AI-Engineered Multispecific Proteins for Obesity: High-Stakes Bet With No Clinical Anchor Yet

Zealand Pharma and Generate Biomedicines Partner for AI-Engineered Metabolic Therapeutics

Zealand Pharma and Generate Biomedicines have announced a strategic collaboration and license agreement to develop multispecific protein therapeutics for obesity and other metabolic diseases. The partnership combines Zealand Pharma's metabolic health expertise with Generate's proprietary generative AI platform for protein engineering. Under the agreement, Generate will receive an upfront payment of USD 30 million, up to USD 35 million in early preclinical milestones, and is eligible for total potential milestones of nearly USD 1.2 billion, plus tiered royalties on worldwide net sales for two initial programs, with an option for a third. Generate will lead research through lead-candidate selection, while Zealand Pharma will handle clinical development, regulatory submissions, and commercialization.

• The collaboration unites Zealand Pharma's deep expertise in metabolic health with Generate Biomedicines' cutting-edge generative AI platform. This integration aims to accelerate the design and development of novel multispecific protein therapeutics, expanding Zealand Pharma's modality capabilities beyond peptides and small molecules into protein-based drugs for complex metabolic conditions.
• Generate Biomedicines stands to receive substantial financial compensation, including an upfront payment of USD 30 million and up to USD 35 million in early preclinical milestones. The agreement also includes potential development, regulatory, and commercial milestones totaling nearly USD 1.2 billion, alongside tiered royalties on global net sales. The collaboration initially covers two programs, with an option to add a third, focusing on obesity and other metabolic diseases.
• The agreement clearly delineates responsibilities: Generate will leverage its AI platform to lead research up to lead-candidate selection. Following this, Zealand Pharma will assume full responsibility for subsequent clinical development, regulatory submissions, and worldwide commercialization of any resulting products. This structure aims to efficiently advance promising drug candidates from discovery through market launch.

Overcoming Current Limitations in Obesity and Metabolic Disease Treatment

Obesity and its associated metabolic comorbidities — including type 2 diabetes, dyslipidemia, arterial hypertension, and cardiovascular disease — present a complex therapeutic landscape where no single intervention delivers universally durable outcomes. Current treatment modalities, spanning lifestyle modification, pharmacotherapy, and bariatric surgery, each carry distinct limitations that constrain their individual and combined effectiveness.

• Weight regain after bariatric surgery remains a significant clinical problem. In the long-term, a substantial proportion of patients after bariatric surgery experience a gradual weight regain that may, in some, reach up to a presurgical body weight, necessitating adjuvant anti-obesity pharmacotherapy to prevent recurrence or further potentiate weight loss.

• Pharmacotherapy requires sustained adherence, and weight regain occurs upon cessation. Pharmacotherapy is scalable and increasingly effective but requires sustained adherence, with loss of treatment-mediated control and weight regain upon cessation. It also does not have as extensive established research on safety as surgery. Semaglutide, for example, reduces muscle size and strength to the same extent as caloric restriction, and lean mass and fat mass rebound to baseline levels at the end of the withdrawal period.

• Optimal timing, duration, and combination of pharmacotherapy with surgery remain undefined. The impact of pharmacotherapy before and after bariatric surgery remains unclear, and the long-term safety and efficacy of these treatments should be assessed. Preoperative weight loss can be achieved through hypocaloric diets, intragastric balloon placement, or the addition of GLP-1 agonists to lifestyle interventions, but studies have not yet confirmed a positive effect on perioperative risk or postoperative weight development.

• Gastrointestinal adverse events limit tolerability and adherence to pharmacological agents. GLP-1 receptor agonists — among the most efficacious current agents — are widely associated with nausea, vomiting, diarrhea, and constipation, driven in large part by their effects on gastrointestinal motility, most notably delayed gastric emptying. Oral small-molecule GLP-1 receptor agonists similarly increase overall adverse events (RR = 2.75, 95% CI: 2.41 to 3.15), with gastrointestinal events predominating, though serious adverse events were not increased (RR = 1.15, 95% CI: 0.93 to 1.43).

