From The Trenches

By Arthur Tzianabos, CEO of Lifordi Immunotherapeutics, as part of the From The Trenches feature of LifeSciVC

As biotech companies tackle the escalating time and cost of clinical trials, the decision to conduct early-stage, proof-of-concept studies outside the U.S. is becoming more common. While this approach is often faster and sometimes cheaper, it should not be the sole driver. Instead, there should be a thoughtful, deliberate set of strategic discussions within the company and its board that considers many factors and requires a careful benefit/risk assessment.

Here, I want to highlight some of the considerations and challenges (expected and unexpected) for those of you who are contemplating this endeavor.  As Bruce Booth noted in his 20 year reflections last fall, “the single most important value creation activity in biotech is excellent clinical trial execution ”─so getting this right is critical.

Conducting clinical trials abroad will impact every function in the organization. Some things can be anticipated, such as managing different time zones, language barriers, diverse patient populations, materials and drug supply, and supply chains. However, it is the unexpected issues that can have the greatest impact on the success of the trial and in some cases the entire company. This harkens me back to the old mantra, “Don’t let the trial fail the drug,” a truth that may be even more fitting when conducting clinical trials in other regions of the world.

Here, I have collected insights from my own experiences and from a few of my biotech colleagues to provide additional perspectives. Deciding which examples to include was not easy. Capturing the ones that can be anticipated but may be underestimated, as well as the unexpected and costly ones, should serve as good reminders and a heads-up for all of us.

Speed and Cost of Trials

The cost of ex-US clinical trials is generally considered to be much less, somewhere on the order of 30-50% cheaper (though not 70-80% as some might think). I have yet to meet a company that conducted trials outside the U.S. and correctly budgeted for travel and other ancillary costs. The frequency with which employees will need to visit clinical sites, key vendors or manufacturing plants can be surprising. Unforeseen situations will arise and the unanticipated, last-minute flights and extended stays are expensive. Additionally, a company may need to provide standard of care (SOC) medicines that are often not available in other countries, and/or offer post-treatment therapies, both of which drive up clinical trial costs. Establishing an in-country presence, whether it is someone ‘parked’ there or hired as your own company representative, may seem like a luxury. However, it is a highly recommended strategy that ultimately can save time and money. When running trials in the EU, companies also need to sponsor a qualified person to ensure the quality of materials and facilities meet GMP standards. Local regulators will examine materials and want either compendial materials or locally sourced ones. A company will need to know and prepare for this ahead of submitting the required investigational product paperwork. Since most clinical plans evolve later on, it is critical to cast a wide net in the process.

Tax incentives for conducting clinical trials in certain territories can be very attractive.  Understanding the details of these is important because they can impact a company’s cash runway and change fundraising plans. Several companies have recounted experiencing ‘tax breaks by two offices.’ This relates to a particular region of the world where one business office distributed the money from tax incentives while another ensured all requirements were met. It became clear that the offices operated independently and didn’t always communicate the same information. As a result, the money received was far less than anticipated, or a significant portion of the money needed to be returned.

These are just a few of the cost considerations when deciding to conduct clinical trials outside the U.S., however, there are other important factors that might influence this decision.

Patient Populations

Studying a more geographically diverse patient population in a clinical trial offers broader safety and efficacy evaluation as well as an opportunity to establish relationships with a global network of regulators and key opinion leaders. This also helps pave the way for broader commercialization beyond the U.S.  Accessing patients might be easier, faster, and less expensive in some places. Accounting for ethnic, genetic, cultural, lifestyle differences, pre-existing conditions and environmental exposures is also required. The “all medicine is local” aspect to this can become apparent during screening and enrollment. In one particular example, it wasn’t until a company had designed the protocol, selected a clinical research organization (CRO), and identified clinical trial sites, that it learned about a genetic mutation native to the area at reasonable background levels that it could predispose healthy participants to screen-fail at a much higher rate. Fortunately, the trial hadn’t started, but it required a protocol amendment, ensuring that a local lab could perform a non-standard test, and increasing the number of patients to be screened.

Holding frequent and detailed meetings with CRO leadership and the teams assigned to your study, clinical investigators and trial site coordinators before a protocol is finalized is critical to avoid unexpected issues like this. It could have saved another company from learning too late that the assay it developed to measure an enzyme involved in fat metabolism did not perform well in a local population whose diet preference was eating sausages at every meal. Companies need to understand how local populations live, receive care, and how those factors might influence trial outcomes.

