Pharmacobiotics
Author: Steve Lucas

Copyright © October 2002 Steve Lucas

The following paper is a working draft of a hypothesis for the establishment of a patient-driven clinical trial model.  This paper is incomplete (particularly with respect to applicable agents for inclusion) and is currently under development.  The author of this paper is not a medical practitioner, he is merely the husband of a rare cancer patient who has proposed a hypothesis for consideration by those interested in the investigation of new approaches to cancer treatment.  It should not be misconstrued as a recommendation for any type of medical treatment.  Reviewers are encouraged to submit their critiques, suggestions, corrections, and opinions.  The paper is copyrighted and the author's permission is required prior to reprinting or excerpting from this document.

Background
Pharmacobiotics – a proposed paradigm shift
The Pharmacobiotics Theory
The analysis behind the Pharmacobiotics Theory
Potential pathways and markers of Antiangiogenic Activity
Potential compounds for inclusion in the tool kit
References

1.     Background

1.1.      Antiangiogenesis - The generally disappointing clinical results from the first two generations of antiangiogenic agent clinical trials has led to widespread rethinking of the action of these drugs and the processes they are intended to inhibit.  Although many of these compounds have entered trials amid significant fanfare, evidence seems to support the conclusion that numerous pathways to tumor angiogenesis need to be inhibited for the patient to obtain the full benefit of the effects of these drugs. 

In a significant number of clinical trials of antiangiogenic agents, like anti-VEGF and MMP inhibitors, the pharmacokinetic analysis of the agent indicates that the intended pathway was, in fact, inhibited, but the patient did not obtain the expected benefit due to the pre-existence of redundant angiogenic pathways that were enabled.

To compensate for the presence of redundant angiogenic pathways/processes, a second wave of clinical trials has begun to focus on “bundling” antiangiogenic agents.  In this paradigm, multiple antiangiogenic agents (generally two or three) are being combined in the hope that a sufficient number of antiangiogenic pathways will be inhibited for the patient to experience physiological benefit.

1.2.      Agent-Driven Clinical Trials – Adding to the difficulty in generating antiangiogenic trial outcomes is the current clinical trial paradigm.  It’s a gross oversimplification, but generally drugs are identified for clinical trials and patients are recruited.  Assuming appropriate patient sample design and stratification, the trial will effectively filter out externalities and determine whether the drug has biological and physiological effects on cancer patients and the tumors they have.  However, antiangiogenic agents have demonstrated that they likely need to be administered in bundles to have the desired curative effect.  This creates significant complications in designing robust trials.  Every agent added to the trial complicates design and the analysis of the data that results from the trial exponentially.

1.3.      How Many Antiangiogenic Agents Will Patients Need? – Given the number of potential angiogenic pathways at a tumor’s disposal to generate required blood supply, it is likely that the number of drugs needed by any given patient to prevent tumor angiogenesis will exceed the ability of clinical trials to design, collect, track and analyze these results.  This really is the crux of the most critical problem facing angiogenic research today – (1) how to identify which agents combination of agents to try and (2) how to track the results to identify the marginal utility of each compound – not just the combination of compounds.  It is proposed here that a significant shift in the design and conduct of clinical trials is needed to explore the full potential of currently available antiangiogenic drugs.

1.4.      Pharmacogenomics – A Partial Answer – Concurrently, there is a significant amount of interest in the field of pharmacogenomics.  This stems from the hope that our deeper understanding of the human genome will enable the development of gene-based medicine.  Genetically-driven cancer treatments based on screening for genetic markers and/or anomalies hold great promise for highly targeted, highly effective cancer treatments.  However, there are a number of significant obstacles that need to be overcome for real-world patients to see any benefit from this research.  Merely knowing the genetic fingerprint of a tumor does not ensure that there are not localized environmental conditions that will also significantly impact treatment effectiveness.  Further, if the genetic profile of tumors seems to be highly specific, there will be a significant amount of work needed to map each kind of tumor. Unfortunately, these are likely relatively distant events that offer little hope to current cancer patients.

