INFECTION AND IMMUNITY, Mar. 2009, p. 1008–1014 Vol. 77, No. 30019-9567/09/$08.00�0 doi:10.1128/IAI.00976-08Copyright © 2009, American Society for Microbiology. All Rights Reserved.

Defects in Innate Immunity Predispose C57BL/6J-Leprdb/Leprdb Miceto Infection by Staphylococcus aureus�

Sunny Park, Jeremy Rich, Frank Hanses, and Jean C. Lee*Channing Laboratory, Department of Medicine, Brigham and Women’s Hospital and Harvard Medical School, Boston, Massachusetts

Received 4 August 2008/Returned for modification 5 September 2008/Accepted 16 December 2008

Foot and ankle infections are the most common cause of hospitalization among diabetic patients, andStaphylococcus aureus is a major pathogen implicated in these infections. Patients with insulin-resistant (type2) diabetes are more susceptible to bacterial infections than nondiabetic subjects, but the pathogenesis of theseinfections is poorly understood. C57BL/6J-Leprdb/Leprdb (hereafter, db/db) mice develop type 2 diabetes due toa recessive, autosomal mutation in the leptin receptor. We established a S. aureus hind paw infection in diabeticdb/db and nondiabetic Lepr�/� (�/�) mice to investigate host factors that predispose diabetic mice to infection.Nondiabetic �/� mice resolved the S. aureus hind paw infection within 10 days, whereas db/db mice withpersistent hyperglycemia developed a chronic infection associated with a high bacterial burden. Diabetic db/dbmice showed a more robust neutrophil infiltration to the infection site and higher levels of chemokines in theinfected tissue than �/� mice. Blood from �/� mice killed S. aureus in vitro, whereas db/db blood was defectivein bacterial killing. Compared with peripheral blood neutrophils from �/� mice, db/db neutrophils demon-strated a diminished respiratory burst when stimulated with S. aureus. However, bone marrow-derived neu-trophils from �/� and db/db mice showed comparable phagocytosis and bactericidal activity. Our resultsindicate that diabetic db/db mice are more susceptible to staphylococcal infection than their nondiabeticlittermates and that persistent hyperglycemia modulates innate immunity in the diabetic host.

Approximately 2 million of the estimated 16 million individ-uals with diabetes in the United States will develop chronicfoot ulcers or infections during the course of their disease (38).An infection is initiated when the skin barrier is breached andbacteria, mostly skin commensals, gain access to the underlyingtissues. Although limb-threatening infections are usuallypolymicrobial, Staphylococcus aureus is a major determinant ofthese infections (21). S. aureus is the predominant pathogen innon-limb-threatening infections, particularly in patients whohave not received antimicrobial therapy (5, 26, 27). The emer-gence of S. aureus strains resistant to multiple antibiotics hasmade treatment of staphylococcal infections especially prob-lematic. Methicillin-resistant S. aureus strains have become in-creasingly prevalent among both nosocomial and community-acquired infections within the United States (22, 36, 43). Theprevalence of methicillin-resistant S. aureus is higher amongdiabetic patients than in the general population (11, 44, 45).Seven S. aureus strains resistant to vancomycin have been iso-lated in the United States, and four of these strains wereisolated from patients with diabetes (42, 50). Complications oftype 2 diabetes such as peripheral neuropathy and vasculopa-thy contribute to delayed wound healing. Although the in-creased susceptibility of the diabetic host to bacterial infectionsis well established, the chronicity of these infections is poorlyunderstood. A consistent defect in the humoral or cell-medi-ated host immune system of diabetic patients has not beendemonstrated. However, deficiencies in the host innate im-mune response are apparent since clinical investigations have

indicated that phagocytes from type 2 diabetic patients are ina heightened state of oxidative stress and have impaired bac-tericidal activity and chemotaxis (6, 15, 41, 48). Clearly, nu-merous pathophysiologic perturbations contribute to the re-currence of soft tissue and bone infections in the lowerextremity of patients with diabetes.

C57BL/6J-Leprdb/Leprdb (hereafter, db/db) mice are a valu-able model of type 2 diabetes since they are hyperglycemic andresistant to insulin, and they experience peripheral neuropa-thy, delayed wound healing, and myocardial disease. In thisstudy we inoculated the hind paws of wild-type Lepr�/� (�/�)and diabetic db/db mice with S. aureus and evaluated thecourse of the ensuing infection in each host type, as well as theresultant host innate immune response to infection. Diabeticmice that were �4 months of age were more susceptible tostaphylococcal infection than age-matched nondiabetic controlanimals. The db/db mice showed a heightened inflammatoryresponse that was characterized by defects in phagocyte func-tion.

MATERIALS AND METHODS

Mice. db/db mice develop type 2 diabetes due to a lack of a functional hypo-thalamic leptin receptor (7, 8). The animals become obese around 3 to 4 weeksof age, elevations of plasma insulin begin at 10 to 14 days of age, and blood sugarlevels start to rise at 4 to 8 weeks. Control nondiabetic mice included wild-type(�/�) and Leprdb/� heterozygous (db/�) littermates. Male mice (8 weeks of age)were purchased from The Jackson Laboratory (Bar Harbor, ME) and werehoused in a modified barrier facility under viral antibody-free conditions. Theanimals were given food and water ad libitum, and blood glucose levels weremeasured by the glucose oxidase test with glucostrips (Bayer, Elkhart, IN).Animal care and experimental procedures complied with guidelines establishedby Harvard Medical School.

Mouse model of S. aureus hind paw infection. S. aureus strain PS80 (strepto-mycin resistant) is a capsular serotype 8 isolate that has been shown to be virulentin a surgical wound infection model (32), a rat model of abscess formation (46),and a hind paw infection model described previously (39). S. aureus PS80 was

* Corresponding author. Mailing address: Channing Laboratory,181 Longwood Avenue, Boston, MA 02115. Phone: (617) 525-2652.Fax: (617) 731-1541. E-mail: jclee@rics.bwh.harvard.edu.

� Published ahead of print on 22 December 2008.

