JOURNAL OF BACTERIOLO-,Y, Juile 1967, p. 1825-1831Copyright ) 1967 Ao11eiricon1 Society for Microbiology

Vol. 93, No. 6Prilnted in U.S.A.

Effect of Streptococcus pyogenes on Tissue CellsROBERT W. QUINN AND P. NYE LOWRY

Delartmnent of Prevenztive Medicinle antd Public Health, Vamiderbilt University School of Medicilne,Nashville, Tennezessee 37203

Received for publication 30 January 1967

Human tissue cell lines from each of the three primary germinal sources, ectoderm(conjunctiva and carcinoma of the buccal mucosa), entoderm (intestine and liver),and mesoderm (heart and monocytes) were inoculated with group A Streptococcuspyogenes, Staphylococcus aureus, and group D streptococci and were then observed.In addition, the effect of these bacteria on mouse fibroblasts was studied. All of thecell lines appeared to be equally susceptible to damage, but damage to the cells byS. pyogenes occurred only when living, actively multiplying bacteria were in contactwith the tissue cells. Streptococcal products in the form of "used" growth mediumhad no observable effect on the cells. Cytopathogenic effects were first noticed aboutthe time one would expect the bacteria to have reached the end of the log phase ofgrowth. No damage to the tissue cells was noted when group A streptococci were

separated from the cells by membrane filter diffusion chambers or dialyzing mem-branes, but a membrane did not protect cells from deleterious effects of staphylococcior group D streptococci. Group A streptococci survived in the tissue culture medium,but multiplication did not occur unless living tissue cells were present.

The inflammation accompanying infection ofthe upper respiratory tract due to group AStreptococcus pyogenes is confined to the sub-mucosa with secondary involvement of lymphoidtissue. This sharp localization suggests that groupA streptococci may possess a certain tissuespecificity or that certain tissue cells may be ableto supply the growth requirements and thereforeare particularly susceptible to damage by strepto-cocci or streptococcal products. Previous experi-ments (2) had shown that human heart cellsgrowing in tissue culture were susceptible todamage by a larger number of strains of group Astreptococci than were monkey kidney or mon-key heart cells, suggesting that human heart cellsmight be especially susceptible to damage by S.pyogenes. Some evidence that human tonsil cellshad an inhibitory effect on the multiplication ofgroup A type 6 streptococci was obtained in thesame experiments (2). The purpose of the presentexperiments was to study further three questions,the first concerning the effect of three differentbacterial species on human tissue cells derivedfrom the three basic germ layers. The threespecies were the etiological agent of streptococcalrespiratory infections, the group A S. pyogenes,Stiaphylococcis ciiireus, which is rarely the primarycause of upper respiratory infection, and group Dstreptococci (enterococcus) which do not causerespiratory infections. The second question in-volved the possible inhibitory effect of some tissue

cells on the growth of group A S. pyogenes. Thethird question was concerned with possiblecauses of damage to cells by S. pyogenes.

MATERIALS AND METHODSTissue cell Iihues. The tissue cell lines used and their

derivation are shown in Table 1. The growth mediumin all cases was Minimal Essential Medium (MEM;BBL), containing 10% calf serum. To observe theeffect of streptococci on nonhuman cells, mouse fibro-blasts were included among the cell lines tested. Thenumber of experiments performed with each cell lineis shown in Table 2.

Bacteria. The bacterial strains were: 03-hemolyticS. pyogenies group A type 12, originally recovered froma patient with a streptococcal upper respiratory infec-tion; group A type 36 streptococci (obtained from theAmerican Type Culture Collection) which did notcause tissue cell damage in previous experiments; astrain of nontypable S. auireus from a patient with astaphylococcal submaxillary abscess; and a group DStreptococcius (enterococcus? isolated from uterinelochia.

