Ontogeny of Acquired Immunity

The Ciba Foundation for the promotion of international cooperation in medical and chemical research is a scieniifc and educational charity established by CIBA Limited - now CIBA-GEIGY Limited - of Basle. The Foundation operates independenily in London under English trust law.

Ciba Foundation Symposia are published in collaboration with Associated Scientific Publishers (Elsevier Scientific Publishing Company, Excerpta Medica. North-Holland Publishing Company) in Amsterdam.

Associated Scientific Publishers, P.O. Box 3489, Amsterdam

Ontogeny of Acquired Immunity

A Ciba Foundation Symposium

1972

Elsevier . Excerpta Medica . North-Holland Associated Scientific Publishers -Amsterdam - London . New York

9 Copyright 1972 Ciba Foundation

All rights reserved. N o part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying and recording, or by any information storage and retrieval system, without permission in writing from the publishers.

ISBN Excerpta Medica 90 219 4005 1 ISBN American Elsevier 0-444-10381-3

Library of Congress Catalog Card Number 72-81001

Published in 1972 by Associated Scientific Publishers, P.O. Box 3489, Amsterdam, and 52 Vanderbilt Avenue, New York, N. Y. 10017. Suggested series entry for library catalogues: Ciba Foundation Symposia.

Printed in The Netherlands by Royal VanGorcum Ltd., Assen

Contents

J. L. GOWANS Chairman’s introduction IX

N. K. JERNE What precedes clonal selection? 1 Discussion 10

A. M. SILVERSTEIN Immunological maturation in the foetus: modulation of the pathogenesis of congenital infectious diseases 17 Discussion 26

populations 35 Discussion 55

J. J . T. OWEN The origins and development of lymphocyte

M. ADINOLFI Ontogeny of components of complement and lysozyme 65 Discussion 8 1

s. DRAY Allotype suppression 87 Discussion 103

D. P. STITES, J. WYBRAN, M. c. CARR and H. H. FUDENBERG

Development of cellular immune competence in man 113 Discussion 130

(‘blocking antibodies’) as mediators of immunological non-reactivity to cellular antigens 133 Discussion 143

K. E. HELLSTR~M and I. HELLSTR~M The role of serum factors

A. E. BEER and R. E. BILLINGHAM Concerning the uterus as a graft site and the foetus as a natural parabiotic organismic homograft 149 Discussion 167

VI Contents

M. J. SELLER Bone marrow transplantation in a genetically determined anaemia in the mouse Discussion 187

175

P. J . LACHMANN Genetic deficiencies of the complement system 193

F. s. ROSEN Defects in immunological development in man 213

D. w. VAN BEKKUM and K. A. DICKE Treatment of immune defi-

Discussion 209

Discussion 2 18

ciency disease with bone marrow stem cell concentrates 223 Discussion 237

H. E . M. KAY Foetal thymus transplants in man 249

A. R. HAYWARD and J. F. SOOTHILL Reaction to antigen by human

Discussion 254

foetal thymus lymphocytes 261 Discussion 268

Index of contributors 275 Subject index 276

Contributors

Symposium on Ontogeny of Acquired Immunity, held at the Ciba Foundation, London, 23rd-25th November 1971

J . L. GOWANS (Chairman) MRC Cellular Immunology Unit, Sir William Dunn School of Pathology, University of Oxford, Oxford OX1 3RE, England

Medical School, London SE1 UP4, England

for Health Research TNO, Rijswijk, The Netherlands

Texas, Southwestern Medical School, 5323 Hines Boulevard, Dallas, Texas 75235, USA

Durham, England

Surrey, England

Sciences, University of Illinois at the Medical Center, PO Box 6998, Chicago, 111. 60680, USA

CH-1211 Geneva 27, Switzerland

ment of Medicine, University of California, San Francisco, Cal. 94122, USA

A. R. HAYWARD Department of Immunology, Institute of Child Health, 30 Guilford Street, London, WClN IEH, England

K. E. HELLSTR~M Department of Pathology, School of Medicine, University of Washington, Seattle, Washington 98 105, USA

M. ADINOLFI Paediatric Research Unit, Guy’s Hospital

D. w . VAN BEKKUM Radiobiological Institute of the Organization

R. E. BILLINGHAM Department of Cell Biology, University of

w. D. BILLINGTON Department of Zoology, The University,

G. CURRIE Chester Beatty Research Institute, Belmont, Sutton,

s. DRAY Department of Microbiology, School of Basic Medical

w. PAGE FAULK Immunology Division, World Health Organization,

H. H. FUDENBERG Section of Immunology and Hematology, Depart-

v1n Contributors

L. A. HERZENBERG Department of Genetics, Stanford University School of Medicine, Stanford, Cal. 94305, USA

J. c. HOWARD MRC Cellular Immunology Unit, Sir William Dunn School of Pathology, University of Oxford, Oxford OX1 3RE, England

J. H. HUMPHREY Division of Immunology, National Institute for Medical Research, Mill Hill, London NW7 lAA, England

N. K . JERNE Basel Institute for Immunology, 487 Grenzacher- strasse, CH-4058, Basel, Switzerland

H. E. M. KAY Department of Clinical Pathology, The Royal Marsden Hospital, Fulham Road, London SW3 556, England

P. J. LACHMANN Department of Immunology, Royal Postgraduate Medical School, London W12 OHS, England

N. A . MITCHISON Tumour Immunology Unit, Department of Zoology, University College, London WCl E 6BT, England

J. J. T. OWEN Tumour Immunology Unit, Department of Zoology, Univrseity College, London WC1 E 6BT, England

P. PORTER Unilever Research Laboratories, Colworth House, Sharnbrook, Bedfordshire, England

F. s. ROSEN Laboratory of Immunology, Department of Pediatrics, Harvard Medical School, Children’s Hospital, 300 Longwood Avenue, Boston, Mass. 02115, USA

MARY J . SELLER Paediatric Research Unit, Guy’s Hospital Medical School, London SE1 UP4, England

A. M. SILVERSTEIN Department of Ophthalmology, The Wilmer Ophthalmological Institute, The Johns Hopkins University School of Medicine, The Johns Hopkins Hospital, 601 N. Broadway, Baltimore, Md. 21205, USA

