Immunology Today, voL 8, No. 4, 1987
A recent workshop* held in the Bavarian alps considered the possibilities for immunological intervention with monoclonal antibodies. While at present the benefits of most attempts to use rnonoclonal antibodies in vivo in man are uncertain, in-vitro and animal experiments have yielded enough positive results to encourage further efforts. Immediate dramatic effects are not to be expected because as one participant noted, we are still at an earlystage in the use of monoclonal antibodies as therapeutic agents. The workshop began by considering the engineeiing of antibody molecules. Several lai:oratories presented alternative strategies, from chemical coupling of mouse Fab to human Fc to make univalent antibodies, via genetic engineering of divalent mouse V region-human C region constructs, to chimaeric 'humouse' antibodies in which only the complementarity determining regions are of mouse origin (M. Neuberger, Cambridge). These constructs have been used to investigate the effect of univalency on modulation and complement dependent lysis (G. Stevenson, Southampton), the invitro biological properties of different human Ig s~,bclasses (L. Sun, Philadelphia) and the affinity and expression of idiotypes of 'humouse' antibodies (Neuberger). As yet, the key question of the immunogenicity in man of these chimaeric constructs has not been tested. Irrespective of the outcome of these early attempts, the potential of genetic engineering was illustrated by the production of a soluble T-cell receptor-immunoglobulin chimaeric molecule (K. Karjalainen, Basel). As an alternative to engineering mouse-human antibodies, several groups have worked on the production of human monoclonal antibodies. Two strategies are employed: transformation with EBV (M. Steinitz, Jerusalem; H. Mitsuya, Bethesda) or cell fusion with mouse (R. Cote, New York; M. Mitchell, Los Angeles), or human (H. ZeiglerHeitbrock, Munich)cell lines. Both methods can work and there is now quite extensive data on antibodies produced in cancer patients. The
Immunological intervention with monoclonal antibodies . . . . . . from P.Ci. Beverleyand G. Riethmuller majority of these are broadly reactive with many cell types and identify intracellular antigens while a minority (5%) are against cell-surface antigens. Many are against glycolipid antigens. The specificity of antibodies generated from cancer and autoimmune patients and normal individuals seems very similar (Cote). The implication is that tumourspecific responses are, at best, few and far between. A variety of strategies for potentiating the effects of antibodies were discussed. Drug-antibody (R. Baldwin, Nottingham) and toxinantibody conjugates (EJ. Wawrzynczak, London; J. Uhr, Dallas)were discussed. Several factors clearly determine the effectiveness of the conjugates, including the nature of the antigen and the epitope to which the antibody is directed, the presence uf both A and B chains of toxins, the st~;bility of chemical linkages and the extent of glycosylation of toxins. Other methods of ootentiation depend on host effects, for example the use of bispecific anti.
.
.
.
~
~
u,lld$
U
LIk,llllVUI
IkC;ll~JC[
r]llLi-
gen and the T-cell receptor (U Staerz, La Jolla; G. Jung, Bethesda) Some clinical protocols have also employed autologous lymphocytes mixed with monoclonal antibodies in an attempt to target a host response (H. Mellstedt, Stockholm; J. Douillard, Bethesda). Targeting of monoclonal antibodies to micrometastases in bone marrow of some cancer patients was reported by G. Schlimock (Augsburg) and G. Riethmuller (Munich). A major part of the meeting dealt with antioidiotypic responses, eitha;r as a subject of deliberate investigation or because a major part of the human (and animal) response to heterologous immunoglobulin is anti-idiotypic. In the mouse mammary tumour virus model, antibodies to the envelope gp52 can be used to raise anti-idiotypes (ABz, the internal image of the antigen) which can act as a protective 'vaccine' in eliciting *An InternationalWorkshopon ImmunologicIn- antibody (AB3) against gp52, terventionwith MonoclonalAntibodieswas heldin delayed-type hypersensitivity and protection against tumour challenge Elman,FRG,15--18October,1986.
