Abstract
Twenty-six multidrug-resistant Pseudomonas aeruginosa isolates were exposed over time to 300 μg of gamma-linolenic acid or arachidonic acid per ml or to the combination of both acids at 150 μg/ml each with ceftazidime and amikacin with or without albumin to observe the in vitro interactions of the antibiotics. Antibiotics and albumin were applied at their levels found in serum. Synergy between acids and antibiotics was found against 13 isolates, and it was expressed after 5 h of growth in the presence of albumin. The results indicate that further application in experimental infection models is merited.
Polyunsaturated fatty acids (PUFAs), which may alter eicosanoid biosynthesis and tumor necrosis factor alpha secretion in experimental bacterial infection models (1, 8, 11), are appropriate candidates for immunomodulation of infections caused by multidrug-resistant isolates. Gamma-linolenic acid (GLA) and arachidonic acid (AA), which are n-6 PUFAs, suppress the growth of gram-negative isolates, and when used in combination they possess a considerable killing effect on Pseudomonas aeruginosa isolates susceptible to antipseudomonal agents (3–5), a phenomenon probably attributed to their peroxidation. The present study is interested in their effects on multidrug-resistant P. aeruginosa and on their interactions with antipseudomonal agents.
Twenty-six P. aeruginosa isolates derived from patients with nosocomial infections were tested. Their epidemiological characteristics and resistance patterns are given in Table 1. GLA and AA ethyl esters (Sigma Chemical Co., St. Louis, Mo.) were dissolved in 99% ethanol (Merck, Darmstadt, Germany) to an initial dilution of 10 mg/ml. Appropriate amounts of the latter dilution were added to 10-ml tubes that contained Mueller-Hinton broth (Oxoid Ltd., London, United Kingdom) and a log-phase P. aeruginosa inoculum (106 CFU/ml). GLA and AA were present at final concentrations of 150 μg/ml (0.55 mM GLA, 0.49 mM AA) and 300 μg/ml (1.1 mM GLA, 0.98 mM AA), and 99% ethanol was present at a concentration of 3% (vol/vol) or 6% (vol/vol). Ceftazidime (Glaxo, Macclsfield, United Kingdom) or amikacin (Bristol, Syracuse, N.Y.), or both, were added to tubes with 300 μg of either acid per ml or to tubes with 150 μg of each acid per ml. Antibiotics were provided as white amorphous powders, and they were applied at a concentration of 16 μg/ml, which is equal to the mean level achieved in serum after administration of conventional doses (6). All experiments were repeated after the addition of albumin (Sigma) at the level at which it is found in serum (4 g%). Tubes that contained 99% ethanol, albumin, ceftazidime, or amikacin or their respective combinations served as controls.
TABLE 1.