J. F. SOOTHILL Department of Immunology, Institute of Child Health, 30 Guilford Street, London WClN IEH, England

R. B. TAYLOR Department of Pathology, University of Bristol Medical School, University Walk, Bristol BS8 1 TD, England

Editors: RUTH PORTER and JULIE KNIGHT

Introduction

J. L. GOWANS

MRC Cellular Immunology Unit, Sir William Dunn School of Pathology, Oxford

One incentive for studying the ontogeny of the immune response is that it may help to illuminate the errors of development which lead to immune deficiency states in man. However, the interaction has not been entirely in one direction, for studies of human disease had already provided hints that the immune system was built up from components derived separately from the thymus and from some bursa1 equivalent, an idea strongly influenced by the anatomical separation of the two components in birds and finally vindicated by the dis- covery of marrow- and thymus-derived lymphocytes in rodents. This con- ference will provide an opportunity to examine whether the simple schemes derived from the study of immune responses in rodents apply to mammals generally and whether they provide a rational basis for the understanding and treatment of deficiency states in man. Another important topic for consider- ation will be the immunological relationship between the mother and the foetus. We must re-examine the privilege enjoyed by the foetus in the light of sugges- tions that blocking factors may be important in masking the immunity which develops in the mother against paternal antigens. This consideration will no doubt, in turn, lead us to discuss the possibility that blocking factors may also play a part in the mechanism of classic immunological tolerance.

We are all very grateful to the Ciba Foundation, and particularly to Dr Ruth Porter, for having conceived and organized this conference. Those fortunate enough to be enjoying the hospitality of the Foundation will be able to bring each other up to date on the impressive record of experimental work which has accumulated on the normal development of the immune response; they will also no doubt be equally impressed at the end of the meeting by the com- plexities which face clinicians when observing the consequences which follow the failure of normal development.

What precedes clonal selection?

NIELS KAJ JERNE

Basel Institute for Immunology, Basel

This paper is not concerned with the ontogeny of the anatomical and morpho- logical arrangements of the cells that belong to or interact with the immune system, nor with the structure and pathway of the signals to which these cells respond. I shall restrict the term ‘immune system’ to the totality of antibody molecules and of lymphocytes that produce such molecules. I shall assume that all immunoglobulins are antibody molecules, including those that somehow function as receptors on the membranes of lymphocytes. I shall consider that lymphocytes of all sizes, thymocytes, antigen-sensitive cells, T cells, B cells, memory cells, plasma cells, etc., all belong to a dynamic population of clones of lymphocytes that interact amongst themselves and can respond to signals mediated by antigens and antibodies. By ‘dynamic’ I mean that the population is in continuous flux: new signals arise from stem cells, some cells are triggered, others are killed, some cells proliferate, some express their potentialities, others are suppressed, and so on. By the ontogeny of the immune system I shall understand all developments of this system from early embryogenesis until the death of the individual.

THE CLONAL SELECTION THEORY

The clonal selection theory (Burnet 1959) states that all antibody molecules synthesized by one lymphocyte are identical, particularly with respect to the specificity of their-combining sites. More precisely, that one lymphocyte expres- ses only two v-genes, one for the variable region of the light chain and one for the variable region of the heavy chain. Furthermore, the theory postulates that a lymphocyte becomes committed to this restricted synthetic expression prior to the arrival of a fitting antigen. Thirdly, it states that the selection of

Ontogeny of Acquired Immunity Ciba Foundation

Copyright 0 1972 Ciba Foundation

2 N . K . Jerne

precommitted cells by antigen can lead to cell proliferation, and thereby to clonal amplification of selected synthetic potentialities. The clonal selection theory has successfully withstood all experimental attempts to disprove it. In fact, many experiments designed to test the theory have added support to its postulates. It has been shown by Dutton & Mishell (1967), Ada & Byrt (1969) and Wigzell & Anderson (1969) that a given antigen can remove a small fraction of antigen-sensitive lymphocytes from a lymphocyte population in vitro, leaving the remaining cells unresponsive to that antigen but capable of responding to other antigens. Several observations indicate that the cells of a clone breed true; that is, that the cells of one clone all produce the same species of antibody molecule. Thus, the continued production of homogeneous anti- bodies to streptococcal or pneumococcal polysaccharides (Krause 1970; Haber 1972) implies the expansion of large clones of cells secreting the same antibody product. Also, by serial transfer of one clone of antigen-sensitive and antibody- producing cells into successive irradiated recipient mice, Askonas, Williamson & Wright (1970) have shown a continued production through many cell generations of identical antibody molecules. This does not mean that mutant cells, or variant cells, synthesizing the product of a modified pair of v-genes, do not arise in a clone. Studies by Oudin (1969) of idiotypic specificities of antibodies at different times during the course of immunization imply that variant antibody molecules of the same idiotype arise and that some variant cells have selective advantages.

THE ANTIBODY REPERTOIRE

I shall assume the basic postulates of the clonal selection theory to be correct. When an antigen confronts the immune system, it impinges upon a repertoire of available antigen-sensitive lymphocytes. Each of these cells displays receptors of one antibody specificity only. The population of cells represents the repertoire of synthetic capabilities of the immune system at a given point in time. The repertoire will be subject to continuous qualitative and quantitative flux. New items will be added by the entry of differentiating stem cells and by mutation, others will be amplified by immunogenic and other mechanisms leading to cell proliferation. On the other hand, items may disappear from the available repertoire by cell death and by tolerogenic and other suppressive mechanisms. What is needed is an expansion of the clonal selection theory with a set of basic concepts concerning the ways in which the repertoire arises and the elements that govern its maintenance and variation.

Selection among the items of a repertoire requires the prior establishment of

What precedes clonal selection? 3

a repertoire. A fundamental choice is needed between two types of theory, (1) a ‘germ-line’ theory claiming that (in spite of rare somatic mutations and variations) the overwhelming part of the available repertoire results from the expression of v-genes already present in the zygote from which the individual has arisen, and (2) a ‘somatic’ theory claiming that (in spite of the expression of a small number of v-genes already present in the zygote) the overwhelming part of the available repertoire results from the selection of cells expressing mutated or modified v-genes that have originated spontaneously in the de- scendants of stem cells before immunogenic stimulation, or among the cells of a clone responding to a stimulus, or both. It is clear that the available repertoire to some extent arises while the immune system functions, and that therefore the function of the system cannot be studied separately from the ontogeny of its repertoire. A discussion of these matters would be facilitated if we had some knowledge about the size of the repertoire; that is, about the order of magnitude of the number of different antibody molecules that the lymphocytes of one ani- mal can produce.

