BMS-986094? You might know this molecule better as Inhibitex’s former nucleoside INX-089. The molecule came to BMS through its $2.5 billion purchase of Inhibitex in 2011, as we wrote last year here at the Haystack.

The molecule belongs to a family of new nucleosides with fairly common structural motifs: a central sugar appended to a nitrogen heterocycle (usually purine- or uracil-based) and an elaborate phosphoramidate prodrug. These new drugs’ similarities may also prove to be their Achilles heel – Idenix Pharmaceuticals announced an FDA-requested partial clinical hold on their IDX-184 lead. This cautious approach aims to protect patients; though the drugs are similar, 184’s main structural difference – a thioester-based, slightly more-polar prodrug – seems to be enough to distance it from the cardiac side-effects seen with BMS-986094.

For a fairly in-depth look at the chemistry behind these inhibitors, as well as dozens of other analogues that never made it to prime time, check out US Patent 7,951,789 B2, issued to Idenix just last year.

Please welcome a new screening technique: ¹³C NMR. How can that work, given the low, low natural abundance of ~1.1% Carbon-13?

Researchers at UT-Southwestern Medical Center have the answer: rig the system. Jamie Rogers and John MacMillan report in JACS ASAP ¹³C-labeled versions of several common drug fragments, which they use to screen new biocatalyzed reactions.

Biocatalysis = big business for the pharma world. The recent Codexis / Merck partnership for HCV drug boceprevir brought forth an enzyme capable of asymmetric amine oxidation. Directed evolution of an enzyme made sense here, since they knew their target structure, but what if we just want to see if microbes will alter our molecules?

Enter the labeled substrates: the researchers remark that they provide an “unbiased approach to biocatalysis discovery.” They’re not looking to

Credit: JACS | UT-Southwestern, 2012

accelerate a certain reaction per se, but rather searching for any useful modifications using the ¹³C “detector” readout. One such labeled substrate, N-(¹³C)methylindole, shows proof-of-concept with their bacterial library, producing two different products (2-oxindole and 3-hydroxyindole) depending on the amount of oxygen dissolved in the broth. NMR autosamplers make reaction monitoring a snap, and in short order, the scientists show biotransformations of ten more indole substrates.

This paper scratches multiple itches for various chem disciplines. Tracking single peaks to test reactions feels spookily close to ³¹P monitoring of metal-ligand catalysis. Organickers, no strangers to medicinally-relevant indole natural products, now have another stir-and-forget oxidation method. Biochemists will no doubt wish to tinker with each bacterial strain to improve conversion or expand scope. The real question will be how easily we can incorporate ¹³C labels into aromatic rings and carbon chains, which would greatly increase the overall utility.

Why have our drugs stopped working?

Recent research from St. Jude’s (Science, 2012, 1110) attempted to answer that question. Using X-ray crystallography, a technique used to see structures at the atomic level, the researchers were able to capture a critical moment when a drug binds to DHPS, its bacterial enzyme target. The scientists could then predict how bacteria evolve to dodge further biocidal bullets.

The antibacterial medicines caught in the act by the St. Jude’s researchers are the sulfa drugs (see right), former front-line treatments many doctors push to the bottom of treatment regimens, due to increasingly resistant bacterial strains. Researchers knew resistance had something to do with the drugs’ mechanism of action; sulfa drugs mimic the binding of PABA – para-aminobenzoic acid, a compound found in many sunscreens (Chemical Note: PABA occurs naturally as bacterial vitamin H¹, and can also be found in yeast and plants. Chemists often borrow naturally-occurring compounds for industrial uses; two prominent examples are vanillin and Vitamin C).

Disruption of this PABA binding shuts down bacterial DNA replication, stopping reproduction. Before now, however, no one had succeeded in growing crystals of the active site that actually showed the drugs interacting with the enzymatic intermediate.

Let’s take one more step back: how does PABA attach itself? The enzyme we’re discussing, DHPS, catalyzes bond formation between PABA and intermediates known as pterins (see picture, left). Earlier researchers believed that this molecular hook-up operated by an S_N2 mechanism, a reaction where the PABA kicks out a small piece of the pterin to form a new C-N bond. We chemists would say that S_N2 means concerted bond formation, meaning that PABA would bind at the same time as the leaving group (OPPi), well, leaves.

