INFFCTION AND IMMUNITY. May 1974. p. 843-850Copyright ( 1974 American Societv for Microhiologs

Vol. 9. No. 5Printed in U.S.A.

Studies on the Effect of Starvation on MycobacteriaW. NYKA

Veterans Administration Hospital. Baltimore, Marnland 21218

Received for publication 29 November 1973

Ten cultures of Mycobacterium tuberculosis, one of' MNcobacterium kansasii(nonsignificant), and one of Mycobacterium phlei were submitted to starvation.As a result they lost first their acid fastness and then all other staining affinitiesbut, in this chromophobic state, they survived for at least 2 years and, after thattime, produced cultures of acid-fast bacilli when transferred onto nutrient media.Chromophobic tubercle bacilli similar to those produced experimentally hadpreviously been demonstrated in caseous lesions of lungs surgically removed frompatients under chemotherapy. Since it has been shown that experimentallyproduced chromophobic bacilli can recover their original biological properties,the opinion is warranted that, under suitable conditions, those in the lung couldalso become reactivated and cause a relapse of the disease.

In previous studies. tubercle bacilli of a newtype have been described in the lungs ot' pa-tients treated by drugs in association withsurgery (7, 8). Although morphologically similarto the acid-fast bacilli, these organisms do notstain with either carbolfuchsin or the counter-stains when applied by the classic Ziehl-Neelsen technique or with any other anilinedyes, and can only be colored by the newlydeveloped oxidative procedures described inthese papers. The demonstration of these "chro-mophobic" bacilli in sections of the lungs ot' allthe patients examined (10) called for theirfurther investigation, aiming at the in vitrodevelopment of live bacilli of the same type.Experiments designed to this effect carried outin this laboratory yielded results which arerecorded in the present study.

MATERIALS AND METHODSBacilli starved in plain agar. Two strains of

virulent tubercle bacilli taken from the stock of theclinical laboratorv of this hospital were used for theexperiment. They were grown in Dubos liquid me-dium with Tween and albumin, and a series ofscrew-capped tubes (150 by 25 mm) of each strain wasprepared. After 4 weeks the cultures were spun downand washed three times with distilled water to removeall traces of the culture medium, the sediment wassuspended in 2 ml of distilled water, and smearpreparations were taken for control. Then a series of50-ml conic glass centrifuge tubes were filled with 15ml of 3% agar in distilled water, covered with alumi-num foil to prevent contamination, sterilized, and putinto a 55 C water bath in order to keep the agar fluid.Next the suspension of bacilli of each tube wastransferred with a sterile pipette into a tube of agar,the mixture was vigorously stirred with the pipette to

secure an even distribution of the bacilli in the jelly,and the tubes were closed with sterilized, new, tightlvfitting rubber caps. After solidification of the agar,one tube was set apart for control and the others wereincubated at 37 C.The control tube was processed as follows. The

block of agar was taken out of the tube with sterileprecautions, transferred into a large petri dish, andcut into slices about 3-mm thick, the largest of whichwere divided into halves or quarters to obtain thenumber of cuts needed for the experiment. One slicewas fixed directly in 10% formol after smear prepara-tions had been made with scrapings from its surface,the remaining slices were each dropped into a screw-capped tube (150 by 25 mm) of Dubos broth, and thewhole series was incubated for 30 days at 37 C. Thetubes were inspected daily and gently shaken individ-ually. and any changes in the agar or the mediumwere noted. The incubated cuts were processed in asimilar way, every other day during the first 2 weeksand thereafter at 4-day intervals until the 30th day.Checks of the medium were taken of tubes incubated14 days or more. The formolized agar blocks weredehydrated with alcohol, embedded in paraffin, andcut in sections of 5-,m thickness which were stained,together with the smear preparations, overnight atroom temperature in carbolfuchsin, decolorized with5% HNO3 in 90% alcohol, counterstained with Loefflermethylene blue, rinsed, dried, and examined. Nosystematic studies with oxidation were carried outbecause trial experiments had shown that oxidizedsections of agar had a tendency to come off the slides.The remaining tubes of plain agar were processed in

the same way as the control tube, at monthly intervalsfor the first 4 months of the experiment, and thenduring the following 8 months at bimonthly intervalsand every 3 months during the 2nd year. To avoiddisruption of the series by contamination of thecultures, very careful technical work was mandatory.

