跳到正文
原文
Nature 综合主刊· Kaitlin Rhee·· 2026-09-30精选AI 评分66

Nature:双功能抗体GTAC经诱导受体转运与降解实现对GPCR的拮抗

GPCR antagonism via rewiring of receptor trafficking and degradation

AI 导读

Nature 发表研究提出 GTAC 双功能抗体平台,通过利用转铁蛋白受体1(TfR1)的快速内化,诱导 GPCR 内吞与溶酶体降解,从而拮抗受体信号。

推荐理由

原文提出以诱导内吞和溶酶体降解实现GPCR拮抗的双功能抗体GTAC平台,读者可了解其在7个GPCR靶点上的临床前表现与设计要点。

正文

Main

G-protein-coupled receptors (GPCRs) comprise the largest receptor family, with more than 800 members that have essential roles in physiology and disease1. Although about one-third of FDA-approved drugs target GPCRs1, therapeutic success is disproportionately concentrated within a small subset—particularly biogenic amine receptors—while many peptide-binding receptors, lipid receptors, orphan receptors and sensory receptors remain underexplored5.

Current GPCR inhibitors function by competing with ligands or stabilizing inactive receptor conformations6, but these mechanisms can be hindered by difficulty in developing sufficiently high-affinity competitive antagonists7 or by ligand-independent, constitutive signalling3. These challenges are often greater for peptide- and protein-binding GPCRs, in which autocrine or paracrine signalling, together with high affinity and local concentrations of ligands, can reduce the effectiveness of competitive inhibition4. Moreover, biologic antagonists are often difficult to engineer because the functional binding pockets are buried within the receptors7. These challenges underscore the need for alternative mechanisms of GPCR modulation.

Endogenous GPCR signalling is dynamically regulated by endolysosomal trafficking, in which internalization, recycling and degradation are tightly coupled to ligand binding, intracellular phosphorylation and β-arrestin recruitment8. Building on this biology, we proposed that induced proximity to an internalizing receptor could synthetically trigger GPCR endocytosis and lysosomal delivery to silence receptor signalling (Fig. 1a).

Fig. 1: GTACs induce potent GPCR depletion and signalling antagonism.

a, Schematic of native and engineered GPCR trafficking (left). Native agonists typically induce GPCR internalization, followed by recycling or degradation. GTACs direct GPCRs for degradation by leveraging rapid, constitutive TfR1-mediated internalization (middle). GTAC design (right): bifunctional antibody with GPCR-binding domains and TfR1-binding domains fused to an Fc. The control antibody lacks TfR1-binding domains. b, Cell-surface GPCR levels (anti-FLAG flow cytometry) after treatment with GTACs or control antibodies. Cell-surface AT1R levels in FLAG–AT1R–GFP+ TfR1+ Expi293 cells after 6 h treatment; cell-surface RXFP1 levels in FLAG–RXFP1+ TfR1+ Expi293 cells after 2 h treatment; cell-surface APJ levels in FLAG–APJ+ TfR1+ HEK293T cells after 2 h treatment; and cell-surface A2AR levels in FLAG–A2AR+ TfR1+ Expi293 cells after 6 h treatment. Data represent mean ± s.d. from n = 3 replicates. c, NFAT-luciferase activation in FLAG–AT1R+ TfR1+ NFAT-Luc+ Expi293 cells, after 2 h pretreatment with 50 nM (left) or varying concentrations (right) of AT1R GTAC or Nb–Fc, followed by 4 h stimulation with varying concentrations (left) or 9 nM (right) ANGII. Data represent mean ± s.d. from n = 4 biological replicates. d, β-Arrestin recruitment in FLAG–AT1R–Tango+ TfR1+ HTLA cells (luminescence) after pretreatment for 2 h with 50 nM (left) or varying concentrations (right) of AT1R GTAC or Nb–Fc, followed by 14 h activation with varying concentrations (left) or 200 nM (right) ANGII. Data represent mean ± s.d. from n = 3 biological replicates for both experiments. e, RLN2-induced cAMP signalling in FLAG–RXFP1+ TfR1+ GloSensor+ HEK293T cells (luminescence) after 2 h pretreatment with 10 nM (left) or varying concentrations (right) of RXFP1 GTAC or control, followed by 30 min stimulation with RLN2 at varying concentrations (left) or 1 nM (right). Data represent mean ± s.d. from n = 3 biological replicates. Diagrams in a, c and e created in BioRender; Zhou, X. https://biorender.com/hymki1y (2026).

Small-molecule PROTACs and molecular glues recruit intracellular proteasomal machinery for target protein degradation9, whereas emerging antibody-based modalities such as LYTAC, AbTAC, PROTAB, KineTAC, EndoTag, LIPTAC and our previously reported TransTAC harness cell-surface effectors for lysosomal protein degradation10,11,12,13,14,15,16,17,18,19. However, applying small-molecule strategies to GPCRs has so far shown limited efficacy, and systematic exploration of induced-proximity approaches for GPCR regulation remains lacking.

Here, we introduce GTACs: bifunctional antibodies that harness the rapid, constitutive internalization of transferrin receptor 1 (TfR1)10,20 for controlling GPCR endocytosis, degradation and signalling (Fig. 1a). GTACs potently depleted and antagonized angiotensin II type 1 receptor (AT1R), a regulator of cardiovascular function21; relaxin family peptide receptor 1 (RXFP1), a key player during pregnancy22 and cancer4; C–C chemokine receptor type 6 (CCR6), a regulator of immune cell trafficking23; and BILF1, a constitutively active virus-encoded GPCR3. AT1R GTACs suppressed angiotensin-potentiated contractions in induced pluripotent stem cell (iPSC)-derived cardiomyocytes; RXFP1 GTACs downregulated metastatic gene programs in ovarian cancer; CCR6 GTACs achieved complete blockade of chemotaxis in primary immune cells and anti-inflammatory activity in vivo; and BILF1 GTACs inhibited constitutive signalling and rescued MHC-I expression. Overall, GTACs exhibited more than 10–100-fold greater potency and improved efficacy over conventional antagonists, demonstrating that induced degradation not only enhances GPCR inhibition but also enables modulation of receptors that are difficult or impossible to target using traditional pharmacology.

We further generalize this strategy to apelin receptor (APJ), a cardiovascular GPCR with multiple ligand isoforms24; adenosine A2A receptor (A2AR), an immune receptor exposed to high local ligand concentrations25; and M3 muscarinic acetylcholine receptor (M3R)26, a GPCR lacking suitable binders. We introduce a universal FLAG-targeting GTAC system for modular, scalable interrogation of GPCR biology. Using imaging and pharmacological approaches, we define mechanistic features of GTACs and identify protein engineering strategies for context-dependent GPCR regulation. Collectively, these studies establish induced endocytosis and degradation as a modular mechanism for GPCR antagonism, expanding the pharmacological toolkit for receptors that are poorly addressed by conventional occupancy-based inhibition.

GTACs potently deplete and inhibit GPCRs

We assembled a panel of GPCRs with diverse ligands, signalling and trafficking properties, including AT1R, RXFP1, APJ and A2AR, to develop the GTAC technology. All receptors contain an N-terminal FLAG tag, with an additional N-terminal HA tag or C-terminal GFP fusion in some cases, and were co-expressed with TfR1 with a v5 epitope tag (TfR1–v5) in HEK293T or Expi293 cells. Tagged overexpression systems are commonly used in GPCR research because low endogenous expression often limits signalling analyses27. GTACs were constructed as bispecific antibodies comprising GPCR binder(s), TfR1 binder(s) and human IgG1 Fc through homodimeric fusion or knobs into holes28 (Fig. 1a and Supplementary Figs. 2–4). Binders used include TfR1 scFv H7 (ref. 29) or nanobody (Nb) VHH-A (ref. 30), AT1R Nb AT118-L (ref. 31), APJ Nb JN241 (ref. 32), A2AR scFv Hu3F6 H5-L5 (ref. 33) and RXFP1 Nb Rx002 (refs. 34,35) (Extended Data Fig. 1a and Supplementary Figs. 2a,b, 3a and 4a).

Cell-surface GPCR depletion was assessed by flow cytometry after treatment with GTACs or GPCR binder controls. In contrast to the controls, GTACs effectively depleted up to 40–90% of cell-surface AT1R, RXFP1, APJ and A2AR with picomolar to single-digit nanomolar IC50s (Fig. 1b). In FLAG–AT1R–GFP+ TfR1+ Expi293 cells, AT1R GTAC also reduced GFP fluorescence, indicating total protein degradation (Dmax = 65%, DC50 ≈ 1 nM) (Extended Data Fig. 1b). AT1R and APJ depletion was evident within 1 h and remained stable for at least 24 h (Extended Data Fig. 1c,d). The fast internalization kinetics were consistent with previous TransTAC studies10. Non-AT1R-targeting GTACs did not significantly deplete AT1R, indicating that depletion is target-specific (Extended Data Fig. 1e).

GTACs also potently antagonized GPCR signalling. AT1R GTACs suppressed angiotensin II (ANGII)-induced NFAT-luciferase reporter activation in FLAG–AT1R+ TfR1+ NFAT-Luc+ Expi293 cells and β-arrestin recruitment in FLAG–AT1R–Tango+ TfR1+ HTLA cells (Fig. 1c,d). The AT1R Nb is a competitive antagonist; however, compared with Nb–Fc, AT1R GTACs were more than 10-fold more potent and achieved greater maximal inhibition, with activity modulated by GTAC geometry and TfR1 binder affinity.

RXFP1 has been implicated as a potential therapeutic target in ovarian cancer but remains difficult to inhibit because of autocrine signalling4, the picomolar affinity of relaxin-2 (RLN2) and a constitutively assembled signalosome that is highly sensitive to ligand36. Existing RLN2 analogue inhibitors have modest potency37, and no antagonists have reached clinical development. We constructed GTACs with an RLN2-competitive Nb we previously generated using yeast surface display34 (Extended Data Fig. 2a,b). The Nb bound RXFP1 on cells (EC50 ≈ 50 nM) but showed limited functional antagonism (Extended Data Fig. 2b). Affinity maturation improved binding (EC50 ≈ 2 nM, Ki = 340 pM), but signalling inhibition remained modest even at micromolar concentrations (Extended Data Fig. 2b–d).

RXFP1 GTACs exhibited nanomolar to sub-nanomolar on-cell binding and efficiently depleted both surface and total RXFP1 in FLAG–RXFP1–GFP+ TfR1+ Expi293 cells (Extended Data Figs. 2f,g and 3a,b). In a doxycycline-inducible cell line, cAMP signalling scaled with surface RXFP1 expression (Extended Data Fig. 2e). Consistent with this, GTAC-mediated RXFP1 depletion resulted in near-complete antagonism of RLN2 signalling in a split-luciferase cAMP assay using FLAG–RXFP1+ TfR1+ GloSensor+ HEK293T cells (Fig. 1e and Extended Data Fig. 2h). The most potent construct (IC50 ≈ 30 pM) was more than 200-fold more potent than the Nb–Fc and induced sustained receptor depletion and signalling inhibition (Fig. 1e and Extended Data Fig. 2i). GTACs did not affect forskolin-induced cAMP signalling, consistent with a target-specific mechanism of action.

We then evaluated GTAC activity in high-grade serous ovarian carcinoma, in which TfR1 is endogenously overexpressed38. Following GTAC pretreatment and RLN2 stimulation, qPCR analysis revealed significant downregulation of RXFP1 target genes, including MMP9 and MMP23B (ECM remodelling), NOTCH1 and NOTCH3 (Notch signalling) and DVL1 (Wnt signalling) (Extended Data Fig. 2j–l). These results indicate that GTAC-mediated RXFP1 depletion effectively suppresses downstream oncogenic transcriptional programs.

Experiments with AT1R, RXFP1, APJ and A2AR show that GTACs are a modular and broadly applicable approach for inducing rapid and sustained GPCR depletion and degradation. GTACs achieved substantially greater signalling inhibition potency and efficacy than the antagonistic Nbs, demonstrating that induced degradation can enhance GPCR antagonism.

FLAG GTAC inhibits constitutive activity

Extracellular FLAG-tagged receptors are widely used in GPCR research. We leveraged the anti-FLAG M2 fragment antibody39 to develop a FLAG-targeting GTAC, providing a targeting strategy that does not interfere with endogenous ligand binding (Fig. 2a and Extended Data Fig. 4a). FLAG GTAC exhibited nanomolar binding and similar internalization and degradation activity compared with AT1R GTAC in FLAG–HA–AT1R–GFP+ TfR1+ Expi293 cells (Fig. 2b and Extended Data Fig. 4b,c). Furthermore, FLAG GTAC inhibited ANGII-induced β-arrestin recruitment, whereas FLAG–Fab did not perturb signalling (Fig. 2c,d). Confocal microscopy showed AT1R distributed at the cell surface in untreated and FLAG–Fab–Fc-treated cells, whereas FLAG GTAC treatment caused intracellular puncta formation and about 60% GFP signal loss after 6 h (Fig. 2e).

Fig. 2: FLAG GTAC as a generalizable tool for GPCR modulation and inhibition of constitutive activity.

a, Schematic of FLAG GTACs depleting FLAG–GPCRs. b, Cell-surface AT1R (left, anti-HA flow cytometry) and total AT1R (right, GFP fluorescence) in FLAG–HA–AT1R–GFP+ TfR1+ Expi293 cells after 6 h FLAG GTAC or Fab–Fc treatment. Data represent mean ± s.d. from n = 3 replicates. c, β-Arrestin recruitment in FLAG–AT1R–Tango+ TfR1+ HTLA cells (luminescence) after 2 h pretreatment with FLAG GTAC or Fab–Fc, then ANGII stimulation (14 h, 200 nM). d, Pretreatment for 2 h with 50 nM FLAG GTAC or Fab–Fc, then ANGII (14 h). Data represent mean ± s.d. from n = 3 biological replicates. e, Total AT1R quantification (left, GFP per cell) and live-cell confocal microscopy of FLAG–AT1R–GFP+ TfR1+ Expi293 cells (right) after FLAG GTAC or Fab–Fc treatment. Scale bar, 20 µm. Centres represent medians; box bounds represent the 25th–75th percentiles; and whiskers represent minima and maxima. Data represent mean ± s.d. from n = 4 cells. f, Schematic showing BILF1 constitutively activating NF-κB and downregulating MHC-I. g, Cell-surface BILF1 (left, anti-HA) and NF-κB reporter activity (right, GFP) in FLAG–HA–BILF1+ TfR1+ NF-κB–GFP+ HEK293T cells after FLAG GTAC or Fab–Fc treatment. Data represent mean ± s.d. from n = 3 biological replicates. h, Surface-bound FLAG antibody (anti-FLAG) on inducible FLAG–BILF1+ Raji cells after FLAG GTAC or control treatment (2 h, 25 nM, 0 °C or 37 °C). Statistical comparisons using unpaired two-tailed t-tests. Data represent mean ± s.d. from n = 3 biological replicates. i, Cell-surface MHC-I in FLAG–BILF1+ Raji after FLAG GTAC or control treatment (left, anti-HLA-ABC, % uninduced; right, % change from induced to uninduced). Statistical comparisons using unpaired two-tailed t-tests. Data represent mean ± s.d. from n = 4 biological replicates. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Exact P values are provided in Supplementary Table 7. Diagrams in a and f created in BioRender; Zhou, X. https://biorender.com/hymki1y (2026).