1008

on March 13, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

cultivated for 24 h at 37°C on Columbia agar supplemented with 2% NaCl. Abacterial suspension was prepared in phosphate-buffered saline (PBS), and thenumber of CFU/ml was confirmed by quantitative plate counts. Mice wereanesthetized subcutaneously with 100 mg of ketamine/kg of body weight and 10mg/kg xylazine. The left hind paw was cleansed with 70% ethanol, and a syringewith a 29.5-gauge hypodermic needle was used to inject 10 �l of the bacterialsuspension into the plantar-proximal aspect of the hind paw. At various timepoints after inoculation, the mice were euthanized by CO2 asphyxiation, and thehind paws were amputated and defleshed. The excised tendon and muscle tissueswere weighed and homogenized in tryptic soy broth with a mechanical homog-enizer. The homogenates were serially diluted and plated quantitatively ontryptic soy agar. The lower limit of detection was �1.6 log10 CFU/g of tissue.

Histological examination of hind paw tissues. Infected hind paws were excisedand fixed in formaldehyde. The tissues were decalcified, embedded in paraffin,and stained with hematoxylin and eosin. Immunohistochemical staining wasperformed on tissue sections that were deparaffinized and pretreated with 10mM sodium citrate buffer (pH 6.0). After tissue samples were rinsed in distilledwater, endogenous peroxidase activity was quenched by treatment with Peroxi-dase Block (Dako, Carpinteria, CA). Mouse neutrophils were detected with ratmonoclonal anti-mouse Ly-6G (Gr-1; BD Pharmingen, San Diego, CA) appliedat a 1:100 dilution for 1 h. After samples were washed in 50 mM Tris-Cl (pH 7.4),rabbit anti-rat antibody was applied at a 1:750 dilution for 1 h. Slides werewashed again and treated with goat anti-rabbit horseradish peroxidase-conju-gated antibody (Envision Plus detection kit; Dako). After another washing step,immunoperoxidase staining was developed with a diaminobenzedene chromogen(Dako), and nuclei were counterstained with hematoxylin. A myeloperoxidase(MPO) assay (30) was used to quantify neutrophil accumulation in the hind pawtissues.

Tissue chemokine measurements. Excised hind paw tissues were homogenizedin lysis buffer (1% Nonidet P-40, 500 mM NaCl, 50 mM HEPES, 1% Igepaldetergent, pH 7.2 to 7.4) containing a mammalian protease inhibitor cocktail(Sigma, St. Louis, MO). Soluble tissue extracts were stored at �70°C beforetesting. Concentrations of CXCL1/KC (keratinocyte-derived chemokine) andCXCL2/macrophage inflammatory protein 2 were measured by a sandwich en-zyme-linked immunosorbent assay (ELISA) from R&D Systems (Minneapolis,MN). The detection limit of the ELISA was 16 ng/ml, and the results werenormalized to the weight of the hind paw tissue samples.

Functional analyses of blood leukocytes. Heparinized blood was collected bynicking the tail vein of mice. Blood killing assays were performed in polypro-pylene tubes containing 200 �l of mouse blood and 100 �l of S. aureus to yielda final concentration of 105 CFU/ml. The samples were incubated at 37°C on arotator, and aliquots were removed for quantitative culture after 0, 60, and 120min. The data were expressed as percent survival of the initial inoculum. Therespiratory burst of the neutrophils was assessed by the oxidation of dihydror-hodamine-123 as described previously (39). Neutrophils were identified andgated on the basis of their forward and side scatter and by staining with phyco-erythrin-conjugated anti-mouse Ly-6G (eBioscience, San Diego, CA). The sam-ples were analyzed on a FACSCalibur flow cytometer (BD Biosciences, San Jose,CA) with CellQuest Pro software.

Functional analyses of bone marrow-derived neutrophils. Bone marrow-de-rived neutrophils were isolated as described by Looney et al. (29). Mice wereeuthanized by CO2 asphyxiation, and their femurs and tibias were flushed withPBS. After hypotonic lysis of erythrocytes, the cell suspension was layered overa 62% Percoll (Sigma) gradient and centrifuged for 30 min at 1,300 � g. Theresulting neutrophil pellet was washed in PBS and suspended in minimum es-sential medium (MEM; Invitrogen, Grand Island, NY) with 1% endotoxin-freebovine serum albumin (BSA). Neutrophils were counted on a hemacytometer,and cell purity was confirmed by Diff Quick staining of cytospin preparations. S.aureus was opsonized with fresh frozen 10% mouse serum for 20 min at 37°C. Forphagocytosis assays, 5 � 106 opsonized bacteria were incubated with 5 � 105

neutrophils in 300 �l of MEM–1% BSA for 30 min at 37°C on a rotator. Cytospinpreparations of each sample were stained with Diff Quick solution and evaluatedunder a light microscope. The percentage of 200 neutrophils with cell-associatedor internalized S. aureus was estimated, and the average number of bacteria perneutrophil was calculated. For the S. aureus killing assays, 2 � 106 opsonizedbacteria were incubated with 2 � 106 neutrophils in 500 �l of MEM–1% BSA at37°C on a rotator. Aliquots were removed at 0, 30, and 60 min; diluted in ice-coldsterile water with 0.05% saponin; vortexed vigorously; and plated on tryptic soyagar plates.

Statistical analyses. Data collected from three groups of animals were ana-lyzed by analysis of variance with Bonferroni posthoc analysis. Comparisonsbetween two groups were analyzed by the unpaired Student’s t test with Welch’scorrection.