The Practice of Medicine and Clinical Endpoints

As a condition of approval, FDA typically requires that companies demonstrate that the clinical data generated ex-US is representative of the patient population and the practice of medicine in the U.S.  While healthcare is managed and reimbursed differently throughout the world, the details are often underappreciated. Depending on the region, health records might be unavailable, incomplete, or unreliable, particularly where care is not tightly monitored. Assessing health and compliance with SOC can be challenging.

Relying solely on CROs to ensure protocols are being followed and all aspects of a trial are well managed is one of the biggest mistakes companies can make. While CROs and trial sites are responsible for overseeing compliance, one company running an ex-US trial in a region where compliance is particularly challenging decided to administer a test to determine whether study participants were taking the SOC medicine during the trial screening period. This would help determine patients who are likely to be compliant during the study and minimize the risk of enrolling the wrong patients. Companies might also pay for crossover treatments, especially for Phase 1 trials, thus providing SOC treatment for a full year post study.

To complement CRO efforts, some companies choose to have ‘boots on the ground’ for key trial events such as first dose(s), especially for first-in-human (FIH) studies. It is expensive but often yields the greatest ROI. As issues arise, extended travel can’t be avoided. It drains employees and organizations when people are away for weeks or a month at a time, particularly for emerging biotechs who are stretched thin and people wear multiple hats. The best advice is to identify potential needs ahead of time. This alleviates pressure, maintains strong relationships inside the company and with ex-US partners, and minimizes clinical trial risks longer term.

Building trust with clinical trial partners in certain foreign countries can be harder and take longer. It is also not something that should be left solely to the CRO. Getting to know the people, cultural norms and communication preferences helps to bridge the distance. It can take months to develop strong enough relationships in some countries compared with others where Zoom meetings are enough to garner the trust needed to get clinical trials up and running.

Establishing clinically meaningful endpoints may vary across regions. There can be different measurement ranges that reflect clinical practice and these complicate data analyses. Understanding exactly how endpoints will be measured and patient outcomes will be recorded is essential for trial success. Language barriers, treatment regimens, diagnostics capabilities ─including imaging and videos, and lifestyle differences can all lead to inconsistent interpretations and assessments. One biotech company with global trials involving over 7,500 patients and ~600 clinical trial sites had to align on big-ticket items with multiple regulatory agencies, clinical sites, and ethics committees across regions. Country-specific amendments were expected and accepted.

Recently, FDA has asked at least two companies for a placebo arm requiring some form of an invasive surgical procedure for their pivotal trials. In each case, these are in rare disease patient populations that require delivery of drugs to the central nervous system.  In one case (with which I am familiar), several European countries deemed this approach unethical, which required additional placebo patients to be added to the study to perform a less invasive surgery.  The lack of alignment between regulatory agencies continues to be a major challenge.

Legal and Regulatory Considerations

As noted above, regulatory requirements vary by country and region, but a clinical development strategy should be global from the outset. Aligning study design and protocols with appropriate regulatory expectations and operational realities is critical to avoid delays, cost overruns, and complicated data interpretations. Regulatory filings outside the U.S. are often less extensive than full investigational new drug  (IND) submissions and review timelines may be shorter, however early engagement with the FDA is important. Even when initial clinical proof-of-concept data will be generated ex-US, conducting a pre-IND meeting before completing GLP/GMP toxicology studies helps ensure the nonclinical package aligns with their feedback and meets FDA expectations. Unfortunately, some companies have encountered significant delays presenting positive ex-US clinical data to the FDA and learned afterwards that additional toxicology work would be required.

International trials must also comply with country-specific legal standards governing data privacy, confidentiality, and cross-border data transfer. In the EU, data protection regulations differ significantly from HIPAA requirements in the U.S. and require companies to appoint a Data Protection Officer to oversee compliance. Understanding the requirements and incorporating them into clinical trial applications and site selection decisions can prevent delays, penalties, and incurring unexpected costs later on.

Manufacturing and Supply Chain Logistics

It is considered good practice for Chemistry, Manufacturing and Controls (CMC) company personnel to be present for drug production runs. While this requires additional time and expense, experienced operators believe this is non-negotiable.  At my company, Lifordi Immunotherapeutics, I had a senior member of my CMC team observe our first production runs in Scotland. Due to a number of issues, my colleague spent several late nights eating dinner from an all-night gas station food market but was thankfully able to manage through the process before it became a major problem for us.

