correlation of histopathologic and bacteriologic changes ... · fections, including septic shock...

INFECTION AND IMMUNITY,0019-9567/97/$04.0010

Sept. 1997, p. 3889–3895 Vol. 65, No. 9

Copyright © 1997, American Society for Microbiology

Correlation of Histopathologic and Bacteriologic Changes withCytokine Expression in an Experimental Murine Model

of Bacteremic Staphylococcus aureus InfectionLEI YAO,1† JOAN W. BERMAN,1,2 STEPHEN M. FACTOR,1,3 AND FRANKLIN D. LOWY2,3,4*

Department of Medicine, Montefiore Medical Center,4 and Departments of Medicine,3 Pathology,1 andMicrobiology and Immunology,2 Albert Einstein College of Medicine, Bronx, New York

Received 24 March 1997/Returned for modification 30 April 1997/Accepted 4 June 1997

Staphylococcus aureus infections are often life threatening. Relatively little is known about the host responseto these infections, in particular, the role played by cytokines. We established a mouse model of bacteremic S.aureus infection to correlate bacteriologic findings and pathologic changes with cytokine gene expression.Bacterial density in blood and tissue was highest at 1 h and minimal by 48 h. Despite the rapid clearance ofbacteria, pathologic abnormalities and inflammatory cytokines were detected after clearance of the bacteria.The number of infiltrating inflammatory cells, as well as the size of inflammatory foci, increased with time.Interstitial accumulation of inflammatory cells and tissue damage, such as microabscesses, edema, andnecrosis progressed following clearance of bacteria from the tissues. Levels of tumor necrosis factor andinterleukin-1 protein in serum were detectable at 1 h and peaked at 4 h. Interleukin-6 protein expressionshowed different kinetics, with low levels detected at 1 h and increasing levels at 72 h postinfection. Tumornecrosis factor and the interleukins were expressed in inflammatory and noninflammatory cells in lung, liver,and heart tissues. Leukocytes in the infected tissues were highly reactive with antibodies to the three cytokines,suggesting that activated leukocytes are a major source of inflammatory cytokines after staphylococcal infec-tion. Expression of interleukin-1 and interleukin-6 in tissue-specific cells and endothelial cells was alsodetected in infected tissues, indicating that cells other than leukocytes contribute to the elevated cytokine levelsin this model. Once initiated, expression of inflammatory cytokines contributes to the pathogenesis of S. aureusdisease.

Despite the availability of effective antimicrobial agents,Staphylococcus aureus continues to cause life-threatening in-fections, including septic shock (2, 24, 33). While a great dealis known about the inflammatory profile that occurs in re-sponse to endotoxin during gram-negative bacterial infections,surprisingly little is known about the host response to staphy-lococcal septicemia or the bacterial components that initiatethis response (2, 32). It has been reported that levels of thecytokines tumor necrosis factor (TNF), interleukin-1b (IL-1),IL-6 and IL-8 in serum are correlated with the severity ofsystemic bacterial infections in humans (4, 5, 15). A similarpattern of cytokines can be elicited by endotoxin alone inexperimental animal models of infection (9, 18). A limitednumber of studies have examined the cytokine response toS. aureus infections (16, 21, 31).

In vitro studies have demonstrated that cytokine geneexpression is induced in monocytes following infection withS. aureus cellular components, including peptidoglycan, lipo-teichoic acid, or enterotoxin (1, 10, 23, 27, 28). Endothelialcells express IL-1, IL-6, and IL-8 following S. aureus infection(35, 36).

In the present study, we examined the cytokines most closelyassociated with mortality—TNF, IL-1, and IL-6. The purposeof this investigation was to demonstrate that an experimentalmouse model of bacteremic S. aureus infection could be used

to correlate cytokine activity in serum and tissue with his-topathologic findings and the density of bacteria in blood andtissue.

MATERIALS AND METHODS

Preparation of bacteria. Wb, a clinical isolate from a patient with endocarditisused in previous studies, was selected for this study (13, 29, 36). Bacteria werestored in nutrient broth (Difco Laboratories, Detroit, Mich.)–15% glycerol at270°C and subcultured on blood agar. Fresh colonies were inoculated intoTodd-Hewitt broth (BBL, Cockeysville, Md.) and grown overnight at 37°C. Thebacteria were collected by centrifugation, washed, and resuspended in saline ata concentration of 5 3 107 CFU/ml measured spectrophotometrically at a wave-length of 620 nm prior to injection. The bacterial inoculum was confirmed bycolony counting.