RESULTS

Susceptibility of diabetic db/db mice to S. aureus infection.The blood glucose levels in 8- to 11-week-old �/� wild-typeand db/� heterozygous nondiabetic control mice were 3.9 to7.3 and 4.7 to 6.3 mmol/liter (71 to 131 and 84 to 113 mg/dl,respectively), respectively. In contrast, db/db mice of the sameage had blood glucose levels that ranged from 25.5 mmol/literto �33.3 mmol/liter (459 to �599 mg/dl). Our initial studieswere aimed at evaluating whether mice with type 2 diabetesshowed greater susceptibility to S. aureus infection of the hindpaw than nondiabetic mice. The animals were challenged withS. aureus strain PS80 at doses ranging from 8 � 104 CFU to8 � 106 CFU. Most of the nondiabetic �/�, db/�, and diabeticdb/db mice inoculated with 8 � 104 CFU of S. aureus resolvedthe infection within 5 days, as estimated by quantitative cul-tures of the homogenized hind paw tissue (data not shown). Incontrast, animals (three mice/group) challenged with 8 � 106

CFU of S. aureus developed an acute, purulent infection, andthe bacterial burden in the infected tissues was uniformly high.The mean log CFU/g of tissue � standard error of the mean(SEM) was 7.48 � 0.98, 7.73 � 0.66, and 7.56 � 0.60 for thedb/db, db/�, and �/� mice, respectively. At an inoculum of8 � 105 CFU, the mean log CFU/g of tissue was highest for thediabetic db/db animals (5.57 � 0.59). The nondiabetic mice hadfewer viable S. aureus cells recovered from the tissues (Table1), although the differences among the groups did not reachstatistical significance by analysis of variance (P � 0.0655). Aninoculum dose of �106 CFU per mouse was considered opti-mal and chosen for further experiments.

Because 8- to 11-week-old db/db mice did not exhibit signif-icantly greater susceptibility to staphylococcal infection thannondiabetic littermates, we considered that animals sustainedin a prolonged diabetic state might be more vulnerable to S.aureus infection. Therefore, we housed the mice for periods upto 35 weeks before hind paw challenge with S. aureus. Asshown in Table 1, db/db mice 20 weeks of age were moresusceptible to S. aureus infection than db/db animals 8 to 11weeks of age. The mean bacterial burden in the infected hindpaws of db/db mice 20 weeks of age was 3 to 4 logs higher thanthat of age-matched db/� or �/� control mice (P � 0.0035).Housing the mice up to 35 weeks did not further enhance thesusceptibility of the db/db animals to staphylococcal infection

TABLE 1. Mean bacterial loads in the hind paw tissues of db/db,db/�, and �/� mice inoculated with �106 CFU of S. aureus

PS80 and euthanized for quantitative cultures on day 5

Mouseage

(wks)

Bacterial load by genotype (log CFU/g oftissue � SEM) P

valuea

db/db (n)b db/� (n)b �/� (n)b

8–11 5.57 � 0.59 (5) 3.19 � 0.68 (7) 4.56 � 0.65 (6) 0.065520 6.71 � 0.98 (4) 2.59 � 0.47 (4) 3.50 � 0.23 (4) 0.003534–35 6.68 � 0.41 (7) 3.79 � 0.57 (4) 4.91 � 0.56 (9) 0.0103

a P values were calculated with one-way analysis of variance. At 20 weeks ofage, the bacterial burden of db/db mice was significantly increased compared tothat of db/� (P 0.05) or �/� (P 0.01) mice (Bonferroni multiple compar-isons test). The numbers of CFU/g of tissue were not significantly differentbetween db/� and �/� mice.

b n, number of mice.

VOL. 77, 2009 S. AUREUS INFECTION IN db/db MICE 1009

on March 13, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

(Table 1), and so mice 16 to 20 weeks of age were utilized forsubsequent experiments.

Figure 1 shows the time course of S. aureus infection in thehind paw of groups of 4 to 10 diabetic db/db or �/� mice 19 to20 weeks of age. Whereas the tissue bacterial burden peaked24 h after bacterial challenge in both groups of mice, thewild-type animals showed a greater ability to clear the infectionthereafter. The db/db animals had significantly more S. aureusrecovered from the infected tissues on both days 5 (P � 0.05)and 10 (P � 0.0168). Furthermore, 6 of 10 �/� mice hadsterile hind paw cultures on day 10 compared with only 1 of 7db/db mice.

Inflammatory response to S. aureus in infected hind paws.Neutrophils play an important role in the initial host responsesto bacterial infection. We evaluated whether the impairedclearance of bacteria from the hind paw tissues of db/db ani-mals was associated with an aberrant inflammatory response.The hind paws of �/� and db/db mice were inoculated with�106 CFU of S. aureus, and the animals were euthanized fortissue harvest after 6, 12, 24, or 48 h. An examination of thestained tissue sections revealed that both groups of animalsresponded to the S. aureus challenge with a robust neutrophilinfiltration to the infected tissue. However, the temporalcourse of the host response indicated that the inflammatoryresponse was more pronounced in the db/db mice 6 and 12 hafter bacterial challenge. Twelve hours after bacterial inocula-tion, the inflammatory response in the infected tissues of dia-betic mice was markedly greater than in nondiabetic �/� mice(Fig. 2A and B), exhibiting an influx of neutrophils (Ly-6Gpositive) to the infection site (Fig. 2C and D). Neutrophilinfiltration was quantified by MPO assays performed on thetissues of mice euthanized at 6, 12, or 24 h. Although neutro-phils could be seen in the hind paw tissue as early as 6 hpostinfection (data not shown), MPO activity was barely de-tectable in the tissue homogenates then (Fig. 2E). Tissue ex-tracts from both mouse groups had detectable MPO activity by12 h, but the hind paw tissues from infected db/db mice showedhigher (P � 0.0154) MPO activity than those of �/� mice,despite similarities in the tissue bacterial burden (Fig. 1). By24 h, the tissue MPO activities were comparable in both mouse

groups (Fig. 2E), confirming the similarities in the influx ofneutrophils by 24 h that we observed histologically (data notshown). These results suggest that db/db mice exhibit a greaterearly inflammatory response to S. aureus infection than thenondiabetic �/� mice. However, the results of quantitativecultures (Table 1 and Fig. 1) show that this early inflammatoryresponse in the db/db animals was not effective in bacterialclearance.

