In malignant melanoma (MM), the tumor marker S100B binds directly to wild-type p53, dissociates the p53 tetramer, enhances hdm2-dependent ubiquitination of p53, and down-regulates p53-dependent tumor suppression functions (1
). As a proof of principle for drug design, inhibiting S100B with small interfering antisense RNA (siRNAS100B) or with several S100B
(SBiXs; X = compound number) was achieved and shown to restore wild-type p53 at the protein level. Inhibiting S100B production was also found to cause an increase in the levels of p53 gene products as necessary to induce cell growth arrest and apoptosis in malignant melanoma (2
). Also encouraging from a drug development standpoint is that ablation of S100B expression in mice via gene targeting produces unremarkable phenotypes with few, if any, problematic physiological consequences (8
With any drug-design program, there is a need to obtain physiological data at an early stage in the process to help determine whether a compound or a series of compounds induce off-target effects and/or cause other unanticipated toxicities. This is particularly important for S100 inhibitors since there are over 20 structurally similar proteins in the S100 protein family, and they each regulate several physiologically important pathways in a cell-specific manner (14
). Thus, an S100 inhibitor could have multiple phenotypes depending on the number of S100 proteins it blocked and the S100 status of the cell-type targeted. In the case of blocking the S100B–p53 protein–protein interaction in malignant melanoma, there exists a separate issue with regard to a feedback loop that is initiated when S100B is inhibited since the gene for S100B itself is up regulated by the tumor suppressor protein, when p53 levels are restored (2
). To address these and other issues, this chapter describes the discovery/development of SBiXs for melanoma therapy with a focus on the importance of doing in vivo screening of lead compounds at an early stage in the drug development process. It is likely that such an approach of performing early in vivo screening could benefit many drug discovery programs.
Small molecule inhibitors have been reported for numerous S100 family members, some of which are in clinical trials. Inhibitors of the S100A10–annexin A2 interaction have not been validated in vivo but their predicted clinical uses include angiogenesis and cancer metastasis therapy (16
). The anti-allergic drug cromolyn disrupts S100P–RAGE interaction and reduces pancreatic tumor formation in animal models (18
). Interestingly, cromolyn also binds other S100 family members (S100A1, S100A12, and S100A13), but its effects on the interaction of these family members with their target proteins have not been fully investigated. Other small molecules that bind S100A1 include pentamidine and propanolol (19
). In the case of S100A4, several phenothiazines block S100A4-mediated depolymerization of myosin-IIA filaments (20
). S100A4 also binds anti-allergic drugs and a modified version of azaxanthone, which is in clinical trials for treatment of metastatic disease states (21
). Two SBiXs developed by our group disrupt S100B–p53 complex formation and prevent unregulated melanoma cell growth and are currently being tested in human and veterinary clinical trials as potential melanoma therapeutics.
For the development of SBiXs, a combination of computer-aided drug design (CADD), high-throughput screening (HTS), structural biology, medicinal chemistry, and in vivo biology/drug testing approaches is employed. The benefits of a structure-based approach are taken advantage of for developing new compounds as well as addressing issues with regard to target specificity since the 3D structures for several S100 and S100–target complexes are published (4
). Likewise, structure/activity relationships (SAR) and SAR by NMR approaches are more efficiently directed when the structures are available (28
). This is especially important for the difficult hurdle of inhibiting a protein–protein interaction such as the S100B–p53 complex (4
). More specifically, modifications on a particular scaffold that depend on the structure of a lead compound bound to S100B are achieved iteratively by using CADD-directed medicinal chemistry; likewise, structural studies can be used to identify multiple sites within the p53-binding cleft on S100B via structural biology approaches (NMR, X-ray), so compounds and combinations of compounds can be rationally linked together synthetically to obtain tight and highly specific S100B binders (i.e., relative to binding S100A1 for example). Such a drug development process is very dependent on the latest new results, so this process occurs in a very iterative manner (). Once promising leads are identified, then it is important that they are screened for their in vivo efficacy and toxicity. Such leads are chosen based on binding affinities (KD
s) and/or specificity both in biochemical assays (i.e., S100B thermodynamic/kinetic data versus other S100s) and in cellular assays (i.e., IC50
s). For the cellular assays, specificity and other potential off-target effects are evaluated by comparing IC50
values of isogenic cell lines plus/minus S100B present (31
). Of the approximately 100–200 lead compounds that undergo extensive in vitro testing each year, eight to ten typically meet the criteria for early-stage in vivo screening.