© 1987,ElsevierPublications,Cambridge 0167- 4919/87/$02.00
(H. Kohler, Buffalo). In man many patients given the 17-1A monoclonal antibody against a colon carcinoma antigen make an anti-idiotypic response (AB2). Fifty per cent of patients given (a mouse monocional) AB2 made antibody (detectable after in-vitro culture) to tumour antigen (AB3). A different use of anti-idiotypes was the attempt to generate a vaccine to the AIDS retrovirus (HIV) by immunization with antibody to its receptor, the CD4 molecule of T cells (R. Kennedy, San Antonio; P. Beverley, London). These anti-idiotypes reduce viral reverse transcriptase and syncytia and bind to envelope glycoprotein (Kennedy). While anti-idiotypes may in future be used to advantage in therapy, they also limit the more straightforward use of mouse antibodies for immunosuppression. In monkeys the duration of effective imn-,unosuppression by CD3, CD4 or CD8 antibodies was limited b,, the antimouse response, much of which was idiotype specific. The possibility of ~w.tLlmlcJ i u l u l . y p u has begun to be explored and preliminary results are encouraging, while there is support for the view that anti-Fc antibody may sometimes potentiate the act.ivity of the mouse antibody (M. Jonker, Rijswijk). Other antibodies with narrower specificity against activatPd T cells (CD25 or anti-lL2R) showed promise in a rat model (T. Diamanstein, Berlin; R. Kirkman, Boston)especially when combined with cyclosporin A. Disappointingly the results in monkeys were less dramatic but illustrated another point (as well as the species effect): the difference it, effectiveness of antibodies against different epitopes of an antigen (Kirkman). In mice antibodies to CD4 were shown to tolerize both to injected rat immunoglobulin and some, but not all, protein antigens injected concurrently (H. Waldmann, Cambridge; J. Goronzy, Stanford). Mice tolerant to L3T4 monoclonal (CD4) could not respond to rat antibodies if these were against soluble antigens but would respond, with an anti-
101
....
idiotypic response, to other rat monoclonals against mouse cell surface antigens (Waldmann). The reason for the high immunogenicity of cell-bound antibody is not yet clear. While these animal experiments sometimes show dramatic effects, what of the human experience? This is less clearcut except in reversal of acute renal allograft rejection episodes because although more than 500 individuals worldwide have been treated with 17-1A antibody, there has not yet been a single controlled trial, so all the results must be considered anecdotal. There was general agreement, however, that if
The use of recombinant DNA techniques, cell fusion and novel bioprocessing in the pharmaceutical industry has assisted the development of many kinds of diagnostic and therapeutic products. Some directly affect the immune system (e.g. interleukir~s, intefferons, turnout necrosis factor, and colony stimulating factors). Others (e.g. peptides, cytokines, growth factors, H2-receptor antagonists, nonsteroidal anti-inflammatory drugs and neurohormonal agents) have immunological reactivity even though they are not designed to be immune modulators. The need to define the immunotoxico!ogical potential of these products during preclinical safety evaluation was among the topics discussed at a recent meeting*.
102
Immunology Today,vol. 8, No. 4, 1987
responses do occur, they often begin well after the monoclonal antibody treatment, suggesting that host responses may be important. Can any general conclusions be drawn ? Perhaps the most important is that we are indeed at the beginning of the use of monoclonal antibodies for immunomodulation. Clearly, they can have powerful effects, alone or with other agents. Equally clearly we need to know much more about many of the areas discussed in this workshop. We need to understand what makes a 'good' or 'bad' target antigen or epitope0 what host effector mechanisms can be recruited and how, what deLer-
mines the half-life of antibodies or conjugates, how to induce tolerance to antibodies or how to manipulate the anti-idiotype response in a favourable way. While it may be a long hard road, the results presented at the workshop encourage the belief that antibodies may one day be Ehrlich's magic bullet.
Peter Beverleyis in the ICRFHuman Tumour Immunology Group, University College London, UK; and G. Riethmulleris at the Institute for Immunology of the Universityof Munich, FRG.
hmunotoxicology and the new biotechnology from JayCavagnaro Active cells expand clonally and both differentiate and proliferate. Effector function may therefore be inhibited while affector function remains intact (L. Loose, Groton).
