Monoclonal. antibodies in cancer therapy
`
`Gert Riethmiiller, Elena Schneider-Gadicke and Judith P Johnson
`
`lnstitut fur lmmunologie, Munchen, Germany
`
`A review of the clinical trials of antibody based cancer therapies reveals
`that this approach can, in rare cases, induce complete remissions in
`individual patients with cancer. Since these trials have usually involved
`patients with large tumor masses, tumor cell inaccessibility is probably
`a major reason for the prevailing failures. Minimal residual disease, the
`stage when tumor cells are few and dispersed, should therefore be
`a more promising target for therapeutic antibodies. This hypothesis is
`supported by a prospective randomized trial on patients with resected
`Dukes C colorectal carcinoma that resulted in increased survival and
`prolonged recurrence-free intervals. Thus, in addition to strategies designed
`to produce more effective, human-derived reagents, efforts need to
`be concentrated on directing passive antibody therapy towards the
`appropriate target.
`
`Current Opinion in Immunology 1993, 5:732-739
`
`Introduction
`
`Passive antibody therapy of cancer is one of the oldest
`and most prominent issues of tumor immunology. As
`early as 1895, a few years after von Behring's and Ki(cid:173)
`taSato's discovery that antisera against diphtheria toxin
`could cure children with diphtheria, Hericourt and Richet
`reported on their attempts to treat cancer patients with
`antisera prepared in dogs and donkeys. Paul Ehrlich,
`with his antisera against plant toxins abrin and ricin,
`had shown the specificity of the newly induced serum
`substances and named them Antikorper (antibody). He
`became particularly intrigued by their potential use as
`specific weapons against cancer cells and coined for them
`the tenn Zauberkugel (magic bullets). Nevertheless, de(cid:173)
`spite these early beginnings, antibody therapy of cancer
`has become a story of unending failures.
`
`In 1975, a turning point seemed to have been reached
`with the invention of the hybridoma technique by Kohler
`and Milstein [ 1] . Monoclonal antibodies (mAbs) with
`their unifonn and well-defined specificity and virtually
`inexhaustible supply, promised to bring a solution to
`the vexing problems of variable specificity and irrepro(cid:173)
`ducibility inherent in polyclonal antisera. Indeed, a spate
`of reports on mAbs with presumed tumor-restricted or
`demonstrated tumor-associated specificity appeared in
`subsequent years. While several of those mAbs were
`used in a clinical setting as valuable diagnostic tools,
`so far none of them has gained recognition as an estab(cid:173)
`lished therapeutic against malignant disease. As reviewed
`by ourselves for solid tumors [2], numerous clinical trials
`have been perfonned with unmodified mAbs without any
`consistent pattern of response. An obvious conclusion
`
`to be drawn from these conspicuous failures was, that
`in spite of their exquisite specificity and their apparent
`ability to target tumor cells, antibodies alone were either
`not sufficiendy cytotoxic or could not adequately harness
`the patients' own effector mechanisms. Consequently, a
`broad research effort was begun to improve the cytotoxi(cid:173)
`city of antibodies by conjugating them with radioactive
`isotopes, cytotoxic drugs or potent toxins. These efforts
`culminated in the development of single chain antibod(cid:173)
`ies, consisting solely of covalently connected VH and VL
`peptides, to which toxin molecules had been fused by
`recombinant DNA techniques (3-5]. However, as acces(cid:173)
`sibility of tumor cells in advanced stages of cancer to
`macromolecules may be strictly limited, this review fa.
`cuses on the minimal residual disease stage as a much
`more promising target for antibody-based therapies.
`
`A decade of clinical trials-some successes
`but more disappointments
`
`Within the last year, several extensive reviews have ap(cid:173)
`peared that describe results of phase I and phase II
`therapeutic trials using both unmodified mAbs as well as
`various antibody conjugates (Table I) (2,6 00,7•,8••,9] .