Epidemiological data for the 26 P. aeruginosa isolates included in the study
| Strain no. | Hospital | Mo and yr of isolation | MIC (μg/ml)a
|
|||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| TIC (64) | PIP (64) | CTZ (8) | AZM (8) | IMP (4) | MER (4) | CIP (2) | GEM (4) | TOB (4) | AMK (16) | |||
| 1b | A | Jan 1993 | >128 | >256 | 16 | 64 | 32 | 32 | >4 | >64 | >64 | 128 |
| 2 | A | Feb 1993 | >128 | >256 | 32 | 32 | >512 | >512 | >4 | >64 | >64 | 128 |
| 3b | B | Feb 1993 | >128 | >256 | 32 | 32 | 16 | 16 | >4 | >64 | >64 | 128 |
| 4b | A | Nov 1992 | >128 | >256 | 32 | 32 | 32 | 16 | 4 | >64 | >64 | 256 |
| 5b | B | Mar 1992 | >128 | >256 | 32 | 256 | 16 | 8 | >4 | >64 | >64 | >512 |
| 6b | A | Jan 1992 | >128 | >256 | 32 | 16 | 16 | 8 | >4 | >64 | >64 | 128 |
| 7b | A | Jun 1992 | >128 | 128 | 16 | 256 | 16 | 16 | >4 | >64 | >64 | 128 |
| 8b | C | Jan 1993 | >128 | >256 | 16 | 16 | 8 | 8 | 4 | >64 | >64 | 128 |
| 9b | D | Mar 1993 | >128 | 128 | >512 | >512 | 16 | 16 | >4 | >64 | >64 | 128 |
| 10b | D | Mar 1993 | >128 | >256 | 256 | >512 | 16 | 32 | >4 | >64 | >64 | >512 |
| 11 | E | Mar 1993 | >128 | >256 | >512 | >512 | >512 | 16 | >4 | >64 | >64 | >512 |
| 12b | E | Jan 1993 | 128 | 256 | >512 | >512 | >512 | >512 | >4 | >64 | >64 | >512 |
| 13b | E | Jul 1993 | >128 | >256 | 256 | >512 | 64 | 32 | >4 | >64 | >64 | 256 |
| 14b | D | Nov 1993 | 128 | 128 | 16 | 32 | 8 | 8 | 4 | 32 | 32 | 16 |
| 15 | D | Nov 1993 | >128 | >256 | >512 | >512 | 32 | 16 | >4 | >64 | >64 | >512 |
| 16b | F | Jan 1993 | >128 | >256 | >512 | >512 | >512 | 256 | >4 | >64 | >64 | >512 |
| 17 | F | Mar 1993 | >128 | >256 | 16 | 32 | 32 | 8 | >4 | >64 | >64 | 64 |
| 18 | E | Jun 1993 | >128 | >256 | >512 | >512 | 16 | 64 | >4 | 16 | 16 | 32 |
| 19 | A | Jan 1994 | >128 | >256 | >512 | >512 | 64 | 32 | >4 | >64 | >64 | 64 |
| 20 | A | Oct 1993 | >128 | >156 | 128 | 64 | 256 | 32 | 4 | 16 | 8 | 32 |
| 21 | A | Jan 1994 | 128 | 128 | >512 | >512 | 32 | 32 | 4 | >64 | >64 | >512 |
| 22b | G | Jan 1995 | >128 | >256 | 32 | 32 | 64 | 32 | 4 | >64 | >64 | 128 |
| 23 | G | Mar 1995 | 128 | 128 | 64 | 16 | 8 | 8 | >4 | >64 | >64 | 128 |
| 24 | G | Apr 1995 | >128 | >256 | >512 | >512 | >512 | >512 | >4 | >64 | >64 | >512 |
| 25 | H | Mar 1994 | >128 | >256 | >512 | >512 | >512 | >512 | >4 | >64 | >64 | >512 |
| 26 | H | Jul 1994 | >128 | >256 | >512 | >512 | >512 | >512 | >4 | >64 | >64 | >512 |
Values in parentheses are susceptibility breakpoints. Abbreviations: TIC, ticarcillin; PIP, piperacillin; CTZ, ceftazidime; AZM, aztreonam; IMP, imipenem; MER, meropenem; CIP, ciprofloxacin; GEM, gentamicin; TOB, tobramycin; AMK, amikacin.
Isolates against which synergistic activity was expressed by one of the tested interactions of PUFAs with ceftazidime and amikacin.
