Vol. 42, No. 1INFECTION AND IMMUNITY, OCt. 1983, p. 341-3490019-9567/83/100341-09$02.00/0Copyright C) 1983, American Society for Microbiology

Effect of Immune Serum on Infectivity of RickettsiatsutsugamushiBARBARA A. HANSON

Department of Microbiology, University of Maryland School of Medicine, Baltimore, Maryland 21201

Received 3 May 1983/Accepted 22 July 1983

Hyperimmune antirickettsial serum was shown to prevent the attachment/penetration stage of Rickettsia tsutsugamushi infection of suspended chickencells. The extent of the inhibition depended on the serum concentration but not onthe presence of complement. The neutralizing activity was reduced by prioradsorption of immune serum with staphylococcal protein A or with intactrickettsiae but was not affected by adsorption with target cells. In the neutraliza-tion tests, there was no cross-reactivity between the Karp and Gilliam strains ofR. tsutsugamushi. Incubation of rickettsiae with immune serum did not alter theircapacity to metabolize glutamate nor grossly damage the permeability barrierfunction of their cytoplasmic membranes. Although the assay method had thecapacity to detect some aggregated infectious organisms, none were found inimmune serum-treated suspensions. It was concluded that immune serum may

inhibit rickettsial infection by blocking a surface component(s) whose function isnecessary for attachment to and/or penetration of target cells.

Rickettsia tsutsugamushi, the etiologicalagent of scrub typhus, persists in recoveredmammalian hosts after overt signs of the diseasehave disappeared (8, 13, 21, 23, 24, 28). Theimportance of circulating antibody in host immu-nity during either the acute or persistent phaseof scrub typhus is unknown. Protection of miceagainst death from acute infection has beenaccomplished experimentally by the transfer ofimmune serum (3, 4, 9, 17, 19, 20, 25) and by Tlymphocytes (6, 17, 20). The in vivo neutralizingcapacity of immune serum from convalescing orhyperimmunized animals has been utilized in theanalysis of R. tsutsugamushi antigenic variants(4, 9). A series of such studies has demonstratedthat serum protection of mice from an otherwiselethal dose of scrub typhus rickettsiae is quitespecific, and it has been possible to defineseveral different strains of R. tsutsugamushi bytesting the extent of cross-reactive protection byimmune sera.By the same token, a strain-specific neutraliz-

ing activity in hyperimmune or convalescentimmune sera also has been shown in infectedcell cultures (1, 16). The manifestations of anti-serum-altered infection analyzed in vitro werethe diminution of rickettsia-induced cytopathiceffects in BS-C-1 cells (1) and plaque reductionin irradiated L cells (16), both ofwhich representthe culmination of several cycles of rickettsialreplication resulting in cytopathology. In a pre-ceding study, Cohn et al. (7) were unable todetect antiserum-mediated inhibition of an early

event in rickettsial infection, namely, the initialattachment/penetration of R. tsutsugamushi intosuspended MB III cells.Important as these investigations have been,

they shed little light on the mechanism of theneutralization or in fact on which stage of theinfection cycle is inhibited. The present studyaddresses these points. The results indicate thatimmune serum prevents the entry of R. tsutsu-gamushi into cultured cells by a means otherthan direct killing of the organisms.

MATERIALS AND METHODSRickettsia. R. tsutsugamushi Gilliam was used at the

143rd egg passage level. The plaque-purified Karpstrain was obtained from J. Osterman, Walter ReedArmy Institute of Research, Washington, D.C., andprepared as 20% (wt/vol) suspensions in brain heartinfusion broth. Plaque assays of infectivity were donein a chicken embryo cell monolayer culture (with 2.5%fetal calf serum in the agarose overlay) by the modified(14) method of Wike et al. (29), except that agarosefeeder layers were added at days 7 and 14, and plaqueswere counted on day 19. Rickettsial bodies (RLB)were counted by the method of Silverman et al. (22).The rickettsial seed suspension used contained anRLB to PFU ratio of 28:1.

Renografin treatment of rickettsiae. When rickettsiaehad to be centrifuged before mixing with target cells, itwas necessary first to remove yolk sac material whichotherwise would cosediment with the organisms incytotoxic quantities. To achieve a rapid, partial purifi-cation of the yolk sac rickettsial suspension, theorganisms were pelleted through Renografin (E. R.Squibb & Sons, Princeton, N.J.) cushions, which

341

on May 4, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

342 HANSON

retarded the sedimentation of most of the yolk materi-al. To accomplish this, the infected yolk sac suspen-sions were diluted two- to fourfold in cold growthmedium (see below), and 1 ml was layered over 0.5 mlof 19% Renografin (wt/vol in medium) in a 1.5-mlcentrifuge tube. After 2 min of centrifugation in amicrofuge at 13,000 x g, the supernatant fluid wasremoved, and the wall of the tube was wiped cleanwith a sterile swab. The rickettsial pellet was resus-pended in a half volume of medium, and the procedurewas repeated. This was followed by an additional washand resuspension in the desired volume of medium.Despite the obvious removal of great quantities of yolkmaterial by this procedure, microscopic examinationrevealed significant residual host contamination. Therecovery of infectious rickettsiae varied from 40 to100%, as determined by the initial and final capacitiesof the suspensions to infect suspended chicken cells(see below).

