Immunomodulation by Non-Absorbable Antibiotics Given by the Intragastric Route Magdalena Leiva, Encarnacion Moreno, Alfonso Ruiz-Bravo, Maria Jimenez-Valera∗

International Journal of Antimicrobial Agents 25 (2005) 252–255

Immunomodulation by non-absorbable antibiotics given by the intragastric route Magdalena Leiva, Encarnacion Moreno, Alfonso Ruiz-Bravo, Maria Jimenez-Valera∗

Department of Microbiology, Faculty of Pharmacy, University of Granada, 18071 Granada, Spain Received 29 July 2004; accepted 6 October 2004

Abstract

To test the role of bacterial fractions released from intestinal flora during immunomodulation by antimicrobial agents, BALB/c mice were treated with the non-absorbable antibiotics polymyxin B or teicoplanin by the intragastric route. The composition of faecal microbiota and the capacity of spleen cells to proliferate in response to B-cell and T-cell mitogens were assessed at several times during the treatment. Both antibiotics lowered the count of some bacteria of the intestinal flora and induced significant modifications in spleen cell ability to proliferate in response to mitogens. Thus, the active fractions released from intestinal bacteria during antibiotic treatments may be able to induce immunomodulating effects. © 2004 Elsevier B.V. and the International Society of Chemotherapy. All rights reserved.

Keywords: Immunomodulation; Antibiotics; Indigenous bacteria; Intestinal microbiota

1. Introduction when passing through the digestive tract, are able to exert immunomodulatory effects, then these effects should be at- Numerous antimicrobial agents are biological response tributed to modulins released by antibiotic-killed bacteria. In modifiers; that is, they are able to modulate multiple inter- and this paper, the effects of two non-absorbable antibiotics on intracellular molecular signals involved in the functioning of the mouse intestinal microbiota and their capacity to modify the immune system [1–4]. Antibiotics may act directly on the response of spleen cells to mitogenic stimuli were studied. cells of the immune system, which has been demonstrated by in vitro experiments with several antibiotic families, such as ␤-lactams [5,6], macrolides [7,8], or fluoroquinolones [9,10]. 2. Methods Additionally, it is conceivable that antibiotics, when given in vivo, may result in lysis of indigenous bacteria, thus releasing 2.1. Mice bacterial fractions having immunomodulatory activity. These are called modulins. Although there are some data suggesting Six to eight-week female BALB/c mice were provided a role for this indirect mechanism in the immunomodulation by the Technical Services of the University of Granada by specific antibiotics [11], conclusive experimental evidence (Granada, Spain). They were maintained under pathogen-free has not been reported. The comparison of immunomodula- conditions. The experiments were approved and supervised tion by antibiotics in both conventional and germ-free an- by the local ethical committee at the University of Granada. imals does not seem a suitable model, since the immatu- rity of the immune system of germ-free animals invalidates 2.2. Antibiotic treatments such comparisons. However, if non-absorbable antibiotics, Polymyxin B was purchased from Sigma Chemical Co. ∗ Corresponding author. Tel.: +34 958 243872; fax: +34 958 246235. (St. Louis, Mo). Teicoplanin was a gift from Aventis (Ro- E-mail address: [email protected] (M. Jimenez-Valera). mainville, France). The drugs were dispersed at the desired

0924-8579/$ – see front matter © 2004 Elsevier B.V. and the International Society of Chemotherapy. All rights reserved. doi:10.1016/j.ijantimicag.2004.10.013 M. Leiva et al. / International Journal of Antimicrobial Agents 25 (2005) 252–255 253 concentrations in distilled water and were administered by diphenyltetrazolium bromide reduction as described by Mos- oesophageal gavage once a day. The drugs were given at the mann [12]. following dosages in milligrams per kilogram of body weight: polymyxin B, 10 and teicoplanin, 40. 2.5. Statistical analysis