Bacilli starved in distilled water. Ten 4-week-old4;3

on May 27, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

INFECT. IMMUNITY

cultures of virulent tubercle bacilli (no. 1, 2, 5, 10, 14.16, 18, 19, 24, 26), one culture of' Mvcobacteriumkansasii (nonsignif'icant) taken from the stock of thebacteriology laboratory of this hospital, and a cultureof Mycobacterium phlei, each grown in two tubes (150by 25 mm) of Dubos broth, were spun down andwashed three times in sterile distilled water. Thesediment of' each tube was suspended in 15 ml of'distilled water and, after a smear preparation wastaken from each suspension, the tubes were screwcapped and incubated at 37 C. being shaken atintervals to keep the sediment in suspension. Tofollow the reaction of' the organisms to adverse envi-ronment, checks were taken f'rom one tube of eachsuspension everx month during the f'irst 6 months,every other month during the f'ollowing 6 months, andevery 3 months during the 2nd year of the experiment,the duplicate tubes being kept in reserve. The prepa-rations were air-dried and f'lamed, and one of' theslides taken from each tube was oxidized overnight in10%. periodic acid. The next day the oxidized serieswas rinsed, first in tap water, and then in distilledwater, and then both series were stained with carbol-f'uchsin in the usual way, examined under a lightmicroscope and, in the case of the nonoxidized series.also under a phase-contrast microscope. Three addi-tional films of' each suspension were examined, oneafter staining with Gram stain and the others af'ternegative staining with India ink or nigrosin. Todetermine the proportion ot' the acid-f'ast, non-acid-f'ast, and chromophobic bacilli. 100 cells were countedin three different fi'elds of the films prepared fromeach suspension, and the average number was calcu-lated. In addition to the microscopy examination,after 6. 12. 18, and 24 months, the suspensions werealso planted on 7H10 agar, the development of thecultures was observed, and at'ter 1 month two smearpreparations were taken from each culture. Thesepreparations were stained with carbolf'uchsin in thesame way as the monthly checks, one directly and theother after oxidation.

RESULTSBacilli starved in plain agar: (i) effect of

starvation. Sections through the nonincubatedcontrol cuts of' agar revealed a bacterial popula-tion in no way diff'erent from that of the originalculture, consisting of' single and agglomeratedacid-fast bacilli interspersed with a few bluecells uniformly scattered throughout the entiresections. In blocks of bacilli starved for 1 monththere was a slight increase in the proportion of'the non-acid-fast forms as compared with thepicture seen in the nonincubated cut. Thisincrease was more accentuated after 2 and 3months of' starvation, at the end of' which timethe bacilli, whether single or agglomerate, wereeither red, blue, or unstained, the latter beingvisible only in light dimmed down by loweringthe condenser. Some of' the stained rods werecolored homogeneously with either fuchsin or

methylene blue; in others the staining wassegmented or spotty and frequently limited toone or both ends of' the cells. Big, dense clumpsof' bacilli were acid f'ast, but in loosely struc-tured clusters there was an admixture of bluebacilli. Taking the staining of the bacterialpopulation of' the control block as a standard, itcan be said that, after 4 months of starvation inplain agar, roughly 60%( of' the bacilli remainedmore or less acid fast, 20%7 became non-acidfast, and about as many remained unstained.With the passing of time the alterations ofstaining became more extensive, so that after 6months 50% of the bacilli were acid fast and theremainder were either non-acid fast or un-stained. At this stage the originally predomi-nant acid-f'ast big clumps consisted of red, blue,and clear areas and had a checkered appear-ance. On close examination these clear areashad a faintly visible fibrillar structure, whichalso characterized smaller aggregates whichwere clear in their entirety, unmistakable evi-dence that they were composed of' unstainedbacilli. Between 6 and 12 months of' starvation,the proportion of' stained to unstained bacillidropped further, being about 10: 90% at the endof' this time. Only an occasional one of thestained bacilli was well colored with eitherfuchsin or methylene blue, some of the othersappeared as red or blue segments of rods,whereas others still could be identif'ied only bytheir polar bodies (Fig. 1). Small bunches werepredominantly unstained and only in some of'the larger unstained clumps could a few acid-f'ast bacilli be recognized. After 15 months nosingle stained cells were found, and after 18months only a few clumps were seen as clear orfaintly blue masses dotted with a few red spots.The last 6 months of the experiment broughtfurther loss of' the staining capacity of thebacilli, so that after 2 years of starvation only afew of the big clear clumps showed one orseveral short, slender pink or red rods, whichwere very likely segments of otherwise un-stained bacilli.