Supply chains and shipping logistics for ex-US trials are complex. It can take a village ─or large network of reliable vendors ─to achieve success. In one instance, a clinical site near a war-torn territory necessitated that materials be flown to a neighboring country and driven across the border to the clinical trial site. Patient samples needed to take a similar route for proper analyses. Maintaining temperatures for drug product, sample storage, and delivery was a top priority. Companies routinely manage these logistics by keeping ahead of things such as new labelling requirements ─a situation that is known to cause delays and incur further expense. Today, rapidly changing import/export rules and imposed tariffs are further complicating supply logistics. New regulations often impact regulatory filings and while the magnitude of changes can be difficult to predict, the potential time and cost to comply must be budget.

Destinations for Ex-US Clinical Trials

Governments, pharma companies, and other industry partners continue to invest in building critical infrastructure overseas including training and creating substantial talent pools. For many years, conducting clinical trials outside the U.S. included locations in Europe, Australia, Japan, and Latin America. Today, Asia-Pacific is the fastest growing region with major investments in China made by pharma companies including a $15B investment by AstraZeneca and a ~$1B investment by Pfizer ─both through 2030, and more recently, a $484M investment in South Korea over the next five years by Roche.

China has increased its relative share of clinical trials by 57% from 2019 to 2023, and in 2024 it was the third most utilized region mainly due to its large patient population demanding new therapies, technological capabilities, cost-effective solutions, and favorable regulatory policies.

While this increase was mostly driven by the explosion of clinical work from China-based companies, trial recruitment in China can also be 5-10 times faster than it is in the U.S. and averages one patient per month for some indications. For these reasons, China is increasingly a topic of discussion by Western companies for FIH studies.

Currently, most US companies are doing FIH trials in Australia (including Lifordi), New Zealand, Belgium or the Netherlands.  Timelines remain rapid in these countries. For healthy participant studies, it takes approximately 2-3 months from final protocol and investigator’s brochure through to dosing first patient. SAD/MAD studies can be completed with 4-week follow-up within about 6 months. Healthy participant data is also readily accepted by the FDA.

Conclusion

As companies weigh the risks and benefits of conducting clinical trials outside the U.S., rapid data generation and potential cost-savings often rise to the top. Other substantial advantages include access to diverse populations and broader access to innovative medicines for patients. Yet these benefits come with expected and unexpected challenges. Companies that invest time and resources to develop local knowledge and build strong relationships are far better positioned to achieve success and avoid costly setbacks. As the old mantra reminds us, careful planning and execution is key—because no company wants the trial to fail the drug.

Comment

By Cody Tranbarger, Entrepreneur In Residence at Atlas Venture, as part of the From The Trenches feature of LifeSciVC  

What if the biggest obstacle to the next generation of I&I therapeutics isn’t finding better targets — but accepting that one target was never going to be enough? It’s a question the industry spent twenty years not asking. The Humira playbook was too lucrative, the growth too easy, and the incentives too misaligned to seriously reckon with biological reality. Now, with the single-target era approaching exhaustion and an efficacy ceiling that fifty antibodies have failed to crack, biopharma has finally embraced multi-drug and multi-target interventions in its quest to reset the bar. 2026 promises to be a defining year for these efforts, and therefore a defining year for the future of I&I.

How’d we get here in the first place? The story starts in 1998 with the approvals of Enbrel and Remicade, the first-ever I&I biologics. The two drugs’ early success catalyzed a “biologics boom” that would define the modern era of immunology. By 2002, both surpassed $1 billion in annual revenue, becoming the first autoimmune-indicated biologics to achieve “blockbuster” status. That same year, Humira was approved for the treatment of Rheumatoid Arthritis (just barely, on December 31st – and for my biotech trivia enthusiasts, that’s only happened twice since). Humira’s approval, the first for a fully human monoclonal antibody, was rightfully celebrated as a major scientific milestone. Wall Street, meanwhile, wasn’t nearly as enthusiastic. Analysts viewed the anti-TNF class as crowded, questioned Humira’s ability to differentiate versus the chimeric antibody Remicade (despite the former’s pre-filled syringe for convenient, Sub-Q, at-home dosing – brings to mind a few oft-criticized players in I&I today, but I’ll leave it there), and criticized Abbott for recklessly overpaying in its $6.9 billion deal for the drug. Abbott itself, despite emphasizing Humira’s multi-indication potential to justify the price tag, was tepid in its forecasts, pegging RA peak sales at just $1 billion.¹

23 years, 11 indications, and $240 billion of revenue later, Humira has proven as important a milestone for the biopharma business model as it was for science. Along the way, Humira became the archetype “pipeline-in-a-product” and dogmatized the business model as a core strategy across the industry. Today’s best-selling I&I biologics – Dupixent, Skyrizi, Stelara, Cosentyx, etc. – all followed this blueprint en route to multi-blockbuster status. Throughout the same 23-year period, an explosion of novel mechanisms and protein engineering technologies has fueled the I&I space through the approval of nearly 50 antibodies that collectively modulate more than 15 distinct targets spanning the full breadth of canonical immune pathways.² Collectively, these medicines have reached tens of millions of patients, and many have redefined standard-of-care in clinical practice.

