Moving Closer to Dynamic Drug Development

liuzishan/Stock.Adobe.com

Modernization, collaboration, and innovation are some of the most important concepts in the drug development sector due to how increasingly complex the therapeutic landscape is becoming. Despite high levels of investment being poured into the biopharma industry each year, many challenges that arise from the use of traditional models and processes still exist within drug development. For example, the emergence of artificial intelligence (AI) can find potential new molecules a thousand times faster than existing models, yet patients are not able to access the benefits because of bottlenecks in the development process, such as project management and collecting data (1).

However, hope is on the horizon as novel solutions and improvements aimed at benefitting the drug development process with the right framework are coming to the fore. Focusing on the patient, digital and technology enablement, and advanced analytics are a few ways that research and development have been given a kickstart in transforming traditional approaches; although, this is just the beginning of a long journey toward a more dynamic, beneficial drug development process (2).

Recent changes in the industry

Traditional drug development is a cost and time-intensive process, which has caused an array of issues in the past. Ted Broering, president of early clinical and bioanalytical solutions at ICON, mentions that apart from the standard objective of establishing drug safety in healthy volunteers, early phase trials are also on the critical pathway of collating data to determine if the drug holds potential for further development and label expansion. In terms of increasing costs and pressures for smaller timelines, he explains how early phase study designs have become the proverbial “kitchen sink” with multiple questions being asked that help determine the development pathway.

“These early protocols are now being written as ‘adaptive’ designs, allowing flexibility to adjust dose levels, to add or clarify sampling timepoints based on emerging data, and trial expansion if needed,” Broering says. “Long gone are the days of a simple ascending dose first in human (FIH) trial, now we can execute adaptive combination designs that would have once been conducted as four different trials—single ascending dose, multiple ascending dose, food effect, and patient arm all in one umbrella protocol.”

In terms of downstream processing, Heidi Jones, market development manager, Process Chromatography at Bio-Rad Laboratories, explains that column packing can be a very time consuming, resource intensive activity, requiring specialized equipment and training. “Additionally, variability in packing quality can lead to inconsistent lot-to-lot purification results,” she says. “Prepacked columns (packed with the appropriate resins, i.e., affinity, ion exchange, mixed mode, and so forth) can mitigate many of these issues and constraints to efficiency. [These prepacked options] provide consistent, FDA compliant columns that meet specific quality criteria” (3).

AI/machine learning types of applications have made a significant impact for analyzing large, linked data sets for accelerating drug development while reducing costs, according to Philip Russmeyer, founder and CEO of FITFILE, which was founded in 2020 by Philip to help deliver safer, faster, and better profiles of record-level health. He further elaborated on recent advances in genomic sequencing and clinical trials, as pharmaceutical companies are increasingly collaborating with academic institutions, government agencies, and non-profit organizations to share resources and non-commercially
sensitive information.

“Furthermore, advances in genomic sequencing are enabling more targeted drug development, facilitating faster drug approvals,” Russmeyer says. “For clinical trials, with record linkage and advanced analytics the right patients can now be quickly and efficiently identified and recruited (4), and site selection has become much easier with the arrival of more granular and dependable information” (5).

Russmeyer also emphasizes that regulatory bodies, like the UK Medicines and Healthcare products Regulatory Agency (MHRA), have recently implemented new policies to streamline the drug approval process. “Under the latest MHRA framework, clinical trials application processes in the UK will be more proportionate, streamlined, and flexible without compromising on safety,” he says.

When asked about the 505(b)(2) pathway, Robert Lee, president at Lubrizol Life Science Health in the contract development and manufacturing organization (CDMO) division, points out that this strategy is effective at accelerating the process, repurposing drugs for different routes of administration, or developing a differentiated dosage form. “Differentiation can be in terms of better biological performance (i.e., lower dose) or a safer product (i.e., lower toxicity or reduced off target exposure). The safety packages are already out there for approved drugs, so there are fewer hurdles to overcome, shrinking the cost and development time significantly,” he explains. “One potential approach to achieving a successful 505(b)(2) repurposing is by utilizing novel excipients in combination with the existing, marketed APIs to make a product that is suitable for new routes of administration.”

Bottlenecks in early drug development

Bottlenecks in data and project areas, along with other hurdles, can impact a drug development process by weeks, months, or years, and its overall influence on the lifespan of a drug development program can magnify more inefficiencies. Delays, added costs, and more can create less availability of new and innovative treatments for patients (1).