(ii) Bacilli in nutrient broth (macro-scopic). The first signs of growth in the culturesof the control cuts and in those of bacilli thathad been starved from 1 to 6 months were ob-served after 6 days of incubation when the me-dium started to become opalescent. Henceforththere was a delay in the development of the cul-tures which increased with the duration of star-vation to which the bacilli had been submitted.Bacilli starved for 10 months started growingafter 9 days in the incubator, those starved 12months started growing after 2 weeks, and

844 NYKA

on May 27, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

EFFECT OF STARVATIMN ON MYCOBACTERIA

_

FIG. 1. Chromophobic bacilli. On the left side,below crystalloid, group of refringent rods of normalshape and size. Note that, although bathed by meth-ylene blue, the cells remained unaffected by that dyejust as they had been by carbolfuchsin. On the rightside, cluster of faintly visible bacilli with prominentpolar bodies. Smear preparation of agar taken after 10months of starvation of the bacilli, stained withcarbolfuchsin, and counterstained with Loeffler meth-ylene blue. x1,100.

those starved 18 months showed signs of growthafter 3 weeks. After 24 months of starvation,the medium began to opacify after 4 weekswithout reaching the usual turbidity even after2 months of incubation.

(iii) Bacilli in nutrient broth (micro-scopic). Microscope examination of the brothin 21-day-old cultures induced by nonincubatedcuts of agar revealed the usual bacterial pop-ulation consisting of acid-fast bacilli inter-spersed with a few blue rods. Checks takenfrom cultures resulting from bacilli starved from1 to 12 months showed no change in either theirmorphology or staining. However, from thenonward there was a slight but progressive de-crease in the density of the opacification of thebroth accompanied by an increase in the pro-

portion of blue cells, noticeable in culturesresulting from bacilli starved for 18 months andstriking in those exposed for 2 years to starva-tion.

(iv) Bacilli inside agar cuts (macroscopic).The macroscope follow-up of the agar blocksthemselves was as follows. Cuts incubated innutrient broth immediately after solidificationof the agar and those seeded with bacilli thathad been starved from 1 to 6 months remainedunchanged during the first days, but after a

week they became opaque at the periphery,remaining clear in the center. During the follow-

ing days the opaque marginal zone resolvedprogressively into a multitude of tiny, grey,ill-defined flakes, distinctly visible with thenaked eye after 10 days of incubation. After 2weeks the flakes became larger and betterdefined, and after 4 weeks they appeared assharply outlined bodies, either rounded, elon-gated, or angular. Densely structured at theperiphery of the blocks, which had been bathedby the broth, they became less numerous andless compact as their distance from the marginincreased. Similar observations were made withbacilli that had been starved for 12 and 18months, but in those instances the first flakeswere observed after 2 and 3 weeks, respectively,and at the end of the incubation those seen inthe latter were smaller than those in the former.Finally, in cuts with bacilli starved between 18to 24 months, flakes appeared only after 1month and were smaller and far less numerousthan those developing from bacilli starved forshorter periods.

