Improving Vaccine and Drug Effectiveness with Artificial Intelligence

Share This Post

In the wake of the COVID-19 pandemic, vaccines and, perhaps more importantly, their effectiveness, has been on virtually everyone’s minds. Amidst the public discourse over the COVID-19 vaccine, though, much of the science around the medical efficacy of these pharmaceutical products have been lost. Despite this, the medical professionals and research experts that strive to improve vaccine and drug efficacy every day have remained hard at work.

This article explains some of the core principles that go into ensuring a pharmaceutical product is indeed medically efficacious, from drug discovery to clinical trials to U.S. Food and Drug Administration (FDA) approval. 

 

Drug efficacy and how it relates to solubility

 

When it comes to the development of a new drug formulation, be that for a vaccine or a medication, a major consideration is protein solubility. Unfortunately, optimizing both can be a big challenge that takes a considerable amount of time and money. 

Dr. Larry DeLucas, president of Soluble Biotech and SVP of Operations at Predictive Oncology (POAI) explained how solubility and stability impact drug and vaccine efficacy.

“When you inject something like the COVID-19 vaccine into someone, you need so many milligrams of proteins injected in a small volume for it to work,” DeLucas said. “You need to concentrate the protein while maintaining its solubility. The protein is very highly concentrated as a result.”

The problem is that highly concentrated proteins or peptides tend to stick to themselves in the vaccine solution. In big clusters like this, it is unlikely that the intended immune system response will occur, reducing the effectiveness of a vaccine or rendering it altogether worthless.

 

Excipients and finding the perfect combinations for solubility

 

To solve this problem, manufacturers turn to additives called excipients, which can improve the solubility of proteins and peptides and keep them from aggregating. Excipients, which could be sugars, amino acids, or polymers, help ensure the solution for a drug or vaccine remains homogeneous and bioavailable, the key to ensuring its effectiveness. 

Unfortunately, there are millions of combinations of excipients to choose from, and every excipient could be introduced in varying concentrations – that makes it nearly impossible to find the best possible combination of excipients. Historically, researchers have had to engage in a timely and costly process of trial and error, examining each excipient in different concentrations and combinations within their pharmaceutical formulation. Naturally, there are only so many they can study before taking their best-educated guess, a far from optimal way to develop drugs.

Moreover, this trial and error process brings with it another major issue: the protein needed for most of these drug formulations is expensive and hard to come by. This makes finding the best combination of excipients and their optimal concentrations extremely costly, requiring companies to dedicate grams of their precious proteins to find the best possible formulation – and that’s before even producing enough to begin pre-clinical trials.

“Companies figure out these combinations by trial and error, but there are literally millions of possible combinations to choose from,” DeLucas said. “It’s time-consuming and expensive. The amount of protein they need for each formulation could be 0.5 grams or more.”

 

How stability can impact FDA clinical trials and approval

 

Adding so many excipients to a drug formulation can affect the stability of the vaccine or drug, a critical aspect of passing regulatory muster. As part of FDA clinical trials, the agency examines how well the vaccine or drug ages to determine its shelf-life and assign it an expiration date that ensures the safety of patients.

“The FDA typically requires over a year of time showing stability at a specified temperature,” DeLucas said. “Part of FDA approval means we understand the right conditions that the drug formulation [the pharmaceutical product] must be kept in until used.”

Once testing is complete, the FDA evaluates the stability data to determine the expiration of a pharmaceutical product. This evaluation includes the formulation’s exposure to various temperature levels, humidity, oxygen, and light, to name a few. The intent is to ensure that patients are administered only medications or vaccines that remain effective and safe even after spending time in transit or storage. It also helps determine optimal storage conditions, such as keeping medications in a cool, dark, dry place, for example. 

Improving the solubility and stability of vaccines and pharmaceutical drugs offers a better chance at a prolonged shelf-life and FDA approval.

“You’ve got to figure out the right combination that keeps it soluble and makes it stable,” DeLucas emphasized. “You can’t have it breaking down quickly.”

