) and SEPAR. Dr. Jordana is a Pulmonary Research Fellow of the Parker B. Francis Foundation (USA).
References 1. Connell, 1. T. 1979. Nasal hypersensitivity. In Comprehensive Immunology, vol 6. S. Gupta and R. A. Good, editors. Plenum Publishing Co., New York. 397-416. 2. Cauna, N., K. H. Hinderer, G. W. Manzetti, and E. W. Swanson. 1972. Fine structure of nasal polyps. Ann. Otol. 81:41-58 3. Kakoi , H., and F. Hiraide. 1987. A histological study of formation and growth of nasal polyps. Acta Otolaryngol. (Stockh.] 103: 137-144. 4. Kaushansky, K., N. Lin, and J. Vf. Adamson. 1988. Interleukin-l stimulates fibroblasts to synthesize granulocyte-macrophage and granulocyte colony-stimulating factors. J. Clin. Invest. sj :92-97. 5. Gauldie, J., C. Richards, D. Harnish. P. Landorp. and H. Baumann. 1987. Interferon 132/B-cell stimulatory factor type 2 shares identity with monocyte-derived hepatocyte-stimulating factor and regulates the major acute-phase protein response in liver cells. Proc. Natl. Acad. Sci. USA 84:7251-7255. 6. Strieter, R. M., S. H. Phan, H. J. Showell et al, 1989. Monokine-induced neutrophil chemotactic factor gene expression in human fibroblasts. J. Bioi. Chern. 264:10621-10626. 7. Jordana, M., C. Vancheri, T. Ohtoshi et al, 1989. Hemopoietic function of the microenvironment in chronic airway inflammation. Agents Actions 28:85-95. 8. Denburg, 1. A., M. Jordana, C. Vancheri et at. 1989. Locally derived hemopoietic cytokines for human basophils, mast cells and eosinophils in respiratory disease. In Mast Cell and Basophil Differentiation and Function in Health and Disease. S. J. Galli and K. F. Austen, editors. Raven Press, New York. 59-69. 9. Jordana, M., J. Schulman, C. McSharry et al. 1988. Heterogeneous proliferative characteristics of human adult lung fibroblast lines and clonally derived fibroblasts from control and fibrotic tissue. Am. Rev. Respir. Dis. 137:579-584.
Vancheri, Ohtoshi, Cox et al.: Neutrophilic Differentiation by Fibroblast-derived GM-CSF
10. Fibbe, W. E., J. Van Damme, A. Billiau et al. 1988. Interleukin I induces human bone marrow stromal cells in long term culture to produce granulocyte colony-stimulating factor and macrophage colony-stimulating factor. Blood 71:430-435. II. Ralph, P., M. K. Warren, M. T. Leeetal. 1986. Inducible production of human macrophage growth factor, CSF-1. Blood 68:633-637. 12. Maniatis, T., E. F. Fritsch, and J. Sambrook. 1982. Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. 187-209. 13. Ullrich, A., C. H. Berman, T. J. Dull, A. Gray, and 1. M. Lee. 1984. Isolation of the human insulin-like growth factor I gene using a single synthetic DNA probe. EMBO J. 3:361-364. 14. Wong, G. G., J. S. Witek, P. A. Temple et al. 1985. Human GM-CSF, molecular cloning of the complementary DNA and purification of the natural and recombinant protein. Science 288:810-815. 15. Feinberg, A. P., and B. Vogelstein. 1983. A technique for radiolabelling DNA restriction endonuclease fragments to high specific activity. Anal. Biochem. 132:6-14. 16. Vancheri, C., J. Gauldie, J. Bienenstock et al. 1989. Human lung fibroblast-derived (GM-CSF) mediates eosinophil survival in vitro. Am. J. Respir. Cell Mol. BioI. 1:289-295. 17. Hutt-Taylor, S. R., D. Hamish, M. Richardson, T. Ishizaka, and J. A. Denburg. 1988. Sodium butyrate and a T lymphocyte cell line-derived differentiation factor induce basophilic differentiation of the human promyelocytic leukemia cell line HL-60. Blood 71:209-215. 18. Metcalf, D. 1986. The molecular biology and functions of the granulocytemacrophage colony-stimulating factors. Blood 67:257-267. 19. Kaushansky, K. 1987. The molecular biology of the colony-stimulating factors. Blood Cells 13:3-15. 20. Zucali, J. R., H. E. Broxmeyer, M. A. Gross, and C. A. Dinarello. 1988. Recombinant tumor necrosis factors a and (3 stimulate fibroblasts to produce hemopoietic growth factors in vitro. J. Immunol. 140:840-844. 21. Seelentag, W., 1. J. Mermod, and P. Vasalli. 1989. Interleukin-l and tumor necrosis factor-a additively increase the levels of granulocyte-macrophage and granulocyte colony-stimulating factor (CSF) mRNA in human fibroblasts. Eur. J. Immunol. 19:209-212. 22. Shabo, Y., J. Lotem, M. Rubinstein et al. 1988. The myeloid blood cell differentiation-inducing protein MGI-2A is interleukin-6. Blood 72: 2070-2073. 23. Takeda, K., T. Hosoi, M. Noda, H. Arimura, and K. Konno. 1988. Effect of fibroblast-derived differentiation inducing factor on the differentiation of human monocytoid and myeloid leukemia lines. Biochem. Biophys. Res. Commun. 155:24-31. 24. Larsen, C. G., A. O. Anderson, E. Appella, J. J. Oppenheim, and K. Matsushima. 1989. The neutrophil-activating protein (NAP-I) is also chemo-
