WHO cooperative studies on the phage-typing of mycobacteria

Bull. Org. mond. Sante 1 1973, 48, 57-63 Bull. Wid HIth Org. J WHO cooperative studies on the phage-typing of mycobacteria 1. Phage lysis of Czechoslovak and Italian strains of Mycobacterium tuberculosis L. gULA, W. B. REDMOND, J. F. COSTER, I. BAESS, J. H. BATES, G. CAROLI, E. MANKIEWICZ, T. MUROHASHI, & E. VANDRA] Although phage-typing of some pathogenic microbes-e.g., Escherichia coli and the genera Salmonella, Staphylococcus, and Streptomyces-is in routine use, little progress has yet been made in the phage-typing of mycobacteria, chiefly because mycobacteriophages possess apolyvalence thatprevents the reliable identification ofindividual species ofmycobac- terium. Furthermore, laboratories cannot compare their results because they employ different methods andculture media andalso because thephages usedcannot be accurately identified. It has been shown that internationally reproducible results can be obtainedfor M. tuberculosis by selecting suitable mycobacteriophages and sensitive culture media, and by using a standardized technique ofphage-typing. Under these conditions, phage lysis may be utilized to classify strains of M. tuberculosis. One of the main activities of the WHO Interna- tional Reference Centre for the Diagnosis of Tuber- culosis, Prague, apart from the search for simpler methods of isolating mycobacteria, is the study of new techniques for their reliable classification. The results of cooperative studies on these subjects have been compared internationally and the distribution of mycobacteria the world over has been reviewed. The pathogenicity and virulence of mycobacteria, their susceptibility to the classical and second-line antituberculosis drugs, and their allergenic and im- munogenic properties have been dealt with by other cooperative studies. The resultant data provided a firm basis for planning the systematic prevention and control of tuberculosis in different countries. Some of these activities have been reported in previous articles in the Bulletin of the World Health Organiza- tion (gula, 1963; gula & Sundaresan, 1963; gula & Langerova, 1963). The study of mycobacteriophages and their pos- sible use for the classification of mycobacteria is closely linked to these previous cooperative studies. Three cooperative studies on the classification of mycobacteria have been undertaken so far. The 1The addresses of the authors are to be found in Annex 1. Requests for reprints should be sent to Dr gula. results of the first two, which were of a preliminary nature, are available from the first author; 2 those of the third study are presented here. The objectives of the present study were: (1) to find suitable methods of preserving and transporting mycobacteriophages; (2) to ascertain the sensitivity of Redmond's RVA and RVB media for growing stock strains used for the replication of phages and for phage lysis tests; and (3) to test Redmond's RTD technique (for a definition of RTD, see below), devised for the prac- tical classification of M. tuberculosis, M. bovis, and M. avium. In view of the extent and complexity of these tasks, they were divided among three of the participating laboratories as follows: Isolation, standardization of mycobacteriophages and classification methods, and preparation of stan- dard media: Tuberculosis Research Laboratory, Vet- erans Administration, Atlanta, Ga., USA; ' Tuberculosis Research Institute, Prague (1965) Pro- ceedings of the Second Symposium on Isolation, Classifi- cation and World-wide Distribution of Mycobacteria: Phage Typification of Mycobacteria, Prague, 18-20 October 1965 (mimeographed document); Sula, L. & Stulcovi, A., ed. (1968) Proceedings of the Fourth Symposium on Isolation, Classification and World-wide Distribution of Mycobacteria: Phage-classification of Mycobacteria, Second Working Group, Utrecht, 29 September-I October 1967, Prague, Tuberculosis Research Institute (mimeographed document). 