← Back to Library

BioByte 142: Addiction treatment with glp-1s, engineering cement-making microbes, and leveraging…

This week's coverage from Decoding Bio doesn't just report on scientific breakthroughs; it argues that we are standing at a rare convergence where biology, materials science, and artificial intelligence are finally ready to solve problems that have long seemed intractable. Pablo Lubroth frames a narrative where the same molecular tools revolutionizing obesity treatment are being repurposed to break the cycle of addiction, while a genetic toolkit for bacteria promises to slash the carbon footprint of the built environment. The most striking claim here is not that these technologies exist, but that they are moving from theoretical possibility to industrial reality with unprecedented speed.

The Biology of Habit and Healing

Lubroth opens with a compelling pivot: the blockbuster drugs designed for weight loss may be the key to treating substance use disorders. He notes that "stories about people struggling with substance use disorders... who have managed to become drug-free for the first time after starting GLP-1 treatments have been spreading rapidly in recent years." This isn't just anecdotal; the piece highlights rigorous data showing that semaglutide significantly reduces heavy drinking days and cravings. The author points out that while first-generation therapies failed to make a dent, "second-generation GLP-1s are much more potent, so these could lead to better results." This distinction is crucial. It suggests that the mechanism of action is real, but the delivery and potency were previously insufficient to overcome the biological resistance of addiction.

BioByte 142: Addiction treatment with glp-1s, engineering cement-making microbes, and leveraging…

The commentary draws a fascinating historical line, noting that in 2015, researchers found a genetic link between GLP-1 receptors and alcohol dependence. Lubroth explains the theory: "alcohol dampens the body's GLP-1 production, so the brain boosts expression of the hormone's receptors 'to preserve sensitivity to the hormone that govern reward and emotion'." This reframes addiction not merely as a failure of will, but as a dysregulated biological feedback loop that modern pharmacology can finally interrupt. However, a counterargument worth considering is whether these drugs address the root psychological causes of addiction or simply suppress the physiological urge, potentially leaving patients vulnerable to relapse if the medication is stopped. Despite this, the momentum is undeniable, with major trials for alcohol-use disorder already underway.

Stories about people struggling with substance use disorders... who have managed to become drug-free for the first time after starting GLP-1 treatments have been spreading rapidly in recent years.

Engineering the Built Environment

Shifting from the human body to the global infrastructure, Lubroth tackles the massive emissions problem of cement production, which accounts for roughly 8% of global emissions. The piece argues that the bottleneck has been biological, not chemical. For years, the most effective microbe for biocement, S. pasteurii, was a "black box" that scientists could not genetically modify. Lubroth writes that the new work by the Cultivarium "moves the needle from process engineering to more rational biology." By creating the first genetic toolkit for this organism, researchers can now "tune urease activity, reduce ammonia byproducts, and engineer crystal formation and sporulation for industrial deployment."

This is a profound shift. Instead of hoping the bacteria behave in a factory setting, scientists can now design them for it. The authors validated this by deleting a specific gene cluster to abolish urease activity, proving they have precise control. Lubroth emphasizes that this breakthrough came from "mining an environmental relative that tolerated the plasmids and then porting functional parts back into the canonical strain." This strategy of borrowing resilience from nature's edge cases to fix industrial workhorses is a powerful lesson for synthetic biology. Critics might note that scaling this from the lab to saline soils and variable pH environments remains a massive engineering hurdle, and the regulatory path for releasing genetically modified organisms into construction sites is uncharted. Yet, the potential for a low-carbon alternative that matches the strength of traditional cement is too significant to ignore.

AI as a Lens for the Invisible

The final major thread explores how artificial intelligence is solving a data scarcity problem in cancer research. Lubroth introduces GigaTIME, a model that generates virtual multiplex-immunofluorescence images from standard, widely available H&E slides. The core argument is that while H&E slides are abundant, they lack the protein-level detail needed to understand the tumor immune microenvironment (TIME). Lubroth explains that the team trained the AI on over 400 paired images to predict 21 protein channels, creating a "virtual population" of nearly 300,000 images.