There are various observations from which an impression of the size of the repertoire may be gained. The following indications suffice for the present discussion. Antibody assays show that the concentration of antibody molecules of a given reactivity in the serum of immunized animals can be several thousand or even several million times higher than their concentration in the serum in normal animals. If we assume that the gamma globulin of normal serum is a mixture of all molecular members of the repertoire, this finding suggests that the repertoire may exceed one million. A similar conclusion can be drawn from experiments by Kunkel (1970) showing that a given human myeloma idiotype occurs with a frequency of less than one in a million among normal serum glob- ulin molecules. Considering the ease with which any rabbit can be induced to produce anti-idiotypic antibodies to the antibody molecules evoked by bacterial antigens in other rabbits (Oudin & Michel 1969; Kelus & Gel1 1968), we could ask whether normal serum may contain antibody molecules reacting with the idiotypic determinants present on other antibody molecules in the same serum. The concentrations a and i molecules per ml of these reactants would be in equilibrium with c complexes per ml. If we permit 1 % of the antibody molecules carrying a given idiotypic determinant to form a complex with a fitting antibody molecule, the relation ai = Kc would permit a = 0.01 K. Considering only antibody molecules of an affinity to this idiotypic determinant corresponding to an equilibrium constant K = 10l2 molecules per ml (or 1.6 x mole), the permissible concentration of this species of antibody molecules would be 1O’O molecules per ml of normal serum. As normal human serum contains about 5 x 1OI6 immunoglobulin molecules per ml, the number of different antibody

4 N . K. Jerne

populations would have to be larger than 5 x lo6. It should be noted that more realistic models assuming the presence of a variety of antibody molecules of different affinity towards any given idiotypic determinant, all lead to estimates of a repertoire higher than 5 x lo6 in man. Such models imply a degenerate network of idiotypic determinants and fitting antibody combining sites : a variety of different antibody molecules would fit any given idiotypic determinant whereas many different idiotypic determinants would fit any given antibody combining site.

It might be thought that the immune system develops tolerance to all idiotypic determinants of its own immunoglobulin molecules so that antibodies to idiotypic determinants present in the same serum do not occur. It should be clear, however, that this would imply an enormous purge of the potential repertoire (Jerne 1960), and would lead to much higher estimates of its size. The concept of a repertoire must be more clearly formulated before attempts can be made to arrive at more meaningful estimates of its size. We must distinguish between the potential repertoire of specificities that could arise given the genetic constitution of the zygote from which the animal develops, and the available repertoire embodied in the cells that can respond to antigens at a given moment in the life-time of the animal. The potential repertoire of animals of one inbred strain may be smaller than that of the entire animal species, because of v-gene polymorphism. The available repertoire at one point in time may be considerably smaller than the total repertoire available to an animal at one time or another during its entire life-time. It seems reasonable to assume that the size of the available repertoire increases during ontogeny and that it will tend towards a maximum in the normally functioning immune system of the adult individual. Furthermore, the question of the relation between T cell repertoire and B cell repertoire needs to be examined.

THE AVAILABLE REPERTOIRE

Rabbits immunized with a strain of Salmonella (Oudin & Michel 1963) or of Bacillus proteus (Kelus & Gel1 1968) all make specific antibodies, but the sets of idiotypic determinants of the antibody molecules produced by any one rabbit differ from those of the antibody molecules produced by any other rabbit. In other words, each rabbit makes use of a different repertoire when responding to the same antigen. Though not inbred, many of these rabbits were of the same allotype. The idiotypes of the antibodies to a given antigen produced by first- generation offspring rabbits were no more similar to those occurring on the antibodies produced by a parent than to those occurring on the antibodies pro-

What precedes clonal selection? 5

duced by unrelated individuals (Kelus & Gel1 1968; J. Oudin & G. Bordenave, personal communication 1971). As an idiotypic determinant represents the antigenic properties of a given pair of variable regions of the polypeptide chains of an antibody molecule, it follows that the v-genes expressed by the responding cells of one rabbit are different from those expressed by the cells of another rabbit responding to the same antigen. These results not only demonstrate the enormous plasticity of the immune system in its ability to use different v-genes for producing different antibody molecules of similar specificity, but they also show that an individual makes use of only a small part of the potential repertoire which its inherited v-genes could have given rise to, A. R. Williamson & W. Kreth (personal communication 1971) have found that individual CBA mice, responding to a hapten (2,4-dinitrophenol, DNP, or 4-hydroxy-3-iodo-5-nitro- phenyl acetic acid, NIP) attached to bovine gamma globulin, each produce more than a hundred different antibodies to the hapten and that the two sets of such antibodies to the same hapten produced by two mice are almost entirely different, so that there will be hardly more than one or two molecular species of antibody that occur in both sets. This experimental demonstration reinforces the conclusion that individual animals make use of widely differing repertoires when responding to an antigen, and that this is true even for the genetically virtually identical animals of the same inbred strain of mice, reared under the same conditions.

REPERTOIRE SUPPRESSION

How are we to interpret the findings (1) that the antibody repertoire available to an individual animal is very large (e.g. > loe), and (2) that each individual responding animal makes use of only a small part of the potential repertoire permitted by its germ-line genes? Two or three possibilities present themselves. The population of lymphocytes may, as it arises, express the entire potential repertoire. In that case, either the immune system does not make use of more than a small part of its available repertoire when responding to an antigen, or the repertoire is reduced drastically by suppressive mechanisms, leaving dif- ferent available repertoires in different individuals. On the other hand, the entire potential repertoire may never be expressed in one individual, but only a sample of it. Or, thirdly, the repertoire actually used by a responding animal may be that which is left over after the expression by its lymphocytes of part of the potential repertoire, after a reduction of this expressed repertoire by suppression, and after a further reduction to the set of cells that antigen actually succeeds in stimulating. Various types of suppression are known. A rabbit

6 N . K. Jerne

of allotypic genotype, say, a,/a, produces antibody molecules of allotype a, as well as antibody molecules of allotype a3. By immunizing the mother with globulin of paternal allotype, or by injecting anti-paternal-allotype antibody neonatally, the expression of the paternal allotype can be suppressed (Dray 1962; Mage 1967). This suppression lasts for many years and shows that practically half the lymphocytes that arise (those attempting to express this allotype) are suppressed. From a large variety of experiments in rabbits and in mice by Jacobson, Herzenberg, Riblet and Herzenberg (1972) it may be con- cluded (1) that the suppressed cells committed to the expression of an allotype are probably not eliminated, since the production of immunoglobulin of this allotype is resumed on transfer of cells from a suppressed animal to an irradiated recipient animal, and (2) that continued allotype suppression is probably effected by the presence of anti-allotypic T cells.

