January 2016 Contents


Challenges in Transdermal Drug Delivery, Big Data, PBPK among Topics at the 2015 Annual Meeting and Exposition

2015 Annual Meeting and Exposition kicked off with keynote address by young cancer research phenom.

By Randolph Fillmore and Keith Freeman

T o say that Jack Andraka is “far from a typical teenager” is a huge understatement. When he was just 15 years old and a high school sophomore, he invented a rapid, inexpensive, early detection test for pancreatic, ovarian, and lung cancers.

Jan 16 Andraka“I did it over my summer vacation,” keynote speaker Andraka told a full house on Sunday afternoon kicking off the American Association of Pharmaceutical Scientists’ (AAPS) 28th Annual Meeting and Exposition held October 25–28 in Orlando. “After a close family friend died of pancreatic cancer, I wanted to know why, and when I found out that 85 percent of pancreatic cancers were diagnosed late, I thought an early detection diagnostic should replace an expensive test, one that had not been revised in over 60 years and did not detect early disease.”

Andraka knew that, besides being inexpensive, the test needed to be rapid, simple, sensitive, and selective. He began by researching 8,000 proteins, which he compared to “searching for a needle in a stack of needles.” When he discovered the one protein that he thought would work as a marker when found in high levels in the blood, he sent out emails to 200 professors explaining what he was doing and asking for research space.

“I received 199 rejections!” said Andraka. “The only professor who expressed interest and eventually offered me lab space was at the Johns Hopkins University. After seven months work, I ended up with a small paper sensor that costs three cents for a test that takes five minutes to run and is able to detect cancer in its early stage, when it is nearly 100 percent curable.”

Andraka’s talk, The Future of Medicine, concluded with his plea to make all scientific journals open access. The pay walls associated with most journals put them out of financial reach of state universities and having to pay $35 to retrieve an article off the Internet is too much for many to afford, he said.

“We now have a knowledge aristocracy—imagine a knowledge democracy,” said Andraka, now 19. “Access to science information should not be a luxury, but a basic human right.”


Avoiding Brain Freeze: Highlights on Contemporary Drug Delivery across the Blood-Brain Barrier, invited attendees to consider the challenges in getting drugs across the blood-brain barrier (BBB), a feat especially challenging for delivering drugs to treat Alzheimer disease and central nervous system cancers.

Jan 16 ladyJohn D. Higgins, Ph.D., senior principal scientist, Merck & Co., Inc., said, “Ninety-eight percent of small molecules do not cross the BBB. One hundred percent of large molecules do not cross the BBB. Yet, there is a great opportunity for delivering drugs to the brain because of extensive vascularization—it is a matter of getting it through the capillaries. While we can get some small molecules across the BBB, the ‘holy grail’ is getting large molecules across.”

There are a number of strategies for getting a molecule past the BBB, he told attendees. The “brute force” method opens up tight cell junctions, but can let good things as well as bad things to pass through. For drug developers, this method might not be the best option. However, there are new and better methods for “passive diffusion” and “active transport.”

Three or four receptors might be of interest, although each has its own complexities and interest. The transferrin receptor, a carrier protein for transferrin, the iron-binding blood plasma glycoprotein, is a candidate, but has limitations. The insulin receptor is another BBB receptor, but is “risky to mess with,” Higgins warned. The low density lipoprotein receptor, part of the multiligand receptor family, also offers some opportunity. The diphtheria toxic receptor may be too toxic, although the receptor is upregulated in inflammatory conditions. He also covered cell-mediated delivery for use in nose-to-brain delivery. “But what’s going to happen to the nose mucosa with daily administration?” he asked.

Nanoparticle approaches can achieve BBB penetration. The question is: “How much drug can we load into a nanoparticle?” The answer, said Higgins, is usually not good.

Jan 16 SessionSilvia Muro, Ph.D., associate professor, Department of Bioengineering and Institute for Bioscience and Biotechnology Research, University of Maryland, discussed drug delivery across cell barriers via “cell adhesion molecule-mediated transport.” Cell adhesion molecules (CAMs) are proteins located on the cell surface and involved in binding with other cells. To cross the BBB, Muro advocated taking advantage of various receptors and vesicular transport pathways. She covered the advantages and disadvantages of some of those vesicular pathways, including the small size of vesicles, an obstacle for transport of drug delivery systems, and the existence of competing ligands in the body. In turn, nonclassical pathways are not well understood but, she explained, hold considerable promise. For example, the CAM pathway is naturally used by leukocytes to traverse the vascular endothelium, which is not restricted to drug carrier size and has no endogenous ligand competitor.