Chemokine levels in infected hind paws. To determinewhether enhanced chemokine production might explain theexcessive neutrophil influx in the hind paw tissues of infecteddb/db mice, we measured CXCL1 and CXCL2 levels in hindpaw tissue extracts from infected mice by an ELISA. Consis-tent with a greater influx of neutrophils to the infected tissues,CXCL1 levels were significantly higher (P � 0.0089) in db/dbmice than in �/� mice 12 h after S. aureus inoculation (Fig.3A). Similarly, tissue CXCL2 levels at 12 h were approximatelytwofold higher in db/db mice than in nondiabetic �/� mice(Fig. 3B), but the difference was not quite significant (P �0.0689). Chemokine levels at 6 h and 24 h were comparable inboth diabetic and nondiabetic mice (Fig. 3), consistent with theMPO activities measured at these same time points (Fig. 2E).

FIG. 1. Time course of hind paw infection in �/� and db/db micechallenged with �106 CFU of S. aureus PS80. The animals were eu-thanized at different time points, and the excised tissue was culturedquantitatively. Each point represents the mean � SEM of culturesperformed on groups of 4 to 10 animals. The bacterial burden in thehind paws of db/db mice (black circles) was higher than that of the �/�mice (white circles) on days 5 (P � 0.05) and 10 (P � 0.0168), asdetermined by the unpaired Student’s t test with Welch’s correction forsamples with unequal variances.

FIG. 2. Inflammation seen at 12 h in the hind paw tissues from �/�(A) and db/db (B, C, and D) mice challenged with �106 CFU of S.aureus. Hematoxylin and eosin staining was used in panels A, B, and C.Immunohistochemical staining of neutrophils with anti-mouse Ly-6Gantibody was used in panel D. Panel C shows a higher magnification(�100) of the boxed region in panel B showing neutrophils as thepredominant cell type. Images in panels A, B, and D are at a magni-fication of �10. (E) MPO activity in homogenates of hind paw tissuesfrom mice (n � 2 to 4) challenged with 106 CFU of S. aureus. MPOactivity was higher (P � 0.0154) at 12 h in the db/db mice (black circles)than in the nondiabetic �/� mice (white circles).

1010 PARK ET AL. INFECT. IMMUN.

on March 13, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

Functional analyses of blood leukocytes. The paradoxicalfinding that the greater inflammatory response of db/db micewas accompanied by an inability to clear the bacterial infectionled us to examine neutrophil function in the db/db mouse.Neutrophils recovered from patients with type 2 diabetes havebeen shown to exhibit impaired bactericidal activity (6, 15, 41,48). To determine whether leukocytes from db/db mice wouldshow a similar defect, we performed S. aureus killing assayswith mouse blood. Blood samples were tested on one to threeoccasions for 22 �/� mice and 37 db/db mice. As shown in Fig.4, blood from �/� mice killed 52% of the S. aureus inoculumafter a 2-h incubation at 37°C. Similar levels of staphylococcalkilling were observed previously in the blood from ICR (out-bred mice derived at the Institute for Cancer Research) andnonobese diabetic (NOD) mice (39, 47). The blood from db/dbanimals killed only 33% of the S. aureus cells (P � 0.0029). Wealong with others have shown that preincubation with up to 33mmol/liter -D-glucose did not impair the in vitro respiratoryburst or bactericidal activity of neutrophils from nondiabeticmice (39, 41).

To determine whether the impaired bactericidal activity ofblood from db/db mice could be attributed to defects in phago-

cyte function, we performed respiratory burst assays. Bloodfrom �/� or db/db mice was incubated for 30 min withMEM–1% BSA or S. aureus, and the respiratory burst wasassessed by the oxidation of dihydrorhodamine and analyzedby flow cytometry. Figure 5A shows the results of a typicalexperiment wherein unstained neutrophils and neutrophilsstained with reduced dye are shown in the R2 region, and thecells that oxidized the dye appear in the R3 region. Neutrophilsfrom the �/� animal showed a marked shift in fluorescence inresponse to S. aureus, indicative of a robust respiratory burst(Fig. 5A). In contrast, neutrophils from the diabetic db/dbmouse responded poorly to the S. aureus stimulus.

The data in Fig. 5B and C summarize the neutrophil re-sponse of 13 db/db mice that were compared with six nondia-betic �/� control mice. The percentage of resting neutrophils(stimulated with MEM–1% BSA only) that oxidized the dyewas higher for db/db mice than �/� mice, but this difference

FIG. 3. Chemokine levels were higher in the hind paw extractsfrom diabetic db/db mice (n � 10) 12 h after inoculation with S.aureus than in those from nondiabetic (�/�) mice (n � 10). P �0.0089 for CXCL1 (A) and P � 0.0669 for CXCL2 (B) at 12 h.Values are means � SEM.

FIG. 4. In vitro bactericidal activity of blood from �/� and db/dbmice against 105 CFU/ml S. aureus. The data (means � SEM) repre-sent the mean values from 22 �/� mice and 37 db/db mice.

FIG. 5. Differential respiratory burst activity of neutrophils from�/� and db/db mice. (A) Representative results from an assay per-formed with mouse blood stimulated for 30 min with MEM-BSA (grayshading) or S. aureus (black shading). Unshaded peaks represent un-stained cells. Cells within the R2 region did not undergo an oxidativeburst. Cells within the R3 region show a shift in fluorescence intensitycharacteristic of an oxidative burst. Neutrophils from the db/db animalresponded poorly to S. aureus compared with neutrophils from the�/� animal. (B) The percentage of blood neutrophils from �/� (n �6) or diabetic db/db (n � 13) mice that responded to stimulation withmedium alone, S. aureus, or PMA is shown. (C) The magnitude of therespiratory burst, shown as the mean fluorescence index (MFI), ofneutrophils activated by S. aureus was significantly lower in db/db micethan in �/� controls.

VOL. 77, 2009 S. AUREUS INFECTION IN db/db MICE 1011

on March 13, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

was not significant (P � 0.1148). All of the animals showed asimilar percentage of neutrophils that responded to the S.aureus stimulus (Fig. 5B); however, the magnitude of the re-sponse (measured by the mean fluorescence index) was signif-icantly lower (P � 0.0494) for the diabetic mice. The additionof 0.1 �M phorbol myristate acetate (PMA), a protein kinaseC (PKC) agonist, induced a similar phagocyte respiratory burstin both db/db and �/� mice (Fig. 5B and 5C).