The immune system in toxicity testing The principal debate at the meetII i~j
vvcl,)
y v ! I~'LI Ir~'l
UI
I IUL
I.I I ~
I I~VV
products of biotechnology can be adequately evaluated by traditional toxicity testing methods or whether no,eel approaches are required. Many toxicologists are wary of giving The major problems in this avalu- an impression of confusion and beation are the compounds' pancyto- lieve that 'there is nothing new'. toxicity and polypharmacological Nevertheless, original thinking is actions (G. Morgan, Philadelp>.ia). needed (H. Ozer, Chapel Hill). The Our knowledge of the biochemical immune system can be a sensitive pharmacology of these preducts is indicator of general toxicity as well limited and the site of action is often as an excellent source of information unknown. The challenge is to design about the mechanisms of action, inthe most specific and relevant ex- cluding toxicity, of the new biotech periment to answer questions of products. toxicity for the target organ. HowTraditional descriptive testing ever, the immune system as a target schemes in animals were developed organ is a complex network of cells for small molecular weight suband tissues and is therefore very stances in the context of regulatory different from the liver, kidney, heart guidelines. In the past, drug toxicity or brain. The immune system is anti- testing focused on physiological, gen driven and responds to a vast biochemical, pharmacological and number of antigens. The rate and formulation safety. Many of the cureffectiveness of the immune re- rent biotechnology products are norsponse can be shifted by xenobiotics. mal constituents of the body but they are administered at pharmaco* Themeetingwasheldat TvsonsCorner,Virginia, logical rather than phys;o~ogical USA,23-24 October,1986. levels, often by an unnatural route,
and may reach tissue sites not normally exposed. Many of the lymphokines may not be toxic but may activate numerous effector mechanisms, so that monitoring optimal activity becomes difficult. If the compounds are acting through indirect mechanisms 'more is certain!y not UI~LLt~I
•
I ll~
l[idXl|rlUrrl
[gl~rdl:eCl
dose (MTD) may be an overdose because the response curve for immunological effects is often bellshaped (C. Pinsky, Frederick). Dose-effect relationships are also problematic when efficacy is inversely related to dose. If immunomodulation is the aim, the optimum dose in a toxicity study may be one that achieves maximum toxicity at the higher levels and by definition maximum inhibition of immunological function (Ozer).
Preclinical testing In the biotech industry toxicology testing is often done early in a product's development (J. Kopplin, Emmeryville). The responsibility of preclinical safety testing thus becomes one of not only establishing the safety and potential toxicity of the test material but also evaluating the pharmacokinetics of the drug and monitoring the pharmacological responses. Such were the goals of the subchronic toxicity study in primates of granuiocyte-macrophage
(~) 1987, Elsevier Publications, Cambridge
0167 - 4919/87/$02.00
A recent workshop* held in the Bavarian alps considered the possibilities for immunological intervention with monoclonal antibodies. While at present the benefits of most attempts to use rnonoclonal antibodies in vivo in man are uncertain, in-vitro and animal experiments have yielded enough positive results to encourage further efforts. Immediate dramatic effects are not to be expected because as one participant noted, we are still at an earlystage in the use of monoclonal antibodies as therapeutic agents. The workshop began by considering the engineeiing of antibody molecules. Several lai:oratories presented alternative strategies, from chemical coupling of mouse Fab to human Fc to make univalent antibodies, via genetic engineering of divalent mouse V region-human C region constructs, to chimaeric 'humouse' antibodies in which only the complementarity determining regions are of mouse origin (M. Neuberger, Cambridge). These constructs have been used to investigate the effect of univalency on modulation and complement dependent lysis (G. Stevenson, Southampton), the invitro biological properties of different human Ig s~,bclasses (L. Sun, Philadelphia) and the affinity and expression of idiotypes of 'humouse' antibodies (Neuberger). As yet, the key question of the immunogenicity in man of these chimaeric constructs has not been tested. Irrespective of the outcome of these early attempts, the potential of genetic engineering was illustrated by the production of a soluble T-cell receptor-immunoglobulin chimaeric molecule (K. Karjalainen, Basel). As an alternative to engineering mouse-human antibodies, several groups have worked on the production of human monoclonal antibodies. Two strategies are employed: transformation with EBV (M. Steinitz, Jerusalem; H. Mitsuya, Bethesda) or cell fusion with mouse (R. Cote, New York; M. Mitchell, Los Angeles), or human (H. ZeiglerHeitbrock, Munich)cell lines. Both methods can work and there is now quite extensive data on antibodies produced in cancer patients. The
Immunological intervention with monoclonal antibodies . . . . . . from P.Ci. Beverleyand G. Riethmuller majority of these are broadly reactive with many cell types and identify intracellular antigens while a minority (5%) are against cell-surface antigens. Many are against glycolipid antigens. The specificity of antibodies generated from cancer and autoimmune patients and normal individuals seems very similar (Cote). The implication is that tumourspecific responses are, at best, few and far between. A variety of strategies for potentiating the effects of antibodies were discussed. Drug-antibody (R. Baldwin, Nottingham) and toxinantibody conjugates (EJ. Wawrzynczak, London; J. Uhr, Dallas)were discussed. Several factors clearly determine the effectiveness of the conjugates, including the nature of the antigen and the epitope to which the antibody is directed, the presence uf both A and B chains of toxins, the st~;bility of chemical linkages and the extent of glycosylation of toxins. Other methods of ootentiation depend on host effects, for example the use of bispecific anti.
