INFECTON AND IMMUNITY, Feb. 1977, p. 410-416Copyright © 1977 American Society for Microbiology

Vol. 15, No. 2Printed in U.S.A.

Reproduction ofthe Eosinopenia ofAcute Infection by PassiveTransfer of a Material Obtained from Inflammatory Exudate

DAVID A. BASS'Nuffield Department of Clinical Medicine, University of Oxford, Oxford, England

Received for publication 21 May 1976

Studies ofthe eosinopenic effect ofacute inflammation were conducted in micepreviously rendered eosinophilic with trichinosis. Exudate removed from apneumococcal abscess contained material (eosinopenic factor [EF]) capable ofcausing eosinopenia of 4- to 24-h duration when injected intraperitoneally intoeosinophilic mice. The material passed through a 0.45-.um filter, but was re-tained by a dialysis membrane capable of retaining protein molecules of greaterthan approximately 30,000 molecular weight. EF was soluble in 7% perchloricacid, was not destroyed by pneumococcal proteolytic enzymes in the presence ofTrasylol, but was inactivated by heating to 56TC for 30 min. EF was detectable inthe exudate after 10 h and had reached its highest concentration after 20 h.When the effect ofEF was expressed as a percent suppression of control eosino-phil levels, there was a geometric dose response. Eosinopenia could not beascribed to steroids present in the preparation, and the EF was effective inadrenalectomized animals. Eosinopenia was not induced by transfer of similarlytreated heat-killed pneumococci, pneumococcal culture filtrate, or normal se-rum. The eosinopenia of acute infection may be the direct effect of a substancepresent at the site of acute inflammation.

An abrupt reduction in the number of circu-lating eosinophils occurs with the onset ofmany acute infections (10, 12, 14). The mecha-nism of this eosinopenia of acute infection hasnot been proven, although it has been assumedto reflect a response to stress that causes adre-nal corticosteroid release (12). Recent studieshave reexamined the phenomenon in a murinemodel (1, 2, 8). Mice were first rendered eosino-philic by infection with trichinosis. Thesuppression of this eosinophilia by acute in-flammatory stimuli was then examined. Eosi-nopenia accompanied induction ofacute inflam-mation by bacterial, viral, or chemical irritantstimuli (2). The sudden drop in eosinophil lev-els occurred independently of adrenal stimula-tion and in adrenalectomized animals (1). Eo-sinophils rapidly accumulated in the vicinity ofacute inflammation, apparently in sufficientnumbers to account for the initial eosinopenicresponse (2). These observations suggested thatthe eosinopenia might be induced by a sub-stance in the inflammatory exudate. The pres-ent study examined this possibility. An extractof an inflammatory exudate has been shown toproduce an eosinopenic response when injectedintraperitoneally into mice with trichinosis.

1 Present address: Department of Internal Medicine,Bowman Gray School of Medicine, Winston-Salem, NC27103.

MATERIALS AND METHODSUnless stated, the experimental methods were

the same as described previously (1, 2).Basic experimental design. Eosinophilia was pro-

duced by administering muscle-stage larvae of Tri-chinella spiralis to C3H/mg mice. Trichinosis inmice produces two peaks of eosinophilia, the firstbeginning 9 to 10 days after inoculation and coincid-ing with the migration of larvae from the intestinaltract, and the second beginning 21 to 22 days afterinoculation, during the period of larval encystationin striated muscle. Eosinophil counts were obtainedon day 22 or 23 of trichinosis. After verification thatthe eosinophilic response was underway, the ani-mals were separated into equivalent groups by theireosinophil levels. Material to be tested was injectedintraperitoneally at 9:00 p.m. on day 22 or 23 oftrichinosis. Eosinophil counts of experimental andcontrol groups were then determined between 9:00and 11:00 a.m. on days 23 and 24 at the expectedpeak of eosinophilia.

