Noninvasive Imaging in Drug Discovery and Development. Where to Begin

by J. Paul Shea, PhD
Director, Molecular Imaging
MPI Research

There has been much publicity and discussion concerning the role of noninvasive imaging techniques in drug discovery and development. I, like many, believe these are process transforming tools that have the potential to fundamentally alter our approach to the challenges we face. As I travel and discuss these technologies with my colleagues in therapeutics discovery and development, a common set of questions emerges. What will be the long term utility of these technologies in therapeutics development? In which technology do I make my initial investment? Where in the discovery - development timeline do these technologies belong? What are the current uses of these technologies in the field?

What will be the long term utility of these technologies in therapeutics development?

Over the past few decades, multiple new technologies have been introduced into the field of drug discovery and development. These have included computer aided drug design, combinatorial chemistry, high throughput screening, high sensitivity analytical tools and a whole host of -omics. In each case, their introductions began a sequence of euphoria, predictions of revolutionary impact, dismay and doubt and, finally, a proper role in the process. It is a logical progression, given that these tools originated in university settings, where initial objectives were more pure than applied science. Accuracy, reproducibility, ruggedness and applicability to the drug development process were initially unanswered questions. Is non-invasive imaging a different story? Consider that anatomical and functional imaging technologies have seen clinical use for decades. Indeed, anatomical technologies such as x-ray, CT (x-ray computed tomography), ultrasound and MRI have become diagnostic staples of current clinical practice. Likewise, functional imaging modalities have found broad clinical use in cardiac function testing and in oncologic disease diagnosis and staging. These decades of experience have progressed an understanding of the utility of these methodologies to clinical assessment, disease progression monitoring and treatment response. Advances in clinical understanding have been accompanied by technological advances in instrument performance and image reconstruction software, resulting in improved image quality, visibility and interpretation. Noninvasive imaging is a mature technology with proven clinical applicability. What is new is the ability to extend this technology to in vivo specimens as small as a mouse, with species scaled sensitivity and resolution comparable to the clinical setting.

The obvious opportunity, then, is the direct translational applicability of these technologies. With the existing clinical infrastructure, it can be a straightforward process to introduce imaging endpoints early in a clinical development plan. From this translational perspective, one can envision the preclinical setting as ideal for establishing the relationship between disease, treatment and imaging endpoint. What technology/biomarker to scan with, what parameters to measure, the optimal scanning protocol, etc. can all be addressed in a well designed preclinical imaging plan. Proper late stage use of imaging endpoints in the preclinical environment not only provides critical clinical planning data, but also firmly establishes a link between the preclinical safety and efficacy database and early clinical data.

The less obvious role, then, becomes imaging’s involvement in the discovery process. While perhaps less obvious, it is also potentially the point of greatest impact on any drug discovery program, that is establishing the validity and relevance of novel biological targets. As discussed below, throughput issues prevent imaging from contributing to large scale compound screening, but do not hinder its ability to probe early biological data. Drug discovery programs, particularly those aimed at unique, novel targets, proceed with a hypothesis of the relationship between target, target modulation and impact on disease. Since functional imaging presents the opportunity to measure physiological processes in the intact, living animal, it can clearly play a role in demonstrating target validity and cross species comparability. Molecular imaging endpoints focused on measuring target modulation effects, i.e. drugability, can contribute greatly to ensuring proper investments in unproven targets.

So, while noninvasive imaging has a clear role in supporting translational research efforts, it also represents a uniquely powerful method of establishing the validity of novel therapeutic intervention. As ongoing clinical research further establishes imaging’s translational role, it will take more time, effort and, most importantly, data to establish imaging’s final utility in the overall process.

In which technology do I make my initial investment?

The commitment to creating an in-house Imaging Center is a large one, particularly if it is to be a multi-modality facility. The power of multimodality imaging makes it difficult to envision a single technology laboratory. In addition to dedicated laboratory space, there are significant instrumentation costs and skilled personnel to hire. If functional imaging is an objective, there will be significant investments in radiochemistry facilities, trained radiochemists and isotope generation. The starting point must be potential applications in your organization’s areas of interest, as early proof of value will support future growth and investment.

Choices for anatomical imaging modalities include CT, MRI and ultrasound. While ultrasound is the low cost entry point, its applicability is limited by depth of penetration, image quality and no opportunity for image fusion with functional imaging platforms. CT represents the intermediate cost option and can visualize bone and, with contrast agent, soft tissue. CT is also the most common anatomical reference data fused with functional data from PET or SPECT instrumentation. Indeed, multiple manufacturers are now offering a single platform capable of acquiring PET/CT or SPECT/CT images. MRI represents the high cost option, offering the highest special resolution and best soft tissue imaging. There is currently no available instrumentation capable of simultaneous MRI/CT or MRI/SPECT images.

Functional imaging modalities include optical (bioluminescence and fluorescence), PET (Positron Emission Tomography) and SPECT (Single Photon Emission Computed Tomography). Optical imaging represents the low cost option and is capable of efficient, high throughput functional assays in rodents. Optical technology is an excellent introduction into the conceptual arena of molecular imaging, but is limited by depth of penetration, 2D images and a lack of applicability in translational studies. The nuclear methodologies PET and SPECT represent high resolution, high sensitivity functional methodologies well suited for translational applications from mouse to man. PET and SPECT each possess relative strengths and weaknesses. PET isotope and tracer availability is more limited than in SPECT, but PET is essential for imaging small molecule drugs, due to its available isotopes. SPECT, on the other hand offers easier access to longer lived isotopes which are well suited for labeling biologics (peptides and antibodies).

