Vol. 60, No. 9INFECrION AND IMMUNrrY, Sept. 1992, p. 3790-37980019-9567/92/093790-09$02.00/0Copyright X) 1992, American Society for Microbiology

Role of Endotoxin in Alterations of Hepatic Drug Metabolismby Diphtheria and Tetanus Toxoids and Pertussis

Vaccine AdsorbedSHERRY ANSHER,1* WALTER THOMPSON,' PHIL SNOY,2 AND WILLIAM HABIG'

Division ofBacterial Products' and Division ofProduct Quality Control,2 Centerfor BiologicsEvaluation and Research, Food and Drug Administration, Bethesda, Maryland 20892

Received 28 April 1992/Accepted 26 June 1992

Administration of diphtheria and tetanus toxoids and pertussis vaccine adsorbed (DTP vaccine) or endotoxin(LPS) resulted in marked alterations in hepatic drug-metabolizing enzymes in endotoxin-responsive (R) andnon-endotoxin-responsive (NR) mice. A single human dose (0.5 ml) of DTP vaccine increased hexobarbital-induced sleep times to 1.6- to 1.8-fold above those of controls in both strains of mice. This effect persisted for7 days. In contrast, BordeteUa pertussis LPS-treated mice showed an increase at 1 day (3.0-fold for R mice and1.5-fold for NR mice), which returned to control levels by day 7. Furthermore, cytochrome P-450 levels weredecreased 30 to 40% 24 h after DTP vaccine administration in both R and NR mice, while after LPSadministration they were decreased 30%o in R mice and less than 10% in NR mice. Both spleen and liver weightsofR and NR mice were increased 7 to 14 days following DTP vaccine administration. However, LPS treatmenthad no apparent effect on liver weights, and spleen weights of R mice were elevated from days 3 to 7.Histopathologic tissue examination showed random, multifocal inflammation with hepatocyte necrosis afterDTP vaccine administration to both R and NR mice and an absence of lesions in LPS-treated mice. PremixingLPS with polymyxin eliminated the increased sleep times, but premixing DTP vaccine with polymyxin did notaffect the increased sleep times. Levels of tumor necrosis factor and interleukin-6 in plasma of R mice weremarkedly increased after DTP and LPS treatment, while NR mice had reduced increases. These results suggestthat LPS contributes to the alterations in R and NR mice seen within the first 24 h of vaccine administrationbut that it is not likely to contribute to the effects observed at later time points.

There has been a great deal of concern about local andsystemic adverse reactions to diphtheria and tetanus toxoidsand pertussis vaccine adsorbed (DTP vaccine) in infants andchildren. Increased reporting of and publicity surroundingadverse reactions have been claimed to result in a drop in thenumber of children being immunized (10, 11). Concomitantwith the decrease in immunizations has been a correspond-ing increase in the number of cases of pertussis (30, 34). Thepertussis component of the DTP vaccine has been implicatedas responsible for the adverse reactions, but the agentsresponsible for toxicity have not been identified. It has beensuggested that endotoxin associated with the vaccine is thecausative agent because of its role in fever induction andsepsis.

Several reports have demonstrated that administrations ofDTP vaccine and other agents cause inhibition of hepaticdrug-metabolizing enzymes in mice (2, 3). For humans, thereare a number of reports on alterations of drug metabolismfollowing immunization with influenza virus vaccine andadministration of endotoxins and cytokines or cytokine-inducing agents (19, 23, 32). The relationship of theseobservations with animals to the reported adverse reactionsin humans following immunization has not been established,nor has the mechanism for the inhibition of drug metabolismbeen identified. There is, however, the suggestion that theimmune system may be involved. Administration of vac-cines has been associated with an increase in interferonlevels in sera of rodents and rabbits, and administration ofinterferon-inducing agents depresses hepatic drug metabo-

* Corresponding author.

lism (2, 3, 26, 27). In mice, the coadministration of interleu-kin-2 (IL-2) and alpha interferon decreases hepatic drugmetabolism in a dose-dependent manner (2). Pretreatment ofmice with gamma irradiation reduces the effects of thecytokines, which is consistent with a role for T cells.However, when irradiated mice were immunized with DTPvaccine, the alterations in hepatic drug metabolism were notchanged from those of nonirradiated controls (2). This sug-gests that there may be an alternate pathway for vaccinetoxicity.

Neither the mechanism by which vaccines inhibit drugmetabolism nor the components of the vaccines responsiblefor these alterations have been established. Previous studiesin our laboratory have shown that aluminum adjuvants, thepreservatives, and diphtheria and tetanus toxoids are not thecomponents responsible for the adverse reactions (3). Ex-tensive studies with diphtheria and tetanus toxoids adsorbed(DT vaccine), which contains the same adjuvant and toxoidcomponents as those present in the DTP vaccine, have failedto show any toxicity. Likewise, injection of DT vaccinemixed with pertussis toxoid or with Bordetella pertussislipopolysaccharide (LPS) failed to induce the alterationsnoted below for DTP vaccine. In contrast, a nonadsorbed,whole-cell pertussis vaccine elicits the same response asDTP vaccine in our mouse model (3).One common component of many bacterial vaccines is

endotoxin. Endotoxins are important factors in the pathoge-nicity of gram-negative organisms. The LPS of B. pertussishas been reported to account for most of the reportedadverse reactions to the vaccine in humans (37). In mice, asingle human dose (0.5 ml) of DTP vaccine increased hex-obarbital-induced sleep times within 12 h and to a maximum

3790

on March 30, 2019 by guest

http://iai.asm.org/

Dow

nloaded from

ENDOTOXIN AND DRUG METABOLISM 3791

at 7 to 10 days following immunization. In contrast, a single50-,ug injection of either B. pertussis or Escheichia coliendotoxin causes maximum sleep time increases on the firstday following injection. Hexobarbital-induced sleep timesreturned to control levels by 7 days in these mice (3).Endotoxins are known to induce a wide variety of biolog-

ical activities. These studies were undertaken to furtherevaluate the role of endotoxin in the vaccine-induced alter-ations of hepatic drug metabolism. The effects of DTPvaccine, polyriboinosinic:polyribocytidylic acid (PIC), en-dotoxin, and other vaccines and components are evaluatedwith endotoxin-responsive (R) and non-endotoxin-respon-sive (NR) mice.