Constitutive ligand-independent receptor activity cannot be inhibited by neutralizing antibodies. One example is BILF1. Expressed during the lytic cycle of Epstein–Barr virus (EBV), BILF1 displays ligand-independent signalling, upregulating ICAM1 through NF-κB pathway activation and downregulating MHC-I, and thereby contributing to viral persistence and immune evasion40 (Fig. 2f). In FLAG–HA–BILF1+ TfR1+ NF-κB–GFP+ HEK293T cells, FLAG GTACs potently depleted up to 70–80% cell-surface BILF1 and suppressed NF-κB-driven GFP expression (IC50 ≈ 0.8–1 nM) (Fig. 2g and Extended Data Fig. 4d). Confocal microscopy showed substantial GTAC-mediated BILF1 internalization in FLAG–HA–BILF1+ TfR1–v5+ HEK293T cells (Extended Data Fig. 5a,b). In FLAG–BILF1+ Raji cells, FLAG GTAC treatment caused BILF1 internalization at 37 °C (ref. 20) (Fig. 2h and Extended Data Fig. 4e). Induction of BILF1 overexpression in the Raji cells reduced cell-surface MHC-I expression, which was restored by FLAG GTAC treatment (Fig. 2i). These results demonstrate the ability of GTACs to functionally inhibit constitutive GPCR activity.

M3R, a target for bronchodilation in the treatment of asthma and chronic obstructive pulmonary disease (COPD)26, at present lacks reported high-affinity and high-selectivity antibodies. Flow cytometry and western blots demonstrated that FLAG GTACs exhibited nanomolar affinity and induced up to about 70% M3R degradation in FLAG–M3R+ TfR1+ Expi293 cells (Extended Data Fig. 4f,g). Moreover, FLAG GTAC antagonized iperoxo-induced β-arrestin recruitment in FLAG–M3R–Tango+ TfR1+ HTLA cells (Imax = 58%, IC50 ≈ 2 nM) (Extended Data Fig. 4h), highlighting its utility for modulating GPCRs that lack specific binders.

Together, these studies demonstrate that FLAG GTACs provide a universal tool for interrogating GPCR biology. By using a binder directed against a functionally inert epitope, FLAG GTACs enable GPCR depletion, antagonism and constitutive activity blockade without relying on a specific GPCR binder.

Context-dependent GTAC design

CCR6 mediates immune cell recruitment to inflamed tissues2 and is expressed on subsets of T cells, B cells, neutrophils and dendritic cells. It is a drug target in autoimmune and inflammatory diseases, including psoriasis, rheumatoid arthritis and ulcerative colitis2,41. However, no approved CCR6-targeting therapies exist, and current antagonists show incomplete blockade or limited selectivity41,42 due to high local chemokine concentrations, sequence homology, a two-epitope ligand binding pocket and low-nanomolar binding affinity for C–C chemokine ligand 20 (CCL20)43.

RXFP1 GTAC and past-generation TransTACs used a high-affinity, transferrin-competitive TfR1 binder (H7) to leverage elevated TfR1 expression in tumours. However, targeting GPCRs in non-cancerous cells may require reduced TfR1 engagement and non-competitive binding to preserve iron homeostasis. Therefore, we developed a second-generation GTAC design, incorporating non-competitive and lower-affinity TfR1 binders (Fig. 3a). Starting from a non-competitive Nb VHH-A, we expressed the Nb variants TfR1HIGH, TfR1INT1, TfR1INT2 and TfR1LOW, which span a range of affinities (Kd ≈ 0.1–120 nM)30 (Fig. 3b and Extended Data Fig. 7a). CCR6-targeting GTACs were constructed by pairing these variants with an N-terminally binding anti-CCR6 Fab (4H11, Kd ≈ 500 pM)44 (Fig. 3d and Supplementary Fig. 5). FLAG–CCR6 was overexpressed in Jurkat cells, which endogenously express high levels of TfR1.

Fig. 3: Context-dependent GTAC design enables selective CCR6 degradation and inhibition.

a, Schematic illustrating target-dependent GTAC engagement. b, Bio-layer interferometry (BLI) analysis of wild-type (TfR1HIGH, Kd = 124 nM) and affinity-reduced (TfR1LOW, Kd = 0.1 nM) TfR1 Nb variants. c, Flow cytometry showing cell surface binding of GTAC affinity variants to CCR6+ and CCR6− Jurkat cells after 10 min incubation on ice. Variants ordered from high to low TfR1 affinity (left to right). Data represent mean ± s.d. from n = 3 biological replicates. MFI, mean fluorescence intensity. d, Schematic of CCR6-targeting GTACs. e, Cell surface staining of FLAG–CCR6 and TfR1 in CCR6+ Jurkat cells after 2 h GTAC or control treatment using flow cytometry. Data represent mean ± s.d. from n = 3 biological replicates. f, Western blot and quantification of total CCR6 and TfR1 levels in CCR6+ Jurkat cells after 18 h GTAC treatment. Statistical comparisons performed using unpaired two-tailed t-tests. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Exact P values are provided in Supplementary Table 7. Data represent mean ± s.d. from n = 3 biological replicates. g, β-Arrestin recruitment measured by Tango assay in CCR6–Tango+ TfR1+ HTLA cells with 1 h GTAC pretreatment and overnight incubation with CCL20. Data represent mean ± s.d. from n = 3 biological replicates. h, Flow cytometry quantification of total F-actin in CCR6+ Jurkat cells with fixation and phalloidin staining. Cells were treated with varying GTAC doses (left) or 10 nM GTAC for 2 h, followed by stimulation with 50 nM CCL20 for 1 min or varying times (right). Data represent mean ± s.d. from n = 3 biological replicates. i, Chemotaxis of CCR6+ luciferase+ Jurkat cells via Transwell migration assay after 2 h treatment and 2 h migration with CCL20 as the chemoattractant. Data represent mean ± s.d. from n = 3 biological replicates. Diagrams in a and d created in BioRender; Zhou, X. https://biorender.com/hymki1y (2026).

To evaluate context-dependent engagement, we established a split-luciferase proximity assay that measures CCR6-dependent GTAC interaction with TfR1 (Extended Data Fig. 7b). H7-fused LgBiT reconstitutes with SmBiT fused to GTACs on GTAC–TfR1 interaction. Compared with the TfR1HIGH GTAC, lower-affinity GTACs showed significantly reduced luciferase signals in CCR6− cells compared with CCR6+ cells, demonstrating CCR6-dependent GTAC–TfR1 interaction (Extended Data Fig. 7c). Consistent with these findings, flow cytometry assays showed that TfR1LOW exhibited the greatest CCR6-dependent on-cell binding, with TfR1INT1 and TfR1INT2 showing intermediate selectivity (Fig. 3c). These results demonstrate that reducing TfR1 affinity enhances CCR6-dependent engagement, validating the context-dependent GTAC design.

In CCR6+ Jurkat cells, both TfR1HIGH and TfR1LOW GTACs induced about 85% CCR6 depletion and degradation with nanomolar IC50s, with more sustained internalization with the TfR1LOW GTAC over 24 h (Fig. 3d–f and Extended Data Fig. 7d,e). TfR1HIGH GTAC partially co-degraded TfR1, while the low-affinity GTACs showed no effects on TfR1 levels (Fig. 3e,f and Extended Data Fig. 7d). GTACs inhibited about 100% of CCL20-induced β-arrestin recruitment and F-actin polymerization, and nearly 100% of Jurkat cell migration (IC50 ≈ 0.4–4 nM), with TfR1LOW GTAC inducing stronger β-arrestin suppression and complete migratory inhibition (Fig. 3g–i and Extended Data Fig. 7f). The antagonist antibody inhibited only 55–60% of F-actin polymerization and chemotaxis, consistent with previous reports41,44. TfR1 knockdown reduced target degradation, confirming the TfR1 dependency of GTACs (Extended Data Fig. 7h).

These results demonstrate that CCR6 GTACs are potent functional inhibitors that surpass the activity of conventional antagonist antibodies, with the context-dependent variant showing particularly promising properties by selectively binding CCR6+ cells while not affecting TfR1 levels. Furthermore, consistent with our other targets, effective CCR6 depletion, degradation and chemotaxis inhibition were achieved by FLAG GTACs (Extended Data Fig. 4i).

GTAC rewires GPCR trafficking

To characterize GPCR trafficking in response to GTACs compared with native ligands, we performed multicolour live-cell imaging in HeLa cells, a cell line widely used for subcellular imaging. Cells were engineered to stably express CCR6–GFP, BFP–TfR1 and mCherry-tagged RAB5 (early endosome), RAB7 (late endosome), RAB11 (recycling endosome) or LAMP1 (lysosome) (Extended Data Figs. 8a and 9a–d). Whereas CCL20 induced receptor internalization without appreciable loss of the CCR6–GFP signal, TfR1HIGH and TfR1LOW GTACs induced near-complete loss of CCR6–GFP signal after 24 h, indicating that GTACs, but not the native agonist, promoted receptor degradation (Extended Data Fig. 8a). GTAC treatment increased CCR6 colocalization with both RAB7 and RAB11 after 4 h, indicating enhanced post-endocytic sorting into both degradative and recycling pathways (Extended Data Fig. 8b and Supplementary Table 5). Consistent with previous western blot results, TfR1LOW GTAC increased TfR1 recycling, as measured by RAB11 colocalization, compared with TfR1HIGH GTAC (Extended Data Fig. 8c). Similar results were observed by immunofluorescence of endogenous RAB proteins (Extended Data Fig. 10a–e).

AT1R showed a similar distinction between GTAC and native ligand responses. In FLAG–AT1R–GFP+ TfR1+ Expi293 cells, both AT1R GTAC and FLAG GTAC induced rapid internalization, followed by loss of AT1R–GFP signal and significant intracellular retention of the receptors after 6 h (Extended Data Fig. 8d,e). By contrast, ANGII induced only transient internalization, followed by subsequent recovery of surface receptor levels. Flow cytometry likewise showed that GTACs drove rapid and sustained receptor degradation, whereas ANGII induced only partial and transient internalization without affecting total receptor levels (Fig. 4c and Extended Data Fig. 1c,f).

In contrast to the cell-surface localization of CCR6 and AT1R, RXFP1 exhibited low membrane expression and was predominantly localized in perinuclear compartments in FLAG–RXFP1–GFP+ TfR1+ HeLa cells, as previously reported for its endoplasmic reticulum and golgi localization45 (Extended Data Fig. 8f). RXFP1 GTAC redistributed the cell-surface pool of RXFP1 to punctate intracellular vesicles. However, depletion of RXFP1 was lower than that observed for the other GPCRs tested (Fig. 1b and Extended Data Fig. 2g). This may reflect the limited surface expression of RXFP1, which could restrict productive RXFP1–TfR1 co-engagement. Moreover, the large intracellular receptor pool may replenish the cell-surface population, limiting the extent of protein depletion.

In line with these distinct expression and trafficking profiles, GTAC activity against RXFP1 showed greater dependence on high-affinity antagonists compared with AT1R or CCR6 (Fig. 3c and Extended Data Figs. 3a and 6d,e), in which the latter were effectively inhibited by non-neutralizing FLAG GTACs. Another feature distinguishing RXFP1 from many other GPCRs is the absence of agonist-induced internalization46 (Extended Data Fig. 6a). Consistent with previous reports, RLN2 did not induce appreciable RXFP1 internalization (Extended Data Fig. 6b). By contrast, RXFP1 GTACs induced receptor internalization, demonstrating that GTACs can rewire receptor trafficking even for GPCRs that do not naturally undergo agonist-induced internalization.

These studies show that GTACs alter GPCR trafficking through mechanisms distinct from those of native agonists. GTACs promoted accelerated GPCR endocytosis and sustained receptor depletion, whereas the native ligands characterized here induced receptor internalization with limited degradation. The results also further show that tuning TfR1 affinity using context-dependent GTACs can optimize TfR1 recycling while maintaining target degradation.

Determinants of GTAC activity

We next used small-molecule inhibitors, GPCR mutants and GTAC variants to elucidate the cellular and molecular factors that influence GTAC activity (Fig. 4a). Proteasome inhibition with MG132 did not significantly affect AT1R or CCR6 downregulation (Fig. 4b,e). By contrast, disrupting lysosomal function with bafilomycin A1 or chloroquine abrogated GTAC-induced degradation, restoring total AT1R and CCR6 levels (Fig. 4b,e and Extended Data Figs. 1g and 7i). Cell-surface levels increased only modestly, indicating that GTACs can still remove receptors from the plasma membrane, even when lysosomal degradation is impaired (Fig. 4b and Extended Data Fig. 1g). Moreover, blocking new protein synthesis with cycloheximide enhanced GTAC-induced AT1R degradation. These results, together with the earlier observation that GTACs enhance both lysosomal trafficking and recycling, indicate that the overall extent of receptor depletion is determined by the combined effects of accelerated receptor internalization, recycling, lysosomal degradation and protein synthesis.

Fig. 4: Cellular and molecular determinants of GTAC-mediated GPCR degradation.

a, Schematics showing potential determining factors of GTAC activity. b, Cell-surface AT1R (anti-FLAG flow cytometry) and total AT1R (GFP fluorescence) in FLAG–AT1R–GFP+ TfR1+ Expi293 cells following 1.5 h pretreatment with 1 µM MG132, 250 nM bafilomycin A1 (Baf) or 25 µg ml−1 cycloheximide (CHX), then AT1R GTAC or Nb–Fc treatment (6 h, 6.25 nM). Data represent mean ± s.d. from n = 3 biological replicates. c, Cell-surface AT1R (anti-HA) and total AT1R (GFP) over 24 h in FLAG–HA–AT1R–GFP+ TfR1+ Expi293 cells after treatment with 1 μM ANGII or 50 nM FLAG GTAC. Data represent mean ± s.d. from n = 3 biological replicates. d, Cell-surface AT1R (anti-HA) in TfR1+ Expi293 cells expressing FLAG–HA–AT1R–GFP with AT1R WT, T282M, or TSTS/A after 1 h treatment with 1 μM ANGII, 50 nM FLAG GTAC or 10 nM AT1R GTAC. Data represent mean ± s.d. from n = 3 replicates. Panels b–d show statistical comparisons using unpaired two-tailed t-tests. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Exact P values are listed in Supplementary Table 7. e, Confocal microscopy of CCR6–GFP+ HeLa cells with 24 h TfR1HIGH GTAC treatment, apart from bafilomycin (n = 8 cells), chloroquine (CQ; n = 6 cells) or MG132 (n = 5 cells). Untreated, n = 8 cells; no inhibitor, n = 6 cells. f, pH-insensitive binding of CCR6 Fab probed with H31A mutation. g, CCR6 Fab light-chain CDR sequence with H31 (yellow). h, Cell surface binding of 5 nM CCR6 Fab–Fc or 125 nM H31 CCR6 Fab–Fc to CCR6+ and CCR6− Jurkat cells. Data represent mean ± s.d. from n = 3 biological replicates. i, Western blot analysis and quantification of total CCR6 levels in CCR6+ Jurkat cells after 18 h H31A CCR6–TfR1HIGH GTAC treatment (n = 3 biological replicates). Diagrams in a and f created in BioRender; Zhou, X. https://biorender.com/hymki1y (2026).

