Clinical Pharmacology— The Quarterback of Drug Development

PEER REVIEWED Ellen Leinfuss; Julie Bullock [DOI: 10.14524/CR-17-0016]

A quarterback’s job is to direct the team toward the end zone Clinical pharmacology accounts for about 50% and score as many points as possible. Considered the leader of of a drug label. Its scope ranges from facilitating the team, the quarterback is often responsible for calling the play; the discovery of new target molecules to deter- mining the effects of drugs in different popula- however, the quarterback isn’t just responsible for knowing how to tions. From both industry-wide and regulatory do his job—he has to know the responsibilities of every player on perspectives, the levers of clinical pharmacology the field to win the game. can address the huge challenges of late-stage attrition and increase the efficiency of drug development in the quest to bring the “ball” into the end zone.

June 2017 20 Clinical Researcher Clinical pharmacology accounts for about 50% of a drug label

FIGURE 1: Role of Clinical Pharmacology in the Drug Label

Warning

Warnings Contraindications and precautions

Special dosage instructions Special populations

Adverse Drug reactions interactions

Clinical Dosage and pharmacology administration PK: ADME

WHAT IS CLINICAL PHARMACOLOGY? in drug development due to scientific challenges in predicting efficacy and safety or characterizing Clinical pharmacology is the science of the sources of drug response variability at early, less relationship between drugs and humans. It focuses expensive stages of discovery. on drug action, and incorporates pharmacolog- Clinical pharmacology can help stakeholders ical principles and techniques into the clinical to address these challenges and improve decision development cycle. It is increasingly leveraged making at critical drug development milestones, at all phases of development with an eye toward whether in early proof-of-concept phases (preclin- delivering the right drug to the right patient with the ical through IIa), or in the later stages where more right dose at the right time. robust risk and efficacy profiles are established Clinical pharmacology contributes to a range of (Phase IIb through III). clinical decisions, and features prominently in the Clinical pharmacology tools, methods, and drug label, as shown in Figure 1. frameworks (e.g., mechanistic or quantitative) span A recent paper1 by a group of scientists from distinct subspecialties, and can have a significant the U.S. Food and Drug Administration (FDA), impact at the interface of nonclinical and clinical academia, and industry clearly articulated how phases. They can greatly reduce uncertainty clinical pharmacology methods and a quanti- related to therapeutic targets, dosing, and patient tative framework can improve the efficiency of populations in which the novel compound may drug development and evaluation. That paper, have the most efficacy.1 “Improving the Tools of Clinical Pharmacology: Goals for 2017 and Beyond,” points to limitations