Animals and infection. Groups of at least four Swiss-Webster female mice (8to 10 weeks old) were used. Mice were injected intravenously via the tail veinwith Wb (107 CFU/mouse) in 0.2 ml of saline. Control mice were injected withthe same volume of saline. Three independent sets of experiments were per-formed. All procedures were approved by the institutional animal use and carecommittee. Animal housing and care were in accord with the guidelines estab-lished by the National Institutes of Health.

Collection of blood and tissues for determination of bacterial density andcytokine levels. At different times (1, 3, 6, 24, 48, and 72 h) after injection, micewere anesthetized by inhalation of Metofane (methoxyflurane; Pitman-Moore,Mundelein, Ill.) and blood was collected. The blood from each infected mousewas plated on heart infusion agar (BBL). Additional blood from the same animalwas centrifuged at 16,000 3 g and 4°C for 8 min. Serum was collected and storedat 220°C for cytokine assays. After bleeding, the mice were sacrificed. Tissuefrom the liver and spleen was collected, weighed, dounced, diluted with saline,and plated into heart infusion agar. The plates were maintained at 37°C for 48 h,and the bacterial colonies were counted. Results were expressed as numbers ofbacterial CFU per 100 ml of blood or milligram of tissue.

Determination of cytokine levels in serum. IL-1 and TNF proteins from thesera of infected mice were assayed by enzyme-linked immunosorbent assay(ELISA) (ELISA kit; Endogene, Cambridge, Mass.). IL-6 protein from sera ofWb-infected mice was assayed by using the IL-6-dependent cell line 7TD1 (ob-tained from J. Van Snick, Ludwig Institute for Cancer Research, Brussels, Bel-gium).

* Corresponding author. Mailing address: Department of Medicine,Division of Infectious Diseases, Montefiore Medical Center, 111 East210th Street, Bronx, NY 10467. Phone: (718) 920-5583. Fax: (718)920-2746. E-mail: [email protected].

† Present address: FDA/CBER, Bethesda, MD 20892.

3889

on April 7, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

http://iai.asm.org/

Collection of tissue for pathologic and immunocytochemical analyses. Follow-ing sacrifice, tissues were rapidly removed, immersed in cold PLP (4% parafor-maldehyde–0.05 M lysine in phosphate-buffered saline); and fixed at 4°C for 3 h.Fixed tissues were transferred to 30% sucrose in phosphate-buffered saline andstored at 4°C overnight. The tissues were then embedded in paraffin blocks andsectioned at a 4 to 5-mm thickness. The sections were mounted onto regularslides for staining with hematoxylin and eosin (H & E) or Gram stain and ontopolylysine-L-coated slides (Sigma) and stored at room temperature until used forimmunocytochemical analysis.

Gram and H & E staining. Gram and H & E staining was performed bystandard methods. The nature and degree of inflammation were graded byS.M.F. on a scale of 0 to 1111 (0, no inflammation; 1, focal interstitialinflammation; 11, more diffuse interstitial inflammation; 111, intense inter-stitial inflammation or microabscesses; 1111, more extensive abscess forma-tion with tissue necrosis). The pathologist was unaware of the treatment group.

Immunocytochemical analysis. Paraffin sections were deparaffinized in xyleneand rehydrated and graded through ethanol (100, 90, 70, and 50%). The sectionswere blocked by incubation with 2% goat serum in Tris-buffered saline (TBS) at37°C for 1 h. The sections were then incubated with primary antibody in TBScontaining 2% goat serum at room temperature for 2 h. After washing threetimes in TBS, a biotinylated secondary antibody (1:500 dilution, 1 mg/ml; VectorLaboratories, Burlingame, Calif.) was applied, followed by a streptavidin-alka-line phosphatase complex (Vectastain-Alkaline Phosphatase ABC kit; VectorLaboratories). The sections were developed by One-Step NBP/BCIP (Pierce,Rockford, Ill.). Sections were examined by light microscopy. The pattern ofcytokine expression was evaluated for intensity of staining. Primary antibodieswere diluted in TBS containing 2% goat serum, and the concentrations usedwere as follows: polyclonal rabbit anti-mouse TNF (Genzyme), 1:200; polyclonalrabbit anti-mouse IL-1 (2 mg/ml; Genzyme), 1:100; polyclonal rabbit anti-mouseIL-6 (1 mg/ml; Genzyme), 1:50. Immunoglobulin G for rabbit polyclonal anti-bodies (1 mg/ml; Endogen Inc.) was used as a negative control for all rabbitantibodies.