Functional analyses of bone marrow-derived neutrophils.Since neutrophils comprise only �28% of mouse peripheralleukocytes (10), we purified neutrophils from the bone marrowof 16- to 20-week-old �/� and db/db mice. At a multiplicity ofinfection (MOI) of 10 (10 CFU per neutrophil), �90% of theneutrophils from diabetic and nondiabetic animals had in-gested S. aureus opsonized with 10% serum after 30 min. Thenumber of bacteria associated with or internalized by eachneutrophil ranged from 9 to 18 for cells recovered from both�/� and db/db mice. If the MOI was decreased to 1 or if thestaphylococci were opsonized with 4% serum, the phagocyticuptake of staphylococci by neutrophils from �/� or db/db miceremained comparable (data not shown).

The bactericidal activity of bone marrow-derived neutro-phils from �/� and db/db mice was assessed with an in vitroopsonophagocytic killing assay. The bacteria were opsonizedwith 10% normal mouse serum and incubated with or withoutneutrophils. As shown in Fig. 6, bone marrow-derived neutro-phils from �/� and db/db mice killed �75% of the S. aureusinoculum within 60 min. Without opsonizing mouse serum, nokilling was observed (data not shown). Additional experimentswherein the bacteria were opsonized with 5% serum or with S.aureus capsule-specific antibodies failed to show differences inthe bactericidal activity of bone marrow-derived neutrophilsfrom diabetic and nondiabetic mice (data not shown).

DISCUSSION

Individuals with type 2 diabetes may develop lower-extrem-ity infections and ulcerations that can be severely debilitating,often leading to functional disability, depression, increased riskof amputation, and associated mortality. Many factors influ-ence the susceptibility of diabetic patients to infection, includ-ing vascular pathology, peripheral neuropathy, and nephropa-thy. Foot and ankle infections are the most common cause ofhospitalization in diabetic patients (37), and S. aureus is a

frequent cause of these infections. Recurrent infection anddelayed wound healing associated with diabetic foot ulcers mayultimately result in lower-extremity amputation.

Previously, we established a chronic S. aureus infection inthe hind paws of NOD mice (39). These animals develop type1 diabetes between 14 and 23 weeks of age. In contrast toage-matched nondiabetic littermates, the diabetic NOD micewere unable to resolve the staphylococcal hind paw infectionover a 10-day period. The diabetic animals exhibited a delayedinflammatory response to S. aureus, and this finding was con-sistent with diminished CXC chemokine levels in the infectedtissues. The impaired killing of S. aureus in the tissues andblood of diabetic mice correlated with a diminished in vitroleukocytic respiratory burst.

Like diabetic NOD mice, db/db mice with type 2 diabeteswere also more susceptible to staphylococcal infection thanage-matched �/� littermates. This susceptibility was more ap-parent in db/db mice that were �20 weeks of age than inyounger animals, perhaps as a result of their sustained hyper-glycemic state. The host response to the S. aureus hind pawinfection in the db/db animals was quite distinct from thatobserved in diabetic NOD mice. Levels of the proinflammatorychemokines CXCL1 and CXCL2, which recruit neutrophils(23, 49), were elevated in the hind paw tissues of db/db mice12 h after bacterial inoculation. Accordingly, db/db miceshowed a greater inflammatory response to S. aureus infectionthan �/� mice. The early neutrophil influx into the infectedhind paw tissues of diabetic mice exceeded that of wild-typemice, and this observation was confirmed by MPO assays.Since others have reported diminished intracellular MPO ac-tivity in neutrophils from patients with type 2 diabetes (40), thisquantitative assay may have actually underestimated the num-bers of neutrophils in the tissues of the db/db animals.

Neutrophils are the major inflammatory cells in acute bac-terial infections. However, the heightened inflammatory re-sponse observed in db/db mice did not result in resolution ofthe staphylococcal infection. Similar observations using otherstaphylococcal infection models have been reported, wherein aneutrophil-rich environment actually exacerbated the infectionas a result of S. aureus survival within neutrophils (16, 31, 32).Consistent with the inefficient clearance of the bacteria fromthe infected hind paw tissues, phagocytic killing of S. aureus inthe blood from db/db mice was impaired (Fig. 5). Severalgroups have reported that patients with diabetes or hypergly-cemia exhibit aberrant neutrophil responses (1, 3, 12, 14, 33,34). Other investigators, however, failed to detect a defect inneutrophil function (18, 28). These discrepancies could beattributed to the metabolic status of the diabetic patients, dif-ferences in methodology, or underlying diseases or complica-tions among the population groups.

The phagocyte respiratory burst is key to neutrophil functionsince reactive oxygen species generated by the NADPH oxi-dase membrane complex effectively kill the engulfed bacteria(17). Individuals with chronic granulomatous disease lackNADPH oxidase activity and suffer recurrent, often life-threat-ening bacterial infections (25). Likewise, mice lacking one ofthe subunits of the NADPH oxidase complex exhibit enhancedsusceptibility to S. aureus infection (13, 35). The respiratoryburst of neutrophils from the blood of db/db mice was aberrantcompared with that of nondiabetic �/� mice (Fig. 5). Whereas

FIG. 6. Serum-opsonized S. aureus cells were incubated in mediumalone (white squares) or at an MOI of 1 in the presence of 2 � 106

bone marrow-derived neutrophils from db/db or �/� mice. Bacterialsurvival is expressed as a percentage of the inoculum, and the datashown are means � SEM. PMN, polymorphonuclear neutrophils.