Clinical Researcher 21 June 2017 FIGURE 2: Clinical Pharmacology Toolkit

According to Kristofer Baumgartner, a strategic communications officer with FDA, MIDD already Population has resulted in shorter trials with fewer patients, PK fewer postmarketing studies, and tailored drug dosing. “The use of modeling and simulation has the potential to make the interpretation of Clinical Trial Exposure- data more efficient, improve the prediction of Simulation Response drug safety and efficacy, and explain variable patient responses,” Baumgartner told BioCentury in September 2016. “MIDD approaches, such as ClinPharm physiologically based pharmacokinetic (PBPK), Quantitative dose- and exposure-response, and disease-drug- Toolkit trial modeling and simulation approaches have been used to inform a variety of drug development and public health decisions, such as drug dosing QSP PBPK and use in certain subpopulations.”2 Key items in the clinical pharmacology toolkit include: • Pharmacometrics Modeling—Population Disease PK, exposure-response, and disease-state Models modeling are used to predict clinical outcomes, provide support for dose recommendations (justification and modification), assess safety and efficacy trends across exposure ranges, and inform “go/no go” decisions. CLINICAL PHARMACOLOGY TOOLS • PBPK—This technology informs key research As clinical and development (R&D) decisions relating to pharmacology has Using quantitative methods—among them, clinical trial design, first-in-human dosing, for- model-informed drug development (MIDD)—to mulation design, dosing in special populations, become increasingly optimize drug development decision-making is and predictions of the likelihood of drug-drug quantitative, it becoming core to the process. These methods, as interactions (DDIs). shown in Figure 2, are key components of clinical • Quantitative Systems Pharmacology (QSP)— has become more pharmacology and medical reviews. An emerging mechanistic modeling approach relevant and critical MIDD frequently provides answers to develop- focused on target exposure, binding, and ment gaps, increases understanding of benefit/risk, expression. It is employed to identify biological for decision-making identifies issues that need further characterization, pathways and disease determinants. QSP purposes. minimizes postmarketing burden, and informs helps understand PD as the disease progresses labeling decisions. Beyond the many ways that through the body. MIDD informs drug development decisions and • Quantitative Systems Toxicology (QST)—QST strengthens the science, it also reduces time and modeling combines toxicity and “omics” data to cost to market via smarter and potentially smaller focus on modes of action and adverse outcome or avoided studies. pathways. From FDA’s perspective, MIDD is a quantitative framework for using knowledge and inference generated from integrated models of compound-, THE RIGHT DRUG/PATIENT/DOSE/TIME mechanism-, and disease-level data to predict Clinical pharmacology, as the quarterback of and extrapolate how drug candidates should or drug development, plays a strategic role across will perform from pharmacokinetic/pharmacody- the entire process. As clinical pharmacology has namic (PK/PD), safety, and efficacy standpoints. become increasingly quantitative, it has become PK is what the body does to the drug (drug concen- more relevant and critical for decision-making tration); PD is what the drug does to the body (drug purposes. Figure 3 affirms the growing importance effect); PK/PD is the drug action over time. and impact of MIDD within the industry and its acceptance by global regulators.

June 2017 22 Clinical Researcher Right Drug continues to grow, new methods, approaches, It is now largely understood that incorrect target tools, and technologies to expedite target selection selection and validation contribute to drug attrition. and validation are needed. Too often, drugs progress through the R&D process The aforementioned QSP, an emerging tool in with a poor understanding of the drug target or the clinical pharmacology, focuses on improving tar- appropriate mechanism of action for the targeted get selection and validation through mechanistic disease condition. A drug candidate may fail to modeling. QSP combines computational modeling interact with the target of interest in humans, and experimental data to examine the relation- or the drug may interact with the target without ships between a drug, the biological system, and benefitting the patient. Furthermore, a drug may not the disease process. This systems-level perspective affect a target’s downstream biochemical pathway integrates quantitative drug data with knowledge after target binding, or unforeseen safety problems of its mechanism of action. may emerge.3 Some believe that up to 40% of new QSP models can be used to address the complex projects closed for efficacy reasons were due to lack and heterogeneous diseases prevalent today. of data linking the target to the disease.4 Leveraging QSP can reduce Phase II attrition At the core of these failures are fundamental by allowing the investigation of a wide range of knowledge gaps regarding intended targets, their what-if scenarios to determine what the likely effi- biological relevance to a given disease, their cacy of the drug is going to be in advance of clinical interactions with the investigational compound, investigation. This can facilitate lead optimization or a compound’s therapeutic mechanism of action. very early on in the discovery process.5 As the sizeable list of potential disease targets

FIGURE 3: Achieving the Right Drug/Dose/Time Using MIDD

Preclinical Phase I Safety Phase II Efficacy Phase III Population Filing

PK/PD PK/PD Safety In Vitro PD Biomarkers Biomarkers Efficacy Animal PK/PD Safety Safety PK Preliminary Efficacy Efficacy Data Generated Data

PBPK, PK/PD QSP Integrated ER/DR ER/DR Preclinical PK/PD ER/DR Safety PopPk ER PopPk Efficacy Analysis Tools

Go/No Go Data-driven dose Go/No Go Target Selection Phase III Dose modifications Regimen Feasibility Intrinsic Factors for safety

Decisions Selection Extrinsic Factors and efficacy

QSP: Quantitative Systems Pharmacology, ER: Exposure-response, DR: Dose-response, PopPK: Population PK