All prepared tubes and one growth control tube were left to incubate at 37°C in a shaking water bath, and bacterial growth was determined at standard time intervals (time zero and after 3, 5, and 24 h of incubation) by removing a 0.1-ml aliquot from each tube. The aliquot was serially diluted 1:10 five times in sterile 0.9% NaCl. Another 0.1-ml aliquot of each dilution was diluted onto MacConkey agar (Becton Dickinson, Cockeysville, Md.). These serial dilutions permitted the avoidance of any antimicrobial carryover effect. At each time interval the change in the log10 number of viable cells compared to the number in the starting inoculum was determined. The activity of a combination of a PUFA and an antibiotic was synergistic when it fulfilled the following criteria, in analogy to the definition applied to antimicrobial agents (7): (i) it was bactericidal, achieving a greater than or equal to 3-log10 decrease in viable cell counts from the baseline count, (ii) it was superior to the activity of ceftazidime combined with amikacin, and (iii) it achieved a greater than 2-log10 decrease in viable cell counts compared to that achieved with the combination of one PUFA with ceftazidime and/or amikacin. A total of 1,092 killing curves were prepared for determination of viable cell counts in duplicate. The change in the log10 cell counts was expressed as the median value, and the values were compared by Wilcoxon's rank sum test. The synergy achieved by the combinations tested was compared by Fisher's exact test. Any P value of <0.05 was considered significant after a Bonferroni correction.
The effects of the combinations tested against multidrug-resistant P. aeruginosa are shown in Table 2. Single acids and antimicrobial agents did not exert any considerable effect on the growth of P. aeruginosa, a phenomenon not influenced by the presence of albumin.
TABLE 2.
Synergy between PUFAs, ceftazidime, and amikacin against 26 multidrug-resistant P. aeruginosa isolates after 5 and 24 h of growth
| Interaction tested | Median log10 change in viable cell counts from the baseline
|
No. (%) of isolates against which the combination was synergistic
|
||
|---|---|---|---|---|
| 5 h | 24 h | 5 h | 24 h | |
| Growth control | +0.96 | +2.08 | ||
| Ceftazidime + amikacinh | −1.69 | −1.00 | 3 (11.5) | 2 (7.7) |
| GLA (300 μg/ml) + amikacinh | −0.12 | +0.35 | 1 (3.8) | 5 (19.2) |
| GLA (300 μg/ml) + ceftazidimeh | 0.00 | −1.15 | 2 (7.7) | |
| GLA (300 μg/ml) + ceftazidime + amikacin | −0.59 | −2.42c | 1 (3.8) | 10 (38.5)a |
| GLA (300 μg/ml) + ceftazidime + amikacin + 4 g% albumin | −1.52 | −1.89 | 10 (38.5) | 8 (30.8) |
| AA (300 μg/ml) + amikacinh | 0.00 | +0.26 | 1 (3.8) | 5 (19.2) |
| AA (300 μg/ml) + ceftazidimeh | −0.22 | −1.36 | 4 (15.4) | |
| AA (300 μg/ml) + ceftazidime + amikacin | −0.65 | −2.00de | 10 (38.5)a | |
| AA (300 μg/ml) + ceftazidime + amikacin + 4 g% albumin | −1.81 | −3.00 | 9 (34.6) | 13 (50.0) |
| GLA (150 μg/ml) + AA (150 μg/ml)h | +0.60 | +1.00 | ||
| GLA (150 μg/ml) + AA (150 μg/ml) + amikacinh | +0.30 | +1.00 | 4 (15.4) | |
| GLA (150 μg/ml) + AA (150 μg/ml) + ceftazidimeh | 0.00 | −1.30 | 3 (11.5) | |
| GLA (150 μg/ml) + AA (150 μg/ml) + ceftazidime + amikacin | −0.65 | −2.65f | 13 (50.0)b | |
| GLA (150 μg/ml) + AA (150 μg/ml) + ceftazidime + amikacin + 4 g% albumin | −1.8 | −2.47 | 9 (34.6) | 12 (46.2) |
| 6% ethanol (99%) | 0.00 | −1.22 | ||
| 6% ethanol (99%) + 4 g% albumin | +1.00 | +2.52 | ||
| GLA (300 μg/ml) + AA (300 μg/ml) | −1.00 | −1.21g | 6 (23.1) | |
| GLA (300 μg/ml) + AA (300 μg/ml) + 4 g% albumin | −1.22 | −2.13 | 8 (30.8) | |
P = 0.05 (after Bonferroni correction) for comparison of GLA or AA with both antimicrobial agents versus GLA or AA with one antimicrobial agent.