2.    Pharmacobiotics – a proposed paradigm shift

2.1.      Background – Supporting Axioms - Cancer patients need a medical paradigm with more immediate and relevant benefits.  I propose a different approach – one that I’ve called Pharmacobiotics. The axioms supporting the theory of Pharmacobiotics are as follows:

 Based on these axioms, it is possible to develop the theory of Pharmacobiotics.

3.    The Pharmacobiotics Theory

Profile-Driven Clinical Trials (The Paradigm Shift)

Since it is likely that the contemporary design of clinical trials cannot support the number of antiangiogenic agents that need to be employed to generate predictable and positive physiological effects in patients, a new paradigm for clinical trials is needed.  Pharmacobiotics proposes that clinical trials for antiangiogenic agents be re-focused to be patient-based instead of agent-based.  Here’s the theory, in steps:

3.1. Identify all angiogenic pathways, antiangiogenic agents and trials  There are numerous single agent trials that have yielded pairwise correlations – agent to outcome.  Most of the results for antiangiogenic agents have been poor.   However, if parallel and/or redundant angiogenesis pathways are present, this is to be expected.

3.2.      Step 2  -  Identify those compounds that have demonstrated pharmacokinetic effects and tolerable toxicity profiles – Since most of the pairwise correlations from single agent trials are not positive, I propose to ignore physiological effect as a success factor, in favor of pharmacokinetic effect and tolerable toxicity.  Simply put, drugs should not be discarded from consideration in a multi-agent trial simply because they were not successful in single agent trials.  Due to the redundancy of angiogenic pathways, single agent trials are almost doomed to fail by design. What we should be looking to establish for each drug is:

  1. Did it manipulate a known/suspected angiogenic pathway as it was intended to (evidenced from blood assay studies)?

  2. Was it tolerable? 

There are qualifications and quantifications within these binary choices – how effective and how tolerable a compound was, but this can be built through a relatively simple sensitivity or cluster analysis.  Any agent that is active on an angiogenic pathway and has tolerable toxicity should be considered for inclusion in multi-agent trials.

3.3.      Step 3  -  Identify predictive factors for drug sensitivity – In lieu of direct evidence of a compounds’ effectiveness, it may be possible to use prognostic and predictive markers to impute the effectiveness of a compound.

3.4.      Step 4  -  Assemble an antiangiogenic tool kit – Assemble as many antiangiogenic agents as feasible drugs into a “tool kit”.  This tool kit represents a list of drugs that ca be bundled to treat a specific tumor, or more optimally, a specific patient.

3.5.      Step 5  -  Mathematical analysis and trial design – Employ a range of mathematical techniques to identify the best candidates from the tool kit for a given situation, whether patient specific or cohort specific.  Please refer to Section 4 for more details.

3.6.      Step 6 - Employ empirical data to hard wire the matrix - Trial results can be used to augment pharmacokinetic pairwise trial results.  Please refer to Section 5 for more details.

4.   The analysis behind the Pharmacobiotics Theory

4.1.      Representational matrix – It may be helpful to think of this problem as an n-dimensional matrix, incompletely populated.  For instance, if there were 50 antiangiogenic compounds in the tool kit, and the effects/side effects of all possible combinations of these 50 drugs were known, the result could be quantitatively captured in a 50 dimensional matrix.  However, there have been no real-world trials to test this hypothesis.  Accordingly, a rational starting point is to assemble as many pairwise trial results as possible to impute multivariate analysis from iterative pairwise analyses.  This level of analysis assumes additive effect, for both biological activity and toxicity tolerance, and assumes no synergy or ordering effect.  Any degree to which multiple compounds have been used in differing configurations could be used to “hardwire” the imputation analysis (In the example, this would mean “filling in” the matrix with “real” trials results.  This “bootstrapping” with empirical evidence will further define the degree of synergy/interaction between and among compounds and determine whether the assumption of additive effects is an appropriate axiom.

4.2.      “Teaching” the matrix – This is the point where the proposal departs from classical correlational and imputational analyses and relies on other feedback mechanisms.  Self-organizing maps and genetic algorigthms may prove useful tolls to enable this feedback.

