Effect of Staphylococcus aureus bacteria and bacterial toxins on meningeal permeability in vitro

Regional Anesthesia and Pain Medicine 24(1): 24-29, 1999

Effect of Staphylococcus aureus Bacteria and Bacterial Toxins on M e n i n g e a l P e r m e a b i l i t y In Vitro Wolfgang C. Ummenhofer, M.D., Ann E. Stapleton, M.D., and Christopher M. Bernards, M.D.

Background and Objectives. Epidural catheterization is associated with a significant

bacterial colonization rate and occasionally frank infection. During epidural space infection, decreased analgesia despite increased epidural opioid doses has been described. One possible explanation for this observation is that bacterial infection decreases meningeal permeability. The purpose of the study was to determine whether Staphylococcus aureus bacteria, the most common organism causing epidural space infection, or S. aureus toxins alter meningeal permeability. Methods. Spinal meninges of M. nemestrina monkeys were mounted in a previously established in vitro diffusion cell model and exposed to S. aureus toxins A, B, and F. Simultaneous transmeningeal fluxes of mannitol and sufentanil were measured before and after toxin exposure and compared to controls. In a second series of experiments, diffusion cells were inoculated with live S. aureus bacteria in suspension and the permeability of sufentanil was investigated. Results. Staphylococcus aureus toxin-A increased the transmeningeal flux of mannitol but not sufentanil. Toxins B and F did not alter the meningeal permeability of either drug. Inoculation with live S. aureus bacteria increased the transmeningeal flux of sufentanil by 115 _+ 21% (P = .032). Conclusions. These data demonstrate that S. aureus alpha-toxin and live S. aureus bacteria can increase meningeal permeability. Thus, clinical observations of decreased epidural analgesia in the face of bacterial infection cannot be explained by decreased meningeal permeability. Reg Anesth Pain Med 1999: 24: 24-29. Key words: epidural analgesia, epidural abscess, infection, meningitis, M. nemestrina, S. aureus.

Epidural catheters left in place long term have been associated with a rate of bacterial colonization as high as 22% (1,2). Even t e m p o r a r y epidural catheter placement has been associated with an increasing incidence of catheter-related infections (3). Catheter-assodated infections usually result from c o n t a m i n a t i o n by skin flora (3-5), with Staphylococcus aureus, the most frequently isolated organism (5). In tact, in reviewing the available literature, Kindler et al. f o u n d that Staphylococcus species were f o u n d in 83% of all catheter-related infections w h e r e an organism was identified and in 60% of all

From the Departments of Anesthesiology and Medidne, University of Washington, Seattle, Washington. This work was performed at the University of Washington, Department of Anesthesiology, Seattle, Washington, and was supported by a grant from the National Institute on Drug Abuse (ROI DA 07313-04A2). Monkey tissues were obtained from the Regional Primate Research Center at the University of Washington, supported by NIH grant RR00166. Accepted for publication August 25, 1998. Reprint requests: Christopher M. Bernards, M.D., Department of Anesthesiology, Box 356540, University of Washington, Seattle, WA, 98195. Copyright © 1999 by the American Society of Regional Anesthesia. 0146-521X199/2401-000655.00/0

24

Staphylococcus aureus Bacteria Effects on Meningeal Permeability •

patients with catheter-related epidural abscess formation (6). Du Pen et al. observed decreased epidural analgesia following epidural space infection, despite increasing opioid doses (4). Although the mechanism by which the opioid's analgesic effect diminishes is unknown, it is likely that the epidural space infection decreases the spinal bioavailability of drugs administered into the epidural space. One way this could occur would be if meningeal permeability, which is the rate-limiting step in movement of drugs between the epidural space and the spinal cord (7), is decreased by bacterial colonization or infection. In fact, S. aureus toxins have been shown to alter the permeability of various tissues (8-11). This study tested the hypothesis that S. aureus decreases permeability through the spinal meninges. To address this question, we used a previously established in vitro diffusion cell model (12-15) to quantitate the transmeningeal flux of mannitol and sufentanil across monkey spinal meninges before and after addition of S. aureus toxins or live S. aureus bacteria. These drugs were studied because they represent very different hydrophobicities and because earlier studies from this laboratory suggest that hydrophilic drugs penetrate the meninges via a different mechanism than do hydrophobic drugs (8-10). In addition, the meninges have been shown to be capable of metabolizing several drugs (1 l, 12), but not mannitol or sufentanil.