17
tactic for T-Iymphocytes. Science 243:1464-1466. 25. Otsuka, H., J. Dolovich, D. Befus, J. Bienenstock, and J. Denburg. 1986. Peripheral blood basophils, basophil progenitors, and nasal metachromatic cells in allergic rhinitis. Am. Rev. Respir. Dis. 133:757-762. 26. Otsuka, H., J. Dolovich, A. D. Befus etal. 1986. Basophil cell progenitors, nasal metachromatic cells, and peripheral blood basophils in ragweedallergic patients. J. Allergy Clin. Immunol. 78:365-371. 27. Ohnishi, M., J. Ruhno, J. Bienenstock, J. Dolovich, and J. A. Denburg. 1989. Hemopoietic growth factor production by human nasal polyp epithelial scrapings: kinetics, cell source and relationship to clinical status. J. Allergy Clin. Immunol. 83: 1091-1100. 28. Collins, S. J., F. W. Ruscetti, R. E. Gallagher, andR. Gallo. 1979. Normal functional characteristics of cultured human promyelocytic leukemia cells (HL-60) after induction of differentiation of dimethylsulfoxide. J. Exp. Med. 149:969-974. 29. Rovera, G., D. Santoli, and C. Damsky. 1979. Human promyelocytic leukemia cells in culture differentiate into macrophage-like cells when treated with a phorbol diester. Proc. Natl. Acad. Sci. USA 76:2779-2783. 30. Fischkoff, S. A., A. Pollak, G. J. Gleich, J. R. Testa, S. Misawa, and T. J. Reber. 1984. Eosinophilic differentiation of the human promyelocytic leukemia cell line, HL-60. J. Exp. Med. 160:179-196. 31. Nagata, S., M. Tsuchiya, S. Asano et al. 1986. Molecular cloning and expression of cDNA for human granulocyte colony-stimulating factor. Nature 319:415-418. 32. Akagawa, K. S., K. Kamoshita, and T. Tokunaga. 1988. Effect of granulocyte-macrophage colony-stimulating factor and colony-stimulating factor-Ion the proliferation and differentiation of murine alveolar macrophages. J. Immunol. 141:3383-3390. 33. Kapp, A., G. Zeck-Kapp, M. Danner, and T. Luger. 1988. Human granulocyte-macrophage colony stimulating factor: an effective direct activator of human polymorphonuclear neutrophilic granulocytes. J. Invest. Dermatol. 91:49-55. 34. Ming Wang, J., Z. Guo Chen, S. Colella et al. 1988. Chemotactic activity of recombinant human granulocyte colony-stimulating factor. Blood 72:1456-1460. 35. Lindenman, A., D. Reidel, O. Wolfgang et al. 1988. Granulocytemacrophage colony-stimulating factor induces interleukin-l production by human polymorphonuclear neutrophils. J. Immunol. 140:837-839. 36. Ohtoshi, T., C. Vancheri, G. Cox et al. 1991. Monocyte-macrophage differentiation induced by human upper airway epithelial cells. Am. J. Respir. Cell Moi. Bioi. In press. 37. Bagby, G. C., E. McCall, K. A. Bergstrom, and D. A. Burger. 1983. A monokine regulates CSA production by vascular endothelial cells. Blood 62:663-668.
We have established primary lines of fibroblasts from nasal polyp (NP) tissues as well as from normal nasal (NN) mucosa and have examined the ability of these cells to release hormone-like peptide messenger molecules (cytokines). Our results show that human upper airway fibroblasts release granulocyte/macrophage colony-stimulating factor (GM-CSF), granulocyte-CSF (G-CSF), and macrophage-CSF (M-CSF) in vitro. We also show that fibroblasts derived from NP tissue express the gene for GM-CSF at a higher level, and release the GM-CSF product in greater amounts, than NN fibroblasts. In addition, we have examined the ability of these fibroblasts and their conditioned medium (CM) to induce differentiation of human hemopoietic progenitor cells. After 7 d, cultures of these cells in RPMI-lO% fetal bovine serum contained 5 ± 2.5% (mean ± SD) neutrophils. In contrast, culture of progenitor cells with fibroblasts resulted in significantly greater neutrophilic differentiation (18 ± 4 %). Culture in fibroblast-CM induced a similar degree of differentiation, and fibroblast-CM from NP fibroblasts elicited greater differentiation compared to CM from NN fibroblasts (17.5 ± 3 versus 12 ± 3%). The neutrophilic differentiation induced by fibroblast-CM can be fully inhibited by preincubating this CM with a monoclonal neutralizing antibody to human GM-CSF. Thus, our results demonstrate: (1) the ability of human upper airway fibroblasts to release GM-, G-, and M-CSF in vitro; (2) that fibroblasts derived from NP tissues express the gene and release the product GM-CSF at greater levels compared to NNfibroblasts; and (3) that fibroblast-derived GM-CSF causes neutrophilic differentiation of human hemopoietic progenitors.