MATERIALS AND METHODS

PIC and E. coli endotoxin (E. coli LPS) (serotype 055:B5)were purchased from Sigma Chemical Co. (St. Louis, Mo.).B. pertussis endotoxin (B. pertussis LPS) was purchasedfrom List Biological Laboratories (Campbell, Calif.). Vac-cines were samples submitted to the Center for BiologicsEvaluation and Research, Food and Drug Administration,for control testing. Endotoxic activity was determined by theDivision of Product Quality Control, Food and Drug Admin-istration, by using the Limulus amebocyte lysate (LAL) testand by rabbit pyrogenicity (20).Animals. Female NR (C3HAHIEJ) and R (C3HIHEN) mice

(6 to 10 weeks old, 10 to 15 g) were purchased from theNational Cancer Institute or Charles River Laboratories.They were provided free access to food and water and weremaintained on a 12-h light-and-dark cycle. Animals wereacclimated several days prior to the start of an experiment.All injections were administered intraperitoneally at a vol-ume of 0.5 ml. PIC or LPS was dissolved in sterile, pyrogen-free saline prior to injection.

Sleep time. Hexobarbital sodium (Sigma Chemical Co.),150 mg, was dissolved in 0.9 ml of 1 N NaOH and diluted to30 ml in water immediately before use. Mice were weighedand injected intraperitoneally with a dose of 100 mg/kg ofbody weight. Sleep time was measured as the time elapsedfrom injection of hexobarbital to the return of the redress-ment reflex (22). A separate group of animals was used foreach time point when sleep times were measured. The sleeptime index refers to the mean sleep time of the experimentalgroup divided by the mean sleep time of the control group.Enzyme assays. Immediately after mice were sacrificed by

cervical dislocation, livers were removed from mice,weighed, and homogenized in 3 volumes of ice-cold isotonicKCl. Cytosolic and microsomal fractions were preparedaccording to the method of Hodgeboom (21). Microsomeswere resuspended in 2 ml of 0.1 M potassium phosphate (pH7.4) containing 20% glycerol, frozen immediately in liquidnitrogen, stored at -70°C, and used within 1 week ofpreparation. No loss of activity was seen upon storage whensamples were handled in this manner.Cytochrome P-450 levels were measured according to the

method of Omura and Sato (31) by using a model A4 50AHewlett-Packard spectrophotometer.Benzphetamine demethylase activity was assayed spec-

trophotometrically according to the method of Astrom et al.(5). Ethylmorphine was obtained from Alltech Applied Sci-ence (Deerfield, Ill.). Ethylmorphine demethylase activitywas assayed by the method of Astrom et al. (5), with thefollowing modifications: a total volume of 0.25 ml was used,to which an equal volume of trichloroacetic acid was addedto stop the reaction. Formaldehyde production was deter-

mined according to the Nash procedure (29). Benzopyrenemonooxygenase activity was determined by the method ofDePierre et al. (13). Benzopyrene was obtained from Ald-rich, and [3H]benzopyrene (40 Ci/mmol) was obtained fromAmersham. [3H]benzopyrene was purified before use asdescribed previously (13).The cytosolic glutathione transferases were assayed with

1-chloro-2,4-dinitrobenzene and 1,2-dichloro-4-nitrobenzeneas substrates (17). Dicoumarol-inhibitable quinone reductasewas assayed according to the method of Ernster (14) with themodifications described by Benson et al. (7) by using 2,6-dichloroindophenol as the electron acceptor in a final volumeof 1.0 ml.

Protein was determined with bicinchoninic acid (38), withbovine serum albumin as a standard.Data were analyzed by Student's t test or analysis of

variance by using Statview (MacIntosh).Plasma. Blood was collected into heparinized tubes by

retro-orbital bleeding immediately before sacrifice. Plasmawas prepared by centrifugation and was stored at -20°Cuntil used. No loss of activity was observed with samplesstored for up to 1 month. The plasma samples were assayed(by Assay Research Inc., College Park, Md.) for tumornecrosis factor (TNF) by using a standard enzyme-linkedimmunosorbent assay (ELISA) kit (Genzyme). IL-6 assayswere performed by bioassay according to the procedure ofAarden et al. by using B9 cells (1) as described in detail byTosato et al. (39). The B9 cells were generously provided byG. Tosato. Activity is expressed in units per milliliter, where1 U represents the dilution of standard necessary to give 50%of maximal activity. The standard was derived from theculture supernatant fluid of an IL-6-secreting cell line.

Pathology. Livers and spleens were removed from DTPvaccine- or endotoxin-treated mice at the times indicated,weighed, rinsed, and placed immediately in buffered forma-lin solution. Histologic examination of the tissues was per-formed by a veterinary pathologist.