Tomasz Kostrzewski, PhD, VP, science and technology at CN Bio Innovations, explains that 90% of drugs that enter clinical trials fail to become approved medicines, and to avoid a “rubbish in–rubbish out” scenario, the problem needs to be addressed ahead of this phase before a significant reduction in attrition rates begins. “Prior drug development decisions regarding target and candidate selection are based upon in-vitro and in-vivo animal models with limited translational relevance as they either lack human complexity, or relevance. Recreating the human immune system poses a particularly significant challenge,” he says. “Within the development phase, animal studies are widely used to further evaluate the safety and efficacy of a drug candidate, to determine optimal dose/schedule for clinical trials and to inform regulatory submissions, but again, interspecies differences mean that the data from these studies often fail to translate into human outcomes.”

Kostrzewski also discusses how drug discovery is moving away from small molecule discovery toward newer drug classes, such as cell and gene therapies, antibodies, and peptides. “This new wave of therapeutics is the future; however, they provide new challenges due to their human-specific mechanism of action which [means] animals [are] less suited for their testing,” he says. “Non-human primates offer a solution, but their use is strictly regulated, costly, unsustainable, and ethically undesirable.”

Supply chain issues are a major hurdle facing early drug development, and Lee describes how challenging backlogs have been seen in the contract business, like contract research organizations (CROs), CDMOs, and API manufacturers. “If your pool of potential service providers shrinks, that’s going to impact the quality and timing of your development program,” he says. “Additionally, there are issues surrounding funding. It’s harder to get capital now, which can add extra pressure when developing new drugs.”

Jones comments that successfully expressing and purifying a potential drug therapeutic can be a challenge in the early drug development process. “Once a viable expression and production system is identified, developing an adequate, (and preferably) robust orthogonal purification process can be challenging and can become the bottleneck in moving a potential drug therapeutic to the next stage of the development process (i.e., for initial stability studies, in-vitro assays, analytics), particularly if expression titers are low, or if the therapeutic is prone to clipping or degradation,” she says.

Interestingly, Broering does not believe that there are specific bottlenecks per se, but that there is a growing demand within industry for early phase studies with a small group of patients as opposed to healthy volunteers. “These early phase patient studies can estimate the pharmacokinetic (PK) levels to see if the drug has the same profile as in healthy subjects and can also inform the design of later phase patient trials,” he comments. “Sponsors are looking for these studies for data comparison to be run under a Phase I setting with very tight timepoint collections which can be challenging.”

Overcoming hurdles in the pipeline

There are several approaches from different areas that the biopharmaceutical industry can take to alleviate the hurdles and avoid future issues that may have hindered companies in the past. According to Russmeyer, health data platforms have the capacity to help solve the “data problem” that drug developers are running into.

“If the right secure platform is deployed, approved persons can access a huge wealth of clinical and non-clinical health data linked at the record-level from a single point of access. This enables researchers to understand complete health profiles and how they relate to the efficacy of new drugs in patient populations over time, so they can more easily measure the impact and monitor safety,” Russmeyer explains. “These platforms’ privacy-preserving technology and high security standards help ensure the safe use of data to drive progress in the industry, ultimately improving outcomes for patients.”

For Kostrzewski, organ-on-a-chip (OOC) technology can help to ease these hurdles by enabling researchers to recreate models of human physiology and disease in the laboratory. “These lab-grown miniature organs and tissues are comprised of physiologically-relevant combinations of primary human cells grown on scaffolds to support 3D growth. They are perfused by media to mimic the blood stream and importantly, can incorporate elements of both the human adaptive and innate immune system,” he explains. “OOC models function and respond to therapeutics in a similar way to humans and deliver clinically relevant end points to facilitate the translation of data between the lab and clinic. It is possible to link models together to simulate processes such as drug absorption and metabolism, or to understand interactions between organs, such as inflammation, which drive disease and cause unexpected toxicities. Importantly, they offer a path forward for the discovery and development of new modality drugs with human specific modes of action in combination with other non-animal-based new alternative methodologies (NAMS) such as in-silico modeling.”

Kostrezewski adds that, going forward, he anticipates that OOC will start to reduce our reliance on costly and expensive animal studies in late discovery early development. “OOC use additionally enables the exploration of a greater experimental space, enabling study design to be fine-tuned to maximize their efficiency and reduce the number of animals used,” he says. “Further along the pipeline within development, OOC can even be used to confirm, or query the results generated in animal studies to overcome concerns about interspecies differences and enable researchers to either proceed into clinical trials with more confidence, or to reengineer drugs to remove any unwanted effects.”

