A Theory of Immunity in Tuberculosis

Florence B. Seibert

Perspectives in Biology and Medicine, Volume 3, Number 2, Winter 1960,

pp. 264-281 (Article)

Published by Johns Hopkins University Press

DOI: https://doi.org/10.1353/pbm.1960.0016

For additional information about this article

https://muse.jhu.edu/article/404648/summary

Access provided at 2 Jan 2020 13:35 GMT from University of Toronto Library
A THEORY OF IMMUNITY IN TUBERCULOSIS

FLORENCEB. SEIBERT, Ph.D., ScD. [Hon.), LL.D. {Hon.)*

This article presents a new theory of immunity in tuberculosis—one

which may in part be applicable to certain other diseases. It has evolved
from facts gathered over a period of thirty years of experimental work,
many of which were puzzling as they emerged. Now, in the light of the
concept here given, they appear to form integral parts ofa logical mosaic.
Immunity in tuberculosis has many strong links with immunity in gen-
eral, and distinguished leaders have often called attention to the discrepancy
between recognizable antibody concentration and actual resistance in this
as well as certain other diseases. Because ofthis discrepancy, humoral anti-
bodies have been thought by many investigators to be oflittle importance
in resistance to tuberculosis and certainly less than other possible mecha-
nisms. However, it will here be emphasized that, without the help of cer-
tain antibodies, the natural defenses ofthe body cannot protect against in-
vasion by the tubercle bacillus. Moreover, at times these antibodies may
be masked.
Polysaccharides have been of great interest in immunology, especially
in the role ofhaptenes. As antigens alone, or when combined with lipids
or proteins, they have also received much attention as immunizing agents.
The important dual role played by the tuberculopolysaccharide in inter-
fering with the natural defenses of the body and, on the other hand, in
eliciting antibodies which are able to eliminate this interference will be
especially stressed in this paper as a vital mechanism in the immune process
* Professor of Biochemistry, Henry Phipps Institute, University of Pennsylvania, Philadelphia,
Pennsylvania. Special Consultant for U.S. Public Health Service. Present address: Box 471, Route 1,
Clearwater, Florida. The experimental work of the author has been aided by grants from the Na-
tional Tuberculosis Association, through its Medical Section; the American Trudeau Society, made
possible by a special bequest from the estate ofSophie Goman; and by research grant No. E-680 from
the National Institute of Allergy and Infectious Diseases of the National Institutes of Health, Public
Health Service.

264 Florence B. Seibert · Immunity in Tuberculosis

Perspectives in Biology and Medicine · Winter 1960
in tuberculosis. Such an interplay offactors bears a strong similarity to the
well-known relationship between the capsular polysaccharides of the
pneumococcus and streptococcus and resistance of the host to these
organisms.
It may also be profitable to consider the possibility of similar mecha-
nisms in some other baffling diseases, such as infection with atypical acid-
fast organisms or certain fungi, Hodgkin's disease, lymphoblastoma, sar-
coidosis, and even cancer, where certain uncommon acid-fast organisms
have recently been found to occur and where polysaccharides have already
had serious consideration for therapy.
1.Consideration ofExperimental Data Available
A. antibodies

i. Presence.—Antibodies related to tuberculous disease present in the

blood stream can be detected by tests utilizing relatively pure tuberculo-
protein and tuberculopolysaccharide antigens which are available. In gen-
eral, they are not found in uninfected rabbits, guinea pigs, or humans (i).
Following natural or experimental infection with virulent or avirulent
organisms, such antibodies may or may not be detected regularly, because
at times they may be present and again they may be absent in the same
animal—a fact which has made their serological detection by many differ-
ent methods an uncertain diagnostic or prognostic procedure. A positive
test undoubtedly means that infection has occurred, but a negative test
does not necessarily mean the opposite.
Under certain circumstances, which will be discussed later, increases in
the gamma or alpha globulins of serum demonstrated by the electropho-
retic techniques were associated with the presence of antibodies. It was
possible to isolate these fractions from certain sera and to demonstrate by
standard procedures the presence ofspecific antibodies in them (2, 3). Not
only have antibodies been found in sera, but they have recently also been
shown to be present in caseous tissue (4). They are probably present in tis-
sues of sensitized animals in general, as evidenced by the Arthus skin re-
action, which is undoubtedly a local antigen-antibody reaction. The tuber-
culin reaction is probably the result of a modified local antigen-antibody
reaction, although its mechanism is still under investigation.
2.Induction.—Antibodies could be induced almost without fail in rab-
bits by repeated injections of purified tuberculoproteins or vaccines made

