the foundations of the disciplinary matrix -...

2. The Foundationsof the Disciplinary Matrix

2. The foundations of the disciplinary matrix

In the garden of Eden, Adam saw

the animals before he named them;

in the traditional school system,

children name the animals before

seeing them.

— Alfred North Whitehead

2.1. Introduction1

The present chapter aims at the characterisation of our current knowledge and perception

of the immune system, and substantiates the idea that this endeavour can be guided by a

consideration of the historical development of discipline. Inspired by evolutionary

biology, this idea is essentially grounded on other epistemological principles. Any

attempt to understand scientific knowledge is inevitably an (auto)reflexive exercise — the

object at stake is our own immunological knowledge, not the natural immune system.

The main difficulty inherent in this exercise is the maintenance of an appropriate degree of

objectivity: the tendency is to remain closed in our own scientific terms of reference. A

historical retrospective, being intuitively akin to a view from the outside, can potentially

overcome this difficulty.

The next two introductory subsections are methodological, and present both the

conceptual and the empirical tools guiding the historical retrospective presented later. The

affiliation to the ideas of Thomas Kuhn (1970, 1990) is acknowledged from the outset;

they constitute one of the most enlightening and comprehensive insights into the

development of science as a social dynamic process. As this chapter tries to illustrate, this

view can successfully guide the interpretation of progress in Immunology.

1 Part of the content of this chapter has been published in: J. Carneiro (1996). The Burnetian

revolution and the foundations of the disciplinary matrix. Revista da Sociedade Portuguesa de

Immunologia. 2 , 15.

– 12 –


2.1.1. Conceptual Elements for a Kuhnian History of Science.

Kuhn holds that the research in mature sciences is structured by what he called (Kuhn,

1990) the essential tension between tradition and innovation. Thus, progress in physics

or chemistry alternated between periods of cumulative development and episodes of non

cumulative revolutions in which the perception and conception of nature was profoundly

changed. This conception of scientific development involves a coherent set of

interconnected elements that Kuhn developed and supported with several examples in the

essay "The structure of scientific revolutions" (Kuhn, 1970). These elements —

understood as conceptual boxes in which we will try to fit the history of Immunology —

are just outlined here.

Let us start by the notion of ordinary science, that refers to the research grounded

on one or several established examples from the past scientific work — laws, theories,

applications, or experimental techniques. It is the commitment to these examples —

exemplars 2 — that ensures the peculiarity and coherence of a scientific community.

Fig.2 .1- Any trained immunologist will

instantaneously recognise the image on the

left as a thymus. The switch from an

immediate perception of a 2D dot-plot to an

immediate perception of a lymphoid organ,

was carried out during the training period as a

student. Together with this expertise, the

student immunologist acquires the implicit

conviction that a lymphoid organ can be

appropriately described by the relative

distribution of surface markers in its

suspension of cells.

2 In first edition of "The Structure of the Scientific Revolutions", Kuhn used the term paradigm.

However, this term was rather ambiguous in that essay. In a more recent elaboration (Kuhn, 1990), he

coined several terms like exemplar and disciplinary matrix to avoid some of those original ambiguities.

The later terminology has been adopted here.

– 13 –


Exemplars provide the community with otherwise inexistent common rules, and

determine the focus and scope of its research activity (see an illustration in fig.2.1).

Ordinary research is an exhaustive elaboration, consolidation and enlargement of its

exemplars, involving three major categories of problems: the determination of the

significant facts, the agreement of facts and theory, and the reformulation of theory. The

solution for these problems is anticipated under the light of the exemplars, but requires

the resolution of all sorts of experimental, conceptual, or mathematical puzzles.

According to Kuhn, the challenge of solving these puzzles is the major motivation for

individual scientists, and the fascinating aspect of ordinary science; the ability to come up

with a solution according to the common canons (consistent with the exemplars ) is what

determines the recognition of the individual by the community. A good example of these

puzzles in Immunology was the molecular basis for the generation of antibody diversity,

which was raised upon the elaboration of clonal selection theory. Referred to in

contemporary literature by the emblematic abbreviation of GOD (for generator of

diversity), the solution of this puzzle granted Tonegawa the Nobel prize in 1984.

But ordinary research is just one complementary aspect of scientific development;

the other is revolutionary or extraordinary research associated with scientific discoveries

and the emergence of new theories. Scientific discoveries happen when non predicted

phenomena become the object of scientists attention; they are the demonstration that

nature does not behave as expected. The assimilation of these anomalous facts, which are

inconceivable in the previous theoretical or practical planes, entails a rearrangement of

both intellectual and technical equipment; a rearrangement that is crucial for scientific

progress. Classical examples of such innovative discoveries are the ones of oxygen, of

electric current, X rays, or the electron. The discovery of idiotypy by Oudin is frequently

referred as an example such discoveries in Immunology (Brussard, 1979).

Theoretical revolutions are structurally similar to the discovery and assimilation

of an anomaly, but have a wider scope; they require the awareness, most frequently by a

small fraction of the community, that a traditional set of exemplars is no longer useful,

– 14 –


i.e. that ordinary science is no longer able to consistently generate and solve its puzzles.

During revolutionary episodes, the set of exemplars shared by the community changes

progressively, as scientists commit to the revolutionary exemplars. Under the light of the

new exemplars, old problems are re-stated or become irrelevant (discarded as non-

scientific), while new problems are generated. A typical case of theoretical revolution —

the advent of Burnet's clonal selection — will be discussed in some detail later.

In an attempt to understand these scientific revolutions, Kuhn argued that the

commitment to new exemplars is analogous to a gestalt switch — the perception of

nature is completely changed. The acquisition of traditional exemplars by student

scientists is associated with an assimilation of the corresponding perception of nature —

for example, the gestalt switch from the perception of "a graphic" to that of "a thymus"

illustrated in fig.2.1. It follows that competing schools — those committed to different

sets of exemplars — have incommensurable ways of conceiving ordinary science: they

differ intrinsically in their perception of observations and experimental results, they are

concerned with only partially overlapping sets of problems, and very frequently the usage

of common words hides incompatible terminologies. Because of this

incommensurability, the ordinary way of solving problems and testing hypothesis by

congruence with nature — the logic of ordinary science founded on exemplars — is no

longer operational. Scientists from competing schools typically try to destroy each others

arguments, using the internal logic inherent to their own exemplars, and giving rise to

long debates in the literature which are nothing than interminable misunderstandings.

Neither of the parties is able to accept or take for granted all the assumptions that the other

needs to validate its perspective, and therefore they can never fully understand each other.

Silverstein (1989) has discussed a classical case of intensive debate between

incommensurable parties in Immunology: the Bordet-Ehrlich dispute on significance of

the sensibilatrice-ambozeptor found in serum normal individuals. Another more recent

example, is the long lasting dispute on the mechanisms of B lymphocyte activation

between the schools of Möller ('the one signal only model') and Cohn ('the two signal

– 15 –


model'), whose research strategies are sharply opposed as pointed out by Sinclair

(1992).

Falling outside the realm of objective scientific logic, the commitment to one or

another set of concurrent exemplars calls into action alternative subjective and social

values. The gradual commitment of a scientific community to new exemplars and their

ultimate dominance is best understood at the social level. The success of "founding"

exemplars will be determined by the persuasive capacity of its early adherents, and the

renewal of the community. Kuhn (1990) coined the term disciplinary matrix to refer to

that assembly of factors, explicit or implicit, that determine the coherence of a scientific

community and the commitment of its members to a coherent set of exemplars.

Heretofore, the term paradigm refers to intersubjective metaphysical or theoretical

principles that are characteristic of a disciplinary matrix.

Closing a circle we are now prepared to restate "the affiliation to the ideas of Kuhn" (in

the introduction to this chapter) as "the commitment to the Kuhnian exemplar of the

history of science".

2.1.2. Empirical Elements for a Structured History of Immunology.

Some previous historical studies on Immunology helped in the acquisition of a very first

picture of the development of the discipline. The excellent essays by Silverstein (1989)

and Moulin (1991) deserve special mention. With this first picture in mind and guided by

the conceptual framework of Kuhn, we plunged into past literature in Immunology.

In order to identify the exemplars — established experimental or theoretical

works — guiding research in immunology, a typical review of literature was performed,

following citations from present into the past. Revolutionary periods correspond to

discontinuities in this citation tree. Literature around those episodes was then the object of

more exhaustive study under several headings: (i) a survey of contemporary articles and

reviews, paying special attention to argumentation — experimental evidence, theoretical

– 16 –


elaboration, and a priori reasoning — and judgements of value used in inter-school

discussions, as well as the evocation of exemplars from other disciplines; (ii) the names

of new journals or the headings of reviews; (iii) the timing and appearance of new text

books; etc. All these strategies allowed a characterisation of the changes in the sets of

problems and puzzles, as well as the identification of the subjective values involved in

the exemplar-paradigm switch.

For example, in order to understand the gradual commitment of the scientific

community to the "Clonal Selection Theory", the Annual Reviews of Medicine, Annual

Reviews of Microbiology, or Annual Reviews of Biochemistry were followed

systematically in the period ranging between 1930-1960. The most persuasive arguments

used by the early defenders — Burnet, Jerne, Medawar and colleagues, Talmage, and

Lederberg — are identified, by definition, as those that caught the attention of the

reviewers of the field. The metaphors and thought experiments of the authors — for

example the dictionary metaphor used by Burnet in his Nobel prize lecture — are

additional source of information about the values involved in the commitment to the new

exemplars.

2.2. A Preview of History of Immunology.

The Basic Structure

The foundation of Immunology as a scientific discipline is commonly attributed to

Pasteur, who seeded the field with the basic postulate that immunity is a defensive

reaction against pathogenic invasions. This postulate — a particular facet of his germ

theory of putrefaction that resulted in successful immunisation protocols — pervades

immunological thinking ever since. The nineteenth century, following Pasteur, was

essentially characterised by an initial descriptive phase in which multiple immunological

phenomena were identified. Metchnikoff described phagocytosis and cellular aspects of

immunity. Behring and Kitasato discovered the existence of antibodies. At the turn of the

– 17 –


century, Ehrlich became a major figure through the development of an entire theory of

humoral immunity that had a decisive influence for the faith of the discipline.

The first half of the twentieth century was characterised by approaches based on

physical chemistry. Research was essentially concerned with the descriptive analysis of

biochemical compounds intervening in acquired immunity and allergy, and the

characterisation of their reactions. The most popular names were chemists like Arrhenius

(Nobel Laureate in Chemistry, 1903), Heidelberger, Svedeberg (Nobel Laureate, 1926),

Haurowitz, Tisselius (Nobel Laureate in Chemistry, 1948), Pauling (Nobel Laureate in

Chemistry, 1954), or physicians like Landsteiner (Nobel Laureate, 1930) whose work

was nevertheless based on chemical approaches. As discussed later (section 2.3),

immunological thinking reflected the perception that the body of an animal, and

subsequently antibody producing cells, were chemical reactors , where the introduction

of an antigen reagent promoted a reaction that had antibody as product. This disciplinary

matrix is referred to here as Immunochemistry of Acquired Immunity.

Since the fifties the discipline has been overwhelmingly dominated by approaches

based on molecular and cellular biology. Research is typically focused on the analysis of

genes, mechanisms of gene expression, molecular structures, cell signals, and cell

differentiation pathways. All the immunological phenomenology is subordinated and

reduced to these observables. Consecrate figures, without being exhaustive, are Burnet

(Nobel Laureate, 1960), Medawar (Nobel Laureate, 1960), Jerne (Nobel Laureate,

1984), Edelman (Nobel Laureate, 1972), Mitchison, Nossal, Tonegawa (Nobel Laureate,

1986), Cohn, Zinkernagel (Nobel Laureate, 1996), etc. Immunological thinking reflects a

very particular perception of reality: molecules are typically treated as if they would carry

information, they are signals that cells process in order to chose alternative programs of

differentiation which are stored in their genome; immunising antigen is the main signal

processed by lymphocytes which are able to make molecular "self nonself

discrimination". Immunopathologies arise from altered genetic programs, altered signal

transduction, or are caused by antigens. This disciplinary matrix is traditionally referred

to as Immunobiology by opposition to Immunochemistry. However, since alternative

– 18 –


"Immunobiologies" exist that are focused on other levels of biological organisation, the

term is equivocal. Therefore, this disciplinary matrix, that has Burnet's Clonal Selection

Theory by main exemplar, is called here Burnetism.

Since the seventies, several research programs are built on exemplars that imply a

break with Burnetism. Not fully coherent with each other, these programs share many

common aspects. They are motivated by the conviction that Burnetism has accumulated

too many anomalies and is no longer able to solve its problems, and that progress in

Immunology requires a major theoretical development that copes with the progress in

other fields of Biology. They are concerned with the immune system qua system. Self-

organisation, internal structure, emergent behaviour, and ontogenesis are the main issues

of research and discussion. The antigen has lost the key status of the cause of all immune

phenomena; physiology and pathology are perceived as internal dynamic states of the

immune system. The nervous system and ecosystems are used in heuristic analogies, and

theoretical discussion is strongly marked by cognitive sciences. The main figures are

Jerne, Vaz, Coutinho, Kearney, Urbain, Avrameas, Cohen. Outsiders from related fields

of Biology associate to this projects, notably Varela.

The structure and scope of these three ways of doing ordinary science and the

corresponding perceptions of reality are discussed in the following sections. The switch

from Immunochemistry to Burnetism — a true scientific revolution that pervaded the

entire community — was paid special attention.

2.3. Immunochemistry of Acquired Immunity

Immunochemistry literature was essentially analytical and devoid of any attempts to draw

synthetic pictures. Both text books and review articles went through exhaustive lists of

reports on the physical chemical properties of antigens, toxins, haptens, or antibodies.

The present retrospective required criteria for the selection of a few issues, that impinge,

in one way or another, in the transition from the Immunochemist to the Burnetist

disciplinary matrix. The text is closely guided by the writings of the immunochemist

– 19 –


"guardians of the temple" , and is therefore representative of the immunological thinking

and construction of scientific objects in those periods. Fig.2.2 presents a chronology of

the main exemplars in Immunochemistry.

Landsteiner (1912)Immunogenic Azotioproteins

Haurowitz (1967)Could Spring Harbor Simposium on Antibodies

1920

1930

1940

1950

1910

1960

Ehrlich (1900)Side chain theory of antibody formation

Pauling (1940)Template theory

Breinl & Haurowitz (1930)Template theory

Burnet & Fenner (1949) Felton(1949)Immunological Paralysis

Chase(1946)Antigen induced desensitisation

Mudd (1932)Templatetheory

Svedeberg (1939)Sedimentation coeficients andMW of Antibodies

Tiselius & Kabat (1939)Antibodies and globulins

Behring & Kitasato(1890)Serologic reaction

Medawar (1946)Acquired immunity to homograft

Heidelberger 19??)Quantitative studies onprecipitin reaction

A Chronology for the Immunochemistry of Acquired Immunity

Fig.2 .2- A chronology of the main exemplars in the disciplinary matrix of the Immunochemistry

of antibody formation.