Effect of bacteria oni tissuie cells. The cell lines weregrown in Falcoln-Cooper plastic tissue culture dishesfor most of the experiments. A hole was cut in thebottom of the tissue culture dish which was then sealedby a cover slip with silicone grease. The well formedin this way facilitated observation through an invertedocular, phase-contrast microscope and allowed the useof the oil immersion lenis. In these experiments, thetissue cells were dropped into the well in the center ofthe dish, and, when growth was adequate, usually in14 to 18 hr, the mediLum wvas changed and streptococci

1825

on Septem

ber 2, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

QUINN AND LOWRY

previously cultured in Todd-Hewitt broth were inocu-lated into the freslh tissue culture medium. The approxi-mate average nLumber of chains of streptococci orindividual staphylococci inl the inoculum for each dishwas 1,800. The number of tissuLe cells used to start aculture was kept withini a range of 20,000 to 30,000cells per dish. The incubation temperature was 35 Cin an attmosplhere of 5',- carboon dioxide in air whiclhhelped maintain the pH of the growth mediuLm be-tween 6.8 to 7.4. To study the effect of bacteria on thetissue cells when the bacteria and cells were separatedby a barrier, a dialyzing membranie sealed with siliconiegrease over the cover slip on the bottom of the tissue-culture dish was used, because it made a tight sealwhich effectively separated cells from bacteria butallowed transfer of nutrienits through the membrane.

TABLE 1. Tissue cell lines,'

Source Deriv-ation

H uman liver (Chang)', EnitodermHuman intestine (Henile) h EntodermHumani conjunctiva (Chang)b EctodermHuman epidermoidal carcinomna Ectoderm

of the mouth'Human heart (Girardi'b MesodermHuman monocytes (Subacute Mesodermmonocytic leukemia blood)-

Mouse fibroblasts

a Growth medium was Minimal Essential Me-dium (BBL), containing 10% calf serum.

I Obtained from Microbiological Associates,Inc., Bethesda, Md.

c Obtained from BBL.

Another metlhod of separating the bacteria fromii thecells was by use of small diffusion chambers coIml-posed of membranie filters (average pore diameter,0.45 ,; Millipore Corp., Bedford, Mass.) on one sideand a dialyzinig membraile oni the other, like thosedescribed by Algire et al. (1). Approximately 1,800chains of type 1 2 streptococci, 33 chains of type 36. orone 2-mmll loopfLul of a 10- tdilution of staphylococcianid group D streptococci, previously cLultured inTodd-Hewitt broth, were initroduced into the chlamuberwhich was then sealed and dropped into the tissue cul-ture dish. This technique allowed bacteria to multiplyin the tissue cLIlture medium withinl the diffusionchamber, separate from the tissue cells. In each ofthese experiments, controls were maintained to furnishevidence that the tissue cells were healthy and mul-tiplying in the absence of bacteria or bacterial products.

Cytopatlhogeniic effects. Cytopathogenic effects weredetermined by methods similar to those describedpreviously (2) whiclh included the occurrence of cyto-plasmic blebs, thickening of the nuclear membrane,coarse cytoplasmic granulation, distortion of nucleoli.cell shrinkage, and failure of cells to stick to the glass.The final phase in observed changes was total or neartotal destruction. The method for determining viabilityof the cells consisted of the addition of 1 ml of l-Itrypan-blue solution to 5 ml of cell suspension. Deadcells turned pale blue, and living cells remainedcolorless.

Possible inhibitory effect of tissue cells oni the growtlthof S. pyogenes. In previous experiments, there wassome evidence of an inhibitory effect of human tonsilcells on streptococcal growth (2). This line of cells wasno longer available but all six of the human cell lineslisted above were tested for possible inhibitory effects.The growth curve for group A type 12 streptococci in

TABLE 2. Number anld type oJ experiments

No. of expt per formed witl

Type of expt \fItuman Human cHonjuaIc- luman hluman Human AMouseliver intestine con carcinoma hieart mono- fibro- Total

(Clhang) (lenle) tiv(Cang) of mouth (Girardi) cvtes blasts_~~(Ciang

ControlCells in dishes with no bac-

teria ... ... .... ...

