SESSION III: FROM ANIMAL STUDIES TO HUMAN CLINICAL TRIALS

LM: MJFF supports numerous PhDs and postdoctoral fellows through our existing programs and initiatives. While we do not offer grants to PhD students, our Target Advancement Program permits postdoctoral to seek funding for validating novel targets for Parkinson’s disease.

KO: Carcinogenicity studies are a large time and monetary commitment, thus need serious risk assessment prior to engaging. Phase 3 is also a huge time and monetary commitment, much more so than carcinogenicity studies. Thus, once a company decides to initiate a Phase 3, all final requirements for supporting an NDA are initiated with due risk assessment. If one begins the carcinogenicity study concurrent with Phase 3, usually there are 6 month data available from a chronic rodent study to inform the 2 year bioassay (there is usually a precursor study here as well), including dose level and any potential increases in tumors (even though may not be designed to statistically answer this question). Unable to locate specific examples as requested.

KO: Cost will vary depending on the indication, clinical plans, study design, type of molecule and data needed to ensure patient safety. No affiliation or benefit to me, but there are a few references to indicate some of the costs associated with nonclinical drug development.

https://www.slideshare.net/webgoddesscathy/preclinical-development-planning-for-emerging-pharma-and-biotech-firms

https://www.contractpharma.com/issues/2013-06/view_features/characterizing-the-cost-of-non-clinical-development-activity/

Control animals are always required and necessary to ensure appropriate interpretation. Historical controls are used to judge any potential outliers or endpoints that may seem odd, just to verify if background effect or otherwise. Historical data may be used to support single dose nonGLP MTD studies.

KO: Have worked successfully with large and small CROs, just depends on the needs of the Sponsor and the capabilities of the CRO. Currently have had success with all CROs I have worked with in the past in the US and beyond. Happy to discuss success stories in a more private forum.

KO: Thank you as well. Great question! Yes, the timelines for toxicology are dependent on the clinical trial study design. Thus if there is a chronic risk/delayed toxicity potential from a drug, the Sponsor must produce data addressing that risk, which may entail designing the most appropriate study with agreement from representative Regulatory authorities. Much of this thinking is ongoing with the newer classes of oncolytic therapies and soon to come CRISPR therapies beyond cell therapy modifications.

KO: Thank you. Agreed, small therapeutic windows can be managed by understanding the potential mechanisms of toxicity and adjusting dose levels, route, frequency and duration to optimize the benefit risk profile for patients. Medicinal chemistry may play a role in limiting off-target effects, but decisions should be informed by data.

KO: Not entirely sure what is meant by underestimate. Subjectivity of some of the data may result in qualitative effects in animals that may translate as more or less severe in other living organisms, but this is currently the best we can do. To what I believe is the point, some data generated from toxicology studies is subjective, which is why the laws (primarily GLP) require an appropriate level of education, experience, and expertise for each of the categories of data being generated, in addition to several other aspects of study conduct and quality of data. As a result, there are several experts and expert Teams involved in the interpretation of GLP toxicology data (e.g. Study directors, technicians, veterinarians, clinical and anatomic pathologists, toxicologists). All involved are responsible for safety of the patient and workers.

JG: 1a. Answers.

Great questions.

1) PET ligand discovery and development rely on several factors which might influence the time for fruitful development. On key is the affinity and target specificity of the tracer. If a small molecule and/or antibody is known with the properties to begin with then the problems to be solved emerges to: i) can this molecule be radiolabeled?; ii) does the tracer have suitable in vivo properties?; iii) does the tracer quantitatively biomark the changes in target receptor in both health and the disease of interest.

2) Assuming that a CSF or blood biomarker will report statistically significant changes between a healthy human versus patient, then in my opinion the intial discovery validation might 2-3 years. Importantly, either biomarker will be required to be confirmed in several hundred patients, which will then take another 1-3 years thereafter. Hence, realistically it is my thinking 6-7 years, which is considered a long time; yet, you have to be sure that your biomarker measure are significantly powered for specific patient populations which will take some time. Additionally, there might be some anomalies that might surface with the data; therefore, you need to be cautious and provide ample time to address these if they emerge.

