Preclinical safety analysis using Artificial Intelligence on the example of Alzheimer’s Disease.

Danai Mavreli
published on September 19, 2020

Introduction

Alzheimer's disease (AD) is a progressive neurodegenerative disorder that currently produces dementia in 5.8 million U.S citizens. This number is projected to reach 13.5 million by 2050 highlighting the urgent need for means to prevent, delay the onset, slow the progression and improve the symptoms of AD (1).

Throughout the years a range of drugs have been investigated as potential treatments for AD such as acetylcholine esterase inhibitors and drugs targeting the amyloid cascade (2). However, these have displayed limited efficacy and adverse effects such as nausea, vomiting, and loss of appetite (3). In 2020, there are 121 unique therapies in clinical trials for AD registered on clinicaltrials.gov (1).

It has been estimated that approximately 80-90% of drug candidates fail before they get tested in humans. The majority of these failures is due to unexpected side effects not predicted in the preclinical stage (4). Therefore, it is essential for researchers to gather as much preliminary information as possible regarding potential side effects of a drug candidate.

Causaly enables researchers to identify safety-relevant information in medical literature regarding a drug candidate. Preclinical experts can include this data in the preclinical study design to minimize the risks of unforeseen toxicities and increase chances of approval.

This blogpost describes the workflow preclinical researchers can follow using Causaly to investigate a potential drug candidate for Alzheimer's disease through targeting alpha 2 adrenergic receptors (α2 ARs), a promising target for AD treatment. We will look at the following 4 key processes:

  • Understand the side effect profile and identify potential off-target effects of α2 AR agonsits
  • Identify the localization profile of α2 ARs in organs and tissues
  • Find side effects reported in animal models
  • Explore drug-drug interactions

Understand the safety profile and identify potential off-target effects of α2 AR agonists

During the preclinical stage of the drug development process, scientists consult the vast amount of published information and databases to obtain as much background information as possible. This is a time-intensive and typically incomplete process due to the high effort associated with reading thousands of documents on α2 AR agonists.

Causaly machine reads over 6,000 articles relevant to α2 AR agonists and extracts 1,400 side effect relationships within seconds gathered from scientific literature, clinical trials and other relevant side effect databases. Results are displayed in an interactive dendrogram. In the case of α2 AR agonists, cerebellar ataxia was identified as a dose-dependent side effect in horses (Figure 1).

blogArtboard-1-copy-4@2x
Figure 1. Overview of side effects associated with α2 AR agonists.

Another major concern during preclinical studies is the assessment of off-target effects. Preclinical toxicities represent more than one third of the causes of drug failure and the majority is attributed to off-target effects (5). Currently, a variety of methods are used to identify off-target interactions including in vivo, in vitro and in silico experiments. However, several limitations exist such as lack of accurate tissue expression information and unreliable computational tools.

The identification of off-target interactions is essential to develop an approach for early prediction of potential adverse events, minimizing the chance of subsequent pitfalls and reducing the cost and time for drug development.

Using Causaly, preclinical experts can find links between a drug candidate and potential off-target side effects. In this case, an indirect connection between the α2 AR agonist Clonidine and heart failure was found, as shown in Figure 2. NPPA gene was identified as a potential off-target effect mediating this side effect (Figure 2).

blogArtboard-1-copy-8@2x
Figure 2. Indirect connections between α2 AR agonists and potential side effects.

Identify the localization profile of α2 ARs in organs/tissues

To further enhance understanding of potential side effects and off-target effects, researchers need to gain an understanding of the localization profile of a target in the human body. In the present case, information regarding localization of α2 ARs is key to anticipate and evaluate potential effects in various organ systems. Using traditional keyword literature search, researchers need to look for the needle in the haystack and read vast amounts of documents to get to these findings.

Causaly gives users quick access to information regarding the presence of α2 ARs in various organ systems. For example a role of these receptors was uncovered in the spinal cord in a rat model treated with intrathecal tramadol (Figure 3). This information can be used to improve preclinical study design and decrease the risk of failure due to unforeseen effects on other organ systems.

blogArtboard-1-copy-5@2x
Figure 3. Localization profile of α2 ARs.

Find side effects reported in animal studies

Another crucial body of information to predict drug safety is evidence found in model organisms. The evaluation of information to predict adverse events and design reliable animal models during the preclinical stage is vital to avoid failure.

