Latest Posts

AI Drug Discovery Hits a New Bottleneck: Experimental Validation

Kathy Vuksanaj

September 8, 2026

AI Drug Discovery Hits a New Bottleneck: Experimental Validation

The bottleneck moves downstream

Drug discovery has traditionally been constrained by the difficulty and cost of identifying promising starting points. AI is loosening that constraint, enabling researchers to explore more molecular possibilities and computationally prioritize designs.

Experimental biology, however, cannot necessarily accelerate at the same pace. Every AI-generated candidate must still confront biological reality. Researchers need to establish whether a molecule produces the desired functional response, behaves as expected in relevant biological systems, and possesses properties compatible with further development. As Chen says, each candidate must undergo “functional screening, developability assessment, safety evaluation, and preclinical testing.”

That shifts the bottleneck downstream. Instead of struggling to generate enough interesting molecules, discovery organizations can face more computationally attractive candidates than their laboratories can efficiently validate.

Chen says this shift is increasing the importance of experimental capabilities including “high-throughput affinity screening, functional and mechanism-of-action assays, developability assessment, immunogenicity testing, advanced in vitro models, and translational in vivo studies.”

A prediction is not a medicine

Computational promise and therapeutic potential are not the same. Predicted affinity or potency might move a candidate forward, but successful medicines must satisfy a much broader set of requirements.

“A promising computational prediction is only the starting point,” says Wenwan Fang, PhD, product manager, discovery at ProBio. A candidate must demonstrate the desired biological activity and mechanism of action while also possessing “favorable safety, pharmacokinetic, and developability characteristics.”

A molecule that performs impressively computationally or in an early assay might still prove unstable, difficult to manufacture at scale, or unsuitable because of immunogenicity or other development risks.

“Ultimately, the most valuable candidates are those that combine strong biological performance with the practical attributes required for successful development and commercialization,” Fang says.

For Chen, that makes prioritization increasingly important. “Success depends not on creating more molecules, but on validating and prioritizing the right ones quickly and efficiently,” he says.

Building validation at AI speed

Keeping pace with AI will likely require more than simply adding laboratory capacity. Companies might need to rethink how validation is integrated into discovery.

“As AI dramatically increases the number of potential drug candidates generated, organizations will need to invest in technologies that accelerate validation rather than discovery alone,” Fang says.

Those investments could include laboratory automation, robotic liquid handling, high-throughput screening platforms, advanced cell-based and functional assays, and integrated data-management systems. Fang also expects growing demand for technologies that assess developability, safety, and manufacturability earlier, alongside more predictive in vitro and in vivo models.

The objective is a tighter feedback loop: AI proposes candidates, experiments test them, and experimental data inform subsequent designs.

Finding molecules that matter

Although better AI models will remain important, Chen sees integration as the larger opportunity. “The greatest competitive advantage is likely to come from tighter integration between computational prediction and experimental validation,” he says. Organizations that create “seamless feedback loops between AI-driven design and high-quality experimental data” will be best positioned to accelerate discovery, reduce development risk, and improve the likelihood of clinical success.

For Fang, the most valuable investments are similarly those that help researchers “rapidly identify which AI-generated candidates are truly worth advancing into development.”

Explore More www.probiocdmo.com/add-aidd-solution.html

ProBioGen Grants Transgene Additional License for AGE1.CR.pIX Cell Line

John Sterling

September 8, 2026

ProBioGen Grants Transgene Additional License for AGE1.CR.pIX Cell Line

ProBioGen and Transgene report that they broadened the scope of their existing collaboration to include an additional license agreement for ProBioGen’s suspension AGE1.CR.plX^Rcell line which is designed to boost Transgene’s manufacturing capabilities with this specialized production technology.

ProBioGen officials describe the cell line as a stable proliferating avian cell line derived from primary cells of a duck embryo. It was developed as an alternative to the use of chicken eggs for large-scale vaccine production. The cell line is available as both a suspension and adherent cell line and has been optimized for viral vaccine production and stability.

The product, which grows in a commercially available, chemically defined medium without animal components and serves as a host for a variety of different virus strains, is aimed at cost-effective production and increased manufacturing productivity.

