RF-TB-1.24.pdf

Media

extracted text
Tuberculosis: Pathogenesis. Protection, and Control
[■dited by Barry R. Bloom
© 1994 American Society for Microbiology. Washington. DC 20005

Chapter 28

Mechanisms of Pathogenesis in Tuberculosis
Graham A. W. Rook and Barry R. Bloom

or in the lungs failed to eliminate the bac­
teria. Thus, if guinea pigs were preimmu­
nized by protocols that gave rise to necro­
tizing skin test reactivity equivalent to the
local necrosis elicited by Koch, they were
rendered more rather than less susceptible
In the 1890s, Robert Koch observed that a
to infection by intramuscular injection of a
primary infection of guinea pigs with My­
small number of virulent organisms. In con­
cobacterium tuberculosis in the skin piotrast, immunization protocols priming small
duced a nonhealing lesion and that reinoc­
tuberculin reactions were protective (Wil­
ulation of the animals after several weeks
produced only a firm, red nodule that ne­ son et al., 1940). These observations led to
crosed and finally healed. These observa­ endless confusion. For instance, is the
tions first suggested the existence of im­ Koch phenomenon an exaggerated version
munity to tuberculosis infection. When of the tissue-damaging process seen routinely in tuberculosis lesions? If so, what is
tuberculous &
guinea r-o
pigs- were challenged
<
intradermally6with a culture supernatant of the relationship between this tissue-damagM. tuberculosis (old tuberculin) or with live ing response and protection? I he Problcm
organisms, there was necrosis both locally is that we do not know the mechanisms
in the challenge site and at a distance in the involved at the cellular or molecular level.
preexisting tuberculous lesion (Koch, Therefore, we do not know whether the
1891). This reaction, now known as “the tissue damage, the Koch phenomenon, and
Koch phenomenon,” protected against vir­ the protection are “excessive” and “regu­
ulent organisms, perhaps among other rea­ lated” manifestations of similar pathways
sons because the local necrosis caused or whether they are the results of qualita­
sloughing of the tissue containing the or­ tively different immunological mechanisms
ganisms, since similar necrosis in deep sites (Dannenberg, 1968). The balance between

KOCH PHENOMENON, TISSUE
DAMAGE, AND PROTECTIVE
IMMUNITY IN THE PATHOGENESIS
OF TUBERCULOSIS

cell-mediated immun
Graham A. W'. Rook • Department of Medical Mi­
crobiology, University College London Medical
School. 67-73 Riding House Street, London W1P 7PP.
United Kingdom.
Barry R. Bloom • Howard
Hughes Medical Institute, Albert Einstein College of
Medicine, BOO Morns Park Avenue. Bronx, Ne

throughout the course of this disease deter­
mines what form the disease takes and may
not be dissimilar to the spectrum in leprosy
that has been correlated to the .balance of
g

lymphocyte subsets (Salgame
(he idea|

York 10461.

I
485

— .1

486

^1
-1

I

I

Rook and Bloom

case, which is the situation for the large cobactericidal activity of macrophages (De­
proportion of tuberculin-positive individu­ nis, 1991a; Chan et al., 1992).
als who have been infected but show no
The situation is still less clear when we
evidence of disease, an early and appropri­
consider human macrophages. To our
ate cell-mediated immune response devel­ knowledge, only two authors have claimed
ops and controls the infection. While the to be able to induce killing of virulent M.
tissue-damaging component is excessive in
tuberculosis by human monocytes or
only a small number of individuals infected
monocyte-derived macrophages in vitro
with M. tuberculosis, it is largely responsi­
(Denis, 1991b; Crowle, 1990). The authors
ble for the clinical manifestations of the
of this chapter have been unable to demon­
disease, and the control of this excess
strate such killing. At best, a slowing of the
would reduce much of the pulmonary de­
rate of intracellular replication by about
struction that occurs. The little that we do
one generation in four was achieved follow­
know about these mechanisms is discussed ing addition of recombinant IFN-y and calin this chapter and in chapter 27.
citriol (Rook et al., 1986b). IFN-y alone
frequently causes increased growth of M.
tuberculosis in human cells (Rook et al..
Activated Macrophages: Discrepancy
1986a, b). Moreover, it seems that human
between Mouse and Human
monocytes arc unable to generate tetra­
The topic oi activated macrophages is hydrobiopterin, which is an essential cofac­
discussed in detail in chapter 27, but in
tor for arginine-dependent NO synthesis
brief, the microbicidal mechanism is uncer­
(Stuehr et al., 1991). We are careful to note
tain, particularly in relation to humans. that these studies have been carried out
Since the work of Mackaness (1968), with blood monocytes, not tissue macro­
mostly with Listeria monocytogenes, dem­ phages. Since other human cell types, e.g.,
onstrated the enhanced nonspecific bacteri­ endothelial cells and liver cells, can pro­
cidal activity of macrophages activated by
duce this NO in large quantities, it is not
mediators released from lymphocytes clear whether human macrophages
are un­
primed and stimulated to specific antigens,
able to do so or whether the correct com­
activated macrophages have usually been bination of cytokines and culture condi­
assumed to be important effectors in myco­ tions or tissue sources has simply not yet
bacterial diseases as well. This view is been found.
compatible with histological evidence, as
outlined above. However, most of the pub­
Cytotoxic T Cells
lished work involves mouse macrophages.
Murine macrophages can be activated to
While 1-cell-derived lymphokines and
inhibit or destroy virulent M. tuberculosis
the activated macrophage represent a nec­
in vitro. This can be achieved with class II
essary condition for protection, it is cer­
major histocompatibility complex (MHC)tainly not the whole story even in the
restricted T-cell lines (Rook et al., 1985) or
mouse. In vitro evidence exists that cyto­
with lymphokines (Rook et al., 1986a;
toxic T cells recognizing mycobacterial an­
Flesch and Kaufmann, 1990). More re­
tigens do develop in both humans and mice
cently, it has become apparent that by
(Kaufmann, 1988; Ottenhoff et al., 1988activation with gamma interferon (IFN-y)
Orme et al., 1992). Transgenic animals
and either lipopolysaccharide or tumor newhose gene for p2-microglobulin is dis­
crosis factor alpha (TNF-a), murine macro­
rupted and who are unable to express class
phages can be triggered to release nitric
I MHC on the cell membrane have greatly
oxide (NO), which is required for the myincieased susceptibility to tuberculosis

Chapter 28

icrophages (De2).
clear when we
iages. To our
-s have claimed
of virulent M.
monocytes or
iages in vitro
)). The authors
able to demoni slowing of the
ition by about
jhieved followIFN-y and cal. IFN-y alone
growth of M.
(Rook et al.,
ns that human
tenerate tetrassential cofacNO synthesis
careful to note
n carried out
tissue macrocll types, e.g.,
ells, can proities, it is not
hages are un­
correct com­
ulture condiimply not yet

hokines and
resent a nec)n, it is cereven in the
.ts that cyto>bacterial anans and mice
et al., 1988;
mic animals
>ulin is dis'xpress class
have greatly
tuberculosis

(Flynn et al., 1992). This implies a role for
CD8' T cells, possibly cytotoxic cells. In­
terestingly, p2-microglobulin “knockout"
mice were unaffected by infection with
BCG or avirulent M. tuberculosis H37Ra.
One way that an antigen known initially to
be taken up into an cndosomal compart­
ment could be presented to MHC class
I-restricted cytotoxic lymphocytes (CTL)
would be by its ability to escape from
phagolysosomes into the cytoplasm, as has
been found for Listeria and Shigella spp. In
mycobacterial infeciton, the issue remains
controversial. There is electron micro­
scopic evidence that virulent M. tuberculo­
sis can escape from the phagosome (Myrvik
et al., 1984; McDonough et al., 1993). In the
latter work, it is noteworthy that in the
7-day macrophage cultures studied, only
the virulent H37Rv and not the avirulent M.
tuberculosis H37Ra or BCG strain was ob­
served in the cytoplasmic compartment. On
the other hand, when rapid freezing tech­
niques are used to preserve membranes,
there is evidence that M. tuberculosis re­
tains host lysosomal membrane antigens
and may still be within a vesicle (Xu et al.,
in press). It is possible that processing of
antigen via the class I pathway is impor­
tant, allowing the host to kill parasitized
cells that are failing to exert bactericidal
effects. This failure could result from an
unusual intracellular location of the organ­
isms or, according to a recent report, from
a failure to present the antigens of such
organisms (Pancholi et al., 1993) (further
discussed below).

y8 T Cells
A large proportion of human peripheral
blood yS T cells, even from PPD-negative
donors, will proliferate in response to my­
cobacteria (Kabelitz et al., 1990; Uyemura
et al., 1991). In vitro, these cells secrete a
pattern of cytokines similar to those of Thl
cells (see below) and are cytotoxic. It is
reasonable to speculate that they may be

• Mechanisms of Pathogenesis

487

involved early in immunity to tuberculosis,
but there is only circumstantial evidence
for this (Barnes et al., 1992). Knockout
mice unable to generate these T cells may
reveal their relevance to the control of
mycobacterial infection.

Lymphokines and Cytokines
Recent experiments involving M. tuber­
culosis infection of transgenic mice are pro­
viding new and interesting information on
the role of some lymphokines and cyto­
kines. As would be expected, knockout
mice with a disruption in the gene for IFN--y
succumb rapidly (within 3 weeks) to M.
tuberculosis infection. In contrast, mice
will eventually die, although after many
weeks, from BCG (Dalton et al., 1993;
Cooper et al., 1993; Flynn et al., 1993).
While knockout mice forTNF-a are not yet
available, it has been possible to treat mice
with neutralizing anti-TNF antibodies. This
has led to dissemination of BCG infection
(Kindler et al., 1989) and rapid death of M.
tuberculosis (Flynn et al., submitted).
Thus, TNF is a necessary condition for
protection, again suggesting the importance
of NO and reactive nitrogen intermediates
in protection. Together, these data suggest
to us that multiple immune mechanisms
are required for protection. On the basis
of available data, we believe that both
MHC class Il-restricted T-cell-derived cy­
tokines, at least IFN-y, macrophage-de­
rived TNF-ot, and MHC class I-restricted
T-cell responses, probably CTL, are neces­
sary conditions for protection; none is suf­
ficient, and the roles of y8 T cells and other
cytokines remain to be elucidated.