RESULTS

Hexobarbital-induced sleep time 24 h after administrationof vaccines or endotoxin. The abilities of DTP and pertussisvaccines to alter hexobarbital-induced sleep time were com-pared with those ofE. coli and B. pertussis LPS preparationsin R and NR mice. The dose of endotoxin used (50 ,ug) waschosen because it contained an endotoxic activity somewhathigher than that found in 0.5 ml of DTP vaccine whenmeasured by the LAL test (typically 50,000 endotoxin units[EU]/ml for LPS versus 10,000 EU/ml for vaccine). Admin-istration of either E. coli or B. penussis endotoxin, at a doseof 50 pg, increased hexobarbital-induced sleep time in Rmice 2.4- to 3.0-fold over that of controls, compared with a1.2- to 1.5-fold increase in NR mice. Administration of DTPvaccine or pertussis vaccine adsorbed caused a 1.6- to1.8-fold increase and a 1.4- to 1.5-fold increase, respectively,in sleep times in R and NR strains of mice (Fig. 1). DTvaccine administration did not significantly increase sleeptimes over that of controls for either strain (1.1- to 1.3-foldincrease). Sleep time was also measured following adminis-tration of a single injection of cholera vaccine, a bacterialvaccine with an endotoxic activity similar to that of DTPvaccine (both have approximately 10,000 EU/ml by the LALtest). There was a 1.5-fold increase in sleep time in R micebut no increase in sleep time in NR mice at 24 h after choleravaccine administration. By 4 days, the increased sleep timein R mice had returned to control levels. Administration of

VOL. 60, 1992

on March 30, 2019 by guest

http://iai.asm.org/

Dow

nloaded from

3792 ANSHER ET AL.

2.4

x

a

z

as 2.0 **

0

ui~or

-J

CO)

1.0

0.0> > 0 co)

TREATMENT

FIG. 1. Hexobarbital-induced sleep time index is shown for Rand NR mice treated with DT vaccine (DT), DTP vaccine (DTP),pertussis vaccine adsorbed (PV), cholera vaccine (CHV), PIC, E.coli LPS (EC LPS), or B. pertussis LPS (BP LPS). Sleep times weremeasured 24 h after administration of agents. Control sleep times foreach group were 33 3 min. Each bar represents the mean sleeptime for four animals per group divided by the mean sleep time forcontrols. Error bars are the pooled standard error for each treat-ment. Significant differences from control: *, P < 0.05; **, P <

0.005; ***, P < 0.001.

PIC increased hexobarbital-induced sleep time 1.5- to 1.7-fold over that of controls in both strains of mice. Nodetectable endotoxic activity was present in the preparationas determined by LAL testing.Time course of hexobarbital-induced sleep time following

vaccine or endotoxin administration. Hexobarbital-inducedsleep times were compared for up to 1 week following a

single injection of B. pertussis LPS or DTP vaccine. Figure2 shows the results of one experiment. At 24 h after a singleinjection of LPS, the sleep time index was 2.3 for R micecompared with 1.5 for NR mice. At 4 days postinjection,hexobarbital-induced sleep times in the NR mice were

indistinguishable from control levels, while in the R mice,there was still a twofold increase over that of controls. By 7days postinjection, however, both groups of mice had sleeptimes that were the same as that of controls. DTP vaccine-treated mice have sleep times at 24 h increased to 1.6-fold(NR mice) and 1.8-fold (R mice) above control levels. Theseincreased sleep times persisted in both strains of mice for atleast 1 week after DTP vaccine administration.Cytochrome P-450 levels and enzyme activities. Spectral

cytochrome P-450 levels were measured 24 h after vaccine,LPS, or PIC administration. As can be seen in Table 1, bothDTP vaccine and PIC decreased P-450 levels significantly,by 30 to 40% in R and NR mice. Administration of B.pertussis or E. coli LPS significantly reduced P-450 levelsonly in R mice by 30 to 45%. In NR mice, the decrease wasabout 10%. The cytochrome P-450 levels returned to controllevels 2 days after LPS administration, while those of theDTP vaccine-treated mice remained reduced for 7 to 10 days(4).The microsomal enzyme activities benzphetamine deme-

xcmz

w

w

w-J

CO)

2.0 [

1.61

1.2

0.80

"******

0**

-w

* El

%--------------19

2 4DAYS AFTER TREATMENT

6

FIG. 2. Duration of hexobarbital-induced sleep time followingadministration of DTP vaccine or LPS. The sleep time index isshown for R (0, 0) and NR (U, l) mice after DTP vaccine (solidline) or LPS (dashed line) administration. Each point represents themean results for four animals per group divided by the mean resultsfor four controls. If no error bar is shown, the standard error wassmaller than the symbol on the graph. Results for one representativeexperiment of the four performed are shown in the graph. Sleep timeof control animals was 30 ± 3 min. Significant differences fromcontrol: *, P < 0.05; **, P < 0.005; ***, P < 0.001.

thylase, benzopyrene monooxygenase, and ethylmorphinedemethylase were also compared 24 h after vaccine or LPSadministration. Figure 3 shows the results of one repre-sentative experiment. As was observed with P-450 levels,both DTP vaccine and PIC decreased benzphetamine dem-ethylase activity about 40% in both strains of mice (Fig. 3A).LPS administration caused a marked decrease in benzphet-amine demethylase activity in R mice (50% of control) butonly a 20 to 30% decrease in NR mice. Benzopyrenemonooxygenase activity was decreased to similar extents inboth groups of mice after PIC and LPS administration (Fig.3B). In DTP vaccine-treated mice, the benzopyrene mono-oxygenase activity was decreased by'30% in R mice, but theNR mice had a slight increase in activity. In contrast,ethylmorphine activity was decreased only in NR mice afterDTP vaccine administration. Activity levels remained un-

TABLE 1. Hepatic cytochrome P-450 levelsa after administrationof DTP vaccine, PIC, or LPS

Level of hepatic cytochrome P450 in:Treatment

NR mice R mice

Expt 1Saline 0.68 ± 0.05 0.54 ± 0.02DTP vaccine 0.41 ± 0.18*** (0.61) 0.37 + 0.01** (0.68)

Expt 2Saline 0.56 ± 0.11 0.64 ± 0.02PIC 0.39 + 0.01* (0.70) 0.41 ± 0.01* (0.63)B. pertussis LPS 0.54 ± 0.04 (0.96) 0.45 ± 0.02** (0.70)E. coli LPS 0.49 ± 0.03 (0.88) 0.36 ± 0.01** (0.56)a Nanomoles per milligram of microsomal protein, measured 24 h after

administration of DTP vaccine, PIC, or LPS. Statistical significances, deter-mined by analysis of variance, for indicated groups versus saline-injectedcontrols are as follows: *, P < 0.05; **,P < 0.005; ***, P < 0.001. Each groupcontained four to six animals. Values in parentheses are fractions of controlvalues.