RESULTS

Bacterial density in blood and tissues following intravenousinjection of S. aureus. Mice were given an intravenous injectionof 107 CFU of S. aureus (Wb). Clinically, the infected micebecame inactive, huddled together in the cage, and were sweat-ing; there were no deaths during the 72-h period of analysis.Blood samples were collected at different time periods (15 minto 72 h). Bacteria were rapidly cleared from the blood, with thegreatest bacterial density observed 15 min after injection. By48 h, few bacteria were cultured from the blood (Fig. 1A). Theconcentrations of S. aureus cultured from the liver and spleenof the infected mice were similar. Bacterial clearance from thetissues was also rapid, with few bacteria detected after 4 h (Fig.

1B). No cultures from the blood or tissues were sterile at 1 or4 h. The blood was sterile (#10 colonies/100 ml) in one of fivemice at 24 h and in five of five mice at 72 h. The liver andspleen were sterile in two of five mice (#5 colonies/mg) at 24 hand in five of five mice at 48 h.

Histopathologic changes following S. aureus intravenous in-fection. Tissue sections of the lung, heart, and liver collectedfrom mice at different time points after intravenous staphylo-coccal injection (1, 4, 24, 48, and 72 h; four or five animals pertime point) were examined (Fig. 2). The pattern and degree oftissue damage varied from animal to animal. Some tissue sec-tions displayed focal damage, while others showed a diffuseinflammatory response. Increased accumulation of inflamma-tory cells in the lung, heart, and liver was observed by 4 h afterinfection. These changes were noted in the liver at 1 h. Whilethe heart showed less leukocyte accumulation at 4 h than thelung and liver, extensive inflammation was detected by 24 h(three animals) with abscess formation. Figure 3 shows H & Eand Gram staining of the same section containing an abscesswith a large collection of bacteria in the myocardium.

Tissue injury was more advanced at 48 and 72 h in the lung,heart, and liver (Fig. 2, 48 h). The histopathologic changesincluded leukocytes migrating away from vessels into the tissueparenchyma, accumulation of inflammatory cells in the inter-stitial spaces, and evidence of interstitial edema in the lung,heart, and liver. The presence of a large number of dispersedleukocytes, as well as aggregates of inflammatory cells, in thesetissues increased with time (Fig. 2).

Small foci of inflammatory cells or microabscesses were seenin the livers at all of the time points examined. Vacuolizationindicative of tissue injury was seen in the liver at 24 h. Ab-scesses were present in the infected hearts after 24 h (data notshown). The damage to the lungs was extensive, with focalpneumonitis and interstitial and intraalveolar edema beginningat 24 h and detected at 48 and 72 h (Fig. 2).

In summary, inflammatory cells were first seen within vesselsand in the adjoining parenchymal tissue. Other changes, in-cluding interstitial edema, formation of microaggregates ofleukocytes within the tissue parenchyma, and tissue necrosis,

FIG. 1. Bacteremia and bacterial density in tissue following S. aureus infection. (A) Bacteremia. Blood samples were collected from Wb-infected mice and platedinto agar at different time points after injection. (B) Bacterial density in tissue. Tissues were processed as described in Materials and Methods. The values are themean 6 the standard error of the mean from three different experiments, and each experimental group contained at least four mice.

3890 YAO ET AL. INFECT. IMMUN.


http://iai.asm.org/

Dow

nloaded from

http://iai.asm.org/

were seen at 48 h or later. The degree of tissue inflammationand damage in the different tissues is shown in Table 1.

Cytokine levels in serum after intravenous infection with S.aureus. Blood samples from S. aureus-infected mice were as-sayed for levels of TNF, IL-1, and IL-6 in serum. Levels of

TNF and IL-1 in serum were highest 4 h after injection anddecreased thereafter (Fig. 4). Low levels of TNF were stilldetected in the serum after 24 h (11 ng/ml). A rapid decreaseof IL-1 in serum was observed at 24 h (7 ng/ml). IL-6 levels inserum differed from those of TNF and IL-1. IL-6 was detected

FIG. 2. Histopathologic changes in S. aureus-infected lung, heart, and liver tissues. H & E-stained sections of infected and uninfected tissues are shown. Thephotomicrographs represent histological sections of lung, heart, and liver tissues from uninfected (uninf) and S. aureus-infected mice at different time points (4, and48 h) postinfection. The pathologic changes are described in the text. Tissue sections from uninfected mice are included for comparison. Arrows: liver at 4 h, necrotichepatocytes; liver at 48 h, aggregation of inflammatory cells around vessels; lung at 48 h, extensive edema. Original magnification, 3105.6. Bar, 100 mm.

FIG. 3. Gram and H & E staining of histological sections of an infected heart illustrating abscess formation. On the left is a gram stained section from an infectedheart. The photomicrograph illustrates a cluster of bacteria in the heart 24 h after injection of S. aureus. Original magnification, 342.2. (B) Comparable H & E stainingof the same section. Original magnification, 342.2. Bar, 100 mm.