Unfortunately, making a formulation more soluble can disrupt its stability. Without guaranteeing both solubility and stability, a drug will not pass FDA approval for clinical trials, which could mean wasted effort even if a company develops an extremely soluble formulation.


How Predictive Oncology is improving drug discovery and development

According to a 2018 study published in the Journal of the American Medical Association (JAMA), the median cost of developing a new drug from research and development through FDA approval to market sale ranged from $765.9 million for nervous system agents to $2.77 billion for antineoplastic and immunomodulating agents. Those costs are even more astronomical when you consider the time, money, and resources poured into drugs that never attain FDA approval.
Obviously, the cost of this process can be immense, but the market value is as well – according to the International Federation of Pharmaceutical Manufacturers and Associations (IFPMA), the combined direct, indirect, and induced effects of the biopharmaceutical industry’s total contribution to the world’s GDP is $1,838 billion – just shy of $2 trillion. 

That’s why companies have tolerated the high-risk investment in drug discovery and development, but today that risk is no longer necessary. That’s because DeLucas and his team are laser-focused on speeding up the time it takes to find viable formulations of drugs that are not only highly soluble but also stable. About 30% of their time is spent researching more effective cancer drug formulations, DeLucas said, but their work has widespread applications in other arenas as well, including vaccination.

“We take a process that usually takes a year and a gram of protein and do it in 3 to 4 months with milligrams of protein,” DeLucas said. 

Through this process Soluble Biotech and POAI can save pharmaceutical companies millions in wasted efforts while helping them quickly identify potentially life-saving vaccines drugs and get them to FDA clinical trials much more quickly, improving their odds of ultimately attaining approval and reaching the market. 

To do this, the team works closely with an advanced neural network and employs innovative approaches in excipient testing in the laboratory. 

 

Self Interaction Chromatography

 

The first piece of the puzzle is determining the potential solubility of a protein in combination with certain excipients in varying concentrations. The method Soluble Biotech employs is called “Self Interaction Chromatography”, and it is designed to minimize wasted protein in the drug discovery process.

“We miniaturized a process known as Self Interaction Chromatography [to determine formulations that optimize the solubility for each protein or peptide therapeutic],” DeLucas said. 

Self Interaction Chromatography involves binding a protein to a chromatography column and then injecting the same protein over the column. The bound protein is randomly bound to the chromatography column thereby ensuring that every orientation of the protein is exposed, as it would be in a solution. 

The speed at which the protein moves through the column tells analysts whether the protein is sticking to or being repelled by the bound protein. Proteins that repel like proteins are less likely to cluster and therefore will be more soluble at higher concentrations. 

Calculating the second virial coefficient

The volume of solution required to run the protein through the column is input into an equation that yields a number that represents a thermodynamic term known as the second virial coefficient. The second virial coefficient represents the sum of all interactions on the surface of a peptide or protein (i.e., charge, dipole, hydrophobic, and even van der Waals interactions). It is considered the gold standard in measuring protein-protein interactions. 

A positive second virial coefficient value indicates net repulsion, or improved solubility, whereas negative values indicate net attraction and a poorer solubility. Using Self Interaction Chromatography is a more efficient method of computing the second virial coefficient compared to the traditional technique, which is known as Static Laser Light Scattering. This method is time consuming, requires a significant amount of protein, and is fraught with experimental challenges.   

Soluble Biotech’s fully automated Self Interaction Chromatography system yields a second virial coefficient value every 30 minutes. Determination of the optimized, final formulation only requires about 60 to 70 milligrams of protein, DeLucas said, whereas other processes use between 0.5 and 1 gram of protein to develop the final formulation. 

While that may not sound like a lot, proteins used in drug and vaccine formulations can be incredibly expensive and difficult to come by – so this step offers companies major savings when it comes to drug discovery.

