[Cancer Biology & Therapy 2:4:Suppl. 1, S134-S139; July/August 2003]; ©2003 Landes Bioscience
`
`Models of Anti-Cancer Therapy
`Human Tumor Xenografts as Predictive Preclinical Models for Anticancer
`Drug Activity in Humans
`Better Than Commonly Perceived—But They Can Be Improved
`
`©2003 Landes Bioscience. Not for distribution.
`
`Robert S. Kerbel
`
`*Correspondence to: Robert S. Kerbel; Molecular & Cell Biology Research;
`Sunnybrook and Women’s College Health Sciences Centre; Toronto-Sunnybrook
`Regional Cancer Centre and Departments of Medical Biophysics and Laboratory
`Medicine/Pathobiology; University of Toronto; Toronto Ontario
`
`Previously published online as a CB&T E-publication at:
`http://www.landesbioscience.com/journals/cbt/toc.php?volume=2&issue=0
`
`KEY WORDS
`
`Chemotherapy, Drug resistance, Pharmaco-
`kinetics
`
`Robert S. Kerbel, Ph.D.
`
`Dr. Kerbel’s research is supported by grants from the National Cancer Institute of
`Canada, the Canadian Institute for Health Research and the National Institutes of
`Health, USA (CA-41223).
`
`ABSTRACT
`It is not uncommon for new anti-cancer drugs or therapies to show highly effective, and
`sometimes even spectacular anti-cancer treatment results using transplantable tumors in
`mice. These models frequently involve human tumor xenografts grown subcutaneously in
`immune deficient hosts such as athymic (nude) or severe combined immune deficient
`(SCID) mice. Unfortunately, such preclinical results are often followed by failure of the
`drug/therapy in clinical trials, or, if the drug is successful, it usually has only modest
`efficacy results, by comparison. Not surprisingly, this has provoked considerable skepticism
`about the value of using such preclinical models for early stage in vivo preclinical drug
`testing. As a result, a shift has occurred towards developing and using spontaneous
`mouse tumors arising in transgenic and/or knockout mice engineered to recapitulate
`various genetic alterations thought to be causative of specific types of respective human
`cancers. Alternatively, the opinion has been expressed of the need to refine and improve
`the human tumor xenograft models, e.g., by use of orthotopic transplantation and there-
`fore promotion of metastatic spread of the resultant ‘primary’ tumors.
`Close inspection of retrospective and prospective studies in the literature, however,
`reveals that human tumor xenografts—even non metastatic ectopic/subcutaneous
`‘primary’ tumor transplants—can be remarkably predictive of cytotoxic chemotherapeutic
`drugs that have activity in humans, when the drugs are tested in mice using pharmacoki-
`netically clinically equivalent or ‘rational’ drug doses. What may be at variance with clinical
`activity, however, is the magnitude of the benefit observed in mice, both in terms of the
`degree of tumor responses and overall survival. It is argued that this disparity can be
`significantly minimized by use of orthotopic transplant/metastatic tumor models in which
`treatment is initiated after the primary tumor has been removed and the distant metastases
`are well established and macroscopic—i.e., the bar is raised and treatment is undertaken
`on advanced, high volume, metastatic disease. In such circumstances, survival should be
`used as an endpoint; changes in tumor burden using surrogate markers or micro-imaging
`techniques can be used as well to monitor effects of therapies on tumor response.
`Adoption of such procedures would more accurately recapitulate the phase I/II/III clinical
`trial situation in which treatment is initiated on patients with advanced, high-volume
`metastatic disease.