“There is no natural ligand in the body to compete with CAM and carriers targeted to CAM, so anything that binds will go through the pathways,”she said.

An additional problem is that some diseases cause dysfunction of the regulatory machinery that controls these pathways. Alzheimer disease is one example.

“Determining which pathways are in place in a particular disease is important,” she said.


Physiologically based pharmacokinetic (PBPK) models are routinely used to predict outcomes of clinical trials. In Physiologically Based Pharmacokinetic Models: Impacting Drug Development and Facilitating Regulatory Approvals, speakers addressed their experiences using PBPK models for predicting drug absorption, drug-drug interactions (DDIs), and to describe PK in pediatric populations.

According to Handan He, Ph.D., leading scientist at Novartis, PBPK models can be applied in the earliest stages of research and development. She explained how Novartis builds PBPK absorption and disposition models and discussed the development, impact, and challenges they have encountered using PBPK models.

Jan 16 audience1“The model is very useful for setting up the formulation strategies,” He said. “However, if you don’t have a lot of data at an early stage, it is difficult to use PBPK.” They have also used the PBPK model for future clinical trials prediction and applied the model to specific populations by using PBPK for DDI prediction strategy. “The predictions,” she said, “worked very well.”

For DDI prediction strategy, He said that Novartis uses a “bottom up” approach to view the PBPK model, and she discussed several oncology drugs that have been approved over the last two years that were developed using PBPK modeling. Their use in routine oncology practice is evidence that PBPK benefits cancer patients, she said.

In some instances, using PBPK modeling to predict outcomes for special populations requires additional work.

“A few years ago we had two compounds we wanted to use to try to predict PK profiles in pregnant women,” explained He. “However, there was no commercial software model, so we had to construct the model ourselves and build in the information. We tested the model for several commercial compounds, and they all turned out fantastic.”

PBPK models for developing pediatric drugs and for drug labeling are areas where the impact of PBPK simulation is becoming more visible, and PBPK models are routinely applied to inform pediatric trials.

“The take-home message is that there will be an increase in the number of opportunities to use PBPK, not only to waive some of the clinical studies, but also using some of the PBPK information for drug labeling,” said He.

Neil Parrot, Ph.D., senior scientist, Modeling and Simulation at F. Hoffmann-La Roche Ltd., discussed Driving Drug Development with Integrated PBPK Approaches. Parrott specifically addressed PBPK for developing adult and pediatric dose predictions.

“When researchers first started applying PBPK, we thought we could use the animal model data and in vitro data to compare the simulations in both and then try to simulate what we would expect,” Parrot explained. While that approach should lead to better predictions of human PK, he suggested that the goal is not just predicting PK but, rather, combining PK information to understand the efficacy of the molecule.

“We convinced ourselves that this method was superior for PK predictions and other benefits, and now routinely employ it in human dose prediction,” said Parrott. “After having done this for a number of years we have now collected examples from 24 molecules, and our predictions are in line with the retrospective evaluation, with similar accuracy.”

In a presentation titled From Research to Regulatory Review: Using PBPK Predictions to Support Regulatory Decisions, Ping Zhao, Ph.D., a pharmacologist with the Food and Drug Administration’s (FDA) Office of Clinical Oncology, laid out how the PBPK model can be applied to assess the impact of drug behavior using the “predict, learn, and confirm” paradigm. Jan 16 awards 

“PBPK can be applied to assess the impact of various factors, either intrinsic or extrinsic, or a combination of the two,” said Zhao. “The model is considered adequate to predict clinical scenarios.”

FDA created their PBPK program in 2013 with the objective of ensuring evidence-based and consistent quality review processes to support decision making, said Zhao. They have been working to harmonize with regulatory bodies outside FDA with regard to PBPK recommendations and are using both in-house research data and published information from sponsors. Together, these databases will be used to create policy that will eventually guide the review process.


John Markowitz, Pharm.D., professor, Pharmacotherapy and Translational Research at the University of Florida College of Pharmacy, presented Dietary Supplement Interactions: What Have We Learned? He explained that much remains unknown about the “sheer number” of supplement products out there and the effects of taking them in combination with prescription drugs. He called this reality “alarming” because 50 percent of adults use botanical drugs and, of those, 70 percent do not tell their health care providers that they are combining supplements with prescribed drugs.