Nasal polyps, a condition of unknown etiology occurring as often in patients with atopy as in the general population, are grape-like formations arising from the posterior nasal and sphenoid sinus mucosae (1). Nasal polyps are characterized by epithelial cell and varying degrees of fibroblast hyperplasia, as well as by the accumulation, in variable numbers, of inflammatory cells, including neutrophils, eosinophils, metachromatic cells, and lymphocytes (1-3). These inflammatory cells can release a variety of enzymes and mediators thought
(Received in original form August 28, 1989 and in revised form June 11, 1990) Address correspondence to: Manel Jordana, M.D., McMaster University, Department of Pathology, Room 4HI7-21, 1200 Main Street West, Hamilton, Ontario L8N 3Z5, Canada. Dr. Vancheri's present address is: Istituto di Malattie dell'Aparato Respiratorio e Tisiologia, Via Passo Gravina 187, 4500 Catania, Italy. Abbreviations: conditioned medium, CM; fetal bovine serum, FBS; granulocyte colony-stimulating factor, G-CSF; granulocyte/macrophage colonystimulating factor, GM-CSF; macrophage colony-stimulatingfactor, M-CSF; 2-mercaptoethanol, 2-ME; normal nasal, NN; nasal polyp, NP; nonadherent mononuclear cells, NAMe. Am. J. Respir. Cell Mol. BioI. Vol. 4. pp. 11-17, 1991
to be involved in the expression of symptoms associated with this condition. Thus, the mechanisms modulating the accumulation and/or activation of those cells in the nasal mucosa are central to the pathogenesis of nasal polyps. Traditionally, the role ascribed to the stromal cell population has been circumscribed to reparative processes. However, it is increasingly apparent that stromal cells such as fibroblasts are also capable of releasing a number of hormone-like messenger peptides (cytokines) and are thus effector cells themselves (4-6). Our laboratory has previously proposed the hypothesis of "microenvironmental control of airway inflammation" to indicate that key aspects of the inflammatory process such as proliferation and activation of mature inflammatory cells as well as differentiation of inflammatory cell progenitors might be modulated at the tissue level by signals originating, at least in part, from the tissue structural cells (7, 8). However, the precise interactions between cells and cytokines involved in this modulation remain to be clarified. We have examined the contribution of human upper airway fibroblasts, derived from nasal polyp (NP) and normal nasal (NN) tissue, to the differentiation of human hemopoietic progenitors. We show here that conditioned media
12
AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 4 1991
from NP fibroblasts induce neutrophilic differentiation of these progenitors to a significantly greater degree than that elicited by NN fibroblasts. We also show that these conditioned media contain granulocyte/macrophage colony-stimulating factor (GM-CSF), granulocyte-CSF (G-CSF), and macrophage-CSF (M-CSF), and that NP fibroblasts release, compared to NN fibroblasts, significantly greater amounts of GM-CSF. Finally, we show that the neutrophilic differentiation induced by fibroblast-conditioned medium can be fully abrogated by preincubating this conditioned medium with a neutralizing monoclonal rat anti-human GM-CSF antibody.
Materials and Methods Cells HL-60 cells were obtained from the American Type Culture Collection (Rockville, MD). Experiments were carried out with cells between the thirtieth and fiftieth passage. Routinely, these cells were subcloned twice weekly to a density of 2.5 x 105 cells/ml with RPMI 1640 (Rosswell Park Memorial Institute) supplemented with 10% heat-inactivated fetal bovine serum (FBS) (GIBCO, Grand Island, NY), 1% penicillin/streptomycin (GIBCO), and 25 mmol/ liter Hepes buffer (Boehringer Mannheim Canada Ltd., Dorval, Quebec, Canada). Cells were kept in a humidified atmosphere at 5 % CO2 and 37° C. Primary lines of human fibroblasts were established in our laboratory using an outgrowth from an explant method as previously described (9). After removal of the explants, fibroblasts were expanded and frozen in liquid nitrogen for future use. In the experiments described here, fibroblasts were always used at a passage earlier than the tenth. Tissue Source Characterization We established nine primary fibroblast lines from NP tissue removed at the time of polypectomy. Four patients had tested positive to one or more skin tests performed in the week prior to the surgical procedure. Four patients tested negative to a battery of skin tests, including 12 antigens, and one patient refused to be skin-tested. Control nasal tissue refers to nasal mucosa tissue from five patients who did not have a history of nasal allergy, had negative skin tests, and did not have nasal polyps. Tissue was obtained at the time of surgery for mechanical obstruction or cosmetic correction. These studies conformed to guidelines for human experimentation and were approved by the Chedoke-McMaster Hospital Ethics Committee. In Vitro Proliferative Characteristics of Nasal Fibroblasts Fibroblasts from confluent dishes were trypsinized and counted. A total of 2 x 105 fibroblasts were then delivered in 4 ml of RPMI-supplemented medium into each well of 6-well plates (Falcon; Becton Dickinson, Oxnard, CA). The cells were incubated for various time intervals at 37° C and 5% CO,. At the end of the incubation period, that is, at days 1, 3, 8, and 11, the cells were trypsinized and counted with a hemocytometer. Two wells were prepared for each data point, and two counts were always performed per well. Cell viability as assessed by trypan blue exclusion was 95 % or greater in all experiments.
Preparation of Conditioned Medium (CM) Fibroblast-CM was obtained from 2 X 106 fibroblasts seeded in 35-mm ¢ wells of 6-well plates (Falcon; Becton Dickinson) with 4 ml of RPMI-supplemented medium (5 x 105/ml). In most experiments, fibroblast-CM was used immediately upon harvesting. In some instances, the CM was aliquoted and frozen at -20° C for future use. Experimental Cultures Fibroblasts, 1.5 X 106 cells, were seeded in 100-mm tissue culture dishes (Corning, Corning, NY) and allowed to adhere overnight. The next day, the medium was washed off and 5 x 105 HL-60 cells in 10 ml of supplemented RPMI were added. After 3 d, 5 ml of fresh medium was added to each dish. In another series of experiments, HL-60 cells were cultured with fibroblast-CM obtained as described above. In this protocol, cells were recovered after 3 d of culture, spun down at 1,000 rpm, and resuspended in either fresh CM or supplemented RPMI. Unless otherwise indicated, fibroblast-CM was used at 50% concentration. Following the same protocol, HL-60 cells were in some instances