Figure 1 Strategy for the discovery and development of SBiXs. An iterative process that includes computer-aided drug design (CADD), high-throughput screening (HTS), structure/activity (SAR) and medicinal chemistry, 3D structure (NMR and X-ray), and a specificity (more ...)
High-throughput biochemical and cellular screens identify two types of lead SBiXs: FDA-approved compounds and new chemical entities. Since there are no effective treatments for melanoma, repurposing of FDA-approved lead SBiXs for melanoma therapy is a high priority. Efficacy evaluation is the primary focus for repurposing approved drugs, since extensive pharmacological/toxicological information is already available for many species, including humans. In fact, phase II human and phase I canine clinical trials for two FDA-approved SBiXs are underway. However, like most drug discovery programs, the vast majority of lead SBiXs are new chemical entities that require extensive development and optimization. In vivo testing in animal models plays a fundamental role in developing new anticancer drugs. Prior to human testing, preclinical tolerability (maximum tolerated dose, MTD), pharmacokinetics (PK), pharmacodynamic (PD), and efficacy assays provide key information that is used to improve advanced leads. For example, a compound may need to be more lipophilic to pass through the cell membrane and reach its target or side groups may need to be added to allow for oral delivery or brain penetration.
In the case of accessible tumors such as melanoma, drugs can be delivered directly to the tumor (intratumoral) without optimization for systemic delivery and/or minimization of toxic effects on normal cells. In addition, intratumoral delivery can achieve significantly higher drug concentrations at the site of action than systemic delivery. In the clinical setting, intratumoral administration has been used to deliver gene therapy constructs, complex biologics, and small molecules to a variety of cancers including adenoviral based p53 genes in head and neck cancer (33
); TNFalpha genes in rectal cancer (34
); interleukin-2 in melanoma (35
); immunostimulant CpG in brain cancers (36
); single treatment of a meta-static squamous cell carcinoma lesion (37
); BCNU in combination with radiotherapy in glioma (38
); and para-toluenesulfonamide in non-small cell lung cancer (39
). Intratumoral delivery has been used in the preclinical evaluation of siRNAs (40
) and immune modulators (41
) as well as to reduce the toxicity of approved agents such as melphalan (43
); 2-deoxy- D -glucose alone and in combination with carboplatin (44
); and paclitaxel/docetaxel in mammary, bladder, prostate, and head and neck cancers (45
). Our S100B inhibitor drug development program includes an intratumoral in vivo screen of lead SBiXs before preclinical testing of advanced leads. While in vivo testing prior to medicinal chemistry optimization and ADME testing is atypical, the availability of in vivo screening data early in the drug discovery process focuses resources on SBiXs with the greatest probability of clinical success.
A major consideration for any in vivo trial is the selection of an appropriate animal model. Mouse models are a mainstay in animal testing because it is feasible to perform studies in a short period of time. However, no single mouse model recapitulates the complex genetics and biology of human melanoma (49
). Therefore, the biological question being asked as well as the advantages and limitations of the various types of models were important criteria in selecting an animal model for intratumoral in vivo screening of SBiXs. Syngeneic transplantation models involve the implantation of well-characterized melanoma cell lines into a syngeneic host and do not recapitulate many aspects of the human disease because characterized mouse lines do not reflect the heterogeneity observed in human tumors. Xenogeneic transplantation models involve the implantation of cell lines or patient-derived cells into an immunocompromised host. Both of these transplantation models can be flank models in which subcutaneous tumors are grown on the backs of mice or orthotopic models in which tumors are grown at the site of origin (i.e., breast cancer cells in the mammary fat pad). Orthotopic models have a better chance to metastasize but are more difficult to monitor. One disadvantage of human xenograft models is the fact that the host is immunocompromised and immunosurvelliance has been shown to play a role in limiting metastasis (53