VOL. 65, 1997 MURINE MODEL OF S. AUREUS INFECTION 3891


http://iai.asm.org/

Dow

nloaded from

http://iai.asm.org/

as early as 1 h after injection and continued to increase until72 h, the last time point assayed (Fig. 4).

Expression of cytokines in tissue after intravenous infectionwith S. aureus. Lung, heart, and liver tissue sections frominfected mice (four animals at each time point) were examinedby immunocytochemistry to assess cytokine gene expression inthe tissues. The staining patterns and intensity of expression ofthese cytokine proteins varied in the different tissues and at thedifferent time points. Expression of all three cytokines wasobserved in all of the infected tissues examined. IL-1, TNF,and IL-6 proteins were not detected in any of the uninfectedtissues at the time points tested. Figure 5 (uninfected) showsan example of uninfected tissue treated with antibodies toTNF. Additionally, the infected tissues from different timepoints exhibited no reactivity with a negative control rabbitimmunoglobulin G antibody (data not shown).

Patterns of TNF expression in tissues. TNF expression (bydifferent cell types) was noted in all tissue sections of theinfected lungs, hearts, and livers (Fig. 5). TNF protein wasdetected in all of the lungs at 1 h. Low levels of TNF were alsopresent in the heart and liver at this time point. The mostintense reactivity for TNF was observed at 24 h. Althoughexpression was decreased by 48 and 72 h, positive TNF reac-tivity of inflammatory cells and tissue-specific cell types, suchas Kupffer cells and alveolar lining cells, was still detected atthese time points. TNF-expressing inflammatory cells were de-tected in blood vessels and tissue spaces at all time points (Fig.5). Intense TNF reactivity of inflammatory cells within thealveoli and alveolar lining cells was detected after 4 h (Fig. 5,TNF, lung). Expression of TNF by intravascular inflammatoryand interstitial cells in the heart was also observed (Fig. 5,heart).

Patterns of IL-1 expression in tissues. IL-1, like TNF, wasdetected in all of the organs examined. IL-1 protein was firstdetected in the lung 1 h after injection. Focal expression ofIL-1 was observed in most tissues 4 h after injection, with theintensity increased at this time. Diffuse IL-1 protein expressionwas detected in most sections of the lung, heart, and liver at 24

and 48 h (Fig. 5). By 72 h after infection, IL-1 expressiondecreased in all tissues.

IL-1 expression by inflammatory, alveolar lining, and endo-thelial cells was first seen in the lung. Expression of IL-1 byendothelial cells in blood vessels was noted after infection.Although arterial endothelial cells showed no reactivity withIL-1 antibody in 4-h lung and heart samples, IL-1 in endothe-lial cells of thin-walled vessels compatible with veins was de-tected (Fig. 5, IL-1, lung). Expression of IL-1 by intravascularand interstitial inflammatory cells was present at all of the timepoints in all of the tissues (Fig. 5, IL-1, heart). Kupffer cells andsinusoidal lining cells around the central vein in the liver alsoexpressed IL-1 (Fig. 5, IL-1, liver).

Patterns of IL-6 expression in tissues. The pattern of IL-6expression was not comparable to that of either IL-1 or TNF(Fig. 5). IL-6 protein was detectable in the lung, heart, andliver as early as 1 h. Increased levels of IL-6 protein in all of thetissues were observed until 24 h and remained at similar in-tensities for up to 72 h.

Inflammatory cells within intravascular and infiltrating cells

TABLE 1. Summary of pathologic changes inS. aureus-infected tissuesa

Time (h)postinfection Organ Mouse 1 Mouse 2 Mouse 3 Mouse 4

1 Lung 0 0 0 0Heart 0 0 0 0Liver 0 1 1 111

4 Lung 11 11 11 11Heart 0 0 1 1Liver 111 11 11 11

24 Lung 111 11 11 1Heart 1111 11 111 111Liver 1 111 111 111

48 Lung 1111 111 111 NDHeart 111 111 111 11Liver 111 111 111 111

72 Lung 1111 111 11 1Heart 111 111 111 11Liver 111 11 1 1

a The results are for four animals per time point. Scale; 0, no inflammation; 1,focal interstitial inflammation; 11, more diffuse interstitial inflammation; 111,intense interstitial inflammation or microabscesses; 1111, more extensive ab-scess formation with tissue necrosis (abscess); ND, not determined.