Preparation of inflammatory exudate. Inflam-matory exudate was obtained by lavage of a subcu-taneous pneumococcal abscess. Production of a 2.5-ml subcutaneous air pouch and inoculation with 2 x10o type 3 pneumococci were performed as describedpreviously (1). This produced an acute exudativeresponse within the subcutaneous air pouch. After20 h, the mice were sacrificed by cervical disloca-tion, hair over the area was shaved, the abscess wasopened, and the exudate was removed with a Pas-teur pipette. Approximately 0.1 ml of exudate couldbe removed from each abscess. In the initial experi-

410

on April 27, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

VOL. 15, 1977

ments, the abscess was then ravaged with 0.5 ml ofpH 7.4 phosphate-buffered saline (PBS) (DulbeccoA, Oxoid, London) containing penicillin G (1,200 U/ml). The exudates and lavage solutions were pooledand held at 40C. They were centrifuged at 1,500 x gfor 10 min and then at 17,000 x g for 30 min, andpassed through an 0.45-pum filter (Millipore, Corp.,Bedford, Mass.). To assure prevention of transfer ofactive pneumococcal infection, all recipients wereroutinely given 4 mg of procaine penicillin G subcu-taneously. Intraperitoneal inoculation of one encap-sulated pneumococcus into a mouse is sufficient tocause a lethal infection (11); none of the mice diedafter the transfer experiments.

In the initial experiments, mice were injectedwith 0.6 ml of the pooled material described above.The biological activity of such a preparation varied,depending on the age of the abscess from which ithad been obtained. The quantity of exudate mate-rial transferred will be espressed as the "abscessdose," one abscess dose being the amount ofmaterialobtained from the lavage of one pneumococcal ab-scess. Unless specified otherwise, all experimentsinvolved transfer to each recipient of one abscessdose of material obtained from a pneumococcal ab-scess of 18- to 20-h duration.

In all experiments, control animals received sub-cutaneous procaine penicillin G and an intraperito-neal injection of PBS that had been treated in amanner identical to that of the inflammatory exu-date material.

Preparation of eosinopenic factor (EF). In exper-iments on heat inactivation, dose response, effect inadrenalectomized animals, and effect on neutrophilsand lymphocytes, inflammatory exudate was pro-cessed as follows. The abscess cavity was ravagedwith 0.5 ml of pH 7.4 PBS containing 1,200 U ofpenicillin G and 500 kallikrein-inactivating units ofthe protease inhibitor Trasylol (FBA Pharmaceuti-cals Ltd., Haywards Heath, Sussex) per ml. Proce-dures were performed at 400. Perchloric acid wasadded to a final concentration of 7%. The resultantsuspension was centrifuged at 17,000 x g for 30 min.The supernatant was neutralized with 20% KOHand held at 40C for 1 h before centrifugation at 1,500x g for 10 min. This supernatant was used in thetransfer experiments and is the preparation referredto as "eosinopenic factor."

RESULTSEosinopenic effect of passive transfer of ex-

udate supernatant. Mice were injected intra-peritoneally with 1 abscess dose of exudate su-pernatant after 0.45-,um filtration, and eosino-phil counts were determined after 14 h. Thiscaused a sharp fall in the numbers of circulat-ing eosinophils (Fig. 1). By 36 h after transfer,the effect had disappeared and the eosinophilcounts were similar to those of the controls.

Ultrafiltration. The material was filteredunder suction through a dialysis membrane tohalf its original volume. Intraperitoneal trans-fer of the filtrate had no significant eosinopenic

EOSINOPENIA OF ACUTE INFECTION 411

effect, whereas that of the residue retained theeosinopenic property (Table 1, experiment 1).

Solubility in perchloric acid. Removal ofmost large proteins in the exudate was accom-plished by precipitation with 7% perchloric acidfollowed by neutralization with potassium hy-droxide. Intraperitoneal transfer of the super-natant so obtained retained the eosinopenic ac-tivity (Table 1, experiment 2).Preservation by inhibitor of proteases. Dur-