`
`A rough assessment of the reported successes and fail(cid:173)
`ures indicates that complete remissions have been most
`often observed in Non-Hodgkin lymphomas with ra(cid:173)
`dioimmunoconjugates, which appear to be superior to
`immunotoxins and unmodified antibodies. Moreover, in
`myeloid leukemia, the combination therapy with high(cid:173)
`dose cytoxan and radiation therapy led to a sizeable
`rate of remissions. Antibody trials on solid tumors, how-
`
`Abbreviations
`CDR----complementarity determining region; GM-CSF--granulocyte·macrophage colony-stimulating factor;
`mAb--monoclonal antibody.
`
`732
`
`© Current Biology Ltd ISSN 0952-7915
`
`Celltrion v. Genentech
`IPR2017-01374
`Genentech Exhibit 2060
`
`

`

`Monoclonal antibodies in cancer therapy Riethmuller et af 733
`
`Table 1. Recent reviews on antibody-based tumor therapy.
`
`Authors
`
`Target
`
`Antibody therapy
`
`No. of trials•
`
`Crossbard et al. [6u]
`
`Leukemia and
`lymphoma
`
`Unmodified antibodies
`lmmunotoxin
`Radioimmunocon jugates
`with or without chemotherapy
`or radiotherapy
`
`Riethmuller and Johnson [21
`
`Melanoma and carcinomas
`
`Unmodified antibodies
`
`Steffens et al. [7•]
`
`Melanoma
`
`Unmodified antibodies
`
`Vitetta et a/. [8•• J
`
`Melanoma and carcinomas
`
`Various immunotoxins
`
`LoBuglio and Saleh [9]
`
`Lymphoma,
`melanoma, ovarian,
`breast and gastrointestinal
`cancer
`
`Unmodified antibodies,
`radioimmunoconjugates,
`immunotoxins
`
`16
`9
`20
`
`12
`
`8
`
`16
`
`16
`9
`10
`
`No. of
`patients
`
`Complete
`responses
`
`161
`179
`211
`
`196
`
`74
`
`375
`
`5
`6
`40
`
`2
`
`3
`
`9
`
`not
`detailed
`
`not
`given
`
`·some trials reviewed by two or more reviewers.
`
`ever, have yielded complete remissions only in the rarest
`cases. This group of tumors comprises mainly cancers
`derived from simple epithelia which develop metastases
`that are composed of host-derived stromal tissue and of
`differentiated epithelial parenchyma growing within the
`envelope of a dense basement membrane. The epithelial
`tissue organization sets carcinomas quite apart not only
`from lymphoma and leukemia but also from melanoma,
`the metastases of which lack the typical coherent tra(cid:173)
`becular or adenomatous formations in which cells are
`connected by desmosomal intercellular junctions. Inter(cid:173)
`estingly, melanoma appears to be one of the more sus(cid:173)
`ceptible neoplasms for antibody-based therapies.
`
`The critical issue of tumor-cell accessibility
`
`The rare complete tumor regressions observed with all of
`the various antibody-based modalities demonstrate that,
`in principle, unmodified antibodies as well as immuno(cid:173)
`conjugates can produce sufficient tumor-directed cyto(cid:173)
`toxicity. Why then do complete remissions only occur
`in rare individual patients? In the absence of any com(cid:173)
`mon immunogenic trait characteristic for the responder
`patients, the suspicion is warranted that peculiar con(cid:173)
`ditions of the individual tumor are responsible for the
`antibody-induced regression. Among these, an abnor(cid:173)
`mal vascularization highly permeable to intravenously
`injected antibodies, a homogeneous expression of the
`target antigen on the relevant clonogenic tumor cells, and
`accessibility as well as vulnerability to direct or indirect
`cytotoxic effects of antibodies, rank very high. Although
`the sporadic nature of antibody-induced regressions may
`be reduced to the rare coincidence of several of these
`factors, several lines of evidence point to the inaccessi(cid:173)
`bility of cancer cells growing in solid tumor parenchyma
`as a leading cause of the observed therapy failures.
`
`The results of the trials themselves (Table 1), showing
`that responses are far more common in hematopoi(cid:173)
`etic malignancies than in solid tumors, underscore this
`reasoning. Furthermore, an impressive amount of exper(cid:173)
`imental data, much of which has been compiled in recent
`years by Jain and his colleagues (10,11], demonstrate that
`macromolecules, including mAbs, have difficulty reaching
`epithelial tumor cells. Elevated interstitial fluid pressure
`in solid tumor nodules is one of the major obstacles in
`the long list of vascular and interstitial barriers imped(cid:173)
`ing the delivery of antibodies to cancer cells (Table 2).