Methods Studies were approved by the University of Washington Animal Care and Use Committee and guidelines of the American Association for Accreditation of Laboratory Animal Care were followed throughout.

Tissue Preparation Monkey (M. nemestrina) tissues were obtained from animals scheduled to be killed as part of the tissue distribution program of the University of Washington Regional Primate Research Center. All animals (n = 26) were anesthetized with thiopental and ketamine before removal of the meningeal specimens. The spinal cords were exposed from T5 to L5 by laminectomy. The spinal cord was removed en bloc and all three meningeal layers were carefully removed from the spinal cord preserving their normal anatomic relationships. From this sheet of intact meningeal tissues, specimens measuring approxi-

Bernards et al. 25

mately 4 cm 2 (2-4 specimens per animal) were cut for mounting in the diffusion cell.

Flux Measurements The intact spinal meninges were placed between two halves of a temperature-controlled (37°C) diffusion cell. Ten milliliters of mock cerebrospinal fluid (CSF) (NaC1 140 mEq, NaHCO 3 25 mEq, MgCI2 0.4 mEq, urea 3.5 mEq, gIucose 4.0 mEq, CaC12 2.0 mEq; pH = 7.38-7.42; 292-298 mOsm) were placed in the fluid reservoirs on either side of the meningeal tissue. Air and CO x (5%) were bubbled through each fluid reservoir to maintain normal pH and to provide oxygen to the meningeal cells. After allowing 20 minutes for the chambers to equilibrate to 37°C, one or two study drugs and the corresponding 3H and/or 14C-labeled radiotracer were added to the donor reservoir on the dura mater side of the diffusion cell. In the toxin experiments, the flux of two different drugs were measured simultaneousIy. The drugs studied were mannitol (2.6 /xM) and sufentanil (2.6 /xM). The radiotracers used were 14C-mannitol (specific activity 56.7 mCi/mmol; radiochemical purity 98.4%, New England Nuclear) and 3H-sufentanil (specific activity 9 Ci/mmoI; radiochemical purity 99%, Janssen Pharmaceutica, Belgium). After adding the study drugs and radiotracers, 100/~L samples were removed from the donor and recipient reservoirs at 5-minute intervals for 60 minutes. The samples were placed in borosilicate scintillation vials for later scintillation counting to determine drug concentration. These experiments served as the baseline measurements of drug flux against which drug flux in the presence of bacterial toxins or live S. aureus bacteria was compared.

Toxin Experiments After 60 minutes, the reservoirs were emptied, rinsed with buffered saline, and refilled with mock CSF. Then 30 or 300 nM S. aureus toxins (Sigma Chemical Co., St. Louis, MO) A, B, or F (Toxic Shock Syndrome Toxin- l) were added to the donor reservoir and the tissue was incubated for 4 hours. At approximately 300 minutes, the study drugs and radiotracers were again added to the donor side of the ceil, and 100-/xL samples were collected from both reservoirs at 5-minute intervals for 60 minutes. These samples were also placed in borosilicate scintillation vials for later scintillation counting to determine drug concentration. In ten experiments, the cells were rinsed as described above and the experiment repeated after

26

Regional Anesthesia and Pain Medicine Vol. 24 No. 1 January-February 1999

adding saline instead of a S. a u r e u s toxin. These experiments served as time controls.

Bacterial Experiments In the second set of experiments (n = 7), the baseline meningeal permeability of sufentanil was determined over 60 minutes, and the diffusion cells were t h e n rinsed as described above. The diffusion cells were t h e n refilled with fresh p r e w a r m e d (37°C) mock CSF, and the d o n o r reservoir was inoculated with 200/xL of a live stationary phase S. a u r e u s bacterial suspension containing 109-10~1 cfu/mL. This bacterial suspension was an isolate obtained from a patient w h o developed an epidural abscess after placement of Harrington Rods. Thus, the wild-type organism used for this study clearly expresses a p h e n o t y p e capable of causing epidural space infection in humans. After a 4 - h o u r incubation, sufentanil and its corresponding radiotracer were added to the d o n o r side, and 100-/,L samples were collected from both reservoirs at 5-minute intervals for 60 minutes to determine the effect of S. a u r e u s on the meningeal permeability of sufentanil. At the end of the experiment, the m o c k C SF was cultured to verify that the bacteria were still viable. As a time control, seven additional experiments were performed as described above, except that S. a u r e u s bacteria were not added to the donor reservoir.

