Journal of Clinical Investigation Vol. 43, No. 11, 1964

Studies of Staphylococcal . I. of Staphy- lococci and Characteristics of Infections in Embryonated Eggs * WILLIAM R. MCCABE t (From the Laboratory, West Side Veterans Administration Hospital, and the Department of , Research and Educational Hospitals, University of Illinois College of Medicine, Chicago, Ill.)

Many of the determinants of the pathogenesis niques still require relatively large numbers of and course of staphylococcal infections remain staphylococci to produce (19). Fatal imprecisely defined (1, 2) despite their increas- systemic infections have been equally difficult to ing importance (3-10). Experimental infections produce in animals and have necessitated the in- in suitable laboratory animals have been of con- jection of 107 to 109 (20-23). A few siderable assistance in clarifying the role of strains of staphylococci have been found that are defense mechanisms and specific bacterial virulence capable of producing lethal systemic infections factors with a variety of other infectious agents. with inocula of from 102 to 103 bacteria (24) and A sensitive experimental model would be of value have excited considerable interest (25-27). The in defining the importance of these factors in virulence of these strains apparently results from staphylococcal infections, but both humans and an unusual antigenic variation (27, 28) which, the usual laboratory animals are relatively re- despite its interest, is of doubtful significance in sistant. Extremely large numbers of staphylo- human staphylococcal infection, since such strains cocci are required to produce either local or sys- have been isolated only rarely from clinical in- temic infections experimentally. Elek was un- fections (29). The insensitivity of these experi- able to produce local purulent lesions in humans mental systems has been a handicap for compari- with inocula of less than 106 staphylococci (11) son of the relative virulence of strains of staphylo- but subsequently demonstrated that the infectious cocci and evaluation of the effect of antibody and inoculum could be reduced by two to four logs other defense mechanisms in a living system. when the infection was combined with concomitant Preliminary studies of staphylococcal infections insertion of a (12). Even larger in a variety of animals indicated that embryonated numbers of staphylococci have been required to eggs were readily infected with pathogenic staphy- produce local lesions in mice (13, 14), guinea lococci (30). The present study describes the pigs (15, 16), and rabbits (17, 18). Various sensitivity of embryonated eggs to 68 strains of traumatic procedures or implantation of foreign positive aureus and 36 bodies also allows the production of infection with strains of coagulase negative Staphylococcus epi- smaller numbers of bacteria in animals (14, 19), dermidis and defines some of the characteristics but despite such measures many of these tech- of this infection. Subsequent publications will de- * Submitted for publication May 4, 1964; accepted scribe the importance of individual bacterial viru- July 9, 1964. lence factors (31) and the protective effect of anti- Reported in part at the joint sessions of the American toxic and antibacterial antibody in the chick em- Society for Clinical Investigation and the American bryo (32). Federation for , April 28, 1963, Atlantic City, N. J. Methods This work was completed during the tenure of a Veterans Administration Clinical Investigatorship. Strains of staphylococci. Forty-two strains of coagu- t Present address: Evans Memorial Department of lase positive were obtained from Clinical Research, Massachusetts Memorial Hospitals, blood cultures and specimens of sputum or purulent and the Department of Medicine, Boston University exudates submitted to the routine bacteriology labora- School of Medicine, Boston University Medical Center. tory. In each instance, the patient from whom the 2146 EXPERIMENTAL STAPHYLOCOCCAL INFECTIONS IN EMBRYONATED EGGS 2147 specimen was obtained was examined to insure that the When recoveries were carried out after injection of a staphylococcus isolated was etiologically related to the large inoculum, 12-day-old eggs in which the volume of clinical infection. Additional strains were obtained allantoic fluid is stated to be 6 ml (33, 34) were used from the anterior nares and of a group of patients for injection. One-tenth ml samples of allantoic fluid without clinical evidence of infection who were attend- were aspirated aseptically after removal of the shell ing an outpatient clinic for the first time. Specimens over the egg sac, plated, the colonies enumerated, and were obtained only from outpatients who had not re- the numbers corrected for the 1 to 60 dilution. Re- ceived within the past 3 months, had not covery with the smaller inoculum was performed simi- been hospitalized during this period, and whose household larly, except as much of the allantoic fluid as possible contacts had not been hospitalized recently. A third was aspirated and 4 ml of PS was injected, mixed, and