TISSUE-DAMAGING MECHANISMS IN
TUBERCULOSIS

Direct Toxicity and Enhanced Susceptibility
of Individual Cells to TNF-ot
M. tuberculosis has some toxicity for
cells in vitro. This is particularly noticeable

488

IL

II

j If

■
I

H'l

Rook and Bloom

with monocytes, which often die if they 4 and 7 days, only the virulent strain of M
take up more than five bacilli. Conversely, tuberculosis, H37Rv, escaped from thes
monocytes or macrophages can take up tightly apposed membrane vesicles and en
very large numbers of M. avium without tered the cytoplasm. These results, will
obvious toxic effects. It had previously electron-dense markers used for lyso
been suggested that M. tuberculosis has the somes, thus confirmed and extended th«
ability to inhibit lysosomal fusion, which earlier observations by Myrvik et al. (1984
could contribute to the survival of the indicating that M. tuberculosis could es
pathogen in macrophages (Armstrong and cape from phagolysosomes into the cyto
Hart, 1975). In those studies, fusion could plasm. However, using immunoelectror
be blocked by agents that prevented lyso­ micrographic techniques designed to pre
somal acidification, and M. tuberculosis serve membranes, Xu et al. (in press) fount
produced ammonia in abundance (Gordon that all bacilli appear to be surrounded witl
et al., 1980). In this context, recent studies host cell membrane and membrane antigen
by Sturgill-Koszycki et al. (in press) indi­
and may not be free within the cytoplasm.
cate that vacuoles formed around M. avium Escape from phagolysosomes would repre­
fail to acidify below pH 6.3 to 6.5. Immusent one of the few biological differences
noclectron microscopy indicated that the between avirulcnt and virulent strains of M.
vacuoles containing the mycobacteria con­ tuberculosis that could be directly related
tain lysosomal proteins but not the protonto virulence, and so the issue is an impor­
ATPase responsible for acidification. Be­
tant one to resolve. Not only is it relevant
cause other lysosomal membrane markers for antigen presentation to MHC class I-rewcie present, this result indicates remark­ stricted CTL, but M. tuberculosis that en­
ably selective fusion of vesicular membrane
ters the cytoplasm of macrophages could
proteins. Because of the difficulty in secur­
exert direct toxic effects on the cells or may
ing M. tuberculosis cultures that are 100%
increase the susceptibility of infected cells
viable, it is always difficult to ascertain with to TNF-ot that is discussed below.
certainty whether fused vesicles contain
M. tuberculosis is readily taken up by a
primarily nonviable organisms and whether wide variety of nonmacrophage cell types
the bacilli found in nonfused vesicles are in vitro (Shepard, 1958; Filley and Rook.
responsible lor most of the damage. Recent 1991; Filley et al., 1992), and such cells are
studies by McDonough et al. (1993) indicate
much less susceptible to the toxicity of the
that lysosomal fusion occurred very rapidly
oiganism. I his has been shown for several
in murine or human macrophages infected cell lines and also for human endothelial
in vitro with live or dead M. tuberculosis or
cells and fibroblasts. This is paradoxical,
BCG. However, the intracellular fates dif­ because unlike M. leprae, M. tuberculosis
fered. BCG essentially remained in fused
is not seen inside such cells in vivo. One
phagolysomes for the entire 7-day observa­
possibility is that in vivo these cells are
tion period. Both H37Rv and H37Ra M.
killed quickly, so that parenchymal cells
tuberculosis strains rapidly appeared to bud infected with bacilli are rarely seen in his­
or extrude from the fused phagolysosome tological sections of tissues. An alternative
to form a unique vesicle, with the organ­
answer may lie in the observation that cells
isms enclosed by a very tightly apposed containing M. tuberculosis are rendered
membrane. Over time, fusion of secondary exquisitely sensitive to killing by TNF-a
lysosomes failed to occur in these tightly
(Filley and Rook, 1991; Filley et al., 1992).
membrane-apposed containing vesicles, Therefore, macrophages infected in vitro
and much of the multiplication of M. tuber­ may be killed by their own production of
culosis occurred in these vesicles. Between
I NF-a, while nonmacrophage cells survive

Chapter 28

rulent strain of M.
•caped from these
ic vesicles and ennese results, with
used for lysoand extended the
lyrvik et al. (1984)
culosis could esics into the cytoimmunoelectron
designed to pred. (in press) found
e surrounded with
nembrane antigen
in the cytoplasm,
mes would repreogical differences
ilent strains of M.
e directly related
ssue is an impormly is it relevant
» MHC class I-re?rculosis that en■.crophages could
i the cells or may
of infected cells
1 below.
ly taken up by a
phage cell types
dlley and Rook,
nd such cells are
ic toxicity of the
town for several
man endothelial
is paradoxical,
M. tuberculosis
Ils in vivo. One
these cells are
renchymal cells
-ely seen in his. An alternative
vation that cells
y arc rendered
ling by TNF-a
ey et al., 1992).
ifccted in vitro
i production of
ge cells survive

in vitro in the absence of TNF-a but are
rapidly killed in vivo, since TNF-a is prob­
ably abundant in lesions, as discussed be­
low. The ability to increase sensitivity to
TNF-a was prominent in virulent strains of
M. tuberculosis but weak in H37Ra and
virtually absent from M. avium and BCG
strains (Filley and Rook. 1991). Finally, if
CTL are engaged, they would have the
capability of lysing parenchymal cells ex­
pressing mycobacterial antigens in associa­
tion with MHC class 1.
The Koch Phenomenon

As outlined above, Koch noted that 4 to
6 weeks after establishment of infection in
guinea pigs, intradermal challenge with
whole organisms or culture filtrate resulted
in necrosis locally and in the original tuber­
culous lesion (Koch, 1891). Similar phe­
nomena occur in humans. The tuberculin
test is frequently necrotic in subjects who
arc or have been tuberculous. This is not an
inevitable consequence of the delayed hy­
persensitivity response to tuberculin, be­
cause necrosis does not occur when the
same test is performed in healthy BCG
recipients or in tuberculoid leprosy pa­
tients. Moreover, Koch sought to exploit
this phenomenon for the treatment of tuber­
culosis and found that injection of larger
quantities of culture filtrate (old tuberculin)
subcutaneously into tuberculosis patients
would evoke necrosis in established tuber­
culous lesions at distant sites (Anderson,
1891). This resulted in necrosis and slough­
ing of the lesions of skin tuberculosis (lupus
vulgaris, usually caused by bovine strains),
but when similar necrosis was evoked in
deep lesions in the spine or lungs, the
results were disastrous and merely pro­
vided further necrotic tissue in which the
bacteria could proliferate. This treatment
was therefore abandoned.
This phenomenon shows parallels with
the Shwartzman reaction. Shwartzman ob­
served that a site primed by an injection of

•

Mechanisms of Pathogenesis

489

gram-negative bacteria (though endotoxin
will substitute) undergoes necrosis if a sec­
ond dose of gram-negative organisms (or
endotoxin) is injected intravenously 24 h
later (Shwartzman. 1937). Several early
workers demonstrated that mycobacterial
lesions will undergo necrosis if the animal is
subsequently challenged intravenously or
subcutaneously with endotoxin-rich bacte­
ria (Bordet, 1931), endotoxin (Shands and
Senterfitt, 1972), or muramyl dipeptide
(Nagao and Tanaka, 1985), and this necro­
sis is accompanied by massive systemic
release of TNF-a (Carswell et al., 1975). It
is thought that the “prepared” inflamma­
tory site is abnormally susceptible to circu­
lating cytokines and activated cells result­
ing from the second challenge injection.
Direct injection of cytokines, particularly
TNF-a, into such sites will cause similar
necrosis (Rothstein and Schreiber, 1988).
Recent studies suggest, however, that the
susceptibility of mycobacterial lesions to
such necrosis differs in one fundamental
way from that studied by Shwartzman. The
injection of mycobacterial components (if
genuinely endotoxin free) will not prepare a
site for TNF-a-mediated necrosis unless
CD4' T-cell reactivity has previously been
primed (Al Attiyah et al., 1992). Moreover,
recent studies have revealed that such sites
undergo necrosis only if the CD4' T-cellmediated response involved is mixed ThlTh2 or ThO. while in contrast, myco­
bacterial immunization schedules leading
exclusively to Thl cytokine release yield
T-cell-mediated inflammatory sites that arc
not sensitive to TNF-a-mediated damage
(Hernandez-Pando and Rook, unpublished
data). Perhaps, therefore, the role of
TNF-a depends on what the T cells are
doing. Is the Koch phenomenon a “T-celldependent” Shwartzman reaction in a sus­
ceptible, mixed Thl-Th2 inflammatory site?
A second set of mediators that must be
considered is reactive oxygen intermedi­
ates and reactive nitrogen intermediates.
Much evidence has shown that macro­