FIG. 4. Levels of TNF, IL-1, and IL-6 proteins in serum after S. aureusinfection. Blood samples from infected mice were collected at different timepoints after injection of Wb bacteria. The sera were assayed for TNF or IL-1 byELISA (TNF and IL-1). IL-6 in serum was assayed by using the IL-6-dependentcell line 7TD1 (IL-6). Sera from uninfected (uninf) mice were used as controls.The values are the mean 6 the standard error of the mean of four animals pergroup in three independent experiments.



http://iai.asm.org/

Dow

nloaded from

http://iai.asm.org/

in all of the infected tissues were positive for IL-6 after 1 h.Like TNF and IL-1, tissue-specific cells, such as alveolar liningcells in the lung and sinusoid lining and Kupffer cells in theliver, expressed IL-6 protein (data not shown). Hepatocytesaround the central vein in the liver were positive for IL-6 (Fig.5, IL-6, liver) at 4 and 24 h. Lung tissue (48 h) exhibitedextensive edema (Fig. 2, lung) and had strong, diffuse IL-6reactivity (Fig. 5, IL-6, lung). In contrast to IL-1, intensivestaining of IL-6 was noted in the vessels, including veins andarteries, in all tissues, particularly at 48 and 72 h (data notshown).

This observation is comparable to our in vitro model, in

which expression of IL-6 by S. aureus-infected endothelial cellspersisted for 72 h (36). The correlation of pathological changesat late time points in different tissues (discussed previously)with persistent high levels of IL-6 in tissues and sera suggeststhat IL-6 plays a contributory role in maintaining the systemicinflammatory response following activation of the proinflam-matory cytokines TNF and IL-1.

DISCUSSION

S. aureus is a virulent pathogen that has the ability to causea variety of potentially life-threatening infections. These infec-

FIG. 5. Immunocytochemical staining of TNF, IL-1, and IL-6 proteins expressed in S. aureus-infected lung, heart, and liver tissues. Uninfected: reactivity ofanti-TNF antibody with uninfected lung, heart, and liver tissues. Infected: reactivity of anti-TNF, anti-IL-1, and anti-IL-6 antibodies with infected tissues. Originalmagnification, 3211.2. Bar, 100 mm. TNF: reactivity of TNF antibody with inflammatory cells and alveolar lining cells in the lung (4 h), intravascular inflammatory cellsin the heart (arrow, 24 h), and Kupffer and inflammatory cells in the liver (4 h). Original magnification, 3211.2. IL-1: reactivity of IL-1 antibody with endothelial cellsin the lung (24 h, arrow); inflammatory cells, Kupffer cells, and sinusoidal lining cells in the liver (24 h, arrow); and intravascular inflammatory cells within the heart(48 h). Original magnification, 3211.2. IL-6: diffuse reactivity of IL-6 antibody in a lung with edema (48 h; original magnification, 3105.6), hepatocytes around thecentral vein of a liver (24 h; original magnification, 3105.6), and inflammatory cells in the interstitial space in a heart (4 h, arrow). Original magnification, 3211.2.



http://iai.asm.org/

Dow

nloaded from

http://iai.asm.org/

tions vary from superficial soft tissue abscesses to septic shock(2, 32). A limited number of studies have examined the patho-genesis of gram-positive septicemia, in particular, the host re-sponse to these infections. Elevated levels of IFN-g, IL-1, andTNF in serum have been observed during staphylococcalinfection in previously reported in vivo studies (11, 16, 21, 31).However, the role of the cytokines in the pathogenesis of S.aureus infections has not been well characterized. The purposeof this study was to use an animal model of S. aureus bactere-mic infection to correlate cytokine gene expression with bac-teriologic and histopathologic findings.

Surprisingly, only a limited number of studies have examinedthe pathology of S. aureus sepsis in humans. A number ofhistopathologic observations from our study are similar to thereported findings on humans with invasive S. aureus disease.The leukocytic infiltration, focal pneumonitis, and edemafound in mice are observed in patients with systemic S. aureusinfections (14, 22, 33, 34). Myocardial abscesses are oftenfound in cases of infective endocarditis (7). Metastic seeding ofother organs with abscess formation is a common complicationof S. aureus bacteremia (22, 33). The pathological changesobserved in the livers of infected mice, such as the presence ofvacuoles and microabscesses, have also been observed in pa-tients with bacterial sepsis. Our histopathologic data suggeststhat this animal model reproduces many of the pathologicalfindings of S. aureus infections in humans and is therefore auseful model for further study of host-pathogen interactions inS. aureus-induced sepsis (14, 22, 33, 34). Other models, includ-ing those of cutaneous or prosthetic device infections, mighthave completely different kinetics of cytokine expression andtissue damage and should be examined (3, 31).

