Intern. Rev. ImmunoI., Vol. 19, pp.

5142 0 2Mw) OPA (Overseas Publishers Association) N.V.

Rcpnnu available dinctly from the publisher Published by license under
Photocopying permitted by lianse only the Harwood Academic Publishers imprint.
part of Tbe Gordon and Breach Publishing Group.
Printed in Malaysia.

TNBS-Colit is
MARKUS NEURATH ,’ IVAN FUSS and WARREN STROBER ’-*
Int Rev Immunol Downloaded from informahealthcare.com by Memorial University of Newfoundland on 10/01/14

of Immunology, 1. Medical Clinic, University of Mainz,

a Laboratory
Main, Germany; Mucosal Immunity Section, LCI, Building 10,
Room 11N238, NIAID, NIH, 10 Center Drive, MSC 1890,
Bethesda, MD 20892-7890, USA

Disease states, such as the occurrence of gastrointestinal inflammation (Crohn’s disease

and ulcerativecolitis), can be secondary to a host of determinants that act in conjunction
to bring about pathologic change. The underlying factors that mediate the development
of such mucosal inflammation has recently been brought to the forefront with the advent
of animal models. The examination of these animal models have given researchers a
better understanding of the mechanisms involved in the pathogenesis of inflammatory
bowel disease. This review discusses one such model, TNBS-colitis, and the insights that
it provides into the occurrence of IBD and its future treatment.
For personal use only.

Keywonis: Inflammatory bowel disease; Experimental colitis; Interleukin-12

INTRODUCTION

In recent years, a number of murine models of chronic colitis have been

developed which are remarkably similar to one or another form of
human inflammatory bowel disease. As such, these models provide an
excellent opportunity to study the immunopathogenesis and possible
treatment of these idiopathic diseases. One such model is TNBS-colitis,
a chronic colitis in mice induced by the intra-rectal administration of
trinitrobenzene sulfonic acid (TNBS). In the following section we shall
discuss our recent studies of this model and show how these studies
have led to new insights into the immunologic mechanism underlying
Crohn’s disease.

*Corresponding author. Tel.: 301-496-6810. Fax: 301-402-2240.

51
 52 M. NEURATH et al.

CHARACTERISTICS OF TNBS-COLITIS

TNBS is a classical skin contactant, i.e., a chemical compound that

induces delayed hypersensitivity reactions when applied to the skin
because it haptenates body proteins with trinitrophenyl (TNP) groups
and renders such self-protein immunogenic to the immune system.
Thus, as shown in Fig. 1, when TNBS is introduced into the colon of
susceptible mice via intra-rectal instillation, it induces T cell-mediated
immune response in the colonic mucosa, in this case leading to a
Int Rev Immunol Downloaded from informahealthcare.com by Memorial University of Newfoundland on 10/01/14

massive mucosal inflammation characterized by the dense infiltration

of T cells and macrophages throughout the entire wall of the large
bowel [11. Moreover, this histopathologic picture is accompanied by
the clinical picture of progressive weight loss (wasting), bloody
diarrhea, rectal prolapse and large bowel wall thickening; in short, a
pattern not dissimilar to that seen in human Crohn’s disease.
One important difference between skin contact hypersensitivity and
TNBS-colitis induced by application of TNBS is that in the skin the
reaction is self-limited whereas in the colon the reaction is persistent.
This is likely to be due to the fact that the immune cells called forth by
TNBS cross-react with ubiquitous mucosal antigens i.e., antigens that
exist in the bacterial microflora, and thus continue to be stimulated
For personal use only.

even after the T”-haptenated proteins have disappeared. Evidence in

support of this view is severalfold. First, if lamina propria T cells in a
mouse with TNBScolitis are transferred to a naive mouse they cause
mild but definite colitis in the recipients; since antigen is not
transferred with such T cells, this reaction is probably due to cross-
reactivity to antigens encountered in the mucosal environment of
the new host [2]. Second, IL-2 deficient mice which have been shown
to develop spontaneous colitis under certain circumstances can be
induced to develop colitis within days by the systemic injection of
TNP-KLH;this suggests that T cells stimulated by TNP can traflic to
the gut where they encounter cross-reactive antigens and induce
colitis [3]. Third and finally, it has recently been shown that T cells
from mice with TNBS-colitis proliferate in response to exposure to
their own flora whereas normal mice do not [4]; this finding implies
that in the normal situation, a mouse is “tolerant” to its own flora and
such tolerance is broken in TNBS-colitis; we shall return to this
important concept below.
 TNBSCOLITIS 53
Int Rev Immunol Downloaded from informahealthcare.com by Memorial University of Newfoundland on 10/01/14