3) Regarding a companion biomarker, I think that the effort should coincide and run parallel over time with the development of a plausible and emergent therapeutic agent. Since establishing and confirming a biomarker will take a few years, which is also be the case for a therapeutic agent, then the sooner you begin to establish a companion biomarker the better. I think the risk here is whether your initial selection of biomarker(s) is suitable to report to the pharmacodynamic effect(s) of the therapeutic as development advances. It might be the case. Or, as you learn more about how your therapeutic effects its actions, it could be that you need to re-think your biomarker entity (i.e., go back and start from beginning to establish a more meaningful and statistically relevant biomarker), which might be needed; if so, this most likely will require more time. Taken together, it is my belief that the best thinking on the biomarker from the outset will serve best in the long run as understanding of the therapy mode of action and outcomes become more defined.

JG: Good question. I do not think digital biomarkers will replace fluid biomarkers. I think that fluid biomarkers reflect measurements of actual pharmacological and/protein levels of interests, whereas digital biomarkers determine other measures (albeit that could include glucose level monitoring and behavioral determinations, etc.). Digital biomarkers are terrific for behavioral-lifestyle measures, which are distinct from fluid measures; yet both might influence one another. It is my thinking that both digital and fluid biomarkers together will report a more refined picture of disease state with greater resolution versus either digital or fluid measures alone. In essence, determined both. The more data, the greater resolution you will have on the state of the patient.

JG: Important question. I think that the better characterized the AD patient is the better the prognosis and/or timeline for understanding changes to the patient might be. For example, let’s say the AD patient has a magnetic resonance (MR) brain assessment and a-beta and phosphorylated-Tau PET imaging scans, adjunct to clinical memory rating, among other measures determined by the physician. In that case, the patient might be defined very well, and be suitable for clinical trial enrollment. The key I think is to design your clinical trial with suitable inclusion versus exclusion criteria, relative to the biomarker measures of the potential patient population. Importantly, there have been a number of studies where for example, multiple (e.g., 3 or more) biomarker measures have been defined for patients, to define and understand where they are on the continuum of neurodegeneration progression – like Mild Cognitive Impairment (MCI) progressing to Alzheimer’s disease (AD). In essence, I think that at last 2 biomarker measures coupled with a clinical assessment are necessary. The more measures that are in hand, the more defined the patient will be, enabling patient cohort stratification for special clinical trail enrollment; hence, resulting in more insightful understanding of disease progression and a realistic patient time line outcome.

JG: This is a great question. I believe the field is progressing as a function of the disease of interest, as different sets of biomarkers might be needed as a function of the disease. Let’s consider Alzheimer’s disease (AD). In this area there have been many significant strides over the past two decades for biomarkers, including quantitative PET imaging definitions of regional cerebral a-beta protein brain load versus clinical symptomology. Further, a similar course has recently developed for phosphorylated tau regional cerebral brain load. Yet, the question of causation of AD remains an open question and biomarking early to later AD progression is a current question of interest. Hence, taken together, I think that several new biomarkers are needed to address: i) AD causation; and ii) early brain pathological changes that elicit AD and related dementias. Biomarkers for causation are less well characterized presently; albeit, a-beta and tau are

pathology players- hence, I think we will have evaluate newly emergent hypotheses about causation in order to consider suitable biomarkers that might take a significant amount of time to validate, if in fact causation is multi-factorial and represents in the clinic decades later after initial manifestation. ii) For early pathological changes, I think several different types of evaluation of regional cerebral alterations to the tri-partite synapse environment might insightfully inform about early manifestations resulting in brain changes. Under the current assumption that early therapy to early brain changes heading towards AD might be important, then I think that biomarking these types of events are plausible and would be meaningful to the field of understanding AD progression. For example, if the tri-partite synapse is composed of pre- and post-synaptic neurons and also astrocytes cells, then I think all aspects of these dynamic cell participants are plausible for biomarker development. In essence, the sky is really the limit here for targets for biomarkers. I wish I could be more specific to answer your question. One area of special interest that I think will report with significant measures are changes to astrocyte cell targets, as astrocytes support the pre- and post-synaptic environment, play key roles in the neuro-vascular unit, interact in the neuroinflammation microglia-astrocyte cell dynamic, and regulate key calcium ion flux and concentrations, among other roles in tri-partite synapse homeostasis continuum. In essence, the current literature suggests many other targets to investigate for early AD. The questions are: i) which ones are the important ones; ii) how much time is given to establish these new biomarkers; and iii) are there sufficient resources (patient samples, money, manpower, etc.) to discover these new biomarkers associated with AD causation and/or early to later temporal pathological AD changes. These same argument hold for dieases other than AD.