I
490

I!
ij

II
J!
11

1

I

Rook and Bloom

phages can be activated by IFN-y to release■ (Rook et al., 1987; Valone et al., i988.
reactive oxygen intermediates, O2, H2O2, Moreno et al., 1989; Silva and Faccioli’
and hydroxyl radicals. Chan et al. (1992)
1988). Blood monocytes (Takashima et a|
showed that these compounds were not
1990) and alveolar macrophages (Rook and
microbicidal for M. tuberculosis. They fur­
Al Attiyah, 1991) from tuberculosis patients
ther showed that IFN-y and TNF-a ac­
release TNF-a “spontaneously” in large
tivated mouse macrophages to release re­
quantities, and the cytokine is present in
active nitrogen intermediates that were
the lung (Barnes et al., 1990). In view of the
mycobactericidal. One mechanism by which weight loss seen in humans, it is interesting
NO is cytotoxic for mammalian cells is by
that cytokine-induced wasting can be
binding to the Fe-S centers present in some
evoked by injecting very small quantities of
critical enzymes, including ribonucleotide
trehalose dimycolate (cord factor) dis­
reductase (Nathan and Hibbs, 1991). The
solved in oil into the peritoneal cavities of
triggering of oxygen radicals and nitric ox­
mice (Silva and Faccioli, 1988). Circulating
ide release by infected macrophages could
levels of TNF-a inhibitors are also high in
in fact damage those and adjacent cells and
the serum of tuberculosis patients (Foley et
contribute to tissue pathology.
al., 1990). These are extracellular domains
Even if these speculations are correct,
of receptors shed in response to TNF-a
they do not provide a complete account of
release. Phus, TNF-a is certainly released
the mechanism of tissue damage in tuber­
in the human disease. Its role is uncertain,
culosis. More important, they leave us un­
but recent studies show that administration
certain as to the link between the acute
of thalidomide, which reduces mRNA lev­
necrotic phenomenon evoked by Koch and
els for TNF-a, to tuberculosis patients
the slowly evolving necrosis leading to ca­
causes rapid symptomatic improvement
seation, liquefaction, calcification, and cav­
and weight gain (Kaplan, 1993), as it did for
ity formation described above.
reducing erythema nodosum in leprosy
We cannot rule out several other possible
(Sampaio et al., 1991). In the mouse, on the
mechanisms. CTL were considered above
other hand. TNF-a is clearly necessary for
as possible components of the protective
protection, though as discussed in the pre­
pathway, but increased activity of such
vious section, this is probably not the
cells could also contribute to the increased
whole story, and protection may depend on
killing of infected macrophages. Indeed, we
the type of T-cell reactivity at the site into
know too little to rule out any effector cell
which TNF is released. Studies by Kindler
type.
et al. (1989) showed that treatment of mice
with polyclonal anti-TNF-a antibodies in­
hibited granuloma formation, leading to dis­
Excessive Cytokine Release
seminated infection with BCG. However,
In view of the ability of M. tuberculosis iin M. tuberculosis-infectcil mice treated
to increase the sensitivity to TNF-a of with neutralizing monoclonal antibodies to
individual cells, the ability of the CD4+ TNF-a, Flynn et al. (submitted) have evi­
1 -cell-mediated response to render a whole
dence that granuloma-like structures
tissue sensitive to the same cytokine, and
formed, but the mice were unable to control
the Shwartzman-like nature of Koch's
the infection. Interestingly, the granulomas
cure for tuberculosis, we must consider in the anti-PNF-a-trcated mice were necro­
the possibility that 1 NF-a in synergy with tizing, while the control mice, which were
other cytokines is a component of tissue able to restrict the growth of M. tuberculo­
destruction in humans. The bacteria pro­
sis, had nonnecrotizing granulomas. Thus,
duce potent triggers of cytokine release
although TNF-a probably plays a role.

I

-

Chapter 28

ne et al., 1988;
a and Faccioli,
akashima et al.,
’ages (Rook and
rculosis patients
□usly’ in large
ie is present in
). In view of the
it is interesting
isting can be
all quantities of
d factor) disneal cavities of
S8). Circulating
ire also high in
tients (Foley et
•llular domains
use to TNF-a
tainly released
e is uncertain,
administration
is mRNA levlosis patients
improvement
9, as it did for
n in leprosy
nouse, on the
necessary for
ed in the preibly not the
ay depend on
the site into
:s by Kindler
Tie nt of mice
ntibodies in­
vading to disi. However,
nice treated
intibodies to
d) have evistructures
le to control
granulomas
were necrowhich were
. tuberculomas. Thus,
iys a role.

these data indicate that high levels of
TNF-a are not necessary for induction of
tissue necrosis in mice. It will be crucial to
resolve the discrepancies between the re­
sults using the polyclonal and monoclonal
anti-TNF antibodies through transgenic
knockout mice unable either to express the
appropriate TNF receptors or to produce
TNF-a itself.

Tuberculin Shock
As will be discussed below, tuberculosis
patients have defective adrenal function,
and some have reduced ability to increase
cortisol levels in response to adrenocorti­
cotropin, so they have little “adrenal re­
serve.” This adrenal deficit may explain a
tendency to go into “tuberculin shock”
(systemic Koch phenomenon?) if chemo­
therapy causes rapid release of bacterial
antigens and cytokine triggers (Scott el al.,
1990). This complication is most often seen
in patients with severe disseminated dis­
ease and in patients with protein malnutri­
tion and liver damage. It may be relevant
that in experimental models, liver damage
greatly enhances susceptibility to the toxic­
ity of TNF-a released in response to bacte­
rial components (Freudenberg and Gal­
anos, 1991).

REGULATION OF PROTECTIVE AND
TISSUE-DAMAGING RESPONSES IN
TUBERCULOSIS
In the absence of certainty about the
effector mechanisms involved, any discus­
sion of the regulation of the protective and
tissue-damaging responses must be some­
what speculative.
Thl and Th2 Cells: Selection of Functional
T-Cell Subsets and Disease Progression

Functional T-cell subsets in mice (Mosmann and Moore, 1991) and humans (Ro­
magnani, 1991; Salgamc et al., 1991) have
been defined by the patterns of lympho-

•

Mechanisms of Pathogenesis

491

kines they produce (see chapter 25 for more
detailed discussion). Type 1 CD4' and
CD8+ CTL produce IFN--y, interleukin-2
(IL-2), and lymphotoxin. Type 2 CD4'
(Th2) cells and CDS' T cells produce pre­
dominantly IL-4. Each subset exerts nega­
tive regulation upon the other (Romagnani,
1991; Maggi et al., 1992), and together with
the major regulatory cytokines IL-12 and
IL-10 produced primarily by macrophages,
they largely determine the type of T-cell
response that ensues. Thus, the cytokine
profile observed in mycobacterium-respon­
sive human cells from healthy donors may
be influenced by the conditions used (Haanen et al., 1991; Barnes et al., 1993) and can
be altered by the addition of cytokines or
cytokine-neutralizing antibodies (Maggi et
al., 1992). In many experimental intracellu­
lar infections, e.g., with Leishmania spp.,
the type 1 response is protective and the
type 2 response leads to disease progres­
sion. In helminth infections, the reverse
may be true. In human leprosy, the type 1
response is associated with resistance and
the type 2 response is associated with the
lepromatous or unresponsive form (Bloom
et al., 1992). One wonders whether the
anergy seen in a quarter of tuberculosis
patients may be related to Th2 function,
and evidence for effects of type 2 cytokines
in tuberculosis has been discussed by Rook
(1991; Rook et al., 1993b).
In many chronic infections or inflamma­
tory diseases, there is a permanent or tran­
sient switch from a Thl pattern of response
to Th2. Such a switch appears to occur, for
example, in schistosomiasis (Grzych et al.,
1991) and syphilis (Fitzgerald, 1992) pa­
tients. Does the evidence for a Th2 compo­
nent in the response of tuberculosis pa­
tients mentioned above suggest a similar
trend in this disease? It is certainly inter­
esting that the necrosis that occurs around
the ova in murine schistosomiasis occurs at
precisely the time when a Th2 response
becomes superimposed on a preexisting
Thl pattern (Grzych et al., 1991). It is not

492

I

(

1

!!