To test whether GTAC activity depends on native GPCR trafficking pathways, we evaluated AT1R mutants that impair β-arrestin recruitment and ligand-induced internalization. Consistent with the imaging results, FLAG GTAC-induced sustained AT1R degradation, whereas ANGII induced transient internalization and no degradation (Fig. 4c). Both AT1R T282M47 and AT1R TSTS/A48 showed reduced ANGII-mediated internalization (Fig. 4c,d). By contrast, both FLAG GTAC and AT1R GTAC efficiently depleted and degraded all AT1R variants, with no significant differences compared with the wild-type receptor (Fig. 4d). These findings indicate that GTAC-mediated internalization and degradation occur independently of β-arrestin-dependent internalization, demonstrating that GTACs operate orthogonally to canonical GPCR signalling pathways.

We next characterized the role of pH-dependent binding. Reduced affinity at endosomal pH may facilitate receptor dissociation from GTAC, thereby reducing co-recycling and enhancing lysosomal degradation. Among the GTACs examined here, the CCR6 Fab contains a histidine residue (H31) in light-chain CDR1, showing pH-dependent binding to CCR6+ Jurkat cells (Fig. 4f,g). Flow cytometry confirmed that the H31A mutation abolished the pH sensitivity, while also reducing overall binding (Fig. 4h). The H31A TfR1HIGH GTAC exhibited about three-fold lower degradation activity compared with the unmodified GTAC (Fig. 4i). These results indicate that pH-dependent binding, altered target binding affinity, or both can influence the degradation activity of the CCR6 GTAC.

We next studied how the geometry, affinity and binder valency of GTACs influence activity across AT1R, RXFP1 and CCR6 (Extended Data Figs. 1h, 2i and 6c–e, Supplementary Figs. 2–5 and Supplementary Tables 1–3). We tested 2:2, 2:1, 1:1 and 1:2 GPCR:TfR1 binding geometries and a range of TfR1 binder affinities. Despite differences in potency, most GTACs achieved substantial receptor depletion and inhibition (Supplementary Tables 1–3). However, the optimal configuration varied by target. RXFP1 GTACs showed improved activity with a 2:1 target:TfR1 binder stoichiometry, whereas AT1R favoured bivalent TfR1 engagement, and CCR6 favoured a 1:1 configuration.

We interrogated the factors that influence TfR1 co-depletion and found that the extent of TfR1 co-depletion depends on both target context and GTAC design. RXFP1 GTACs depleted RXFP1 with minimal impact on TfR1 levels (Extended Data Fig. 6f,g). For AT1R, bivalent TfR1 engagement increased AT1R depletion but also promoted partial TfR1 co-depletion, whereas monovalent TfR1-binding formats reduced this effect (Extended Data Fig. 1h). For CCR6, lowering TfR1 affinity allowed target degradation without TfR1 loss (Fig. 3e–i and Extended Data Fig. 7d,g). Moreover, comparison of target:TfR1 expression suggested that TfR1 co-depletion also depends on the relative receptor abundance (Extended Data Fig. 6h).

These studies highlight the modularity and robustness of the GTAC platform, as effective receptor depletion and antagonism were achieved across diverse molecular configurations. Meanwhile, although geometry, valency and affinity can be tuned to optimize target degradation efficacy and TfR1 preservation, the preferred design varies across GPCR targets.

ANGII inhibition in hiPSC cardiomyocytes

AT1R is among the most extensively targeted GPCRs; small-molecule angiotensin receptor blockers (ARBs) are widely used to treat hypertension, heart failure and chronic kidney disease49. However, many patients exhibit persistent or treatment-resistant hypertension despite combination therapies, including ARBs and angiotensin-converting enzyme (ACE) inhibitors, suggesting that more complete suppression of AT1R signalling through receptor depletion may be beneficial. Furthermore, ARBs and ACE inhibitors are contraindicated during pregnancy because they cross the placenta and cause on-target fetal toxicity, motivating interest in protein-based therapeutic modalities that do not cross the placental barrier50.

To evaluate whether GTAC-mediated receptor degradation can inhibit endogenous AT1R signalling, we used human iPSC-derived cardiomyocytes. It is well established that acute ANGII–AT1R signalling modulates mechanical contractions in cardiomyocytes, whereas chronic signalling drives hypertrophy and pro-fibrotic remodelling51. We therefore tested whether AT1R GTACs could block acute ANGII responses in iPSC cardiomyocytes using electrochromic optical recording (ECORE), a label-free method that simultaneously measures electrical activity and mechanical contraction52 (Fig. 5a,b).

Fig. 5: GTACs show robust activity in primary cells, in iPSC-derived cardiomyocytes and in vivo.

a, Schematic of hiPSC-cardiomyocytes cultured on electrochromic polymer thin films. b, Single representative ECORE recording trace of action potential and mechanical signal from one cell. c, Mechanical contraction across 1 µM ANGII injection with and without 18 h pretreatment with 70 nM AT1R GTAC. d, Beat interval before and after 1 µM ANGII or DMSO injection, with and without AT1R GTAC pretreatment. Bars represent median; error bars represent interquartile range; n = individual spots within the same culture. e, CCL20-mediated chemotaxis of CCR6+ Jurkat cells after 2 h TfR1LOW GTAC or PF-07054894 treatment. Statistical comparisons performed using unpaired two-tailed t-tests. f, Flow cytometry validation of naive CD4+ CD45+ T cell isolation and CXCR3− CCR6+ CD4+ TH17 cell differentiation. g, Flow cytometry staining of CCR6 in primary TH17 cells after 2 h GTAC and control treatment. h, CCL20-mediated chemotaxis of TH17 cells using Transwell assay after 2 h GTAC or control treatment. i, Flow cytometry showing binding of TfR1HIGH or TfR1LOW GTACs to CD3+ pan-T cells or TH17 cells. In e, g–i, data represent mean ± s.d. from n = 3 biological replicates. j, Schematic of in vivo experimental design. k, Body weight following IMQ and TfR1HIGH GTAC or control administration. l, Time course of ear swelling, erythema, scaling and PASI scores. m, qPCR analysis of Il17a, Il1b and Cxcl2 mRNA in the ear skin, normalized to Gapdh. n, Flow cytometry staining of human (h)CCR6 on neutrophils and inflammatory monocytes in the ear skin. In l–n, statistical comparisons are performed using two-tailed Mann–Whitney tests; data represent mean ± s.e.m. from n = 4 (untreated, 3 for Il17a), 6 (TfR1HIGH GTAC) or 8 (CCR6 Fab–Fc) ear skin samples from a single experiment. PASI, psoriasis area severity index. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Exact P values are provided in Supplementary Table 7. Diagrams in a and j created in BioRender; Zhou, X. https://biorender.com/hymki1y (2026).

Source data

ANGII stimulation increased cardiomyocyte beating rate, shortening the beat interval substantially (Fig. 5c). Pretreatment with AT1R GTAC suppressed this response by about two-fold at 7 nM and completely abolished the response at 70 nM (Fig. 5d). Moreover, ANGII induced substantial mechanical beat-to-beat variation in contraction waveforms and intervals. AT1R GTAC inhibited 85–95% of this response (Extended Data Fig. 11a,b). Although the AT1R Nb–Fc control partially suppressed these effects through Nb-mediated blockade of ANGII binding, AT1R GTAC produced more complete inhibition (Extended Data Fig. 11c,d). Amplitude, per-beat time to peak and per-beat electromechanical delay were not significantly affected by ANGII (Extended Data Fig. 11e–g). These studies demonstrate that GTAC-mediated receptor depletion provides more complete AT1R pathway suppression than competitive inhibition in a human iPSC-derived cardiac model.

CCR6 antagonism in TH17 cells and in vivo

Finally, we evaluated the CCR6 GTACs in additional endogenous systems and compared them with clinical-stage molecules. The TfR1LOW CCR6 GTAC achieved greater inhibition of chemotaxis than PF-07054894, a small-molecule CCR6 antagonist that advanced to phase Ib clinical trials for ulcerative colitis42 (Fig. 5e). The CCR6 Fab also exhibited greater selectivity for CCR6 over CCR7 and CXCR2, two chemokine receptors inhibited by PF-07054894 (Extended Data Fig. 7j), which may help to limit off-target effects, including impaired immune tolerance and neutropenia53.

Primary human T helper 17 (TH17) cells endogenously express CCR6 and are the predominant responders to CCL20-mediated chemotaxis in autoimmune diseases23. We differentiated human TH17 cells from naive CD4+ T cells (Fig. 5f). TfR1LOW GTAC and H7 GTAC induced up to 95% CCR6 internalization with sub-nanomolar potency after 2 h treatment (Fig. 5g). H7 GTAC induced receptor degradation in CCR6+ Jurkat cells over 18 h (Extended Data Fig. 12a–c). GTAC treatment for 2 h inhibited about 90% of CCL20-induced TH17 cell chemotaxis, in contrast to the parental CCR6 antagonist (Imax = 55%) (Fig. 5h). Moreover, TfR1LOW GTAC preferentially bound CCR6+ TH17 cells over CD3+ pan-T cells, in contrast to TfR1HIGH GTAC (Fig. 5i), demonstrating that the context-dependent GTAC ensures target-dependent binding in primary cells.

We next evaluated GTAC activity in vivo using an imiquimod (IMQ)-induced psoriatic inflammation model in human (h) CCR6 transgenic C57BL/6N mice, in which CCR6 activation directs IL-17-producing cell migration to sites of inflammation54. The transgenic model was used because the CCR6 binder does not cross-react with mouse CCR6 (mCCR6). We validated that mouse CCL20 induced chemotaxis in hCCR6+ Jurkat cells, although with lower potency compared with hCCL20 (Extended Data Fig. 12d). TfR1HIGH GTAC achieved the strongest binding to mTfR1 and the most potent depletion of hCCR6 in hCCR6+ mouse MC38 cells (Extended Data Fig. 12e–f).

Imiquimod (topical application to the ears) and TfR1HIGH GTAC or CCR6 Fab–Fc (intraperitoneal injection, 10 mg kg−1) were administered to mice for 6 consecutive days and did not induce weight changes (Fig. 5j,k). GTAC significantly reduced clinical scores and histological epidermal hyperplasia compared with the CCR6 Fab–Fc (Fig. 5l and Extended Data Fig. 12g), suppressed the expression of Il17a, Il1b and Cxcl2 in the skin (Fig. 5m), and depleted hCCR6, but not hCCR6 mRNA, on neutrophils and inflammatory monocytes, which are key contributors to psoriasis-like inflammation through CCR655,56 (Fig. 5n and Extended Data Fig. 12h,i). Despite the intrinsic antagonist activity of the CCR6 Fab–Fc, GTAC produced greater therapeutic effects, demonstrating that receptor degradation provides additional pharmacological activity beyond antagonism alone.

To assess potential effects on TfR1 biology, mouse TfR1 (mTfR1) immunofluorescence was measured in critical organs following CCR6 TfR1HIGH GTAC treatment. TfR1 levels were not reduced by GTAC treatment in the spleen or brain (Extended Data Fig. 13a–c). Liver mTfR1 expression was low and unchanged by GTAC treatment (Supplementary Fig. 10). Furthermore, the non-competitive TfR1HIGH or TfR1LOW Nbs did not impair TfR1-mediated iron uptake, as measured by intracellular labile Fe2+ using the FerroOrange assay (Extended Data Fig. 13d). In vivo, the antibodies remained in circulation for up to 3 days (TfR1HIGH GTAC) or 10 days (H7 GTAC, TfR1LOW GTAC and CCR6 Fab–Fc) (Extended Data Fig. 13e).

These findings demonstrate activity of CCR6 GTACs in primary cells and in vivo, supporting their potential as a new strategy for targeting this pathway. These studies, however, were performed in a humanized model with reduced mouse CCL20 activity towards hCCR6 and reduced GTAC binding to mTfR1. Future studies using cross-reactive or mouse surrogate GTACs will further evaluate their therapeutic potential.

Discussion

Extracellular targeted protein degradation has emerged as a new mechanism to regulate secreted and membrane proteins17,18,57. Here we describe a comprehensive study of extracellular GPCR degrader engineering. Across seven targets, GTACs consistently achieved target degradation and potently inhibited GPCR signalling in engineered cell systems, cancer models, iPSC-derived cardiomyocytes and human primary cells, exhibiting one to two orders of magnitude greater potency than conventional antagonists. These findings establish induced endocytosis and lysosomal degradation as a generalizable mechanism for potent and durable GPCR antagonism.

For RXFP1, the exquisite sensitivity to RLN2 poses formidable obstacles to conventional antagonism. RXFP1 GTACs inhibited cAMP signalling with >200-fold greater potency than a sub-nanomolar affinity Nb antagonist34. These studies result in a picomolar-potency biologic antagonist of RXFP1 and show that targeted degradation can overcome challenges posed by high ligand sensitivity. For CCR6, GTACs induced complete receptor degradation and functional inhibition both in primary immune cells and in vivo. Chemokine receptors have remained challenging to inhibit because of high local chemokine concentrations and high-affinity, multi-epitope ligand–receptor interactions43. We show that targeted degradation offers a means to overcome these limitations.

We also extend extracellular degradation to BILF1, a GPCR with ligand-independent signalling that cannot be addressed using competitive antagonists. GTACs effectively suppressed BILF1 signalling and reversed its downstream effects on NF-κB signalling and MHC-I expression. These findings provide one of the first demonstrations that targeted degradation can inhibit constitutive receptor signalling and may enable targeting of other classes of proteins with similar signalling properties.

Our work identifies new protein engineering principles for degrader design. Reducing effector affinity created a context-dependent degrader that may be applicable to other target classes and endocytic effectors. Consistent with recent studies58,59, we found that pH-sensitive binding may enhance degradation by facilitating target release in the endosomes. However, the robust activity of multiple GTACs lacking pH-sensitive binders suggests that this is an optional design feature to optimize on a target-by-target basis. Our studies also advance TfR1-targeting biologic engineering through the use of non-competitive TfR1 binders and affinity engineering. While these approaches can minimize perturbation of native TfR1 biology, the long-term effects of harnessing the TfR1 pathway warrant further evaluation. Cross-reactive or surrogate GTACs will enable more comprehensive assessment of efficacy, pharmacology and safety in vivo.

Apart from therapeutic applications, GTACs provide a platform for probing GPCR signalling and function without requiring genetic manipulation. Given the widespread use of FLAG-tagged GPCR constructs60, FLAG GTACs offer a useful research tool when target-specific binders are unavailable. GTACs also provide an opportunity to dissect spatially organized GPCR signalling8. GTACs are not dependent on β-arrestin recruitment mechanisms, allowing unbiased investigation of how receptor mutations and cellular pathways regulate post-endocytic trafficking and signalling.

GTACs do not eliminate the need for high-quality GPCR binders but reduce the requirement for intrinsic antagonistic activity. As antibody discovery and computational design continue to advance, access to GPCR binders should become increasingly tractable. The GTAC mechanism therefore has the potential to greatly expand the range of targetable epitopes and druggable GPCRs. Occupancy-based inhibition and degradation, however, need not be mutually exclusive mechanisms. In some contexts, combining ligand blockade with the GTAC mechanism can further strengthen inhibition.

Overall, our findings position GTACs as a distinctive addition to the repertoire of GPCR-modulating strategies. Although further studies are needed to define the translational scope of this approach, the principles established here provide a foundation for expanding extracellular targeted degradation as both a therapeutic modality and a tool for probing GPCR biology.