IN-ECT. IMMUN.

on March 30, 2019 by guest

http://iai.asm.org/

Dow

nloaded from

ENDOTOXIN AND DRUG METABOLISM 3793

[2 NR140 A *R

120

100

80 +

60 + + +

40

20 00

0LDTP PIC B. PERT E. COU140 B

-Jo 120I-100z*o so0LI. 600

4020

DTP PIC B. PERT E. COU140: C120-

DTP PIC B. PERT E. COU

TREATMENTFIG. 3. Microsomal enzyme activities 24 h following administra-

tion of DTP vaccine, PIC, or LPS. Each activity is depicted as thepercentage of that of the control. This represents the mean result forthe experimental group divided by the mean result for the controlgroup. Each group contained three samples, and the experiment wasperformed three times. Pooled standard errors for each group areshown. Significant differences: *, P < 0.05; **, P < 0.01; +, P <0.005; + +, P < 0.001. (A) Benzphetamine demethylase activity in Rand NR mice. The mean control activity was 31.0 ± 0.4 for NR miceand 30.5 ± 2.2 for R mice. The activity is expressed in nanomolesper minute per milligram of microsomal protein. (B) Benzopyrenemonooxygenase activity 24 h after administration of vaccine, PIC,or LPS. The mean control activity was 7.23 ± 0.34 for NR mice and7.49 ± 0.14 for R mice. Activity was expressed as nanomoles perhour per milligram of protein. (C) Ethylmorphine demethylaseactivity in R and NR mice. The activity was expressed as micro-grams of formaldehyde formed per hour per milligram of microsomalprotein. Control activities were 14.7 ± 0.67 for NR mice and 19.2 +1.4 for R mice.

changed or slightly increased following all other treatmentsfor both strains of mice (Fig. 3C).

Cytosolic enzyme activities were also measured 24 h aftervaccine, PIC, or endotoxin administration. Consistent withour data with C57BL/6 mice (3), no changes in glutathionetransferase or quinone reductase activity were evident at thistime point (data not shown).

Cytokine levels after administration of PIC, vaccine, orendotoxin. Cytokines were measured in plasma 1 to 24 h afteradministration of vaccine, LPS, or PIC. These results areshown in Fig. 4 and 5 for TNF and IL-6, respectively. Bothcytokines show increased levels in the plasma, which aretransient. Each cytokine had returned to control values by 8h; thus, the graphs extend only to 8 h. Basal levels of TNFfollowing injection of saline are about 10 pg/ml, and those of

R8000

6000 * LP

4000

2000Lh

Sr 2 4 6

2000 r

NRLPS

1000 1

8

2000r DTP

2 4 W6

DTP

1000 11000

00 DT

0L.

IL 1000z

O-0 2 4 6 8

2000

1000

iK a a2 4 6 8

200 rDT

1000 _

uupl p -

2000

0f 2 4 -6 8

CHV

1000 1

2 4 6

20 r Plc

1000 I

8 2 4 6 8

2000 -PIC

1000

n.6 8

Time (Hours)

2 4 6 8

FIG. 4. TNF levels in plasma of mice after administration of B.penrussis LPS, vaccine, or PIC. TNF concentration was calculatedby using a standard recombinant murine TNF. Results for onerepresentative experiment of four are shown.

IL-6 are about 10 to 20 U/ml. TNF was increased in R miceto more than 6,000 pg/ml 1.5 h following a single injection ofB. pertussis LPS (Fig. 4). This was more than sixfold higherthan the level of TNF measured after DTP vaccine admin-istration (Fig. 4). In the NR mice, this dose of LPS inducedTNF to levels of 500 pg/ml, similar to results with both DTPand cholera vaccines in this mouse strain. PIC administra-tion induced TNF to 500 pg/ml in both strains of mice. Incontrast to the sharp peak of TNF after LPS administration,TNF levels were elevated over a period of 4 h in R mice afterDTP vaccine administration. There was also a broader peakof TNF in both strains of mice after administration of PIC.

IL-6 levels in R mice were increased following adminis-tration of LPS, DTP and cholera vaccines, and PIC but notDT vaccine. The increases seen after administration of LPSand DTP vaccine were more similar in peak shape andmagnitude than those of TNF (Fig. 5). The peak IL-6 levelwas measured at 4 h with LPS at 12,000 U/ml, while for DTPvaccine the peak was at 3 h and the level was 7,500 U/ml. Incontrast, PIC induced IL-6 to 500 U/ml and cholera vaccine

a 2 A a a

r 2 4 1

VOL. 60, 1992

1% 0 0

n

. CHV

on March 30, 2019 by guest

http://iai.asm.org/

Dow

nloaded from

3794 ANSHER ET AL.

R

1200

800

400

10 FLPS

0o 4

2 4 6 8

12000 DTP

8000[_

400

E0 2000

%-W 1000~

(0

04

-

2000

1000

120C

80

40t

1204

80C

401

2 4 6 8

200DT

100

2 4 6 8

20CcHv

104

2 4 6 1

2000 r Pic

1000

204

101

2 4 6 8

Time (HouFIG. 5. IL-6 levels in plasma of mice al

pertussis LPS, vaccine, or PIC. One unit ofamount of IL-6 needed to give 50% of msamples for this experiment were the same,,legend to Fig. 4.

