Journal of Clinical InvestigationVol. 43, No. 11, 1964

Studies of Staphylococcal Infections. I. Virulence of Staphy-lococci and Characteristics of Infections in Embryonated

Eggs *WILLIAM R. MCCABEt

(From the Research Laboratory, West Side Veterans Administration Hospital, and theDepartment of Medicine, Research and Educational Hospitals, University of Illinois

College of Medicine, Chicago, Ill.)

Many of the determinants of the pathogenesisand course of staphylococcal infections remainimprecisely defined (1, 2) despite their increas-ing importance (3-10). Experimental infectionsin suitable laboratory animals have been of con-siderable assistance in clarifying the role of hostdefense mechanisms and specific bacterial virulencefactors with a variety of other infectious agents.A sensitive experimental model would be of valuein defining the importance of these factors instaphylococcal infections, but both humans andthe usual laboratory animals are relatively re-sistant. Extremely large numbers of staphylo-cocci are required to produce either local or sys-temic infections experimentally. Elek was un-able to produce local purulent lesions in humanswith inocula of less than 106 staphylococci (11)but subsequently demonstrated that the infectiousinoculum could be reduced by two to four logswhen the infection was combined with concomitantinsertion of a foreign body (12). Even largernumbers of staphylococci have been required toproduce local lesions in mice (13, 14), guineapigs (15, 16), and rabbits (17, 18). Varioustraumatic procedures or implantation of foreignbodies also allows the production of infection withsmaller numbers of bacteria in animals (14, 19),but despite such measures many of these tech-

* Submitted for publication May 4, 1964; acceptedJuly 9, 1964.

Reported in part at the joint sessions of the AmericanSociety for Clinical Investigation and the AmericanFederation for Clinical Research, April 28, 1963, AtlanticCity, N. J.

This work was completed during the tenure of aVeterans Administration Clinical Investigatorship.

t Present address: Evans Memorial Department ofClinical Research, Massachusetts Memorial Hospitals,and the Department of Medicine, Boston UniversitySchool of Medicine, Boston University Medical Center.

niques still require relatively large numbers ofstaphylococci to produce infection (19). Fatalsystemic infections have been equally difficult toproduce in animals and have necessitated the in-jection of 107 to 109 bacteria (20-23). A fewstrains of staphylococci have been found that arecapable of producing lethal systemic infectionswith inocula of from 102 to 103 bacteria (24) andhave excited considerable interest (25-27). Thevirulence of these strains apparently results froman unusual antigenic variation (27, 28) which,despite its interest, is of doubtful significance inhuman staphylococcal infection, since such strainshave been isolated only rarely from clinical in-fections (29). The insensitivity of these experi-mental systems has been a handicap for compari-son of the relative virulence of strains of staphylo-cocci and evaluation of the effect of antibody andother defense mechanisms in a living system.

Preliminary studies of staphylococcal infectionsin a variety of animals indicated that embryonatedeggs were readily infected with pathogenic staphy-lococci (30). The present study describes thesensitivity of embryonated eggs to 68 strains ofcoagulase positive Staphylococcus aureus and 36strains of coagulase negative Staphylococcus epi-dermidis and defines some of the characteristicsof this infection. Subsequent publications will de-scribe the importance of individual bacterial viru-lence factors (31) and the protective effect of anti-toxic and antibacterial antibody in the chick em-bryo (32).

Methods

Strains of staphylococci. Forty-two strains of coagu-lase positive Staphylococcus aureus were obtained fromblood cultures and specimens of sputum or purulentexudates submitted to the routine bacteriology labora-tory. In each instance, the patient from whom the

2146

EXPERIMENTALSTAPHYLOCOCCALINFECTIONS IN EMBRYONATEDEGGS

specimen was obtained was examined to insure that thestaphylococcus isolated was etiologically related to theclinical infection. Additional strains were obtainedfrom the anterior nares and skin of a group of patientswithout clinical evidence of infection who were attend-ing an outpatient clinic f or the first time. Specimenswere obtained only f rom outpatients who had not re-ceived antibiotics within the past 3 months, had notbeen hospitalized during this period, and whose householdcontacts had not been hospitalized recently. A thirdgroup of 36 strains of coagulase negative Staphylococ-ctus epidermidis was obtained from a variety of speci-mens submitted to the hospital laboratory and from thenares and skin of outpatients.

Tube and slide coagulase determinations and mannitolfermentation were performed on all isolated strains toinsure their identification. Strains of staphylococci weremaintained on tryptose phosphate (TP) agar slantsat 0° C after isolation and identification until egg viru-lence studies were performed the following week.Immediately before use, a single colony selected froma subculture of the original agar slant was inoculatedinto TP broth for the infectious inoculum. Several 0.1-ml samples of the latter were transferred to 2.5 ml ofTP broth in screw-capped vials and stored in a - 200 Cdeep freeze for subsequent use.