group of 36 strains of coagulase negative Staphylococ- removed. The latter procedure was repeated nine times, ctus epidermidis was obtained from a variety of speci- and the pooled allantoic fluid and saline washings were mens submitted to the hospital laboratory and from the filtered through a Millipore H.A. filter (0.45 Ah). After nares and skin of outpatients. filtration, the filter was removed, placed on a blood Tube and slide coagulase determinations and mannitol agar plate, incubated for 48 hours, and the number of fermentation were performed on all isolated strains to bacterial colonies determined. insure their identification. Strains of staphylococci were growth curves. Allantoic fluid was collected maintained on tryptose phosphate (TP) agar slants aseptically from 15 eggs, pooled, and 4.9 ml dispensed at 0° C after isolation and identification until egg viru- into sterile capped tubes for determination of growth lence studies were performed the following week. curves. One-tenth ml of a 10' dilution of the organism Immediately before use, a single colony selected from to be tested was added and the allantoic fluid incubated a subculture of the original agar slant was inoculated for 18 hours. Duplicate determinations were made at into TP broth for the infectious inoculum. Several 0.1- each sampling time. Concomitant growth curves in TP ml samples of the latter were transferred to 2.5 ml of broth were performed in a similar fashion. TP broth in screw-capped vials and stored in a - 200 C growth curves. Lots of 64 10-day-old embry- deep freeze for subsequent use. onated eggs were injected with 0.1 ml of a 10' and 10' Preparation and injection of inocula. Serial tenfold dilution of coagulase positive staphylococci and 0.1 ml of dilutions of overnight TP broth cultures of staphylococci a 10' dilution of coagulase negative staphylococci. were made to provide inocula of approximately 1,000 to Four eggs from each of the three groups were sacrificed 99,999 (10' to 10') and 1 to 100 colony forming units at 4, 8, 18, and 24 hours and daily thereafter for 10 days. (cfu). The smaller inocula consisted of less than 10 Only eggs containing living embryos were sacrificed to cfu in over one-half the studies. The size of the inoculum insure that acceleration of growth and invasion of the was determined by plating 0.1 ml of 10-', 10', and 10- embryo "post-mortem" did not obscure the results. dilutions of the overnight culture on TP agar and enu- The shell over the air sac was removed; allantoic fluid merating the number of colonies after overnight incu- was aspirated aseptically and serially diluted for plate bation. Lots of 20 10-day-old fertile eggs were injected counts. The remainder of the allantoic fluid was re- intra-allantoically with 0.1 ml of 10-s and 10' dilutions moved, and 1.5% saponified cresylic acid' was added to of each strain studied, and the eggs were candied daily the shell containing the intact yolk sac, amnion, and to determine viability. A group of ten controls was in- embryo. After 10 minutes, the cresylic acid solution jected with a similar volume of sterile 0.85% saline and the yolk were removed. The amnion and embryo (PS) in each . Subcultures were obtained were extracted with sterile forceps and placed in a from allantoic fluid, yolk, and embryonic blood in the sterile petri dish. The amnion was removed and the first sets of and, when indicated, in sub- embryo washed in the cresylic acid solution and rinsed sequent studies to insure that maternal of with sterile PS. The embryo was opened by a ventral infection or subsequent contamination had not occurred. incision, and the heart and liver were removed asepti- Eggs. Fertilized eggs were obtained from a com- cally. These were ground in a glass tube with a Teflon mercial source that did not use an dietary tissue grinder, and 0.2 ml of the bloody supernatant supplement. Tube dilution assays for antibacterial ac- fluid was removed. Bacterial titers were determined by tivity were performed on allantoic and amniotic fluid on plating 0.1 ml of undiluted fluid and serial tenfold several occasion with negative results. Eggs were incu- dilutions of the fluid. bated for 10 davs at 370 C in a rocker incubator before infection and- then incubated upright at 38.5° C with Results 60% relative humidity for an additional 10 days after infection. Lethality of coagulase positive and coagulase Recovery of inoculum. Studies to determine the num- negative staphylococci. The 10-day fatality rates ber of cfu that could be recovered immediately after in- for 10-day-old eggs injected intra-allantoically with jection were carried out by two techniques. Eggs were inocula of two different sizes of 68 strains of co- injected as described above, allowed to stand for 15 min- utes, sealed with paraffin, shaken gently, and sacrificed. 1 Cresylone, Parke, Davis & Co., Detroit, Mich. 2148 WILLIAM R. McCABE GOAGULASE POSITIVE STAPHYLOCOCCI COAGULASE v Isolated from Nores' NEGATIVE Isolated from Infections of Outpatients STAPHYLOCOCCI (Nx42 ) (Nu26) (N= 36) INOCULUm. 105- DACTERIA 100 mn AI1 l I 75 - 1%- 50-