i Hl;I
h

Rook and Bloom

impossible that a similar mixed pattern of Tayoub, 1986; Rook et al., unpublished
response lies behind the necrosis seen in data). This may have serious conse­
tuberculosis (Barnes et al., 1993), and as
quences, because the desulfated form,
pointed out above, it is these “mixed” DHEA, is a genuine antiglucocorticoid, the
responses that are TNF-a sensitive in one specific receptors for which are found in T
murine model.
cells (Meikle et al., 1992). DHEA enhances
Current dogma states that the cytokines Thl activity (Suzuki et al., 1991; Daynes et
released by Thl cells enhance Thl activity al., 1990) and inhibits the effects of gluco­
and inhibit Th2. Why, then, docs the Thl -> corticoids, including their tendency to sup­
I h2 shift occur? We should remember that
press Thl lymphocytes and enhance Th2
the Thl/Th2 ratio is not determined only by (Blauer et al., 1991; Daynes et al., 1990;
cytokines. The release of prostaglandins Fischer and Konig, 1991). For instance, a
horn activated macrophages downrcgulates single dose of DHEA given before dexa­
Thl cells (Phipps et al., 1991), and there are
methasone can block the ability of the latter
two striking endocrine changes in tubercu­ to cause depletion of thymocytes and tem­
losis patients that would indeed be ex­
porary unresponsiveness of peripheral T
pected to have this effect.
cells to mitogens (Blauer et al., 1991).
(i) Formation of calcitriol in mycobacte­ Therefore, the T cells of tuberculosis pa­
rial lesions. The macrophages of tuberculo­
tients may be chronically exposed to glu­
sis patients, following activation by 1FN-7,
cocorticoid effects unopposed by the an­
express an active la-hydroxylasc and rap­ tiglucocorticoid influence of DHEA. This
idly convert 25(OH)-vitamin D3 to calcitriol
may not only encourage a Thl-to-Th2
(Rook et al., 1986b; Rook, 1988). Their T
switch but even contribute to the fall in
cells may also produce this enzyme CD4' T-cell count and in CD4/CD8 ratio
(Cadranel et al., 1990). This is a potent
that is well documented in tuberculosis
phenomenon, leading occasionally to leak­
patients (Singhal et al., 1989; Ainslie et al.,
age of calcitriol into the periphery and to 1992) and of course in human immunodefi­
hypercalcemia, though it has in the past ciency virus-infected patients, in whom
been difficult to understand its role in the DHEA is also low and correlates directly
disease (Rook, 1988). It now seems possi­
with CD4 counts (Wisniewski et al., 1993).
ble that this is a feedback mechanism that
This has been developed further as a hy­
tends to downregulate Thl and enhance
pothesis elsewhere (Rook et al., 1993b).
Th2 responses, because the active vitamin
At present, no information exists on the
D3 metabolite, 1,25(OH)2 cholccalciferol
patterns of lymphokines and T-cell subsets
(calcitriol), inhibits production of IFN--y in healing and progressing lesions in tuber­
and IL-2 and increases production of IL-4
culosis. It is clear that there is anergy
and IL-5 (Daynes et al., 1991).
associated with a significant proportion of
(ii) Adrenal dysfunction in tuberculosis.
tubeiculosis patients and that the prognosis
The pituitary-adrenal axis is disturbed in for these patients prior to chemotherapy
patients with tuberculosis (Sarnia et al.,
1990; Ellis and Tayoub. 1986; Barnes et al.,’ was generally poorer than for those with
tuberculin hypersensitivity. That picture is
1989a). There is a striking reduction in
suggestive of the type 2 (Th2) predomi­
adrenal function, reflected by low levels of
nance in many patients with lepromatous
essentially all steroid metabolites in 24-h leprosy. We suggest the hypothesis that in
urine collections (Rook et al., unpublished the proportion of individuals infected with
data). Most significantly, patients have M. tuberculosis who progress to primary
very low or absent levels of dehydroepi­
progressive or recrudescent forms of dis­
androsterone sulfate (DHEA-S) (Ellis and <
ease, the relative ratios of type 1 (Thl) to

JI

Chapter 28

, unpublished
erious consculfated form,
□corticoid, the
are found in T
IEA enhances
’91; Daynes et
ects of gluco■dency to supenhance Th2
et al., 1990;
□r instance, a
before dexa:y of the latter
ytes and ternperipheral T
t al., 1991).
•erculosis pa□osed to glu1 by the anDHEA. This
Thl-to-Th2
o the fall in
)4/CD8 ratio
tuberculosis
kinslie et al.,
immunodefis, in whom
ates directly
et al., 1993).
ler as a hy1993b).
xists on the
-cell subsets
ms in tuber' is anergy
oportion of
ie prognosis
emothcrapy
those with
it picture is
I) predomiepromatous
iesis that in
fected with
to primary
rms of dis1 (Thl) to

type 2 (Th2) are altered. We propose fur­
ther, on the basis of the requirement for
MHC class I-restricted T-cell function for
resistance to tuberculosis in mice and the
finding that protection can be engendered
by the 65-kDa antigen presented in trans­
fected mammalian cells (Silva and Lowric,
in press), that the number of CDS' CTL
may similarly be critical.

Roles of Different Antigenic Components of
M. tuberculosis
Is there any evidence that protective cellmediated immunity and the Koch phenom­
enon or tissue-damaging pathways repre­
sent responses to different components of
the organism? It has long been the dream of
many workers to create a novel vaccine,
perhaps a modified BCG, that primes bac­
tericidal cell-mediated immunity but not
necrosis. The practical advantages would
be even greater if nonnecrotizing skin test
responsiveness could also be dissociated
from induction of protective immunity.
Such a vaccine, if achievable, would have
many potential benefits (adapted from Dannenberg (1990]). (i) Vaccinated individuals
would not be appreciably tuberculin posi­
tive, so that tuberculin testing of such per­
sons would still be a useful procedure for
diagnosing infection with virulent tubercle
bacilli, (ii) A vaccine that evoked little or no
delayed hypersensitivity could be given
more than once to create high levels of
immunity (especially in high-risk groups),
(iii) If available, such a vaccine might re­
place isoniazid in preventive therapy for
persons who recently became tuberculin
positive, obviating hepatotoxicity and pos­
sibly bacillary resistance to the drug, (iv)
Such a vaccine could be given to patients
with active disease in order to boost the
bactericidal pathways (immunotherapy).
Tuberculin and BCG cannot be used for
immunotherapy, because they evoke the
Koch phenomenon, as Robert Koch dis­
covered to his cost (Anderson, 1891).

•

Mechanisms of Pathogenesis

493

How realistic a proposal is this? First, as
discussed in detail in chapter 28, there is
unambiguous evidence that tuberculin skin
test reactivity can be dissociated from pro­
tection in humans. For example, while
greater than 85% skin test conversion was
observed in most of the controlled trials of
BCG against tuberculosis, protection var­
ied from 0 to 77%. Similarly, in the British
Medical Research Council trials, one lot of
vole bacillus vaccine produced very few
skin test conversions yet protected well
against tuberculosis. This and other studies
with different strains of BCG (Fine, 1989)
suggest that tuberculin positivity may not
correlate with protective immunity in hu­
mans. Al present, there is no evidence that
different antigens are involved. If, as sug­
gested by Kaufmann (1988) and Flynn et al.
(1992, 1993), both type 1 CD4 T-cell func­
tion and lymphokines and CD84 MHC
class I-restricted T cells are required for
protection, it might be expected that differ­
ent antigens and certainly different T-cell
epitopes would be involved. On the other
hand, there is good evidence that the necro­
sis-inducing components can be separated
from at least some potentially protective
antigens, as explained below.
The tissue-damaging responses are
evoked by tuberculin, which is a very crude
culture supernatant from old autolysing
bacterial cultures precipitated by trichloro­
acetic acid or ammonium sulfate. Such su­
pernatants contain fragments of essentially
all the antigens of M. tuberculosis. There­
fore, it is not meaningful to speak of remov­
ing the “tuberculin” antigens from BCG.
However, the Koch phenomenon (mani­
fested as a necrotic skin test reaction) ap­
pears to be targeted preferentially toward
the species-specific epitopes of M. tubercu­
losis. When tuberculosis patients arc skin
tested with antigens derived from other
mycobacteria, they do not exhibit necrotic
responses. In fact, their responses to the
common, cross-reactive mycobacterial an­
tigens or epitopes (which must include the

494

r
i

I

III

J

Rook and Bloom

heat shock proteins [HSPs]) are diminished
(Kardjito et al., 1986). Moreover, there is
strong evidence that cell-mediated immu­
nity to mycobacteria can be initiated, per­
haps mediated, by responses to the com­
mon, cross-reactive epitopes. The more
obvious reasons for this assertion are as
follows, (i) BCG protects against leprosy as
well as or better than it does against tuber­
culosis (Fine ct al., 1986). (ii) Small positive
tuberculin reactions in non-BCG recipients
correlate with protection (Palmer and
Long, 1966; Fine, 1994). These reactions
are caused by contact with environmental
mycobacteria, (iii) A common cross-reac­
tive protein (the 65-kDa HSP) from M.
leprae that is expressed in murine macro­
phages can engender protection against tu­
berculosis in a murine model (Silva and
Lowric, in press), (iv) Tuberculosis pa­
tients may lose their skin test responses to
common cross-reactive antigens (Kardjito
et al., 1986).
At present, the simplest way to achieve a
preparation with suitable adjuvant proper­
ties that contains a broad spectrum of com­
mon but not species-specific antigens of M.
tuberculosis appears to be use of related
mycobacterial species. Preliminary results
of immunotherapy with an autoclaved
member of the fast-growing subgenus M.
vaccae are encouraging (Onyebujoh and
Rook, 1991; Rook et al., 1994). For immu­
notherapy as an adjunct to chemotherapy,
such preparations are at present the only
possibility. It is dangerous to administer
species-specific components of M. tubercu­
losis to patients, since they may trigger the
Koch phenomenon and tuberculin shock
(Anderson. 1891).
Tuberculin Skin Test: Significance in
Apparently Healthy People

One must consider at the outset, from the
presently available evidence, that the tu­
berculin test, like the lepromin skin test,
represents the consequences primarily of

type 1 (Thl) CD4 T cells in response to
mycobacterial antigens. If CDS CTL are
necessary for protection, their presence is
unlikely to be readily detected by skin
testing with tuberculoproteins that do not
have access to the cytoplasmic compart­
ment of antigen-presenting cells and arc not
presented in the context of MHC class I
antigens.
In countries with a high standard of liv­
ing, a positive test can be a diagnostic clue,
but in developing countries, much tubercu­
lin skin test positivity is due to frequent
contact with ubiquitous environmental spe­
cies. In such populations, a positive test
has little validity as a diagnostic tool. On
the other hand, a negative test (unless the
patient has evidence of advanced disease or
AIDS and is anergic) renders tuberculosis
unlikely. Nevertheless, epidcmiologically,
even in developing countries, the test has
some predictive power in healthy people.
As already mentioned, small nonnecrotiz­
ing responses in people who have not re­
ceived BCG correlate with a significantly
decreased risk of developing tuberculosis,
while large reactions (Koch phenomena?)
correlate with an increased risk of disease,
perhaps because many such individuals are
in fact already infected (Palmer and Long,
1966; Fine, 1994).
A large or necrotic tuberculin reaction,
remaining years after the primary disease
has healed, probably signifies that a few
dormant bacilli are still present in inappar­
ent caseous foci. Such bacilli may be re­
leased from time to time and then rapidly
destroyed, which gives a booster effect to
the whole immune system, including to the
level of tuberculin sensitivity.
Individuals who have been infected with
the tubercle bacillus can in time become
tuberculin negative with or without antimi­
crobial treatment. In some of these individ­
uals, the tubercle bacillus may have been
eradicated. In many, a recall of tuberculin
sensitivity is produced by the antigens in
tuberculin (purified protein derivative) in-