265
from heat-killed BCG and other organisms with or without the addition
ofan adjuvant (5). Antibodies so induced by tuberculoproteins which had
been isolated from unheated culture filtrates, and designated as A, B, and C
proteins because oftheir different physicochemical properties, were specific
for the antigens injected (6). There were, however, some cross-reactions
between the proteins, possibly due to the great difficulty of their clean
separation (7, 8).
By repeated injections with the large protein-free tuberculopolysac-
charide II, specific antibodies to it were also induced (9). However, the
smaller molecular polysaccharide I was never able to elicit antibodies but
acted only as a haptene. It readily precipitated antibodies in sera from rab-
bits sensitized by vaccine made from the whole bacillus. Sera from such
rabbits usually reacted with both types of polysaccharide.
3. Detection.—The most direct method for detecting tuberculous anti-
bodies was by the precipitin reaction. A very simple and practical qualita-
tive test involving the use of only about 0.5 cc. of serum has been em-
ployed (1). Added to each of four tubes of 0.1 cc. serum was 0.1 cc. of
purified protein antigen containing decreasing amounts of antigen—re-
spectively, 50, 10, 5, and ? gamma. To a fifth control tube 0.1 cc. saline
was added. Decreasing amounts of precipitate were obvious with this
range ofdilutions. The method can, ofcourse, be made more sensitive by
using closer decrements ofantigen, and even to the limit of antigen dilu-
tion where no precipitate occurs. However, until the protein antigens are
isolated in even purer state than they now exist (6), the quantitative tech-
nique ofHeidelberger (10), as it is applied to pure egg or serum-albumin
antigen-antibody systems, cannot be used, because precipitation often
occurs indefinitely on repeated additions of antigens.
Precipitation of antisera with tuberculopolysaccharide occurred usually
with a much higher antigen dilution, there being no precipitate on the
addition of 50 gamma, but increasing amounts of precipitate with de-
creasing amounts ofantigen. Thus the heaviest precipitate was given by 1
gamma or less of antigen (11). The same effect—solubility of polysac-
charide antigen-antibody complex in antigen excess—was conspicuous in
the case of the polysaccharide antigen in contrast to the protein antigens.
Differentiation could be made between the presence of antibodies to
tuberculoprotein and to tuberculopolysaccharide by this simple four-tube
test with these antigens. It is possible that a quantitative precipitin tech-

266 Florence B. Seibert · Immunity in Tuberculosis

Perspectives in Biology and Medicine · Winter i960
nique for use with the pure tuberculopolysaccharide antigen may be de-
vised in the future.
Complement fixation reaction was given mainly by certain isolated
tuberculoproteins, as, for example, PPD and C type of protein, but not
by A or B type proteins. Tuberculopolysaccharide, under similar circum-
stances, did not bind complement and, in fact, seemed to inhibit fixation
by PPD protein (12). The negative results with A and B proteins may also
have been due to the high content of polysaccharide in these proteins. In
fact, the unsatisfactory results with complement fixation tests in earlier
investigation may also have been due to this inhibition by polysaccharide
present in the impure antigens then used.
The hemagglutination test also has fallen short of expectations. The
antigen used in this reaction has been claimed to be polysaccharide (13),
and it undoubtedly is to a large extent, but protein also plays some role.
Agar-gel double diffusion techniques would seem to be the most prom-
ising methods for differentiating different protein from polysaccharide
antigen-antibody systems, but, until purer antigens are available, these
methods also will probably not be very helpful. By this technique polysac-
charide antigen-antibody bands were near the serum-agar interface (8),
and there were two zones of precipitation—which is consistent with the
fact that at least two types of polysaccharide have been found (9). How-
ever, the type of antibody to protein is very difficult to identify, because
several bands appeared with each type ofpurified protein. In spite of this,
some suggestive differences appeared (8), so more intensive work in this
field may be fruitful. It is true that, if infection occurred in such a way as
to cause the presence of antibodies in the serum demonstrable by other
methods, bands were exhibited by this agar-gel technique, but, as will be
pointed out in the next section, an imbalance in the proportion ofantigens
to antibodies could mask their presence.
Another more or less indirect method for detecting the presence ofanti-
bodies was by means ofelectrophoresis (3). For example, following repeat-
ed injections with tuberculoprotein into rabbits, it was found that in one
case the gamma globulin increased and in another the alpha2 globulin. In
both cases, when antiserum was precipitated with the antigen which had
been injected, a specific precipitate occurred. Moreover, when these pre-
cipitates were removed and the remaining supernatants studied again in
electrophoresis, there was a reduction in the respective gamma or alpha2

267
globulin peak, indicating that antibodies had been removed from them.
Furthermore, when the gamma globulin was isolated from these antisera
by electrophoresis or by method io of Cohn's procedures (14), it gave a
heavy antigen-antibody precipitate with the specific antigen.
4.Correlation with resistance.—One of the most significant results ob-
served was that not all rabbits reacted similarly to repeated injections with
BCG vaccine (1). Practically all rabbits were ultimately able to elicit de-
tectable antibodies to tuberculoprotein, although some more readily than
others. Some rabbits also exhibited high concentrations of antibodies to
tuberculopolysaccharide following repeated injections with BCG, while
others showed none or only low concentrations of these antibodies (11).
Resistance to subsequent infection with virulent Ravenel strain of bovine
tubercle bacilli appeared to be greatest in those rabbits which had been
able to produce antibodies to tuberculopolysaccharide and least in those
not able to develop them.
Those rabbits able to produce antibodies only to tuberculoprotein could
be made to produce even more ofthem following repeated injections with
protein, but, when they were challenged with virulent tubercle bacilli,
they developed massive caseation and died more quickly than untreated
rabbits. This was observed frequently (11,15) and, at first, posed one ofthe
puzzles concerning the relationship of antibodies to resistance. For ex-
ample, it was shown by Emmart and Seibert (16) that a gamma globulin
fraction, isolated from sera ofrabbits sensitized with tuberculoprotein and
shown to contain antibodies, inhibited the development of tubercles on
the chick membrane. It also inhibited the growth oftubercle bacilli in the
test tube, while the albumin fraction stimulated growth. However, the
same rabbits from which the sensitized sera were taken were not resistant
to challenge with virulent organisms, in spite ofthe presence ofantibodies
in their sera. Now it can be argued that, since these antibodies were mainly
antiprotein antibodies, it is evident that resistance was associated not
merely with the fact that antibodies were produced but with appearance
of the right kind of antibodies—namely, antipolysaccharide antibodies.
5.Effect on growth of tubercle bacilli.—Since certain antibodies obviously
play an important role in resistance, what is the nature ofthis relationship?
Early experiments cited above, showing that growth of tubercle bacilli
and even tubercle formation on the chick membrane were interfered with
by addition ofBCG antisera and by antibody-containing gamma globulin

