A Proposal for a New Direction to Treat Cancer

J. theor. Biol. (1998) 195, 111–128 Article No. jt980792

A Proposal for a New Direction to Treat Cancer S R Oncologic, Inc., 5920 San Pablo Avenue, Oakland, CA 94608, U.S.A. (Received on 16 April 1998, Accepted in revised form on 13 July 1998)

A new approach is proposed that has the potential to be a successful therapy for most disseminated cancers because it can circumvent the problems posed by three characteristics which are universally expressed by cancer cells: heterogeneity, plasticity, and the lack of a cancer specific or cancer associated characteristic which is not also shared by some normal cells. Analysis shows that almost all current and research approaches for treating disseminated cancers have the same fundamental strategy: they rely on an agent interacting individually and effectively with each cancer cell. We call all these approaches ‘‘lock and key’’ strategies to emphasize the need for this individual agent to cell interaction. The three characteristics preclude current approaches from successfully treating most disseminated cancers because they operate by a ‘‘lock and key’’ strategy which (a) only kills cancer cells expressing a single particular trait, (b) allows other cancer cells to adapt and survive the treatment, and (c) also kills the normal cells which express the same particular trait. The heterogeneity and plasticity of cancer cells can only be circumvented by an attack which is microregional (not cell by cell) and destructive (not killed by conventional endogenous or exogenous cytotoxic agents). All cells in each microregion must be destroyed, including those which do not express an exploitable trait. The proposed approach can achieve such microregional destruction by the delivery to, and long term immobilization of, a large number of radio-isotopes. The proposed approach exploits the additive contribution of multiple mechanisms to enhance tumor specificity of the microregions. Given that all targeting and killing agents are ‘‘imperfect’’, this is the only way specificity can be enhanced. The biological basis of these specificity enhancing mechanisms are well-known. However, they are ignored by current therapies because most of them can only be exploited in the context of the proposed approach. Some of the mechanisms reflect characteristics, such as heterogeneity, genetic instability, and tumor progression which are the result of the micro-evolutionary process of tumor development. These are virtually always present in, and virtually specific to, cancer. Others reflect the somewhat ‘‘imperfect’’ cancer associated characteristics of structures, including cancer cells, extracellular structures, and non-malignant cells, within the tumor mass. The additive contribution of the multiple mechanisms gives the process the potential to destroy all the cancer cells with minimal non-tumor toxicity. The cornerstone of the proposed approach is a novel class of soluble chemicals. They can be administered intravenously to subjects, circulate throughout body fluids and are enzymatically converted into an insoluble material when the chemicals reach targeted sites. In this paper, these chemicals are called ‘‘soluble precipitable reagents’’ (SPR) to describe their ability to be converted from a soluble to an insoluble material. The insoluble material is called platform to indicate that it has the ability to bind various agents. The SPR chemicals enable a three-step process to be constructed which can deliver and retain a large number of radio-isotope atoms in tumor tissue. 0022–5193/98/021111 + 18 $30.00/0

7 1998 Academic Press

. 

112

In step 1, a binary reagent comprised of an SPR attached to an imperfect cancer targeting agent is administered. The binary reagent is endocytosed and transported into lysosomes where the targeting agent moiety is digested and the detached SPR is converted by natural intracellular lysosomal enzymes into a platform. As will be discussed, a very large number of platform molecules can be made to accumulate inside targeted cells. In step 2, a supersensitive fraction of the cancer cells, including some which had accumulated platform in step 1, are killed by the administration of a very low dose of an anti-cancer agent. Very few, if any, normal cells will be killed by the very low dose. The death of the cells relocates the accumulated platform into the extracellular tissue fluid. After a lapse of time, any platform which is still retained in the extracellular tissue fluid can proceed to step 3. In step 3, the platform is used by one of two different methods to generate supra-lethal radiation fields. This is achieved in the first method by the binding and long term retention of a large number of isotope-carrying molecules to the extracellular platform. In the second method, a bispecific reagent, having a non-mammalian enzyme moiety is bound to the platform. The bound enzyme converts a subsequently administered free (non-targeted) radioactive SPR into an insoluble radioactive second precipitate which remains adjacent to the bound enzyme for a long time. Both methods, thus, result in the immobilization and long term retention of a large number of isotope atoms in the tumor tissue, thereby generating intense radiation fields which cause microregional destruction of thousands of neighboring cells. 7 1998 Academic Press

Introduction to the Cancer Problem    1. Problem posed by heterogeneity of cancer cells Fifty years of intense research has shown that there is a wide heterogeneity in every characteristic or trait that has been measured in cancer cell populations (Rubin, 1990). One manifestation of the heterogeneity is that only a fraction of the cancer cells in any tumor population express a trait which makes them vulnerable to any particular agent. Cancer cells not expressing this trait survive treatment by the agent and can become the dominant cell type in the tumor population. No single agent can be expected to target, treat or kill every cancer cell. Principle which must be satisfied to circumvent heterogeneity. Cancer treatment can only succeed if it is capable of killing all cancer cells in the tumor, including those that do not express an exploitable trait. Circumventing the problems posed by heterogeneity requires the attack to be microregional and not cell by cell. Theoretically, it may be possible that this constraint need not apply if multiple agents were used to match the particular and different traits present on each subset of the heterogeneous population of cancer cells. The combinatorial diversity of the immune system (Weiss & Rajewsky, 1990) may have the

potential to generate multiple different agents to match the different traits. However, apart from a few cancers, for example some hematological cancers, this potential is unlikely to be realized even if the immune response is augmented by vaccines and cytokines. Some cancer cells simply may not have an appropriate antigen to which the immune system can respond. Others may not express the antigen adequately through the major histocompatibility complex. Malignant cells, particularly metastatic foci when compared with the primary tumor lesion, are frequently observed to have absent or low expression of this histocompatibility complex (Cordon-Cardo et al., 1991; Elliott et al., 1989; Gopas et al., 1989). 2. Problem posed by plasticity of cancer cells Cancer cells, like all organisms, can adapt and become resistant, via multiple genetic and epigenetic (Prehn, 1994) mechanisms, to many untoward environments which can include an excess or deficiency of a natural agent in the system, or the addition of an artificial agent, such as a cytotoxic drug. Principle which must be satisfied to circumvent plasticity. Circumventing the problem posed by plasticity requires the attack to be so intense that it overwhelms all the cellular mechanisms,

      including those responsible for adaptation. In short, the cells must be destroyed. As will be discussed later in the paper, destruction of cells is unlikely to be achievable by soluble drugs or agents, even if they are produced locally within the tumor. Destruction of cells, however, can be readily achieved by immobilizing a large number of radio-isotope atoms in each microregion and retaining them in their immobilized location for an extended time. A recent approach (which will be described later) targets radiolabeled antibodies to extracellularly relocated DNA of necrotic tissues in an attempt to generate intense radiation fields and achieve microregional destruction. 3. Problem posed by normal cells expressing the same characteristics as cancer cells Cancer cells in any one tumor do not express an exploitable agent sensitive trait that is not also expressed by some normal cells (or that is not expressed in a form that prevents cross-reaction to the applied therapeutic agent). Unless a unique cancer specific trait is discovered, there will always be some normal cells that express the same trait as the cancer. This reflects the accumulating evidence that most ‘‘cancer associated traits’’ are not new functions, but rather, are due to a dysregulation of existing normal functions. For example, it has been amply shown that antigens expressed on melanomas reflect the differentiation state of their normal counterparts (Coulie et al., 1994; Houghton, 1994; Kawakami et al., 1994). The host response may represent an apparent autoimmune recognition to differentiation antigens that are expressed as a direct consequence of the malignant state, or perhaps because the malignant state is ‘‘seen’’ as a form of tissue injury. Principle which must be satisfied to circumvent shared characteristics. Since there is no ‘‘perfect’’ exploitable cancer trait or targeting mechanism, and since most scientists do not believe that one will ever be discovered, it follows that the only way to improve specificity is by exploiting the additive contribution of multiple ‘‘imperfect’’ mechanisms. The degree of specificity that can be achieved depends on the number of mechanisms and the effectiveness of each one of them. The practicality of exploiting multiple mechanisms