Sera. Hyperimmune sera, prepared in rabbits by 8 to12 injections of rickettsiae-infected mouse livers andspleens by W. T. Walsh, have been described in earlierstudies (11). Control rabbit sera were taken from thesame rabbits just before the first immunization. Rabbitsera were heat inactivated (56°C for 30 min) and filtersterilized before use. Guinea pig serum (GPS) (BBLMicrobiology Systems, Cockeysville, Md.), used as asource of complement, contained approximately 500hemolytic units/ml before dilution.For serum absorption, sera diluted 1/5 in culture

medium were incubated with equal volumes of Reno-grafin-treated rickettsiae, trypsinized chicken embryocells (CEC), or washed, fixed Staphylococcus aureus(Pansorbin; Calbiochem; La Jolla, Calif.) for 30 min atroom temperature. The sera were then separated fromthe cells by centrifugation (13,000 x g for 2 min) andreexposed in the same manner to second aliquots ofthe same absorbants. The cell-free sera were stored forless than 2 weeks at -20°C before use.

Infection of cultured cells. Rickettsial infectivity wasdetermined by microscopic examination of their ca-pacity to associate with target cells. CEC were tryp-sinized from primary monolayer cultures, washed,suspended in growth medium (Dulbecco modified Ea-gle medium [GIBCO, Grand Island, N.Y.] with half-strength vitamins and amino acids and supplementedwith 10% fetal calf serum and 15 mM HEPES buffer[N-2-hydroxyethylpiperazine-N'-2-ethanesulfonicacid] [pH 7.3]), mixed with an equal volume of rickett-siae, and routinely incubated at 34°C for 60 min withfrequent agitation. At the end of this period, the cellswere washed by centrifugation and distributed to slidechambers (Lab-Tek Products, Naperville, Ill.), wherethey were allowed to settle for 3 to 5 h at 34°C. Thistime period was sufficient to allow attachment andsome spreading of the cells on the glass surface butwas not long enough to reflect significant increases inrickettsial numbers due to growth. Therefore, the 3- to5-h incubation only permitted measurement of attach-ment/penetration of the organisms and not their repli-cation. The slides then were fixed and stained withGiemsa for counting the number of organisms associ-ated with 100 cells in each of two duplicate cultures.For rickettsial growth curves, cultures were infectedsimilarly, but slides were examined at 24-h intervalsafter the initial infection. The results are presented asaverage number of rickettsiae per cell (R/c) and per-

cent cells infected (Pi). The R/c values were highlyreproducible; analysis by a paired t test showed thatcounts from duplicate cultures did not differ signifi-cantly from each other (P = 0.3; N = 102). In all cases,the reduction in R/c was accompanied by a similarreduction in Pi. All straight lines presented in thegraphs were arrived at by regression analysis.Serum neutralization. Rickettsiae in growth medium

were incubated with various concentrations of serumat room temperature for 30 min (pretreatment) beforebeing added to an equal volume of suspended CEC toassess their infectivity as described above. The serumconcentrations referred to in the text are those concen-trations which existed during the incubation of treatedrickettsiae with cells and thus represent a twofolddilution of that concentration present during the 30-min preincubation period.Glutamate oxidation. This metabolic test was per-

formed with Renografin-treated rickettsiae as de-scribed previously (12), except that 1.5 ml of rickettsi-al suspension was placed in each 10-ml flask. L-[U-14C]glutamic acid (290 mCi/mmol) was obtained fromAmersham Searle (Des Plaines, Ill.) and used at a finalconcentration of 3.4 x 10-6 M.

Electron microscopy. For electron microscopic ex-amination, Renografin-treated rickettsial suspensionsin the desired diluent were mixed with an equalvolume of double-strength fixative and processed aspreviously described (12).

RESULTSEffect of immune serum on infection of CEC.

The first step in the rickettsial replication cycleis entry into host cells. The means by which thisis accomplished in cells other than professionalphagocytes is not entirely clear, but an inducedphagocytosis has been suggested (27). Rickettsi-al metabolic capacity as well as target cell viabil-ity are required (7, 27). To determine the effectof rabbit immune serum on this early stage ofrickettsial infection, R. tsutsugamushi Karp wasincubated in growth medium with various serumadditions for 30 min at 24°C, and this mixturewas then added to an equal volume of suspendedCEC. The final concentrations in the infectionmixture were 4 x 106 PFU/ml, 107 CEC/ml, 5%rabbit serum, 5% GPS, and 5% fetal calf serum.The mixtures were incubated at 34 or 0°C, andportions were removed at brief intervals formicroscopic enumeration of CEC-associatedrickettsiae.As shown in Fig. 1A, the addition of nonim-

mune rabbit serum (NRS) and either fresh orheated GPS had no effect on the ability ofrickettsiae to associate with CEC. In contrast,the immune serum had a marked inhibitoryeffect on the infectivity of the rickettsiae, asevidenced by a lower initial rate of uptake aswell as a fourfold lower final level of infection.The slight apparent difference between the twogroups of immune serum-treated rickettsiae wasnot due to the heating of the GPS, because this

INFECT. IMMUN.

on May 4, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

NEUTRALIZATION OF SCRUB TYPHUS RICKETTSIAE 343

_, 2.0wC<)

w49a) 1.5

wI-I- -

1.0

A. 34°C 0A

B0

0.51

B. 0°C

90 15 30 60 90INCUIATION TIME (nminutes)

-J

w

w

I--wC.)

50

10

5

.1

C.

0

0

0

A

72

HOURS AFTER INFECTIONFIG. 1. Effect of serum additions on infection of CEC by R. tsutsugamushi Karp. Pretreated rickettsiae (0.4

PFU/cell) were mixed with suspended CEC at 34°C (A) or 0°C (B) in culture medium with no additions (O), 5%NRS + 5% heated GPS (A), 5% NRS + 5% fresh GPS (0), 5% anti-Karp serum + 5% heated GPS (A), or 5%anti-Karp serum + 5% fresh GPS (0). Regression lines in (A) and (B) were calculated from the pooled data forthe 0- to 60-min time period. (C) After 90 min, the infected cells were cultured at 34°C for determination ofrickettsial growth capacity.

difference was not seen in subsequent, similarexperiments. Equivalent results were obtainedin a parallel incubation at 0°C; although therewas a very slight increase in cell-associatedrickettsiae at this temperature in the controlmixtures, none was seen when immune serumwas present (Fig. 1B).

After the 90-min incubation of the rickettsiaewith suspended CEC at 34°C, the mixtures werediluted 100-fold in growth medium containing noserum other than 10% fetal calf serum, distribut-ed to chamber slides, and incubated at 34°C forobservation of rickettsial replication. The resultsshowed that, although the immune serum had a

VOL. 42, 1983

on May 4, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

DO - As a means of further characterizing the rick-ettsial neutralization, the effect of serum dilu-tions on the level of infection after a 60-minincubation of serum-treated rickettsiae with sus-

DO - :\",;¢ pended CEC was determined. A preliminaryexperiment showed that incubation in NRS inconcentrations ranging from 0.5 to 5% had noeffect on rickettsial uptake (not shown); there-

50_-\ fore, the control mixtures in subsequent testscontained only one NRS concentration. The

NAsvexperiment presented in Fig. 2 demonstratedN\" that the inhibition of rickettsial infection was

dependent on the antiserum concentration (theN" 50% inhibitory dose was about 0.5% for thisA serum) and confirmed that in this system, guinea

pig complement does not affect neutralization bythe immune serum.

Specificity of neutralization. Neutralization10 .05 0.15 0.5 1.5 5 tests with specifically absorbed immune or con-

trol serum indicated that the immune serum-IMMUNE SERUM CONCENTRATION (%) mediated inhibition of infection was due to a

J. 2. Effect of antiserum concentrations and rickettsia-antibody interaction (Table 1). Thus,lement on attachment/penetration of suspended neutralizing activity was removed by prior incu-:y R. tsutsugamushi Karp. Pretreated rickettsiae bation with Pansorbin (anti-immunoglobulin G)FU/cell) were mixed for 60 min with CEC in or a substantial quantity of Renografin-treatede medium containing various amounts of im- Karp, but not by prior exposure of the serum toserum and 5% heated GPS (A) or 5% fresh GPS CEC Moreover, dose-response experiments-ontrol (5% NRS + 5% heated or fresh GPS) demonstrated that the immune serum-mediatedions resulted in R/c values of 1.26 and 1.11, inhibition of R. tsutsugamushi infection was;tively.

strain specific. Rickettsiae of the Karp strainng inhibitory effect on the initial infection, were not at all neutralized by concentrations ofrickettsiae which were able to enter CEC anti-Gilliam which greatly inhibited Gilliam up-presence of anti-Karp serum replicated at take (Fig. 3A), and doses of anti-Karp whichame rate as the NRS-treated organisms were highly effective against the homologous1C). Therefore, this experiment demon- strain did not alter Gilliam infection of CEC-d that immune rabbit serum inhibited the (Fig. 3B).of rickettsiae into CEC but not the subse- Mechanism of antiserum-mediated neutraliza-replication of the few remaining infectious tion. Two types of experiments were consistentisms, and furthermore, that, under the with the idea that rickettsial killing was not aitions described, guinea pig complement necessary contributing factor for the neutraliza-IO additional inhibitory effect. tion of R. tsutsugamushi. First, it was demon-

IC

FIGcomplCEC t(1.8 P

culturmune

(0). (infectirespec

strikiithosein thethe s(Fig.strateentryquentorgancondihad n

TABLE 1. Effect of prior absorption on neutralizing capacity of anti-Karp serum'

Absorbant Rickettsiae/cell (avg)bExpt Absorbant per ml of

serum NRS Anti-Karp % Neutralization

1 None 3.28 0.62 81.81 Karp 108 RLB 3.81 0.67 82.41 CEC 107 cells 3.29 0.45 86.31 Protein A 0.3 mlc 2.51 1 37de 45.42 None 2.00 0.26 87.02 Karp 109 RLB 2.57 1.65d 27.9

a In experiment 1, infecting rickettsiae were an untreated yolk sac suspension, 5 x 106 PFU/ml (1.4 x 108RLB/ml). In experiment 2, infecting rickettsiae were a Renografin-treated yolk sac suspension, approximately 3x 106 PFU/ml (8.4 x 107 RLB/ml).

b In each infection, the CEC concentration was 2 x 106/ml, and the final rabbit serum concentration was 2.5%.c Packed cell volume of Pansorbin.d Compared to unabsorbed anti-Karp, P < 0.01.e Not significantly different from NRS control.

0-J

-

w0

-J

"I-.w5:?

344 HANSON INFECT. IMMUN.

2c

on May 4, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

NEUTRALIZATION OF SCRUB TYPHUS RICKETTSIAE

A. ANTI-GILLIAM + GILLIAM (A,,A)KARP (0,0)

3----0

A

A%a"s~~\

B. ANTI - KARP + KARPM()

0.15 0.5 1.5 5 0.15 0.5 1.5 5

IMMUNE SERUM CONCENTRATION (%)

FIG. 3. Strain specificity of neutralization: effect of anti-Gilliam (A) and anti-Karp (B) sera on attachment/penetration of suspended CEC by R. tsutsugamushi Karp (circles) or Gilliam (triangles). Pretreated rickettsiae (2to 3 PFU/cell) were mixed for 60 min with CEC in culture medium containing various amounts of antisera. (A)Two separate experiments (open and closed symbols) with different antisera. Control (5% NRS) infectionsresulted in R/c values of 0.83 (A; Karp), 1.94 (A; Gilliam), 4.05 (B; Karp), and 2.62 (B; Gilliam).

strated that preincubation of rickettsiae withimmune serum did not subsequently alter theircapacity to metabolize [14C]glutamate (Table 2).Therefore, an inhibition of this function, thecapacity for which apparently is necessary forR. tsutsugamushi penetration of target cells (7),could not be invoked to explain the neutraliza-tion. Second, the permeability barrier functionof the cytoplasmic membranes of antibody-treat-ed rickettsiae remained intact, at least insofar astheir ability to exclude sucrose was concerned.This was shown by suspending immune serum(10%)-pretreated organisms in hypertonic (30%[wt/vol]) sucrose and fixing them 10 min later forelectron microscopic examination. Controlsconsisted of pretreatment with NRS and anisosmotic sucrose-phosphate glutamate (5) solu-tion. The capacity of organisms to exclude su-crose would be manifested by their plasmolysis,whereas if sucrose were able to diffuse freelyinto the cytoplasm of (damaged) organisms, nodifferential osmotic pressure would occur, andno plasmolysis would be seen (12). This experi-ment clearly showed that antiserum-treated rick-ettsiae were as readily plasmolysed in hyperton-ic sucrose as were the NRS-treated controls(Fig. 4), thus demonstrating that antibody doesnot grossly damage this rickettsial cytoplasmicmembrane function. Figure 4 also comparesNRS- and antibody-treated Karp rickettsiaewhen they were fixed in isosmotic sucrose-

phosphate glutamate. No effects of antiserum onrickettsial ultrastructure were apparent.To assess the possibility that the rickettsial

neutralization is due to clumping of the organ-isms, the distribution of serum-treated rickettsi-ae among the infected cells was analyzed. Un-like the plaque assay, in which the number ofplaques counted may merely represent the num-ber of infectious clumps present in the inoculum,our uptake assay examines the infecting capaci-ty of individual organisms. By this method, theinfection of a particular cell by more than oneviable rickettsia is not counted as a single event,

