Mike Nally, Generate Biomedicines CEO

Nearly a year after raising a staggering $370 million in Series B financing just before the market for privately funded startups turned bearish, Generate Biomedicines says it has more than begun to live up to the ambitious goal summarized in its very name.

Since garnering its eye-popping financing in November 2021, the protein therapeutics developer has focused on building its pipeline, which is on track for 14 preclinical programs this year.

“That will ultimately lead toward multiple INDs in 2023,” Generate CEO Mike Nally told GEN Edge. The company is heading into IND-enabling studies for its first two programs—one in COVID-19, the other undisclosed—and starting to scale up manufacturing with CMO partners.

In addition, Nally said Generate has also expanded its platform beyond creating antibodies and enzymes into bispecifics and cell therapies. In the second half of this year, he says, “we’ll work more in terms of fleshing out the right bispecific constructs.”

“What bispecifics and cell therapy represent, in some respects, is just more complex biology. The advantage the computer has is it can co-optimize multiple parameters simultaneously, in a way that the human mind starts to struggle with,” Nally said. “As you go from an antibody to a bispecific to cell therapy, there’s a continuum of complexity that you’re navigating, and we think the computer helps us in that arc.”

Nally said Generate wants to carve out a distinct presence in three disease areas: oncology, immunology, and infectious disease: “We think the technology lends itself well there, there are areas of good learning where there’s good early predictive markers of success that could help us validate the platform and address major unmet needs.”

That platform, the company asserts, can generate medicines across modalities that include antibodies and peptides, enzymes, and stealth proteins by uncovering generalizable principles of protein complexes, allowing the prediction of novel binders for desired targets. The company says it has generated femtomolar (fM) binders that have outperformed its upside target for optimization, and structurally confirmed its de novo binders across multiple targets.

The platform integrates protein science expertise with structural biology, artificial intelligence, and machine learning algorithms designed to analyze many millions of known proteins and peptides, looking for statistical patterns linking amino-acid sequence, structure, and function, data science, and computational infrastructure—all supplemented with Generate’s own proprietary experimental data. The platform aims to identify specific biological processes involved in disease that are capable of modulation through a variety of protein modalities.

Through that approach, which the company calls generative biology, Generate aims to engineer de novo proteins that address an existing or emerging therapeutic need. The company can generate in silico antibodies on demand that can be engineered to target defined epi­topes or ago­nize cell sur­face receptors. It can also generate functional and agonistic antibodies, as well as antibodies to integral membrane proteins, multiprotein complexes, and other long difficult to hit targets.

Best-in-Class Ambition

“Not only can we generate de novo proteins, but we can make it a best-in-class protein,” Nally said. “We’ve pushed on affinity, taking low picomolar binders down to the femtomolar level, levels that to be honest I didn’t know were possible. That obviously changes the underlying potency as well as the dosing profile for potential medicines.

Generate has gone from creating variant sets of around 100 proteins last fall to millions of proteins at a time, largely through investments in automation and microfluidics.

“What you could see for the first time in drug discovery is economies of scale, where you’re migrating from the artisanal craft that has been the bedrock of pharmaceutical innovation over the past 50 to 100 years, to a place where all of a sudden, you get better with every protein you generate,” Nally observed. “You make biology more engineerable, and ultimately, typically over the course of history that’s when industrial revolutions occur.”

One infectious disease Generate has worked to address is COVID-19, where Generate has focused on the S2 domain of the SARS-CoV-2 Spike protein—as a result generating the first potent pan-sarbecovirus S2 antibody.

“Upon the emergence of [the Omicron variant] last Thanksgiving [2021], we said: Could we find regions of the spike protein that are more immune cryptic, less immune dominant than the receptor-binding domain, which has been where most of the therapeutic antibodies have been targeted,” Nally recalled. “When we dissected the Spike protein, what we found was both the NTD supersite, as well as the RBD had been evolving rapidly. The NTD non-supersites in the S2 domain were much more highly conserved, not just for SARS-CoV-2, but across all cervical viruses.”

“If you could find that potent and broadly neutralizing antibody that bound to those regions, you potentially could be protective against all of the variants of concern, at least until there’s pressure put on those regions,” Nally added.

The company’s work has extended beyond Omicron, to its BA.5 subvariant, where neutralization has been observed.

“Antibodies tend to bind to certain epitopes, and some epitopes are just more hidden from the immune system. The ability to direct these binders to specific locations is one of the real advantages we think our technology has,” Nally said. “We’re not left to the will of the immune system. We can actually direct things to desired epitopes.”

“If we can find something that could help certain populations—think of the immunocompromised population, for which vaccination doesn’t provide the same level of protection—an antibody-based regimen could be really meaningful for those populations,” he added.

Blockbuster partnership

Of Generate’s 14 pipeline programs, nine have identified targets: Five are internal targets the company has worked on, while the other four identified targets stem from its up-to-$1.9 billion collaboration with Amgen, launched in January.

The companies are partnering to discover and create protein therapeutics for five clinical targets to be selected by Amgen across several therapeutic areas and multiple modalities.

“We’ve started on three of them. We just learned of their decision on the fourth one, so we’ll start on the fourth one in short order,” Nally said.

Neither Amgen nor generate is disclosing the targets. “What we found is that there’s been complimentary areas of interest, where the areas of their expertise are additive to some of the areas where we’ve been focusing on,” Nally added.

Amgen has agreed to pay Generate $50 million upfront, and a collective $1.9 billion plus royalties for the five programs—to consist of up to $370 million in milestone payments plus royalties per program. Amgen also retains an options to nominate up to five additional programs, at additional cost committed to participating in a future financing round for privately-held Generate.