113

requires that they do not put an additional burden on the patient.    Disseminated cancer could be successfully treated if highly radioactive, insoluble, and non-digestible beads were accurately implanted in every cubic millimeter of the primary and metastatic tumor tissue. An intense radiation field would extend beyond each bead and destroy thousands of cancer cells in the immediate microregion. In this paper we define intense radiation fields and microregional destruction as fields which can destroy all cells in the microregion, including those not expressing any particular agent sensitive trait. Accurate implantation would also prevent the therapy from producing significant non-tumor toxicity. However, such an ideal implantation is not possible by physical and surgical methods, and is the only reason why bead implantation is not a successful treatment for disseminated cancers. On the other hand, soluble radioactive agents are distributed throughout the body fluids and can reach the tumor tissue virtually no matter where it is located. However, unlike implanted beads, soluble radioactive agents cannot achieve a high enough concentration in any one microregion, and are not retained for long enough, to generate the necessary intense radiation fields to cause cell destruction. Logic demands that a therapeutic process has the advantages of both soluble agents and insoluble beads. Such a logic is satisfied during the successful treatment of disseminated thyroid cancer by radio-iodide. Radio-iodide is a soluble agent which is distributed throughout body fluids, and is converted into a quasiinsoluble radioactive bead (colloid) by normal and malignant thyroid cells which enables it to be retained for a long enough time to generate intense radiation fields. Since the treatment of thyroid cancer faces the same cancer driven obstacles as any other cancer, and yet can often be treated successfully, it suggests that the same logic should be used to treat other cancers. The proposed approach uses a novel chemical to construct a three-step process which mimics, for non-thyroid cancers, the treatment of thyroid

. 

114

cancer with radio-iodide. Like radio-iodide, these novel chemicals are soluble when administered and precipitate by the catalytic action of enzymes in targeted cells. In this paper, these chemicals are called ‘‘soluble precipitable reagents’’ (SPR) to describe their ability to be converted from a soluble to an insoluble material. The insoluble material is called a platform to indicate that it has the ability to bind various agents.

Step 1 Continued administration of SPR binary reagent

The Proposed Three-step Approach Current therapies for disseminated cancer use soluble agents which have an immediate pharmacological effect on cells. In contrast, the strategy of the proposed approach is a three-step process which uses the SPR chemicals to create a biologically neutral, or relatively neutral, insoluble platform that is later used to create intense radiation fields by immobilizing and

Endocytosis

Enzyme digests targeting agent & converts SPR into platform

SPR chemical

Targeting agent

Platform accumulation inside cells increases with time

Binding site on SPR Number of platform molecules can be over 1000x greater than the number of receptors

Step 2 Administer very low dose of cell killing lysing agent

Cell membrane integrity lost

A large number of platform molecules from each cell are relocated to the extracellular tissue fluid Step 3 (first method) Administer isotope carrying molecules

Targeting moiety specific for binding site on SPR

Each platform binds and retains a large number of isotope carrying molecules

Generates supra-lethal radiation fields that extend beyond each platform to cause microregional destruction of thousands of cells

F. 1. Three-step oncologic process.

      retaining a large number of radio-isotope atoms in tumor tissue. The three-step process is depicted in Fig. 1. Step 1: a soluble binary reagent is administered intravenously to cause the accumulation of an insoluble platform inside targeted cells. The soluble binary reagent is comprised of the novel SPR chemical attached, either directly or via a lysosomal enzyme sensitive link, to a protein targeting agent for cancer cells. (Although the so-called targeting agents have some preferential ability to target cancer cells, they do not target all cancer cells and they also target some normal cells. The way in which the proposed approach circumvents this imperfection is described later.) The binding of the binary reagent to the specific receptors on the target cells activates receptor-mediated endocytosis which transports the binary reagent into the acidic, enzyme rich, lysosomal compartments inside the cell (Geuze et al., 1986; Murphy, 1988; Pearse & Bretscher, 1981) where the targeting agent is digested and the SPR is converted by lysosome enzymes into an insoluble platform. Step 1 of the proposed approach is a minor modification of published results which describe the delivery of a wide variety of different agents to the lysosomes of targeted cells using binary reagents and exploiting endocytosis, The delivered agents include cytotoxic drugs, toxins, dyes, antidotes to toxic drugs, non-digestible carbohydrates and molecules carrying radio-isotopes (Wu et al., 1983, 1985; Firestone, 1994; Rushfeldt & Smedsrod, 1993; Pittman et al., 1983; Jansen et al., 1993; Pittman & Steinberg, 1978; Pittman et al., 1979a, b). In all these published examples, the detached chemical in the lysosome remains soluble. In the proposed approach, the detached SPR is converted into an insoluble platform. There are two classes of SPR chemicals. The first class are inherently soluble. When they are detached from the targeting agent by lysosome enzymes, a second set of lysosome enzymes converts them into highly reactive intermediates which dimerize and create an entirely new molecule which is highly insoluble—no solublizing agent is required for this class of SPR. The second class of SPR chemicals are inherently

115

insoluble, but are solublized by their attachment to the protein targeting agent, and in some cases also by their attachment to a solublizing polymer. In the latter case, the solublizing polymer is either digestible, or is made of small segments of non-digestible polymers with intervening digestible segments so that the small non-digestible segments can escape from the cell. A number of candidate SPR chemicals are available in each class, and apart from their antigenicity, they have similar characteristics. This enables the proposed approach to sequentially use different SPR chemicals to avoid potential problems which might arise should the subject develop an immune response to any one of them, even though they are only administered for a short time and even though, unlike most other approaches, the treatment process of short duration. The candidate platforms made from the SPR chemicals are insoluble in both water and lipids, are relatively neutral to living cells, can act as an affinity matrix by having binding sites, and can be made to have an ordered structure, such as a linear polymer or micelle, so that the binding sites are spaced at defined distances to reduce steric hindrance. The binding sites can be chosen from a large library of candidates, and can be artificial (not found in mammalian fluid or cells) to ensure that the complementary ligand can bind to them with high affinity and specificity and with minimal binding to natural structures in the body. For the following reasons, a very large number of platform molecules can be made to accumulate inside cells: (a) the accumulation occurs via receptor-mediated endocytosis, which is a natural cellular process, (b) the accumulation proceeds via a time dependent, cumulative process, which can continue to operate as long as the SPR binary reagent is available, (c) the platform itself is non-digestible, and stable, which enables it to be retained (like the insoluble material in tattoos) inside the cell for a long time, and (d) the platform is insoluble and relatively non-toxic, which prevents it from osmotically or pharmacologically interfering with the ability of the cell to continue the accumulation process. The SPR can be prepared with a trace label in such a way that the platform retains this label.

116

. 