Calculation of Permeability Coefficients Drug flux was determined from drug concentration data by plotting the a m o u n t of drug in the recipient reservoir at each time point. The slope of the line relating concentration versus time data was determined by least squares linear regression and is equal to the test drug's flux t h r o u g h the meninges. Because of the unavoidable delay in reaching steady-state flux following addition of the study drugs, flux was determined from the samples collected b e t w e e n 20 and 60 minutes (before adding toxin or bacteria) and b e t w e e n 315 and 360 minutes (after adding toxin or bacteria), respectively. Thus, the reported flux values represent steady state and not initial flux conditions.

counting mock CSF w i t h o u t added radiotracer and was subtracted from the deputations per m i n u t e of each sample. After converting depurations per m i n u t e to millimoles, linear regression was used to determine drug flux (millimoles/minute) before and after adding the different toxins or bacterial suspension.

Statistical Analysis Differences in drug flux before and after addition of toxins were assessed by two-way, repeated measures, analysis of variance (ANOVA) with time and toxin as the factors. W h e n ANOVA yielded a P value of less t h a n .05, Dunnett's test was p e r f o r m e d as a post-hoe test to compare the flux after toxin with the control flux. Power analysis was performed for all comparisons. Differences in sufentanil flux with and w i t h o u t live S. a u r e u s bacteria were assessed using a two-tail, paired t-test. Differences for all comparisons were considered statistically significant at the P -< .05 level. All data are reported as m e a n _+ standard error.

Results There were no differences in baseline flux bet w e e n the control groups and any of the experimental groups for either drug or either t r e a t m e n t (toxin or live bacteria). Table 1 shows the flux of mannitol and sufentanil before and after adding a toxin or saline control. Administration of S. a u r e u s toxin A significantly increased mannitol flux t h r o u g h the spinal m e n i n ges (power = 80%), but did not affect sufentanil flux (power = 84%). Toxin B had n o statistically significant effects on sufentanil (power = 91%) or mannitol flux (power = 5 %). Similarly, toxin F did not alter sufentanil (power = 16%) or mannitol flux (power = 9%). After a 4 - h o u r incubation with live S. a u r e u s bacteria, transmeningeal flux ot sufentanil increased a statistically significant 115% from 2.8 + 0.3 to 5.8 _+ 0.7 p M / m i n / c m 2 (P = .032). In contrast, the control group increased from 3.8 _+ 0.9 to 4.3 _+ 0.9 p M / m i n / c m 2 which was not significant.

Drug Analysis Hydrofluor scintillation fluid (5-10 mL) was added to each sample, and the samples were c o u n t e d in a Packard liquid scintillation counter (Tricarb 2000) for 10 minutes or until the standard deviation of depurations per m i n u t e was -<2%. Background radioactivity was determined by

Discussion The goal of this study was to determine w h e t h e r bacteria or its toxins alter the drug p e r m e ability characteristics of spinal meningeal tissue. S t a p h y l o c o c c u s a u r e u s was chosen because it is the S. a u r e u s

Staphylococcus aureus Bacteria Effects on Meningeal Permeability

•

Bernards et al.

27

Table 1. Mannitol- and Sufentanil-Fluxes Before and After S. aureus Enterotoxin Administration Mannitol Flux Toxin