– 20 –


2.3.1. The Physical-Chemical Nature of Antibodies

Antibodies were originally characterised by their appearance in the serum of animals after

immunisation and detection by serologic reactions (Behring & Kitasato, 1890). The

immediate concern of the community was the chemical analysis of the seric compounds

carrying antibody functions.

The very first achievements were the establishment that serologic reactions were

specific for the eliciting antigen. In this stage, the work of Landsteiner in the twenties

was most relevant, and his strategy of probing antibody specificity against chemically

modified proteins became one of the main exemplars of Immunochemistry (Heidelberger,

1932; Haurowitz, 1953; Bibel, 1988). Additional characterisation of antibodies was

performed by their functional susceptibility to hydrolytic reactions, enzymatic activity,

heat lability, etc.; treatments that were correlated with candidate biochemical

compositions. Precipitin and flocculation reactions were the object of quantitative studies

with special attention being paid to basic principles of physical chemistry —

stoichiometry, applicability of mass action laws, effects of insolubilisation, etc. The

works, consecrate in the thirties, were those of Arrhenius (1907), Heidelberger &

Kendall (refs. in (Heidelberger, 1932; Heidelberger, 1938)), and Marrack (1934).

The invention of ultracentrifugation (Svedberg, 1939) and electrophoresis

(Tiselius & Kabat, 1939) represented a major break through in the purification of

antibodies, that would remain otherwise based on fractional precipitation recipes. Serum

fractions, with clear sedimentation or migration coefficients, could be subjected to finer

biochemical analysis and antibody functional assays. Hence, Svedeberg's group

established the sedimentation coefficients of antibodies, and reported the first estimations

of their molecular weight. Tiselius & Kabat (1939) separated electrophoretically the

antibodies from other compounds reporting that they migrated between the β- and γ-

globulin fractions, or in the later fraction.

– 21 –


Overall, by the late thirties, compelling evidence had accumulated for the protein

nature of antibodies, and refractory claims on "protein free fractions" were regarded

with suspicion, recalling that the assays for detecting antibody function were much more

sensitive that those for detection of proteins (Heidelberger, 1938; Bordet, 1939). The

main concern shifted progressively from the nature of antibody to its origin and

formation, that was still terra incognita.

2.3.2. The chemistry of antibody formation.

Antibody as modified serum globulin.

In the very first volume of the Annual Reviews in Biochemistry from 1932,

Heidelberger stated that "most workers in this field [Immunochemistry] now hold the

view that antibodies are modified serum globulins". It might be useful, to trace further

back the origins of this view, that pervaded the literature until late fifties.

The first widely accepted theory for antibody formation had been Ehrlich's side

chain theory (Ehrlich, 1900), that held that antibodies were preformed nutrient receptors

in the cell surface that would be secreted into the blood by "over compensation" of their

consumption by specific reaction with antigen. Ehrlich's theory was discredited by

Landsteiner's demonstration of the antigenicity of diverse modified proteins, since it had

become "difficult to assume that the animal body should have preformed receptors for all

these artefacts of the chemical laboratory" (Haurowitz, 1950). A new view prevailed,

that subordinated the specificity and diversity of antibodies to the eliciting antigen. Thus,

antigen itself was the cause of the specificity antibody molecule. Essentially two

concurrent hypothesis existed to explain this specific modification. The first, proposed by

Buchner (ref. in Heidelberger (1938)), held that antigen itself or antigen products would

incorporate the antibody. Evidence — the comparative analysis of antigen and antibody

compositions, and the disproportion of antibodies elicited by antigen — was in the

majority against this hypothesis (Heidelberger, 1938; Bordet, 1939). The theory that

– 22 –


finally gained the consensus of the immunochemists was the "more reasonable" proposal

by Breinl & Haurowitz (1930) and later by Mudd (1932), that held that:

The synthesis of normal globulins may occur in such a way that the spatial

configuration of the cellular protoplasm impresses upon the globulin the

spatial3, chemical, and species-specific properties characteristic for the

animal in question. Penetration of injected antigen or its partial

degradation products to the sites of synthesis could disturb the spatial

relations which normally exist and distort them. This distortion might

reasonably occur in a manner characteristic of the foreign material, so that

any new globulin synthesised in this distorted manner would bear a certain

spatial relation to the antigen. Hence if the new globulin molecules would

again encounter the antigen in the circulation, or in vitro, interaction might

be possible.

—Heidelberger (1938)

Note that, although this template theory had no direct experimental support, at the

time it provided a reasonable mechanism for antibody formation, it accounted for the

existence of normal circulating globulins and its physico-chemical similarity with

antibody, and postulated that the specificity in the reaction with antigens was due to a

precise re-ordering or re-arrangement of the aminoacid residues that rendered the two

molecules complementary. The elucidation of the obscure aspects of the theory

represented all sorts of seductive Kuhnian puzzles — in both theoretical and

experimental grounds — that the community of immunochemists would try to solve in

the following years.

The most popular, and for a long time regarded as the most sound refinement of

the template theory, was due to Pauling (1940). Based on his authoritative knowledge on

bonding and molecular structure — Pauling introduced quantum mechanics into

chemistry and later discovered the alpha helix structure of proteins — he proposed a

mechanism of antibody formation (fig.2.3) that as he emphasised could account for a

variety of empirical observations such as "the heterogeneity of immune sera", "the

3 In this and in the following quotations, the original emphasis by the authors is indicated by non

italic characters, while our emphasis is indicated by underlined characters.

– 23 –


bivalence of antibodies and multivalency of antigens", "the antibody-antigen molecular

ratio in precipitates", and "antigenic power". In Pauling's theory there was no particular

reference to cells.

Fig.2 .3- "Antibodies differ from normal serum globulin only in the way in which the

two end parts of the globulin polypeptide chain are coiled, these parts, as a result of their

amino-acid composition and order, having accessible a very great many configurations with

nearly the same stability; under the influence of the antigen molecule they assume

configurations complementary to surface regions of the antigen." — The figure and

the legend are from Pauling (1940) .

Concerning the connection of theory and experimental observations the

polypeptide chains and the conformational adaptation of antibodies and normal serum

globulins were under the spot light. The aminoacid composition of antibody and normal

globulin was identical, and they were serologically indistinguishable (Raffel, 1954).

Although sequencing of an entire antibody molecule was not feasible, the report by Porter

(1950) on the sequence identity of segments of both normal globulins and antibody, was

in perfect agreement with the expectation that they were the same polypeptide chain

(Haurowitz & Crampton, 1950). Attempts were made to mimic in vitro the formation of

antibody under antigen template, by denaturing and renaturing globulins or antibodies in

the presence of antigens. The results were suggestive but not conclusive. Typically, an

antibody would lose reactivity to eliciting antigen and acquire reactivity to other antigens,

but without specificity.

– 24 –


The long persistence of immunity became a major line of research. Hence, the

first report on the half-life of antibody molecules in the serum by Schoenheimer et al.

(1942), who estimated its value as 2 weeks, required a continuous supply of fresh

antibodies. According to the consensus template theories, long lasting immunity was

interpreted by the persistence of antigen molecules in antibody-producing cells. The

hypothesis was exhaustively investigated using labelled or coloured antigens —

polysaccharide antigens were shown to persist for long periods (Heidelberger et al.,

1950), while protein antigens were more rapidly eliminated — the results remaining

controversial and puzzling (Haurowitz, 1953).

From this controversy emerged an alternative conception of antibody production,

that started to shift the focus of research from molecules to cells: antigenic stimulus

would lead to a permanent change in globulin-producing cells (Dixon et al., 1952;

Haurowitz, 1953; Raffel, 1956). The "adaptive enzyme system" of Burnet & Fenner

(1949), that we will discuss in some detail later, was included in this general class of

theories by the literature of that time (Haurowitz, 1953). These hypothesis were

investigated, namely by transfer of antibody producing tissues from immunised to

unmanipulated animals, but again, no definitive conclusion could be drawn (Haurowitz,

1953). The most consistent finding was that the second animal would only produce

antibodies if the eliciting antigen was a live microorganism, that could very well be

transferred with the tissues.

2.3.3. Immunology as Immunochemistry

The immunochemist conceptions on the nature and the formation of antibodies, that were

briefly overviewed, were not only the interpretation for acquired immunity, but

influenced the understanding of all other immunologic phenomena, and therefore the

focus of research. The scope and significance of this reductionist subordination of

immunology to chemistry, is illustrated here by the immunochemical account of

biological issues like the role of cells in acquired immunity, immunopathology,

– 25 –


immunologic paralysis, and the transplantation problem. These issues emerged in the

literature during the late stages of the immunochemistry period, and some contributed to a

switch in the disciplinary matrix, and the advent of what is referred to here as Burnetism.

CELLULAR ASPECTS OF IMMUNITY

Around 1950, the significance of cells was restricted their active role in local

inflammation, namely by extracellular and intracellular degradation of microbial

pathogens and immunising material, and to their passive role as a reactor for globulin-

antibody synthesis4.

The early notion, due to Ehrlich's side chain theory, that antibody formation

occurred in any cell in the body had been essentially abandoned, and antibody synthesis

was believed to be restricted to "reticular tissues" (Taliaferro, 1949) . The importance of

the spleen was demonstrated already around 1930, by the impact of splenectomy on

acquired immunity (refs. in Taliaferro (1949)). In the fourties, Ehrich & Harris (refs in

(Taliaferro, 1949; Raffel, 1954) established that antibodies were also produced locally in

the lymph nodes.

The finer analysis of the cell type producing antibody gave rise to several non

mutually exclusive hypotheses, implicating phagocytes (based on the "reasonable"

assumption that the same cell that degraded particulate antigen should produce

antibodies), lymphocytes (Raffel, 1954), plasma cells (Fagraeus, 1948), and appropriate

combinations (Taliaferro, 1949; Raffel, 1954). The issue, however, could not be clarified

by the available technology — histology based on cell morphology; transfers of cells and

assays for hypersensitivity or acquired immunity; correlation between histology and

serum antibodies; etc.— and the multicellular hypothesis remained quite popular (Raffel,

1954).

4 Note that globulin- antibody producing cells were conceived as a chemical reactor that produced

normal globulins in the absence of antigen, and specific antibody in the presence of antigen or upon an

antigenic inprint.

– 26 –


IMMUNOPATHOLOGY

One of the unique characteristics of the immunochemistry period was the

conviction that antibodies specific for body constituent molecules (autoantibodies) could

not exist. The dogma was known as Ehrlich's "horror autotoxicus", although, as pointed

out by Silverstein (1989), the original formulation by Ehrlich was not synonymous of an

absolute inability to produce autoantibodies, but expressed the notion that they were

somehow prevented from reacting with autologous antigens. The truth of the matter is

that autoantibodies were not easily conceivable under "template theories". At the time, it

was believed that in the absence of immunisation, antibody-globulin polipeptide chain

would fold into species-specific normal globulin, not into antibody. The molecular

composition of the body was what "catalised" the folding of the polipeptide chain into

normal globulins; the immunisation required that some other molecules would be

brought into the body or the cell in order to "catalyse" the folding of the polipetide chain

into a different product: the antibody molecule.

A growing interest of the community for pathologies involving antibodies —

allergic hypersensitivity and anaphylaxis — would lead to a pragmatic reconsideration of

the horror autotoxicus dogma. The 1950 review by Grabar (1950) was particularly

significant in this context:

It seems that sufficient indications exist to warrant further investigations into

the possibility of the existence of auto-antibody. It deserves much more

attention, as it would explain numerous pathological states.

—Grabar (1950)

In the fifties, the possible involvement of autoantibody and autoimmunisation was

investigated in many diseases, such as disseminated lupus, rheumatoid arthritis,

thrombopenic purpura, with unknown ethiology but displaying "immunologic aspects"

(Cooke et al., 1955). Animal models of autoimmunisation were also generated, following

the discovery of adjuvants by Freund and colleagues (reviewed in Freund (1947)), that

were used to elicit autoantibodies by injection of autologous, homologous or

heterologous tissue homogenates, extracts, or proteins. The final result of these research

– 27 –


programs is rather well known: in a 1958 review on "Autoimmunity in disease" (!),

Dixon (1958) concluded that horror autotoxicus was definitively "buried" .

The speculative accounts for the pathogenesis of autoimmunity in the literature of

that time are worth closer examination since they reveal the deep influence of

immunochemistry exemplars. Indeed, in a clear analogy with Landsteiner's chemically

modified proteins, autoantibodies were believed to be elicited by some modification of

body proteins into new "foreign" proteins (Grabar, 1950). The main examples under

discussion were therefore drug induced "autoimmunologic disorders", as can be

illustrated by the following account for the pathogenesis of thrombopenic purpura:

Ackroyd (...) subjected a patient with thrombopenic purpura attributable to

Sedormid (...) to exhaustive and critical analysis. (...) From these findings

the author surmises that Sedormid-platelet combination acts as antigen to

produce anti-platelet auto-antibody.

—Cooke et al. (1955)

Some authors, however, were aware that immunochemistry exemplars were failing to

account for autoimmune pathologies. For example, the hypothesis that lupus lesions were

due to allergic sensitisation was regarded by Klemperer et al. (cited in Cooke et

al.(1955)) with scepticism, arguing that it meant "only to explain an obscure structural

alteration by an equally obscure pathogenic mechanism".

IMMUNOLOGIC PARALYSIS

Some instances of induced immunologic unresponsiveness were known from the

beginning of the century (refs. in Chase (1959)), but they were never given special

attention. In the fourties, the issue regained interest following several reports on

unexpected decreases of immunologic response. Chase (1946) reported the inhibition of

experimental drug allergy and anaphylaxis by prior feeding with the sensitising agent.

Felton (Felton & Ottinger, 1942; Felton, 1949) reported on the phenomena of

"immunologic paralysis": mice injected with excessive amounts of pneumococcal

polysaccharide remained specifically unresponsive to immunising doses of this antigen.

– 28 –


Felton's work became one of the main immunochemical exemplars in the fifties (Raffel,

1956; Chase, 1959).

The candidate explanations for "immunologic paralysis" were that excessive

amounts of antigen persisting in the antibody-producing cell inhibited the reaction of

antibody synthesis, or neutralised all the antibody being synthesised (Raffel, 1954).

These hypotheses were founded and consistent with the demonstration by Felton that

polysaccharide remained in cells for long periods. They were, nevertheless,

controversial since, as already mentioned, the persistence of the antigen was evoked by

Heidelberger to explain the long persistence of polysaccharide specific antibodies in the

serum.

THE HOMOTRANSPLANTATION PROBLEM

The emergence of the transplantation problem in the post-war period was more

the consequence of contemporary social context, rather than the result of an elaboration of

immunochemistry exemplars. Medawar, who contributed notably to the field, was

personally motivated by his frustrating attempts to treat patients with extensive burns

during the war (Bibel, 1988). Under the cold-war environment, the medical

consequences of exposure to radiation were being investigated (Byars & Randall, 1953).