FIGURE 1 Histologic analysis of the colon in mice with TNBS-induced colitis. Photo-
micrograph of HE-stained cross section (xS0) of a colon of a SJL/J mouse with
TNBS induced colitis. (See Color Plate I.)
For personal use only.

IMMUNOLOGIC CHARACTERIZATIONOF TNBS-COLITIS

In our initial studies of TNBS-colitis, we sought to characterize the

types of T cells that differentiate in the colon in response to TNBS
stimulation [l]. We thus extracted purified CD4+ T cells from the
colon and spleen of mice with TNBS-colitis and then stimulated the
extracted cells with polyclonal stimulants such as anti-CD3/anti-CD28
to determine their capacity to secrete cytokines. In such studies, a
predominance of IFN-y secretion would be indicative of a Thl T cell
response whereas that of IL-4 secretion would be indicative of a Th2
T cell response. As shown in Fig. 2, we found that the T cells in the
TNBS-colitis lesions produce greatly increased amounts of IFN-y and
decreased amounts of IL-4 following stimulation in vitro (as compared
to T cells in controls), a clear indication that in TNBS-colitis the re-
levant immune response is in fact a Thl T cell reaction. In other studies
 54 M. NEURATH et al.

500 1 T
Int Rev Immunol Downloaded from informahealthcare.com by Memorial University of Newfoundland on 10/01/14

--
z 4
a
ethanolgroup
TNBScolitis group

2 3

E
rn
g 2

f
1
For personal use only.

n
lnda aCD3/ am28

FIGURE 2 Cytokine production of stimulated and unsiimulated LP CD4+ T celk in

TNES-induced colitis. LP CD4+ T cells were isolated from TNBS- and control ethanol-
treated mice on day 7, cultured for 2 days in the absence or presence of anti-CD3 and
anti-CD28 and culture supernatants were analyzed for concentration of 1FN-y and
IL4. Data represent three independent experiments done in triplicate.. Standard errors
are indicated.

we established that the T cells were also producing high amounts of

TNF-a, another Thl T cell cytokine.
If TNBS-colitis is indeed a Thl T cell-mediated inflammation as
suggested by the findings described above, it should be associated with
increased production of the cytokine that plays a central role in Thl
T cell differentiation, IL-12. We therefore performed in situ immuno-
histologic studies in which tissue from mice with and without TNBS-
colitis were stained with fluorochrome-labeled antibodies specific for
 TNBS-COLITIS 55

IL-12 (as well as control antibodies). With this approach we showed

that TNBS-colitis is characterized by the presence of vastly increased
amounts of tissue-associated IL- 12, strongly implying that IL- 12
synthesis is increased in the inflamed tissue. These results, plus those
obtained from studies recounted below showing that anti-IL-12
abrogates TNBS-colitis, provide very solid evidence that TNBS-colitis
is an IL-12-driven Thl T cell-mediated inflammation.
On the basis of the above information, TNBS-colitis is best explained
by the immunologic mechanism depicted in Fig. 3 which can be
Int Rev Immunol Downloaded from informahealthcare.com by Memorial University of Newfoundland on 10/01/14

described as follows: In the first step of the process, the application of

TNBS to the rectal mucosa results in the presentation of TNP to TNP-
specific (and cross-reactive) T cells, which as a result become activated
and express the CD40L signaling molecule; the latter via its interaction
with the CD40 molecule on the APCs then allow the T cells to back
stimulate the APCs to produce IL-12. In the next step of the process,
the IL- 12 produced by APCs induces T cells in the lamina propria to
undergo Thl T cell differentiation and thus cells which secrete IFN-y
and TNF-a. The latter, in turn, bring about two effects: on the one
hand, these cytokines act “up-stream” on APCs to induce increased
amounts of IL-12 and thus to magnify the Thl T cell reaction; on the
other hand, these cytokines act “down-stream” on macrophages to
For personal use only.

induce the latter to produce a host of pro-inflammatory cytokines such

as IL- I/?, IL-6 and (additional) TNF-a. Finally, in the last step of the

11-12

(3f IL-1p

FIGURE 3 Induction of Thl Responses: origin of intammation in TNBS-colitis.

