Cellophane TransferApplication to the Study of Activity of Combinations of Antibiotics

Y. A. CHABBERT1 AND J. C. PATTE2

Laboratoire de 1'Hopital, Institutt Pasteur, Paris, France

Received for publication October 6, 1959

The bactericidal activity of combinations of anti-biotics have been studied by two methods; in onie,li(quid media are used (Kirby, 1944; Hobby andDawson, 1946; Price et al., 1949; Bigger, 1950; Blisset al., 1952; Jones and Finland, 1956) and in the second,solid media are preferred. A simplified method usingli(quid media, suggested by Jawetz and co-workers(1955), can be used in the clinical laboratory. It hasbeen further modified by us (Chabbert, 1953b). Inthis modification (the disc tube method) which we useroutinely, the percentage of survivors is estimated bystreaking on a solid medium.

These simplified techniques permit the study of onlya small number of arbitrarily selected concentrationsof antibiotics, but the methods utilizing diffusion onsolid media permits one to obtain gradients of concen-trations (Bonifas, 1952; Chabbert, 1953a; Knapp andJacherts, 1955; Dye, 1956; Weil and Kotsevalov, 1956).Dye (1956) suggested the use of two strips of filterpaper, deposited at right angles on the surface of theagar to permit the study of the bacteriostatic effect ofvarying concentrations of two different antibiotics.To study the bactericidal effect of combinations of

antibiotics, the method of Dye had to be modified topermit the transfer of survivors to an antibiotic-freemedium. This transfer can be done by the velveteenreplica plating method of Lederberg and Lederberg(1952) (Elek and Hilson, 1954; Pilkington et al., 1956;Manten, 1956) or by growing the organisms to bestudied on cellophane. This method proposed by Gratia(1944) and Heatley (1947) has been modified furtherin our work (Chabbert, 1953a). Recently we described amethod permitting the growth of bacteria on a dia-phragm of cellophane (Chabbert, 1957). The combina-tion of this method with that of Dye (1956) permittedus to study the bactericidal effect of combinations ofantibiotics in the clinical laboratory.

MATERIALS AND METHODS

1. Preparation of cellophane diaphragm. Cellophanesheets 0.03 mm thick are cut up into squares measuring150 by 150 mm. One square sheet is heated in a moistatmosphere in a vessel and tightly stretched with arubber band over a Pyrex glass ring 80 mm in diam and

I Chef de Laboratoire de l'Institut Pasteur.2 Assistant de Laboratoire de l'Institut Pasteur.

25 mm high. This forms a tambourine which is thenlaid down in a Petri dish, 100 mm in diam, the cello-phane resting on a filter paper slightly moistened withwater. The whole is wvrapped in sulfurized paper andsterilized in an autoclave. The tambourine can be usedas long as a certain degree of humidity remains in thedish.

2. Preparation of the dishes. Strips of blotting paper(200 g per m2) 70 by 5 mm were impregnated with thefollowing concentrations of antibiotics and then airdried at 37C: penicillin 50 OU per ml; streptomycin,1000 ,ug per ml; chloramphenicol, 1000 ,ug per ml;tetracycline, 100 ,ug per ml; erythromycin; 200 jAg perml; and novobiocin, 100 ,ug per ml. These conceintrationswere chosen because they gave, after the incubationtime used, gradients of conicentrations equivalent tothose found in body fluids and tissues. These deter-minations were made by utilizing for each antibiotic aseries of strains of known sensitivity as determinied bydilution assays in agar.The strips were deposited at right angles to each

other on the surface of Petri dishes containing 30 ml ofagar, such as brain heart infusion agar, and the anti-biotics were allowed to diffuse 24 hr at 37 C.

3. Seeding and setting the tambourine. The blotting

Figure 1. Cellophane tambourine showing growth of thepersisters.

193

on June 5, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

Y. A. CHABBERT AND J. C. PATTE

paper strips are then removed. Seeding of the drum'scavity was with 1 ml of a 24-hr culture (brain heartinfusion) diluted in distilled water 1:10 to 1:100.Excess liquid was removed with care and the drumwas deposited on the agar. The dish was dried at 37 C,opened, and turned upside down in the incubator on aperforated metallic sheet (figure 1).