(v) Bacilli inside agar cuts (microscopic).Sections through control blocks of agar incu-bated in nutrient broth taken during the firstfew days revealed the usual bacterial popula-tion consisting of single and aggregate predomi-nantly acid-fast bacilli. After a week there was aslight increase in the number of blue bacilliclose to the margin of the section, which after 2weeks was strewn with a multitude of smallcolonies of blue bacilli interspersed with a fewacid-fast rods. After 3 weeks the colonies hadincreased in size and after a month they hadreached their full development. At that time thelarge colonies strewn throughout the peripheryof the sections and the increasingly smaller oneslocated in their depth, as well as the flakeslocated in the center, also consisted of bluebacilli with a small admixture of red cells.Sections through blocks seeded with bacillistarved for 1 month and then incubated innutrient broth showed, during the first week, nochanges in the bacterial population other thanthose due to starvation, i.e., an increased num-ber of blue bacilli. With the passing of time thebacilli continued to grow, and, after 2 weeks,colonies of predominantly blue bacilli devel-oped which continued to increase in size andwhich, after 4 weeks in the incubator, resembledin all respects those described in the previousseries. There was no significant difference in thedevelopment of bacilli starved for 1 month andthose kept under starvation for 12 months, butfrom then onward there was a progressive de-crease in the number and size of the colonies.Disc'ernible after 18 months, these alterations

845VOL. 9, 1974

on May 27, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

INFECT. IMMUNITY

became further accentuated in sections throughcuts of agar seeded with bacilli starved for 18 to24 months. However, even the few small colo-nies found in sections through cuts of bacillistarved for 2 years and incubated in nutrientbroth for 1 month or longer also consisted ofblue cells with an occasional red rod enmeshedin them.

It is worth mentioning that the irregularshape and ill-defined jagged outline of the smallcolonies developing deep in the agar cuts weredue to the pattern of growth of the peripheralbacilli, many of which were projecting into thesurrounding agar. Some of these cells werearranged into segmented rows, each segmentbeing the length of one average sized tuberclebacillus, which penetrated deep into the sur-rounding jelly (Fig. 2, inset). It is of interestthat the constituent bacilli in these rows wereeither red or blue and that there was no definiteorder in their location within the rows, a red rodbeing in continuity with one or two blue onesand vice-versa, with either a red or a blue cell atthe beginning or end of the row.During processing, some of the colonies in-

serted into the agar were partly or entirelydisintegrated and their bacterial content wasspilled out from their original site and scatteredthroughout the surrounding jelly. These spillsconsisted of morphologically normal and per-fectly stained blue bacilli and a few equally

well-stained, acid-fast rods. Cocco-bacillary andcoccoid forms were few, filamentous forms havenever been observed, and there was no trace ofclumping of the bacilli (Fig. 3).

Bacilli starved in distilled water: (i) M.tuberculosis. The most striking alteration ofthe tubercle bacilli starved in distilled waterwas the rapid decline of their acid fastness.Noticeable after 1 month, it was very extensiveafter 3 months of exposure, when only about 1/:3of the bacilli were more or less well stained withcarbolfuchsin, an about equal number werecolored with methylene blue, and the remainingones were unstained. Most of the single acid-fast bacilli were homogeneously stained a deepor pale red, the others were segmented orgranular, and the general picture of the bacte-rial population was reminiscent of that seen inaged cultures on solid medium, the more so asthe same irregularity of staining was observedamong the non-acid-fast bacilli. The chromo-phobic forms, visible only in dimmed light orunder the phase microscope, appeared as homo-geneous, translucent streaks, identifiable bytheir sharp outline, normal length, and charac-teristic curvature (Fig. 1). Small bunches ofbacilli were predominantly non-acid fast, biggerclusters consisted of both acid-fast and non-acid-fast bacilli, and big dense clumps wereeither predominantly or entirely acid fast. Notinfrequently clusters and clumps of bacilli con-

Ip

I

I_1AYVeF

.~

'I

FIG. 2. Section through an agar cut processed after 10 months of starvation of the bacilli and after 1 month ofincubation in nutrient broth. Colonies of blue bacilli scattered throughout the marginal zone of the section.Note the great difference in number, shape, size, and structure between the marginal colonies and thosedeveloping in depth of the cut. The free space surrounding the big dark clump is due to retraction of the agar.Stained as before. x450. Inset: One of the small blue colonies developing in the center of the cut, with rows ofbacilli protruding into the agar. Same section, same staining. x1,100.