Methods

Plasmid construction

Plasmids were constructed using standard Gibson cloning methods. DNA fragments of APJ, BILF1, TfR1 scFv H7 (ref. 29), TfR1 VHH-A (ref. 30), APJ VHH JN241 (ref. 32), A2AR VH/VL domains Hu3F6 H5-L5 (ref. 33), CCR6 VH/VL domains from 4H11 (ref. 44) and Fc variants (knobs-into-holes Fc for heterobispecific constructs28, G236R/L328R/D356E/L358M and H435A variants to abolish FcγR and FcRn interactions61,62,63 and N297G variant for increased stability and developability64) were synthesized by Integrated DNA Technologies or Twist Bioscience. Binder sequences and/or Fc variants were subcloned into a pFuse-hIgG1 vector (InvivoGen) for mammalian expression.

The pcDNA-FLAG–AT1R and pCDNA-FLAG–RXFP1 plasmids were obtained as previously described31,65. AT1R VHH and RXFP1 VHH were developed as previously described31,66. ADORA2A-DuET was a gift from Thomas Sakmar (Addgene plasmid 213174; http://n2t.net/addgene:213174; RRID:Addgene_213174). The pCDH-TfR-mApple lentiviral vector was a gift from Angus Johnston (Addgene plasmid 179384; http://n2t.net/addgene:179384; RRID:Addgene_179384). GPCR sequences were fused with FLAG tag, HA tag and/or GFP and subcloned into pCDH or Tet-On lentiviral vectors for stable dox-inducible expression (Supplementary Fig. 8). The TfR1 sequence was fused with a v5 tag and/or BFP and subcloned into pCDH lentiviral vector for stable overexpression.

NFAT-luciferase reporter was a gift from Toren Finkel (Addgene plasmid 10959; http://n2t.net/addgene:10959; RRID:Addgene_10959). The pGloSensor cAMP plasmid was obtained from Promega. CCR6-Tango was a gift from Bryan Roth (Addgene plasmid 66243; http://n2t.net/addgene:66243; RRID:Addgene_66243). Notably, the C-terminal tail also included a V2 vasopressin receptor tail to enhance β-arrestin recruitment. The NF-κB GFP reporter was synthesized by Twist Bioscience and subcloned into the pCDH vector. The AT1R sequence was subcloned into the Tango vector for the β-arrestin recruitment assay60. RAB5, RAB7, RAB11, LAMP1, CCR7 and CXCR2 sequences were synthesized by Twist Bioscience and subcloned into pLLM lentiviral vectors.

Cell lines and culture

HEK293T, Lenti-X 293T, 293GP (Takara Bio), MC38 and HeLa cells and a stable pTetOn-FLAG–AT1R adherent Expi293 cell line were cultured in Dulbecco's modified Eagle's medium (DMEM) (Thermo Fisher Scientific) with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin, incubated at 37 °C with 5% CO2. Jurkat, MDA-MB-231, OVCAR-8, PC9 and Raji cells were cultured in RPMI-1640 (Thermo Fisher Scientific) with 10% FBS and 1% penicillin/streptomycin, incubated at 37 °C with 5% CO2. All other Expi293 cells (Thermo Fisher) were cultured in Expi293 Expression Medium (Thermo Fisher Scientific), incubated with gentle shaking at 37 °C with 8% CO2. OVCAR-8 was generously provided by the Robert Rottapel lab (Princess Margaret Cancer Centre, University of Toronto). PC9 cells with doxycycline-inducible TfR1 knockdown were generously provided by the Pasi Jänne lab (Dana-Farber Cancer Institute). All other cell lines were originally sourced from ATCC and were authenticated by the supplier using standard genotyping methods. Cells were not tested for mycoplasma. Human iPSCs were obtained from Stanford Cardiovascular Institute Biobank and used per Institutional Review Board/Stem Cell Research Oversight Panel guidelines, adhering to federal, state and Stanford University human stem cell research policies (protocol no. SCRO-745). Human peripheral blood mononuclear cells were isolated from whole blood collars (Crimson Core, Brigham and Women's Hospital).

Protein expression

Antibodies were expressed in Expi293 cells (Thermo Fisher Scientific) using FectoPRO (Thermo Fisher Scientific) with the manufacturer’s standard protocol. Antibodies were purified using Protein A affinity chromatography and buffer-exchanged into PBS. Proteins were kept at 4 °C for up to 1–2 weeks or flash frozen for storage at −80 °C. The purity and integrity of proteins were confirmed by SDS-PAGE electrophoresis (Supplementary Fig. 7).

Engineered cell generation

Lentivirus

Lenti-X 293T cells were transfected with the carrier plasmid and lentiviral helper plasmids (psPax2 and pMD2.G) using a standard Lipofectamine3000 protocol (Thermo Fisher Scientific). The following day, the medium was replaced. Viral supernatant was harvested 48–72 h post-transfection, centrifuged at 15,000g for 1 min to remove cell debris, and immediately used or stored at −80 °C. Cells were infected with lentivirus diluted 1:5 with 8 μg ml−1 polybrene. The medium was replaced the next day, and cells were subsequently cultured with selection antibiotics (2 μg ml−1 puromycin, 200 μg ml−1 hygromycin, or 100 μg ml−1 zeocin) or sorted.

AT1R NFAT activation

Adherent Expi293 cells with stable dox-inducible FLAG–AT1R were obtained as previously described31 and were further transduced with BFP-TfR1-v5. For AT1R degradation and imaging, adherent Expi293 cells were generated by transitioning from shaking at 37 °C with 5% CO2 to shelf incubation at 37 °C with 5% CO2, while gradually adapting cells from Expi293 medium to DMEM with 10% FBS + 1% Pen-Strep over 2–3 weeks. Adherent Expi293 cells were transduced with FLAG–AT1R–GFP and BFP–TfR1–v5 using lentivirus and sorted to obtain homogeneous AT1R and TfR1 expression levels.

Other AT1R variants

Cells were transduced with TfR1 constructs, sorted and transfected with AT1R variants using Lipofectamine3000.

β-Arrestin recruitment

HTLA cells were transduced with TfR1 using lentivirus as described above, then transfected with AT1R or CCR6 Tango constructs.

APJ

HEK293T cells were transduced with TetOn-FLAG–APJ and TfR1–v5 using lentivirus and cultured with puromycin and zeocin to obtain a stable cell line.

BILF1

Raji cells were seeded in a six-well plate and transduced with TetOn-FLAG–BILF1 through lentiviral spinfection (centrifugation at 750g, 37 °C, for 2 h). Cells were expanded and subsequently sorted to obtain a homogeneous stable cell population with high BILF1 expression.

RXFP1 and CCR6 imaging

HeLa cells were transduced with FLAG–RXFP1–GFP or FLAG–CCR6–GFP, BFP–TfR1–v5 and endosome/lysosome markers (RAB5, RAB7, RAB11 and LAMP1) fused to mCherry with lentivirus. PC9 cells with doxycycline-inducible TfR1 knockdown were transduced with FLAG–CCR6–GFP. Cells were sorted to obtain homogeneous GFP or mCherry fluorescence levels.

CCR6

Jurkat cells were infected with doxycycline-inducible FLAG–CCR6 lentivirus diluted 1:7 with 8 μg ml−1 polybrene, spinning at 890g at 30 °C for 2 h. The medium was changed the next day with 20% FBS RPMI-1640, and cells were further cultured with 2 μg ml−1 puromycin.

CCR6 Fab binding to CCR6 compared with CCR7 and CXCR2

Lenti-X 293 T cells transfected in a 10 cm plate with 15 μg of the CCR6-TANGO plasmid using Lipofectamine3000. Retrovirus 6–7 × 106 293GP cells were plated in a 10 cm plate. The following day, cells were co-transfected with 9 μg  MMLV Firefly Luciferase-mNeonGreen and 4.5  μg RD114 packaging plasmid. Media was replaced after 18–24 h, and retroviral supernatant was collected twice at 48 h and 72 h after transfection. A non-tissue culture-treated six-well plate was coated with 25 μg ml−1 retronectin (Takara) in PBS overnight. After blocking with 2% BSA in PBS, 1.5 ml of Firefly Luciferase-mNeonGreen retroviral supernatant was added. The plate was then centrifuged at 2,000g at 32 °C for at least 2 h. Viral supernatant was then removed, and 0.5 × 106 doxycycline-inducible FLAG–CCR6 cells were plated in each well for transduction before expansion and cell sorting for mNeonGreen expression.

Western blots

M3R

BFP–TfR1-v5+ Expi293 cells were transfected with FLAG–M3R in a six-well plate at 70–80% confluency with Lipofectamine3000 (described above). The next day, cells were trypsinized and seeded in 12-well plates. After specified treatments, cells were harvested in 2 mM EDTA in PBS, pelleted by spinning at 500–800g for 1–5 min, and incubated with RIPA lysis buffer (Thermo Fisher) on ice for 30–60 min. The lysates were centrifuged at 20,000g, and the clarified supernatant was sonicated. Samples were loaded into SDS-PAGE gels (Bio-Rad) in equal amounts with 4× Laemmli sample buffer (Bio-Rad) with BME.

CCR6

Doxycycline-inducible CCR6+ Jurkat cells were treated with 250 ng μl−1 doxycycline. After 24 h, 9 × 105 cells were seeded in each well of a 12-well plate. After specified treatments, cells were harvested as described above. Cell pellets were resuspended in 1% digitonin, 150 mM NaCl and 100 mM Tris pH 7.5 and 1× protease inhibitor (Thermo Fisher) and incubated on ice for 45 min. Lysates were clarified, sonicated and loaded onto a gel as described above.

All gels were run at 150 V for 60 min and transferred onto LF-PVDF membranes using Trans-Blot Turbo Western Blotting Transfer System (Bio-Rad). Membranes were blocked for 15–30 min with 5% milk in TBST at room temperature, washed with TBST, stained with primary antibody overnight at 4 °C, washed with TBST, stained with secondary antibody (anti-mouse IgG or anti-rabbit IgG) for 1 h at room temperature, washed with TBST and imaged with a LICOR imager and Image Studio software (v.5.2.5). Bands were quantified with ImageJ software (v.2.16). Uncropped blots are shown in Supplementary Fig. 1. Western blot biological replicates are shown in Supplementary Fig. 9. Antibodies used were rabbit anti-β-actin (Cell Signaling Technology, 4970S, 1:1,000), mouse M1 anti-FLAG (expressed from hybridoma and purified in-house, 1:500, requires 2 mM CaCl2), anti-CCR6 (Thermo Fisher, 14-1969-82, 1:300), anti-TfR1 (Thermo Fisher, 13-6800, 1:1,000), IRDye 800CW goat anti-mouse IgG (LICOR, 926-32210, 1:10,000), IRDye 680RD goat anti-rabbit IgG (LICOR, 926-68071, 1:10,000).

Flow cytometric receptor depletion assays

AT1R

Stable FLAG–AT1R–GFP+ BFP–TfR1–v5+ Expi293 cells were seeded in 24-well plates and induced with 1 μg ml−1 dox for 24 h. After specified treatments at 70–80% confluency, cells were collected with 300 μl cold EDTA (1 mM in PBS) for flow cytometry analysis.

AT1R mutant comparisons

BFP–TfR1–v5+ Expi293 cells were transfected with FLAG–HA–AT1R–GFP, AT1R WT, T282M or T332A/S335A/T336A/S338A using Lipofectamine3000 and seeded for the assay as described above.

RXFP1, A2AR and AT1R variant comparisons

Expi293 cells were seeded in a 125 ml flask at 3 × 106 cells per ml and transfected with FLAG-tagged and/or GFP-fused GPCR and TfR1 using FectoPRO (described above). The next day, cells were enhanced with 360 μl valproic acid (300 mM) and 720 μl glucose (20%). Two days after transfection, cells were seeded in 24-well plates.

APJ

TetOn-FLAG–APJ+ TfR1–v5+ HEK293T stable cells were seeded in a 24-well plate and induced with 50 ng ml−1 dox for 24 h. After specified treatments, cells were harvested for flow cytometry analysis as described above.

BILF1

HEK293T cells were transfected with TetOn-FLAG–HA–BILF1, TfR1–v5 and NF-κB–GFP reporter (internalization and signalling assays were performed on the same cells) in a six-well plate. The next day, cells were trypsinized and seeded into 12-well plates, induced with doxycycline, treated with GTAC or controls for the specified duration and harvested for flow cytometry analysis.

Temperature comparisons

TetTon-FLAG–BILF1+ Raji cells were seeded at 50,000 per well in a 96-well plate, induced with 1 µg ml−1 doxycycline for 24 h. After specified treatments on ice or at 37 °C, cells were harvested for flow cytometry.

CCR6

Doxycycline-inducible CCR6+ Jurkat cells were induced with 25 ng µl−1 doxycycline for 24 h. A total of 2.5 × 105 cells were seeded in each well of a 96-well flat-bottom plate and were treated with GTAC for 2 h or 24 h before being analysed using flow cytometry. A total of 6 × 104 PC9 cells with constitutive CCR6 expression and doxycycline-inducible TfR1 knockdown were seeded in a 24-well plate and were induced with varying doxycycline concentrations for 48 h. After specified GTAC treatments, cells were harvested for flow cytometry. For assessing degradation, 2 × 105 CCR6-GFP+ PC9 cells were seeded in a 12-well plate, treated with GTAC the next day for 18 h and harvested for flow cytometry.

Flow cytometry and fluorescence-activated cell sorting

Harvested cells were centrifuged at 300g for 5 min. Pellets were washed 1× with fluorescence-activated cell sorting (FACS) buffer (ice-cold PBS + 2% BSA). Cells were either analysed (flow cytometry) or sorted (FACS) directly, or incubated with primary antibody in FACS buffer for 20–30 min on ice. Cells were washed 3× and resuspended in FACS buffer for flow analysis.

On-cell binding

Cells were treated with Ab/GTAC diluted in PBS and were incubated on ice for 10–20 min, stained with secondary antibody for 10–20 min and washed 3× before being resuspended for flow analysis. Flow cytometry was performed using a CytoFLEX cytometer (Beckman Coulter, V0-B2-R2) and CytExpert software (v.2.3.1.22). FACS was performed using a CytoFLEX SRT cell sorter (Beckman Coulter, V0-B2-Y5-R3) and CytExpert SRT software (v.1.1.0.10007). Data were analysed with FlowJo (v.10.8.0). Representative gating strategies for live cells and single cells are shown in Supplementary Fig. 6.

Primary antibodies

The primary antibodies were Alexa Fluor 647 anti-FLAG Tag (M1) (1:500 dilution, purified from hybridoma supernatant and conjugated in-house), Alexa Fluor 647 anti-HA.11 Epitope Tag (1:1,000 dilution, BioLegend 682404), Alexa Fluor 488 anti-V5 Tag (1:1,000 dilution, Thermo Fisher 37-7500-A488), FITC anti-human CD71 (1:500 dilution, BioLegend 334104), and Alexa Fluor 647 anti-human CD71 (1:500 dilution, BioLegend 334118).

Secondary antibodies

The secondary antibodies were Alexa Fluor 647 Protein A (1:500 dilution, Thermo Fisher P21462) and Alexa Fluor 488 Protein A (1:500 dilution, Thermo Fisher P11047).