4. Culture and transfer. After 6 or 18 hr, the drumwas removed and deposited on a Petri dish containing30 ml of the agar medium best suited to the growth ofsurvivors. Photographic records could be kept for theclinician.

All the commonly used antibiotics passed throughthe cellophane. Cellophane did not carry over detect-able amounts of antibiotics at the concentrationsrecommended here. This was determined by seedingsensitive bacteria in cellophane drums which had beenkept on antibiotic-containing media and then depositingthem on an antibiotic-free medium.

RESULTS AND DISCUSSION

I. ReadingA. Activities of single antibiotics. According to the

scheme of Dye (1956), the phenomenon wouild beconstant if taken on a parallel to the strip of paper.

It is commonly admitted that the bactericidalactivity of antibiotics is roughly proportional to theirconcentration: the number of survivors decreases whenthe concentration increases.Such phenomena seldom held true when using

cellophane transfer. Photographs never showed aregular decrease in the number of survivors whenapproaching the strips.

Observations may be classified in three groups:

Q~ CL

lk

$~~x- :s0. -_

(a) The percentage of persisters was very nearlyconstant, whatever the concentration. The populationwas homogeneous over the whole gradient (see theactivity of penicillin, figure 4).

(b) The percentage of persisters was decreasing, notconstantly, but by successive steps: the density sud-denly changed from one value to a lower one or to nil(see the activity of tetracycline in figure 2b).

(c) The population exhibited a zone phenomenon.This is pictured by alternating areas of total bactericidalactivity and areas showing an homogeneous density ofsurvivors (see the activ;ity of erythromycin in figure 3).

It may be predicted that the effects of various com-binations will vary depending on the type of actionexhibited by the antibiotics forming the combination.

B. Discussion of types of activities caused by combina-tions. The effect of the combination should be read inthe zone where the two antibiotics acted simultaneously,that is, in the right angle formed by the strips of anti-biotics. For any point of this zone, the effect of eachantibiotic, acting alone, could be known by observingthe effect produced on the bacteria in the zone whereeach antibiotic acts alone, without being influenced bythe other member of the pair.

Antibiotics exhibiting an activity like that of typea, and often of type b, will usually produce, when incombination, easily interpreted activities. One singletype of action will be observed throughout the zone ofsimultaneous action. Such activities may be brokendown as follows (figure 2):

(a) Additivity: The density of persisters variesconstantly when passing from the area of activity ofthe first antibiotic to that of the second through thearea of dual activity. In the case of total bactericida

~ Chloraamphenicol

Figure 2. Demonstration of antagonism, additivity, and synergism between antibiotics. a (left). Staphylococcus strain no. 231.Combination: novobiocin and chloramphenicol. Transfer after 24 hr. b (center). Staphylococcus strain no. 343. Combination: chlor-amphenicol and tetracvcline. Transfer after 24 hr. c (right). Staphylococcus strain no. 121. Combination: ristocetin and streptomycin.Transfer after 4 hr.

194 [VOL. 8

on June 5, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

CELLOPHANE TRANSFER AND ANTIBIOTIC ACTIVITY1

activity, an additive curve in the form of a hyperbole,will be formed exactly as in bacteriostasis (see figure 2b).

(b) Synergism: A decrease in the density of persistersis observed in the area where both antibiotics are active.This decrease will usually be characterized by an areaof no growth giving rise to a characteristic notch (figure2c).

(c) Antagonism: The density of persisters is in-creased in the area of dual activity (figure 2a).

It is apparent that the geometrical patterns formedby the survivors are very similar to those we havedescribed for bacteriostatic activity (Patte et al., 1958).However, it should be emphasized that bacteriostasisseems to be entirely independent of bactericidal activ-ity. Since the transfer technique permitted the succes-sive study of both phenomena, we have noticed thatthey were not usually equivalent. Moreover, it shouldfrequently be possible to prescribe an antibiotic with

Figure S. Effect of combination of tetracycline and erythromycin on Staphylococcus strain no. 302

{ i.4.

Figure 4. Effect of combination of ristocetin and penicillin on Streptococcus group D Ly with transfer made after 24 hr.