2979 -57- 58 L. SULA ET AL. Serology and lyophilization ofmycobacteriophages: National Institute of Public Health, Utrecht, Nether- lands; Biochemistry and ultramorphology ofmycobacterio- phages: WHO International Reference Centre for the Diagnosis of Tuberculosis, Prague, Czechoslovakia. As phage-typing is much more complicated than other methods of classification, special attention was paid to the standardization of methods so as to achieve reproducibility of results. For this purpose it is essential to limit variations in local investigation techniques to a minimum, and this was attempted in the present study by using a uniform testing tech- nique, with standardized phages and media prepared at one of the participating laboratories (Tuberculosis Research Laboratory, Atlanta). Table 1. Mycobacteriophages and their host strains Phage Host strain DS6A H37Rv GS4E H37Rv AG-1 M. kansas/i (A21) BK-1 H37Rv D-34 F-130 BG-1 Me-153 MATERIALS AND METHODS Phages Table 1 shows the phages used and their host strains, which were prepared in liquid suspension at Atlanta and dispatched to the other participating laboratories in duplicate: one batch in flame-sealed ampoules and the other in screwcapped bottles, in concentrations designated as 1 x RTD and 10 x RTD.1 Only the phage BK-1 was not distrib- uted with its original host strains. One lot of the phages was returned by each laboratory to Atlanta, where the reduction in the viable unit count caused by transport between Atlanta and the cooperating laboratories was investigated. The results of that study will be reported separately. As the use of liquid suspensions is not particularly desirable (since this leads to a rapid loss of viability of 1 For the purposes of this study, " RTD " was defined as the highest phage dilution that would produce confluent lysis in the host bacteria. the organisms), the same phages were lyophilized in the Utrecht laboratory, using a mixture of 5 % sodium glutamate and 0.5% gelatine as the vehicle of lyo- philization. The lyophilized phages were dispatched from Utrecht direct to the other cooperating labora- tories, together with instructions for their reconstitu- tion. Media Redmond's RVA and RVB media were prepared in dehydrated form by Baltimore Biological Labora- tories and sent to the cooperating laboratories for reconstitution by the following techniques: RVA agar medium: Dissolve 29 g of RV agar base in 900 ml of distilled water to which 8 ml of glycerol have been added. Autoclave at a pressure of 0.01 kN/m2 (15 lb/in2) for 20 minutes. Cool to 47°C and add 100 ml of oleic acid albumin complex.* Mix gently. Dispense 30 ml into each Petri dish. * Oleic acid albumin complex: 3.5 g of bovine albumin (fraction V), 95 ml of distilled water, 5 ml of oleic acid solution (0.12 ml of oleic acid in 10 ml of 0.05 N NaOH), and 0.35 ml of sodium hydroxide. RVB liquid medium: Dissolve 18 g of RV broth base in 900 ml of distilled water to which 8 ml of glycerol and 3 ml of TW3 * have been added. Dispense in 180-ml amounts into small flasks and autoclave at a pressure of 0.01 kN/m2 (15 lb/in2) for 15 minutes. Cool and add 20 ml of bovine albumin solution ** to each flask. * TW3: 1 part of Tween 80 in 3 parts of distilled water. Autoclave at a pressure of 0.01 kN/m2 for 15 minutes. After autoclaving, shake tubes while still hot to remix the Tween 80 and water. Keeps indefinitely. ** Bovine albumin solution: 2.5 g of bovine albumin (fraction V), 0.70 ml of sodium hydroxide, and 100 ml of distilled water. Sterilize by Seitz filtration. Strains For the first two series of trials, strains of M. tu- berculosis, M. bovis, and M. kansasii, which had been maintained for long periods under laboratory condi- tions, were used. As frequent passages on artificial media might have caused substantial changes in the biochemical and other properties of the mycobac- teria and, as a result, altered their sensitivity to mycobacteriophages, only freshly isolated myco- bacteria were used for the study. The M. tuberculosis strains were dispatched to the other cooperating institutes from Prague (36 strains) and Pisa (20 strains), and the bovine strains from Pisa PHAGE-TYPING OF MYCOBACTERIA 59 (20 strains). Most of the latter were isolated from abattoir samples. The avian strains, isolated mainly from Dutch pigs, were sent to the other cooperating laboratories from Utrecht. All strains were grown in RVB medium for 5 days. The cultures were shaken daily in order to achieve a homogeneous suspension. Phage-typing technique Phage-typing was carried out according to the technique described by Redmond & Ward (1966). The various strains, which had been grown in a homogeneous suspension on RVB medium, were inoculated on RVA-agar plates in such