• Patients with extreme obesity face disproportionately higher surgical risk with comparatively reduced benefit. Individuals with extreme obesity (BMI ≥ 50 kg/m²) have a high risk of surgical complications, mortality, comorbidities, and reduced weight loss following bariatric surgery, while lifestyle changes and anti-obesity medications are less effective than surgery in this population.

• No pharmacological agent fully replicates the multifactorial physiological impact of surgery. While obesity medications cannot fully replicate the multifactorial physiological impacts of metabolic/bariatric surgery, surgery induces superior short- and long-term weight loss via profound hormonal, neurological, and metabolic shifts, resulting in durable outcomes without ongoing intervention — though it remains difficult to scale.

• Evidence supporting specific drug combinations and sequencing strategies remains insufficient. Observational studies in the post-bariatric surgery population consistently demonstrate benefit from medical weight management, with most evidence highlighting liraglutide, topiramate, and phentermine/topiramate, but an evidence-based consensus has not yet been achieved because of the heterogeneity of populations studied and the studies themselves.

How Zealand and Generate Are Advancing Emerging MoAs

The past three years have seen a marked expansion beyond GLP-1 receptor monotherapy, with multi-hormonal and amylin-based mechanisms emerging as the most clinically advanced new directions in obesity and metabolic disease pharmacotherapy. Rational peptide engineering has enabled single molecules to engage multiple receptor pathways simultaneously, producing weight-loss efficacy that in some cases approaches that of bariatric surgery.

• Amylin receptor agonism (AMYRs 1–3): Amylin, a pancreatic β-cell hormone co-secreted with insulin, activates heterodimeric receptors composed of the calcitonin receptor (CTR) and receptor activity-modifying proteins (RAMP1, RAMP2, or RAMP3). Activation of these receptors in the brain regulates blood glucose, suppresses glucagon secretion, slows gastric emptying, and promotes satiation through both homeostatic and hedonic feeding circuits. Long-acting analogs — including cagrilintide, eloralintide, petrelintide, and NN1213 — are now in Phase II/III trials. Eloralintide, an AMY1R-selective agonist, demonstrated week-4 mean percent changes from baseline in body weight of -2.5% (p<0.01) and -4.4% (p<0.001) at single doses of 4 mg and 12 mg, respectively, versus +0.6% for placebo, with a pharmacokinetic profile supporting once-weekly dosing.

• Dual amylin and calcitonin receptor agonism (DACRAs): Parallel activation of both CTR and AMYRs is considered a more effective weight-loss strategy than selective amylin receptor engagement alone. Cagrilintide, the leading DACRA, adopts a "bypass" binding mode in both AMYR and CTR complexes, with key structural features including an F23 residue anchoring the peptide at the receptor transmembrane bundle, an E14-R17 intramolecular salt bridge enhancing helical stability, and C-terminal P37 interaction with the receptor extracellular domain. KBP-series DACRAs have additionally demonstrated improved insulin sensitivity and increased muscle-specific glucose uptake independent of weight loss in diabetic rat models. BGM1812, a novel DACRA optimized from petrelintide via structure-based drug design, showed significantly enhanced efficacy in both in vivo and in vitro studies.

• GLP-1/amylin combination and unimolecular multi-agonism (CagriSema, amycretin): Combining an amylin analog with a GLP-1 receptor agonist produces additive appetite reduction through separate but complementary mechanisms. CagriSema (cagrilintide co-formulated with semaglutide) has achieved greater effects in clinical trials than either component alone. Amycretin, a unimolecular AMYR/CTR/GLP-1R multi-agonist, has been developed for both weekly injection and as a once-daily tablet.

• GIP/GLP-1/glucagon triple agonism (retatrutide): Retatrutide (LY3437943) simultaneously targets GLP-1, GIP, and glucagon receptors. Phase 2 trials report weight reductions described as comparable to bariatric surgery, with additional benefits for metabolic comorbidities such as NASH and cardiovascular disease. The glucagon receptor component contributes to energy expenditure and hepatic lipid clearance; mechanistic work in long-acting glucagon analog models shows that glucagon receptor signaling enhances energy expenditure in obese — but not lean — mice through persistent cAMP/PKA signaling via PDE4B/4D downregulation.