Chapter 28

n response to
3D8 CTL are
ir presence is
cted by skin
is that do not
mic compartlls and are not
MHC class I

jected for skin testing. When retested with
intermediate-strength purified protein de­
rivative, a person who was negative 3
weeks earlier may now be tuberculin posi­
tive as a result of the booster effect of
tuberculin itself (Thompson et al., 1979).

andard of livagnostic clue,
meh tubercue to frequent
mmcntal spepositive test
•Stic tool. On
st (unless the
ed disease or
tuberculosis
miologicaily,
the test has
dthy people,
nonnecrotizhave not resignificantly
tuberculosis,
ihenomena?)
k of disease,
dividuals are
r and Long,
lin reaction,
lary disease
that a few
t in inappar­
may be re­
hen rapidly
ter effect to
uding to the

ifected with
me become
lout antimiese individhave been
f tuberculin
antigens in
ivative) in-

i

I

•

Mechanisms of Pathogenesis

495

pressive effects on type 1 T-cell function
and macrophage activation.
Persisting Viable Tubercle Bacilli

In human beings, after even an inappar­
ent tuberculous infection heals, the lungs
may contain one or more small encapsu­
lated caseous foci. In such foci, tubercle
Significance in Patients with Active
bacilli may persist in a dormant and nonme­
Tuberculosis
tabolizing state, insusceptible to steriliza­
The size of the dermal tuberculin reac­ tion by antimicrobial agents. The bacilli
tion in patients is of limited prognostic may remain viable in the host for life and
significance (Lurie, 1964), though individu­ cause active disease when resistance is
als who are very ill with tuberculosis may lowered by old age, corticosteroids, immu­
show a negative tuberculin skin test. How­ nosuppressants, AIDS, or other factors. It
ever, when they are recovering, they again is the presence of these bacilli that necessi­
show a positive skin test. Several mecha­ tates prolonged (6-month) courses of che­
nisms could be involved, including com­ motherapy, with the resulting problems of
partmentalization of T cells, Th? cell dom­ cost, compliance, and drug resistance. It is
inance, or suppressive cytokines from not certain that drugs able to kill dormant or
infected macrophages. Regarding compart­ stationary-phase mycobacteria can be de­
mentalization, tuberculous lesions may col­ vised, since most microbicidal agents de­
lect most of the relevant circulating T cells, pend on actively metabolizing or dividing
so that few are available to participate in cells. Persistence of viable tubercle bacilli
the tuberculin reaction. This concept re­ may also be the reason the positive tuber­
ceives support from the fact that lympho­ culin reaction is usually maintained for life.
cytes in bronchoalveolar lavage and pleural Each time the bacillus multiplies, the im­
fluids (and presumably in other diseased mune system may be stimulated.
tissues) contain a greater proportion of an­
In addition, tubercle bacilli may possibly
tigen-specific T cells, secrete greater quan­ persist within macrophages as forms with
tities
and show a greater, unusual cell walls (Stanford, 1987), and
. .
___ of lymphokines,
tendency to proliferate (in the presence of thcre are rcports of mycobacterial genomic
specific antigen) than T lymphocytes in the material in the tissues of some patients with
peripheral blood (Barnes et al., 1989b, sarcoidosis (Bocart et al., 1992; Fidler et
1990). Tuberculin-negative patients with al., 1993), in spite of the absence of bacteria
active tuberculosis also have a greater num­ detectable by conventional means. It is
ber of monocytes and lymphocytes in their conceivable that similar forms exist in tu­
peripheral blood that exert apparently sup­ berculosis patients, and new studies with in
pressive effects in vitro (Elinor and Wallis, situ polymerase chain reactions should help
1989). Production of prostaglandin E2 (Ell- us explore this point.
ner and Wallis, 1989), or possibly IL-10 or
transforming growth factor p, by these
Mycobacteria and Idiopathic or
monocytes may contribute to their suppres­
“Autoimmune” Diseases
sive effects. Finally, as mentioned above,
Tuberculosis patients have a spectrum ot
the balance of type 1 (Thl) to type 2 (1 h2)
autoantibodies
that is remarkably similar to
T-cell subsets may be critical, and a switch
that
seen
in
rheumatoid
arthritis patients
from type 1 to type 2 responses (or to mixed
(Shoenfeld
and
Isenberg,
1988). They also
responses) could result in anergy or sup-

496

Rook and Bloom

have a striking change in the glycosylation
of the immunoglobulin G heavy chain, and
this is also characteristic of patients with
rheumatoid arthritis or Crohn’s disease and
a subset of patients with sarcoidosis (Rook
et al., 1993a). The immunotherapy dis­
cussed above leads to a rapid fall in percent
agalactosyl immunoglobulin G in tubercu­
losis patients (Rook et al., 1994), though the
significance of this remains to be eluci­
dated.
The known capacity of mycobacterium­
containing adjuvants (Freund's complete
adjuvant) to facilitate experimental induc­
tion of autoimmunity and evidence that
mycobacterial disease can be accompanied
by a sterile arthropathy (reviewed by Rook
et al. [1993a]) have reawakened speculation
that some autoimmune syndromes may be
cryptic infections or may be triggered by
past encounters with mycobacterium-like
organisms (Rook et al., 1993a). Interest was

I

I

'I :

i
J
i

I

■i

IJ

Fretd-s 'co^X- a^e

sponse may focus its attention on HSPs
because they are so conserved and may be
induced by the stress of infection or inflam­
mation. This may enable rapid recognition
of any pathogen. (Inadvertent autoimmu­
nity is a consequent risk.) (iii) Self-HSPderived peptides presented by MHC mole­
cules may be targets for cytotoxic cells.
This would enable the immune response to
detect and eliminate stressed autologous
cells, which might facilitate recognition of
transformed or infected cells.
At present, the evidence for recognition
of HSPs by T cells from synovia ofjoints of
patients with inflammatory arthritides or
from thyroid tissue of patients with thyroid­
itis (reviewed by Young [1992]), despite the
best eflforts of several laboratories, have
mostly been unsupportive of the HSP
cross-reactivity hypothesis as the basis for
autoimmunity, and sporadic reports of re­
activity to the human homolog in juvenile

J —- - — ~

recognize the mycobacterial 65-kDa HSP
(van Eden et al., 1988).
HSPs arc involved in assembly, folding
and transport of other cellular proteins and
aie expressed under various conditions of
stress. 'I hese functions are fundamental to
the survival of all life-forms, particularly
under stressful conditions, when synthesis
of HSPs may increase while synthesis of
other proteins is reduced. HSPs are highly
conserved throughout evolution, and there
is striking sequence homology between the
HSPs of microorganisms and those of
higher animals. These facts, together with
the demonstration that T cells mediating an
experimental autoimmune disorder recno
nize the mycobacterial HSP65 have led to
the following hypotheses, for each of which
there is some evidence (compare reviews
by Young and Elliott (1989], Polla [1991],
and Cohen and Young [1991]). (i) The im­
mune response to bacterial HSPs may

!

cross-rcact with host HSPs, leading to au­
toimmune disease, (ii) The immune re­

-- ------------ et al.,
1991) or in Per^z/zm-associatcd reactive ar­
thritis (Hermann et al., 1991) are uncon­
vincing. rely on few cell lines, and fail to
exclude the possibility of contamination of
the recombinant protein with the Esche­
richia coli products.
In tuberculosis, such autoimmune reac­
tions could contribute to both caseous ne­
crosis and liquefaction, but there is at
present no direct evidence for this. In fact,
desensitization to mycobacterial antigens
prevented cavity formation in animals (Yamamura et al., 1974), indicating that if there
is any autoimmune component, it is depen­
dent on initiation by a mycobacterium-spe­
cific cellular immune response. We suggest
u L
■
'
°U8h
eV'dCnCe
----- f— °r. recognition
of HSPs by CD4 Th cells is negligible,
possibility must be considered that com­
mon cross-reactive HSP antigens could be
recognized by MHC class I-restricted CTL.
The logic is that HSPs are expressed as
cytoplasmic antigens and would be ex-

Chapter 28

’s, leading to auFhe immune re­
tention on HSPs
rved and may be
fection or inflam­
rapid recognition
rtent autoimmu•) (iii) Self-HSP1 by MHC molecytotoxic cells.
mne response to
ssed autologous
e recognition of
Ils.
' for recognition
lovia of joints of
y arthritides or
its with thyroid92]), despite the
■oratories, have
3 of the HSP
as the basis for
c reports of re□log in juvenile
-Meeder et al.,
ted reactive ar91) are uncontes, and fail to
•ntamination of
ith the Esche-

jimmune reacth caseous ncit there is at
>r this. In fact,
terial antigens
n animals (Ya­
ng that if there
nt, it is depen>acterium-spee. We suggest
or recognition
is negligible,
ed that comgens could be
stricted CTL.
expressed as
'ould be ex­