The cumulative innovation in I&I has been so substantial that, for a subset of indications, some argue we’ve reached “endgame” status. In Psoriasis, for instance, Bimzelx’s 60%+ PASI-100 sets a daunting benchmark. Among the majority, however, unmet need remains. Novel mechanisms have improved patient outcomes by enabling class switching and multi-line intervention but have largely failed to deliver step-change efficacy. In most indications, particularly those characterized by greater severity and heterogeneity, the clinical high-water mark has stubbornly plateaued. In Inflammatory Bowel Disease (IBD), the bar has scarcely budged from ~25-30% Clinical Remission (all benchmarks based on placebo-adjusted, induction period endpoints). In Rheumatoid Arthritis (RA), ACR50 has plateaued at ~30-40%. In Axial Spondyloarthritis (AxSpa), ~30-35% ASAS40, in Hidradenitis Suppurativa (HS), ~40-45% HiSCR50, and so on… Or, more succinctly, the “Efficacy Ceiling.”

Underlying biology helps explain this ceiling, to a point. Autoimmune diseases are not monolithic entities driven by single pathways. Rather, they are complex ecosystems of dysregulated cytokines, aberrant cell populations, and tissue-specific factors that vary across patients and even within the same patient over time. Blocking one pathway can provide meaningful benefit but rarely addresses the full constellation of disease-driving mechanisms. After many millions of years in a co-evolutionary arms race, redundancy is a feature, not a bug.

Where biology’s explanation stops, market structure and commercial incentives take over. For the last two decades, biologics in I&I have enjoyed a rising tide – in a set of markets with very low biologics penetration historically, the primary growth lever was simply expanding access to the class. When “good enough” is lower risk and highly lucrative, the rational economic actor naturally allocates fewer resources toward high-risk, paradigm-shifting projects. Today, however, the incentives have shifted – biologic penetration has plateaued at ~60% in the largest indications, biosimilar adoption is finally gaining momentum, and fierce competition amongst an ever-growing arsenal of options has fragmented the branded market. The remaining I&I white space sits behind the glass ceiling, and much of biopharma has resolved to break through.

Most of those players have since coalesced around biologics combinations and/or multi-specific antibodies as their preferred strategy. Look no further than Big Pharma commentary during this year’s JP Morgan Healthcare Conference. J&J, Sanofi, and UCB all highlighted co-formulations and/or multi-specifics as a core pillar of their I&I strategy. Abbvie positioned Skyrizi as the anchor asset for I&I combinations, and projected sustained long-term growth for the franchise. Regeneron unveiled an IL-4 x IL-13 bispecific at the center of its Dupixent LCM strategy, notable given leadership’s frequent disparagement of “me-better” assets (including, with as much earnest zeal as ever, during this very same presentation).

Moreover, the industry pipeline reflects those comments – in a late 2025 analysis of >100 clinical-stage Big Pharma I&I assets, ~70% were biologics, and within this subset, ~25% were combinations or bispecifics.³ Not everyone is sold, however. Equally interesting are those still on the sidelines, a list that includes some of the largest I&I franchises today, such as Novartis, Roche, Eli Lilly, and AstraZeneca.

So, who’s right? 2026 should reveal much of the answer. In contrast to a smattering of one-off datapoints in recent years, this year’s dense calendar of clinical catalysts should finally provide a sufficient corpus of data to enable a more grounded debate as whether combos and multi-specifics are the next chapter in I&I or merely an addendum to an already crowded therapeutic narrative. To contextualize this inflection point, one must understand how we got here, and what we’ve learned along the way.

Target & Modality Selection – Insights from the Bench to Bedside

Not all combinations are created equal. History is littered with examples of biologically rational combinations that failed to translate, most attributable to underestimated toxicity overlap or overestimated mechanistic orthogonality. The rationale for supporting a target combination should be built from a foundation of first principles: have the targets / pathways been shown to contribute independently to disease pathophysiology? Is there evidence of non-redundancy? Can mechanistic synergy be demonstrated preclinically?