Turns out that picture’s not quite right: it’s more S_N1-like, which means that the pterin first forms a positively-charged, enzyme-stabilized species! As you can imagine, this is no small feat, since the reaction works at physiological pH, in water, which could hydrate the intermediate (but doesn’t). Nope – instead, this charged molecule sits around waiting for a PABA – or a sulfa drug – to bind to it. When PABA binds, the complex exits the enzyme, but when the drug binds, it locks up the active site.

So how do these models help us to understand resistance?

The group noticed something odd: sulfathiazole (STZ) and sulfamethoxazole (SMX), two standard sulfas, both bound in the normal PABA cavity of DHPS. Unlike PABA, however, they hang their heterocyclic rings “outside” the normal pocket. The researchers built upon earlier observations by another group (Proc. Natl. Acad. Sci. U.S.A., 2010, 20986), speculating that the resistance might not have to do with the active site at all: it’s the external region, where the heterocycle bumps into the protein. To cheat death, all the bacterium needs to do is mutate an amino acid from this “outside” region (nearby proline and phenylalanine residues, see picture), which shuts down drug binding.

Could we design better drugs based on this model? Sure, we could synthesize a complimentary heterocycle, one that binds to the “outside” of mutant

DHPS bound to STZ - Source: Science / St. Jude's

enzymes (more polar for certain mutations, less for others). Another option? Cut the drug down to size: sulfonilamide, the grandfather of the sulfa drugs, should fit almost as snugly in the cavity as PABA, which might function perfectly against resistant bugs.

A new review, currently out on J. Med. Chem. ASAP, lays out all the computational backing behind this venture. Three computational chemists (David Huggins, Woody Sherman, and Bruce Tidor) break down five binding events from the point-of-view of the drug target: Shape Complementarity, Electrostatics, Protein Flexibility, Explicit Water Displacement, and Allosteric Modulation….whew!

Selectivity Strategies for Rational Design | Credit: Huggins, Sherman, Tidor; J. Med. Chem.

Note: Before we dive into this article, let’s clarify a few terms computational drug-hunters use that bench chemists think of differently: ‘decoy’ – a test receptor used to perform virtual screens; ‘ligand’ – the drug docking into the protein; ‘affinity / selectivity’ – a balance of characteristics, or how tightly something binds vs. which proteins it binds to; ‘allosteric’ – binding of a drug molecule to a different site on an enzyme than the normal active site. Regular readers and fans of compu-centric chem blogs such as The Curious Wavefunction and Practical Fragments will feel right at home!

We’ll start at the top. Shape complementarity modeling uses small differences in a binding pocket, such as a methylene spacer in a residue (say, from a Val to Ile swap) to dial-in tighter binding between a target and its decoy. The authors point out that selectivity can often be enhanced by considering a drug that’s literally too big to fit into a related enzymatic cavity. They provide several other examples with a ROCK-1 or MAP kinase flavor, and consider software packages designed to dock drugs into the “biologically active” conformation of the protein.

Electrostatic considerations use polar surface maps, the “reds” and “blues” of a receptor’s electronic distribution, to show how

Affinity Optimization - Black dot represents the optimal minimum energy between Coulombic forces (green) and desolvation penalty (blue) | Credit: Huggins, Sherman, Tidor; J. Med. Chem.

molecular contacts can help binding to overcome the desolvation penalty (the energy cost involved in moving water out and the drug molecule in). An extension of this basic tactic, charge optimization screening, can be used to test whole panels of drugs against dummy receptors to determine how mutations might influence drug binding.

Because target proteins move and shift constantly, protein flexibility, the ability of the protein to adapt to a binding event, is another factor worth considering. The authors point out that many kinases possess a “DFG loop” region that can shift and move to reveal a deeper binding cavity in the kinase, which can help when designing binders (for a collection of several receptors with notoriously shifty binding pockets – sialidase, MMPs, cholinesterase – see p. 534 of Teague’s NRDD review).