P = 0.018 (after Bonferroni correction) for comparison of both acids at 150 μg/ml in combination with both antimicrobial agents versus both acids at 150 μg/ml in combination with one antimicrobial agent.
P = 0.05 (after Bonferroni correction) for comparison of GLA at 300 μg/ml with both antimicrobial agents versus GLA at 300 μg/ml with one antimicrobial agent.
P was not significant NS (after Bonferroni correction) for comparison of AA at 300 μg/ml with both antimicrobial agents versus AA at 300 μg/ml with ceftazidime.
P = 0.006 (after Bonferroni correction) for comparison of AA at 300 μg/ml with both antimicrobial agents versus AA at 300 μg/ml with amikacin.
P = 0.002 (after Bonferroni correction) for comparison of GLA and AA with both antimicrobial agents versus GLA and AA with one antimicrobial agent.
P < 0.001 (after Bonferroni correction) for comparison of GLA and AA versus the control (6% ethanol [99%]).
Similar results were obtained in the presence of albumin.
The strains included in the present study differed from each other since they were isolated in different time periods and at different hospitals and the MICs of 10 antimicrobial agents for the isolates were different (Table 1). A significant alteration of the growth of multidrug-resistant P. aeruginosa is achieved by the combination of PUFAs, with which synergistic activity against 30.8% of isolates was found (Table 2), but not by a single acid or by the interaction of a single acid with either ceftazidime or amikacin, with which synergistic activity was not found against more than 19.2% of isolates. However, when both ceftazidime and amikacin interact with GLA or AA at 300 μg/ml, synergistic activity was found against 38.5% of isolates after 24 h of growth, raising up to 50% the proportion of isolates against which each PUFA at 150 μg/ml used in combination interacted with both antibiotics. Even earlier synergist activity, after 5 h of growth, is obtained in the presence of albumin, which is extremely important, since PUFAs are transported in blood bound to albumin (10). Although few data on the levels of PUFAs in plasma exist, the concentration of eicosapentaenoic acid under conditions of stress like sepsis reaches 140 μM (2). Since eicosapentaenoate is usually found at much lower concentrations than AA and both GLA and AA coexist in plasma (although it is presumed that the 150-μg/ml concentration should be more easily achievable in plasma than the 300-μg/ml one), the results presented here should be of clinical significance. That assumption is further supported by the application of ceftazidime, amikacin, and albumin at the levels achieved in serum. The bactericidal effect of GLA in combination with AA has also been demonstrated in studies by our group with isolates susceptible to antipseudomonal agents (5).
It has been shown that the action of PUFAs on the bacterial cell is regulated through their peroxidation (5), leading to cell killing, a process reversed in the presence of the antioxidant vitamin E. For all isolates tested, the MICs of antimicrobial agents with diverse chemical structures were elevated, and synergy was irrespective of the MICs of ceftazidime for the isolates (Table 1). Since the multidrug-resistance of these isolates might result in defective permeation of antibiotics through the bacterial outer membrane (9), the products of the peroxidation of PUFAs attack the cell wall, thus making it more permeable, and this might be a suitable hypothesis of the mechanism of the synergistic activities of these acids with ceftazidime and amikacin. It should be emphasized that the effect of PUFAs is not affected by albumin, since binding of PUFAs to albumin does not involve bonds in which the peroxidation process occurs (10).
The present study indicates that n-6 PUFAs might render ceftazidime and amikacin active against multidrug-resistant P. aeruginosa acids. Since the intravenous administration of the lithium salt of GLA to patients with end-stage human immunodeficiency virus infection has been well tolerated and has been accompanied by a transient increase in CD4+ cell counts (12), PUFAs might be applied as adjuvants of antimicrobial chemotherapy in experimental infection models.