A

NR to 1,200 U/ml. DT vaccine administration did not induceIL-6 above control levels in either mouse strain. IL-6 was

DO LPS induced in NR mice following administration of LPS, DTPand cholera vaccines, and PIC. About 200 U/ml was de-tected 2 h after administration of either PIC or cholera

Do vaccine. DTP vaccine administration caused an increase in2 _ IL-6 to 2,000 U/ml, approximately one-third of that seen in2 4 6 8 the R mice.

Effects of PXN on hepatotoxicity of DTP vaccine or LPS.00

DTP DTP vaccine (0.5 ml) or B. pertussis LPS (50 jig) was mixedoo and incubated with 50 ,ug of polymyxin B sulfate (PXN) for

24 h before administration to mice. The LAL test-measur-o0 - able endotoxic activity was reduced from 125,000 to 125

o ; ^ EU/ml after incubation of LPS with PXN. DTP vaccine was0 2 4 6 8 reduced from a rabbit endotoxic activity of 300,000 EU/ml to

19,200 EU/ml after mixing with PXN (a 16-fold reduction).30 - DT These LAL tests were done with 1:100 dilutions of material

because this dilution was necessary for the rabbit pyrogen)o - test. Rabbit pyrogen testing showed a consistent decrease in

pyrogenicity as a result of premixing the vaccine with PXN.The mean temperature increase for three rabbits was 1.4°Cs12 4 6 8 for a 1:100 dilution of the DTP vaccine and 0.7°C for the

DO same dilution of the premixed PXN and DTP vaccine.CHV TNF and IL-6 were measured in R and NR mice 2 and 4 h

after administration of LPS or DTP vaccine alone or afterLb- premixing with PXN as described above. These results are

shown in Fig. 6 and 7 for TNF and IL-6, respectively. There0 bleft!E is a twofold decrease in IL-6 units when the PXN-premixedI12 4 6 8 DTP vaccine is administered to both NR and R mice. In

°°0 PIC contrast, PXN treatment of LPS reduced the increase in IL-6by 5-fold in NR mice at 4 h and by more than 80-fold in Rmice at 4 h.loo - Histologic examination of livers and spleens after adminis-tration of vaccine or LPS. Livers and spleens from R and NR

o° 246 8 mice treated with either DTP vaccine or LPS were examinedfor pathologic changes. Tissues were examined at days 1, 3,5, 7, 10, and 14 postinoculation. The livers and spleens of the

rs) LPS-treated mice were indistinguishable from those of sa-fter administration of B. line-injected controls at each time point. This was true foractivity is defined as the both R and NR mice. Administration of DTP vaccine re-aximal activity. Plasma sulted in a 1.5-fold increase in the liver weight and a 3-foldas those described in the increase in the spleen weight at day 7 postinoculation in both

R and NR mice. These two strains of mice exhibited similar

B

20001-

--PXN

+PXN

1500j

1000-

5001-

oV1IJl,Wj7 WY. WZ M2HR 4HR

NR

2HR 4HR

R

2HR 4HR

NR

2HR 4HR

R

DTP LPSFIG. 6. TNF levels after administration of DTP vaccine or B. pertussis LPS alone (-PXN) or premixed with PXN (+PXN). TNF levels

were measured 2 and 4 h after administration of the indicated agents as described in the legend to Fig. 4.

DO -

ol-

n

INFEcr. IMMUN.

2500r

-'e.

on March 30, 2019 by guest

http://iai.asm.org/

Dow

nloaded from

ENDOTOXIN AND DRUG METABOLISM 3795

A B

of M xz#KY.E.:X.f,.XI2HR 4HR 2HR 4HR

NR RLPS

FIG. 7. IL-6 levels after administration of DTP vaccine or B. pernwssis LPS alone (-PXN) or after premixing with PXN (+PXN). Thetreatment is the same as that described in the legend to Fig. 6. Units are the same as in Fig. 5.

histopathologic changes of the liver and spleen followingvaccine administration. By day 7, the livers of DTP-inocu-lated mice were characterized as having moderate to severerandom multifocal inflammation, with associated hepatocytenecrosis. The inflammation was primarily neutrophilic, withfewer macrophages and lymphocytes (Fig. 8a and b). Thecapsular surface of the liver had multiple granulomas (Fig.9). There were milder hepatic lesions beginning on day 3.Lesions began to resolve between days 10 and 14. Both Rand NR mice treated with DTP vaccine had splenic hyper-trophy of up to three times the normal spleen size. Theseaffected spleens also had subcapsular aggregates of neutro-

phils and less severe capsular inflammation than that whichappeared on the liver capsule (Fig. 10).

DISCUSSION

The studies reported in this article were undertaken toassess the contribution of endotoxin to adverse reactions inmice immunized with DTP vaccine. Previous studies in ourlaboratory (2-4) have demonstrated that DTP vaccine altershepatic drug-metabolizing enzymes in C57BL/6 mice. In thisstudy, R and NR mice were used to further examine theseobservations. The rationale for this approach was that if the

FIG. 8. (a) Liver of DTP vaccine-inoculated mouse at day 7. There are many random, multifocal areas of inflammation. (b) Highmagnification (x 100) of inflammation shown in panel a. The focus consists of mostly neutrophils accompanied by necrotic hepatocytes andmuch nuclear debris.