25 -

N.4 INOCULUM. I- A0OACTERIA 100 -4 a .1. II1. kz 75 h I 50 - 25 mor I Emus MDONE' m .1 .... I..IR- 5 10 15 20 25 30 35 42 5 10 IS 20 26 5 10 I5 20 25 30 36 FIG. 1. CHICK EMBRYO LETHALITY OF COAGULASE POSITIVE STAPHYLOCOCCI ISOLATED FROM INFECTIONS COMPARED WITH THAT OF STRAINS ISOLATED FROM NASAL CARRIERS AND COAGULASE NEGATIVE STAPHYLOCOCCI.

agulase positive Staphylococcus aureus and 36 Pathogenic staphylococci produced fatality rates strains of coagulase negative Staphylococcus epi- ranging from 50 to 100%, mean 89.2 14.9, dermidis are illustrated in Figure 1. Two groups when 103 or 104 cfu was injected, and from 40 to of coagulase positive staphylococci were studied. 1007%, mean 75.8 18.5, with inocula of less than Forty-two strains were isolated from unequivocal 100 cfu. Fatality rates of 509o% or greater oc- human infections in which they were the sole curred with all strains isolated from human in- etiologic agent of the infections. These strains fections with the larger inocula, and with 40 of are referred to, henceforth, as pathogenic staphylo- 42 of these strains with the smaller inocula. The cocci. A second group of 26 strains isolated from lethality of the 26 strains of coagulase positive the nares and skin of uninfected outpatients rep- staphylococci isolated from outpatients was be- resents staphylococci of uncertain pathogenicity. tween 15 and 100%, mean 72.8 26.9, with the The latter group, whose over-all human patho- large inocula and from 5 to 100%o, mean 60.0 genicity, although uncertain, would be anticipated 29.3, with the small inocula. Both the large and to be. less than that of strains isolated from definite small inocula of pathogenic strains were signifi- human infections, was selected to sample strains cantly more lethal (tat = 4.5, p < 0.001; t66 = characteristic of those prevalent in the community. 3.95, p < 0.001) than were equivalent inocula of The third group, coagulase negative Staphylococ- strains isolated from outpatients. Although strains cus epidermidis, represented strains that are rarely that had produced clinical infections were more human . virulent for embryonated eggs in general than EXPERIMENTAL STAPHYLOCOCCAL INFECTIONS IN EMBRYONATED EGGS 2149

N~bmm BACTEREMIA Mild to Moderate Severe BACTEREIAT WOUND INFECTIONS IOSTEOMYELITIS FIG. 2. LETHALITY OF STRAINS OF STAPHYLOCOCCUS AUREUS ISOLATED FROM INFECTIONS OF VARIOUS TYPES AND DEGREES OF SEVERITY. strains isolated from uninfected patients, consider- were differentiated by the following criteria: oc- able overlap in egg virulence was observed with currence in patients without serious underlying individual strains of both types. , extensive inflammatory reaction, leuko- Coagulase negative strains were significantly cytosis greater than 17,000 per mm3, and less lethal than coagulase positive staphylococci in excess of 1020 F for 5 days or longer. All isolated either from septic patients or uninfected other wound infections were classified as being of outpatients. Fatality rates did not exceed 20%o mild to moderate severity when two or fewer of either with inocula of 1 to 100 or 108 to 104 cfu, the above criteria were present. The latter in- and the mean fatality rate with the larger inocula, fections usually consisted of stitch and 11.0 + 6.1%, did not differ significantly (t66 = infections producing moderate purulent drainage 1.65, p = 0.1) from the fatality rate with the and wound induration with only slight systemic smaller inocula, 8.3 + 5.8%. The mean fatality symptoms. No further attempt was made to rate produced by 1 to 100 cfu of coagulase positive classify the severity of infections except in these staphylococci isolated from outpatients was sig- five instances. Staphylococcal pneumonia was nificantly greater (t86 = 10.1, p < 0.001) than associated with bacteremia in two patients; these that produced by 100 to 1,000 times as many co- strains are shown in both categories. Three pa- agulase negative staphylococci. tients with primary hematogenous Lethality of strains isolated from infections of were observed, and their strains are grouped varying severity. Figure 2 demonstrates the fa- with strains isolated from patients with bactere- tality rates observed in embryonated eggs after mia. The mean fatality rate from infections with injection of 1 to 100 cfu of strains of pathogenic strains causing mild to moderate wound infections staphylococci isolated from different types of was 62.6 + 17.0% and was significantly less clinical infections. Strains isolated from patients (t21 = 3.5, p < 0.01; t28 = 3.6, p < 0.01; t28= with staphylococcal bacteremia, pneumonia, and 4.7, p < 0.001) than the fatality rate of infections wound infections were studied. Five patients had with strains isolated either from patients with particularly extensive and severe wound infections, severe wound infections (87%), pneumonia and the strains isolated from these patients are (86%), or bacteremia (87%). Even when strains distinguished from strains causing less severe in- associated with severe wound infections were fections in Figure 2. Severe wound infections combined with those that produced mild to mod- 2150 WILLIAM R. McCABE

FATALITIES caw#FECT/ os 105_104 A) % )POSfT07V /-/00 *- 100 _ STAP ou TPATIENTs-1-04 A---A ifax /-/00 A---A COAG. NEGATVE STA. i0o-io ZoZ ft1#/-/00 @-. SALNE CONTROLS x- x

