A targeted bispecific TGFBR2 antagonist antibody demonstrates cell selectivity and enhanced potency on human fibroblasts

Curation statements for this article:
  • Curated by eLife

    eLife logo

    eLife Assessment

    This study presents a well-executed and valuable proof of concept relating to the targeting of TGFBR2 in human disease. The evidence supporting the claims of the authors is largely solid, although it falls considerably short of a conclusive pre-clinical package of data, and as such the claims relating to clinical impact need to be downplayed. Additionally, flow cytometry data quantification should be included to allow numerical contrast between receptor abundance and functional potency. The work will be of interest to biologists and translational scientists working on cancer or lung disease.

This article has been Reviewed by the following groups

Read the full article

Abstract

Beyond its critical roles in development and tissue homeostasis, TGFβ signaling promotes key aspects of cancer progression and is a primary driver of fibrosis. Although blocking TGFβ signaling has great therapeutic potential for cancer and fibrotic diseases such as idiopathic pulmonary fibrosis (IPF), preclinical and clinical studies revealed that non-specific alteration of the pathway can have severe adverse consequences; therefore, inhibiting TGFβ signaling in a cell-type specific manner may avoid systemic toxic effects while preserving potential therapeutic effects. The parasitic helminth Heligmosomoides polygyrus has evolved cell-type-targeted modulators of TGFβ signaling. With insights from the development of other targeted signaling modulators and using the worm proteins as a guide, we sought to develop a human-fibroblast-targeted TGFBR2 antagonist. Here, we report mechanistic insights into the targeted worm TGFBR2 antagonist TGM6 and fusion proteins containing the TGM6 targeting domains. We created a bispecific antibody TGFBR2 antagonist that binds PDGFRA as a targeting receptor and demonstrates cell selectivity and enhanced potency in fibroblasts. Our findings suggest a viable path for developing targeted TGFβ signaling antagonists as therapeutics for cancer and tissue fibrosis.

Article activity feed

  1. eLife Assessment

    This study presents a well-executed and valuable proof of concept relating to the targeting of TGFBR2 in human disease. The evidence supporting the claims of the authors is largely solid, although it falls considerably short of a conclusive pre-clinical package of data, and as such the claims relating to clinical impact need to be downplayed. Additionally, flow cytometry data quantification should be included to allow numerical contrast between receptor abundance and functional potency. The work will be of interest to biologists and translational scientists working on cancer or lung disease.

  2. Reviewer #1 (Public review):

    Summary:

    This manuscript addresses a genuinely important and long-standing problem in the field: how to exploit the therapeutic potential of TGFβ pathway blockade while avoiding the systemic toxicity that has repeatedly derailed clinical candidates. The biologic rationale is sound, the stepwise molecular design is elegant, and the central finding - a ~300-fold potency enhancement in PDGFRA-expressing human fibroblasts with no enhancement in PDGFRA-negative HEK293 cells - is striking and clearly presented. As a proof-of-concept study, the work is significant. However, the evidence base is entirely in vitro, rests primarily on a single immortalized human fibroblast line, and leaves the mechanism of the potency enhancement incompletely resolved. The study should be viewed as a well-executed first step, not as validation of a therapeutic strategy, and the findings' generalizability and mechanistic claims require further substantiation before the conclusions drawn can be fully accepted.

    Strengths:

    The paper is well-motivated. The failure of broad TGFβ pathway inhibitors in the clinic, including the recent halt of the BEACON-IPF bexotegrast trial (NCT06097260), lends urgency to the idea of cell type-targeted antagonism, and the authors situate their work clearly in that context. The incremental design logic is a particular strength: they first characterize the natural bispecific TGM6, use it to establish mechanistic principles (LYTAC-like receptor degradation, tolerance for lower-affinity TGFBR2 binders), then systematically replace the immunogenic worm domains with antibody fragments, culminating in a fully human-sequence bispecific. This stepwise approach makes the reasoning transparent and persuasive. The choice of PDGFRA as a fibroblast targeting handle is well-justified by recent literature (Tsukui et al., 2024), and the use of multiple orthogonal readouts - luciferase SBE reporter, flow cytometry for surface TGFBR2, RT-qPCR for fibrosis-relevant genes, and western blotting for pSMAD2/3 is commendable. The ~300-fold potency shift in fibroblasts, absent in HEK293 cells, is the paper's headline result and is reproduced across independent readouts in a compelling manner.