 

Using artificial intelligence to boost solubility

 

DeLucas and his team are able to take the results of the Self Interaction Chromatography test and plug them into a formula to calculate the 2nd Virial Coefficient, a measure of a protein’s solubility. With this knowledge in hand, the team turns to a neural network to determine the ideal blend of excipients 

After identifying the optimal excipients and their concentrations using the Self Interaction Chromatography system, the solution excipient conditions and measured solubility data are automatically input into a neural network which operates using a process known as “back propagation.”

In back propagation, the neural network takes about 5% of the existing data and sets it aside without analyzing. Then, leveraging the remaining data, it weights the different excipients and their concentrations and rapidly runs a series of refinements until it can accurately predict the data that is set aside. Over time, the algorithm learns and improves – once it reaches a point where the accuracy of its predictions can no longer improve, it then predicts 4,000 new excipient combinations and concentrations to determine the most soluble formulations for the vaccine or drug. 

DeLucas and his team take the top 25 recommendations the neural network returns and test them for solubility and stability. If the formulations meet the team’s expectations and correlate well to the AI’s predictions, they can move forward into development and pre-clinical trials – no trial and error needed.

This offers pharmaceutical companies a major advantage over those who employ the conventional methods of trial and error with excipients, only to find out stability has been irreparably impacted by the time their formulation reaches pre-clinical trials. By then, immense amounts of time and money have been wasted. But with POAI’s model, companies can cut the waste and get straight to what works for both solubility and stability.

Want to learn more about how artificial intelligence supports Predictive Oncology’s mission of eliminating cancer? Read more on CoRE and PeDAL, our proprietary AI algorithms.

 

Beating the odds with Predictive Oncology and Soluble Biotech

 

Pharmaceutical companies have long been playing on the same playing field when it comes to drug discovery and development – and that playing field is rife with pitfalls and missteps. Fortunately, Predictive Oncology brings with it the capability to streamline drug discovery and development efforts by leveraging AI to more easily identify and optimize drug formulations from start to finish – that includes improving the odds they make it successfully through FDA clinical trials.

Related Content:

Lawrence J. DeLucas, Ph.D

SVP, OPERATIONS,
Predictive Oncology
President, Soluble Biotech
At Predictive Oncology

Dr. DeLucas is the Vice president of Operations for Predictive Oncology and President and co-founder of Soluble Biotech, Inc. DeLucas is currently working to complete development of GMP facilities at Soluble Biotech and at TumorGenesis. In addition, he oversees Soluble Biotech’s solubility and stability contracts for numerous pharmaceutical/biotech companies.

Before Predictive Oncology

From 1981-2016 Dr. DeLucas was a faculty member at the University of Alabama at Birmingham (UAB) where he served as a Professor in the School of Optometry, Senior Scientist and Director of the Comprehensive Cancer Center X-ray Shared Facility, and Director of the Center for Structural Biology. Dr. DeLucas received five degrees from UAB culminating in a Doctor of Optometry degree and a Ph.D. degree in Biochemistry. He also received honorary Doctor of Science degrees from The Ohio State University, Ferris State University, the State University of New York (SUNY), and the Illinois College of Optometry. He has published 164 peer-reviewed research articles in various scientific journals, co-authored and edited several books on protein crystal growth and membrane proteins and is a co-inventor on 43 patents involving protein crystal growth, novel biotechnologies and structure-based drug design. DeLucas was a payload specialist NASA astronaut and member of the 7-person crew of Space Shuttle Columbia for Mission “STS-50”, called the United States Microgravity Laboratory-1 (USML-1) Spacelab mission. Columbia launched on June 25, 1992, returning on July 9.  In 1994 and 1995, Dr. DeLucas served as the Chief Scientist for the International Space Station at NASA Headquarters in Washington, D.C. In 1999, Dr. DeLucas was recognized as one of the scientists who could shape the 21st century in an article published by “The Sunday Times” of London titled “The Brains Behind the 21st Century.”  In 2004, he was recognized as a Top Ten Finalist for the Entrepreneur of the Year award from the Birmingham Business Journal. 

