Effects of granulocyte-macrophage colony stimulating factor (GM-CSF) in vivo on cytokine production and proliferation by spleen cells

Abstract

GM-CSF is a potent stimulator of haematopoietic cells as well as some functions of granulocytes and macrophages. GM-CSF has many clinical uses; however, little is known about the effects of GM-CSF treatment in vivo on the responses of tissue lymphocytes in terms of secretion of Th-1 and Th-2 cytokines. We investigated this issue by measuring the responses of spleen cells from mice 24 h after treatment i.p. with saline or rmGM-CSF. GM-CSF at 16·7–50·0 µg/kg significantly increased (P < 0·01) spleen cellularity 2–2·5-fold and enhanced proliferative responses of non-stimulated (no mitogen) as well as concanavalin A (Con A)-stimulated spleen cells. Secretion of IFN-γ by Con A (2·5 µg/ml)-stimulated spleen cells was significantly (P < 0·01) increased from 1·8 µg/ml by control spleen cells to 5·2 μg/ml by GM-CSF spleen cells. IL-10 production was greater (0·25 μg/ml, P < 0·05) by Con A-stimulated spleen cells from GM-CSF-treated mice compared to control spleen cells (0·06 μg/ml). By contrast, there were no significant differences in IL-4 production by Con A-stimulated spleen cells from the different groups. These results show that GM-CSF treatment increases spleen cellularity and primes lymphocytes for enhanced responses. The enhanced production of Th-1 cytokines by primed lymphocytes may partially explain the beneficial role of in vivo administration of GM-CSF in several clinical situations.

Introduction

Granulocyte-macrophage colony stimulating factor (GM-CSF) is a 22-Kda glycoprotein cytokine [1–7] secreted by mononuclear leucocytes [8] which regulate the proliferation and differentiation of myeloid cells [9] and the activation of mature granulocytes and monocytes [10,11]. There are several clinical benefits resulting from administration of GM-CSF. GM-CSF can increase the number of leucocytes in patients with AIDS [12], aplastic anaemia [13] and chemotherapy-induced neutropenia [14]. In addition to the effects of GM-CSF cited above, in vitro GM-CSF can induce secretion of several inflammatory cytokines by targeted cells, e.g. IL-1, TNF, M-CSF and G-CSF [15]. Here we report on the in vivo effects of GM-CSF administration on priming leucocytes for enhanced proliferative responses to stimulation and production of Th-1 and Th-2 cytokines.

Materials and methods

Reagents

Recombinant mouse GM-CSF was purchased from R&D Systems Inc., Minneapolis, MN, USA or Endogen, Woburn, MA, USA. In preliminary experiments the activity of both preparations was found to be the same, and they could be used interchangeably. ELISA kits for IL-4, IL-10, and IFNγ were purchased from Endogen Company. Woburn, MA, USA. [methyl]-[³H]-thymidine, specific activity 185GBq/mmol 5·0 Ci/mmol, was obtained from Nycomed Amersham, Buckinghamshire, UK. RPMI-1640, fetal bovine serum (FBS) and concanavalin A (Con A) were purchased from Sigma Chemical Co., St Louis, MO, USA.

Cytokines

Groups of male CD-1 mice (Charles Rivers, Hollister, CA, USA) 6–8 weeks of age were treated i.p. with saline (0·2 ml/mouse) or rmGM-CSF (0·5–1·5 µg/mouse, i.e. 16·7–50·0 µg/kg). The doses of GM-CSF spanned a range from previous experiments where GM-CSF in vivo was able to reverse dexamethasone suppression of alveolar macrophages. Spleens were removed 24 h after treatment, weighed and single cell suspensions prepared. Spleen cells were counted with a haemocytometer and the total number of spleen cells per spleen calculated. Spleen cells (2·5 × 10⁶/ml RPMI-1640 + 10% FBS) were dispensed 0·2 ml per microtest plate well and were cultured with or without Con A at 37°C in 5% CO₂ + 95% air for 20, 24 or 26 h. Cultured supernatants were collected, stored at −80°C until tested for IFN-γ, IL-4 and IL-10 using ELISA kits.