268 Florence B. Seibert · Immunity in Tuberculosis

Perspectives in Biology and Medicine · Winter i960
isolated from sera of sensitized rabbits (16), led to the supposition that
some antibodies may have a direct destructive effect upon the bacillus.
Later results (17) showed that this was probably not the case, since normal
serum and, in fact, gamma globulin isolated from normal serum had even
more inhibitory effect on the growth of tubercle bacilli than did the anti-
body-containing fractions. It is probable, therefore, that naturally existing
antimicrobial substances, which will be discussed later, may be the chief
direct inhibitors of tubercle bacilli. Since antibody probably coats bacilli,
as postulated by Bordet (18) and Mudd and Anderson (19), it is possible
that its direct effect upon the bacillus may be in the nature ofimmobilizing
it as an antibody-bacillary complex in a manner similar to that found by
Taliaferro (20), who pointed out that an antibody to the parasite Trypano-
soma lewisi specifically inhibited its reproduction by coating its surface and
thus depressing all metabolic activities. Still another type of resistance in
the immunological process may be given by antibodies, namely, by im-
mobilizing inhibitors to the immune mechanism. This will be discussed
later.

B. antigen-antibody complexes

i. In vitro.—When rabbit BCG antiserum was mixed with the proper

concentration of a solution of tuberculoprotein or tuberculopolysaccha-
ride, a white percipitate occurred (1), its extent depending upon the con-
centration of antibodies present. In general, an optimal concentration of
tuberculoprotein antigen was about 50 gamma/0.1 ml, while that of
tuberculopolysaccharide antigen was considerably less, only 1 gamma/
0.1 ml, as mentioned under methods of detection.
The polysaccharide-antigen-antibody complex could be completely
resolved in antigen excess (4), and this may explain why no precipitate
occurred when 50 gamma polysaccharide/0.1 ml was used as precipitating
antigen. On the other hand, the protein-antigen-antibody complex was
only partially resolved when concentrations even greater than 50 gamma
antigen/0.1 ml were added. This is in contrast to some other proteins,
such as egg albumin, which causes better resolution ofits complex in anti-
gen excess, as found by Opie (21).
2. In vivo.—It was mentioned above that antibodies had been identified
in caseous tissue (4). It would then also be reasonable to expect to find
tuberculous antigens in tissues, especially where tubercle bacilli are de-

269
stroyed. Such free antigen was actually demonstrated in suspensions of
caseous materials by the use of potent antisera (4).
If both antigen and antibody are present at the same time, it would be
logical to expect the presence ofantigen-antibody complexes. Such a com-
plex might well form a reasonable portion of white tuberculous caseous
material, along with cellular debris, etc. In support of this contention are
experiments (4, 15) showing that guinea pigs or rabbits which have been
repeatedly injected with tuberculoprotein either prior to and/or following
infection developed massive caseation ofthe lungs and died rapidly. Anti-
bodies to tuberculoprotein could always be demonstrated in their blood
stream. Moreover, Opie's classical work (22) showed that precipitation
of antibody by its specific antigen occurred locally at the site of entry of
antigen, causing a local inflammatory reaction. Therefore, there is no
reason to doubt that such precipitates are formed wherever antigen meets
antibody in the tissues, with deposition of white precipitates.
3. Role in liquefaction.—Precipitates consisting of tuberculopolysaccha-
ride and its antibody were resolved in vitro in excess polysaccharide as
mentioned above. The question whether such a resolution could also occur
in vivo was answered in the affirmative (4) when it was shown that the
density of a suspension of caseous material was decreased by addition of
extra tuberculopolysaccharide under controlled conditions. Resolution of
this nature, occurring locally in caseous material in the presence of excess
antigen coming from the breakdown of tubercle bacilli, could well repre-
sent part ofthe liquefaction process. Increased enzymatic digestion of cel-
lular debris accompanied by the release ofnucleic acid and toxic products
must also be considered as a contributing factor in liquefaction. However,
a counteracting influence was demonstrated by Weiss (23), who showed
that tuberculopolysaccharide inhibited the action of the endocellular en-
zyme, Cathepsin II, of tuberculous tissue, thus decreasing autolysis.
c. antimicrobial substances

