|
Chapter 1
(Excerpt)
The Beginnings
By tradition, stories are best begun with the phrase "It was a dark and stormy night..." but I cannot now remember whether the colony-stimulating factor (CSF) story began for me in the morning or afternoon. It was certainly not night, and I doubt strongly whether it could have been raining.
For my temerity in wishing to do cancer research in an immunological institute, The Walter and Eliza Hall Institute of Medical Research in Melbourne, Australia, I had been exiled from the main Institute building in 1958 to some small laboratories in the Institute's animal house. This was a noisome building in an adjacent veterinary precinct, accessible from the Institute by a long tunnel under the Royal Melbourne Hospital, which then passed through the laundry block and around the hospital rubbish dump. For someone highly allergic to all animals, particularly mice, it was an unfortunate location in which to spend eight years, but despite this, our little group was happy and productive, if somewhat isolated and disregarded.
We were funded entirely by the Anti-Cancer Council of Victoria and our self-chosen scientific brief was to attempt to understand the nature of the leukemias and the mechanisms involved in the development of these diseases. This had considerable bearing on the curious manner in which we were to undertake some of our work on the CSFs. The particular approach we adopted was to explore the possibility that, like tumors of endocrine target tissues, the leukemias might arise at least in part as a consequence of an imbalance in the regulatory factors controlling blood cell proliferation. This notion was to prove to have some validity, but the approach had an obvious initial requirement—to establish the existence and nature of these unknown regulatory factors that were being postulated as controlling blood cell formation.
When these studies began in the mid-1950s only one regulator of hematopoiesis had been identified with any sort of experimental validation. This was erythropoietin, an agent believed to regulate red cell formation. Even for erythropoietin, the evidence supporting its reality and possible mode of action was minimal by today's standards.
One of the most promising alternative leads in the early 1950s was that the thymus seemed to be essential for the development of lymphoid leukemia in mice. The thymus might not only be producing lymphocytes but might also act as a regulator of lymphocytes throughout the body. Some of my early work, which others were never able to repeat, suggested that thymus extracts, when injected into baby mice, could elevate blood lymphocyte levels. This work led me into a decade of exploring the consequences of thymectomy and an analysis of how cell production in the thymus appeared to be regulated. This latter question remains to be properly resolved, and in the early 1960s, with the necessity of using whole animal models, it was certainly not feasible to dig very deeply into the possible operating mechanisms. My small group persisted, but we were very conscious of the fact that whole animal work had severe limitations and that we were confronted by a technically impassable barrier. We were eager for some simpler system that would allow hematopoietic regulators to be characterized and were ready with a long list of questions for resolution using such a system. Indeed, we were in a thoroughly primed but frustrated state.
A visit in 1964 by a colleague was to make a dramatic impact on the direction of our future activities. Ray Bradley, who was working across the street from the Institute at the University of Melbourne, came to our laboratories one day with a small metal tray. He and I had collaborated in former years on a number of studies exploring the repopulation of thymus and spleen grafts, but what he was carrying was something radically different. He had brought along for my enlightenment and comments a set of carefully handled glass petri dishes containing semisolid agar cultures of mouse bone marrow cells in which extraordinary colonies had developed.
Ray Bradley's astonishing bone marrow colonies.
I challenge anyone on first seeing such colonies not to be astonished and intrigued. These colonies are three-dimensional populations of cells of wonderfully variable shapes and sizes, and look like galaxies as approached by a fast-moving spaceship. Their beauty continues to fascinate me, even after 35 years.
Ray Bradley in later years.
Bone Marrow Cultures: A Background
The context in which the arrival of these colonies occurred underscored their dramatic impact....
Ch. 1 | Ch. 2 | Ch. 8 | Ch. 9 | Ch. 16 | TOP
Chapter 2
(Excerpt)
Project Goals and Approaches
Casting aside the depressing possibilities that colony formation might be due to viral transformation or to the trivial action of certain metabolites, Bill Robinson and I chose to take the most optimistic view that the colony-stimulating activity being supplied to the cultures by added cells or conditioned medium was in fact a mandatory, tissue-specific physiological regulator of granulocytes and macrophages. This was to remain a hopeful but unverifiable premise throughout almost 20 years of subsequent work.
