Subverting traditional medical research: developing anticancer drugs with artificial intelligence

Release date: 2016-04-18

In November 2013, more than 100 cancer patients – including pancreatic cancer, breast cancer, liver cancer and brain tumors – participated in a clinical trial of a new drug, BPM 31510, discovered using an algorithm.

The story of BPM 31510 begins with the extraction of biological data from healthy tissue and cancer tissue samples from more than 1,000 patients. These data are then processed by artificial intelligence algorithms that analyze the data and suggest possible drug treatments. We basically reversed the way we do medical science research. Niven R Narain, 38-year-old co-founder and president of Boston Pharmaceuticals startup Berg, who developed BPM 31510, said that instead of using preconceived assumptions to guide our experiments and generate specific types of data, let the patient's biological data guide us to make assumptions. .

According to Narain, developing an effective anti-cancer drug is a very difficult process. The cost of completing a R&D and production process for a pharmaceutical company is as high as $2.6 billion, and the time cost is 12 to 14 years. Only one percent of the final anticancer drugs that entered clinical trials proved effective. The price is high and the development process is unforgivable, Narain said. "If you are in any other industry, we will all be unemployed.

In Narain's view, the problem lies in the way pharmaceutical companies develop anticancer drugs. Scientists first came up with the hypothesis that, for example, a particular abnormal protein is responsible for a particular type of cancer. He said the pharmaceutical company then tested the hypothesis by screening the target protein from millions of compounds for a few of the possible chemical reactions and potential drugs. This is a textbook description of this research method. This process is called “hit-and-miss” by Narain, a dart-type method for treating diseases. He believes he can do it in different ways of thinking, with half the time and money costs. Better.

Berg co-founder and president Niven R Narain

BPM 31510 is a compound made with enzymes that play a key role in cellular metabolism and is the first therapeutic recommended by Berg's artificial intelligence. The development of this drug actually began in 2009, when the startup purchased cancer tissue samples from more than 1,000 patients from various medical schools across the United States. These samples contain more than 40 different cancer cells—multiple types of breast cancer, prostate cancer, liver cancer, kidney cancer, lung cancer, brain cancer—and healthy tissue samples taken from the same patient.

Narain didn't want to limit his research to a type of cancer. Instead, he wants to develop a model that can see the commonality of different cancers. Berg's researchers used these samples for in vitro cell culture, allowing them to perform further experiments. To mimic the microenvironment that cells experience in humans, they culture these cells in different levels of sugar and oxygen Petri dishes and continue to identify and detect the lipids, metabolites, proteins and enzymes they produce. Track their evolution. "We were testing how the data changed with the environment," Narain said. In addition to the genomic information, we have 14 trillion data points in a single tissue sample.

This torrent of biological data has helped Berg scientists map the cascade of molecular reactions occurring in cells—from genes to proteins to lipids and metabolites—in unprecedented detail. They then used artificial intelligence to compare data from healthy and diseased cells. It is impossible to (manually) process all of that data, and it is not even possible to understand it logically. Narain says you need to use artificial intelligence to find out how normal cell evolution is interrupted, how it can lead to disease and what the underlying treatments are. Many people say that this is not the way to develop drugs. My answer to this is: It is true, because this is the way to develop drugs.

At the age of 25, Niven R Narain came to the conclusion that "the science driven by hypothesis is flawed" when he was an oncologist at the Miller School of Medicine at the University of Miami. On July 22, 2003, he conducted an experiment using melanoma cells in a petri dish. He treated it with a cream containing CoQ10, an enzyme located in the mitochondria, the source of energy for cells. To his surprise, almost all the cancer cells in the Petri dish died the next day. He then repeated the experiment in mice with melanoma. After 30 days, the tumors were reduced by an average of 55%. But Narain has no clue as to why this process happened.

That was two years after the completion of the Human Genome Project, Narain said, when people believed that cancer was primarily genetically driven. After the genome was sequenced, most people believe that all we have to do is find the genes responsible for cancer, regulate them, and cure most cancers. Narain became a research assistant in skin oncology and therapy research at Miller School of Medicine, working under the direction of Chinese researcher Sung "Bob" Hsia; Hsia was co-discoverer of vitamin K and estrogen, Nobel Prize Winner Edward Doisy has worked together.