4.3.           Not multi-agent trials – “Mega”-Agent trials and the need to minimize side effects – The goal of Pharmacobiotics should be to identify as many angiogenic pathways as possible and identify the compounds that inhibit each of these pathways with the least known or expected side effects.  It is highly likely that this will result not in multi-agent trials, but in “mega”-agent trials.  As the number of agents allowed I a trial increases, the complexity of the information brought in from the matrix increases as well.   Trying to simultaneously maximize predicted biological effect from each compound while simultaneously minimizing side effects from these compounds will become increasingly difficult mathematically.

In this instance, it is likely that the problem of bundling large numbers of antiangiogenic compounds will devolve into a simple mathematical minimization problem.  This is why the abandonment of clinical outcome as a determinant of success is critical.  For this proposal to work, nominal inhibition of an angiogenic pathway needs to be a condition of inclusion of a drug in a trial, but minimization of side effects should be the driving force.  There is some evidence for this approach.  Many of the MMP inhibitor trials indicated a fairly discrete threshold above which the drug “worked” pharmacokinetically and below which it didn’t.

4.4.      Complicating factors – Depending upon the narrowness of angiogenic pathways as evidenced in existing clinical trial information – the degree to which certain drugs work in only very limited and specific situations – there may need to be numerous matrices produced to avoid correlational contamination (such as covariance).  Or it may be possible to specify the activity range of a compound as an attribute in the matrix, thereby keeping it intact, but only utilizing part of the matrix for a given query.

4.5.      Employing predictive systemic data to hard wire the matrix - Trial results which can be used to augment pharmacokinetic pairwise trial results include the following:

4.5.1.  Drug metabolization and genetic markers - Genetic differences do determine how individuals metabolize drugs. Polymorphisms have been identified in more than 20 human drug-metabolizing enzymes that can determine whether an individual will fail to respond to a drug or suffer an exacerbated clinical response. Pharmacogenomics is making dramatic progress in developing tests to predict which patients are likely to benefit from a medicine and which patients are likely to suffer a toxic side effect. DNA-based tests designed for clinical use will give physicians the ability to predict patient response to a broad range of drug therapies. (1)

4.5.2.  Polymorphisms in Drug-Metabolizing Enzymes - A relatively small number of drug-metabolizing enzymes (DMEs) are responsible for metabolizing the majority of drug therapies in clinical use today. There are a relatively small number of relevant polymorphisms within these enzymes, and many of them can result in lack of therapeutic effect or in exacerbated clinical response. (1)

Genetic polymorphism in DMEs gives rise to 3 distinct subgroups of people who have measurable differences in their ability to metabolize drugs to either inactive or active metabolites. Individuals capable of efficient drug metabolism are called extensive metabolizers (EMs). Individuals with deficiencies in metabolism, which typically require mutation or deletion of both alleles of a gene, are termed poor metabolizers (PMs). Conversely, overexpression due to gene amplification results in ultra-rapid metabolizers (UMs).

Standard doses of drugs with a steep dose-response curve or a narrow therapeutic range may produce adverse drug reactions, toxicity, or decreased efficacy in PMs. When taken by UMs, the standard dose may be inadequate to produce the desired effect. (1)

Genetic variation is only 1 factor in human drug response, which also depends on concomitant drug therapies, environment, lifestyle, health status, and disease status. Unlike TPMT and CYP2D6, where 1 gene divides the population into 2 or 3 distinct groups (i.e., monogenic traits), response to most medications involves interaction between multiple genes (i.e., polygenic traits), which are more difficult to discover and distinguish from environmental factors. (1)

The TPMT genetic test has been well documented in the effective clinical management of patients with acute lymphoblastic leukemia (ALL). (2) Adjusting the dose of 6-mercaptopurine by a 10- to 15-fold decrease compared with conventional doses makes thiopurine as tolerable and effective in TPMT-deficient patients as it is in patients with normal activity levels.

In ALL therapy, the DNA-based TPMT diagnostic is now being used extensively. At St. Jude Children's Research Hospital, for example, ALL patients are routinely tested for TPMT activity to optimize therapy. TPMT testing is being evaluated for applications in azathioprine treatment for Crohn's disease, rheumatoid arthritis, and renal transplantation.