With fairly acceptable, and not always reliable, cultures we nonetheless felt able to do what I at least had longed to do for years—to begin a search for possible hematopoietic regulators, but now using the marrow cultures as the basic bioassay procedure. The rationale behind the use of this technique was that the cultures did seem to be able not only to detect such a postulated regulator, but to be capable of providing a workable quantitative assay for it because colony numbers and size increased as the content of active material added to the cultures was increased. By counting colony numbers, some quantitative estimate of the amount of active material added seemed possible.
Motivations Behind the Work
Despite our confidence in the importance of agar cultures as a bioassay system and our desire to characterize the CSF, our publications in the next few years are a curious mixture of studies, many of which had little or no apparent bearing on the pursuit of candidate regulatory factors.
The reasons behind this odd behavior were complex....
Ch. 1 | Ch. 2 | Ch. 8 | Ch. 9 | Ch. 16 | TOP
Chapter 8
(Excerpt)
The Discovery and Purification of Multi-CSF (IL-3)
In 1968, Noel Warner's group in our Unit had been inducing the development of a large series of primary murine tumors, mostly plasmacytomas, by the injection of mineral oil or pristane into responsive BALB/c mice. One of these primary animals (WEHI-3) appeared highly unusual in that the enlarged lymph nodes were greenish in color. Analysis showed that this tumor was in fact a rare myelomonocytic leukemia and transplantation studies led to the derivation of four very different sublines of transplantable tumors (WEHI-3AD), one (WEHI-3D) being remarkable in being tetraploid and in generating tetraploid mature neutrophil progeny. WEHI-3B was of more interest to Malcolm Moore and me because it was capable of forming small, if rather unhealthy, colonies in agar culture, particularly when stimulated by CSF-containing material. WEHI-3B leukemic cells were later established by Chris Wyss in our laboratory as a continuously growing cell line. At some stage in this process the line altered its in vitro growth characteristics to become the autonomous cell line able to form large colonies in agar—the cell line used by us to monitor the purification of G-CSF.
With cultures of the original WEHI-3B tumor cells in 1968, I had observed a curious phenomenon in which coculture of the tumor cells with normal marrow cells led to colony formation demonstrable as originating from the cultured normal cells. This phenomenon was remarkable for two reasons. We had never observed before that coculture of tumor cells with normal cells could result in the stimulation of colony formation by normal cells, suggesting that this cell line had an unusual capacity to produce CSF. More remarkable was the curious appearance of the small numbers of colonies developing. These were composed of uniformly dispersed cells in globular clouds. I had never before observed colonies with this peculiar morphology and pointed out in the published description of these studies that a novel CSF and/or cell type must be involved. On occasion, Malcolm and I had also observed the development of small clusters of giant cells with what could have been multiple or multilobed nuclei. They could have been megakaryocytes, but we hesitated to reach this conclusion because no specific stain for megakaryocytes was then available (now acetylcholinesterase would be used for this purpose) and because aging macrophages in colonies can undergo cell fusion with the formation of large multinucleate cells.
These observations had been set aside as curiosities for possible further exploration at some future date. In retrospect, they represented the first detection of Multi-CSF (IL-3), because at very low concentrations this agent is quite distinctive in stimulating the formation by marrow cells of small numbers of dispersed colonies whose cells still remain only partially characterized.
A quite separate stream of studies was begun in our laboratories in 1973 by an American sabbatical leave worker, John Parker, who had a background of working with lectin-stimulated proliferation of T lymphocytes. Tom McNeill, another postdoc who had been in our laboratory in 1968, had recently observed that the lectin phytohemagglutin seemed able to induce spleen cells to produce CSF. In simultaneous studies, John Parker continued this general line of enquiry by investigating in more detail whether lectins of this type could provoke T lymphocytes to produce CSF. Lectins proved able to stimulate spleen and lymph node populations to produce CSF and he established that it was indeed T
lymphocytes in the mitogen-stimulated spleen cell populations that were the source of the CSF. He settled on pokeweed mitogen as the most effective lectin to use for spleen cell activation and, of the various mouse strains tested, BALB/c spleen cells seemed to be the most active and reliable. Apart from being a novel source of highly active CSF- containing conditioned medium, pokeweed mitogen-stimulated spleen-conditioned medium (SCM) proved able to stimulate the formation of the same curious dispersed colonies as had developed in the earlier cocultures of marrow cells with WEHI-3B cells. We wrote a paper describing these dispersed colonies in which the referees forced me to change the name of the colonies to eosinophil-like colonies. We did have some Giemsa-stained cell preparations from these colonies in which cells were present that could have been eosinophils, but I was certainly not happy to refer to them as eosinophil colonies. This assignation of the name "eosinophil" to these dispersed colonies was quite wrong. They are certainly not composed of eosinophils and later work with an eosinophil-specific stain (Luxol Fast Blue) made it clear that while SCM does in fact stimulate eosinophil as well as granulocyte and macrophage colonies to develop, mouse eosinophil colonies have a very different gross morphology, usually looking like small granulocytic colonies.