5000 r

40001

-PXN

+PXN

3000j

20001

looo0

-Ux;..x.;p2HR 4HR

NR

2HR 4HR

RDTP

VOL. 60, 1992

/1,

r7//77A

on March 30, 2019 by guest

http://iai.asm.org/

Dow

nloaded from

3796 ANSHER ET AL.

FIG. 9. Liver capsule showing granulomatous inflammation on the capsule. These foci have central cores of necrosis surrounded by a layerof neutrophils and an outer zone of histiocytes and fibroplasia. Magnification, x 10.

alterations seen after administration of DTP vaccine werecaused by endotoxin alone, then the hepatic changes ob-served would be similar to those caused by LPS in R and NRmice. If, on the other hand, the alterations were due tocomponents other than LPS, then R and NR mice wouldshow more similar responses to vaccine than to LPS.As these studies demonstrate, endotoxin cannot fully

reproduce the hepatic changes that follow DTP vaccineadministration in mice. When LPS was administered to bothstrains of mice, there was a marked difference in hexobar-bital-induced sleep times. LPS administration produced amaximum increase in hexobarbital-induced sleep time at 24 hto 2.4-fold above that of controls in R mice and an increaseto 1.5-fold above that of controls in NR mice. By day 7

#~~~~ *a '* - ,a -,,; , ._.

4b~~ ~ ~ ~ ~ ~ ~ *

FIG. 10. Spleen from a DTP vaccine-inoculated mouse. Chronic inflammation on the capsular surface and subcapsular foci of neutrophilicinffiltration are shown.

INF-ECT. IMMUN.

on March 30, 2019 by guest

http://iai.asm.org/

Dow

nloaded from

ENDOTOXIN AND DRUG METABOLISM 3797

postinjection, the sleep times had returned to control levels.In contrast, DTP vaccine administration produced a 1.6- to1.8-fold increase in sleep time in both R and NR mice at 24h, which persisted at the same level for at least 7 days. Ourprevious studies with C57BL/6 mice had shown an increasein sleep time to 1.2- to 1.3-fold above that of controls at 24 h,which gradually increased to 2-fold at 7 to 10 days postin-jection (3). Another strain of mice, B6F1, is completelyrefractory to DTP vaccine administration, while its responseto endotoxin is normal (unpublished observations). Thesestrain differences are unexplained at this time.

Spectral cytochrome P-450 levels and microsomal enzymeactivities also demonstrated differences in responses of Rand NR mice to LPS and DTP vaccine administration. DTPvaccine administration reduced cytochrome P-450 levels at24 h 30 to 40% in both strains of mice, whereas LPS (E. colior B. pernussis) caused a significant reduction in P-450 levelsonly in R mice (Table 1). The observed effects of LPSadministration are consistent with other reports of alter-ations in hepatic drug metabolism in rodents (8, 12, 15, 28,35, 36). However, this is the first report describing alter-ations in hexobarbital-induced sleep time, cytochrome P-450levels, and other microsomal enzyme activities by LPS andvaccine administration in R and NR mice.To further understand the differences between results with

LPS and DTP vaccine, PIC was administered to R and NRmice. PIC is thought to inhibit drug metabolism by inductionof alpha/beta interferon in response to its administration (33,40). It has been well documented that administration of PICdecreases hepatic drug-metabolizing enzymes in severalstrains of mice. When hexobarbital-induced sleep time mea-surements and cytochrome P-450 levels were examined, PICadministration produced effects similar to those of vaccineadministration. Sleep times were elevated 1.6- to 1.8-foldabove controls for both strains of mice, while P-450 levelswere decreased in a manner similar to those seen after DTPvaccine administration. This suggests that these two strainsof mice have similar responses to other active compoundseven though they differ dramatically in their responses toendotoxin.

Studies with PXN were designed so that neutralization ofthe endotoxin present in the DTP vaccine would allow for astraightforward evaluation of the other components in thevaccine. Unfortunately, as described in the text, while PXNcompletely neutralized soluble LPS, a significant amount ofdetectable LPS remained in the DTP vaccine. This is con-sistent with earlier studies in which unadsorbed pertussisvaccine mixed with PXN failed to show a reduction inendotoxic activity at doses of PXN comparable to those usedin this study (6). Although the PXN treatment should reducethe endotoxic activity, the failure to completely eliminatemeasurable endotoxin complicates the interpretation of thesleep time data. It is not possible to rule out the possibilitythat residual LPS contributes to the failure of DTP vaccinepremixed with PXN to reduce sleep time to control levels, asseen when LPS is premixed with PXN. Hexobarbital-in-duced sleep times were measured for mice injected with DTvaccine premixed with LPS to simulate bound endotoxin.This failed to cause increases in sleep time that persisted forlonger than 24 h (as seen with LPS alone). Other whole-cellvaccines, such as cholera vaccine, only increased sleep timefor 24 h, while fluid pertussis vaccine caused increased sleeptimes similar to that caused by DTP vaccine (reference 3 andunpublished observations).

Histopathologic examination supported the thesis thatendotoxin alone does not account for the alteration in

hepatic function caused by DTP vaccine. Purified LPScaused no lesions in either R or NR mice. Additionally, thesemouse strains had lesions similar in severity and type ofinflammation at 7 days following vaccine administration.Both R and NR DTP vaccine-inoculated mice had extensivemultiple foci of inflammation in the liver, which consisted ofmainly neutrophils accompanied by hepatic necrosis. Theextent of the lesions was severe enough to result in the lossof the ability to metabolize hexobarbital, thus resulting inincreased sleep times. The severity of lesions at day 7postinoculation coincides with our finding that sleep timesremained elevated in DTP vaccine-inoculated mice at 7 days.The livers from mice at days 10 and 14 after DTP vaccineinoculation showed gradual resolution of lesions. Again, thiscoincides with the return of normal sleep times.The observed differences in the responses of R and NR