The long intracellular retention time of the platform enables the amount and location of its accumulation to be followed by scan and/or biopsy when there is a zero radiation background in the body fluids. Step 2: a very low dose of an anti-cancer agent is administered which kills a small fraction of the cancer cells that are super-sensitive. The death of these cells relocates the accumulated platform to the extracellular tissue fluid where it is retained. The dose of the agent is so low that very few (if any) normal cells are killed. As will be discussed later, this selective killing (together with platform accumulation) is one of the main mechanisms which determine the location of the final anti-cancer attack. It is not the main factor in eliminating the majority of the cancer cells. After a period of time is allowed to lapse, platform which is retained in the extracellular tissue fluid is available to proceed to step 3. Step 3: the platform in the extracellular fluid is used by either of two different methods as a launching pad from which the intense radiation fields are generated. In the first method, radiation fields are generated by binding a large number of cell impermeant isotope-carrying molecules to the artificial binding sites on the platform. In the second method, a bispecific reagent, having a non-mammalian enzyme moiety, is bound to the platform. The bound enzyme converts a subsequently administered free (non-targeted) radioactive SPR into an insoluble radioactive second precipitate which remains in situ adjacent to the bound enzyme for an extended period of time. Both methods of step 3, thus, generate intense radiation fields. The resultant microregional destruction circumvents the problems posed by the heterogeneity and plasticity of the cancer cells because all cells in the microregion are destroyed irrespective of their biological status. The binding of the isotope carrying molecule and the bispecific reagent to relocated extracellular platform is analogous to the in vivo binding of radiolabeled antibody to extracellularly relocated DNA (Chen et al., 1990), myosin (Khaw et al., 1984 ), and cytokeratin (O’Brien &

Bolton, 1995) when the cells were killed. However in these published results, the antibody binds to a natural, digestible structure of the body and is removed by natural processes. In contrast, the agents administered in step 3 bind to the sites on the artificial platform which is stable and non-digestible and for this reason the binding can be more specific, have a higher affinity, and can remain bound for a much longer period of time. This more stable binding is analogous to the stable binding of antibodies to slow turn-over extracellular structures such as the glomerular membrane of the kidney where the antibody remains bound for months (Yong & Rhodes, 1990). The Proposed Approach Mimics the Successful Treatment of Thyroid Cancer with Radio-iodide Even though some thyroid cancer cells do not have the trait that enables them to take up and store the radio-iodide therapeutic agent, the treatment of thyroid cancer with radio-iodide is successful in a high percentage of cases (Adelstein & Kassis, 1987; Fitzgerald et al., 1950). Since the treatment of thyroid cancer essentially faces the same problems as other cancers and yet can often be successfully treated, the proposed approach was designed to mimic the operating conditions that operate during this treatment. The SPR technology enables the three-step process to replicate, for non-thyroid cancers, the operating conditions that enable the treatment of thyroid cancer to succeed. 1. The process is similar In both the thyroid and the proposed approach, a soluble agent is administered which distributes throughout the body fluids and can reach the disseminated cancer tissue where it is converted via multiple steps and natural intracellular enzymes into a relatively stable material (slow turn-over colloid in the thyroid, and insoluble platform in the proposed approach). In both cases the accumulation proceeds in a time dependent cumulative manner and can continue for as long as is necessary. In the thyroid model, the accumulated colloid is immediately radioactive. In the proposed approach the platform is not immediately radio-

      active (except trace labeled) and only later becomes radioactive by either of the two methods described for step 3 of the process. 2. The quantitative parameters that achieve microregional destruction are similar The ability to achieve intense radiation fields and microregional destruction is determined by the number of radioactive isotope atoms which are deposited in each microregion and their retention time in this location. Calculations (in part from preliminary data) show that the first and second method of step 3 can achieve similar conditions as those which enable the treatment of thyroid cancer to create intense radiation fields. In the thyroid model, assume that there are 106 thyroid cancer cells per cubic millimeter, that approximately 50% of the cancer cells can take up the radio-iodide, and that each of these cells can take up and convert 104 radio-iodide atoms into a colloid. There will be 5 × 109 isotope atoms per cubic millimeter, thereby generating a radiation field of approximately 250 mCi per gram of tissue. The time taken for this uptake and conversion is approximately 2–4 hr, once taken up, the isotopes are retained with a biological half-life of a few days. Since isotope continuously leaves the cancer, systemic toxicity is produced both during the uptake and release process. In the proposed approach, the number of isotope atoms which are ultimately deposited depends on how much platform is accumulated inside cells in step 1 and how many isotope atoms are immobilized in step 3. As discussed, a very large number of platform molecules can be made to accumulate inside cells in step 1. The accumulation of the platform reflects the memory of a cell’s biological activity over time. Preliminary results have shown that (a) the continued administration of the SPR binary reagent results in the accumulation of a large number of platform molecules inside targeted cells, (b) the process of forming and accumulating the platform inside cells was relatively not toxic, and (c) the platform is retained inside cells for a very long time. The hepatoma cell line, HepG2, known to express the specific asialoglycoprotein receptor, was grown in tissue culture

117

medium. The HepG2 cells were cultured in duplicate cultures for 10 days in plastic ware (Falcon) at 37°C under 5% carbon dioxide and 95% air in medium containing 5 mM concentration of the lactosilated polylysine-SPR. At the end of the culture period, the cells were washed three times in balanced salt solution and harvested. The cells were incubated with 0.1 normal sodium hydroxide for 30 min at room temperature, then dissolved in liquid scintillation fluid, and finally centrifuged in 2 ml centrifuge tubes. The insoluble material was seen as a small pellet made up of small particles approximately 0.1 mm in diameter. Calculations from this experiment showed that each targeted cell accumulated 1000 times as many molecules of SPR-insoluble material as the number of specific asialoglycoprotein receptors. The results above show that over 10 days, targeted cells can accumulate, in their lysosomes, 1000 times as many molecules of platform as there are specific receptors on the surface of the cell at any given time. Assuming that there are 106 cancer cells per cubic millimeter, that only 5% of cancer cells both accumulate platform during step 1 and relocate platform in step 2, that each cell has 104 receptors (a conservative assumption), and that the receptor recycle rate is 5 per hour. Using these assumptions, calculations show that each of these cells can accumulate at least 107 platform molecules in 200 hr (104 receptors multiplied by receptor turn-over rate of 5 per hour multiplied by 200 hr of accumulating time) and therefore there would be 5 × 1011 platform molecules per cubic millimeter of tumor tissue. Assume that only 10% of the platform molecules can bind the agent administered in the first or second method of step 3. In the first method, the platforms would bind 5 × 1010 isotope atoms per cubic millimeter which will remain bound for longer than the retention time of radio-iodide during the treatment of thyroid cancer. Given both of these parameters, the first method of step 3 has the potential to generate a radiation field that is at least 10 times as intense as in the thyroid model. The second method of step 3 (compared with the first) has several advantages: (a) the second method can generate radiation fields that are

118

. 

very much more intense because the number of isotope atoms which are immobilized is proportional to the size of the platform (like the first method) multiplied by the turnover number of the enzyme (the turnover number of an enzyme is the number of soluble molecules that the enzyme can convert into an insoluble precipitate per unit time); (b) the isotope atoms remain in their immobilized location for a longer period. Even if the radioactive precipitate is phagocytosed by macrophages, it will be retained inside the macrophage cells within the tumor mass. It is likely that the movement of the ‘‘hot’’ macrophages within the tumor will distribute the radioactivity more evenly throughout the tumor which would enhance the probability of destroying more cancer cells in the tumor (macrophages in tissues are not thought to return to the general circulation, though they can move to the regional lymph glands). In contrast, in the first method, if the platform with its bound isotope carrying molecule is phagocytosed by macrophages, the attachment of the isotope to the platform will be disrupted by the combined effect of low pH and the action of lysosome enzymes. This will cause the isotope to escape from the cell thereby decreasing the intensity of the radiation fields in the tumor and increasing the systemic radiation toxicity; (c) furthermore, less systemic toxicity is likely to be produced by the second method because it is much easier to make a small radioactive SPR which can be converted into a precipitate, compared with making a small isotope carrying molecule which will bind to the platform. The candidate radioactive SPR has ideal characteristics of pool distribution and excretion kinetics because it is small and can only be converted into an insoluble precipitate by the catalytic action of the non-mammalian enzyme moiety of the bound bispecific reagent. An analogue of this candidate SPR is known to circulate in the blood without causing significant systemic toxicity. The radiation fields generated by the first, and particularly by the second method, of step 3 can be hundreds or even thousands of times more intense than those produced when treating thyroid cancer. The fields can be intense enough to cause microregional destruction (radionecrosis) of thousands of cancer cells in each

