Kinetics Layout 1 23/06/2011 13:26 Page 39

Kinetics_Layout 1 23/06/2011 13:26 Page 39

Drug Discovery

The need for high throughput kinetics early in the drug discovery process

Generally kinetics are only investigated late in a programme by which time many compounds, slightly less potent but with a much better kinetic profile, have already been discarded. This paper argues that having kinetic information at the early stages of drug discovery can bring numerous benefits to the process.

he majority of drug discovery projects rely is not possible to predict these from the KD as Dr Wilma Keighley upon estimates of compound affinity to a shown in Table 1. Ttarget protein to guide lead seeking medic- So, are on and off-rates critical information or inal chemistry in early stages when high compound merely nice to know? The overwhelming answer is numbers are being screened. At this stage it is now that they are critical. Andersson et al1 recognise quite common to test at IC50 level not only against kinetic binding data as a decisive element in drug the primary target but also against several selectiv- discovery, while the importance of drug-target res- ity and toxicity targets, so choosing compounds for idence time (as quantified by the dissociative half progression against multiple criteria. Nevertheless life of the drug-target binary complex) is empha- any efforts to measure the on and off-rates that sised for its potential impact on duration of effect underlie the affinity (to either the primary target or and target selectivity by Copeland et al2 and selectivity target) are left to much later stages of the Zhang and Monsma3. Hopkins (www.hop- campaign when fewer compounds are tested. This kinslab.org) puts this firmly into the perspective of is simply due to the high cost and low throughput the patient ‘understanding the binding kinetics can of the specialist biosensor instruments typically help shape the clinical profile of a drug that are used to determine binding kinetics. important to patients: efficacy, safety, duration of However, evidence is plentiful that compounds action, greater tolerability, indication and thera- with the same affinity but different on and off-rates peutic differentiation’. The following examples can have a very different biological activity profile. show why this is so. The affinity-driven triage strategy currently used therefore results in superior compounds being Slow off-rates rejected early in the discovery process. Improved Increased residence time gives methods and technologies enabling cost-effective, better duration of action convenient and high throughput access to kinetic A survey of the literature identifies a number of data would create opportunities for a much drugs, across a variety of target families, having improved early drug discovery paradigm. The exceptionally long duration of action, which can attributes of an ideal technology are discussed. be ascribed to very slow off-rates when compared Affinity (KD) for a target protein, derived from to other drugs of the same class. Examples include 4 5 6 IC50, is used regularly to rank compound perform- Aprepitant , Amlodipine , Maraviroc , 7 8 9 ance. KD is dependent on both the association (on) Candesartan , Saquinavir and Tiotropium . and dissociation (off) rates of the compound and it In all of these cases, affinity would have been no