Antigen-presenting cells (APCs) present antigen to T cells which, in turn, express
CD4OL and back-stimulate the APC via CD40. This leads to IL-12 secretion and the
induction of IFN? by the T cell. IFN-7 stimulates macrophages (macs) to produce
the mediators of idammation, T N F - a , IL-6 and IL-lp IFN-7 and TNF-a act in
synergy to increase further production of IL-12 in a positive feedback response.
 56 M.NEURATH et al.

process, these latter substances cause the influx of cells giving rise to the
inflammation of TNBS-colitis.

TREATMENT OF TNBS-COLITIS

One of the predictions of this proposed sequence of events underlying

TNBS-colitis is that the latter should be treatable by the systemic
administration of antibodies (or other agents) that interfere with the
Int Rev Immunol Downloaded from informahealthcare.com by Memorial University of Newfoundland on 10/01/14

sequence at any one of several stages: in particular, it should be

treatable by the systemic administration of anti-IL-12. To formally test
this possibility we administered anti-IL-12 to mice either at the same
time as TNBS was administered per rectum or after two to three weeks
following such administration. As shown in Fig. 4, these treatment
regimens were remarkably effective in that they either completely
prevented the development of TNBS-colitis (when given at the time of
induction of the disease) or led to dramatic resolution of the TNBS-
colitis (when given when the lesion was fully developed). These studies,
in showing that a murine model resembling Crohn’s disease is treatable
For personal use only.

FIGURE 4 Histologic analysis of the colon in mice with TNBS-induced colitis given
anti-IL-12. Photomicrograph of HE-stained cross-section (xS0) of a colon of a SIL/J
mouse with TNBSinduced colitis after treatment with anti-IC12 antibodies. A strik-
ing reduction in inflammatory activity is evident. (See Color Plate 11.)
 TNBSCOLITIS 57

with anti-IL-12, imply that Crohn’s disease itself is also treatable in

this fashion. On the basis of this possibility, we are currently planning
to test “humanized” anti-IL-12 in clinical trials of patients with crohn’s
disease (see further discussion below).
While the blocking of IL-12 activity with anti-IL-12 may be an
efficient way of interrupting the Thl T cell activation pathway
necessary for TNBScolitis, it is certainly not the only way the latter
can be accomplished. As indicated above, the critical APC-T cell inter-
actions leading to Thl T cell differentiation requires T cell expression
Int Rev Immunol Downloaded from informahealthcare.com by Memorial University of Newfoundland on 10/01/14

of CD4OL on activated T cells and signaling of APC via CD40 for

production of IL-12. Thus, it is theoretically possible to interfere with
Thl T cell differentiation by blocking the CD40LXD40 interaction.
To test this possibility, we administered anti-CD40L antibody to mice
at the time of TNBS-colitis induction with intra-rectal TNBS adminis-
tration [5]. Such treatment did indeed prevent colitis induction as well
as the increased IFN--y production in the lamina propria associated
with colitis and thus is a second avenue available for the prevention of
TNBS-colitis. Whether anti-CD40L can also be used to treat ongoing
TNBS-colitisas can anti-IL-12 awaits further study.
Another way the Thl pathway can be experimentally thwarted in
the context of the TNBScolitis model would be to inhibit more distal
For personal use only.