BacteriaStreptococcus A-12.Streptococcus A-36........Streptococcus DStaphylococcus alureis

Algire chamberStreptococclis A-12, S. aoureuis

or Streptococcius D..Dialyzing membraneStreptococcus A-12, S. alulreutsor Streptococcuts D.

"Used' medium.

Total...........

10

4 71 11)I

1I I

5l 6

II 5 5

2) 3

27 34

8 7 8

4 5 41 1 12 21 1

4 5 D

4

9 3I 2

25 27 28

5

3

1

1

4 I1

4 ]

19 5_1

46

28

6

106

2712

165

7

l1826 J BAC TERIO1-

I

I

I

on Septem

ber 2, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

EFFECT OF S. PYOGENES ON TISSUE CELLS

MEM in the presence of and in the absence of tissuecells was determined. The tissue cells were grown inglass cuvettes 30,000 to 40,000 cells per cuvette in 4 mlof medium, rather thani in tissue culture dishes becauseof the difficulty of estimating bacterial growth indishes. After 14 hr of incubation, the time at which itwas established that the tissue cells were healthy,multiplying, and adhering to the wall of the cuvette,the mediumi was inoculated with approximately 3,400chains of group A type 12 streptococci and then in-cubated at 35 C. Bacterial growth was determined byoptical density and percentage of light transmission,measured in a photoelectric colorimeter at 600 m,usuch that light did not pass through the layer of tissuecells which was confined to a narrow band on one sideof the cuvette. A different cell type (human embryoniclung) was employed to test for possible effects ofstreptococci (A-12) on turbidity due to changes in thecalf serum contained in the MEM. Controls consistedof cuvettes with medium plus calf serum and mediumplus cells and calf serum. The calf serum was veryfaintly turbid, and the effect on percentage of lighttransmission was very slight. At the end of the growthperiod, the pH was adjusted to that in the controlcuvette. The cuvettes were inverted once or twice todisperse the clumps of bacteria, and then the percent-age of light transmission was determined. The colorim-eter scale was set at 1001- light transmission by use ofthe control cuvette containing medium plus calf serumas the standard. The cuvettes were then centrifuged at2,300 rev/min for 15 min, and the percentage of lighttransmission was again determined. The growth curvefor group A type 6 hemolytic Streptococcuts in Todd-Hewitt broth has been determined (R. W. Quinn, andR. S. Quinn, unlpliblished clata) by optical density,percentage of light transmission, and numerous serialcolony counts at 30-min intervals; it is similar to theLIsLual characteristic growth curve of other bacteria.However, an explanation is necessary to indicate whystreptococcal bacterial growth was measured by de-crease in percentage of light tranismission rather thanby viable counts. Actually, the latter is extremely dif-ficult to determine. Wheni streptococci are streaked ona blood-agar plate and colony counts are made, eachcolony originates from a viable chaini of streptococcirather than from a single coccus, so the count isrepresentative of viable chains, not of viable singlecocci. As the number of chainis begins to increaseraoidly during the log phase of growth, higher andhigher dilutions of the medium are necessary to pre-vent the plates from being overgrown with countlesscolonies. During the serial-dilutioni process, eventhough it is done as gently as possible to obtain propermixing, chains are Llnavoidably broken. Each of thebroken parts of the clhain may form a separate colony,thus resultillg in many more colonies than would havebeen obtained before the chains were broken. Thegrowth curve for hemolytic streptococci does notmeasure the number of single cocci but ratlher thenumber of chains. Because of the considerable teclhnii-cal difficulty attendant on determining viable counts,estimates of grow,th based on light transmission weremade instead.