ing the early experiments, this substance wasfound to be very labile. Holding a preparationat room temperature for 1 h or at 4°C for 10 hwas regularly accompanied by loss of activity.This lability could be an inherent property ofthe substance itself, or it could be due to theaction of degradative enzymes, either endoge-nous or the autolytic enzymes produced bypneumococci (13). Trasylol, a protease inhibitorextracted from bovine lung, has been shown toinhibit a wide variety of proteases, includingtrypsin, chymotrypsin, and plasmin, and theactivation of plasminogen and kallikrein (per-sonal communication from the manufacturer).The exudate material was prepared as previ-ously with the addition of Trasylol, (500 kalli-krein-inactivating units/ml) to the buffer usedin the lavage of the abscess. A similar prepara-tion was made without the use ofTrasylol. Bothwere held at 40C for 24 h before intraperitonealtransfer into mice with trichinosis. The prepa-ration wihout Trasylol had lost its eosinopenia-inducing effect; that with Trasylol retained ac-tivity (Table 1, experiment 3).As noted above, all further experiments used

1600 r

1200

1000 r

8000_

600_

400

200

.1 % Saline controls

0. 45 p frit eofaeudat \

\ / 9~~~(10)

23 24 25DAYS OF TRICHINOSIS

FIG. 1. Effect ofintraperitoneal transfer offiltrateofpneumococcal abscess exudate to mice during eo-sinophilia of trichinosis. Each recipient received theamount obtained from savage of one abscess or asimilar volume of buffered saline (at time of arrow).Eosinophil counts (mean + standard error) weredetermined 12 h later. Numbers in parentheses indi-cate numbers of mice per point.

7777:7-77r7.-r 77.77:77-7!i.!.:. :i. i:, !:..i il.l:...ill.i...:.:Ii:::I..i:.l:'.i.i.:i.::::.'..i:I ii.:::

:,; .:

EE

Ln

on April 27, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

TABLE 1. Effect ofprocedures upon the ability of exudate supernatant to produce eosinopenia afterintraperitoneal injection in mice with trichinosisa

No. of Eosinophils/mm3 (mean ± standard error)Expt Material injected mice

Before injection 14 h after injection1 Controls 5 540 ± 55 570 ± 50

Ultrafiltrate 5 510 ± 52 501 ± 54 NSUltrafiltration residue 5 570 ± 32 105 ± 24 <0.01

2 Controls 5 880 ± 52 826 ± 180Exudate supernatant 5 932 ± 61 160 ± 47 <0.01Perchloric acid supernatant 5 910 ± 58 282 ± 64 <0.01

3 Controls 6 800 ± 48 1,050 ± 146EF held 24 h without Trasylol 6 804 ± 82 989 ± 120 NSEF held 24 h with Trasylol 6 778 ± 60 306 ± 58 <0.01

4 Controls 6 580 ± 62 1,250 + 360EF 6 571 37 410 97 <0.01EF heat inactivated at 560C for 30 min 6 576 ± 42 927 ± 205 NS

a All mice received an intraperitoneal injection of 0.6 ml of experimental or control material at 7 to 9:00p.m. on day 22 or 23 of trichinosis and had blood sampled for eosinophil counts 14 h later.

b Comparison of experimental and control groups. NS, Not significant.

preparation of the exudate material with addi-tion of Trasylol, precipitation with 7% per-chloric acid, and neutralization with potassiumhydroxide. This material is the EF.Heat inactivation. One-half of a preparation

of EF was placed in a water bath at 560C for 30min. This mild heat treatment markedly sup-pressed the eosinopenic activity (Table 1, ex-periment 4).Dose-response relationship. EF was pre-

pared from pneumococcal abscesses of 10- and20-h duration to provide preparations of differ-ent concentrations. Each was serially dilutedand injected into mice on day 22 of trichinosis.Controls were given similarly treated PBS.Eosinophil counts were determined 15 h aftertransfer. The degree of eosinopenia producedwas expressed as the percent suppression of themean of control eosinophil counts. When theabscess dose of EF was compared with the per-cent suppression, a log dose-log response graphprovided a linear regression (Fig. 2). Moreover,the lines formed by the 10- and 20-h prepara-tions were parallel through the concentrationsshown. In greater concentration there was littlefurther suppression, and in weaker concentra-tion suppression was undetectable. The EF ob-tained from an abscess of 10-h duration hadapproximately half the activity of that seen inthe 20-h preparation.Lack of steroid mediation. Adrenal corticos-

teroids may accumulate at the site of inflam-mation (5) and may be present in sufficientquantity to produce an eosinopenic effect ontransfer. A pool of exudate material obtained

8100

°'o _-'?sz (5) (5)0o60-

0. 1 . . . . . .