). In the case of SBiXs, S100B’s effects on immune responses in the CNS are well documented although little is known about its role in peripheral immune responses. Nonetheless, human xenograft models are usually considered to be superior to genetically modified mouse models because the composition of the resulting tumor mimics the heterogeneity observed in patients. However, the development of multi-allelic genetically engineered mouse models that mimic spontaneous tumorigenesis and heterogeneity as well as target validation in simulated clinical trials using standard of care chemotherapeutics confirm the utility of genetically modified mouse models in developing cancer therapeutics (54
The RAS-induced INK4a/ARF−/−
mouse melanoma model (55
) was chosen for in vivo screening of lead SBiXs because it has (1) an intact S100B–p53 signaling pathway (elevated S100B and wild-type p53), (2) an intact immune system, (3) tumors which are amenable to intratumoral delivery, and (4) a proven record in developing new melanoma therapies (55
). The Tyr::RASG12V
line is bigenic and contains two genomic mutations on an FVB background: a mutated H-rasG12V
transgene on the Y chromosome and inactivated INK4a/ARF alleles on chromosome 4. This model is not commercially available and we maintain a breeding colony that generates experimental animals as well as breeders for the individual Tyr::RASG12V
and INK4a/ ARF−/−
lines. Founders for both lines were obtained from the National Cancer Institute Mutant Mouse Resource (Frederick, MD). At 2–3 months of age experimental Tyr::RASG12V
males develop spontaneous cutaneous melanomas in the pinna of the ears (30%), torso (23%), and tail (20%) without distant metastasis (55
). Our intratumoral in vivo screening protocol for modified SBiXs is a longitudinal design with a study period of 3–8 weeks and uses the relative tumor proliferation rate (the tumor volume at a particular treatment interval/tumor volume at the time of treatment initiation) as the primary outcome (). Although this trial design is not optimized for garnering tolerability, PK or PD information, the gross/histological pathology, SBiX levels, and p53 pathway reactivation in the tumors are monitored and this information is useful in selecting advanced leads that will proceed to preclinical testing.
Figure 2 Intratumoral in vivo screening trial. The histogram depicts the mean tumor responsiveness for intratumoral injection of SBi931 (red bars) or vehicle only (black bars) every other day for 21 days. Asterisk denotes p < 0.05 when compared to vehicle (more ...)
In the case of modified leads with predicted ADME properties favorable for systemic administration, concurrent tolerability (MTD) and pharmacokinetic (PK) assays are also conducted. These trials provide valuable pharmacological information that cannot be garnered from local administration trials including potential effects of the route of administration on tumor responsiveness (57
). MTD and PK studies are conducted in the same species/strain and employ the same dosing scheme that will be used in subsequent efficacy studies. For example, if the expected SBiX dosing scheme is 2 weeks of consecutive intraperitoneal (IP) administration in female nu/nu mice, then the MTD experiment should be dosed IP daily × 14 in female nu/nu mice. The route of administration, oral (PO), intraperitoneal (IP), intravenous (IV), or subcutaneous (SC), is determined by the predicted chemical properties of the SBiX. In MTD experiments, animals are monitored post cessation of dosing for delayed toxicity for a minimum of 1 week and optimally 2 weeks. In mouse models, the best indicator for toxicity is weight. MTDs for novel cancer drugs are considered an LD10
, the dose at which 10% of mice in one group die or lose 20% of their body weight (58
) (). The therapeutic index for cancer drugs is very small in that the MTD can be equivalent to the effective dose. The goal of PK studies is to select a dose and route of administration that allow the plasma/target tissue drug levels to exceed the IC50
value from cellular assays and/or KD
from biochemical screens. SBiX levels in the tumor tissue provide useful information about tumor penetrance, half-life in target tissue, and the compound’s ability to leave the bloodstream. Also, SBiX levels in brain tissue assess brain penetration and potential for use in metastatic disease therapy and primary brain cancers. The data from the intratumoral in vivo screening, MTD, and PK trials described below determines if modified leads are suitable for preclinical testing, require additional medicinal chemistry optimization and further testing prior to preclinical testing, or are eliminated.
Figure 3 Systemic tolerability testing/MTD. The graph depicts the percent body weight loss verses time for groups (n = 3 mice) receiving daily IP doses (mid, low, or high) of a drug for 7 consecutive days. The decrease in mean body weight in the high-dose group (more ...)