Proliferative responses of spleen cells

Spleen cells from different groups of mice were suspended to 2 × 10⁶/ml CTCM and were dispensed 0·2 ml per round-bottom microtest plate wells. Sets of quadruplicate cultures were incubated with or without Con A for 48 h at 37°C in 5% CO₂ + 95% air, then 0·01 ml of [³H]-thymidine (0·1 mCi/ml) was added per culture. After incubation for another 24 h cultures were harvested onto Whatman, GF/C, glass microfibre filters with a multi-well cell harvester. Dried filter disks were placed in 7-ml polyethylene vials, 5 ml of scintillation fluid (Scintisafe Plus, Fisher Chem. Co., Fairlawn, NJ, USA) and counts per minute (cpm) measured with a TM Anayltic Mark V liquid scintillation counting system.

Statistics

Student's t-test was used to compare statistical differences between groups. Where appropriate, Bonferroni's adjustment to the t-test was used. Statistical significance was set at P < 0·05.

Results

Effect of in vivo GM-CSF on spleen cell proliferation

GM-CSF (Endogen) 0·75 µg to 1·5 µg/mouse (25·1–50 µg/kg) given i.p. resulted in spleen cells that had significantly (P < 0·01) increased proliferation without stimulation (no Con A) compared to spleen cells from saline-treated mice (Table 1). Moreover, GM-CSF treatment primed spleen cells for significantly (P < 0·01) enhanced proliferative responses to the T cell mitogen Con A. The enhancement of spleen cell responses to Con A was seen over a range of Con A concentrations, 1·0–0·1 µg/ml (Table 1). When Con A at 5 µg/ml was used there were no significant differences in proliferative responses between spleen cells of mice treated with GM-CSF and saline.

Table 1.

Effect of GM-CSF in vivo on spleen cell responses to Con A

Conditions Con A µg/ml	Saline Spleen cells cpm ± s.d.	GM-CSF Spleen cells cpm ± s.d.	P-value*
0·0	2164 ± 205	4491 ± 672	0·01
1·0	84 772 ± 8685	116 555 ± 18 200	0·05
0·5	22 916 ± 3983	51 853 ± 8410	0·01
0·1	2515 ± 446	5269 ± 1177	0·01

^*Responses of spleen cells from saline-treated mice compared with responses of spleen cells from GM-CSF (1·5 µg/mouse)-treated mice.

Similar results were obtained in two other experiments where non-stimulated spleen cells from saline-treated mice gave 1473 ± 272 and 2513 ± 1052 cpm and spleen cells from GM-CSF-treated (1·0 and 0·75 µg/mouse) mice gave 4115 ± 1237 and 5167 ± 905 cpm (P < 0·01), respectively. Responses to Con A at 1·0 µg/ml by spleen cells after saline treatment were 38 431 ± 11 184 and 90 800 ± 2441 cpm, whereas responses by spleen cells after GM-CSF treatment were significantly greater (P < 0·01), 63 242 ± 11 824 and 108 467 ± 4696.

Effect of GM-CSF in vivo on spleen cellularity

In four experiments the number of spleen cells obtained per spleen was determined. A significantly (P < 0·01) greater number of spleen cells per spleen (113·0 ± 14·0 × 10⁶) were obtained from spleens of GM-CSF (Endogen) (0·5–1·25 µg/mouse)-treated mice compared to spleen cells numbers per spleen from saline-treated mice (80·0 ± 7·0 × 10⁶).

Secretion of IFNγ by Con A-stimulated spleen cells

Spleen cells from saline-treated mice were cultured with or without Con A for 24 h then cell-free supernatants were harvested. Spleen cell supernatants with or without Con A 1·0-µg/ml did not contain detectable amounts of IFNγ (< 0·037 µg/ml). However, Con A at 2·5 and 5·0 µg/ml induced secretion of IFNγ in a dose-dependent manner (Fig. 1). For example, Con A at 2·5 and 5·0 µg/ml induced secretion of 1·85 µg/ml and 7·50 µg/ml of IFNγ, respectively.

Fig. 1.

Effect of GM-CSF in vivo on IFNγ production by spleen cells. Secretion of IFNγ (µg/ml) by spleen cells from saline- or GM-CSF-treated mice (0·75 µg/mouse) stimulated with Con A is shown on the vertical axis. The mean ± standard deviation from the means of two experiments is given.