4.5.3.  Gene expression - Two genomics tools employed in pharmacogenomics research are: (1) identifying inherited variation in DNA and genes, and (2) monitoring gene expression by assaying which genes are "on" (expressed), or "off" (not expressed). (3)

4.5.4.  DNA Variation - Much of the effort in pharmacogenomics has been focused on DNA variation, with a goal of identifying which inherited variation in DNA correlates with differences in patients' responses to drug treatments. (4)  However, despite extensive research efforts in the area, there are few candidate DNA variants that correlate with drug response. Some of the best-characterized examples are in genes that code for drug-metabolizing enzymes, such as cytochrome P450 and thiopurine methyltransferase (TPMT). (5)

Further, even if variation in drug response is largely genetic, it is not clear whether this effect results from single genes or multiple genes. If the effect is due to the interaction of multiple gene products, it may be impossible to statistically determine the underlying causes because of the small sample sizes in clinical studies.[5,6] For example, despite extensive research efforts to determine genetic causes of complex diseases (i.e., diseases resulting from the effects of multiple genes), only a handful of candidate genetic loci have been identified. (5) Thus, determining the DNA variants that correlate with drug response could prove problematic if the response is due to multiple genes.

Since clinical trials typically try to minimize the number of subjects involved in the study, an analysis of a subset of the patients would require even greater initial numbers of subjects, resulting in greatly increased costs of clinical trials.

Instead of a statistical imputation analysis, a non-statistical approach may prove relevant.  Self organizing maps, neural networks or genetic algorithms may prove not only relevant, but advantageous, depending upon the amount of noise in the patterns that emerge from the data.

4.5.5.  Monitoring Gene Expression Using DNA Micro Arrays - Another genomics approach with emerging potential to improve drug treatment monitors gene expression in cells, tissues, and tumors. When a gene is "turned on," it produces an RNA intermediate. The amount of RNA produced by the gene generally corresponds to the amount of gene product produced by the gene. It is therefore possible to estimate the amount of gene expression by measuring the amount of RNA produced. (3) The ability to measure gene expression has been facilitated in recent years by the availability of DNA micro arrays or DNA chips, which consist of a matrix of DNA spots on a glass slide, each one corresponding to a single gene. (6-9)  Using DNA micro arrays, it is possible to monitor expression of tens of thousands of genes simultaneously, throwing open the possibility of unprecedented molecular characterization.  (11,12)

Gene expression studies can detect not only inherited variation in gene expression, but also responses to environmental influences or stochastic changes in the patient. They are therefore potentially good tools for: (1) identifying additional potential drug target candidates; (2) providing more accurate classifications of disease; (3) predicting clinical outcome and presymptomatic disease; and (4) monitoring and predicting drug responses. (3)

To identify new drug targets, DNA micro arrays are used to compare the expression of genes in two different cell, tissue, or tumor types. The genes that are more highly expressed in one cell type, but not in the other, are candidates for key determinants of the difference. For example, in one study, metastatic vs. nonmetastatic cells were compared using DNA micro arrays. (10)  Some genes, such as RhoC, were preferentially expressed in the metastatic, but not in the nonmetastatic cells. It was subsequently shown that RhoC is necessary and sufficient to cause metastasis, leading to a better understanding of the process by which cells become metastatic. This type of analysis can lead to improved drug development, since genes that are preferentially expressed in disease tissue are typically good candidates for drug targets.

4.5.6.  NCI Cell Line Analysis - The National Cancer Institute has evaluated a set of 60 cell lines (NCI60) derived from different tissue cancer types to determine their response to over 70,000 chemical compounds, including all common chemotherapies. (13) Analysis of 8000 genes using DNA micro arrays permitted determination of the tissue of origin of these tumor cell lines based solely on their patterns of gene expression. (14)

This type of classification offers the potential for more precise tumor classifications as well as better determination of appropriate drug treatments.

4.6.      Specific predictive data to hard wire the matrix

4.6.1.     ER-Beta/ER-Alpha – Co-expression of ER-Beta was found with the majority of breast cancer tumors, but ER-Alpha correlated with Progr\esterone receptor (PR).(25)

4.7.      Employing empirical data to hard wire the matrix - There may be pragmatic, empirical information that can be used to hard wire some of the patterns.   Key to this process will be gaining access to clinicians and PI’s to add their knowledge to this system.