cultured with various concentrations of recombinant human cytokines. In addition, in some specific experiments in which we examined the effect of a rat anti-human GM-CSF and a mouse anti-human G-CSF antibody, the fibroblast-CM was preincubated with the antibody for 2 h at 37° C before being added to the HL-60 cells. All experiments were terminated at day 7 of culture. At that time, HL-60 cells were recovered from each dish after at least two washes, spun down at 1,000 rpm for 10 min, and resuspended in 2 ml of medium. The cells were counted with a hemocytometer, and viability was assessed by using 0.1% trypan blue exclusion (GIBCO). Morphology Cytocentrifuge preparations were made using a ShandonSouthern Cytospin (Johns Scientific, Toronto, Ontario, Canada). In all experiments, smears were stained with the DiffQuik'" modification of the May-Giemsa technique (American Scientific Products, McGraw Park, IL). Assay of Colony-forming Units Mononuclear cells were isolated from heparinized peripheral blood using Ficoll-Hypaque density sedimentation (specific gravity, l(177) and washed twice with McCoy's 5A medium (GIBCO), supplemented with 15% FBS, 1% penicillin-streptomycin, and 5 X 10-5 2-mercaptoethanol (2-ME). Nonadherent mononuclear cells (NAMC) were obtained by removal of adherent cells after incubation in plastic tissue culture flasks (Falcon ~024; Becton Dickinson) for 2 h at 37° C. A total of 106 NAMC per 35-mm culture dish (Falcon 3001) were cultured in 0.9% methylcellulose containing Iscoves modified Dulbecco's medium (GIBCO) supplemented with 20% FBS and 5 x 10-5 M of 2-ME with or without fibroblast-CM. Granulocyte colonies of ~ 40 cells were counted under an inverted microscope (Olympus, Tokyo,Japan) at day 14. Reagents A rat anti-human monoclonal neutralizing anti-GM-CSF antibody was generously provided by Dr. John S. Abrams
Vancheri, Ohtoshi, Cox et al.: Neutrophilic Differentiation by Fibroblast-derived GM-CSF
(DNAX Research Institute, Palo Alto, CA). In a colony assay, 1:100 dilution of antibody is capable of neutralizing up to 5 Dlml recombinant GM-CSF (10 pg/ml). A mouse antihuman monoclonal neutralizing anti-G-CSF antibody was kindly donated by Dr. Bruce W. Altrock (Amgen Corp., Thousand Oaks, CA); the neutralizing activity of this antibody in a CFV-GM bone marrow assay is 100,000 neutralizing Uzrnl. Immunoassay of GM-, G-, and M-CSF in Fibroblast-CM The content of human GM-CSF in fibroblast-CM was detected by means of a specific immunoassay performed by Dr. John S. Abrams. The limit of sensitivity of this assay is 30 pg/rnl. Human G-CSF content was quantitated in a monospecific sandwich immunoassay by Dr. Bruce W. Altrock, and the limit of sensitivity of this assay is 0.2 ng/ml (10). Human M-CSF content was assessed using a radioimmunoassay by Dr. Peter Ralph (Cetus Corp., Emeryville, CA); the limit of sensitivity of this assay is 0.8 ng/ml (11). RNA Preparation and Analysis Fibroblast RNA was isolated following essentially the same procedure as previously described (12). Briefly, fibroblasts were lysed with 5 M guanidium isothiocyanate and total RNA was recovered by ultracentrifugation through a 5.7 M CsCI gradient, resuspended in diethylpyrocarbonate-treated water, and then denatured with formaldehyde. The RNA was then transferred onto a nitrocellulose blot and prehybridized at 42° C for4t06hin4x SSC (IX SSC = 0.15 mM NaCl/15 rnM Na citrate, pH 7.0), 2x Denhardt's, 50 mM Na phosphate, pH 7.0,20%formamide, 10% dextran sulphate, 1 mM Na pyrophosphate, 50 ttg/ml ATP, and 50 ttg/ml sonicated salmon sperm DNA (13). GM-CSF mRNA was detected using a human cDNA probe kindly provided by Dr. G. Wong (14). The probe was labeled to 109 cpm/ug using hexanucleotide primers and 32P-dCTP (15). A total of 107 counts per minute of probe were added to 10 ml of prehybridization solution, and the blot was hybridized at 42° C overnight. Afterwards, the blot was washed in 0.2 X SSC/OJ% NaDodS04 , dried, and exposed for 48 h to Kodak X-Omat X-ray film with intensifying screens. Statistical Analysis Statistical comparisons were made using a nonpaired Student's t test.
25
Fibroblast-induced Neutrophilic Differentiation HL-60 cell cultures in RPMI containing 10% FBS showed a relatively small amount of spontaneous neutrophilic differentiation (5 ± 2.5%). In contrast, a statistically significant greater number of neutrophils was found (18 ± 4 %) when the HL-60 cells were cocultured with NP-derived fibroblast monolayers (Figure 1). A similar result was obtained by culturing HL-60 cells with NP fibroblast-CM, indicating that a soluble factor released by the fibroblasts was responsible for the effect. As shown in Figure 2, CM derived from different cultures of NN fibroblasts induced neutrophilic differentiation above the level of spontaneous HL60 differentiation but significantly less than NP fibroblast-
T
20
-ET Z ~
15
10 5
CONTROL 9) ( n
=
NP (n
= 3)
Figure 1. Fibroblast-induced neutrophilic differentiation of HL-60 cells. The empty bar depicts the neutrophilic differentiation (percentage of neutrophils in the differential count) of HL-60 cells cultured in RPMI-lO% FBS and harvested at day 7. Results represent mean ± SD of nine experiments. The solid bar depicts the neutrophilic differentiation of HL-60 cells cocultured with a monolayer of nasal polyp (NP) fibroblasts. Results represent mean ± SD of three experiments in which three different primary fibroblast lines were used.
CM (12 ± 3% versus 17.5 ± 3%, respectively). The effect offibroblast-CM on neutrophilic differentiation was dose dependent (Figure 3). We carried out additional experiments in which the effect of fibroblast-CM, specifically NP-CM, on HL-60 cells was evaluated at day 10 of culture as opposed to day 7. We found that there were no significant differences between these two time points in regard to the extent of neutrophilic differentiation (data not shown). Finally, Table 1 shows the results of a detailed morphologic characterization of HL-60 cells cultured with CM from NP fibroblasts. The data, which are combined from four experiments using four different NP-
25
20
-ET
15
Z
~
Results
13
10
S.D.
5
0 CONTROL ( n
= 9)
NN (n
= 5)
NP (n
= 4)
Figure 2. Fibroblast-conditioned medium (CM)-induced neutrophilic differentiation of HL-60 cells. The empty bar shows, as in Figure 1, the neutrophilic differentiation of HL-60 cells cultured in RPMI-lO% FBS. The solid bars depict the neutrophilic differentiation ofHL-60 cells cultured with fibroblast-CM from normal nasal (NN) or nasal polyp (NP) fibroblasts and harvested at day 7. Results represent mean ± SD of five and four experiments, respectively. In each experiment, a different primary fibroblast line was used to generate the CM.
