Lantern Pharma Making Headway in Using AI to Revive Old Drugs, Develop New Compounds
by Charlotte Hu – Precision Oncology News
Jul 8, 2020
NEW YORK – Lantern Pharma, armed with $26.3 million from a recent initial public offering, is using its RADR artificial intelligence platform to create gene signatures that can guide patient enrollment in drug trials, as well as discover new compounds for biomarker-selected populations.
The Dallas, Texas-based firm said that the drug response predicting platform has thus far collected over 450 million data points, and it is using that data to revive two drugs that were once shelved by other drugmakers: the first for metastatic hormone-refractory prostate cancer patients and the second for never smokers with non-small cell lung cancer. It is also garnering preclinical data on a novel compound for glioblastomas and solid tumors bearing a certain genomic signature discovered by the RADR platform.
Lantern Pharma was founded in 2013 by Arun Asaithambi, Gregory Tobin, Peter Nara, and Biological Mimetics, a Maryland-based drug development company that owns a minority stake in Lantern, with the goal of rescuing drugs that had failed late-stage clinical trials. Nara and Tobin, who both previously worked at the National Cancer Institute, were inspired by the successful foray of another small company into rescuing and repurposing once-abandoned drugs.
“There was a company called Cougar Biotech that took a drug that had been abandoned and pointed it specifically at metastatic hormone refractory prostate cancer and then sold the drug to J&J for a billion plus dollars,” said Panna Sharma, who became CEO of Lantern Pharma in 2018. The idea behind the company’s initial business strategy was to “pick up the rights and buy drugs that had failed in late-stage and then use modern genomics to better target that drug to the right cancer subtype, and then eventually [bring it] to the right patients.”
Lantern focuses on the scientific strategy, using its AI platform and all the data elements added to it, regulatory strategy and approach. The company doesn’t operate a wet lab, but partners with others that do to conduct genomic sequencing, drug sensitivity testing, and preclinical models. Sharma cited Memorial Sloan Kettering, Fox Chase Cancer Center, and Georgetown University as partners in this regard.
Seven years ago, when Lantern was founded, artificial intelligence was becoming increasingly used to comb through large swaths of genomic data for emerging patterns. Last month, Lantern announced that its platform has collected over 450 million data points and is well on its way to collecting 1 billion, a year ahead of schedule. The company is leveraging the data to identify abandoned drugs that could be re-trialed in newly defined populations and optimize the development of new drugs. More data points mean that the platform can make more accurate in silico models to support these drug development efforts, said Sharma.
Lantern goes after four main types of data to power its RADR platform: RNA sequencing data, drug sensitivity data, biomarker data, and patient information. The team will gather data from relevant microarray or RT-PCR assays, drug sensitivity data, including analysis using cell lines and other models, and a variety of biomarker data. Patient information Lantern collects relates to treatment history, medical background, and any relevant lifestyle characteristics such as being a smoker.
Lantern gathers the data from various sources, including clinical trials, presentations at medical conferences, peer-reviewed publications, and statistical analyses that were previously done on the drug that Lantern has purchased for rescue. “When we buy the rights to a drug that we’re interested in, we’re buying all the historical data that is associated with that,” Sharma explained.
Lantern partners with others, such as the NCI, Genomics England, and Fox Chase, to get access to proprietary data on specific drugs or cancers of interest. It purchases the patient data utilized by the RADR system from hospitals or healthcare centers or gains access to it through the Institutional Review Board.
The platform uses this real-world data on patients, particularly whether they are responding or not responding to specific drugs, to identify the cancer types an agent might be particularly effective in, Sharma noted. The company then uses this information to make decisions about which drugs to license and identify the types of compounds that might work best for a particular group of patients.
Lantern’s research collaborators can also aid the company in generating its own internal data through sequencing campaigns and new clinical trials.
“Now that we’re public and have raised some capital, we are doing a whole series of pretty large-scale RNA sequencing campaigns in certain tumor areas that we’re interested in … [such as] in glioblastoma, across different types: some that are MGMT methylated and some that are not,” Sharma said. “We might get in cell line [data] or biopsy material, xenograft, whatever systems or models that we think are relevant.”
Second chance hopefuls
In a filing with the US Securities and Exchange Commission, Lantern stated that as of April 2020, it owns or controls the rights to over 115 active patents and patent applications across over 14 patent families that underlie the drug candidates it is developing. Specifically, Lantern is working to revive LP-100 and LP-300, two drugs that previously failed late-stage trials.
LP-100, also known as irofulven, was a compound which exploited cancer cells’ DNA repair dysfunction. It was first developed by scientists at the University of California, San Diego and subsequently licensed to MGI Pharma. When MGI was acquired by Eisai in 2008, the drug was shelved after failing a Phase III trial in metastatic pancreatic cancer. The trial did not use a biomarker to try to predict best responders to LP-100, said Sharma.