For an introduction into the power of functional imaging, optical imaging would be hard to argue against. For the most general range of applications, CT combined with PET or SPECT would be the most appropriate, with PET offering small molecule studies and SPECT more applicable in a biologics oriented laboratory. For the uncertain, the anxious or the under funded, commercial and university outsourcing opportunities exist that would allow the incorporation of imaging endpoints into any preclinical development program.

Where in the discovery - development timeline do these technologies belong?

The discovery – development timeline has very different requirements as the compound selection process progresses. In early lead discovery, the number of compounds screened can be on the order of 105 or 106. While high throughput screening is impedance matched to these numbers, imaging technology is not. Animal handling, image collection times and image processing all contribute to the limited throughput of imaging technology. Imaging is not an appropriate general screening tool. It does, however, represent an ideal platform for early systems biology measurements. A drug discovery program is a triad of therapeutic, disease and target. Novel targets resulting from genomic or proteomic studies often have an uncertain relationship to a specific disease. Additionally, a perfectly acceptable target from a biological perspective is of little value if it is not drugable, that is, if its action cannot be modulated by an external therapeutic. Imaging is an ideal platform to assess these target biology questions. By focusing functional imaging endpoints on presumed target effects, the validity of the target can be probed in normal and diseased models, across species and, potentially, in early clinical trials. The longitudinal, whole system datasets generated by noninvasive imaging provides a platform for investigating the time course relationship between target, disease state and therapeutic. Such data can be extremely valuable in interpreting early clinical results.

On the other end of the timeline, the translational potential of imaging technologies is evident. As development programs mature, key questions of pharmacodynamics, efficacy and toxicity often emerge. These key questions and the data generated to answer them often determine the success or failure of the program. Is the dose-response curve in man similar to that in the preclinical model species? What is the dose relationship between preclinical models and clinical disease state? What stage of disease, if any, is optimal for therapeutic intervention? Is a side effect as likely to be seen in man as in the toxicology species? What are my clinical patient selection criteria? How uniform is my disease population? A well designed preclinical imaging program has the potential to establish quantifiable endpoints upon which to answer these types of translational questions.

Early or late in the process, noninvasive imaging has the potential to contribute uniquely and concretely to the therapeutics discovery process.

What are the current uses of these technologies in the therapeutics discovery field?

Noninvasive imaging techniques are seeing increasing utilization in the therapeutics discovery field. Interest in the technology has been spurred by the FDA Critical Path initiative and Exploratory IND guidelines. Large Pharma is making investments in preclinical and clinical imaging centers to facilitate the inclusion of imaging endpoints throughout the development process.

In oncology research, preclinical imaging is being used both for early efficacy verification and for later stage clinical planning. Using primarily functional imaging (optical, PET), researchers are establishing noninvasive imaging endpoints as measures of efficacy in primary tumor treatment and metastatic disease. In our laboratory, we have used PET imaging in mice to assist clinicians in determining the optimal radiotracer and optimal imaging protocol for use in early clinical efficacy trials.

Anatomical methodologies (CT, MRI) are being used to assess efficacy in bone diseases and other anatomically defined diseases. CT, in particular, has been used to assess compound efficacy in osteoarthritis, osteoporosis, and bone healing. In our laboratory, we are investigating the use of CT scanning as a tool for fetal skeletal anaysis.

Compound distribution and pharmacokinetics are areas of concentration for nuclear methodologies. These methods provide unique, continuous distribution datasets that can be extended into the clinic. Using appropriately labeled compounds, targeted antibody therapies can be rapidly assessed in preclinical and clinical settings. Labeled small molecules can provide early clinical distribution data to support compound selection, i.e. Phase 0 studies in man. Labeled small molecules are also seeing wide application in CNS research, where brain receptor occupancy studies are being used to establish clinical dosing regimens. In our laboratory, we have used PET distribution studies to make a rapid assessment of compound behavior in the rat. These data provide a time and cost-effective means to establish key timepoints and organs for more traditional and/or detailed distribution studies.

Summary

Noninvasive imaging technologies are seeing increasing utilization in the drug discovery and development process. Across therapeutic areas, across program timelines, imaging endpoints are showing promise as quantifiable measures of compound efficacy and disease response to treatment. Access to these technologies is increasing, through the establishment of in-house imaging resources and CRO’s. Choose a challenge, select a technology and start collecting data.

J. Paul Shea is the Director of Molecular Imaging Services at MPI Research, located in Mattawan, Michigan. After receiving his PhD from the University of Washington in medicinal chemistry in 1983, Dr. Shea completed a postdoctoral fellowship at Vanderbilt University School of Medicine’s Center for Molecular Toxicology. His industry experience as a key decision-maker includes big pharma (Bristol Myers Squibb), radiopharmaceuticals (PETNET Pharmaceuticals), drug discovery (Molecumetics, Ltd.) and consulting (Arthur D. Little, Inc.). He is also the founder of Luminare Molecular, LLC, a company that provides molecular imaging services to pharmaceutical and biotech drug development programs. Dr. Shea is the Summit Facilitator and a Session Chair for the upcoming scientific meeting, Advances in Therapeutic Discovery and Drug Development: Molecular Imaging in Translational Research, which is being sponsored by MPI Research in collaboration with the Society of Non-Invasive Imaging in Drug Development (SNIDD) and the Academy of Molecular Imaging (AMI).