25

DAYS AFTER INFECTION FIG. 3. CUMULATIVE FATALITY RATES AFTER ALLANTOIC INFECTION WITH STRAINS OF STAPHYLOCOCCUS AUREUS AND STAPHYLOCOCCUS EPI- DERMIDIS. crate infections, their over-all lethality was sig- rate increased progressively until the final day of nificantly less than that of strains isolated from , with 60%o of the total fatalities oc- pneumonia and bacteremia. Lethality did not curring after the second day. A similar but less differ materially, however, for strains isolated striking relation between the size of the inoculum from patients with severe wound infections, pneu- and rapidity of fatalities also occurred with strains monia, or bacteremia. isolated from uninfected outpatients. The effect Size of inoculum and rapidity of death. The of the size of the inoculum on the time of death increase in fatality attained with a 1,000-fold in- was even more apparent when individual strains crease in inoculum size over that observed with were considered. More virulent strains that pro- a small inoculum often was not so great as might duced fatality rates of 85% or greater with a have been anticipated. This finding appeared to small inocula produced 84% of their fatalities in reflect resistance of a small proportion of the eggs the first 2 days after infection with 103 bacteria, to fatal infection in addition to differences in the whereas less lethal strains produced only 47% of rate of killing by inocula of 100 and 103 to 104 their fatalities within this time (X2 = 7.3, p < staphylococci during the period of observation. 0.01). The effect of the latter may be seen in part in the Lethality of coagulase negative staphylococci cumulative daily death rates with the two sizes of and culture media. Since coagulase negative inocula of the three groups of staphylococci il- staphylococci produced very few fatalities in the lustrated in Figure 3. Larger inocula of patho- comparative studies, their virulence was assessed genic staphylococci were more rapidly fatal and more critically by comparing the lethality of five produced 70% of the total fatalities within the strains with that of the culture media. Tenfold first 2 days after infection, and only a small pro- dilutions of overnight TP broth cultures were portion of deaths occurred later in the 10-day pe- made in sterile PS, and 0.1 ml of undiluted cul- riod. With smaller inocula, there was an initial tures and 10-1, 10-2, 10-3 dilutions were injected lag before deaths were apparent, and the death intra-allantoically into ten eggs per dilution with EXPERIMENTAL STAPHYLOCOCCAL INFECTIONS IN EMBRYONATED EGGS 2151

INOCULUIMV. /-/OX /OS SACTERIA

.44 1 N.1 k

K4

0 2 3 4 5 6 7 MONTHS FIG. 4. OF FATALITY RATES IN EMBRYONATED EGGS OF FOUR STRAINS OF STAPHYLOCOCCUS AUREUS STUDIED OVER A 7-MONTH PERIOD. particular care being taken to minimize trauma to many instances. To insure that a large proportion the eggs. Sterile TP broth was diluted and in- of the injected organisms was not lost, recoveries jected in a similar fashion, and the number of were carried out with two different sizes of in- deaths was determined daily for 10 days for all oculum as described previously. The results of five groups. The inoculum varied from 1 x 107 four such recoveries are shown in Table II. In to 2 x 108 cfu with undiluted cultures. The num- each instance, recoveries approximated 60% of ber of survivors in each group is shown in Table the bacteria injected and *were well within the I. No significant differences were observed range of error of the experimental method. among any of the groups, and undiluted overnight Reproducibility of fatality rates. The virulence cultures of coagulase negative staphylococci were of individual strains was reproducible over a pe- no more lethal than the media in which they were riod of several months. Serial assessments of grown. the lethality of three virulent strains and one Recovery of injected staphylococci. Some leak- nonvirulent strain, shown in Figure 4, were made age occurred after the injection of bacteria in at intervals of as long as 7 months. Fatality

II TABLE I TABLE Comparative lethality of coagulase negative staphylococci Recovery of injected bacteria from allantoic fluid and sterile culture media No. of bacteria Per cent Strains of coagulase negative staphylococci Injected Recovered recovered Sterile Dilution broth F R B K S Cfu* cfu 5% Undiluted 9/10* 9/10 8/10 10/10 9/10 8/10 3.1 X 104 1.8 X 104 58 10-1 10/10 10/10 10/10 9/10 10/10 10/10 4.5 X 104 3.0 X 104 67 10- 9/10 10/10 10/10 10/10 10/10 9/10 7 4 59 10-5 10/10 10/10 9/10 9/10 10/10 10/10 7 4 59

* No. of survivors at 10 days/no. of eggs injected. * cfu = colony forming units. 212N2ILLS LIAMt R. McCABE

INOCULUM . /-/Ox /O3 BACTERIA * COAG. POSIr/VE STAPH. * COAG. NEGAT/VE STAPH.