`A recent report demonstrates that interstitial pressures as
`high as 33 mmHg can be directly measured in individual
`head and neck tumors in situ (12) and similar values have
`been determined for subcutaneously growing metastases
`of melanoma and primary cervical cancers (10,13). Ad(cid:173)
`ditional barriers for the free diffusion and convection
`of antibodies include the basement membrane envelop(cid:173)
`ing the epithelial tumor trabeculae, the intercellular tight
`junctions and the long distances extravasated antibodies
`must travel through the dense intrastitial mesh of proteo(cid:173)
`glycans in order to reach their cellular targets.
`The positive in vivo labelling data obtained by numerous
`immunoscintigraphic studies and the ex vivo autoradio(cid:173)
`graphic analysis of labelled tumor biopsies do not refute
`this view of the ability of antibodies to penetrate into tu(cid:173)
`mor tissue, as the majority of these studies attest to a
`heterogeneous uptake of the antibody by the tumor tis(cid:173)
`sue [14).
`Tumor cell accessibility is a parameter that cannot easily
`be assessed in model systems. The many reports of com(cid:173)
`plete cures obtained after antibody treatment of nude
`mice transplanted with human tumors may convince
`the experimental novice, but not the skeptical clinician,
`as they frequently fail to work in patients. Marked dif(cid:173)
`ferences exist between the vasculature of spontaneous
`
`

`

`734 Cancer
`
`Table 2. Barriers and factors impeding free access of monoclonal
`antibodies to cancer cells growing in solid tumors.
`
`Heterogeneous or poor vascularization of tumors, reduction of total
`vascular surface area compared with normal tissue
`Elevated interstitial fluid pressure in tumor nodules
`Shallow or reversed transvascular pressure gradient leading to
`decreased transvascular .::onvection and diffusion
`Long transport distances for extravasated macromolecules in
`interstitium of tumors
`Radially outward directed interstitial fluid convection
`Basement membranes surrounding epithelial tumor tissue
`Shed or released tumor antigen present in peritumorous extracellular
`matrix
`lntercellular tight junctions in tumorous epithelia.
`
`Adapted from [111.
`
`autochthonous tumors and transplants of these tumors
`[ 10,11 J, indicating that xenotransplantation models pri(cid:173)
`marily measure antibody effector function.
`
`If accessibility of tumor cells is indeed a major reason
`for the overall disappointing results of the clinical tri(cid:173)
`als and if accessibility is negatively correlated with the
`volume of the tumor mass, the question must be asked
`whether therapy trials on patients with advanced malig(cid:173)
`nant disease, i.e. with bulky epithelial tumor masses or
`with leukemic or lymphoma cells in excess of 1012 cells,
`will ever show a therapeutic efficiency of antibodies.
`
`A much more appropriate target for assessing the efficacy
`of antibody therapy may well be minimal residual disease,
`a stage in which, after resection of all macroscopic tumor
`the remaining cancer cells are very few and dispersed as
`individual cells or small clusters in the interstitium of var(cid:173)
`ious distant organs.
`
`Minimal residual disease- a target within
`reach
`
`Minimal residual disease is present in roughly half of the
`patients with curatively resected solid tumors. Previously,
`the presence of hidden metastasis in these patients could
`be inferred only retrospectively from overt relapses oc(cid:173)
`curring several years after curative surgery. In the last few
`years, however, novel immunocytochemical methods that
`allow the detection of small numbers of carcinoma cells
`in bone marrow have become available [15,16.,17]. A<-,
`these cells do not express proliferation associated anti(cid:173)
`gens they appear to be in a state of dormancy [ 18, 19•].
`Several studies have now shown that the presence of
`these micrometastatic cells during early stages of tumor
`dissemination can serve as a strong predictor of a later
`clinical relapse [ 20 .. ,21,22].