Many substances with antimicrobial activity for tubercle bacilli have

been found to exist naturally in varying amounts in various tissues in the
body. These have been listed and discussed by Dubos (24), by Hirsch (25),
and also by Long (26). Their importance for resistance of the host cannot
be doubted.
i. Location.—A potent crystalline antimycobacterial substance, identi-

270
Florence B. Seibert · Immunity in Tuberculosis
Perspectives in Biology and Medicine · Winter i960
fied as spermine phosphate, was isolated from extracts of beef kidney by
Hirsch and Dubos (27), but later (28) it was found to exert its effect on
tubercle bacini only in the presence ofa certain serum enzyme. Some basic
proteins such as histones and protamines and a basic polypeptide from
thymus (29) also inhibited growth oftubercle bacilli. Another inhibitor is
the carbohydrase lysozyme, shown to occur normally in serum, leuco-
cytes, and tissues (30-34). Recently a non-basic tuberculostatic protein has
been isolated by Myrvik and Soto-Figueroa (35) from bovine spleen.
Urine also contains an antimicrobial substance, according to Bjornesjö
(36), and there have been numerous other reports of similar but less well-
defined substances in many organ and tissue extracts.
A study of normal serum for its antituberculous activity revealed the
fact that the gamma portion of serum contained most, if not all, of this
activity, while the albumin portion had a stimulatory effect on growth of
tubercle bacilli (16, 17).
The inhibitory substance in gamma globulin was probably not proper-
din, since magnesium was not required in the reaction, and the activity
was not decreased at 56o C. Of all the antimycobacterial substances de-
scribed above, the one with an electrophoretic mobility similar to that of
gamma globulin is lysozyme. It is probable that this substance naturally
present in serum may be an important and accessible inhibitor for tubercle
bacilli. It is also possible that all the substances mentioned, and perhaps
many more still unidentified, play an important role in the body's defense,
depending upon where the tubercle bacillus lodges.
2. Mechanism of interference with growth.—While it is well recognized
that there are naturally occurring antimycobacterial substances in the body,
their modes of interfering with growth oftubercle bacilli are little under-
stood.
The effect of spermine or spermidine on growth oftubercle bacilli was
thought by Hirsch (28) to be due to some product released during the
reaction between spermine and an amine oxidase found in the alpha globu-
lin of serum. He also thought that the inhibitory effect of some basic pro-
teins and other basic substances was due to their ability to compete for
substances essential for growth of tubercle bacilli (25).
An oxidation product ofascorbic acid was believed by Myrvik and col-
leagues (37) to be the inhibitory substance found in urine by Bjornesjö
(36) and others.

271
 Myrvik, Weiser, and Kelly (38) also believed from their studies that the
action of lysozyme on tubercle bacilli was dependent upon autolytic en-
zymes ofthe cell rather than upon its basic properties. On the other hand,
it was suggested by the author (17) that the polysaccharide shown by
Meynell (39) to be on the surface of the tubercle bacillus is the point of
attack by lysozyme, since Epstein and Chain (40) had previously shown
that polysaccharide was a vulnerable site for other bacteria, and Kaiser (41)
and Kerby and Eadie (42) demonstrated that the carbohydrate heparin in-
hibited lysozyme activity. In our laboratory recent studies (43) have shown
that lysozyme attacks tuberculopolysaccharide molecules. For example, in
the case of one type of polysaccharide, the effect was to depolymerize a
large non-dialyzable polysaccharide to a smaller dialyzable one. The other
polysaccharide, on the other hand, was precipitated to an insoluble
complex.
Perhaps the best evidence that the activity exerted on tubercle bacini by
the antimicrobial substance in normal gamma globulin is through the
polysaccharide constituent ofthe bacillus lies in the fact that the inhibitory
effect could be decreased by addition offree tuberculopolysaccharide (17),
This will be discussed in the next section.

D. polysaccharide antigens

When the tubercle bacillus is broken down and its antigens are released
into the body cells or fluids, responses of the body to such foreign sub-
stances can well be expected. It was noted above that one of the main re-
sponses of the body is an attempt to produce antibodies and that protein
appears to be most successful in this respect under all circumstances. Poly-
saccharide also is capable of such stimulation when it is in the form of a
large molecule (9) or associated with protein or as polysaccharide in a large
molecular complex, as in lipopolysaccharide (44).
Polysaccharide in culture filtrates exists chiefly as small molecules. Nev-
ertheless, they are large enough to be held back to some extent by ultra-
filters. In this form it is not able to stimulate the production ofdemonstrable
antibodies, but it is capable of acting as a haptene and thus of giving pre-
cipitates with antipolysaccharide antibodies.
i. Interference activities of tuberculopolysaccharide.—If more of the poly-
saccharide is added to a tube containing the precipitated polysaccharide-
antigen-antibody complex, as mentioned before (4), the precipitate dis-