846 NYKA

on May 27, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

EFFECT OF STARVATION ON MYCOBACTERIA

a

I .1

/

-. '\I

FIG. 3. Same section as before. Single blue cellsinterspersed with about 10 red rods, shaken out fromthe colonies during processing, and scattered through-out the agar. Note the homogeneous structure of thecells and the absence of clumping. Same staining.xl,100.

sisted of red, blue, and unstained areas, havingthe characteristic checkered appearance de-scribed in the agar series. It is worth emphasiz-ing that, at this early stage of starvation, theacid fastness of strain number 26 was slightlybetter preserved than that of the other strains.After 6 months, roughly 20% of the bacilli were

more or less acid fast, whereas after 9 monthsonly 10% retained carbolfuchsin, about as manywere blue, and the remainder were unstained.After 18 months only about 20 red cells could becounted in a high-power magnification field,whereas in suspension number 26 roughly 20%of the bacilli were normally acid fast. In smears

of nine suspensions of bacilli starved for 2 years,about 10 cells per high-power magnificationfield were counted. Strain numbers 19 and 24showed an occasional additional bunch of shortacid-fast rods and strain number 2 showed an

occasional additional bunch of normal-sizedbacilli, either fully or weakly acid fast. How-ever, in films of suspension number 26 about10% of the bacterial population was still acidfast.

(ii) M. kansasii. This microogranism, whichappeared better stained than M. tuberculosis atthe end of 3 months of starvation, showedgeneral weakening of its staining after 6months, but about 50% of the cells still ap-peared faintly acid fast. After 18 months, 30% ofthe organisms were weakly acid fast, and after 2

years 20% of the cells were red, 30% were blue,and 50% remained unstained.

(iii) M. phlei. The alterations of the acidfastness of M. phlei went parallel to those of M.tuberculosis except for strain number 26.During the first 18 months of the experiment

there was no parallel between the loss of acidfastness and the viability of the tubercle bacilliin all suspensions, which yielded cultures ofacid-fast bacilli in the usual time when plantedon 7H10 agar, and only during the final 6months were signs of a decline in their viabilityobserved. After 2 years of starvation, subcul-tures of strain numbers 1, 10, 14, 19, 24, and 26showed a delay of 1 week in the appearance ofgrowth on nutrient agar as compared with theiroriginal development on this same medium,whereas strain numbers 2, 5, 16, and 18 showeda delay of 2 weeks and produced less abundantcultures than the first group. In contrast to M.tuberculosis, M. kansasii and M. phlei showedtheir normal growth throughout the entirecourse of the experiment. In subcultures of allthe suspensions under study the bacilli weresimilar both in morphology and staining to theiroriginal cultures, and all responded well tooxidation.

Three months of starvation did not alter thereaction to oxidation of any of the suspensionsunder study, but from then onwards only theweakly acid-fast and non-acid-fast cells re-sponded to the treatment, whereas the greatmajority of the chromophobic ones remainedunaffected.Attempts to stain the entire bacterial popula-

tion present in the smears in order to gain ageneral idea of the number of bacilli actuallypresent in the films were not very successfulbecause of the extraordinary resistance of thechromophobic forms to aniline dyes. To date,the most satisfactory results have been obtainedwith Gram stain after which the bacilli, allgram-negative, were faintly stained with saf-ranin. More helpful than the Gram mixture orphase contrast was negative staining with Indiaink or nigrosin. Although not differential, thisprocedure gave valuable information about thenumber and morphology of the bacilli which,even after 2 years of starvation, were no differ-ent from those grown on the usual culturemedia.