Incubation of spleen cells from GM-CSF (Endogen)-treated mice (0·75 µg/mouse) without Con A, or Con A at 1·0 µg/ml, did not result in secretion of measurable amounts of IFNγ. On the other hand, spleen cells from GM-CSF-treated mice stimulated with Con A at 2·5 and 5·0 µg/ml significantly (P < 0·01) increased the secretion of IFNγ, compared to production of IFNγ by Con A-stimulated spleen cells from saline-treated mice (Fig. 1). For example, 5·25 ± 0·35 and 15·2 ± 2·0 µg IFNγ/ml, respectively, compared to 1·85 and 7·50 µg IFNγ/ml by matched controls (Fig. 1). Lower doses of GM-CSF (0·06, 0·12, 0·25 or 0·5 µg/mouse) did not produce this effect.

When another source of GM-CSF (R&D) was used in experiments, results similar to the above were obtained with 0·75 µg/mouse. Con A (5 µg/ml)-stimulated spleen cells from GM-CSF-treated mice produced 7·25 ± 0·77-fold greater amounts of IFNγ (P < 0·01) than similarly cultured spleen cells from saline-treated mice.

Production of IL-4 and IL-10 by spleen cells

GM-CSF (R&D) was used in a set of experiments where secretion of IL-4 and IL-10 by spleen cells was studied. Spleen cells from saline or GM-CSF (0·5 µg/mouse, i.e. 16·7 µg/kg)-treated mice were cultured, with or without Con A at 5·0 µg/ml. In these experiments supernatants were collected after 20 and 26 h of incubation.

Although detectable amounts of IL-4 were measured (20–50 pg/ml) in culture supernatants, there was no significant difference in IL-4 production by Con A-stimulated spleen cells from the different groups (four groups, two experiments).

There were no significant differences in IL-10 secretion by Con A-stimulated spleen cells from the different groups after 20 h of incubation (Fig. 2). However, IL-10 secretion was significantly (P < 0·05) greater by Con A (5 µg/ml)-stimulated spleen cells from GM-CSF-treated mice (250 ± 75 pg/ml) at 26 h than IL-10 (75 ± 37 pg/ml) by saline spleen cells (Fig. 2).

Fig. 2.

Effect of GM-CSF in vivo on spleen cell production of IL-10. Secretion of IL-10 by spleen cells from saline- or GM-CSF-treated mice (0·5 µg/mouse) after stimulation with Con A for 20 h or 26 h is given on the vertical axis. The mean ± standard deviation from the means of two experiments is given.

Discussion

The multiple doses of GM-CSF (1·0–8·0 µg/kg/day, 14 days) used clinically [16] is much smaller than the single dose of GM-CSF (16·7–50·0 µg/kg) used in our studies. Additional studies are necessary to determine whether small multiple doses of GM-CSF would produce the effects reported here.

Increased spleen cell cellularity resulting from treatment with GM-CSF is not surprising, because GM-CSF is noted to be a potent stimulator of haematopoietic cells [9]. Although we measured a significant increase in spleen cells per spleen from GM-CSF-treated mice, we did not compare GM-CSF spleen cell phenotypes with saline spleen cell phenotypes.

The proliferation of spleen cells from GM-CSF-treated mice was significantly elevated in cultures without or with Con A. Since proliferation of lymphocytes is IL-2-dependent, our results indicate that GM-CSF treatment produced signals that turned on IL-2 production in spleen cell cultures. The signalling mechanism between GM-CSF receptor-bearing cells and priming of lymphocytes for these responses remains to be determined.

Increased IFNγ production by Con A-stimulated spleen cells from GM-CSF-treated mice compared to control spleen cells is a novel finding. Although there are numerous in vitro studies showing GM-CSF potentiation of phagocyte antimicrobial activity [10,11,17–20] there are few reports on the in vivo effect of GM-CSF in the non-compromised host relative to resistance to infection. Enhanced IFNγ production by stimulated lymphocytes after in vivo GM-CSF treatment reported here suggests that this could be a factor for increased resistance to certain pathogens.

We found that IL-10, but not IL-4, secretion was significantly increased by Con A-stimulated spleen cells from GM-CSF-treated mice compared to similarly cultured saline spleen cells. Increased production of IL-10 was seen only after 26 h of incubation. The delayed production of IL-10, a Th-2 cytokine with immunosuppressive activity for Th-1 cells [21,22], suggests an operative feedback mechanism in the spleen cell cultures. The significance of these results in terms of clinical use of GM-CSF needs further assessment at this time.