k 0 0|B .14

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

75 ~.11K 50

25 00L 7 days 8 days 10 days 13 days 15 days AGE OF EMBRYO AT INFECTION FIG. 5. SUSCEPTIBII ITY OF EMBRYONATED EGGS OF V-ARIOt-S AG;ES TO LETHAL INFECTION WX ITH STAPHYLOCOCCUS AI RElS AND STA.INYloCOCCv5s EPIDERM IS. rates did not vary more thani 10%/. even w\heni lent at all ages, although the differences were studies were iatlde after proloigedi intervals. A most evident in embryos 10 days of age or older. numl)er of other straills not shown in this Figutre (;GroZL'tli of staplilococci in allan toxic fluid in have l)eei stuldieci On sev-eral ccaISions With vitro. ( _-ro-wth ctirves in allantoic fluid inl Vitro equally reproulucible results. \were determined for three strains of coagulase SelSCeptibility of ciil)r\vos of different agts. negative sta)hylococci and three strains of coagu- The age of the embryo exerted a significant in- lase positive staphylococci and are compared with fluence on its susceptibility to lethal infection with oroxvth curves of each in TP broth in Figure 6. both coagulase positive and negative staphylococci. Multiplication of coagulase Ilegative staphylococci The five-dav fatality rates produced by three in allantoic fluid lagged behind that observed in strains of coagulase positive and by three strains broth and was less than that of coagulase positive of coagtilase negative staphylococci for 7- 8- 10- strains in either type of mieditim during the first 13-, and 15-day-old chick embryos are shown in 6 hours. This difference failed to persist, how- Figure 5. Seven- aid, to a lesser degree, 8-day- ever, and both types attained similar population old embryos were susceptible to lethal infections densities of 10S bacteria per ml by 18 hours in both wvith coagulase negative strains but became re- nedia. sistalnt b1 the tenth day. The age of the embryo (;roowtl of staplivlococci in vivo. In zvivo growth odiflied the virulence of coagulase positive strains curves also were determined in both allantoic in a similar fashion. Almost uniform fatality oc- fluid and the embryo after injection of both a curred in 7- or 8-day old embryos; 13- and 15-day- large (2 x 103 cfu) anId small (2 cfu) inoculum of old embryos were less susceptible. The greater coagulase positive staphylococci, and a large in- suscel)tibilitv of the younger emnbryos did not ocultim (9 X 103 cfu) of coagulase negative obscure the differences in virulence between co- staphylococci. These growth curves plus the cum- agulase positive and coagulase negative strains. ulative dlailx fatality rates of the strains studied The former strains were significantly more virmi- are illustrated in Figlure 7. Points on the graph EXPERIMENTAL STAPHYLOCOCCAL INFECTIONS IN EMBRYONATED EGGS22153 represent the mean of counts obtained from four eggs in each of three experiments. Growth curves of the large inoculum of coagulase posi- tive and coagulase negative staphylococci were quite 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 with 103 cfu of pathogenic staphylococci precluded as- sessment of growth beyond this period, although this measurement was attempted with lots of 64 eggs on three separate occasions. After the in- jection of a small number of coagulase positive staphylococci (2 cfu), 48 hours elapsed before FATALITY RATE FCUMAULATIVE their numbers reached those attained by the larger inoculum, but after this period the micro- 75 ------50 . bial population persisted essentially unchanged in 2F7-8 16 24 2 3 4 5 6 both allantoic fluid per and the (108 ml) embryo 0 162 234 67 8 9 IC (103 to 104) for the remainder of the study. In ',- HOURS -' DAYS contrast, after an initial rise, a progressive de- FIG. 7. GROWTH CURVES AND CUMULATIVE FATALITY crease in the number of coagulase negative staphy- RATES OF STAPHYLOCOCCUS AUREUS AND STAPHYLOCOCCUS lococci was observed in both the embryo and EPIDERMIDIS IN ALLANTOIC FLUID AND EMBRYONIC TISSUE allantoic fluid until the number of coagulase nega- IN vivo. Each point on the growth curves represents the tive staphylococci present on day 3 had de- mean of counts obtained from four eggs in each of three creased by 99.7% in the allantoic fluid, and by experiments. 99.2% in the embryo by the tenth day. All three experiments gave similar results, with coagulase progressive decrease in numbers of coagulase positive staphylococci attaining similar numbers negative staphylococci after days 3 and 4 also in 48 hours and persisting at this level, while the occurred in each study. Invasion of the embryo was observed only when bacterial counts in the allantoic fluid ex- 8 ceeded 107 cfu per ml. Involvement of the embryo appeared to be related to increasing fatalities with 7 coagulase positive staphylococci, but coagulase negative staphylococci were capable also of in- 6 the embryo without \ 6 vading producing fatalities. Discussion t3 These investigations were undertaken in an attempt to find a more sensitive experimental host - Br ,COAGM/ASEX ~ ~ ~~~ * @othC-A for the study of staphylococcal infections. Embry- 42 f-SrfiWAfMj** Fluid COAG NFGAT/-Vj-fm--- Broth onated eggs seem to possess many features that -- C-A Fluid would be desirable in such an experimental model. They are extremely susceptible to infection with small numbers of pathogenic staphylococci, and O 2 4 6 8 10 12 14 16 18 INCUBATION TIME, HOURS the lethality of individual strains is consistent and a several The FIG. 6. GROWTH CURVES OF STAPHYLOCOCCUS AUREUS reproducible over period of months. AND STAPHYLOCOCCUS EPIDERMIDIS IN ALLANTOIC PLUID sensitivity to infection also may be modified by AND BROTH MEDIA. utilizing the differences in susceptibility of em- 2154 WILLIAM R. McCABE bryos of various ages if greater or lesser lethality Embryonated eggs have been used predomi- is desired. In addition, a distinct relationship be- nantly for the propagation and study of tween the pathogenicity of staphylococci for hu- and rickettsiae, although Goodpasture recognized mans and their virulence for chick embryos was their potential value for investigation of bacterial observed, confirming earlier preliminary findings infections (35) soon after he popularized their (30). Strains of Staphylococcus aureus definitely use for cultivation of viruses (36). One of his responsible for human infections were extremely early studies with this model described the lesions lethal and almost uniformly produced fatality produced by infection of allantoic membranes and rates of 50% or greater when less than 100, and the ability of a variety of bacteria, including usually less than 10 cfu, were injected. Other staphylococci, to invade the membrane and pro- strains of indeterminant potential pathogenicity duce lethal infections (35). Subsequently, Frap- to man, isolated from uninfected carriers, ex- pier and Sonea observed that staphylococci in- hibited, individually, variable degrees of virulence jected subcutaneously into 18- or 19-day-old em- for the chick embryo but as a group were sig- bryos produced lethal infections (37), and nificantly less lethal than strains isolated from Knothe studied a variety of allantoic infections clinical infection. Fatality rates from infection including those produced by staphylococci (38). with strains that are avirulent or noninvasive for The extreme sensitivity of chick embryos to humans, coagulase negative Staphylococcus epi- staphylococci was noted in both of these investi- dermidis, were no greater than those occurring gations and more recently by Wiley, who deter- after injection of culture media similar to that mined the lethality of four strains of Staphylo- used for their propagation. Further correlation coccus aureus in embryonated eggs (39). How- between the severity of the clinical infections and ever, the limited number of strains evaluated and the virulence of individual strains for chick em- the failure to further extend these bryos was also observed. Staphylococci isolated has hampered recognition of the potential value- from patients with bacteremia or pneumonia were of this model. more lethal than strains isolated from wound in- Explanation of the marked differences in viru- fections, and strains isolated from more severe lence of coagulase positive and coagulase nega- wound infections produced greater fatality rates tive staphylococci for embryonated eggs is not than strains isolated from less severe wound readily apparent from these studies. Both types infections. of organisms grew equally well in allantoic fluid The excellent correlation observed between in vitro and initially in vivo, and both invaded the pathogenicity and virulence of staphylococci the embryo with equal facility. but subsequently for humans and their lethality in chick embryos their behavior differed. Pathogenic staphylo- does not imply, however, that this system can be cocci persisted at essentially the same microbial utilized to predict the pathogenicity of staphylo- population attained after logarithmic growth in cocci for humans. The parameters used to char- both allantoic fluid and the embryo throughout acterize clinical behavior of staphylococci are the entire infection. In contrast, although coagu- relatively imprecise, and the potential patho- lase negative staphylococci reached a population genicity of individual staphylococci cannot be pre- similar in magnitude to that of pathogenic staphy- dicted accurately by any currently available tech- lococci, they were nonlethal, and a progressive- niques. Precise quantification and assignment of decrease in their numbers occurred after the- relative degrees of virulence will require thorough third day until only 0.3% of the maximal popula- epidemiologic investigation of the frequency and tion attained in the allantoic fluid and only 0.8% types of human infection in relation to contact attained in the embryo remained by the tenth day and acquisition of individual strains as well as after infection. A variety of and exo- concomitant