SCM was to prove to have some remarkable properties. In 1974, while working for a year in Lausanne, I used SCM as reliable material for stimulating the formation of colonies that we showed were composed of megakaryocytes. To SCM therefore goes the credit of providing the first stimulating material for growing megakaryocyte colonies. In 1976, Greg Johnson and I discovered its ability to stimulate multipotential and erythroid colony formation in cultures of mouse fetal liver cells in the absence of added erythropoietin. No doubt small amounts of erythropoietin may have been present in the human plasma used in the cultures, but it was certainly a novelty to be able to grow large red erythroid colonies without the addition of erythropoietin. Furthermore, the multipotential colonies were an astonishing addition to the hematopoietic colony types now able to be grown in vitro because they could contain cells of at least five lineages.
These developments had been proceeding at a time when the major projects in the laboratory were the purification of lung GM-CSF and the imminent commencement of attempts to purify G-CSF. Although our group now contained 10 scientists and students, it was not large enough to contemplate with equanimity the additional task of purifying what could have been several novel lineage-specific colony-stimulating factors in SCM.
We had formed the view that material like SCM was likely to contain multiple regulatory factors. These potentially might include one that was specific for megakaryocytes, one for multipotential cells, one for eosinophils, and one for erythroid cells. Tests on WEHI-3B-conditioned medium revealed that it had a comparable range of biological actions to SCM, but with one consistent difference - it seemed relatively inefficient in stimulating erythroid colony formation or at least the formation of bright red, fully hemoglobinized erythroid colonies. Was the leukemia-derived material different from that being produced by mitogen-activated normal T-lymphocytes? This seemed a valid enough speculation at the time.
As usual, we slipped gradually into an ever more demanding project to examine what factors might be in SCM or WEHI-3B conditioned medium. At first, this embryonic project was assigned to a research assistant under the guidance of Tony Burgess, then later to a new Ph.D. student, Rob Cutler, under the supervision of Nick Nicola. Greg Johnson and I shared the various marrow and fetal liver assays required.
Because we were convinced that multiple regulatory factors must be responsible, including almost certainly GM-CSF, our initial efforts were not really undertaken with purification as the goal but to provide evidence for the existence of potentially separable lineage-specific molecules.
Other workers, including Mike Dexter and his group in Manchester and Malcolm Moore and his group in New York, held an opposite view that material like SCM and WEHI-3B-conditioned medium might well contain a novel multipotential factor. If so, this was the Holy Grail of hematopoietic regulators, likely not only to be multipotential, but possibly also to be able to stimulate the proliferation and expansion of stem cell populations. The possibilities for clinical exploitation of such a molecule in marrow transplantation had much appeal for these groups.
Frank Lee, Mike Dexter, Don Metcalf, and Nick Nicola
at a CIBA Symposium in London.
As is often the case with controversies, both views turned out to have some validity. SCM does contain a multiplicity of hematopoietic regulators, but it also contains one multipotential regulator that met many of the expectations of those agreeing with the Manchester/New York view. For example, I found that mouse SCM had a truly extraordinary action in cultures of human marrow cells, in stimulating the exclusive formation of eosinophil colonies. This remains the most dramatic example yet encountered of selective colony formation by unfractionated bone marrow cells. We carried out a few biochemical experiments to establish the general properties of the active molecule responsible for stimulating human eosinophil colony formation but took this story no further after publication of these findings.
What had been discovered was later to become known as interleukin-5 (IL-5), and the murine version of IL-5 happens to be fully active on human cells, in contrast to most murine regulators, which have no action on human cells. In the now large literature on IL-5, the initial publication is rarely referred to. By the tough rules in this field, credit is reserved for those making definitive advances like purification or cloning of a regulator. It is not sufficient to discover and partially characterize a molecule. In retrospect, we regret letting this discovery pass undeveloped, but we were attempting to cope with several major projects simultaneously and were about to enter an even more demanding project. We felt that we did not have sufficient available resources and something had to be dropped. At least, I was left with the joy of having seen and worked with some truly remarkable cultures and retain a fond private regard for IL-5 as a briefly loved but abandoned stepchild.