Florence B. Seibert · Immunity in Tuberculosis

272
Perspectives in Biology and Medicine · Winter i960
solves in antigen excess due to the reversibility of the antigen-antibody
reaction, similarly to many other colloidal complexes. In the same way, if
added in advance, no precipitate will appear. Thus only at the exact equiva-
lence point can maximum precipitation be obtained, and this fact may
play an important role in the body where free tuberculopolysaccharide is
involved.
Complement appeared not to be bound by tuberculopolysaccharidewhen
tested directly by the usual block titration method ofBCG antisera. How-
ever, when it was present at the time the fixation test was performed with
PPD-S or C type protein, it appeared to interfere with fixation by these
proteins, as mentioned earlier. The mechanism of this interference was
studied (12), and it was found that tuberculopolysaccharide really did bind
complement through the first stage of fixation at 370C. but released it
when the red-blood-cell amboceptor system was introduced. However, if
a temperature of 40 C. was used instead of 37o C, complement remained
bound by polysaccharide. Therefore, the effect of the presence of free
polysaccharide in the system would be ofthe nature ofan interference with
the normal function of complement.
Free tuberculopolysaccharide (17) also inhibited the natural antimicro-
bial activity existing in gamma globulin ofnormal serum. This led to the
conclusion stated above that the site on the tubercle bacillus which is vul-
nerable to attack by the antimicrobial substance may probably be poly-
saccharide, which Meynell (39) concluded is located on the surface of the
bacillus.

II. Significant Substances in Serum

Substances in serum bearing some relationship to resistance in tuber-
culosis will be briefly mentioned.
It was found during studies of serum changes which took place as the
tuberculous process progressed that significant changes occurred in all
electrophoretically identifiable components (45-47). Gamma globulin in-
creased, and at least some ofthis increase was due to accumulation ofanti-
bodies which could be identified in isolated gamma globulin fractions.
Beta globulin also increased, especially in rabbits which developed exten-
sive renal tuberculosis (45), probably due to accumulation oflipids, as was
shown by Longsworth in cases of renal pathology (48).
An increase in alpha3 globulin was found to be a definite indication ot

273
disease activity and, in fact, a bad prognostic sign (49). It was increased
wherever there was tissue destruction. Certain antibodies were also located
in this serum fraction (3), and Cole and Favour (50) even thought an anti-
body involved in the allergic reaction may be present. Moreover, it was
found that a substance of polysaccharide or nucleic acid nature accom-
panied this alpha3 globulin, since it gave a reaction with tryptophane in
perchloric acid (49). It could readily be evaluated by means ofa simple test
based on this chemical reaction. Parallel with these changes the amount of
serum albumin was found to be reduced in proportion to the increase in
globulin fractions. This decrease could be expected to be of benefit in the
process of resistance, since albumin is a stimulant for growth of tubercle
bacilli.

III. Suggestion ofa Mechanism ofImmunity in Tuberculosis

Based on the accumulated facts referred to above, the following theory
ofa mechanism of immunity in tuberculosis has evolved. The hypothesis
is presented as a stimulus to efforts at strengthening some ofthe connecting
links.
Naturally occuring antituberculous substances, among which may be
lysozyme, form the first line of defense of the body when tubercle bacilli
enter. Their great value lies in their ability to attack and destroy the bacil-
lus, and the individual who naturally manufactures them possesses one of
the greatest assets for survival.
However, presence ofgreat quantities ofthese protectors will be oflittle
avail if other interdependent elements in the body are not in proper bal-
ance. When the bacillus is disrupted, its various antigens, tuberculopro-
teins, polysaccharides, etc., are freed to circulate in the cells or serum. If,
as outlined earlier in this paper, free tuberculopolysaccharide is able to in-
terfere with the very antimicrobial agent which attacked the bacillus, as
well as with other immunological mechanisms, then the presence of too
much polysaccharide would be harmful to the body. At certain times,
therefore, the rapid disintegration ofthe tubercle bacilli by the antimicro-
bial substance is undesirable.
On the other hand, if antibodies, especially to this polysaccharide, can
be formed in sufficient quantity—as they can be in rabbits which have been
shown to resist the disease well—then free polysaccharide will be combined
in an antigen-antibody complex and thus be removed from exerting its

Florence B. Seibert · Immunity in Tuberculosis

274
Perspectives in Biology and Medicine · Winter i960
potentially harmful ability to block the antimicrobial substance which
functions in native resistance. The right kind of antibodies, the antipoly-
saccharide antibodies, do not directly destroy tubercle bacilli, but they play
a vitally important indirect part in immunity by immobilizing free poly-
saccharide. Thus the benefit of immunization with BCG or possibly some
of the lipopolysaccharide fractions (44) may be derived from the produc-
tion of these antibodies.
Tuberculoprotein antigens appear to play an unimportant or possibly
harmful role in immunity. It is true that antibodies to them are readily
formed, but their antigen-antibody complexes, in contrast to polysaccha-
ride-antibody complexes, are less or not at all resolvable in antigen excess
and would tend to accumulate in the form of massive caseation. Lurie has
shown that such masses of caseous material are prominent in susceptible
rabbits (51), and we have found them to be conspicuous in rabbits sensi-
tized with the protein (4, 15). On the other hand, resolution of the poly-
saccharide-antibody complex in caseous material of resistant rabbits may
contribute to the process ofliquefaction, which in certain respects is bene-
ficial and in others may be harmful.
With such resolution, the formed elements of the caseous mass—such
as cellular debris, etc.—are naturally also released, to be acted upon by
enzymes and phagocytes. With destruction of such cellular elements, re-
lease of cellular proteins and extra nucleic acid derivatives could be antici-
pated and would probably be reflected in an increase in polysaccharide-
containing substances of the serum. Such substances have been shown to
accompany the alpha2 globulin fraction (49). This fact can be useful in
diagnosis, as will be discussed later.
The presence ofmuch antibody could also be responsible for immobili-
zation in contrast to destruction of the tubercle bacillus, since the latter is
known to be coated by antibody. In such form it is possible that the bacilli
may be deposited in dormant condition until some time later when excess
antibody could resolve the complex and free the bacilli to circulate again,
as appears to be the case in endogenous infection.
It is the presence and interplay of tuberculin antigens, antibodies, and
naturally occurring antimicrobial agents that determine the degree of re-
sistance to tuberculosis. None of these factors can be considered by itself.
Past failure to understand that a careful balance ofthese factors is necessary
to effect favorably the underlying mechanism of immunity has probably