.
.
.
~
~
u,lld$
U
LIk,llllVUI
IkC;ll~JC[
r]llLi-
gen and the T-cell receptor (U Staerz, La Jolla; G. Jung, Bethesda) Some clinical protocols have also employed autologous lymphocytes mixed with monoclonal antibodies in an attempt to target a host response (H. Mellstedt, Stockholm; J. Douillard, Bethesda). Targeting of monoclonal antibodies to micrometastases in bone marrow of some cancer patients was reported by G. Schlimock (Augsburg) and G. Riethmuller (Munich). A major part of the meeting dealt with antioidiotypic responses, eitha;r as a subject of deliberate investigation or because a major part of the human (and animal) response to heterologous immunoglobulin is anti-idiotypic. In the mouse mammary tumour virus model, antibodies to the envelope gp52 can be used to raise anti-idiotypes (ABz, the internal image of the antigen) which can act as a protective 'vaccine' in eliciting *An InternationalWorkshopon ImmunologicIn- antibody (AB3) against gp52, terventionwith MonoclonalAntibodieswas heldin delayed-type hypersensitivity and protection against tumour challenge Elman,FRG,15--18October,1986.
© 1987,ElsevierPublications,Cambridge 0167- 4919/87/$02.00
(H. Kohler, Buffalo). In man many patients given the 17-1A monoclonal antibody against a colon carcinoma antigen make an anti-idiotypic response (AB2). Fifty per cent of patients given (a mouse monocional) AB2 made antibody (detectable after in-vitro culture) to tumour antigen (AB3). A different use of anti-idiotypes was the attempt to generate a vaccine to the AIDS retrovirus (HIV) by immunization with antibody to its receptor, the CD4 molecule of T cells (R. Kennedy, San Antonio; P. Beverley, London). These anti-idiotypes reduce viral reverse transcriptase and syncytia and bind to envelope glycoprotein (Kennedy). While anti-idiotypes may in future be used to advantage in therapy, they also limit the more straightforward use of mouse antibodies for immunosuppression. In monkeys the duration of effective imn-,unosuppression by CD3, CD4 or CD8 antibodies was limited b,, the antimouse response, much of which was idiotype specific. The possibility of ~w.tLlmlcJ i u l u l . y p u has begun to be explored and preliminary results are encouraging, while there is support for the view that anti-Fc antibody may sometimes potentiate the act.ivity of the mouse antibody (M. Jonker, Rijswijk). Other antibodies with narrower specificity against activatPd T cells (CD25 or anti-lL2R) showed promise in a rat model (T. Diamanstein, Berlin; R. Kirkman, Boston)especially when combined with cyclosporin A. Disappointingly the results in monkeys were less dramatic but illustrated another point (as well as the species effect): the difference it, effectiveness of antibodies against different epitopes of an antigen (Kirkman). In mice antibodies to CD4 were shown to tolerize both to injected rat immunoglobulin and some, but not all, protein antigens injected concurrently (H. Waldmann, Cambridge; J. Goronzy, Stanford). Mice tolerant to L3T4 monoclonal (CD4) could not respond to rat antibodies if these were against soluble antigens but would respond, with an anti-
101
....
idiotypic response, to other rat monoclonals against mouse cell surface antigens (Waldmann). The reason for the high immunogenicity of cell-bound antibody is not yet clear. While these animal experiments sometimes show dramatic effects, what of the human experience? This is less clearcut except in reversal of acute renal allograft rejection episodes because although more than 500 individuals worldwide have been treated with 17-1A antibody, there has not yet been a single controlled trial, so all the results must be considered anecdotal. There was general agreement, however, that if
The use of recombinant DNA techniques, cell fusion and novel bioprocessing in the pharmaceutical industry has assisted the development of many kinds of diagnostic and therapeutic products. Some directly affect the immune system (e.g. interleukir~s, intefferons, turnout necrosis factor, and colony stimulating factors). Others (e.g. peptides, cytokines, growth factors, H2-receptor antagonists, nonsteroidal anti-inflammatory drugs and neurohormonal agents) have immunological reactivity even though they are not designed to be immune modulators. The need to define the immunotoxico!ogical potential of these products during preclinical safety evaluation was among the topics discussed at a recent meeting*.