microregion irrespective of their biological status and even if they have not proceeded through steps 1 and 2 of the process. In this way the proposed approach mimics the killing of those thyroid cancer cells that do not take up radio-iodide by the radiation fields generated by the thyroid cancer cells that do take up the isotope. Microregional destruction is the only way to defeat the problems posed by the heterogeneity and plasticity of the cancer cell population, and it circumvents the necessity of finding a uniquely exploitable trait common to all cancer cells. 3. Achieving insignificant systemic radiation toxicity Even though radio-iodide atoms circulate in the blood during uptake and are slowly released from the colloid during the treatment of thyroid cancer, only insignificant systemic radiation toxicity is usually produced because radio-iodide has ideal pool distribution and excretion characteristics. Although the isotope carrying molecule used in step 3 is unlikely to have the same ideal distribution characteristics as radioiodide, this deficiency can be compensated for by the other more advantageous parameters. It is unlikely that the proposed approach will produce significant systemic toxicity for the following reasons: (a) the large number of binding sites on the platform increases the rate of transfer of the isotope carrying molecule from the blood and its immobilization on, or adjacent to, the platform (binding to the platform in the first method and conversion to a precipitate in the second method); (b) in both methods, the long retention time of the immobilized isotopes increases the ratio of tumor vs. systemic radiation dose—particularly if an isotope with a relatively long physical half-life is used; (c) in both methods, the isotope carrying molecule can be chosen from a wide variety of candidates to have the best pool distribution and circulation kinetics. 4. Overlapping microregions of radiation Overlapping microregions of destruction are required to eliminate all the cancer cells in the tumor. In the thyroid model, the microregions often overlap because many thyroid cancers are

      sufficiently well differentiated that cancer cells with positive iodide trapping traits are close enough together so that the fields they generate overlap. Achieving this necessary overlapping for most solid tumors treated by the proposed approach may require a second administration of the isotope-carrying molecule which will produce a second round of radiation fields. For the following reasons, the number of second round radiation fields is likely to be much greater than the first round: (a) all cancer cells which have accumulated platform and which are in the outer peripheral rim of the first round of radiation fields will be destroyed, and as a consequence will relocate their platform to the extracellular tissue fluid of the rim; (b) there will be no convective flow or viable macrophages in the rim because blood capillaries, lymphatics, and macrophages in the microregion and rim will, like the cancer cells, also be destroyed. As a consequence, the platforms relocated in the tissue fluid of the rim will remain in situ for an extended time because they cannot easily ‘‘escape’’ from the extracellular tissue fluid; (c) although insoluble particles and platform can only move by convective flow, small soluble isotope carrying molecules can move by diffusion into the convective free rim. The ability of the proposed approach to generate a second round of radiation fields greatly increases the chance that the microregions will overlap and that the therapy will destroy all the cancer cells in the tumor. One can view the few supersensitive cancer cells which are biologically killed in step 2 of the process, and even the cancer cells that are destroyed in the first round of radiation fields, as simply being ‘‘markers’’ for the main destruction achieved during the second round. Since thousands of cancer cells are destroyed around each platform in the first and second round fields, and since the number of second round fields will be greater than the first ones, the therapy is not compromised even if the number of cells that both accumulate platform in step 1 and are supersensitive and killed in step 2, is small. The practical ability to carry out second round radiation fields depends on producing a low level of systemic radiation toxicity in the first round, and this in turn, largely depends on the

119

characteristics of the isotope carrying molecule and how rapidly these molecules are immobilized. Second round fields are analogous to the second administration of radio-iodide which is occasionally necessary to treat thyroid cancer. 5. The thyroid model and the proposed approach achieve tumor specificity in different ways In the thyroid model, the radiation fields are located specifically and exclusively in normal and malignant thyroid tissue because radio-iodide is a ‘‘near perfect’’ and exclusive targeting agent for this tissue. For this reason, the proposed approach cannot mimic the method by which the thyroid model achieves the location specificity of the radiation fields. Tumor Specificity of the Proposed Approach Since there is no equivalent ‘‘near perfect’’ targeting mechanism for non-thyroid cancer (most scientists do not believe that such a mechanism will ever be discovered) the proposed approach uses the additive contribution of multiple mechanisms to improve the tumor specificity of the radiation fields. These mechanisms, which do not put a toxic burden on the patient, operate in each step of the process to maximize the differences in the number and size of the platforms in tumor tissue vs. normal tissue prior to step 3. The difference in the amount of platform available in step 3 is determined by where and how much platform is accumulated in step1, where and how much is relocated to the extracellular tissue fluid in step 2, and where and how much is retained in the extracellular tissue fluid prior to step 3. Finally, for any given platform distribution, the specificity of the radiation fields is determined by the specificity of the isotope immobilizing process.  1           Preferential accumulation of platform can be achieved in cancer cells that over-express high affinity binding endocytosing receptors. The micro-evolutionary process and tumor progression (Cheng & Loeb, 1993; Foulds, 1969, 1975; Hill, 1990) make it likely that cancer cells

120

. 

having such receptors for growth promoting or survival ligands will become a dominant cell type. The preferential accumulation of platform can be further enhanced if the binary reagent is administered at a continuous low dose infusion over time. This simulates the low concentration of natural circulating ligands in body fluids which is a major factor contributing to their specificity. The low dose of the binary reagent is analogous to the phenomenon of low dose administration of antigen resulting in the immune system generating high binding antibodies because the antigen preferentially binds to, and causes the proliferation of, immune cells expressing high binding receptors for the antigen. The low concentration is feasible in the context of the proposed approach because the platform can be made to accumulate slowly and is only used later when sufficient platform has accumulated and has been retained. In contrast, the agents used in lock and key strategies must be given at a high dose to achieve a relatively high concentration so that they will exert their immediate pharmacological action.  2           A very low dose of an anti-cancer agent is administered in step 2. This low dose will kill the small fraction of cancer cells which are super-sensitive and causes the SPR-platform to be relocated to the extracellular tissue fluid. The dose of the agent is so low that very few, if any, normal cells will be killed. The goal of the selective killing (together with platform accumulation) is to be one of the markers for the location of the final anti-cancer attack. It is not the main factor in directly killing the majority of the cancer cells. The presence of these super-sensitive cells in a tumor, and not in normal tissues, reflects the genetic instability (Loeb, 1998) and resultant heterogeneity (Rubin, 1990) of the cancer cells (features which are considered to be the hallmarks of the disease). The presence of super-sensitive cancer cells may be one of the most common features of tumor populations, but they are totally ignored by current research,

presumably because these ‘‘easy to kill’’ cells are of no value in the context of lock and key approaches in which the problem of resistant and multi-drug resistant cells is a critical problem. However, when used in the proposed approach as a tumor specific marker determining the location of the anti-cancer attack, the presence and ‘‘easy killing’’ of these super-sensitive cancer cells may be the most exploitable feature of cancer. In fact, the administration of low doses of current therapeutic agents can be considered as a non-toxic enhancement of the ongoing natural killing of super-sensitive cancer cells by the defense systems of the body (Wyllie, 1992). In the proposed approach, the tumor specificity depends primarily on the radiation fields being generated only if the same cancer cell both accumulates platform in step 1 and relocates the intracellular platform to the extracellular tissue fluid in step 2. This double requirement should enhance specificity because it requires the same cell to have two different and independent traits. The first trait is the expression of endocytosing receptors that enable the cell to accumulate platform, and the second trait is the cell’s super-sensitivity to being killed by low doses of anti-cancer agents which relocate the platform to the extracellular tissue fluid. As described, the therapeutic process is not compromised if there is a very low number of cells with this double set of characteristics because it is possible to generate a second round of radiation fields. The agents used in step 2 of the proposed approach cause lysis, which destroys the integrity of the cell membrane and results in the intracellular contents, including the platform, to be relocated to the extracellular tissue fluid (which is the only location where the platform can generate radiation fields). Therefore, lysed cells can generate radiation fields. In contrast, the death of normal cells which occurs by apoptosis as part of their natural cellular turn-over does not cause relocation of the platform to the extracellular tissue fluid. Apoptosis in vivo results in the formation of impermeable membrane-bound vesicles which are rapidly engulfed by adjacent parenchymal or professional phagocytes (Stewart, 1994; Wyllie, 1992). This prevents the release of their intracellular constituents, including accumulated

