INFECTION AND IMMUNITY, Jan. 1991, p. 351-3560019-9567/91/010351-06$02.00/0Copyright C 1991, American Society for Microbiology

Rickettsial Effects on Leukotriene and Prostaglandin Secretion byMouse Polymorphonuclear Leukocytes

T. STUART WALKER* AND CATHY S. HOOVERt

Muncie Center for Medical Education, Indiana University School of Medicine and Ball State University, andDepartment of Biology, Ball State University, Muncie, Indiana 47306

Received 15 June 1990/Accepted 8 October 1990

Typhus rickettsiae were incubated with mouse exudative polymorphonuclear leukocytes (PMN), andsupernatants were examined for leukotriene B4 (LTB4) and prostaglandin E2 (PGE2) secretion by radioimmu-noassay. PMN incubated with native rickettsiae secreted significantly more LTB4 and PGE2 than did thoseincubated with buffer alone. Autacoid secretion was dependent on both the time of PMN incubation withrickettsiae and the number of rickettsiae present in the incubation suspension. Rickettsial stimulation of LTB4secretion was associated with rickettsial hemolytic activity; treatments which inactivated the rickettsialhemolysin abolished the ability of rickettsiae to stimulate PMN LTB4 secretion. Trifluoperazine, which did notalter the rate of phagocytosis of rickettsiae by PMN, stimulated rickettsial effects on secretion of both LTB4 andPGE2 but inhibited the PMN LTB4 response to A23187. This suggested that the PMN response to rickettsiaeand to the calcium ionophore involved differing mechanisms of activation. Finally, rabbit antirickettsialantiserum, which inhibited rickettsial hemolysis and was opsonic, did not block the effects of rickettsiae onPMN LTB4 secretion.

Few investigators have examined interactions betweentyphus rickettsiae and polymorphonuclear leukocytes(PMN). This is probably because of the putative relativeimportance of the longer-lived macrophage in modulating theoutcome of acute typhus (7, 12, 36). Nevertheless, rickett-siae must interact with PMN during the course of infection,and the results of these interactions may play a significantrole in the course of rickettsial disease. The studies ofWisseman and Tabor (35) demonstrated that there was aPMN response early during the course of infection withRickettsia prowazekii and that both the extent of the re-sponse and the rates of phagocytosis were greater in immunethan in nonimmune individuals. In vitro studies by Wisse-man et al. (32, 33) with killed rickettsiae and by Walker andWinkler (22) with live rickettsiae showed that rickettsiaewere phagocytized poorly in the absence of specific anti-rickettsial antibody. Indeed, Walker and Winkler (22) re-ported that hemolytically active rickettsiae damaged PMNsufficiently to release cytoplasmic lactic dehydrogenase andthat increasing the number of extracellular rickettsiae byincreasing the multiplicity of infection (MOI) resulted in arapid increase in the amount of PMN lactic dehydrogenasereleased into the external environment while increasing onlymodestly the numbers of rickettsiae phagocytized. Theyhypothesized that rickettsiae in the suspension colliding withPMN activated the rickettsial hemolytic mechanism and thatthe repeated action of this phospholipase activity (29-31)damaged the integrity of the PMN.PMN are known to respond to membrane phospholipase A

activity by secreting arachidonate-derived autacoids (6, 11)such as leukotriene B4 (LTB4) and prostaglandin E2 (PGE2).Because the extensive PMN membrane damage reported byWalker and Winkler (22) could result in generation of phar-macologically active autacoids, we examined the ability of

* Corresponding author.t Present address: Biochemicals Building, Boehringer-Mannheim

Biochemicals, Indianapolis, IN 46250.

PMN exposed to a high rickettsial MOI to secrete LTB4 andPGE2. We report that such PMN secreted increasedamounts of both autacoids and that the PMN autacoidresponse corresponded to expression of the rickettsial he-molytic activity. We also report that rickettsiae elicitedLTB4 and PGE2 by a mechanism which differed from thatresponsible for the ability of the calcium ionophore A23187to elicit autacoid generation.

MATERIALS AND METHODS

Rickettsial growth and preparation. Rickettsia prowazekii,Madrid E strain, was propagated in 6-day-embryonatedantibiotic-free chicken eggs after inoculation from a seedpool (yolk sac passage 272). Rickettsial suspensions wereprepared from heavily infected yolk sacs by a modification(25) of the methods of Bovarnick and Snyder (3) andWisseman et al. (34). Rickettsial viability was monitored bya modification of the hemolytic procedure of Snyder et al.(16). Rickettsiae were used immediately after harvest orwere frozen at -80°C for later use. Rickettsialike bodieswere enumerated by a modification (20) of the method ofSilberman and Fiset (15). The number of hemolyticallyactive rickettsiae was determined by a modification of theantibody hemolysis method of Walker and Winkler (21).Because some rickettsial viability was lost upon freezing andthawing, most results were expressed as a function of thenumber of hemolytically active rickettsiae.The diluent for rickettsial suspension during the purifica-

tion procedure was a sucrose-phosphate-glutamate solution(SPG), described by Bovarnick et al. (2), and the diluent forexperiments was SPG containing 0.01 M MgCl2 and 0.2%D-glucose (SPGMgGlu).