Sufentanil Flux

pre

4h

pre

4h

Individuals

30 n M 300 n M

4 . 2 6 _+ 0.68 3.96 _+ 0.37

8.82 _+ 0.80* 8.50 ± 0.82*

3.87 ± 0.19 4 . 2 7 ± 0.32

3.81 ± 0 . I 7 3.45 ± 0.43

n = 5 n = 5

30 n M 300 n M

4.01 _+ 0.92 3.77 ± 1.07

6.43 -+ 1.23 6.26 -+ 1.94

5.10 ± 0.45 4 . 6 7 ± 0.81

6,57 -+ 0.95 3,26 -+ 0 . 4 8

n = 5 n = 5

30 n M 300 n M Control

3.71 ± 1.63 2 . 5 4 -+ 0.75 3.39 ± 0 . 7 0

6.99 ± 2 . 7 2 4 , 3 6 -+ 1.26 5.25 _+ 1,10

4 . 8 7 _+ 0.66 3.18 ± 0 . 6 4 4.35 ± 0.38

7.11 ± 1.12 4 . 6 8 -+ 0.96 5.31 _+ 0.53

n = 5 n - 7 n = 10

A

B

F

F l u x = p M / m i n / c m 2. * P < 0.5. V a l u e s a r e m e a n _+ SE.

most c o m m o n cause of epidural space infections following epidural catheterization (5). The arachnoid m a t e r is the principal permeability barrier presented by the spinal meninges, accounting for more t h a n 90% of the meningeal resistance to drug diffusion (7). The arachnoid is composed of multiple overlapping layers of flattened epitheliallike cells connected to one a n o t h e r by frequent tight junctions and occluding junctions. Thus, m o v e m e n t t h r o u g h the arachnoid requires that drug molecules traverse multiple cell m e m b r a n e s or alternatively pass b e t w e e n arachnoid cells by traversing the tight junctions. Bacterial toxins m a y impact meningeal permeability by altering either of these diffusion routes. Staphylococcus aureus alpha-toxin (toxin A) was the first bacterial toxin to be identified as a m e m brane pore f o r m e r (13), and it has b e e n s h o w n to permeabilize rabbit portal veins (14) and to cause protracted vascular leakage in perfused rabbit lungs (15). Thus, it is not surprising that toxin A increased the permeability of the arachnoid mater. Although it is u n k n o w n w h y the presence of pores increased mannitol flux, possibilities include mannitol m o v e m e n t t h r o u g h pores or disruption of cell to cell contacts which normally exclude the m o v e m e n t of hydrophilic substances b e t w e e n cells. We speculate that sufentanil's flux was not altered because it traverses the arachnoid cell m e m b r a n e s by preferentially partitioning into the hydrophobic cell m e m b r a n e , thus the opening of hydrophilic pathways does not alter its flux. Staphylococcus aureus beta-toxin (toxin B) is an e n z y m e which belongs to a class k n o w n as sphingomyelinases (16,17). Toxin B hydrolyses both sphingomyelin and lysophosphatidylcholine. Hydrolysis of sphingomyelin, which is a p r o m i n e n t c o m p o n e n t of m a n y cell membranes, is responsible for toxin B's ability to lyse red blood cells (16).

However, toxin B had no effect on meningeal permeability of the studied drugs. This is perhaps surprising given the ability of toxin B to lyse some cell types. Potential explanations for the absence of effect include use of an inadequate dose a n d / o r inadequate study duration. However, the doses chosen and the duration of exposure are as great or greater t h a n those s h o w n to be effective for S. aureus toxins in other in vitro systems (13-15,18,19). Perhaps a more likely explanation is that meningeal tissue does not express the proteins or glycosphingolipids to which toxin B binds. Toxin F (also k n o w n as TSST-1) is an exotoxin responsible for toxic shock syndrome. Toxin F produces toxic shock s y n d r o m e by stimulating T-cells to elaborate large a m o u n t s of cytokines including t u m o r necrosis factor, interferon-gamma, and interleukin-2 (20,21). Because the in vitro model used in this study does not include T-cells, it is not surprising that this toxin had no effect o n permeability. It should be n o t e d that the statistical p o w e r for analyses involving toxin F's effect on mannitol and sufentanil flux, and toxin B's etfect on mannitol flux was low (Table 1). Thus, it must be kept in mind that it is possible that a type II error was made, i.e., that there really is an effect of these toxins but that we did not perform a sufficient n u m b e r of studies to detect the difference. In contrast, for experiments with toxin A's effect o n mannitol and sufentanil flux and toxin B's effect on sufentanil flux, the statistical p o w e r was high (->80%); therefore, we are confident of our conclusions regarding them. Interestingly, the variability in the flux measurements was m u c h greater for mannitol (average coefficient of variation = 53 _+ 7%) t h a n for sufentanil (average coefficient of variation = 29 + 4%). This difference in variability is not the result of technical errors (e.g., sampling errors, measure-