Drug Discovery World Summer 2011 39 Kinetics_Layout 1 23/06/2011 13:26 Page 40

Drug Discovery

Table 1: For the same affinity, on and off-rates can vary enormously and cannot be predicted apparent for this drug over other mixed M1/M2/ (illustrated here using mock data) M3 antagonists. Changes in off-rate leading to large differences AFFINITY K-ON M-1 K-OFF MIN-1 SPEED in duration of action are most obvious where a -1 (KD) MIN compound series is subject to rapid clearance. Slowly dissociating antagonists, where the half life 1 x 10-9 105 0.0001 (6900 min) Exceptionally slow of the antagonist-receptor complex exceeds that of on/off the free antagonist are likely to maintain a longer receptor protection in vivo than fast dissociating 1 x 10-9 106 0.001 (690 min) Slow on/off antagonists. Where the half life of the free antagonist is longer than the half life of the complex, 1 x 10-9 107 0.01 (69 min) Fast on/slow off longer occupation by the slowly dissociating antagonist is only relevant if the receptor is subject to large fluctuations in free agonist concentration1. 1 x 10-9 108 0.1 (6.9 min) Very fast on/fast off Improved selectivity due to 1 x 10-9 109 1 (0.69 min) Exceptionally fast differential off-rates on/off Since many compounds bind to a number of proteins in addition to their intended target, improved selectivity for the chosen target can be a strong differentiator across a class of drugs. By monitoring guide to efficacy duration and it is a general find- not only the dissociation rate of the compound ing that drugs with slow-off-rates are found by from its intended target but also from other poten- serendipity. We can see the impact of slow-off tial selectivity or ADME targets, it should be pos- clearly in Table 2 by comparing data from some sible to identify compounds with preferred charac- M3 antagonists. teristics where a faster off-rate from the unintend- In the case of the Tiotropium, the molecule ed targets is observed in the presence of slow off- exhibits very long lasting effects with very little rate for the intended target. Table 3 gives some variation in bronchiodilatation between peak and examples of these findings. trough (the time between administration and next administration) which rendered it the first truly Fast on rates once-a-day bronchodilator10. Plasma levels of Improved in vivo efficacy where Tiotropium at trough are in the low pg/ml range agonist concentrations are high and hence very unlikely to explain the sustained Despite a 50,000-fold difference in in vitro Ki val- effectiveness of the drug, which is instead ascribed ues for two thrombin inhibitors, hirudin and mala- to its slow dissociation from the M3 receptors in gatran, comparable plasma concentrations resulted the lung9. In addition to long duration of action, in comparable antithrombotic effects in vivo. In the ability to retain effectiveness at lower blood contrast, comparable in vitro Ki values and in vivo levels endows a large therapeutic window. EC50 scores were obtained when malagatran was Additionally, since Tiotropium has a faster dissoci- compared with inogatran. These differing profiles ation rate from M1 and M2 muscarinic receptors have been explained by the concentration of than from M3 receptors, an improved selectivity is thrombin in the thrombus and the rate of attaining an effective inhibitory concentration of antagonist drug. For the inhibitors tested (eight in total) there Table 2: Muscarinic M3 antagonists with similar affinities but very different off-rates9,10 was an inverse relationship between k-on values in

KD/KI ON RATE OFF-RATE vitro and the slope of the dose response curves in X109M-1 MIN-1 MIN-1 vivo with inhibitors with k-on values of <1x107 M- 1 s-1 having steep slopes with Hill coefficients >1 Atropium 0.2 nM 1.5 0.27 (2 mins) and better therapeutic index due to rapidly reach- ing critical concentration for efficacy. Contrarily, Ipatropium 0.2 nM 0.5 0.07 (10 mins) the association time for inhibition of thrombin by Clidinium 0.3 nM 10 0.02 (30 mins) slow binding inhibitors is too long to give effective antithrombotic effects at low plasma concentra- Tiotropium 8 pM 0.16 0.0015 (34.7 hrs) tions. Where higher plasma concentrations are achieved, the association time decreases resulting