“Based on the [Series B] raise that we had, and based on the upfront from Amgen, we have a cash runway that puts us into 2024,” Nally said. “So, we’re able to prosecute a pretty ambitious agenda with a good underlying financial position.”

The collaboration with Generate aligns with Amgen’s interest in designing improved protein-based drugs that use smaller, simpler protein antibodies to bind to multiple targets. Last year, Amgen acquired single-chain antibody developer Teneobio for up to $2.5 billion—$900 million in upfront cash, plus up to $1.6 billion in milestones—and developed a Digital Biologics Discovery group that combines biology, automation, and protein engineering.

“We thought Amgen was a great partner—they’re one of the preeminent protein engineering houses in the world. But more than that, they bring complementary capabilities to us,” Nally said. “They have a rich history in disease biology and large-scale manufacturing of large molecules. They’ve excelled in that over their history. And [they have] some great clinical development commercial capabilities. So, we thought pairing our ML [machine learning] drug discovery technology with those capabilities would be really beneficial.”

Nally is also CEO-Partner with Flagship Pioneering, the venture capital and business acceleration firm whose best-known spinout is Moderna. Flagship founded Generate in 2018, and solely funded the $50-million Series A financing through which Generate emerged from stealth mode and formally launched in September 2020. Generate has raised more than $420 million since its founding.

Growing field

As GEN reported earlier this year, Generate is among companies specializing in the growing field of AI-based drug discovery. Other drug and platform developers with a similar focus include:

  • Absci, which in January inked an up-to-$610 million collaboration with Merck & Co. to develop drugs using Absci’s AI-powered Integrated Drug Creation™ Platform, and Bionic Protein™ non-standard amino acid technology to produce enzymes tailored to Merck’s biomanufacturing applications. In April, Absci opened an AI research lab in New York City.
  • Cyclica, a “neo-biotech” whose platform leverages AI and computational biophysics. Toronto-based Cyclica is working to discover drug candidates for targets selected by Seoul-based Samjin Pharmaceuticals through a collaboration announced August 2.
  • Deep Genomics, whose AI Workbench platform is designed to enable RNA therapies for almost any gene in any genetic condition. Deep Genomics completed a $180-million Series C financing last year.
  • Healx, whose Healnet indication expansion platform analyzes millions of drug and disease data points to find novel connections that could be turned into new treatments for rare diseases. In July, Healx moved into 3,500 square feet within the Science Village lab building of Chesterford Research Park in Cambridge, U.K.
  • Insilico Medicine, whose platform enabled its antifibrotic small-molecule inhibitor ISM001-055, a treatment for idiopathic pulmonary fibrosis, to advance from target discovery to a Phase I trial (NCT05154240) in under 30 months—the company’s first pipeline candidate to reach the clinic. Insilico said August 10 it completed a second closing bringing its Series D financing to $95 million, led by Prosperity7, the diversified growth fund of Aramco Ventures.
  • X-Chem, which last year acquired Glamorous AI, the developer of an AI-powered software as a service (SaaS) drug discovery engine, for an undisclosed price. In April, X-Chem launched XD3, designed to combine data from its DNA-encoded library (DEL) screening, its analytical and predictive ArtemisAI platform, and deep medicinal chemistry expertise.

Adding space

Generate’s growth trajectory is reflected in its move, completed in May, from Cambridge, MA, to its own 71,000-square-foot headquarters in suburban Somerville, MA, where it is one of four Flagship Pioneering spinouts to occupy a space of more than 270,000 square feet on eight floors at 101 South Street. The building is the first in Somerville to be purpose-built as Class-A lab space, part of the 1.8 million-square-foot Boynton Yards mixed-use lab/residential/cultural development, located less than a mile from Cambridge’s Kendall Square and MIT.

Additionally, Generate is fitting out another 70,000+ square feet in Andover, MA, for an R&D facility with four cryo-electron microscopes, “to really start adding to the world’s capacity in protein-protein interactions.”

“We’re often asked, how does what you’re doing compared to what AlphaFold does?” Nally said, referring to the AlphaFold Protein Structure Database, developed by DeepMind Technologies, UK-based AI subsidiary of Google owner Alphabet. AlphaFold helps researchers predict the 3D structure and folding pattern of proteins from their amino-acid sequence.

“We’re generating proteins that nature has never sampled before, and in doing so, the only way to know whether your prediction or your generation is accurate is actually to build those proteins. The cryo-EM piece allows us to determine whether the structures that we’re predicting and generating off of are actually manifesting in the ways we want.”

Generate has expanded its staff, which has multiplied from 30-35 employees when Nally became CEO in March 2021, and from 80 employees when Nally spoke with GEN Edge last November. Over the summer, Generate hired its 200th employee, and is on track to hit 500 employees over the next two years.

Building the organization, he said, has entailed growing Generate’s computational core, protein production, and preclinical testing operations: “We’re now starting to add into the clinical development core, and then manufacturing CMC [chemistry, manufacturing and controls], as well as the broader SG&A [selling, general and administrative expenses] infrastructure.”

Generate has also grown its headcount by expanding its leadership team, recently announcing the appointment of Alexandra Snyder, MD, as chief medical officer and Jason Silvers, MD, JD, as chief financial officer.

Snyder was previously a principal at private equity firm Two River, and earlier was the Head of Translational Oncology at Merck & Co. Silvers is a 20-year Goldman Sachs veteran who co-managed the firm’s healthcare investment banking for Europe, the Middle East and Africa.

“The company is scaling well, and we continue to have ambitious growth targets on the headcount side. And I think we’re making good progress toward those,” Nally said. “We’re on a pretty good track as long as the technology continues to at a minimum meet if not exceed our expectations. We recognize the opportunity we have before us, and we continue to build out the organization.”