28

Regional Anesthesia and Pain Medicine Vol. 24 No. 1 January-February 1999

m e n t errors), because the flux of b o t h drugs was m e a s u r e d simultaneously, thus a n y technical errors w o u l d be identical for b o t h drugs. Thus, w e suspect that the variability is "real," i.e., a p r o p e r t y of the m e n i n g e a l tissue and not an e x p e r i m e n t a l artifact. Although it is not clear w h y the variability is so m u c h greater for mannitol, we suspect that it reflects differences in the routes a n d m e c h a n i s m s by w h i c h hydrophilic a n d h y d r o p h o b i c drugs traverse the meninges. Live S. aureus bacteria also increased the m e n i n geal permeability of sufentanil. Although the m e c h a n i s m for this increased m e n i n g e a l permeability is not clear, toxin elaboration or bacterial invasion of arachnoid cells a n d resultant cell d e a t h or breakd o w n of tight junctions b e t w e e n arachnoid cells are possible m e c h a n i s m s . Staphylococcus aureus produces multiple toxins in addition to those investigated in this study w h i c h are capable of causing tissue d a m age by a variety of m e c h a n i s m s including m e m b r a n e pore f o r m a t i o n causing cell lysis, m e m b r a n e lipid disruption, b a s e m e n t m e m b r a n e disruption, and m a n y others. Thus, a l t h o u g h the individual toxins investigated in this study w e r e chosen because t h e y are a m o n g the best characterized a n d m o s t p o t e n t S. aureus toxins, t h e y are by n o m e a n s the only toxins produced. A l t h o u g h we do not k n o w w h i c h toxins w e r e elaborated b y the organism used for this study, w e do k n o w that the organism is highly pathogenic in h u m a n s because it was isolated f r o m an epidural abscess. Thus, we believe the results are qualitatively predictive of the in vivo effect of live S. aureus bacteria. The concentration of S. aureus used for these studies is greater t h a n the range used for in vivo rabbit models of meningitis (105-108 cfu/mL). This was done because the 4 - h o u r duration of these studies did not allow for as m u c h bacterial g r o w t h as occurs over days in in vivo models. However, the duration of exposure is adequate to allow for S. aureus a d h e r e n c e and toxin elaboration (Ann E. Stapleton, M.D., u n p u b l i s h e d observations). I m p o r tantly, m o c k CSF cultured at the e n d of experim e n t s grew S. aureus colonies in n u m b e r s c o m p a rable to the initial inoculations, thus indicating that the bacteria w e r e still viable and w e r e n o t killed b y the conditions of the study. An i m p o r t a n t c o m p o n e n t of epidural space infection not included in this in vitro investigation is the presence of a n intact i m m u n e system. The inflamm a t i o n that attends bacterial infection in vivo and the resultant release of cytotoxic lysosomal contents a n d i n f l a m m a t o r y mediators f r o m host white blood ceils m a y produce quantitatively different results in vivo. For example, in the intact organism,

toxin F w o u l d be able to induce p r o d u c t i o n and release of t u m o r necrosis factor, i n t e r f e r o n - g a m m a , and interleukin-2 f r o m T-cells. Thus, the effects of epidural space infection on drug permeability in vivo m a y well be w o r s e t h a n w h a t w e h a v e d e m o n strated in vitro. In conclusion, D u P e n et al. (5) n o t e d decreased epidural analgesia a n d a c o n s e q u e n t n e e d to increase the dose of epidural opioid coincident w i t h epidural space infection. Our results suggest that this clinical p h e n o m e n o n is not the result of decreased m e n i n g e a l permeability. Other m e c h a n i s m s m u s t be i n v o k e d to explain DuPen's clinical observation e.g., increased pain sensation secondary to inflammation, decreased drug bioavailability because of nonspecific binding to material in the abscess or to changes in local pH, or increased drug clearance secondary to h y p e r e m i a induced b y inflammation.