MGI Pharma stopped the Phase III pancreatic cancer trial in April 2002 because the comparator drug fluorouracil showed a greater than expected survival benefit. However, by then, the drug’s safety and tolerability were well understood, since it had already been through 41 clinical studies and more than 1,500 patients had received it. “Drugs that fail these late-stage trials aren’t necessarily bad drugs, [but] just not enough is known about them in terms of … which patients are most likely to benefit,” Sharma said.
LP-100 was also tested in a Phase II trial of male hormone-refractory prostate cancer patients in 2006. “The response rate to LP-100 was 86 percent greater in men with hormone-refractory metastatic prostate cancer versus those given the control, which was chemotherapy plus prednisone. They tended to have almost a year of extra life” in the LP100 arm, Sharma noted. “By all measures, that is a perfect candidate for Phase III. However, that wasn’t the trial that went to Phase III.”
Lantern acquired LP-100 knowing that it had activity in ovarian, pancreatic, and prostate cancer. Specifically, in prostate cancer, the company saw an opportunity to molecularly define the responder population based on a combination of genomic factors. Lantern partnered with Denmark-headquartered Oncology Venture to develop an 86-gene signature that correlates to response in metastatic hormone-refractory prostate cancer.
Oncology Venture is developing a propriety drug response prediction tool, dubbed DRP, that can scope out which patients are likely to benefit from certain drugs.
According to Lantern’s SEC filing, it holds an exclusive license for the development and commercialization of LP-100. However, in 2015, Lantern and Oncology Venture inked a development agreement around LP-100, and the drug is currently undergoing a Phase II clinical trial in Europe for androgen receptor-targeted and docetaxel-pretreated metastatic castration-resistant prostate cancer patients.
Patients are screened using a LP-100-specific response biomarker signature with Oncology Venture’s DRP, and patients identified as likely responders to LP-100 are recruited in the trial. The study recruited its first patient in 2018 and Lantern expects to announce results in 2021.
The other drug Lantern is repurposing, LP-300, was acquired from BioNumerik Pharmaceuticals in a 2018 deal, which granted Lantern the rights to domestic and international patents and trademarks, as well as related technology and data relating to the compound for human therapeutic treatment indications. LP-300, also known as dimesna, had been studied across 10 clinical trials, mainly in lung and breast cancer. It failed a Phase III trial in non-small cell lung cancer, in which LP-300 was combined with paclitaxel and cisplatin.
“It didn’t reach the kind of increase in two-year survival that [BioNumerik] would have liked to have seen,” Sharma said. “But there was a group of 66 patients that had remarkable response to the drug.” These patients, who tended to have a 13-month improvement in overall survival over those receiving standard-of-care paclitaxel and cisplatin, were likely to be women and never smokers.
In a virtual abstract presented at ASCO’s annual meeting in May, preclinical studies of LP300 showed that the drug was able to inhibit the activities of the EGFR, MET and ROS1 kinases. Further, researchers found that LP-300 has the potential to boost the inhibitory activities of targeted tyrosine kinase inhibitors such as erlotinib (Astellas Pharma/Roche’s Tarceva) and crizotinib (Pfizer’s Xalkori).
Lantern is expecting to enter the drug next year in a Phase II trial for patients who are never smokers, have non-small cell lung adenocarcinomas, and have a genomic signature identified by its RADR platform that correlates to LP-300 response.
Developing new drugs
During the ASCO virtual meeting this year, Lantern Pharma also presented data on a 10-gene signature that conferred response to a novel compound, LP-184.
LP-184 is a next-generation alkylating acylfulvene agent that has been shown to be less toxic and more active than other compounds in the fulvene drug class. The agent has shown activity in tumors that overexpress PTGR1, according to Sharma. “PTGR1 and PTPN14 are two major [tumor] drivers. And when those two genes are expressed, this drug lights up in those tumors.”
Lantern’s RADR platform has identified multiple solid tumor indications that highly express PTGR1, including prostate, ovarian, kidney, liver, lung, pancreatic and thyroid cancers. LP184 also has good blood-brain barrier permeability, said Sharma, which supports the company’s plans to develop the drug in a second program for glioblastomas and other central nervous system cancers.
The team is aiming to conduct additional preclinical studies to gauge the toxicity of the drug and its activity in various cancer cell lines, as well as with fresh patient biopsies. Sharma said Lantern plans to publish biomarker data for LP-184 later this year in solid tumors and glioblastoma, which the firm will use to ultimately select patients for a Phase I trial in the next two years, pending regulatory discussions.