But these shifting proteins also swim in a sea of water and other cytoplasmic goodies. This means that drug designers, whether they like it or not, must account for explicit water molecules. The authors even suggest a sort of “on-off” switch for including the bound water molecules, but contend that more efforts should be directed to accurate modeling of water in these protein settings.

Finally, the authors weigh the effects of allosteric binding, the potential for a modeled molecule to be highly selective for a site apart from where the protein binds its native ligand. The authors consider the case of a PTP1B ligand that binds 20Å away from the normal active site, at the previously mentioned “DFG loop.” Since this binding hadn’t been seen for related phosphatases, it could then be used to control selectivity for PTP1B.

In each section, the authors provide examples of modeling studies that led to the design of a molecule. Two target classes recur often throughout the review: HIV protease inhibitors (saquinavir, lopinavir, darunavir) and COX-2 inhibitors (celecoxib), which have all been extensively modeled.

Two higher-level modeling problems are also introduced: the substrate-envelope hypothesis, which deals with rapidly mutating targets, and tailoring molecules to take rides in and out of the cell using influx and efflux pumps in the membrane. Since different cell types overexpress certain receptors, we can use this feature to our advantage. This strategy has been especially successful in the development of several cancer and CNS drugs.

Overall, the review feels quite thorough, though I suspect regular Haystack readers may experience the same learning curve I did when adapting to the field-specific language that permeates each section. Since pictures are worth a thousand words, I found that glancing through the docking graphics that accompany each section helped me gain a crucial foothold into the text.

So what did BMS get for their money?

INX-089, Inhibitex’s lead molecule, has a common antiviral motif: a nucleoside core (the 5-membered ring sugar attached to a nitrogen heterocycle) with an amino acid based prodrug hanging off the left-hand side. Clinically-tested antivirals sharing this basic setup include IDX-184 and NM-283, both from Idenix, and PSI-352938, from Pharmasset  (For an overview of the varied structures currently in development for HCV, see Lisa’s 2010 C&EN story).

INX-089 bears a close resemblance to Pharmasset’s lead nucleotide inhibitor PSI-7977. That’s not a mistake, believes ‘089 discoverer Chris McGuigan, of the Welsh School of Pharmacy. In a recent article (J. Med. Chem. 2010, 53, 4949), McGuigan himself comments “The Pharmasset nucleoside [is] rather parallel to our early work on anti-HIV ProTides.”

Wait, what are ProTides?

Both INX-089 and PSI-7977 aren’t themselves the active viral inhibitor, but phosphoramidate “ProTide” prodrugs: compounds broken down by the body into the active drug (Chem Note: PSI-7977 has single-enantiomer S_p chirality at phosphorus, while INX-189 is a mixture of diastereomers).

Once in the body, enzymes cleave the phosphoramidate group to a phosphate (PO₄^2-). Kinases attach two more phosphate groups, and viruses let this dressed-up molecule inside, where the nucleotide warhead inhibits HCV by interfering with RNA replication (Antimicrob. Agents Chemother. 2011, 55, 1843).

A few comments on the drug itself: The similarity of the ProTide portion (left-hand side) of the molecule to PSI-7977 really is striking: swap in an isobutyl ester and a phenyl, and it’s the same beast! The more interesting switch comes on the upper-right (“eastern”) part of the structure: a protected guanosine ring. This ring harks back to guanine, one of the four common nucleic acids found in DNA.

Source: J. Med. Chem., Pharmasset

PSI-7977, meanwhile, shows off a uracil, a nucleic acid found in RNA, not DNA.

Although it’s tempting to think such similar compounds all dock into the NS5B polymerase at the active site (in the yellow “palm” of the hand-shaped enzyme), don’t be too sure – a recent paper by Pharmasset scientists (J. Med. Chem. 2012, Just Accepted) shows quite a few “Finger,” “Palm,” and “Thumb” sites.  It’s not yet clear whether all nucleoside drugs bind to the active site in the same way. The authors also remark that, due to fast replication and mutation, potentially resistant strains of HCV pop up daily.