• GLP-1/GIP/Y receptor triple agonism: A novel unimolecular triple agonist, GP-38, was engineered to simultaneously activate GLP-1, GIP, and the appetite-suppressing Y receptor. In diet-induced obese mice, GP-38 significantly outperformed GLP-1 receptor monotherapy in improving glucose homeostasis and reducing body weight, with semaglutide and cagrisema used as controls.

• Structural and peptide engineering strategies enabling these MoAs: Lipidation for long-acting properties, reversible albumin binding, targeted methylation strategies, and hybrid peptide construction (fusing N-terminal and C-terminal fragments from different endogenous peptides) have been central to advancing these mechanisms into viable drug candidates. Selective AMY1R agonism — as exemplified by eloralintide — may offer a gastrointestinal tolerability advantage over non-selective DACRAs, given that calcitonin receptor engagement and exposure kinetics influence whether reduced food intake reflects physiological satiation or aversive signaling.

Unlocking Novel Targets with Generative AI for Metabolic Diseases

Research into obesity and metabolic disease is advancing well beyond single-target approaches, with multiple receptor systems, neuropeptide pathways, and intracellular signaling axes now under active investigation. The convergence of gut hormone biology, central appetite regulation, and adipocyte-level mechanisms is reshaping the therapeutic landscape for type 2 diabetes and obesity.

• GLP-1/GIP/Glucagon Triple Receptor Agonism: Unimolecular triple agonists targeting GLP-1, GIP, and glucagon receptors have demonstrated supraphysiological potency in rodent models, with efficacy superior to dual coagonists and best-in-class monoagonists in reducing body weight, enhancing glycemic control, and reversing hepatic steatosis. The mechanism integrates synergistic glucagon action to increase energy expenditure, GLP-1 action to reduce caloric intake and improve glucose control, and GIP action to potentiate the incretin effect and buffer against the diabetogenic effect of inherent glucagon activity. The GLP-1/GIP co-agonist tirzepatide has been approved by the U.S. Food and Drug Administration (FDA) for the treatment of type 2 diabetes, facilitating weight loss of up to 22.5% in non-diabetic individuals living with obesity — comparable to certain types of bariatric surgery.

• Dual GCGR/GLP-1R Agonism (BI 456906): BI 456906, a 29-amino acid peptide derived from glucagon incorporating potent GLP-1 activities, represents a novel dual GCGR/GLP-1R agonist. It contains a C18 diacid mediating albumin binding to enable once-weekly subcutaneous dosing. GCGR agonism aims to enhance body weight-lowering effects by increasing energy expenditure in addition to the anorectic action of GLP-1R agonists. Glucose-lowering efficacy has been demonstrated in a Phase II trial in patients with type 2 diabetes mellitus and obesity, associated with clinically meaningful body weight loss.

• Combination Neurohormonal Strategies — Amylin and PYY[3-36]: Combined administration of amylin and PYY[3-36] produced greater anorexigenic and weight-reducing effects than either peptide alone in diet-induced obesity rodent models. A single intraperitoneal injection of amylin (10 µg/kg) plus PYY[3-36] (1000 µg/kg) reduced food intake for 24 hours, while 14-day subcutaneous infusion suppressed food intake and body weight to a greater extent than monotherapy. Food intake suppression was synergistic and body weight reduction was additive, with amylin plus PYY[3-36] (but not monotherapy) increasing 24-hour fat oxidation.

• AgRP/PBN CGRP Neural Circuit Modulation: Agouti-related protein (AgRP) neurons in the hypothalamus and calcitonin gene-related peptide (CGRP)-expressing neurons in the parabrachial nucleus (PBN) represent a functionally interacting circuit governing appetite. AgRP neuron stimulation reduces Fos expression in PBN CGRP neurons and can overcome appetite suppression mediated by amylin, cholecystokinin (CCK), and lithium chloride. Projections from AgRP neurons to the PBN are sufficient to increase feeding, identifying this circuit as a potential target for modulating pathological appetite drive.