I wish to stress this suppressive effect involving the antigenic properties of anti- body molecules, because these may play an important role in the development, maintenance, and shift of the repertoire available to an individual. We might generalize, tentatively, that both certain concentrations of antibodies, as well as the emergence of certain T cells, exhibiting antibody combining sites directed against antigenic determinants of antibody molecules (allotypes, idiotypes), can suppress the ‘expression’ of such molecules by B cells. If T cells can suppress such B cells, the target of this type of suppression would seem to be the antigenic determinants of the receptor molecules of these B cells, since these are the only targets that distinguish different B cells. Furthermore, it would seem that these targets are recognized by the combining sites of the T cell receptors. It is conceivable that the expression of many idiotypic determinants is normally suppressed in this same way, and that the available repertoire is correspondingly reduced. Conversely, we may conclude that antibodies (or B cell receptor molecules), by their allotypic determinants, suppress T cells of certain spe- cificities that would emerge under conditions of allotype suppression. This could be taken as an example of induction of tolerance by antigens, including idiotypic antigenic determinants. Thus, Iverson & Dresser (1970) have shown that a mouse myeloma protein can be made immunogenic by attachment of hapten to the molecule and can provoke the formation of anti-idiotypic antibody in normal mice of the inbred strain in which the myeloma had arisen, whereas the injection of unaltered myeloma protein into such mice leads to tolerance to its idiotypic determinants.

The above examples (which could be multiplied) show that lymphocytes committed to the expression of a given antibody molecule A can be suppressed (1) by other antibodies, either humoral or functioning as receptors on other lymphocytes, possessing combining sites directed against the antigenic deter-

What precedes clonal selection? 7

minants of A, and (2) by other antibodies possessing antigenic determinants fitting the combining sites of A. It therefore seems likely that antibodies arising from antigenic stimulation of a set of lymphocytes suppress other lymphocytes, and that the entire system represents a complex interacting ‘network’ of expression and suppression of potentialities. The available repertoire would represent the balance resulting from this continuing process.

In these considerations, I have left out all the many forms of induced toler- ance, as well as other known examples of suppression of potentialities, such as the fact that immune responsiveness to a given antigen can be suppressed by passive IgG antibody directed against the same antigen, and the finding by Askonas & Williamson (1972) that established clones of cells producing a given antibody can prevent the same antigen from stimulating other cell clones. Another example is self-tolerance which implies (most obviously in F1 animals that are heterozygous for histocompatibility antigens) that part of the potential repertoire of the parental genes is suppressed. All in all, it is clear that the immune system exerts self-control by suppressive mechanisms, and that these suppressive actions restrict the available repertoire.

SOURCE OF THE REPERTOIRE

If, conceptually, we were to place the potential repertoire in the germ-line - that is, if we assume that all structural v-genes for the antibodies that an individual may potentially express are already present in the DNA of the zy- gote - then we would be tempted to conclude that only a small fraction of these are actually expressed in the available repertoire of an individual. Other- wise, genetically identical or related animals would be expected to produce, at least in part, identical antibody molecules to the same antigen.

This repertoire restriction appears to make a germ-line hypothesis untenable. We must admit that the number of v-genes required to encode an available antibody repertoire is already uncomfortably large, if it has to be located in the germ-line genome. The situation becomes worse if we consider that a germ-line theory would require the presence, in the genome of the zygote, of the entire potential repertoire, which is far larger. A collection of genes can be kept intact in evolution only if each gene is used and if its absence impairs survival to some degree. It is hard to believe that the presence of every gene in the large set that is required to encode the potential repertoire is essential. We cannot be quite certain of this, however, for even if a given light chain v-gene is not expressed in combination with any of several heavy chain v-genes, it may find expression in combination with other heavy chain v-genes. In spite of this consideration, the

N . K. Jerne

fact that the potential repertoire is much larger than the available repertoire suggests that many v-genes can be dispensed with. As unessential genes will be lost by mutation, it seems that we must look for a different genetic solution to the problem of the repertoire, a solution that must start out with a number of essential v-genes for antibodies in the zygote that is much smaller than the number of such genes expressed by the lymphocyte populations present in a set of genetically identical immunocompetent animals. The clue to such a solution is the astonishing difference between the repertoires available to dif- ferent individuals that are genetically similar and have grown up under similar circumstances. This suggests an element of randomness either in the choice of cells responding to a given antigen in an animal, or (more likely) in the ontogeny of the repertoire. This could be supplied by a somatic mutation mechanism or by some other somatic gene modification mechanism (Gally & Edelman 1970) that would potentially permit a repertoire enormously much larger than the available repertoire actually arising in a given individual animal.

BOUNDARY CONDITIONS

To obtain diversity of repertoire by somatic mutation there must be selection pressures which favour lymphocytes that have acquired a suitable v-gene mutation over non-mutant lymphocytes of that clone. If lymphocytes that are suppressed (as a result of the phenotypic expression of the pair of v-genes to which they are committed) proliferate in some restricted manner, then mutant cells from such a clone could escape suppression by expressing a different antibody combining site or different idiotypic determinants on their receptor molecules and on the globulins that they can secrete. The many forms of suppression prevalent in the immune system may thus engender repertoire diversity, and this process may be concomitant with the functioning of the system throughout life.

This hypothesis would be more complete if it could account not only for the final degree of diversity of the repertoire available to a mature animal, but also for the starting point and the initial stages from which the repertoire must expand. The specificities of the antibodies actually encoded in the small set of germ-line v-genes and expressed by differentiating stem cells should be iden- tified, as well as the factors impinging initially and favouring the proliferation of mutant cells. A theory postulating a fundamental orientation of the repertoire towards a set of histocompatibility antigens of the animal species has been formulated (Jerne 1971). It assumes that the antigens of this set that are

What precedes clonal selection? 9

present on the surface of the lymphocytes of an animal provide the initial signals for suppression of cell clones committed to express v-gene pairs for antibodies directed against these antigens, thereby favouring mutant cells. This suggestion tries to account also for the phenomenon of allo-aggression and for the mapping of dominant immune responsiveness genes near the major histocompatibility locus.

