Modality Series 06 | New Variables in Deep B-Cell Depletion for Autoimmune Diseases

Originally published in Chinese on October 20, 2026. This commentary reflects information available as of October 20, 2026.

HKeyBio’s take. What in vivo T-cell engagers (TCEs) fundamentally change is not the mechanism through which T cells eliminate B cells, but how therapeutic TCE exposure is generated inside the patient. Messenger RNA (mRNA) packaged in lipid nanoparticles (LNPs) enables the patient’s own cells to produce TCE protein transiently, creating gradually rising exposure that persists for a period of time. ABO2203 provides preliminary human evidence and remains in early clinical development. Its potential as a product for chronic autoimmune diseases will depend on whether in vivo expression can generate exposure sufficient for deep B-cell depletion and remain consistent, predictable, reproducible, and capable of declining as intended.

On October 2, 2026, Abogen Biosciences announced a licensing and option agreement with Novartis covering RNA-encoded therapies. Under the agreement, Novartis would receive an exclusive worldwide license to ABO2203, together with options to license additional programs from Abogen’s RNA platform. The agreement provided for a US$575 million upfront payment and up to approximately US$7.2 billion in potential milestone payments if all options across the programs were exercised and the agreed development, regulatory, and commercial milestones were met. At announcement, the transaction remained subject to customary closing conditions, including required regulatory clearances. Beyond the headline transaction value, the distinguishing feature is the modality: ABO2203 delivers mRNA encoding a CD19 × CD3 TCE rather than a premanufactured TCE protein.

01 What exactly is an in vivo TCE?

Strictly speaking, conventional protein TCEs also act in vivo. In this article, “in vivo TCE” specifically refers to administration of nucleic acid encoding a TCE, allowing the patient’s cells to synthesize and secrete the molecule for a limited time. ABO2203 delivers mRNA via LNPs. The resulting CD19 × CD3 TCE recruits endogenous T cells to CD19-positive B cells, enabling targeted depletion. The licensing announcement identifies this RNA-encoded modality as the focus of the Novartis agreement. [1]

This modality must also be distinguished from in vivo CAR-T. In vivo CAR-T delivers genetic information encoding a chimeric antigen receptor (CAR) into T cells, converting some of them into CAR-expressing effector cells. In vivo TCE does not genetically engineer T cells; instead, the expressed TCE enables immune synapse formation between endogenous T cells and target cells. Both in vivo strategies can bypass the conventional ex vivo cell-manufacturing process, but their active therapeutic entities, pharmacokinetics, persistence, and sources of risk differ.

Table 1. Key distinctions among three T-cell-redirecting modalities

These distinctions determine how candidates should be assessed. For a protein TCE, the administered dose directly represents the amount of protein introduced. For an in vivo TCE, the mRNA dose is only the starting input. Actual exposure depends on which cells take up LNPs, translation efficiency, expression duration, secretion efficiency and the half-life of the TCE protein itself. Accordingly, microgram quantities of mRNA cannot be compared directly with doses of protein TCEs or CAR-T cells.

02 What has ABO2203 demonstrated so far?

On September 18, Cell published an online report of ABO2203 in refractory secondary immune thrombocytopenia (ITP). All three treated patients experienced rapid and complete peripheral B-cell depletion, with sustained depletion also observed in bone marrow. At six months, investigators reported sustained platelet recovery, improved serologic markers and reduced disease activity. Reported adverse events were grade 1 or 2, and no cytokine release syndrome (CRS) was observed. B cells returning after depletion were predominantly transitional and naïve populations, providing an early window into immune reconstitution.[2]

These observations mark an initial feasibility milestone: an mRNA-LNP system can generate pharmacologically active TCE protein in humans and achieve deep B-cell depletion in both blood and bone marrow. The study also suggests a potentially important pharmacokinetic (PK) feature. Cellular uptake and translation introduce a delay before TCE exposure develops; that delay may provide a PK buffer against early cytokine release. Comparative animal studies and early human PK and safety observations provide initial support for this mechanism. Larger cohorts will further clarify the safety characteristics of this exposure pattern relative to protein TCEs. [2]

An abstract presented at EULAR 2026 provided broader preliminary information. As of December 16, 2025, 15 patients with refractory autoimmune diseases had been enrolled. All nine patients who received a starting dose of 120 μg showed measurable TCE expression. The reported time to peak serum TCE concentration was four days, with a terminal half-life of 7.3 days. The abstract reported no CRS, immune effector cell-associated neurotoxicity syndrome (ICANS), or hepatotoxicity. These interim observations from an ongoing early-stage, single-arm study complement the six-month outcomes in the three ITP patients reported in Cell. [3,4]

These findings provide important early clinical feasibility evidence. The three ITP cases contribute human data on deep B-cell depletion, platelet recovery, and B-cell reconstitution. Larger cohorts and longer follow-up will allow a more complete assessment of durable remission, infections, immunoglobulin changes, and immune recovery. As additional cohorts in lupus, Sjögren’s syndrome, and systemic sclerosis mature, they will further clarify pharmacologic consistency and disease-specific responses. The Novartis–Abogen collaboration reflects industry interest in this approach and its broader platform potential, while bringing additional attention to subsequent clinical development and platform expansion. [1]

03 New opportunities and variables created by in vivo production

The appeal of in vivo TCEs is that conventional TCE manufacturing and dosing challenges are translated into the problem of controlling expression within the body. Protein TCEs often use step-up dosing to mitigate cytokine-associated risk at treatment initiation. Because mRNA must first enter cells and be translated, TCE concentrations may rise more gradually. If this exposure pattern proves reliably reproducible, it could potentially reduce complex step-up regimens and inpatient monitoring while preserving flexibility for repeat dosing.