appraisal of predisposing host factors. has been purported to be responsible for Until such information is available, definitive cor- the virulence of coagulase positive staphylococci relation between the similarity of behavior of and may play an important role in infections staphylococci in humans and animals will not be caused by these organisms in the chick embryo. possible. If do exert a significant influence on EXPERIMENTAL STAPHYLOCOCCAL INFECTIONS IN EMBRYONATED EGGS 2155 the course of the infection, production of strains of Staphylococcus aureus isolated from in vivo rather than injection of preformed exo- clinical infections almost uniformly produced is responsible. Previous studies have dem- fatality rates of 50% or greater. The virulence of onstrated that the lethality after injection of individual strains was reproducible over pro- washed and unwashed staphylococci does not dif- longed periods and tended to parallel their clini- fer materially and that heat-killed cultures are cal behavior. Staphylococci isolated from clinical nontoxic to the embryo (30). Alpha hemolysin infections were more lethal than strains obtained is often listed as being one of the more important from healthy carriers, strains isolated from se- of the toxins elaborated by staphylococci because vere staphylococcal infections were more lethal of its hemolytic, dermonecrotic, leukocytic, and than strains isolated from milder infections, and lethal properties. Only negligible amounts of strains of Staphylococcus epiderinidis were es- a-hemolysin are produced by the cultural tech- sentially nonlethal for chick embryos. The age niques utilized for the preparation of the infec- of the embryo also modified lethality, embryos of tious inoculum, but production in vivo does oc- less than 10 days of age being more susceptible to cur in infected chorioallantoic fluid (31, 40). fatal infection with both coagulase positive and co- Recent investigations by Goshi, Cluff, and Nor- agulase negative staphylococci than older embryos. man have suggested that the leukocytic action of Differences in growth between coagulase posi- a-hemolysin produced in vivo inhibits phagocyto- tive and coagulase negative staphylococci during sis and allows enhanced proliferation of coagulase infection were observed but did not explain the positive staphylococci in infections in rabbits (41). marked differences in virulence of the two spe- A similar effect of a-hemolysin may have been cies. Both grew equally well in allantoic-fluid the mechanism allowing persistence of large num- in vitro and in vivo and invaded the embryo. bers of coagulase positive staphylococci in the Coagulase positive staphylococci persisted at es- embryo at a time when the microbial population sentially the same numbers that were attained of coagulase negative staphylococci was progres- after maximal growth in allantoic fluid and in the sively decreasing. Production of a-hemolysin does embryo, whereas the numbers of coagulase nega- not entirely explain the lethality of staphylococci tive staphylococci progressively decreased after in this system, however. Studies, yet to be re- initial logarithmic growth until less than 1% re- ported, have demonstrated that antitoxin does not mained viable 10 days after infection. The mecha- significantly alter the lethality of the infection al- nism of death for the embryo, the role of exotoxins, though it completely inactivates a-hemolysin pro- and explanation of the differences in virulence of duced in vivo (32). In addition, strains of staphy- coagulase positive and coagulase negative staphy- lococci, lethal to embryonated eggs, have been lococci remain to be determined. studied which fail to produce a-hemolysin either in vivo or in vitro (31). A similar situation is Acknowledgment found in human infections where a-hemolysin, al- The technical assistance of Mrs. Lois Swanson is though produced by most staphylococci isolated gratefully acknowledged. from human infections, is not a prerequisite for human pathogenicity (42, 43). Other exotoxins References may be equally or more important than a-hemoly- 1. Mudd, S. Staphylococcal disease: imperfectly un- sin in this experimental system, and attempts are derstood aspects. Arch. environm. Hlth 1962, 4, in progress to delineate their role and the exact 1. cause of death of the embryo. 2. Mudd, S. Practicability of enhancing specific re- sistance to staphylococcic infection. J. Amer. med. Ass. 1960, 173, 1360. Summary 3. Finland, M., W. F. Jones, Jr., and M. W. Barnes. Changes in the occurrence of serious bacterial in- Studies with embryonated eggs demonstrated fections since the introduction of antimicrobial that staphylococci pathogenic for man are ex- agents. Trans. Ass. Amer. Phycns 1959, 72, 305. tremely lethal to this host. Intra-allantoic infec- 4. Rogers, D. E. The current problem of staphylococ- tion with less than 100 colony forming units of cal infections. Ann. intern. Med. 1956, 45, 748. 2156 WILLIAM R. McCABE