For clinically validated mechanisms, an integrated analysis of basic biology, human genetics, preclinical models, and clinical data tends to spit out clear answers to these questions; and in my view, the answers are often “no.” If the explosion of cytokine target combinations across biopharma pipelines is any indication, that will be a controversial take, but I think it’s important to keep the bar high, because that signal is much harder to see through the biological complexity and heterogeneity of real-world patients.

From there, practical translational considerations can be overlaid to inform modality selection. As important as the targets themselves are the tools employed to modulate them. The relative virtues of the two dominant strategies in I&I – antibody co-formulations and bispecific antibodies – are hotly debated, often pitted against each other in a zero-sum, winner-take-all competition. More likely, in my view, is that both ultimately find a place in the armamentarium, with their respective territories defined by those sets of targets for which each is uniquely positioned. For example, the relative abundance and localization of the two targets are critical variables. Traditional bispecific antibody formats may be well-suited for a combination of two soluble targets typically observed at comparable concentrations in the serum, whereas a bispecific may underperform its co-formulated monoclonal constituents for a combination in which one soluble target’s concentration exceeds the other’s by several orders of magnitude, or one in which the TMDD of a membrane-bound target precludes saturation of its soluble counterpart.

Of course, this choice also has downstream implications for discovery and development. On average, bispecifics are harder to discover (incremental degrees of freedom necessitate careful engineering and multiparametric coordination), harder to enhance (Fc mods like YTE perform inconsistently), and harder to make (light chain mispairing reduces yield and necessitates sophisticated purification processes). Conversely, antibody co-formulations are much harder to develop – a consequence of statutory requirements to demonstrate the contribution of each component.

J&J’s Phase 2b DUET studies in IBD illustrate this dynamic perfectly. To comply with the above, each trial has six treatment arms: placebo, each monotherapy, and three doses of the combination. To maintain power, they enrolled ~575 and ~700 patients, more than twice the size of an average single-agent Phase 2 in IBD historically; and even now, at any clinical remission delta below 20%, the trials are less than 80% powered for dual superiority.⁴ Naturally, enrollment scope cascades through timelines – both DUET trials enrolled in ~22 months, ~6-9 months longer than historical benchmarks – and cost – at least twice that of historical benchmarks, conservatively. For more granular look at cost, keep an eye on Spyre’s P&L in 2026 and beyond. The Phase 2 SKYLINE platform trial looks a lot like DUET-UC, and those costs will become increasingly visible as the placebo-controlled portion begins enrolling later this year.

The best target combinations are those for which scientific rationale and translational tractability are airtight – when genetics, mechanistic biology, preclinical models, and interventional pharmacology all converge on orthogonal, non-redundant pathways that are imminently druggable with well-understood tools. Today, this framework also benefits from the efforts of early pioneering efforts to bring antibody combinations and multi-specifics to I&I patients. Though many failed, and most of those that haven’t remain early and immature, valuable translational insights have already emerged from a decidedly mixed set of early clinical outcomes.

You might be surprised (I was) to learn that randomized, controlled trials of antibody combinations in I&I date as far back as the early 2000s. Between 2004 and 2007, results were published from three multicenter, 100-plus patient RCTs testing combinations of Enbrel (anti-TNF), Kineret (anti-IL-1Rα), and Orencia (CLTA4-Ig) for RA. Across the board, it was worst case outcome – the combinations not only failed to demonstrate superior efficacy, they were also associated with 2-3x higher rates of SAEs and serious infections.⁵ Not a great start, but the data points weren’t universally negative – around the same time, a smaller RCT testing the addition of Tysabri (anti-α4 integrin) to a steady-state regimen of Remicade (anti-TNF) in Crohn’s patients failing to achieve remission with monotherapy. Most efficacy measures numerically favored the combination, particularly among high-risk patients; and importantly, no incremental safety risk was observed.⁶ Despite that, one small, inconclusive win stood little chance against the new narrative. The prevailing view was so negative that ACR went as far as updating treatment guidelines with an explicit recommendation against dual biologic therapy in RA based on unfavorable benefit-risk.⁷ Although experimentation from the backwater would continue to emerge through case reports and single-arm studies, the biologic combination hypothesis had largely been put on ice.