Preparation and injection of inocula. Serial tenfolddilutions of overnight TP broth cultures of staphylococciwere made to provide inocula of approximately 1,000 to99,999 (10' to 10') and 1 to 100 colony forming units(cfu). The smaller inocula consisted of less than 10cfu in over one-half the studies. The size of the inoculumwas determined by plating 0.1 ml of 10-', 10', and 10-dilutions of the overnight culture on TP agar and enu-merating the number of colonies after overnight incu-bation. Lots of 20 10-day-old fertile eggs were injectedintra-allantoically with 0.1 ml of 10-s and 10' dilutionsof each strain studied, and the eggs were candied dailyto determine viability. A group of ten controls was in-jected with a similar volume of sterile 0.85% saline(PS) in each experiment. Subcultures were obtainedfrom allantoic fluid, yolk, and embryonic blood in thefirst sets of experiments and, when indicated, in sub-sequent studies to insure that maternal transmission ofinfection or subsequent contamination had not occurred.

Eggs. Fertilized eggs were obtained from a com-mercial source that did not use an antibiotic dietarysupplement. Tube dilution assays for antibacterial ac-tivity were performed on allantoic and amniotic fluid onseveral occasion with negative results. Eggs were incu-bated for 10 davs at 370 C in a rocker incubator beforeinfection and- then incubated upright at 38.5° C with60% relative humidity for an additional 10 days afterinfection.

Recovery of inoculum. Studies to determine the num-ber of cfu that could be recovered immediately after in-jection were carried out by two techniques. Eggs wereinjected as described above, allowed to stand for 15 min-utes, sealed with paraffin, shaken gently, and sacrificed.

When recoveries were carried out after injection of alarge inoculum, 12-day-old eggs in which the volume ofallantoic fluid is stated to be 6 ml (33, 34) were usedfor injection. One-tenth ml samples of allantoic fluidwere aspirated aseptically after removal of the shellover the egg sac, plated, the colonies enumerated, andthe numbers corrected for the 1 to 60 dilution. Re-covery with the smaller inoculum was performed simi-larly, except as much of the allantoic fluid as possiblewas aspirated and 4 ml of PS was injected, mixed, andremoved. The latter procedure was repeated nine times,and the pooled allantoic fluid and saline washings werefiltered through a Millipore H.A. filter (0.45 Ah). Afterfiltration, the filter was removed, placed on a bloodagar plate, incubated for 48 hours, and the number ofbacterial colonies determined.

In vitro growth curves. Allantoic fluid was collectedaseptically from 15 eggs, pooled, and 4.9 ml dispensedinto sterile capped tubes for determination of growthcurves. One-tenth ml of a 10' dilution of the organismto be tested was added and the allantoic fluid incubatedfor 18 hours. Duplicate determinations were made ateach sampling time. Concomitant growth curves in TPbroth were performed in a similar fashion.

In vivo growth curves. Lots of 64 10-day-old embry-onated eggs were injected with 0.1 ml of a 10' and 10'dilution of coagulase positive staphylococci and 0.1 ml ofa 10' dilution of coagulase negative staphylococci.Four eggs from each of the three groups were sacrificedat 4, 8, 18, and 24 hours and daily thereafter for 10 days.Only eggs containing living embryos were sacrificed toinsure that acceleration of growth and invasion of theembryo "post-mortem" did not obscure the results.The shell over the air sac was removed; allantoic fluidwas aspirated aseptically and serially diluted for platecounts. The remainder of the allantoic fluid was re-moved, and 1.5% saponified cresylic acid' was added tothe shell containing the intact yolk sac, amnion, andembryo. After 10 minutes, the cresylic acid solutionand the yolk were removed. The amnion and embryowere extracted with sterile forceps and placed in asterile petri dish. The amnion was removed and theembryo washed in the cresylic acid solution and rinsedwith sterile PS. The embryo was opened by a ventralincision, and the heart and liver were removed asepti-cally. These were ground in a glass tube with a Teflontissue grinder, and 0.2 ml of the bloody supernatantfluid was removed. Bacterial titers were determined byplating 0.1 ml of undiluted fluid and serial tenfolddilutions of the fluid.

Results

Lethality of coagulase positive and coagulasenegative staphylococci. The 10-day fatality ratesfor 10-day-old eggs injected intra-allantoically withinocula of two different sizes of 68 strains of co-

1 Cresylone, Parke, Davis & Co., Detroit, Mich.

2147

WILLIAM R. McCABE

GOAGULASEPOSITIVE STAPHYLOCOCCIv

Isolated from Nores'Isolated from Infections of Outpatients

(Nx42 ) (Nu26)

COAGULASENEGATIVE

STAPHYLOCOCCI(N= 36)

INOCULUm. 105- DACTERIA

mn AI1 l I

INOCULUM. I- A0OACTERIA

a .1. II1. I

5 10 15 20 25 30 35 42 5 10 IS 20 26 5 10 I5 20 25 30 36

FIG. 1. CHICK EMBRYOLETHALITY OF COAGULASEPOSITIVE STAPHYLOCOCCI ISOLATED FROM

INFECTIONS COMPAREDWITH THAT OF STRAINSNEGATIVE STAPHYLOCOCCI.

agulase positive Staphylococcus aureus and 36strains of coagulase negative Staphylococcus epi-dermidis are illustrated in Figure 1. Two groupsof coagulase positive staphylococci were studied.Forty-two strains were isolated from unequivocalhuman infections in which they were the soleetiologic agent of the infections. These strainsare referred to, henceforth, as pathogenic staphylo-cocci. A second group of 26 strains isolated fromthe nares and skin of uninfected outpatients rep-

resents staphylococci of uncertain pathogenicity.The latter group, whose over-all human patho-genicity, although uncertain, would be anticipatedto be. less than that of strains isolated from definitehuman infections, was selected to sample strainscharacteristic of those prevalent in the community.The third group, coagulase negative Staphylococ-cus epidermidis, represented strains that are rarelyhuman pathogens.