`
`INTRODUCTION
`One of the greatest challenges faced by developers of new drugs and treatment strategies
`for cancer is the obvious need to test them in preclinical in vivo models that have a good
`probability of being predictive of similar activity in humans. For more than half a century
`the laboratory mouse has been the primary species in which experimental cancer treat-
`ments have been tested. Until about 25 years ago syngeneic transplantable mouse tumors
`were used most commonly for such preclinical therapy studies, and still are, especially for
`immunotherapy experiments in which an intact immune system is required. The discovery
`that human tumor cell lines, and sometimes even primary biopsy human tumor specimens,
`can give rise to progressively growing, and potentially lethal cancers in immune deficient
`mice gradually resulted in a shift towards the use of human tumor xenografts for the study
`of virtually all other types of anti-cancer drugs and treatment strategies.1 Essentially every
`clinically approved anti-cancer drug was tested using these models, and showed positive
`anti-cancer effects before being evaluated in early, and then late phase clinical trials.
`Nevertheless, these successes have been overshadowed by highly visible failures in which a
`particular new anti-cancer drug, or treatment strategy, demonstrated remarkable
`
`S134
`
`Cancer Biology & Therapy
`
`2003; Vol. 2 Issue 4, Suppl. 1
`
`Ex. 1065-0001
`
`

`
`HUMAN TUMOR XENOGRAFTS AS PREDICTIVE PRECLINICAL THERAPY MODELS
`
`anti-tumor effects using a transplantable tumor model in mice, only
`to be followed by failure in the clinical trial setting2 (“failure” in this
`case being defined here as having little or no survival benefit, regardless
`of whether it was found to be safe, or not, in humans).
`Perhaps the most spectacular and recent example of this was the
`study by Boehm, O’Reilly et al.3 who reported stunning effects of
`endostatin on three different transplantable tumors subcutaneously
`grown in syngeneic mice: the Lewis Lung carcinoma, the B16
`melanoma and the T41 fibrosarcoma.3 Cycles of daily endostatin
`treatment, an antiangiogenic protein drug, caused repeated and total
`regressions of established tumors. There was no evidence of relapse
`involving emergence of drug resistant variant/mutant subpopulations.
`Leaving aside the question of whether this result is reproducible
`(most other published studies of successful endostatin therapy show
`much more modest growth delays, but not overt tumor regressions),
`this result sparked enormous interest in both the scientific literature4
`and lay press.5 It fueled unprecedented rapid initiation of phase I
`clinical trials in the United States, the results of which were recently
`reported.6,7 The results of these trials showed the drug to be safe
`(which is the primary purpose of phase I trials) but there was cer-
`tainly no evidence of the type of spectacular preclinical responses
`that had been observed in any of the treated patients.6,7 The
`inevitable result has been the disappointment expressed not only
`about the drug itself, but about antiangiogenic therapy in general. In
`fairness, the results of other clinical trials involving antiangiogenic
`therapy such as the humanized monoclonal antibody to vascular
`endothelial cell growth factor (VEGF) known as bevacizumab
`(tradename: Avastin), which was tested in a randomized phase III
`trial as a third line therapy combined with Xeloda in advanced
`metastatic breast cancer, have also contributed significantly to this
`sense of current disappointment. But even in this case the disap-
`pointment stems, in part, from the many impressive results of prior
`preclinical studies utilizing a variety VEGF targeting of antiangio-
`genic drugs and approaches in a variety of mouse tumor models.
`In 1999, Dr. Judah Folkman was quoted in a Newsweek maga-
`zine article as saying that a mouse study does not belong on the front
`page of the New York Times.8 This makes considerable sense, and
`was a logical follow up to a quote he made in the May 3, 1998
`Sunday New York Times article: “if you are a mouse and have
`cancer, we can take good care of you”.5 This statement would also
`seem to be logical, but as explained in this review, it is not necessarily
`so, and can be seriously challenged. Simply put, if you are a mouse
`with advanced, high-volume metastatic disease we probably cannot
`take good care of you.
`The apparent lack of predictability of results often obtained using
`transplantable mouse or human tumors in normal or immune deficient
`mice has convinced many investigators to move away from such
`models and instead use spontaneously arising tumors, in particular
`genetically manipulated transgenic/knockout mice where the
`tumors which arise have mutations thought to be causative of the
`respective human cancers.9,10 Alternatively, other investigators have
`suggested that transplantable tumor models can be made much
`more predictive by orthotopic transplantation which frequently
`facilitates metastatic
`spread—especially of human
`tumor
`xenografts11,12—and thus testing the effects of a given therapy on
`either (or both) the primary tumor growing in a physiologically
`relevant site (as opposed to an ectopic site) and distant metastatic
`disease.