The concept of biologic combinations and multi-targeting emerged again in the mid-2010s with the advent of bispecifics in I&I. In much the same way as the trials above, these pioneering efforts universally fell short while teaching the field important lessons along the way. Two programs, Covagen’s COVA322 and AbbVie’s ABT-122, entered the clinic around the same time. Both targeted TNF and IL-17A, informed by preclinical validation and observations of elevated IL-17 and Th17 cells in TNF-treated patients. The former failed too quickly to test the hypothesis. The latter was evaluated in separate RA and Psoriatic Arthritis (PsA) Phase 2s, head-to-head against Humira. In neither case did ABT-122 outperform Humira. Clean attribution of that outcome, however, proved challenging. In particular, high rates of neutralizing ADAs were thought to have precluded a true test of the biological hypothesis.⁸ A subsequent post hoc exposure-response analysis across both trials determined that, at comparable molar exposures, there was no differentiation in efficacy, and therefore no detectable contribution of IL-17 suppression.⁹

In the aftermath of ABT-122, the bispecific hypothesis largely sat dormant in I&I, despite significant advancement in other therapeutic areas. It wasn’t until Bimekizumab (anti-IL-17A+F), now Bimzelx, established definitive proof-of-concept for dual cytokine neutralization in 2018-19 that the concept of multi-specificity was revisited en masse.

Shortly after Bimekizumab blew open the door for multi-specifics, J&J’s VEGA trial readout did the same for antibody combinations. VEGA evaluated the combination of anti-IL-23 Guselkumab (Tremfya) and anti-TNF-α Golimumab (Simponi) vs. each monotherapy in predominantly biologic-naïve Ulcerative Colitis patients. This combination, in addition to leveraging the trust and reliability of IBD’s two grizzled veterans, capitalized on a clean and compelling biological hypothesis. IL-23 had increasingly been implicated in the development of anti-TNF resistance, and a host of preclinical models had shown clear mechanistic synergy of intervening downstream via TNF and upstream via IL-23 simultaneously. First announced in 2022, VEGA finally notched a win for antibody combinations in I&I. At week 12, the combination achieved 83% clinical response, 8-12% higher than the monotherapy arms. More impressively, the combination achieved clinical remission of 47%, 22-23% higher than the monotherapy arms.¹⁰ Unlike first-generation strategies, the combination possessed comparable safety profile to the monotherapies. The ceiling had been broken, emphatically and safely.

Momentum only built further from there. In 2023, provocative early Ph1b Asthma data from Sanofi’s Lunsekimig (IL-13 x TSLP) established a compelling first proof point in Th2-driven disease. Later that year and into 2024, J&J’s acquisition of two peri-clinical bispecific assets for a cumulative $2B drove sentiment to a fever pitch.^11,12 Companies were started (probably too many), billions of dollars traded hands in collaborations and licensing agreements (mostly to China), and the cytokine bingo card was soon full. The die had been cast.

2025 marked the first year with multiple, meaningful multi-specific readouts, a representation of the maturation of this broader theme. It was also a reality check, with several of the assets that had fueled the industry’s excitement failing to live up to expectations. First, the full Phase 1b Asthma dataset for Sanofi’s Lunsekimig (IL-13 x TSLP) revealed a much more equivocal package than assumed. The previously disclosed and highly impressive FeNO signal was, in fact, the outlier. The drug failed to achieve broader anti-inflammatory effects such as deeper eosinophil suppression or improvements in lung function, raising doubts about its ability to raise the bar in its ongoing Asthma Phase 2.¹³

Subsequently, J&J terminated a Phase 2b Atopic Dermatitis trial evaluating JNJ-4939 (IL-4Rα x IL-31) after an interim analysis failed to meet the internal threshold for efficacy (likely superiority versus Dupixent).¹⁴ With how well-understood the itch-scratch dynamics are in AD, IL-4Rα and IL-31 were generally considered among the most rational, low risk target pairs. J&J’s willingness to spend $1.25B to acquire the asset in the first place exemplifies that. There’s much more to learn in the fullness of time, but the membrane x soluble target pair compatibility may have played a role. At the very least, this outcome was a timely reminder that the consensus low-hanging fruit probably isn’t as easy to pick as it seems.

On the other hand, the space also enjoyed a few notable successes. UCB’s Galvokimig (IL-13 x IL-17A/F) produced one of the more striking proof-of-concept data points in 2025. Intended to suppress both the Th2 pathway (via IL-13) and Th17 pathway (via IL-17A/F), Galvokimig aims to address a broader and more heterogenous AD population. Initial Phase 2a data presented at EADV 2025 garnered a lot of interest, for good reason – at 12 weeks, Galvokimig achieved 53% and 43% placebo-adjusted EASI75 and EASI90, respectively.¹⁵ The latter compares favorably Dupixent’s Phase 2b performance and suggests the mechanism may be accomplishing exactly what it intended, but such a small trial should not be overinterpreted.