JG: This is a timely question. I think that there are several blood measures that are suitable for reporting on AD. Yet, the caveat is that the blood measures alone might not be sufficient. Therefore, I think you will need several other biomarker determinations adjunct with the blood measures. Those other biomarker measures might include: i) clinical rated memory assessments; ii) Magnetic Resonance (MR) brain scans; iii) a-beta and/or phosphorylated tau brain PET imaging scans; and iv) perhaps one or two digital biomarker measures. Thus, it is my thinking that a blood test alone will not sufficiently report on AD (it might suggest AD), as I think it will be best to have multiple biomarker measures coupled with clinical assessments to make a more definitive diagnosis of AD.

ADF: The choice to go to Australia is complex. It is truly a balance between time, cost and operational complexity. The start up in Australia is simpler than the US and often faster. The submission to the IRB includes the protocol and IB and we have an Australian toxicologist review our package and submit a report  (this is not required but often recommended). There are also some cost advantages, as we have an office there. That being said, operations can be challenging. They do speak English, which is helpful when leaving the US. The state of medical care is also similar to the US. I would recommend that you fly to meet the study sites and CRO in person and continue with these visits, approximately annually. We have formed some very good relationships. The relationships include, sites, drug depot and labs for safety and PK.

ADF: I think that this may be a good idea.  There are some considerations, are you in the population that you are going to study, and if not will healthies respond to the biomarker in a manner sufficiently indicative.  Also, your numbers will be small so don’t over or under read into them.  That being said, it may give you an early indication, which will not only help you design future studies but also may give you additional confidence to spend.  Finally, the operations around some biomarkers may be challenging and it is an opportunity to start working with these logistics before your numbers are too large.  Things like videos for processing and clear shipping instructions if not put carefully in place can result in missing samples or samples that cannot be analyzed.  This may lead to ambiguous data and a great deal of cost.

ADF: We debate this quite a bit internally.  With the small numbers, it is difficult to truly interpret a signal.  In my opinion, the benefit of utilizing placebo vs historical control is of little value. That being said , safety data during this phase of study is often difficult to interpret and you may set yourself up for a regrets analysis if there is no placebo. I would also weigh the expectation of the community. If in the therapeutic area you are studying placebo is often used, you may want to continue with the use of placebo. You will most likely have your data reviewed with KOLs and you would like the focus to be on the data and not on the use of placebo.

ADF: I would say it depends.  It will take you more time and cost as your screen fail rate may be very high. It will also depend on the manner in which you want to use the biomarker in the future – will you want this data to be generalized to the broader population or is your target only those patients with the biomarker. In early phase 2 trials this type of enrollment, regardless of your future plans, may give you the confidence to move into later phases, vs the alternative where a population without the biomarker may not produce the specificity that you require to move forward. 

ADF: In my opinion, this depends on the phase of study and the use of the data. During phase 2, post hoc analyses, are often performed as this is truly your exploratory phase. This is the period of study when you have good numbers and you are best able to truly understand your data. These post hoc analyses are documented in your final report as are your deviations from your statistical analysis plan. Should you want to submit this Phase 2 as a pivotal or registration trial, you may have challenges with the agency, as you will most likely lose some trial integrity. During phase 3, especially if you have a SPA or the like, this would not be advised.  From a regulatory perspective, it would be most unusual to deviate from your statistical plan without losing integrity. 

ADF: In my opinion, I would include these early. I realize that you may not have the time or numbers to get a “p value”, however, you will most likely see trends and gain experience with these endpoints. This experience has two forms. The first, the response/ability of the measure to reflect the therapeutic effect of your drug, the second is operational.  Should you have measures on a diary, in the clinic? The operations of these measures, when we deviate from in clinic paper assessments, becomes complex and I would recommend that you gain experience with vendors such as eDiary or the like as early on as possible.

I apologize, but I don’t feel that I am familiar enough with the data to really present a valuable opinion.