275
been one of the chief causes of delay in beneficial therapy. Moreover, the
fact that antibodies may or may not be present at any particular time, de-
pending upon the state of balance of the immune factors enumerated,
would seem to prevent any hope for a specific and dependable serological
test for either diagnosis or prognosis. Is there, then, any other way ofdeter-
mining the prospects of the outcome of tuberculosis in a particular
individual?

IV. The Prognosis of Tuberculosis

One could look for breakdown products ofbody cells or tubercle bacilli,
which may be helpful in diagnosis. For example, it was possible to identify
the presence of tuberculoprotein and tuberculopolysaccharide in caseous
material by the use ofhighly potent antisera (4). Such a test is impractical,
and no report has been found where such products have been detected in
serum.

Consden and Glynn (52) were able to identify a-e-diaminopimelic acid

in some lung lesions oftuberculous patients. Since this amino acid has not
been found in hydrolysates of normal tissue fluids of mammals but was
identified in Mycobaterium tuberculosis by Asselineau, Choucroun, and
Lederer (53) and by Work (54), and in a certain purified tuberculin protein
fraction by Seibert (6), its presence in sarcoid lesions along with that of a
mycolic acid similar to wax C of M. tuberculosis led Nethercott and
Strawbridge (55) to believe that these lesions were previously tuberculous.
But the fact that a-e-diaminopimelic acid was identified by Seibert and
colleagues (56) in an acid-fast organism found to be prevalent in tumors
makes the conclusion somewhat questionable. Moreover, it is unknown at
present whether this amino acid may or may not also be present in many of
the atypical acid-fast organisms so widely found in recent years.
However, breakdown products of body cells may furnish a useful test
for following the progress of tuberculous infection. Leucocytes of sensi-
tized animals were demonstrated by Fabrizio (57) in tissue cultures to be
agglutinated and disintegrated by the C type of tuberculoprotein, a pro-
tein which is released when the tubercle bacillus is broken down. As the
rate of destruction of body cells is accelerated under these circumstances,
their disintegration products will tend to accumulate. Since it has been
shown (46, 49) that an increase of such products of cellular destruction is
found in the alpha2 globulin fraction ofserum and can be measured by the

276 Florence B. Seibert · Immunity in Tuberculosis

Perspectives in Biology and Medicine · Winter i960
TA (tryptophane-perchloric acid) method, analysis ofthis fraction should
prove to be prognostically helpful—as it was indeed in the study of re-
sistance of rabbits challenged after sensitization with different tuberculo-
protein fractions or BCG (58). For example, when the values found by
the TA method began to rise, the weight of the animal invariably de-
creased, and in an average of 90 days the rabbit was dead. The lag period
for these changes was greatly increased in rabbits previously sensitized
with BCG.
Moreover, in a series oftuberculous patients with graded and carefully
diagnosed lesions, content ofalpha2 globulin and its accompanying poly-
saccharide component paralleled the extent of the disease (46). Marker
(59) also found in patients with primary tuberculosis that the alpha2
globulin content of serum was high in proportion to the severity of the
disease and decreased with improvement. Confirmation of the high clini-
cal significance ofalpha2 globulin values in prognosis is found in papers by
Meyer et al. (60) and Kreis et al. (61).
An increase in the gamma globulin does not necessarily indicate in-
creased resistance in spite of the fact that the fraction contains antibodies
(16). An increase may indicate that antibodies are being formed, but no
differentiation between antiprotein and antipolysaccharide antibodies is
made except by the serological test. In animals able to produce both kinds
ofantibodies, if one is produced, both are likely to be formed. Therefore,
it is possible that increased gamma globulin may parallel improvement in
the disease; but in my rabbits (45), and in cases ofhuman tuberculosis (46)
as well, terminal disease was also accompanied by a very large gamma
globulin, which presumably contained large quantities of anti-protein
antibodies. A decrease in this fraction in general paralleled improvement
in the tuberculous involvement in patients serially followed by Zitrin et
al. (62).
It would seem that the TA method for measuring tuberculous disease
activity should be a very helpful prognostic test, although it is capable of
measuring cellular destruction in other diseases as well. It can be very help-
ful in conjunction with determinations of total protein by the Kingsley
method (63) and gamma globulin by the Kunkel method (64)—certainly
in prognosis, but also even in diagnosis (46). Zitrin et al. (59) found serial
analyses of the alpha2 and gamma globulins to be of great value in deter-
mining the course of treatment in childhood tuberculosis.