`
`Because of their low number, their presence in mes·
`enchymal interstitium and lack of epithelial structures,
`these visible micrometastatic cells can be considered
`as ideal targets for therapeutic antibodies. Indeed, a
`previous study demonstrated that intravenously injected
`
`mAbs directed against a membrane-associated glycopro(cid:173)
`tein could be targeted to individual tumor cells in bone
`marrow [ 23].
`
`Therefore, with these deliberations in mind, a multicen(cid:173)
`ter randomized clinical trial involving 189 patients with
`resected colorectal carcinoma was initiated in 1985 and
`was completed in December of 1992 ( G Riethmuller, E
`Schneider-Gadicke, G Schlimok et al., unpublished data).
`Following surgery, the patients, all of whom had Dukes C
`stage carcinoma, were randomized to a control arm, i.e.
`observation only, and to a treatment group. The treat(cid:173)
`ment group received 500 mg of mAb 17- IA within two
`weeks of surgery followed by four subsequent monthly
`infusions of 100 mg of antibody. After a median follow-up
`of 5 years, therapy with antibody was found to have de(cid:173)
`creased the overall death rate by 30% and reduced the
`recurrence rate by 27%. These data contrast with the
`results of numerous, non-randomized trials with 17-IA
`antibody in advanced tumors where anecdotal remissions
`were observed only in a few patients and no benefit for
`survival could be secured. Interestingly, in this adjuvant
`study, the reduction in recurrence rate was found to be
`restricted to the development of distant metastases, while
`local relapses were not reduced by the treatment. This
`altered pattern of recurrences can be interpreted such
`that local satellite tumors were already too big and/or in(cid:173)
`accessible to the antibodv, in contrast with the distant
`micrometastases which ~ere destroyed by it. This trial
`shows that by carefully selecting the stage of tumor
`growth at which therapy is initiated, antibody therapy
`of colorectal carcinoma is comparable with other ad(cid:173)
`juvant therapies (Table 3) [24,25]. However, because
`of the remarkably low toxicity of unmodified antibody,
`this therapy can be administered to patients follo\\ing
`curative surgery without exposing them to the current
`hazards of adjuvant chemotherapies.
`
`As to the contentious issue of target antigens most suited
`for antibody therapy, it is notable that the antigen rec(cid:173)
`ognized by 17- IA mAb is by no means a tumor-specific
`antigen as it is widely expressed on various normal sim(cid:173)
`ple epithelia including, not only small and large intestine,
`but also bile ducts, kidney tubules and epithelial cells of
`thyroid and prostate [26]. The antibody, a murine IgG2a,
`has been administered to more than 300 patients with
`advanced disease [27]. Both the the lack of toxicity and
`efficacy (some minor transient gastrointestinal effects ex(cid:173)
`cepted) of doses of antibody up to 12 g may be due to
`the poor delivery of the antibody to cells shielded by a
`dense basement membrane and other vascular and inter(cid:173)
`stitial barriers [ 28]. Thus, one may arrive at the conclu(cid:173)
`sion that absolute tumor specificity of an antibody is less
`important than homogeneous expression of the relevant
`antigen on as many tumor cells as possible, as long as
`their normal counterparts and the stem cells from which
`they are derived are either less accessible or do not ex(cid:173)
`press the antigen. It appears from the reviews in Table l
`that numerous therapy trials have been performed with
`antibodies, immunotoxins and radioimmunoconjugates
`recognizing absolutely normal differentiation antigens,
`e.g. in B-cell lymphoma, without intolerable toxicity for
`the recipient. The 17- IA antibody has a remarkably low
`affinity and induces only intermediate antibody-depen-
`
`

`

`Table 3. Comparison of adjuvant therapies in colorectal carcinoma.
`
`Monoclonal antibodies in cancer therapy Riethmuller et al.