40 Drug Discovery World Summer 2011 Kinetics_Layout 1 23/06/2011 13:26 Page 41

Drug Discovery

in a steeper dose response curve but at the price of laboratory species. If on/off-rates vary in human a narrowed therapeutic interval11. In examples versus animal receptors, an expected margin of such as this, because the need to achieve critical safety observed in one species could disappear due concentration in a short time is the crucial factor in to different kinetic profiles in the human at the rel- determining efficacy, on-rate is a much better pre- evant receptors – or vice versa: a compound reject- dictor of efficacy than Ki. ed due to poor margin of safety or selectivity in an animal species could be, in fact, adequately selec- How could we apply on and off-rate tive for the desired over the unwanted effects in properties to enable better drug human. Such variations would be particularly discovery? important where efficacy and safety studies are In vitro screening conducted in different species. Understanding of kinetics at early stages of Drug Is a fast Koff of use? It is possible to envisage that Discovery, even at the screening phase, ie before iv dosage forms of a drug with fast off-rate would many compounds have been rejected, gives access enable a fast wash out of drug – short acting anaes- to more diverse chemical space, more scope for thetic perhaps, or other use where tight control on intellectual property (IP) and patents, and better the duration of effect would be valuable. defined biology. Receptor selectivity can be driven by off-rate and not necessarily affinity and affinity Manipulation of kinetic profiles and off-rate SAR may diverge with large differ- An understanding of kinetics, at an early stage in a ences in dissociation rates within a chemical series lead seeking programme, could give real assistance being not uncommon. In addition to kinetics at the by knowing in which areas to focus where a com- primary target, this principle can apply equally to pound falls short of required effectiveness. Is there off target effects in respect of selectivity and safety such a thing as a ‘good kinetic profile’? Geitmann12 issues. Where SAR diverges it may be possible to describes a study of HIV-1 protease inhibitors achieve a greatly improved safety profile by select- where structural manipulation was able to produce ing compounds from the series based on their a range of on and off-rates such that the compound on/off-rates at proteins of concern. properties could be used to map inhibitors into clusters that related their properties to their poten- In vivo screening tial utility as drugs. Note here the different proper- Generally, better in vivo profiles are apparent with ties of association and dissociation rates. slow off compounds, less variation in receptor Association rates are related to the structure of the occupancy between peak and trough plasma con- compound and to the dosage, while dissociation centrations leading to a smoother efficacy curve. rates are solely dependent on the molecular struc- Slow off-rates also are more likely to lead to a ture. However, receptor kinetics are of even more once-a-day dose projection, which can be an use if we view them in conjunction with the phar- important differentiator. macokinetic properties of the compounds. Where in Understanding on and off-rates may help us rec- vivo performance is sub-optimal, by considering all oncile currently puzzling mismatches in efficacy, of this data we would understand the reason for the selectivity or safety between humans and animal poor efficacy and how this might be improved.

Table 3: Selectivity conferred by differing kinetics at receptor subtypes

COMPOUND RECEPTORS DATA LITERATURE REFERENCE

Tiotropium M3 versus M2, M1 M3 off-rate much slower Disse, 19999 muscarinic receptor than M2 or M1 off-rate

NPY/PYY NPY1 versus NPY2 Y1 off-rate = 1 hour Dautzenberg, 200518 Y2> 24 hours

ICI 137,798 Beta 1 versus Beta 2 Beta 2 off-rate 30 min Keith, 198919 adrenoceptor Beta 1 > 6 hours

SCH 527123 CXCR1 versus CXCR2 CXCR1 off-rate 10 min Gonsiorek, 200720 CXCR2 20 hours

Drug Discovery World Summer 2011 41 Kinetics_Layout 1 23/06/2011 13:26 Page 42

Drug Discovery

Table 4: What does a kinetic profile deliver?

PROFILE HOW DOES THE PROS CONS COMPOUND BEHAVE?

Fast on/Fast off Rapidly reaches effective Low dose required for Duration is determined concentration at effect hence good by clearance rate of free receptor, rapidly therapeutic index. Could drug hence frequent dissociates be preferable if dosing often required bioavailability is poor

Slow on/Slow off Takes time to build to Long duration allows Slow on-rate, if plasma effective concentration at once daily dosing, stable concentration is low, may receptor, slow effects between peak and mean that drug never dissociation once bound trough; Long duration achieves sufficient even if free drug is receptor occupancy for subject to rapid clearance efficacy