Using Causaly, preclinical experts can collect and incorporate this pre-existing information into the design of a drug development strategy during the preclinical stages and address regulatory concerns ahead of time and reduce costs.

For α2 AR agonists, a focused search to identify side effects reported in animal studies revealed depression-like behavior and affected neuron activity in rats receiving clonidine (Figure 4).

blogArtboard-1-copy-6@2x
Figure 4. Side effects of α2 AR agonists in animal models.

Explore drug-drug interactions

Drugs commonly used by the general population can often interact with new candidate substances and cause wide-ranging side effects. Aspirin is one of the most frequently used drugs worldwide. It has been estimated that nearly half of adults aged 70 years and older, approximately 10 million people, take Aspirin daily (5). Alzheimer’s Disease is also more frequently seen in older patients. It is therefore critical to test for any drug-drug interactions with the α2 AR agonist that can either affect its safety or efficacy.

Typically, finding these interactions is difficult for preclinical safety researchers, because more than 20,000 drugs have been approved (6) and searching using traditional literature review methods is very time intensive and error prone.

In the example of the α2 AR agonists, over 600 potential drug-drug interactions were identified. Specifically, Aspirin was found to interact with Clonidine, reducing its efficacy as shown in Figure 5. These insights can form the basis to generate hypotheses and influence the study design.

alpha2-adrenergic-agonists-and-aspirin
Figure 5. Drug-drug interactions between α2 AR agonists and aspirin.

Summary

In the field of preclinical safety, gathering all available information regarding potential side effects of a drug candidate is vital for increased chances of approval for clinical studies and subsequent market launch. In the case of α2 AR agonists, with over 6,000 publications available and safety-relevant information “hidden” in research articles it is important to collect all relevant data to optimize preclinical safety studies. Causaly facilitates this process and enables experts to minimize the risks of unforeseen toxicity effects in the next stages of the drug development process in an efficient and timely manner.

References

1. Cummings J et al. Alzheimer's disease drug development pipeline: 2020. Alzheimers Dement (N Y). 2020;6(1):e12050.
2. Alzheimer’s Foundation. Medications for Memory. https://www.alz.org/alzheimers-dementia/treatments
3. Casey DA et al. Drugs for Alzheimer's disease: are they effective?. P T. 2010;35(4):208-211.
4. Seyhan, A.A. Lost in translation: the valley of death across preclinical and clinical divide – identification of problems and overcoming obstacles. transl med commun 4, 18 (2019).
5. Birkebak J et al. Pharmaceutical industry perspective on combination toxicity studies: Results from an intra-industry survey conducted by IQ DruSafe Leadership Group, Regulatory Toxicology and Pharmacology. 2019;102:40-46.
6. O'Brien CW et al. Prevalence of Aspirin Use for Primary Prevention of Cardiovascular Disease in the United States: Results From the 2017 National Health Interview Survey. Ann Intern Med. 2019;171(8):596-598.
7. Fact Sheet: FDA at a Glance.
https://www.fda.gov/about-fda/fda-basics/fact-sheet-fda-glance

Request access for your organisation

If you are currently working on preclinical safety and would like access for your organisation, visit www.causaly.com or email contact@causaly.com to request a demo or trial.

Causaly vs PubMed®: 2x as many relevant articles identified by Causaly using the same data
use case

Causaly vs PubMed®: 2x as many relevant articles identified by Causaly using the same data

Causaly AI finds more relevant articles than PubMed alone, using its advanced machine-reading technology.

Target identification and validation using AI for literature-based insights: Causaly & Pierre Fabre Joint Webinar

Target identification and validation using AI for literature-based insights: Causaly & Pierre Fabre Joint Webinar

Causaly and Pierre Fabre joint webinar Causaly and Pierre Fabre co-hosted a joint webinar on the 28th of October, addressing how...

Full-text vs Abstract advantage: Causaly identifies 3x as many relevant articles by machine-reading the full-text
Application

Full-text vs Abstract advantage: Causaly identifies 3x as many relevant articles by machine-reading the full-text

Causaly enables regulatory experts to reduce time spent scanning research literature, while at the same time increase the yield from full-text articles which typically are not selected due to unsuspecting abstracts.

Sign up for Causaly newsletter