Extensive skin lesions and destruction of hematopoietic tissues were anticipated problems

that could potentially be successfully overcome by transplantation (Byars & Randall,

1953). The progress of plastic surgery was such that tissue autografting no longer

represented a problem; the transplantation of skin from the same species — then called

homotransplantation — remained, however, "completely unsatisfactory" (Cooke et al.,

1955).

The reappraisal of homotransplantation as an immunological problem is mainly

due to the work of Medawar, notably upon the demonstration that skin graft survival time

was inversely proportional to the quantity of tissue, and that secondary grafts were

rejected faster than primary grafts, that suggested a phenomena of acquired immunity

(Medawar, 1946a; Medawar, 1946b; and refs. therein).

– 29 –


Although Little and his successors Snell and Görer (reviewed by Medawar

(1958)) had already established the genetic basis for the transplantation of tumours, the

solution envisaged for the transplantation problem was oriented to the possibility of

inducing immunologic paralysis. Hence, the issues of transplantation tolerance induction

and immunologic paralysis were always associated and discussed as similar phenomena

in the reviews of the field (Raffel, 1956; Chase, 1959).

2.4. Burnetism

The emergence of Burnetism, between 1949 and 1967, was a critical period in the

history of Immunology, and is the root of its currently dominant disciplinary matrix. As

in other fields of Biology, the focus of attention switched from the chemistry of

molecules to the genetics and molecular biology of cells. A chronology of the main

exemplars of Burnetism and some events directly or indirectly involved in its

consecration are presented in fig. 2.4.

2.4.1. A Change in Immunological Concerns.

The Transplantation Tolerance Trojan Horse

Burnet was a true biologist who felt that the preponderance of immunochemistry and its

enlargement to all questions of biology was illegitimately leading to a "pseudo-

knowledge" (Burnet, 1969; Bibel, 1988; Moulin, 1991). He asserted these positions in

both the 1941 and 1949 versions of "The production of antibodies", where he restated the

problem of antibody production in a "general biological background" (Burnet, 1941;

Burnet & Fenner, 1949).

In 1941 Burnet had been unable to immunise chicken embryos against influenza

virus, and related this finding to the classical epidemiological observations of high

infantile mortality by infectious disease and with the observation that infants after

exposure to tuberculosis or vaccination remain negative for the tuberculin reaction.

– 30 –


Burnet & Fenner (1949)Production of antibodies

Jerne (1955) The natural selection theory of antibody formation

Talmage (1957) Allergy and Immunology

Burnet (1957)

A modification of Jerne's theory ... clonal selection.

Billingham, Briant, & Medawar (1953) 'Actively acquired tolerance' to foreign cells

Burnet (1958)The clonal selection theory of acquired Immunity

The Nobel Prize, 1960 , awarded to Burnet and Medawar

1950

1955

1960

1965

Watson & Crick (1953)Genetical implications of the structure of DNA

Could Spring Harbor Simposium on Antibodies is opened by Burnet"The clonal selection theory is accepted withoit reservations by the community"

Burnet (1969)Cellular Immunology

Crick (1958)The biological replication of macromolecules

Lederberg (1959)Genes and antibodies

Burks, Goldstine, &Von Neuman (1946)... logic design of an electronic computing instrument

1945

The first International Phage Symposium

Prague Conference on Antibody"Signs of rapidly increasing interest in the clonal selection ideas"

Cronology for the Emergence of Burnetism

Fig.2 .4- A chronology of the founding exemplars within the Burnetian disciplinary matrix. The arrows

indicate direct influences.

– 31 –


If in contemporary literature these observations were rather vaguely interpreted as the

result of a "transient immaturity" of antibody production, Burnet regarded them as

problem requiring a scientific explanation. His genius lay in postulating a theory for

antibody production, that tied together the inability to immunise animals in the early

stages of ontogenesis, and the widely accepted non antigenicity of body components.

Evoking the concept of adaptive enzyme — at that time the interpretation of

enzyme induction by a substrate (see for example Monod (1947)) — Burnet & Fenner

(1949) postulated that during embryogenesis, or in an immediate post-embryonic phase,

the first enzymatic units involved in the production of antibodies would adapt to one or

another of the molecular "self-markers" , becoming thereafter involved in the disposal of

effete and damaged body components. Following the introduction of foreign antigen into

cells carrying these "self-replicating enzymatic systems" , the units would adapt to adsorb

the antigen and the enzymatic system would multiply. Circulating antibodies were

understood as "partial replicas of the intracellular enzymatic system", that would lose

replicating and enzymatic potential, but not the peculiar adsorptive properties. In order to

account for the long persistence of acquired immunity, the hypothesis required that

"when cells carrying the adaptative enzyme multiply, the descendant cells will also carry

the new character".

It is important to note, once again, that the "modified-enzyme" mechanism was

already cellular (although not yet clonal), like many other contemporary "template

theories". At variance with those, however, it provided a coherent synthesis of

previously unrelated observations and principles in immunochemistry, embryology,

physiology, and epidemiology. More than its mechanistic details, this general framework

caught the attention of immunologists. The following prediction of the theory was

particularly consequent.

If in embryonic life expendable cells from a genetically distinct race are

implanted and established, no antibody response should develop against the

foreign cell antigen when the animal takes on independent existence.

—Burnet & Fenner (1949)

– 32 –


Indeed, the group of Medawar turned towards the homotransplantation problem with this

new insight. In an article that became famous, Billingham et al. (1953) reported the

induction of homotransplantation tolerance by intraembryonic injection of cells, and

discussed this finding as an experimental confirmation of the major prediction of Burnet

& Fenner (1949). "The electrifying success of Medawar's group in establishing tolerance

to skin homotransplantation" (Chase, 1959), and its intelligibility under Burnet's

proposal synergised to persuade the community (Raffel, 1956). As matter of fact, Burnet

and Medawar shared the Nobel Prize in Physiology and Medecine 1960 for having

"unveiled a fundamental law governing the development and maintenance of" immunity

"by penetrating analysis of existing data and brilliant deduction, and by painstaking

experimental research" (Gard, 1960).

The general conception of immunity proposed by Burnet became increasingly

popular after 1953, and consequently any convincing mechanism for antibody formation

had to accounted for its main axiom:

The failure of antibody production against autologous cells demands the

postulation of an active ability of the reticulo-endothelial cells to recognise

"self" pattern from "not-self" pattern in organic material taken into their

substance. The first requirement of an adequate theory of antibody

production is to account for this differentiation of function ...

—Burnet & Fenner (1949)

Anyone can recognise this axiom as one of the main exemplars of current immunology,

that acquired the reputation of the science of "self non-self discrimination" (Klein, 1982).

Note, however, that this axiom is both historically and conceptually distinct from the

principles of clonal selection, that are addressed in next sections.

2.4.2 Towards Selective Antibody Formation

Burnet himself (1967) acknowledges Jerne as the "only begetter" for the origin and

development of the clonal selection principles. In 1955 Jerne (1955) proposed a

revolutionary solution to the problem of antibody formation, based on antigen selected

– 33 –


amplification of circulating antibody variants; this mechanism not only satisfied the axiom

postulated by Burnet, but "made the self-marker ideas (...) look clumsy and artificial"

(Burnet, 1967).

The core of the "natural selection theory of antibody formation" as formulated by

Jerne himself:

The role of the antigen is neither that of a template nor that of an enzyme

modifier. The antigen is solely a selective carrier of spontaneously

circulating antibody to a system of cells which can reproduce antibody.

Globulin molecules are continuously being synthesised in an enormous

variety of different configurations.(...) The introduction of an antigen into

the blood or into the lymph leads to the selective attachment to the antigen

surface of those globulin molecules which happen to have a complementary

configuration.

—Jerne (1955)

The natural selection theory was so revolutionary and far-reaching in its consequences

that its origins and nature are worth closer examination.

Jerne joined Delbruck's phage group in 1955, where he studied antibodies

specific for bacteriophages, apparently because "immunology was not then an 'in'

subject" (Jerne, 1966). It was in this special environment that Jerne's ideas flourished.

His observations that avidity of antibodies seemed to increase during the course of an

immunisation (both in the case of antitoxin or anti-T4 antibodies) had "Darwinian

overtones", and encouraged the elaboration of the natural selection theory of antibody

formation.

Jerne (1966) recalled that during many preliminary discussions with his peers, the

theory had great difficulties in being accepted. For example, Pauling, with whom

Delbrück arranged a meeting, understood the selective principles being exposed, and

rejected them "probably within five seconds". Watson rejected the mechanism of

antibody reproduction by a succinct and unambiguous: "it stinks!". Despite all these

"blows" the article was written and published. The very same difficulties might be

reflected in absence of any mention to the article in any Annual Reviews until 1957.

– 34 –


For different reasons, both traditional immunochemists and the founding fathers

of molecular biology found unacceptable features in Jerne's theory; as a whole, however,

the proposal of applying Darwinian principles to immunity had something intriguingly

seductive and revolutionary. Under the dominant "template theories" the external antigen

was the source of antibody complementarity and specificity. Antibody diversity was

therefore derived from antigenic diversity, and acquired immunity required no genetic

mechanism. In Jerne's proposal, the diversity of antibody variants preceded specificity;

specificity arose from the selection of antibodies variants that were already

complementary for the antigen. The generation of antibody variants and their selection by

antigen, was in the same conceptual framework of the post-mutational selection of

bacterial variants established by Luria and Delbrück. Jerne was seeding Immunology

with the founding principles of molecular biology, that would be consolidated after the

establishment of the central dogma , and pervaded the whole of biology.

2.4.3 The Clonal Selection Theory of Acquired Immunity

Talmage (1957) compared the antibody formation model of Burnet & Fenner with that of

Jerne, arguing that the natural selection theory gives a "simpler and more definitive

explanation for the absence of auto-antibodies". Talmage also pointed out the logical

similarity between Jerne's ideas and the side chain theory of Ehrlich, and suggested that it

would be more appropriate if the "multiplying unit in the antibody response is the cell

itself". It might be relevant to recall here that Talmage belonged to the cellular school of

"template theories" (Dixon et al., 1952).

The appropriate refinement of the natural selection theory would be carried by

Burnet, who placed the selective principles on a cellular and clonal basis within his

general scheme of immunity. Burnet (1957) published the outline of this clonal selection

hypothesis in the Australian Journal of Science. Apparently this obscure journal was

chosen to ensure priority, but to minimise the impact on Burnet's reputation in case of a

failure of the revolutionary selection principles. The fully developed hypothesis was only

– 35 –


presented next year, in the Abraham Flexner Lectures in Vanderbilt University (Burnet,

1958). The major postulate of the clonal selection hypothesis was that:

Each type of [gamma globulin molecule] is the product of a clone of

mesenchymal cells, and it is the essence of the hypothesis that each cell

automatically has available on its surface representatives reactive sites

equivalent to those of the globulin they produce.

—Burnet (1957)

Antigen entering in blood or tissue fluids, would attach to the surface of lymphocytes

carrying complementary reactive sites, and activate them to settle in appropriate tissues,

and undergo local proliferation. Their progeny should include forms capable of

spontaneously producing soluble antibodies, and forms maintaining functions identical to

their parental cells. The net result should be both a change in the serum population of

gamma globulin molecules such that those reacting with antigen should be in excess (i.e.

antibody response), and an increase in the number of circulating lymphocytes of the

expanded clones that could ensure a rapid and extensive response to a subsequent entry

of the same antigen (i.e. a secondary immune response).

It is worth emphasising that the original clonal selection theory was placed in the

realm of the ontogenesis of the individual animal. Both the generation of antibody

diversity and the elimination of lymphocytes reacting with body determinants was

postulated to take place, once and for all, during embryogenesis; the actual repertoire of

lymphocyte clones was therefore conceived as settled for the life time of the animal.

The theory requires at some stage in early embryonic development a genetic

process for (...) 'randomisation' of the coding responsible for part of the

specification of gamma globulin molecules (...). At this stage, (...) any

clones of cells which carry reactive sites corresponding to body determinants

will be eliminated.

—Burnet (1957)

The outline of the clonal selection theory is illustrated in fig. 2.5.

– 36 –


... ...... ...

G.O.D.

1 2 3 4 111 112 623 1245 n

1 2 3 4 111 112 623 1245 n... ...... ...

Ag A

Ag B

Stem Cell

Antigen Independent D

ifferentiation

Ear

ly S

tage

s of

Ont

ogen

y (B

urne

t)

111 111 111 111 1245 1245 1245 1245

Ig 1245Ig 111

111 111 1245 1245

Antigen D

ependent Differentiation

Fig.2 .5- Illustration of Clonal Selection Theory. In the original proposal by Burnet, both the

generation of diversity (GOD) and the elimination of cells recognising autologous antigens (A)

were assumed to occur in the early stages of the ontogeny of the animal. Lederberg proposed a

switch from the standpoint of the individual organism to the standpoint of the individual

lymphocytes.

– 37 –


2.4.4. The Early Elaboration of Clonal Selection

Lederberg (1959), who had been a guest in Burnet's laboratory two years before,

published a formulation of the clonal selection hypothesis in 9 postulates. The aim was to

give it a sound foundation on the "genetic doctrines developed in studies of microbial

populations". One important new feature was the switch from the realm of ontogenesis of

the animal as a whole, to the realm of the differentiation of individual antibody-forming

cells. For example, the absence of antibodies against body components, previously

interpreted as resulting from the elimination of corresponding cells during

embryogenesis, was now interpreted as a lifelong process (see also fig.2.5):

A6. The immature antibody-forming cell is hypersensitive to an antigen-

antibody combination: it will be suppressed if it encounters the homologous

antigen at this time.

—Lederberg (1959)

Note that this subordination of the biology of the animal to the biology of individual cells,

although not due to Burnet himself, represents one of the major reductionist features of

what is called here Burnetism. Lederberg clearly identified as the most questionable

proposal his hypothesis that genetic diversity was due to hypermutability by random

assembly of the DNA of globulin gene during certain stages of cellular proliferation.

From 1957 to 1967 the clonal selection theory was, directly or indirectly,

consolidated by several lines of research (Burnet, 1967).

Edelman (1959) and Porter (1959) characterised the basic structure of antibodies.

Edelman & Gally (1962) demonstrated that Bence-Jones proteins were light chains

chemically identical to those in myeloma globulins from the same patients. Following this

lead, Hilschman & Craig (1965) "electrified the immunological community" with the

report that Bence-Jones proteins, and thus antibody light chains, had a diverse N-terminal

and an invariant C-terminus. Specificity and antibody folding was ascribed to aminoacid

sequence by several methods of denaturing and reconstituting antibodies (reviewed in

Freedman & Sela (1966)), supporting the first postulate of Lederberg (1959).