`
`In this commentary, two major points are made:
`1. growth and testing of human tumors in subcutaneous tissue sites
`that are ectopic for a given type of cancer have provided relevant
`and predictive information to the clinic, provided that clinically
`relevant, pharmacokinetic parameters (especially dosing) are
`employed; and,
`2. orthotopic transplants are nevertheless potentially valuable when
`used to generate metastases—but that therapy should be initiated
`at a point when the metastases are well established and macroscopic
`in nature (i.e., high volume metastatic disease)—and not just on
`low-volume (occult) minimum residual disease, which is what
`almost all previous studies have utilized when testing therapies on
`metastatic disease.
`Also highlighted is the need for continuous vigilance with respect to
`the nature and origin of the cell lines used for transplantable tumor
`studies.
`
`RETROSPECTIVE STUDIES OF CHEMOTHERAPEUTIC
`DRUGS USING SUBCUTANEOUS/ECTOPIC HUMAN
`TUMOR XENOGRAFTS SHOWING A HIGH DEGREE
`OF CLINICAL RELEVANCE
`Nomura, Inaba and colleagues of the Cancer Chemotherapy
`Centre, Japanese Foundation for Cancer Research, Tokyo, have
`published a series of important and insightful studies which show
`clearly the remarkable potential of ectopic human tumor xenografts
`for predicting the pattern of activity of conventional cytotoxic
`chemotherapeutic drugs in humans.13-17 Prior to undertaking their
`studies many other published reports showed that the majority of
`chemotherapeutic drugs have significant anti-tumor effects on a
`particular type of human cancer, even though most of the drugs tested
`were known not to have such activity on the respective tumor type
`in the clinical situation. In other words, the results of preclinical
`xenograft models were not retrospectively predictive of clinical activity.
`However, Nomura, Inaba and colleagues reasoned this could be due
`to inappropriate drug dosing. It turns out that the maximum tolerated
`dose (MTD) of most chemotherapeutic drugs that be given to mice
`is higher (4–5 times) than in humans. In some cases, the MTD is
`lower in mice than in humans, and in some cases (e.g., adriamycin)
`it is the same. Thus, in many cases, if one uses the MTD of a given
`chemotherapeutic drug for mice, the blood levels of drug will be
`significantly higher than can be attained in humans, leading to false
`positive tumor responses in mice.
`To study this hypothesis, Nomura, Inaba and colleagues tested a
`large number of independent cell lines (e.g., generally eight to
`twelve) for each type of cancer tested. They reasoned this was similar
`in nature to the number of patients in a typical phase I clinical trial,
`and as such, would minimize the risks associated with obtaining a
`false positive or false negative response when testing just a single or
`few cell lines. In other words, one looks for an overall pattern of
`response in mice to different drugs that may be similar to what is
`seem a population of cancer patients. Each tumor cell line was grown
`as subcutaneous xenograft in a number of athymic nude mice. These
`mice were subsequently treated with at least 5 or 6 different
`chemotherapeutic drugs, tested as monotherapies, where some of the
`drugs were known to be clinically active on the particular type of
`human cancer being tested, and some not. The drugs were adminis-
`tered to some groups of tumor-bearing mice using the MTD of the
`drug for mice, whereas in another group the pharmacokinetically
`clinically equivalent dose (CED) or “rational dose” for humans was
`used.
`
`www.landesbioscience.com
`
`Cancer Biology & Therapy
`
`S135
`
`Ex. 1065-0002
`
`

`
`HUMAN TUMOR XENOGRAFTS AS PREDICTIVE PRECLINICAL THERAPY MODELS
`
`Figure 1. Results of an experiment in which large, established (0.75 cm3) human neuroblastomas (NB) were treated with a metronomic low-dose vinblas-
`tine schedule, or DC101 (an anti-VEGFR-2 monoclonal antibody) or a combination of the two drugs. The dosing of the drugs is indicated in the lower
`figure. Note that the metronomic/ maintenance regimen was preceded by an induction regimen of the same drug to try and rapidly debulk the tumor mass
`before initiating the metronomic low-dose chemotherapy schedule. Progression of disease was seen in the single treatment groups, whereas slow but even-
`tually complete tumor regression was noted in the combination group in which the therapy was continued for 7 months, which was possible by the lack of
`toxicity of this regimen. Taken from Klement, G. et al. “Continuous low-dose therapy with vinblastine and VEGF receptor-2 antibody induces sustained tumor
`regression without overt toxicity” J Clin Invest 2000; 105:R15-R24.
`
`Analysis of the data for a large number of tumor types including
`lung, glioma, breast and gastric cancers showed that the pattern of
`response obtained when the mouse MTD was used was not associated
`with clinical pattern of responsiveness—most or all drugs showed
`activity. In other words, there was a high rate of false positives. In
`striking contrast, when the clinically equivalent or rational dose was
`used, the pattern of response in mice was similar to the activity of
`the respective drugs in the respective human cancer.13-18
`These results were obtained using over 60 different established
`human cancer cell lines, all of which were injected subcutaneously.
`In no case was orthotopic injection of a cell line undertaken.
`
`PROSPECTIVE STUDIES USING SUBCUTANEOUS HUMAN
`CHILDHOOD TUMOR XENOGRAFTS
`Houghton and colleagues at St. Judes Children’s Hospital in
`Memphis have also undertaken an exhaustive series of pharmacoki-
`netic investigations in which a variety of pediatric malignancies were
`tested as subcutaneous xenografts in nude mice with respect to
`response to a variety of chemotherapeutic drugs. In particular, the
`relationship between systemic exposure and tumor response was
`evaluated, with emphasis on topoisomerase inhibitors such as
`irinotecan or topotecan.19-24 These studies showed that a panel of
`neuroblastoma xenografts was highly sensitive to irinotecan, espe-
`cially when administered using protracted schedules with lower
`
`doses of drug. For example, irinotecan was administered intra-
`venously (i.v.) daily 5 days per week for 2 consecutive weeks (defined
`as one cycle) and compared to more protracted low-dose schedules
`where cycles were repeated every 21 days for a total of three courses.
`In the latter the total amount of drug was 5–10 mg/kg and was given
`using a daily schedule for 5 days, which was repeated 2 out of every
`3 weeks for 9 weeks. Complete responses were observed in most of
`four of five xenografts using the intensive one cycle 40 mg/kg MTD
`schedule but the tumors tended to regrow. In contrast, with one
`exception, all neuroblastomas tested showed complete responses
`(CRs) which did not regrow during therapy when the protracted
`low-dose schedules were used involving a total dose of 10 mg/kg or
`5 mg/kg.23 Estimation of the lowest effective dose using the pro-
`tracted i.v. schedule indicated that neuroblastomas respond to daily
`doses as low as 1.25 mg/kg.23 It is interesting to consider these
`results in the light of those obtained by other investigators using a
`variety of similar protracted low-dose “metronomic” chemotherapy
`regimens as a putative antiangiogenic therapy, where increased effi-
`cacy and reduced toxicity have been frequently noted using such
`schedules, compared to the MTD of the same drug.25-31
`The preclinical studies of Houghton and colleagues were directly
`translated to the clinic where the same protracted schedule was used
`and found to be well tolerated in children with refractory solid
`tumors; in addition encouraging, if not remarkable, rates of clinical
`responses were observed as well—16 of 23 patients experienced
`
`S136
`
`Cancer Biology & Therapy
`
`2003; Vol. 2 Issue 4, Suppl. 1
`
`Ex. 1065-0003
`
`

`
`HUMAN TUMOR XENOGRAFTS AS PREDICTIVE PRECLINICAL THERAPY MODELS
`
`Figure 2. Results of an experiment published recently (Man et al. “Anti-tumor and anti-angiogenic effects in mice of low-dose (metronomic) cyclophosphamide
`administered continuously through the drinking water.” Cancer Res., 62: 2731-2735, 2002) in which a human breast cancer cell line was injected “ortho-
`topically” into the mammary fat pads of severe combined immunodeficient (SCID) mice, which allowed the tumor to metastasize to the lungs, liver and lymph
`nodes of the mice. Therapy was initiated when the “primary” intramammary fat pad tumor attained a size of 200 mm3 and the disease had metastasized
`in a microscopic fashion only. Mice were then administered cyclophosphamide through their drinking water on a continuous non-stop basis at an estimat-
`ed dose of 25 mg/kg per day, or treated with the DC101 anti-VEGFR-2 monoclonal antibody. In addition, another group of mice were given cyclophos-
`phamide in the MTD fashion, i.e., at 150 mg/kg once every two days over a 6-day period (indicated by the vertical arrows). This MTD regimen was high-
`ly toxic to the SCID mice and resulted in death within one to two weeks. In contrast, mice given the same drug metronomically showed no signs of toxicity
`despite receiving up to 3 times the cumulative maximum tolerated dose given acutely.
`
`stable disease and 5 showed partial responses.24 These results show
`that preclinical xenograft models, even those involving ectopic/sub-
`cutaneous transplants, can provide useful predictors of the activity
`and responses of some pediatric cancers to topoisomerase I
`inhibitors such as irinotecan. A more detailed overview and discus-
`sion of the testing of new agents in childhood cancer models, both
`xenografts and transgenic oncomouse models was recently published
`by Houghton et al.32
`
`IMPROVING HUMAN TUMOR XENOGRAFT MODELS
`FOR PREDICTING THE RELATIVE BENEFIT OF ANTI-CANCER
`DRUGS IN HUMANS—THE IMPORTANCE OF TREATING
`(ADVANCED) METASTATIC DISEASE
`While the results summarized above are encouraging, and clearly
`show the potentially predictive value of human tumor xenografts,
`there is an aspect of the results in many of the preclinical studies that
`is nevertheless troubling: the excellent, if not remarkable, nature of
`the tumor responses in mice, as such responses are infrequently
`observed in cancer patients even though the drug being tested may
`be active against its respective human counterpart. For example, as
`discussed above, Houghton et al. observed complete responses of
`established solid neuroblastoma xenografts in a high proportion of
`cases using various irinotecan dosing schedules, especially protracted
`low-dose protocols.23 However, such dramatic responses were not
`
`observed in the respective clinical trial of 23 patients, which included
`five children with neuroblastoma.24 It is this aspect of experimental
`therapy studies in mice that can be frustrating as it often attracts
`considerable attention (e.g., the endostatin studies of Boehm,
`O’Reilly et al. discussed above) and expectation. This disparity has
`caused considerable skepticism about what to expect in the clinic on
`the basis of prior preclinical therapy studies. However, this skepticism
`may not always be justified when one takes into account, in retrospect,
`a crucial and fundamental difference between virtually all published
`experimental mouse therapy studies and corresponding clinical trials,
`and it is this: in most phase I, II and III clinical trials the patients
`being treated have advanced, high-volume metastatic disease whereas
`most mouse studies do not test the effects of therapy on advanced
`metastatic disease, but rather on a primary tumor transplant or
`spontaneously arising primary tumor, or microscopic, low-volume
`metastatic disease (Lee Ellis, personal communication). With respect
`to treatment of metastatic disease, typically, in such experiments,
`tumor cells are injected intravenously to generate lung or liver tumor
`colonies (“artificial metastases”), and therapy is initiated within one
`or a few days after injection of the cells—if not before tumor cell
`injection! This constitutes a form of adjuvant (or prophylactic)
`therapy, on microscopic, low-volume metastatic disease. Alternatively,
`growing primary tumors may be surgically removed, and treatment
`then initiated within a few days when the spontaneous metastases
`that have formed are microscopic in size. Thus, there is a much less
`demanding therapeutic situation for mice than for humans, when it
`
`www.landesbioscience.com
`
`Cancer Biology & Therapy
`
`S137
`
`Ex. 1065-0004
`
`

`
`HUMAN TUMOR XENOGRAFTS AS PREDICTIVE PRECLINICAL THERAPY MODELS
`
`tively non-toxic and effective way of giving chemotherapy, and com-
`bining it with a targeted antiangiogenic drug.27,35-38
`Figure 2 shows the results of a similar experiment in which a
`human breast cancer (MDA-MB-231) was injected orthotopically in
`the mammary fat pads of female SCID mice, and then treated
`continuously with an oral low-dose regimen of cyclophosphamide
`administered continuously through the drinking water, combined
`with the same antiangiogenic drug, DC101.29 A control using an
`MTD regimen of cyclophosphamide was also used. In terms of
`survival, the best treatment regimen was the combination of the
`metronomic oral low-dose cyclophosphamide/ DC101, and the
`survival benefit was obvious. However, in this model, while the
`orthotopic breast cancer can metastasize, the metastases remain
`largely microscopic because of the retention of the primary tumor
`and the timing of the initiation of treatment. Thus, treatment of
`low-volume, metastatic disease was undertaken.
`More recent experiments have involved ‘raising the therapeutic
`bar’, so to speak. In Figure 3 a tumor cell line, called MDA-MB-435,
`supposedly a well known breast cancer cell line used extensively in
`breast cancer research, was injected into the mammary fat pads of
`SCID mice and allowed to grow for about one month. The resultant
`primary tumors were then surgically removed and initiation of treat-
`ment with oral low-dose cyclophosphamide and/or DC101 was
`delayed for about 10 days to allow establishment of extensive macro-
`scopic metastases in the lungs and draining lymph nodes of the
`SCID mice, as well as diffuse metastatic spread in the liver (data not
`shown). Using survival as an endpoint, neither DC101 alone or oral
`low-dose cyclophosphamide alone had had impact on survival; the
`combination did have an effect, but the magnitude of the benefit
`was rather modest in comparison to the sort of results shown in
`Figures 1 and 2. Of considerable interest, however, was the finding
`that a metronomic low-dose vinblastine protocol—0.33 mg/kg
`given intraperitoneally three times a week—alone caused complete
`resolution of advanced metastatic disease and greatly prolonged
`survival of the mice. Eventually, the mice had to be sacrificed
`because tumors recurred at the site of surgical removal and grew
`progressively in spite of the success of the therapy on distant metastatic
`disease (unpublished observations).
`It is of course difficult to compare the results of each experiment
`since different tumor cell lines and different treatment regimens
`were used. Indeed, the MDA-MB-435 ‘breast’ tumor cell line has
`recently been implicated to be a melanoma, based on gene and protein
`expression profiling,39,40 results which we have confirmed using the
`MDA-MB-435 line discussed in Figure 3. Nevertheless, the results
`of Figure 3 do suggest that treatment of advanced metastatic disease
`in mice will give results that may turn out to be much more reflec-
`tive, i.e., predictive, of the clinical situation typically encountered
`when testing new drugs in phase I, II or III clinical trials. The vin-
`blastine therapy results also point to the possibility that we cannot
`always assume that the response of a primary tumor will mirror the
`effects of the same therapy on distant metastases—this is obvious.
`What is not so obvious, and surprising, is that the response of metas-
`tases may be significantly better than the primary tumor in some
`cases. We would anticipate that this would be the exception rather
`than the rule; nevertheless this has ramifications for anti-cancer
`screening and drug testing, if correct.
`
`Figure 3. Effect of various therapy regimens on survival of SCID mice with
`advanced, metastatic cancer at the time of initiation of therapy. SCID mice
`were inoculated with MD-MBA-435 human tumor cells. The inoculation was
`into the mammary fat pads which facilitated distant metastatic spread,
`provided the primary tumors are surgically excised. This was done approx-
`imately 4 weeks after tumor cell inoculation, after which therapy was initi-
`ated approximately 8-9 days later. The cyclophosphamide was given
`continuously through the drinking water at an estimated dose of 20 mg/kg
`per day, whereas vinblastine or taxol at the indicated dose were injected at
`low-dose twice a week.
`
`comes to comparing most preclinical trials to the clinical trial coun-
`terparts. Perhaps much of the disparity in results between the two is
`related to this variable since it is well known that high-volume
`advanced metastatic disease is generally much more difficult to treat
`than low-volume adjuvant disease. Add to this the fact that many
`patients entered into clinical trials had been treated previously with
`other therapies and have relapsed with refractory disease. Heavily
`pretreated and resistant patients are often less responsive to a new
`therapy, and usually have advanced metastatic disease at the time of
`entry into a clinical trial.33 How often have investigators in the past
`tested a new drug or therapy in mice where this dire clinical situa-
`tion is recapitulated? The answer is rarely—if ever.
`To illustrate the point about treating (advanced) metastatic
`disease, some recent results obtained in this laboratory are shown.
`Figure 1 shows the results of an experiment in which a metronomic
`low-dose vinblastine protocol, in combination with an antiangio-
`genic drug, called DC101 (an anti-VEGF receptor-2 blocking
`antibody) was used to treat large, established human neuroblastoma
`xenografts obtained after subcutaneous injection of SK-NM-C
`cells.27 The results showed a remarkable anti-tumor effect could be
`obtained with the combination—sustained and complete tumor
`regressions. In effect, the mice were cured since the therapy was
`continuously maintained for 7 months,27 and surprisingly, tumors
`did not resume growth when the treatment was finally terminated
`(unpublished observations). However, because the tumors were
`injected subcutaneously (i.e., ectopically) they did not metastasize,
`and therefore the much more demanding clinical situation of treating
`advanced metastatic neuroblastoma metastases was not duplicated in
`the mouse studies. The preclinical study was not intended to predict
`clinical activity—as implied by a headline proclaimed on the front
`page of a prominent national Canadian newspaper,34 but to illus-
`trate the principle of metronomic low-dose chemotherapy as a rela-
`
`S138
`
`Cancer Biology & Therapy
`
`2003; Vol. 2 Issue 4, Suppl. 1
`
`Ex. 1065-0005
`
`

`
`HUMAN TUMOR XENOGRAFTS AS PREDICTIVE PRECLINICAL THERAPY MODELS
`
`CONCLUSIONS
`In light of these results one might want to rethink Dr. Folkman’s
`quote “if you are a mouse and have cancer, we can take good care of
`you”.5 One may argue this applies to mice with rapidly growing,
`transplanted, subcutaneous, encapulated/non-metastatic tumors. In
`contrast, mice with high-volume, advanced, metastatic disease in
`sites such as the lungs, liver and brain may not be so easy to take care
`of, similar to their human counterparts. The vinblastine results do
`however provide some basis for optimism, and emphasize the need
`to begin testing models which involve advanced metastatic disease.
`This, incidentally, is one of the limitations of many of the current
`transgenic oncomouse models, as they usually do not spontaneously
`metastasize.41,42 Moreover, monitoring the effects of therapies on
`metastatic disease in mice is becoming easier and less subjective with
`the growing use of small animal non-invasive micro-imaging
`research tools43 and non-invasive biochemical techniques, e.g.,
`measuring secreted tumor-specific protein markers that can be
`introduced into tumor cell lines.29,44 It is also time to reexamine
`some of the current dogmas regarding mouse models of cancer. First,
`human tumor xenografts can be surprisingly predictive of clinical
`activity, and in some cases this includes subcutaneous/ectopic trans-
`plants. The wisdom of the rush towards exclusive use of much more
`expensive transgenic oncomouse models for drug therapy testing can
`be questioned, especially when such tumors fail to express the most
`critical element of malignant disease—ability to metastasize, and the
`fact that less expensive transplantable tumor models are available
`which work—if used appropriately.