195

on June 5, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

Y. A. CHABBERT AND J. C. PATTE

no bacteriostatic activity of its owni, but which has asynergistic activity when in combination with anotherantibiotic. Such synergism wN-as rather frequent in thecase of streptomycin (figure 5) and we shall discuss itlater.Very often the type of activity of the combination

differed within the zone of dual activity from one pointto another. The phenomenon is then complex and mayaccount for the discrepancies observed between liquidmedia techniques and cellophane transfer. Such com-plex effects may be illustrated as follows:

(a) In the case illustrated in figure 3, the two anti-

biotics (tetracycline and erythromycin) showed a verycomparable zone effect.

If one takes a point of total bactericidal activity,the parallels (drawn in the picture) would lead to zonesof persisters for each antibiotic. This original pointwould seem to be due to synergism. Conversely, anotherpoint chosen in the zone of persisters would havecoordinates leading to zones where the two antibioticsacting alone would give total bactericidal activity. Thiszone of persisters would appear to be due to antagonism.Such are the kinds of effects one may encounter whenusing, in a liquid medium, concentrations chosen

Figure 5. Effect of combination of penicillin and streptomycin on Streptococcuis group D, no. 7 and no. 10 with transfer madeafter 24 hr.

[VOL. 8196

on June 5, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

CELLOPHANE TRANSFER AND ANTIBIOTIC ACTIVITY

haphazardly. Therefore, we prefer to consider thedifferent areas in terms of the hyperbole which we havedefined as a typical additive curve. These phenomenawould be classified as "additive" effects.

(b) One antibiotic could also overpower the otherone in the zone of combined activity, especially at highconcentrations. If the former antibiotic was the lessbactericidal one, the combination would be antagonisticto the latter antibiotic in that particular area (figure 4).

This was the case for the ristocetin-penicillin com-bination in the above picture. When alone, penicillinyielded a small number of of persisters, while ristocetinyielded many. When in combination, even with largequantities of penicillin, high concentrations of ristoce-tin and penicillin acted jointly, the bactericidal effectbecame important: an area of synergism showed up.Beyond this concentration the observed effects werethat of penicillin alone.The study of such complex activities offered a good

opportunity to show the advantage of a techniquesuch as cellophane transfer over liquid media tech-niques: we were actually able to set up combinationsat all the concentrations, and the results of such asso-ciations could be seen all at once.

II. Comparison between Results Givenby Different Techniques

We have tested bactericidal activity by three differenttechniques: a technique in liquid medium with a squaredesign of experiment; the velveteen replication ofLederberg and Lederberg, (1952) and Elek and Hilson,(1954); and cellophane transfer. In these ways we havecompared results from 208 combinations on thefollowing bacteria:

Staphylococci: Three penicillin-sensitive strains andtwo penicillinase-producing strains, sensitive to all

TABLE 1

Comparison of the results obtained by three different methods usedin the study of bactericidal effects of combinations of twoantibiotics. Percentages of combinations falling in each group

Staphylo- Strepto- Gram-Type of Agreement Between cocci, 120 cocci negative Total, 208

Method Combina- Group D, Bacteria, Combi-tos60 Combi- 28 Combi- nationstos nations nations

Comparison betweensquare liquid tech-nique and diffusiontechniques

Concordance ....... 49.49 83.5 58.5 60.5Discordance ............ 1 0 0 0.5Meaningless ............ 50 16.5 41.5 39

Comparison between twodiffusion methods

Concordance............ 57 80 64 65Discordance ............ 6.7 10 0 7Meaningless ............ 36.3 10 36 28

other antibiotics, have been studied with all combina-tions of penicillin, streptomycin, chloramphenicol,tetracycline, erythromycin, novobiocin, and bacitracin.

Eight strains resistant to several antibiotics havebeen studied with two or three combinations only.Group D streptococci: Six strains isolated from

patients suffering from acute endocarditis have beenstudied with combinations of penicillin, streptomycin,chloramphenicol, tetracycline, and erythromycin.