At that time Hsia was nearly 80 years old, Narain recalls, he would say: "I don't have these new genetic technologies, but don't forget that biochemistry is the foundation of the human body. This is the end of a study of health continuation or disease onset. Narain did not forget these words. He argues with scientists who believe that the genome is the answer to all medical problems. In Narain's view, the genome is only part of a larger chemical reaction involving proteins, metabolites, and cellular lipid studies.

When Narain told his results about the Hsia of CoQ10 for decades, his mentor retorted that this lipid, enzyme and mitochondrial tissue could not kill cancer cells, especially melanin. tumor. You screwed up this experiment.

A Berger examines a patient sample

Narain repeated this experiment. Other research groups also replicated this experiment. This is not an accident. We didn't pass the gene, so it made me wonder what happened. It must go far beyond genomics. It’s not a good thing to say that at the time, but you have to follow this data.

He came up with a hypothesis and tested it, collected the data and analyzed it. Meaningless. But he did come to the conclusion that hypothesis-driven science is not the best way to deal with complex diseases, especially cancer.

In March 2005, Mitch Gray, CEO of a private equity firm in Nashville, Tennessee, visited the University of Miami's Dermatology Department at Narain. Gray bought the right to use the brand name Sea & Ski, a popular sunscreen brand in the 1980s, and he was looking for new skin products. Narain had a bad cold that day, and he was planning to go home when his chairman called him to attend a meeting with Gary and give a quick demonstration of his work. I am not feeling well and really don't want to go there, so I just flipped my slides there. Narain said that when I briefly mentioned the work of CoQ10, Gray asked me to stop and explain further. I said that because he came here for skin care products, this will not be of interest to him. Gray replied: "Please allow me to decide what I am interested in. He interrupted me.

When Narain finished his presentation, Gray pulled him aside and called Narain to meet his Nashville team and his partner Carl Berg. Berg was a vending machine repairman who invested in commercial real estate in Silicon Valley in the 1960s. Now he is a billionaire and co-founder of Mission West Properties in Cupertino, which controls more than 700,000 square meters of Silicon Valley real estate. He is also an early investor in more than 100 technology companies, including Sun MicroSystems and NetLogic.

In 2006, Gray and Berg received licenses for products including NaQ's CoQ10, which they have been working on. The three of them founded a new company, Cytotech Labs, where Narain served as a scientific consultant and remained in Miami to continue product development. Six months later, he called and told Berg some news. He has tested this product on many different cancer cells. It kills cancer cells. “He told me: “I think there are some things in your product that may have an impact on cancer.” Berg recalls that I am very concerned about this because I have been told that every cancer cell is different. If this works for many cancers, then we really extend it. I said, "Call me when you have cured someone."

Berg A sample of more than 60,000 patients with cancer, diabetes and neurological diseases in a cold store at temperatures as low as -80 °C

One year after obtaining the patent for the drug, an intellectual property expert contacted Narain. For privacy reasons, Narain did not mention his name. The man explained to Narain that his eight-year-old son was dying of life with Ewing's sarcoma, a rare and apparently incurable lymph node cancer. They have tried many treatments, both traditional and experimental. But all failed, and now his son can only live for a few months. He wants to try Cytotech's new drug. Narain explained that the ointment has not been tested for sarcoma. "He told me: "I know, but the method you use is different. I want to try it." Narain recalls, "I don't want to create false hopes, but he is a very nervous father, can't accept the answer." After they met, Narain agreed, and also clearly told the boy's father and doctor that he and his company did not believe that this method had a reasonable basis for treating his son's cancer. "I did that because I couldn't tell you what the father's eyes looked like," Narain said. "That's a desperate look."

To continue, Cytotech must first obtain approval from the US Food and Drug Administration (FDA). When Cytotech's regulatory team submitted the documents to the FDA, they were approved within three hours.

For treatment, the doctor applies the cream to the child's arm and to the position of the lymph nodes behind the knee and neck. "The lymph nodes don't absorb the oil into the body," Narain said. "If you want to infiltrate the cream into your skin, I think if you use it enough on the lymph nodes, then maybe there will be some entry into the system and circulation." This is the basic principle, but let me point it out very clearly: you can never cure such a solid tumor. The probability of this method being effective is zero."