“This additional license reflects our confidence in the technology to prepare for future clinical needs at scale,” said Alessandro Rive, chairman and CEO of Transgene.

“We value Transgene’s continued confidence in our AGE1.CR.pIX platform,” added Volker Sandig, PhD, CSO at ProBioGen. “A continuous cell line qualified as a GMP master cell bank, grown in chemically defined and animal-component-free medium, means every batch begins from the same fully characterized starting material.”

Transgene focuses on designing and developing targeted immunotherapies for the treatment of cancer. The company’s clinical-stage programs consist of a portfolio of viral vector-based immunotherapeutics. TG4050, the first individualized therapeutic vaccine based on the myvac^® platform is Transgene’s lead asset in the adjuvant treatment of head and neck cancers.

TG4070, a second individualized vaccine candidate derived from the myvac platform, is in Phase I clinical development in combination with nivolumab in adjuvant non-small lung cancer (NSCLC). The company has other viral vector-based assets, including BT-001, an oncolytic virus based on the Invir.IO^® viral backbone, which is in clinical development.

Anticancer Candidate Strengthens Bones and Prevents Weight Gain in Postmenopausal Mice

Sophia Ktori

September 8, 2026

Anticancer Candidate Strengthens Bones and Prevents Weight Gain in Postmenopausal Mice

The results of a preclinical study by researchers at the University of East Anglia suggest that an experimental anticancer drug could stop osteoporosis and help women prevent weight gain after the menopause. Their study in mice found that the small molecule drug CADD522 not only protected against osteoporosis but could also reduce body fat and reverse some of the metabolic changes linked to menopause. Currently in development for cancer therapy, CADD522 is a small molecule inhibitor of the transcription factor RUNX2, which is helps to drive the growth and spread of several cancers.

Research lead Darrell Green, PhD, at UEA’s Norwich Medical School, said, “We have uncovered an entirely new way of tackling the disease. We found that a drug originally developed to stop cancer could help millions of women facing the twin challenge of fragile bones and midlife weight gain. We hope our work could lead to a new generation of osteoporosis treatments that tackle bone loss while also addressing some of the wider metabolic consequences of menopause.”

Green is senior and co-corresponding author of the team’s report in npj Drug Discovery, titled “RUNX2 inhibitor CADD522 improves bone microarchitecture and lipid metabolism in post-menopausal bone loss.” In their paper the team concluded that their collective studies “… identify RUNX2 inhibition as a therapeutic strategy that simultaneously improves skeletal integrity and metabolic homeostasis, supporting further development of CADD522 for osteoporosis and other RUNX2-driven diseases.”

Osteoporosis is a metabolic bone disorder characterized by low bone mass, structural deterioration and increased fracture risk. The disorder affects one in three women and one in five men aged over 50 years globally, the authors wrote, but there are limited therapeutic options. “… existing antiresorptive and anabolic therapies remain limited by safety concerns, contraindications and poor long-term adherence,” the authors stated. “Thus, new treatments with few side effects and broader applicability remain a clinical priority.”

Green said: “Osteoporosis affects around one in three women over the age of 50, leaving sufferers vulnerable to painful fractures that can seriously impact quality of life. Current treatments exist, but many are plagued by side effects, safety concerns or inconvenient dosing schedules that make long-term use difficult.

The researchers had previously evaluated the small molecule RUNX2 antagonist CADD522 in several preclinical cancer models. “RUNX2 is critical for in utero skeletogenesis and cancer metastasis,” they explained. Unexpectedly, studies indicated that CADD522 reduced cancer-induced bone disease, “… suggesting potential utility in osteoporosis.”

For their newly reported study Green and colleagues investigated whether RUNX2 inhibition could protect against post-menopausal bone loss, in an ovariectomy-induced mouse model, to mimic the hormonal changes seen after menopause.

The team found that animals treated using CADD522 for eight weeks showed significant improvements in bone health. Scans revealed increased bone volume and better preservation of the delicate honeycomb-like structures inside bones that are crucial for strength and resilience. Blood tests suggested the drug stimulated new bone growth, without interfering with the body’s normal process of breaking down and rebuilding bone.