BILF1 signalling and functional assays

NF-κB activation

HEK293T cells were transfected with TetOn-FLAG–HA–BILF1, TfR1–v5 and NF-κB–GFP reporter in a six-well plate. The next day, cells were trypsinized and seeded into 12-well or 24-well plates, induced with doxycycline, subjected to specified treatments and harvested for flow cytometry analysis.

MHC-I expression

TetOn-FLAG–BILF1+ Raji cells were seeded at 25,000 per well in a 96-well plate, induced with 1 µg ml−1 doxycycline for 24 h, treated with GTAC or controls for 24 h, and harvested for flow cytometry analysis.

RXFP1 cAMP assays

HEK293T cells were transfected with FLAG–RXFP1, TfR1–v5 and Promega GloSensor plasmid (1 μg each per well) using Lipofectamine3000. The next day, cells were trypsinized, seeded in a 96-well plate (50,000 cells per well) and incubated 24 h. Cell culture medium was removed and replaced with Ab or GTAC diluted in CO2-independent medium with 10% FBS + 5% d-luciferin reagent: 40 mg ml−1 d-luciferin potassium salt (GoldBio eLUCK-500) in 10 mM HEPES buffer. Cells were incubated for 2 h (unless otherwise specified) in the dark. Luminescence was measured with a CLARIOstar PLUS microplate reader (BMG Labtech, used for all plate-reader assays unless otherwise stated) to obtain the pre-read; recombinant human RLN2 (Fisher Scientific 3596RN025) was added immediately, and another measurement was taken after 30 min. Unless otherwise specified, Ab or GTAC dose–response curves were generated, and IC50 values were obtained using 1 nM RLN2 to activate signalling; RLN2 dose–response curves and EC50 values were obtained after treating cells with 10 nM Ab/GTAC.

Time course

HEK29T cells were transfected with FLAG–RXFP1, TfR1–v5 and GloSensor as described above. After specified treatments, cell culture medium was then removed and replaced with d-luciferin reagent diluted in CO2-independent medium, and the assay proceeded as described above, using 1 nM RLN2 to activate signalling.

RXFP1 expression compared with cAMP correlation curve

A FLAG–RXFP1+ TfR1+ HEK293T stable cell line was generated with lentivirus. Cells were seeded in a 24-well plate (50% confluency) and induced with varying doxycycline concentrations for 24 h before harvesting for flow cytometry. In parallel, cells were transfected with GloSensor plasmid as described above. The next morning, cells were trypsinized, seeded in a 96-well plate (50,000 cells per well) and induced with varying doxycycline concentrations for 24 h before proceeding with the cAMP assay as described above, using 1 nM RLN2 to activate signalling.

Affinity-matured RXFP1 Nb (RX002i4)

HEK293T cells were seeded in white, clear-bottomed 96-well plates, coated with poly-d-lysine at 40,000 cells per well. The cells were transfected with FLAG–RXFP1 and Promega GloSensor plasmids for at least 24 h (20 ng each per well) using FuGENE according to the manufacturer’s protocol. Post-transfection, the cells were incubated in CO2-independent medium with 10% (v/v) FBS and 2 mg ml−1 of d-luciferin (40 μl per well) and incubated for 2 h at room temperature in the dark. Luminescence was measured as a pre-read using a SpectraMax M5 microplate reader before cells were treated with RX002i4 (at 1 mM per well) or PBS for 20 min. Following this, RLN2 was added to the wells in a dose-dependent manner for 30 min, and luminescence was measured.

RXFP1 Nb binding and competition with RLN2

Expi293F TetR cells were transfected with FLAG–RXFP1 or empty pcDNA-Zeo-tetO vector using FectoPro according to the manufacturer’s instructions. Cells were enhanced with 0.4% glucose and 3 mM sodium valproic acid 24 h post-transfection. Protein expression was induced with 4 μg ml−1 doxycycline 48 h post-transfection. Cells were aliquoted into a V-bottomed, 96-well plate (100,000 cells per well) 24 h post-induction and harvested by spinning at 200g for 5 min at 4 °C. Cells were washed once and blocked for 20 min in HEPES-buffered saline containing 1% (v/v) FBS and 2 mM calcium chloride (flow cytometry buffer) at 4 °C, with shaking.

Direct binding

Cells were incubated with a dilution series of RX002i4 Nb for 60 min at 4 °C, with shaking. For competition, cells were incubated with 200 nM SE301 (Fc-tagged RLN-2) and a dilution series of RX002i4 Nb for 60 min at 4 °C, with shaking. Cells were washed once and incubated with M1 anti-FLAG antibody labelled with Alexa Fluor-488 (1:200) and Alexa Fluor-647 anti-V5 (1:1,000) or Alexa Fluor-647 anti-human IgG Fc (1:1,000) for 30 min. Cells were washed once and analysed on a CytoFLEX flow cytometer (Beckman Coulter) using CytExpert software.

RXFP1 signalling qPCR analysis

OVCAR-8 (500,000 cells per well) were seeded in a six-well plate and incubated overnight. Cells were starved for 6 h and subsequently treated with Ab or GTAC (100 nM) for 18 h. Cells were induced with 60 ng ml−1 rhRLN2 and incubated for 8 h. Cell lysate was harvested with Trizol (Thermo Fisher 15596026) (700 μl per well). RNA extraction was performed with Qiagen kits (Qiagen 217084, 79254, 79656), RIN scores >8.0 were confirmed for quality control, and cDNA was synthesized from 1 μg RNA per sample with Superscript IV (Life Technologies 18090010) following standard protocol. qPCR was run with Taq polymerase, dNTPs and SYBR Safe Green in-house master mix. GAPDH and ACTB primers were designed and previously validated in-house. MMP9, MMP23, NOTCH1, NOTCH3 and DVL1 primer sequences were generously provided by the Robert Rottapel Lab (Princess Margaret Cancer Centre, University of Toronto), with sequences listed in Supplementary Table 6.

AT1R NFAT assays

A stable adherent Expi293 cell line with inducible FLAG–AT1R and constitutive TfR1 was seeded in a six-well plate and incubated overnight. Cells were transfected at 70–80% confluency with NFAT-luciferase (1 μg per well using Lipofectamine3000) and incubated 16 h. Cells were trypsinized, seeded in a 96-well plate (50,000 cells per well) and induced with 1 μg ml−1 dox for 24 h. Cells were treated with Ab or GTAC for 2 h, and subsequently ANGII (MedChemExpress HY-13948) was added (9 nM ANGII was used to generate Ab or GTAC dose–response curves and obtain IC50 values). Four hours later, cell culture medium was removed and replaced with Promega Bright-Glo reagent diluted 1:1 in 10 mM HEPES buffer (40 μl per well). Luminescence was measured with a plate reader.

β-Arrestin recruitment assay

AT1R

TfR1+ HTLA cells were transfected with AT1R–Tango (1 μg per well using Lipofectamine3000). The next day, cells were trypsinized, seeded in a 96-well plate (50,000 cells per well) and induced with 1 μg ml−1 dox. After 24 h of transfection, cells were starved and treated with Ab or GTAC for 2 h before adding ANGII (200 nM ANGII was used to generate Ab or GTAC dose–response curves and obtain IC50 values). Then, 14 h later, cell culture medium was removed and replaced with Promega Bright-Glo reagent diluted 1:1 in 10 mM HEPES buffer (40 μl per well). Luminescence was measured using the Bright-Glo Luciferase Assay System (Promega) on a plate reader.

CCR6

TfR1+ HTLA cells were transfected in a 10-cm plate with 15 μg of the CCR6–TANGO plasmid using Lipofectamine3000 (Thermo Fisher Scientific). Cells were trypsinized 18 h after transfection and were seeded in serum-free DMEM into poly-d-lysine-coated 96-well clear-bottom plates with 4.5 × 104 cells per well. After 6 h, cells were pre-treated with 10 nM of CCR6 GTACs for 1 h followed by incubation with both the GTACs and CCL20 for 18 h. Luminescence was measured as described above.

F-actin polymerization

Doxycycline-inducible CCR6+ Jurkat cells were treated with 25 ng μl doxycycline. After 24 h, 2.5 × 105 cells were seeded in a 96-well flat-bottom plate and were incubated with GTACs for 2 h. A total of 10 nM of GTACs were used for experiments with only one dose. Cells were then resuspended in a V-bottom plate with 50 nM CCL20 followed by staining and fixation with phalloidin staining buffer after the indicated time. Phalloidin staining buffer consists of 1% BSA, 0.2% Triton X-100, 4% paraformaldehyde and 1 U of Flash Phalloidin Green 488 (BioLegend). Cells were stimulated for 1 min for experiments with only one time point. Cells were washed 3× and resuspended in cold 0.1% Triton X-100 in PBS for flow cytometry analysis.

Chemotaxis assay

Doxycycline-inducible CCR6+ luciferase+ Jurkat cells were treated with 25 ng μl−1 doxycycline. After 24 h, 4 × 105 cells were seeded in a 96-well flat-bottom plate and were incubated with GTACs for 2 h in 100 μl of serum-free RPMI-1640. Each 100 μl cell solution was then transferred to hanging cell culture inserts with 5 µm pores (Millipore Sigma). These inserts were placed within a 24-well plate that contained 600 μl of 100 nM CCL20, and cells were allowed to incubate for 2 h. Then, migrated cells were collected from the outer chambers of the 24-well plate and were transferred to a white 96-well flat-bottomed plate. A total of 150 µg ml−1 d-luciferin (GoldBio) was added to the cells to allow for quantification using luminescent readout with a plate reader. For chemotaxis of TH17 cells, 50 μl of migrated cells were mixed with 50 μl of Precision Count Beads (BioLegend 424902) to enable quantification of cell number using flow cytometry.

Imaging and colocalization

AT1R live imaging

FLAG–AT1R–GFP+ BFP–TfR1–v5+ Expi293 cells were used. For RXFP1 live imaging assays, FLAG–RXFP1–GFP+ BFP–TfR1–v5+ RAB11–mCherry+ HeLa cells were used. For both assays, 24-well glass-bottom optical imaging plates were pre-coated with poly-l-lysine (30 min at room temperature) and washed 3× with PBS; cells were subsequently seeded in coated plates (100,000 cells per well, 500 μl total culture volume), induced with 1 µg ml−1 dox, and incubated for 24 h. Cells were treated with Ab, GTAC or ligand for the specified time before removing the cell culture medium and adding imaging solution (500 μl PBS + 10% FBS + 2% Pen-Strep).

BILF1 immunofluorescence

FLAG–HA–BILF1+ TfR1–v5+ HEK293T cells were used. For CCR6 immunofluorescence assays, FLAG–CCR6–GFP+ BFP–TfR1–v5+ HeLa cells were used. For both assays, cells were seeded in 24-well glass-bottom optical imaging plates pre-coated with poly-l-lysine as previously described. The next day, cells were treated with GTAC or controls for the specified time. Cells were then fixed with ice-cold methanol for 10 min, washed with 70% ethanol and PBS, stained with primary antibody diluted in PBS with 3–5% BSA at 4 °C overnight, washed 3×, and stained with secondary antibody for 1–2 h at room temperature. The staining solution was washed and replaced with PBS before imaging.

Antibodies

The antibodies used were anti-RAB5 rabbit mAb C8B1 (CST 3547, 1:200); anti-RAB7 rabbit mAb D95F2 (CST 9367, 1:200); anti-RAB11 rabbit mAb D4F5 (CST 5589, 1:200); anti-LAMP1 rabbit mAb D2D11 (CST 9091, 1:200); and Alexa Fluor 647 donkey anti-rabbit IgG antibody (BioLegend 406414, 1:1,000).

CCR6 live imaging

GTACs were conjugated to Alexa Fluor 647 using Zip Alexa Fluor Rapid Antibody Labelling Kits (Thermo Fisher Scientific). HeLa cells (4 × 104) were plated in poly-d-lysine-coated 24-well glass-bottom plates 2 days before imaging. Cells were treated with GTAC for specified durations and washed 3× with PBS before imaging. For inhibitor studies, cells were pre-treated for 1 h before GTAC treatment with MG132 (MedChemExpress), Chloroquine (MedChemExpress) or Bafilomycin (MedChemExpress).

CCR6 immunofluorescence

Cells were plated in glass-bottom plates and treated with GTACs similarly to previously described before fixation and staining (as described for BILF1). All imaging was performed using a Zeiss 980 confocal/Spectral/Airyscan Microscope (Zen v.3.4, Blue edition) through the DFCI Molecular Imaging Core. Images were acquired using 405 nm, 488 nm and 594 nm lasers, processed and analysed using ImageJ (v.2.16.0) and Fiji (v.2.9.0).

Jurkat cell split-luciferase assay

Doxycycline-inducible CCR6+ Jurkat cells were either treated with 25 ng μl−1 doxycycline for 24 h or left untreated. Jurkat cells (1 × 105) were seeded in 96-well V-bottom plates and were spun down at 300g for 5 min. Cells were washed with PBS and were then resuspended with 10 nM Fc-H7-LgBit for 10 min on ice. After washing in PBS 3×, cells were resuspended with 10 nM GTAC-smBit for 10 min on ice and were washed another 3× in PBS. Cells were then resuspended in Nano-Glo Luciferase Assay Substrate (Promega, N1110, 1:100) in PBS and moved to a white-bottom 384-well plate, and luminescence was measured with a plate reader.

ECORE with iPSC-cardiomyocytes

ECORE sample fabrication

Samples were prepared as described previously67. In brief, a 3D-printed well was glued to a UV–ozone-cleaned, ITO-coated glass slide (resistance 70–100 Ω). A PEDOT:PSS film was electropolymerized onto the ITO from a 10 mM 3,4-ethylenedioxythiophene (EDOT) (483028, Sigma-Aldrich) with 2 w/w% poly(sodium 4-styrenesulfonate) (PSS) (molecular weight 70,000, Sigma-Aldrich) aqueous solution by applying a constant 100 µA for 40 s, producing a 28 mm2, about 60 nm-thick film.

ECORE optical setup

The ECORE optical hardware was identical to our previous report67 except for the laser source. A 200-mW, 532-nm diode-pumped solid-state laser (Spectra-Physics) was single-mode-fibre-coupled and split by polarizing beam-splitter cubes into a sample and a reference beam; half-wave plates set their polarization (sample beam s-polarized) and intensity ratio. The sample beam was focused to an approximately 30 µm spot onto a BK-7 equilateral prism and the laser incident angle set to 68°. The reflected sample and reference beams were combined on a differential photodetector, suppressing common-mode noise. Clampfit 11 and Python (v.3.12) were used for electrochromic optical recording data analysis.

hiPSC cardiomyocyte culture

The ECORE samples were disinfected with 70% EtOH for 2 h. After rinsing with water three times, the polymer film was coated with 1 mg ml−1 poly-l-lysine in 1× phosphate-buffered saline (PBS) solution (21-040-CM, Corning) for 30 min, followed by 3× PBS rinsing. The film was then coated with 0.5% glutaraldehyde in PBS for 15 min, followed by 3× PBS rinsing. The film was coated with 1:200 Matrigel (REF 35623, Corning) in DMEM/F12 (REF 10565-018, Gibco) at 37 °C for 1 h. hiPSC-derived cardiomyocytes were maintained in Matrigel pre-coated 24-well plates in RPMI 1640 medium (REF 11875-093, 500 ml, Gibco) with 1× B-27 supplement (REF 17504-044, 10 ml, Gibco). The culture medium was refreshed every 2–3 days. Five days before the experiment, cells were transferred onto the polymer films. Cardiomyocytes were dissociated with TryPLE select 10× (REF A12177-01, Gibco) at 37 °C for 7 min, then 4 × 104 cells were plated to each coated polymer sample in RPMI 1640 medium with 1× B-27 supplement and 10% KnockOut Serum Replacement (KSR) (REF A31815-01, Gibco). The samples with cells were cultured in a 5% CO2 incubator at 37 °C. The medium was changed to RPMI 1640 medium with 1× B-27 supplement 24 h after cell seeding and refreshed every 2–3 days.