Sanofi’s Brivekimig (TNF-α x OX40L) grabbed headlines at the same conference. The rationale underlying Brivekimig is the simultaneous modulation of two distinct but complementary inflammatory nodes – TNF drives acute tissue damage and immune cell recruitment, while OX40L-OX40 signaling promotes effector T-cell survival and proliferation that sustains chronic inflammation. The initial data from an HS Phase 2a impressed, achieving 29% and 32% placebo-adjusted HiSCR50 and 75, respectively, comparable to those achieved by Bimzelx in a similar setting.¹⁶ Again, caveats abound – small trial, biologic-naïve patients, high placebo response – but a strong start, nonetheless.

The 2026 Calendar – Key Catalysts & Potential Learnings

Against this backdrop, 2026 is shaping up to be a defining year for the combination / multi-specific thesis in I&I. There are several key readouts we’ll be watching closely:

J&J’s JNJ-78934804 (IL-23 x TNF-a) DUET-CD & DUET-UC – the largest, most robust test of the combination thesis to date. Building on VEGA’s proof-of-concept, the long-awaited DUET readouts will determine if a similar benefit can be achieved in a refractory population, and with a much more patient-friendly fixed-dose co-formulation. Many were hopeful to see these data last year, but J&J has remained cagey on timing. With both trials reaching their Primary Completion Dates in May 2025, data are almost certainly imminent.

AbbVie’s Skyrizi Combo Platform Trial (IL-23 x a4b7 & IL-23 x IL-1A/B) – relative to the attention DUET receives, the Skyrizi Platform trial is massively underappreciated. It’s a DUET-esque trial – 500 patients, biologic refractory – likely to provide a robust read on two of the most important target pairs in IBD. One of these, IL-23 x a4b7, will have much broader implications, particularly for the fate of Spyre’s pipeline.

Apogee’s APG279 (IL-13 + OX40L) in Atopic Dermatitis – first of all, any Phase 1b that’s set-up to demonstrate superiority over King Dupixent is a must-watch. That said, within the Th2 category, APG279 also stands out as one of the only co-formulation efforts amongst a sea of multi-specifics.

UCB’s Donzakimig (IL-13 x IL-22) in Atopic Dermatitis – beyond the buzz of Galvokimig, UCB’s bispecific portfolio includes Donzakimig, which pairs IL-13 with IL-22, a cytokine involved in skin barrier integrity. IL-22 hasn’t been a popular target for multispecifics, despite independent proof-of-concept for anti-IL-22 antibodies in AD and clear pathway orthogonality. Phase 2a AD data later this year will put this hypothesis to the test for the first time, and that’s always worth watching.

Sanofi’s Lunsekimig (IL-13 x TSLP) in Asthma & Atopic Dermatitis – on the back of Sanofi’s creative disclosure choices for the Phase 1 Asthma data, data from a large, multi-dose Phase 2b are sure to separate signal from noise. The latter is of less interest given TSLP’s lackluster data in AD historically.

Pfizer’s Tilrekimig (IL-4 x IL-13 x TSLP) & PF-07264660 (IL-4 x IL-13 x IL-33) in Asthma & Atopic Dermatitis – Pfizer’s “trispecifics” have been a topic of intrigue following Tilrekimig’s progression into the second stage of the ongoing Phase 2 platform trial in AD. Pfizer has thus far declined to disclose any of the data supporting this decision, but they learned enough to move Tilrekimig into a separate Phase 2b in Asthma. 2026 should bring clarity.

Others – Bambusa’s BBT001 (IL-4Rα x IL-31) in Atopic Dermatitis (a first clean second shot on goal for a failed pair of targets), Aclaris’ ATI-052 (IL-4Ra x TSLP) in Asthma and AD (additional insight into the performance of membrane-bound x soluble bispecific constructs), and Zura’s Tibulizumab (IL-17A x BAFF) in HS (one of few programs testing parallel T and B cell modulation).

Through the I&I Looking Glass – Fad or Future?

2025’s mixed track record injected a healthy dose of humility into what had become an increasingly consensus bull case. The setbacks and successes alike reinforce the importance of target selection, yet remind us that biology is unpredictable and attrition is inevitable.