“ Soluble Biotech is continually demonstrating to pharmaceutical and biotech companies the significant value of its novel HSC technology for optimizing protein therapeutic formulations to treat a variety of chronic and infectious diseases. ”

Education
  • Five degrees from Univ. of Alabama at Birmingham (UAB): B.S. Chemistry, M.S. Chemistry, B.S. Physiological Optics, O.D. Optometry, Ph.D Biochemistry

  • Published 164 peer-reviewed research articles in various scientific journals

  • 1993-2016: Director of the UAB Comprehensive Cancer Center X-ray Shared Facility, and Director of the Center for Structural Biology

  • NASA Astronaut, flew on Columbia Space Shuttle

  • 1994-1995: Appointed Chief Scientist for the International Space Station at NASA HQ
  • Arlette Uihlein, MD, FCAP, FASCP

    Dr. Arlette Uihlein is Senior Vice President of Regulatory Affairs and Quality for Predictive Oncology and Site Leader of Helomics, serving as the Vice President of Operations, Pathology Services and Medical Director of Helomics® Clinical and Research Labs since 2011. Dr. Uihlein is Board Certified in Anatomic and Clinical Pathology, Cytopathology and Family Medicine. Dr. Uihlein completed her Pathology Residency at Allegheny General Hospital, where she served as Chief Resident in Pathology and completed Fellowships in Cytopathology and Surgical Pathology. During that time, she conducted extensive clinical research involving molecular pathology diagnostic and predictive markers, imaging of solid tumors, and novel applications of cellular tumor markers. While serving as Medical Director at Helomics, a CLIA and New York State certified lab, Dr. Uihlein has published research in molecular assay development, lab automation, and tissue and cell processing. She is a Designated Civil Surgeon for the U.S. Dept. of Justice and a certified Medical Review Officer for the Department of Transportation. She is a Fellow of the College of American Pathologists and the American Society of Clinical Pathology, NYSDOH Certificate Qualified, and a member of ASCO.

    “ At Helomics we’re delivering better-informed decision making saving pharma time and money, while providing cancer patients with appropriate therapies.”

     

     

    Education

    Medical College of Ohio
    Doctor of Medicine

    Baldwin-Wallace University
    BS, Biology

    Richard Gabriel, BS, MBA

    SVP, RESEARCH & DEVELOPMENT
    Predictive Oncology
    Site Leader, TumorGenesis
    At Predictive Oncology
    My role at Predictive Oncology is to bring the business sense to managing Research and Development programs at all our companies. To seek new ways and opportunities to commercialize exciting new technologies that we have built, licensed, acquired, or are developing through our own research and development. The success of any company is to get the research off the bench and to the customers. That is what I do at POAI and help the other companies as well.
    Before Predictive Oncology
    Prior to starting his first company in 1984 and registering with the FDA a pilot plant facility to make pharmaceutical actives, Mr. Gabriel managed a $50 million product line for W.R. Grace, developed new marketing and sales strategies for Ventron a Division of Morton Thiokol, research work at Ashland Chemical for pressure sensitive adhesives and plant scale-up. Since then, he ran a genetics company, built three GMP/Research facilities, and helped 5 drugs reach their markets in AIDS and cancer. Real expertise in cGMP process scale-up and compliance. Completely understand the needs of an API manufacturing facility and build processes that are scalable, environmentally acceptable, and safe. 3 FDA inspections with no 483’s, ISO certification, DEA registration, DoD compliance, NCI contractor and inventor. Has also broad-based experience in start-up companies and how to make them operational and profitable. 7 years of Team set-up, R&D management, and implementation for 165-person (85 PhD’s and Engineers) company (Pharm-Eco) and lecturer on cGMP and Teams within the Pharmaceutical Industry.