1012 PARK ET AL. INFECT. IMMUN.

on March 13, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

only �20% of resting neutrophils from control mice wereactivated in the presence of medium alone, �40% of restingneutrophils from the blood of db/db mice were activated. Sev-eral studies have demonstrated that resting neutrophils fromdiabetic patients exhibit an augmented oxidative burst activitycompared with resting neutrophils from nondiabetic controls(2, 18, 24). Chronic exposure to low levels of reactive oxygenspecies may contribute to vascular complications and the de-velopment of chronic low-grade inflammation commonlyfound in diabetic hosts (41). In our study, these “primed”neutrophils exhibited a blunted response to stimulation with S.aureus. Similar numbers of neutrophils responded to S. aureusin both mouse groups, but the magnitude of the oxidative burstfor neutrophils from the db/db mice was only about 60% of thewild-type response, consistent with results from Shurtz-Swirskiand colleagues (41), who stimulated human neutrophils in vitrowith zymosan. The measured rate of superoxide release fromthe neutrophils of type 2 diabetic patients was significantlylower than that of age- and sex-matched normal control sub-jects. Consistent with the findings of Zykova et al. (51), weobserved no difference in the PMA-induced respiratory burstin neutrophils from db/db and �/� mice. PMA enters neutro-phils and activates PKC by directly binding to its diacylglycerol(natural activator) site (19). Because the S. aureus-inducedrespiratory burst also requires activated PKC (20), the de-creased respiratory burst of db/db mice in response to S. aureusstimulus may derive from a defect upstream of PKC in theprocess of NADPH oxidase activation. Alternatively, activa-tion of NADPH oxidase by PMA versus S. aureus may involvedifferent signaling pathways or different isotypes of PKC.

To further investigate the phagocytic and bactericidal activ-ities of neutrophils from nondiabetic and diabetic mice, wecollected bone marrow-derived neutrophils from each mousegroup. Notably, neutrophils from the bone marrow of diabeticdb/db mice ingested and killed opsonized S. aureus as well asneutrophils from �/� mice. Likewise, the results of respiratoryburst assays comparing �/� and db/db bone marrow-derivedneutrophil responses to S. aureus were similar (data notshown). Since neutrophils from the bone marrow are consid-ered fully functional (4), the diminished bactericidal activityand respiratory burst characteristic of neutrophils in the bloodmay be due to modulation of neutrophil function by cytokinessecreted by blood mononuclear cells. Zykova et al. (51)reported that stimulated macrophages from db/db mice pro-duced lower levels of tumor necrosis factor alpha and inter-leukin-1� than cells from db/� mice. Blood glucose concen-trations do not directly affect the respiratory burst or thebactericidal activity of neutrophils (39, 41). However, Shurtz-Swirski (41) did show a correlation between the rate of super-oxide release and blood levels of hemoglobin A1c. Prolongedexposure of neutrophils to elevated blood glucose concentra-tions may cause the formation of advanced glycation end prod-ucts that may compromise neutrophil function (9). Treatmentof neutrophils with advanced glycation end product albuminresulted in increased phagocytic activity; however, the treat-ment abrogated the production of S. aureus-induced reactiveoxygen species, thereby protecting ingested bacteria from in-tracellular killing (9). Whether this mechanism or others con-tribute to the defective neutrophil function of db/db mice iscurrently under investigation in our laboratory.

In summary, our results suggest that defects in innate im-munity predispose diabetic db/db mice to staphylococcal infec-tion. db/db mice showed a heightened inflammatory responseto S. aureus, but the recruited neutrophils failed to resolve thebacterial infection, a result that correlated with an impairedneutrophil respiratory burst. The factors that contribute to theimpaired function of neutrophils derived from the blood andtissues of diabetic mice merit further investigation.

ACKNOWLEDGMENTS

This research was supported by NIH grant AI29040. J.R. was therecipient of a Juvenile Diabetes Research Foundation award (3-2004-546), and J.R. and S.P. were supported by NIH Infectious DiseaseTraining Grant T32 AI07061. F.H. was supported by a grant fromGerman Academy of Sciences Leopoldina.

We gratefully acknowledge Rachel McLoughlin for discussions andhelpful suggestions.

REFERENCES

1. Alba-Loureiro, T. C., C. D. Munhoz, J. O. Martins, G. A. Cerchiaro, C.Scavone, R. Curi, and P. Sannomiya. 2007. Neutrophil function and metab-olism in individuals with diabetes mellitus. Braz. J. Med. Biol. Res. 40:1037–1044.

2. Bagdade, J. D., R. K. Root, and R. J. Bulger. 1974. Impaired leukocytefunction in patients with poorly controlled diabetes. Diabetes 23:9–15.

3. Bhattacharya, S. K., S. Shastri, P. Mahajan, S. V. Madhu, A. K. Tripathi,G. P. Rauniar, B. P. Das, and K. R. Paudel. 2007. Polymorphonuclearleukocyte function in type-2 diabetes mellitus patients and its correlationwith glycaemic control. Nepal Med. Coll. J. 9:111–116.

4. Boxio, R., C. Bossenmeyer-Pourie, N. Steinckwich, C. Dournon, and O.Nusse. 2004. Mouse bone marrow contains large numbers of functionallycompetent neutrophils. J. Leukoc. Biol. 75:604–611.

5. Caputo, G. M., J. Ulbrecht, P. Cavanagh, and P. J. Juliano. 2000. The roleof cultures in mild diabetic foot cellulitis. Infect. Dis. Clin. N. Am. 9:241–243.

6. Chang, F. Y., and M. F. Shaio. 1995. Respiratory burst activity of monocytesfrom patients with non-insulin-dependent diabetes mellitus. Diabetes Res.Clin. Pract. 29:121–127.

7. Chen, H., O. Charlat, L. A. Tartaglia, E. A. Woolf, X. Weng, S. J. Ellis, N. D.Lakey, J. Culpepper, K. J. Moore, R. E. Breitbart, G. M. Duyk, R. I. Tepper,and J. P. Morgenstern. 1996. Evidence that the diabetes gene encodes theleptin receptor: identification of a mutation in the leptin receptor gene indb/db mice. Cell 84:491–495.

8. Chua, S. C., Jr., W. K. Chung, X. S. Wu-Peng, Y. Zhang, S. M. Liu, L.Tartaglia, and R. L. Leibel. 1996. Phenotypes of mouse diabetes and rat fattydue to mutations in the OB (leptin) receptor. Science 271:994–996.