ISOLATED FROMNASAL CARRIERS AND COAGULASE

Pathogenic staphylococci produced fatality ratesranging from 50 to 100%, mean 89.2 14.9,when 103 or 104 cfu was injected, and from 40 to

1007%, mean 75.8 18.5, with inocula of less than100 cfu. Fatality rates of 509o% or greater oc-

curred with all strains isolated from human in-fections with the larger inocula, and with 40 of42 of these strains with the smaller inocula. Thelethality of the 26 strains of coagulase positivestaphylococci isolated from outpatients was be-tween 15 and 100%, mean 72.8 26.9, with thelarge inocula and from 5 to 100%o, mean 60.029.3, with the small inocula. Both the large andsmall inocula of pathogenic strains were signifi-cantly more lethal (tat = 4.5, p < 0.001; t66 =

3.95, p < 0.001) than were equivalent inocula ofstrains isolated from outpatients. Although strainsthat had produced clinical infections were more

virulent for embryonated eggs in general than

100

75 -

50-

25 -

100

75 h

50 -

1%-

N.4

-4

kz

25

2148

I morEmusMDONE' m.1 .... I..IR-

EXPERIMENTALSTAPHYLOCOCCALINFECTIONS IN EMBRYONATEDEGGS

N~bmm BACTEREMIAMild to Moderate Severe PNEUMONIA BACTEREIAT

WOUNDINFECTIONS IOSTEOMYELITISFIG. 2. LETHALITY OF STRAINS OF STAPHYLOCOCCUSAUREUSISOLATED FROM

INFECTIONS OF VARIOUS TYPES AND DEGREESOF SEVERITY.

strains isolated from uninfected patients, consider-able overlap in egg virulence was observed withindividual strains of both types.

Coagulase negative strains were significantlyless lethal than coagulase positive staphylococciisolated either from septic patients or uninfectedoutpatients. Fatality rates did not exceed 20%oeither with inocula of 1 to 100 or 108 to 104 cfu,and the mean fatality rate with the larger inocula,11.0 + 6.1%, did not differ significantly (t66 =

1.65, p = 0.1) from the fatality rate with thesmaller inocula, 8.3 + 5.8%. The mean fatalityrate produced by 1 to 100 cfu of coagulase positivestaphylococci isolated from outpatients was sig-nificantly greater (t86 = 10.1, p < 0.001) thanthat produced by 100 to 1,000 times as many co-

agulase negative staphylococci.Lethality of strains isolated from infections of

varying severity. Figure 2 demonstrates the fa-tality rates observed in embryonated eggs afterinjection of 1 to 100 cfu of strains of pathogenicstaphylococci isolated from different types ofclinical infections. Strains isolated from patientswith staphylococcal bacteremia, pneumonia, andwound infections were studied. Five patients hadparticularly extensive and severe wound infections,and the strains isolated from these patients are

distinguished from strains causing less severe in-fections in Figure 2. Severe wound infections

were differentiated by the following criteria: oc-currence in patients without serious underlyingdisease, extensive inflammatory reaction, leuko-cytosis greater than 17,000 per mm3, and feverin excess of 1020 F for 5 days or longer. Allother wound infections were classified as being ofmild to moderate severity when two or fewer ofthe above criteria were present. The latter in-fections usually consisted of stitch abscesses andinfections producing moderate purulent drainageand wound induration with only slight systemicsymptoms. No further attempt was made toclassify the severity of infections except in thesefive instances. Staphylococcal pneumonia wasassociated with bacteremia in two patients; thesestrains are shown in both categories. Three pa-tients with primary hematogenous osteomyelitiswere observed, and their strains are groupedwith strains isolated from patients with bactere-mia. The mean fatality rate from infections withstrains causing mild to moderate wound infectionswas 62.6 + 17.0% and was significantly less(t21 = 3.5, p < 0.01; t28 = 3.6, p < 0.01; t28=

4.7, p < 0.001) than the fatality rate of infectionswith strains isolated either from patients withsevere wound infections (87%), pneumonia(86%), or bacteremia (87%). Even when strainsassociated with severe wound infections werecombined with those that produced mild to mod-

2149

WILLIAM R. McCABE

FATALITIES caw#FECT/ os 105_104A) % 07V)POSfT /-/00 *-100 _ STAP ou TPATIENTs -1-04 A---A

ifax /-/00 A---A

COAG. NEGATVESTA. i0o-io ZoZft1#/-/00 @-.

SALNECONTROLS x- x

25

DAYS AFTER INFECTION

FIG. 3. CUMULATIVE FATALITY RATES AFTER ALLANTOIC INFECTIONWITH STRAINS OF STAPHYLOCOCCUSAUREUS AND STAPHYLOCOCCUSEPI-DERMIDIS.

crate infections, their over-all lethality was sig-nificantly less than that of strains isolated frompneumonia and bacteremia. Lethality did notdiffer materially, however, for strains isolatedfrom patients with severe wound infections, pneu-monia, or bacteremia.