REFERENCES
- 1.Blok W L, Vogels M T E, Curfs J H A J, Eling W M C, Buurman W A, van der Meer J W M. Dietary fish-oil supplementation in experimental gram-negative infection and in cerebral malaria in mice. J Infect Dis. 1992;165:898–903. doi: 10.1093/infdis/165.5.898. [DOI] [PubMed] [Google Scholar]
- 2.Falconer J S, Ross J A, Fearon K C H, Hawkins R A, O'Riordain M G, Carter D C. Effect of eicosapentaenoic acid and other fatty acids on the growth in vitro of pancreatic cell lines. Br J Cancer. 1994;69:826–832. doi: 10.1038/bjc.1994.161. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Giamarellos-Bourboulis E J, Grecka P, Dionyssiou-Asteriou A, Grammatikou M, Giamarellou H. Do Escherichia coli susceptibilities to various antibiotics decrease in the presence of polyunsaturated fatty acids? A preliminary report. J Chemother. 1994;6:39–43. doi: 10.1080/1120009x.1994.11741127. [DOI] [PubMed] [Google Scholar]
- 4.Giamarellos-Bourboulis E J, Grecka P, Dionyssiou-Asteriou A, Giamarellou H. In vitro inhibitory activity of gamma-linolenic acid on Escherichia coli strains and its influence on their susceptibilities to various antimicrobial agents. J Antimicrob Chemother. 1995;36:327–334. doi: 10.1093/jac/36.2.327. [DOI] [PubMed] [Google Scholar]
- 5.Giamarellos-Bourboulis E J, Grecka P, Dionyssiou-Asteriou A, Giamarellou H. In vitro activity of polyunsaturated fatty acids on Pseudomonas aeruginosa: relationship to lipid peroxidation. Prostaglandins Leukotrienes Essent Fatty Acids. 1998;58:283–287. doi: 10.1016/s0952-3278(98)90037-0. [DOI] [PubMed] [Google Scholar]
- 6.Giamarellos-Bourboulis E J, Grecka P, Giamarellou H. Comparative in vitro interactions of ceftazidime, meropenem and imipenem with amikacin on multiresistant Pseudomonas aeruginosa. Diagn Microbiol Infect Dis. 1997;29:81–86. doi: 10.1016/s0732-8893(97)00123-5. [DOI] [PubMed] [Google Scholar]
- 7.Hindler J. Tests to assess bactericidal activity. In: Eisenberg H D, editor. Clinical microbiology procedures handbook—1992. Washington, D.C.: American Society for Microbiology; 1992. pp. 5.16.14–5.16.24. [Google Scholar]
- 8.Jeffery B N M, Cortina M, Newsholme E A, Calder P C. Effects of variations in the proportions of saturated, monounsaturated and polyunsaturated fatty acids in the rat diet of spleen lymphocyte functions. Br J Nutr. 1997;77:805–823. doi: 10.1079/bjn19970077. [DOI] [PubMed] [Google Scholar]
- 9.Masuda N, Sakagawa E, Ohya S. Outer membrane proteins responsible for multiple drug resistance in Pseudomonas aeruginosa. Antimicrob Agents Chemother. 1995;39:645–649. doi: 10.1128/AAC.39.3.645. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Montgomery R, Conway T W, Spector A A, Chapell D. Lipoproteins. In: Montgomery R, Conway T W, Spector A A, Chapell D, editors. Biochemistry, a case-oriented approach. 6th ed. St. Louis, Mo: The C. V. Mosby Co.; 1996. pp. 356–389. [Google Scholar]
- 11.Turek J J, Schoenlein I A, Watkins B A, van Alstine W G, Clark L K, Knox K. Dietary polyunsaturated fatty acids modulate responses of pigs to Mycoplasma hyopneumoniae infection. J Nutr. 1996;126:1541–1548. doi: 10.1093/jn/126.6.1541. [DOI] [PubMed] [Google Scholar]
- 12.Winther M, Schlech W, Conway B, Singer J. Phase I open label study of the safety, pharmacokinetics and potential efficacy of LiGLA in advanced HIV infection. Prostaglandins Leukotrienes Essent Fatty Acids. 1997;57:247. [Google Scholar]