Using Technologies of the Future to Disrupt Healthcare (1): Personalized Medicine Today

A few weeks ago, at the University of Tokyo, an elderly lady was treated for Leukemia by Watson. Watson is a cognitive supercomputer and through, it’s machine learning models, it explored over 20 million cancer research writings within 10 minutes, enabling it to prescribe a personalized treatment. She made a full recovery.

This is a great example of how technology around us can bring a positive influence to our communities. But, what was the science behind this story? Personalized medicine.

Personalized medicine is an emerging science in the field of medicine and genomics. Lately, it has been making a buzz in the scientific innovation community. Why, you ask? Other exponential technologies such as Artificial Intelligence (AI) and Quantum Computing.

In this article series, I will discuss:

1. Personalized Medicine, what it is and how it works today
2. The Implementation of Artificial Intelligence, which will cover other applications of AI, as well as how it can be incorporated into Personalized Medicine
3. An Introduction to Quantum Computing and how it can be used in the Medical and Genomics fields

Personalized Medicine Today

Think of personalized medicine as the large umbrella that covers a range of topics such as predictive disease testing (looking at one’s genome for mutations that could potentially cause the occurrence of a specific disease in a patient) and pharmacogenomics (using genomics to predict which drug or medication would be optimal to improve an individual’s health). The main advantage personalized medicine provides for patients is the fact that it takes into account for individual variability in genes, environment and lifestyle. This approach allows doctors and researchers to predict more accurately which treatment and prevention strategies for a particular disease will work in which groups of people. Each of these sub-topics play an important role in making sure that each patient receives the correct personalized therapy instead of a generalized solution that works for the majority of the population, as is explained below through predictive testing and pharmacogenomics.

Predictive Testing

Predictive disease (or genetic) testing is using the process of genome sequencing to look for mutations in a person’s DNA that are linked to disease. Since these mutations normally begin to show their symptoms much after birth, predictive testing can be useful for anyone such as people who have a family history of a genetic disorder or people who show no signs of illness at all.

Why it matters?

Predictive testing plays an important role in personalized medicine as it helps genetic advisors and caregivers analyze your blood and genotypes to find what mutations you have that can cause health problems in the future. Through careful planning and therapy, your chance of getting the disease and recovering can increase by a large margin through a genetic testing and analysis by experts. In the past year alone, over 2 million patients died from a disease, despite following their prescribed treatments. If accounting for the individual variability in genes, environment and lifestyle when prescribing treatments was implemented on a large scale, think about how many lives could have been saved in the past and how many could be saved in the future.

Genome Sequencing

As previously stated, genome sequencing is important in the world of predictive testing. So what is it? It’s goal to separate pieces of DNA that differ in length by one base (there are four bases: adenine (A), thymine (T), cytosine (C) and guanin (G)). Click here to read more about these bases and the basics of DNA. In order to this, a technique known as electrophoresis is used. It involves copying a piece of DNA multiple times and dividing into four batches, which are chemically modified and fluorescently dyed to serve as markers for each base (four bases so four batches). Next, DNA is run through an array of 96 gel-filled compartments in a machine called a capillary sequencer and these machines read the base sequence as the DNA moves through the gel. As the DNA pieces emerge from the gel, they move past a laser that causes the dye molecules to fluoresce. A fluorescent light detector “reads” the color of the DNA strips and the results are displayed on a connected computer screen. After analyzing the genome, genetic advisors can offer a diagnosis.

There are two types of predictive testing, both which are based on genome sequencing — presymptomatic and predispositional testing. I’ll go through them one-by-one:

Presymptomatic Testing

If the disorder is known in the family, presymptomatic testing is offered to “at-risk” symptomless patients to predict whether the individual carries the corresponding gene. The reasons behind having this testing is reassurance, future life planning, screening or preventative treatment. Positive test results in this case mean that the patient will certainly develop symptoms of the disease in the future and through this testing, genetic advisors can advise a possible, generalized therapy plan of how to proceed in terms of changes to one’s lifestyle. The amount of time required for this testing depends on what gene is being searched for as each gene varies in terms of size, effect and host differences. Such differences result in the time to find a certain gene vary.

For example, when looking for mutations in the CFTR (Cystic Fibrosis Transmembrane Regulator) gene, which is responsible for cystic fibrosis (the buildup of thick, sticky mucus that can damage many of the body’s organs due to the CFTR protein not regulating the proper flow of chloride and sodium in and out of the cell membranes in the lungs), testing is a quick process as mutations in most segments of the gene are easily identified. On the other hand, larger genes such as the dystrophin gene in Duchenne Muscular Dystrophy (DMD) which causes degenerative muscle loss, may take much longer to look for because of their size and because the mutation is slightly different in each individual.

Predispositional Testing

Predispositional testing is often used to test for cancer. In many ways, it is similar to presymptomatic testing in terms of gene sequencing, analyzing and prescribing but it differs in terms of results. In predispositional testing, if you get positive results, you are not certain to contract a disease but your genetic makeup makes you more likely than others to get the disease while positive results in presymptomatic testing mean that a patient is certain to get a disease. One common application is looking for breast cancer. Predisposition testing looks for a germ-line mutation in a cancer susceptibility gene, which is a major indicator that a woman could fall victim to breast cancer in the future. Generally, predisposition testing focuses on finding altered genes and analyzing if the carried mutation increases the likelihood of a patient getting cancer.