Green said: “This is particularly important because many existing osteoporosis drugs work by suppressing bone loss, which can sometimes lead to complications when used for long periods.”

The biggest surprise came when the investigators looked beyond bone health, Green continued. “The mice receiving CADD522 weighed less than their untreated counterparts despite eating the same amount of food. They also had less body fat and fewer fat deposits accumulating inside their bone marrow—a process that is commonly seen after menopause and is linked to declining bone health.” The authors wrote in summary, “A remarkable finding was that the skeletal improvements occurred alongside reduced peripheral and marrow adiposity and selective remodeling of lipid metabolism.”

The team also examined brain tissue and found the drug appeared to reverse several menopause-related changes in fatty acids. Levels of beneficial omega-3 fats, including DHA, remained largely intact, while a number of other lipid abnormalities shifted back towards healthier patterns. “Rather than inducing indiscriminate metabolic disruption, RUNX2 inhibition therefore appears to restore specific lipid pathways perturbed by menopause,” the team noted. “Although the relationship between altered brain lipid composition and cognitive function was not examined in the present study, the selective remodeling of brain lipids following CADD522 treatment raises the possibility that RUNX2 inhibition may influence neurological adaptations to menopause in addition to preserving skeletal integrity.”

Green added: “We didn’t directly test for memory or thinking ability, but our work raises questions about whether this drug could one day help address wider menopause-related health problems.” The team suggests that further studies should determine whether the metabolic effects demonstrated translate into improvements in cognition or other measures of brain health.

The prospects for future clinical applications also received a boost from safety testing. Experiments in mice, rats and dogs found CADD522 could be taken orally and was well tolerated. “Cross-species pharmacokinetic and toxicological studies demonstrated oral bioavailability, favorable short-term tolerability and target engagement despite rapid systemic clearance, while cellular thermal shift assays confirmed direct engagement of RUNX2,” they noted.

The team also found the drug appeared to be metabolized more slowly in human tissue than in rodents, potentially improving its performance in people. “This is still in the early stages and has so far only been tested in animals but we hope that the benefits will translate to humans to ultimately reduce fracture rates,” added Green.

In conclusion, the authors wrote, “In summary, our findings reveal an unexpected role for RUNX2 inhibition in protecting against post-menopausal bone loss while simultaneously improving systemic metabolic homeostasis.” They suggest that their study “… establishes a translational framework for further development of CADD522 and identifies RUNX2 as a therapeutically tractable regulator of skeletal and metabolic adaptation in adulthood.”

Regenerative Medicine in the Bronze Age: MSCs and MSC-EVs Leading the Way

Kathy Vuksanaj

September 8, 2026

Regenerative Medicine in the Bronze Age: MSCs and MSC-EVs Leading the Way

MSC Therapies: Promise, Frustration, and a Push Toward Biomarkers to Drive Efficacy

MSC therapies show strong biological promise, but translation remains slow because efficacy and consistency are difficult to demonstrate. A key priority is aligning therapy design with MSC biology. Autoimmune diseases, such as progressive multiple sclerosis, are strong applications because treatment options are limited and early studies suggest MSCs may help reset immunity. Experts recommend biomarker-driven trials to identify appropriate patients earlier and improve evaluation. MSCs are often misunderstood and underutilized, making biomarker-driven design and global harmonization essential for success.

Lonza Scientists Discussing Data — Credit: Lonza

MSCs are attractive because they can be used in allogeneic therapies: donor cells can be cryopreserved, thawed, and used off-the-shelf to treat larger patient populations, including autoimmune and cardiovascular indications. Developers are moving toward 3D culture systems to increase production, simplify manufacturing, and enable MSC-EV generation. However, shifting from 2D to 3D culture systems can alter confluency, marker profiles, EV productivity, and other critical quality attributes. Teams must optimize media, serum or HPL, additives, and reagents to maintain CQAs at scale.