      platform, into the extracellular space and prevents radiation fields from later being generated. Radiation fields are not generated even if normal cells have accumulated platform and have died as a result of their natural cell turnover. Thus, apoptotic death of normal cells will tend not interfere with the tumor specificity of the radiation fields.      3                      Normal epithelial cells exfoliate into the external environment when they die or are killed (Croitoru & Riddell, 1993). This exfoliation has been demonstrated for epithelial cells of the gastrointestinal tract (Ishikawa et al., 1993), kidney (Goligorsky et al., 1993; Walker, 1994), bladder (Rebel et al., 1994), lung (Bogdanffy et al., 1994; Finotto et al., 1993), and others. Exfoliation causes dead normal epithelial cells and their intracellular contents to be discharged into the external environment where they obviously cannot generate radiation fields. Therefore, even if any normal epithelial cells accumulate platform in step 1 (because they have the same endocytosing receptor as the cancer) and are killed in step 2 (because they were super-sensitive), they will not produce radiation fields. In contrast, epithelial cancer cells (with a few exceptions, such as superficial bladder cancers), are not located on the boundary between the internal and external environment and they cannot exfoliate into the external environment. When these epithelial cancer cells are killed, they relocate their accumulated platform to the extracellular tissue fluid (this being the only location that enables radiation fields to be generated). Similarly, killed endothelial cells are discharged into the blood stream and not into the extracellular tissue fluid (Baillie et al., 1995; Yuzawa et al., 1994). Once in the blood stream, these cells and any platform they had previously accumulated, will be quickly engulfed by the phagocytic systems of the lung, liver, and spleen. In this intracellular location, the platform will not be able to generate radiation fields.

121

It should be noted that endothelial cell damage and death are common and serious side-effects of immunotoxins (Soler-Rodriguez et al., 1993) and of interleukin-2 therapy (Fujita et al., 1994). The following example demonstrates the contribution of epithelial cell exfoliation and endothelial cell discharge into the blood stream in determining the tumor specificity of the proposed approach. An immunotoxin was made by attaching a pseuodomonas exotoxin A (PE) with its cell receptor binding portion deleted, to an antibody, MAbB3 to make a binary reagent. The MAbB3 binds to 90% of metastatic colon cancer but does not react with normal colon. It also reacts with other metastatic cancers including many breast, ovarian, esophageal, stomach, bladder and prostate cancers. The antibody is not entirely cancer specific and reacts with a limited number of normal epithelial tissues including, glands of the stomach, the superficial epithelium of bladder, trachea, and esophagus. The immunotoxin could cause toxicity in these normal tissues. However, in the proposed approach, the same MAbB3 antibody could be used to make the SPR binary reagent to accumulate platform in step 1. As described above, even the platform accumulation occurred in these normal cells, and even if some of these normal cells were killed in step 2, the specificity of the therapy would not be compromised. Similarly the antibody portion of the immunotoxin may also react with B3 antigens on endothelial cells and cause vascular leak syndrome. However, as described above, the proposed approach would not be compromised because killed endothelial cells are discharged into the bloodstream and the released platform would be quickly phagocytosed by the macrophages in the liver, lung and spleen. Once inside these macrophages, the platform would not be able to generate a radiation field in step 3. A review of immunotoxins has recently been published which describes and references many targeting antibodies that have been used as binary reagents to deliver toxins to cancers (Pastan,1997). It should be noted that many of these antibodies could also be used to deliver the SPR without causing the difficulties experienced when they are used to deliver toxins.

122

. 

In addition, platform which has been relocated to the extracellular fluid of cancer tissue will be retained in this location for a longer period compared with the extracellular fluid of normal tissue. This difference in retention time creates a temporal window that can be exploited to enhance the tumor-specificity of the radiation fields by appropriately timing the initiation of step 3. The lack of lymphatics in cancer tissue explains why trypan blue (which binds to albumin) injected into tumor tissue is retained in situ for up to 5 days, compared with only hours when injected into normal tissues (Matsumura & Maeda, 1986). Given the longer retention of the soluble macromolecules in this example, it can be predicted that, compared with normal tissue, the retention time of the insoluble platform in the cancer tissue is likely to be much longer. Convective flow in normal tissues transports injected particles into the lymphatic channels and then to the regional lymph glands where they are effectively engulfed by the resident macrophages lining the sinusoids (Ikomi et al., 1995). By analogy, platform that has been relocated to the extracellular tissue fluid will also be transported to lymph glands and be engulfed inside the macrophage cells where the now intracellular platforms cannot generate radiation fields. In contrast to normal tissues, the lack of lymphatics in cancer tissue (Fallowfield & Cook, 1990; Jain, 1987; Seymour, 1992) reduces the possibility of this transportation and engulfment process, and as a consequence, the platforms will be retained in cancer tissue for a much longer time than in normal tissue. All the specificity enhancing mechanisms described above operate without adding any extra burden to the patient and without administering any additional agents. The proposed approach uniquely benefits from the body’s natural functions, and particularly those functions which have been altered by the presence of the tumor. The combined effect of these multiple mechanisms results in the presence of a very large amount of platform in the extracellular fluid of tumor tissue, and very little, if any, platform in the extracellular fluid of normal tissues at the time the isotope carrying molecule is administered in step 3.

     3                         Specificity is not degraded by the first method of step 3 because the binding sites on the platform are foreign to mammalian structures and the ideal binding sites can be chosen from a wide variety of candidates. For these reasons, it is possible to make the isotope carrying molecule so that it will bind with high specificity and affinity to the sites on the platform, without binding significantly to cells or structures in the body. As a consequence of this exclusive binding, the radiation fields will be generated only around the platforms. Since the platforms will be located preferentially, and perhaps virtually exclusively, in the tumor, the radiation fields themselves will be located virtually exclusively in the tumor. Specificity is not degraded by the second method of step 3 because the radioactive SPR is only converted into a radioactive precipitate by the non-mammalian enzyme moiety of the bispecific agent, which itself is bound only to the platform.          -  In the proposed approach, the tumor specificity achieved by the natural mechanisms can be greatly enhanced by the administration of a variety of non-toxic agents. For example, targeting molecules specific for homogeneous receptors (for example, cell lineage receptors N) on a set of normal cells can be used to deliver a variety of non-toxic agents that inhibit one or more of the essential steps of the radiation generating process from operating. Because the normal cells of any one tissue are relatively homogeneous and express N, they can be targeted and the radiation generating process can be inhibited in most of these N positive cells. Specifically, the administration (prior to step 2) of an enzyme targeted to N can cleave the binding epitopes on the platform inside these cells. This would prevent the agents in step 3 from binding to the platform and thus would prevent radiation fields from being generated.