14
AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 4 1991
TABLE 2
20
Cytokine content offibroblast-conditioned medium (Fib-CM) 15
Fib-CM
GM-CSF*
G-CSFt
M-CSF+
(pg/ml)
(ng/mll
(ng/mll
-GZ
10
Normal nasal tissue
194 (n
Nasal polyp tissue
5
663 (n
5
10
25
50
±
69
= 5)
± 258 = 12)
1.3
± 0.9
(n
= 5)
4.8 (n
± 0.8 = 2)
0.6
± 0.3
5.0
± 0.0
(n
= 5)
(n
= 2)
* Mean ± SEM; sensitivity of the assay is 30 pg/m!. t Mean ± SEM; sensitivity of the assay is 0.2 ng/ml. t Mean ± SEM; sensitivity of the assay is 0.8 ng/m!.
90
Fib-CM Concentration ("10)
Figure 3. Dose response of fibroblast-conditioned medium (CM)induced neutrophilic differentiation of HL-60 cells. The empty bar represents differentiation of cells cultured in RPMI alone. Results depict mean + SD of three experiments using CM from different nasal polyp fibroblast cultures.
CM, illustrate that the majority of the induced cells (66.2 %) were either band or segmented, that is, mature neutrophils. Immunoassay of GM-, G-, and M-CSF in Fibroblast-CM Table 2 shows the human CSF content in CM from a number of individual primary lines of fibroblasts from different tissue sources. Fibroblast-CM contained detectable amounts of G- and M-CSF, but no significant differences related to tissue source (NN- versus NP-derived fibroblasts) were seen. Fibroblast-CM also contained GM-CSF, and the amounts present in conditioned media from NP fibroblasts were significantly greater than those present in CM from NN fibroblasts (663 ± 255 versus 194 ± 70 pg/ml [mean ± SEM)). Proliferative Characteristics of Nasal Fibroblasts Table 3 shows the growth characteristics of three primary lines of NN fibroblasts and of three primary lines of NP fibroblasts. It is apparent that the number of cells in culture of either type of fibroblasts is.very similar at days 1, 3, and 8. The data also show that the cell number of NN fibroblasts reaches a plateau, in contrast to NP fibroblasts, between days 8 and 11; in fact, the difference in cell number between both types of fibroblasts is statistically significant at day 11.
TABLE 1
Effect of Specific Neutralizing Antibodies on the Neutrophilic Differentiation Induced by Fibroblast-CM Figure 4 shows the effect of a neutralizing rat anti-human monoclonal neutralizing antibody on the neutrophilic differentiation induced by fibroblast-CM. The dose range of the antibody used for these experiments was similar to that which fully abrogates the effect of fibroblast-CM on eosinophil survival in vitro (16). It can be seen that this anti-GlvlCSF antibody inhibits the neutrophilic differentiation induced by fibroblast-CM in a dose-dependent fashion with full inhibition at a 1:50 dilution. Under the same experimental conditions, an anti-G-CSF antibody at concentrations up to 1:100 did not inhibit fibroblast-CM-induced neutrophilic differentiation of HL-60 cells. Effect of Fibroblast-CM on Colony Growth Table 4 shows the effect of a number of fibroblast-CM on colony growth in a methylcellulose assay system. The positive control in these experiments was CM from the Mo human T lymphocyte cell line, which contains GM-CSF (17). Mo-CM at 5 % concentration induced the emergence of about 24 colonies of which 4.5 were GM colonies. CM from five different primary lines of NP fibroblasts, used at a concentration of 15 %, induced an average of 32 colonies of which more than 12 were GM colonies. Rat monoclonal antiGM-CSF antibody at a concentration of 1:50 (vol/vol) inhibited colony formation induced by fibroblast-CM (Table 5). RNA Analysis Figure 5 illustrates that a low-intensity signal is seen in lanes where total RNA obtained from cultures of NN fibroblasts was loaded. The intensity of the signal is stronger in those
Fibroblast-induced neutrophilic differentiation of HL-60 cells Stage
Immature neutrophils Myelocytes Metamyelocytes Mature neutrophils Total
%
Morphology
Band Segmented
± 0.5* ± 1.2 7.0 ± 2.3 5.5 ± 1.0 18.8 ± 3.1 1.1 5.2
TABLE 3
Relative %
(6.1 (27.7 (36.3 (29.9
Proliferative characteristics of nasal fibroblasts in vitro
± 3.0)t ± 2.6) ± 7.0) ± 7~4)
(100%)
* Results are expressed as mean ± SD of four experiments in which four different nasal polyp-conditioned media were used. t The numbers in parentheses represent the proportion of cells at each stage to the total number of neutrophils.
Days in Culture Day I
NN-Fib* (n = 3) 2.1 NP-Fibt (n = 3) 2.1
* NN-Fib
± 0.2t ± 0.3
Day 3
6.8 5.5
± 2.6 ± 1.3
Day 8
14.1 11.6
± 3.9 ± 4.6
= normal nasal fibroblasts. t NP-Fib = nasal polyp fibroblasts. t Cell number (X 10'). Results expressed as mean ± SD.
Day II
13.8 18.0
± 2.9 ± 4.7
Vancheri, Ohtoshi, Cox et al.: Neutrophilic Differentiation by Fibroblast-derived GM-CSF
15
TABLE 5
15
Effect of anti-GM-CSF antibody on colony growth induced by fibroblast-conditioned medium (Fib-CM)*
S.D.