mice following LPS and vaccine administration stronglysuggest that LPS is not the sole component responsible forthe hepatic alterations to the DTP vaccine. NR mice differfrom R mice at a single gene locus on chromosome 4 (28).The most recent evidence suggests that the LPS locus iscomposed of genes which encode cellular components in-volved in signal transduction (28). It has been shown that NRmice have normal levels of LPS-binding proteins and CD14receptors (16, 18, 24). Thus, if one bypasses the receptor forLPS and introduces a toxic substance that exerts its effectslater in the pathway, the results with R and NR mice wouldbe expected to be similar (36). This is apparently the casewhen DTP vaccine and PIC are administered. In vitro,similar results have been obtained by using phorbol myristicacid in macrophages derived from R and NR mice. Thisprotein kinase C activator induces similar changes in thesestrains of mice (34).TNF, IL-1, and IL-6 levels are known to be increased

following administration of endotoxin in animals (25) andfollowing septicemia in humans (9). TNF and IL-6 levelswere measured to determine the role these cytokines mightplay in toxicity. In contrast to the hexobarbital-inducedsleep time results, LPS, DTP vaccine, and cholera vaccinecaused marked increases in TNF and IL-6 levels in R mice.NR mice showed only a small increase in these cytokines inresponse to these agents. DT vaccine administration did notincrease TNF or IL-6 levels in either mouse strain, and PICinduced small increases in both cytokines in both strains ofmice. The similarity in the inductions of TNF and IL-6 byLPS and DTP vaccine in R mice suggests a role for endo-toxin in the early response of these mice to vaccine admin-istration. The shape of the TNF curve following DTP vac-cine administration was unexpected. It may be that theadjuvant prolongs stimulation of TNF release. This effectwas not seen with cholera vaccine or the other agents tested.Neutralization of the endotoxin by PXN reduced the ele-vated levels of cytokines in the plasma of vaccine-treatedmice as well as in the plasma of LPS-treated mice. Thecytokine levels were not reduced as much when PXN waspremixed with DTP vaccine as when PXN was premixedwith LPS. Again, this could be due to incomplete neutrali-zation of the endotoxin in the vaccine. However, the findingof reduced levels of cytokines in the plasma of mice treatedwith PXN premixed with DTP vaccine further supports arole for endotoxin in the early stage of vaccine-inducedalteration of hepatic drug metabolism.

Further studies are in progress to identify the componentsof the DTP vaccine which contribute to the alterations inhepatic drug metabolism described in this article as well as toidentify ways to more effectively neutralize the endotoxin.

VOL. 60, 1992

on March 30, 2019 by guest

http://iai.asm.org/

Dow

nloaded from

3798 ANSHER ET AL.

REFERENCES1. Aarden, L., E. DeGroot, D. Schaap, and P. Lansdorp. 1987.

Production of hybridoma growth factor by human monocytes.Eur. J. Immunol. 17:1411-1416.

2. Ansher, S., R. Puri, W. Thompson, and W. Habig. 1992. Theeffects of interleukin-2 and a-interferon administration on he-patic drug metabolism in mice. Cancer Res. 52:262-266.

3. Ansher, S., W. Thompson, and W. Habig. 1991. Vaccine-induced alterations in hepatic drug metabolism. Vaccine 9:277-283.

4. Ansher, S., W. Thompson, and W. Habig. 1992. The role ofendotoxin in DTP vaccine-induced hepatotoxicity. J. Cell. Bio-chem. Suppl. 16C:169.

5. Astrom, A., J. Meier, and J. DePierre. 1983. Characterizationof the microsomal cytochrome P-450 species induced in rat liverby 2-acetylaminofluorene. Cancer Res. 43:342-348.

6. Bannatyne, R., and R. Cheung. 1981. Reducing the endotoxicactivity of pertussis vaccine. J. Hyg. Camb. 87:377-381.

7. Benson, A., M. Hunkeler, and P. Talalay. 1980. Increase ofNAD(P)H:quinone reductase by dietary antioxidants: possiblerole in protection against carcinogenesis and toxicity. Proc.Natl. Acad. Sci. USA 77:5216-5220.

8. Bertini, R., M. Bianchi, A. Erroi, P. Villa, and P. Ghezzi. 1989.Dexamethasone modulation of in vivo effects of endotoxin,tumor necrosis factor, and interleukin-1 on liver cytochromeP-450, plasma fibrinogen, and serum iron. J. Leukocyte Biol.46:254-262.

9. Calandra, T., J. Gerain, D. Heumann, J.-D. Baumgartner, andM. Glauser. 1991. High circulating levels of interleukin-6 inpatients with septic shock: evolution during sepsis, prognosticvalue, and interplay with other cytokines. Am. J. Med. 91:23-29.

10. Centers for Disease Control. 1991. Diphtheria, tetanus, andpertussis: recommendations for vaccine use and other preven-tive measures. Morbid. Mortal. Weekly Rep. 40:1-28.

11. Cody, C. L., L. J. Baraff, J. D. Cherry, S. M. Marcy, and C. R.Manclark. 1981. Nature and rates of adverse reactions associ-ated with DTP and DT immunizations in infants and children.Pediatrics 68:650-660.

12. Coto, J., and J. Williams. 1989. Effect of endotoxin to differen-tially affect cytochrome P-450 monooxygenase activities ofuntreated rats and animals induced with phenobarbital or 3-me-thylcholanthrene. Int. J. Immunopharm. 11:623-628.

13. DePierre, J., K. Johannesen, M. Moron, and J. Seidegard. 1978.Radioactive assay of aryl hydrocarbon monooxygenase andepoxide hydrolase. Methods Enzymol. 52:412-415.

14. Ernster, L. 1967. DT diaphorase. Methods Enzymol. 10:309-317.

15. Ferluga, J., A. Kaplun, and A. Allison. 1979. Protection of miceagainst endotoxin-induced liver damage by anti-inflammatorydrugs. Agents Actions 9:566-574.

16. Gregory, S., D. Zimmerman, and M. Kern. 1980. The lipid Amoiety of lipopolysaccharide is specifically bound to B cellsubpopulations of responder and nonresponder animals. J.Immunol. 125:102-107.