    “ Patients are always first, is our driving force. Oncology is a tough space, and we are determined to bring the best validated science to help cancer patients and as our CEO says, ‘Eliminate Cancer.’ That takes teamwork and a lot of smart hard-working people, our team members at POAI are up to the challenge. ”

     

     

    Education
    Suffolk University
    Executive MBA Program

    Ohio Dominican College
    BS, Chemistry

    Ohio State University
    Microbiology and Virology

    University of Cincinnati
    Associates Degree, Liberal Arts

    Mark A Collins, Ph.D

    SCIENTIFIC/CUSTOMER DEVELOPMENT
    Predictive Oncology
    Chief Technical Officer, Helomics
    At Predictive Oncology

    Mark is currently Chief Technical Officer at Predictive Oncology. Using the power of AI, Mark is responsible for leveraging Helomics’ vast repository of physical and digital tumor samples, to build multi-omic predictive models of tumor drug response and outcome. Such models can be applied to the discovery of new targeted therapies for cancer as well as used in clinical decision support to help oncologists individualize  treatment.

    Before Predictive Oncology

    Dr. Mark Collins embarked on a career in the pharmaceutical industry following his postdoctoral work. Pursuing a passion for both biology and computing, Mark has held multiple executive roles in a variety of discovery, informatics and bioinformatics functions within global pharma, and founded three startup software companies in the artificial intelligence (AI) machine learning (ML) and drug discovery space. Mark relocated to the USA in 2001 to work for Cellomics (now part of Thermo Fisher Scientific), where he played a pivotal role in establishing the High-Content cell analysis market, building, and commercializing several key informatics and bioinformatics products.

    Since leaving Thermo Fisher, Mark has focused on developing and commercializing informatics solutions for clinical and translational research, specifically in the specimen tracking, ‘omics data management and NGS analysis space, through key roles at BioFortis, Global Specimen Solutions and Genedata

    “ I have been pursuing a vision since the late 1990s that AI will help deliver better patient therapies. I firmly believe at POAI we will deliver on that vision. ”

    Education
    University of Surrey, UK
    Ph.D., Microbiology

    University of Wolverhampton, UK
    Undergraduate Degree, Applied Science

    Bob Myers, BBA, MBA

    CHIEF FINANCIAL OFFICER
    Predictive Oncology
    Site Leader, Skyline Medical
    At Predictive Oncology

    Executive Officer, Compliance Officer, Corporate Secretary, and member of the Senior Leadership Team. Responsible for Finance, Administration, Human Resources, Investor Relations, and IT. Skyline Medical Site Leader.

    Before Predictive Oncology

    Numerous years as CEO/Controller consultant including medical devices companies. Executive positions with CES Computer Solutions, Computer Accomplishments, Hi-Tech Stationary & Printing, Capital Distributors Corp, International Creative Management American Express, Showtime Entertainment and public accounting with Laventhol & Horwath, CPA’s.

    “ It’s a privilege to work with a highly talented team to pursue oncology advances, while protecting and increasing shareholder value. ”

    Education

    Adelphi University
    MBA, Finance

    Hofstra University
    BBA, Public Accounting 

    J. Melville (“Mel”) Engle, BS, MBA

    CHIEF EXECUTIVE OFFICER
    & CHAIRMAN OF THE BOARD
    Predictive Oncology
    At Predictive Oncology
    Mr. Engle became POAI’s CEO in 2021 and was appointed to the POAI Board of Directors in 2016. He was elected Chairman of the Board in 2020.
    Before Predictive Oncology

    Between 2012 and 2021, he was CEO of Engle Strategic Solutions, a consulting and coaching company focused on CEO issues. From 2009 to 2012, he was CEO and Chairman of Thermogenesis, a cell separation company. From 2002 to 2007, he was Regional Head/Director, North America at Merck Generics and CEO of Dey Laboratories, a respiratory company. From 1996 to 2001, he was CEO and Chairman of Anika Therapeutics, an orthobiologics company. From 1980 to 1995, he was with Allergan, Inc., an eye and skin care company, where he served as CFO, Managing Director (living in Toronto), and other positions with the firm.

    “ Eliminating cancer is our overall goal.  This is not an easy task.  We are working hard every day to make this happen, using our novel technology and talented team of employees. ”

    Education

    University of Southern California
    School of Business
    MBA, Finance

    University of Colorado Boulder
    BS, Accounting