He said, “Many botanical products are used and marketed as combinations of botanicals.”

The majority of today’s studies looking at drug-supplement-botanical interactions are phase 1 metabolism studies and phase 2 studies, he told attendees and suggested that in vivo animal model studies are “pretty poor models” for drug-drug interaction studies. Using them for studying drug-botanical interactions is even “hazier.”

Jan 16 AliceIn vivo human studies looking at drug-supplement interactions could be carried out but would have many limitations and long, expensive time lines. Additionally, most studies do not measure botanical product metabolism.

According to Mark Menolascino, M.D., who practices functional medicine at the Meno Clinic, “everything goes back to nutrition” and “when your nutrition intake is shifted, your health is shifted.”

In discussing the potential benefits of genetic knowledge, its promise for personalized medicine, and its relationship to functional medicine applications, he said, “You are not your nature, you are your nurture. You are not fixed with your genetic markers. And that is a powerful message.”

Looking at today’s health issues, he predicted that the generation just being born will live shorter lives than their parents, and that one in three children born today will have diabetes by the time they are adults. “It’s shocking what’s going on,” he stated.

J. Christopher Gorski, Ph.D., principal research scientist, Mylan Pharmaceuticals, discussed Drug-Supplement Interactions: Drug Development Considerations and Case Studies. He explained why the study of botanical-drug interactions is important, noting that herbal supplements are a $6.5 billion industry, yet botanicals are not subject to FDA approval.

Even if botanicals were subject to FDA approval, there are challenges in determining the potency of herbal supplements. Factors such as sunlight, soil, temperature, planting time, harvesting, and rainfall patterns not only influence potency, but also create variations, and both these factors make herbals difficult to study. Also, because different parts of a plant could be used in different supplements, efficacy could be impacted by not only those variations, but also by variations in processing.

Using St. John’s wort as an example of a variable active ingredient, he said, “There are also multiple varieties of St. John’s wort formulations to consider, such as tablets, oils, drops, and leaves. When studying drug-botanical interactions, which formulation do you pick?”


How can we use big data to predict and/or prevent drug-related adverse events (AEs) and drug-drug interactions (DDIs)? That was the question tackled by Lang Li, Ph.D., professor, Medical and Molecular Genetics, Indiana University; Darrell Abernethy, M.D., Ph.D., associate director for Drug Safety in the Office of Clinical Pharmacology at FDA; and Tal Lorberbaum, M.A., Columbia University, in Big Data in the Pharmaceutical Industry: Opportunities and Challenges.

Jan 16 exhibitionLi, who addressed big data issues related to DDIs, told attendees that each year there are 195,000 hospital visits and 74,000 emergency room visits related to AEs caused by drugs or by DDIs. Three to five percent of all AEs are caused by drug AEs or interactions. He also noted that the drug inter-action-adverse events issue is “not a new topic for this community. The DDI issue has been problematic—and studied—for a long time.”

In his work on DDIs, Li used the medical records of 12 million patients and other big data sources, he said, such as public domain abstracts on PubMed, in FDA databases, and pharmacogenetics studies, to look at and compare thousands of drug profiles to predict AEs by looking for drug interaction “signals.”

“In our studies, we asked if we could scale up by looking not at a few drugs, but at thousands, and to be more efficient and scientific by using methods different than used in traditional drug interaction research,” explained Li. Accordingly, they investigated a great number of drug interaction pairs and their clinical impact and looked specifically at statins, popular drugs in the data and known to cause myopathy. A second investigation involved DDIs related to diabetes.

“Once we had data, we had to be able to look for clinical impact,” he said.

Abernethy discussed a current FDA program that links pharmacological drug action data to identify and predict potential adverse drug effects and interactions and evaluate potential signals for adverse drug events. He laid out several questions regarding adverse events and drug safety. Can we predict AEs? What information and analytical tools do we need? How would that data be evaluated, developed, and shared? What are the phenotypes that can be observed, and how can they be effectively described?

“The integration of information is one of the biggest challenges,” he told attendees. “Information on adverse events comes from a variety of sources. The amount of information available in the public literature is almost limitless; however, the information that would have an impact on what we are looking at may be rather small, so it has to be ferreted out. Drug companies also have a great amount of information that would be very specific and more relevant to a particular problem.”