Process changes in media, microcarriers, or feed schedules can dramatically affect MSC performance, increasing interest in AI-driven optimization of consumables and reagents. As 3D systems advance, developers still rely on serum or HPL to support cell health and consistency, while the industry increasingly demands chemically defined, serum-free systems. Better serum-free and HPL-free solutions are needed for clinical development of MSC therapies and MSC-EVs. Together, these trends show MSC therapies becoming more engineered, precise, and clinically defined, while many MSC developers expand into MSC-EV solutions and fuel excitement around both modalities.

MSC-EVs: Increasing Interest in a Novel Therapeutic Solution

The shift of MSC-EVs from “cells as drugs” to “cells as biological triggers” raises an interesting central question: can developers deliver therapeutic impact without administering the cells themselves?

With more than 500 MSC-derived EV clinical trials underway, the field continues to grow. MSC-EVs lead the EV and exosome space because of the strong foundation built during decades of MSC research. However, key questions remain around classification, safety, efficacy, scale-up, and commercialization. MSC-EVs remain in a regulatory grey zone, with agencies still determining whether they are biologics, cell-based therapies, or a new class. Regulators are likely to focus on process consistency, EV identity, and release testing in later-stage development.

Purity and characterization remain central because upstream and downstream processes can change EV composition. Each indication must show how cell source, processing, and purification affect the final product. With no commercial EV therapeutics approved so far, developers must establish precedent while maintaining consistency through scale-up. As MSC-EV developers scale into larger trials and commercial manufacturing, GMP processes and reagents can change EV outputs and must be addressed early. Early conversion to GMP media and consumables can reduce risk, lower costs, and clarify efficacy, purity, and consistency priorities. MSC-EVs will also face competition from EVs derived from NK cells, macrophages, HEK293 cells, plants, and other sources. These alternatives may offer lower-cost production or different applications, while shared learnings can advance the broader EV field.

MSCs and MSC-EVs Drive Progress and Advance Regenerative Medicine

Immunotherapies have defined a recent Golden Age in cell and gene therapy. As manufacturing matures and refinements become more incremental, the next major wave of innovation may come from regenerative medicine. MSCs and MSC-EVs hold major promise, but many companies still lack the infrastructure and timelines that attract major pharma investment. This places the field in a biotech Bronze Age: rich with opportunity in rare disease, organ failure, allogeneic therapies, and scalable EV platforms, but still building the evidence, biology, and manufacturing precedent needed for broader success. Continued progress in scale-up, process control, and modality definition could unlock a future Golden Age for MSC and MSC-EV therapeutics.

Jessica Pickrell; Associate Director, Media Solutions, Lonza

Further Reading

1. Wang CK, Tsai TH, Lee CH. Regulation of exosomes as biologic medicines: Regulatory challenges faced in exosome development and manufacturing processes. Clin Transl Sci. 2024; Aug 17(8):e13904. doi: 10.1111/cts.13904. PMID: 39115257; PMCID: PMC11307316.

2. Bourcier AJ & Kirkor ZM. Regulatory, ethical, and clinical barriers to exosome use in interventional pain medicine. Intl Pain Medicine 2026; 5(1):100746. doi: 10.1016/j.inpm.2026.100746.

3. nternational Society for Cell & Gene Therapy (ISCT). 2026. Roundtable on Regulation and Policy, ISCT 2026 Annual Meeting, Dublin, Ireland.

Lonza Group Ltd. and its affiliates (collectively and individually, “Lonza”) make efforts to include accurate and up-to-date information. However, Lonza makes no representations or warranties, express or implied, including as to accuracy or completeness of information. All trademarks belong to Lonza, and are registered in the USA, EU and/or CH, or used in common law, or belong to third-party owners and are used for only informational purposes. All third-party copyrights have been reproduced with permission from their owners. The user bears the sole responsibility for determining the existence of any third-party rights and obtaining any necessary licenses and approvals. For more information, including regarding legal disclaimers, Lonza’s intellectual property rights, and how Lonza collects, uses and protects personal information: www.lonza.com/legal, www.lonza.com/about-us/strategy/intellectual-property and www.lonza.com/privacy. © 2026 Lonza. All rights reserved

To learn more, visit bioscience.lonza.com/lonza_bs/US/en/mscs-mesenchymal-stem-cells.