a way that the whole surface of the medium was evenly spread. After inoculation the plates were placed in an incu- bator to allow the inoculum to dry somewhat. This required 24 hours on the average. Each strain was exposed to both concentrations of all the phages-i.e., the phages transported in ampoules and bottles and those that had been lyophilized. The latter were reconstituted at the participating labora- tories in such a way as to obtain identical numbers of particles of each phage, corresponding to the dose designated as 1 x RTD. After 1 ml of sterile distilled water had been added to each ampoule, the stock phage suspensions were diluted as follows in order to obtain a 1 x RTD dose of each phage: Phage Dilution DS6A 1:100 GS4E 1:100 AG-1 (undiluted) BK-1 1:100 D-34 1:10 BG-1 1:100 In all the cooperating laboratories except Prague, the plates were evaluated within 5 days of the inocula- tion of strains. In Prague, however, the reading was made after 30 days. RESULTS Only the results for the Czechoslovak and Italian strains of M. tuberculosis are presented here,' since the growth of bovine strains was too sparse to allow of reliable phage-typing. Avian mycobacteria, on the 1 Space did not permit the inclusion of two of the tables submitted with this article. The omitted tables have been deposited in the WHO Library and single photocopies may be obtained by professionally interested persons on request to: Chief Librarian, World Health Organization, 1211 Geneva 27, Switzerland. other hand, grew well but proved to be resistant to all the phages except AG-1. There were no significant differences between the Italian and Czechoslovak strains as regards suscep- tibility to phages. However, there were marked dif- ferences in the lytic effects of different phages on the various strains. In Prague, phages DS6A and AG-1 proved to be the most effective, as they produced lysis in all the strains of mycobacterium. Phage BG-1 was only slightly less effective. The other phages failed even to approach the lytic effects of these three. However, when the test dose was increased tenfold, the number of phage-sensitive strains increased. The higher dose therefore appears to be more satisfactory for phage typing than the lower one. The results obtained in Prague were confirmed by the other laboratories, except that the relatively ineffective phages GS4E, BK-1, and D-34 produced an even lower degree of lysis than in Prague, espe- cially in Pisa and Tokyo. This applied particularly to phage D-34, which lysed 16 strains in Prague, com- pared with 1 and 4, respectively, at Copenhagen and Montreal and 0 at Budapest, Little Rock, Tokyo, and Utrecht. The higher degree of lysis produced by these phages in Prague may be explained by the prolongation of the observation period to 30 days at that laboratory. During this time, the agar medium dried out to some extent and, as a result, a number of lytic plaques that had previously escaped notice in the normally hydrated medium became clearly vis- ible. Whether or not lysis occurs in M. tuberculosis exposed to phages depends on the number of par- ticles contained in the phage suspensions. As this number could be affected by the conditions in which phages are transported from one laboratory to an- other, an additional comparative investigation was carried out in Prague in 14 strains of M. tuber- culosis to ascertain whether there was any differ- ence between the lytic action of phages transported in flame-sealed ampoules and screw-capped bottles and that of lyophilized phages. The results are set out in Table 2. The table shows, on the whole, no substantial differences between the lytic effects of the lyophilized phages and those of the phages des- patched in ampoules and bottles, except that lyo- philized phage BG-1 produced lysis in only 5 of the 14 test strains, whereas the phages sent in ampoules and bottles lysed all 14. However, the latter gave unsatisfactory results in Tokyo, where only phages BG-1 and D-34 produced confluent lysis with the dose designated as 1 x RTD. The table also shows L. §ULA ET AL. o _ that, for some phages, the dose called 10 x RTD >. LOID was more effective than that designated 1 x RTD, without, however, inducing nonspecific lysis. Con- vincing proof of the latter was obtained by renewed isolation of the phages from lytic plaques. The degree of