`
`735
`
`Therapy
`
`mAb 17·1A versus control
`Colorectal cancer. stage Ill
`Riethmi.iller el al., unpublished data
`
`Levamisole + fiuoroacil versus control
`Colon cancer, stage Ill
`Moertel et al. (241
`
`% Reduction in mortality rate % Reduction in recurrence rate
`(with 95% confidence interval)
`(with 95% confidence interval)
`
`31 (1-54)
`
`25 (1-45)
`
`33 (10-50)
`
`41 (23- 54)
`
`Radiation + fiuoroacil + methyl·CCNU versus radiation alone
`Rectum cancer, stage II or Ill
`Krook et al. (25)
`
`29 (7-45)
`
`34 (12- 50)
`
`dent cell-mediated cytotoXJoty (29,30] . Whether these
`peculiar characteristics of the antibody are essential for
`its clinical efficacy is unknown so far. However, an argu(cid:173)
`ment can be made that low affinity antibodies penetrate
`the solid tumor more deeply (30,31] . This argument is
`contended by Schlom et al. (32] who in contrast favor
`high affinity antibodies as more efficient therapeutics.
`
`murine antibodies. This technique allows the isolation
`of high affinity, antigen-specific Fabs or Fvs, even from
`naive human B cells [ 40 J. Furthermore, such antibodies
`may be generated from 'semi-synthetic libraries', which
`are produced by replacing the CDR3 region of a single
`human lg with random oligonucleotides [41•,42•] .
`
`New perspectives for antibody therapy
`
`If minimal residual disease, a stage so frequent in patients
`with the most common solid tumors, is such a promising
`target for antibody-based strategies, then a further refine(cid:173)
`ment of antibodies indeed makes sense.
`
`A5 the target patient population is quite healthy and as
`at least half of them are already cured by surgery and/ or
`local radiation therapy alone, the risk/benefit assessment
`of experimental therapies becomes critical. Thus, for the
`development of adjuvant therapies, unmodified antibod(cid:173)
`ies with their low toxicity profile have clear advantages
`over immunotoxins or radioconjugates. In order to ob(cid:173)
`tain steep transvascular concentration gradients towards
`mesenchymal tissue compartments, higher doses of an(cid:173)
`tibodies may be rc.-1uired which, in turn, will favor the
`induction of counterproductive immune responses in pa·
`tients. Indeed, in virtually all the trials cited in Table 1, hu(cid:173)
`man antibodies to the murine lg reagents were produced.
`Although the number of reported anaphylactic reactions
`has been low, it is clear that for prolonged therapy reg(cid:173)
`imens, the immunogenicity of antibodies should be as
`low as possible.
`
`A number of clinical trials have clearly shown that re(cid:173)
`placement of the Fe region with human sequences can
`substantially reduce the immunogenicity of murine anti·
`bodies (33-36] . The least immunogenicity is expected to
`be obtained when only the complementarity determining
`regions (CDRs) of the murine antibody remain, and re(cid:173)
`cent studies suggest that it may be possible to significantly
`simplify the production of these reagents (36,37•].
`
`The use of antibodies derived entirely from humans and
`isolated from combinatorial libraries in bacteriophage
`[38,39] will most likely soon replace such engineered
`
`Table 4. Cell-directed effects of unmodified antibodies.
`
`Activation or stimulation of cells
`e.g. Signalling via receptor aggregation, mimicry of agonists
`such as cytokines. hormones. adhesion ligands
`Inactivation of cells
`Negative signalling
`Blockade of functions of receptors or ion channels
`Modulation of receptors/adhesion molecules
`Induction of differentiation
`Elimination of cells by
`Complement mediated cytolysis
`Opsonisation. sequestration and phagocytosis
`Induction of apoptosis, directly via anti-Fas (Apo-I) antibodies
`or indirectly via antibody-dependent cellular cytotoxicity
`Induction of cytotoxic T-cells against murine lg. processed and
`presented by antibody labelled target cells
`Induction of anti·idiotypic antibodies (ab3) with anti-cellular
`activity
`
`The therapeutic efficacy of mAbs may be further in(cid:173)
`creased by a miniaturization of the antibody molecule.
`By linking the VH and VL sequences of such an antibody
`together on a single transcript, single chain antibodies
`(sFvs) can be produced. Because of their small size
`single-chain antibodies are deemed to penetrate more
`rapidly into tissues and interstitial spaces [ 43] . Single(cid:173)
`chain antibodies can be easily engineered and produced
`in bacteria. A variety of effector moieties such as toxins,
`cytotoxic drugs, growth factors, functional receptor do(cid:173)
`mains, and cytokines as well as lg Fe regions can be fused
`to these mini-antibodies.