Possihlh (c1sesC5 of (Icoflicge 1o tissute cec/s bY S.

p) ogeIies. To test the possibility that tissue cell damagemight be due to depletion of essential nutrients or totoxic products in the growth mediulll, medium inwhich group A type 12 S. py/rogenes and cells hadgrown together for 32 hr was incubated with 125 unitsof penicillin and 125 units of streptomycin for 2 hr tokill the streptococci. MediuLml from group A type 36S. pyogelnes was also used. The bacteria were removedby centrifugation and this 'used" medium was em-ployed as a change of medium for cell lineshomologous with the cells whichi had been inoculatedwith the two strains of streptococci. One set of experi-ments utilized "used" medium from which the killedstreptococci had not been removed. At the time ofaddition of antibiotics to kill the streptococci, thetissue cells of each of the six human cell lines showedtotal destruction, indicating that whatever was re-sponsible for cell death was very effective.

RESULTSEffect of bacteria on tissue cells. These experi-

ments confirm previous observations (2) thatdamage to human cells growing in tissue culturemedia by group A S. pyogenes occurs only whenthe cells are in direct contact or close proximitywith living, multiplying bacteria. Figure 1 showshuman monocytes before inoculation with strep-tococci. Figure 2 shows the remains of the samecells after being incubated for 25 hr with groupA type 12 streptococci. Streptococcal chains canbe seen on top of the cellular debris. Althoughdifficult to photograph in sharp focus, countlesschains of streptococci "blanketing" the tissuecells were seen through a phase-contrast micro-scope. Complete destruction occurred in all thehuman cell lines within 36 to 38 hr by both types12 and! 36 group A streptococci, S. aureuts, andgroup D enterococci when growing in directcontact. Total destruction of human heart cellsby group D streptococci was delayed until atleast 44 hr. Separation of the two strains of groupA streptococci from the cells by a dialyzing mem-brane or a membrane filter prevented damage tothe tissue cells. Large numbers of group A strep-tococci growing in the dishes had no observableeffect on cells within 48 hr when separated bv amembrane, nor did smaller numbers containedwithin a diffusion chamber. The findings forgroup A streptococci difTered from those forstaphylococci and group D streptococci, becausecell damage did occur when these bacteria wereseparated by a dialyzing membrane from the cells.No cytopathogenic effects were observed whenstaphylococci or group D streptococci were en-closed in a membrane-filter chamber. This puz-zling difference between the effect of staphylococciand group D streptococci and group A strepto-cocci is unexplained. Damage to mouse fibroblastsby group A and group D streptococci wasminimal even after 36 to 92 hr in direct contact.

1 827VOL. 93, 1967

on Septem

ber 2, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

QUINN AND LOWRY

Fic I Normal lunum i0ocytes sI(hctt 1O0C1i(iC lelikeln1't Wlood (I(/lo' 24 hr1ofi(4'lubation in A lunmulEssential .Alidium (it5 C.

Inhibitlor cfcCI otSnecCi tiShe C/ll/s on /)uCtCe-iolgrowt,th. Group A streptococci, types 12 and 36,inoculated into the MEM in the absence of tissuecells survived, but there was no evidence thatmnultiplication took place. Growtlh of bacteria inthe presence of cells in a tissue culture dish isdramatic alnd easily observed. It is noteworthythat ther-e was at striking difVerenice between thenlumber-s of streptococcal chaills gI-owin in thepresence of cells in a tissue cultur-e tube aLnd thenumber-s of those grown in a t'issue culture dish.Growth in the dishes was extremely heavy; thestreptococci seemiied to formll aL lblanket' overthe cells Fig. 2). This was chalrLccter-istic of allcell lines. In the tubes, growth l'was flocculent, asis usually seen in Todd-Hewitt broth, but wasmuclh lighter than in the dishes, with only a fewclumps of bacteria adhering to the cuvette walls.The reasons for this difference in growth are notapparent. There was only slight variation in the