According to Lee, novel excipients are perfect for the 505(b)(2) pathway to reduce both time and cost to approval. “With 505(b)(2) you want to have a differentiated, better dosage form than what was commercially available, and maybe a different route of administration. Novel excipients offer fresh intellectual property and can be offered on an exclusive basis to help protect the asset,” he explains. “Particularly in the early stages of drug development, selecting a CDMO with a wide range of capabilities and expertise can improve the probability and accelerate the process of reaching the target product profile. Flexibility is also important, particularly with high value products like biologics, you need to make sure there are no lower limitations on batch size.”

A variety of chromatography tools are available to assist drug developers in the purification process, Jones mentions. “From 96-well plate format for resin scouting, to process scale prepacked columns, researchers can perform design of experiment [DOE] studies through to process scale-up while eliminating the time spent manually packing columns, along with the need for specialized equipment, specialized training, as well as reducing packing variation, given that prepacked columns are required to adhere to specific and stringent quality controls,” she says (3).

Patient-centric technical development

Patient-centricity has been driving change within the drug development area for some time, and it continues to grow in importance. Technical development activities, like digitization of clinical trials, can help achieve this goal and accelerate the overall drug development process.

Ute Berger, president of development and commercialization solutions at ICON, explains the concept of digitization of clinical trials in more detail, which, if executed in the right way, has the potential to improve patient-centricity of trials and reduce potential failures. “Decentralized clinical trials have been shown to overcome some of the long-term industry challenges in drug development that have had a significant impact in timelines, such as patient recruitment and retention,” she elaborates. “According to Tufts Center for the Study of Drug Development [CSSD] 41% of clinical trials fail to meet planned enrollment estimates and other industry figures suggest 30% patient drop out from trials” (6).

Berger further explains that there are some basic fundamentals in ensuring the technology is patient centric, including: ensuring the study design and technology interaction is developed to reduce rather than increase patient burden; ensuring patients have the option rather than making it mandatory to be monitored remotely, whether that is through digital health technologies or data capture at in-home visits; and most importantly, providing appropriate support both technical and clinical through specialized concierge services. “Concierge services can help with scheduling, welcoming patients, managing consent, responding to patient concerns, collecting reported events or retrieving equipment, all which increase patient satisfaction and lead to higher retention, to avoid delays,” she says.

The most important pillar of patient-centricity for Russmeyer is bringing the voices of real patients into the earliest phases of technical development, from study design to formulation and delivery decisions. “Analyzing data from patient focus groups and surveys will support drug companies to approach the technical development with a better understanding of how to improve patient engagement and participation in clinical trials, and how to improve adherence to and outcomes from new medications,” he says. “In addition, using safely anonymized data united across sources in a way that doesn’t require consent can fold in a patient-centric approach at a much larger, clinically valuable scale.”

Lee mentions that one might think there is enough of a challenge to commercialize a new molecular entity via a 505(b)(1) path so that patients can receive treatment as soon as possible without looking at novel delivery systems. However, new technologies are emerging all the time, with a major focus on patient-centricity upfront. “Drug–device combination products are a great example, for instance in the case of an autoinjector, where you could potentially self-administer at home, saving time by not having to go to the doctor’s office,” he explains. “Many drug developers are looking to new devices to overcome patient challenges of conventional delivery, and in fact we have an established department dedicated to drug-eluting devices such as vaginal rings, intraocular implants, and subcutaneous implants.”

AI to advance processes

More automated solutions, such as AI, are being widely implemented to improve efficiencies. The value of these types of technologies in the biopharmaceutical drug development area has become a tool for collaboration and innovation for many.

Russmeyer explains how AI can be deployed to analyze vast data sets much more quickly than has ever been possible before, which can accelerate drug discovery in several ways—for example, from identifying novel drug targets and potential drug candidates to identifying promising candidates for further development. “AI can also help streamline the drug development process by identifying potential safety issues and predicting how drugs will behave in the body. This can help reduce risks to trial participants, and the number of failed or aborted trials,” he adds. “It can assist in designing and optimizing drugs by predicting their properties, such as solubility, bioavailability, and toxicity. This can help drug developers make more informed decisions about drug design, leading to more effective and safer drugs.”

AI has become more than just an automation solution, in the words of Berger. “AI-powered capabilities, including pattern recognition and evolutionary modeling, are essential to gather, normalize, analyze, and harness that growth and fuel modern therapy development,” she says. “AI has many applications across the full drug development value chain. In pre-clinical R&D, AI is being used to identify novel drug candidates without the need for traditional wet chemistry. In clinical research, AI innovations are driving transformations in clinical trial execution, including improved trial design, the development of novel digital endpoints, and the use of real-world data to identify optimal research sites and patient populations.”