      Alternatively, these targeting molecules could be used to deliver, to N positive cells, an antidote to the low dose of the anti-cancer agent administered in step 2. This would protect N positive cells from relocating their platform to the extracellular tissue fluid in step 2, which in turn, would prevent radiation fields from being generated around them in step 3. Used in this way, targeting molecules could prevent the radiation generating process from operating in most normal cells. This could markedly reduce non-tumor toxicity. Even if a few of the heterogeneous cancer cells have the receptor N and the radiation generating process was inhibited in these few, the therapeutic process would not be compromised in its ability to destroy all cancer cells because each radiation field destroys thousands of cancer cells, and because the frequency of second round radiation fields is much higher than first round fields. Thus, the therapeutic improvement achieved by preventing radiation fields in normal tissues can be gained without compromising the ability of the process to destroy all the cancer cells. The strategy of preventing radiation fields in normal tissues exploits the marked homogeneity of normal cells compared with the marked heterogeneity of cancer cells. Lock and key strategies cannot use the strategy of protecting normal cells because some cancer cells would also be protected, and, without microregional destruction, such protection would be counterproductive. For example, folinic acid is used as an antidote to systemic methotrexate. This does protect normal cells and reduces systemic toxicity, but it is highly likely that the folinic acid would also protect some cancer cells. Specificity of the proposed approach can be further enhanced because it can exploit virtually universal and specific characteristics of cancer cells such as genetic instability (Loeb, 1998) and tumor progression (Foulds, 1969, 1975). The proposed approach can exploit these characteristics, as described below, because the platform is stable and remains inside lysosomes for an extended period of time. At the beginning of the therapeutic process, some cancer cells and some normal cells express the endocytosing receptor X (which is responsible for the accumulation of the

123

intracellular platform) and the platform will contain a binding site Y (to which the isotope carrying molecule or the bispecific reagent in step 3 can later bind, and which can only be cleaved from the platform by a non-mammalian enzyme). After the intracellular accumulation of platform is completed, and after a period of time has been allowed to lapse to allow the genetic instability (Loeb, 1998) and the ‘‘tumor progression’’ of the cancer cells to be expressed, (Cheng & Loeb, 1993; Foulds, 1969, 1975; Hill 1990; Loeb, 1998), it is likely that some cancer cells will have arisen which have lost X. However, since normal cells are genetically and epigenetically stable they will continue to express X. Prior to step 2, targeting agents which are X specific can be used (as described above) to deliver agents (enzymes, antidotes, etc.) to X positive normal cells and X positive cancer cells but not to the few X negative cancer cells. Either of these delivered agents will prevent radiation fields from being generated around X positive cells while still permitting radiation fields to be created around X negative cells that were originally X positive. Exploiting the unique change in the few cancer cells from being X positive to X negative will greatly increase tumor specificity. Although the number of cells losing the receptor X in a relatively short time is likely to be very small, their location should be highly specific to the cancer. The potential problem posed by the low frequency of radiation fields in the cancer can be circumvented by the generation of a large number of second round of radiation fields. For example, X can be used to deliver an agent which protects X positive normal cells from being killed in step 2. In this scenario, the binding sites on the platform in cancer cells will be intact, the first round radiation fields will destroy thousands of cells which will relocate their platform to the extracellular tissue fluid. These relocated platforms will be able to generate second round radiation fields. In this way, virtually all cancer cells which had accumulated platform in step 1 will be able to generate second round fields. Thus, tumor specificity and overlapping radiation fields are both achievable. Lock and key strategies attempt to achieve specificity by finding an agent that acts effectively

124

. 

on a large number of cancer cells and only on a small number of normal cells. The proposed approach is not limited by these constraints. For example, if five different agents are available to target cancer cells or to kill them, the proposed approach can select and use the agent even if it is specific to only a few cancer cells and even if it is specific to a large number of particular normal cells that cannot generate radiation fields. These latter normal cells include epithelial or endothelial cells that inherently cannot generate radiation fields, and/or those normal cells which have a common homogeneous receptor, such as the common receptor on B cells (Scheinberg, 1995) or hepatocytes (Geuze et al., 1986), which can be targeted to inhibit the radiation fields from developing around them. Nonetheless, as described earlier, even if a few cancer cells are also protected the final therapeutic outcome is not compromised. The aggressiveness and tumor specificity of any therapeutic process or approach is greatly influenced by the various intra-tumor structures. The characteristics of intra-tumor extracellular structures and non-tumor cells are influenced by the characteristics of the cancer cells. Since the latter are heterogeneous, it is likely that the characteristics of the former will also be heterogeneous. As a consequence, any process, agent, or mechanism that is effective in one location of the tumor may be ineffective in another location of the same tumor. In order to kill cancer cells in sub-optimum locations, for example areas of poor blood supply or having a particular cytokine microenvironment, the therapeutic process must be able to generate an attack that is many times in excess of what is necessary to kill cancer cells in therapeutically optimum locations. The proposed approach can achieve radiation fields that are sufficiently intense to satisfy this excess requirement as shown by the calculations described earlier and similarly, ‘‘excess’’ tumor specificity can be readily achievable. It is important to note the proposed approach exploits multiple mechanisms to enhance tumor specificity in each step, some of the mechanisms operate naturally and others require non-toxic intervention. Since the final outcome of the therapy is the result of the individual effective-

ness and the additive contribution of the mechanisms, the proposed approach has the potential to be highly specific and to be capable of future improvements. This emphasizes one of the critically important features of the proposed approach: multiple steps provide multiple opportunities to improve the final result without putting a burden on the patient. In this way, the proposed approach is analogous to the multiple mechanisms the body uses to achieve a function in a specific and controlled manner. In contrast, the therapeutic outcome of lock and key strategies that operate by a single step cannot be improved significantly. Discussion Three strategies have been developed which attempt to circumvent the obstacles facing lock and key strategies. They operate by converting targeted cancer cells into ‘‘production sites’’ for soluble cytotoxic agents which diffuse into the immediate microregion and attempt to kill neighboring non-targeted cancer cells. These strategies include: antibody dependent enzyme pro-drug therapy or ADEPT (Rodrigues et al., 1995; Smith et al., 1997), vector dependent enzyme pro-drug therapy or VDEPT (Trinh et al., 1995), and therapy by gene insertion designed to over-produce natural peptides (Weichselbaum et al., 1994). These three strategies represent a recognition of the need for the anti-cancer attack to kill cancer cells that cannot be targeted. However, these approaches are actually all special cases of lock and key strategies and for the following particular reasons they are likely to fail: (a) since the soluble agents diffuse into the blood and cause a level of systemic toxicity which is proportional to the toxicity produced locally, the local concentration cannot be made to exceed a certain level. This level must be lower if the tumor is large because the larger the tumor, the greater the number of drug producing locations; (b) it is likely that some cancer cells will be resistant (and even adapt to become super-resistant) to the marginally higher concentration of the locally produced cytotoxic agents; (c) the cells on which the agent is generated (ADEPT), or the cells which produce the cytotoxic agent (VDEPT),

      will be exposed to the highest concentration of the agent and are likely to be among the first cells to be killed which prevents further production of cytotoxic agent; (d) the location of the attack cannot be made sufficiently specific because the first step of each of these strategies uses a single ‘‘imperfect’’ targeting agent to convert them into ‘‘production sites.’’ Like the proposed approach, a recent strategy, called tumor necrosis therapy (TNT), has been developed which recognizes that it is impossible for any therapy to succeed if it depends on killing cancer cells which have a particular characteristic or trait. TNT also recognizes that destruction of cells can best be achieved by immobilizing a large number of radio-isotope atoms in each microregion of the tumor and retaining them in their immobilized location for an extended time. The TNT binds a radiolabeled antibody to extracellularly relocated DNA of necrotic or dying tissue (Chen et al., 1989, 1990). Data from clinical trials have produced encouraging results and have shown that TNT can achieve a radio-isotope uptake by the tumor 5–8 times larger than radiolabeled antibodies targeted to cell receptors as in conventional approaches. Instead of using naturally present DNA as in TNT, the proposed approach uses an artificial platform which has previously accumulated inside targeted cells. The following analysis suggests that this seemingly minor difference enables the anti-cancer attack in the proposed approach to be much more aggressive, more tumor specific, and to have the potential to keep improving: (a) compared to TNT, the proposed approach has the potential to deposit a larger number of radio-isotope atoms in the tumor and to retain them for longer. The number of radio-isotope atoms which can bind to the DNA is proportional to the fixed amount of this material in the cell. The retention time of the isotopes is limited because the DNA is a digestible material. In the proposed approach, the number of isotopes which bind is proportional to the number of platform molecules, and as has been described, this number can be made very large because it is proportional to a time dependent cumulative process which can continue until the