0 40--z

W030 0 60~~~

z 20 -~0 5U,

0.1 0.2 0.3 0.4 0.6 0.8 1.0

"ABSCESS DOSE" OF EOSINOPENIC FACTOR

FIG. 2. Dose-response effect of the EosinopenicFactor preparation. One abscess dose is the amountofmaterial obtained by lavage ofone mouse pneumo-coccal abscess. Eosinophil counts were determined 15h after intraperitoneal injection ofEF to mice withtrichinosis and expressed as the percent suppressionof eosinophil counts of controls (mean + standarderror).

from 20 abscesses was found to contain 32 ng ofcorticosterone per abscess as determined by thecompetitive protein-binding corticosterone as-say described previously (1). The exudate corti-costerone concentration was, therefore, 32 ,g/100 ml, approximately the level of corticoster-one in the serum of mice 20 h after pneumococ-cal infection (1). If one could assume immediatedistribution (to produce the greatest possibleeffect) in the recipient, this would produce anincrease of 1.6 ,g/100 ml in the serum corticos-terone concentration of the recipient mice, less

412 BASS INFECT . IMMUN .

on April 27, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

VOL. 15, 1977

than that produced by the intraperitoneal injec-tion itself, to which the control mice were sub-jected. It could not explain the eosinopenia ob-served.The possibility of mediation by adrenal stim-

ulation ofrecipient mice was ruled out by trans-fer of the EF to adrenalectomized mice. Micewere adrenalectomized on day 21 of trichinosis.Eosinophil counts were determined on day 22.The mice were divided into equivalent groupsand were given an intraperitoneal injection ofeither EF or similarly treated PBS on the eve-ning of day 22. Eosinophil counts were againdetermined 14 h later. The eosinopenic effect ofthe EF preparation was not prevented by priorbilateral adrenalectomy (Fig. 3).

Effects on circulating neutrophils and lym-phocytes. The possible effects of EF on thenumbers of circulating neutrophil leukocytesand lymphocytes were examined. Eosinophilicmice were divided into 12 groups, with at least6 mice per group. They received 1 abscess doseofEF (or treated PBS), and blood samples wereobtained after the desired intervals, in all casesbetween 9 and 11:00 a.m. Eosinophil, total leu-kocyte, and differential counts were deter-mined. The response is a compilation of the 12groups and required three preparations of exu-date material (Fig. 4). There may have beensome variation in the concentration of the exu-date material, especially since that used for the4-h determination was held for 8 h before trans-fer. Nevertheless, the results suggest that theeosinopenia is significant by 4 h, continues for24 h, and is back to control levels by 36 h. Therewas a reduction of circulating lymphocytes by36% of controls at 8 h after transfer. There wasno significant difference between neutrophilcounts of the control animals and of the EFrecipients.

ADRENALECTOMY

600 r

500

E%EE

-i

z

0

400 _

300 _

200 _

---4 (5 ;Saline

(5I E. F.

pa 0. 025

21 22 23

DAYS OF TRICHINOSIS

FIG. 3. Eosinophil counts (mean + standard er-

ror) ofadrenalectomized, trichinous mice after intra-peritoneal injection (second ) of either EF or simi-larly treated PBS.

EOSINOPENIA OF ACUTE INFECTION 413

Lack of effect of killed pneumococci, pneu-mococcal culture filtrate, or normal mouseserum. Since the exudate was obtained from apneumococcal abscess, possible mediation by aproduct ofthe bacteria was examined. A filtrateof a 20-h pneumococcal culture was prepared.Intraperitoneal injection of 0.6 ml, containing0.1 ml ofculture filtrate diluted as in the prepa-ration of exudate supernatant, did not result ineosinopenia (Table 2). Pneumococci were killedby heating to 100°C for 10 min. Intraperitonealinjection of 107 heat-killed pneumococci also didnot produce a significant reduction ofthe eosin-ophilia of trichinosis (Table 2).

