BioByte 077: Body axis maps, efficiency of pharma R&D, focused research organizations in atomistic simulation, GLP-1 directed NMDA antagonism
Welcome to Decoding Bio, a writing collective focused on the latest scientific advancements, news, and people building at the intersection of tech x bio. Happy decoding!
Hope everyone is having a great Spring thus far. We will be taking a brief hiatus for just a couple of weeks to work on some new initiatives - don’t miss us too much! We’ll be back soon, better than ever.
What we read
Our body axis maps are getting redrawn [Ground Truths, 2024]
The way we feel, behave and act is governed by intricate circuits in our body. Eric Topol covers recent papers (see below) that shed light on new findings in this field, paving the way to novel therapies and means to promote health and treat disease.Â
A body–brain circuit that regulates body inflammatory responses [Nature, May 2024]
Scientists from Columbia University Have identified cells in the brainstem that can sense immune cues from the body and act as master regulators of the body’s inflammatory response. This is significant as scientists have know that the brain plays a role in immune response coordination but the exact mechanism has been elusive. Specifically, the team found that cytokines communicate with specific populations of vagal neurons and in response, the brain can modulate the course of the subsequent inflammatory response. The studies were performed in mice.
Paternal microbiome perturbations impact offspring fitness [Nature, May 2024]
It is clear that germline mutations are passed on between generations, but recently researchers at EMBL have produced evidence suggesting that paternal microbiome perturbations may also be passed on to the offspring. Negative perturbations to the microbiome including antibiotics and laxatives produced offspring with shortened lifespan and lower birthweight is some cases.
Brain-muscle communication prevents muscle aging by maintaining daily physiology [Science, May 2024]
Researchers examined the role of the circadian lock in the peripheral tissue and its involvement in muscle maintenance.
Efficiency, effectiveness and productivity in pharmaceutical R&D [Roland et al., Nature Reviews Drug Discovery, 2024]
This analysis used data for a cohort of 14 major pharma companies to model snapshot views of R&D productivity reflecting the inter-company differences across early and late stage R&D activities. R&D productivity is assessed as the ratio of R&D effectiveness (average NPV per product) to R&D efficiency (cost per product approval).
In summary:
The median cost per new product approval was $2.8B, with the most efficient company achieving $1.7B per approval, and the least efficient $6.9B.Â
The median company NPV per approval was $5.4B, ranging from $2.4B to $20.4B.
The median company productivity was 2.2m ranging from 0.5 to 8.0.
The authors went on to show the impact of top- and bottom-quartile clinical phase cycle times on R&D efficiency and which levers companies could focus on to improve productivity:
Roadmapping Report: Focused Research Organizations in Atomistic Simulation [Sam Holton, May 2024]
Why it matters: Atomistic simulations have the potential to revolutionize drug discovery, materials design, and our understanding of chemical processes at the molecular level. However, the field has not seen the same rapid progress as other computational disciplines like machine learning. This roadmapping report identifies key bottlenecks, such as the lack of high-quality datasets and standardized benchmarks, that are holding the field back.Â
A recent perspective piece from Convergent Research provides an overview for how Focused Research Organizations could advance the field of atomistic simulation. Atomistic simulation involves computational methods to model atomic interactions based on physical and chemical principles, using techniques like molecular mechanics and quantum mechanical methods. In contrast, machine learning-based methods use data-driven algorithms to predict properties by learning from large datasets, offering faster predictions once trained. These approaches complement each other, with atomistic simulation providing detailed accuracy and ML enhancing speed and efficiency.
Despite decades of research, the field of molecular simulation has not experienced an acceleration akin to the machine learning revolution. The essay argues that this is due to a lack of high-quality datasets, open-source code sharing, and competitive benchmarks to coordinate research efforts. Collecting valuable datasets and establishing open challenges is proposed as a way for an organization to catalyze this transition.
Two sources of novel data are highlighted: 1) experimental measurements of atomic motions in proteins using structural techniques like cryoEM and NMR. These data could validate and refine simulation force fields (functions used to describe the potential energy of a system of atoms which are essential for simulating the dynamics and thermodynamics of chemical systems) beyond what is possible with theoretical calculations alone. 2) User interaction data from browser-based tools that enable citizen scientists to conduct basic atomistic simulations. These data could train an AI agent to autonomously explore chemical problems.
GLP-1-directed NMDA receptor antagonism for obesity treatment [Petersen et al., Nature, 2024]
Why it matters: GLP-1-targeted agonists have taken the world by storm over the past few years, with exceptional data supporting their role in treating several chronic conditions including diabetes, obesity, hypercholesterolemia, and many more. The success of GLP-1 agonists have rapidly led others to explore ways to build upon this mechanism for improved treatment of metabolic-related disease. In this paper, Petersen et al., leverage a bimodal molecule combining NMDA receptor antagonism with GLP-1 receptor agonism to reverse obesity, hyperglycemia, and dyslipidemia in rodents, and was more efficacious than GLP-1 agonists alone.
GLP-1 agonists stimulate GLP-1 receptor-expressing neurons in the brain that are integrated to peripheral circulation (brainstem and hypothalamus), which activates neural circuits that regulate appetite and reward. Here, the authors generated a drug consisting of a GLP-1 agonist linked to an NMDA receptor antagonist. NMDA receptors bind to glutamate and have a critical role in regulating synaptic plasticity (and have been linked to obesity in prior human genome studies). Of note, indiscriminate inhibition of NMDA receptors can lead to significant side effects including hyperthermia and hyperlocomotion, which has led to failures in this space in the past. However, this bimodal compound facilitates specific inhibition; the NMDA receptor antagonist, dizocilpine, is only activated after being bound to the GLP-1 receptor and internalized by the cell. Thus, only GLP-1 receptor expressing neurons will be subject to NMDA receptor inhibition.
Indeed, the bimodal compound was superior to GLP-1 receptor agonists, and was better at reducing body weight in rats and mice. The specific delivery to GLP-1 receptor expressing neurons was critical for this effect, as linking the NMDA receptor antagonist to other peptides was not more effective than other compounds. Of note, rodents treated with the bimodal compound had minimal side effects associated with both GLP-1 agonists and NMDA receptor antagonists. Given that the bimodal compound reduced weight even more than calorie-restricted animals, they hypothesized that the drug influences something beyond just reducing food intake (a la GLP-1 agonists). The authors found that the bimodal compound successfully altered gene expression in both GLP and NMDA related signaling pathways, and more strongly activated reward-regulating areas of the brain while also upregulating genes related to synaptic plasticity.Â
Although this work shows promise in rodents, significant work will be necessary to establish the utility of dual GLP1 and NMDA receptor antagonism. However, this adds to the exciting body of work focused on therapeutic interventions for obesity.Â
What we listened to
Notable Deals
UK biotech raises £35M to develop solid tumor drugs based on machine learning
Progentos emerges with $65M for multiple sclerosis trials with IP from Frequency
AltruBio, after tweaking checkpoint drug, raises $225M for midphase ulcerative colitis program
ADC biotech Pheon raises $120M to test three assets in the clinic
Blackstone-backed biotech launches with up to $300M for immune drug
Carolyn Bertozzi-founded Lycia Therapeutics raises $106M to enter the clinic
What we liked on socials channels
Field Trip
Did we miss anything? Would you like to contribute to Decoding Bio by writing a guest post? Drop us a note here or chat with us on Twitter: @ameekapadia @ketanyerneni @morgancheatham @pablolubroth @patricksmalone