Bacteria were rapidly cleared from blood and tissues follow-ing intravenous challenge. These findings are consistent withearlier experimental animal studies, in which rapid clearanceof infected bacteria was observed in mice infected with S.aureus (25, 26). However, the relationships among bacterialclearance, tissue damage, and cytokine expression was not ex-amined. In this mouse model of S. aureus infection, we foundsignificant pathologic changes during and after the eliminationof bacteria from the blood and tissues. Extensive leukocyteinfiltration, abscess formation, and edema were noted wellafter clearance of bacteria from the blood and tissues.

Previous in vivo studies demonstrated that TNF and IL-1 areinitiators of the inflammatory cascade following bacterial in-fection and that overexpression of TNF and IL-1 plays animportant role in the development of sepsis after gram-nega-tive bacterial infection or after lipopolysaccharide (LPS) ad-ministration (6, 19). In this mouse model of S. aureus infection,we found rapid induction of TNF and IL-1 in serum and tissuefollowing infection. Rozalska and Wadstrom (21) reported asimilar pattern for levels of TNF and IL-1a in serum in theirexperimental murine model of S. aureus infection. The associ-ation of the pathologic changes in the infected tissues with theevidence of cytokine expression suggests that cytokines con-tribute to this process. Cytokine expression in the tissues cor-related with leukocyte migration into the affected tissue andwith evidence of progressive tissue damage. Leukocytes werefirst noted within vessels and subsequently in the tissue paren-chyma. As noted, the pathologic changes continued to progressfollowing clearance of bacteria from these sites.

The kinetics of TNF and IL-1 expression in different tissuesvaried. Cytokine expression appeared earlier in the lungs thanin the heart and liver, suggesting that tissues respond differ-ently to S. aureus infection. This may be associated with thebacterial density or the rate of bacterial clearance from these

tissues. It is also supported by previous studies in which S.aureus infected and survive differently in various organs (3, 20).

The result of sustained levels of IL-6 for up to 72 h in bloodalso correlated with the evolution of pathologic changes intissues. High levels of IL-6 have been observed in patients andanimals with Gram-positive bacterial sepsis (4, 16). Our invitro study also found that infection of endothelial cells with S.aureus induced large and persistent amounts of IL-6 mRNAand protein (36). The immunocytochemical results showedevidence of IL-6 in the vessels of all of the tissues, as well assinusoids in the liver, after 48 h, suggesting that endothelialcells contributed to the late production of IL-6 in vivo. Thesecells may therefore play an important role in the persistenthigh levels of IL-6 and the pathologic changes seen in thevessels.

Studies of gram-negative bacterial infections in humans orstudies of animals treated with LPS show increased levels ofinflammatory cytokines in serum as well as tissue injury (8, 17,19, 30). Although the inflammatory response after treatmentwith LPS is similar to the results of this study, the cause ofcytokine expression and the contribution to the pathologicchanges in our model are not known. CD14, the LPS receptor,is required for LPS function. Whether the same or a similarmechanism is involved in cytokine induction by S. aureus com-ponents is still under investigation (12).

The results of this study provide a baseline for subsequentstudies on the role of inflammatory cytokines in the pathogen-esis of staphylococcal diseases, including sepsis. The studydemonstrates that following infection, bacteria are rapidlycleared from the blood, liver, and spleen. Despite this, there iscontinued cytokine expression and tissue damage. While it ispossible that bacterial components, such as peptidoglycan, re-main as stimuli, it is likely that the inflammatory process, onceinitiated by bacteria, is amplified by cytokines. This is the firstanimal model used to study in detail the correlation of cytokinegene expression with S. aureus infection. It does not establisha causal relationship between the expression of cytokines andthe histopathologic changes noted. To clearly establish such arelationship, studies would need to be performed with specificanticytokine antibodies or mice with cytokine gene knockouts.Further study will focus on the cytokine gene expression in-duced by S. aureus components and prevention of some of theconsequences of gram-positive sepsis by down regulation ofinflammatory cytokine expression.

ACKNOWLEDGMENTS

We gratefully acknowledge the expert secretarial support of Evan-gelina Vega and Lynette Bradberry Watson.

F.L. was supported by a grant-in-aid from the American HeartAssociation. J.W.B. was supported by a grant-in-aid from the Ameri-can Heart Association, New York City Affiliate.

REFERENCES

1. Bhakdi, S., T. Klonisch, P. Nuber, and W. Fischer. 1991. Stimulation ofmonokine production by lipoteichoic acids. Infect. Immun. 59:4614–4620.

2. Bone, R. C. 1993. How gram-positive organisms cause sepsis. J. Crit. Care8:51–59.

3. Bunce, C., L. Wheeler, G. Reed, J. Musser, and N. Barg. 1992. Murine modelof cutaneous infection with gram-positive cocci. Infect. Immun. 60:2636–2640.