He noted that standardization of databases would also be useful and that a number of companies are already involved with FDA in bringing that information together.

Lorberbaum spoke on the value of using clinical trials data to better understand mechanisms of AEs. He said there are 200,000 trials ongoing globally; we can learn a lot from them, especially about drug-related AEs. One problem, however, is that when you control for the characteristics and demographics of patients taking part in clinical trials, you might not see the same heterogeneity in patients who are ultimately taking the drug once it is on the market. Additionally, side effects may be rare or not caught during clinical trials; this issue is especially important when it comes to evaluating DDIs. Jan 16 career 

“Data on DDIs is especially important for older patients since the Centers for Disease Control estimates that many older Americans are taking at least two prescription drugs at the same time,” he said.

He explained his efforts working with large data sets to mine for and predict DDIs and to establish data sets to predict drug safety. He pointed to a number of big data sources for information about drug adverse events, including FDA databases and electronic health records. Lorberbaum also said that there are drawbacks to using observational data, which tend to be “very messy” and lead to false positives. How can we account for and control for these biases? Building a side effect profile is one way, he said, and offered an example of an experiment carried out with animal models to observe DDIs.


In Scientific and Regulatory Challenges in the Development of Generic Transdermal Drug Delivery (TDD) Systems, Tim Peterson, M.S., TDD product development leader at 3M, and Yellela S.R. Krishnaiah, Ph.D., research chemist, FDA’s Office of Product Quality Research, emphasized that the technologies used for optimal skin contact and high performance in transdermal drug delivery are challenging and highly dependent on choice of both adhesive and matrix.

Peterson said that the topic of transdermal drug delivery is “dear to my heart,” especially the design considerations and in vitro quality testing. 3M has developed about 30,000 adhesive products since 1925, and those with substrates designed to adhere to the skin are the most challenging.

Key aspects to consider for optimal transdermal drug delivery are good contact with the skin with a pressure-sensitive adhesive (PSA) with an elastic and viscous character; a sufficient reservoir for the drug; biocompatibility with the skin; functionality good for the release of the drug; and a good matrix for the drug, one that will not interfere with or react with the drug’s active pharmaceutical ingredient (API).

Jan 16 Mentoring“Adhesives have been the same for years, but with diverse chemistry for the polymers,” said Peterson.

He listed a number of properties in skin that make adherence difficult: elasticity up to 50 percent, a low energy surface, and hair, pores, and dead skin cells. Accordingly, design considerations include choosing an adhesive that has high solubility characteristics and choosing among the variety of characteristics in polymers right for each API’s compatibility.

“The choice of adhesive matters with the choice of API and its characteristics,” he cautioned, and discussed methods to measure solubility by the “crystal growth method,” explaining how crystals are absorbed. “The matrix has to have adequate solubility to maintain delivery, yet not deplete the reservoir before the wear period is over as excess solubility can be a problem that leads to reduced delivery.”

The bottom line for Peterson was that the ability to modulate drug solubility within the polymer is as important as the overall compatibility among the elements.

Krishnaiah discussed at length one transdermal drug delivery problem—cold flow. “Cold flow occurs when medication migrates beyond the edge of the transdermal drug delivery patch,” he explained. “Among the causes of cold flow is the degree of PSA plasticization, which can be either too high or too low, with the loss of medication as a result.”

FDA is concerned about cold flow and other defects, such as leaky reservoirs and adhesion failures (the most common defect, accounting for one third of all patch defects).

Jan 16 FoundationKrishnaiah presented the internal research on the development of a method for measurement of cold flow in patches (a model transdermal system used in hormone replacement therapy). He noted that the problem of cold flow can reduce the drug flux, and drug transfer can occur from the cold flow region of patches. The cold flow measurement tools can be useful in selecting the best adhesive, excipient, and drug substance ratios and formulations, as well as developing appropriate quality control methods and potentially justifying acceptance criterion for cold flow. He emphasized that cold flow should be included in stability protocol along with the results at the time of new and abbreviated new drug application submissions.

Generic transdermal drug delivery systems, he said, are required to have the same API, dose, strength, and bioavailability as the prescription systems. As problems often arise with adhesion and skin irritation, he suggested that manufacturers adopt quality-by-design approaches for developing transdermal products.

Randolph Fillmore is the director of Florida Science Communications Inc. (FSC), and Keith Freeman is a staff science writer for FSC specializing in health and health care policy.