An ability to modulate on and off-rates would be So, since having kinetic information would be particularly valuable where poor pharmacokinetics really helpful to a discovery programme, why is it are observed. Where bioavailability is low, a high that kinetics based screening doesn’t already hap- on rate is essential to get the drug on to the bind- pen on a regular basis? ing site. Improving affinity by decreasing the off- A number of technologies are available to gener- rate will not help since dissociation rate is only ate kinetic information, and can be broken down important if there is a drug-target complex in the into two broad types: those utilising a label such as first place (in most cases, the drug will have to a radiochemical or fluorophor, and label free tech- achieve some minimum receptor occupancy level nologies which include surface plasmon resonance, to achieve efficacy). Contrarily, where the effec- biolayer interferometry and resonant acoustic pro- tiveness of a drug series is compromised by rapid filing (for a more comprehensive review of available A clearance, a compound with a very slow off-rate technologies see Comley13) Nonetheless, the fact can still be efficacious, by maintaining receptor that most laboratories use non kinetic measures to occupancy longer than the half life of the free drug make an early selection decisions and make full would suggest. Decreasing the off-rate will be kinetic measurements only on compounds that sur- much more important than increasing the affinity vive this pre selection process14 is ample evidence (eg by increasing the on-rate) in this circumstance. that existing methodology does not currently Slow functional reversibility is associated with long enable a wider use of kinetic data. Some more R lasting in vivo efficacy, whereas the efficacy of accessible technologies, eg localised surface plas- compounds with rapid reversibility is determined mon resonance (LSPR) currently used to obtain by their pharmacokinetics. affinities have, in the principle, the capability of d To summarise, there is no ‘best ‘ kinetic profile; generating kinetic data but this is not yet developed. it depends on the objectives of the discovery pro- Where laboratories generate kinetic data cur- p gramme, as illustrated in Table 4. Clearly if the rently, it is likely that they will be using Biacore or kinetic profile of compounds are not known then a similar sensor-based system in their work, the ability to select compounds on a rational basis Biacore having by far the largest market share is lost. (29% in 2008 according to Comley, now likely to approach 50%). Until recently, Biacore enjoyed Why are kinetics not investigated complete dominance of the kinetics-based label- earlier in drug discovery? free market due the virtual absence of credible Because kinetics generally are investigated only late alternatives and the Biacore name became synony- in a programme after a lot of decisions have been mous with both label-free and SPR. Yet despite this made and triage carried out, the fact is that many market dominance, and clear market interest in compounds slightly less potent than those selected label-free methods, most label-free screening has for progression, but with a much better kinetic remained a niche and specialist approach. Label- profile, may already have been discarded. free competitors to Biacore came to market over

42 Drug Discovery World Summer 2011 Kinetics_Layout 1 23/06/2011 13:26 Page 43

Drug Discovery

the past several years and most are concentrating and is the key limitation – the cost per data point on an instrument-based approach – there are in an average automated system is estimated at 15- already many technical reviews covering new 30 times higher for SPR (excluding instrument instruments (see, for example, Rich & Myszka15) cost) than traditional in vitro methods. High capi- which can be subdivided into Biacore-like tal cost is an additional hurdle which inhibits many Instrumentation and those with a substantively dif- laboratories from entering this field and may have ferent approach. Corning, BioRad, ForteBio and contributed to the withdrawal of the TTPLabtech all have commercial offerings which FujifilmPharma offering from the SPR market. require instrumentation that is expensive and ded- Some companies, eg SensiQ, have chosen to reduce icated to the purpose. Despite using different tech- the level of automation in their equipment so as to nologies, all the systems aim to address similar reduce the cost to own, at the expense of through- application areas as Biacore. Some offer higher put or convenience, although innovations in the throughput and improved ability to screen frag- way the equipment is used may mitigate this16. ments whereas others provide benefits such as the ForteBio has adopted an alternative approach to use of crude samples. However, as a consequence the Biacore-like implementation with the Octet of high instrument cost, the need for expert users family of instruments that dip optical microfibres (see Rich & Myszka15 for the pitfalls that can directly into the wells of standard plates rather catch the unwary and even expert users) and limit- than requiring complex microfluidics. With similar ed throughput, none of these dedicated instrument- flexibility to Biacore and a reasonable throughput, based approaches has resulted in a high uptake of the Octet offers a real alternative at a reduced (but use of kinetic analyses early in the drug discovery still substantial) capital outlay. As a further inno- process. Currently the ratio of cost to throughput vation, ForteBio has attempted to broaden the compares unfavourably with other technologies scope of use of its equipment so that it has utility

Address Even the Most Diĸ cult Assays

TM Label-free screening for: RapidFire enables label-free • oncology • epigeneƟ cs detecƟ on of naƟ ve analytes for • metabolic disorders • cardiovascular disease previously intractable targets • anƟ -infecƟ ves • neurological disorders • inﬂ ammaƟ on

T: 781.928.2700 | F: 781.998.0054 | www.BIOCIUS.com | [email protected]