4. Calandra, T., J. Gerain, D. Heumann, J. D. Baumgartner, and M. P.Glauser. 1991. High circulating levels of interleukin-6 in patients with septicshock: evolution during sepsis, prognostic value, and interplay with othercytokines. The Swiss-Dutch J5 Immunoglobulin Study Group. Am. J Med.91:23–29.

5. Casey, L., R. A. Balk, and R. C. Bone. 1993. Plasma cytokine and endotoxinlevels correlate with survival in patients with the sepsis syndrome. Ann.Intern. Med. 119:771–778.

6. DeForge, L. E., and D. G. Remick. 1991. Kinetics of TNF, IL-6, and IL-8



http://iai.asm.org/

Dow

nloaded from

http://iai.asm.org/

gene expression in LPS-stimulated human whole blood. Biochem. Biophys.Res. Commun. 174:18–24.

7. Dressler, F. A., and W. C. Roberts. 1989. Infective endocarditis in opiateaddicts: analysis of 80 cases studied at necropsy. Am. J. Cardiol. 63:1240–1257.

8. Fischer, E., M. A. Marano, A. E. Barber, A. Hudson, K. Lee, C. S. Rock, A. S.Hawes, R. C. Thompson, T. J. Hayes, and T. D. Anderson. 1991. Comparisonbetween effects of interleukin-1 alpha administration and sublethal endotox-emia in primates. Am. J. Physiol. 261:R442–R452.

9. Hesse, D. G., K. J. Tracey, Y. Fong, K. R. Manogue, M. A. Palladino, Jr., A.Cerami, G. T. Shires, and S. F. Lowry. 1988. Cytokine appearance in humanendotoxemia and primate bacteremia. Surg. Gynecol. Obstet. 166:147–153.

10. Heumann, D., C. Barras, A. Severin, M. P. Glauser, and A. Tomasz. 1994.Gram-positive cell walls stimulate synthesis of tumor necrosis factor alphaand interleukin-6 by human monocytes. Infect. Immun. 62:2715–2721.

11. Hinshaw, L. B., T. E. Emerson, Jr., F. B. Taylor, Jr., A. C. Chang, M. Duerr,G. T. Peer, D. J. Flournoy, G. L. White, S. D. Kosanke, and C. K. Murray.1992. Lethal Staphylococcus aureus-induced shock in primates: prevention ofdeath with anti-TNF antibody. J. Trauma 33:568–573.

12. Kusunoki, T., E. Hailman, T. S. Juan, H. S. Lichenstein, and S. D. Wright.1995. Molecules from Staphylococcus aureus that bind CD14 and stimulateinnate immune responses. J. Exp. Med. 182:1673–1682.

13. Lowy, F. D., J. Fant, L. L. Higgins, S. K. Ogawa, and V. B. Hatcher. 1988.Staphylococcus aureus-human endothelial cell interactions. J. Ultrastruct.Mol. Struct. Res. 98:137–146.

14. Lyons, C. 1942. Bacteremic staphylococcal infection. Surg. Gynecol. Obstet.74:41–46.

15. Marty, C., B. Misset, F. Tamion, C. Fitting, J. Carlet, and J. M. Cavaillon.1994. Circulating interleukin-8 concentrations in patients with multiple or-gan failure of septic and nonseptic origin. Crit. Care Med. 22:673–679.

16. Matuschak, G. M., C. Munoz, N. A. Epperly, R. S. Britton, D. Walsh, D. R.Schilly, T. L. Tredway, T. A. Khan, B. R. Bacon, and A. J. Lechner. 1994.TNF-alpha and IL-6 expression in perfused rat liver after intraportal candi-demia vs. E. coli or S. aureus bacteremia. Am. J. Physiol. 267:R446–R454.

17. Matuschak, G. M., M. E. Mattingly, T. L. Tredway, and A. J. Lechner. 1994.Liver-lung interactions during E. coli endotoxemia. TNF-alpha: leukotrieneaxis. Am. J. Resp. Crit. Care Med. 149:41–49.

18. Natanson, C., R. L. Danner, R. J. Elin, J. M. Hosseini, K. W. Peart, S. M.Banks, T. J. MacVittie, R. I. Walker, and J. E. Parrillo. 1989. Role ofendotoxemia in cardiovascular dysfunction and mortality. Escherichia coliand Staphylococcus aureus challenges in a canine model of human septicshock. J. Clin. Invest. 83:243–251.

19. Remick, D. G., R. M. Strieter, M. K. Eskandari, D. T. Nguyen, M. A. Genord,C. L. Raiford, and S. L. Kunkel. 1990. Role of tumor necrosis factor-alphain lipopolysaccharide-induced pathologic alterations. Am. J. Pathol. 136:49–60.