Size of inoculum and rapidity of death. Theincrease in fatality attained with a 1,000-fold in-crease in inoculum size over that observed witha small inoculum often was not so great as mighthave been anticipated. This finding appeared toreflect resistance of a small proportion of the eggsto fatal infection in addition to differences in therate of killing by inocula of 100 and 103 to 104staphylococci during the period of observation.The effect of the latter may be seen in part in thecumulative daily death rates with the two sizes ofinocula of the three groups of staphylococci il-lustrated in Figure 3. Larger inocula of patho-genic staphylococci were more rapidly fatal andproduced 70% of the total fatalities within thefirst 2 days after infection, and only a small pro-portion of deaths occurred later in the 10-day pe-riod. With smaller inocula, there was an initiallag before deaths were apparent, and the death

rate increased progressively until the final day ofobservation, with 60%o of the total fatalities oc-curring after the second day. A similar but lessstriking relation between the size of the inoculumand rapidity of fatalities also occurred with strainsisolated from uninfected outpatients. The effectof the size of the inoculum on the time of deathwas even more apparent when individual strainswere considered. More virulent strains that pro-duced fatality rates of 85% or greater with asmall inocula produced 84% of their fatalities inthe first 2 days after infection with 103 bacteria,whereas less lethal strains produced only 47% oftheir fatalities within this time (X2 = 7.3, p <0.01).

Lethality of coagulase negative staphylococciand culture media. Since coagulase negativestaphylococci produced very few fatalities in thecomparative studies, their virulence was assessedmore critically by comparing the lethality of fivestrains with that of the culture media. Tenfolddilutions of overnight TP broth cultures weremade in sterile PS, and 0.1 ml of undiluted cul-tures and 10-1, 10-2, 10-3 dilutions were injectedintra-allantoically into ten eggs per dilution with

2150

EXPERIMENTALSTAPHYLOCOCCALINFECTIONS IN EMBRYONATEDEGGS

INOCULUIMV. /-/OX /OS SACTERIA

.44

1

N.1k

K4

0 2 3 4 5 6 7

MONTHS

FIG. 4. REPRODUCIBILITY OF FATALITY RATES IN EMBRYONATEDEGGS OF

FOUR STRAINS OF STAPHYLOCOCCUSAUREUS STUDIED OVER A 7-MONTHPERIOD.

particular care being taken to minimize trauma tothe eggs. Sterile TP broth was diluted and in-jected in a similar fashion, and the number ofdeaths was determined daily for 10 days for allfive groups. The inoculum varied from 1 x 107to 2 x 108 cfu with undiluted cultures. The num-

ber of survivors in each group is shown in TableI. No significant differences were observedamong any of the groups, and undiluted overnightcultures of coagulase negative staphylococci were

no more lethal than the media in which they were

grown.

Recovery of injected staphylococci. Some leak-age occurred after the injection of bacteria in

TABLE I

Comparative lethality of coagulase negative staphylococciand sterile culture media

Strains of coagulase negative staphylococciSterile

Dilution broth F R B K S

Undiluted 9/10* 9/10 8/10 10/10 9/10 8/1010-1 10/10 10/10 10/10 9/10 10/10 10/1010- 9/10 10/10 10/10 10/10 10/10 9/1010-5 10/10 10/10 9/10 9/10 10/10 10/10

* No. of survivors at 10 days/no. of eggs injected.

many instances. To insure that a large proportionof the injected organisms was not lost, recoverieswere carried out with two different sizes of in-oculum as described previously. The results offour such recoveries are shown in Table II. Ineach instance, recoveries approximated 60% ofthe bacteria injected and *were well within therange of error of the experimental method.

Reproducibility of fatality rates. The virulenceof individual strains was reproducible over a pe-riod of several months. Serial assessments ofthe lethality of three virulent strains and onenonvirulent strain, shown in Figure 4, were madeat intervals of as long as 7 months. Fatality

TABLE II

Recovery of injected bacteria from allantoic fluid

No. of bacteriaPer cent

Injected Recovered recovered

Cfu* cfu 5%3.1 X 104 1.8 X 104 584.5 X 104 3.0 X 104 67

7 4 597 4 59

* cfu = colony forming units.

2151

2N2ILL LIAMt R. McCABE

INOCULUM. /-/Ox /O3 BACTERIA

k 0 0|B.14

K INOCULUM. /-/O BACTERIA~.11 1Qz 10U -

75

K 50

25

* COAG. POSIr/VE STAPH.* COAG. NEGAT/VE STAPH.

00L7 days 8 days 10 days 13 days 15 days

AGE OF EMBRYOAT INFECTIONFIG. 5. SUSCEPTIBII ITY OF EMBRYONATEDEGGS OF V-ARIOt-S AG;ES TO

LETHAL INFECTION WX ITH STAPHYLOCOCCUSAI RElS AND STA.INYloCOCCv5sEPIDERMIS.

rates did not vary more thani 10%/. even w\ henistudies were iatlde after proloigedi intervals. Anuml)er of other straills not shown in this Figutrehave l)eei stuldieci On sev-eral ccaISions Withequally reproulucible results.