Preparation of mouse PMN. Mouse exudative PMN wereprepared by harvesting peritoneal exudates from BALB/cmice that had been stimulated 5 h previously by intraperito-neal injection with 2.5 ml of a 3% sterile thioglycolatesolution (1). The exudates were collected from killed mice,suspended in a modified Hanks solution (9) containing 0.01%

351

Vol. 59, No. 1

on January 29, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

352 WALKER AND HOOVER

bovine serum albumin and 0.1% glucose (HBG), washedtwice in this buffer, and then suspended in HBG. The cellswere enumerated in a hemacytometer, and the percentage ofPMN in the suspension was determined by staining the cellsfor ao-naphthylacetate esterase and for naphthol AS-D chlo-roacetate esterase; exudates used in these experiments con-tained >90% PMN as determined by histochemical staining.

Phagocytosis of rickettsiae. Rickettsiae labeled with [oa-32P]ATP were suspended in SPGMgGlu, and the phagocyto-sis of labeled rickettsiae by PMN was determined as previ-ously described (19, 20, 22). This method, which is based onthe ability of typhus rickettsiae to transport ATP by obligateexchange (26), distinguishes intracellular and adherent rick-ettsiae and takes into account the loss of radioactivity fromextracellular rickettsiae during incubation as well as accu-mulation of nonrickettsial labeled material from the incuba-tion medium by the target cells. Results are expressed as thenumber of adherent or intracellular rickettsiae per PMN.The experiments were performed with triplicate samples.PMN secretion of autacoids. PMN were incubated with

rickettsiae at 37°C at various MOI for up to 120 min, and theamounts of LTB4 and PGE2 present in the supernatants weredetermined by using commercially available radioimmunoas-says. Because preliminary results indicated that PMN had tobe incubated with high multiplicities of rickettsiae for detec-tion of changes in autacoid generation by small numbers ofPMN over a short time frame, most experiments examinedthe secretion of LTB4 and PGE2 by PMN incubated withapproximately 500 rickettsiae per PMN. Briefly, 1 x 106PMN in a tube were suspended in 1 ml of SPGMg containing5 x 108 rickettsiae. SPGMg was used for these experimentsbecause we wished to maximize the effects of the rickettsiaeon the PMN; SPGMg has been shown (2) to optimallypreserve rickettsial activity in vitro. The suspension wasincubated for 60 min at 37°C with tumbling. At the end of theincubation period, the suspension was centrifuged at 4°C for10 min (850 x g) to pellet the PMN, and a sample of thesupernatant was examined for the presence ofLTB4 or PGE2by radioimmunoassay. Experiments were typically con-ducted in triplicate. Controls incubated in parallel weretreated identically to the experimental samples but containedno rickettsiae. Preliminary results indicated that PMN incu-bated with similar numbers of Escherichia coli did notsecrete increased amounts of LTB4 and that rickettsiaeincubated alone (no PMN) did not secrete LTB4.

Statistics. Results were assessed for statistical significanceby using a commercially available statistics package (NCSS,Kaysville, Utah). Typically, data were examined by analysisof variance; when significant variation among groups wasdetected, individual groups were compared by using Dun-can's post-hoc test.

Materials. Eggs for rickettsial culture were obtained fromTruslow Farms, Chestertown, Md. The radioimmunoassayfor LTB4 was obtained from Amersham Corp., ArlingtonHeights, Ill., and that for PGE2 was obtained from DupontNEN, North Billerica, Mass. [a-32P]ATP was obtained fromICN Chemical and Radioisotope Division, Irvine, Calif.BALB/c AnNHsd BR mice were obtained from HarlanSprague Dawley Corp., Indianapolis, Ind. Kits for determin-ing a-naphthyl acetate esterase and naphthol AS-D chloro-acetate esterase were obtained from Sigma Chemical Co.,St. Louis, Mo. Other chemicals and their sources were asfollows: bovine serum albumin (fraction V), N-ethylmaleim-ide (NEM), 2-mercaptoethanol, atractyloside, ATP, trifluo-perazine (TFPZ), p-bromophenacyl bromide (Pbpb), andA23187, Sigma Chemical Co., St. Louis, Mo.; Triton X-100,

TABLE 1. Effect of rickettsiae on PMN leukotriene secretion

Constituents of PMN incubation Amt of LTB4 pb(no. of expt) secreted (pg)a

Buffer only (28) 262 + 46 NARickettsiae (25) 493 ± 64 <0.01NEM-inactivated rickettsiae (4) 240 + 82 NSFormalin-inactivated rickettsiae (3) 152 ± 58 NS

a Expressed as picograms of LTB4 secreted by 1 x 106 PMN in 60 minstandard error of the mean.

b Probability that the mean for each treatment is significantly different fromthe value obtained when PMN were incubated with buffer alone (by Duncan'spost-hoc test). NA, Not applicable; NS, not significant.

Dupont NEN, N. Billerica, Mass.; formaldehyde, FisherScientific Co., Chicago, Ill.

RESULTS

Rickettsial stimulation of PMN LTB4 secretion. Rickettsiaewere incubated with exudative mouse PMN for 60 min at37°C at an MOI of 500, and supernatants were examined forthe presence of LTB4. PMN incubated with rickettsiaesecreted significantly more (P < 0.01; Duncan's post-hoctest) LTB4 than did those incubated with buffer alone (Table1). This rickettsial effect on PMN occurred only when PMNwere incubated with hemolytically active rickettsiae; PMNincubated with hemolytically inactive rickettsiae (rickettsiaeinactivated by treatment with NEM or Formalin) secreted nomore LTB4 than did sham-incubated PMN. The stimulatoryeffect of native rickettsiae on PMN LTB4 secretion wasrelated to the time of PMN exposure to rickettsiae (r =0.953) (Fig. 1). Furthermore, the amount of LTB4 secretedby PMN was directly related (r = 0.986) to the number ofhemolytically active rickettsiae present in the incubationsuspension (Fig. 2).Mechanism of rickettsial effects on PMN. Rickettsiae are

known to damage host cells by expression of a phospholi-pase A activity (28-31). It is not known, however, whetherthis activity is due to a rickettsial phospholipase or whether

z PMN Incubated with:

750 Rlckettslae *Buffer, alone o

n 500

3250 -

I-

30 60 90 120

Time (minutes)FIG. 1. PMN secretion of LTB4 as a function of time. Mouse

exudative PMN were suspended in buffer alone or in buffer contain-ing typhus rickettsiae (MOI = 500) and were incubated at 37°C withtumbling. At various times, the supernatants were examined for thepresence of LTB4. The results presented are means and standarddeviations from five experiments.

INFECT. IMMUN.

on January 29, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

RICKETTSIAL EFFECTS ON PMN 353

z2 1000 P<.0la. ~~r=.986x

P<.05P<.05

500 ,

IL0 250 500 1000

MOIFIG. 2. Relationship of rickettsial MOI to PMN LTB4 secretion.

PMN were incubated with the indicated MOI of rickettsiae for 60min at 37°C with tumbling, and supernatants were examined for thepresence of LTB4. The results presented are means and standarddeviations from eight experiments. The P values indicated arecomparisons of LTB4 secretion at each MOI with the value obtainedin the absence of rickettsiae (by the paired t test).

rickettsiae activate a host membrane phospholipase. Weexamined the mechanism of rickettsial effects on PMN byincubating PMN and rickettsiae in the presence of an inhib-itor (23) of calmodulin activity (TFPZ) or an inhibitor (18) ofhistidine-containing phospholipases (Pbpb). Because the cal-cium ionophore A23187 is known to stimulate PMN LTB4secretion by activation of a calmodulin-dependent endoge-nous phospholipase A (5), some PMN were stimulated withA23187 in the presence or absence of these inhibitors. Table2 presents the results of these studies. Although TFPZ andPbpb had no detected effect on the secretion of LTB4 byresting PMN, TFPZ strongly potentiated (P < 0.01; Dun-can's post-hoc test) the effect of rickettsiae on PMN LTB4secretion while significantly inhibiting (P < 0.01) the PMNresponse to A23187. In comparison, Pbpb inhibited therickettsial effect on PMN but had no discernible effect onionophore-mediated LTB4 secretion. Because treatment ef-fects on rickettsiae might be due to reduced rickettsialviability or to reduced rickettsial interaction with PMN, we

TABLE 2. Effects of treatments on PMN LTB4 secretion

Constituents of PMN Autacoid secretionaincubation (no. of expt) LTB4 PGE2

Buffer alone (7) 252 ± 36 4,078 ± 517TFPZ (7) 269 ± 62 1,676 ± 750Pbpb (7) 200 ± 66 2,029 ± 422

Rickettsiae (5) 404 ± 57 5,941 ± 834Rickettsiae + TFPZ (5) 1,104 ± 311 3,101 ± 188Rickettsiae + Pbpb (5) 129 ± 59 2,684 ± 190

A23187 (3) 2,565 ± 387 3,180 ± 82A23187 + TFPZ (3) 751 ± 206 NDbA23187 + Pbpb (3) 2,276 ± 610 ND

a Expressed as picograms secreted by 1 x 106 PMN in 60 min at 37°C +standard error of the mean.

b ND, Not determined.

TABLE 3. Phagocytosis of rickettsiae by mouse PMN in thepresence of effectors

Constituents of PMN incubation No. of rickettsiae/cellb(no. of expt)a Total Intracellular

Rickettsiae (7) 19.7 ± 3.1 8.9 ± 1.8NEM-inactivated rickettsiae (2) 13.6 ± 4.0 9.1 ± 2.5Rickettsiae + antiserumc (2) 380.6 ± 13.1 91.0 ± 6.5Rickettsiae + TFPZ (2) 16.8 ± 0.2 7.7 ± 3.0Rickettsiae + Pbpb (2) 12.7 ± 1.8 8.3 ± 1.6

a Incubation was carried out for 60 min at 37°C.b Results are expressed as the mean number of rickettsiae associated with

each PMN ± standard error of the mean.' Native rickettsiae in the presence of rabbit antirickettsial antiserum.

examined PMN phagocytosis of rickettsiae in the presenceand absence of these effectors. Neither inactivation of rick-ettsiae nor incubation of rickettsiae and PMN in the pres-ence of TFPZ or Pbpb significantly altered the number ofrickettsiae phagocytized in 60 min (Table 3). When PMNwere incubated with inactivated rickettsiae, or with nativerickettsiae in the presence of Pbpb, however, fewer rickett-siae adhered to the PMN (4.5 and 4.4 adherent rickettsiae,respectively, per PMN) than when PMN were incubatedwith native rickettsiae alone (10.8 adherent rickettsiae perPMN). This is consistent with our observation that both ofthese treatments significantly inhibited rickettsial hemolysis(Table 4) and that rickettsiae treated with NEM or Pbpb didnot transport lysine normally; ricksettial lysine transportwas inhibited 62% + 6.8% in the presence of 2 ,uM Pbpb.Therefore, the Pbpb effect on rickettsial stimulation of LTB4secretion may have been at least partially related to theeffect of this substance on rickettsial viability.Because PMN are known to secrete the immunosuppres-

sive prostaglandin PGE2 (11), we examined rickettsial effectson PMN PGE2 secretion. PMN incubated with native rick-ettsiae secreted significantly more (P < 0.05; Duncan'spost-hoc test) PGE2 than did those incubated with bufferalone (Table 2). Unlike in the studies described above,however, PMN did not respond to either A23187 or thrombin(data not shown) by secreting additional amounts of PGE2.This may have been due to the relatively stimulated state ofthe exudative PMN. Both TFPZ and Pbpb significantlyinhibited (P < 0.05; Duncan's post-hoc test) resting PMNsecretion of PGE2. Although it appeared that both treat-ments also inhibited rickettsial stimulation of PGE2 secre-tion, when the effects of these treatments on PGE2 secretionby resting PMN were taken into account, TFPZ significantlystimulated PGE2 secretion (P < 0.01) and Pbpb reducedPGE2 secretion modestly (P < 0.05). Therefore, the patternsof rickettsia-stimulated PMN LTB4 and PGE2 secretion in

TABLE 4. Effect of treatments on rickettsial hemolytic activityConstituents of incubation with sheep Hemolysis'erythrocytes (no. of expt.)Rickettsiae....................... 100NEM-inactivated rickettsiae (6) ................ ....... 8.3 + 4.0Rickettsiae + TFPZ (6) ....................... 81.2 + 6.7Rickettsiae + Pbpb (5) ....................... 60.4 + 16.4Rickettsiae + antiserumb (4) ........... ............ 6.0 + 2.3

a Expressed as a percentage of the amount of hemolysis observed whensheep erythrocytes were incubated with native rickettsiae alone ± standarderror of the mean.

b Native rickettsiae in the presence of rabbit antirickettsial antiserum.

VOL. 59, 1991

on January 29, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

354 WALKER AND HOOVER

TABLE 5. Effect of antirickettsial antiserum on PMN secretion ofarachidonate-derived autacoids

Constituents of PMN Amt of LTB4 Amt of PGE2incubation secreted (pg) secreted (pg)

(no. of expt)Y (no. of expt)aBuffer alone 196 ± 31 (4) 5,340 ± 447 (3)Rickettsiae 433 ± 78 (4) 7,369 ± 391 (3)Rickettsiae + NRS 236 ± 55 (4) 6,679 ± 293 (3)Rickettsiae + antiserum' 260 ± 78 (4) 6,407 ± 110 (3)

a Results are expressed as the mean amount of autacoid secreted by 1 x 106PMN in 60 min ± standard error of the mean.

b Native rickettsiae in the presence of rabbit antirickettsial antiserum.

the presence of these effectors were generally similar, al-though not identical.

Effect of antibody on rickettsial stimulation of LTB4 andPGE2 secretion. Because rabbit antiserum raised againstrickettsiae (but not preimmune serum) blocked rickettsialhemolysis (Table 4) and was opsonic (Table 3), we incubatedPMN with rickettsiae in the presence of preimmune serum(normal rabbit serum [NRS]) or hyperimmune antirickettsialantiserum and assessed LTB4 and PGE2 secretion; theresults are presented in Table 5. Both NRS and antirickett-sial antiserum reduced the absolute response ofPMN to thepresence of rickettsiae. However, when the fold responsesof PMN LTB4 or PGE2 secretion in the presence of rickett-siae suspended in buffer alone or in buffer containing NRS orimmune serum were compared with those of matched appro-priate controls, there were no significant differences amongthe three groups; NRS reduced LTB4 secretion by PMN inthe absence of rickettsiae by about 50%. Therefore, theantirickettsial antiserum used in this study blocked hemoly-sis and opsonized without altering rickettsial effects on PMNLTB4 or PGE2 secretion.

DISCUSSION

In our earlier studies (22) with rickettsiae and rabbit PMN,we demonstrated that although rickettsiae were phagocy-tized poorly, they could damage PMN. We further demon-strated that this damage probably occurred "from without"and was associated with the expression of the rickettsialhemolytic activity. We have extended these studies to dem-onstrate that rickettsiae incubated at a high MOI with mouseexudative PMN caused the secretion of increased amountsof the arachidonate-derived autacoids LTB4 and PGE2. Asbefore, rickettsiae were phagocytized poorly by the PMN.We used extremely high rickettsial multiplicities to acceler-ate the effects of the rickettsiae on the PMN, yet even aftera 60-min incubation, PMN had phagocytized only approxi-mately 9 rickettsiae per PMN, with an additional 11 adherentrickettsiae per cell. This stands in stark contrast to ourearlier demonstration (22) that E. coli or type 4 gonococcisimilarly incubated with PMN were rapidly phagocytized.The mechanism of this relative resistance of rickettsiae tophagocytosis is not understood, but it appears that the closeinteraction of a small number of rickettsiae with the PMN isable to significantly change PMN autacoid secretion.What is the cause of this rickettsial effect on PMN? We

considered the possibility that rickettsiae stimulated activa-tion of the PMN membrane phospholipase in a fashionanalagous to that of other cell-activating substances such asthe calcium ionophore A23187 (6) or thrombin (4). It appearsthat this was not the case. LTB4 secretion in response to

A23187 was sensitive to the inhibitory action of TFPZ, aninhibitor of calmodulin activity (23); in contrast, the LTB4response to rickettsiae was potentiated by this effector.Additionally, although PMN did not respond to eitherA23187 or thrombin by secreting additional amounts ofPGE2, they did secrete PGE2 in the presence of addedrickettsiae. Therefore, it appears that the PMN responses torickettsiae and to ionophores were not identical. These datasuggest that the phospholipase responsible for autacoidsecretion in the presence of rickettsiae was not the same asthat activated by A23187. Because there were differences inthe extent of TFPZ and Pbpb effects on PMN LTB4 andPGE2 responses, and because A23187 and thrombin did notstimulate PMN PGE2 secretion, we also suspect that themeans by which rickettsiae stimulated the secretion of thetwo autacoids were not identical. Studies are needed tofurther characterize the differential effects of rickettsiae onPMN secretion of LTB4 and PGE2.

It is possible that interaction with a rickettsial particleresulting in phagocytosis is sufficient to activate the endog-enous PMN phospholipase. We report that although therewas no significant difference in the numbers of rickettsiaephagocytized when PMN were incubated with living orkilled rickettsiae, hemolytically inactive rickettsiae did notstimulate LTB4 secretion. In our preliminary studies (datanot presented), we could not detect any significant changesin PMN LTB4 secretion when PMN were incubated withliving E. coli cells under conditions identical to those used toincubate PMN and rickettsiae. Therefore, it seems unlikelythat PMN phagocytosis of rickettsial particles was respon-sible for increasing the amount of LTB4 secreted by thePMN. This possible association of a putative bacterial phos-pholipase with stimulation ofPMN leukotriene generation isnot unique, although it is unusual; we cite two such studiesfor comparison. First, Suttorp et al. (17) reported thatstaphylococcal alpha-toxin stimulated leukotriene genera-tion in rabbit PMN. They reported that, as they varied thealpha-toxin concentration from 0.156 to 1.56 puM, LTB4secretion increased from 1 to 30 ng per 2 x 107 PMN.Although direct comparisons are difficult because of thedifferences between our system and theirs (including thesource of PMN), the LTB4 secretion levels they reportedwere similar to ours; corrected for cell number, they re-ported 50 to 1,500 pg of LTB4 secreted per 1 x 106 PMN. Incontrast to our study, they proposed that alpha-toxin stim-ulated LTB4 secretion by acting as a calcium channel,activating calmodulin-dependent PMN phospholipases. Inthe second study, Scheffer et al. (14) incubated humanperipheral PMN with E. coli strains that produced variousamounts of soluble hemolysin. They were able to use verylow bacterial multiplicities because each bacterium secretedmultiple hemolysin molecules; the actual ratio of hemolysinmolecules to PMN was not determined. They reported thatthe ability of E. coli to stimulate LTB4 secretion was relatedto the amount of hemolysin present, although it was alsoaffected by other E. coli products. Depending on the strainand the incubation conditions, E. coli hemolytic variantsstimulated PMN to secrete between 1 and 18.6 ng of LTB4per 2 x 107 PMN; converting this as above to reflect thenumbers ofPMN we used, this was 50 to 930 pg of LTB4 per1 x 106 PMN. Thus, the rickettsial effects on PMN secretionwere roughly similar in extent to the effects of staphylococ-cal alpha-toxin and hemolytic E. coli. However, the mech-anisms by which the individual agents exerted their actionappeared to be dissimilar.Wisseman and co-workers (7, 32, 33) reported that anti-

INFECT. IMMUN.

on January 29, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

RICKETTSIAL EFFECTS ON PMN 355

body plays an important role in destruction of typhus rick-ettsiae by macrophages and is opsonizing for PMN. We havealso reported the opsonizing effect of specific antibody forPMN both in our earlier study (22) and in this communica-tion. In our earlier study we reported that this opsonizingantibody did not ameliorate the toxic effect of rickettsiae onrabbit PMN. This observation is now extended by the resultsof the present study, which demonstrate that hyperimmuneantirickettsial antiserum does not inhibit the effect of rick-ettsiae on PMN LTB4 secretion, even though antirickettsialantiserum blocked rickettsial hemolysis. It does not seemlikely that the inability of antiserum to block rickettsialeffects on PMN was simply because the antibody waspresent at too low a concentration to be effective, becausewhen antibody and rickettsiae were incubated with erythro-cytes at the same ratios as those used in these experiments,hemolysis was blocked. Rather, it may point to a fundamen-tal difference in the means by which the two host cells arerecognized by rickettsiae. Although Ramm and Winkler (13)reported that the binding site for rickettsiae on sheep eryth-rocytes could be mimicked by cholesterol-palmitate-contain-ing vesicles and that rickettsial binding to erythrocyte ghostswas inhibited by amphotericin B, we have been unable todemonstrate a similar cell recognition specificity pattern byrickettsiae for other host cells. It may be that antirickettsialantibody blocks hemolysis by binding to an epitope whichrecognizes lipid-containing membrane sites on host cells.This antibody should be opsonic as a result of increasedimmune adherence, but may fail to halt rickettsial damage toPMN if the antibody does not bind directly to either the cellrecognition epitope on the rickettsia or the putative rickett-sial phospholipase. Alternatively, antirickettsial antibodymay fail to directly block rickettsial adherence to any cell.Winkler (27) reported that when antirickettsial antiserumwas added to suspensions of rickettsiae and erythrocytes,hemolysis was blocked by >90%; phase-contrast micros-copy revealed that the rickettsiae were agglutinated by theantibody. However, when antibody was added to preformedrickettsia-erythrocyte complexes stabilized by NaF, andthen the NaF was neutralized, the rickettsia-erythrocytecomplexes lysed normally. Nevertheless, rickettsiae re-leased from the complexes were unable to readsorb toerythrocytes, and subsequent hemolytic rounds wereblocked. This may indicate that agglutination is the keyantihemolytic mechanism of antirickettsial antibody and thatwhen such antibody opsonizes rickettsiae, at least somerickettsiae are brought into proper juxtaposition with thePMN membrane to allow further rickettsial damage to thePMN. Therefore, the difference between the effects ofantibody on rickettsial interactions with erythrocytes andPMN may be the lack of Fc receptors on erythrocytes.Further studies are needed to further describe the effects ofantibody on rickettsial interactions with host cells.Does this phenomenon of rickettsial stimulation of PMN

LTB4 secretion have physiologic significance? Certainly themultiplicities of rickettsiae used were far beyond thoseexpected to occur in vivo. However, our earlier studies (22)suggested that the number of rickettsiae being phagocytizedat these high multiplicities may not be much larger thanthose observed at much lower MOI. LTB4 is a potentautacoid and has been implicated as a chemotactic signal forPMN, causing PMN adherence to the microvascular endo-thelium (10). It is a powerful agent in promoting increasedvascular permeability (8, 10, 24); its presence in increasedamounts could result in systemic effects. The relativelymodest PMN response in patients suffering from acute

typhus (36) may, however, preclude assigning PMN LTB4secretion a major role in the pathogenesis of the acute phaseof typhus. This phenomenon may, however, serve as amodel for rickettsial interactions with macrophages, a hostcell of great significance during rickettsial infections. Furtherstudies are needed to determine whether rickettsiae alsostimulate macrophages to secrete arachidonate-derived au-tacoids. Finally, studies are needed to assess circulatinglevels of LTB4 and other arachidonate-derived autacoids inpatients suffering from acute typhus.

ACKNOWLEDGMENTS

This work was supported by Public Health Service grant Al 23115from the National Institute of Allergy and Infectious Diseases.We thank Eva Lantz for drawing the artwork.

REFERENCES1. Baron, E. J., and R. A. Proctor. 1982. Elicitation of peritoneal

polymorphonuclear leukocytes from mice. J. Immunol. Meth-ods 49:305-313.

2. Bovarnick, M. R., J. C. Miller, and J. C. Snyder. 1950. Theinfluence of certain salts, amino acids, sugars, and proteins onthe stability of rickettsiae. J. Bacteriol. 59:509-522.

3. Bovarnick, M. R., and J. C. Snyder. 1949. Respiration of typhusrickettsiae. J. Exp. Med. 89:561-565.

4. Brotherton, A. F. A., and J. C. Hoak. 1982. Role of Ca2" andcyclic AMP in the regulation of the production of prostacyclinby the vascular endothelium. Proc. Natl. Acad. Sci. USA79:495-499.

5. Feinstein, M. B., and R. I. Sha'afl. 1983. Role of calcium inarachidonic acid metabolism and in the actions of arachidonicacid-derived metabolites, p. 337-376. In W. Y. Cheung (ed.),Calcium and cell function. Academic Press, Inc., Orlando, Fla.

6. Ford-Hutchinson, A. W., M. A. Bray, M. V. Doig, M. E. Shipley,and M. J. H. Smith. 1980. Leukotriene B, a potent chemotacticand aggregating substance released from polymorphonuclearleukocytes. Nature (London) 286:264-265.

7. Gambrill, M. R., and C. L. Wisseman, Jr. 1973. Mechanisms ofimmunity in typhus infections. II. Multiplication of typhusrickettsiae in human macrophage cell cultures in the nonimmunesystem: influence of virulence of rickettsial strains and ofchloramphenicol. Infect. Immun. 8:519-527.

8. Malik, A. B., M. B. Perlman, J. A. Cooper, T. Noonan, and R.Bizios. 1985. Pulmonary microvascular effects of arachidonicacid metabolites and their role in lung vascular injury. Fed.Proc. 44:36-42.

9. Martin, S. P., and R. Green. 1958. Methods for the study ofsurviving leukocytes. A. Preparation of cell suspension. Meth-ods Med. Res. 7:136-138.

10. McMillan, R. M., and S. J. Foster. 1988. Leukotriene B4 andinflammatory disease. Agents Actions 24:114-119.

11. Moncada, S., and J. R. Vane. 1979. Pharmacology and endoge-nous roles of prostaglandin endoperoxides, thromboxane A2,and prostacyclin. Pharmacol. Rev. 30:293-331.

12. Popov, V. L., S. V. Prozorovsky, 0. A. Vovk, N. K. Kekcheeva,N. S. Smirnova, and 0. I. Barkhatova. 1987. Electron micro-scopic analysis of in vitro interaction of Rickettsia prowazekiiwith guinea pig macrophages. II. Macrophages from immuneanimals. Acta Virol. 31:59-64.

13. Ramm, L. E., and H. H. Winkler. 1976. Identification ofcholesterol in the receptor site for rickettsiae on sheep erythro-cyte membranes. Infect. Immun. 13:120-126.

14. Scheffer, J., W. Konig, J. Hacker, and W. Goebel. 1985. Bacte-rial adherence and hemolysin production from Escherichia coliinduces histamine and leukotriene release from various cells.Infect. Immun. 50:271-278.

15. Silberman, R., and P. Fiset. 1968. Method for counting rickett-siae and chlamydiae in purified suspensions. J. Bacteriol. 95:259-261.

16. Snyder, J. C., M. R. Bovarnick, J. C. Miller, and R. S. Chang.1954. Observations on the hemolytic properties of typhus rick-

VOL. 59, 1991

on January 29, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

356 WALKER AND HOOVER

ettsiae. J. Bacteriol. 67:724-730.17. Suttorp, N., W. Seeger, J. Zucker-Reimann, L. Roka, and S.

Bhakdi. 1987. Mechanism of leukotriene generation in polymor-phonuclear leukocytes by staphylococcal alpha-toxin. Infect.Immun. 55:104-110.

18. Volwerk, J. J., W. A. Pieterson, and G. W. deHaas. 1974.Histidine at the active site of phospholipase A2. Biochemistry13:1446-1454.

19. Walker, T. S. 1984. Rickettsial interactions with human endo-thelial cells in vitro: adherence and entry. Infect. Immun.44:205-210.

20. Walker, T. S., and H. H. Winkler. 1978. Penetration of culturedmouse fibroblasts (L cells) by Rickettsia prowazekii. Infect.Immun. 22:200-208.

21. Walker, T. S., and H. H. Winkler. 1979. Rickettsial hemolysis:rapid method for enumeration of metabolically active typhusrickettsiae. J. Clin. Microbiol. 9:645-647.

22. Walker, T. S., and H. H. Winkler. 1981. Interactions betweenRickettsia prowazekii and rabbit polymorphonuclear leuko-cytes: rickettsiacidal and leukotoxic activities. Infect. Immun.31:289-296.

23. Weiss, B., W. Prozialeck, M. Cimino, M. S. Barnette, and T. L.Wallace. 1980. Pharmacological regulation of calmodulin. Ann.N.Y. Acad. Sci. 356:319-345.

24. Williams, T. J., P. J. Jose, C. V. Wedmore, M. J. Peck, andM. J. Forrest. 1983. Mechanisms underlying inflammatory ede-ma: the importance of synergism between prostaglandins, leu-kotrienes, and complement-derived peptides. Adv. Prostaglan-din Thromboxane Leukotriene Res. 11:33-37.

25. Winkler, H. H. 1974. Inhibitory and restorative effects ofadenine nucleotides on rickettsial adsorption and hemolysis.Infect. Immun. 9:119-126.

26. Winkler, H. H. 1976. Rickettsial permeability. An ADP-ATPtransport system. J. Biol. Chem. 251:389-3%.

27. Winkler, H. H. 1977. Rickettsial hemolysis: adsorption, desorp-

tion, readsorption, and hemagglutination. Infect. Immun. 17:607-612.

28. Winkler, H. H., and R. M. Daugherty. 1989. Phospholipase Aactivity associated with the growth of Rickettsia prowazekii inL929 cells. Infect. Immun. 57:36-40.

29. Winkler, H. H., and E. T. Miller. 1980. Phospholipase A activityin the hemolysis of sheep and human erythrocytes by Rickettsiaprowazekii. Infect. Immun. 29:316-321.

30. Winkler, H. H., and E. T. Miller. 1982. Phospholipase A and theinteraction of Rickettsia prowazekii and mouse fibroblasts (L-929 cells). Infect. Immun. 38:109-113.

31. Winkler, H. H., and E. T. Miller. 1984. Activated complex ofL-cells and Rickettsia prowazekii with N-ethylmaleimide-insen-sitive phospholipase A. Infect. Immun. 45:577-581.

32. Wisseman, C. L., Jr. 1961. Interaction of rickettsiae and phago-cytic host cells. II. Chemotactic action of typhus rickettsiae onhuman polymorphonuclear leukocytes in vitro. J. Immunol.87:468-471.

33. Wisseman, C. L., Jr., J. Glazier, and M. J. Grieves. 1959.Interaction of rickettsiae and phagocytic host cells. I. In vitrostudies of phagocytosis and opsonization of typhus rickettsiae.Arch. Inst. Pasteur Tunis 36:339-360.

34. Wisseman, C. L., Jr., E. B. Jackson, F. E. Hahn, A. C. Ley, andJ. E. Smadel. 1951. Metabolic studies of rickettsiae. I. Theeffects of antimicrobial substances and enzyme inhibitors on theoxidation of glutamate by purified rickettsiae. J. Immunol.67:123-126.

35. Wisseman, C. L., Jr., and H. Tabor. 1964. Interaction ofrickettsiae and phagocytic host cells. IV. Early cellular re-sponse of man to typhus rickettsiae as revealed by the skinwindow technique with observations on in vivo phagocytosis. J.Immunol. 93:816-825.

36. Wolbach, S. B., J. L. Todd, and F. W. Palfrey. 1922. Theetiology and pathology of typhus. The League of Red CrossSocieties. Harvard University Press, Cambridge, Mass.

INFECT. IMMUN.

on January 29, 2020 by guesthttp://iai.asm

.org/D

ownloaded from