277
 V. Questions Raised by Hypothesis
The hypothesis presented raises many questions which experiments can
answer.

Why do different people as well as different species differ in their re-

sponse to infection with the tubercle bacillus? Do they naturally possess
different amounts of antimicrobial substances? Do they differ in their
ability to produce the needed anti-polysaccharide antibodies? Or both?
Do people with diabetes, who are known to be more susceptible to this
disease, lack one or both of these characteristics? If so, is the lack due to a
hormonal imbalance? Could either one be influenced by hormonal treat-
ment?
Could cortisone treatment be good in one case and bad in another be-
cause it affects the production of one kind ofantibody more than another?
How would one increase the amount of antimicrobial substances or the
anti-polysaccharide antibodies in those individuals lacking them? Could it
be done by treatment with hormones, since Lurie's work suggested that
treatment with triiodothyronine increased resistance of his rabbits (65)?
Could it be done by transfer of sera from animals with known resistance
(66)? Or is there a "core" of susceptible individuals, as suggested by Long
(67), who never respond favorably to any treatment because they cannot
be helped to overcome their native deficiencies, which may be due to
some of the mechanisms here discussed?
One may also ask whether the antimicrobial substances here mentioned
will attack atypical or avirulent acid-fast organisms as well as virulent or-
ganisms. Since the polysaccharides of certain acid-fast bacilli were shown
(68)to cross-react with each other, will they cause similar cross-reacting
inhibitions, as discussed in this paper?
More also should be known about the mechanism oflysis ofthe tubercle
bacillus by the antimicrobial substances.
Substances have been demonstrated in the serum, especially in advanced
tuberculosis, which are thought to be products of cellular disintegration,
such as nucleic acid and certain polysaccharides. May these not be respon-
sible for the decreased allergy in terminal tuberculosis, since it was shown
(69)that the addition of nucleic acid to tuberculoprotein decreased its
reactivity? Is it possible that certain ofthese substances may even stimulate
the growth of the tubercle bacillus in a more direct manner than is done

Florence B. Seibert · Immunity in Tuberculosis

278 Perspectives in Biology and Medicine · Winter i960
by the polysaccharide? Unpublished results of Patino (70) in my labora-
tory showed that certain concentrations of nucleic acid, when added to
tubercle bacillus cultures, stimulated growth of the bacillus. What effect
does the presence oftuberculopolysaccharide have on the allergic reaction
with tuberculoprotein?
REFERENCES

i. F. B. Seibert, E. E. MaLER, U. Busemann, M. V. Seibert, E. Soto-Figueroa, and

L. Fry. Am. Rev. Tubera, 73 : 547, 1956.
2.F. B. Seibert andJ. W. Nelson. Proc. Soc. Exper. Biol. & Med., 49:77, 1942.
3.F. B. Seibert and J. W. Nelson. J. Am. Chem. Soc, 65:272, 1943.
4.F. B. Seibert. J. Infec. Dis., 99:76, 1956.
5. ----------.J. Immunol., 65:297, 1950.
6.F. B. Seibert, E. Soto-Figueroa, and E. H. DuFour. Am. Rev. Tuberc, 71 :704,
1955·
7.F. B. Seibert, C. Crumb, and E. H. DuFour. J. Infec. Dis., 89:252, 1951.
8.F. B. Seibert and E. Soto-Figueroa. Am. Rev. Tuberc, 75:601, 1957.
9.F. B. Seibert, M. Stacey, and P. W. Kent. Biochim. et biophys. acta, 3 :632, 1949.
10.M. Heidelberger and F. E. Kendall. J. Exper. Med., 61 1563, 1935.
11.F. B. Seibert and M. V. Seibert. J. Infec. Dis., 101 :i09, 1957.
12.F. B. Seibert and E. Soto-Figueroa. Proc. Soc. Exper. Biol. & Med., 92:811, 1956.
13.G. Middlebrook and R.J. Dubos. J. Exper. Med., 88:521, 1948.
14.E.J. Cohn, F. R. N. Gurd, D. M. Surgenor, B. A. Barnes, R. K. Brown, G. De-
rouaux, J. M. Gillespie, F. W. Kahnt, W. F. Lever, C. H. Liu, D. Mittelman,
R. F. Mouton, K. Schmid, and F. Uroma. J. Am. Chem. Soc, 72:465, 1950.
15.F. B. Seibert. Proc. Soc. Exper. Biol. & Med., 30:1274, 1933.
16.E. W. Emmart and F. B. Siebert. J. Immunol., 50:143, 1945.
17.F. B. Seibert. J. Infec. Dis., 103:52, 1958.
18.J. Bürdet. Traité de l'immunité. Paris: Masson, 1920, 1939.
19.S. Mudd and T. F. Anderson. J. Immunol., 42:251, 1941.
20.W. H. Taliaferro. Bact. Rev., 12:1, 1948.
21.E. L. OpiE. J. Immunol., 8:19, 1923.
22. ----------. Ibid., 9:231, 1924.
23.C. Weiss and N. Halliday. Proc. Soc. Exper. Biol. & Med., 57:299, 1944.
24.R. J. Dubos. Biochemical determinants of microbial diseases. Cambridge, Mass.:
Harvard University Press, 1954.
25.J. G. Hirsch (ed.). Ciba Foundation symposium on experimental tuberculosis. Lon-
don: J. & A. Churchill, 1955.
26.E. R. Long. The chemistry and chemotherapy of tuberculosis. 3d ed. Baltimore:
Williams & Wilkins Co., 1958.
27.J. G. Hirsch and R.J. Dubos. J. Exper. Med., 95:191, 1952.