– 38 –


The specificity of antibodies produced by single cells generated considerable

debate and controversy as can be seen by the inconclusive review by Nossal & Mäkëla

(1962). The clarification of the issue required studies on single cells, using several

techniques like immunofluorescence, autoradiography, bacterial immobilisation,

agglutination, and chaining, bacteriophage neutralisation test. When closing the 1967

Symposium, following the report by Makëla (1967), Jerne (1967) concluded that the one

cell—one antibody rule applied to heavy and light chains and probably also to the

antibody combining site.

As often happens in research, before the one cell-one antibody controversy was

solved, the studies on single cells gave fundamental information on the nature of antibody

forming cells, on stages of lymphocyte differentiation, on the proportion of antibody

producing cells, etc. Plasma cells were definitively identified as the antibody producers.

A non negligible role would be played from then on by the haemolytic plaque forming

cell assay (Jerne & Nordin, 1963). The development of in vitro cell cultures, in vivo

isotope labelling , transfer of cells into irradiated recipients, and mitotic inhibitors allowed

the establishment of the pathway precursor lymphocyte - (antigen) target lymphocyte -

proliferating lymphocyte - plasma cell or memory lymphocyte. (see Papermaster, 1967;

and refs. therein).

The genetic basis for the diversity of antibodies was a matter of speculative

debate, and the concurrent hypotheses ranged from the hypermutable genes and point

mutations during evolution of multiple genes, to translational control of antibody

synthesis, errors introduced in DNA repair, or somatic recombination of multiple

antibody genes (for the main references at that time see (Smithies, 1967); for a general

understanding of the state of affairs in this period see Crick (1967)). The solution for the

generator of diversity, a key feature of clonal selection theory that granted it the

abbreviation of GOD, became progressively one of the major puzzles of Burnetism.

The general scheme of differentiation dependent antigen responsiveness proposed

by Lederberg (1959), was taken by Mitchison (1964) to interpret low and high zone

induction of immunologic paralysis. The classical immunochemical explanation of

– 39 –


paralysis by the accumulation of antigen in the responding cell was briefly discussed and

refuted; Felton's paralysis — that had been, as pointed out before, a major exemplar of

immunochemistry — was mentioned but discarded as probably no longer relevant. In the

new disciplinary matrix, the work by Mitchison became the new exemplar for antigen

mediated tolerance (see for example Nossal et al. (1967)).

In this period, there was a quite popular hypothesis on antibody formation, that

attributed a fundamental role to macrophage (Fishman & Adler, 1967; Gottlieb et al.

1967). It held that the macrophage would phagocyte the antigen and transfer specific

RNA to plasma cells, that would use it to produce the antibody proteins. Respecting the

central dogma, this hypothesis retained nevertheless some classical features, notably the

understanding of the antibody producing cell as a mere chemical reactor.

2.4.5. Fashionable Metaphors and Sources of Inspiration.

Information, IO-machines, and computers.

From 1955 to 1967 there is a considerable change in the language of immunologists, that

reveals much of the contemporary scientific fashion. Notions from cybernetics,

information theory, and computer science started to be commonly used, and inevitably

inspired immunological thinking. The influence of information theory was not peculiar to

Immunology; it represents one of the major tenets of molecular biology. The central

dogma (Crick, 1970) postulated that information — sequential code — passed

unidirectionally from nucleic acids to proteins.

Lederberg (1959) refers to the classical "template theories" as the "instructionist

theories" explaining that "antigen conveyed the instructions" for the synthesis of

complementary antibody, in a clear violation of the central dogma. The selective theory is

contrasted by the fact that "the information required to synthesise a given antibody is

already inherent in the organism". The new language and ideas are also evident in Jerne's

"Immunological speculations" from 1960, that start with the following citations:

– 40 –


I shall always regard the differentiation between self and not-self as crucial

to all immunological theory. F.M.Burnet

Or, aucune machine n'est capable de choix, sauf selon des crìteres précis

prédéterminés et inscrits dans son programme. E.Delavenay

—Jerne (1960)

Also Burnet, in his Nobel Lecture, discussed the essence of clonal selection theory under

the heading of "Immunological Information" . Following the then "popular convention of

discussing coding" , he illustrates the mechanism of "differentiation between self and not-

self" by a dictionary metaphor (inspired by Jerne's "Immunological speculations") in

which a computer plays the role of the generator of diversity.

One can easily recognise here the foundations of the metaphoric conceptualisation

of cells as IO-machines with a DNA-encoded program of cell differentiation and gene

expression, that process information inherent to molecular signals. Note that these new

metaphors imply a profound revolution in the perception of reality. Haurowitz (1967),

unable to adopt this perception of reality, was still comparing template theories and

selective theories in the classical terms of physico-chemical reactions.

2.4.6 Consecration of Burnetism

In the 1967 Cold Spring Harbour Symposium on "Antibodies" the majority of the

participants were committed to the clonal selection principle. Burnet (1967) and Jerne

(1967) made respectively the opening and closing sessions, in a clear recognition of their

contribution to the recent face-lifting of Immunology. Only Haurowitz (1967), in a

typical Kuhnian resistance, tried to convince the audience that most of the experimental

observations supporting clonal selection, could still be accommodated by a refinement of

his classical "template theory" of antibody formation.

Confirming another aspect of the Kuhnian views on scientific activity, in 1969

Burnet published the first text book providing a general interpretation of immune

phenomenology under the clonal selection principles — Cellular Immunology (Burnet,

– 41 –


1969). The following extract of the Preface of the book summarises clearly the reasons

for the dominance of Burnetism over the classical immunochemistry.

It was a point of view that seemed to be in accord with the way biology was

developing, and even in a rather crude early form provided a better frame

for the facts of immunity than any variant of the dominant instructive theory.

It was not enthusiastically received and has been 'disproved' on several

occasions. But it is the only theoretical approach to immunology which has

been able to adapt itself successfully to the accelerating flood of new

techniques and new facts coming from experimental immunology.

—Burnet (1969)

In 1970 a new journal was created — Cellular Immunology — to account for "the

remarkable growth in the scope and volume of significant contributions to the cellular

aspects of the immune response" (Lawrence, 1970).

2.4.7 Immunology as Burnetism

The core of current perceptions of immunity — genetic generation of diversity, antigen

driven immune responses, self nonself discrimination by induction of cellular-tolerance to

body components — was already fully developed during the one and a half decades that

were just reviewed. Proceeding with an examination of how this basic set of exemplars

of Burnetism has been elaborated and enlarged in the last 30 years is clearly outside the

scope and limits of this historical retrospective. Some discoveries entailed some structural

readjustments of the basic set of exemplars — the establishment of the two independent

B- and T- lymphocyte systems, and their co-operation for antibody formation, the

discovery of MHC restriction, the characterisation of somatic mutation and affinity

maturation. Other sets of observations, discoveries in stricto senso, were not truly

accommodated by elaboration of Burnetism and remain anomalous. Dominant or

infectious tolerance, idiotypic interactions, and natural autoreactivity, are the basic

– 42 –


examples of such anomalies. The scientific innovations entailed by these anomalies are

the essence of Jerneism, and are discussed in section 2.5.

An historical outline of the accounts for the mechanism of tolerance to body

antigens and its relation to autoimmune pathologies, illustrates how the refinement of

Burnetism proceded without major changes of its key postulates. The issue is

particularly relevant in the context of the contrast between Burnetism and Jerneism.

CELLULAR-TOLERANCE AND AUTOIMMUNE PATHOLOGIES

The intriguing puzzle represented by the mechanism of tolerance to body

components, changed progressively from neonatal deletion of "forbidden clones" in the

original proposal of Burnet in 1957, to deletion in early phases of lymphocyte

differentiation as suggested by Lederberg, to Mitchison's "low zone tolerance" in the

seventies, to central deletion in the eighties (Kappler et. al., 1987; MacDonald et al.,

1988; von Boehmer et al., 1989), which has been "complemented" with several hill

defined mechanisms of peripheral inactivation or deletion of clones, like clonal abortion

(Nossal & Pike, 1975), clonal anergy (Pike et al., 1983; Rammensee et al., 1989;

Schwartz et al., 1989; Webb & Sprent, 1990), clonal exhaustion (Webb et al., 1990;

Moskophidis et al., 1993; Rocha et al., 1995), etc. The target of tolerance induction

changed from B-lymphocytes to T-lymphocytes, upon the discovery of their synergy for

immune responses (Claman et al., 1966; Cooper & Good, 1966; Mitchell & Miller,

1968), the carrier effect (Mitchison et al., 1970; Mitchison, 1971), and MHC restriction

(Katz et al., 1973; Rosenthal & Shevach, 1973; Zinkernagel, 1974; Bevan, 1977;

Zinkernagel et al., 1978). It became common sense that it is sufficient to tolerise helper

lymphocytes. This is a major qualitative postulate that made it possible to disregard as

major scientific problems the otherwise paradoxical findings that natural autoreactive

antibodies exist (as discussed below) and that autoreactive B-lymphocytes arise during

somatic mutations.

– 43 –


Overall, the mechanism(s) of tolerance remains central to immunological thinking

and its status in the theoretical context remains exactly the same. In a nutshell, the whole

phenomenon of tolerance is reduced to its cellular levels of description . The main idea of

underlying tolerance is that individual cells bearing autoreactivity are potentially harmful,

and therefore must be, and normally are, prevented from being activated. Any means are

good as long as the end is the same: cellular-tolerance, the cellular IO-machinery is

rendered non operational. Even the regulation of potentially malignant autoreactivity, that

imposed itself as a consistent empirical fact, is still seen as one of the many candidate

mechanisms by which otherwise malignant autoreactive cells are "shut down".

The conceptual and practical consequences of this simple and reductive postulate

are multiple. Autoimmune pathologies must involve by definition the activation of

autoreactive cells. A brief overview of the fashionable accounts for their etiology during

the last fifty years illustrates the argument. In the sixties, the popular explanation for

autoimmunity was the accidental exposure of organ specific antigens that had been

secluded from the unknown mechanism of induction of cell-tolerance. The demonstration

of circulating levels of candidate target antigens during autoimmune pathologies

supported this interpretation. However, upon the demonstration that target antigens could

be found in the circulation of healthy individuals, it became obvious that the explanation

was not general (Torrigiani et al., 1969). The relatively low concentration of the target

antigen in circulation in normal individuals as compared to patients, contributed to

establish Mitchison's "low zone tolerance" as the candidate explanation for cell-tolerance

to the body antigens (Torrigiani et al., 1969). In the same period, B-T co-operation and

the carrier effect became accepted facts, and entailed new perspectives about the etiology

of autoimmune diseases. Since the popular explanations were that self-tolerance involved

essentially T-lymphocytes, autoimmunity arose by any conceivable mechanism that

provided an alternative carrier to autologous antibody determinants (Weigle, 1971;

Allison, 1971): antibody crossreactivity between somatic and microbial antigens,

microbial adjuvants overcoming the need of T lymphocytes, the classical drug induced

autoimmunity, etc.. Presently, the list of mechanisms for the etiology of autoimmunity

– 44 –


maintains the very same items, and is supplemented by the failure of signalling

transduction or signal processing mechanisms with a probable genetic basis. Thus,

currently, the most popular animal models of autoimmunity are mutant strains — NOD

(Wicker et al., 1995), lpr (Singer et al., 1994; Nagata & Suda, 1995), etc. In the IO-

machine metaphor of the cell this amount to saying that autoimmunity results from either

altered inputs or altered processing of the normal Inputs.

In the context of these views on individual cell-tolerance and pathogenic

autoreactivity it is unavoidable that the therapy of autoimmune diseases is essentially

based on immunosuppressive agents like cyclosporin, corticoids, FK506, etc.

2.5. Jerneism

A significant number of research programs in Immunology are presently constructed

upon a common set of exemplars that remain anomalous and paradoxical in the context of

Burnetism. The commitment to these exemplars reflects a perception of reality, which is

different from traditional one. Jerneism refers to this revolutionary way of doing

Immunology. As in other fields of Biology, the focus of attention switches from the

molecular biology of cells to the organisation of biological systems. The main exemplar

is the idiotypic network hypothesis of Jerne (1974), that contained already many of the

basic ideas that have been subsequently developed by different groups. It is worth

emphasising that Jerneism is not a disciplinary matrix in stricto senso. Although it has

many features of a potential scientific revolution, notably being based on non ortodox

exemplars, so far it has failed to persuade the entire community. A chronology of some

of its exemplars are presented in fig. 2.6.

– 45 –


Jerne (1960)"Immunological spcculations"

Jerne (1974) "Towards a network theory of antibody formation"

Oudin & Cazenave(1971)Antibodies with different spceificites share common idiotypes

1965

1970

1975

1980

Cronology of Jerneism

1960

Lindenman(1973) Speculations on idiotypes and homobodies

Hart et al. (1972)Idiotype suppression by antiidiotypic antibody administration

Gershon & Kondo (1970)Suppressor cells and infectious tolerance

Kunkel et al (1963)Oudin & Michel (1963) Antibodies idiotypes

Mitchison (1969)Helper cells and carrier eff'ect

Claman et al (1966)Synergism of thymus-marrow cells in antibody production

Mitchel & Miller(1968)T and B cell subsets

Richter (1975); Hoffman (1975) The first mathematical models of immune networks

Fig.2 .5- A chronology of the founding exemplars of Jerneism.

2.5.1 The Seeds of an Idea ?

In his "Immunological speculations", Jerne (1960) proposed an "immunological

terminology", in order to clarify conceptual discussions. The interaction between the

complementary regions in antibody and antigen molecules was called parataxis. The

binding sites in the antigen and the antibody were respectively the epitope and the

– 46 –


paratope ; antibody molecules would carry two identical paratopes and, being themselves

antigenic, a set of characteristic epitopes. Epitopes carried by the molecules of a given

individual animal were called the idiotopes. After a few other definitions that time has

lost, Jerne proceeded to an examination of the question "are paratopes epitopes?", well

aware of its conceptual significance. His inflexible logic lead him to conclude that

symmetry in parataxis would bring into existence reactions amongst antibodies of the

same individual — idiotypic interactions and networks in 1960! . Recognising the

conceptual challenge, he made a pragmatic retreat:

To avoid these conceptual difficulties, the assumption is made that

paratopes and epitopes represent two different classes of stereospecific

regions. A paratope is not potentially immunogenic. Paratopes are not

epitopes but anti-epitopes; keys are keys, and locks are locks.

—Jerne (1960)

In the next decade, the concept of idiotypic networks evolved from a repulsive conceptual

difficulty to a major conceptual break through. A few events contributed to this switch,

and are discussed in next section. Curiously enough, the network hypothesis (Jerne,

1974) maintained this original assumption that "parataxis" was asymmetric.

2.5.2 The Melting Pot of the Network Hypothesis

THE ANTIGENICITY OF INDIVIDUAL ANTIBODY MOLECULES:

ALLOTYPY AND IDIOTYPY

Oudin is one of the first hand references in idiotypic networks (Jerne, 1974;

Brussard, 1979; Kearney, 1993). In 1953, he was injecting antigen-antibody complexes

to rabbits, in order to generate large amounts of antibody specific for the antigen. Against

this expectation, Oudin found not only antibodies against the antigen but also against the

immunoglobulin of the animal that was source of the first antibody. This "exploration of

the unexpected" (Brussard, 1979) was the basis for the concept of allotypes, i.e.