Antigenic determinants other than histocompatibility antigens can be imagined to act in the early modulation and further expansion of the repertoire. We might consider the large number of idiotypic determinants of the maternal antibodies, which are among the earliest foreign antigens encountered in on- togeny. Different antibody combining sites are associated with different idio- typic determinants, and the immune system is capable both of producing antibodies and of developing tolerance to such determinants. It follows that any antibody combining site may fit some idiotypic determinant on another antibody molecule, and (perhaps only incidentally) may also fit some other antigen not belonging to the system itself. Modulation of the repertoire through idiotypic determinants would make the system more autonomous and self-contained and less dependent on encounters with foreign antigens. The final boundaries of the repertoire in a mature animal may reflect the optimal complexity of a functional network of antibody combining sites and idiotypic determinants. This boundary would depend also on the life-time and the num- ber of cell generations allotted to a clone derived from one stem cell. Also, it seems rather obvious that a final limitation is imposed by the size of the system. Thus, a human individual having 10l2 lymphocytes and lozo circulating antibody molecules should be able to develop a larger repertoire than a mouse, for which these numbers are a thousand times smaller. These ideas suggest that the final repertoire available to an animal is only one of many potential possibilities, and that it arises by a Darwinian process on a reduced time scale: ontogeny mimics phylogeny.

References

ADA, G. L. & BYRT, P. (1969) Nature (Lond.) 222, 1291 ASKONAS, B. A. & WILLIAMSON, A. R. (1972) Nature in press ASKONAS, B. A., WILLIAMSON, A. R. & WRIGHT, B. E. G. (1970) Proc. NatI. Acad. Sci. U.S.A.

BURNET, F. M. (1959) The Clonal Selection Theory of Acquired Immunity, Cambridge University

DRAY, S. (1962) Nature (Lond.) 195, 677 DUITON, R. S. & MISHELL, R. I. (1967) J. Exp. Med. 126, 443

67, 1398

Press, London

10 Discussion

GALLY, J. A. & EDELMAN, G . M. (1970) Nature (Lond.) 227, 341 HABER, E. (1972) Ann. N.Y. Acad. Sci. 190,285-304 IVERSON, G. M. & DRESSER, D. W. (1970) Nature (Lond.) 227, 274 JACOBSON, E. B., HERZENBERG, LA. , RIBLET, R. & HERZENBERG, L. A. (1972) J. Exp. Med.

JERNE, N. K. (1960) Ann. Rev. Microbiol. 14, 341 JERNE, N. K. (1971) Eur. J . Immunol. 1, 1 KELUS, A. S. & GELL, P. G . H. (1968) J. Exp. Med. 127, 215 KRAUSE, R. M. (1970) Adv. Immunol. 12, 1 KUNKEL, H. G. (1970) Fed. Proc. Fed. Am. SOC. Exp. Biol. 29, 55 MAGE, R. G . (1967) Cold Spring Harbor Symp. Quant. Biol. 32,203 OUDIN, J. (1969) Eehringwerk-Mitfeilungen 49, 77 OUDIN, J. & MICHEL, M. (1963) C.R. Hebd. Seances Acad. Sci. D 257, 805 OUDIN, J. & MICHEL, M. (1969) C.R. Hebd. Seances Acad. Sci. D 268, 230 WIGZELL, H. & ANDERSON, B. (1969) J. Exp. Med. 129,23

135, 1163

Discussion

Billingham: Dr Jerne mentioned the serial transfer of committed clones of cells from one irradiated mouse to another. Can this process be continued indefinitely?

Jerne: B. A. Askonas, A. R. Williamson and W. Kreth (personal communi- cation 1971) have serially transferred the same clone of antibody-forming cells into seven or eight successive mice, and they calculate that this is likely to correspond to 80 or 90 cell generations. After that the clone dies out during further transfers in which new clones will make their appearance. It seems that as long as the original clone is active, it suppresses the responses, to the antigen, of other cells among the transferred population. The 80 to 90 generations remind one of Hayflick’s phenomenon (Hayflick & Moorhead 1961): the limited lifespan of a clone of diploid cells in culture.

Gowans: Is there any evidence that large animals have bigger repertoires of responses than small animals? And are there any species which have rather few lymphocytes and surprisingly large repertoires?

Jerne: I do not know of any good evidence on the first point. On your second question, L. Du Pasquier (personal communication 1971) working in our Institute in Basel has studied the response of amphibian larvae to various antigens and finds, for example, that a tadpole at a stage of development when it contains less than 200 OOO lymphocytes can produce specific antibodies to each of several bacteriophages. At that stage, the tadpole spleen contains about 10 OOO lymphocytes. Cultures of about 30 OOO tadpole spleen cells will respond to antigen in vitro. The antibody produced was found to be specific but of low

What precedes clonal selection? 11

affinity. It neutralizes bacteriophage, for instance, in a reversible manner. Mitchison: I accept the drift of the argument that the number of clones is very

large, but Dr Jerne’s exposition may have overestimated it for two reasons. I think Kunkel’s data may underestimate the frequency of idiotypes. More recent measurements by Iverson (1970) and earlier estimates by Hurez et al. (1968) give frequencies of idiotypes in serum that are higher by at least an order of magnitude. Iverson’s values were obtained in mice by inhibition of binding of antigen labelled with lz5I. In Seligmann’s estimates (see Hurez et al. 1968), there was a good deal of variation from one idiotype to another, but some of them were in the range estimated by Iverson.

The second way in which I would question these numbers is that you assume that the anti-DNP and anti-NIP responses are typical in the degree of variation of antibodies produced. The fact that DNP and NIP are popular haptens is no coincidence. They are popular for just the reasons which would lead one to expect there to be a very large number of reactive cells, and I suspect that other determinants, such as sugar determinants or amino acid determinants, might induce much lower numbers of clones. In the last year or two, as people have focused attention on restricted antibody responses, these have turned out to be more frequent than one would have anticipated from the data obtained with nitrobenzenes and their derivatives.