• CGRP Peripheral Signaling in Metabolism: Intraperitoneal CGRP administration in C57Bl/6 wild-type mice at 2 nmol significantly reduced food intake, respiratory quotient (RQ), VCO₂, and VO₂, and lowered total energy expenditure (TEE). CGRP treatment also induced significantly lower plasma levels of glucagon and leptin but higher levels of amylin, confirming a role in appetite and metabolic hormone regulation beyond its established cardiovascular and pain functions.

• Hippo-YAP1/FGF21 Axis in Adipocytes: Deletion of MOB1 — a hub of the Hippo core molecule complex that negatively regulates YAP1/TAZ activation — specifically in mature adipocytes (aMob1DKO mice) conferred resistance to diet-induced obesity. YAP1 activation in mature white adipose tissue increased basal lipolysis, "beiging," and energy expenditure, suppressed ROS production and inflammation, improved insulin sensitivity and glucose tolerance, and reduced ectopic fat accumulation. FGF21, which improves lipid metabolism, was upregulated directly via YAP1 activation, and many phenotypes in aMob1DKO mice were dependent on FGF21, identifying a YAP1-FGF21 axis in adipocytes as a potential therapeutic target.

• Multi-Target Lipid Metabolism Modulation via EFAX: The ethyl acetate fraction of Amomum xanthioides (EFAX) ameliorated hepatic steatosis in a high-fat diet mouse model through activity on multiple hepatic proteins including sterol regulatory element-binding protein (mSREBP)-1c, peroxisome proliferator-activated receptor (PPAR)-α, AMP-activated protein kinase, and diglyceride acyltransferases (DGATs), suggesting a multi-target approach to nonalcoholic fatty liver disease (NAFLD) spanning steatosis to fibrosis.

Generative AI Powers Next-Gen Obesity Therapeutics

The landscape of metabolic disease treatment is undergoing a profound transformation, driven by an urgent need for more effective therapies to combat the global obesity epidemic. While current GLP-1 receptor agonists have revolutionized weight management, the scientific community is actively exploring next-generation approaches, particularly those involving multispecific agonism, to achieve even greater weight loss and broader cardiometabolic benefits. This is where the strategic collaboration between Zealand Pharma and Generate Biomedicines becomes particularly compelling.

By combining Zealand's deep expertise in metabolic health with Generate's cutting-edge generative AI platform for protein engineering, this partnership aims to unlock new possibilities in drug discovery. The power of generative AI lies in its ability to design novel protein structures and binders de novo, moving beyond traditional methods to create highly specific and potent therapeutic candidates. This technology is particularly adept at tackling complex challenges, such as designing proteins that can bind to flexible targets or modulate multiple signaling pathways simultaneously, which is crucial for developing effective multispecific agonists. The literature highlights the potential of such approaches, from dual GLP-1R/GIPR agonists like tirzepatide to triple agonists and biased agonists optimized for specific signaling pathways, all promising enhanced efficacy.

However, this ambitious endeavor is not without its inherent complexities. Designing truly functional de novo proteins with exquisite control over their dynamics and multi-state interactions remains a significant scientific hurdle. The journey from an in silico design to a clinically viable drug also requires robust and scalable experimental validation, a process that can often become a bottleneck. Furthermore, carefully optimizing the pharmacological profile of novel multispecifics, including their biased signaling properties and receptor internalization kinetics, is critical to ensure a favorable balance of efficacy and safety in human trials. Despite these challenges, the integration of AI into drug discovery represents a paradigm shift, offering the potential to accelerate the development of highly tailored, potent, and differentiated therapies that could redefine the standard of care for obesity and other metabolic diseases. This collaboration underscores a future where computational design plays an increasingly central role in delivering groundbreaking medicines.