`
`The type of linker used to couple the antigen-binding do(cid:173)
`mains to effector domains is also of critical importance.
`
`

`

`736 Cancer
`
`By coupling an anti-tumor antibody to do,cirubicin using
`a linker that is stable in plasma but acid labile and, there(cid:173)
`fore, set free after internalization in lysosomes, Trail et al
`[ 44• J were able to dramatically increase the effectiveness
`of the immunoconjugate.
`
`Unmodified antibodies, which may be much more rea(cid:173)
`sonable agents for treatment of minimal residual disease,
`rely on the various natural effector mechanisms of the
`host and the manifold ways they may interfere with cell
`function (Table 4). These functional characteristics of an(cid:173)
`tibodies are generally determined by their Fe receptors,
`which can also now be exchanged at will. Several of the
`antibodies used in clinical trials appear to work by acti(cid:173)
`vating human complement. Activation of complement in
`vivo has been observed to occur in patients treated with
`a IgG2a antibody directed against the ganglioside GD2
`[ 45•]; this was shown by a decrease in C4, C3c and C3a
`during treatment. In addition, an IgG3 antibody directed
`against the Lewis Y carbohydrate epitope, which has been
`shown to be very effective in the activation of human
`complement in vitro [ 46], has recently been shown to
`result in a reduction or eradication of antigen positive
`tumor cells in the bone marrow of five out of seven pa(cid:173)
`tients treated for two weeks with 6 x 100 mg antibody
`( G Schlimok, H Loibner, I Fackler-Schwalbe, K Pan tel,
`G Riethmuller, unpublished data). The efficiency of the
`complement cascade may now be further increased by
`blocking the membrane proteins CD59, C8bp and decay
`accelerating factor that control the activity of homologous
`complement components [ 47] .
`
`Another important anti-tumor effect of unmodified anti(cid:173)
`bodies is antibody dependent cellular cytotoxicity, which
`is mediated by various effector cells including neutrophils
`[48) . A recent study of 17-lA mAb in patients with ad(cid:173)
`vanced colorectal carcinoma suggests that cellular effec(cid:173)
`tor functions can be enhanced by the additional admin(cid:173)
`istration of cytokines such as granulocyte-macrophage
`colony-stimulating factor ( GM-CSF) [ 49•]. Cellular ef(cid:173)
`fector functions may also be directed to the target by
`antibody-coupled cytokines (50] or by bi-specific an(cid:173)
`tibodies designed to activate and orient cytotoxic cells
`to the tumor (51,52•].
`
`Antibodies recognizing particular epitopes on function(cid:173)
`ally important cell surface molecules may also be ef(cid:173)
`fective in tumor therapy, even without the engagement
`of conventional host effector mechanisms. For exam(cid:173)
`ple, antibodies directed to the Fas antigen have been
`shown to induce apoptotic cell death [ 53,54), and cer(cid:173)
`tain antibodies directed against the Her-2/ neu cell sur(cid:173)
`face receptor can induce differentiation of the tumor
`cells, which results in decreased growth rate both in
`vitro and in vivo ( 55•] . Antibodies against the lg idio(cid:173)
`type of B-cell lymphomas have been used in some of the
`most successful clinical trials [ 6•• ,8 00 ]. Such antibodies
`can also induce regulatory changes in experimental B(cid:173)
`cell lymphoma such that aggressive growth is abrogated
`and the tumor cells revert to a non-cycling dormant state
`(56•].
`
`Passive antibody therapy versus active
`immunization strategies
`
`A notable consequence of the general disappointment
`with antibody-based strategies is the recent surge of
`interest in active immunotherapy of cancer [57--60].
`The identification and cloning of tumor associated cell
`surface antigens as well as of peptides recognized by
`MHC restricted T lymphocytes open up the possibil(cid:173)
`ity of specifically immunizing patients against defined
`antigens. In addition, vaccination with genetically engi(cid:173)
`neered tumor cells that are transduced with lymphokine
`genes has yielded impressive results in transplanted tu·
`mar models [60), and recently a protective vaccination
`against B-cell lymphomas was shown to be improved
`when the lymphoma-specific lg was fused with GM-GSF
`(61) .