RNAV8 Bio Joins ARPA-H Team to Pioneer Programmable RNA Medicines

John Sterling

September 8, 2026

RNAV8 Bio Joins ARPA-H Team to Pioneer Programmable RNA Medicines

RNAV8 Bio (pronounced Renovate Bio) has been selected to receive funding from the Advanced Research Projects Agency for Health (ARPA-H) for PROPEL (Programmable RNA for Optimal Precision in Therapeutic Efficacy and Localization).

RNAV8, a biofirm applying AI-driven design and laboratory validation to engineer more predictable mRNA medicines, will be part of a team led by the Rouskin Lab at Harvard Medical School with the Weissman Lab at MIT/the Whitehead Institute. The award, a one-year ARPA-H pilot of up to $4.4 million, will fund a program to turn RNA’s natural folding behavior into a precise, drug-tunable control layer for a new generation of RNA medicines. PROPEL is led by ARPA-H program manager Shannon Greene, PhD.

“The promise of mRNA has always been that it’s programmable, but in practice the relationship between a RNA’s sequence and chemistry and what it actually does has been hard to predict,” said Devan Shah, founder and CEO of RNAV8 Bio. “Our work with the Rouskin and Weissman labs is aimed squarely at that gap: learning the rules that connect sequence to function, and building them into logic-gated constructs that express where and when they should.

“Just as important, every element we discover stays answerable in the format a medicine actually ships in. We’re proud to help carry this science from a regulatory element toward a real therapeutic.”

Within PROPEL, RNAV8 Bio focuses on the therapeutic format itself. It screens protein output from pools of transfected mRNA, engineers UTRs in the delivery format, and validates results at the cargo level, for example driving cell-type-selective expression of therapeutic payloads such as gene-editing enzymes or CAR constructs.

PROPEL pursues a different kind of control

Most medicines act wherever their chemistry carries them, and most genetic therapies work by permanently changing DNA. PROPEL pursues a different kind of control. An RNA molecule’s untranslated regions (UTRs) fold into structures, and those structures set how much protein the message produces. When a small molecule binds one of those folds, the structure rearranges and the output changes, turning a dose into a dial on protein expression, without altering the genome. Bacteria use this logic openly, in elements called riboswitches; in human cells it remains largely uncharted.

“For decades we’ve read RNA sequence as a set of instructions for making a protein. What we can now read, at the scale of thousands of sequences at once, is how that sequence folds, and how a small molecule can change the fold and, with it, the output,” said Silvi Rouskin, PhD, assistant professor of microbiology, Harvard Medical School. “Human cells almost certainly already use this kind of structural control; we’ve simply never had the tools to find it systematically. PROPEL is our attempt to map it and put it to work.”

PROPEL pursues this along two lines of work that share a single screening pipeline. Discovery searches human RNA for naturally occurring elements whose structure responds to a metabolite or an FDA-approved drug, native regulation reachable with existing, well-tolerated molecules.

Engineering installs regulatory elements into therapeutic mRNA so that a single construct is selective by cell type through its sequence and tunable through its structure. Both rely on high-throughput mapping of how a RNA’s folds shift when a ligand binds, which flags the elements that regulate protein output before the team tests which ones a small molecule can move.

The Weissman Lab contributes massively parallel screens of human untranslated-region elements across cell types, measuring which ones set expression where.

“The untranslated regions flanking a message are among the most powerful and least exploited levers on how much protein a cell makes,” said Jonathan Weissman, PhD, professor of biology, MIT and member, Whitehead Institute. “By screening these elements across cell types, we can begin to tell which ones set expression where, turning a vague notion of ‘regulation’ into a defined, reusable parts list. That is what makes a control layer generalizable rather than a one-off trick.”