DISCUSSIONThe most conspicuous alteration of the starv-

ing bacilli was the progressive loss of acidfastness. This peculiarity of staining was firstobserved in lepra bacilli (6) and a year later also

847VOL. 9, 1974

a jp

I,

I"I

on May 27, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

INFECT. IMMUNITY

in the newly discovered tubercle bacilli (4).Since that time great efforts have been made todetermine its nature and signif'icance, evi-denced by the long list of' publications on thesubject, which are analyzed in classic referencebooks of bacteriology. At the present time thisspecial staining affinity is generally ascribed tomycolic acid (1). However, this opinion hasbeen questioned on the grounds that mycolicacid is too weakly acid f'ast to account for thebrilliant red staining of the bacilli (11), and thatit is an oversimplification to ascribe their acidfastness to any single substance developed bythe mycobacterial cells (5).The loss of acid fastness became noticeable

very soon after the beginning of starvation,when an increasing number of bacilli turnednon-acid fast and stained blue, and then lost alltheir staining affinities and became chromo-phobic, a state reached by the majority of themafter about 6 months of exposure. From thenonwards the transformation progressed at anincreasingly slower pace, so that even after 2years of exposure a few bacilli still exhibitedsome degree of acid fastness. This sequence ofevents, in addition to showing the slow, progres-sive changes in the staining affinities of M.tuberculosis, also brings the demonstration thatthe non-acid-fast or blue bacilli are not artifactsor degenerating cells as has been suggested bysome authors (12), but are intermediate stagesbetween the acid-fast and the chromophobicbacilli.A key to the understanding of the chemistry

of these alterations is provided by the demon-stration that weakly acid-fast and even chromo-phobic mycobacteria can be made fully acidfast by oxidation (8), and that acid-fast bacillican be made chromophobic by reduction andhave their acid-fastness restored by oxidation(9). These experiments provide evidence thatoxygen plays a decisive role in the resistance of'mycobacteria to acid destaining, and that acid-fastness hinges on the oxygen potential of thecells. It can be postulated, therefore, that thewaning of the acid fastness of starving bacilli iscaused by a progressive inhibition of theirmetabolic activity which depresses and eventu-ally arrests the oxidative processes accompany-ing it, resulting in depletion of their oxygenpotential.The oxidation-reduction procedure (9) also

provides information about the significance ofacid fastness. Since the oxygen involved in acidfastness can be removed from the bacilli byreduction and restored to them by oxidationwithout any noticeable damage to the cells, it is

not likely to be chemically bound with theirsubstance but rather loosely connected with it.It seems, therefore, not to be directly concernedwith the growth of' the bacilli but rather withtheir respiration. This hypothesis finds supportin the observations of' earlier writers on therespiration of tubercle bacilli, which led them tobelieve that the consumption of oxygen and thesynthetic processes involved in the growth of' themicrobes depend on diff'erent metabolic systems(5).There were signif'icant differences in the re-

sistance of the acid f'astness of the variousstrains of mycobacteria to starvation, and therewas no parallel between the loss of' acid fastnessof starved bacilli and their viability. Tuberclebacilli starved in agar were but slightly dimin-ished in their viability at a time when only 10%of them were acid f'ast, whereas those kept indistilled water yielded normal cultures after 18months, when only 20 red cells were counted permicroscope field in the inoculum, and viablebacilli have been demonstrated after 2 years of'starvation in all the suspensions of bacilli understudy. In the distilled water series, nine strainsof tubercle bacilli and M. phlei lost their acidfastness more readily than strain number 26 of'M. tuberculosis, which showed 10' of' acid-fastbacilli after 2 years but initiated growth of'subcultures at about the same time as the otherstrains of M. tuberculosis. M. kansasii, on theother hand, with 20% of' acid-f'ast bacilli at theend of the experiment, and M. phlei, with only afew red cells left, both yielded subcultures asoriginally. This shows that there were greatbiochemical and metabolic differences not onlybetween the various species of' mycobacteriaunder study, but also between individual cellsof the same suspension, some becoming chro-mophobic during the f'irst months of' starvationand others showing some degree of acid f'astnesseven after 2 years of' exposure. Similar dif'fer-ences are regularly observed not only in agedcultures of tubercle bacilli, as has previouslybeen mentioned, but also in routinely stainedfilms of regular cultures, which always showvariable proportions of' acid-f'ast, non-acid-f'ast,and chromophobic bacilli, thus reflecting themetabolic state of' the cells at the time of'processing.