References 1. Barreto RS. Bacteriological cultures of indwelling epidural catheters. Anesthesiology 1962: 23: 643641. 2. Hunt JR, Rigor BM, Collins J. The potential for contamination of continuous epidural catheters. Anesth Analg 1977: 56: 222-225. 3. Pegues DA, Carr DB, Hopkins CC. Infectious complications associated with temporary epidural catheters. Clin Infect Dis 1994: i9: 970-972. 4. Cooper AB, Sharpe MD. Bacterial meningitis and cauda equina syndrome after epidural steroid injections. Can J Anaesth 1996: 43: 471-474. 5. Du Pen SL, Peterson DG, Williams A, Bogosian AJ. Infection during chronic epidural catheterization: Diagnosis and treatment. Anesthesiology 1990: 73: 905-909. 6. Kindler CH, Seeberger MD, Staender SE. Epidural abscess complicating epidural anesthesia and analgesia. An analysis of the literature. Acta Anaesthesiol Scand i998: 42(6): 614-620. 7. Bernards C, Hill H. Morphine and alfentanil permeability through the spinal dura, arachnoid and pia mater of dogs and monkeys. Anesthesiology 1990: 73: i214-1219. 8. Bernards C. Effect of (hydroxypropyl)-/3-cyclodextrin on the flux of morphine, fentanyl, sufentanil, and alfentanil through the spinal meninges of the monkey. J Pharm Sci 1994: 83: 620-622. 9. Bernards C, Kern C. Palmitoyl camitine increases the transmeningeal flux of hydrophilic but not hydrophobic compounds in vitro. Anesthesiology 1996: 84: 392-396. 10. Ummenhofer WC, Bernards CM. Acylcarnitine chain length influences carnitine-enhanced drug flux through the spinal meninges in vitro. Anesthesiology 1997: 86(3): 642-648.

Staphylococcus aureus Bacteria Effects on Meningeal Permeability • 11. Kern C, Bemards CM. Ascorbic acid inhibits spinal meningeal catechol-o-methyl transferase in vitro, markedly increasing epinephrine bioavailability. Anesthesiology I997: 86(2): 405-409. I2. Ummenhofer WC, Brown SM, Bernards CM. Acetylcholinesterase and butyrylcholinesterase are expressed in the spinal meninges of monkeys and pigs. Anesthesiology 1998: 88(5): I 2 5 9 - 1 2 6 5 . 13. Bhakdi S, Tranum-Jensen J. Alpha-toxin of Staphylococcus aureus. Microbiol Rev 1991: 55: 733-75I. 14. Trinlde-Mulcahy L, Siegman M J, Butler TIM. Metabolic charaaeristics of alpha-toxin-permeabilized smooth muscle. Am J Physiol 1994: 266: 1673-1683. 15. W a l m r a t h D, Griebner M, Grimminger F, Galanos C, Schade U, Seeger W. Endotoxin primes perfused rabbit lungs for enhanced vasoconstrictor response to staphylococcal alpha-toxin. A m Rev Respir Dis 1993: 148: 1179-1186. 16. Walev I, Weller U, Strauch S, Foster T, Bhakdi S. Selective killing of h u m a n monocytes and cytokine release provoked by sphingomyelinase (beta-toxin) of Staphylococcus aureus. Infect I m m u n 1996: 64(8): 2974-2979.

Bernards et al.

29

17. Dziewanowska K, Edwards VM, Deringer JR, Bohach GA, Guerra DJ. Comparison of the beta-toxins from Staphylococcus aureus and Staphylococcus intermedius. Arch Biochem Biophys 1996: 335(1): 102-108.

18. Lee P, Vercellotti GM, Deringer JR, Schlievert PM. Effects of staphylococcal toxic shock syndrome toxin 1 on aortic endothelial cells. J Infect Dis 1991: 164: 711-719. 19. Jonas D, Walev I, Berger T, Liebetrau M, Palmer M, Bhakdi S. Novel path to apoptosis: Small transm e m b r a n e pores creased by staphylococcal alphatoxin in T lymphocytes evoke internucleosomal DNA degradation. Infect I m m u n 1994: 62: 1 3 0 4 1312. 20. Saha B, Jaklic B, Harlan DM, Gray GS, June CH, Abe R. Toxic shock syndrome toxin-1-induced death is prevented by CTLA41g. J I m m u n o l 1996: 157(9): 3869-3875. 21. Wahlsten JL, Ramakrishnan S. Separation of function between the domains of toxic shock syndrome toxin-1. J I m m u n o l 1998: 160(2): 8 5 4 - 8 5 9 .