Commentary

Pharmacology in cerebrovascular disease research: Pharmacokinetics–pharmacodynamics to the rescue

Journal of Cerebral Blood Flow & Metabolism 2016, Vol. 36(3) 639–644 ! Author(s) 2015 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/0271678X15619429 jcbfm.sagepub.com

Giora Z Feuerstein

Abstract Pharmacological tools (compounds) have an important role in validation of biological molecular targets for their role in disease processes and their prospective development for therapeutic objectives. Effective utilization of such pharmacological tools impacts on the veracity of the information by which decisions regarding target validation is reached, and investment in clinical development is committed. This commentary addresses frequent gaps in effective utilization of pharmacological principles and practices of pharmacokinetics and pharmacodynamics in experimental pre-clinical research, which if more rigorously implemented could contribute to eventual success in drug development for stroke and neurotrauma.

Keywords Translational research, drug discovery, pharmacokinetics, pharmacodynamics, stroke, brain ischemia, drug trials Received 22 July 2015; Revised 15 October 2015; Accepted 3 November 2015

Introduction Basic research in biology and medical sciences is a high priority in many developed societies where significant portion of the ‘gross domestic product’ is devoted to healthcare. Discovery and development of new drugs are cardinal to the pursuit of improved healthcare; indeed, medicines are a significant part of health costs and pharmaceutical R&D budgets in 2014 totalled over $50 billion. Yet, realizing innovative, highly effective and safe therapeutics in general, and in stroke and neurotrauma in particular, continue to be disappointing. While recognizing the particular challenges in conducting clinical studies in these latter diseases, less attention has been given to scientific conducts throughout the discovery research (pre-clinical), which lay the foundation for new therapeutic. This commentary focuses on a specific element, pharmacology, and its derivatives – pharmacokinetics (PD) (the way compounds are transformed in the body) and pharmacodynamics (PK) (the gamut of responses to the administered compound) as possible contributing factors to successful translational studies. The reasons for insufficient attention to the

intricacies of pharmacological principles in translational research1 is beyond the objectives of this commentary yet, if properly conducted, better prospects in clinical translational medicine could be envisioned. This commentary highlights strategic and tactical pursuits of pharmacological studies in pre-clinical translational research.

PK/PD discipline in experimental cerebrovascular disease research Pharmacological small organic molecules and biologicals are used in biology and medical research for two main purposes. First, as part of ‘target validation’ where manipulation of biological targets by pharmacological means is used to discern discrete molecular

FARMACON LLC, Translational Medicine Consulting, PA, USA Corresponding author: Giora Z Feuerstein, FARMACON LLC, Translational Medicine Consulting, 540 Hoffman, Dr Bryn Mawr, PA 19010, USA. Email: [email protected]

640

Journal of Cerebral Blood Flow & Metabolism 36(3)

pathways that execute biological functions prevalent in academic research. In this perspective, pharmacology is ancillary to genetic and molecular biology techniques applied for similar objectives. Second, more prevalent in the pharmaceutical industry, pharmacology is part of compound development and validation as candidate for drug development. Whether the former or latter objectives are in mind, the conduct for properly understanding drug–target relationships remain the same. Either purpose commands discrete understanding of the physical–chemical nature of drug interactions with the biological target (Table 1), which serve for rationalization of PK–PD studies. The specific techniques to detail these attributes are available in standard texts in the art.1,2 These attributes, briefly summarized in Table 1, are largely founded on dose/concentration–response studies. As such studies report probabilistic data, they require proper statistical analyses and algorithms. Such are also PK/PD data, which command attention to ‘power standards’, confidence intervals (CI) and statistics-based modeling.1,2 Figure 1 highlights the key attributes of PK information, which governs all PD outcomes – efficacy, safety and tolerability. The simplified formula presented in Figure 1(a) calls attention to three elements: (1) Cmax (maximal concentration of drug); (2) ‘AreaUnder the PK Curve’ (AUC), which is expected to secure sufficient target engagement by the compound throughout the dosing period and (3) Time (T) period the target is sufficiently engaged (as calculated from parameters listed in Table 1, e.g. Kd) to generate the

