|
1 Laboratory of Immunobiology, Department of Immunology,
Pasteur Institute, Paris, France
2 Department of Immunology, Clinica Puertade Hierro, Madrid, Spain
A Introduction
After the introduction of Jerne's network theory 10 years ago,
modern views of the immune system have been strongly influenced
by the ideas of autonomy previously developed by those working on
complex systems such as the brain. This "new immunology" attempts
to describe the normal immune system as it is, the fundamental characteristics
of its organization and, in the words of Francisco Varela, "the
landscape of its eigen behaviours." Within this perspective, the
study of the internal activity of the immune system, of the generation
and decay of its cellular and molecular components, is likely to
provide more relevant information as to the physiology of immune
activity than the study of immune responses to the injection of
large doses of antigen, as in classical approaches. Obviously, we
aim to describe and understand the basis for this inner life of
the immune system, embedded as it is in a multitude of molecular
components that constitute the "self' of the individual, and exposed
to the surrounding "noise" of the environment. We do not use artificial
priming with antigen, neither antibodies nor antiidiotypes as surrogates
of antigens in the induction of immune responses. On the other hand,
the interpretation of such internal activity, as well as any hypothesis
on the internal mechanisms inducing effector cells and their specificity,
must necessarily rely on detailed knowledge of lymphocyte physiology:
the mechanisms by which lymphocytes are turned on and turned off
and how their proliferation and maturation to effector functions
are regulated as well as the nature of the functionally relevant
molecules expressed at the surface of these cells. In a meeting
on human leukemia, what we have to say risks being completely out
of place. We neither work with human cells, nor are we concerned
with leukemia. Moreover, our perspectives and approaches may well
lead us nowhere. The profound motivation to follow them is our dissatisfaction
with current approaches to biologic systems, particularly within
immunology, and their failure to solve problems (such as that of
cancer and autoimmune pathology) within classical frameworks of
thought and experimentation. We have provided detailed references
in a previous publication [ I ].
B. OrganizationalClosure in the Normal Immune System
I Internal Activity
It has now become clear that the normal immune system does not
need environmental stimulation to be directed into relatively high
levels of activity. This activity can be measured not only at the
level of production and decay of lymphocytes, but also in the generation
of effector cells. Thus, germ-free mice maintained on low molecular
weight, chemically defined diets, possess in their spleens numbers
of immunoglobulin-secreting cells that are quite similar, if not
superior, to those found in the spleens of conventionally bred and
infected mice. Interestingly, such mice have no plasma cells in
the lym ph nodes, an observation which actually provides a control
for the "antigen-free" state of these mice. The ontogenic development
of "natural plasma cells" in the spleen of these mice is perfectly
comparable to that of mice born from normal mothers and bred in
an infected environment with a normal diet [2]. It would appear,
therefore, that at least within the spleen compartment, the immune
system shows an activity of its own, which leads to the generation
of large numbers of high rate immunoglobulin-secreting cells. Such
an environment-independent activity is also observed in the other
lymphocyte compartments. As we have recently shown, normal mice
contain in their spleens activated effector cells of both the helper
and the suppressor type, in roughly the same numbers as background
effector cells of the B lymphocyte lineage. It is yet to be determined
whether or not such "natural" generation of effector T cells is
truly internal, by studying "antigen-free" mice, but we think it
very likely that this is indeed the case. It is obvious. however,
that even those carefu]]y maintained animals are not antigen free,
as the immune system is in contact with its own antigens and with
those constituting the soma.
II. A Formal Network of Idiotypes
We have therefore turned experimentally to the study of the mutual
interactions among the molecular and cellular components of the
immune system, as well as between the immune system and other molecules
found in the internal environment. To this end, we have isolated
large collections of hybridomas, both of B and T lymphocytes, from
newborn untreated animals, aiming at obtaining representative samples
of the lymphocytes that have been activated in the internal environment.
The analysis of these collections has al ready provided the first
experimental evidence for the existence of a formal idiotypic network
in the developing immune system. Thus, natural antibodies isolated
from a single newborn mouse show an astonishingly high frequency
of mutual reactions, demonstrating a degree of connectivity and
redundancy in the developing immune system which might be expected
on theoretical grounds as an attribute of stable networks, We are
thus far unable to decide on the functional relevance of such idiotypic
interactions. Experiments are at present being carried out, trying
to analyze these aspects.
III. Mutual Influence of the T and B Cell Repertoires
The detailed study of these collections of naturally activated
cells in normal mice has led us to an interesting conclusion: the
specificity repertoire of background-activated cells is unique to
each individual. even if newborn mice from the same litter of an
inbred cross are compared, On the other hand, we can of course expect
that the diversity repertoire of individual mice is submitted to
genetic constraints, that is, limited to the possibilities allowed
by gene families such as immunoglobulins, MHC, and T cell receptors.