9. Collison, K. S., R. S. Parhar, S. S. Saleh, B. F. Meyer, A. A. Kwaasi, M. M.Hammami, A. M. Schmidt, D. M. Stern, and F. A. Al-Mohanna. 2002.RAGE-mediated neutrophil dysfunction is evoked by advanced glycationend products (AGEs). J. Leukoc. Biol. 71:433–444.

10. Cotter, M. J., K. E. Norman, P. G. Hellewell, and V. C. Ridger. 2001. A novelmethod for isolation of neutrophils from murine blood using negative im-munomagnetic separation. Am. J. Pathol. 159:473–481.

11. Dang, C. N., Y. D. Prasad, A. J. Boulton, and E. B. Jude. 2003. Methicillin-resistant Staphylococcus aureus in the diabetic foot clinic: a worsening prob-lem. Diabet. Med. 20:159–161.

12. Delamaire, M., D. Maugendre, M. Moreno, M. C. Le Goff, H. Allannic, andB. Genetet. 1997. Impaired leucocyte functions in diabetic patients. Diabet.Med. 14:29–34.

13. Ellson, C. D., K. Davidson, G. J. Ferguson, R. O’Connor, L. R. Stephens, andP. T. Hawkins. 2006. Neutrophils from p40phox�/� mice exhibit severe de-fects in NADPH oxidase regulation and oxidant-dependent bacterial killing.J. Exp. Med. 203:1927–1937.

14. Gallacher, S. J., G. Thomson, W. D. Fraser, B. M. Fisher, C. G. Gemmell,and A. C. MacCuish. 1995. Neutrophil bactericidal function in diabetesmellitus: evidence for association with blood glucose control. Diabet. Med.12:916–920.

15. Geerlings, S. E., and A. I. Hoepelman. 1999. Immune dysfunction in patientswith diabetes mellitus (DM). FEMS Immunol. Med. Microbiol. 26:259–265.

16. Gresham, H. D., J. H. Lowrance, T. E. Caver, B. S. Wilson, A. L. Cheung, andF. P. Lindberg. 2000. Survival of Staphylococcus aureus inside neutrophilscontributes to infection. J. Immunol. 164:3713–3722.

17. Hampton, M. B., A. J. Kettle, and C. C. Winterbourn. 1998. Inside theneutrophil phagosome: oxidants, myeloperoxidase, and bacterial killing.Blood 92:3007–3017.

18. Hand, W. L., D. L. Hand, and Y. Vasquez. 2007. Increased polymorphonu-clear leukocyte respiratory burst function in type 2 diabetes. Diabetes Res.Clin. Pract. 76:44–50.

VOL. 77, 2009 S. AUREUS INFECTION IN db/db MICE 1013

on March 13, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

19. Heyworth, P. G., and J. A. Badwey. 1990. Protein phosphorylation associatedwith the stimulation of neutrophils. Modulation of superoxide production byprotein kinase C and calcium. J. Bioenerg. Biomembr. 22:1–26.

20. Kampen, A. H., T. Tollersrud, and A. Lund. 2004. Flow cytometric measure-ment of neutrophil respiratory burst in whole bovine blood using live Staph-ylococcus aureus. J. Immunol. Methods 289:47–55.

21. Karchmer, A. 2002. Microbiology and treatment of diabetic foot infections,p. 207–219. In A. Veves, J. M. Giurini, and F. W. LoGerfro (ed.), Thediabetic foot: medical and surgical treatment. Humana Press, Totowa, NJ.

22. Klevens, R. M., M. A. Morrison, J. Nadle, S. Petit, K. Gershman, S. Ray,L. H. Harrison, R. Lynfield, G. Dumyati, J. M. Townes, A. S. Craig, E. R.Zell, G. E. Fosheim, L. K. McDougal, R. B. Carey, and S. K. Fridkin. 2007.Invasive methicillin-resistant Staphylococcus aureus infections in the UnitedStates. JAMA 298:1763–1771.

23. Kobayashi, Y. 2008. The role of chemokines in neutrophil biology. Front.Biosci. 13:2400–2407.

24. Larijani, B., P. Shooshtarizadeh, N. Mosaffa, and R. Heshmat. 2007. Poly-morphonuclear leucocyte respiratory burst activity correlates with serum zinclevel in type 2 diabetic patients with foot ulcers. Br. J. Biomed. Sci. 64:13–17.

25. Liese, J., S. Kloos, V. Jendrossek, T. Petropoulou, U. Wintergerst, G.Notheis, M. Gahr, and B. H. Belohradsky. 2000. Long-term follow-up andoutcome of 39 patients with chronic granulomatous disease. J. Pediatr. 137:687–693.

26. Lipsky, B. A., R. E. Pecoraro, S. A. Larson, M. E. Hanley, and J. H. Ahroni.1990. Outpatient management of uncomplicated lower-extremity infectionsin diabetic patients. Arch. Intern. Med. 150:790–797.

27. Lipsky, B. A., R. E. Pecoraro, and L. J. Wheat. 1990. The diabetic foot. Softtissue and bone infection. Infect. Dis. Clin. N. Am. 4:409–432.

28. Llorente, L., H. De La Fuente, Y. Richaud-Patin, C. Alvarado-De La Bar-rera, A. Diaz-Borjon, A. Lopez-Ponce, I. Lerman-Garber, and J. Jakez-Ocampo. 2000. Innate immune response mechanisms in non-insulin depen-dent diabetes mellitus patients assessed by flow cytoenzymology. Immunol.Lett. 74:239–244.

29. Looney, M. R., X. Su, J. A. Van Ziffle, C. A. Lowell, and M. A. Matthay. 2006.Neutrophils and their Fc gamma receptors are essential in a mouse model oftransfusion-related acute lung injury. J. Clin. Investig. 116:1615–1623.

30. Magnotti, L. J., J. S. Upperman, D. Z. Xu, Q. Lu, and E. A. Deitch. 1998.Gut-derived mesenteric lymph but not portal blood increases endothelial cellpermeability and promotes lung injury after hemorrhagic shock. Ann. Surg.228:518–527.

31. McLoughlin, R. M., J. C. Lee, D. L. Kasper, and A. O. Tzianabos. 2008.IFN-gamma regulated chemokine production determines the outcome ofStaphylococcus aureus infection. J. Immunol. 181:1323–1332.