Intraperitoneal injection of protein mightproduce a nonspecific inflammatory reaction,which in turn might cause a reduction in theeosinophil count. This possibility was examinedby intraperitoneal injection of normal mouseserum. Eosinophilic mice received 0.05, 0.1, or0.2 ml of normal mouse serum, a preparation ofEF, or a similar volume of buffered saline (Ta-ble 2). Only the injection of 0.2 ml of serum

EOS NOPENI C

FACTOR

6000 -F

NE UTROPH L S

2000

9000

7000 11 -

LYMPHOCYTES/

m35000-1

3000

900

700 _-

EOS NOPH LS/mm3 500

300 _

100

0 4 8 15 24 36

HOURS

FIG. 4. Effect of one abscess dose of eosinopenicfactor (0) or similarly treated saline (-) on circulat-ing eosinophil, lymphocyte, and neutrophil countsduring the peak eosinophilic response to T. spiralisin mice. Time of intraperitoneal injection varied sothat all samples were taken between 9 and 11:00 a.m.Each point (mean ± standard error) is the mean ofaseparate group of six to eight mice.

on April 27, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

414 BASS

produced a significant suppression of eosinophilcounts, and this suppression was considerablyless than that seen after injection of the EF.Preparations ofEF contained an average of 0.54mg of protein per abscess dose, equivalent toroughly 0.01 ml of serum protein. The amountof protein present in the EF was thus less than5% of that required to produce a significantnonspecific eosinopenic effect.Peritoneal exudate induced by EF. Lavage

of the peritoneal cavity at the time of bloodsampling provided an indicator of any perito-neal reaction produced by the injection of EF.Eosinophilic mice received a dose of EF intra-peritoneally. Twelve hours later, blood sampleswere taken, the peritoneal cavity was ravagedby injection of 3 ml of PBS, and cell counts ofthe lavage fluid were determined (Table 3).Such a procedure can provide only a roughassessment of the total exudate; however, thelavage volume is large in relation to the smallamount of spontaneous peritoneal fluid, and itscontent should be indicative of the relative cel-lular constitution of the peritoneal reaction.Little change was observed after EF adminis-tration (Table 3). The slight increase in eosino-phils is probably inadequate to explain the per-sistent blood eosinopenia. The apparent in-crease in neutrophils is not statistically signifi-

INFECT. IMMUN.

cant. Since the magnitude of acute inflamma-tion sufficient to induce eosinopenia is notknown, the possibility that EF might causeeosinopenia by production of acute peritonealinflammation must be considered; however, thesmall increase in peritoneal neutrophils sug-gests that such a mechanism is unlikely.

DISCUSSIONThe characteristic eosinopenia that accom-

panies many acute infections was first men-tioned by Ehrlich in 1880 and well described byZappert in 1893 (14). By 1914 Schwarz (10) wasable to cite over 100 references confirming theoccurrence ofan absolute eosinopenia as a regu-lar event during the acute phase ofpneumonia,staphylococcal and streptococcal suppurativedisease, erysipelas, epidemic meningitis (pre-sumably meningococcal), typhus, typhoid, mea-sles, varicella, rubella, cholera, and dengue.The most recent clinical review ofthis phenom-enon was that by Weiner and Morkovin in 1952(12). They reaffirmed the regular observation ofa decrease in absolute eosinophil counts duringthe first several days ofmany acute infections,but found the degree and duration difficult topredict. They felt that the best correlation wasseen with the severity of the patient's symp-toms. Higher fevers were generally associated

TABLE 2. Effect of intraperitoneal injection ofpneumococcal culture filtrate, killed pneumococci, or normalserum on eosinophil counts in mice with trichinosis

Eosinophils/mm3 (mean + standard error)Expt Material injected No. of mice P

Before i.p.a injection 14 h after injection1 Controls 8 871 ± 67 950 ± 56

Culture filtrate 6 862 ± 56 881 ± 88 NSbKilled pneumococci 6 837 ± 70 816 ± 74 NSEF 6 843 ± 102 81 ± 27 <0.01

2 Controls 10 981 ± 97 1,372 ± 1280.05 ml of serum 5 950 ± 147 1,325 ± 122 NS0.1 ml of serum 5 973 ± 124 1,152 ± 98 NS0.2 ml of serum 5 1,080 ± 102 730 ± 117 <0.01EF 5 980 + 53 270 + 46 <0.01

a i.p., Intraperitoneal.b NS, Not significant.