needed PD duration for the expected biological impact and medical benefit. This formula is agnostic to routes of drug deliveries, yet its delineation varies according to the accessible body fluids and tissues. For CNS drugs, cerebrospinal fluid (CSF) and brain drug levels (e.g. via PET/SPECT) should be obtained in both pre-clinical and clinical development. Figure 1(b) exemplifies a PK curve annotated for the key elements described in Figure 1(a). The AUC must include CI (vertical lines in green) securing exposure that accommodates a high percentage of study subjects (humans or animals) over the dosing period. Furthermore, the extent by which the AUC exceeds the Kd should accommodate targets that require high engagement (e.g. 95% inhibition) to generate optimal efficacy. Temporal prospect of compound-target engagement that sustains the needed AUC/Kd throughout the dosing period (T/Kd) is also an essential variable. This latter parameter might carry significant weight in determining biotherapeutic responses.3 Failure to sustain the needed AUC/Kd and/or T/Kd and/or CI (confidence interval), could well be a reason for ‘false negative’ results in clinical development. Furthermore, excess AUC/maximal tolerated dose (MTD) (Figure 1(b)) could also contribute to drug development failure due to adverse events. Figure 1(c) illustrates optimization of PK distribution over the dosing period that does not require change of dose. All essential parameters (Cmax, AUC/Kd and T/Kd) are optimized simply via modeling of distribution of exposure within the Kd and MTD lines. The dash/dot line (in black, Figure 1(b) and (c))

Table 1. Attributes of antagonists compound–target interactions.1 Attribute

Advantage

Disadvantage

Potency

Slow off rate; PK advantage; lesser adverse effects, lower ‘cost of goods’ Safety, tolerability Target validation High extent of blockade; Safety Limited excess pharmacology

In extreme-formulation challenges

Specificity (selectivity) Mode: Competitive Mode: Insurmountable

Mode: Non-competitive Target location: Circulation confined

Target location: Intracellular

based adverse effects Prolonged pharmacology Limited dosing Easy access of compound to target Low volume of distribution: safety Larger volume of distribution

PK: pharmacokinetics; BBB: blood–brain barrier; CNS: central nervous system

Limited broad spectrum potential May require prolonged PK Potential limited efficacy

Extended pharmacology beyond needed Low CNS entry across BBB

Barrier for compound access to target impacts efficacy; efflux

Feuerstein

641

Figure 1. Pharmacokinetic determinants to optimize precision dosing regimens. (a) Description of pharmacokinetic determinants of compounds that bear directly on the pharmacodynamics consequences. f: free D concentration; T: Time; Cmax/min: points of maximal/minimal drug concentration; AUC: Area Under the PK Curve; Kd, Ki, IC50: standard basic parameters of drug interaction with the target; x1-n: multiples of the Kd or Ki or IC50 needed to secure sufficient drug-target engagement. Yellow dashed line denotes the Kd (Ki, IC50) parameter of the drug, while dash/dot lines the maximal tolerated drug exposure. (b) PK Example of single IV bolus drug dose. Part B points to two parameters (ratios) that are important considerations to secure optimized drug effect: AUC/ Kd and Time (T)/Kd. The example illustrates sufficient target engagement throughout 2 h (only 8.3% of a daily dosing period) suggesting need for multiple IV injections or continuous infusion. Part B also denotes (green vertical lines) the importance of determination of the confidence intervals of the PK curves; for a median point at the Kd line suggests that half of the population studied will be below targeted exposure and will require adjustment. (c) Example of exposure optimization of same drug that does not require change of drug dose yet secures 20 h of AUC/Kd (83%) of the dosing period. This optimization scheme also illustrates potential safety advantage in cases where the drug has the propensity for adverse effect at the Cmax period, which is mitigated when drug is delivered as per scheme in part C.

Step 1

Step 2

Literature searches

Pilot

Step 3 MAD MODELING

SAD

Step 4 Efficacy pilot

Pilot MAD Selecon of Animal mode & Study design

Analysis Definive

SAD MTD/NOAEL Tolerability

Analysis

Definive MAD

Definive Efficacy

MTD/NOAEL

PK ~ fCmax + fAUC/Kd + T/Kd Figure 2. Orderly process to assess PK coordinates within safety/tolerability boundaries toward efficacy studies. The objective of this figure is to provide ‘gold standard of operation’ of PK studies in animal models (in preclinical phase) prior to performing efficacy studies. ‘Pilot study’ refers to small scale (N¼3–4) study aimed to quickly ‘scope out’ ranges of potential suitable exposures. ‘Double arrow lines’ (red) denote potential re-iterative process. The process depicted illustrates an orderly process whereby PK parameters are linked to the tolerability of the compound in the targeted species intended for efficacy studies. In this manner, the PK selected for the efficacy (POC) studies is likely to be free from confounding issues such as adverse effects that could impact on efficacy variables. SAD: single ascending dose; MAD: multiple ascending doses; MTD: maximum tolerated dose; NOAEL: no adverse effect level; POC: proof of concept study.