`
`Active immunization, however, relies on an intact im(cid:173)
`mune system and this is often compromised in advanced
`stages of cancer. Even more importantly, it requires that
`the target cell maintain MHC expression, proper antigen
`processing capability and the expression of any of a va(cid:173)
`riety of additional accessory molecules. However, loss or
`downregulation of such molecules is a common trait of
`human tumors. Most spontaneous human tumors and the
`micrometastatic cells found in minimal residual disease
`have lost expression of one or more MHC class I prod(cid:173)
`ucts [ 62,63]. Interestingly, this may even be an early event
`in some tumors as it is observed in about half of benign
`colorectal adenomas [ 64 .. ]. Human tumor cells defec(cid:173)
`tive in peptide processing and transport have also been
`identified [ 65•] and several studies now suggest that the
`expression of co-stimulatory molecules, such as B7, by
`the tumor cells may also be necessary for induction of
`immunity [66] . Clearly then, for active immunization to
`be effective, not only must the patient's immune system
`be more or less intact, but the tumor cells themselves
`need to express an entire array of gene products. These
`manifold and complex requirements for successful vacci(cid:173)
`nation stand in stark contrast with passive antibody ther(cid:173)
`apy, the only demand of which is that the tumor cells
`continue to express the target antigen.
`
`Outlook or "Jester do oft prove prophets"
`(King Lear)
`
`As long as the focus of current research is centered on
`the design of ever-new antibody constructs, employing
`the whole armamentarium of synthetic biology, and not
`centered on the judicious selection of more appropriate
`clinical targets and carefully designed therapeutic trials,
`one can foresee that another decade will be spent on
`'misguided missiles' [8 00 ] directed towards unassailable
`targets. For the adjuvant therapy situation, i.e. for the
`treatment of hidden metastatic cells, unmodified anti(cid:173)
`bodies or antibody derivates that rely on natural effector
`mechanisms offer clear advantages over immunotoxins,
`because the intended cytotoxic reaction will be restricted
`to the target site where the antibody has bound. More(cid:173)
`over, cellular and humoral components of effector sys-
`
`

`

`Monoclonal antibodies in cancer therapy Riethmuller. et al. ' 737
`
`terns may be decreased or even absent in normal tissues
`such as simple epithelia shielded by a dense basement
`membrane. Thus, despite extensive crossreactivity be(cid:173)
`tween benign and malignant tissue, an operational speci(cid:173)
`ficity may be achieved in vivo even with a broadly cross(cid:173)
`reacting anti-epithelial antibody. In addition, the use of
`such differentiation antigens with their more homoge(cid:173)
`nous expression on cancer cells may circumvent the
`formidable problem of antigenic heterogeneity so of(cid:173)
`ten encountered with more restricted or tumor-specific
`antigens.
`
`Humanization of rodent antibodies as well as generation
`of human mAbs from recombinant libraries will without
`doubt allow the best adaptation of therapeutic antibodies
`to the natural effector mechanisms. A major drawback of
`the naked antibody scenario is that it looks too simple
`and, therefore, runs against the current fashion for so(cid:173)
`phisticated immunoconjugates. Irrespective of the type
`of applied immunotherapeutics, the emphasis towards
`minimal residual cancer will require that the diagnosis
`of micrometastatic cells is refined. In the arduous area
`of adjuvant therapies, the pace of progress in immuno(cid:173)
`logical as well as chemical cancer treatment will critically
`depend on the availability of surrogate markers that al(cid:173)
`low a quick and reliable assessment of the particular
`therapeutic manoeuvre. The immunocytochemical diag(cid:173)
`nosis of micrometastatic epithelial cells in bone marrow
`of patients with various cancers is slowly gaining ground
`in the clinic [ 17]. Although the demonstration of their
`prognostic significance does not prove that they are
`the actual progenitors of later arising metastases, they
`clearly provide evidence for the disseminative capabil(cid:173)
`ity of an individual tumor. Furthermore, it has been
`suggested by Schlimok et al. [ 17 J that the elimination
`of such cells might give valuable information on the cy(cid:173)
`toreductive efficacy of a particular antibody. The further
`establishment of micrometastatic cells in bone marrow
`as surrogate targets can be envisaged as a crucial step
`towards a more rational design of immunotherapies of
`minimal residual disease. As long as primary prevention
`of cancer will remain an utopic goal the secondary pre(cid:173)
`vention of metastatic disease by immunological means is
`a worthwhile and realistic option.