2.3. Quantification of bacterial populations in feces The differences between control and antibiotic-treated groups were analysed using Student’s t-test. A P value of Fresh faeces were weighed, homogenised in sterile saline less than 0.05 was considered significant. solution, and suitable dilutions of the homogenates were plated onto culture media agar plates for viable bacteria counts. The following culture media (Difco Laboratories Inc., 3. Results Detroit, MI) were used: Tryptic soy agar (TSA) for total aerobic bacteria, MacConkey agar for enterobacteria, MRS agar 3.1. Intestinal microbiota changes and for lactobacilli and Brewer agar, under anoxic atmosphere immunomodulation following treatment with polymyxin B (GasPak, BBL Microbiology Systems, Cockeysville, MD) in anaerobic jars, for anaerobic bacteria. Results were expressed The impact of polymyxin B on faecal bacterial counts as the number of colony-forming units (CFU) per g of fresh is shown in Table 1. No relevant changes in the aerobic faeces. flora, anaerobic flora and lactobacilli were observed during the study period, but the number of enterobacteria dropped 3 2.4. Spleen cell proliferation assay below the lowest detectable number of bacteria (<10 CFU per g of faeces) after 24 h following the start of treatment Spleens were removed aseptically and homogenised and continued to be undetectable for the administration in RPMI 1640 medium supplemented with 10% heat- period. inactivated foetal calf serum, 50 ␮M 2-mercaptoethanol, We examined the effect of treatment with polymyxin on penicillin G (100 U/ml), streptomycin (100 ␮g/ml), ampho- splenocyte proliferation in response to mitogens (Table 2). tericin B (0.25 ␮g/ml), 1 mM sodium pyruvate and 2 mM A 5-day treatment suppressed the Con A-driven response l-glutamine (Sigma). Splenocytes were sedimented by cen- by 69% (P < 0.025), whereas the response to LPS was un- trifugation, resuspended in red blood cell lysing buffer affected. A 12-day treatment enhanced the response to LPS (Sigma) for 10 min, washed, and resuspended in fresh by 66% (P < 0.005) and suppressed the response to Con A medium. Cell suspensions were distributed (5 × 105 viable by 20% (P < 0.025). When the treatment was extended for 19 cells per well) into 96-well tissue culture clusters with flat- days, no significant changes were observed. bottom wells (Costar, Cambridge, Mass.). Salmonella Typhi lipopolysaccharide (LPS; Sigma) was used at 2.5 ␮g/ml as B- 3.2. Intestinal microbiota changes and cell mitogen and concanavalin A (Con A; Sigma) was used immunomodulation following treatment with teicoplanin at 1 ␮g/ml as T-cell mitogen; these mitogen concentrations have been shown to induce optimum splenocyte proliferation The effect of teicoplanin on the intestinal flora is shown in our assay conditions [11]. After incubation at 37 ◦Cin5% in Table 3. Both aerobic and anaerobic bacterial counts were CO2 for 3 days, proliferation of spleen cells was measured clearly increased. Dramatic increases in CFU counts on Mac- by colorimetric reading of 3-(4,5-dimethylthiazol-2-yl)-2,5- Conkey plates were also observed. The macroscopic and mi-

Table 1 Changes of bacterial populations in faeces during treatment with polymyxin B given by the intragastric routea Days from initiation of polymyxin B Log colony counts per gram of faeces

TSAb MRSc Brewerd MacConkeye 0 8.3 ± 0.22 8.5 ± 0.15 9.3 ± 0.13 6.2 ± 0.45 1 8.6 ± 0.20 8.7 ± 0.12 9.1 ± 0.12 <3 2 8.7 ± 0.23 8.7 ± 0.14 9.1 ± 0.13 <3 4 8.3 ± 0.18 8.6 ± 0.15 9.2 ± 0.12 <3 10 8.4 ± 0.25 8.6 ± 0.17 9.2 ± 0.12 <3 16 8.1 ± 0.23 8.7 ± 0.15 9.1 ± 0.12 <3 a Polymyxin B (10 mg/kg) was given by oesophageal gavage once a day. Data represent the mean log10 ± standard deviation of viable bacteria per gram of faeces from four animals. b TSA Tryptic Soy Agar (total aerobic bacteria). c MRS (lactobacilli). d Brewer Agar (anaerobic bacteria). e MacConkey Agar (enterobacteria). 254 M. Leiva et al. / International Journal of Antimicrobial Agents 25 (2005) 252–255

Table 2 Mitogen-driven proliferation of spleen cells from mice treated with polymyxin Ba Mitogen Treatment of mice Optical density (570–630 nm)

Day 5 Day 12 Day 19 LPS None 0.242 ± 0.0277 0.137 ± 0.0269 0.153 ± 0.0259 Polymyxin B 0.230 ± 0.0299 0.228 ± 0.0184b 0.194 ± 0.0723 ConA None 0.447 ± 0.1547 0.456 ± 0.0455 0.480 ± 0.0910 Polimyxin B 0.137 ± 0.0572c 0.366 ± 0.0264c 0.364 ± 0.0664 a Polymyxin B (10 mg/kg) was given by oesophageal gavage once a day. Results are mean ± standard deviation for four animals. b P < 0.005. c P < 0.025.

Table 3 Changes of bacterial populations in faeces during treatment with teicoplanin given by intragastric routea Days from initiation of teicoplanin Log colony counts per gram of faeces

TSA MRS Brewer MacConkey 0 8.1 ± 0.22 8.4 ± 0.20 8.8 ± 0.32 5.8 ± 0.35 2 9.9 ± 0.24 10.1 ± 0.20 10.1 ± 0.21 9.7 ± 0.30 410± 0.19 10 ± 0.21 10 ± 0.25 9.6 ± 0.22 10 9.5 ± 0.20 9.3 ± 0.25 9.4 ± 0.26 9.4 ± 0.29 a Teicoplanin (40 mg/kg) was given by oesophageal gavage once a day. Data represent the mean log10 ± standard deviation of viable bacteria per gram of faeces from four animals.

Table 4 the immunomodulation reported here may be attributed to Mitogen-driven proliferation of spleen cells from mice treated with a the disturbance of the ecological balance of intestinal flora teicoplanin caused by the administration of antimicrobial agents. How- Mitogen Treatment of mice Optical density (570–630 nm) ever, immunomodulation disappeared when the treatment Day 5 Day 12 with polymyxin B was continued for 19 days, although the in- LPS None 0.278 ± 0.0330 0.330 ± 0.0357 testinal microbiota remained disturbed. This suggests that im- Teicoplanin 0.318 ± 0.0244 0.273 ± 0.1139 munomodulation is due to modulins released from antibiotic- ConA None 0.308 ± 0.0246 0.342 ± 0.0482 susceptible bacteria, since it is reasonable to accept that sus- Teicoplanin 0.495 ± 0.0501b 0.248 ± 0.0530c ceptible bacteria are suppressed when antibiotic treatment is a Teicoplanin (40 mg/kg) was given by oesophageal gavage once a day. long enough. A similar explanation has been proposed for the Results are mean ± standard deviation for four animals. loss of immunosuppressive activity of ciprofloxacin in mice b P < 0.002. when the antibiotic treatment is extended for a week [11]. c P < 0.05. Besides the loss of effect with prolonged treatment, im- croscopic study of colonies revealed the presence of several munomodulation by both polymyxin B and teicoplanin was types of Gram-positive bacilli and cocci on Brewer plates clearly influenced by the duration of the antibiotic administra- from faeces of untreated mice. During the teicoplanin admin- tion period. This may be the consequence of a time-dependent istration period, these Gram-positive bacteria were replaced pattern of release of helper and suppressive cytokines as de- with a Gram-negative, facultatively anaerobic bacilli that was scribed for some experimental models of immunomodulation able to growth on MacConkey agar. by bacterial modulins [13,14]. Table 4 shows the effect of teicoplanin on splenocyte pro- Animal models and clinical studies have shown that the liferation in response to mitogens. The 5-day treatment en- antibiotic treatment of bacteraemia may cause swift clinical hanced the response to Con A by 61% (P < 0.002), whereas deterioration, including septic shock, due to rapid bacterial the 12-day treatment suppressed it by 27% (P < 0.05). The lysis with release of biologically active cell wall components LPS-driven response was not modified. into the circulation [15]. The antibiotic-mediated release of cell wall components, such as endotoxin from Gram-negative bacteria promotes production of pro-inflammatory cytokines 4. Discussion that trigger systemic inflammatory reactions [16]. However, there is no evidence that antibiotic-killed bacteria belonging The data presented here demonstrate that non-absorbable to the intestinal flora can be a source of biologically active antibiotics are able to exert immunomodulatory effects when fractions inducing immunomodulation at the systemic level. given by the intragastric route. These effects are associ- To our knowledge, this is the first report to demonstrate that ated with changes in the intestinal microbiota. Given the antimicrobial agents may exert immunomodulatory effects influence of intestinal bacteria on the immune functions, through the release of modulins from indigenous bacteria. M. Leiva et al. / International Journal of Antimicrobial Agents 25 (2005) 252–255 255

Polymyxin B is known to neutralize the biological activity [5] Tufano MA, Cipollaro de l’Ero G, Ianniello R, Baroni A, Galdiero of endotoxin [17,18], and a similar function has been found F. Antimicrobial agents induce monocytes to release IL-1 alpha, IL- in teicoplanin [19]. Although we do not known the relative 6, and TNF, and induce lymphocytes to release IL-4 and TNF tau. Immunopharmacol Immunotoxicol 1992;14:769–82. contribution that endotoxin made on the immunomodulation [6] Dabrowski MP, Stankiewicz W. Desirable and undesirable im- described here, it is reasonable to accept that the endotoxin munotropic effects of antibiotics: immunomodulating properties of role was diminished by the neutralizing ability of both drugs. cefaclor. J Chemother 2001;13:615–20. The endotoxin contribution should be more effective in the [7] Morikawa K, Oseko F, Morikawa S, Iwamoto K. Immunomodula- case of treatment with antibiotics lacking in neutralizing ac- tory effects of three macrolides, midecamycin acetate, josamycin, and clarithromycin, on human T-lymphocyte function in vitro. An- tivity. However, the immunomodulation due to fractions from timicrob Agents Chemother 1994;38:2643–7. indigenous bacteria is very difficult to demonstrate with ab- [8] Ichiyama T, Nishikawa M, Yoshitomi T, et al. Clarithromycin in- sorbable antibiotics, which reach blood levels that are high hibits NF-kappaB activation in human peripheral blood mononuclear enough to cause immunomodulation by direct action on the cells and pulmonary epithelial cells. Antimicrob Agents Chemother immunocompetent cells. The use of non-absorbable antibi- 2001;45:44–7. [9] Riesbeck K, Sigvardsson M, Leanderson T, Forsgren A. Superin- otics administered to mice by the intragastric route provides duction of cytokine gene transcription by ciprofloxacin. J Immunol an useful model to study the effect of bacterial modulins re- 1994;153:343–52. leased from the intestinal flora. [10] Dalhoff A, Shalit I. Immunomodulatory effects of quinolones. Lancet Infect Dis 2003;3:359–71. [11] Jimenez-Valera M, Sampedro A, Moreno E, Ruiz-Bravo A. Modi- Acknowledgments fication of immune response in mice by ciprofloxacin. Antimicrob Agents Chemother 1995;39:150–4. [12] Mosmann T. Rapid colorimetric assay for cellular growth and sur- This work was supported by Ministerio de Ciencia y Tec- vival: application to proliferation and cytotoxicity assays. J Immunol nolog´ıa and the European Regional Development Fund (SAF Methods 1983;65:56–63. 2002-01606), and by the Junta de Andalucia (Research Group [13] Johnson AG, Tomai MA, Chen Y, Odean MJ. A comparison of the CVI201). M. Leiva was supported by a grant from the Minis- immunomodulating properties of two forms of monophosphoryl lipid terio de Educacion, Cultura y Deporte (Beca-colaboracion). A analogues. J Immunother 1991;10:398–404. [14] Odean MJ, Johnson AG, Mohrman M, Hasegawa A. Reversal of lipopolysaccharide-analog-induced antibody suppression by anti- transforming growth factor ␤ and indomethacin. Infect Immun References 1995;63:2371–3. [15] Lepper PM, Held TK, Schneider EM, Bolke E, Gerlach H, Traut- [1] Ritts RE. Antibiotics as biological response modifiers. J Antimicrob mann M. Clinical implications of antibiotic-induced endotoxin re- Chemother 1990;26(Suppl. C):31–6. lease in septic shock. Intensive Care Med 2002;28:824–33. [2] Jimenez-Valera M, Moreno E, Ruiz-Bravo A. Immunomodulation by [16] Morrison DC, Ryan JL. Endotoxin and disease mechanisms. Ann antimicrobial agents. In: Pandalai SG, editor. Recent research devel- Rev Med 1987;38:417–32. opments in antimicrobial agents and chemotherapy, vol. 2. Trivan- [17] Riffikind D, Palmer JD. Neutralization of endotoxin toxicity in chick drum: Research Singpost; 1997. embryos by antibiotics. J Bacteriol 1966;92:815–9. [3] Labro MT. Interference of antimicrobial agents with phagocyte func- [18] Tufano MA, Rossano F, Catalanotti P, et al. Immunobiological ac- tions: immunomodulation or immuno-fairy tales? Clin Microbiol Rev tivities of Helicobacter pylori porins. Infect Immun 1994;62:1392–9. 2000;13:615–50. [19] Foca A, Matera G, Berlinghieri MC, Liberto MC, De Sarro [4] Hamilton-Miller JM. Immunopharmacology of antibiotics: direct and GB. Teicoplanin reduces in-vitro reactivity and murine lethality indirect immunomodulation of defence mechanisms. J Chemother of Salmonella minnesota R595 lipopolysaccharide. J Antimicrob 2001;13:107–11. Chemother 1992;29:443–6.