`
`Acknowledgments
`I thank Ms. Cassandra Cheng for her outstanding secretarial assistance,
`and Shan Man for her excellent technical assistance in the experiments
`reported in Figures 2 and 3. Dr. Lee Ellis pointed out the lack of prior exper-
`imental/preclinical studies in which high-volume metastatic disease is treated.
`
`References
`1. Shimosato Y, Kameya T, Nagai K, Hirohashi S, Koide T, Hayashi H, Nomura T.
`Transplantation of human tumors in nude mice. J Natl Cancer Inst 1976; 56:1251-60.
`2. Takimoto CH. Why drugs fail: of mice and men revisited. Clin Cancer Res 2001;
`7:229-30.
`3. BoehmT, Folkman J, Browder T, and O'Reilly MS. Antiangiogenic therapy of experimen-
`tal cancer does not induce acquired drug resistance. Nature 1997; 390:404-7.
`4. Kerbel RS. A cancer therapy resistant to resistance. Nature 1997; 390:335-6.
`5. Kolata G. A cautious awe greets drugs that eradicate tumors in mice. The New York Times,
`May 3, 1998.
`6. Twombly R. First clinical trials of endostatin yield lukewarm results. J Natl Cancer Inst
`2002; 94:1520-1.
`7. Eder JP Jr, Supko JG, Clark JW., Puchalski TA, Garcia-Carbonero R, et al. Phase I clinical
`trial of recombinant human endostatin administered as a short intravenous infusion repeat-
`ed daily. J Clin Oncol 2002; 20:3772-84.
`8. Kalb C. Hype, hope, cancer. Newsweek , 73. 12-28-1998.
`9. Jackson-Grusby L. Modeling cancer in mice. Oncogene 2002; 21:5504-14.
`10. Van Dyke T, Jacks T. Cancer modeling in the modern era: progress and challenges. Cell,
`2002; 108:135-44.
`11. Fidler,I.J. Rationale and methods for the use of nude mice to study the biology and thera-
`py of human cancer metastasis. Cancer Metastasis Rev 1986; 5:29-49.
`12. Hoffman RM. Orthotopic transplant mouse models with green fluorescent protein-
`expressing cancer cells to visualize metastasis and angiogenesis. Cancer Metastasis Rev
`1998; 17:271-7.
`13. Maruo K, Ueyama Y, Inaba M, Emura R, Ohnishi Y, Nakamura O, et al. Responsiveness
`of subcutaneous human glioma xenografts to various antitumor agents. Anticancer Res
`1990; 10:209-12.
`14. Inaba M, Tashiro T, Kobayashi T, Sakurai Y, Maruo K, Ohnishi Y, et al. Responsiveness of
`human gastric tumors implanted in nude mice to clinically equivalent doses of various anti-
`tumor agents. Jpn J Cancer Res 1988; 79:517-22.
`15. Inaba M, Kobayashi T, Tashiro T, Sakurai Y, Maruo K, Ohnishi Y, et al. Evaluation of anti-
`tumor activity in a human breast tumor/nude mouse model with a special emphasis on
`treatment dose. Cancer 1989; 64:1577-82.
`
`16. Tashiro T, Inaba M, Kobayashi T, Sakurai Y, Maruo K, Obnishi Y, et al. Responsiveness of
`human lung cancer/nude mouse to antitumor agents in a model using clinically equivalent
`doses. Cancer Chemother Pharmacol 1989; 24:187-92.
`17. Inaba M, Kobayashi T, Tashiro T, Sakurai Y. Pharmacokinetic approach to rational thera-
`peutic doses for human tumor-bearing nude mice. Jpn J Cancer Res 1988; 79:509-16.
`18. Thompson J, Zamboni WC, Cheshire PJ, Richmond L, Luo X, Houghton JA, et al.
`Efficacy of oral irinotecan against neuroblastoma xenografts. Anticancer Drugs 1997; 8: 313-22.
`19. Vassal G, Boland I, Santos A, Bissery MC, Terrier-Lacombe MJ, Morizet J, et al