inflammatory cytokine effects, either by the administration of anti-

IFN-y or anti-TNF-a. In studies relevant to this possibility we found
that anti-IFN-7 also inhibited the development of colitis, although
such inhibition was not as effective as that achieved with anti-IL-12
[6,7]. Thus, while prevention of TNBS-colitis was inhibitable with a
single injection of anti-IL-12, prevention of colitis with anti-IFN-7
required multiple injections and was, in general, not as complete. In
addition, wherein mice cured of colitis with anti-IL-12 were not subject
to re-induction of colitis with sub-colitis-inducing doses of TNBS, mice
cured of colitis with anti-IFN-y were subject to such colitis-induction
[7]. This suggests that anti-11-12-treated mice no longer have cells
reactive with colitis whereas anti-IFN-?-treated mice have such cells.
Important confirmation of this possibility has recently come from the
observation that mice with TNBS-colitis treated with anti-IL-12 dis-
play large numbers of apoptotic cells in the inflamed colons whereas
the same mice treated with anti-IFN--y display only modest numbers of
apoptotic cells at this site. Overall, then, these studies strongly suggest
 58 M. NEURATH ef al.

that the efficacy of anti-IL-12 in TNBS-colitis is due to the fact that

the antibody not only neutralized anti-IL-12 but also rids the body of
the disease-causingThl T cells.
In parallel studies exploring the role that anti-TNF-a might play in
the treatment of TNBS-colitis we have demonstrated that this anti-
body also is effective in treating established colitis, and that mice
expressing a transgenic TNF-a gene treated with TNBS per rectum
rapidly develop lethal colitis. In further studies we have shown that
isolated macrophages from normal mice with TNBS-colitis treated
Int Rev Immunol Downloaded from informahealthcare.com by Memorial University of Newfoundland on 10/01/14

with anti-TNF-a not only produced less TNF-a but also less IL-10
and IL-6 in culture. This last result implies that the effect of anti-TNF-
a is not simply on blocking the peripheral effect of TNF-a (on pro-
inflammatory TNF-a production) but also in reducing the capacity of
TNF-a to provide positive feedback to IL-12 synthesis, thereby
leading to decreased IL-12 of the inflammation and the abatement of
all pro-inflammatory cytokine levels, Whether this effect on IL- 12
production is the main mechanism of its therapeutic efficacy remains
to be seen, as does the efficacy of this antibody in comparison with
anti-IL-12.
In a final and novel approach to the therapy of TNBS-colitis, we
have also inhibited “distal” TNF-a production by administration of a
For personal use only.

substance which down-regulates the activity of a transcriptional factor

necessary for TNF-a gene transcription. NF-KB [S]. In these studies
we affected such down-regulation by the administration (either by
intra-rectal instillation or by systemic injection) of an oligonucleotide
that blocks functional NF-KB p65 chain production at the mRNA
transcription level to mice with TNBS-colitis. This approach to the
treatment of TNBS-colitis might be especially useful in the treatment
of human (Crohn’s) colitis because the action of the anti-sense
oligonucleotide are very specific, and are likely to be unassociated with
harmful side effects.

THE MECHANISM OF TNBS-COLITIS

So far in our discussion we have discussed the “how” of TNBS-colitis

but not the “why”. In our approach to this question, we were aware of
the fact that mucosal responses are normally under the strict control
 TNBS-COLITIS 59

of counter-regulatory responses known collectively as ”oral tolerance”

[9]. By such responses, encounters between the mucosal immune
system and mucosal antigens normally result in tolerance rather than
immunity and thus the mucosal immune system is spared of responses
that could lead to autoimmunity and/or inflammation. Thus, it seemed
possible that TNBScolitis arises because the administration of TNBS
per rectum bypasses or subverts the oral tolerance mechanism.
In our studies of the role of oral tolerance (or lack thereof) in
TNBS-colitis, we took note of the fact that oral tolerance is now
Int Rev Immunol Downloaded from informahealthcare.com by Memorial University of Newfoundland on 10/01/14

known to be due to two independent but interacting processes, the

induction of suppressor T cells producing TGF-P (and perhaps other
suppressive cytokines) and the induction of clonal deletion and/or
anergy [9,10]. In first of these, the induction of suppressor T cells, is
the dominant mechanism of tolerance at low antigen doses and
appears to be dependent on the processing and presentation of antigen
to T cells in the Peyer’s patches. In contrast, the second of these
processes, the induction of clonal deletion or anergy, is the dominant
mechanism of tolerance at higher antigen doses and is dependent on
the occurrence of “processed” antigen which induces deletion and or
anergy not only in mucosal tissues but also in other tissues as well.
Emerging evidence shows that this deletional tolerance is independent
For personal use only.