Jerne: I have some data on antibody to P-galactosidase that are relevant to this discussion. If lymphoid cells are mixed with P-galactosidase, cells carrying receptor molecules which recognize an antigenic determinant of this enzyme molecule will bind the enzyme. After washing, those cells can be counted by localized enzymic activity. Rotman & Cox (1971) have worked out this tech- nique and find, in their mice, that the fraction of P-galactosidase-binding lym- phoid cells is about 40 per million. A P-galactosidase molecule must have several or many different antigenic determinants. F. Melchers (personal communica- tion 1971) in our Institute works with E.coli mutants producing inactive P-galactosidase molecules that become enzymically active when antibody to wild-type P-galactosidase is added. There is evidence that this activation in- volves only one particular antigenic determinant of the enzyme molecule. By staining populations of normal mouse lymph node cells with the mutant enzyme, Melchers finds that the fraction of cells carrying receptors fitting the particular antigenic determinant is of the order of one per million, or less. After immuni- zation with wild-type (3-galactosidase, and application of isoelectric focusing to the serum, the band pattern of antibodies that can activate mutant P-galactosi- dase can be developed by staining with this mutant enzyme. Melchers finds that similar mice of one inbred strain produce sera showing different band patterns. Many mice produce patterns showing about three different antibodies to the

12 Discussion

particular antigenic determinant of the enzyme, but the data so far collected indicate that the antibodies produced by different individuals are obviously different. These results seem to support estimates of repertoire size that are larger than implied by Dr Mitchison.

Fudenberg: If a particular idiotype reacts with a number of different anti- bodies, would you not expect that this would lead to the removal of the anti- gen-antibody complexes from the circulation, so it wouldn’t be evident, using the same animal?

Jerne: This is the question: do we become tolerant to all the idiotypes occur- ring on the antibody molecules we produce? After immunogenic stimulation by any antigen, particular antibodies carrying certain idiotypic determinants expand to large concentrations. I would imagine that we become tolerant to those idiotypes, as Iverson & Dresser (1970) have shown by injectinga particular mouse myeloma protein into mice. But as long as cells expressing immuno- globulins of a certain idiotype are not stimulated, the concentration in the blood of molecules carrying this idiotype may remain low enough to permit coexis- tence with molecules carrying anti-idiotypic combining sites.

Mitchison: The idea that things go clonally before as well as after the immune response has always been an assumption, and the most recent evidence seems to be against it. The experiments on the selective removal of cells sensitive to a particular determinant or a particular antigen by passage through a column have always worked well with primed cells but less well with unprimed cells. In fact, under the conditions of those experiments, you may be removing quite a large fraction of the population, leaving plenty of room for the possibility that cells have many specificities. The fact that immunoglobulin class and allotype specificities, which sort themselves out clonally after stimulation, are now known not to have sorted themselves out before stimulation, raises the possibility that the same may be true of variable sequences as well (Greaves & Hogg 1971).

Jerne: How long before stimulation do you mean? Were the cells already at an antigen-sensitive stage of differentiation?

Mitchison: Yes. Jerne: So you are leaving open the possibility of pluripotentiality of a cell at

Mitchison: Yes. Humphrey: If one examines the cells from spleen or lymph node of a mouse

which can bind radioactive antigens in vitro, one finds that there are relatively many (of the order of one per thousand) which bind a few molecules but very few which bind many molecules - say 5000 or more. I have looked at three different systems in normal mice and have tried to estimate the frequency of cells which bind 5000 or more molecules (assuming for this purpose a molecular

an early antigen-sensitive stage?

What precedes clonal selection? 13

weight of about lo5 daltons). It works out at about one cell in 100000 for keyhole limpet haemocyanin or spidercrab haemocyanin, for tetanus toxoid, or for a synthetic polypeptide p(T,G)-PA-pL heavily substituted with iodine in the tyrosine groups (TIGAL). Each of these antigens must carry a variety of antigenic determinants. If one can assume that a cell with receptors specific for a given antigen is one which can bind it firmly, these results suggest that there are at least 100 OOO different kinds of lymphoid cells in a normal mouse with different receptors. Of course it is difficult to be sure that the mice had never encountered the antigenic determinants involved. This is practically unanswer- able unless the experiment is repeated in germ-free mice. But the figures don’t greatly disagree with Dr Jerne’s thesis, even as modified by Dr Mitchison. However, the agreement is less obvious if one takes into account the more numerous cells which bind less antigen.

Gowans: The definition of an antigen-binding cell is obviously arbitrary. How does the frequency of such cells change as the scoring is allowed to include cells with progressively smaller numbers of grains?

Humphrey: It is more like one in 10000 if you take 50 grains as the limit instead of more than 200. With five grains as the limit the number would be one in 500 to one in 2000. These numbers are of the same order as those reported by others.

Fudenberg: May I raise a point of heresy? Let us draw an analogy between the virgin lymphocyte and an ovum. Depending on which sperm hits that ovum out of the 10 million that come across, you get an end product which has certain characteristics. At the same time (unless within a finite period another sperm hits, in which case you get twins) you get a unique product. Can it be that every lymphocyte in the virgin animal is totipotential, but as soon as it is hit by an antigen, a shut-off process occurs so that it is no longer susceptible to other antigens, and its proliferation will result in a clone of cells sensitive only to that antigen? Are there any experiments to exclude this possibility?

Jerne: As you pointed out yourself, this is heresy! Dr Humphrey said that we never know whether an animal has already encountered antigens carrying determinants similar to those of the antigen we choose for experimentation. In the spleen of a normal mouse you can find about 100 plaque-forming cells producing antibodies to sheep red cells, or close to one cell in a million. Nordin (1968) has shown that germ-free mice living on an antigen-free diet have practi- cally the same proportion of cells in their spleens forming plaques against sheep red cells. This suggests that this potentiality arises without previous stimulation by foreign antigens.

As for Dr Mitchison’s remarks about DNP and NIP (and we may include sheep red cells), it is curious that practically any antigen you choose to work

14 Discussion

with turns out to be a peculiar antigen! Perhaps, from the data of Askonas, Williamson and Kreth, and of Melchers, that I mentioned earlier in this discus- sion, one could arrive at an estimate of the repertoire in mice of different anti- body molecules reacting with a given antigen, no matter how peculiar. If we also had an estimate of the percentage of spontaneous myelomas producing an immunoglobulin reacting with this antigen, we might assume, as a first approxi- mation, that the particular repertoire estimated is the same percentage of the total repertoire.

Lachmann: I’d like to mention the ‘idiotype paradox’. If one takes the ex- treme view that every individual antibody molecule carries its own idiotypic determinant (or even several, since the anti-idiotype antibodies precipitate) the number of potential anti-idiotype antibodies must be equal to or even greater than the number of potential idiotypes, and one runs into the problem of where all the antibodies to NIP, to DNP, to sheep red cells and so on come from, if all antibodies are actually directed against the v-region of y-globulin. There are two ways of resolving this paradox, and I think Dr Jerne is using both. One is to say that it is true that all antibodies are indeed directed against the v-regions of y-globulin, and that the antigenic repertoire is determined by cross-reactivity with the universe of v-regions. The alternative explanation is that the idiotype system is greatly degenerate, and that a very large number of ‘different’ antibody molecules (as defined by combining site) share one idiotype. This second ex- planation is not compatible with a control mechanism based on the suppression of anti-idiotype, because each new ‘antibody’ would switch off many clones and that must lead to agammaglobulinaemia!