4.8.      Trial Development - Trials can be developed in two ways – cohort-based and profile-based:

4.8.1   Cohort-based trials – In this instance, patients with similar profiles (similar blood markers, genetic markers, and/or tumor location/histology) are assembled and a pattern recognition run against the matrix of tool kit agents.  This will result in a combination of compounds that “best fits” the tumor profile of the cohort.  Further, it is likely possible to establish an order to the compounds such that the relevance of each compound to the overall treatment is established.  If intolerable side effects are encountered, the compound that correlates least with the profile can be eliminated to reduce side effects, unless a hard-wired matrix entry indicates a specific side effect was caused by a specific compound in prior trials.

4.8.2   Profile-based trials – In this instance, a profile is created for each patient and each receives an individualized antiangiogenic agent bundling.  This kind of trial is the most complex from a design standpoint, but the most optimal from a patient’s standpoint.  Analyzing data from this kind of trial and feeding it back into a matrix will be challenging, and likely will require

4.9.      Potential Reaction to this Proposal

Pros – This approach relies more on developing logical extensions to completed trials on already developed drugs, with clinical and real-world insight added.  This can be turned into a mathematical problem with pattern recognition and simulation components.  This approach is patient-focused and offers the opportunity to test readily existing, but relatively ineffective drugs in new combinations to provide immediate benefits to patients.

Cons – If this approach is completely successful and we are able to develop a profile for each patient and a separate protocol (with maybe a dozen or more drugs) for each, there may be too many different combinations to determine if an individual compound worked, even if the concept, as a whole, worked. It might take a long time for enough similar individual patient profile test results to yield enough information to make assessments about each individual compound.

Accordingly, This process needs to focus on drugs already in the IND pipeline.  The data generation would be too slow and too difficult to unravel to make assessments on pre-clinical drugs.  Since this process is patient-focused, that is of little concern, but drug companies may not want to include non-FDA approved compounds in this kind of trial since the relationship of one drug (among many) to an outcome will have to be imputed, to be replaced by statistical data over time – possibly a long time.

5.    Potential Pathways and Markers of AntiangiogenicActivity

Endothelial Growth Factor Receptors - At least 3 distinct pathways have been identified, including Ras, Akt/PI3-K, and STAT. Tumorigenesis has been linked to overexpression of EGFR (found in approximately 70% of solid tumors) as well as to ligand overexpression. Current treatment modalities against EGFR/ligand overexpression are either tyrosine kinase inhibitors or monoclonal antibodies to

HER2 - HER-2 is a cellular proto-oncogene coding for a transmembrane receptor belonging to the epidermal growth factor receptor family.  It is an effective predictor of response to trastuzumab, an anti-HER-2 antibody.

HER3

HER4 – Serum basic fibroblastic growth factor (bFGF)

6.    Potential compounds for inclusion into the tool kit

6.1       Herceptin and the HER2 receptor – It is the degree of expression of the HER2 receptor gene, rather than genetic variation of of DMEs that the determinant of Herceptin response in a patient. HER2, a receptor for hormones that stimulate tumor growth, is overexpressed in approximately one fourth of breast cancer patients. Overexpression of the HER2/neu oncogene is correlated with poor prognosis, increased tumor formation and metastasis, and resistance to chemotherapeutic agents. Trastuzumab, a cloned antibody that blocks the receptor, has significant benefit when used as an adjunct to conventional chemotherapy. HER2 testing predetermines patients who overexpress HER2 and who will respond to trastuzumab. (15)

6.2       Antileukemics 6-mercaptopurine and 6-thioguanine, and the immune suppressant azathioprine amd TMPT - TPMT is responsible for the metabolism of thiopurine medications: the antileukemics 6-mercaptopurine and 6-thioguanine, and the immune suppressant azathioprine. TPMT activity is essential for normal metabolism of these drugs and determines both efficacy and toxicity. Patients with inherited TPMT deficiency suffer severe, potentially fatal hematopoietic toxicity when exposed to standard doses of thiopurine drugs.

A pharmacogenomic test, developed at St. Jude Children's Research Hospital, enables physicians to predetermine patients' TPMT activity levels based on whether or not they have inherited the alleles associated with TPMT deficiency. The test classifies patients according to normal, intermediate, and deficient levels of TPMT activity. Concordance between genotype and phenotype approaches 100%.

Patients classified as normal in activity -- about 90% of whites and blacks -- are treated with conventional doses. Lower doses are tailored to avoid toxicity in deficient and intermediate patients, who represent about 10% of each of these populations. Approximately 1 in 300 whites and blacks are TPMT deficient. Although this polymorphism is relatively rare, TPMT-deficient patients suffer exaggerated, potentially life-threatening toxic responses to normal doses of azathioprine and thiopurine drugs. (16)

6.3             Cetuximab (IMC-225) - a chimeric monoclonal antibody that has a high degree of affinity for the EGFR, leading to block of ligand binding and subsequent prevention of tyrosine kinase activation, has been shown previously to produce partial responses in patients with colorectal cancer. (17)

For Irinotecan and IMC-225:            Partial response rate = 22.5%             Stable disease = 7%

Grade 3 and 4 toxicity was less with the combination of irinotecan and IMC-C225 than in previously single-agent administration of irinotecan, with diarrhea and neutropenia being the most significant (22% and 14%, respectively).

Among patients who developed a skin rash, the response rate was 29% vs a response rate of 3% among patients who did not developed skin rash. Whether a skin rash can serve as a marker for potential tumor response is a consideration that should be determined.  Also, IMC-C225 could potentially be used to reverse drug resistance."Times New Roman"">

6.4.      OSI-774 - Phase 1 data established 150mg/m2 as the recommended daily dose for OSI-774, a selective oral EGFR tyrosine kinase inhibitor.  Side effects well tolerated, with over 50% of patients developing an acneiform rash (less than grade 2), and fewer patients experienced mild headaches and asymptomatic hyperbilirubinemia. In vivo activity was measured by 18F-fluordeoxyglucose-positron emission tomography (FDG-PET) scan. Two of 14 patients showed objective clinical response, while 2 others maintained stable disease. Additionally, activity was confirmed by immunohistochemical stains, showing downregulation of EGFR and EGFR phosphorylation. Responses occurred among a variety of tumor types, including renal cell, head and neck, non-small cell lung and colorectal cancers. (18,19)

In multicenter phase 2 trials of heavily treated locally advanced or metastatic squamous cell carcinoma of the head and neck, EGFR expression seemed to correlate with severity of the disease.  Employing the phase 1 dose (150mg/m2), partial responses were observed in 7 patients (5.6%), while 49 patients (39.5%) maintained stable disease for a combined median of 5.8 months, which is comparable to median duration of response with more toxic regimens (5-FU/cisplatin +/- interferon) in a meta-analysis based on historic control data. One-year survival data were also comparable. Toxicity was primarily cutaneous (74.2%; only 11% grade 3), with diarrhea a distant runner up (30%; 3.2% grade 4).

6.5.           ZD-1839 - Breast cancers that overexpress both EGFR and HER2/neu behave aggressively. HER2 requires phosphorylation, leading to tyrosine kinase activation. ZD1839, which is known to activate EGFR, was introduced into several breast tumor cell lines. Subsequently, HER2 phosphorylation was measured and found to be decreased, suggesting that EGFR acts as a ligand for the phosphorylation of HER2, supporting the idea of cross-talk between multiple intracellular networks and more specifically, EGFR 's role as a key regulator for HER2 expression. (20)

In murine models, the addition of ZD1839 to trastuzumab enhanced the antitumor effect, resulting in decreased tumor growth and proliferation. This finding prompted the Eastern Cooperative Oncology Group (ECOG) to formulate a trial that will combine ZD1839 with trastuzumab, looking at the degree of HER2/neu expression, in addition to the combination drug effect on HER3 and HER4 overexpression.

6.6             Thalidomide - Single-agent thalidomide has little or no activity in patients with heavily pretreated breast cancer. (21)  Evaluation of circulating angiogenic factors and pharmacokinetic studies failed to provide insight into the reason for the lack of efficacy.

The dose was reduced because of somnolence to 600 mg for five patients and to 400 mg for two and was increased for one to 1,000 mg and for four to 1,200 mg. On the 200-mg arm, 12 patients had progressive disease at or before 8 weeks and two had stable disease at 8 weeks, of whom one was removed from study at week 11 because of grade 3 neuropathy and the other had progressive disease at week 16. Dose-limiting toxicities included somnolence and neuropathy. Adverse events that did not require dose or schedule modifications included constipation, fatigue, dry mouth, dizziness, nausea, anorexia, arrhythmia, headaches, skin rash, hypotension, and neutropenia.

Thalidomide is a generally well-tolerated drug that may have antitumor activity in a minority of patients with recurrent high-grade gliomas. (22)

Considerations

Beyond the traditional chemotherapy targets of DNA, topoisomerases, and microtubules, antiangiogenic agents affect growth factors, receptor tyrosine kinases, integrins, cell adhesion molecules, proteases, membrane glycoproteins, enzymes, and cyclooxygenase-2 (COX-2) expression.

6.7.      Recombinant human endostatin (rHE) - Escalated doses (ranging from 15mg/m2 to 600mg/m2) of rHE by intravenous infusion over 20 minutes on a daily basis were administered to 25 patients with a variety of solid tumors, including melanomas, sarcomas, as well as breast, lung, colorectal, head and neck, renal, and thyroid carcinomas. No dose-limiting toxicity was observed at any dose level and pharmacokinetics were linear. There were trends of antitumor activity and apoptosis among individual patients, and initial PET studies appeared to be useful. (23)

6.8.      Recombinant human angiostatin (rHA) - Phase 1 data on the safety profile, pharmacokinetic and pharmacodynamic profiles of recombinant human angiostatin (rHA) were tested on 19 patients given dose escalations (ranging from 15mg/m2 to 240mg/m2) of rHA by 10 minutes intravenous infusion on a daily basis until disease progression. Colorectal, soft tissue sarcoma, breast, head and neck, and ovarian tumors were included. (24)

Each patient received an average of 53 doses of drug across all dose levels. No dose-limiting toxicity was identified and there was only 1 grade 3 toxicity (lymphopenia). Other mild toxic effects included nail bed changes, hand/foot syndrome, and maculopapular skin rash. Fourteen of the patients who had undergone recent surgeries experienced no delay in wound healing and no abnormal coagulation studies were recorded. Serum levels of rHA reflected administered dose levels in a linear fashion.

Posttreatment tissue analysis showed a decline in tumor vascularity although modulation of blood flow and capillary sprout formation, as measured by contrast-enhanced ultrasound and magnetic resonance imaging (MRI), was not as clear. Vascular endothelial growth factor (VEGF) expression was also decreased in some treated specimens.

7.    References

  1. Lazarou J, Pomeranz BH, Corey PN. Incidence of adverse drug reactions in hospitalized patients: a meta-analysis of prospective studies. JAMA. 1998;279:1200-1205.

  2. Bittner M, Meltzer P, Chen Y, et al. Molecular classification of cutaneous malignant melanoma by gene expression profiling. Nature. 2000;406:536-540.

  3. Rioux PP, Basilion JP. What's new in pharmacogenomics? Available at http://www.medscape.com/viewarticle/408599

  4. Risch NJ. Searching for genetic determinants in the new millennium. Nature. 2000;405:847-856.

  5. Altshuler D, Daly M, Kruglyak L. Guilt by association. Nat Genet. 2000. 26:135-137.

  6. Clark EA, Golub TF, Lander ES, Hynes EO. Genomic analysis of metastasis reveals an essential role for RhoC. Nature. 2000;406:532-535.

  7. Golub TR, Slonim DK, Tamayo P, et al. Molecular classification of cancer: class discovery and class prediction by gene expression monitoring. Science. 1999;286:531-537.

  8. National Cancer Institute. Developmental Therapeutics program (DTP). Available at: http://dtp.nci.nih.gov.

  9. Ross DT, Scherf U, Eisen MB, et al. Systematic variation in gene expression patterns in human cancer cell lines. Nat Genet. 2000;24:227-235.

  10. Perou CM, Sorlie T, Eisen MB, et al. Molecular portraits of human breast tumours. Nature. 2000;406:747-752.

  11. Alizadeh AA, Eisen MB, Davis ER, et al. Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling. Nature. 2000;403:503-511.

  12. Scherf U, Ross DT, Waltham M, et al. A gene expression database for the molecular pharmacology of cancer. Nat Genet. 2000;24:236-244.

  13. Saltz L, Rubin M, Hochster H, et al. Cetuximab (IMC-C225) plus irinotecan (CPT-11) is active in CPT-11-refractory colorectal cancer (CRC) that expresses epidermal growth factor receptor (EGFR). Program and abstracts of the American Society of Clinical Oncology 37th Annual Meeting; May 12-15, 2001; San Francisco, California. Abstract 7.

  14. Rowinsky EK, Hammond L, Siu L, et al. Dose-schedule-finding, pharmacokinetic (PK), biologic, and functional imaging studies of OSI-774, a selective epidermal growth factor receptor (EGFR) tyrosine kinase (TK). Program and abstracts of the American Society of Clinical Oncology 37th Annual Meeting; May 12-15, 2001; San Francisco, California. Abstract 5.

  15. Senzer NN, Soulieres D, Siu L, et al. Phase 2 Evaluation of OSI-774, a potent oral antagonist of the EGFR-TK in patients with advanced squamous cell carcinoma of the head and neck. Program and abstracts of the American Society of Clinical Oncology 37th Annual Meeting; May 12-15, 2001; San Francisco, California. Abstract 6.

  16. Moulder SL, Yakes M, Bianco R, Arteaga C. Small molecule EGF receptor tyrosine kinase inhibitor ZD1839 (IRESSA) inhibits HER2/neu (erb-2) overexpressing breast tumor cells. Program and abstracts of the American Society of Clinical Oncology 37th Annual Meeting; May 12-15, 2001; San Francisco, California. Abstract 8.

  17. Baidas SM, Winer EP, Fleming GF, et al, Phase II Evaluation of Thalidomide in Patients with Metastatic Breast Cancer
    J Clin Oncol. 2000;18(14):2710-2717.

  18. Fine HA, Figg WD, Jaeckle K, et al, Phase II Trial of the Antiangiogenic Agent Thalidomide in Patients with Recurrent High-Grade Gliomas,J Clin Oncol. 2000;18(4):708-15.

  19. Herbst RS, Tran HT, Mullani NA, et al. Phase I clincal trial of recombinant human endostatin (rHE) in patients with solid tumors: pharmacokinetic, safety and efficacy analysis using surrogate endpoints of tissue radiologic response. Program and abstracts of the American Society of Clinical Oncology 37th Annual Meeting; May 12-15, 2001; San Francisco, California. Abstract 9.

  20. DeMoraes ED, Fogler WE, Grant D, et al. Recombinant human angiotensin (rhA): a phase I clinical trial assessing safety, pharmacokinetics, and pharmacodynamics. Program and abstracts of the American Society of Clinical Oncology 37th Annual Meeting; May 12-15, 2001; San Francisco, California. Abstract 10.

  21. Fuqua SAW, Schiff R, Parra I, et al. Expression of estrogen receptor beta protein in human breast cancer: correlation with clinical parameters. Program and abstracts of the 23rd Annual San Antonio Breast Cancer Symposium; December 6-9, 2000; San Antonio, Texas. Abstract 123. Breast Cancer Res Treat. 2000;64:41.

  22. Isola JJ, Tanner M, Holli K, et al. Amplification of topoisomerase II alpha is a strong predictor of response to epirubicin-based chemotherapy in HER-2/neu positive metastatic breast cancer. Program and abstracts of the 23rd Annual San Antonio Breast Cancer Symposium; December 6-9, 2000; San Antonio, Texas. Abstract 21. Breast Cancer Res Treat. 2000;64:31.

  23. Coon JC, Marcus E, Gupta-Burt S, et al. Amplification of topoisomerase IIa or c-erbB-2 predicts response to doxorubicin and docetaxel in locally advanced breast cancer. Program and abstracts of the 23rd Annual San Antonio Breast Cancer Symposium; December 6-9, 2000; San Antonio, Texas. Abstract 309. Breast Cancer Res Treat. 2000;64:78.