growthi curve for- group A type 12 streptococci incontact with each of the six human cell lines. Thelatent or log plhase lasted about 3 hr, after whichia slight but steady decrease in percentage of lighttransmission occur-r-ed. Figure 3 shows the growthcurve for group A type 12 S. p)yogenes in humanhiear-t cells gr-owling in MEM or Eagle's BasicMedium. Even tlhouglh the evidence of gr-owth ofbacteriac balsed on decr-ease in percentage of lighttransmission ws nlot striking it was greater thanth,at for eaclh of the other cell lines. Figur-e 3repr-esetnts the greatest amount of gr-owth ob-served in an ot' the exper-imilents. Multiplicationcontinued slowly unitil about 24 hr, at whlich timeno further increcase in optical density or decreasein percentaige of light tracnsmission occurr-ed. Thegrowth rate of streptococci in the presence ofliving humnan tissue cells in a tissue culture tubeis much slower aLnd the total numbers of bacter-iaare significantly fewer than when the bacteria are

1 8218 .1. BA( AHOll()V

on Septem

ber 2, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

EFFECT OF S. PYOGENES ON TISSUE CELLS

FIG. 2. Severely d(liiaaged IhiumlIanl 1o10n1ocytes, cell detrituts, anid clhailis of streptococci. Photograph takeln of thesamiie cells ais in Fig. 1, 25 hr after inoculation with groutp A/ tYpe 12 streptococci anld after incluhaltioni cit 35 C. Ar-rowvs poilmt to clilciils of streptococci.

grown in ordinary bacterial culture medium,such as Todd-Hewitt broth. The important pointis that multiplication of the two strains of groupA streptococci in Eagle's Basic Medium andMEM probably did not occur in the absence ofliving tissue cells, indicating that living tissuecells were essential for bacterial multiplicationin this system.The experiments to demonstraLte a possible

effect of streptococci on calf ser-umnand onturbidity showed that in each tube after centrif-ugation the percentage of light transmission waseven greater than that of the control, indicatingthat turbidity other than that due to bacteria didnot influence the measurements. The mediumactually became slightly clearer after the cellsand bacteria had incubated. Colony counts atthe time of inoculaLtion of the medium plus cellswitlh bacteria and at the end of the growth period

indicated great multiplication, even though thedecrease in percentage of light transmission wcasno more than 20% in most of the experiments.The experiments designed to detect cell dam-

age, possibly the result of toxic products ornutritional deficiencies in the medium, showedthat cell growth and multiplication proceedednormally in "used" medium and that no deleteri-ous effects on the cells occurred withiln 72 hr.

DISCUSSIONHuman tissue cells (liver, intestine, conjunctivl,

carcinoma of the buccal mucos(a. heart, andmonocytes) growing in tissue cultutre mediumwere completely destroyed by group A strepto-cocci, S. aium'euls, and group D streptococci whichwere actively multiplying in close proximity,regardless of the germ layer origini of the cells.Daimage to the cells by types 12 and 36 group A

1 829VO L 9)3, 1967

on Septem

ber 2, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

QUINN AND LOWRY

z

0

z

I

z

LLC)

Lu

0oo

95

90

85

80

75

70

65

60

55X

50

5 9 12 14 16 18 20 22 24

TIME IN HOURS

FIG. 3. Growth curve for groutp A type 12 Strepto-coccus pyogeiies in tile presenice oJ Ilunuall hleart cellsgrowivng in Minimal Essen,tial Medium (dutplicate ex-periments').