642 that denotes the MTD indicates risk at the Cmax zone (AUC over black dash/dot black line, Figure 1(b)) is effectively eliminated simply by proper utilization of mathematical models.4,5 In summary, while Figure 1(b) offers several causes of ‘false negative’ results by the way a certain PK/PD study is done while Figure 1(c) provides remedies to all three possibilities and therefore optimizes drug development success.

Tactical approach in performing PK studies in view of commencing efficacy studies in pre-clinical phases Successful PK studies require proper consideration of the compound tolerability within the desired PK needed to secure efficacy. Figure 2 provides ‘step-wise rules of engagement’ towards PK studies. Following Step 1, selection of the proper species and experimental design, pre-clinical PK studies commence (step 2) via ‘pilot guesstimate’ of potential doses to be delivered systemically as single ascending dose (SAD). The brief and limited ‘Pilot’ SAD allows for a more precise, calculated SAD study that aims to define the MTD and calculation of a ‘no observed adverse events level’ (NOAEL). Unless the drug is intended for single dose only, this phase proceeds to define the PK in ‘multiple ascending dosing’ (MAD) regimens (step 3) towards defined MTD and NOAEL. In both steps (2 and 3), an opportunity presents itself to observe the animals for potential adverse effects. Simple monitoring scales of neuro-behavioral variables such as sedation, convulsions, tremor, gastro-intestinal and respiratory functions could help define the dosing regimen free of adverse effect while delineating the efficacy.6 If the efficacy species differs from those used in standard preclinical safety studies, consideration must be given to implications of such variances.

Convergence of PK–PD information for drug candidate selection Figure 3(a) and (b) presents schemes that detail asymmetry in possible PK/PD efficacy translation. In both parts of Figure 3, the upper matrix delineates 3D relationships that optimize efficacy studies: drug exposure, extent of biological target occupancy and functional modulation and the ‘medical’ consequences (animal models, e.g. stroke scale, infarct size). Underneath, a 2D matrix divided into quadrants (A–D) segregates compound efficacy across target occupancy, specificity and safety/tolerability boundaries. Quadrant A qualifies for specificity and safety of compound within target occupancy; quadrant B denotes efficacy generated by compound of poor selectivity; quadrant C denotes drug efficacy beyond target occupancy that is associated with adverse effect

Journal of Cerebral Blood Flow & Metabolism 36(3) (‘off target’) and quadrant D denotes efficacy beyond target occupancy without apparent adverse effects. Figure 3(a) displays a dose–response emanating from quadrant A that generates optimal PD within target occupancy and tolerability. The display in quadrant A is a desired PK/PD, but also suggests that some specificity is invariably lost when high occupancy requires very high exposure. Figure 3(b) displays three lines that similarly reach robust PD responses yet in three different contexts. Quadrant B line (green) carries risks by virtue of poor specificity and need for very high occupancy, which calls for potential adverse effect including ‘excessive Pharmacology’. Line depicted in quadrant C (black) also generates robust PD response yet with observed adverse effects, which in the pre-clinical models could be ‘tolerable’ but not in patients. Line depicted in quadrant D (brown) generates desired PD response free of adverse effects, while preserving apparent target selectivity. However, the excessive exposure required to manifest the PD substantially exceeds the biological target occupancy. Such situation may suggest that efficacy that emanates from unknown target(s) might not exist or of sufficient congruency to humans. The three cases depicted in Figure 3(b) could lead to ‘false positive’ in pre-clinical animal studies ultimately facing failures or poor outcomes in clinical development.