– 47 –


genetically determined antigenic determinants from immunoglobulin molecules that vary

between groups of the same species.

The relevance of the question whether or not individual antigenic determinants

existed in each antibody molecule was strongly reinforced by antibody allotypy. The

issue intrigued the community, and remained controversial (reviewed by Kunkel et

al.(1963)). Kunkel et al. (1963) and again Oudin & Michel (1963) reported on the

specific antigenicity of individual antibodies, respectively in humans or rabbits. Oudin

(1966) called idiotypic to this antigenicity that was peculiar "firstly to antibodies against

one given antigen, and secondly to the individual or perhaps to one group of individuals,

within which the idiotypic specificities of the antibodies is not the same as it is in the

other individual groups".

The characterisation of idiotypy of antibodies originated many research programs,

namely that of Kellus & Gell (1968) that are systematically associated to the discovery of

idiotypy. Experimental research was essentially based on immunoadsorption,

immunoelectrophoresis and the Ouchterlony (1948, 1968) analysis of antibody

crossreactivities, that redeployed as a major experimental exemplar.

In this context, several findings came to be very important, notably the

suppression of idiotypic specificities in adult mice by administration of anti-idiotypic

antibodies (Hart et al., 1972), and the observation that immunisation generated both

antigen-specific and antigen-non specific antibodies that shared the same idiotypes (Oudin

& Cazenave, 1971). The two later observations were essentially unexpected and defied

any possible explanation under Clonal Selection Theory. Jerne built upon these anomalies

of the dominant theory to elaborate the network hypothesis.

THE ERA OF SUPPRESSION!

Gershon & Kondo (1970, 1971) surprised the immunological community in 1970

with a report that the B lymphocyte responses against erythrocytes could be specifically

inhibited by co-operation with T-lymphocytes. At that time, regulatory circuits were quite

fashionable in Biology. Following the lead of Gershon, regulatory suppression became

– 48 –


quite popular, and was the candidate interpretation of many other experimental systems,

notably allotype or idiotype suppression, genetic inability to respond to particular

antigens, or low-zone tolerance. The following extract of the preface for the Proceedings

of a Conference on The Immune System, Genes, Receptors and Signals illustrates rather

well the spirit of the community in 1974:

It is the era of suppression! This meeting saw a common acceptance of a

new paradigm that the outcome of immune cellular interactions was due to

competition between synergistic and suppressive T-cells.

—Sercarz et al. (1974)

The suppressor cell was considered metaphorically as the new "conductor of the

immunological orchestra", replacing GOD in main subject of immunological discussions

(Sercarz et al., 1974).

Note that negative regulation has more fundamental implications for

immunological theory, than the discovery of helper cells. The synergy between T and B

lymphocytes for the production of antibodies was rather immediately accommodated in

the scaffold of the Clonal Selection Theory, with no reformulation of its major rules.

Namely, the immune response would still increase with the size of the (cor)responding

clones; and the elimination of cells or precursors (cor)responding to the body

determinants would always eliminate the possibility of autoimmune responses. The two

signal model of lymphocyte activation of Cohn and Langman (for a good retrospective

see (Cohn, 1994)) is an explicit elaboration on this basic idea.

The situation changes dramatically with the introduction of suppressor cells.

Regulatory circuits can potentially be included in Clonal Selection Theory, by regarding

each cluster of antigen responding cells — helper, suppressor, and B lymphocytes — as

absolutely independent units of selection. However, the conceptual difficulties and

pitfalls are many, and stem from the fact that the simple direct relation between the

antigenic stimulus and acquisition of immunity, that was so seductive in Clonal Selection

Theory, is lost. On the one hand, the increase in the size of the (cor)responding

suppressor clone leads to a decrease in the immune response; and on the other hand, the

– 49 –


elimination of suppressor clones or precursors (cor)responding to body determinants can

paradoxically potentiate an autoimmune response. The paradoxical and counter intuitive

nature of suppression in a Burnetist context is probably the reason why its "guardians of

the temple", Cohn and Langman, have always regarded with scepticism the role of

suppression in "self nonself discrimination" (Cohn, 1986; Langman, 1987).

Jerne was well aware that specific suppression was an anomaly from a Burnetist

perspective. The network hypothesis was not a simple refinement of Clonal Selection

Theory, but represented a revolutionary conception of the Immune System in which

specific suppression had a major role.

SYSTEM THEORY AND SELF-ORGANISATION

In the late sixties, large, dynamic and complex systems, in which many identical

units interact with each other, became the object of scientific interest. These systems were

often social or biological, and the main paradigm was the nervous system, with 1010

neurones. The regain of interest in system theory — for which von Bertalanffy (1928,

1952) had provided the basis already in 1928 — was entailed by the recent developments

of cybernetics, and also by the possibility to treat problems of such dimension due to the

advent of computers.

In was in this context that Von Föerster (1962) developed the concept of a self-

organised system, as an independent system governed by internal rules that is able to

adapt to changes in its environment. Self-organisation was illustrated and guided research

programs in several fields of biology. The work of Weiss in embryology was particularly

significant, and his demonstration that single cell suspensions from chick-embryos in

advanced stages of differentiation could reconstitute complete organs became a major

exemplar (Weiss & Taylor, 1960).

In the early seventies the notion of self-organised system bursted from every

corner of Biology. Eigen (1971) was talking about self-organised matter; Maturana and

Varela (1980) were developing the concept of autopoiesis ; etc. The idiotypic network

– 50 –


theory of Jerne, and the concern with the "eigen-behaviour" of the immune system, were

well integrated in this general intellectual movement.

2.5.3 Towards a Network Theory.

The Eigen-Behaviour of the Immune System

In the early seventies, Jerne had developed the conviction that the Clonal Selection

Theory was no longer able to cope with the pace of accumulation of experimental

observations, and that Immunology needed a major theoretical development. He

sustained this conviction in the closing session of a colloquium on "The Genetics of

Immunoglobulins and of the Immune Response":

I think there is now a need for a novel and fundamental idea that may give

a new look to immunological theory, similar to the impact the idea of

selection had on theoretical developments in the period 1950-1970.

—Jerne (1974)

Upon a consideration of the potential diversity of the repertoire of antibodies and

lymphocyte receptors, Jerne came to the basic conclusion that,

Formally, the immune system is an enormous and complex network of

paratopes that recognize sets of idiotopes, and idiotopes that are recognized

by sets of paratopes. (...) this network interwines cells and molecules.5

—Jerne (1974)

Jerne addressed the potential responses of lymphocytes to signals involving the paratopes

(either repressive or activatory) or the idiotopes (repressive) in their receptors,

maintaining, as mentioned before, the asymmetry in the idiotypic-paratopic interaction.

He proceed to express the popular conviction that "the essence of the immune system is

the repression of its lymphocytes", reviewing its experimental support. So far the

5 The last sentence deserves emphasis because Jerne has been unfairly accused of being concerned

only with circulating antibody, neglecting lymphocytes in general, and T lymphocytes in particular.

– 51 –


proposal was not very different from some of its contemporaries, notably Lindenman's

speculations on idiotypic interactions and regulation (Lindenmann, 1973), or Gershon's

attribution of a major functional significance to suppression (Gershon, 1974).

However, Jerne proceeded into major discontinuity with Burnetism. Based on the

estimation of the concentrations of individual paratopes and idiotopes in circulation, Jerne

reached the conclusion that the network that could be conceptually outlined has

unavoidable functional existence even in the absence of immunising antigen.

The immune system, even in the absence of antigens that do not belong to

the system, must display a eigen-behaviour mainly resulting from paratope-

idiotope interaction within the system. By its eigen-behaviour the immune

system achieves a dynamic steady state as its elements interact between

themselves, and as some elements decay and new ones emerge.

—Jerne (1974)

Jerne explicited the meaning of eigen-behaviour (fig.2.8) and further demonstrated that

the network hypothesis offered simple "alternatives to current explanations" for several

unrelated phenomena, and allowed an interpretation for findings that the consensus

theory could not account for — the dynamics and regulation of the immune response,

low-zone tolerance, the inhibition of response to antigen by administration of specific

IgG antibody (then called "7S inhibition"), the antigen-competition phenomenon, the

Weigle phenomenon, the Oudin-Cazenave phenomenon, the non-specific component of

the response, and anamnestic responses or immunologic memory.

The article was concluded with a discussion of the "striking phenotypic" analogy

between the immune and the nervous systems. This conclusion not only revealed his

sources of inspiration, but opened immunological thinking to many paradigms from

cognitive sciences — representationalism, internal images, autopoiesis, etc.

Two aspects in the proposal of Jerne were fundamentally revolutionary and

innovative. They are distinct but closely related, and pointing them out here might help in

the understanding the subsequent history of the Immune Network Hypothesis. The first

aspect is that the properties of the immune system, the immunological phenomena, cannot

be reduced to the description of isolated molecules, cells, or clones; they involve

– 52 –


organisational properties of a dynamic network of molecules and cells. The second aspect

is the notion that the immune system has an eigen-behaviour, an internal structure and

dynamics that is not subordinated to external immunisation. In the context of this eigen-

behaviour, the antigen is not the cause of every immunological phenomena; its new

status is that of a perturbation that can potentially entail a change from one eigen-state to

another eigen-state, from one internal state to another internal state.

Fig . 2 .8 - "The eigen-behaviour of the immune system. The immune system contains a

set p1 of combining sites (paratopes) on immunoglobulin molecules and on cell receptors

which recognize a given epitope (E) of an antigen. This recognizing set includes the

potentially responding lymphocytes. The molecules of set p1 also display a set i1 of

idiotopes. Apart from recognizing the foreign epitope, the set p1 like wise recognizes a

set i2 of idiotopes which thus constitutes, within the immune system, a kind of internal

image of the foreign epitope. This set i2 occurs in molecular association with a set p2 of

paratopes. Likewise, the i1 is recognized within the immune system, by a set of p3 of

paratopes which represent anti-idiotypic antibodies. Beside the recognizing set p1i1 there

is a parallel set pxi1 of immunoglobulins and cell receptors which display idiotopes of

the set i1 in molecular association with combining sites that do not fit the foreign

epitope. As a first approximation, the arrows indicate a stimulatory effect when idiotopes

are recognized by paratopes on cell receptors and a suppressive effect when paratopes

recognize idiotopes on cell receptors. Successive groups of ever larger sets encompass the

entire network of the immune system" — The illustration and the legend are

from Jerne (1974) .

– 53 –


2.5.4 A parenthesis.

Idiotypic Network as an Elaboration of Burnetism.

Jerne argued that the network hypothesis was not a refusal of clonal selection, but was a

comprehensive view of the immune system in which the selection of clones was

reinterpreted and integrated in the wide context of an idiotypic network. The reaction of

the immunological community was clearly the opposite. Note that one of the heuristic

values of the Clonal Selection Theory was that, for the purposes of immunisation, the

entire immune system could be reduced to the isolated cells or clones (cor)responding to

the antigen. Most of the research programs exploring the potential of idiotypic

interactions and suppression, applied the symptomatically the very same reductionist

principle — concerned with responses, vaccination, memory, or tolerance to immunising

antigen, both theoretical or experimental research studied how individual antigens

triggered and animated the (cor)responding regulatory network. This pragmatic attitude

was also reflected in the examples that Jerne used to illustrate and promote the idiotypic

networks.

In this specific context, local networks, regulating specific immune responses,

substituted single lymphocyte clones in the traditional scaffolding of Burnetism. Those

research programs are therefore outside what we call Jerneism; and their discussion has

no place here unless a brief comment. The conceptual paradoxes and difficulties

undermining any Burnetist perception of regulatory networks have already been

mentioned. The truth of the matter is that those difficulties were never overcome by

appropriate elaboration of the Clonal Selection Theory that accommodated networks, and

the interplay between theory and experimentation remained overwhelmingly unfruitful.

Nevertheless, these studies had the consequence which was the accumulation of several

consistent observations which remain problematic or paradoxical for Burnetism. In this

context, the existence of antibody and lymphocyte autoreactivity in normal healthy

individuals is particularly relevant as we will discuss below.

– 54 –


2.5.5 Eigen-Behaviour and Cognitive Properties of the Immune System.

The concept of eigen-behaviour of the immune system and its analogy with the nervous

system — the core of what is called here Jerneism — inspires many research programs.

The experimental and theoretical approaches are systematically discussed in the respective

literature as requiring a discontinuity with the traditional views of Burnetism. The titles of

several view point and reviews articles are rather illustrative of this feature — "Self and

Nonsense" (Vaz & Varela, 1978), "From horror autotoxicus to gnothi seauton"

(Avrameas, 1991), "The Cognitive Paradigm Challenges the Clonal Selection Principle"

(Cohen, 1992b). This revolutionary immunological thinking is hopefully reflected in the

following discussion of a few theoretical and experimental exemplars shared in Jerneism.

THE COGNITIVE IMMUNE SYSTEM: SELF OR NONSENSE

Strong in his ideas on system theory, self-organisation and cognitive science,

Varela became interested in Immunology in the late seventies. Hence, Vaz & Varela

(1978) presented an elaboration of the idiotypic network hypothesis within the general

framework of autopoiesis (Maturana & Varela, 1980). The proposal situated the eigen-

behaviour on the main stream of research on idiotypic networks. Moreover, it was

probably the very first article that discussed Immunology in the context of serious

paradigms from cognitive science. The advantages and difficulties in such enterprise were

identified, notably those involving the interplay between the cognitive domains of the

immune system and that of the immunologist observer.

Several aspects of the theoretical framework were particularly innovative and

would rapidly reveal both theoretical and experimental consequences. The first was a

conception of the organisation of the immune system that was almost a logical inversion

of the "self nonself discrimination" axiom.

– 55 –


We must replace the notion of the lymphoid system as a collection of

unconnected lymphocyte clones carrying receptors directed outwards —

towards unpredictable encounters, with the notion of a network of

interacting lymphocytes where the receptors are directed inward, making

the activities of the whole lymphoid system curl and close onto itself.

—Vaz & Varela (1978)

The second aspect of the theory was the postulate that direct correlations exist between

the network structure and the immune function.

The higher the number of influences converging upon a node, the higher

must be its immunologic inertia, in the sense that the larger portions of the

network will be disturbed when the node is disturbed.

—Vaz & Varela (1978)

The conceptual shift is not trivial and must be emphasised. A suppressor cell intervening

in regulatory circuits as proposed by the Gershon school, was intrinsically suppressive,

i.e. it could be isolated as such and it displayed identified surface markers (for a

significant example see (Jandinski et al., 1976)). According to the Vaz & Varela (1978)

proposal, a suppressor cell is only be defined by its status in the network organisation

and corresponding to any cell "which resists displacements in the organisation of the

network which can be necessary for the production of intense immune responses" . In the

parlance that we introduced in chapter 1, a suppressor cannot be isolated as such, it is a

relational object.