1. HCV Takes Off – New treatments for Hepatitis C have really gained momentum. An amazing race has broken out to bring orally available, non-interferon therapies to market. In October, we saw Roche acquire Anadys for setrobuvir, and then watched Pharmasset’s success with PSI-7977 prompt Gilead’s $11 billion November buyout.  And both these deals came hot on the heels of Merck and Vertex each garnering FDA approval for Victrelis and Incivek, respectively, late last spring.

2. Employment Outlook: Mixed – The Haystack brought bad employment tidings a few times in 2011, as Lisa reported. The “patent cliff” faced by blockbuster drugs, combined with relatively sparse pharma pipelines, had companies tightening their belts more than normal. Traffic also increased for Chemjobber Daily Pump Trap updates, which cover current job openings for chemists of all stripes. The highlight, though, might be his Layoff Project.  He collects oral histories from those who’ve lost their jobs over the past few years due to the pervasive recession and (slowly) recovering US economy.. The result is a touching, direct, and sometimes painful collection of stories from scientists trying to reconstruct their careers, enduring salary cuts, moves, and emotional battles just to get back to work.

3. For Cancer, Targeted Therapies – It’s also been quite a year for targeted cancer drugs. A small subset of myeloma patients (those with a rare mutation) gained hope from vemurafenib approval. This molecule, developed initially by Plexxikon and later by Roche / Daiichi Sankyo, represents the first success of fragment-based lead discovery, where a chunk of the core structure is built up into a drug with help from computer screening.From Ariad’s promising  ponatinib P2 data for chronic myeloid leukemia, to Novartis’s Afinitor working in combination with aromasin to combat resistant breast cancer. Lisa became ‘xcited for Xalkori, a protein-driven lung cancer therapeutic from Pfizer. Researchers at Stanford Medical School used GLUT1 inhibitors to starve renal carcinomas of precious glucose, Genentech pushed ahead MEK-P31K inhibitor combinations for resistant tumors, and Incyte’s new drug Jakifi (ruxolitinib), a Janus kinase inhibitor, gave hope to those suffering from the rare blood cancer myelofibrosis.

4. Sirtuins, and “Stuff I Won’t Work With  – Over at In the Pipeline, Derek continued to chase high-profile pharma stories. We wanted to especially mention his Sirtris / GSK coverage (we had touched on this issue in Dec 2010). He kept up with the “sirtuin saga” throughout 2011, from trouble with duplicating life extension in model organisms to the Science wrap-up at years’ end. Derek also left us with a tantalizing tidbit for 2012 – the long-awaited “Things I Won’t Work With” book may finally be coming out!

5. Active Antibacterial Development – In the middle of 2011, several high-profile and deadly bacterial infections (Germany, Colorado, among others) shined a spotlight on those companies developing novel antibacterials. We explored front -line antibiotics for nasty Gram-negative E.coli, saw FDA approval for Optimer’s new drug Fidiclir (fidaxomicin) show promise against C. difficile  and watched Anacor’s boron-based therapeutics advance into clinical testing for acne, and a multi-year BARDA grant awarded to GSK and Anacor to develop antibacterials against bioterrorism microorganisms like Y. pestis.

6. Obesity, Diabetes, and IBS – Drugs for metabolic disorders have been well-represented in Haystack coverage since 2010. Both Carmen and See Arr Oh explored the vagaries of Zafgen’s ZGN-433 structure, as the Contrave failure threatened to sink obesity drug development around the industry. Diabetes drugs tackled some novel mechanisms and moved a lot of therapies forward, such as Pfizer’s SGLT2 inhibitors, and Takeda’s pancreatic GPCR agonist. Ironwood and Forest, meanwhile, scored an NDA for their macrocyclic peptide drug, linaclotide.

7. The Medicine Show: Pharma’s Creativity Conundrum – In this piece from October, after Steve Jobs’ passing, Forbes columnist Matt Herper both eulogizes Jobs and confronts a real ideological break between computer designers and drug developers. His emphasis? In biology and medical fields, “magical thinking” does not always fix situations as it might in computer development.

We hope you’ve enjoyed wading through the dense forest of drug development with Carmen, Aaron, Lisa, and See Arr Oh this past year. We here at The Haystack wish you a prosperous and healthy 2012, and we invite you to come back for more posts in the New Year!