TABLE 3. Cellular content ofperitoneal savage after EF administration a

No of Blood eosinophils/mm3 Peritoneal lavage cells/mm3Material injected :mice Before injection 14 h after injection Eosinophils Neutrophils Total leukocytes

Controls. 5 881 ± 53 824 ± 292 89 ± 27 752 ± 138 8,460 ± 1,057EF 5 932 ± 61 161 ± 49 136 ± 29 1,032 ± 227 9,575 ± 628

a All mice received an intraperitoneal injection of Trasylol-perchloric acid-treated PBS (controls) orexudate (EF) at 9:00 p.m. on day 23 of trichinosis. Blood eosinophil counts were obtained 12 h later. Resultsare for peritoneum ravaged with 3 ml of PBS and cell counts obtained on resultant fluid. Results areexpressed as mean + standard error.

on April 27, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

VOL. 15, 1977

with more profound eosinopenia, yet the eosino-phils returned to normal before subsidence ofthe fever in several patients. The eosinopeniadid not consistently correlate with the presenceof neutrophilia or with the erythrocyte sedi-mentation rate. Thus, although in generalthose patients with more symptoms, higher fe-vers, and more leukocytosis had greater andlonger suppression of the eosinophil counts,none of these individual variables could relia-bly predict the eosinophil behavior in an indi-vidual patient. The eosinopenia of acute infec-tion usually occurs even in patients with apreexistent eosinophilia; however, there havebeen occasional reports of instances of refrac-tory eosinophilia, e.g., in patients with Loffier'sendomyocarditis where acute infection was un-able to suppress the eosinophilia (3, 9).Eosinopenia occurs as part of the response to

stress (4). This response may be duplicated byadministration of glucocorticosteroids or adeno-corticotrophic hormone (7). A similar eosino-penic response occurs after injection of epineph-rine, although this response is also apparentlydependent upon the presence of low, "permis-sive" levels of adrenal glucocorticoids (6). Theassumption arose that the eosinopenia of acuteinfection represented a nonspecific manifesta-tion ofthese hormonal responses to the stress ofacute infection. This hypothesis was acceptedby Weiner and Morkovin (12), although theirseries included observations of eosinopenia dur-ing localized infections such as pharyngitis,which would not be expected to exert a strongadrenal stimulation. Recent studies in eosino-philic mice demonstrated that the eosinopenicresponse to acute inflammation preceded therise of serum corticosterone and occurred nor-mally in adrenalectomized animals, withoutany requirement for maintenance "permissive"corticosteroid replacement (1). Thus, the eosi-nopenia of acute infection occurs independentlyof the hormonal reactions to nonspecific stress;its cause must arise in another aspect of thehost response to acute inflammation. The pres-ent study demonstrates that passive transfer ofmaterial obtained from acute inflammatory ex-udate may duplicate such an eosinopenic re-sponse in recipient animals.The pneumococcal infection of a subcuta-

neous air pouch provided a convenient source ofexudate produced by a pathogen free of knowntoxin production. After removal of bacteria andcellular debris by 0.45-.um filtration, intraperi-toneal injection of 0.1 ml of exudate into eosino-philic mice was followed by a fall in numbers ofcirculating eosinophils which was significantwithin 4 h and lasted for 24 h after transfer.