Enterobacteriaceae: Eight strains of Escherichia coliwere tested using combinations made up of streptomy-cin, chloramphenicol, and tetracycline.Pseudomonas aeruginosa: Neomycin-polymyxin B

combinations have been studied with six strains.We shall not describe in detail the results obtained

with each one of these combinations. Briefly, for eachmethod, we classified the results as synergistic, indif-ferent, or antagonistic. In a number of cases, one oreven several techniques did not give conclusive resultsbecause both antibiotics were totally bactericidal underthe experimental conditions used. Such results wereconsidered as "meaningless." In other cases, agreementor disagreement between methods was established.Results in percentages are grouped in table 1.Good agreement between the different methods was

noted. Results are usually comparable, and onemethod will suffice to describe the phenomenon.A real advantage of the cellophane transfer method

appeared to be its sensitivity. Even very small percent-ages of survivors could be demonstrated, and it waspossible to transfer from 50 to 100 times as manypersisters as the velveteen replication technique underthe same conditions. This fact was established bycolony counts.

III. Practical Applications of the CellophaneTransfer Method

1. Detailed study of an antibiotic. The cellophanetransfer technique is now used in our laboratory toascertain the activity of new antibiotic when used incombination. It has been applied to the tetracycline-oleandomycin combination. Both antibiotics showed aneat zone effect and the combination gave a resultvery similar to that shown in figure 3. The patternobserved was of the additive type as defined by uspreviously. It was very similar to that observed pre-viously during the study of bacteriostasis of thisassociation (Patte et al., 1958) and very different of thesynergism observed with the combinations penicillin-streptomycin.

2. Clinical importance. This technique is also used forroutine purposes to study ungrouped as well as groupD streptococci producing acute endocarditis. Theseinfections may be treated immediately by a streptomy-cin-penicillin combination, but cellophane transfers(one after 6 hr contact, another after 24 hr) may allow

1971960]

on June 5, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

Y. A. CHABBERT AND J. C. PATTE

the physician to approximate the doses of penicillin tobe used, and to decide whether streptomycin is to beused or not. Two transfers are made to permit a betterevaluation of the treatment needed. A strain whichleaves no survivors after 6 hr will require a less inten-sive treatment than one which would leave survivorsafter 24 hr.

Effects may be classified into three groups, corre-sponding to three different types of subsequent recom-mended treatment.

(a) Streptomycin and penicillin, a very active com-bination, allowed few survivors after a 6 hr contactand none after 24 hr. If there were survivors, synergismwas easily shown. It would be wise to prescribe penicil-lin-procaine for example, at a moderate dose (a fewmillion units intramuscularly). Combination withstreptomycin would then not be absolutely necessary(figure 5).

(b) Penicillin alone was much less inhibitory thanthe combination and so was streptomycin. Both allowmany persisters, even after 24 hr. But combinationshowed a very marked synergism (see figure 5, top).The treatment would then be: penicillin at high doses(several tens of million units, when necessary withbenemide) given in perfusion, plus streptomycinintramuscularly.

(c) The two previously detailed types made up 95per cent of the cases. But, in the 5 per cent remaining,antagonism was observed. Then penicillin must beprescribed at a high concentration, without streptomy-cin (see figure 5, bottom).

This classification is obviously theoretical and manyintermediate cases may be encountered.Agreement between results in vitro and therapeutic

results has been studied by Martin et al. (1958) in 81cases of streptococcal endocarditis. Apart from resultswhich were difficult to make out for various reasonsand which represents 18 per cent of the cases, goodagreement was observed in 70 per cent. In 12 per centof the cases, a discrepancy between clinical and labo-ratory tests was found.

Finally, we also used cellophane transfer inculturefrom cases of septicemia due to resistant staphylococcior gram-negative bacteria which were only slightlysensitive to antibiotics. However, the effects were fartoo complex and varied according to the strains. Thenumber of cases has not yet been sufficient to allow astatistical comparison between laboratory results andtherapeutic trials.

SUMMARYThe growth of bacteria within a cellophane tam-

bourine permitted the demonstration of the bactericidaleffect of antibiotics in combinations. By diffusion fromtwo strips of paper each containing a single antibioticand deposited at right angles to one another on an

agar medium, gradients of antibiotics could be formed.In the first step, those concentrations of antibioticswere permitted to act on the microorganisms inoculatedon the cellophane tambourine put on the agar. Later,the tambourine was transferred onto a new antibiotic-free agar medium so that any survivors might grow inthe zone of growth inhibition on the cellophane.