lysis produced in the host strains by co the phages sent in liquid suspension from Atlanta and in lyophilized form from Utrecht is shown in Table 3. It is apparent from the data that there was a a= < * * considerable decrease in the titre of the phages transported from the USA to Japan, whereas most of oU 3 O -the phages sent from Utrecht in lyophilized form produced confluent lysis. These results demonstrate that the better method of preparing phages for x <: < < transport-and one that guarantees a sufficiently o a high degree of phage viability-seems to be to 0 lyophilize them according to the technique developed le _ * *in Utrecht (H. W. B. Engel & J. F. Coster, unpub- o X lished data). However, according to the results pre- xC wsented in Table 3, neither the phages in suspensionC? - >nor all the lyophilized phages retained their full titre _', _ during transport from the two distributing labora- o tories, and it must be assumed that they lost their co > E mtitre to varying degrees depending on the duration of c _~ X transport and the temperature. In future, therefore, when phages are to be sent out from a central n 0E <: _ _ laboratory, it would be desirable to dispatch them o z L undiluted. The RTD can then be determined by the DE o laboratories that are to use the phages. In this way, o z . O better standardization of the technique would be00 xQ co achieved, since a decrease in titre during transport C. a c] __ would be eliminated. C', 0 a, DISCUSSION ~~~~~0~ c a6.4 a. In view of the divergent results that various labo- C) N c ratories have obtained when testing the susceptibility 0o a of mycobacteria to phages, this method has not yet .c. a: _been widely used for classification purposes. TheX nS X ' Xreasons for these differences are several. First, there CA~~~~~~cato m are no standardized and sufficiently specific myco- > o _ bacteriophages that would enable a reliable differen- tiation of mycobacterial species to be made. This is co because most phages are polyvalent. o I- D If it were possible to surmount this obstacle by using a whole battery of different phages simulta- neously, as first described by Baess (1966, 1969) and LA fin le later by Bates & Fitzhugh (1967), who succeeded in differentiating several subgroups of M. tuberculosis c o on the basis of their susceptibility to phages, substan- 0 cn 0 x xtial success would have been achieved in the tax- co XCao_O onomy of M. tuberculosis. These mycobacteria form 60 PHAGE-TYPING OF MYCOBACITERA Table 3. Lysis produced in the host strains by phages in lyophilized form and by 2 different doses of phages in liquid suspension (results from Tokyo) Host Phages strain DS6A AG-1 BG-1 GS4E BK-1 D-34 in lyophilized form H37Rv ++ ++ ++ ++ +- H37Ra ++ ++ ++ ++ BCG ++ ++ - A21 ± ++ F-130 - - - - - + Me-153 - - + in liquid suspension 1OxRTD 1xRTD 1OxRTD IxRTD 1OxRTD IxRTD 1OxRTD 1xRTD 1OxRTD IxRTD lOxRTD 1xRTD H37Rv ++ + + - + + + - + - - - A21 - - + i F-130 - - - - - - - - - ++ ++ Me-1 53 ++ ++ + +, confluent lysis; +, semiconfluent lysis; ±, less than 10 independent plaques; , no lysis. a homogeneous group and it has not been possible so far to differentiate them by serological or cyto- chemical tests, in contrast to M. avium, for which serotyping according to the method of Schaefer (1965) has become a routine laboratory method. In order to reduce the variability of results to the absolute minimum, an attempt was made in the present study to standardize the basic trial condi- tions. The strains of M. tuberculosis used in the tests were accurately identified and verified as such. All the strains had been freshly isolated from persons in whom tuberculosis had been newly detected and who had not been previously treated. The strains were susceptible to the classical antituberculosis drugs and they were nitrate-positive and peroxidase-posi- tive. They were normally virulent, causing progres- sive disease in guineapigs into which a dose of 50000-500000 viable microbes had been injected subcutaneously. On both Lowenstein-Jensen and Ogawa media, they showed typically eugonic growth. The strains isolated in Pisa were subjected, in all the laboratories except Prague, to the same tests as had been carried out in Prague, with similar results. Thus both lots of strains may be designated reliably as M. tuberculosis. The myco- bacteriophages used in the trial had