of suppressor T cell development and the presence of suppressor

cytokines. Finally, we also noted that with respect to oral tolerance
mediated by suppressor T cells, Th2 T cell cytokines (IL-4) appear to
favor suppressor T cell development whereas Thl T cell responses
(IFN-y) appears to inhibit the latter [9,11].
Collectively, the above facts suggested to us that if TNBS could be
administered to mice in a form that elicited oral tolerance, particularly
the suppressor T cell-mediated form of tolerance, one could prevent
TNBS-colitis. Accordingly, we administered TNP-haptenated colonic
protein (a form of TNP-linked antigen that would not induce inflam-
mation in the upper GI tract as a result of feeding) to mice orally at the
time when we administered TNBS per rectum in the usual way used to
induce TNBS-colitis [I 11. We found that such oral TNBS-protein
administration completely prevented the development of TNBS-colitis
and that the mice receiving oral TNBS-protein, instead of develop-
ing a Thl T cell response in the intestine (i.e., cells producing JFN-y)
develop a Th2 response (i.e., cells producing IL-4). Perhaps more
 60 M. NEURATH et al.

importantly, we found that T cells were now found in both the Peyer’s
patches and lamina propria, which upon stimulation with TNBS
in vitro, produced the suppressor cytokine TGF-P. Finally, we showed
that the TGF-P-producing T cells were in fact the cells responsible for
lack of development of colitis: administration of anti-TGF-P antibody
to the animals receiving both TNBS protein per 0 s and TNBS per
rectum resulted again in mice with colitis.
The demonstration that the induction of TNBS-colitis by TNBS
administration per rectum is prevented by the administration of
Int Rev Immunol Downloaded from informahealthcare.com by Memorial University of Newfoundland on 10/01/14

TNBS-protein by mouth and the induction of tolerance to TNBS (as

illustrated in Fig. 5 ) strongly suggests that TNBS-colitis induced by
intra-rectal TNBS administration is in fact due to a failure of oral
tolerance development. One possible mechanism of such failure is that
intra-rectal TNBS induces a massive Thl T cell response directly in the
For personal use only.

11-12

FIGURE 5 Mechanism of colitb induction and tolerance in TNBS-colitis. TNBS

administration intra-rectally and formation of TNP-haptenated colonic protein lead to
induction of a polarized Thl T-cell response in the lamina propria that is not coun-
ter-regulated by TGF-P-producing T cells; the resultant unfettered IFN-y production
leads to activation of macrophages, production of inflammatory cytokines and devel-
opment of colitis (left). Oral administration of TNP-haptenated colonic protein leads
to induction of T cells producing TGF-P, which shuts down the excessive Thl T e l l
response of TNBS-colitis (right). TNP = trinitrophenyl;TGF-P = transforming growth
factor-beta; IFNq = interferon-gamma.
 TNBSCOLITIS 61

lamina propria which is unaccompanied by the generation of TGF-P-

producing suppressor T cells. In addition, in this situation any and all
suppressor T cells developing in the Peyer’s patches following the
intra-rectal TNBS administration is quickly suppressed by IFN-y
production which, as mentioned above, is known to inhibit such
suppressor cell development.

CONCLUSION
Int Rev Immunol Downloaded from informahealthcare.com by Memorial University of Newfoundland on 10/01/14

Taken together, these studies suggest that a ying/yang relation between

Thl (IL-l2/12/IFN-.y) responses on the one hand and TGF-/3 T cell
responses on the other governs whether or not inflammation develops
in the gastrointestinal tract. Thus, in normal individuals, the mucosal
response set point is shifted towards tolerance and non-responsive-
ness so that inflammation does not develop in relation to ordinarily
harmless exposure to mucosal antigens. In contrast, in individuals with
inflammatory bowel diseases i.e., Crohn’s disease, the mucosal res-
ponse set point is shifted toward responsiveness and we have the
development of inflammation. Evidence in favor of this view is
inherent in the recent finding that in Crohn’s disease, just as in TNBS-
For personal use only.