GM Colonies Total Colonies
10
Negative control! (n = 2) Positive control! (n = 2) Fib-CM§ (n = 3) Fib-CM§ + anti-GM-CSF 1:50 (n
-ez at 5
* Results expressed
=
1.0 2.5 7.5 3) 0.5
± ± ± ±
0.0 0.5 2.1 0.3
3.5 12.5 22.3 0.8
± ± ± ±
0.5 1.5 4.7 0.2
as mean ± SEM.
t Medium alone. t CM at a 5 % concentration from the Mo human T lymphocyte cell line, CONTROL FIb-eM
which contains GM-CSF. § Fib-CM was used at a 15% concentration.
FIb-eM + anti GM-CSF ab
1:500
1:200
1:50
Figure 4. Effect of a neutralizing rat anti-human GM-CSF antibody on the neutrophilic differentiation of HL-60 cells induced by fibroblast-conditioned medium (CM). The empty bar shows the neutrophilic differentiation of HL-60 cells cultured in RPMI-lO% FBS. The solid bar depicts the neutrophilic differentiation induced by fibroblast-CM. The hatched bars illustrate the neutrophilic differentiation induced by this fibroblast-CM when preincubated with the anti-GM-CSF antibody.
lanes where RNA recovered from NP fibroblasts was loaded. In all instances, the RNA was obtained from the same number of cultured fibroblasts, plated at the same cell density, and incubated for 48 h.
Discussion There is increasing evidence that the tissue microenvironment represents a compartment where inflammatory and immune cell function is modulated. One component of this activity involves the release of soluble signals (cytokines) by the structural cells. These cytokines can interact with both immature and mature inflammatory cells and thereby modulate their differentiation, proliferation, and activation, thus participating in the regulation of the inflammatory response (18, 19). In regard to the effector capability of fibroblasts, Zucali and associates have shown that supernatants from recombinant human (rh) interleukin-l (IL-l)-stimulated human lung fibroblasts contain GM-CSF activity (20). Kaushansky and colleagues have shown that quiescent dermal fibroblasts constitutively release low amounts of G-, GM-, and M-CSF in culture and that rh-IL-l markedly enhance the transcription and release of both G- and GM-CSF (4). A similar observation has been made by Seelentag and cowork-
TABLE 4
Effect ofconditioned medium from nasal polyp fibroblast (Fib-CM) on colony growth* GM Colonies
Total Colonies
1.2 ± 0.2 4.5 ± 1.6 4.3 ± 0.5
3.2 ± 0.8 23.8 ± 7.6 15.2 ± 4.0
Negative control! (n = 3) Positive controlt (n = 3) Fib-CM§ (n = 5)
* Results
expressed as mean ± SEM.
t Medium alone. t CM at a 5% concentration from the Mo human T lymphocyte cell line,
which contains GM-CSF. § Fib-CM was used at a 5 % concentration.
ers; they have also shown that IL-l and tumor necrosis factor (TNF) have an additive stimulatory effect on G- and GMCSF production by skin fibroblasts (21). In addition to the colony-stimulating factors, fibroblasts express and release the gene product interleukin-6 (IL-6) (5), which has now recognized hemopoietic functions (22, 23) as well as a molecule, now referred to as interleukin-8 (IL-8) (24), which is chemotactic for neutrophils (6) and lymphocytes (24). With regard to upper airway inflammation, we have previously shown that a potential mechanism for inflammatory cell accumulation is the ingress and local differentiation of inflammatory cell progenitors (25-27). In the studies we report here, we sought to investigate whether human upper airway fibroblasts release cytokines that could stimulate the differentiation of hemopoietic progenitors and whether such activities were related to disease expression. To address this question, we examined the effect of fibroblasts and fibro-
5
2.5
JJRNA
NP NP NP NN NN NC Figure 5. Detection of GM-CSF-specific mRNA (slot-blot) in human upper airway fibroblasts. The first three lanes show the level ofGM-CSF expression of tot a! RNA extracted from three different primary lines of nasal polyp (NP) fibroblasts. Lanes 4 and 5 were loaded with RNA from two different lines of normal nasal (NN) fibroblasts. NC represents the negative control, which was calf liver.
16
AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 4 1991
blast-CM from both NP and NN tissues on the differentiation of the human HL-60 myeloid leukemia cell line, which is capable of differentiating into mature cells of different lineages, depending on the inducing agent (28-30). Thus, this assay system is useful to study biochemical and molecular events associated with differentiation of progenitor cells into functionally mature inflammatory cells. HL-60 cells differentiated toward neutrophils when cocultured with a monolayer of human upper airway fibroblasts (Figure I). Fibroblast-CM induced a similar degree of neutrophilic differentiation of HL-60 cells, indicating that a soluble factor present in the CM was responsible for the effect, and the differentiating activity induced by CM from NP-derived fibroblasts was significantly greater than that induced by CM from fibroblasts obtained from NN tissues (Figure 2). This difference appeared permanent in that it was maintained in fibroblasts that had gone through a number of passages in vitro. The neutrophilic differentiation induced by fibroblast-CM was not specific for the HL-60 target because, as shown in Table 3, this CM induced the emergence of GM colonies from a preparation of nonadherent peripheral blood mononuclear cells containing hemopoietic progenitors. Since the type of differentiation seen suggested the presence of colony-stimulating factors in the CM, we then determined the content of such factors in the CM using specific radioimmunoassays. As shown in Table 2, G-, GM-, and M-CSF were detected in fibroblast-CM from both polyp and control tissues, and the CM from NP fibroblasts contained, compared to that from fibroblasts derived from NN tissues, significantly greater amounts ofGM-CSF Table 3 shows that the cell number of NN and NP fibroblasts was very similar during the time at