Clinical Researcher 23 June 2017 Right Patient with various drug characteristics (e.g., DDIs, food We know that individuals react differently to interactions) to recommend when and how that identical therapeutics. In fact, drugs used today are individual receives the drug. effective in only 25–60% of patients. Using clinical Leveraging clinical pharmacology tools pharmacology and its quantitative methods throughout drug development is a key to informed (specifically PBPK) has allowed us to move from drug development decision-making. “one size fits all” in drug development to serving stratified populations. Between drug approval GETTING IT RIGHT FOR CHILDREN— and postmarketing analysis, we can now address A CASE STUDY special populations, including pregnant women, In pediatrics, traditional development methods pediatric and geriatric patients, and those with can be problematic for ethical and logistical rea- renal and hepatic impairment. sons; after all, children are not small adults. Most Clinical pharmacology tools facilitate the importantly, children are a heterogeneous popu- understanding of disease mechanisms and how lation, especially neonates and infants, as organ they relate to such realms as an individual’s maturation affects drug exposure and response. genetics, metabolism, and environmental factors. In this case example, there was no precedent Diagnostics are increasingly being used to screen for regulatory acceptance of an accelerated patients for genetic variations or biomarkers, and pathway for a small molecule antiviral drug to treat models are being leveraged that predict disease respiratory syncytial virus infection in infants. susceptibility and disease progression. The prom- The approach was to conduct a healthy volunteer ise of precision medicine is becoming a reality. single-ascending dose/multiple-ascending dose Right Dose study, and then in parallel begin collecting PK Poor dose selection will in turn often lead to data in infants. Specifically, after dosing about 60 failed Phase III trials, delays/denials of regulatory healthy volunteers, the next patient who received Clinical pharmacology submissions and/or changes in doses postap- this drug was an infant. and MIDD have already proval, additional postmarketing commitments, Preclinical models, disease and PK/PD modeling and further requirements for development in other and simulation, and clinical PK, viral kinetics, and had an unquestionable age and/or ethnic groups, all of which ultimately safety data from healthy adults were used to define eliminate/invert any initial gains. Taking the above impact on drug exposure-response relationships and inform the into account, poor dose selection is one of the main dosing strategy. To provide initial estimates of PK 6 development. causes for the soaring cost of drug development. parameters and the impact of dose on inhibiting viral Clinical pharmacology technologies sit replication, standard allometric scaling of clearances However, for squarely in the middle of this challenge, with in volumes, including incorporating a renal matura- innovation in this MIDD leveraged for determining first-in-human tion factor, was used; this was utilized to estimate a doses, establishing dose-response, and inform- Phase I starting dose in pediatrics that had demon- discipline to continue, ing dosing regimen decisions. MIDD provides strated sufficient safety margins in preclinical species we not only need a framework for regulatory decisions and dose and in exposures obtained in adults. optimization, as too high a dose can cause toxicity The starting dose was also expected to be in the to integrate these and too low a dose decreases the chance of demon- low-therapeutic range because this model initially tools earlier in drug strating efficacy. Specifically, PK/PD, adaptive was purely based on adult data, without any pediatric trial design simulation, and in silico DDI studies information. The model was updated and adjusted development, we also have repeatedly demonstrated their value in dose over the course of the study, as data were collected. must cultivate our optimization. The case demonstrates how the use of clinical Right Time pharmacology tools was essential in supporting the future pharmacology By combining a quantitative analysis of the right continued development of this new drug, which leaders by nurturing drug, with the right patient, and the right dose, stands to benefit some of our smallest patients. clinical pharmacology will facilitate developing the their scientific acumen right dosing schedule for a patient to achieve and ADVANCING CLINICAL PHARMACOLOGY along with their maintain therapeutic benefit. Clinical pharmacol- While we have discussed the amazing impact that ogy technologies will allow prescribers to integrate clinical pharmacology and its quantitative tools communication skills. an individual’s genotypic and phenotypic data have already had on reducing drug attrition, the