Predispositional testing is much more complex than what it seems so far. Why? Cancer cells (the products of unregulated cell growth in the body) are actually present in each person’s body. Certain factors in our lifestyle such as what we eat, metabolism and the amount of radiation we intake can trigger cancer cells to start abnormally dividing and concentrating in one area of the body. But we’ll save the science behind cancer and how it can potentially be treated for a different article.

Pharmacogenomics

Pharmacogenomics focuses on how genetic variants influence drug responses in patients by looking at the connection between gene expression (the process of transcribing DNA into RNA and translating RNA into proteins) and pharmacokinetics (how various medications move in the body) and pharmacodynamics (how medications affect the human body and why such effects occur). The picture below shows that differences in a chromosome can create variations of the ATCG sequence, which are simply referred to genetic variants.

So in terms of genomics, what causes genetic variances from person to person? There are multiple enzymes. Some of the major ones include:

• Uridine 5'-diphosphate (UDP)-glucuronyltransferase is a phase II enzyme that catalyzes the glucuronidation (major mechanism for the formation of water-soluble substrates) of various endogenous compounds such steroid hormones and exogenous compounds such as prescription medications. Accumulation of this enzyme due to a decrease glucuronidation can potentially lead to gastrointestinal and hematopoietic toxicity.
• CYPs (Cytochrome P450s) are a gene family of hemoproteins which use many small and large molecules as substrates in enzymatic reactions. They also are responsible for the metabolisms of over 70% of medications and drugs.
• TPMT (Thiopurine methyltransferase) gene catalyzes the formation of a covalent compounds with Sulphur and mercaptopurine or azathiopurine (different types of thiopurine). These compounds are used to suppress the immune system for treatment of patients with leukemia or after organ transplantation. Since TMPT falls under a 6-MP (mercaptopurine) metabolism which is known to have significant toxicological consequences to the affected individual, particularly related to liver dysfunction.
• VKORC1 (Vitamin K Epoxide Reductase Complex Subunit 1) gene, when combined with the CYP2C9 gene (one of the CYPs) is responsible for identifying any potential possibilities of bleeding when Warfarin (a drug used to treat blood clots and prevent strokes in people with artificial heart valves) is used.

Such variances mean that a generalized solution for a health concern won’t treat everyone so in comes the process of applied pharmacogenomics:

1. Get genetic data
2. Find the patient-specific genetic variants
3. Translate such genetic variants into likely DNA variations
4. Figure out the patient’s genotype with the information gained from the DNA variations
5. Translating the genotype into phenotypes
6. Determining what medication the patient will take

To understand this rather simple process, take look at the following example:

Genetic testing has dramatically reduced the number of people suffering side-effects to Abacavir, a HIV medicine (not to be confused with AIDS as HIV is the virus that leads to AIDS). Statistically, Abacavir is a highly effective treatment for HIV but around 5–8% of its users suffer severe side-effects such as rash, fatigue and diarrhea.

These symptoms suggested that the patients who suffered from side effects of the drug were suffering a ‘hypersensitivity’ reaction, which means their immune systems were producing an exaggerated response to the drug, like an allergy. This suggested that the genes were controlling their immune system’s responses might be responsible for the side effects they were experiencing.

In 2002, the genetic variant which was the key factor in hypersensitivity to abacavir was identified as the HLA-B*5701 allele in major histocompatibility complex (MHC), a set of cell surface proteins essential for the acquired immune system to recognize foreign molecules.

The HLA-B*5701 allele occurs at a frequency of roughly 5% in European populations, 1% in Asian populations and less than 1% in African populations. Screening for the HLA-B*5701 allele before treatment has become standard nowadays and if a patient is found to have the HLA-B*5701 allele, Abacavir is avoided and alternative HIV treatments such as Lamivudine, Zidovudine and Emtricitabine (these are proven to work on individuals who have the HLA-B*5701 allele) are given.

The Bigger Picture

Of course both of these advancements in the medical field are great and personalized medicine is expanding today like never before. However, there is still one problem: humans are performing such tasks manually. This takes up human energy, time and a lot of extra effort. Instead, by incorporating exponential technologies such as Artificial Intelligence and Quantum Computing in healthcare, especially personalized medicine, a lot of time and effort could be saved, as well as gradual benefits in the health of patients.

Key Takeaways

• Personalized medicine’s main advantage is that it takes into account for individual variability in genes, environment and lifestyle when prescribing treatments
• Predictive testing is using genome sequencing to analyze an individual’s potential to contract a disease because of their genetic makeup
• Pharmacogenomics revolves around studying how an individual would react to a certain drug or medication based upon genetic variances
• Because all of this is done manually, it can be very time and energy consuming

If you enjoyed this article, like it, share it with your network and stay tuned for my next two articles which will discuss how AI and quantum computing can be incorporated into the healthcare and personalized medicine fields.