Usually, the bactericidal effect of an antibiotic wasnot proportional to its concentration, and zone effectswere frequent. For antibiotic combinations, aside fromsimple effects which were classified as favorable orunfavorable, complex effects, varying with the con-centrations, were often shown. In this way the tech-nique easily permitted the study of the action of singleantibiotics and the interaction of pairs of antibiotics.

Results were compared with those obtained by othertechniques for a total of 208 antibiotic combinations.Good agreement between the different methods wasnoted where inclusive results for both methods couldbe obtained, but the cellophane transfer technique wasthe most sensitive.

In a clinical laboratory, cellophane transfer may helpto establish approximately the doses of pencillin to beused in the treatment of streptococcal endocarditis,and assist in the determination of whether streptomy-cin is to be associated or not.

REFERENCES

BIGGER, J. W. 1959 Synergism and antagonism as displayedby certain antibacterial substances. Lancet, 259, 46-50.

BLISS, E. A., WORTH, P. T., AND LONG, P. H. 1952 Studiesof combinations of antibiotics in vitro and in experimentalinfections in mice. Bull. Johns Hopkins Hosp., 90, 149-169.

BONIFAS, V. 1952 D6termination de l'association synergiquebinaire d'antibiotes et de sulfamides. Experientia, 8,234-235.

CHABBERT, Y. A. 1953a Action des associations d'anti-biotiques sur les germes aerobies. Ann. inst. Pasteur,84, 545-561.

CHABBERT, Y. A. 1953b Technique simplifiee pour l'6tudede l'action bactericide des associations d'antibiotiques.Ann. inst. Pasteur, 85, 122-125.

CHABBERT, Y. A. 1957 Une technique nouvelle d'etude del'action bact6ricide des associations d'antibiotiques: letransfert sur cellophane. Ann. inst. Pasteur, 93, 289-299.

DYE, W. E. 1956 An agar diffusion method for studyingthe bacteriostatic action of combinations of antimicrobialagents. In Antibiotics annual 1955/1956, pp. 374-382.Medical Encyclopedia, Inc., New York, New York.

ELEK, S. D. AND HILSON, G. R. F. 1954 Combined agardiffusion and replica plating techniques in the study ofantibacterial substances. J. Clin. Pathol., 7, 37-44.

GRATIA, A. 1944 Application de la culture sur cellophanea l'etude et a la production des antagonismes microbiens.Compt. rend. soc. biol., 138, 893-894.

HEATLEY, N. G. 1957 A simple plate method for multipletests of antibacterial activity of many bacteria againstother bacterial strains. J. Gen. Microbiol., 1, 168-170.

HOBBY, G. L. AND DAWSON, M. H. 1946 The effect of sub-

[VOL. 8198

on June 5, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

DETECTION OF POLLUTED WATER, MILK, AND ICE CREAM

fonamides on the action of penicillin. J. Bacteriol., 51,447-456.

JAWETZ, E., GUNNISON, J. B., COLEMAN, V. R., AND KEMPE, C.1955 A laboratory test for bacterial sensitivity to com-bination of antibiotics. Am. J. Clin. Pathol., 25, 1016-1031.

JONES, W. J., JR., AND FINLAND, M. 1956 Antibiotic com-binations: antistreptococcal and antistaphylococcal ac-tivity of plasma of normal subjects after ingestion oferythromycin or penicillin or both. New Engl. J. Med.,255, 1019-1024.

KIRBY, W. M. M. 1944 Bacteriostatic action of sulfonamide-penicillin and urea-penicillin mixtures in vitro. Proc.Soc. Exptl. Biol. Med., 57, 149-151.

KNAPP, W. AND JACHERTS, D. 1955 Uber die qualitativeAuswertung des Streifentestes zur Beurteilung der syner-gistischen oder antagonistischen Wirkung zweier Anti-biotikapraparate. Zentr. Bakteriol., Parasitenk., Abt.I., Orig., 162, 474-478.