streptococci was prevented when the bacteriaand cells were separated by a dialyzing membraneor a membrane filter. These results confirm theprevious findings (2) that damage to tissue cellsin tissue culture by types 6 and 12 group A strep-tococci occurs only when streptococci are inintimate contact with the cells; they also confirmthe well-known fact that group A Streptococcusis an extracellular parasite. Here, the similarityamong the two types of group A streptococci, thestrain of S. aurelis, and group D enterococcusends, since, with staphylococci and group D en-

terococci separated from the cells by a dialyzingmembrane, damage to the cells did occur. Thereason for this difference is not apparent. If celldamage is due to toxic molecules of staphylococcior group D streptococci, the molecules are ap-parently small enough to pass through a dialyzingmembrane. The lack of cell damage when thesetwo species of bacteria were enclosed in a smallchamber might be associated with the relativelysmall numbers of bacteria, the damaging effectsoccurring only with large numbers of bacteria.It might be postulated that the limited numberof group A streptococci contained in the smalldiffusion chambers are not enough to cause

damage to the cells. However, much larger num-

bers of streptococci multiplying in the dishes, butseparated by a membrane, did not damage the

cells. Apparently, large numbers of group Astreptococci are not a factor, unless there is directcontact with cells.The possibility that tissue cell damage is due to

depletion of essential nutrients does not seemlikely, since cells and streptococci will developand multiply normally in "used" medium inwhich the streptococci have been killed and thecells have ceased to multiply and are dying. Oncethe streptococci were killed by antibiotics, itmade no difference whether the bacteria wereremoved or left in the medium; the cells grewnormally. The fact that cells grew and multipliednormally in the "used" medium is evidence thata toxic substance or lack of nutrients in the "used"medium is not responsible for the damage. Oneis tempted to explain the cell damage which beginsto occur at about the time the bacteria havereached the end of the log phase of growth, bysaying that the streptococci are better competitorsfor the available nutrients simply because of theiroverwhelming numbers, for, as soon as the strep-tococci are killed or removed, the cells then areable to grow normally. Stevens (3) has postulatedan explanation for the cause of death in bacterialinfections: "the chief cause of death is depriva-tion of oxygen to critical host tissues secondaryto direct use of this oxygen by the bacteria."Although they do not furnish proof, the presentexperiments do suggest such a hypothesis. Thisquestion will be studied further.The results, contrary to previous findings (2),

suggest that tissue cells of any of the lines studiedenhance the growth of streptococci. Althoughstreptococci survive, it is doubtful whether multi-plication occurs in tissue cell growth medium(MEM) in the absence of cells; when clls arepresent, bacterial growth is excellent, indicatingthat the cells furnish something necessary forgrowth of streptococci. With few exceptions,streptococci pathogenic for humans do notsurvive in nature outside the human host, oneexception being the case of bovine mastitis dueto group A streptococci. Man is the only reservoirof importance. These experiments furnish furtherevidence of the same kind of dependent relation-ship since streptococci did not multiply in theabsence of living cells.The observation that nonhumrian cells tmouse

fibroblasts) were only slightly damaged by groupA streptococci reinforces what is known of thespecies specificity of Streptococcuts, since it israrely found outside the human host except in thecase of bovine mastitis transmitted from man tocow or in mice infected in the laboratory.

I'IO J BACTER 1L.

on Septem

ber 2, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

VOL. 93, 1967 EFFECT OF S. PYOGENES ON TISSUE CELLS 1831

ACKNOWLEDGNMENT bers. I. Survival of homografts in immunlllllizecdmice. J. Natl. Cancer Inst. 15 :493-507.

This investigation was supported by grant Nonr 2. LOWRY, P. N., AND R. W. QUINN. 1964. Effect of2837(00) from the Office of Naval Research. hemolytic streptococci on tissue cells Proc Soc.

Exptl. Biol. Med. 116:46-51.LITERATUIRE CITED 3. STEVENS, K. M. 1964. The cause of death in acute

1. ALGIRE, G. H., J. M. WEAVER, AND R. T. PREHN. systemic bacterial inifections and in cancer.1954. Growtlh of cells in vivo in diffusion cham- Lancet 1:1254-1256.

on Septem

ber 2, 2020 by guesthttp://jb.asm

.org/D

ownloaded from