By the end of its initial phase, the collaboration aims to produce resources that do not exist today: a map of how thousands of human RNA sequences respond to small molecules, with structural models; a ranked catalogue of human UTR elements that set translation cell type by cell type; engineered UTRs that reach meaningful selectivity in a therapeutic format; and RNA sequences, natural or engineered, shown to switch structure inside human cells. Together these are meant to form a generalizable, disease-agnostic foundation for RNA medicine, according to the scientific team.

PROPEL is a program supported by ARPA-H that develops a programmable, drug-tunable control layer for RNA medicines.

Measuring Cell Line Development Quality Within Days of Single-Cell Cloning, Not Months Later

Kathy Vuksanaj

September 8, 2026

Measuring Cell Line Development Quality Within Days of Single-Cell Cloning, Not Months Later

Clean Cells Broadens Viral Safety Testing Portfolio with NGS

John Sterling

September 8, 2026

Clean Cells Broadens Viral Safety Testing Portfolio with NGS

France-based Clean Cells launched its next-generation sequencing (NGS)-based safety testing for viral matrices, such as viral vectors and recombinant preparations. The company also announced that it had acquired French bioinformatics CRO Xegen, which specializes in NGS data analysis.

The acquisition is to help consolidate, develop, and secure Clean Cells’ bioinformatics capabilities to support every NGS method and de-risk the most technically complex part of the workflow, according to Clean Cells CEO Laurent Claisse, who added that a secured bioinformatics backbone, the most critical part of NGS, means reliable, auditable data handling.

Xegen co-founder and CEO, Julien Paganini, PhD, will join Clean Cells as expert project manager of bioinformatics, analytical methods development, and validation, along with his team of two staff.

“With the launch of our NGS viral safety testing and the acquisition of Xegen’s Next-Generation Sequencing bioinformatics expertise, we provide a strong and secure analytical pipeline, with greater control of the full workflow from sample to interpreted result. We can provide our international clients with a more integrated NGS offering to support key questions from identity to viral safety,” said Claisse.

“Sequencing is the technology but turning it into a quality control solution involves much more than simply generating reads. It requires a robust analytical strategy to transform complex data into interpretable, traceable and regulatory-ready results. By joining Clean Cells, we can combine our established bioinformatics structure and deep expertise with its laboratory and regulatory capabilities to build a fully integrated solution and make NGS a powerful and practical tool for quality control in regulated environments,” added Paganini.

A Clean Cells official pointed out that this new offering comes at a critical time, as European regulatory frameworks are making NGS increasingly relevant in checking that biologic medicines are not contaminated with unwanted viruses. Clean Cells’ strategic move strengthens its position in the viral safety testing market in step with European regulations, notably ICH Q5A(R2) (2024) and European Pharmacopoeia chapter 2.6.41, which are accelerating NGS adoption from a niche technology to a mainstream viral safety testing tool, explained Elodie Ribert, R&D deputy director, head of analytical development and validation.

“The regulatory ground has shifted. With ICH Q5A(R2) and European Pharmacopoeia chapter 2.6.41, NGS is now a recognized route for adventitious virus detection. It answers a real problem for viral matrices, where classical assays struggle because no anti-serum exists,” said Ribert. “The sequencing alone is not the whole story. A signal still must be investigated and confirmed, and that is where having the full quality control panel alongside it makes a difference.”

Going forward, Clean Cells states that it will extend the validated NGS viral safety method from acellular viral matrices to cell banks (MCB/WCB), a process currently in progress. It will also build additional viral safety expert capacity and business development enablement.

12 3...2,144 Page 1 of 2,144

A council of separate assays

One framework, one language

Catching what no one thought to ask

When the process changes

What MAM does not solve

Analytical continuity by design

Thaw-and-use target cells

Off-the-shelf primary effector cells

The bottleneck moves downstream

A prediction is not a medicine

Building validation at AI speed

Finding molecules that matter

A modular protein adaptor

Stable retargeting for systemic delivery

One platform, multiple opportunities

MSC Therapies: Promise, Frustration, and a Push Toward Biomarkers to Drive Efficacy

MSC-EVs: Increasing Interest in a Novel Therapeutic Solution

MSCs and MSC-EVs Drive Progress and Advance Regenerative Medicine

PROPEL pursues a different kind of control

Podcast