SelSCeptibility of ciil)r\vos of different agts.The age of the embryo exerted a significant in-fluence on its susceptibility to lethal infection withboth coagulase positive and negative staphylococci.The five-dav fatality rates produced by threestrains of coagulase positive and by three strainsof coagtilase negative staphylococci for 7- 8- 10-13-, and 15-day-old chick embryos are shown inFigure 5. Seven- aid, to a lesser degree, 8-day-old embryos were susceptible to lethal infectionswvith coagulase negative strains but became re-sistalnt b1 the tenth day. The age of the embryo

odiflied the virulence of coagulase positive strainsin a similar fashion. Almost uniform fatality oc-curred in 7- or 8-day old embryos; 13- and 15-day-old embryos were less susceptible. The greatersuscel)tibilitv of the younger emnbryos did notobscure the differences in virulence between co-agulase positive and coagulase negative strains.The former strains were significantly more virmi-

lent at all ages, although the differences weremost evident in embryos 10 days of age or older.

(;GroZL'tli of staplilococci in allan toxic fluid inv it ro. ( _-ro-wth ctirves in allantoic fluid inl Vitro\were determined for three strains of coagulasenegative sta)hylococci and three strains of coagu-lase positive staphylococci and are compared withoroxvth curves of each in TP broth in Figure 6.Multiplication of coagulase Ilegative staphylococciin allantoic fluid lagged behind that observed inbroth and was less than that of coagulase positivestrains in either type of mieditim during the first6 hours. This difference failed to persist, how-ever, and both types attained similar populationdensities of 10S bacteria per ml by 18 hours in both

nedia.(;roowtl of staplivlococci in vivo. In zvivo growth

curves also were determined in both allantoicfluid and the embryo after injection of both alarge (2 x 103 cfu) anId small (2 cfu) inoculum ofcoagulase positive staphylococci, and a large in-ocultim (9 X 103 cfu) of coagulase negativestaphylococci. These growth curves plus the cum-

ulative dlailx fatality rates of the strains studiedare illustrated in Figlure 7. Points on the graph

21 S

EXPERIMENTALSTAPHYLOCOCCALINFECTIONS IN EMBRYONATEDEGGS2

represent the mean of counts obtained from foureggs in each of three experiments. Growthcurves of the large inoculum of coagulase posi-tive and coagulase negative staphylococci werequite similar for the first 3 days with both reach-ing numbers of 108 cfu per ml or greater in al-lantoic fluid and 103 to 104 cfu in the embryo.The high early fatality rate in eggs injected with103 cfu of pathogenic staphylococci precluded as-sessment of growth beyond this period, althoughthis measurement was attempted with lots of 64eggs on three separate occasions. After the in-jection of a small number of coagulase positivestaphylococci (2 cfu), 48 hours elapsed beforetheir numbers reached those attained by thelarger inoculum, but after this period the micro-bial population persisted essentially unchanged inboth allantoic fluid (108 per ml) and the embryo(103 to 104) for the remainder of the study. Incontrast, after an initial rise, a progressive de-crease in the number of coagulase negative staphy-lococci was observed in both the embryo andallantoic fluid until the number of coagulase nega-tive staphylococci present on day 3 had de-creased by 99.7% in the allantoic fluid, and by99.2% in the embryo by the tenth day. All threeexperiments gave similar results, with coagulasepositive staphylococci attaining similar numbersin 48 hours and persisting at this level, while the

8

7

6\ 6

t3

42

O 2 4 6 8 10 12 14INCUBATION TIME, HOURS

16 18

FIG. 6. GROWTHCURVES OF STAPHYLOCOCCUSAUREUS

AND STAPHYLOCOCCUSEPIDERMIDIS IN ALLANTOIC PLUIDAND BROTH MEDIA.

FATALITY RATE FCUMAULATIVEMORTALITY RATE

75 ------

50 .

2F 8 16 24 2 3 4 5 67-

0 162 234 67 8 9 IC',- HOURS-' DAYS

FIG. 7. GROWTHCURVES AND CUMULATIVE FATALITYRATES OF STAPHYLOCOCCUSAUREUSAND STAPHYLOCOCCUSEPIDERMIDIS IN ALLANTOIC FLUID AND EMBRYONICTISSUEIN vivo. Each point on the growth curves represents themean of counts obtained from four eggs in each of threeexperiments.

progressive decrease in numbers of coagulasenegative staphylococci after days 3 and 4 alsooccurred in each study.

Invasion of the embryo was observed onlywhen bacterial counts in the allantoic fluid ex-ceeded 107 cfu per ml. Involvement of the embryoappeared to be related to increasing fatalities withcoagulase positive staphylococci, but coagulasenegative staphylococci were capable also of in-vading the embryo without producing fatalities.