279
28.J. G. Hirsch. J. Exper. Med., 97:345, 1953.
29.R.J. Dubos andJ. G. Hirsch. J. Exper. Med., 99:55, 1954.
30.A. Fleming. Proc Roy. Soc, London, s.B. 93:306, 1922.
31.A. Fleming and V. D. Allison. Brit.J. Exper. Path., 3:252, 1922.
32.J. Smolens andJ. Charney. J. Bact., 54:101, 1947.
33.Q. N. Myrvik and R. S. Weiser. Am. Rev. Tuberc, 64:669, 1951.
34.Q. N. Myrvik, R. S. Weiser, and H. Agar. Am. Rev. Tuberc, 67:217, 1953.
35.Q. N. Myrvik and E. Soto-Figueroa. Am. Rev. Tuberc, 78:93, 1958.
36.K. B. Bjornesjö. Acta tuberc. scandinav., 25:425, 447, 457, 1951; 27:116, 126, 1952.
37.Q. N. Myrvik, R. S. Weiser, B. Houglum, and L. R. Berger. Am. Rev. Tuberc,
69:406, 1954·
38.Q. N. Myrvik, R. S. Weiser, and M. C. Kelly. Am. Rev. Tuberc, 68:564, 1952.
39.G. G. Meynell. J. Path. & Bact., 67:137, 1954.
40.L. W. Epstein and E. Chain. Brit.J. Exper. Path., 21 1339, 1940.
41.E. Kaiser. Nature, 171:607, 1953.
42.G. P. Kerby and G. S. Eadie. Proc. Soc. Exper. Biol. & Med., 83:111, 1953.
43.F. B. Seibert. Fed. Proc, 17 (Part I) ¡308, 1958.
44.N. Choucroun. Am. Rev. Tuberc, 56:203, 1947.
45.F. B. Seibert and J. W. Nelson. Am. Rev. Tuberc, 47:66, 1943.
46.F. B. Seibert, M. V. Setbert, A. J. Atno, and H. W. Campbell. J. Clin. Invest.,
26:90, 1947.
47.K.Jahnke and W. Scholtan. Beitr. Klin. Tuberk., 105:249, 1951.
48.L. G. Longsworth, J. Shedlovsky, and D. A. MacInnes. J. Exper. Med., 70:399,
1939·
49.F. B. Seibert, M. L. Pfäff, and M. V. Seibert. Arch. Biochem., 18:279, 1948.
50.L. Cole and C. B. Favour. J. Exper. Med., 101:391, 1955.
51.M. B. Lurœ, P. Zappasodi, and C. Tickner. Am. Rev. Tuberc, 72:297, 1955.
52.R. Consdbn and L. E. Glynn. Lancet, p. 943, May 7, 1955.
53.J. AssELiNEAU, N. Choucroun, and E. C. Lederer. Biochim et biophys. acta, 5:
197, 1950.
54.E. Work. Biochem. J., 49:17, 1951.
55.S. E. Nethercott and W. G. Strawbridge. Lancet, p. 1132, Dec. ?, 1956.
56.F. B. Seibert, E. Soto-Figueroa, E. E. Miller, and M. V. Seibert. Growth, 18:145,
1954·
57.A. M. Fabrizio. Am. Rev. Tuberc, 65:250, 1952.
58.F. B. Seibert and M. V. Seibert. Am. Rev. Tuberc, 62:67, !95°·
59.C. M. Zitrin, E. M. Lincoln, R. Carretero, and E. Melly. A.M.A. J. Dis. Child.,
98: 330,1959-
60.A. Meyer, H. Kaufmann, and J. Gelin. Rev. tuberc. (Paris), 19:178, 1955.
61.B. Kreis, ?. F. Jayle, P. Clerc, and H. Lutter. Rev. tuberc. (Paris), 19:187, 1955.
62.C. M. Zitrin, E. M. Lincoln, A. Saifer, and S. Lewkowicz. Am. Rev. Tuberc,
74:15.1956.
63.G. R. Kingsley. J. Lab. & CUn. Med., 27:840, 1942.

280 Florence B. Seibert · Immunity in Tuberculosis

Perspectives in Biology and Medicine · Winter i960
64.H. G. Kunkel. Proc. Soc. Exper. Biol. & Med., 66:217, 1947.
65.M. Lurie and G. S. Ninos. Am. Rev. Tuberc, 73:434, 1956.
66.C. M. Zitrin and O. Wass-Höckert. Am. Rev. Tuberc, 76:256, 1957.
67.E. R. Long. Am. Rev. Tuberc, 78:499, 1958.
68.D. M. Tennent and D. W. Watson. J. Immunol., 45:179, 1942.
69.F. B. Seibert. Faraday Soc. Dis., 13:251, 1953.
70.T. Patino. Unpublished Master's thesis. Univ. of Pennsylvania, 1947.

281