With these numbers in hand, ARIAD enters a tight race, already populated by headliners like Gleevec (imatinib), which in 2001 made a splash as a first-line CML therapy. Drugs such as Gleevec and ponatinib belong to the family of tyrosine kinase (TK) inhibitors, which dock with a mutated protein called Bcr-Abl. This protein (actually a fusion of two distinct proteins via a chromosomal mishap) triggers disease by accelerating blood cell creation, leading to uncontrolled growth and eventually CML.

Since cancers constantly evolve, new mutations in the TK active site had rendered Gleevec ineffective for certain variations of CML. Many of the PACE trial patients had previously tried newer TK inhibitors, such as Sprycel (dasatinib, BMS) and Tasigna (nilotinib, Novartis), and found that their CML had become resistant due to a single amino acid mutation in the kinase active site, which swapped a polar residue (threonine) for a carbon chain (isoleucine). So, ARIAD chemists decided to develop a drug that borrowed the best points from the earlier therapies, but capitalized on this mutation (A pertinent review in Nature Chemical Biology covers early examples of “personalized” cancer drugs developed for disease variants).

So, how did they accomplish this particular act of molecular kung-fu?  For that, we hit up the literature and go all the way back to . . . 2010. As explained in a development round-up (J. Med. Chem., 2010, 53, 4701), most approved Bcr-Abl inhibitors share several traits: densely-packed nitrogen heterocycles linked to a toluyl (methyl-phenyl) amide, then a highly polar end group, such as piperazine or imidazole. Since the mutation axed a threonine residue, the hydrogen-bond donor adjacent to the ring in earlier drugs was no longer necessary. So, chemists replaced it with a vinyl group.

A computer analysis designed to achieve better binding and drug-like properties suggested an alkyne linker might fit into the mutated active site even better than a vinyl group, so that’s ultimately what ARIAD installed. The program also suggested moving an exocyclic amino group into the aromatic (forming an uncommon imiadzo-[1,2-b]-pyridazine, green in picture). Borrowing the best stuff from other therapies, ARIAD’s chemists also wove in the “flipped” amide and -CF₃ motifs (both blue) from nilotinib, as well as the methylpiperazine (red) from imatinib.

With computational rendering (Cancer Cell, 2009, 16, 401) ARIAD scientists could overlay both imatinib and ponatinib in the mutated enzyme’s active site (see picture, right). Notice that unlike imatinib, ponatinib avoids bumping into isoleucine 315. Ponatinib also gets a little extra binding oomph by poking its CF₃ group into a hydrophobic pocket near the bottom of the active site.

After the uproar generated by the announcement, I wondered what, exactly, was this contentious molecule, and what did it do?

In the US, hospitals administer Plan B as two small pills, each with a 750 μg dose of the synthetic hormone levonorgestrel. First approved by the FDA in 1999, levonorgestrel prompted several companies, among them generic manufacturers Barr, Watson, and Teva, to jump in as suppliers in the ensuing decade. According to a 2011 Teva patent, Plan B is most effective when taken within 72 hours of when a person’s first-line contraceptive fails. The FDA estimates its success rate at 80-90%.

Levonorgestrel binds to the same receptors as other sex hormones (think estradiol or progesterone), and prevents ovulation or impairs fertilization of egg cells. Some researchers believe that Plan B prohibits already-fertilized eggs from adhering to the endometrium (uterine inner wall), which might prevent further embryonic development leading to pregnancy. In fact, a large dose of 17-α-ethinylestradiol (EE) – the main ingredient in most birth control pills – can sometimes be used “off-label” to achieve the same effect.

The uncertainty over whether Plan B actually terminates pregnancies brings it onto similar ground with mifepristone (RU-486) and diethylstilbestrol (DES). These two drugs, previously popular options for emergency contraception, have mixed public perception today; many associate RU-486 with abortion, and DES with endocrine disorders and tumor formation in offspring.

Chemistry Note: It’s humbling to watch Mother Nature re-use the same chemical templates over and over, and that small changes in the overall steroid structure lead to huge biochemical consequences. Like Batman, with his never-ending supply of utility-belt gadgets, the steroid core structure can be tweaked in seemingly endless ways to produce biologically active molecules. I would have to devote (several) more posts to just how many modifications, but think about the effects simple oxidation (bile acids), ring expansion (cortistatins), or conjugation (sulfonated sterols) have on biological processes.