colitis, there is heightened reactivity to one’s own bacterial microflora

and thus a loss of tolerance to these mucosal constituents [12]. Future
study will be focused on defining the precise mechanisms that
determine the all-important mucosal response set point, i.e., the
genetically-determined factors that determine whether or not tolero-
genic responses or immunogenic response in the mucosa will be
dominant. Such studies will help to define the fundamental factors
which determine why some individuals develop Crohn’s disease and
others do not. Meanwhile, work must go forward directed at attempts
to treat Crohn’s disease by addressing the outcome of these funda-
mental abnormalities, the Thl T cell final common pathway.

References

[I] M. Neurath, I. Fuss,B. Kelsall, E. Stuber and W. Strober. ‘Antibodies to interleukin

12 abrogate established experimental colitis in mice.” J. Exp. Med. 182 1281-1290,
1995.
 62 M. NEURATH et al.

121 M. Neurath, I. Fuss, B. Kelsall, D. Presky, W. Waegall and W. Strober.

‘Experimental granulomatous colitis in mice is abrogated by induction of TGF-/3-
mediated oral tolerance.” J. Exp. Med. 183:26052616,1996.
[3] R. Ehrhardt, B. Ludviksson, B. Gray, M.Neurath and W. Strober. ’Induction and
prevention of colonic inflammation in IL-2 deficient mice.” J. Immunol. 158: 566-
573, 1996.
[4] R. Duchmann, E. Schmitt, P. Knolle, K.H. Meyer nun Buschenfelde and
M. Neurath. ‘Tolerance towards resident intestinal flora in mice is abrogated in
experimental colitis and restored by treatment with interleukin-10 or antibodies to
interleukin-12.”Eur. J. Immunol.26 934938,1996.
[q E. Stuber, W. Strober and M. Neurath. “Blocking the CD40L-CD40 interaction
in vivu specifically prevents the priming of T helper 1 cells through the inhibition of
interleukin 12 secretion.” J. Exp. Med. 183 693698,1996.
[q M. Neurath, I. Fuss,M. Pasparakis, L. Alexopoulou, S. Haralambous, K.H. Meyer
Int Rev Immunol Downloaded from informahealthcare.com by Memorial University of Newfoundland on 10/01/14

zum Buschenfelde, W.Strober and G. Kollias. ’Predominant pathogenic role of

tumor necrosisfactora in experimentalcolitis in mice.” Eur. J. I m ~ c n o l27
.
[7l I. Fuss,T. Marth, G.Pearlstein, M.Neurath and W. Strober. “Anti-interleukin-12
regulates apoptosis of activated Thl T cells in murine experimental granulomatous
( T N B S ) colitis.” Gartroenteroloy 112 A978,1997.
[8] M. Neurath, S. Pettersson, K.H. Meyer zum Buschenfelde and W. Strober. “Local
administration of antisense phosphorothioate oligonucleotides to the p65 subunit
of NF-kappa B abrogates established experimental colitis in mice.” Nut. Med. 2:
99&1004,1996.
[9] H. Weiner. “Oral tolerance: immune mechanisms and treatment of autoimmune
disease.” Zmmunobgy Today 18:335-343,1997.
[lo] W.Strober. ‘OralTolerance.” J. Clin. Immunol. 18:1-30,1998.
[ll] R. Seder, T. M a d , M.Sieve, W. Strober, J. Letterio, A. Roberts and B. Kelsall.
“Factors involved in the differentiation of TGF-pproducing cells from naive CD4+
T cells: IG4 and IFN? have opposing effects while TGF-/3 positively regulates
its own production.” J . Immunol. 160: 571%5728,1998.
For personal use only.

[12] R. Duchmann, I. Kaiser, L. Hennann, W.Mayet, K. Ewe and K.H. Meyer zum
Buschenfelde. Tolerance exists towards resident intestinal flora but is broken in
active inflammatory bowel disease (IBD).“ Clin. Exp. Immwrol. 102 445447,1995.