Finally, they proposed a reasonable explanation for the disproportionate impact of

the genetic background and the manipulations in early ontogenesis on the composition

and function of the immune system; these findings, not easily accounted for by

contemporary immunological theories, were interpreted by the extreme sensitivity of self-

organised systems, with recursive eigen-states, to initial conditions or perturbations.

IDIOTYPIC NETWORKS:

GENETICS, ONTOGENESIS , DYNAMICS, AND CONNECTIVITY

As already referred, the Idiotypic Network Hypothesis inspired many experiments

based on idiotype-directed manipulations (reviewed by Kearney (1993)). Although,

– 56 –


many of these were designed to analyse specific immune responses, and not the eigen-

behaviour of the immune system, they became nevertheless exemplars of Jerneism. Only

the most classical findings are listed here without any pretension to be complete.

Several studies revealed that the antibody responses to certain haptens are

characterised by strain specific idiotype dominance (Blomberg et al., 1972; Lieberman et

al., 1974; Hansburg et al., 1977; Crews et al., 1981). The chronic elimination of

idiotypic response or dominance by administration of small amounts of complementary

anti-idiotypic antibodies in neonatal periods was reported (Augustin & Cosenza, 1976;

Strayer & Köhler, 1976), and extended to several other systems (reviewed by Kearney &

Vakil (1986)).

From the follow up of idiotypes and anti-idiotype titers in multiple immunisation

protocols and autoimmune pathologies the picture emerged that these complementary

antibodies tend to circulate in reciprocal concentrations (Kelsoe & Cerny, 1979; Bona,

1982; Rodkey & Adler, 1983), as easily predicted by the network hypothesis. Lundqvist

et al. (1989) studied the dynamics of natural serum antibodies in mice reporting erratic

fluctuations of the titers of idiotype and anti-idiotypes, that were extremely sensitive to

the administration of minimal amounts of their counterparts.

Holmberg et al. (1984) and Vakil & Kearney (1986) independently reported on

the high internal connectivity in the neonatal B lymphocyte from the spleen and foetal

liver. These neonatal antibodies revealed also a considerable degree of multireactivity

reacting against panels of autologous antigens (Dighiero et al., 1985).

IMMUNOLOGIC ACTIVITY IN NON-IMMUNISED ANIMALS

Already since the late sixties animals were raised without intestinal microflora —

"germ free" — and fed chemically defined ultrafiltered diet — "antigen free". These were

originally designed to investigate nutritional, biochemical, and metabolic effects of the

interactions between the host and the microflora (Reddy, 1971), but it became soon

obvious that they represented a good model to study the impact of background

immunisation on the immune system (reviewed in (Wostmann et al., 1971)). In this

– 57 –


period, the basic observables were the serum titers of γ-globulins by

immunoelectrophoresis, the size and morphology of spleen, lymph nodes and Peyer's

patches, the frequency of PFCs, and the blood leukocyte counts. In general, the immune

system of germ free and antigen free animals was regarded as "small and relatively

inactive but functionally intact" as expected by the Clonal Selection Theory (Pollard &

Nordin, 1971); the issue was essentially disregarded as a relevant field of research.

In late seventies and in the eighties, the activity in germ free and antigen free

animals regained obvious interest under the light of the hypothesis of an autonomous

immune network. Several groups (Kashimoto et al., 1978; Hooijkaas et al., 1984;

Pereira et al., 1986) confirmed the previous serologic observations, namely that while the

levels of IgM were essentially the same in conventional, germ free or antigen free, the

titers of IgG were considerably reduced with the decrease in the antigen load.

Comparable observations were extended to the lymphocyte compartments, namely

antigen free animals were then shown to be essentially devoid of germinal centers and

lymph nodes, but had frequencies of "naturally" activated B and T lymphocytes non

significantly reduced as compared to those found in germ free or conventional animals

(Pereira et al., 1986).

The theoretical implications of these new findings were driven notably by

Coutinho and colleagues (Pereira et al., 1986; Huetz et al., 1988; Coutinho, 1989; Varela

& Coutinho, 1991), who emphasised their incompatibility with the traditional Burnetist

view that lymphocyte activation requires external immunisation, and their consistency

with the autonomous, self-centered, immune network. Pereira et al. (1986) interpreted

the assembled observations as suggestive that "cells involved in the autonomous activity

represent a compartment of the immune system basically distinct from the antigen-

dependent one" (one can recognise here the seminal ideas of what would mature into the

Central and Peripheral Immune System).

The different facts emerging in early seventies and in late eighties from essentially

the same crude observations, represents a typical illustration of the critical role that the

commitment to exemplars plays in the establishment the significant facts.

– 58 –


AUTOIMMUNITY IN PHYSIOLOGY AND DISEASE

The conception of physiological autoreactivity is one of the hallmarks of

Jerneism. Historically, it emerged from the convergence of several lines of research,

namely the investigation of the presence of autoantibodies in normal healthy

individuals, the research on spontaneous and experimental autoimmune pathologies and

its putative regulation by suppressor cells or anti-idiotypic antibodies.

Although the presence of autoantibodies in healthy individuals was reported

several times since the beginning of the century (reviewed by Grabar (1983) and Mackay

(1983)), the issue only regained interest with the works of Avrameas and colleagues in

the early eighties (Gilbert et al., 1982; Dighiero et al., 1983; Dighiero et al., 1985), who

established that the existence of autoantibodies in healthy individuals was the rule rather

than the exception. The contrast between this finding and the Burnetist theory was

immediately emphasised, notably by Grabar (1983) (who revived his old postulate of

physiologic role for autoantibodies), Mackay (1983) , and Cohen & Cooke (1986).

As already mentioned, in the seventies Gershon's suppressive circuits and

dominant tolerance became a major exemplar for experimental research and manipulation

of autoimmune pathologies. Cohen & Wekerle (1974) were already addressing the

questions like "do T lymphocytes with specificity for self antigens exist in healthy

individuals?", what is their relation to autoreactive effectors in organ specific autoimmune

diseases?, and "how is self-recognition controlled?". The popularity and success of these

empirical research programs is rather patent in the emergence in reviews and text books

of "a failure in regulatory circuits" as a candidate mechanism for the etiology of

autoimmune diseases (for example in (Nossal, 1983) or (Roitt, 1980) ).

In the framework of Jerneism, physiological autoreactivity described in healthy

individuals is understood precisely as the natural way by which pathologic autoimmunity

is avoided (Cohen & Cooke, 1986; Huetz et al., 1988). Pathology is understood as a

deviation from healthy physiology, and not the result of some external germ or cause.

The theoretical implications are obvious, but even more significant, are the consequences

– 59 –


in terms of clinical application. On the one hand, the understanding of the onset of

disease requires, as a major prerequisite, the adequate characterisation of the healthy

physiology, which is another way of saying the eigen-behaviour. On the other hand,

again in contrast with classical immunosupression, the therapeutic strategies envisaged

are based on an appropriate stimulation of the immune system in order to re-enforce and

restore the healthy state (Rossi et al., 1989).

HETEROGENEITY OF THE COGNITIVE PARADIGMS:

A PSEUDO-DISCIPLINARY MATRIX

The assembly of experimental exemplars just mentioned, represents a common

basis for several research programs, which also share the nervous system metaphor and

the inspiration from cognitive science. However, it is important to note that these research

programs are not fully homogeneous in terms of the paradigms from cognitive science.

Representationalism is rather patent and explicitly assumed in the perception of the

immune system by Cohen (1992b) — the immune system has a primitive representation

of foreign microorganism in the genome, while the immunological homunculus is

nothing else than a representation of the dominant antigens in the body (Cohen, 1992a).

Many authors follow this very same line of thought, which is characteristic of

mathematical studies on the memory capacity of the immune networks (reviewed in next

Chapter).

Varela and colleagues are systematically in a different register. The operation of

the autopoietic immune network brings forth its own cognitive domain and categories; it

does not store or represent any information already present in the antigenic world. This

contrast between autopoietic cognition and representation is explicitly developed, notably,

in the final section of Varela et al. (1988) — "Enacting vs. Representing" .

The heterogeneity in the cognitive paradigms imported into Immunology reflects

an actual heterogeneity in contemporary cognitive science itself.

– 60 –


2.5.6 The Central and Peripheral Immune Systems.

In late eighties Coutinho and colleagues (Huetz et al., 1988; Coutinho, 1989) attempted a

synthesis of many experimental observations on the autonomy and ontogenesis of the

immune network, that was guided by several studies on mathematical models. The result

of this synthesis was the concept that the immune system is organised as Central and

Peripheral Immune Systems, that embody complementary sets of lymphocytes.

We consider a central immune system (CIS), composed of activated

lymphocytes expressing a repertoire characterised by high levels of

connectivity both with V-regions in the same set, and with other somatic

structures (autoreactivity). This organisation results in typical dynamics

which are very different from immune response dynamics and ensure

recursivity in the continuous selection of newly arising clonal reactivities,

which are connected and autoreactive. The majority of the lymphocytes in a

normal individual, however, integrate the peripheral immune system (PIS),

composed of resting cells which find no productive complementarities in the

internal environment, and, therefore, embody a repertoire which is devoid of

V-region connectivity and autoreactivity. Lack of network organisation in

this set, however, allows it to perform clonal immune responses, elicited

obviously by molecular patterns absent from the internal environment.

—Huetz et al.(1988)

This hypothesis is particularly interesting for its pragmatism — it provides a

comprehensive synthesis of otherwise incommensurable Burnetian and network

descriptions of immune phenomena (Coutinho, 1989; Varela & Coutinho, 1991). On the

one hand, the Central Immune System accommodates most of the phenomenology that

was referred to above — autonomous activity in non immunised animals, preimmune

selection of the lymphocyte repertoires, physiological autoreactivity, and dominant

tolerance — on which the Clonal Selection Theory or its later elaborations do not

pronounce. On the other hand, it not only allows a rather useful accommodation of many

exemplars of Burnetism, but it explains why these research programs seem to be so

productive in the description of immune responses and memory. The essence of the

– 61 –


Peripheral Immune System, conceptualised as a set of disconnected resting lymphocytes,

can indeed be captured by the Clonal Selection Theory. Thus, once the Central and

Peripheral dualism is ensured, each peripheral disconnected cell or its clone can be

operationally treated as independent of the rest of the immune system. Note, however,

that its conceptual status has changed considerably — previously it was a cell produced in

order to respond to an antigenic stimulus, whereas it is now a cell that does not engage in

interactions with the Central Immune System, and therefore can respond independently to

eventual antigenic stimuli.

As already mentioned, the literature embodying Jerneism is typically built around

the argument that it provides a better alternative for the interpretation of the

phenomenological dualism that is traditionally discussed in terms of Burnet's axiom of

"self nonself discrimination". Unfortunately, these alternative interpretations are typically

too qualitative, descriptive and vague to support a solid research program. The

organisational dualism between Central and Peripheral Immune Systems is a good

exception since it provides an explicit and concrete account for a phenomenological

dualism, that is based in the direct mapping between network structure and

immunological function.

In the original proposals there was no hint on how the distinction between Central

and Peripheral Immune Systems could come about, i.e. how the immune network

restricts its domain of operation to a fraction of the potential repertoire. The problem was

identified clearly and its solution remains a major theoretical problem.

2.6. Recapitulating ...

The aim of this chapter was the characterisation of our current knowledge and perception

of the immune system. Although limited and not exhaustive, this brief historical outline

of well-established theories, of a few decisive observations, and of the overall

perceptions of immune phenomena is enough to confirm that progress in Immunology

– 62 –


has not been a continuous accumulation of discoveries. Hence, three ways of organising

immunological thinking and practice were identified — Immunochemistry, Burnetism,

and Jerneism (some aspects are compared in Table 2.1). These three forms of carrying

out science, are not only founded on alternative exemplars, but embody incommensurable

perceptions of reality. Some major switches in the significant facts were already illustrate

(notably by the status of Gershon's paralysis and by the immunologic activity in antigen-

free animals). A consideration of the perception and scientific relevance of normal or

natural serum immunoglobulins will further illustrate the extent of their

incommensurability.

Table 2.1

Disciplinary

Matrix

Immunochemistry

(1900-1960)

Burnetism

(1960-...)

Jerneism

(1970-...)

Level

of

Description

Physical Chemistry

of

Immune Reactions

Molecular biology

of

Molecules and Cells

Organisation of

Molecule and Cell

Populations

Conception of the

Behaviour of

Immune System

Heteronomous

Antigen driven

(Antigen as reagent)

Heteronomous

Antigen driven

(Antigen as cell signal)

Autonomous

Organism centered

Operational

Metaphors

Body or Cells as

Chemical Reactors

Cells as IO-Machines

(Computer)

Self-Organisation

(Brain, Ecosystem)

As discussed previously, normal (or natural) serum (immuno)globulins had a

major experimental and theoretical status in the Immunochemistry matrix — they were

easily revealed and quantified by the analytical methods; they shared with antibody all the

observable physical chemical properties, with the exception of the reactivity to antigen;

and they were believed to be the product of a continuous reaction that antigen, as an

additional reagent, could drive towards the production of specific antibody. Internal to

immunological theory and models, with a different status from that of antibodies elicited

by immunisation, normal (immuno)globulins were one of the most relevant

immunological facts. In the Burnetist matrix, normal or natural serum immunoglobulins

became an irrelevant observation not internal to the theory — by definition they have to

– 63 –


result from contingent environmental immunisations. The finding that the serum levels of

IgM are essentially independent of the antigenic load must be, and therefore has been,

regarded as irrelevant. No one with a Burnetist common sense would spend time

analysing (normal or) natural serum immunoglobulins, expected as they are to have the

same nature and origin as any other antibody, and to have a purely contingent

repertoire.However, (normal or) natural serum immunoglobulins are a fundamental

immunologic fact in Jerneism, where they are back in the core of theory. They constitute

the idiotypic network, they reflect and participate in the self-organisation of the immune

system. The structure of their interactions, the repertoire they embody and their

population dynamics are amongst the most relevant properties of the immune system.

When an Immunologist looks at a band in a gel electrophoresis or an ELISA

titration of natural serum immunoglobulins, what he perceives is strictly dependent on the

Immunological matrix he was trained in. An Immunochemist used to see a species

specific globulin continuously synthesised by globulin-antibody producers; a Burnetist

sees a miscellaneous collection of antibodies resulting from contingent environmental

immunisations, and a Jerneist sees a stable network of interacting immunoglobulins

embodied in a self-organising immune system. They are simply not seeing the same

things! The same conclusion would be reached if a similar exercise was made with any

other observation or concept.