June 2017 24 Clinical Researcher lack of consistent knowledge and usage across the in silico development paths. The virtual drug References industry could slow its impact. In response to this development program is always ahead in each 1. Zineh I, Abernethy D, Hop CECA, Bello A, McClellan issue, members of FDA, the industry, and academia “time zone” compared to the in vivo drug MB, Daniel GW, Romine have recommended several areas for focus and development program. This paradigm confers MH. 2017. Improving improvement1: a self-learning process that guides in vivo drug the tools of clinical pharmacology: goals for • Erase the preclinical/clinical divide: Drug development; thus, the clinical program gains 2017 and beyond. Clin development is often depicted as being linear; the ability to move faster, more predictably, and Phrmcl & Thrps 101:22–4. a drug is identified in discovery, advances more reliably. doi: 10.1002/cpt.530 2. BioCentury. September 5, into preclinical studies, progresses into the 2016:22–4. clinic, and then finally enters the realm of CONCLUSION 3. Cartwright ME, Cohen patient care. We must abandon this thinking. S, Fleishaker JC, Madani In reality, the information flow is circular, and Clinical pharmacology and MIDD have already S, McLeod JF, Musser B, had an unquestionable impact on drug develop- Williams SA. 2010. Proof collaborating across historically siloed disci- of concept: a PhRMA plines and sharing information allows teams to ment. However, for innovation in this discipline to position paper with recommendations for make better decisions about drug candidates. continue, we not only need to integrate these tools earlier in drug development, we also must cultivate best practice. Clin Phrmcl The “learn-confirm” archetype, although & Thrps 87:278–85. doi: introduced many years ago, is increasingly our future pharmacology leaders by nurturing 10.1038/clpt.2009.286 being leveraged, thus creating a continuum their scientific acumen along with their communi- 4. Center for Health Policy cation skills. at Brookings. 2015. of learning and knowledge that builds across Improving productivity in the drug development cycle. Some observers Case studies articulated in “plain language” pharmaceutical research and development: the role suggest that investing in MIDD is essential to are compelling and should be encouraged by R&D management. Pharmacometrics education of clinical pharmacology leveraging preclinical and early-stage clinical and experimental data to inform late-stage decision making.1 and real-world training opportunities within medicine. https://www. drug development also need to expand to provide brookings.edu/events/ • Advance the use of QSP approaches: QSP has scientists with hands-on education. improving-productivity-in- enormous potential for improving pharmaceu- pharmaceutical-research- In short, this discipline sits at the exciting and-development/ tical R&D productivity. QSP models predict how intersection of biological sciences and computing 5. Toon S, van der Graaf drugs modify cellular networks in space and capabilities—a ripe area for innovation and ethical P. 2016. Quantitative time, and how they impact and are impacted by systems pharmacology societal impact. After all, clinical pharmacology is brings value to drug human pathophysiology. One barrier that must the quarterback of drug development. development. Clin Ldr. be overcome for QSP to be effective is the lack https://www.clinicalleader. of common tools. Piet van der Graaf, PharmD, com/doc/quantitative- systems-pharmacology- PhD, vice president and head of Certara QSP, brings-value-to-drug- says, “To take QSP to the next level, we also development-0001 need a QSP infrastructure, i.e. software and 6. European Medicines Agency report from Dose information technology tools, that both com- Finding Workshop. April 7, panies and regulators can use and understand. 2015. Without an infrastructure, converting QSP from a purely academic discipline into one that industry actually uses will be very difficult.” • Increase the number of vetted biomarkers: Both internal and regulatory decisions require using trusted biomarkers. Many biomarkers are developed as “one-offs” that cannot be repur- posed for other drug programs. Precompetitive partnerships have been critical in increasing the number of vetted biomarkers to support clinical development. • Enhance the “totality of evidence” approach: Ellen Leinfuss is chief Julie Bullock is senior director Meeting the regulatory requirements for commercial officer at Certara. for consulting services at efficacy entails using an array of data from Certara, and contributed to more than 14 new molecular clinical pharmacology studies. We recommend entity approvals during her building parallel, but connected, in vivo and 10-year FDA career.

Clinical Researcher 25 June 2017