which the CM was collected (48 h). Thus, the differences in GM-CSF content in CM that we sec reflect differences in either GM-CSF output per cell or in the proportion of cells releasing GM-CSF at a given time, but not differences in the total number of cells in the culture. We also examined NP and NN fibroblasts for the presence of GMCSF-specific mRNA and, as seen in Figure 5, NP fibroblasts express the GM-CSF signal at a higher level than NN fibroblasts. In the light of these findings, we next examined the effect of a neutralizing monoclonal rat anti-human GM-CSF antibody with the fibroblast-CM and, as shown in Figure 4, this antibody completely inhibited the neutrophilic differentiation induced by NP fibroblast-CM. G-CSF has been shown to induce the emergence of neutrophil colonies in semisolid assay systems (31). However, in our experimental system, an anti-human G-CSF antibody was not capable of abrogating the neutrophilic differentiation induced by fibroblast-CM. Together, these results indicate that fibroblast-derived GMCSF is fully capable of mediating the neutrophilic differentiation of HL-60 cells induced by fibroblast-CM. From these findings, we conclude that human upper airway fibroblasts express the signal and release the gene product GM-CSF and that fibroblast-derived GM-CSF induces neutrophilic differentiation of human hemopoietic progenitors. Our data also show that fibroblasts derived from a disease condition such as nasal polyps elicit, in comparison with NN fibroblasts, an enhanced neutrophilic differentiation, release increased amounts of GM-CSF and express the mRNA GM-CSF at a higher level. Since the method of isolation and culture were identical and these differences were
maintained over a number of passages, the findings suggest that NP fibroblasts might be upregulated (activated) in vivo and that this could define a distinct fibroblast phenotype associated with disease. However, final proof depends upon studies examining the expression and content of GM-CSF directly in whole tissue, or the proportion of cells expressing GM-CSF in tissues from various sources. The observations we have documented here further support the concept that fibroblasts are effector cells themselves. In particular, the enhanced release of GM-CSF by fibroblasts derived from NP tissue has implications beyond neutrophilic differentiation of hemopoietic progenitors. We have recently shown that lung fibroblast-derived GM-CSF supports in vitro human eosinophil survival (16). In addition, GM-CSF is capable of stimulating proliferation of mature cells such as macrophages (32), as well as activation (33) and chemotaxis (34), including the release ofIL-l (35), of neutrophils. The observation that fibroblasts, as well as human upper airway epithelial cells (36), and endothelial cells (37) are capable of releasing cytokines, among them GM-CSF, involved in multiple aspects of the inflammatory process, lends much support to the notion that structural cells contribute to an inductive microenvironment that could significantly modulate the inflammatory response. Furthermore, the observations in NP fibroblasts would suggest that disregulations of structural cells' effector function might be an important determinant in chronic airway inflammation. Acknowledgments: We are much indebted to Drs. Bruce Altrock and Peter Ralph for their generous supply of reagents and for examining the content of specific cytokines in the supernatants. We are also indebted to Drs. D. Hitch and P. Lapp from the ENT Department who provided us with tissues for study. The excellent technical help of Mr. Jerry Schulman and Mrs. Karen Howie is gratefully acknowledged. This work has been funded by the Medical Research Council (Canada) and the Council for Tobacco Research Inc. (USA). Dr. Cox is a Research Fellow of the Canadian Lung Association. Dr. Xaubet is supported by grants from Hospital Clinic i Provincial de Barcelona, Spain: FISS (88/2331) and SEPAR. Dr. Jordana is a Pulmonary Research Fellow of the Parker B. Francis Foundation (USA).
References 1. Connell, 1. T. 1979. Nasal hypersensitivity. In Comprehensive Immunology, vol 6. S. Gupta and R. A. Good, editors. Plenum Publishing Co., New York. 397-416. 2. Cauna, N., K. H. Hinderer, G. W. Manzetti, and E. W. Swanson. 1972. Fine structure of nasal polyps. Ann. Otol. 81:41-58 3. Kakoi , H., and F. Hiraide. 1987. A histological study of formation and growth of nasal polyps. Acta Otolaryngol. (Stockh.] 103: 137-144. 4. Kaushansky, K., N. Lin, and J. Vf. Adamson. 1988. Interleukin-l stimulates fibroblasts to synthesize granulocyte-macrophage and granulocyte colony-stimulating factors. J. Clin. Invest. sj :92-97. 5. Gauldie, J., C. Richards, D. Harnish. P. Landorp. and H. Baumann. 1987. Interferon 132/B-cell stimulatory factor type 2 shares identity with monocyte-derived hepatocyte-stimulating factor and regulates the major acute-phase protein response in liver cells. Proc. Natl. Acad. Sci. USA 84:7251-7255. 6. Strieter, R. M., S. H. Phan, H. J. Showell et al, 1989. Monokine-induced neutrophil chemotactic factor gene expression in human fibroblasts. J. Bioi. Chern. 264:10621-10626. 7. Jordana, M., C. Vancheri, T. Ohtoshi et al, 1989. Hemopoietic function of the microenvironment in chronic airway inflammation. Agents Actions 28:85-95. 8. Denburg, 1. A., M. Jordana, C. Vancheri et at. 1989. Locally derived hemopoietic cytokines for human basophils, mast cells and eosinophils in respiratory disease. In Mast Cell and Basophil Differentiation and Function in Health and Disease. S. J. Galli and K. F. Austen, editors. Raven Press, New York. 59-69. 9. Jordana, M., J. Schulman, C. McSharry et al. 1988. Heterogeneous proliferative characteristics of human adult lung fibroblast lines and clonally derived fibroblasts from control and fibrotic tissue. Am. Rev. Respir. Dis. 137:579-584.