20. Ricci, M. A., R. Mehran, N. V. Christou, F. Mohamed, A. M. Graham, andJ. F. Symes. 1991. Species differences in the clearance of Staphylococcusaureus bacteremia. J. Invest. Surg. 4:53–58.

21. Rozalska, B., and T. Wadstrom. 1993. Interferon-gamma, interleukin-1 andtumor necrosis factor-alpha synthesis during experimental murine staphylo-coccal infection. FEMS Immunol. Med. Microbiol. 7:145–152.

22. Samuelson, J., and F. von Lichtenberg. 1994. Infectious diseases, p. 335–337.In R. S. Cotran, V. Kuman, S. L. Robbins, and F. J. Schoen (ed.), Robbinspathologic basis of disease, 5th ed. The W. B. Saunders Co., Philadelphia,Pa.

23. See, R. H., and A. W. Chow. 1992. Staphylococcal toxic shock syndrome toxin1-induced tumor necrosis factor alpha and interleukin-1 beta secretion byhuman peripheral blood monocytes and T lymphocytes is differentially sup-pressed by protein kinase inhibitors. Infect. Immun. 60:3456–3459.

24. Sheagren, J. N. 1984. Staphylococcus aureus. The persistent pathogen.N. Engl. J. Med. 310:1437–1442.

25. Smith, I. M., A. P. Wilson, E. C. Hazard, W. K. Hummer, and M. E. Dewey.1960. Death from staphylococci in mice. J. Infect. Dis. 107:369–378.

26. Smith, J. M., and R. J. Dubos. 1956. The behavior of virulent and avirulentstaphylococci in the tissues of normal mice. J. Exp. Med. 103:87–107.

27. Strieter, R. M., K. Kasahara, R. M. Allen, T. J. Standiford, M. W. Rolfe, F. S.Becker, S. W. Chensue, and S. L. Kunkel. 1992. Cytokine-induced neutro-phil-derived interleukin-8. Am. J. Pathol. 141:397–407.

28. Timmerman, C. P., E. Mattsson, M. L. Martinez, G. L. De, S. J. Van, H. A.Verbrugh, J. Verhoef, and A. Fleer. 1993. Induction of release of tumornecrosis factor from human monocytes by staphylococci and staphylococcalpeptidoglycans. Infect. Immun. 61:4167–4172.

29. Tompkins, D. C., V. B. Hatcher, D. Patel, G. A. Orr, L. L. Higgins, and F. D.Lowy. 1990. A human endothelial cell membrane protein that binds Staph-ylococcus aureus in vitro. J. Clin. Invest. 85:1248–1254.

30. Ulich, T. R., K. Z. Guo, B. Irwin, D. G. Remick, and G. N. Davatelis. 1990.Endotoxin-induced cytokine gene expression in vivo. II. Regulation of tumornecrosis factor and interleukin-1 alpha/beta expression and suppression.Am. J. Pathol. 137:1173–1185.

31. Vaudaux, P., G. E. Grau, E. Huggler, F. Schumacher-Perdreau, F. Fiedler,F. A. Waldvogel, and D. P. Lew. 1992. Contribution of tumor necrosis factorto host defense against staphylococci in a guinea pig model of foreign bodyinfections. J. Infect. Dis. 166:58–64.

32. Verhoef, J., and E. Mattsson. 1995. The role of cytokines in Gram-positivebacterial shock. Trends Microbiol. 137:136–140.

33. Waldvogel, F. A. 1995. Staphylococcus aureus (including toxic shock syn-drome), p. 1754. In G. L. Mandell, J. E. Bennett, and R. Dolin (ed.),Principles and practice of infectious diseases, 4th ed. Churchill Livingstone,New York, N.Y.

34. Wollenman, O. J., Jr., and M. Finland. 1943. Pathology of staphylococcalpneumonia complicating clinical influenza. Am. J. Pathol. 19:23–41.

35. Yao, L., F. D. Lowy, and J. W. Berman. 1996. Interleukin-8 gene expressionin Staphylococcus aureus-infected endothelial cells. Infect. Immun. 64:3407–3409.

36. Yao, L., V. Bengualid, F. D. Lowy, J. J. Gibbons, V. B. Hatcher, and J. W.Berman. 1995. Internalization of Staphylococcus aureus by endothelial cellsinduces cytokine gene expression. Infect. Immun. 63:1835–1839.

Editor: V. A. Fischetti



http://iai.asm.org/

Dow

nloaded from
http://iai.asm.org/

correlation of histopathologic and bacteriologic changes ... · fections, including septic shock...

Documents