    Weaknesses:

    Mechanistic interpretation of LYTAC (Lysosomal Targeting Chimera) like activity. The authors propose that TGM6 and the bispecific act by co-engaging a co-receptor that drives lysosomal degradation of TGFBR2. The flow cytometry data showing rapid, near-complete surface TGFBR2 depletion is visually striking. However, the inhibitor experiments (Figures 1I, 1J) are imperfect support for this conclusion. Dynasore (a dynamin inhibitor) only partially restores surface TGFBR2, consistent with but not proof of endocytosis-mediated depletion. More problematically, the proteasome inhibitor MG132 is used to support lysosomal involvement, but proteasome inhibition does not directly block lysosomal degradation of transmembrane receptors. A lysosomal inhibitor (e.g., bafilomycin A1 or chloroquine) would be the appropriate reagent; its absence leaves the mechanistic claim inadequately tested. The paper also does not directly demonstrate co-internalization of TGFBR2 with either PDGFRA or LRP1 (the latter identified as the D4D5 receptor in van Dinther et al., 2026, cited by the authors).

    The potency enhancement mechanism for the PDGFRA bispecific is insufficiently dissected. The authors attribute the 300-fold potency enhancement to cell-targeting and a LYTAC-like mechanism, and flow cytometry (Figure 3D) does show greater TGFBR2 depletion with the targeted molecule. However, simple bivalent avidity effects from simultaneously engaging two receptors on the same cell surface are not formally excluded. A monovalent anti-PDGFRA control (e.g., anti-PDGFRA Fab alone, or a bispecific with a non-TGFBR2-binding VH) would help partition the targeting and degradation contributions. Similarly, no ternary complex (simultaneous binding to PDGFRA and TGFBR2) is demonstrated biochemically.

    Limited cell model diversity. The human fibroblast data all derive from a single commercially available immortalized fetal lung fibroblast line (MRC5). Whether the results generalize to primary adult human pulmonary fibroblasts, fibroblast subtypes from IPF patients, or fibroblasts from other tissues is entirely unknown. This substantially limits the translational claims in the Discussion.

    Selectivity concerns PDGFRA as the target receptor. PDGFRA is expressed not only on pulmonary fibroblasts but also on pericytes, stellate cells, some hematopoietic progenitors, and certain tumor cell types. The paper tests selectivity only against HEK293 cells, which lack many fibroblast-relevant properties beyond PDGFRA expression. Whether the bispecific engages and functionally modulates TGFβ signaling in other PDGFRA-expressing cell types is not addressed.

    The VH271-12-TGM6-D4D5 data in human cells (Figure 2F) are only duplicates. This is explicitly acknowledged, but it means the most mechanistically informative intermediate construct is under-replicated. The authors should provide p-values.

    Functional readout limitations. The RT-qPCR panel covers only two ECM genes (COL1A1 and ELN), and there is notable variability in the data (Figure 3E). The western blot for ACTA2 (Figure 3F) also shows wide individual-experiment scatter. These data are directionally consistent with the reporter assay but are not themselves quantitatively convincing.

    Conclusion:

    The core in vitro proof-of-concept claim that a PDGFRA-targeted TGFBR2 bispecific can achieve cell-selective, potency-enhanced TGFβ pathway antagonism in human fibroblasts is supported by the data as presented. However, the mechanistic framing (LYTAC-like degradation) is incompletely evidenced, the generalizability beyond MRC5 cells is untested, and there are no in vivo data of any kind. The Discussion's forward-looking statements about therapeutic application to IPF and cancer should be substantially tempered.

  3. Reviewer #2 (Public review):

    Summary:

    In this study, the authors aim to provide proof-of-concept evidence to support the development of a targeted anti-TGF beta therapy for the treatment of solid tumours and fibrotic diseases. By mimicking a natural strategy employed by a murine parasite, they aim to provide a foundational design for a precision, cell-targeted therapeutic to circumvent the prohibitive toxicity associated with systemic TGF-beta inhibition.

    Strengths:

    (1) The authors demonstrate the design of a molecule that can achieve an in vitro selectivity of around 300-fold and show a functional effect of inhibition on fibrotic signalling and myofibroblast phenotype. A 300-fold cellular potency window would be expected to be of clinical and therapeutic relevance if it could be replicated across tissues.