Summary Drug discovery and development is a highly complex scientific process that requires translation of information from artificial systems and ‘proxy models’ to humans. Experimental (pre-clinical) stages of ‘proof of concept’ studies based on pharmacological agents should commit to rigorous scientific conducts aimed to detail and validate compounds as drug candidates with high degree of confidence. Since information detailed in this commentary is invariably generated within the pharmaceutical industry, investigators interested to deploy compounds for POC studies or participate in drug-candidates development should be cognizant of practices used in target validation with special attention to the PK/PD discipline. In particular, thorough understanding of mathematical detailing of drug exposures needed to generate sustained medical benefits within afforded safety and tolerability are paramount. Exceeding these ‘high bars’ we need to set for ourselves in the discovery (pre-clinical models) phase, scientists need to add to their research teams expertise in PK and detailed statistically/modeling analyses, as well as gain access to animal safety and tolerability data. Following these suggestions can be expected to optimize the veracity of the experimental data-base, including positive results in preclinical animal models, to eventually being realized in the clinic.

Feuerstein

643

(a) Target function inhibition (%)

Exposure [M]

100

ED 80 ED 50 ED 20

Disease Modulation; ‘Medical’ Benefits

0

poor selectivity

-

C

0B

“Off Target” Adverse Effects

0 High selectivity

-

+

A

0

D -

Target occupancy

1

+

Beyond target occupancy

(b) Target function Modulation (%)

Exposure [M]

100

ED 80

ED 50

ED 20

Disease Modulation ‘Medical’ Benefits

0

poor selectivity

-

0B

C

“Off Target” Adverse Effects

0 High selectivity

-

+ 0

A

D -

Target occupancy

1

+

Beyond target occupancy

Figure 3. Convergence of PK–PD information. (a) and (b) aim to provide the reader insights on qualities desired for PD responses in respect to predictive translation parameters for clinical development or POC studies. (a) and (b) Each displays two separate matrixes: upper panel describes a 3D display where drug exposures as correlated with target functional impact and the PD results (e.g. ‘clinical’ response, such as neurological function improvement by treatment). The lower panel is a 2D display of drug exposure in respect to target occupancy (from 0 to 1, one being 100%, e.g. receptor binding assay or PET) against selectivity (high selectivity (þ); poor selectivity ()) of the drug. In summation, this graphics could assist scientists to evaluate the desired PD impact that has best quality to serve for the objective of the pharmacology part of the study. (a) An ideal case is depicted where a dose–response (quadrant A) conforms well with the principles whereby PD is generated within a sufficient specificity to the target, and its occupancy well correlates to target biological function modulation. (b) calls attention to several possibilities that an apparent ‘successful PD’ might harbor ‘costs’ that could risk realizing successful translation either for lack of specificity that will not be tolerated in humans (quadrant B), non-specific toxicities (quadrant C) or PD unrelated to the target (panel D).

644 Declaration of conflicting interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding The author(s) received no financial support for the research, authorship, and/or publication of this article.

References 1. Oates JA. The science of drug therapy. In: Brunton LL, Lazo JS and Parker KL (eds) Goodman and Gilman’s The pharmacological basis of therapeutics [Chapter 5], 11th ed. New York: McGraw-Hill, 2006, pp.117–136. 2. Labreche MJ, Graber CJ and Nguyen HM. Recent updates on the role of pharmacokinetics-pharmacodynamics in antimicrobial susceptibility testing as applied to clinical practice. Clin Infect Dis 2015; 61(9): 1446–1452.

Journal of Cerebral Blood Flow & Metabolism 36(3) 3. McLaughlin PJ and Zagon IS. Duration of opioid receptor blockade determines bio therapeutic response. Biochem Pharmacol 2015; 97: 236–246. 4. Crump KS, Guess HA and Deal KL. Confidence intervals and test of hypotheses concerning dose response relations inferred from animal carcinogenicity data. Biometrics 1977; 33: 437–451. 5. Cosmatos D, Feuerstein G and Chow S-C. Strategic concepts in translational medicine. In: Cosmatos D and Chow S-C (eds) Translational medicine strategies and statistical methods. Boca Raton, FL: CRC Press, Taylor & Francis, A Chapman and Hall Book, 2008, pp.9–32. 6. Zhou J, Zhuang J, Li J, et al. Long-term post-stroke changes include myelin loss, specific deficits in sensory and motor behaviors and complex cognitive impairment detected using active place avoidance. Plos One 2013; 8: 1–15.