LEDERBERG, J. AND LEDERBERG, E. M. 1952 Replica platingand indirect selection of bacterial mutant. J. Bacte-riol., 63, 399-406.

MANTEN, A. 1956 The actiorn of antibiotic combinationson pathogenic staphylococci in vitro. Antibiotics &Chemotherapy, 6, 480-486.

MARTIN, R., CHABBERT, Y. A., SUREAU, B., MARTIN, L.,MORNET, J., AND DUMAS, J. R. 1958 Interet et valeurdu laboratoire dans la conduite du traitement de l'endo-cardite subaiguie A streptocoques. 31 eme Congr. Frang.Med. Paris, p. 139-196 (Masson Ed.).

PATTE, J. C., HIRSCH, H., AND CHABBERT, Y. A. 1958 ftudedes courbes d'effet bact6riostatique des associationsd'antibiotiques. Ann. inst. Pasteur, 94, 621-635.

PILKINGTON, T. R. E., ELEK, S. D., AND JAWELL, P. 1956The action of six antibiotics singly and in combination onenterococci isolated from cases of subacute bacterialendocarditis. J. Lab. Clin. Med., 47, 562-571.

PRICE, C. W., RANDALL, W. A., WELCH, H., AND CHANDLER, V.1949 Studies of the combined action of antibiotics andsulfonamides. Am. J. Public Health, 39, 340-344.

WEIL, A. J. AND KOTSEVALOV, 0. 1956 The ratio of con-centrations as a factor in the in vitro effect of combina-tions of antibiotics. Zentr. Bakteriol., Parasitenk., Abt.I., Orig., 165, 459-468.

Comparative Studies on the Detection of Pollution of Water, Milk,and Ice Cream

F. M. RAMADAN

Walter Pollution Unit, National Research Centre, Dokki, Cairo, U. A. R.

Received for publication October 19, 1959

While notable advances favor the use of fecal strep-tococci as an index of pollution in water (Leininger andMcClesky, 1953; Morris and Weaver, 1954) and in foodexamination (Brown and Gibbons, 1950; Barnes et al.,1956; Mossel et al., 1957), critical reports seem to favorthe indicative value of the coli-aerogenes group (Parr,1938; Kelly and Arcisz, 1954; Gilereas and Kelly, 1955;Thomas et al., 1958).Masinova and Cledova, (1957) studied the biochem-

ical properties of Escherichia coli in water and soil forover 2 years. They ascertained that the organisms re-

covered varied considerably under different environ-mental conditions in such manner that the typical E.coli population changed into the intermediate typesand sporadically into Aerobacter aerogenes. They con-cluded that in hygienic analysis of water and soil, thecharacter of fecal pollution must be assessed in relationto the whole group of coliform bacteria and that in-dividual species of the group may be used to evaluatethe duration of fecal pollution.

In a previous report Cooper and Ramadan (1955)studied the merits of fecal streptococci as indicators ofpollution. These studies indicated that typical Strep-tococcus faecalis was recovered only from human feces

and that strains of Streptococcus bovis, capable of hy-drolysing starch, were characteristic of bovine feces.Other members of the enterococcus group, for example,S. faecalis var. liquifaciens, S. faecalis var. zymogenesand a variety of atypical S. faecalis strains were en-countered in human, bovine, and sheep exereta. S.bovis, failing to hydrolyse starch (Orla-Jensen, 1942),and Streptococcus durans were occasionally recoveredfrom human feces. By adopting a special telluritemethod for isolating streptococci from exeretal matter,it was possible to differentiate between the human- andanimal-derived strains quantitatively according to threedevised tests; namely, the Janus green reduction test,the heat-resistance test, and the heat and tellurite-resistance test. Thus, fecal streptococci passing thetwo heat-resistance tests with or without the Janusgreen reduction test were derived entirely (100 per cent)from human feces. Strains failing to pass all three testsaccounted for 98.5 per cent of the animal-derivedstrains. Fecal streptococci showing reactions varyingfrom those stated were found to be less characteristicof a particular source. Truly, such a hypothesis mayharden into a conclusion only with extended investiga-tion covering streptococci from other farm animals and

1960] 199

on June 5, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from