The sex hormones have been puzzling synthetic chemists for nearly 100 years; in fact, two prominent chemists spent large portions of their careers perfecting the introduction of a single methyl group into the steroid core! Levonorgestrel claims “second-generation” hormone status; next-gen progestins, such as desogestrel, do away completely with C-3 oxygenation, and sport a new alkene at C-11. These new atomic decorations lead to improved side-effect profiles and lower the overall EE dose in combined pill formulations.

Update (6:05PM, Dec 9, 2011) – Changed “mg” to “μg”  (Thanks, Ed!) 

So what’s the secret sauce behind BG-12? Many pharmaceuticals are small molecules with multiple heteroatoms and aromatic rings, but not BG-12: it’s just dimethyl fumarate! A search for ‘fumarate’ on pubs.acs.org returned >4800 hits, which gives you an idea of its common use in several organic reactions: [3+2] cycloadditions, Diels-Alder reactions, and Michael additions. Interestingly, dimethyl fumarate is the all-E stereoisomer; the Z-configuration, where the two esters are on the same side of the central double bond, goes by the tagline ‘dimethyl maleate’ and does not seem to possess anti-MS effects.

Very small molecules such as BG-12 (molecular weight = 144) are notoriously tough to use as drugs: they hit lots of enzymatic targets, not just the intended ones, and tend to have unpredictable side effects (see Derek Lowe’s 2005 article regarding the FDA “approvability” of several common drugs today). Toss in BG-12’s alkylating behavior to boot (fumarates can interact with nucleophilic amines or sulfides at multiple sites, including enzyme active sites), and you have to wonder how it functions in the body. Well, so do scientists. A 2011 review implicates up to 3 potential biochemical mechanisms – the Nrf2 pathway Lisa mentioned in the 2009 piece, T-helper phenotype 2 interleukin upregulation (IL-4, IL-10, IL-5, which “change gears” for immune system functioning), and CD62E inhibition, which controls adhesion of blood cells to inflammation sites.

Side notes: Flavoring chemists have added fumaric acid, the parent diacid of BG-12, to industrially-prepared foodstuffs such as baking powder and fruit juices since the 1930s. A darker side of dimethyl fumarate emerges when you consider its non-medicinal use: certain furniture companies applied it to new upholstered chairs and sofas to stop mold growth. This unfortunately caused several cases of severe skin irritation, which a 2008 exposé in London’s Daily Mail likened to actual burns.

The data seems to have convinced Roche, which acquired Anadys last Monday in all-cash deal analysts say represented a 260% premium over Anadys’s Friday stock closing price. Roche, no stranger to the HCV battle, hopes to integrate setrobuvir into a potential oral drug cocktail with its current suite of polymerase and protease inhibitors.

Setrobuvir interacts with N5SB polymerase at the allosteric “palm” binding site, located in the center of the baseball-mitt shaped enzyme. The drug’s sulfur-nitrogen heterocycle – a benzothiadiazine – is the key to virus inhibition; Anadys has installed the motif in all their HCV inhibitors, going back to their 2005 patents.

Chemists have known about the virus-targeting properties of this heterocycle for a while, but most derivatives have been culled in pre-clinical testing (see J. Antimicrob. Chemoth. 2004, 54, 14-16 for a brief review). Interestingly, chemists initially prepared benzodiathiazines, such as those in Merck’s chlorothiazide (c. 1957, according to the Merck Index), as diuretics, which found use in diabetic treatment. Over the next 40 years, modified medicines treated conditions ranging from epilepsy and cognitive therapy to hypertension and transcriptase regulation. Tweaked benzodiathiazines first showed anti-HIV and anti-CMV activity in the mid-1990s.

One final advantageous wrinkle in this structure: unlike PSI-7977, setrobuvir is not nucleoside-derived. This feature changes its binding behavior, pharmacokinetics, and even its intellectual property strategies, since many current antiviral therapies mimic the nucleosides found in RNA and DNA chains.