`
`Acknowledgement
`
`the support of Deutsche
`The authors gratefully acknowledge
`Krebshilfe and Dr. Mildred-Scheel-Stiftung fur Krebsforschung, Bonn,
`Gennany.
`
`References and recommended reading
`
`Papers of particular interesc, published within the annual period of
`review, have been highlighted as:
`•
`of special interest
`••
`of outstanding interest
`
`I.
`
`2.
`
`KOHLER G, MHSfEIN C: Continuous Culture of Fused Cells
`Secreting Antibody of Predefined Specificity. Nature 1975,
`256:495--497.
`
`RIETiiMOIJ.ER G, JOHNSON JP: Monoclonal Antibodies in the
`Detection and Therapy of Micrometastatic Epithelial Can(cid:173)
`cers. Curr Opin lmmunol 1992, 4:647-655.
`
`3.
`
`4.
`
`5.
`
`PAI LI-I, PA.~TAN I: lmmunotoxin Therapy for Cancer. J Am
`Med A5Soc 1993, 269:78-81.
`
`HUSTON JS, MCCARTNEY J, TAI MS, MOTTOlA-HARsntORN C, JIN
`D, WARREN F, KECK P, OPPERMAN H: Medical Applications of
`Single-Chain Antibodies. Int Rev Immunol 1993, 10:19S-217.
`
`BRINKMANN U, GAU.a M, BRINKMANN E, KUNWAR S, PASTAN I: A
`Recombinant Immunotoxin that is Active on Prostate Can(cid:173)
`cer Cells and is Composed of the Fv Region of Monoclonal
`Antibody PRl and a Truncated Form of Pseudomonas Ex(cid:173)
`otoxin. Proc Natl Acad Sci USA 1993, 90:547-551.
`
`6.
`••
`
`GROSSBARD ML, PRESS OW, APPLEBAUM FR, BERNSTEIN JD,
`NADLER LM: Monoclonal Antibody-Based Therapies of
`Leukemia and Lymphoma. Blood 1992, 80:863-878.
`Extensive review of all clinical trials of unmodified as well as conjugated
`antibodies in the treatment of leukemia and lymphoma. In addition, a
`thorough discussion of the pitfalls and promises of this approach.
`
`7.
`•
`
`STEFFENS TA, BAJORIN DF, HOUGHTON AN: Immunotherapy
`with Monoclonal Antibodies in Metastatic Melanoma. World
`J Surg 1992, 16:261-269.
`A review of the Phase I trials of mAb treatment in melanoma patients
`with emphasis on antibodies directed to gangliosides.
`VITETTA ES, THORPE PE, UHR JW: Immunotoxins: Magic Bui-
`8.
`lets or Misguided Missiles. lmmunol Today 1993, 14:252-259.
`••
`A review of the completed or ongoing Phase I and Phase II trials using
`immunocoxin therapy with a discussion of the problems that have been
`encountered, including toxicity, immunogenicity and accessibility to the
`tumor cells.
`
`9.
`
`10.
`
`I I.
`
`I 2.
`
`13.
`
`14.
`
`I 5.
`
`LOBUGUO AF, SAJEH MN: Monoclonal Antibody Therapy of
`Cancer. Crit Rev Oncol 1992, 13:271-282.
`
`JAIN RK: Vascular and Interstitial Barriers to Delivery of
`in Tumors. Cancer Met Rev 1990,
`Therapeutic Agents
`9:253-266,
`
`JAIN RK: Determinants of Tumor Blood Flow: A Review.
`Cancer Res 1988, 48:2641-2658.
`
`GUTMANN R, LEUNIG M, FEYH J, GOETZ AE, MESSMER K,
`KA.5TENBAUER E, JAIN RK: