Stephen W. Fesik

An inhibitor of Bcl-2 family proteins induces regression of solid tumours.

Proteins in the Bcl-2 family are central regulators of programmed cell death, and members that inhibit apoptosis, such as Bcl-XL and Bcl-2, are overexpressed in many cancers and contribute to tumour initiation, progression and resistance to therapy. Bcl-XL expression correlates with chemo-resistance of tumour cell lines, and reductions in Bcl-2 increase sensitivity to anticancer drugs and enhance in vivo survival. The development of inhibitors of these proteins as potential anti-cancer therapeutics has been previously explored, but obtaining potent small-molecule inhibitors has proved difficult owing to the necessity of targeting a protein–protein interaction. Here, using nuclear magnetic resonance (NMR)-based screening, parallel synthesis and structure-based design, we have discovered ABT-737, a small-molecule inhibitor of the anti-apoptotic proteins Bcl-2, Bcl-XL and Bcl-w, with an affinity two to three orders of magnitude more potent than previously reported compounds. Mechanistic studies reveal that ABT-737 does not directly initiate the apoptotic process, but enhances the effects of death signals, displaying synergistic cytotoxicity with chemotherapeutics and radiation. ABT-737 exhibits single-agent-mechanism-based killing of cells from lymphoma and small-cell lung carcinoma lines, as well as primary patient-derived cells, and in animal models, ABT-737 improves survival, causes regression of established tumours, and produces cures in a high percentage of the mice.

Discovering High-Affinity Ligands for Proteins: SAR by NMR

A nuclear magnetic resonance (NMR)-based method is described in which small organic molecules that bind to proximal subsites of a protein are identified, optimized, and linked together to produce high-affinity ligands. The approach is called “SAR by NMR” because structure-activity relationships (SAR) are obtained from NMR. With this technique, compounds with nanomolar affinities for the FK506 binding protein were rapidly discovered by tethering two ligands with micromolar affinities. The method reduces the amount of chemical synthesis and time required for the discovery of high-affinity ligands and appears particularly useful in target-directed drug research.

ABT-263: A Potent and Orally Bioavailable Bcl-2 Family Inhibitor

Overexpression of the prosurvival Bcl-2 family members (Bcl-2, Bcl-xL, and Mcl-1) is commonly associated with tumor maintenance, progression, and chemoresistance. We previously reported the discovery of ABT-737, a potent, small-molecule Bcl-2 family protein inhibitor. A major limitation of ABT-737 is that it is not orally bioavailable, which would limit chronic single agent therapy and flexibility to dose in combination regimens. Here we report the biological properties of ABT-263, a potent, orally bioavailable Bad-like BH3 mimetic (K(i)s of <1 nmol/L for Bcl-2, Bcl-xL, and Bcl-w). The oral bioavailability of ABT-263 in preclinical animal models is 20% to 50%, depending on formulation. ABT-263 disrupts Bcl-2/Bcl-xL interactions with pro-death proteins (e.g., Bim), leading to the initiation of apoptosis within 2 hours posttreatment. In human tumor cells, ABT-263 induces Bax translocation, cytochrome c release, and subsequent apoptosis. Oral administration of ABT-263 alone induces complete tumor regressions in xenograft models of small-cell lung cancer and acute lymphoblastic leukemia. In xenograft models of aggressive B-cell lymphoma and multiple myeloma where ABT-263 exhibits modest or no single agent activity, it significantly enhances the efficacy of clinically relevant therapeutic regimens. These data provide the rationale for clinical trials evaluating ABT-263 in small-cell lung cancer and B-cell malignancies. The oral efficacy of ABT-263 should provide dosing flexibility to maximize clinical utility both as a single agent and in combination regimens.

Promoting apoptosis as a strategy for cancer drug discovery

Apoptosis is deregulated in many cancers, making it difficult to kill tumours. Drugs that restore the normal apoptotic pathways have the potential for effectively treating cancers that depend on aberrations of the apoptotic pathway to stay alive. Apoptosis targets that are currently being explored for cancer drug discovery include the tumour-necrosis factor (TNF)-related apoptosis-inducing ligand (TRAIL) receptors, the BCL2 family of anti-apoptotic proteins, inhibitor of apoptosis (IAP) proteins and MDM2.

Structural Basis for the Inhibition of Caspase-3 by XIAP

The molecular mechanism(s) that regulate apoptosis by caspase inhibition remain poorly understood. The main endogenous inhibitors are members of the IAP family and are exemplified by XIAP, which regulates the initiator caspase-9, and the executioner caspases-3 and -7. We report the crystal structure of the second BIR domain of XIAP (BIR2) in complex with caspase-3, at a resolution of 2.7 A, revealing the structural basis for inhibition. The inhibitor makes limited contacts through its BIR domain to the surface of the enzyme, and most contacts to caspase-3 originate from the N-terminal extension. This lies across the substrate binding cleft, but in reverse orientation compared to substrate binding. The mechanism of inhibition is due to a steric blockade prohibitive of substrate binding, and is distinct from the mechanism utilized by synthetic substrate analog inhibitors.

Structural basis for binding of Smac/DIABLO to the XIAP BIR3 domain

The inhibitor-of-apoptosis proteins (IAPs) regulate programmed cell death by inhibiting members of the caspase family of enzymes. Recently, a mammalian protein called Smac (also named DIABLO) was identified that binds to the IAPs and promotes caspase activation. Although undefined in the X-ray structure, the amino-terminal residues of Smac are critical for its function. To understand the structural basis for molecular recognition between Smac and the IAPs, we determined the solution structure of the BIR3 domain of X-linked IAP (XIAP) complexed with a functionally active nine-residue peptide derived from the N terminus of Smac. The peptide binds across the third β-strand of the BIR3 domain in an extended conformation with only the first four residues contacting the protein. The complex is stabilized by four intermolecular hydrogen bonds, an electrostatic interaction involving the N terminus of the peptide, and several hydrophobic interactions. This structural information, along with the binding data from BIR3 and Smac peptide mutants reported here, should aid in the design of small molecules that may be used for the treatment of cancers that overexpress IAPs.

14-3-3 proteins and survival kinases cooperate to inactivate BAD by BH3 domain phosphorylation.

The Bcl-2 homology 3 (BH3) domain of prodeath Bcl-2 family members mediates their interaction with prosurvival Bcl-2 family members and promotes apoptosis. We report that survival factors trigger the phosphorylation of the proapoptotic Bcl-2 family member BAD at a site (Ser-155) within the BAD BH3 domain. When BAD is bound to prosurvival Bcl-2 family members, BAD Ser-155 phosphorylation requires the prior phosphorylation of Ser-136, which recruits 14-3-3 proteins that then function to increase the accessibility of Ser-155 to survival-promoting kinases. Ser-155 phosphorylation disrupts the binding of BAD to prosurvival Bcl-2 proteins and thereby promotes cell survival. These findings define a mechanism by which survival signals inactivate a proapoptotic Bcl-2 family member, and suggest a role for 14-3-3 proteins as cofactors that regulate sequential protein phosphorylation events.

Specificity of short interfering RNA determined through gene expression signatures

Short interfering RNA (siRNA) is widely used for studying gene function and holds great promise as a tool for validating drug targets and treating disease. A critical assumption in these applications is that the effect of siRNA on cells is specific, i.e., limited to the specific knockdown of the target gene. In this article, we characterize the specificity of siRNA by applying gene expression profiling. Several siRNAs were designed against different regions of the same target gene for three different targets. Their effects on cells were compared by using DNA microarrays to generate gene expression signatures. When the siRNA design and transfection conditions were optimized, the signatures for different siRNAs against the same target were shown to correlate very closely, whereas the signatures for different genes revealed no correlation. These results indicate that siRNA is a highly specific tool for targeted gene knockdown, establishing siRNA-mediated gene silencing as a reliable approach for large-scale screening of gene function and drug target validation.

Solution structure of the antiapoptotic protein bcl-2

The structures of two isoforms of Bcl-2 that differ by two amino acids have been determined by NMR spectroscopy. Because wild-type Bcl-2 behaved poorly in solution, the structures were determined by using Bcl-2/Bcl-xL chimeras in which part of the putative unstructured loop of Bcl-2 was replaced with a shortened loop from Bcl-xL. These chimeric proteins have a low pI compared with the wild-type protein and are soluble. The structures of the two Bcl-2 isoforms consist of 6 α-helices with a hydrophobic groove on the surface similar to that observed for the homologous protein, Bcl-xL. Comparison of the Bcl-2 structures to that of Bcl-xL shows that although the overall fold is the same, there are differences in the structural topology and electrostatic potential of the binding groove. Although the structures of the two isoforms of Bcl-2 are virtually identical, differences were observed in the ability of the proteins to bind to a 25-residue peptide from the proapoptotic Bad protein and a 16-residue peptide from the proapoptotic Bak protein. These results suggest that there are subtle differences in the hydrophobic binding groove in Bcl-2 that may translate into differences in antiapoptotic activity for the two isoforms.

Heteronuclear three-dimensional nmr spectroscopy. A strategy for the simplification of homonuclear two-dimensional NMR spectra

Recently, three-dimensional Fourier transform techniques have been suggested (I3) as a possible method to resolve the spectra of complicated systems. Analogous to the comparison of 2D with 1 D NMR, three-dimensional NMR spectroscopy allows an increase in resolution and offers additional correlations for the elucidation of scalarand dipolar-coupled spin networks as compared to 2D NMR. Homonuclear 3D NMR experiments have been reported which emp loy tailored soft pulses (I, 2) to excite only parts of the proton spectral window or which use the narrow frequency band of J couplings for the third dimension (3) in order to reduce the experimental accumulation time. In this communication, we propose heteronuclear 3D NMR spectroscopy to aid in the simplification of two-dimensional proton NMR spectra of isotopically labeled compounds. For proteins, labels can easily be introduced by adding isotopically labeled nutrients to the growth med ium of bacterial expression systems (e.g., Ref. (4)). The experiments described comprise a combination of heteronuclear mu ltiple-quantum correlation (HMQC) (5, 6) with homonuclear COSY and NOESY and may therefore be named HMQC-COSY and HMQC-NOESY. Heteronuclear 3D experiments hold considerable promise for application to larger systems because one of the coherence transfer steps involves scalar couplings much larger than the ‘H linewidths. F igure 1 shows the pulse sequences of the HMQC-COSY and HMQC-NOESY experiments. Analogous to the HMQC experiment, heteronuclear mu ltiple-quantum coherence created by the first two RF pulses evolves during tI effectively with the single-quantum heteronuclear chemical shift. This coherence is converted to proton single-quantum coherence ant iphase with respect to the heteronucleus by the second hetero pulse, which is refocused at the end of the second 7 period. On ly those protons which are coupled to the heteronucleus are frequency labeled during tl . Therefore, the further frequency labeling during t2 under the homonuclear Hamiltonian needs to cover only the heteronuclear coupled protons, thus allowing nonselective RF pulses to be used throughout the sequence. Protons not coupled to the heteronuclei, which may be considered axial peaks in ol, are suppressed by an inversion of the first hetero pulse concomitant with an inversion of the receiver phase (6). Depending on the experiment, the signals arising from heteronuclear coupled protons are converted to COSY (Fg 1A) or NOESY (Fii 1B) type responses. In order to enhance sen-