The result is a dataset that allows researchers to see patterns invisible to the naked eye. The piece notes that the model "showed far more promising performance on recapitulating cell-level patterns rather than just pixel level details." This is a critical distinction; the AI isn't just making pretty pictures, it's reconstructing biological reality. By combining these generated images with clinical data from over 14,000 patients, the team identified synergistic protein associations that single-protein analyses missed. Lubroth highlights that this approach "yielded more significant associations than TCGA," the traditional gold standard of cancer genomics. This suggests that the future of precision oncology may lie not in generating more expensive data, but in intelligently expanding the data we already have. A valid concern, however, is the potential for AI hallucination—if the model infers a protein presence that isn't there, it could lead to flawed clinical conclusions. The authors acknowledge the need for more diverse datasets to ensure these models work across all populations, a necessary step before widespread clinical adoption.

GigaTIME shows the promise of integrating rare, but highly informative data sources like mIF with more widespread H&E images to produce accurate models of the TIME and connect that to clinical information.

Bottom Line

Pablo Lubroth's curation effectively demonstrates that the next wave of scientific progress will be defined by convergence: repurposing metabolic drugs for addiction, genetically editing microbes for climate solutions, and using AI to democratize high-resolution pathology. The strongest part of this argument is the evidence that these are no longer theoretical concepts but active, funded, and rapidly maturing fields. The biggest vulnerability lies in the translation gap—moving from successful lab trials to scalable, regulated, and equitable real-world deployment remains the formidable challenge that will define the next decade. Readers should watch closely for the upcoming clinical trial readouts on GLP-1s for addiction and the first field tests of the engineered biocement, as these will be the true stress tests for these revolutionary claims.

Deep Dives

Explore these related deep dives:

  • Glucagon-like peptide-1

    The article extensively discusses GLP-1 receptor agonists for addiction treatment, but readers may not understand the underlying hormone biology, how GLP-1 functions in the brain's reward system, and why manipulating this pathway affects both appetite and addictive behaviors

  • Tumor microenvironment

    The GigaTIME paper section assumes familiarity with tumor immune microenvironments. Understanding the complex ecosystem of immune cells, fibroblasts, and signaling molecules that surround tumors provides essential context for why AI-generated imaging of these environments matters for cancer treatment

Sources

BioByte 142: Addiction treatment with glp-1s, engineering cement-making microbes, and leveraging…

by Pablo Lubroth · · Read full article

Welcome to Decoding Bio’s BioByte: each week our writing collective highlight notable news—from the latest scientific papers to the latest funding rounds—and everything in between. All in one place.

Dewey Albinson, Cement Works (1933)

What we read.

Blogs.

Will blockbuster obesity drugs revolutionize addiction treatment? [Elie Dolgin, Nature News Feature, December 2025]

Stories about people struggling with substance use disorders, such as opioid, cigarettes, cocaine and alcohol addiction, who have managed to become drug-free for the first time after starting GLP-1 treatments have been spreading rapidly in recent years.

Earlier this year, Christian Hendershot, a psychologist at USC, reported that weekly injections of semaglutide reduced alcohol consumption:

Semaglutide treatment did not affect average drinks per calendar day or number of drinking days, but significantly reduced drinks per drinking day (β, −0.41; 95% CI, −0.73 to −0.09; P =.04) and weekly alcohol craving (β, −0.39; 95% CI, −0.73 to −0.06; P =.01), also predicting greater reductions in heavy drinking over time relative to placebo (β, 0.84; 95% CI, 0.71 to 0.99; P =.04).

In 2015, a team of researchers uncovered the first evidence of the connection between GLP-1 and alcohol dependence, where there was a common variant of the GLP-1R gene was linked to heavier drinking. The same team used post-mortem brain tissue and discovered that GLP-1R expression was elevated in reward-related regions of the brain.The mechanistic hypothesis is that alcohol dampens the body’s GLP-1 production, so the brain boosts expression of the hormone’s receptors “to preserve sensitivity to the hormone that govern reward and emotion”.

However, since then, the first slew of trials with first-generation GLP-1 therapies like exenatide and dulaglutide did not have an impact on alcohol or smoking dependence. One study, however, involving people with opioid-use disorder saw a 40% reduction in cravings with liraglutide.

Second-generation GLP-1s are much more potent, so these could lead to better results. There are several trials in progress, testing high dose semaglutide for alcohol-use disorder; reading out next year. Eli Lilly, for instance, has launched a 300 person trial to test tirzepatide for the treatment for alcohol-use disorder.

You Gotta Make It ‘Til You Make It [Lee, Off the Media, December 2025 (based on original paper by Ortiz et al., bioRxiv, December 2025)]

Cement production is responsible for roughly 8% of global emissions. Creating lower-carbon alternatives that match its strength and durability is a pressing priority for decarbonizing the built environment. Microbially-induced ...