all been prepared as ready test suspensions at the same laboratory (Atlanta), which dispatched them to the cooperating laboratories either in screwcapped bottles or flame- sealed ampoules. No substantial differences in sen- sitivity between these phages were found, and the same applies to the phages that had been lyophilized in 5% sodium glutamate and 0.5% gelatine. In Utrecht, where the lyophilization was carried out, only phage GS4E had a reduced titre after 6 months' storage in the dark and at room temperature, where- as all the others remained stable during the observa- tion period. As GS4E is not particularly effective at producing lysis in M. tuberculosis, its stability after lyophilization is not really significant for the present study, even though it may be assumed that this property could be enhanced by selecting a more suitable vehicle for lyophilization. Another factor contributing to standardization was the use, by all the participating laboratories, of the same dehydrated Redmond RVA and RVB media. Some of the laboratories also prepared their own media according to Redmond's recipe and conducted comparative trials with them. In Copen- hagen, differences were found in the sensitivity of RVB media, on which some strains, either sensitive or resistant to phage BK-1, failed to grow. Other- wise, the results obtained with these media were satisfactory in all the cooperating laboratories except 5 61 L. §ULA ET AL. Pisa, where 7 out of the 36 Czechoslovak strains failed to grow altogether or grew so sparsely that the appropriate tests could not be performed. As all the strains exhibited abundant growth on Lowenstein- Jensen medium, the question of an optimum medium for phage-typing tests remains open. Comparison of the results obtained by the partici- pating laboratories leads to the conclusion that phages DS6A, AG-1, and BG-1 were the most suitable for tests of phage lysis in M. tuberculosis in the experimental conditions of the present study, since they produced lysis in all the tested Czecho- slovak and Italian strains when used in either of the doses designated as 1 x RTD and 10 x RTD. Evi- dence exists from another study (Sula, 1968), in which strains of M. tuberculosis from Africa (Tunis), Asia (Japan), and South America (Brazil and Vene- zuela) were used, that the above-mentioned phages produce lysis. Furthermore, those studies revealed that mycobacteriophage D-29 has an equally strong lytic effect. However, phages GS4E, BK-1, BG-1, and D-34 did not produce + + + (almost confluent) lysis in M. tuberculosis strains when spotted on a mycobacterial lawn in either of the doses called 1 x RTD and 10 x RTD. This factor is highly important because it allows of subdividing the spe- cies M. tuberculosis, since about 20-30% of strains undergo almost confluent lysis when tested in this manner (Bates & Mitchison, 1969). The efficacy of the other phages used in the present cooperative studies was low and also, judging by the experience of the participating laboratories, variable. Therefore the results obtained with them cannot yet be used for a reliable differentiation of the phage-susceptibility or phage-resistance of M. tuberculosis. On the other hand, all the cooperating laboratories found that phage GS4E-which in Redmond's view exerts a lytic effect on M. tuberculosis and is ineffective against M. bovis-was not sufficiently specific (i.e., it did not consistently cause lysis in M. tuberculosis and failed to produce it in M. bovis). Thus, out of a total of 35 genuine strains of M. tuberculosis, as many as 29 were resistant to this phage in Copenhagen and, in Budapest, 25 out of 29 test strains were GS4E- resistant. Phages DS6A and AG-1 were the most effective also against Italian bovine strains, some of which were lysed in addition by phage GS4E, as was the case with some strains of M. tuberculosis. It is therefore not possible to draw a reliable distinction between the two species on the basis of their phage- susceptibility by using the above-mentioned phages and Redmond's tables, as previously communicated by Murohashi (1963) and Caroli et al. (1967). The Dutch avian mycobacteria investigated in the present study were predominantly resistant to the phages in question, with the exception of phage AG-i-which bears out the general experience with this species. There appears to be an urgent