Jerne: I agree with your general exposition of this argument, but not with your conclusion. Suppression and stimulation are likely to depend on the equilibrium constants of combining sites and determinants occurring on cell receptor molecules. What you said first was just what I wanted to point out: that the whole system may be self-contained by the interaction of combining sites and antigenic determinants of molecules belonging to the system itself, with only incidental interaction with outside antigens.

Lachmann: I agree that the first resolution of the paradox is elegant even if at first sight it seems implausible. With respect to the second explanation, I think your argument is based on the failure to detect anti-idiotype in normal serum and this would surely imply that the level of detection and the level of switching off must be the same. Therefore, if a large number of antibody combining sites share one idiotype, and if the generation of a new antibody directed against this idiotype switches off all antibodies that can react with it, this should eventually eliminate the whole y-globulin system.

Jerne: That was exactly my argument for assuming idiotypic determinants and

What precedes clonal selection? 15

corresponding antibody combining sites both to be present in the blood simultaneously, because a removal of either might purge the system to an im- possible extent.

References

GREAVES, M. F. & HOGG, N. M. (1971) Prog. Zmmunol. 1, 111-126 HAYFLICK, L. & MOORHEAD, P. S. (1961) Exp. Cell Rex 25,585-621 HUREZ, D., MESHAKA, G., Mmsco, C. & SELIGMANN, M. (1968) J. Zmmunol. 100, 69-79 IVERSON, G. M. (1970) Nature (Lond.) 227, 273 IVERSON, G. M. & DRESSER, D. W. (1970) Nature (Lond.) 227,274 NORDIN, A. A, (1968) Proc. SOC. Exp. Biol. Med. 129, 57 ROTMAN, B. & Cox, D. R. (1971) Proc. Natl. Acad. Sci. U.S.A. 68, 2377

Immunological maturation in the foetus : modulation of the pathogenesis of congenital infectious diseases

ARTHUR M. SILVERSTEIN

The Wilmer Institute, The Johns Hopkins University School of Medicine, Baltimore

It has become increasingly clear during recent years (Sterzl & Silverstein 1967; Solomon 1971) that the acquisition of immunological competence during mammalian ontogenesis may occur long before birth in some species. In the hu- man, active immunological response by the foetus to transuterine infection has provided new tools for assessing the role of microorganisms in the development of congenital malformation (Alford er al. 1968; Woodside & Mitchell 1968). Of no less importance, however, is the opportunity provided by the study of natural and experimental infections of the mammalian foetus to identify those aspects of host-parasite interaction which may contribute to the pathogenesis of certain infectious disease processes.

It is the purpose of this paper to explore some of the mechanisms whereby the immunological status of the host may profoundly affect the pathogenicity or apparent virulence of an infectious agent, in terms of intrauterine infection within a foetal host in immunological transition. The attainment of immuno- logical competence by the developing foetus will be seen in some instances to ‘turn off’ an infectious disease process, in other instances to render highly pathogenic an otherwise innocuous agent, and in certain cases to contribute to a change in the very nature of the disease produced by a given agent. Finally, some recent data will be discussed which bear upon the possible involvement of immunological tolerance in foetal infection and disease.

ONTOGENESIS OF THE IMMUNE RESPONSE

The principal phenomenological characteristics of immunological differen- tiation have been reviewed extensively elsewhere (Sterzl & Silverstein 1967; Solomon 1971) and need only be summarized here. In some species, immunol-

Ontogeny of Acquired Immunity Ciba Foundation

Copyright 0 1972 Ciba Foundation

18 A . M . Silverstein

ogical competence is substantially achieved only after birth (most notably among the common laboratory rodents). In such species as the sheep, cow, rhesus monkey, and man, maturation of most immunological capabilities occurs in ufero, and the newborn is more-or-less well prepared to employ his own immunological resources on his own behalf. While adequate data are lacking for most species, the rabbit furnishes one example of a maturation of the immunological apparatus occurring in the immediate perinatal period.

The second point of significance to this discussion is that the immunological apparatus does not appear to mature as a single all-encompassing event which renders the young animal thenceforth capable of active response against all types of antigens. On the contrary, the process presents as a series of discrete events involving the attainment of competence first to one antigen, then to another, and so on in a stepwise fashion finally to encompass the universe of potential immunogens. Each species thus far studied appears to manifest a unique maturational sequence, involving a timing and hierarchical order of antigens peculiar to the species or, in the case of inbred mice, peculiar even to the strain. While the fundamental nature of the differentiative events leading to competence to respond specifically to a given antigen is not understood, the timing for each antigen would appear to be under extremely precise control. In some species (sheep, pig, opossum), the first active responses appear ex- tremely early in development when the total lymphoid mass of the animal is extremely limited; competence to other antigens, in the lamb at least, appears only after birth in a host whose lymphoid population is many orders of magni- tude larger.

Another significant characteristic of the earliest active immunological responses in hitherto immunologically virgin animals is the degree of maturity of these responses. In our own experience with antibody formation (Silverstein et al. 1966) and skin allograft rejection (Silverstein et al. 1964) in the foetal lamb, and with antibody formation in the foetal rhesus monkey (Silverstein et al. 1970), we have been repeatedly impressed by the absence of any sign of im- maturity or inadequacy on the part of the foetus. Thus, once the young animal is able to respond at all to a given antigen, its response appears to be adult-like with respect both to heterogeneity of immunoglobulin product and to the appearance of both humoral and cellular immune components (Osburn & Silverstein, unpublished data). These observations may require some modi- fication in view of the demonstration that the newborn rabbit (Sterzl & Trnka 1957) and piglet (Tlaskalovi et al. 1970) may require appreciably higher anti- genic doses for a given response than does the adult of each of these species.

The implications of the several ontogenetic observations described above with respect to the currently tantalizing question of the generation of immuno-

Immunological modulation of congenital infections 19

logical diversity have been discussed at length elsewhere (Silverstein & Prender- gast 1970).