125

required number has accumulated. The isotope will be retained for a longer time because it is bound to a stable, non-digestible platform. Both these parameters give the proposed approach the potential to generate more intense radiation fields than TNT; (b) TNT antibody may cross-react with other natural structures in the body because it binds to the natural epitopes on the DNA. In contrast, the agents administered in step 3 of the proposed approach bind to the artificial site on the platform. Since these sites can be chosen from a library of candidates, the agent can bind to them with a high affinity and specificity, and not bind to natural structures in the body; (c) TNT uses a single ‘‘cancer associated characteristic’’—namely the presence of dead or dying cells—to locate the radiation fields. Although dead cells are found preferentially in tumor tissue, reduction in tumor specificity may result from the presence of dead or dying normal cells (which occur as part of their natural turnover) throughout the body. The location of the radiation fields in the proposed approach requires a cell to have two independent characteristics (the cell must have the appropriate receptor to accumulate platform and it must be supersensitive to being killed in step 2). This double requirement by itself enhances specificity. However, as described, the tumor specificity in the proposed approach can be further improved by the delivery of targeted agents (enzymes and antidotes) to homogeneous receptors on normal cells which will cleave the binding sites on the platform contained in these normal cells or protect them. Although there have been significant advances in molecular biology, in biotechnology, and in the understanding of important cellular processes, they have not been translated into a successful (or even a greatly improved) therapy for most solid cancers. In a 1997 article entitled ‘‘Cancer Undefeated,’’ Bailar gave a statistical analysis of the results of cancer therapy over the last decade. He asserted that there has been virtually no improvement in the mortality rate of most cancers (Bailar & Gornik, 1997) although there has been some advance in treating hematological cancers (Scheinberg,

126

. 

1995), neuro-blastomas (van Noesel et al., 1997), and testicular cancers (van Basten et al., 1997). Analysis shows that almost all current and research approaches for treating disseminated cancers have the same fundamental strategy: they rely on an agent interacting individually and effectively with each cancer cell via its appropriate agent sensitive trait and it is for this reason that they have failed to substantially improve. We called all these approaches ‘‘lock and key’’ strategies to emphasize the need for this individual agent–cell interaction. They are defeated by the problems posed by the three universal characteristics of cancer cells described earlier. These approaches include chemotherapy, immunotoxins, immunotherapy, signal transduction, apoptosis, cell cycle control, and various forms of gene therapy. The claim has been made that the new approaches, based on modern molecular biology, will be able to kill more cancer cells with less non-tumor toxicity. This claim is unlikely to be true—by virtue of their essential strategy, they will only kill a fraction of the cancer cells, therefore, there will be a tendency (as in the past) for the therapist to keep increasing the dose in an attempt to kill all the cancer cells. This higher dose causes the systemic toxicity common to all these approaches. Despite the inherent limitations of lock and key based strategies, ‘‘euphoric’’ announcements are repeatedly made in the media and by companies for the latest of these strategies. Only much later, the same strategy is shown to fail. For example, ‘‘magic bullets’’ were once considered a major break-through, but are now described in articles as the ‘‘bullet that flopped.’’ Although significant improvements were made in the design and manufacture of these bullets, the improvements did not address the fundamental reasons why they originally failed so it is not surprising that they continue to fail. A similar history is being repeated for the latest lock and key strategies. For example, attempts to treat cancer by correcting P53 gene error, are likely to fail because the P53 gene error is not found in every cancer cell, genes cannot be inserted into, or corrected in, every cancer cell having the error, and other gene errors are also present which contribute to the malignant state.

In 1998, Klein wrote an article from which two comments are quoted. ‘‘Biochemists, immunologists, cytogeneticists and virologists have kept searching for the decisive event that had to be discovered before the riddle of cancer could be solved—a common metabolic disturbance, the mirage of a common cancer antigen, the universal virus, a general chromosomal imbalance, are examples of the all-encompassing theories that were advocated with particular fervor’’. This was followed by a later comment: ‘‘it has been said that the investment into cancer research has been a waste, that science did not live up to its expectations, and that cancer supports more people than it kills’’ (Klein, 1998). A consideration of the biological basis of the three problems facing cancer therapy, and of the need for ‘‘excess’’ specificity enhancing and killing potential reveals that the failure of lock and key approaches reflects their essential strategy (the mirage of a common metabolic disturbance), and is not due to the exact composition or action of any therapeutic agent. The implication of this conclusion is inescapable: if these strategies continue to dominate the field, future efforts in cancer therapy will continue to fail, and Bailar will surely be able to write another of his ‘‘Cancer Undefeated’’ articles. It is evident that a new direction is needed—one that abandons the simple lock and key strategies and embraces the complexities that reflect the microevolutionary process by which the tumor develops and progresses. The proposed approach represents such a new direction and it has the potential to circumvent the many cancer driven obstacles facing the problem of constructing a successful therapy for cancer.

REFERENCES A, S. J. & K, A. L. (1987). Radiobiologic implications of the microscopic distribution of energy from radionuclides. Int. J. Rad. Appl. Instrum. (B) 14, 165–169. B, C. T., W, M. C. & B, N. J. (1995). Tumor vasculature—a potential therapeutic target. Br. J. Cancer 72, 257–267.  , J. P., S K, H., S, D. T., P, E.,  D, M. F. & H, H. J. (1997). Current concepts about testicular cancer. Eur. J. Surg. Oncol. 23, 354–360.

      B, M. S., D-  M, H. C., B, R. B., F, V. J., C, T. C., T, T. R., V, M. B. & R, R. W. (1994). Chronic toxicity and oncogenicity inhalation study with vinyl acetate in the rat and mouse. Fundam. Appl. Toxicol. 23, 215–229. C, F. M., T, C. R. & E, A. L. (1989). Tumor necrosis treatment of ME-180 human cervical carcinoma model with 131I-labeled TNT-1 monoclonal antibody. Cancer Res. 49, 4578–4585. C, F. M., E, A. L., L, Z. & T, C. R. (1990). A comparative autoradiographic study demonstrating differential intra-tumor localization of monoclonal antibodies to cell surface (Lym-1) and intracellular (TNT-1) antigens. J. Nucl. Med. 31, 1059–1066. C, K. C. & L, L. A. (1993). Genomic instability and tumor progression: mechanistic considerations. Adv. Cancer Res. 60, 121–156. C-C, C., F, Z., D, M., M, C., E, L. & F, M. (1991). Expression of HLA-A,B,C antigens on primary and metastatic tumor cell populations of human carcinomas. Cancer Res. 51, 6372–6380. C, P. G., B, V., V P, A., W, T., S, C., T, E., D P, E., L, C., S, J. P., R, J. C. & B, T. (1994). A new gene coding for a differentiation antigen recognized by autologous cytologic T lymphocytes on HLA-A2 melanomas. J. Exp. Med. 180, 35–42. C, K. & R, R. H. (1993). Reduce, reuse, and recycle: shedding light on shedding cells. Gastroenterology 105, 1089–1097. E, B. E., C, D. A., R, A. M. & W, A. (1989). Perspectives on the role of MHC antigens in normal and malignant cell development. Adv. Cancer Res. 52, 181–245. F, M. E. & C, M. G. (1990). Lymphatics in primary cutaneous melanoma. Am. J. Surg. Pathol. 14, 370–374. F, S. R, V., D V, A., M, G., F, L. M. & M, P. (1993). Identification of epithelial cells in bronchoalveolar lavage. Hum. Exp. Toxicol. 12, 43–46. F, R. A. (1994). Low density lipoprotein as a vehicle for targeting antitumor compounds to cancer cells. Bioconjugate 5, 105–113. F, P. F., F, F. W. & H, R. F. (1950). Concentration of I131 in thyroid cancer shown by radioautography. Cancer 3, 86–105. F, L. (1969). Neoplastic Development I. New York: Academic Press. F, L. (1975). Neoplastic Development II. New York: Academic Press. F, S., P, R., Y, Z. X., T, W., Y, M. & F, V. J. (1994). Interleukin-1 alpha reduces the severity of the vascular leak syndrome produced by interleukin-2 and interleukin-2 plus interferon alpha. Toxicol. Pathol. 22, 381–397. G, H. J., V D D, H. A., S, C. F., S, J. W., S, G. J. & S, A. L. (1986). Receptormediated endocytosis in liver parenchymal cells. Inter. Rev. Exp. Path. 29, 113–171. G, M. S., L, W., R, L. & S, E. E. (1993). Integrin receptors in renal tubular epithelium: new insights into pathophysiology of acute renal failure. Am. J. Physiol. 264, 1–8.