102
Immunology Today,vol. 8, No. 4, 1987
responses do occur, they often begin well after the monoclonal antibody treatment, suggesting that host responses may be important. Can any general conclusions be drawn ? Perhaps the most important is that we are indeed at the beginning of the use of monoclonal antibodies for immunomodulation. Clearly, they can have powerful effects, alone or with other agents. Equally clearly we need to know much more about many of the areas discussed in this workshop. We need to understand what makes a 'good' or 'bad' target antigen or epitope0 what host effector mechanisms can be recruited and how, what deLer-
mines the half-life of antibodies or conjugates, how to induce tolerance to antibodies or how to manipulate the anti-idiotype response in a favourable way. While it may be a long hard road, the results presented at the workshop encourage the belief that antibodies may one day be Ehrlich's magic bullet.
Peter Beverleyis in the ICRFHuman Tumour Immunology Group, University College London, UK; and G. Riethmulleris at the Institute for Immunology of the Universityof Munich, FRG.
hmunotoxicology and the new biotechnology from JayCavagnaro Active cells expand clonally and both differentiate and proliferate. Effector function may therefore be inhibited while affector function remains intact (L. Loose, Groton).
The immune system in toxicity testing The principal debate at the meetII i~j
vvcl,)
y v ! I~'LI Ir~'l
UI
I IUL
I.I I ~
I I~VV
products of biotechnology can be adequately evaluated by traditional toxicity testing methods or whether no,eel approaches are required. Many toxicologists are wary of giving The major problems in this avalu- an impression of confusion and beation are the compounds' pancyto- lieve that 'there is nothing new'. toxicity and polypharmacological Nevertheless, original thinking is actions (G. Morgan, Philadelp>.ia). needed (H. Ozer, Chapel Hill). The Our knowledge of the biochemical immune system can be a sensitive pharmacology of these preducts is indicator of general toxicity as well limited and the site of action is often as an excellent source of information unknown. The challenge is to design about the mechanisms of action, inthe most specific and relevant ex- cluding toxicity, of the new biotech periment to answer questions of products. toxicity for the target organ. HowTraditional descriptive testing ever, the immune system as a target schemes in animals were developed organ is a complex network of cells for small molecular weight suband tissues and is therefore very stances in the context of regulatory different from the liver, kidney, heart guidelines. In the past, drug toxicity or brain. The immune system is anti- testing focused on physiological, gen driven and responds to a vast biochemical, pharmacological and number of antigens. The rate and formulation safety. Many of the cureffectiveness of the immune re- rent biotechnology products are norsponse can be shifted by xenobiotics. mal constituents of the body but they are administered at pharmaco* Themeetingwasheldat TvsonsCorner,Virginia, logical rather than phys;o~ogical USA,23-24 October,1986. levels, often by an unnatural route,
and may reach tissue sites not normally exposed. Many of the lymphokines may not be toxic but may activate numerous effector mechanisms, so that monitoring optimal activity becomes difficult. If the compounds are acting through indirect mechanisms 'more is certain!y not UI~LLt~I
•
I ll~
l[idXl|rlUrrl
[gl~rdl:eCl
dose (MTD) may be an overdose because the response curve for immunological effects is often bellshaped (C. Pinsky, Frederick). Dose-effect relationships are also problematic when efficacy is inversely related to dose. If immunomodulation is the aim, the optimum dose in a toxicity study may be one that achieves maximum toxicity at the higher levels and by definition maximum inhibition of immunological function (Ozer).
Preclinical testing In the biotech industry toxicology testing is often done early in a product's development (J. Kopplin, Emmeryville). The responsibility of preclinical safety testing thus becomes one of not only establishing the safety and potential toxicity of the test material but also evaluating the pharmacokinetics of the drug and monitoring the pharmacological responses. Such were the goals of the subchronic toxicity study in primates of granuiocyte-macrophage
(~) 1987, Elsevier Publications, Cambridge
0167 - 4919/87/$02.00