IMMUNOLOGICAL SUPPRESSION OF PATHOGENICITY

Since the earliest days of immunology, it has been well established that enhancement of both specific and non-specific host defence mechanisms would reduce the apparent pathogenicity of an agent, whereas interference with these responses would render the agent more noxious. The latter observation has received striking confirmation from the study of the consequences of human and experimental immunological deficiency states (Bergsma 1968). The former process may be considered the ‘normal mode’ of immunity in infectious diseases, and offers a superabundance of examples. It may be well, however, in the context of a discussion of congenital infectious diseases, to give yet another example which will also serve to underline again the dangers that may accom- pany the use of live virus vaccines during pregnancy.

Bluetongue virus of sheep in the wild form results in the severe involvement of the vascular endothelium of the adult animal (Moulton 1961). After attenuation by repeated egg passage, the virus is rendered non-pathogenic and finds ex- tensive use as a highly efficacious vaccine. Vaccination of the pregnant ewe, however, results in transplacental passage of the virus and severe central nervous system disease of the foetus with necrotizing encephalitis (Young & Cordy 1964; Osburn et al. 1971a) and retinopathy (Silverstein et al. 1971). Depending upon the time of infection and its intensity, the affected foetuses may come to term blind and with varying degrees of hydranencephaly or porencephal y.

But most significant is the fact that the foetus shows susceptibility to this live ‘attenuated’ virus only during the first half of gestation, during which time the virus is found to replicate freely and to inflict its damage upon susceptible cells. After midgestation, the foetal host no longer supports the damaging viral disease process, but rather appears to find the virus as innocuous as does the adult animal. While some degree of viral replication still occurs in the foetus as it does in the adult, active plasma cell differentiation is now observed to occur and is followed thereafter by the appearance of circulating virus-neutral- izing antibody and clearance of the virus (Osburn et al. 1971b). It would appear, therefore, that one of the chief elements in the shift from high patho- genicity in the young foetus to lack of pathogenicity in the older foetus is the acquisition by the developing host of immunological competence to defend itself against the viral agent. As may be true of many foetal infectious processes,

20 A . M. Silverstein

however, any simple explanation may be complicated by other developmental variables, such as the possible maturational loss of susceptibility to viral infection and damage by brain and retinal cells, the contribution of which cannot adequately be evaluated.

THE IMMUNOLOGICAL ACTIVATION OF PATHOGENlClTY

By far the best-studied example of a disease process in which the triggering of an immune response by an otherwise innocuous agent results in extensive pathology is that of viral lymphocytic choriomeningitis of mice. This model has been so widely described that only its major aspects need be repeated here. In brief, the virus is found to infect and multiply within the brain of the im- munologically incompetent foetal or neonatal host (Hotchin 1962; Mims 1966) or the immunosuppressed adult (Nathanson & Cole 1970; D. H. Gilden, G. A. Cole, A. A. Monjan & N. Nathanson, personal communication 1971) without the production of pathological lesions. Infection of the immunological- ly normal host (Hotchin 1962) or cellular reconstitution of the immunologically defective viral carrier (Cole et al. 1971) results in fatal immunopathological disease, with viral antigens serving as the target for immunological attack and host cells and tissues suffering the indirect consequences.

A somewhat comparable if less well understood analogue of the situation may exist in human congenital infectious processes. Congenital syphilis in the human foetus is predominantly a chronic inflammatory disease characterized by widespread lymphocytic and plasmacytic infiltrates and a variety of secondary changes (Silverstein & Lukes 1962). Little evidence exists that the Treponema pallidum exerts a direct deleterious effect upon the human cell either by release of toxic substances or by specific cytopathogenicity; rather, most aspects of the disease process seem ascribable to the consequences of host response to this agent (World Health Organization 1970).

Congenital syphilis in the human is seldom seen before the fifth or sixth month of gestation, an observation often ascribed to changes in the placenta at that time which permit the treponeme to pass from the infected mother to the foetus. It is of no little interest, therefore, that the organism has occasionally been found in foetal tissues before this age in the absence of the inflammatory disease so characteristic of congenital syphilis. The implications of a benign presence of Treponema pallidum within the tissues of the very young foetus prompted us to suggest (Silverstein 1962) that the organism might cross the placenta during early gestation more frequently than was supposed, but at a time when the absence of pathological response would prevent both foetal

Immunological modulation of congenital infections 21

embarrassment and subsequent abortion as well as discourage the pathologist from undertaking the difficult search for the organism itself. This led to the speculation that the gestational time at which the full disease is first seen may only mark the development of foetal ability to mount an adequate host response to the organism, and that a significant component of the pathogenicity of this disease might be immunological in nature.

THE IMMUNOLOGICAL MODULATION OF PATHOGENESIS

The term modulation is employed here to indicate a marked change in the basic nature of a disease process as a result of maturation of the host’s im- munological capabilities. Thus the same agent may produce in the immunolog- ically competent foetus a qualitatively different disease process from that seen in the immature host. This situation is perhaps best exemplified by the dif- fering pathological pictures of rubella infection of the human in early and late gestation. The disease of the first trimester is one of multiple system congenital anomalies (Singer et al. 1967) in the absence of an inflammatory response (Tondury 1962; Tondury & Smith 1966). The rubella virus shows few signs of marked cytopathogenicity, but rather exerts its effect by interfering with cell di- vision at critical stages of organogenesis in the very young foetus to produce dysgenic malformations (Rawls & Melnick 1966). Later in gestation the disease takes a completely different form, now presenting as a group of widely distrib- uted chronic inflammatory lesions involving lymphocytes, plasma cells, and the formation of germinal centres in organized lymphoid tissue (Singer et al. 1967). Here the host is contributing substantially to the pathology, with the resulting hepatitis, leptomeningitis, iridocyclitis, and otitis, among others, responsible for a completely different set of signs and symptoms. Of course, most congenital rubella appears as a combination of these two distinct disease entities, but at least one study of late gestation rubella (Hardy et al. 1969a) points up well the purely inflammatory nature of this late gestation process.

Another less widely appreciated instance of the changing nature of a disease process with the changing status of the foetal host occurs in congenital brucel- losis of the foetal lamb. Infection of the pregnant ewe or of the foetus directly prior to about 90 days of gestation results in a chronic disease characterized by reticuloendothelial hyperplasia with granuloma formation, with no significant lymphocytic or plasmacytic contribution to the inflammatory event and no activation of organized lymphoid tissue (Osburn 1968). Infection of the foetus after this time leads to a pathological picture characterized by severe chronic inflammatory reactions with now a lymphoreticular hyperplasia involving