TABLE 2. Effect of antiserum on glutamateoxidation by R. tsutsugamushi Karp

Serum' cpm/109 RLBb

NRS ..... 10, 783 ± 2,682cAnti-Karp .................. 11,218 ± 1,215

a Renografin-treated rickettsiae were preincubatedfor 30 min at 24°C in growth medium with 25% NRS oranti-Karp and then diluted in glutamate oxidationbuffer (12) to result in 4.2% serum during metabolicassay.

b Counts per minute of [14C]glutamate converted toCO2 during 2 h of incubation at 32°C; counts perminute in Renografin-treated uninfected yolk sac con-trol suspensions were subtracted.

c Mean ± standard error of two separate experi-ments.

200

100

c 500

0

01w

4

I-I- 10

C.)i

5

VOL. 42, 1983 345

on May 4, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

346 HANSON

*Wt'b1

d

lFIG. 4. Effect of immune serum on ultrastructure of serum-treated R. tslutsijgainiishli Karp fixed in sucrose-

phosphate glutamate (a and b) or hypertonic sucrose (c and d). (a and c) 10% NRS; (b and d) 10% anti-Karp. Noteplasmolysis in both (c) and (d) compared to (a) and (b) controls.

but rather each cell-associated rickettsia is enu-

merated singly. Thus, up to a certain limit,aggregation of infectious organisms should notresult in a decrease in the calculated averagenumber of rickettsiae per cell. To illustrate this,Fig. 5A presents a histogram showing the distri-bution of Renografin-treated rickettsiae in cellsinfected after 60 min of incubation at 34°C, whenthe rickettsiae used as the inocula were firstcentrifuged and resuspended with either theusual vigorous pipetting to break up any clumps(controls) or very gentle pipetting to leaveclumps of organisms in the suspension. It can beseen that the distribution of control rickettsiae ininfected cells is strongly shifted to the left, withonly 2% of the infected cells containing more

than four organisms. This pattern is typical ofthis level of infection (R/c, 0.57). The effect ofrickettsial clumping can be seen by the altereddistribution of organisms in the second infectedcell culture, which was not as strongly clusteredand in which 21% of the infected cells containedmore than four organisms. It also should benoted that infection with the suspension of par-tially clumped organisms did not result in a

lower average number of rickettsiae per cell

(0.82 compared to 0.57) and in fact may haveenhanced it. Other standard uptake experi-ments, using higher infecting doses, occasional-ly have revealed up to 23 rickettsiae in a singlecell, presumably at least partially an effect ofrickettsial clumping.On the other hand, in no case was this phe-

nomenon observed when antibody-inhibitedrickettsiae were incubated with CEC. Figure 5Bcompares typical distributions of NRS- and anti-serum-treated Karp after 60 min of incubationwith CEC. With a higher infecting dose thanused in the previous experiment (Fig. 5A), thedistribution of control rickettsiae was also differ-ent, the mode being two rickettsiae per cell and35% of the cells containing four or more rickett-siae. In contrast, in the presence of 5% anti-Karp serum, no cells contained more than threerickettsiae, giving no indication of clumping.When the infection was done in 1.5% antiserum,the distribution of rickettsiae fell between that ofthe 5% NRS and 5% antiserum, with 13% of thecells containing more than four organisms.Again, the distribution did not reflect the occur-rence of any rickettsial aggregation. As theconcentration of immune serum was lowered,

INFECT. IMMUN.

jrJ

I

..,.AdL.j

on May 4, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

NEUTRALIZATION OF SCRUB TYPHUS RICKETTSIAE 347

A.

C,)-J-JCC

wI~-

LLz

L-0

9 10 >10RICKETTSIAE / CELL

80 r

70 F

Cl)-J

CM

-i

w

w

ILzUa.0

60 F

50 F

40 F

30 F

201-

01-

B.

1 2 3 4 5 6 7 8 9 10RICKETTSIAE / CELL

.l0

FIG. 5. Distribution of R. tsutsuigamiushi Karp among infected cells after 60 min of incubation with suspendedCEC. (A) Effect of degree of rickettsial clumping. Renografin-treated rickettsiae were thoroughly (E) orincompletely (U)) dispersed in growth medium before mixing with CEC at a ratio of about 1 PFU/cell. Thecontrol infection resulted in values of 0.57 R/c and 28.5 Pi. The infection with clumped rickettsiae resulted in 0.82R/c and 31.0 Pi. The distributions were significantly different (P < 0.005) in the chi-square test. (B) Effect ofantiserum. Pretreated rickettsiae (3.3 PFU/cell) were mixed for 60 min with CEC in growth medium containing5% NRS (0), 5% anti-Karp serum (U), or 1.5% anti-Karp serum (E1) to result in values of 3.04 R/c and 74 Pi, 0.19R/c and 14 Pi, and 0.66 R/c and 29 Pi, respectively. The distributions of rickettsiae treated with 5 and 1.5%immune sera significantly differed from the NRS controls (P < 0.005 and < 0.05, respectively).

the rickettsial distributions became more andmore similar to the controls (not shown). There-fore, the distribution patterns of antibody-treat-ed rickettsiae among infected cells could bepredicted solely on the basis of the number ofinfectious rickettsiae added.