EOSINOPENIA OF ACUTE INFECTION 415

The active material was retained by a dialysismembrane, was soluble in 7% perchloric acid,and was protected from destruction by autolyticenzymes by a protease inhibitor. Althoughthese data provide only a minimal characteriza-tion of the material, they suggest that the ac-tive agent may be a large glycoprotein.The active material is termed "eosinopenic

factor" for convenience, although induction ofeosinopenia may not be its only biological func-tion. The mechanism of action of EF is notknown. The rapid eosinopenic response sug-gests a direct effect on the eosinophil, perhapsby increasing eosinophil adhesiveness andthereby inducing intravascular margination ofthe cells. This effect might also occur by activa-tion of other endogenous substances yet to bedefined. Such speculations must await verifica-tion after further purification of the EF andstudy of its effects on eosinophil leukocytes.The data suggest that the eosinopenic re-

sponse is not merely a manifestation of an in-flammatory response to the material injectedintraperitoneally. Injection of 20 times the pro-tein present in 1 abscess dose of EF producedminimal eosinopenia. Conversely, transfer of25% of the material obtained from one abscessproduced a systemic eosinopenic response ofover 12-h duration in recipient mice. The mate-rial is thus active at concentrations consider-ably below those found in vivo at an inflamma-tory site.The roles of the eosinophil, EF, and the eosi-

nopenia of acute infection remain undefined.Further study of those phenomena distinctiveto eosinophils may suggest new approaches tothe role of these cells. One such phenomenon,the eosinopenia of acute infection, is duplicatedby passive transfer of a material obtained frominflammatory exudate. Clarification of thecharacter and physiological effects of this mate-rial should provide insight into one of the mostdistinctive aspects of eosinophil behavior.

ACKNOWLEDGMENTSI gratefully acknowledge the guidance ofPaul B. Beeson,

who directed this study. I thank Sarah Davies for excellenttechnical assistance.

This work was supported by grants from the MedicalResearch Council and the Commonwealth Fund.

LITERATURE CITED1. Bass, D. A. 1975. Behavior of eosinophil leukocytes in

acute inflammation. I. Lack ofdependence on adrenalfunction. J. Clin. Invest. 55:1229-1236.

2. Bass, D. A. 1975. Behavior of eosinophil leukocytes inacute inflammation. II. Eosinophil dynamics in acuteinflammation. J. Clin. Invest. 56:870-879.

3. Castleman, B. 1970. Case records of the MassachusettsGeneral Hospital. N. Engl. J. Med. 283:476-485.

on April 27, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

416 BASS

4. Dalton, A. J., and H. Selye. 1939. The blood pictureduring the alarm reaction. Folia Haematol. 62:397-407.

5. Dougherty, T. F., H. E. Brown, and D. L. Berliner.1958. Metabolism ofhydrocortisone during inflamma-tion. Endocrinology 62:455.

6. Henry, W. L., L. Oliner, and E. R. Ramey. 1953. Rela-tionship between actions of adrenocortical steroidsand adrenomedullary hormones in the production ofeosinopenia. Am. J. Physiol. 174:455-458.

7. Hills, A. G., P. H. Forsham, and C. A. Finch. 1948.Changes in circulating leukocytes induced by theadministration of pituitary adrenocorticotrophic hor-mone in man. Blood 3:755-768.

8. Morgan, J. Elizabeth, and P. B. Beeson. 1971. Experi-mental observations on the eosinopenia induced byacute infection. Br. J. Exp. Pathol. 52:214-220.

9. Roberts, W. C., D. G. Liegler, and P. P. Carbone. 1969.

INFECT. IMMUN.

Endomyocardial disease and eosinophilia. A clinicaland pathologic spectrum. Am. J. Med. 46:28.

10. Schwarz, E. 1914. Die Lehre von der allgemeinen undortlichen Eosinophilie. Ergeb. Allg. Pathol. Pathol.Anat. 17:137-787.

11. Wamoscher, L. 1927. Uber Pneumokokkeninfektionenbei verminderter individueller Resistenz. Z. Hyg. In-fektionskr. 107:240.

12. Weiner, H. A., and D. Morkovin. 1952. Circulatingblood eosinophils in acute infectious disease and theeosinopenic response. Am. J. Med. 13:58-72.

13. White, B. 1938. The biology of the pneumococcus: thebacteriological, biochemical and immunological char-acters and activities of Diplococcus pneumoniae.Commonwealth Fund, New York.

14. Zappert, J. 1893. Uber das Vorkommen der Eosinophi-len Zellen im menschlichen Blute. Z. Klin. Med.23:227-308.

on April 27, 2021 by guest

http://iai.asm.org/

Dow

nloaded from