•

Mechanisms of Pathogenesis

497

pected to be presented in association with human CTL, the nature of the antigens they
MHC class I antigens. Since any patho­ recognize, and their relation to protection,
gen—viral, bacterial, or protozoal—that in­ tissue damage, and autoimmunity.
vaded the cytoplasm of infected cells would
Acknowledgments. G.A.W.R. acknowledges sup­
induce the same conserved antigens, there port from the World Health Organization, the British
would be existing memory for CTL activity Medical Research Council, and the Wellcome Trust.
that could provide an early response. Cyto­
toxic CD4 T cells in humans have been
REFERENCES
reported to kill macrophages pulsed with
the mycobacterial 65-kDa HSP (Ottenhoff Ainslie, G. M., J. A. Solomon, and E. I). Bateman.
et al., 1988) and stressed autologous cells in
1992. Lymphocyte and lymphocyte subset numbers
in blood and in bronchoalveolar lavage and pleural
the absence of mycobacterial antigen, pre­
fluid in various forms of human pulmonary tubercu­
sumably through cross-reactive recognition
losis at presentation and during recovery. Thorax
of the autologous HSP (Koga et al., 1989),
47:513-518.
although this has not yet been confirmed. Al Attiyah. R., C. Moreno, and G. A. W. Rook. 1992.
Finally, the results of Silva and Lowrie (in
TNF-alpha-mediated tissue damage in mouse foot­
pads primed with mycobacterial preparations. Res.
press) indicate that expression of the 65Immunol. 143:601-610.
kDa heat shock cognate protein of BCG in
Anderson, M. C. 1891. On Koch’s treatment. Lancet
mouse macrophages provided effective im­
1:651-652.
munization against challenge with M. tu­ Armstrong, J. A., and I’. D. Hart. 1975. Phagosome­
lysosome interactions in cultured macrophages in­
berculosis and suggest that there must be
fected with virulent tubercle bacilli: reversal of the
some immune responses against HSP65
usual nonfusion pattern and observations on bacte­
that can provide protection. While defini­
rial survival. J. Exp. Med. 142:1-16.
tive evidence for a role of HSPs in cithei
Barnes, I). J.. S. Naraqi, P. Temu, and J. R. Turtle.
protection or autoimmunity is lacking, the
1989a. Adrenal function in patients with active tu­
available data are consistent with the hy­
berculosis. Thorax 44:422-424.
pothesis that if HSPs are involved, rather Barnes, P. F., J. S. Abrams, S. Lu, P. A. Siding, T. H.
Rea, and R. L. Modlin. 1993. Patterns of cytokine
than being important at the level of MHC
production by mycobacterium-reactive human T
class Il-restricted CD4 cells, they would
cell clones. Infect. Iinmun. 61:197-203.
preferentially engage CD8 MHC class Barnes, P. F., S.-J. Fong, P. J. Brennan, P. E.
Twomey, A. Mazumder, and R. L. Modlin. 1990.
I-restrictcd CTL. They could then exert
Local production of tumor necrosis factor and IFNcytotoxic activity on infected macrophages
gamma in tuberculous pleuritis. J. Immunol. 145:
containing more bacilli than they are able to
149-154.
kill and presumably could liberate the ba­ Barnes,
P. F., C. L. Grisso, J. S. Abrams, H. Band,
cilli to enable them to be phagocytosed by
T. H. Rea, and R. L. Modlin. 1992. Gamma delta T
lymphocytes in human tuberculosis. J. Infect. Dis.
infiltrating macrophages that may be acti­
165:506-512.
vated by the IFN--y released locally by the
CD8+ CTL or CD4+ Thl cells in granulo­ Barnes, P. F., S. I). Mistry, C. L. Cooper, C. Pirmez,
T. H. Rea, and R. L. Modlin. 1989b. Compartmen­
mas and be more effective in killing at lower
talization of a CD4+ T lymphocyte subpopulation in
multiplicities of infection. CTL might also
tuberculous pleuritis. J. Immunol. 142:1114-1119.
kill somatic cells in the vicinity either in­ Blauer. K. L., M. Poth, W. M. Rogers, and E. W.
Bernton. 1991. Dehydroepiandrosterone antago­
fected with M. tuberculosis or stressed to
nises the suppressive effects ol dexamethasone on
express endogenous HSPs in association
lymphocyte proliferation. Endocrinology 129:3174—
with MHC class I. Events that can be
3179.
interpreted in this manner arc seen in his­ Bloom, B. R., R. L. Modlin, and P. Salgame. 1992.
Stigma variations: observations on suppressor I
topathological studies (Dannenberg, 1991).
cells and leprosy. Annu. Rev. Immunol. 10:453-488.
Clearly, it will be important to undertake
experiments to establish the existence of Bocart. I)., I). Lecossier, A. de Lassence, 1). Valeyre,

&

498

Rook and Bloom

J.-P. Battesti, and A. Hance. 1992. A search for
human 60 kD heat shock protein by mononuclear
mycobacterial DNA in granulomatous tissues from
cells from patients with juvenile chronic arthritis
patients with sarcoidosis using the polymerase chain
Lancet 337:1368-1372.
reaction. Am. Rev. Respir. Dis. 145:1142-1148.
Denis, M. 1991a. Interferon-gamma-treated murine
Bordet, P. 1931. Contribution & I’dtude de I’allergie
macrophages inhibit growth of tubercle bacilli via
C.R. Soc. Biol. 107:622-623.
the generation of reactive nitrogen intermediates
( adranel, J., M. Garabedian. B. Milleron, H. Guillozo,
Cell. Immunol. 132:150-157.
G. Akoun, and A. J. Hance. 1990. 1,25(011)2 D3
Denis, M. 1991b. Killing of Mycobacterium tuberculo­
production by T lymphocytes and alveolar macro­
sis within human monocytes: activation by cyto­
phages recovered by lavage from normocalcaemic
kines and calcitriol. Clin. Exp. Immunol. 84:200patients with tuberculosis. J. Clin. Invest. 85:1588206.
1593.
Ellis, M. E., and F. Tayoub. 1986. Adrenal function in
Carswell, E. A., L. J. Old, R. L. Kassel. S. Green, W.
tuberculosis. Br. J. Dis. Chest 80:7-12.
Fiore, and B. Williamson. 1975. An endotoxin-in­
Ellner, J. J., and R. S. Wallis. 1989. Immunologic
duced serum factor that causes necrosis of tumours
aspects of mycobacterial infections. Rev. Infect
Proc. Natl. Acad. Sci. USA 72:3666-3670.
Dis. U(Suppl. 2):S455-S459.
< han, J., Y. Xing, R. S. Magliozzo, and B. R. Bloom.
Fidler. H. M., G. A. W. Rook, N. M. Johnson, and J.
1992. Killing of virulent Mycobacterium tuberculo­
McFadden. 1993. Mycobacterium tuberculosis DNA
sis by reactive nitrogen intermediates produced by
in tissue affected by sarcoidosis. Br. Med J 306activated murine macrophages. J. Exp. Med 175.546-549.
1111-1122.
Filley, E. A., H. A. Bull, P. M. Dowd, and G. A. W.
Cohen, 1. R., and D. B. Young. 1991. Autoimmunity,
Rook. 1992. The effect of Mycobacterium tubercu­
microbial immunity and the immunological homun­
losis on the susceptibility of human cells to the
culus. Immunol. Today 12:105-110.
stimulatory and toxic effects of tumour necrosis
Cooper, A. M., I). K. Dalton, T. A. Stewart, J. P.
factor. Immunology 77:505-509.
Griffin, D. G. Russell, and I. M. Orme. 1993. Dis­
Filley, E. A., and G. A. W. Rook. 1991. Effect of
seminated tuberculosis in interferon-y gene-dis­
mycobacteria on sensitivity to the cytotoxic effects
rupted mice. J. Exp. Med. 178:2242-2248.
of tumor necrosis factor. Infect. Immun. 59:2567Crowle, A. J. 1990. Intracellular killing of mycobacte­
2572.
ria. Res. Microbiol. 141:231-236.
Fine. P. E. M. 1989. The BCG story: lessons from the
Dalton, D. K., A. Pitts-Meek, S. Keshaw, I. S. Figari,
past and implications for the future. Rev. Infect
A. Bradley, and T. A. Stewart. 1993. Multiple de­
Dis. IKSuppl. 2):S353-S359.
lects of immune cell function in mice with disrupted
Fine, P. E. M. 1994. Immunities in and to tuberculosis:
interferon-gamma genes. Science 259:1739-1742
implications for pathogenesis and vaccination, p.
Dannenberg, A. M„ Jr. 1968. Cellular hypersensitivity
53-74. In K. P. W. J. McAdam and J. D. H. Porter
and cellular immunity in the pathogenesis of tuber­
(cd.), Tuberculosis: Back to the Future: Proceed­
culosis: specificity, systemic and local nature, and
ings of the London School of Hygiene and Tropical
associated macrophage enzymes. Bacterial Rev
Medicine 3rd Annual Public Health Forum. John
32:85-102.
Wiley, Chichester.
Dannenberg, A. M. 1990. Killing intracellular bacteriaFine, P. E. M., J. M. Ponnighaus, N. Maine, J. A.
dogmas and realities. Res. Microbiol. 141:193-196
Clarkson, and L. Bliss. 1986. The protective efficacy
Dannenberg, A. M. 1991. Delayed-type hypersensitiv­
of BCG against leprosy in northern Malawi. Lancet
ity and cell-mediated immunity in the pathogenesis
ii:499-502.
of tuberculosis. Immunol. Today 12:228-233.
Fischer, A., and W. Konig. 1991. Influence of cyto­
Daynes, R. A., B. A. Araneo, T. A. Dowell, K. Huang,
kines and cellular interactions on the glucocorticoidand D. Dudley. 1990. Regulation of murine lymphoinduced Ig (E, G. A. M) synthesis of peripheral
kine production in vivo. III. The lymphoid tissue
blood mononuclear cells. Immunology 74:228-233
microenvironment exerts regulatory influences over
Fitzgerald, T. J. 1992. The TH1/TH2 switch in syphi­
I helper cell function. J. Exp. Med. 171:979-996
litic infection: is it detrimental? Infect. Immun
Daynes, R. A., A. W. Meikle, and B. A. Araneo. 1991.
60:3475-3479.
Locally active steroid hormones may facilitate com­
Flesch, I. E. A., and S. H. E. Kaufmann. 1990. Acti­
partmentalization of immunity by regulating the
vation of tuberculostatic macrophage functions by
types of lymphokines produced by helper T cells.
gamma
interferon, interleukin-4, and tumor necrosis
Res. Immunol. 142:40-45.
factor. Infect. Immun. 58:2675-2677.
De Graeff-Meeder, E. R., R. van der Zee. G. T.
Flynn, J. L., J. Chan, K. J. Triebold, I). K. Dalton,
Rijkers, H. J. Schuurman, W. Kuis, J. W. Bijlsma,
T. A. Steward, and B. R. Bloom. 1993. An essential
B. J. Zegers, and W. van Eden. 1991. Recognition of
role for IFN-y in resistance to Mycobacterium tu-

i

Chapter 28

i by mononuclear
chronic arthritis.

ia-treated murine
iberclc bacilli via
en intermediates.
terium tuberculo'.ivation by cyto-

•nmunol. 84:200-

Irenal function in
7-12.
<9. Immunologic
ns. Rev. Infect.