Eight months later, the child has no signs of cancer. "He is still alive today and lives very well," Narain said. "We don't know what happened. This thing really brings us a view that this may be really a very special molecule. When I heard the news, I cried. No scientists expected this result."

At the end of 2008, Narain moved to Cambridge, Massachusetts, and the company was renamed Berg Pharma. Narain is frustrated with everything he sees in the clinic. “The way patients are treated is often unrelated to biology and their disease,” he said. “I know I have to start over and put biology at the forefront.” After that, Narain encouraged Berg to develop data-driven use of high-throughput Molecular screening and artificial intelligence to find drugs. Berg expressed interest. He is tired of real estate business and decided to invest in companies that have the potential to change the world. Narain's idea convinced him.

“I’ve been avoiding biotechnology for the past 50 years, because I think FDA is just another additional risk that I don’t want to take,” Berg said. “But I think it’s so revolutionary, I’m willing to take this risk. ."

In early 2010, the team completed the first iteration of its cancer model: an unprecedented detail depicting the mapping of cascades of biological events leading to cancer from physical health. “The output of the machine is like a map of the airline, including large hubs that account for most of the traffic, such as New York and London, and routes to and from them,” Narain said. “The hub in the map represents the excess of cancer tissue relative to healthy organizations. Or defective molecules. The molecules in these hubs become our medicine. This is the difference between artificial health and the disease that artificial intelligence tells us. The platform marks unwanted or defective molecules in diseased cells, so we Fill these systems or enzymes back into the system, or we reduce or suppress them to get the system back to normal."

The resulting cancer mapping has 5 hubs. The largest of these is composed of a group of enzymes in the mitochondria. CoQ10 is one of them. "When I saw that hub, it was a moment that made me feel good," Narain said. "That is a time for verification. Remember, I have been thinking about it for many years, and it’s not related to genomics. Artificial intelligence tells me that the fulcrum of cancer is mitochondria. Of course, in my opinion, according to all my previous experiments, this makes perfect sense."

When Narain added CoQ10 back to in vitro cultured cancer tissue, most of the cancer cells began to die. Using Berg's intensive molecular data-tracking, Narain finally understood why. Mitochondria, in which CoQ10 is active and help generate energy in cells, not only provide energy to cells, but also control the ability of cells to die.

"Cell cancer cells can turn off mitochondria and produce energy from lactic acid rather than oxygen," Narain said. "So cancer cells also lose the ability to die." Sending CoQ10 back to the mitochondria reverses this effect and turns cancer cells into Ordinary cells. "We re-teach them how to die."

On a cloudy afternoon in December 2015, Niven R Narain asked WIRED to visit Berg Pharma's lab in Framingham, 30 km west of Boston. The area enclosed by white walls does not look like an eight-year-old pharmaceutical company. It is more like a large biomedical research institute with more than 200 employees – including doctors, physicists and computer scientists. - Work in a large room with a mass spectrometer and a large server. “What you see here is the most rigorous, high-capacity and high-capacity laboratories in the country,” Narain explained. “We have more samples here than any academia or industry lab.”

Berg also researched drugs for the treatment of Parkinson's disease. This refrigerator is filled with samples of patients

Narain was born in Guyana and grew up in the Bahamas, but his gentle and calm voice did not cover his Caribbean style. He wore a dark buttoned shirt and striped jacket. His dark hair was neatly divided, and his face was dotted with goatee. As he walks slowly along the long corridor, he stops to chat with his colleagues, and his behavior is cheerful and infectious.

Berg Pharma is currently investigating more than 200 drugs, many of which are still in preclinical studies, most of which target cancer, diabetes and neurological diseases. Narain expects the company to approve at least one drug for multiple types of cancer in the next few years and will also commercialize a new diagnostic method for prostate cancer. Berg also has drug research projects for neurological diseases such as Alzheimer's disease and Parkinson's disease. "I think neurology is the next "cancer" that is difficult to cure," Narain said. "Many pharmaceutical companies have closed their nerve projects because most of the trials have failed. Our biological mechanisms for this disease no idea."