AT1R pharmacology experiments with iPSC-cardiomyocytes

ANGII was dissolved in PBS with 10% DMSO to make 10 mg ml−1 stock solutions and kept at −80 °C. It was diluted in HEPES-buffered Tyrode’s solution to 5 μM on the day of experiment. During ECORE recording, 200 μl of 5 μM ANGII was added to the sample well with 800 μl of HEPES-buffered Tyrode’s solution, yielding a final drug concentration of 1 μM. DMSO (10%) in PBS is diluted and added to samples in the same way as ANGII as a control experiment.

TH17 cell differentiation

Human peripheral blood mononuclear cells were isolated from whole blood collars (Crimson Core, Brigham and Women’s Hospital) by Ficoll gradient centrifugation and subsequent lysis of red blood cells with ACK Lysis Buffer (Thermo Fisher A1049201). MojoSort Human CD4 Naïve T Cell Isolation Kits (BioLegend 480041) were then used to isolate naive CD4+ T cells from peripheral blood mononuclear cells following the manufacturer’s protocol. Isolated cells were analysed for purity by flow cytometry using Alexa Fluor 647 anti-CD45 (BioLegend 304056, 1:50), Alexa Fluor 488 anti-CD4 (BioLegend 317408, 1:50). Primary human naive CD4+ T cells were cultured in X-VIVO 15 medium (Lonza Bioscience, 04-418Q) supplemented with 5% FBS, 4 μM N-acetyl-l-cysteine, 0.05 mM 2-mercaptoethanol, and CD3/CD28 activation beads (BioLegend 422604) at a 1:1 bead:cell ratio. For differentiation to TH17 cells, 25 ng ml−1 IL-6 (BRB Preclinical Biologics Repository L0406011), 10 ng ml−1 IL-1β (GoldBio 1110-01B-10), 15 ng ml−1 IL-23 (Thermo Fisher 200-23-10UG), 5 ng ml−1 TGF-β1 (Thermo Fisher 100-21-10UG), 10 µg ml−1 anti-IFN-γ (BioLegend 506502) and 10 µg ml−1 anti-IL-4 (BioLegend 500802) were added to culture for 5 days. TH17 cells were analysed for purity by flow cytometry using PE/Cyanine7 anti-CXCR3 (BioLegend 353720, 1:50), Alexa Fluor 647 anti-CCR6 (BioLegend 353415, 1:50) and Alexa Fluor 488 anti-CD4 (BioLegend 317408, 1:50).

Biolayer interferometry

Biolayer interferometry measurements were performed through the Harvard Medical School Center for Macromolecular Interactions Core using an Octet RED384 (ForteBio) instrument with Octet BLI Discovery, v.13.0 software. A total of 4 µg ml−1 biotinylated human TfR1 (Sino Biological, 11020-H07H-B-100) or mouse TfR1 (ACROBiosystems, TFRM824925UG) extracellular domains were immobilized on optically transparent streptavidin biosensors (ForteBio) and loaded for roughly 120 s until a 1–2 nm signal was achieved. After washing with buffer, monomeric TfR1 VHH variant binders fused to an Fc region were used as the analyte for association and dissociation measurements. Tris-buffered saline with 0.1% Tween-20 detergent was used as the buffer. Data were analysed using the ForteBio Octet Analysis Studio v.13.0 software, and kinetic parameters were determined using a 1:1 monovalent binding model.

Intracellular iron measurements

MDA-MB-231 cells were seeded in a 96-well flat-bottom plate. The next day, they were treated with 10 nM of TfR1-targeting constructs or 100 µM of ammonium iron (II) sulfate (Sigma, 215406) for 72 h. Cells were washed 3× with PBS, then treated with 1 μmol l−1 FerroOrange working solution (Dojindo Laboratories, F374) and were incubated at 37 °C for 30 min. Fluorescence intensity was then measured by a plate reader.

Animal experiments

All experiments were performed according to the Guide for the Care and Use of Laboratory Animals and were approved by the National Institutes of Health and the Committee on Animal Care at Harvard Medical School (IS-00003882). Male C57BL/6N-Ccr6tm1(CCR6)Bcgen/Bcgen mice were purchased from Biocytogen. All mice were maintained under specific-pathogen-free conditions. All mice were kept in a 12–12 h light–dark cycle. The main room temperature was set at 20–26 °C, and humidity was 40–65%. Mice had ad libitum access to food and water. Sample size was ensured to meet a minimum of n = 3 mice per group. Randomization and blinding were not included for this study.

IMQ-induced mouse psoriasis model and antibody administration

A daily dose of 10 mg of IMQ-containing cream (5%, Taro Pharmaceutical Industries) was applied to the ears of mice aged 10 weeks for 6 consecutive days. Ear thickness was measured daily with a thickness gauge (Mitsutoyo), and the difference from day 0 was expressed as a change in ear swelling. Psoriasis area severity index (PASI) scores were also evaluated daily for each mouse as previously reported54. CCR6 Fab–Fc or TfR1HIGH GTAC was administered daily at a dose of 10 mg kg−1 for 6 consecutive days, and IMQ-containing cream was applied after 5–10 min of antibody administration.

Flow cytometry of mouse immune cells

Ear skin was cut and split into dorsal and abdominal sides, and the dorsal side was incubated for 55 min at 37 °C with 500 μl of digestion buffer (13 ml of complete RPMI 1640, 5 mg Liberase TL (Roche), 3.3 mg DNase I (Sigma) and 10 μg ml−1 brefeldin A (Sigma)). Ear skin was then incubated with 10 mmol EDTA-2Na (Corning) for 5 min at 37 °C. The cell suspensions were filtered using a 40-μm cell strainer (MTC Bio), stained with indicated antibodies, and then subjected to flow cytometric analysis. The gating strategy is shown in Supplementary Fig. 6. Flow cytometry was performed using a Symphony A5 (BD) and analysed with FlowJo (Treestar).

RNA extraction and quantitative PCR analysis

Total RNA was isolated from the ear skin, 24 h after the last IMQ application, using Trizol (Invitrogen) and RNeasy kits (Qiagen). cDNA was synthesized using oligo dT (iScript cDNA synthesis kit, Bio-Rad). Quantitative real-time reverse transcriptase PCR was performed with SYBR Green I (Roche) and an ABI QuantStudio5 (Thermo Fisher Scientific) according to the manufacturer’s instructions. All primers were obtained from Integrated DNA Technologies (Supplementary Table 6). For each sample, reactions were performed in technical triplicate and validated against a control reaction lacking reverse transcriptase. Results were normalized to levels of GAPDH mRNA.

Histology and epidermal measurements

Ear skin samples were fixed with 10% formalin (Sigma) and embedded in paraffin. Sections of 5 μm thickness were prepared and stained with haematoxylin and eosin. Images of stained sections were obtained with an Olympus VS200 Widefield Slide Scanner. Epidermal thickness was calculated as a measure of the total area of the epidermis divided by the total length of the epidermis using OlyVIA 3.8 (Olympus).

Plasma half-life measurements

Male C57BL/6NJ mice (Taconic Biosciences) aged 6 weeks were injected with either 10 mg kg−1 (body weight) CCR6 GTAC or 10 mg kg−1 (body weight) CCR6-Fab-Fc control (n = 3 for each group). Blood samples were collected from the facial vein using ethylenediaminetetraacetic acid capillary tubes on days 1, 3, 7 and 10. Plasma was obtained by centrifuging the blood at 2,000g at 4 °C for 15 min. Then, 2 µl of plasma was diluted in 10 µl of PBS and subjected to standard western blot analysis as previously mentioned with an anti-human IgG antibody (LICOR 926-32232, 1:5,000). With the same samples, a stain-free gel was also run and imaged to determine total protein levels.

Tissue preparation and cryostat sectioning

Organs were removed and fixed by immersion in 4% paraformaldehyde (in PBS). Tissue samples were then transferred into sucrose gradient solutions and incubated overnight at each concentration (15%, 30%: w/v = g ml−1). Samples were then sealed in OCT compound (Optimal Cutting Temperature compound, Scigen) and frozen quickly at −80 °C. Frozen 30 μm sections were made on the cryostat (Leica CM1950), and the slices were stored free-floating in an ethylene-glycol-based cryoprotectant solution at −20 °C until used for downstream immunofluorescence experiments.

Immunofluorescence staining

Tissue slices were washed briefly in PBS-X (pH 7.4, supplemented with 0.05% Triton X) and blocked for 1 h with 5% normal donkey serum (AB7475, Abcam) diluted in PBS-X at room temperature. Samples were then incubated overnight at 4 °C with primary TfR1 antibody (13-6800, Invitrogen, 2.5 μg ml−1) diluted to 2.5 μg ml−1 in blocking buffer. After washing three times in PBS-X, samples were incubated with a secondary antibody cocktail containing PBS-X, 1% normal donkey serum, 1,000× diluted donkey anti-mouse 647 (A32787, Invitrogen) and 5,000× diluted DAPI stain (AB228549, Abcam) for 2 h at room temperature. Stained slices were washed three times in PBS-X and once in PBS for 5 min, then mounted onto slides and cover-slipped for imaging on the Zeiss AxioScan7.

Data and statistical analysis

All graphing and statistical analyses were performed in GraphPad Prism (v.10.4.2) or Microsoft Excel (v.16.98). Unpaired two-tailed t-tests were used for statistical analyses unless otherwise noted. A two-tailed P value was used to determine statistical significance for all analyses. The following convention was used to assign significance: not significant (NS), P ≥ 0.05; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. Sample sizes followed the common standard of biological replicates of n = 3 or more, with n = 2 in a few cases, and were not determined using statistical methods. All schematics were created with BioRender.com, Affinity Designer (v.2.6.3), and Adobe Illustrator (v.30.x). Raw flow cytometry plots were created with FlowJo (v.10.8.0). All figures were assembled in Affinity Designer (v.2.6.3).

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

The animal datasets are provided as source files. Uncropped western blot images and biological replicates are shown in Supplementary Figs. 1 and 9, flow cytometry gating strategies in Supplementary Fig. 6, and raw immunofluorescence images in Supplementary Fig. 10. Sample sizes and Pearson correlation coefficients for image analyses are listed in Supplementary Tables 4 and 5, qPCR primer sequences in Supplementary Table 6, and statistical P values in Supplementary Table 7. All other datasets are available on Zenodo68 (https://doi.org/10.5281/zenodo.21964503). Source data are provided with this paper.

References

  1. Hauser, A. S., Attwood, M. M., Rask-Andersen, M., Schiöth, H. B. & Gloriam, D. E. Trends in GPCR drug discovery: new agents, targets and indications. Nat. Rev. Drug Discov. 16, 829–842 (2017).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  2. Gómez-Melero, S. & Caballero-Villarraso, J. CCR6 as a potential target for therapeutic antibodies for the treatment of inflammatory diseases. Antibodies 12, 30 (2023).

    Article  PubMed  PubMed Central  Google Scholar 

  3. Paulsen, S. J., Rosenkilde, M. M., Eugen-Olsen, J. & Kledal, T. N. Epstein-Barr virus-encoded BILF1 is a constitutively active G protein-coupled receptor. J. Virol. 79, 536–546 (2005).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  4. Burston, H. E. et al. Inhibition of relaxin autocrine signaling confers therapeutic vulnerability in ovarian cancer. J. Clin. Invest. 131, e142677 (2021).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  5. Lorente, J. S. et al. GPCR drug discovery: new agents, targets and indications. Nat. Rev. Drug Discov. 24, 458–479 (2025).

    Article  PubMed  CAS  Google Scholar 

  6. Manglik, A. & Kruse, A. C. Structural basis for G protein-coupled receptor activation. Biochemistry 56, 5628–5634 (2017).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  7. Downey, M. L. & Peralta-Yahya, P. Technologies for the discovery of G protein-coupled receptor-targeting biologics. Curr. Opin. Biotechnol. 87, 103138 (2024).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  8. Eichel, K. & von Zastrow, M. Subcellular organization of GPCR signaling. Trends Pharmacol. Sci. 39, 200–208 (2018).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  9. Li, K. & Crews, C. M. PROTACs: past, present and future. Chem. Soc. Rev. 51, 5214–5236 (2022).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  10. Zhang, D. et al. Transferrin receptor targeting chimeras for membrane protein degradation. Nature 638, 787–795 https://doi.org/10.1038/s41586-024-07947-3 (2024).

    Article  ADS  PubMed  PubMed Central  CAS  Google Scholar 

  11. Huang, B. et al. Designed endocytosis-inducing proteins degrade targets and amplify signals. Nature 638, 796–804 (2025).

    Article  ADS  PubMed  CAS  Google Scholar 

  12. Banik, S. M. et al. Lysosome-targeting chimaeras for degradation of extracellular proteins. Nature 584, 291–297 (2020).

    Article  ADS  PubMed  PubMed Central  CAS  Google Scholar 

  13. Cotton, A. D., Nguyen, D. P., Gramespacher, J. A., Seiple, I. B. & Wells, J. A. Development of antibody-based PROTACs for the degradation of the cell-surface immune checkpoint protein PD-L1. J. Am. Chem. Soc. 143, 593–598 (2021).

    Article  ADS  PubMed  PubMed Central  CAS  Google Scholar 

  14. Pance, K. et al. Modular cytokine receptor-targeting chimeras for targeted degradation of cell surface and extracellular proteins. Nat. Biotechnol. 41, 273–281 (2023).

    Article  PubMed  CAS  Google Scholar 

  15. Zhao, F. et al. Hijacking extracellular targeted protein degrader-drug conjugates for enhanced drug delivery. J. Am. Chem. Soc. 147, 39912–39925 (2025).

    Article  ADS  PubMed  PubMed Central  CAS  Google Scholar 

  16. Marei, H. et al. Antibody targeting of E3 ubiquitin ligases for receptor degradation. Nature 610, 182–189 (2022).

    Article  ADS  PubMed  PubMed Central  CAS  Google Scholar 

  17. Rhee, K. & Zhou, X. Two in one: the emerging concept of bifunctional antibodies. Curr. Opin. Biotechnol. 85, 103050 (2024).

    Article  PubMed  CAS  Google Scholar 

  18. Wells, J. A. & Kumru, K. Extracellular targeted protein degradation: an emerging modality for drug discovery. Nat. Rev. Drug Discov. 23, 126–140 (2024).

    Article  PubMed  Google Scholar 

  19. Ahn, G. et al. Elucidating the cellular determinants of targeted membrane protein degradation by lysosome-targeting chimeras. Science 382, eadf6249 (2023).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  20. Harding, C., Heuser, J. & Stahl, P. Receptor-mediated endocytosis of transferrin and recycling of the transferrin receptor in rat reticulocytes. J. Cell Biol. 97, 329–339 (1983).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  21. Kawai, T. et al. AT1 receptor signaling pathways in the cardiovascular system. Pharmacol. Res. 125, 4–13 (2017).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  22. Chen, T.-Y. et al. The relaxin family peptide receptor 1 (RXFP1): an emerging player in human health and disease. Mol. Genet. Genomic Med. 8, e1194 (2020).