17. Habig, W., M. Pabst, and W. Jakoby. 1974. Glutathione-S-transferases, the first enzymatic step in mercapturic acid forma-tion. J. Biol. Chem. 249:7130-7139.

18. Haeffner-Cavaillon, N., J.-M. Cavaillon, and L. Szabo. 1985.Cellular receptors for endotoxin, p. 1-24. In L. J. Berry (ed.),Handbook of endotoxin, vol. 3. Elsevier/North Holland Bio-medical Press, New York.

19. Hannan, S., J. May, D. Pratt, W. Richtsmeier, and J. Bertino.1988. The effect of whole virus influenza vaccination on theo-phylline pharmacokinetics. Am. Rev. Respir. Dis. 137:903-906.

20. Hochstein, D., R. Elin, J. Cooper, and S. Wolff. 1973. Furtherdevelopments of limulus amebocyte lysate test. Bull. Parenter.Drug Assoc. 27:139.

21. Hodgeboom, G. 1955. Fractionation of cell components of

animal tissues. Methods Enzymol. 1:16-19.22. Jay, G. 1955. Variation in response of various mouse strains to

hexobarbital (Evipal). Proc. Soc. Exp. Biol. Med. 90:378-380.23. Kramer, P., and C. McClain. 1981. Depression of aminopyrine

metabolism by influenza vaccination. N. Engl. J. Med. 305:1262-1264.

24. Lei, M.-G., and D. C. Morrison. 1988. Specific endotoxiclipopolysaccharide-binding proteins on murine splenocytes. I.Detection of lipopolysaccharide-binding sites on splenocytesand splenocyte subpopulations. J. Immunol. 141:996-1005.

25. LeMay, D., L. LeMay, M. Kluger, and L. D'Alecy. 1990. Plasmaprofiles of IL-6 and TNF with fever-inducing doses of lipopoly-saccharide in dogs. Am. J. Physiol. 259:R129-R132.

26. Lery, L., Y. Rotivel, A. Guiliermain, and J. Descotes. 1989.Activity of tetanus vaccines on the enzymatic hepatic drugmetabolism, p. 428-433. In G. Nistico, B. Bizzini, B. Bytch-enko, and R. Triau (ed.), Eighth International Conference onTetanus. Pythagora Press, Rome.

27. Mannering, G., K. Renton, R. El-Azhary, and L. Deloria. 1980.Effects of interferon-inducing agents on hepatic cytochromeP-450 drug metabolizing enzymes. Ann. N.Y. Acad. Sci. 350:314-331.

28. Michalek, S., and J. Katz. 1990. Mechanisms of immunomodu-lation by endotoxin: an overview, p. 355-364. In A. Nowotny, J.Spitzer, and E. Ziegler (ed.), Cellular and molecular aspects ofendotoxin reactions, vol. 1. Elsevier Science Publishers B.V.,Amsterdam.

29. Nash, T. 1953. The colorimetric estimation of formaldehyde bymeans of the Hantzsch reaction. Biochem. J. 55:416-421.

30. Nencioni, L., M. Pizza, M. Bugnoli, T. De Magistris, A. DiTommaso, F. Giovannoni, R. Manetti, I. Marsili, G. Matteucci,D. Nucci, R. Olivieri, P. Pileri, R. Presentini, L. Villa, J. G.Kreeftenberg, S. Silvestri, A. Tagliabue, and R. Rappuoli. 1990.Characterization of genetically inactivated pertussis toxin mu-tants: candidates for a new vaccine against whooping cough.Infect. Immun. 58:1308-1315.

31. Omura, T., and R. Sato. 1964. The carbon monoxide-bindingpigment of liver microsomes. J. Biol. Chem. 239:2370-2378.

32. Renton, K., J. Gray, and R. Hall. 1980. Decreased elimination oftheophylline after influenza vaccination. Can. Med. Assoc. J.123:288-290.

33. Renton, K., and G. Mannering. 1976. Depression of hepaticcytochrome P-450-dependent monooxygenase systems with ad-ministered interferon inducing agents. Biochem. Biophys. Res.Commun. 73:343-348.

34. Romanus, V., R. Jonsell, and S.-O. Berquist. 1987. Pertussis inSweden after the cessation of general immunization in 1979.Pediatr. Infect. Dis. J. 6:364-371.

35. Shibayama, Y., S. Asaka, and K. Nakata. 1991. Endotoxinhepatotoxicity augmented by ethanol. Exp. Mol. Pathol. 55:196-202.

36. Shinomiya, H., H. Hirata, and M. Nakano. 1991. Purificationand characterization of the 65-kDa protein phosphorylated inmurine macrophages by stimulation with bacterial liposaccha-ride. J. Biol. Chem. 146:3617-3625.

37. Sidey, F. M., B. L. Furman, and A. C. Wardlaw. 1989. Effect ofhyperreactivity to endotoxin on the toxicity of pertussis vaccineand pertussis toxin in mice. Vaccine 7:237-241.

38. Smith, P., R. Crohn, G. Hermanson, A. Mallia, F. Gartner, M.Provenzano, E. Fujimoto, N. Goeke, B. Olson, and B. Klenk.1985. Protein assay using bicinchoninic acid. Anal. Chem.150:76-85.

39. Tosato, G., J. Tanner, K. D. Jones, M. Revel, and S. E. Pike.1990. Identification of interleukin-6 as an autocrine growthfactor for Epstein-Barr virus-immortalized B cells. J. Virol.64:3033-3041.

40. Zerkle, J., and A. Wade. 1984. Depressant effects of theinterferon inducer polyribosinic:polyribocytidylic acid on cy-tochrome P-450 hemoproteins. Pharmacology 28:1-11.

INFEC-F. IMMUN.

on March 30, 2019 by guest

http://iai.asm.org/

Dow

nloaded from