Johnson, and J.
iberculosis DNA
?r. Med. J. 306:
1, and G. A. W.
:terium tubercuian cells to the
umour necrosis

1991. Effect of
ytotoxic effects
imun. 59:2567-

essons from the
e. Rev. Infect.
to tuberculosis:
vaccination, p.
J. D. H. Porter
lure: Proceedle and Tropical
i Forum. John

. Maine, J. A.
tective efficacy
lalawi. Lancet

lence of cytoducocorticoidof peripheral
y 74:228-233.
vitch in syphifect. Immun.
n. 1990. Actifunctions by
umor necrosis

D. K. Dalton,
. An essential
bacterium tu-

•

Mechanisms of Pathogenesis

499

intracellular microbial infections. Immunol. Today
9:168-174.
Kindler, V., A.-P. Sappino, G. E. Grau, P.-F. Piguet,
and P. Vassalli. 1989. The inducing role of tumor
necrosis factor in the development of bacterial gran­
ulomas during BCG infection. Cell 56:731-740.
Koch, R. 1891. Fortsetzung uber ein Heilmittel gegen
Tuberculose. Dtsch. Med. Wochenschr. 17:101-102.
Koga, T., A. Wand-Wurttenberger, J. DeBruyn, M. E.
Munk. B. Schoel, and S. H. Kaufmann. 1989. T cells
against a bacterial heat shock protein recognise
tion.
stressed macrophages. Science 245:1112-1115.
Foley, N., C. Lambert, M. McNicol, N. M. 1. Johnson,
Lurie, M. B. 1964. Resistance to Tuberculosis: Exper­
and G. A. W. Rook. 1990. An inhibitor of the
imental Studies in Native and Acquired Defensive
toxicity of tumour necrosis factor in the serum of
Mechanisms. Harvard University Press, Cam­
patients with sarcoidosis, tuberculosis and Crohn's
bridge, Mass.
disease. Clin. Exp. Immunol. 80:395-399.
Mackaness, G. B. 1968. The immunology of antituber­
Freudenberg, M. A., and C. Galanos. 1991. Tumor
culous immunity. Am. Rev. Respir. Dis. 97:331-344.
necrosis factor alpha mediates lethal activity of
Maggi. E., P. Parronchi, R. Manetti, C. Simonelli.
killed gram-negative and gram-positive bacteria in
M. P. Piccinni, F. S. Rugui, M. De Carli, M. Ricci,
D-galactosamine-treated mice. Infect. Immun. 59:
and S. Romagnani. 1992. Reciprocal regulatory ef­
2110-2115.
fects of 1FN-gamma and IL-4 on the in vitro devel­
Gordon, A. H., P. D’Arcy Hart, and M. R. Young.
opment of human TH 1 and TH2 clones. J. Immunol.
1980. Ammonia inhibits phagosome-lysosome fu­
148:2142-2147.
sion in macrophages. Nature (London) 286:79-81.
McDonough, K. A., Y. Kress, and B. R. Bloom. 1993.
Grzych, J. M., E. Pearce, A. Cheever, Z. A. Caulada,
Pathogenesis of tuberculosis: interaction of Myco­
P. Caspar, S. Henry, F. Lewis, and A. Sher. 1991.
bacterium tuberculosis with macrophages. Inject.
Egg deposition is the stimulus for the production of
Immun. 61:2763-2773.
TH2 cytokines in murine schistosomiasis mansoni.
Meikle, A. W., R. W. Dorchuck, B. A. Araneo, J. D.
J. Immunol. 146:1322-1340.
Stringham, T. G. Evans, S. L. Spruance, and R. A.
Haanen, J. B., R. de Waal-Malefijt, P. C. Res, E. M.
Daynes. 1992. The presence of a dehydroepiandros­
Kraakman, T. H. Ottenhoff, R. R- de Vries, and H.
terone-specific receptor-binding complex in murine
Spits. 1991. Selection of a human T helper type
T cells. J. Steroid Biochem. Mol. Biol. 42:293-304.
l-like T cell subset by mycobacteria. J. Exp. Med.
Moreno, C., J. Taverne, A. Mehlert, C. A. Bate, R. J.
Brealey, A. Meager, G. A. W. Rook, and J. H. L.
174:583-592.
Hermann, E., A. W. Lohse. R. Van der Zee, W. Van
Playfair. 1989. Lipoarabinomannan from Mycobac­
Eden, W. J. Mayet, P. Probst, T. Poralla, K. IL
terium tuberculosis induces the production ot tu­
Meyer zum Buschenfelde, and B. Fleischer. 1991.
mour necrosis factor from human and murine mac­
Synovial fluid-derived Yersinia-reactive T cells re­
rophages. Clin. Exp. Immunol. 76:240-245.
Mosmann, T. R., and K. W. Moore. 1991. The role of
sponding to human 65-kDa heat-shock protein and
IL-10 in crossregulation of TH1 and fH2 responses.
heat-stressed antigen-presenting cells. Eur. J. Im­
Immunol. Today 12(3):A49—A53.
munol. 21:2139-2143.
Myrvik, Q. N., E. S. Leake, and M. J. Wright. 1984.
Kabelitz, D., A. Bender, S. Schondelmaier, B. Schocl.
Disruption of phagosomal membranes of normal
and S. IL E. Kaufmann. 1990. A large fraction of
alveolar macrophages by the H37Rv strain of My­
human peripheral blood gamma/delta positive 1
cobacterium tuberculosis: a correlate ol virulence.
cells is activated by Mycobacterium tuberculosis
Am. Rev. Respir. Dis. 129:322-328.
but not by its 65-kD heat shock protein. J. Exp.
Nagao, S., and A. Tanaka. 1985. Necrotic inflamma­
Med. 171:667-679.
tory reaction induced by muramyl dipeptide in guin­
Kaplan, G. 1993. Therapy trials in tuberculosis pa­
ea-pigs sensitized by tubercle bacilli. J. Exp. Med.
tients. Robert-Koch-Symposium 1993 on Progress

herculosis infection. J. Exp. Med. 178:2249-2254.
Flynn, J. L., M. M. Goldstein, K. J. Triebold, B.
Koller, and B. R. Bloom. 1992. Major histocompat­
ibility complex class i-restricted T cells are required
for resistance to Mycobacterium tuberculosis infec­
tion. Proc. Natl. Acad. Sci. USA 89:12013-12017.
Flynn, J. L., K. Pfeffer, R. Schreiber, K. J. Triebold,
M. M. Goldstein, T. W. Mak, and B. R. Bloom.
TNF-a is necessary for resistance to Mycobacte­
rium tuberculosis in mice. Submitted for publica­

in Tuberculosis Research. Berlin. (Abstract.)
Kardjito, T., J. S. Beck, J. M. Grange, and J. L.
Stanford. 1986. A comparison of the responsiveness
to four new tuberculins among Indonesian patients
with pulmonary tuberculosis and healthy subjects.
Eur. J. Respir. Dis. 69:142-145.
Kaufmann, S. H. E. 1988. CD8+ T lymphocytes in

162:401^112.
Nathan, C. F., and J. B. Hibbs, Jr. 1991. Role of nitric
oxide synthesis in macrophage antimicrobial activ­
ity. Curr. Opin. Immunol. 3:65.
Onyebujoh, P., and G. A. W. Rook. 1991. Letter.
Lancet 338:1534.
Orme, I. M., E. S. Miller, A. D. Roberts, S. K. Furney,