Currently, Burnetism and Jerneism mark the tone of the scientific discussion of

both theorical models and experimental observations. The two incommensurable

perceptions they embody are only rarelly made explicit in literature, that remains in this

regard ambiguous and equivocal. According to Kuhn the coexistence of

incommensurable paradigms is a symptom of periods of crisis in a discipline, and this

crisis is currently acknowledged by both traditional burnetian and innovative jernean.

literature. The historical retrospective allows us to trace the origins of this crises to the

accumulation of anomalies in the dominant burnetian matrix. These were the basis of

jerneism and are now the cumulative product of its research. The other side of the coin is

that jernean research, being concerned with levels of description which are not addressed

– 64 –


by traditional empirical methods, has difficulties in establishing a proper relation between

theory and experimentation, and consequently in convincing a sceptical scientific

community. Whether or not a scientific revolution will emerge from this crises is

something that obviously will only be known a posteriori.

Concluding this historical exercice, some aspects that characterised the emergence

of Burnetism and may be generic for other scientific revolution, can be stressed. Firstly,

the consolidation of Burnetism involved some special situations in which the interplay

between theory and experimentation was fully operational and fruitful, the best example

being the synergy between Burnet and Medawar. Secondly, the newly developed theory

provided a solution for major problems or puzzles steming from the previous disciplinary

matrix, namely the allotransplantation problem and the production of antibodies, but

reformulated them in completely different terms. Thirdly, the alterations in immunological

thinking and practice followed the wider developments in Biology, represented by the

emergence of information theory and of molecular biology.

2.7. Final Remarks.

The emergence of three different perceptions of the immune system along the history of

Immunology was the object of this chapter. Hopefully, the reader has acquired the notion

that current immunological thinking is determined by two incommensurable conceptions

of immunity that were referred here as Burnetism and Jerneism. The former is the

traditional disciplinary matrix. It is rooted in the Clonal Selection Theory of Burnet.

During the last four decades it has accumulated many anomalies in the Kuhnian sense,

notably the established finding that a physiologic (in the sense of nonpathologic)

autoreactivity exists. Jerneism is precisely built upon the major anomalies of Burnetism.

Its classic exemplar is the Idiotypic Network Hypothesis of Jerne, that restated the most

solid propositions of Clonal Selection Theory — notably the one cell-one receptor axiom,

and the generation of diversity —, and accommodated them in the context of the internal

– 65 –


organisation of the immune system. The most recent theoretical elaboration on this

exemplar is the working hypothesis that the immune system is organised into a Central

and a Peripheral subsystems; a hypothesis that stems from the recent history of

Immunology as a radically simple but comprehensive view. The postulate of an

organisational dualism allows a comprehensive restatement of the main exemplars of

both Burnetism and Jerneism in a common framework, and in so doing creates a major

problem — How can the ontogenesis of an immune network give rise to an

organisational dualism? The present thesis is precisely an elaboration on the Central and

Peripheral Immune Systems, with special attention being paid to the nature of this

dualism.

– 66 –


References

Allison, A.C. (1971). Unresponsiveness to self antigens. Lancet. 2 , 1401.

Arrhenius, S. (1907). Immunochemistry. The application of the principles of physical

chemistry to the study of the biological antibodies. New York: Macmillan.

Augustin, A. & Cosenza, A. (1976). Expression of new idiotype following neonatal

idiotypic suppression of a dominant clone. Eur.J. Immunol. 6 , 497.

Avrameas, S. (1991). Natural autoantibodies: from 'horror autotoxicus' to 'gnothi

seauton'. Immunol. Today. 12 , 154.

Behring, E. & Kitasato, S. (1890). Ueber das Zusdandekommen der Diphtherie—

Immunität und der Tetanus-Immunität bei Thieren. Deutsche Medizinische

Wochenschrift. 16 , 1113.

Bevan, M.J. (1977). In a radiation chimera, host H-2 antigens determine immune

responsiveness of donor cytotoxic cells. Nature. 269 , 417.

Bibel, D.B., (1988). Milestones in Immunology. A historical exploration. Madison,WI:

Science Tech Publishers.

Billingham, R.E., Brent, L. & Medawar, P.B. (1953). 'Actively acquired tolerance' of

foreign cells. Nature. 172 , 603.

Blomberg, B., Geckeler, W.R. & Weigert, M. (1972). Genetics to the antibody response

to dextran in mice. Science. 177 , 178.

Bona, C.A. (1982). Inverse fluctuations of idiotypes and anti-idiotypes during immune

response. In: Regulation of Immune Dynamics. (DeLisi & Hiernaux, eds.) pp. 75.

Boca Raton: CRC Press.

Bordet, J. (1939). Traité d'immunité dans les maladies infectieuses. Paris: Masson et

Cie. Éditeurs.

Breinl, F. & Haurowitz, F. (1930). Chemische untersuchungen des praezipitates aus

haemoglobin und antihaemoglobulin-serum und bemerkungen uerber die natur der

antikoerper. Zs. physiol. Chem. 192 , 45.

Brussard, A. (1979). Jacques Oudin et son oeuvre. Ann. Immunol. (Inst. Pasteur). 130

C, 101.

– 67 –


Burnet, F.M. (1941). The production of antibodies. Melbourne: Macmillan.

Burnet, F.M. (1957). A Modification of Jerne's Theory of Antibody Production using

the Concept of Clonal Selection. Aust. J. Sci. 20 , 67.

Burnet, F.M. (1958). The Clonal Selection Theory of Acquired Immunity. Nashville:

Vanderbuilt University Press.

Burnet, F.M. (1967). The impact of ideas on Immunology. Could Spring Harbour

Symp. Quant. Biol. 32 , 1.

Burnet, F.M. (1969). Cellular immunology. Victoria and New York: Melbourne

University Press. Cambridge University Press.

Burnet, F.M. & Fenner, F. (1949). The production of antibodies. Melbourne:

Macmillan.

Byars, L.T. & Randall, P. (1953). Plastic surgery: wound care and skin grafting. Ann.

Rev. Med. 4 , 373.

Chase, M.W. (1946). Inhibition of experimental drug allergy by prior feeding the

sensitizing antigen. Proc. Soc. Exptl. Biol. Med. 61 , 257.

Chase, M.W. (1959). Immunological Tolerance. Annu. Rev. Microbiology. 13 , 349.

Claman, H.N., Chaperon, E.A. & Triplett. (1966). Thymus-marrow cell combinations,

synergism in antibody production. Proc. Soc; Exp. Biol. Med. 122 , 1170.

Cohen, I.R. (1992a). The cognitive paradigm and the immunological homunculus.

Immunol. Today. 13 , 490.

Cohen, I.R. (1992b). The cognitive principle challenges clonal selection. Immunol.

Today. 13 , 441.

Cohen, I.R. & Cooke, A. (1986). Natural autoantibodies might prevent autoimmune

disease. Immunol. Today. 7 , 363.

Cohen, I.R. & Wekerle, H., Brent & Holborrow, (1974). Autoimmunity and self-

recognizing T lymphocytes. In: Progress in Immunology II. (Brent,

L.,Holborrow, J., eds.) 5 pp. 5. North-Holland Publishing Company.

Cohn, M. (1986). The concept of functional idiotype network for immune regulation

mocks all and comforts none. Ann. Immunol.(Inst. Pasteur.) 137C, 64.

– 68 –


Cohn, M. (1994). The wisdom of hindsight. Annu Rev Immunol. 12 , 1.

Cooke, R.A., Smith, J.M. & Skaggs, J.T. (1955). Allergy and Immunology. Ann. Rev.

Med. 6 , 125.

Cooper, M.D. & Good, R.A. (1966). The functions of the thymus system and the bursa

system. J. Exp. Med. 123 , 76.

Coutinho, A. (1989). Beyond clonal selection and network. Immunol. Rev. 110 , 63.

Crews, S., Griffin, J., Huang, H., Calame, K. & Hood, L. (1981). A single VH gene

segment encodes the immune response to phosphorylcholine: somatic mutation is

correlated with the class of the antibody. Cell. 25 , 59.

Crick, F. (1970). Central dogma of molecular biology. Nature. 227 , 561.

Crick, F.H.C. (1967). General discussion on theories of antibody variability. Cold

Spring Harbor Symp. Quant. Biol. 32 , 169.

Dighiero, G., Lymberi, P., Holmberg, D., Lundkvist, I., Coutinho, A. & Avrameas, S.

(1985). High frequency of natural auto-antibodies in normal new born mice. J.

Immunol. 134 , 765.

Dighiero, G., Lymberi, P., Mazié, J.-C., Rouyre, S., Butler-Browne, G.S., Whalen,

R.G. & Avrameas, S. (1983). Murine hybridomas secreting natural monoclonal

antibodies reacting with self antigens. J. Immunol. 131 , 2267.

Dixon, F.J. (1958). Allergy and immunology. Autoimmunity in disease. Annu. Rev.

Med. 9 , 257.

Dixon, F.J., Talmage, D.W. & Maurer, P.H. (1952). Radiosensitive and radioresistant

phases in the antibody response. J. Immunol. 68 , 693.

Edelman, G.M. (1959). Dissociation of gamma globulin. J. Amer. Chem. Soc. 81 ,

3155.

Edelman, G.M. & Gally, J.A. (1962). The nature of Bence-Jones proteins. Chemical

similarities to polypeptide chains of myeloma globulins and normal gamma

globulins. J.Exp. Med. 116 , 207.

Ehrlich, P. (1900). On immunity with special reference to cell life. Proc. Royal Soc.,

London. 66 , 424.

– 69 –


Eigen, M. (1971). Selforganisation of matter and the evolution of biological

macromolecules. Die Naturwissenschaften. 10 , 465.

Fagraeus, A. (1948). The plasma cellular reaction and its relation to the formation of

antibodies in vitro. J. Immunol. 58 , 1.

Felton, L.D. (1949). The significance of antigen in animal tissues. J.Immunol. 61 , 107.

Felton, L.D. & Ottinger, B. (1942). Pneumococcus polysaccharide as a paralyzing agent

on the mechanism of immunity in white mice. J. Bacteriol. 43 , 94.

Freedman, M.H. & Sela, M. (1966). Recovery of antigenic activity upon reoxidation of

completely reduced polyalanyl rabbit immunoglobulin-G. J.Biol.Chem. 241 ,

2383.

Freund. (1947). Some aspects of active immunization. Ann. Rev. Microbiology. 1 , 291.

Fishman, M. & Adler, F.L. (1967). The role macrophage-RNA in the immune response.

Cold Spring Harbor Symp. Quant. Biol. 32 , 343.

Gard, S. (1960). Introduction speech of Nobel Prize for Physiology and Medecine 1960.

Gershon, R.K. (1974). T cell regulation: the second law of thymodynamics. In: The

Immune System. Genes, Receptors, Signals. (Sercarz et al., eds.) pp. 471. New

York: Academic Press.

Gershon, R.K. & Kondo, K. (1970). Cell interactions in the induction of tolerance: the

role of thymic lymphocytes. Immunology. 18 , 723.

Gershon, R.K. & Kondo, K. (1971). Infectious Immunological Tolerance.

Immunology. 21 , 903.

Gilbert, B.G., Dighiero, G.P. & Avrameas, S. (1982). Naturally occuring antibodies

against nine common antigens in human sera. I. Detection, isolation, and

characterisation. J. Immunol. 128 , 2779.

Gottlieb, A.A., Glisin, V.R., & Doty, P. (1967). Studies on macrophage RNA involved

in antibody production. Proc. Natl. Acad. Sci. USA. 57 , 1849.

Grabar, P. (1950). Immunochemistry. Annu. Rev. Biochemistry. 19 , 453.

Grabar, P. (1983). Autoantibodies and the physiological role of immunoglobulins.


– 70 –


Hansburg, D., Briles, D.E. & Davie, J.M. (1977). Analysis of the diversity of the

murine response to dextran B1355. II. Demonstration of multiple idiotypes with

variable expression in several strains. J. Immunol. 119 , 1406.

Hart, D.A., Wang, A.-L. & Pawlak, L.L. (1972). Suppression of idiotypic specificities

in adult mice by administration of antiidiotypic antibody. J. Exp. Med. 135 , 1293.

Haurowitz, F. (1950). The chemistry and function of proteins. New York and London:

Academic Press.

Haurowitz, F. (1953). The immunological response. Annu. Rev. Microbiology. 7 , 389.

Haurowitz, F. (1967). The evolution of selective and instructive theories of antibody

formation. Cold Spring Harbor Symp. Quant. Biol. 32 , 559.

Haurowitz, F. & Crampton, C.F. (1950). The role of the nucleus in the protein

synthesis. Exptl. Cell. Research. 19 , 45.

Heidelberger, M. (1932). Immunochemistry. Annu. Rev. Biochem. 1 , 655.

Heidelberger, M. (1938). Protein constitution and immunological behaviour. Cold


Heidelberger, M., Dilapi, M.M., Siegel, M. & Walter, A.W. (1950). Persistence of

antibodies in human subjects injected with pneumococcal polysaccharides.

J.Immunol. 65 , 535.

Hilschmann, N. & Craig, L.C. (1965). Amino acid sequence studies on with Bence-

Jones proteins. Proc. Natl. Acad. Sci. USA. 53 , 1403.

Holmberg, D., Forsgren, S., Ivars, F. & Coutinho, A. (1984). Reactions amongst IgM

antibodies derived from normal, neonatal mice. Eur. J. Immunol. 14 , 435.

Hooijkaas, H., Benner, R., Pleasants, J.R. & Wostmann, B.S. (1984). Isotypes and

specificities of immunoglobulin produced by germ-free mice fed chemically defined

ultrafiltered "antigen-free" diet. Eur. J. Immunol. 14 , 1127.

Huetz, F., Jacquemart, F., Pena-Rossi, C., Varela, F. & Coutinho, A. (1988).

Autoimmunity: The moving boundaries between physiology and pathology. J.

Autoimmunity. 1 , 507.

Jandinski, J., Cantor, H., Tadakuma, T., Peavy, D.L. & Pierce, C.W. (1976).

Separation of helper T cells from suppressor T cells expressing different Ly

– 71 –


components. I. Polyclonal Activation: suppressor and helper activities are inherent

properties of distinct T-cell subclasses. J. Exp. Med. 143 , 1382.

Jerne, N.K. (1955). The Natural-Selection Theory of Antibody Formation. Proc. Natl.

Acad. Sci. USA. 41 , 849.

Jerne, N.K. (1960). Immunological Speculations. Ann. Rev. Microbiology. 14 , 341.

Jerne, N.K. (1966). The natural selection theory of antibody formation; Ten years later.

In: Phage and The Origins of Molecular Biology. (Cairns et al., ed.) pp. 301. Cold

Spring Harbor: Cold Spring Harbor Laboratory of Quantitative Biology.

Jerne, N.K. (1967). Waiting for the end. Cold Spring Harbor Symp. Quant. Biol. 32 ,

591.

Jerne, N.K. (1974). Towards a network theory of the immune system. Ann. Immunol.

(Inst. Pasteur). 125C, 373.

Jerne, N.K. & Nordin, A.A. (1963). Plaque Formation in agar by single antibody-

producing cells. Science. 140 , 405.