5. Hospital acquired staphylococcal disease. Proceedings 23. Smith, J. M., and R. J. Dubos. The behavior of viru- of the National Conference, U. S. Public Health lent and avirulent staphylococci in the tissues of Service, Bureau of State Services, Communicable normal mice. J. exp. Med. 1956, 103, 87. Center, Atlanta, Ga. 1958, 73, 1082. 24. Hunt, G. A., and A. J. Moses. Acute infection of 6. Smith, I. M. Staphylococcal infections. Chicago, mice with Smith strain of Staphylococcus aureus. Year Book, 1958. 1958, 128, 1574. 7. Rogers, D. E. Staphylococcal infections. Disease a 25. Tompsett, R. Relation of clumping factor produced Month 1958, 1. by staphylococci to their and intra- 8. Finland, M. Antibiotics for staphylococcal infec- cellular survival in Antimicrobial Agents and tions. Med. Clin. N. Amer. 1958, 42, 1179. Chemotherapy-1961, M. Finland and M. S. Sav- 9. Howe, C. W. Postoperative wound infections due age, Eds. Detroit, American Society of Microbiol- to Staphylococcus aureus. New Engl. J. Med. 1954, ogy, 1962, p. 67. 251, 411. 26. Koenig, M. G. Factors relating to the virulence of 10. Blowers, R., G. A. Mason, K. R. Wallace, and M. staphylococci. I. Comparative studies on two Walton. Control of wound infection in a thoracic colonial variants. Yale J. Biol. Med. 1962, 34, 537. surgery unit. Lancet 1955, 2, 786. 27. Koenig, M. G., M. A. Melly, and D. E. Rogers. 11. Elek, S. D. Experimental staphylococcal infections Factors relating to the virulence of staphylococci. in the skin of man. Ann. N. Y. Acad. Sci. 1956, II. Observations on four mouse-pathogenic strains. 65, 85. J. exp. Med. 1962, 116, 589. 12. Elek, S. D., and P. E. Conen. The virulence of 28. Tompsett, R. Role of phagocytosis in immunity to Staphylococcus pyogenes for man. A study of the clumping factor-negative staphylococci in Anti- problems of wound infection. Brit. J. exp. Path. microbial Agents and Chemotherapy-1962, J. C. 1957, 38, 573. Sylvester, Ed. Detroit, American Society of Mi- 13. Smith, I. M., A. P. Wilson, E. Ch. Hazard, W. K. crobiology, 1963, p. 100. Hummer, and M. E. Dewey. Death from staphylo- 29. Fisher, S. Observations on an antistaphylococcal cocci in mice. J. infect. Dis. 1960, 107, 369. mouse protective antibody in human sera. Aust. 14. James, R. C., and C. J. MacLeod. Induction of J. exp. Biol. med. Sci. 1961, 39, 413. staphylococcal infections in mice with small inocula 30. McCabe, W. R. Lethality of pathogenic and non- introduced on sutures. Brit. J. exp. Path. 1961, pathogenic staphylococci in embryonated eggs. 42, 266. Proc. Soc. exp. Biol. (N. Y.) 1962, 110, 452. 15. Dold, H. Das Verhalten der wichtigsten aEroben 31. McCabe, W. R. Studies of staphylococcal infections. manchenpathogenen Bakterien in der Haut des II. Role of exotoxins in infections in embryonated Meerschweinchens (Nachweis toxischen Staphylo- eggs. In preparation. kokken-, Streptokokken-, Coli-, und Proteus- 32. McCabe, W. R. Studies, of staphylococcal infections. Typen durch den Intrakutanversuch. Zbl. Bakt., III. Protective effect of staphylococcal antibody I. Abt. Orig. 1927, 102, 1. and antitoxin in infections in embryonated eggs. 16. Smith, W., J. H. Hale, and M. M. Smith. The role In preparation. of coagulase in staphylococcal infections. Brit. J. 33. Beveridge, W. I. B., and F. M. Burnet. The cultiva- exp. Path. 1947, 28, 57. tion of viruses and rickettsiae in the chick embryo. 17. Panton, P. N., and F. C. 0. Valentine. Staphylococ- Spec. Rep. Ser., med. Res. Counc. (Lond.) no. 256. cal infection and re-infection. Brit. J. exp. Path. 34. Romanoff, A. L. The Avian Embryo, Structural and 1921, 10, 257. Functional Development. New York, Macmillan, 18. Johnson, J. E. III, L. E. Cluff, and K. Goshi. Stud- 1960. ies on the pathogenesis of staphylococcal infec- 35. Goodpasture, E. W., and K. Anderson. The prob- tion. I. The effect of repeated skin infections. lem of infection as presented by bacterial invasion J. exp. Med. 1961, 113, 235. of the chorio-allantoic membrane of chick embryos. 19. Goshi, K., L. E. Cluff, J. E. Johnson III, and C. R. Amer. J. Path. 1937, 13, 149. Conti. Studies on the pathogenesis of staphylo- 36. Goodpasture, E. W. Use of embryo chick in in- coccal infection. II. The effect of non-specific in- vestigation of certain pathological problems. Sth. flammation. J. exp. Med. 1961, 113, 249. med. J. (Bgham, Ala.) 1933, 26, 418. 20. Cowan, S. T. Staphylococcal infections in rabbits; 37. Frappier, A., and S. Sonea. L'infection staphylococ- antibacterial and non-specific immunity. J. path. cique de l'embryon de poulet inocule par voie sous- Bact. 1939, 48, 545. cutanee. Un. med. Can. 1954, 83, 382. 21. Flaum, A. Studies in staphylococci and staphylococ- 38. Knothe, H. Das verhalten von aeroben Bakterien in cal immunity. Acta path. microbial. scand. (suppl. bebrutenen Huhnerei. II. Metterlung. Das Ver- 35) 1938, 35, 1. halten von Staphylokokken, Meningokokken, Strep- 22. Gorrill, R. H. Experimental staphylococcal infec- tokokken, Pneumokokken, Bacterium tularense, tions in mice. Brit. J. exp. Path. 1951, 32, 151. Pasteurella pestis, und Brucella abortus Bang im, EXPERIMENTAL STAPHYLOCOCCAL INFECTIONS IN EMBRYONATED EGGS 2157

bebrilteten Huhnerei. Zbl. Bakt., I Abt. Orig. Mechanism of immunity conferred by anti-alpha 1952, 158, 453. hemolysin. Bull. Johns Hopk. Hosp. 1963, 112, 39. Wiley, B. B. A new virulence test for Staphylococ- 31. cus aureus and its application to encapsulated 42. Lack, C. H., and D. G. Wailling. A study of 435 strains. Canad. J. Microbiol. 1961, 7, 933. strains of Staphylococcus pyogenes with reference 40. Jones, M. The production of staphylococcus toxin to factors which may contribute to pathogenicity. in the chorio-allantoic fluid of the embryonated J. Path. Bact. 1954, 68, 431. egg. J. Bact. 1946, 51, 789. 43. Elek, S. D., and E. Levy. Distribution of haemoly- 41. Goshi, K., L. E. Cluff, and P. S. Norman. Studies on sins in pathogenic and non-pathogenic staphylo- the pathogenesis of staphylococcal infection. VI. cocci. J. Path. Bact. 1950, 62, 541.