In summary, the acid t'astness of mycobacte-ria seems to be due to the oxygen potential of'the bacterial cells, which is the result of' theactive oxidative processes accompanying theirmetabolism and varies with their metabolicactivity. It is primarily related to the respirationof the cells, which is closely connected with

848 NYKA

on May 27, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

EFFECT OF STARVATION ON MYCOBACTERIA

their general metabolism. Actively metaboliz-ing mycobacteria with uninhibited respirationare fully acid fast, less active cells are weaklyacid fast or non-acid fast, and those witharrested metabolism are chromophobic. Theopinion is warranted, therefore, that acid-faststaining, currently used for coloring mycobacte-ria, is also a test indicative of the metabolicstate of the cells.The most important result of the starvation

experiments is the demonstration that metabol-ically inactivated tubercle bacilli lose theirstaining capacity but remain alive for a longperiod of time and recover their original biologi-cal properties when transferred into a favorableenvironment. Their remarkable resistance toadverse conditions and their reversibility are offundamental importance for the bacteriology.pathology, and treatment of tuberculosis. Inbacteriology, they show that M. tuberculosis isa two-phase microorganism which, dependingon its environment, can appear either in itsmetabolically active acid-fast form or in itsinactive chromophobic form. This biologicalcharacteristic is reflected in the pathogenesis oftuberculosis. During the acute, exudative phaseof the infection, which provides a favorableenvironment for their development, the bacilliare acid fast, but in the caseous lesions, anotoriously poor culture medium characteriz-ing the chronic phase of the disease, theybecome predominantly chromophobic in bothtreated and untreated lungs (2). The indiffer-ence to drugs of the chromophobic bacilli hasbeen proven by their demonstration in all thelungs of treated patients examined, no matterwhat drugs or drug regimens were used, what-ever the duration of the treatment or whetherthe bacilli were originally drug sensitive or drugresistant (10). In the chemotherapy of tubercu-losis the biphase nature of M. tuberculosis isreflected in the well-known favorable influenceof the drugs in fresh exudative lesions, wherethey interfere with the metabolism of the acid-fast bacilli, and their limited success in chronicproductive tuberculosis in which the metabolic-ally inactive chromophobic forms predominate(3). Finally, the reversibility of the bacilli comesinto play when, in suitable conditions, thechromophobic forms can recover their originalproperties and cause relapse of the disease.

It is generally believed that only the acid-fastbacilli, the vegetative form of M. tuberculosis,can reproduce on artificial media. This opinionis refuted by the development of bacilli starvedin plain agar after transfer of the cuts intonutrient broth. Soon after immersion the chro-

mophobic bacilli enclosed in the agar began togrow, but instead of yielding acid-fast bacillithey developed predominantly blue forms. Themain stages of this development are illustratedby Fig. 1 to 3 taken from color films which,unfortunately, had to be reproduced in blackand white. The agar technique makes it possibleto follow the development of the colonies stepby step, showing such features as the multi-plication of the bacilli by transverse fission andthe development and composition of the bluecolonies. The frequently observed disruption ofblue colonies by processing and the dispersion oftheir bacterial content indicate that blue bacilligrowing in the conditions described do notdevelop any cord substance, which is one of thefeatures differentiating them from colonies ofacid-fast bacilli.The in vitro development of blue bacilli is a

demonstration of the decisive role of nutritionnot only in the growth of tubercle bacilli butalso in the development of their acid fastness.When a block of agar seeded with bacilli madechromophobic by starvation is transferred intonutrient broth, two different cultures developconcurrently in the same tube. One is thatinduced by bacilli spilled off by the broth fromthe surface of the cut and the other by thosedeveloping inside the jelly. Both originate fromchromophobic bacilli but, whereas the formerare in direct contact with the medium and areprovided with all the nutrients they need forboth growing and developing their special stain-ing affinities, and become acid fast, the latterare separated from it by the agar envelopingthem and are either reached by the broth inan insufficient amount or receive it altered inits composition through loss of some ingredi-ents and, therefore, grow predominantly non-acid fast. That this difference is due to nutri-tion and not to the direct action of the agar orto lack of contact with the air is demonstratedby as yet unpublished studies on the growth ofbacilli incorporated into solid nutrient media,which develop small colonies of acid-fast bacilliinside the jelly. These observations show theextreme selectivity of the growth requirementsof the bacilli. They also indicate that, on thestandard media now in use, mycobacteria donot grow acid fast because they are inherentlyso, but because they find there the nutrientsthey need for both their growth and the de-velopment of acid fastness. It can be assumed,therefore, that in adequately adjusted mediait might be possible to grow M. tuberculosis invitro not only in its acid-fast form but also inits non-acid-fast and even chromophobic forms.