    Article  PubMed  PubMed Central  Google Scholar 

  23. Meitei, H. T., Jadhav, N. & Lal, G. CCR6-CCL20 axis as a therapeutic target for autoimmune diseases. Autoimmun. Rev. 20, 102846 (2021).

    Article  PubMed  CAS  Google Scholar 

  24. Kuba, K., Sato, T., Imai, Y. & Yamaguchi, T. Apelin and Elabela/Toddler; double ligands for APJ/Apelin receptor in heart development, physiology, and pathology. Peptides 111, 62–70 (2019).

    Article  PubMed  CAS  Google Scholar 

  25. Vijayan, D., Young, A., Teng, M. W. L. & Smyth, M. J. Targeting immunosuppressive adenosine in cancer. Nat. Rev. Cancer 17, 709–724 (2017).

    Article  PubMed  CAS  Google Scholar 

  26. Alagha, K. et al. Long-acting muscarinic receptor antagonists for the treatment of chronic airway diseases. Ther. Adv. Chronic Dis. 5, 85–98 (2014).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  27. Lobingier, B. T. et al. An approach to spatiotemporally resolve protein interaction networks in living cells. Cell 169, 350–360.e12 (2017).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  28. Wang, Q. et al. Design and production of bispecific antibodies. Antibodies 8, 43 (2019).

    Article  PubMed  PubMed Central  Google Scholar 

  29. Neiveyans, M. et al. A recycling anti-transferrin receptor-1 monoclonal antibody as an efficient therapy for erythroleukemia through target up-regulation and antibody-dependent cytotoxic effector functions. MAbs 11, 593–605 (2019).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  30. David, M., Cohen, R. et al. Novel single-domain antibodies targeting a unique transferrin receptor 1 epitope for cross-species delivery of drugs in the central nervous system. J. Control. Release 388, 114344 (2025).

  31. Skiba, M. A. et al. Antibodies expand the scope of angiotensin receptor pharmacology. Nat. Chem. Biol. 20, 1577–1585 https://doi.org/10.1038/s41589-024-01620-6 (2024).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  32. Ma, Y. et al. Structure-guided discovery of a single-domain antibody agonist against human apelin receptor. Sci. Adv. 6, eaax7379 (2020).

    Article  ADS  PubMed  PubMed Central  CAS  Google Scholar 

  33. Liu, S., Zhao, X. & Changyun, H. U. Anti-adenosine receptor (a2aR) antibodies. World Patent WO2022072601A1 (2023).

  34. Paul, S. et al. Machine learning enables efficient and effective affinity maturation of nanobodies. Preprint at bioRxiv https://doi.org/10.64898/2026.01.11.698911 (2026).

  35. Zhao, X. & Changyun, H. U. Anti-adenosine receptor (a2ar) antibodies and the use thereof. World Patent WO2023192489A2 (2023).

  36. Halls, M. L. & Cooper, D. M. F. Sub-picomolar relaxin signalling by a pre-assembled RXFP1, AKAP79, AC2, β-arrestin 2, PDE4D3 complex. EMBO J. 29, 2772–2787 (2010).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  37. Hossain, M. A. et al. Development of novel high-affinity antagonists for the relaxin family peptide receptor 1. ACS Pharmacol. Transl. Sci. 6, 842–853 (2023).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  38. Basuli, D. et al. Iron addiction: a novel therapeutic target in ovarian cancer. Oncogene 36, 4089–4099 (2017).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  39. Beugelink, J. W., Sweep, E., Janssen, B. J. C., Snijder, J. & Pronker, M. F. Structural basis for recognition of the FLAG-tag by anti-FLAG M2. J. Mol. Biol. 436, 168649 (2024).

    Article  PubMed  CAS  Google Scholar 

  40. Zuo, J. et al. The Epstein-Barr virus-encoded BILF1 protein modulates immune recognition of endogenously processed antigen by targeting major histocompatibility complex class I molecules trafficking on both the exocytic and endocytic pathways. J. Virol. 85, 1604–1614 (2011).

    Article  PubMed  CAS  Google Scholar 

  41. Robert, R. et al. Essential role for CCR6 in certain inflammatory diseases demonstrated using specific antagonist and knockin mice. JCI Insight 2, e94821 (2017).

    Article  PubMed  PubMed Central  Google Scholar 

  42. Li, W. et al. A novel C-C chemoattractant cytokine (chemokine) receptor 6 (CCR6) antagonist (PF-07054894) distinguishes between homologous chemokine receptors, increases basal circulating CCR6+ T cells, and ameliorates interleukin-23-induced skin inflammation. J. Pharmacol. Exp. Ther. 386, 80–92 (2023).

    Article  PubMed  CAS  Google Scholar 

  43. Schall, T. J. & Proudfoot, A. E. I. Overcoming hurdles in developing successful drugs targeting chemokine receptors. Nat. Rev. Immunol. 11, 355–363 (2011).

    Article  PubMed  CAS  Google Scholar 

  44. Lissilaa, R. et al. Antibodies that bind to ccr6 and their uses. World Patent WO2016059253A1 (2016).

  45. Kern, A. & Bryant-Greenwood, G. D. Characterization of relaxin receptor (RXFP1) desensitization and internalization in primary human decidual cells and RXFP1-transfected HEK293 cells. Endocrinology 150, 2419–2428 (2009).

    Article  PubMed  CAS  Google Scholar 

  46. Callander, G. E., Thomas, W. G. & Bathgate, R. A. D. Prolonged RXFP1 and RXFP2 signaling can be explained by poor internalization and a lack of β-arrestin recruitment. Am. J. Physiol. Cell Physiol. 296, C1058–C1066 (2009).

    Article  PubMed  CAS  Google Scholar 

  47. Cao, Y. et al. Angiotensin II type 1 receptor variants alter endosomal receptor-β-arrestin complex stability and MAPK activation. J. Biol. Chem. 295, 13169–13180 (2020).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  48. Thomas, W. G., Motel, T. J., Kule, C. E., Karoor, V. & Baker, K. M. Phosphorylation of the angiotensin II (AT1A) receptor carboxyl terminus: a role in receptor endocytosis. Mol. Endocrinol. 12, 1513–1524 (1998).

    Article  PubMed  CAS  Google Scholar 

  49. Kantor, E. D., Rehm, C. D., Haas, J. S., Chan, A. T. & Giovannucci, E. L. Trends in prescription drug use among adults in the United States from 1999-2012. JAMA 314, 1818–1830 (2015).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  50. Bullo, M., Tschumi, S., Bucher, B. S., Bianchetti, M. G. & Simonetti, G. D. Pregnancy outcome following exposure to angiotensin-converting enzyme inhibitors or angiotensin receptor antagonists: a systematic review. Hypertension 60, 444–450 (2012).

    Article  PubMed  CAS  Google Scholar 

  51. Mehta, P. K. & Griendling, K. K. Angiotensin II cell signaling: physiological and pathological effects in the cardiovascular system. Am. J. Physiol. Cell Physiol. 292, C82–C97 (2007).

    Article  PubMed  CAS  Google Scholar 

  52. Alfonso, F. S. et al. Label-free optical detection of bioelectric potentials using electrochromic thin films. Proc. Natl Acad. Sci. USA 117, 17260–17268 (2020).

    Article  ADS  PubMed  PubMed Central  CAS  Google Scholar 

  53. Jurcevic, S. et al. The effect of a selective CXCR2 antagonist (AZD5069) on human blood neutrophil count and innate immune functions. Br. J. Clin. Pharmacol. 80, 1324–1336 (2015).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  54. van der Fits, L. et al. Imiquimod-induced psoriasis-like skin inflammation in mice is mediated via the IL-23/IL-17 axis. J. Immunol. 182, 5836–5845 (2009).

    Article  PubMed  Google Scholar 

  55. Schön, M. P., Broekaert, S. M. C. & Erpenbeck, L. Sexy again: the renaissance of neutrophils in psoriasis. Exp. Dermatol. 26, 305–311 (2017).

    Article  PubMed  Google Scholar 

  56. Singh, T. P. et al. Monocyte-derived inflammatory Langerhans cells and dermal dendritic cells mediate psoriasis-like inflammation. Nat. Commun. 7, 13581 (2016).

    Article  ADS  PubMed  PubMed Central  CAS  Google Scholar 

  57. Ahn, G., Banik, S. M. & Bertozzi, C. R. Degradation from the outside in: targeting extracellular and membrane proteins for degradation through the endolysosomal pathway. Cell Chem. Biol. 28, 1072–1080 (2021).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  58. Kougentakis, C. et al. Catalytic degradation of circulating targets with FcRn-mediated cycling LYTACs. Preprint at bioRxiv https://doi.org/10.1101/2025.01.12.632472 (2025).

  59. Crook, Z. R. et al. CYpHER: catalytic extracellular targeted protein degradation with high potency and durable effect. Nat. Commun. 15, 8731 (2024).

    Article  ADS  PubMed  PubMed Central  CAS  Google Scholar 

  60. Kroeze, W. K. et al. PRESTO-Tango as an open-source resource for interrogation of the druggable human GPCRome. Nat. Struct. Mol. Biol. 22, 362–369 (2015).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  61. Edwards, J. M. et al. Enhancement of antibody-dependent cellular cytotoxicity and phagocytosis in anti-HIV-1 human-bovine chimeric broadly neutralizing antibodies. J. Virol. 95, e0021921 (2021).

    Article  PubMed  PubMed Central  Google Scholar 

  62. Stapleton, N. M. et al. Human IgG lacking effector functions demonstrate lower FcRn-binding and reduced transplacental transport. Mol. Immunol. 95, 1–9 (2018).

    Article  PubMed  CAS  Google Scholar 

  63. Hamblett, K. J. et al. Altering antibody-drug conjugate binding to the neonatal Fc receptor impacts efficacy and tolerability. Mol. Pharm. 13, 2387–2396 (2016).

    Article  PubMed  CAS  Google Scholar 

  64. Jacobsen, F. W. et al. Engineering an IgG scaffold lacking effector function with optimized developability. J. Biol. Chem. 292, 1865–1875 (2017).

    Article  PubMed  CAS  Google Scholar 

  65. Erlandson, S. C. et al. The relaxin receptor RXFP1 signals through a mechanism of autoinhibition. Nat. Chem. Biol. 19, 1013–1021 (2023).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  66. Erlandson, S. C. et al. Engineering and characterization of a long-half-life relaxin receptor RXFP1 agonist. Mol. Pharm. 21, 4441–4449 (2024).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  67. Zhou, Y. et al. Ultrasensitive label-free optical recording of bioelectric potentials using dioxythiophene-based electrochromic polymers. Nat. Commun. 16, 6776 (2025).

    Article  ADS  PubMed  PubMed Central  CAS  Google Scholar 

  68. Rhee. K. & Shue, L. GTAC. Zenodo https://doi.org/10.5281/zenodo.21964502 (2026).

Download references

Acknowledgements

We thank the members of the Zhou lab for valuable input throughout this project. We especially thank J. Duque-Jimenez and D. Zhang for assistance with initial cloning and antibody production; I. Algov for work on FLAG constructs; and Z. Ma for generating luciferase-expressing CCR6+ Jurkat cells. We also thank J. Smith for insights into chemokine receptor biology and GPCR assays; V. Calvillo-Miranda for discussions of M3R biology and assays; D. McGlamery and other members of the Adelman lab for qPCR support; A. Jyothidasan and H. Rockman for discussions of AT1R signalling; J. La Rose and R. Rottapel for discussions of RXFP1 in ovarian cancer and for providing RXFP1 qPCR primers and OVCAR-8 cells; F. Facchinetti of the Pasi Jänne lab for TfR1 knockdown PC9 cells; and K. Huang for guidance on confocal microscopy.

Funding

K.R. acknowledges support from the Chemical Biology NIH T32 Training grant (T32GM139775), the NIH Training in Pharmacological Sciences grant (2T32GM132089-06), the 2024 Landry Cancer Biology Research Fellowship, and the 2026 Cancer Biology Fellowship from the Jean Strouse Sharf and Lisa Sharf Green Cancer Research Fund. L.S. acknowledges support from the NIH Training in Pharmacological Sciences grant (2T32GM132089-06). A.A. acknowledges support from fellowships awarded by the Daiichi Sankyo Foundation of Life Science, the Japanese Dermatological Association and the Mochida Memorial Foundation for Medical and Pharmaceutical Research. P.S., Q.L. and B.C. acknowledge support from NIH grants (R01NS121934 and R01CA311218-01). J.O.-O. acknowledges support from the HHMI Hanna H. Gray Fellows Program. K.A. acknowledges support from NIH grant (R01GM134539). M.A.S. acknowledges support from NIH grant (K99HD110612). A.C.K. acknowledges support from NIH grants (R01CA260415, R01AR079489 and R01AR081264), and the Smith Family Foundation. X.Z. acknowledges support from the NIH (DP2GM154013 and R00EB030587), the V Foundation for Cancer Research (T2025-003), the Helen Gurley Brown Foundation and the DFCI Accelerator Program. X.Z. is a Damon Runyon-Rachleff Innovator supported (in part) by the Damon Runyon Cancer Research Foundation (DRR-91-26).

Author information

Author notes

  1. Meredith A. Skiba

    Present address: Department of Biological Chemistry, University of Michigan School of Medicine, Ann Arbor, MI, USA

  2. These authors contributed equally: Kaitlin Rhee, Lawrence Shue

Authors and Affiliations

  1. Department of Cancer Biology, Dana-Farber Cancer Institute, Boston, MA, USA

    Kaitlin Rhee, Lawrence Shue, Dingjingyu Zhou, Aoxing Cheng & Xin Zhou

  2. Department of Biological Chemistry and Molecular Pharmacology, Blavatnik Institute, Harvard Medical School, Boston, MA, USA

    Kaitlin Rhee, Lawrence Shue, James Osei-Owusu, Aoxing Cheng, Apoorva Baluapuri, Edward P. Harvey, Karen Adelman, Meredith A. Skiba, Andrew C. Kruse & Xin Zhou

  3. Department of Immunology, Blavatnik Institute, Harvard Medical School, Boston, MA, USA

    Akimasa Adachi & Jun R. Huh

  4. Wu-Tsai Neuroscience Institute, Stanford University, Stanford, CA, USA

    Pengwei Sun, Qingyue Li & Bianxiao Cui

  5. Department of Radiology, Stanford University, Stanford, CA, USA

    Pengwei Sun

  6. Department of Neurology, Ann Romney Center for Neurologic Diseases, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA, USA

    Yi Ran Xu

  7. Department of Chemistry, Stanford University, Stanford, CA, USA

    Qingyue Li & Bianxiao Cui

  8. The Eli and Edythe L. Broad Institute, Cambridge, MA, USA

    Apoorva Baluapuri & Karen Adelman

  9. Department of Pharmacology, University of Michigan School of Medicine, Ann Arbor, MI, USA

    Meredith A. Skiba

Authors

  1. Kaitlin Rhee
  2. Lawrence Shue
  3. Akimasa Adachi
  4. Pengwei Sun
  5. James Osei-Owusu
  6. Dingjingyu Zhou
  7. Aoxing Cheng
  8. Yi Ran Xu
  9. Qingyue Li
  10. Apoorva Baluapuri
  11. Edward P. Harvey
  12. Karen Adelman
  13. Meredith A. Skiba
  14. Bianxiao Cui
  15. Jun R. Huh
  16. Andrew C. Kruse
  17. Xin Zhou

Contributions

K.R. and L.S. performed all experiments and data analyses unless otherwise stated. K.R. was responsible for work related to AT1R, RXFP1, APJ, A2AR, BILF1, M3R and FLAG; L.S. was responsible for work on affinity-tuned context-dependent GTAC design, CCR6, BILF1 and FLAG. K.R. and L.S. contributed equally to this work, and the order of names is determined alphabetically. A.A. performed the CCR6 in vivo work and data analyses with guidance from J.R.H.; P.S. and Q.L. performed the iPSC ECORE study with guidance from B.C.; J.O-O. and E.P.H. conducted RXFP1 Nb discovery and affinity maturation with guidance from A.C.K.; D.Z. assisted with antibody production and assays, characterized CCR6 GTAC on mouse cells, and jointly characterized TfR1 affinity-tuned antibodies. A.C. cloned the TfR1 binder mutants and jointly characterized TfR1 affinity-tuned antibodies. Y.R.X. and D.Z. performed TfR1 tissue immunofluorescence assays. K.R. and A.B. performed RXFP1 GTAC qPCR assays and data analyses with input from K.A.; M.A.S. provided AT1R binder information and input on signalling assays. K.R., L.S., A.A., J.R.H., A.C.K. and X.Z. co-wrote the manuscript with input from all authors. A.C.K. provided scientific input on GPCR biology. X.Z. supervised the study.