Kashimoto, K., Handa, H., Umehara, K. & Sasaki, S. (1978). Germfree mice reared on

an "antigen-free" diet. Laboratory Animal Science. 28 , 38.

Kappler, J.W., Roehm, N., & Marrack, P. (1987). T cell tolerance by clonal elimination

in the thymus. Cell. 49 , 273.

Katz, D.H., Hamaoka, T., Dorf, M.E., Mauser, P.H. & Benacerraf, B. (1973). Cell

interactions between histocompatible T and B lymphocytes. The H-2 gene complex

determines successful physiologic lymphocyte interactions. Proc. Natl. Acad. Sci.

USA. 70 , 2624.

Kearney, J. & Vakil, M. (1986). Idiotype directed interaction during ontogeny play a

major role in the establishment of the adult B cell repertoire. Immunol. Rev. 94 ,

39.

Kearney, J.F. (1993). Idiotypic Networks. In: Fundamental Immunology. (Paul, ed.)

pp. 887. New York: Raven Press.

Kelsoe, G. & Cerny, J. (1979). Reciprocal expression of idiotypic and antiidiotypic

clones following antigen stimulation. Nature. 279 ,

– 72 –


Kelus, A.S. & Gell, P.G.H. (1968). Immunological analysis of rabbit anti-antibody

systems. J. Exp. Med. 127 , 215.

Klein. (1982). Immunology: The Science of Self-Nonself Descrimination. New York:

Wiley.

Kuhn, T.S. (1970). The structure of scientific revolutions. Chicago: University of

Chicago Press.

Kuhn, T.S. (1990). La tension essentielle. Tradition et changement dans les sciences.

Paris: Éditions Gallimard.

Kunkel, H.G., Mannik, M. & Williams, R.C. (1963). Individual antigenic specificity of

isolated antibodies. Science. 140 , 1218.

Langman, R.E. (1987). The self-nonself discrimination is not regulated by suppression.

Cell. Immunol. 108 , 214.

Lawrence, H.S. (1970). Editorial Cellular Immunology. Cellular Immunology. 1 , 1.

Lederberg, J. (1959). Genes and Antibodies. Science. 129 , 1649.

Lieberman, R., Potter, M., Munshinski, E.B., Humphrey, W.J. & Rudikoff, S. (1974).

Genetics of a new Ig VH (T15 idiotype) marker in the mouse regulating natural

antibody to phosphorylcholine. J. Exp. Med. 139 , 983.

Lindenmann, J. (1973). Speculations on idiotypes and homobodies. Ann. Immunol.

(Inst. Pasteur). 124C, 171.

Lundkvist, I., Coutinho, A., Varela, F. & Holmberg, D. (1989). Evidence for a

functional network amongst natural antibodies in normal mice. Proc. Natl. Acad.

Sci. USA. 86 , 5074.

MacDonald, H.R., Pedrazzini, T., Schneider, R., Louis, J.A., Zinkernagel, R.M., &

Hengartner, H. (1988). Intrathymic elimination of Mlsa-reactive (Vβ6+) cells

during neonatal tolerance induction to Mlsa-encoded antigens. J. Exp. Med. 167 ,

2005.

Mackay, I.R. (1983). Natural antibodies to the fore — forbidden clones to the rear.


Makëla, O. (1967). The specificity of antibodies produced by single cells. Cold Spring

Harbor Symp. Quant. Biol. 32 , 423.

– 73 –


Marrack, J.R. (1934). The chemistry of Antigens and Antibodies. Special Report Series

no. 194. London: Medical Research Council.

Maturana, H.R. & Varela, F.J. (1980). Autopoiesis and Cognition: the Realization of the

Living. Dordrecht: Reidel.

Medawar, P.B. (1946a). Immunity to homologous grafted skin. I. The suppression of

cell division in grafts transplanted to immunized animals. Brit. J. Exptl. Pathol.

27 , 9.

Medawar, P.B. (1946b). Immunity to homologous grafted skin. II. The relationship

between antigens of blood and skin. Brit. J. Exptl. Pathol. 27 , 15.

Medawar, P.B. (1958). The Immunology of Transplantation. The Harvey Lectures. 52 ,

144.

Mitchell, G.F. & Miller, J.F.A.P. (1968). Immunological activity of thymus and

thoracic-duct lymphocytes. Proc. Natl. Acad. Sci. USA. 59 , 296.

Mitchison, N.A. (1964). Induction of immunological paralysis in two zones of dosage.

Proc. Roy. Soc. B. 161 , 275.

Mitchison, N.A. (1971). The carrier effect in the secondary response to hapten-protein

conjugates. I. Cellular cooperation. Eur. J. Immunol. 1 , 18.

Mitchison, N.A., Rajewsky, K. & Taylor, R.B. (1970). Cooperation of antigenic

determinants and of B cells in the induction of antibodies. In: Developmental

aspects of antibody formation and structure. ed.) 2 . pp. 547. New York: Academic

Press.

Monod, J. (1947). The phenomenon of enzymatic adaptation and its bearings on

problems of genetics and cellular differentiation. Growth Symposium. 11 , 223.

Moskophidis, D., Laine, E. & Zinkernagel, R.M. (1993). Peripheral clonal deletion of

antiviral memory CD8+ T cells. Eur. J. Immunol. 23 , 3306.

Moulin, A.M. (1991). Le dernier language de la médecine. Histoire de l'Immunologie de

Pasteur au Sida. Paris: Presses Universitaires Francaises.

Mudd, S. (1932). A hypothetical mechanism of antibody formation. J. Immunol. 23 ,

423.

– 74 –


Nagata, S. & Suda, T. (1995). Fas and Fas ligand: lpr and gld mutations. Immunol.

Today. 16 , 39.

Nossal, G.J. (1983). Cellular mechanisms of immunologic tolerance. Annu. Rev.

Immunol. 1 , 33.

Nossal, G.J. & Makëla, J. (1962). Elaboration of antibodies by single cells. Annu. Rev.

Microbiol. 53.

Nossal, G.J., Shortman, K.D., Miller, J.F.A.P., Mitchell, G.F., & Haskill, J.S.

(1967). The target cell in the induction of immunity and tolerance. Cold Spring

Harbor Symp. Quant. Biol. 32 , 369.

Nossal, G.J. & Pike, B.L. (1975). Evidence for the clonal abortion theory of B-

lymphocyte tolerance. J. Exp. Med. 141 , 904.

Ouchterlony, Ö. (1948). In vitro method for testing the toxin-producing capacity of

diphtheria bacteria. Acta Pathologica et Microbiologica Scandinavica. 25 , 186.

Ouchterlony, Ö. (1968). Handbook of Immunodiffusion and Immunoelectrophoresis.

Ann Arbor: Ann Arbor Science Publishers.

Oudin, J. (1966). The genetic control of antibody synthesis. Proc. Roy. Soc. London.

B. 166 , 207.

Oudin, J. & Cazenave, P.-A. (1971). Similar Idiotypic Specificities in Immunoglobulin

Fractions with different Antibody Functions or Even without Detectable Antibody

Function. Proc. Natl. Acad. Sci. USA. 68 , 2616.

Oudin, J. & Michel, M. (1963). Une nouvelle forme d'allotypie des globulines γ du

sérum de Lapin, apparemment liée à la function et à la spécificité anticorps. C.R.

Acad. Sci. (Paris). 257 , 805.

Papermaster, B.W. (1967). The clonal differentiation of antibody-producing cells. Cold


Pauling, L. (1940). A theory of the structure and process of formation of antibodies. J.

Amer. Chem. Soc. 32 , 2643.

Pereira, P., Forni, L., Larsson, E.-L., Cooper, M., Heusser, C. & Coutinho, A.

(1986). Autonomous activation of B and T cells in antigen-free mice. Eur. J.

Immunol. 16 , 685.

– 75 –


Pike, B.L., Abrams, J. & Nossal, G.J. (1983). Clonal anergy: inhibition of antigen-

driven proliferation among single B lymphocytes from tolerant animals, and partial

breakage of anergy by mitogens. Eur. J. Immunol. 13 , 214.

Pollard, M. & Nordin, A. (1971). Immune Capabilities of Germ-Free Animals. In:

Progress in Immunology. First International Congress of Immunology (Amos, B.,

ed.) pp.1295.Washington: Academic Press.

Porter, R.R. (1950). A chemical study of rabbit antiovalbumin. Biochem. J. 46 , 473.

Porter, R.R. (1959). The hydrolysis of rabbit gamma globulin and antibodies with

crystalline papain. Biochem. J. 73 , 119.

Raffel, S. (1954). Immunity. Hypersensitivity. Serology. New York: Appleton-Century-

Crofts, Inc.

Raffel, S. (1956). Immunity. Annu. Rev. Med. 7 , 385.

Rammensee, H.G., Kroschewski, R. & Frangoulis, B. (1989). Clonal anergy induced

in mature Vbeta6+ T lymphocytes on immunizing Mls-1b mice with Mls-1a

expressing cells. Nature. 339 , 541.

Reddy, B.S. (1971). Nutrition Society Symposium. Nutritional and biochemical aspects

of host-microflora interactions. Introductory Remarks. Fed. Proc. 30 , 1172.

Rocha, B., Grandien, A. & Freitas, A.A. (1995). Anergy and exhaustion are

independent mechanisms of peripheral T cell tolerance. J. Exp. Med. 181 , 993.

Rodkey, L.S. & Adler, F.L. (1983). Regulation of natural antiallotype antibody

responses in idiotypic network-induced auto-antiidiotypic antibodies. J. Exp. Med.

157 , 1920.

Roitt, I.M. (1980). Essential Immunology. Oxford: Blackwell Scientific Publications.

Rosenthal, A.S. & Shevach, E.M. (1973). Function of macrophages in antigen

recognition by guinea pig T lymphocytes. I. Requirement for histocompatible

macrophages and lymphocytes. J. Exp. Med. 138 , 1194.

Rossi, F., Dietrich, G. & Kazatchkine, M.D. (1989). Anti-idiotypes against

autoantibodies in normal immunoglobulins: evidence for network regulation of

human autoimmune responses. Immunol. Rev. 110 , 135.

– 76 –


Schoenheimer, R., Rattner, S., Rittenburg, D. & Heidelberger, M. (1942). The

interaction of the blood proteins of the rat with dietary nitrogen. J.Biol.Chem.

144 , 541.

Schwartz, R.H., Mueller, D.L., Jenkins, M.K. & Quill, H. (1989). T-cell clonal anergy.

Cold Spring Harbor Symp. Quant. Biol. 54 , 605.

Sercarz, E.E., Williamson, A.R. & Fox, C.F., (1974). The Immune System. Genes,

Receptors, Signals. New York: Academic Press.

Silverstein, A.M. (1989). A History of Immunology. San Diego: Academic Press.

Sinclair, N.R. (1992). Can B-cell signalling be understood? Scand. J. Immunol. 35 ,

125.

Singer, C.C., A.C., C., Marshak-Rothstein, A., Martinez, C. & Abbas, A.K. (1994).

Apoptosis, Fas and systemic autoimmunity: the MRL-lpr/lpr model. Curr. Opin.

Immunol. 6 , 913.

Smithies, O. (1967). The genetic basis of antibody variability. Cold Spring Harbor

Symp. Quant. Biol. 32 , 161.

Strayer, D.A. & Köhler, H. (1976). Immune response to phosphorylcholine. II. Natural

"auto"-anti-receptor antibody in neonatal BALB/c mice. Cell. Immunol. 25 , 294.

Svedberg, T. (1939). The ultra-centrifuge and the study of high molecular weight

compounds. Nature. 139 , 1051.

Taliaferro, W.H. (1949). The cellular basis of immunity. Ann. Rev. Microbiology. 3 ,

159.

Talmage, D.W. (1957). Allergy and Immunology. Ann. Rev. Med. 8 , 239.

Tiselius, A. & Kabat, E.A. (1939). An electrophoretic study of immune sera and purified

antibody preparations. J. Exp. Med. 69 , 119.

Torrigiani, G., Doniach, D. & Roitt, I.M. (1969). Serum Thyroglobulin Levels in

Healthy Subjects and in Patients with Thyroid Disease. J. Clin. Endocr. 29 , 305.

Vakil, M. & Kearney, J.F. (1986). Functional characterization of monoclonal auto-anti-

idiotype antibodies isolated from the early B cell repertoire of BALB/c mice. Eur. J.

Immunol. 16 , 1151.

– 77 –


Varela, F. & Coutinho, A. (1991). Second generation immune networks. Immunol.

Today. 12 , 159.

Varela, F., Coutinho, A., Dupire, B. & Vaz, N. (1988). Cognitive networks: immune,

neural and otherwise. In: In Theoretical Immunology. (Perelson, ed.) 3 . pp. 359.

Redwood City: Addison-Wesley Publ. Co., Inc.

Vaz, N.M. & Varela, F.G. (1978). Self and nonsense: an organism-centered approach to

immunology. Med. Hypothesis. 4 , 231.

Von Bertalanffy, L. (1928). Modern theories of development. London: Oxford

University Press.

Von Bertalanffy, L. (1952). Problems of Life. London: Watts.

Von Boehmer, H., Teh, H.S., & Kisielow, P. (1989). The thymus selects the useful,

neglects the useless and destroys the harmful. Immunol. Today. 10 , 50.

Von Förster, H. (1962). Principles of self-organisation. New York.

Webb, S., Morris, C. & Sprent, J. (1990). Extrathymic tolerance of mature T cells:

clonal elimination as a consequence of immunity. Cell. 63 , 1249.

Webb, S.R. & Sprent, J. (1990). Induction of neonatal tolerance to Mls a antigens by

CD8+ T cells. Science. 248 , 1643.

Weigle, W.O. (1971). Recent observations and concepts in immunological

unresponsiveness and autoimmunity. Clin. Exp. Immunol. 9 , 437.

Weiss, P. & Taylor, A.C. (1960). Reconstitution of complete organs from single-cells

suspensions of chick-embryos in advanced stages of different. Proc. Natl. Acad.

Sci. USA. 46 , 1170.

Wicker, L.S., Todd, J.A. & Peterson, L.B. (1995). Genetic control of autoimmune

diabetes in the NOD mouse. Annu. Rev. Immunol. 13 , 179.

Wostmann, B.S., Pleasants, J.R. & Bealmear, P. (1971). Dietary stimulation of immune

mechanisms. Fed. Proc. 30 , 1779.

Zinkernagel, R.M. (1974). Restriction by H-2 gene complex of transfer of cell-mediated

immunity to Listeria monocytogenes. Nature. 251 , 230.

– 78 –


Zinkernagel, R.M., G.N., C., Aalthage, A., Cooper, S., Klein, P.A. & Klein, J.

(1978). On the thymus in the differentiation of "H-2 self-recognition" by T cells:

Evidence for dual recognition? J. Exp. Med. 147 , 882.

– 79 –

the foundations of the disciplinary matrix -...

Documents