VOL. 9, 1974 849

on May 27, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

INFECT. IMMUNITY

Various morphologically abnormal forms ofmicrobes such as the L-forms and others havebeen described during the past decades but,since their meaning has not been determined, itis difficult to tell whether the chromophobictubercle bacilli belong to that group. It iscertain, however, that striking similarities existbetween these and the abnormal forms whichare found in the late or subclinical stage ofinfectious diseases such as syphilis, brucellosis,and typhus, which the present writer has alsohad the opportunity of studying. For instance,in a plaque of a syphilitic aorta, forms ofspirochaetes were found which were so differentfrom the classic shape of Treponema pallidumthat their true nature could only be determinedthanks to the presence of a few recognizableparasites scattered among them. Their abilityto maintain a subclinical infection for manyyears and their indifference to drugs suffice toshow the importance of these abnormal formswhich, to underline their development by adap-tation to an increasingly adverse environment,could be called "adaptation forms," or "a-forms" for short. Since the transformation ofthe original microorganisms into their a-forms isdue to inhibition of their metabolism, as a resultof which they lose some of their non-essentialfeatures and become indifferent to drugs, mi-crobes should be investigated and tested notonly in their actively metabolizing state but, todetermine all their potentialities, also whentheir metabolic activity is at its lowest. Suchstudies are particularly indicated at the present

time when the increasing resistance of themicrobes to drugs threatens to develop into aserious challenge to the chemotherapy of infec-tious diseases.

LITERATURE CITED

1. Anderson, R. J. 1932. The chemistry of the lipoids oftubercle bacilli. Physiol. Rev. 12:166.

2. Canetti, G. 1955. The tubercle bacillus in the pulmonarvlesions of man, p. 177. Springer Publishing CompanyInc., New York.

3. Chun, Y. E., C. H. J. Chang., and W. Nyka. 1972. Newobservations concerning the development of resistantbacilli in pulmonary tuberculosis. Clin. Med. 79:15-20.

4. Ehrlich, P. 1882. Neue Methode zur Entdeckung desTuberkelbazillus. Deut. Med. Wochenschr. 8:269.

5. Long, E. R. 1958. The chemistry and chemotherapy oftuberculosis, 3rd. ed., p. 8 and 64. The Williams andWilkins Company, Baltimore.

6. Neisser, A. 1881. Weitere Beitraege zur Aetiologie derLepra. Virchows Arch. 83:514.

7. Nyka, W. 1963. Studies on mycobacterium tuberculosisin lesions of the human lung. A new method of stainingtubercle bacilli in tissue sections. Amer. Rev. Resp.Dis. 88:670-679.

8. Nyka, W. 1967. Method for staining both acid-fast andchromophobic tubercle bacilli with carbolfuchsin. J.Bacteriol. 93:1458-1460.

9. Nyka, W. 1971. Influence of oxidation and reduction onthe acid-fastness of mycobacteria. Infect. Immunity4:513-515.

10. Nyka, W., and E. F. O'Neill. 1970. A new approach to thestudy of non acid-fast mycobacteria. Ann. N.Y. Acad.Sci. 174:862-871.

11. Rich, A. 1951. The pathogenesis of tuberculosis, 2nded., p. 15. Charles C Thomas, Springfield, Ill.

12. Yegian, D., and K. R. Porter. 1944. Some artifactsencountered in stained preparations of tubercle bacilli.I. Non-acid-fast forms arising from mechanical treat-ment. J. Bacteriol. 48:83-92.

850 NYKA

on May 27, 2021 by guest

http://iai.asm.org/

Dow

nloaded from