    (2) The mechanistic investigations around receptor internalisation and degradation are solid and are supported by a paper that has been published since this manuscript was submitted (van Dinther et al. 2026).

    (3) The paper gives some insights into factors that may further influence selectivity. The authors note that whilst the non-target HEK293 cells are 100-fold less sensitive than the MCF5 lung fibroblasts to TGF-beta inhibition with the targeted molecule, the HEK293 are more sensitive to the untargeted ligand. This raises the likelihood that expression levels of the therapeutic target may play into the selectivity of the molecule as well as the expression distribution of the targeting protein.

    Weaknesses:

    (1) The cell line chosen to demonstrate selectivity is an embryonic line with characteristics of neural progenitor cells and, as such, not relevant to the intended patient populations for TGF-beta inhibitors. This study could have had more impact if it had investigated selectivity against the cell types relevant to the known cardiovascular side effects associated with TGF-beta inhibitors.

    (2) The authors mention that selectivity may be influenced by expression levels of the target proteins but do not compare expression levels of TGFBR2 and PDGFRA in HEK293 and MCF5 lung fibroblasts directly. Furthermore, there is little investigation into the relative impact of different binding affinities against both TGFBR2 and PDGFRA; the paper could have been more compelling if this concept had been expanded on. A deep understanding of relative binding kinetics and target protein expression levels will be critical to the successful development of a cell-targeted TGF-beta inhibitor.

    (3) The molecule developed is aimed exclusively at targeting pathogenic fibroblasts, but pulmonary fibrosis is a complex disease involving multiple cell types that respond to TGF-beta. This includes the epithelial alveolar type II cell, which is known to be a driver of disease. A truly disease-modifying therapeutic will need to tackle multiple cell types in IPF whilst minimising the impact on protective TGF-beta functions.

    Impact:

    A targeted TGF-beta therapeutic that enables selective inhibition of the pathological functions of TGF-beta whilst maintaining normal homeostatic signalling would have far-reaching benefits for patients. This study demonstrates for the first time that a naturally occurring helminth strategy for TGF-beta modulation can be adapted for human tissue targeting in vitro. It raises interesting potential as a therapeutic strategy, though the data as presented currently is far from that required for a convincing preclinical data package. The pharmacology will be complex, and it is of note that the helminth strategy to modulate the pathway local to the site of infection utilises multiple proteins and multiple targeting domains. Systemic modulation is an even higher bar, requiring detailed knowledge of the system kinetics and extensive in vitro profiling of normal and disease tissue responses. The next steps would also include an understanding of the therapeutic window required in vivo.

  4. Reviewer #3 (Public review):

    Summary:

    The authors aim to develop a cell-targeted approach to inhibiting transforming growth factor beta signalling, motivated by the potential toxicity associated with systemic inhibition of this pathway. They first investigate the naturally occurring helminth protein TGM6, which inhibits signalling in mouse fibroblasts and causes rapid loss of its receptor from the cell surface. They then use this biological principle to engineer an antibody-based bispecific antagonist that simultaneously binds TGFBR2 and PDGFRA, the latter being expressed by pulmonary fibroblasts relevant to fibrosis.

    The principal finding is that targeting the antagonist to PDGFRA substantially increases its activity in human MRC5 pulmonary fibroblasts. The targeted molecule inhibits TGF-beta-induced signalling with an IC50 of approximately 120 pM compared with approximately 35 nM for the untargeted molecule, corresponding to an almost 300 fold difference in potency. This enhancement is not observed in PDGFRA-negative HEK293 cells. The targeted molecule is also associated with reduced surface TGFBR2 and decreases downstream transcriptional and protein responses to TGF beta.

    The study provides a valuable proof of principle that engagement of a second cell surface receptor can markedly enhance the activity and potentially the cellular selectivity of a TGFBR2 antagonist. This could provide a useful general strategy for developing inhibitors of signalling pathways for which systemic inhibition is limited by toxicity. The evidence for the central phenomenon is solid, although the mechanism responsible for the striking increase in potency and the extent to which this can be attributed specifically to PDGFRA engagement and receptor depletion remain incompletely established.