Narain walked into a room full of refrigerated cabinets containing more than 60,000 cell samples from medical patients. The samples inside are stored in Eppendorf tubes. They are marked with: Parkinson's disease, Alzheimer's disease, diabetes, cancer, lupus, inflammation. "We have cell samples and cell secretions," Narain said. "There is a lot of information we can extract. The genome is the template for the human body, but the genes produce RNA, which in turn produces proteins and metabolites and lipids. That is Our raw materials."

Berg is working with the US Department of Defense to launch a diagnostic test for prostate cancer in 2017, and the company can use thousands of tissue samples from its employees. "Prostate cancer is one of the biggest problems facing the military. They want us to come up with a better diagnostic test than the existing prostate-specific antigen (PSA) method that is not very good." Narain said, "We Four biomarkers were found that were not only more predictive than PSA, but they could be separated into more positive forms to let us know if patients might need surgery. There were no markers in the past to do this.” In addition, Berg also Harvard Medical School has teamed up on Project Survival to discover the first biomarker for pancreatic cancer until the late stages of cancer, which is often difficult to detect. Of course, we don't know anything about Berg's method—Narain calls it Interrogative Biology—that's cancer-specific. Berg is also studying tissues from patients with Parkinson's disease (since 2013) and Alzheimer's disease (since 2014).

Berg is using the same data-driven approach for clinical trials of its BPM 31510. It collects urine, serum and plasma from the patient during each cycle of drug infusion, which is done regularly over a 60-day cycle. These samples were then analyzed by 12 machine mass spectrometers: 5 for proteomics, 5 for metabolism, and 2 for lipidomics. (Their genes are sequenced at Mount Sinai Hospital in New York.) "I can tell you what dinner you had eaten ten days ago because it is still in your lipids." Michael, deputy head of the System Medicine department at Berg Kiebish said, "If you pick up something, you will say, "Okay, what is this?" We are picking up something that has never been detected. This is the real virgin land. We will find some people who will never It's something that's linked to the disease, because we measured it for the first time. There is still a broad and open field in health data, and most people are still focusing on genomics. We have to go beyond it. ."

Berg's scientists are cultivating cancer cells to make sure they keep growing

Leonardo Rodrigues, Deputy Director of Advanced Analysis at Berg, agrees. “Genomics is very important, but it brings us a static picture of what is happening in our body,” he said. “It’s like trying to determine the quality of a restaurant by looking at the recipes used in the restaurant. Genomics There are many things happening outside. We can capture these causal events to get a dynamic picture of what happened. That will lead to better drugs and biomarkers.” Every time the team collects samples from patients, they use them. CAT and PET scans measure tumors. They also measure the concentration of the drug in the patient's plasma and keep track of their diet and sleep patterns. Narain said: "It is very likely that we are trying to do this for the first time and have been approved by the FDA."

The current trial involved more than 100 participants from the Cornellville Medical Center in New York, the MD Anderson Cancer Center at the University of Texas, and the Palo Alto Medical Foundation. Although still in its infancy, early results were encouraging, suggesting that the drug is effective in treating high-metabolism solid tumors in patients with severe disease with minimal side effects. “We saw a reduction in the tumor and we saw that the patient felt energetic and a lot better,” Narain said. “We even saw the tumor disappear in the first phase of the trial, which is for this type of patient. Very rare."

Narain clearly remembers such a special case. The patient was a male medic who had bladder cancer. His doctor was one of Cornell's leading gastrointestinal experts, but the patient refused all treatments.

I wrote in my notebook: "The first evidence of no signs of tumor activity in the patient during the Phase 1B trial." Narain said, "The solid tumor has completely disappeared. It's amazing. It verified everything for us. That day. I feel very proud."

Source: Heart of the Machine Synced

IV Solutions for Antitumor and Chemotherapy Antiemetic

Iv Solutions For Antitumor And Chemotherapy Antiemetic,Tropisetron Hydrochloride Injection,Ondansetron Hydrochloride Injection,Granisetron Hydrochloride Injection

Shandong Qidu Pharmaceutical Co., Ltd. , https://www.qdyypharma.com