Discussion

These investigations were undertaken in anattempt to find a more sensitive experimental hostfor the study of staphylococcal infections. Embry-onated eggs seem to possess many features thatwould be desirable in such an experimental model.They are extremely susceptible to infection withsmall numbers of pathogenic staphylococci, andthe lethality of individual strains is consistent andreproducible over a period of several months. Thesensitivity to infection also may be modified byutilizing the differences in susceptibility of em-

X ~ ~ ~~~ - Br@oth,COAGM/ASEf-SrfiWAfMj** * C-A Fluid

COAGNFGAT/-Vj-fm--- Broth-- C-A Fluid

2153

WILLIAM R. McCABE

bryos of various ages if greater or lesser lethalityis desired. In addition, a distinct relationship be-tween the pathogenicity of staphylococci for hu-mans and their virulence for chick embryos wasobserved, confirming earlier preliminary findings(30). Strains of Staphylococcus aureus definitelyresponsible for human infections were extremelylethal and almost uniformly produced fatalityrates of 50% or greater when less than 100, andusually less than 10 cfu, were injected. Otherstrains of indeterminant potential pathogenicityto man, isolated from uninfected carriers, ex-hibited, individually, variable degrees of virulencefor the chick embryo but as a group were sig-nificantly less lethal than strains isolated fromclinical infection. Fatality rates from infectionwith strains that are avirulent or noninvasive forhumans, coagulase negative Staphylococcus epi-dermidis, were no greater than those occurringafter injection of culture media similar to thatused for their propagation. Further correlationbetween the severity of the clinical infections andthe virulence of individual strains for chick em-bryos was also observed. Staphylococci isolatedfrom patients with bacteremia or pneumonia weremore lethal than strains isolated from wound in-fections, and strains isolated from more severewound infections produced greater fatality ratesthan strains isolated from less severe woundinfections.

The excellent correlation observed betweenthe pathogenicity and virulence of staphylococcifor humans and their lethality in chick embryosdoes not imply, however, that this system can beutilized to predict the pathogenicity of staphylo-cocci for humans. The parameters used to char-acterize clinical behavior of staphylococci arerelatively imprecise, and the potential patho-genicity of individual staphylococci cannot be pre-dicted accurately by any currently available tech-niques. Precise quantification and assignment ofrelative degrees of virulence will require thoroughepidemiologic investigation of the frequency andtypes of human infection in relation to contactand acquisition of individual strains as well asconcomitant appraisal of predisposing host factors.Until such information is available, definitive cor-relation between the similarity of behavior ofstaphylococci in humans and animals will not bepossible.

Embryonated eggs have been used predomi-nantly for the propagation and study of virusesand rickettsiae, although Goodpasture recognizedtheir potential value for investigation of bacterialinfections (35) soon after he popularized theiruse for cultivation of viruses (36). One of hisearly studies with this model described the lesionsproduced by infection of allantoic membranes andthe ability of a variety of bacteria, includingstaphylococci, to invade the membrane and pro-duce lethal infections (35). Subsequently, Frap-pier and Sonea observed that staphylococci in-jected subcutaneously into 18- or 19-day-old em-bryos produced lethal infections (37), andKnothe studied a variety of allantoic infectionsincluding those produced by staphylococci (38).The extreme sensitivity of chick embryos tostaphylococci was noted in both of these investi-gations and more recently by Wiley, who deter-mined the lethality of four strains of Staphylo-coccus aureus in embryonated eggs (39). How-ever, the limited number of strains evaluated andthe failure to further extend these observationshas hampered recognition of the potential value-of this model.

Explanation of the marked differences in viru-lence of coagulase positive and coagulase nega-tive staphylococci for embryonated eggs is notreadily apparent from these studies. Both typesof organisms grew equally well in allantoic fluidin vitro and initially in vivo, and both invadedthe embryo with equal facility. but subsequentlytheir behavior differed. Pathogenic staphylo-cocci persisted at essentially the same microbialpopulation attained after logarithmic growth inboth allantoic fluid and the embryo throughoutthe entire infection. In contrast, although coagu-lase negative staphylococci reached a populationsimilar in magnitude to that of pathogenic staphy-lococci, they were nonlethal, and a progressive-decrease in their numbers occurred after the-third day until only 0.3% of the maximal popula-tion attained in the allantoic fluid and only 0.8%attained in the embryo remained by the tenth dayafter infection. A variety of enzymes and exo-toxins has been purported to be responsible forthe virulence of coagulase positive staphylococciand may play an important role in infectionscaused by these organisms in the chick embryo.If exotoxins do exert a significant influence on

2154

EXPERIMENTALSTAPHYLOCOCCALINFECTIONS IN EMBRYONATEDEGGS

the course of the infection, production of exotoxinin vivo rather than injection of preformed exo-toxin is responsible. Previous studies have dem-onstrated that the lethality after injection ofwashed and unwashed staphylococci does not dif-fer materially and that heat-killed cultures arenontoxic to the embryo (30). Alpha hemolysinis often listed as being one of the more importantof the toxins elaborated by staphylococci becauseof its hemolytic, dermonecrotic, leukocytic, andlethal properties. Only negligible amounts ofa-hemolysin are produced by the cultural tech-niques utilized for the preparation of the infec-tious inoculum, but production in vivo does oc-cur in infected chorioallantoic fluid (31, 40).Recent investigations by Goshi, Cluff, and Nor-man have suggested that the leukocytic action ofa-hemolysin produced in vivo inhibits phagocyto-sis and allows enhanced proliferation of coagulasepositive staphylococci in infections in rabbits (41).A similar effect of a-hemolysin may have beenthe mechanism allowing persistence of large num-bers of coagulase positive staphylococci in theembryo at a time when the microbial populationof coagulase negative staphylococci was progres-sively decreasing. Production of a-hemolysin doesnot entirely explain the lethality of staphylococciin this system, however. Studies, yet to be re-ported, have demonstrated that antitoxin does notsignificantly alter the lethality of the infection al-though it completely inactivates a-hemolysin pro-duced in vivo (32). In addition, strains of staphy-lococci, lethal to embryonated eggs, have beenstudied which fail to produce a-hemolysin eitherin vivo or in vitro (31). A similar situation isfound in human infections where a-hemolysin, al-though produced by most staphylococci isolatedfrom human infections, is not a prerequisite forhuman pathogenicity (42, 43). Other exotoxinsmay be equally or more important than a-hemoly-sin in this experimental system, and attempts arein progress to delineate their role and the exactcause of death of the embryo.

Summary

Studies with embryonated eggs demonstratedthat staphylococci pathogenic for man are ex-tremely lethal to this host. Intra-allantoic infec-tion with less than 100 colony forming units of

strains of Staphylococcus aureus isolated fromclinical infections almost uniformly producedfatality rates of 50% or greater. The virulence ofindividual strains was reproducible over pro-longed periods and tended to parallel their clini-cal behavior. Staphylococci isolated from clinicalinfections were more lethal than strains obtainedfrom healthy carriers, strains isolated from se-vere staphylococcal infections were more lethalthan strains isolated from milder infections, andstrains of Staphylococcus epiderinidis were es-sentially nonlethal for chick embryos. The ageof the embryo also modified lethality, embryos ofless than 10 days of age being more susceptible tofatal infection with both coagulase positive and co-agulase negative staphylococci than older embryos.

Differences in growth between coagulase posi-tive and coagulase negative staphylococci duringinfection were observed but did not explain themarked differences in virulence of the two spe-cies. Both grew equally well in allantoic -fluidin vitro and in vivo and invaded the embryo.Coagulase positive staphylococci persisted at es-sentially the same numbers that were attainedafter maximal growth in allantoic fluid and in theembryo, whereas the numbers of coagulase nega-tive staphylococci progressively decreased afterinitial logarithmic growth until less than 1% re-mained viable 10 days after infection. The mecha-nism of death for the embryo, the role of exotoxins,and explanation of the differences in virulence ofcoagulase positive and coagulase negative staphy-lococci remain to be determined.

Acknowledgment

The technical assistance of Mrs. Lois Swanson isgratefully acknowledged.

References1. Mudd, S. Staphylococcal disease: imperfectly un-

derstood aspects. Arch. environm. Hlth 1962, 4,1.

2. Mudd, S. Practicability of enhancing specific re-sistance to staphylococcic infection. J. Amer. med.Ass. 1960, 173, 1360.

3. Finland, M., W. F. Jones, Jr., and M. W. Barnes.Changes in the occurrence of serious bacterial in-fections since the introduction of antimicrobialagents. Trans. Ass. Amer. Phycns 1959, 72, 305.

4. Rogers, D. E. The current problem of staphylococ-cal infections. Ann. intern. Med. 1956, 45, 748.

2155

WILLIAM R. McCABE

5. Hospital acquired staphylococcal disease. Proceedingsof the National Conference, U. S. Public HealthService, Bureau of State Services, CommunicableDiseases Center, Atlanta, Ga. 1958, 73, 1082.

6. Smith, I. M. Staphylococcal infections. Chicago,Year Book, 1958.

7. Rogers, D. E. Staphylococcal infections. Disease aMonth 1958, 1.

8. Finland, M. Antibiotics for staphylococcal infec-tions. Med. Clin. N. Amer. 1958, 42, 1179.

9. Howe, C. W. Postoperative wound infections dueto Staphylococcus aureus. New Engl. J. Med. 1954,251, 411.

10. Blowers, R., G. A. Mason, K. R. Wallace, and M.Walton. Control of wound infection in a thoracicsurgery unit. Lancet 1955, 2, 786.

11. Elek, S. D. Experimental staphylococcal infectionsin the skin of man. Ann. N. Y. Acad. Sci. 1956,65, 85.

12. Elek, S. D., and P. E. Conen. The virulence ofStaphylococcus pyogenes for man. A study of theproblems of wound infection. Brit. J. exp. Path.1957, 38, 573.

13. Smith, I. M., A. P. Wilson, E. Ch. Hazard, W. K.Hummer, and M. E. Dewey. Death from staphylo-cocci in mice. J. infect. Dis. 1960, 107, 369.

14. James, R. C., and C. J. MacLeod. Induction ofstaphylococcal infections in mice with small inoculaintroduced on sutures. Brit. J. exp. Path. 1961,42, 266.

15. Dold, H. Das Verhalten der wichtigsten aErobenmanchenpathogenen Bakterien in der Haut desMeerschweinchens (Nachweis toxischen Staphylo-kokken-, Streptokokken-, Coli-, und Proteus-Typen durch den Intrakutanversuch. Zbl. Bakt.,I. Abt. Orig. 1927, 102, 1.

16. Smith, W., J. H. Hale, and M. M. Smith. The roleof coagulase in staphylococcal infections. Brit. J.exp. Path. 1947, 28, 57.