500

Rook and Bloom

J. P. Griffin, K. M. Dobos, D. Chi, B. Rivoire, and
P. J. Brennan. 1992. T lymphocytes mediating pro­
tection and cellular cytolysis during the course of
Mycobacterium tuberculosis infection: evidence for
different kinetics and recognition of a wide spectrum
of protein antigens. J. Immunol. 148:189-196.
Ottenhoff, T. H., B. K. Ab, J. I), van Embden, J. E.
Thole, and R. Kiessling. 1988. The recombinant
65kDa heat shock protein of Mycobacterium bovis
Bacillus Calmette-Gucrin/M. tuberculosis is a target
molecule for CD4+ cytotoxic T lymphocytes that
lyse human monocytes. J. Exp. Med. 168:1947—
1952.
Palmer, C. E., and M. W. Long. 1966. Effects of
infection with atypical mycobacteria on BCG vacination and tuberculosis. Am. Rev. Respir. Dis.
94:553-568.
Pancholi, P., A. Mirza, N. Bhardwaj, and R. M.
Steinman. 1993. Sequestration from immune CD4+
T cells of mycobacteria growing in human macro­
phages. Science 26(1:984-986.
Phipps, R. P., S. H. Stein, and R. L. Roper. 1991. A
new view of prostaglandin E regulation of the im­
mune response. Immunol. Today 12:349-352.
Polla, B. S. 1991. Heat shock proteins in host-parasite
interactions. Immunol. Today 12(3):A38-A41.
Romagnani, S. 1991. Human TH1 and TH2 subsets:
doubt no more. Immunol. Today 12:256-257.
Rook, G. A. W. 1988. The role of vitamin D in
tuberculosis. Am. Rev. Respir. Dis. 138:768-770.
(Editorial.)
Rook, G. A. W. 1991. Mobilising the appropriate T-cell
subset: the immune response as taxonomist. Tuber­
cle 72:253-254.
Rook, G. A. W., et al. Unpublished data.
Rook, G. A. W., and R. Al Attiyah. 1991. Cytokines
and the Koch phenomenon. Tubercle 72:13-20.
Rook, G. A. W., B. R. Champion, J. Steele, A. M.
Varey, and J. L. Stanford. 1985. I-A restricted
activation by T cell lines of anti-tuberculosis activity
in murine macrophages. Clin. Exp. Immunol. 59:
414—420.
Rook, G. A. W., P. M. Lydyard, and J. L. Stanford.
1993a. A reappraisal of the evidence that rheuma­
toid arthritis, and several other idiopathic diseases,
are slow bacterial infections. Ann. Rheum. Dis.
52(Suppl. l):S30-S38.
Rook, G. A. W., P. Onyebujoh, and J. L. Stanford.
1993b. TH1—»TH2 switch and loss of CD4 cells in
chronic infections; an immuno-endocrinological hy­
pothesis not exclusive to HIV. Immunol. Today
14:568-569.
Rook, G. A. W., P. Onyebujoh, E. Wilkins, H. M. Ly,
R. Al Attiyah, G. M. Bahr, T. Corrah, H. Hernan­
dez, and J. L. Stanford. 1994. A longitudinal study of
% agalactosyl IgG in tuberculosis patients receiving
chemotherapy, with or without immunotherapy. Im­

munology 81:149-154.
Rook, G. A. W., J. Steele, M. Ainsworth, and B. R.
Champion. 1986a. Activation of macrophages to
inhibit proliferation of Mycobacterium tuberculosis:
comparison of the effects of recombinant gamma­
interferon on human monocytes and murine perito­
neal macrophages. Immunology 59:333-338.
Rook, G. A. W., J. Steele, L. Fraher, S. Barker, R.
Karmali, J. O'Riordan, and J. Stanford. 1986b.
Vitamin D3, gamma interferon, and control of pro­
liferation of Mycobacterium tuberculosis by human
monocytes. Immunology 57:159-163.
Rook, G. A. W., J. Taverne, C. Leveton, and J. Steele.
1987. The role of gamma-interferon, vitamin D3
metabolines and tumor necrosis factor in the patho­
genesis of tuberculosis. Immunology 62:229-234.
Rothstein, J., and H. Schreiber. 1988. Synergy be­
tween tumor necrosis factor and bacterial products
causes hemorrhagic necrosis and lethal shock in
normal mice. Proc. Natl. Acad. Sci. USA 85:607611.
Salgame, P. R., J. S. Abrams, C. Clayberger, H.
Goldstein, R. L. Modlin, J. Convit, and B. R. Bloom.
1991. Differing lymphokine profiles of functional
subsets of human CD4 and CDS T cell clones.
Science 254:279-282.
Sampaio, E. P., E. N. Sarno, R. Galilly, Z. A. Cohn,
and G. Kaplan. 1991. Thalidomide selectively inhib­
its TNFa production by stimulated human mono­
cytes. J. Exp. Med. 173:699-703.
Sarnia, G. R., I. Chandra, G. Ramachandran, P. V.
Krishnamurthy, V. Kumaraswami, and R. Prabhakar. 1990. Adrenocortical function in patients
with pulmonary tuberculosis. Tubercle 71:277-282.
Scott, G. M., P. G. Murphy, and M. E. Gemidjioglu.
1990. Predicting deterioration of treated tuberculo­
sis by corticosteroid reserve and C-reactive protein.
J. Infect. 21:61-69.
Shands, J. W., and V. C. Senterfitt. 1972. Endotoxininduced hepatic damage in BCG-infected mice. Am.
J. Pathol. 67:23-40.
Shepard, C. C. 1958. A comparison of the growth of
selected mycobacteria in HcLa, monkey kidney,
and human amnion cells in tissue culture. ./. Exp.
Med. 107:237-246.
Shoenfeld, ¥., and D. A. Isenberg. 1988. Mycobacteria
and autoimmunity. Immunol. Today 9:178-182.
Shwartzman, G. 1937. Phenomenon of Local Tissue
Reactivity and Its Immunological. Pathological,
and Clinical Significance, p. 461. Paul B. Hoeber,
New York.
Silva, C. L., and L. H. Faccioli. 1988. Tumor necrosis
factor (cachectin) mediates induction of cachexia by
cord factor from mycobacteria. Infect. Immun. 56:
3067-3071.
Silva, C. L., and D. B. Lowrie. A single mycobacterial
protein (hsp60) expressed by a transgenic antigen-

Chapter 28

th, and B. R.
'Crophages to
tuberculosis:
•nant gamma­
nurine perito3-338.
•». Barker, R.
iford. 1986b.
mtrol of provw by human

and J. Steele,
vitamin D3
in the patho2:229-234.
Synergy berial products
lal shock in
JSA 85:607tyberger, H.
B. R. Bloom,
f functional
cell clones.

Z. A. Cohn,
lively inhibman mono­
dran, P. V.
d R. Prabin patients
71:277-282.
iemidjioglu.
' tuberculoive protein.
Endotoxin­
mice. Am.

growth of
jy kidney,
e. J. Exp.
cobacteria
'8-182.
cal Tissue
'hological.
I. Hoeber.
>r necrosis
ichexia by
nmun. 56:

obacterial
c antigen-

presenting cell vaccinates mice against tuberculosis.
Immunology, in press.
Singhal, M., J. N. Banavalikar, S. Sharma, and K.
Saha. 1989. Peripheral blood T lymphocyte subpop­
ulations in patients with tuberculosis and the effect
of chemotherapy. Tubercle 70:171-178.
Stanford, J. L. 1987. Much’s granules revisited. Tu­
bercle 68:241-242.
Stuehr, D. J., H. J. Cho, N. S. Kwon, M. F. Weise, and
C. F. Nathan. 1991. Purification and characteriza­
tion of the cytokine-induced macrophage nitric ox­
ide synthase: an FAD- and FMN-containing fla­
voprotein. Proc. Natl. Acad. Sci. USA 88:77737777.
Sturgill-Koszycki, S., P. H. Schlesinger, P.
Chakraborty, P. L. Haddix, H. L. Collins, A. K. fok,
R. D. Allen, S. L. Gluck, J. Heuser, and D. G.
Russell. Mycobacterium phagosomes fail to acidify
through exclusion of the vesicular proton-ATPase.
Science, in press.
Suzuki. T., N. Suzuki, R. A. Daynes, and E. G.
Engleman. 1991. Dehydroepiandrosterone enhances
IL2 production and cytotoxic effector function of
human T cells. Clin. Immunol. Immunopathol. 61:
202-211.
Takashima, T., C. Ueta, I. Tsuyuguchi, and S. Kishimoto. 1990. Production of tumor necrosis factor
alpha by monocytes from patients with pulmonary
tuberculosis. Infect. Immun. 58:3286-3292.
Thompson, N. J., J. L. Glassroth, D. E. Snider, Jr.,
and L. S. Farer. 1979. The booster phenomenon in
serial tuberculin testing. Am. Rev. Respir. Dis.
119:587-597.
Uyemura, K., R. J. Deans, H. Band. J. Ohmen, G.

•

Mechanisms of Pathogenesis

501

Panchamoorthy, C. T. Morita, T. H. Rea, and R. L.
Modlin. 1991. Evidence for clonal selection of gamma/delta T cells in response to a human pathogen. J.
Exp. Med. 174:683-692.
Valone, S. E., E. A. Rich, R. S. Wallis, and J. J. Ellner.
1988. Expression of tumor necrosis factor in vitro by
human mononuclear phagocytes stimulated with
whole Mycobacterium bovis BCG and mycobacte­
rial antigens. Infect. Immun. 56:3313-3315.
van Eden, W., J. E. R. Thole, R. van der Zee, A.
Noordzij, J. D. A. van Embden, E. J. Yamamura,
E. J. Hensen, and I. R. Cohen. 1988. Cloning of the
mycobacterial epitope recognised by T lymphocytes
in adjuvant arthritis. Nature (London) 331:171-173.
Wilson, G. S., H. Schwabacher, and I. Maier. 1940.
The effect of the desensitisation of tuberculous
guinea-pigs. J. Pathol. Bacteriol. 50:89-109.
Wisniewski. T. L., C. W. Hilton, E. V. Morse, and F.
Svec. 1993. The relationship of serum DHEA-S and
cortisol levels to measures of immune function in
human immunodeficiency virus-related illness. Am.
J. Med. Sci. 305:79-83.
Xu, S., S. Sturgill-Koszycki, T. van Heyningen, A.
Cooper, D. Chatterjee, I. Orme, P. Allen, and 1). G.
Russell. Intracellular trafficking and the Mycobacterm/n-infected macrophage. J. Immunol., in press.
Yamamura, Y., Y. Ogawa, II. Maeda, and Y. Yama­
mura. 1974. Prevention of tuberculous cavity forma­
tion by desensitization with tuberculin-active pep­
tide. Am. Rev. Respir. Dis. 109:594-601.
Young, D. B. 1992. Heat-shock proteins: immunity and
autoimmunity. Curr. Opin. Immunol. 4:396-400.
Young, R. A., and T. J. Elliott. 1989. Stress proteins,
infection, and immune surveillance. Cell 59:5-8.

Position: 4731 (1 views)