Drug Discovery World Summer 2011 43 Kinetics_Layout 1 23/06/2011 13:26 Page 44

Drug Discovery

beyond kinetic measurements, thus mitigating the ands can be conducted by a number of specialist cost to own with increased versatility. providers. Where an iodinated product is pre- Looking back to label based-methods, the ferred, the relatively short half life limits users to Motulsky Mahan method has been useful histori- local suppliers, hence in Europe PerkinElmer (cov- cally utilising competition between cold and radio- ering the range of materials previously supplied by labelled compounds (although it would be equally NEN) and Quotient Bioresearch Ltd (which has applicable to fluorescence labels) to directly meas- acquired the Amersham business from GE) would ure on and off-rates of compounds. This method likely be first choice. Where a tritiated compound has been especially applied to GPCR receptors. would be satisfactory, US-based suppliers such as With Biacore-like methods, the protein of interest RC TriTec, ViTrax, Tjaden and Moravek could be must be immobilised on the sensor chip and this considered. has thus far proved difficult with GPCR receptors As a final comparator, a newer label-free tech- (previously achieved only for the more accessible nique, localised SPR (LSPR), can be considered case of the oestrogen receptor, strides are being alongside these more established methods. LSPR made in screening of membrane bound GPCRs shares the advantages of the label-based methods using SPR17 but this still remains a challenging and the ForteBio Octet in being microtitre plate area). This methodology, although laborious, finds based – a big advantage for a laboratory conduct- favour in some laboratories due to familiarity with ing a range of assay types, but unlike ForteBio and the techniques employed and no requirement for all other biosensor instruments, requires no spe- additional specialist knowledge or equipment. cialised equipment (measurement is via a standard Throughputs achieved can be substantial, limited absorbance reader). To date LSPR has been used to only by the stamina of the scientist. A majority of measure affinities, but since the method can follow laboratories using this methodology will rely upon the association of compound with protein it radiochemical labelling due to the difficulties in should, in principle, be possible to derive the off- achieving fluorescence labelling that does not rate from the measured affinity and on-rate. While induce steric hindrance to the interaction being there is no evidence yet of proof of principle, such measured. Custom production of radiolabelled lig- an approach would be very attractive particularly

Table 5: Key features of various kinetic measurement technologies

TECHNOLOGY VENDOR CAPITAL COSTS CONSUMABLE FORMAT THROUGHPUT SKILLS BASE OTHER COSTS PER DATA REQUIRED DATA POINT** POINTS/DAY

Surface Biacore, $0.5-1 Million $1.5-$3 (mainly Flow cell based Theoretical Specialist skill Plasmon SensiQ, Fujifilm (varies with the cost of the maximum 4,800 for data Resonance Pharma and extent of sensor) In practice, less interpretation others autom’n) than 200 per and for sensor day regeneration

Localised Pharma- Nil*1 $0.2-$0.75 Microtitre plate Any competent Kinetic data Surface diagnostics (throughput based screening possible in Plasmon dependent) scientist theory – Resonance practical proof currently absent

Biolayer ForteBio $400-600K $1.5-$3 Microtitre plate Specialist skills interferometry based for data interpretation

Radiochemical Configure in Nil*2 Medium Microtitre plate >5,000 Any competent Time (Motulsky house based screening consuming and Mahan scientist laborious. method) Radioactivity usage

*assumes that the laboratory has access to general laboratory equipment. 1 absorbance reader, 2 scintillation counter. ** author’s estimates