In terms of OOC, Kostrzewski reveals that it is possible to reap immediate benefits by integrating computational in-silico tools. “There are three obvious opportunities, firstly the use of in-silico modeling upfront to inform drug design. By doing so companies can reduce the numbers of molecules that need to be screened. This opens up the opportunity to use more predictive human organ assays, such as OOC, earlier in the development process and speeds up the hit to lead process,” he explains. “The second use of in-silico modeling for [DOE] ahead of OOC to identify the right parameters to test, or to ascertain a concentration range for target engagement in an OOC model that is human-relevant. Following an OOC screening study, in silico physiologically based pharmacokinetic [PBPK] modeling can then be used to extrapolate data to predict the human response for ultimate data translatability. Overall, this combined approach provides significant accuracy and efficiency gains and with addition of AI and machine learning the data generated from many OOC studies could be combined and used to help further predict human responses to different interventions and therapeutic treatment regimens.”

The future of drug development

There are some approaches, solutions, and technologies already available that can help jumpstart a drug developer transition to a more agile and flexible way of working. For example, single use and continuous processing together are increasingly used in manufacturing facilities, which Jones explains enables manufacturing facilities to “pop-up” as needed with very little required infrastructure to house equipment.

“The single use aspect of continuous processing reduces batch-to-batch cross contamination, as well as providing a built-in sterile manufacturing environment. Prepacked columns are a great compliment to this type of manufacturing model, given that they are equipped with the features necessary to plug and play into continuous process systems, and as mentioned previously, provide consistent results and stringent quality controls,” Jones says. “I expect that we will see continued advancements in the use of AI in prediction software geared towards all aspects of the drug development process, including the prediction of purification processes.”

Additionally, more and more developers are reaching out to CDMOs to complement in-house competencies and resources, as it gives them a more flexible approach, remarks Lee. “This is especially true in the early stages of development, where CDMOs with wide-ranging expertise and techniques can enable a higher probability of achieving the target product profile,” he says. “Knowledge and experience with multiple delivery routes can help identify the most effective delivery system, then during scale-up, leveraging a CDMO can reduce risk, expense, and timescales of in-house investment.”

Real-world data has gained significant momentum in drug development, enabling sponsors to harness insights that were previously unavailable to support their product development and positioning. Berger raises the point that tokenization is well established in the commercial setting as a way of accessing multiple data sources for insights.

“However, clinical trial tokenization, which links patient-level data from diverse sources without compromising patient privacy and allowing sponsors to incorporate real-world data into clinical trial analysis and gain a comprehensive picture of the patient journey across the healthcare system and into the follow-up stage, has not yet been widely adopted,” Berger adds. “Both large pharma and biotech companies are exploring, piloting, and evaluating the potential of this solution which combines technology, data, and analytics and delivers valuable insights. If adopted widely, our approach to clinical development could be transformed.”

References

• Conroy, D. Accelerating Drug Development to Keep Pace with Drug Discovery. Forbes Technology Council, Council Post, Mar. 10, 2023.
• Agrawal, G.; Keane H.; Prabhakaran, M.; Steinmann, M. The Pursuit of Excellence in New-Drug Development. McKinsey & Company, Article, Nov. 1, 2019.
• Kronbetter, L.; Khaled, S. (2023) Refining Process Separation Procedures with Prepacked Chromatography Columns. International Labmate, Article, April 12, 2023.
• Baillie, K.; Mueller, T.; Pan, J.; et al. Use of Record Linkage to Evaluate Treatment Outcomes and Trial Eligibility in a Real-World Metastatic Prostate Cancer Population in Scotland. Pharmacoepidemiol Drug Saf., 2020, 29 (6), 653–663.
• Hurtado-Chong, A.; Joeris, A.; Hess, D.; Blauth, M. Improving Site Selection in Clinical Studies: A Standardised, Objective, Multistep Method and First Experience Results. BMJ Open, 2017, 7 (7), e014796.
• Getz, K. Enrollment Performance: Weighing the “Facts.” Applied Clinical Trials, Article, May 1, 2012.

About the Author

Jill Murphy is an Editor for BioPharm International.

Article Details

BioPharm International
Volume 36, No.6
June 2023
Pages 10-14

Citation

When referring to this article, please cite it as Murphy, J. Moving Closer to Dynamic Drug Development. BioPharm International 36 (6) 2023.