We now know that SCM also contains GM-CSF, IL-6, stem cell factor, leukemia inhibitory factor (LIF), Flk ligand, IL-9, and doubtless other as yet unidentified regulators....
Ch. 1 | Ch. 2 | Ch. 8 | Ch. 9 | Ch. 16 | TOP
Chapter 9
(Excerpt)
The End of the Beginning:
An Appraisal of Progress So Far
With the resolution of the multipotential regulator controversy and inclusion of interleukin-3 (IL-3, or Multi-CSF) into the gang of four CSFs able to regulate granulocyte-macrophage production, the discovery and purification phases for the murine colony-stimulating factors ended.
Beginning with the striking phenomenon of granulocyte-macrophage colony formation by mouse bone marrow cells when stimulated by organ fragments or medium conditioned by various tissues, we had convinced ourselves that the phenomenon was due to stimulation by a glycoprotein produced by various organs and that this glycoprotein, termed colony-stimulating factor, was detectable in the serum and urine with its concentration varying in situations involving perturbations of granulocyte-macrophage populations.
We had come to recognize that granulocyte macrophage colony formation was a more complex phenomenon than at first seemed apparent, and, with the later development of similar culture techniques for eosinophil, megakaryocyte, erythroid, and multipotential colony formation, a rather complete analytical system had emerged for analyzing in detail the biology of hematopoietic populations.
Techniques for growing various erythroid colonies had been developed in Toronto, beginning in 1961, by Stephenson, Axelrad, Tepperman, and colleagues. These systems were clearly allowing the detection and characterization of erythropoietin, the regulator of more mature erythroid precursors, but this was not a subject of central interest to us. It was best left to our Canadian and U.S. colleagues.
We had concentrated on the granulocyte-macrophage colony- forming system, later recognizing from work of Paul Chervenick that eosinophil colonies also developed in such cultures.
While we often found ourselves culturing human marrow-derived colonies, we had tended to focus our efforts on murine cultures and their stimulation by CSF. Even though our first efforts to purify CSF had involved human urine, for a long period after this we concentrated on murine CSFs. We then progressively had recognized that more than one CSF existed, turning our efforts in sequence from M-CSF to GM- CSF then G-CSF and finally Multi-CSF (IL-3).
Throughout this 15-year period we had few competitors who were making sustained contributions to the purification and characterization of these murine CSFs. As a consequence, through Richard Stanley, first in Melbourne then in Toronto and New York, M-CSF had been purified and its actions established. We had purified and characterized GM-CSF and G-CSF while the efforts of several groups to characterize murine Multi-CSF had been abruptly truncated by the work of Jim Ihle and his co-workers in purifying IL-3 (Multi-CSF).
Our own view was that these four CSFs represented a group of functionally related regulators corresponding for the granulocyte-macrophage lineages to the single regulator, erythropoietin, for erythropoiesis. We had few firm grounds for confidence in portraying these four CSFs as the regulatory system for granulocyte and macrophage populations. We had buried under the carpet some worrying observations that there might well be differing forms of GM-CSF and that not all M-CSFs seemed to have identical properties. G-CSF and Multi-CSF seemed less of a worry in this regard.
Why did we believe that the CSFs were a functionally related family? The major reason, I suppose, was that they stimulated somewhat similar, if individually distinct, granulocyte-macrophage colony formation by bone marrow cells, the defining property for the name CSF. But was this simply a self-fulfilling terminology? The molecules were of quite different sizes when glycosylated and the small amounts of amino acid sequence data obtained indicated no obvious homology between the four. However, the four CSFs were beginning to reveal a common pattern of polyfunctionality, to be discussed shortly, which tended to unify them. Moreover, various tissues seemed able to simultaneously produce at least three of these CSFs and in response to comparable inducing signals. The odd man out in the group was Multi-CSF, readily able to be produced by lymphocytes in vitro but not detectable in vivo in normal mice. Also of likely relevance in unifying these CSFs were observations made by Nick Nicola and Francesca Walker that binding of one CSF to membrane receptors had predictable consequences for the ability of the cells to bind other CSFs, suggesting some sort of close functional interaction between these four regulators.
Having said all this, it must be admitted that it was somewhat presumptuous to give these four regulators the common name of CSF, implying that they were in fact a tight family of regulators. It was not until considerably later when the CSF genes and their receptors had been cloned that more cogent reasons emerged for associating at least two of these CSFs - GM-CSF and Multi-CSF - into a paired group. It certainly helped us in the early 1980s to talk and write about the CSFs as a functional family and was a quite defensible simplification to allow the regulatory biology of granulocyte-macrophage populations to be introduced to general audiences. It was however not much more than a stratagem for erecting some signposts in what had previously been uncharted territory.
I do not recall that we spent much time formally considering the possibility that there would be four matching human CSFs for the four murine CSFs. Despite this lack of clear formulation, we obviously must have had such a working understanding. We had, after all, begun with human M-CSF, which, although having some odd functional deficiencies, was obviously related to murine M-CSF. When we were working with human placental-derived CSFs we had no hesitation in identifying human analogues for GM-CSF and G-CSF. We had not encountered a human molecule corresponding to Multi-CSF at that time nor had any other group produced evidence for such a molecule, so this CSF remained somewhat in limbo.
In the later stages of our work on the murine CSFs, other groups had begun characterizing human CSFs, mainly being produced by tumor cell lines. Not having access to these particular cell lines, we fairly wisely had not been tempted to extend still further our range of activities but nonetheless followed these developments with proprietary interest. The initial results from these groups were fairly messy, given by then the quite detailed information available on how to purify murine CSFs and the stringent requirements to be met before purification could be achieved.
Ultimately, the group achieving success with the purification of human GM-CSF was that of Judy Gasson and David Golde at UCLA using the Mo leukemia cell line. In parallel, the two groups achieving success finally with the purification of G-CSF were those of Shige Nagata and colleagues in Tokyo and of Karl Welte, Erich Platzer, and Malcolm Moore in New York when finally the material was passed to Larry Souza's group at Amgen. These successes can be dated approximately to 1985-1986, sometime after completion of purification of the murine CSFs. I remain uncertain whether anyone actually ever purified human Multi-CSF from native sources....
Ch. 1 | Ch. 2 | Ch. 8 | Ch. 9 | Ch. 16 | TOP
Chapter 16
(Excerpt)
The Bottom Lines
Not a day passes now when CSFs are not in use in experimental hematology laboratories for one type of experiment or another. In our case, we now most often seem to be using the CSFs to investigate the behavior of granulocyte-macrophage populations from mice with some form or other of gene manipulation. Certainly, not a day passes when patients somewhere are not receiving injections of CSF or transplants of CSF-elicited peripheral blood stem cells.
The CSFs have become part of the fabric of experimental hematology and clinical medicine. For our group, which uncovered the CSFs and stuck with them during their subsequent difficult gestation and development, this has been a satisfying outcome. With the passage of time, members of our team have come and gone. Of the 522 refereed papers and six books on the CSFs and related matters so far coming from studies in which we participated, there are in fact 329 scientists and clinicians as authors (see Appendix).
The Cancer Research Unit, 1996.
The cost of this work is difficult to calculate, but, in the case of GM-CSF, our largest single project in the area, we have estimated that it cost us in the Hall Institute in the region of U.S. $30,000,000. It has become the fashion of many governments these days to hope for financial returns from appropriate medical research. For the Hall Institute and Ludwig Institute our combined royalties from GM-CSF approximate $2,000,000 per year and, for a patent with a 10- to 15-year life, we will not recoup the money we spent on the GM-CSF project. For our G-CSF work we receive only nominal royalties from Australian sales of G-CSF by one company
This is an abysmal way of assessing medical research. Medical research includes, but is not merely, the intellectual exercise of discovering how the body and its various tissues develop and function. The primary objectives of medical research have always been to discover the nature of human diseases and, with this knowledge, to develop methods for disease prevention or cure. Medical research has therefore always been goal-oriented - the mastery of disease through discovery. Those undertaking medical research have become familiar with having to justify their research in a defensive manner - justifying to donors or government agencies why effort and money has been, or will be, expended properly in medical research. Justification, rather than exploitation, therefore becomes incorporated early into the ethos of all young research workers....
Ch. 1 | Ch. 2 | Ch. 8 | Ch. 9 | Ch. 16 | TOP
|