44 Drug Discovery World Summer 2011 Kinetics_Layout 1 23/06/2011 13:26 Page 45

Drug Discovery

to first time users and to third world laboratories 2009, Wilma was Senior Director for New References due to the absence of requirement for major capi- Technologies at Pfizer’s European research head- 1 Andersson, K et al (2006). tal investment. This same advantage would allow a quarters leading a team responsible for identifying, Expert Opin. Drug Disc., 1, 439-446. much broader application of kinetic measurements assessing and implementing new methods to 2 Copeland, RA et al (2006). in laboratories currently hindered by throughput broaden scope and improve screening efforts. For Nature Rev.Drug Disc., 5, 730- limitations. this she built on her previous experience forming 739. Thus, issues with current methods and technolo- and leading a compound profiling department 3 Zhang, R and Monsma, F gies include the requirement for major capital tasked with supporting all projects at the site with (2009). Curr. Opin. Drug Disc. Dev., 12, 488-496 investment, consumables costs well in excess of in vitro potency and selectivity data. Most impor- 4 Lindström, E et al (2007). those for ‘traditional’ screening technologies, rela- tantly, her team pioneered the use of frozen cells to JPET., 322, 1286-1293. tively low throughput, lack of flexibility, specialist support primary screening and innovative combi- 5 Rigby, JW et al (1988). skills required to run equipment and interpret the nations of automation and manual effort to max- J.Cardiovasc.Pharmacol., 12, data output, lack of body of evidence for fidelity imise flexibility in compound management and Suppll.6, S144. 6 Napier, C et al (2005). (particularly true of the newer techniques). No one screening. With experience in Lean and 6-sigma Biochem. Pharmacol., 31, 163- technology is free or guilty of all of these issues, as methodologies, Wilma has been responsible for 72. summarised in Table 5. improving cost-effectiveness in early drug discov- 7 Gradman, AH (2002). J. ery across a number of assignments including assay Human Hypertension 16, S9- To enable a much wider use of kinetics in early development, compound profiling, high through- S16. 8 Markgren, P-O et al (2002). project decision making what would an ideal put screening and compound management. J.Med.Chem., 45, 5430-5439. technology look like? www.wilmakeighleyconsulting.co.uk 9 Disse, B et al (1999) Life So here’s the challenge to innovators and vendors. Sci., 64, 457-464. We want to be able to generate kinetic data on the 10 Noorda, JA van et al confirmed hits from HTS (so for a HTS of 1 mil- (2000). Thorax 55, 289-294. 11 Elg, M et al (1997). lion entities that is 1,000-10,000 compounds Thromb.Haemost., 78, 1286- depending on selected hit threshold – typically 0.1- 1292. 1%), and on all compounds synthesised in 12 Geitmann, M et al (2006). response to the HTS hits for all projects (ie lead J.Med.Chem.,49, 2367-2374. seeking phase of a project), and on all members of 13 Comley, J. (2008). Drug Discovery World Fall edition, a fragment library. Where we have a project that 28-49. has a clear need for a definite kinetic profile (eg 14 Boettcher, A et al (2010). .J slow off for a long lasting bronchodilator) we want Biomol. Screen. 15, 1029-41. to conduct screening of large subsets or even of the 15 Rich, L and Myszka, DG whole file in kinetic format. Therefore, we would (2010). J. Mol. Recognit. 23, 1- 64. like a simple (any bio-scientist can do it), cost- 16 Rich, RL et al (2010). Anal. effective (no more expensive than conventional Biochem., 407, 270-277. screening technologies, say, $0.1 per data point) 17 Navratilova, I et al (2011). and reasonably high throughput (>10,000 data ACS Med. Chem. Lett., Article points per day – remember that each compound ASAP DOI:10.1021/ m/2000017. may require multiple data points!), assay technolo- 18 Dautzenberg, FM and gy to help us to achieve this. And we would like to Naysari, S (2005). Pharmacol., be able to attempt this without recourse to special- 75, 21-29. ist instrumentation requiring high capital outlay, 19 Keith, RA et al (1989). and to be compatible with current compound stor- JPET., 248, 240-248. 20 Gonsiorek, W et al (2007). age formats. Is that too much to ask? DDW JPET., 322, 477-485.

Other references Dr Wilma Keighley is an independent consultant Abbas, A et al (2011). Biosens. working with biotech and academia to help design, Bioelectron., 26, 1815-1824. develop and position their output for use within Navratilova, I and Hopkins AL (2010). ACS Med. Chem. Letts big Pharma. She has a strong record of success in 1, 44-48. helping gain financial support for research projects from UK and European funding bodies, a typical example of which would be BBSRC follow on funding which emphasises the commercial potential of research. Before her independent work in

Drug Discovery World Summer 2011 45