While I suspect the list of truly synergistic, efficacy ceiling-breaking target pairs is shorter than most expect, I firmly believe combinations and multi-specifics will be a defining theme of the next decade in I&I. Those that truly raise the bar will reshape treatment paradigms, redefine efficacy expectations, and create significant value for patients and investors alike.

The broader Atlas team has explored these dynamics equally deeply. The fund’s recently announced investment in Caldera reflects conviction in this broader theme, the TL1A x IL-23 hypothesis specifically, and the capacity of an extraordinary, experienced team to out-execute in a hyper-competitive space.

2026 will not provide all the answers, but it will be defining. For the first time, the data will be dense enough, the trials large enough, and the target diversity broad enough to move beyond anecdote and toward empiricism. The next chapter of I&I is being written as we speak – stay tuned.

References

1. Abbott Laboratories. Annual Report 2002: Form 10-K for the Fiscal Year Ended December 31, 2002. U.S. Securities and Exchange Commission, 19 Feb. 2003.
2. The Antibody Society.YAbS: Database for Therapeutic Antibodies.
3. Stifel Biotechnology Equity Research. 2025 I&I Strategic Pipeline Deep Dive: Mapping Out Key Areas of Focus and White Space Opportunities. Stifel, 10 Sept. 2025.
4. European Union Clinical Trials Register. A Phase 2b Randomized, Double-blind, Active-and Placebo-controlled, Parallel-group, Multicenter Study to Evaluate the Efficacy and Safety of Induction and Maintenance Combination Therapy with Guselkumab and Golimumab in Participants with Moderately to Severely Active Ulcerative Colitis (EudraCT No. 2021-005528-39).
5. Furer, Victoria, and Ori Elkayam. “Dual Biologic Therapy in Patients with Rheumatoid Arthritis and Psoriatic Arthritis.” Rambam Maimonides Medical Journal, vol. 14, no. 2, 30 Apr. 2023.
6. Sands, Bruce E., et al.“Safety and Tolerability of Concurrent Natalizumab Treatment for Patients with Crohn’s Disease Not in Remission While Receiving Infliximab.” Inflammatory Bowel Diseases, vol. 13, no. 2, 2007, pp. 2–11.
7. Singh, Jasvinder A., et al.“2012 Update of the 2008 American College of Rheumatology (ACR) Recommendations for the Use of Disease-Modifying Anti-Rheumatic Drugs and Biologics in the Treatment of Rheumatoid Arthritis (RA).” Arthritis Care & Research (Hoboken), vol. 64, no. 5, May 2012, pp. 625–639.
8. Genovese, Mark C et al. “ABT-122 in Patients With Rheumatoid Arthritis With an Inadequate Response to Methotrexate: A Randomized, Double-Blind Study.” Arthritis & rheumatology vol. 70,11 (2018).
9. Khatri, Amit et al. “Exposure-response analyses demonstrate no evidence of interleukin 17A contribution to efficacy of ABT-122 in rheumatoid or psoriatic arthritis.” Rheumatology (2019).
10. Feagan, Brian G., et al.“Guselkumab Plus Golimumab Combination Therapy Versus Guselkumab or Golimumab Monotherapy in Patients with Ulcerative Colitis (VEGA): A Randomised, Double-Blind, Controlled, Phase 2, Proof-of-Concept Trial.” The Lancet Gastroenterology & Hepatology.
11. Johnson & Johnson. “Johnson & Johnson Strengthens Pipeline to Lead in Atopic Dermatitis with the Completion of the Acquisition of Yellow Jersey Therapeutics, Gaining Ownership of NM26.”  23 Nov. 2023.
12. Johnson & Johnson. “Johnson & Johnson to Acquire Proteologix Inc. to Lead in Atopic Dermatitis Treatment.” 9 Aug. 2023.
13. Deiteren, Annemie, et al.“A Proof-of-Mechanism Trial in Asthma with Lunsekimig, a Bispecific NANOBODY Molecule.” European Respiratory Journal, vol. 65, no. 4, 2025.
14. Mahatole, Siddhi. “Johnson & Johnson Halts Mid-Stage Trial of Experimental Eczema Drug.” Reuters, 26 Dec. 2025.
15. UCB. Galvokimig EADV Data Presentation. UCB, 29 Sept. 2025.
16. Sanofi. EADV: Sanofi’s Brivekimg Achieved Positive Results in Hidradenitis Suppurative in Phase 2a Study. 17 Sep. 2025.

Comment