need to isolate new phages effective against M. tuberculosis before at- tempting the systematic phage-typing of this group of mycobacteria. Mycobacteriophage BK-1 is worthy of particular attention. Baess was able, by means of this phage, to separate 205 strains of M. tuberculosis into 62 strains susceptible to it and 143 strains resistant to it. Out of 88 groups of patients connected epidemiologically, 22 groups excreted only phage- sensitive strains and 61 groups only phage-resistant strains. All 25 of the bovine strains isolated by Baess were resistant to phage BK-1-a finding that tallies with the results obtained with this phage in the present study in respect of the lysis of Italian bovine myco- bacteria. However, BK-1 proved to have little effi- cacy against Czechoslovak and Italian strains of M. tuberculosis, so that a detailed analysis of these strains could not be performed. One of the possible explanations for this discrepancy is that Redmond used a host strain different from that used by Baess, thus substantially altering the lytic properties of the phage. On the other hand, the strong lytic effect of phage BG-1 on M. tuberculosis points to the close relationship of the latter with M. kansasii, which is the host strain for phage AG-1. REtSUMIt ETUDES COLLECTIVES DE L'OMS SUR LE TYPAGE DES MYCOBACTERIES PAR LES PHAGES: 1. LYSE PAR LES PHAGES DE SOUCHES TCHECOSLOVAQUES ET ITALIENNES DE MYCOBACTERIUM TUBERCULOSIS Grace a la collaboration de plusieurs laboratoires spe- Les etudes ont porte sur les methodes les mieux adaptees cialises, on a recherch6 les possibilites d'utilisation des de conservation et de transport des mycobacteriophages, phages en vue de la classification des mycobacteries. sur la sensibilit6 des milieux de culture et d'epreuve et 62 PHAGE-TYPING OF MYCOBACTERIA 63 sur la technique des tests. Les essais ont ete effectu6s sur des souches de Mycobacterium tuberculosis en pro- venance de Tchecoslovaquie et d'Italie. Les mycobact6rio- phages ont et6 pr6par6s soit en suspension soit sous forme Iyophilis6e. Les resultats d'ensemble montrent la n6cessit6 d'uni- formiser au maximum les conditions des 6preuves si l'on veut utiliser les phages pour le typage des mycobact6ries. Les milieux RVA et RVB de Redmond ont donne de bons r6sultats dans tous les laboratoires sauf un. Les phages DS6A, AG-1 et BG-1 se sont r6vel6s les plus appropries aux 6preuves de lyse de Myco. tuberculosis, mais il est urgent d'isoler de nouveaux phages actifs contre cette mycobact6rie avant d'envisager un typage syst6matique du groupe. II semble que le meilleur proc&16 de pr6para- tion des phages en vue du transport est la lyophilisation qui assure un degr6 maximal de viabilit6. REFERENCES Baess, I. (1966) Amer. Rev. resp. Dis., 93, 622-623 Baess, I. (1969) Acta path. microbiol. scand., 76, 464-474 Bates, J. H. & Fitzhugh, J. K. (1967) Amer. Rev. resp. Dis., 96, 7-10 Bates, J. H. & Mitchison, D. A. (1969) Amer. Rev. resp. Dis., 100, 189-193 Caroli, G. et al. (1968) In: Proceedings of the Symposium on Bacteriophages of Mycobacteria, Bilthoven-Utrecht, 20-30 September 1967, Utrecht, National Institute of Public Health of the Netherlands, p. 32 Murohashi, T. et al. (1963) Amer. Rev. resp. Dis., 88, 664-669 Redmond, W. B. & Ward, D. M. (1966) Bull. Wld Hlth Org., 35, 563-568 Schaefer, W. B. (1965) Amer. Rev. resp. Dis., 92, 85-93 gula, L. (1963) Bull. Wld Hlth Org., 29, 589-606 gula, L. & Langerova, M. (1963) Bull. Wld Hlth Org., 29, 579-588 gula, L. & Sundaresan, T. K. (1963) Bull. Wid Hlth Org., 29, 607-623 Sula, L. et al. (1968) Rozhl. Tuberk., 28, 302-310 Annex I COOPERATING LABORATORIES WHO International Reference Centre for the Diagnosis of Tuberculosis Institute of Hygiene and Epidemiology Srobirova 48 Prague 10 Czechoslovakia (Dr L. gula) Tuberculosis Research Laboratory, Veterans Administra- tion Atlanta, Ga. USA (Dr W. B. Redmond) National Institute of Public Health Utrecht Netherlands (Dr J. C. Coster & Dr H. W. B. Engel) State Serum Institute Copenhagen Denmark (Dr I. Baess ) Medical Research Centre Little Rock, Ark. USA (Dr J. H. Bates) Institute of Hygiene University of Pisa Italy (Professor G. Caroli) Royal Edward Chest Hospital Montreal Canada (Dr E. Mankiewicz) National Institute for Tuberculosis Tokyo Japan (Dr T. Murohashi) Koranyi State Institute for Tuberculosis Budapest Hungary (Dr E. Vandra)