17. Panton, P. N., and F. C. 0. Valentine. Staphylococ-cal infection and re-infection. Brit. J. exp. Path.1921, 10, 257.

18. Johnson, J. E. III, L. E. Cluff, and K. Goshi. Stud-ies on the pathogenesis of staphylococcal infec-tion. I. The effect of repeated skin infections.J. exp. Med. 1961, 113, 235.

19. Goshi, K., L. E. Cluff, J. E. Johnson III, and C. R.Conti. Studies on the pathogenesis of staphylo-coccal infection. II. The effect of non-specific in-flammation. J. exp. Med. 1961, 113, 249.

20. Cowan, S. T. Staphylococcal infections in rabbits;antibacterial and non-specific immunity. J. path.Bact. 1939, 48, 545.

21. Flaum, A. Studies in staphylococci and staphylococ-cal immunity. Acta path. microbial. scand. (suppl.35) 1938, 35, 1.

22. Gorrill, R. H. Experimental staphylococcal infec-tions in mice. Brit. J. exp. Path. 1951, 32, 151.

23. Smith, J. M., and R. J. Dubos. The behavior of viru-lent and avirulent staphylococci in the tissues ofnormal mice. J. exp. Med. 1956, 103, 87.

24. Hunt, G. A., and A. J. Moses. Acute infection ofmice with Smith strain of Staphylococcus aureus.Science 1958, 128, 1574.

25. Tompsett, R. Relation of clumping factor producedby staphylococci to their phagocytosis and intra-cellular survival in Antimicrobial Agents andChemotherapy-1961, M. Finland and M. S. Sav-age, Eds. Detroit, American Society of Microbiol-ogy, 1962, p. 67.

26. Koenig, M. G. Factors relating to the virulence ofstaphylococci. I. Comparative studies on twocolonial variants. Yale J. Biol. Med. 1962, 34, 537.

27. Koenig, M. G., M. A. Melly, and D. E. Rogers.Factors relating to the virulence of staphylococci.II. Observations on four mouse-pathogenic strains.J. exp. Med. 1962, 116, 589.

28. Tompsett, R. Role of phagocytosis in immunity toclumping factor-negative staphylococci in Anti-microbial Agents and Chemotherapy-1962, J. C.Sylvester, Ed. Detroit, American Society of Mi-crobiology, 1963, p. 100.

29. Fisher, S. Observations on an antistaphylococcalmouse protective antibody in human sera. Aust.J. exp. Biol. med. Sci. 1961, 39, 413.

30. McCabe, W. R. Lethality of pathogenic and non-pathogenic staphylococci in embryonated eggs.Proc. Soc. exp. Biol. (N. Y.) 1962, 110, 452.

31. McCabe, W. R. Studies of staphylococcal infections.II. Role of exotoxins in infections in embryonatedeggs. In preparation.

32. McCabe, W. R. Studies, of staphylococcal infections.III. Protective effect of staphylococcal antibodyand antitoxin in infections in embryonated eggs.In preparation.

33. Beveridge, W. I. B., and F. M. Burnet. The cultiva-tion of viruses and rickettsiae in the chick embryo.Spec. Rep. Ser., med. Res. Counc. (Lond.) no. 256.

34. Romanoff, A. L. The Avian Embryo, Structural andFunctional Development. New York, Macmillan,1960.

35. Goodpasture, E. W., and K. Anderson. The prob-lem of infection as presented by bacterial invasionof the chorio-allantoic membrane of chick embryos.Amer. J. Path. 1937, 13, 149.

36. Goodpasture, E. W. Use of embryo chick in in-vestigation of certain pathological problems. Sth.med. J. (Bgham, Ala.) 1933, 26, 418.

37. Frappier, A., and S. Sonea. L'infection staphylococ-cique de l'embryon de poulet inocule par voie sous-cutanee. Un. med. Can. 1954, 83, 382.

38. Knothe, H. Das verhalten von aeroben Bakterien inbebrutenen Huhnerei. II. Metterlung. Das Ver-halten von Staphylokokken, Meningokokken, Strep-tokokken, Pneumokokken, Bacterium tularense,Pasteurella pestis, und Brucella abortus Bang im,

2156

EXPERIMENTALSTAPHYLOCOCCALINFECTIONS IN EMBRYONATEDEGGS

bebrilteten Huhnerei. Zbl. Bakt., I Abt. Orig.1952, 158, 453.

39. Wiley, B. B. A new virulence test for Staphylococ-cus aureus and its application to encapsulatedstrains. Canad. J. Microbiol. 1961, 7, 933.

40. Jones, M. The production of staphylococcus toxinin the chorio-allantoic fluid of the embryonatedegg. J. Bact. 1946, 51, 789.

41. Goshi, K., L. E. Cluff, and P. S. Norman. Studies on

the pathogenesis of staphylococcal infection. VI.

Mechanism of immunity conferred by anti-alphahemolysin. Bull. Johns Hopk. Hosp. 1963, 112,31.

42. Lack, C. H., and D. G. Wailling. A study of 435strains of Staphylococcus pyogenes with referenceto factors which may contribute to pathogenicity.J. Path. Bact. 1954, 68, 431.

43. Elek, S. D., and E. Levy. Distribution of haemoly-sins in pathogenic and non-pathogenic staphylo-cocci. J. Path. Bact. 1950, 62, 541.

2157