127

G, J., R-Z, B., B-E, M., H, G. & S, S. (1989). The relationship between MHC antigen expression and metastasis. Adv. Cancer Res. 53, 89–115. H, R. P. (1990). Tumor progression: potential role of unstable genomic changes. Cancer Metastasis Rev. 9, 137–147. H, A. N. (1994). Cancer antigens: immune recognition of self and altered self. J. Exp. Med. 180, 1–4. I, F., H, G. K. & S-S, G. W. (1995). Mechanism of colloidal uptake into the lymphatic system: base study with percutaneous lymphography. Radiology 196, 107–113. I, T., T, A., S, I. J. & S, J. (1993). Epidermal growth factor protects gastric mucosa against ischemia-reperfusion injury. J. Clin. Gastroenterol. 17(Suppl. 1), 104–110. J, R. K. (1987). Transport of molecules in the tumor interstitium: a review. Cancer Res. 47, 3039–3051. J, R. W., K, J. K.,  B, T. J. C. & M, D. K. (1993). Coupling of the antiviral drug Ara-AMP to lactosminated albumin leads to specific uptake in rat and human hepatocytes. Hepatology 18, 146–152. K, Y., E, S., D, C. H., R, P. F., R, S. L., T, S. L., M, T. & R, S. A. (1994). Cloning of the gene coding for a shared human melanoma antigen recognized by autologous T cells infiltrating into tumor. Proc. Nat. Acad. Sci. U.S.A. 91, 3515–3519. K, B. A., M, J. A., M, G., S, H. W., G, H. K. & H, E. (1984). Monoclonal antibody to cardiac myosin imaging of experimental myocardial infraction. Hybridoma 3, 11–23. K, G. (1998). Foulds’ dangerous idea revisited: the multistep development of tumors 40 years later. Adv. Cancer Res. 72, 1–23. L, L. (1998). Cancer cells exhibit a mutator phenotype. Adv. Cancer Res. 72, 25–56. M, Y. & M, H. (1986). A new concept for macromolecular therapeutics in cancer therapy: mechanism of tumor accumulation of proteins and antitumor agent SMANCS. Cancer Res. 46, 6387–6392. M, R. F. (1988). Processing of endocytosed material. Adv. Cell Biol. 2, 159–180.  N, N. M., H, K., H-C, F. G. & E, R. M. (1997). Neuroblastoma 4s: a heterogeneous disease with variable risk factors and treatment strategies. Cancer 80, 834–843. O’B & B (1995). P, I. (1997). Targeted therapy of cancer with recombinant immunotoxins. Biochem. Biophys. Acta (Netherlands) 1333, c1–6. P, B. M. F. & B, M. S. (1981). Membrane recycling by coated vesicles. Ann. Rev. Biochem. 50, 85–101. P, R. C. & S, D. (1978). A new approach for assessing cumulative lysosomal degradation of proteins or other macromolecules. Biochem. Biophys. Res. Com. 81, 1254–1259. P, R. C., A, A. D., C, T. E. & S, D. (1979a). Tissue sites of degredation of low density lipoprotein: application of a method for determining the fate of plasma proteins. Proc. Nat. Acad. Sci. U.S.A. 76, 5345–5349.

128

. 

P, R. C., G, S. R., A, A. D. & S, D. (1979b). Radiolabeled sucrose covalently linked to protein. J. Biol. Chem. 254, 6876–6879. P, R. C., C, T. E., G, C. K., G, S. R., T, C. A. & A, A. D. (1983). A radioiodinated, intracellularly trapped ligand for determining the sites of plasma protein degredation in vivo. Biochem. J. 212, 719–800. P, R. T. (1994). Cancers beget mutations versus mutations begat cancers. Cancer Res. 54, 5296–5230. R, J. M., T, C. D., V, M., D, A., Z, E. C. & V  K, T. H. (1994). E-cadherin expression determines the mode of replacement of normal urothelium by human bladder carcinoma cells. Cancer Res. 54, 5488–5492. R, M. L., P, L. G., K, C. E., W, C., M, J., O, G., W, W. L., N, A., B, B. & C, P. (1995). Development of a humanized disulfide-stabilized anti-p185HER2 Fv-betalactamase fusion protein for activation of a cephalosporin doxorubicin prodrug. Cancer Res. 55, 63–70. R, H. (1990). The significance of biological heterogeneity. Cancer Metastasis Rev. 9, 1–20. R, C. & S, B. (1993). Distribution of colon cancer cells permanently labeled by lecitinmediated endocytosis of a trap label. Cancer Res. 53, 658–662. S, D. A. (1995). Adult leukemia in 1995: new directions. Lancet 346(8973), 455–456. S, L. W. (1992). Passive tumor targeting of soluble macromolecules and drug conjugates. Crit. Rev. Therapeutic Drug Carrier Systems 9, 135–187. S, G. K., B, S., B T. A., C, M., H, J., L, R. M., M, J., M, C. P., M, R., R, P. H., W, L. M. & W, L. A. 3rd. (1997). Toward antibody-directed enzyme prodrug therapy with the T268G mutant of human carboxypeptidase A1 and novel in vivo stable prodrugs of methotrexate. J. Biol. Chem. 272, 15804–15816. S-R, A. M., G, M. A., O-M, N., U, J. W. & V, E. S. (1993).

Ricin A chains and ricin A chain immunotoxins rapidly damage human endothelial cells, implications for vascular leak syndrome. Exp. Cell Res. 206, 227–234. S, B. W. (1994). Mechanisms of apoptosis: integration of genetic, biochemical, and cellular indicators. J. Nat. Cancer Inst. 86, 1286–1296. T, Q. T., A, E. A., M, D. M., K, V. C. & H, B. E. (1995). Enzyme/prodrug gene therapy: comparison of cytosine deaminase/5-flurocytosine versus thymidine kinase/ganciclovir enzyme/prodrug systems in a human colorectal carcinoma cell line. Cancer Res. 55, 4808–4812. W, P. D. (1994). Alterations in renal tubular extracellular matrix components after ischemia reperfusion injury to the kidney. Lab. Invest. 70, 339– 345. W, U. & R, K. (1990). The repertoire of somatic antibody mutants accumulating in the memory compartment after primary immunization is restricted through affinity maturation and mirrors that expressed in the secondary response. J. Exp. Med. 172, 1681–1689. W, G. Y., W, C. H. & S, R. J. (1983). Model for specific rescue of normal hepatocytes during methotrexate treatment for hepatic malignancy. Proc. Nat. Acad. Sci. U.S.A. 80, 3078–3080. W, G. Y., W, C. H. & R, M. I. (1985). Acetaminophen hepatotoxicity and targeted eescue: a model for specific chemotherapy of hepatocellular carcinoma. Hepatology 5, 709–713. W, A. H. (1992). Apoptosis and the regulation of cell numbers in normal and neoplastic tissues: an overview. Cancer Metastasis Rev. 11, 95–103. Y, L. C. J. & R (1990). The clearance of heterologous antibodies in experimental antibasement membrane antibody mediated glomerulonephritis. Exp. Pathol. 39, 79–87. Y, Y., B, J. R., B, J., C, P. R., E, C., F, A., G, G., S, D. & A, G. (1994). Antibody-mediated redistribution and shedding of endothelial antigens in the rabbit. J. Immunol. 150, 5633–5646.