DISCUSSIONThese experiments have shown that antirick-

ettsial antibody can inhibit R. tsutsugamushiinfection of CEC by preventing rickettsial entry

into the host cells. Those few organisms whichdo enter cells in the presence of immune serumsubsequently replicate at an undiminished rate.The neutralization of rickettsial uptake wasstrain specific. This finding is consistent withearlier reports that antiserum-mediated inhibi-tion of in vivo or in vitro pathological effects candistinguish R. tsutsugamushi strains (1, 3, 4, 9,16, 19).The effectiveness of rabbit hyperimmune sera

in preventing rickettsial entry into CEC wassubstantial. The neutralizing activity of seven

VOL. 42, 1983

K&Ma nm kgm rn kl%N 0

on May 4, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

348 HANSON

different sera diluted to 5% ranged from 68 to97% (mean, 86.5%). Our results are in contrastto an earlier report by Cohn et al. (7), in which itwas stated that convalescent human, monkey,or rabbit serum had no effect on the penetrationof MB III lymphoblasts by R. tsutsugamushiKarp. This discrepancy could be due to differ-ences in either the sera or the cells used in thetwo studies; the MB III cells were not activelyphagocytic, but their lymphoid origin could con-fer properties not possessed by fibroblasts. Sim-ilarly, conflicting results have been reported onthe effect of antibody on Rickettsia prowazekiiuptake by non-professionally phagocytic cells,wherein convalescent human serum was notinhibitory (32) but hyperimmune rabbit serumwas inhibitory (26). It is likely that the effect ofantiserum on rickettsial infection depends heavi-ly on the specificities and/or avidities of theantibodies in that serum. Significnt to this pointis the recent finding that different clones ofmouse monoclonal antibody directed against amajor surface protein of R. prowazekii had dif-ferent effects on rickettsial entry into CEC, tothe extent that some enhanced uptake and oneinhibited it (E. V. Oaks, Ph.D. thesis, Universi-ty of Maryland, Baltimore, 1982). The target celltype also must be considered when evaluatingantiserum effects on rickettsial infection (32):immune serum-mediated enhancement of uptake(opsonization) and preparation for intracellularkilling in macrophages have been found consis-tently with several species of Rickettsia, includ-ing R. tsutsugamushi (2, 10, 15; A. P. Andreseand C. L. Wisseman, Jr., Proc. Electron Micros-copy Soc. Am. 29:39-40, 1971; W. H. Meyer IIIand C. L. Wisseman, Jr., Abstr. Annu. Meet.Am. Soc. Microbiol., 1980, D10, p. 39). Like-wise, antiserum-mediated opsonization of rick-ettsial phagocytosis by polymorphonuclear cellsalso has been reported (18, 30, 31). One of thesestudies (18) showed that a lower proportion ofantibody-coated R. tsutsugamushi escaped fromphagosomes, but that, because of the opsoniza-tion, a higher absolute number of organisms wasfound free in the cytoplasm after antibody treat-ment.Complement did not augment the rickettsial

neutralization observed in our studies. Howev-er, complement probably would enhance theeffects of antibodies with lowered avidity bystabilizing the immune complexes formed (com-pare reference 1). Although complement maynot appear to play a role in some in vitro assays,it may still be very important in natural infec-tions, when antibody of lower avidity might beexpected, when multiple rounds of replicationoccur (so that a few rickettsiae which escape theinitial neutralization may still overwhelm thehost), and when antibody clearing mechanisms

of inhibition other than that described here maybe of equal or greater importance.

Consistent with the lack of necessity for com-plement in our inhibition studies was our failureto detect gross alterations in reckettsial metabol-ic and permeability barrier functions. It seemsmost likely that antibody blocks rickettsial up-take not by direct killing of the organisms but byaffecting the attachment/penetration process.The possibility that the neutralization of R.tsutsugamushi uptake is due solely to antibody-mediated aggregation appears unlikely. We haveshown that our uptake assay system detectsclumps but that the parameter (R/c) measuredwas not reduced in the face of some degree ofrickettsial aggregation. Furthermore, no indica-tion of rickettsial clumping was seen when anti-body-treated rickettsiae were mixed with hostcells. It follows that if rickettsial clumps wereformed in the presence of antibody, they werenot infectious, and this contrasts with the infec-tious aggregates obtained after incomplete dis-ruption of a rickettsial pellet. Therefore, im-mune serum must prevent rickettsial entry by ameans either other than, or in addition to, parti-cle aggregation.The above findings support the view that

antibody inhibits a specific event required in theattachment/penetration process. This might bevia a mechansim which sterically hinders oralters the structural configuration of a criticalrickettsial surface component, be it a membraneattachment site or an enzyme. Alternatively, theserum might reduce the organisms' capacity toenter cells by altering nonspecific attractionsbetween infectious agent and target cell. Thesemechanisms could be invoked for antibody-ag-glutinated organisms as well as for individualrickettsiae. It is difficult to narrow our specula-tion without additional clarification of the entryprocess, including the apparent penetration ofaggregated organisms. It is hoped that the infor-mation presented will provide a means of furtherstudying rickettsial attachment/penetration.

ACKNOWLEDGMENTS

This work was supported by Public Health Service grant Al-17743 from the National Institute for Allergy and InfectiousDiseases.

I thank Kathryn Sheridan for her excellent and devotedtechnical assistance and David Silverman for performing theelectron microscopy and preparing the micrographs.

LITERATURE CITED

1. Barker, L. F., J. K. Patt, and H. E. Hopps. 1%8. Thetitration and neutralization of Rickettsia tsutsugamushi intissue culture. J. Immunol. 100:825-830.

2. Beaman, L., and C. L. Wisseman, Jr. 1976. Mechanismsof immunity in typhus infections. VI. Differential opsoniz-ing and neutralizing action of human typhus rickettsia-specific cytophilic antibodies in cultures of human macro-phages. Infect. Immun. 14:1071-1076.

INFECT. IMMUN.

on May 4, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

NEUTRALIZATION OF SCRUB TYPHUS RICKETTSIAE 349

3. Bell, E. J., B. L. Bennett, and L. Whitman. 1946. Antigen-ic differences between strains of scrub typhus as demon-strated by cross-neutralization tests. Proc. Soc. Exp.Biol. Med. 62:134-137.

4. Bennett, B. L., J. E. Smadel, and R. L. Gauld. 1949.Studies on scrub typhus (tsutsugamushi disease). IV.Heterogeneity of strains of R. tsutsugamushi as demon-strated by cross-neutralization tests. J. Immunol. 62:453-461.

5. Bovarnick, M. R., J. C. Miller, and J. C. Snyder. 1950.The influence of certain salts, amino acids, sugars, andproteins on the stability of typhus rickettsiae. J. Bacteriol.59:509-522.

6. Catanzaro, P. J., A. Shirai, L. D. Agniel, Jr., and J. V.Osterman. 1977. Host defenses in experimental scrubtyphus: role of spleen and peritoneal exudate lymphocytesin cellular immunity. Infect. Immun. 18:118-123.

7. Cohn, Z. A., F. M. Bozeman, J. M. Campbell, J. W.Humphries, and T. K. Sawyer. 1959. Study on growth ofrickettsiae. V. Penetration of Rickettsia tsutsugamushiinto mammalian cells in vitro. J. Exp. Med. 109:271-292.

8. Fox, J. P. 1948. The long persistence of Rickettsia orienta-lis in the blood and tissues of infected animals. J. Im-munol. 59:109-114.

9. Fox, J. P. 1949. The neutralization technique in tsutsuga-mushi disease (scrub typhus) and the antigenic differentia-tion of rickettsial strains. J. Immunol. 62:341-352.

10. Gambrill, M. R., and C. L. Wisseman, Jr. 1973. Mecha-nisms of immunity in typhus infections. III. Influence ofhuman immune serum and complement on the fate ofRickettsia mooseri within human macrophages. Infect.Immun. 8:631-640.

11. Hanson, B., and C. L. Wisseman, Jr. 1981. Heterogeneityamong Rickettsia tsutsugamushi isolates: a protein analy-sis, p. 503-514. In W. Burgdorfer and R. L. Anacker(ed.), Rickettsiae and rickettsial diseases. AcademicPress, Inc., New York.

12. Hanson, B. A., C. L. Wisseman, Jr., A. Waddell, and D. J.Silverman. 1981. Some characteristics of heavy and lightbands of Rickettsia prowazekii on Renografin gradients.Infect. Immun. 34:596-604.

13. Kohis, G. M., C. A. Armbrust, E. A. Irons, and C. B.Philip. 1945. Studies on tsutsugamushi disease (scrubtyphus, mite-borne typhus) in New Guinea and adjacentislands: further observations on epidemiology and etiolo-gy. Am. J. Hyg. 41:374-396.

14. Murphy, J. R., C. L. Wisseman, Jr., and L. B. Snyder.1976. Plaque assay for Rickettsia mooseri in tissue sam-

ples. Proc. Soc. Exp. Biol. Med. 153:151-155.15. Nacy, C. A., and J. V. Osterman. 1979. Host defenses in

experimental scrub typhus: role of normal and activatedmacrophages. Infect. Immun. 26:744-750.

16. Oaks, S. C., Jr., F. M. Hetrick, and J. V. Osterman. 1980.A plaque reduction assay for studying antigenic relation-ships among strains of Rickettsia tsutsugamushi. Am. J.Trop. Med. Hyg. 29:998-1006.

17. Osterman, J. V., A. Shirai, and P. J. Catanzaro. 1978.Mechanisms of immunity in experimental scrub typhus, p.333-344. In J. Kazar, R. A. Ormsbee, and I. N. Taraser-

ich (ed.), Rickettsiae and rickettsial diseases. Veda, Brati-slava.

18. Rikihisa, Y., and S. Ito. 1983. Effect of antibody on entryof Rickettsia tsutsugamushi into polymorphonuclear leu-kocyte cytoplasm. Infect. Immun. 39:928-938.

19. Robinson, D. M., and D. L. Huxsoll. 1975. Protectionagainst scrub typhus infection engendered by the passivetransfer of immune sera. Southeast Asian J. Trop. Med.Public Health 6:477-482.

20. Shirai, A., P. J. Catanzaro, S. M. Phillips, and J. V.Osterman. 1976. Host defenses in experimental scrubtyphus: role of cellular immunity in heterologous protec-tion. Infect. Immun. 14:39-46.

21. Shirai, A., T. C. Chan, E. Gan, and D. L. Huxsoll. 1979.Persistence and reactivation of Rickettsia tsutsugamushiinfections in laboratory mice. Jpn. J. Med. Sci. Biol.32:179-184.

22. Silverman, D. J., P. Fiset, and C. L. Wisseman, Jr. 1979.Simple differential staining technique for enumeratingrickettsiae in yolk sac, tissue culture extracts, or purifiedsuspensions. J. Clin. Microbiol. 9:437-440.

23. Smadel, J. E., E. B. Jackson, and A. B. Cruise. 1949.Chloromycetin in experimental rickettsial infections. J.Immunol. 62:49-65.

24. Smadel, J. E., H. L. Ley, Jr., F. H. Diercks, and J. A. P.Cameron. 1952. Persistence of Rickettsia tsutsugamushiin tissues of patients recovering from scrub typhus. Am. J.Hyg. 56:294-302.

25. Topping, N. H. 1945. Tsutsugamushi disease (scrub ty-phus). The effects of an immune rabbit serum in experi-mentally infected mice. Public Health Rep. 60:1215-1220.

26. Turco, J., and H. H. Winkler. 1982. Differentiation be-tween virulent and avirulent strains of Rickettsia prowaze-kii by macrophage-like cell lines. Infect. Immun. 35:783-791.

27. Walker, T. S., and H. H. Winkler. 1978. Penetration ofcultured mouse fibroblasts (L cells) by Ric kettsiae proiva-

zekii. Infect. Immun. 22:200-208.28. Webster, W. J. 1940. Typhus in the Simla hills. IX.

Laboratory observations, chiefly on human XK strains.Indian J. Med. Res. 27:657-666.

29. Wike, D. A., G. Tallent, M. G. Peacock, and R. A.Ormsbee. 1972. Studies of the rickettsial plaque assaytechnique. Infect. Immun. 5:715-722.

30. Wisseman, C. L., Jr., J. R. Gauld, and J. G. Wood. 1963.Interaction of rickettsiae and phagocytic host cells. III.

Opsonizing antibodies in human subjects infected withvirulent or attenuated Rickettsia proivazekii or inoculatedwith killed epidemic typhus vaccine. J. Immunol. 90:127-131.

31. Wisseman, C. L., Jr., J. Glazier, and M. J. Grieves. 1959.Interaction of rickettsiae and phagocytic host cells. I. Invitro studies of phagocytosis and opsonization of typhusrickettsiae. Arch. Inst. Pasteur Tunis 36:339-360.

32. Wisseman, C. L., Jr., A. D. Waddell, and W. T. Walsh.1974. Mechanisms of immunity in typhus infections. IV.Failure of chicken embryo cells in culture to restrictgrowth of antibody-sensitized Ric kettsia proviwazekii. In-fect. Immun. 9:571-575.

VOL. 42, 1983

on May 4, 2021 by guest

http://iai.asm.org/

Dow

nloaded from