Within these potentialities, however, the individuality of immune
systems defined by available repertoires appears to be established
somatically by the connectivity between its cellular elements. Bearing
in mind such organizational closure and internal activity, we have
paid particular attention to the influences that antibody repertoires
might have on T cell repertoires and the reverse, that is, the influences
that T cell repertoire,s of both the helper and cytolytic type (as
we]] as MHC genes) might have on natural antibody repertoires. Three
clear-cut examples already exist, supporting the existence of these
mutual influences which determine the specificity of the internal
activity of the immune system. One such example is already a few
years old. It described the H-2 and immunoglobulin allotype control
of a function of the normal immune system which consists in reproducing
increased circulating levels of a given idiotype upon injection
of nanogram amounts of same idiotype. Since idiotypes that had this
property of "autoreproduction" are consistently found as natural
antibodies in the serum of normal mice, we have concluded that natural
antibody repertoires may well be under the influence of H-2 and,
consequently, are at least in part selected on the basis of T cell
specificities and activities. We have further suggested that among
natural antibodies (i.e., products of cells that were internally
activated), a relatively high frequency of idiotypic profiles resembling
MHC products could be expected on the basis of the predominant anti-H-2
specificities in the T cell compartment. This hypothesis has been
recently confirmed by isolating natural idiotypes which are internal
images of selfMHC antigens. As a mirror image of this type of influence,
we have also found that helper cell repertoires, particularly the
expression of idiotypes on clonally distributed receptors, is controlled
not only by MHC-Iinked genes, but also by immunoglobulin heavychain
genes. We have further shown that the influence of immunoglobulin
genes on such repertoires is indirect and results from internal
com plementarities established between the two repertoires, because
helper cell idiotypic repertoires are profoundly altered in mice
deprived from birth of the antibodylB cell system. We conclude from
all these observations that a normal immune system is characterized
by a high degree of internal activity which results from mutual
specific complementarities between T and B cell repertoires. Effector
cells are induced in the internal environment and themselves regulate
the levels of activity in the normal system and determine, within
the "noise" that surrounds the immune system, what makes sense to
it and can therefore perturb its equilibrium and modify its activity.
C. The Normality of Autoreactivity
A large body of evidence has accumulated over the last few years
indicating the existence of immune reactivities directed to other
components of the immune system itself. Thus, antibodies, helper
cells, and cytolytic T cells have been shown to recognize idiotypic
determinants on other antibodies or on T lymphocytes. Furthermore,
normal autoreactivity of T lymphocytes that appears to be stimulated
by self-I-A under some conditions, has led to a large number of
descriptions of what is called autologous MLR. It appears, therefore,
that autoreactivity is a normal component within the immune system
itself, as one would expect from a complex autonomous system that
is self-organized. For a number of years, quite independently from
these observations, a considerable number of reports have dealt
with the existence in normal individuals of lymphocyte precursors
in the B cell lineage with specificities for determinants expressed
on other proteins of the "self' internal environment. The prevalent
concept, however, is that in the absence of effective helper activity
which is thought to be eliminated (T cell tolerance to self-determinants
is a widely accepted concept), such B cell precursors will not be
induced to antibody formation in the normal immune system. Autoimmunity
has invariably been considered as pathologic and the approaches
to its pathogenesis have been the search for either the abnormal
expression (qualitative or quantitative) of a self-antigen, or the
abnormal occurrence of one or more lymphocyte clones that should
have been "forbidden." More recently, however, considerable evidence
has accumulated for the existence of autoreactive antibodies in
the pool of natural circulating immunoglobulin. In the analysis
or natural antibodies in newborn mice, we have observed that a very
large fraction of these internally induced antibodies show extensive
reactions with selfantigens. Other observations in adult individuals,
both mice and humans, have led Avrameas and his collaborators to
inrer the invariable presence of autoreactive antibodies in the
normal serum or these species. It follows that the presence of autoantibodies
is not correlated with autoimmune pathology, a conclusion that had
already been suggested by some workers in the field of autoimmunity.
It becomes important, therefore, to separate the physiology from
the pathology of autoreactivity, and to evaluate its physiologic
relevance. It also appears to us that the study of the internal
activity of the normal immune system, which is formally more similar
to pathologic situations due to autoreactivity, may be more likely
to lead us to the solution of these problems than the study of immune
responses, developed within systemic strategies and clonal patterns
of lymphocyte behavior which are definitely very different from
those that can be observed in the normal physiology of the immune
system.
Reference
1. Coutinho et al. (1984) Immunol Rev 79:151-168
2. Benner R, personal communication
|