Vancheri, Ohtoshi, Cox et al.: Neutrophilic Differentiation by Fibroblast-derived GM-CSF
10. Fibbe, W. E., J. Van Damme, A. Billiau et al. 1988. Interleukin I induces human bone marrow stromal cells in long term culture to produce granulocyte colony-stimulating factor and macrophage colony-stimulating factor. Blood 71:430-435. II. Ralph, P., M. K. Warren, M. T. Leeetal. 1986. Inducible production of human macrophage growth factor, CSF-1. Blood 68:633-637. 12. Maniatis, T., E. F. Fritsch, and J. Sambrook. 1982. Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. 187-209. 13. Ullrich, A., C. H. Berman, T. J. Dull, A. Gray, and 1. M. Lee. 1984. Isolation of the human insulin-like growth factor I gene using a single synthetic DNA probe. EMBO J. 3:361-364. 14. Wong, G. G., J. S. Witek, P. A. Temple et al. 1985. Human GM-CSF, molecular cloning of the complementary DNA and purification of the natural and recombinant protein. Science 288:810-815. 15. Feinberg, A. P., and B. Vogelstein. 1983. A technique for radiolabelling DNA restriction endonuclease fragments to high specific activity. Anal. Biochem. 132:6-14. 16. Vancheri, C., J. Gauldie, J. Bienenstock et al. 1989. Human lung fibroblast-derived (GM-CSF) mediates eosinophil survival in vitro. Am. J. Respir. Cell Mol. BioI. 1:289-295. 17. Hutt-Taylor, S. R., D. Hamish, M. Richardson, T. Ishizaka, and J. A. Denburg. 1988. Sodium butyrate and a T lymphocyte cell line-derived differentiation factor induce basophilic differentiation of the human promyelocytic leukemia cell line HL-60. Blood 71:209-215. 18. Metcalf, D. 1986. The molecular biology and functions of the granulocytemacrophage colony-stimulating factors. Blood 67:257-267. 19. Kaushansky, K. 1987. The molecular biology of the colony-stimulating factors. Blood Cells 13:3-15. 20. Zucali, J. R., H. E. Broxmeyer, M. A. Gross, and C. A. Dinarello. 1988. Recombinant tumor necrosis factors a and (3 stimulate fibroblasts to produce hemopoietic growth factors in vitro. J. Immunol. 140:840-844. 21. Seelentag, W., 1. J. Mermod, and P. Vasalli. 1989. Interleukin-l and tumor necrosis factor-a additively increase the levels of granulocyte-macrophage and granulocyte colony-stimulating factor (CSF) mRNA in human fibroblasts. Eur. J. Immunol. 19:209-212. 22. Shabo, Y., J. Lotem, M. Rubinstein et al. 1988. The myeloid blood cell differentiation-inducing protein MGI-2A is interleukin-6. Blood 72: 2070-2073. 23. Takeda, K., T. Hosoi, M. Noda, H. Arimura, and K. Konno. 1988. Effect of fibroblast-derived differentiation inducing factor on the differentiation of human monocytoid and myeloid leukemia lines. Biochem. Biophys. Res. Commun. 155:24-31. 24. Larsen, C. G., A. O. Anderson, E. Appella, J. J. Oppenheim, and K. Matsushima. 1989. The neutrophil-activating protein (NAP-I) is also chemo-
17
tactic for T-Iymphocytes. Science 243:1464-1466. 25. Otsuka, H., J. Dolovich, D. Befus, J. Bienenstock, and J. Denburg. 1986. Peripheral blood basophils, basophil progenitors, and nasal metachromatic cells in allergic rhinitis. Am. Rev. Respir. Dis. 133:757-762. 26. Otsuka, H., J. Dolovich, A. D. Befus etal. 1986. Basophil cell progenitors, nasal metachromatic cells, and peripheral blood basophils in ragweedallergic patients. J. Allergy Clin. Immunol. 78:365-371. 27. Ohnishi, M., J. Ruhno, J. Bienenstock, J. Dolovich, and J. A. Denburg. 1989. Hemopoietic growth factor production by human nasal polyp epithelial scrapings: kinetics, cell source and relationship to clinical status. J. Allergy Clin. Immunol. 83: 1091-1100. 28. Collins, S. J., F. W. Ruscetti, R. E. Gallagher, andR. Gallo. 1979. Normal functional characteristics of cultured human promyelocytic leukemia cells (HL-60) after induction of differentiation of dimethylsulfoxide. J. Exp. Med. 149:969-974. 29. Rovera, G., D. Santoli, and C. Damsky. 1979. Human promyelocytic leukemia cells in culture differentiate into macrophage-like cells when treated with a phorbol diester. Proc. Natl. Acad. Sci. USA 76:2779-2783. 30. Fischkoff, S. A., A. Pollak, G. J. Gleich, J. R. Testa, S. Misawa, and T. J. Reber. 1984. Eosinophilic differentiation of the human promyelocytic leukemia cell line, HL-60. J. Exp. Med. 160:179-196. 31. Nagata, S., M. Tsuchiya, S. Asano et al. 1986. Molecular cloning and expression of cDNA for human granulocyte colony-stimulating factor. Nature 319:415-418. 32. Akagawa, K. S., K. Kamoshita, and T. Tokunaga. 1988. Effect of granulocyte-macrophage colony-stimulating factor and colony-stimulating factor-Ion the proliferation and differentiation of murine alveolar macrophages. J. Immunol. 141:3383-3390. 33. Kapp, A., G. Zeck-Kapp, M. Danner, and T. Luger. 1988. Human granulocyte-macrophage colony stimulating factor: an effective direct activator of human polymorphonuclear neutrophilic granulocytes. J. Invest. Dermatol. 91:49-55. 34. Ming Wang, J., Z. Guo Chen, S. Colella et al. 1988. Chemotactic activity of recombinant human granulocyte colony-stimulating factor. Blood 72:1456-1460. 35. Lindenman, A., D. Reidel, O. Wolfgang et al. 1988. Granulocytemacrophage colony-stimulating factor induces interleukin-l production by human polymorphonuclear neutrophils. J. Immunol. 140:837-839. 36. Ohtoshi, T., C. Vancheri, G. Cox et al. 1991. Monocyte-macrophage differentiation induced by human upper airway epithelial cells. Am. J. Respir. Cell Moi. Bioi. In press. 37. Bagby, G. C., E. McCall, K. A. Bergstrom, and D. A. Burger. 1983. A monokine regulates CSA production by vascular endothelial cells. Blood 62:663-668.