At the same time, in vivo production introduces additional variables. LNP uptake and translation efficiency can vary among patients, so an identical mRNA dose may not produce identical TCE exposure. Discontinuing a protein infusion stops further administration, whereas expression from delivered mRNA continues according to its biological kinetics. Its actual duration is jointly governed by mRNA stability, the cells receiving the payload, and the half-life of the TCE protein. Repeat administration also requires assessment of LNP-associated innate immune responses, changes in delivery efficiency, anti-drug antibodies, and organ exposure.

Autoimmune therapy does not seek indefinite T-cell activation. An ideal product would achieve sufficiently deep B-cell depletion within a limited period, followed by declining TCE exposure and T-cell activation, while allowing assessment of the composition of the reconstituting B-cell population. A fuller assessment of product value therefore integrates peak exposure and B-cell nadir with predictable expression profiles, reproducible tissue depletion, recovery after treatment ends, and sustained clinical remission.

04 The next competitive frontier in expression control and evidence

ABO2203 has expanded the competitive landscape for autoimmune TCEs from molecular architecture to delivery and expression control. Future programs will need to demonstrate four connected layers of evidence, extending beyond reductions in peripheral blood B cells:

1. Carrier level. Which tissues and cells take up LNPs? How efficient is delivery? Are complement activation, innate immune responses or organ toxicity observed? Does biodistribution change after repeated dosing?

2. Transcript level. What are the distribution, integrity and clearance kinetics of mRNA across tissues? Can between-patient differences in translation be controlled?

3. Protein level. Does the TCE synthesized in vivo have the correct structure and potency? How do exposure, immunogenicity and cytokine release relate to one another?

4. System level. How do the depth and recovery of B-cell depletion in blood, bone marrow and lymphoid tissues relate to T-cell status, immunoglobulins, infection risk and remission within an interpretable PK/PD framework?

Preclinical evaluation cannot simply replicate the assessment package used for a conventional protein therapeutic. Cross-species reactivity at CD19 and CD3, LNP delivery differences, and species-specific mRNA translation can all affect pharmacological relevance. Concurrent measurements of mRNA, TCE protein, immune-cell subsets, cytokines and histopathology are needed to distinguish carrier-related, expression-related and TCE-mediated effects.

HKeyBio focuses on building a comprehensive, layered evidence package around each in vivo TCE candidate. Evaluation can draw on human cell systems, suitable mouse models, and nonhuman primate (NHP) studies with relevant target and CD3 pharmacology, as appropriate for the candidate. Study design should encompass delivery, expression, B-cell depletion, T-cell activation, cytokines, immunoglobulins, tissue distribution, and recovery. The choice of model and endpoints should be guided jointly by the carrier, target, route of administration, and development stage.

In vivo TCEs are no longer purely an engineering concept. ABO2203 provides early human feasibility evidence, and the Novartis collaboration has brought greater attention to the approach. The next phase of competition will extend beyond producing a TCE inside the body to controlling its level, duration, tissue distribution, and immune effects within a therapeutic window suited to patients with autoimmune diseases. This first clinical chapter is only beginning.

References

[1] Abogen Biosciences. Abogen Announces Licensing and Option Agreement with Novartis to Advance mRNA-Encoded T-Cell Engager and Other Potential RNA Programs. 2 October 2026. Source

[2] Wang M, et al. mRNA-encoding CD19-targeting T cell engager for refractory immune thrombocytopenia. Cell. Published online 18 September 2026. doi:10.1016/j.cell.2026.08.039. Source

[3] Wang M, et al. Preclinical to clinical translation of ABO2203, a messenger RNA (mRNA) encoding a CD3 × CD19 T-cell engager (TCE), for refractory autoimmune diseases (AIDs): Preliminary results from the first-in-human clinical trial. EULAR 2026; POS0378. Ann Rheum Dis. 2026;85(Suppl 1):s603. Abstract identifier: annrheumdis-2026-eular.B.404. Source

[4] ClinicalTrials.gov. A Study of mRNA Encoding CD19 CD3 T Cell Engager ABO2203 in Patients With Refractory Autoimmune Diseases. NCT06747156. Source

[5] Wang L, et al. ABO2203, a messenger RNA encoding a CD3 × CD19 T cell engager, for relapsed/refractory B cell non-Hodgkin lymphoma: Preliminary results from first-in-human study. Cancer Res. 2026;86(8 Suppl):CT032. doi:10.1158/1538-7445.AM2026-CT032. Further reading on the separate oncology program. Source

Evidence note. ABO2203 studies in autoimmune diseases remain early-stage and single-arm. The three ITP cases published in Cell, interim observations from 15 patients in the EULAR abstract, and commercial transaction disclosures are presented as distinct types of evidence. This article is intended for R&D information exchange and does not constitute clinical, investment, or business decision advice.

Author: HKeyBio Marketing Department
Reviewed by: Jing Yan
Edited by: Wang Shuyan

About HKeyBio

HKeyBio is a preclinical contract research organization specializing in autoimmune and allergic diseases and providing end-to-end in vivo pharmacology services. Its core technical team has more than 20 years of experience across the autoimmune drug development process. The company is ISO 9001 certified and operates an integrated research facility in Suzhou Industrial Park covering molecular biology, cell biology, pathology, small-animal studies, and analytical testing, alongside NHP research sites in Suzhou and Guangxi.

HKeyBio has nearly 80 employees, and its project leads hold master’s or doctoral degrees. The team’s collective experience includes more than 500 autoimmune disease IND submissions, supporting clients in addressing complex challenges throughout drug development and submission.

· 50+ NHP models of autoimmune and allergic diseases

· 500+ animal models of autoimmune and allergic diseases

· Experience with 500+ autoimmune disease IND submissions