32. McLoughlin, R. M., R. M. Solinga, J. Rich, K. J. Zaleski, J. L. Cocchiaro, A.Risley, A. O. Tzianabos, and J. C. Lee. 2006. CD4� T cells and CXCchemokines modulate the pathogenesis of Staphylococcus aureus wound in-fections. Proc. Natl. Acad. Sci. USA 103:10408–10413.

33. Mowat, A., and J. Baum. 1971. Chemotaxis of polymorphonuclear leukocytesfrom patients with diabetes mellitus. N. Engl. J. Med. 284:621–627.

34. Nielson, C. P., and D. A. Hindson. 1989. Inhibition of polymorphonuclearleukocyte respiratory burst by elevated glucose concentrations in vitro. Di-abetes 38:1031–1035.

35. Pollock, J. D., D. A. Williams, M. A. Gifford, L. L. Li, X. Du, J. Fisherman,S. H. Orkin, C. M. Doerschuk, and M. C. Dinauer. 1995. Mouse model ofX-linked chronic granulomatous disease, an inherited defect in phagocytesuperoxide production. Nat. Genet. 9:202–209.

36. Popovich, K. J., R. A. Weinstein, and B. Hota. 2008. Are community-asso-ciated methicillin-resistant Staphylococcus aureus (MRSA) strains replacingtraditional nosocomial MRSA strains? Clin. Infect. Dis. 46:787–794.

37. Ramsey, S. D., K. Newton, D. Blough, D. K. McCulloch, N. Sandhu, andE. H. Wagner. 1999. Patient-level estimates of the cost of complications indiabetes in a managed-care population. Pharmacoeconomics 16:285–295.

38. Reiber, G. E., L. Vileikyte, E. J. Boyko, M. del Aguila, D. G. Smith, L. A.Lavery, and A. J. Boulton. 1999. Causal pathways for incident lower-extrem-ity ulcers in patients with diabetes from two settings. Diabetes Care 22:157–162.

39. Rich, J., and J. C. Lee. 2005. The pathogenesis of Staphylococcus aureusinfection in the diabetic NOD mouse. Diabetes 54:2904–2910.

40. Sato, N., K. Kashima, Y. Tanaka, H. Shimizu, and M. Mori. 1997. Effect ofgranulocyte-colony stimulating factor on generation of oxygen-derived freeradicals and myeloperoxidase activity in neutrophils from poorly controlledNIDDM patients. Diabetes 46:133–137.

41. Shurtz-Swirski, R., S. Sela, A. T. Herskovits, S. M. Shasha, G. Shapiro, L.Nasser, and B. Kristal. 2001. Involvement of peripheral polymorphonuclearleukocytes in oxidative stress and inflammation in type 2 diabetic patients.Diabetes Care 24:104–110.

42. Sievert, D. M., J. T. Rudrik, J. B. Patel, L. C. McDonald, M. J. Wilkins, andJ. C. Hageman. 2008. Vancomycin-resistant Staphylococcus aureus in theUnited States, 2002–2006. Clin. Infect. Dis. 46:668–674.

43. Skiest, D. J., K. Brown, T. W. Cooper, H. Hoffman-Roberts, H. R. Mussa,and A. C. Elliott. 2007. Prospective comparison of methicillin-susceptibleand methicillin-resistant community-associated Staphylococcus aureus infec-tions in hospitalized patients. J. Infect. 54:427–434.

44. Stanaway, S., D. Johnson, P. Moulik, and G. Gill. 2007. Methicillin-resistantStaphylococcus aureus (MRSA) isolation from diabetic foot ulcers correlateswith nasal MRSA carriage. Diabetes Res. Clin. Pract. 75:47–50.

45. Tentolouris, N., G. Petrikkos, N. Vallianou, C. Zachos, G. L. Daikos, P.Tsapogas, G. Markou, and N. Katsilambros. 2006. Prevalence of methicillin-resistant Staphylococcus aureus in infected and uninfected diabetic foot ul-cers. Clin. Microbiol. Infect. 12:186–189.

46. Tzianabos, A. O., J. Y. Wang, and J. C. Lee. 2001. Structural rationale for themodulation of abscess formation by Staphylococcus aureus capsular polysac-charides. Proc. Natl. Acad. Sci. USA 98:9365–9370.

47. Watts, A., D. Ke, Q. Wang, A. Pillay, A. Nicholson-Weller, and J. C. Lee.2005. Staphylococcus aureus strains that express serotype 5 or serotype 8capsular polysaccharides differ in virulence. Infect. Immun. 73:3502–3511.

48. Wykretowicz, A., B. Wierusz-Wysocka, J. Wysocki, A. Szczepanik, and H.Wysocki. 1993. Impairment of the oxygen-dependent microbicidal mecha-nisms of polymorphonuclear neutrophils in patients with type 2 diabetes isnot associated with increased susceptibility to infection. Diabetes Res. Clin.Pract. 19:195–201.

49. Zhang, X. W., Q. Liu, Y. Wang, and H. Thorlacius. 2001. CXC chemokines,MIP-2 and KC, induce P-selectin-dependent neutrophil rolling and extravas-cular migration in vivo. Br. J. Pharmacol. 133:413–421.

50. Zhu, W., N. C. Clark, L. K. McDougal, J. Hageman, L. C. McDonald, andJ. B. Patel. 2008. Vancomycin-resistant Staphylococcus aureus isolates asso-ciated with Inc18-like vanA plasmids in Michigan. Antimicrob. Agents Che-mother. 52:452–457.

51. Zykova, S. N., T. G. Jenssen, M. Berdal, R. Olsen, R. Myklebust, and R.Seljelid. 2000. Altered cytokine and nitric oxide secretion in vitro by mac-rophages from diabetic type II-like db/db mice. Diabetes 49:1451–1458.

Editor: R. P. Morrison

1014 PARK ET AL. INFECT. IMMUN.

on March 13, 2021 by guest

http://iai.asm.org/

Dow

nloaded from