    Strengths:

    The major strength of the study is the magnitude and apparent cellular specificity of the enhancement in antagonist activity. In MRC5 fibroblasts, the PDGFRA-targeted antagonist is almost 300-fold more potent than the corresponding untargeted molecule, whereas targeting provides no apparent advantage in PDGFRA-negative HEK293 cells. This provides persuasive evidence for the underlying concept that engagement of an additional cell surface receptor can markedly enhance TGFBR2 antagonism.

    The experimental progression is also a strength. The authors move logically from characterisation of the naturally occurring TGM6 protein, through replacement of its TGFBR2 binding component with an antibody-derived binder, to replacement of the helminth targeting domains with an antibody against PDGFRA. The intermediate construct retains activity in mouse fibroblasts and, importantly, also inhibits TGF-beta-induced signalling in human MRC5 fibroblasts. This provides a coherent link between the biological observations concerning TGM6 and the final fully antibody-based molecule.

    The use of several complementary readouts also strengthens the conclusions. In addition to a signalling reporter assay, the authors examine receptor abundance at the cell surface and downstream effects on extracellular matrix-associated gene expression and ACTA2 protein. The revised manuscript has further strengthened these aspects by improving quantification and reporting of biological replication and by adding the protein-level downstream readout.

    Finally, the authors appropriately identify an important potential limitation of ligand competitive TGFBR2 antagonism. HEK293 cells are more sensitive than the intended fibroblast population to the untargeted antagonist, despite lacking the targeting receptor. This observation is important because it indicates that cellular context can substantially influence sensitivity independently of the targeting strategy and highlights an issue that will need consideration in further development of this approach.

    Weaknesses:

    The principal weakness is that the causal contribution of PDGFRA engagement to the increased potency has not been directly demonstrated. The central evidence comes from comparison of MRC5 fibroblasts with PDGFRA-negative HEK293 cells. These are biologically distinct cell types and, importantly, differ in their sensitivity to the untargeted TGFBR2 antagonist. An isogenic system would provide a more rigorous test of the proposed mechanism. For example, deletion of PDGFRA in MRC5 cells, ideally followed by restoration of its expression, could establish whether the marked potency advantage of the targeted molecule is specifically dependent on PDGFRA engagement.

    There is also an important apparent disparity between the magnitude of the functional effect and the changes in surface TGFBR2. The approximately 300-fold difference in functional potency is striking, whereas the differences in the flow cytometry distributions appear considerably more modest. The flow cytometry experiments are largely presented as representative distributions rather than quantitative measurements of receptor abundance across biological replicates. Quantification of median fluorescence intensity or relative surface receptor abundance, together with appropriate analysis across replicates, would allow the magnitude and reproducibility of receptor depletion to be assessed more clearly. It would also help determine whether receptor removal is sufficient to explain the very large functional effect.

    Related to this, the precise mechanism underlying the enhanced potency remains uncertain. TGM6 causes rapid loss of TGFBR2 from the cell surface, and inhibitors of internalisation and protein degradation partially reduce this effect. These findings are consistent with altered receptor trafficking and degradation but do not directly establish lysosomal degradation. Moreover, the disparity between the modest apparent changes in surface receptor abundance and the large change in functional potency raises the possibility that increased local concentration, avidity, or other consequences of PDGFRA engagement contribute substantially to the enhanced activity. The description of the mechanism as "LYTAC like" should therefore remain appropriately qualified.

    The biological validation of the final construct is relatively limited. The human experiments use MRC5 cells, an established foetal lung fibroblast line, and downstream effects are assessed principally through a signalling reporter, two extracellular matrix-associated transcripts and ACTA2 protein. These experiments establish inhibition of TGF beta signalling but do not demonstrate a broader change in fibroblast phenotype or an antifibrotic effect. Confirmation in primary adult pulmonary fibroblasts, particularly cells derived from fibrotic lungs, would increase the biological relevance of the findings.

  5. A targeted bispecific TGFBR2 antagonist antibody demonstrates cell selectivity and enhanced potency on human fibroblasts https://www.biorxiv.org/content/10.64898/2026.08.11.744117v1

  6. A targeted bispecific TGFBR2 antagonist antibody demonstrates cell selectivity and enhanced potency on human fibroblasts https://www.biorxiv.org/content/10.64898/2026.08.11.744117v1