Corresponding authors

Correspondence to Andrew C. Kruse or Xin Zhou.

Ethics declarations

Competing interests

X.Z. is a co-founder, equity holder and advisor of atida bio; a scientific founder, equity holder and advisor of VincenTx; and has received research funding from Merck unrelated to this work. A.C.K. is a co-founder and consultant for Tectonic Therapeutic and Seismic Therapeutics, and for the Institute for Protein Innovation, a nonprofit research institute. K.A. is a consultant to Odyssey Therapeutics, serves on the Scientific Advisory Board of CAMP4 Therapeutics and has received research funding from Novartis unrelated to this work. K.R., L.S., D.Z., J.O-O., A.C., A.C.K. and X.Z. have filed patent applications covering the GTAC technologies (PCT/US2025/020274, PCT/US2025/020271, PCT/US2025/060865, PCT/US26/36667 and two unpublished provisional applications).

Peer review

Peer review information

Nature thanks Ross Bathgate, Michelle Halls, Roshanak Irannejad and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Extended data figures and tables

Extended Data Fig. 1 Characterization of GTAC binding, receptor degradation, kinetics, and specificity.

(a) Schematics of AT1R GTAC, AT1R Nb-Fc, FLAG GTAC, and FLAG Fab-Fc constructs (left). AT1R GTACs and Nb-Fc binding to FLAG-AT1R+ TfR1+ or FLAG-AT1R+ Expi293 cells (right, Protein A flow cytometry). Data represent mean ± standard deviation from n = 3 replicates. (b) Cell surface AT1R (left, anti-FLAG), and total AT1R (right, GFP) in FLAG-AT1R-GFP+ TfR1+ Expi293 cells after 1 h or 6 h treatment with AT1R GTAC or Nb-Fc. Data represent mean ± s.d. from n = 3 biological replicates. (c) Cell surface AT1R (left, anti-FLAG) and total AT1R (right, GFP) in FLAG-AT1R-GFP+ TfR1+ Expi293 cells over a 24 h time course after treatment with 10 nM AT1R GTAC or 10 nM Nb-Fc. Data represent mean ± s.d. from n = 3 biological replicates. (d) Cell surface APJ (anti-FLAG) in FLAG-APJ+ TfR1+ HEK293T cells after treatment with 50 nM APJ GTAC or Nb-Fc over 24 h. Data represent mean from n = 2 biological replicates. (e) Cell surface AT1R (left, anti-FLAG) and total AT1R (right, GFP) in FLAG-AT1R-GFP+ TfR1+ Expi293 cells after 6 h treatment with 25 nM AT1R, RXFP1, and CCR6 GTACs and control antibodies. Statistical comparisons via unpaired two-tailed t tests. Data represent mean ± standard deviation from n = 3 biological replicates. (f) Cell surface AT1R (left, anti-FLAG) and total AT1R (right, GFP) in FLAG-AT1R-GFP+ TfR1+ Expi293 cells following treatment with AngII for 1 h, 6 h, or 24 h. Data represent mean ± s.d. from n = 3 biological replicates. (g) Cell surface AT1R (left, anti-FLAG) and total AT1R (right, GFP) in FLAG-AT1R-GFP+ TfR1+ Expi293 cells after 6 h treatment with AT1R GTAC or Nb-Fc, following 1.5 h pretreatment with either 250 ng/mL bafilomycin A1, 25 μg/mL cycloheximide, or no pretreatment, measured by flow cytometry. Data represent mean ± s.d. from n = 3 biological replicates. (h) Cell surface AT1R (left, anti-FLAG) and cell surface TfR1 (middle, anti-v5) in FLAG-AT1R-GFP+ TfR1-v5+ Expi293 cells after treatment with 10 nM AT1R GTAC or Nb-Fc across 24 h. Same data represented in line plot (right). Data represent mean ± s.d. from n = 3 biological replicates. Statistical comparisons via unpaired two-tailed t tests. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Exact p-values are listed in SI Table 7. Diagram in a created in BioRender; Zhou, X. https://biorender.com/hymki1y (2026).

Extended Data Fig. 2 Engineering and characterization of RXFP1-targeting GTACs.

(a) Schematic showing competition between high-affinity RXFP1 Nb and the native ligand RLN2 for RXFP1 binding. (b) On-cell binding of parent RXFP1 Nb (left) and high-affinity RXFP1 Nb (right) to Expi293 cells transfected with FLAG-RXFP1 or empty vector, measured by flow cytometry. Data represent mean ± s.d. from n = 3 biological replicates. (c) Flow cytometry competition assay showing RLN2-Fc binding to Expi293 cells transfected with FLAG-RXFP1 or empty vector, titrated with high-affinity RXFP1 Nb. Data represent mean ± s.d. from n = 3 biological replicates. (d) Cellular cAMP levels in FLAG-RXFP1+ GloSensor+ HEK293T cells measured by Glosensor assay after 20 min pre-treatment with 0 or 1 μM high-affinity RXFP1 Nb, followed by 30 min stimulation with RLN2. Data represent mean ± s.d. from n = 3 biological replicates. (e) Correlation between doxycycline-inducible RXFP1 cell surface expression (anti-FLAG flow cytometry) and RLN2-induced cAMP (Glosensor assay after 30 min induction with 1 nM RLN2) in FLAG-RXFP1+ TfR1+ GloSensor+ HEK293T cells. Data represent mean ± standard deviation from n = 3 biological replicates. (f) Monovalent and bivalent RXFP1 GTAC construct designs. (g) Cell surface RXFP1 (left, anti-FLAG) and total RXFP1 (right, GFP) in FLAG-RXFP1-GFP+ TfR1+ Expi293 cells after 6 h treatment with 100 nM RXFP1 GTAC or Nb-Fc. Statistical comparisons via unpaired two-tailed t tests. Data represent mean ± s.d. from n = 3 biological replicates. (h) cAMP response in FLAG-RXFP1+ TfR1+ GloSensor+ HEK293T cells after 2 h pre-treatment with 10 nM bivalent RXFP1 GTAC, followed by 30 min stimulation with RLN2 (left) or forskolin (right). Data represent mean ± standard deviation from n = 3 biological replicates. (i) Cell surface RXFP1 (left, anti-FLAG) in FLAG-RXFP1+ TfR1+ Expi293 cells after 10 nM treatment with bivalent RXFP1 GTAC across 30 h. cAMP levels (right, Glosensor) in FLAG-RXFP1+ TfR1+ GloSensor+ HEK293T cells following pre-treatment with 10 nM bivalent RXFP1 GTAC or Nb-Fc over 30 h, then stimulation for 30 min with 1 nM RLN2. Data represent mean ± s.d. from n = 3 biological replicates. (j) Schematic showing downstream RXFP1–cAMP signaling pathways implicated in tumor progression, including activation of PKA-dependent transcriptional programs involved in ECM remodeling, cell adhesion, and metastasis. (k) Bivalent RXFP1 GTAC and Nb-Fc binding to OVCAR-8 (Protein A flow cytometry). Data represent mean ± standard deviation from n = 3 biological replicates. (l) Quantitative PCR measurements of RXFP1-regulated genes in OVCAR-8 ovarian cancer cells. Cells were serum-starved, untreated or treated overnight with 100 nM Bivalent GTAC or control, then stimulated with 10 nM RLN2 for 8 h. mRNA levels of MMP9, MMP23B, NOTCH1, NOTCH3, and DVL1 were measured by qPCR and normalized to GAPDH and the value for untreated cells set at 1. β-actin is shown as a control. Statistical comparisons via unpaired two-tailed t tests. Data represent mean ± s.d. from n = 3 replicates. Statistical comparisons via unpaired two-tailed t tests. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Exact p-values are listed in SI Table 7. Diagrams in a, f and j created in BioRender; Zhou, X. https://biorender.com/hymki1y (2026).

Extended Data Fig. 3 Binding properties of RXFP1 GTAC variants.

(a) Bivalent RXFP1 GTACs and Nb-Fcs, with varying affinity for RXFP1, binding to FLAG-RXFP1+ Expi293 or FLAG-RXFP1+ TfR1+ Expi293, measured by flow cytometry (Protein A). Data represent mean ± s.d. from n = 3 biological replicates. (b) Monovalent RXFP1 GTACs and Nb-Fcs, with varying affinity for TfR1, binding to FLAG-RXFP1+ Expi293 or FLAG-RXFP1+ TfR1+ Expi293 (Protein A). Data represent mean ± s.d. from n = 3 biological replicates. Diagrams in a created in Adobe Illustrator (v.30.x) and b created in BioRender; Zhou, X. https://biorender.com/hymki1y (2026).

Extended Data Fig. 4 Characterization of FLAG-targeting GTACs across multiple GPCRs.

(a) Schematic showing FLAG GTAC construct designs. (b) FLAG GTAC and Fab-Fc binding to FLAG-AT1R+ Expi293 or FLAG-AT1R+ TfR1+ Expi293 cells (Protein A flow cytometry). Data represent mean ± s.d. from n = 3 biological replicates. (c) Cell surface AT1R (left, anti-HA) and total AT1R (right, GFP) in FLAG-HA-AT1R-GFP+ TfR1+ Expi293 cells after 1 h treatment with FLAG GTAC or FLAG Fab-Fc. Data represent mean ± s.d. from n = 3 biological replicates. (d) Cell surface BILF1 (left, anti-HA) and NF-κB reporter activity (right, GFP) in FLAG-HA-BILF1+ TfR1+ NF-κB-GFP+ HEK293T cells after 24 h treatment with 10 nM TfR1 scFv-Fc, FLAG Fab-Fc, or FLAG GTAC. Statistical comparisons via unpaired two-tailed t tests. Data represent mean ± s.d. from n = 3 biological replicates. (e) FLAG GTAC and Fab-Fc binding to Raji cells (Protein A) with or without inducing FLAG-BILF1 overexpression. Data represent mean ± s.d. from n = 3 biological replicates. (f) FLAG GTAC and Fab-Fc binding to FLAG-M3R+ TfR1+ Expi293 cells (Protein A). Data represent mean ± s.d. from n = 3 biological replicates. (g) Western blot of FLAG-M3R and β-actin (left) in FLAG-M3R+ TfR1+ Expi293 cells after 8 h treatment with FLAG GTAC or Fab-Fc. Quantification of total M3R expression (right). Statistical comparisons via unpaired two-tailed t tests. Data represent mean ± s.d. from n = 3 biological replicates (SI Fig. 9 shows replicate gel images). (h) β-arrestin recruitment in FLAG-M3R-Tango+ TfR1+ HTLA cells measured by Tango assay after 2 h pre-treatment with FLAG GTAC or Fab-Fc followed by 16 h stimulation with 10 nM iperoxo. Data represent mean ± s.d. from n = 3 biological replicates. (i) Western blot of CCR6 in CCR6+ Jurkat cells after 18 h GTAC treatment (left). Flow cytometry of cell surface FLAG-CCR6 in CCR6+ Jurkat cells after 2 h FLAG Ab/GTAC treatment (middle). Chemotaxis of FLAG-CCR6+ luciferase+ Jurkat cells measured by Transwell migration assay after 2 h FLAG Ab/GTAC treatment & 2 h migration with CCL20 as the chemoattractant (right). Data represent mean ± s.d. from n = 3 biological replicates. Statistical comparisons via unpaired two-tailed t tests. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Exact p-values are listed in SI Table 7. Diagram in a created in BioRender; Zhou, X. https://biorender.com/hymki1y (2026).

Extended Data Fig. 5 FLAG GTACs induce intracellular accumulation of BILF1.

(a) Immunofluorescence microscopy of FLAG-HA-BILF1+ TfR1-v5+ HEK293T cells (anti-HA, anti-v5 staining) after 2 h or 24 h treatment with 25 nM FLAG GTAC or Fab-Fc. Scale bar: 20 µm. (b) Quantification of intracellular BILF1 (left, intracellular fraction of total HA signal per cell). Center lines: medians; box bounds: 25th-75th percentiles; whiskers: minima and maxima. Statistical comparisons via unpaired two-tailed t tests. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Exact p-values are listed in SI Table 7. Data represent mean ± standard deviation from n = 6 cells. Schematic of FLAG- and HA-tagged BILF1 and FLAG targeting constructs (right) utilized for this assay. Diagram in b created in BioRender; Zhou, X. https://biorender.com/hymki1y (2026).

Extended Data Fig. 6 Design principles and optimization of RXFP1-targeting GTACs.

(a) Schematic of native RXFP1 trafficking, showing poor RLN2-induced internalization. (b) Cell surface RXFP1 in FLAG-RXFP1+ TfR1+ Expi293 (left, anti-FLAG flow cytometry) after 6 h incubation with varying doses of RLN2. Data represent mean ± standard deviation from n = 3 biological replicates. Schematic (right) showing the lack of RLN2-induced RXFP1 internalization. (c) RXFP1 GTAC design with parent or high-affinity RXFP1 Nb; monovalent RXFP1 GTAC; and optimized bivalent RXFP1 GTAC. (d) Cell surface RXFP1 (left, anti-FLAG), in FLAG-RXFP1+ TfR1+ Expi293 cells after 2 h treatment with low-affinity parent RXFP1 GTAC or Nb-Fc. cAMP levels (right, Glosensor) in FLAG-RXFP1+ TfR1+ GloSensor+ HEK293T cells after 2 h pre-treatment with GTAC or Nb-Fc, followed by 30 min stimulation with 1 nM RLN2 (right). Data represent mean ± s.d. from n = 3 biological replicates. (e) Cell surface RXFP1 (anti-FLAG) in FLAG-RXFP1+ TfR1+ Expi293 cells after 2 h treatment with high-affinity GTAC or Nb-Fc (left). cAMP levels in FLAG-RXFP1+ TfR1+ GloSensor+ HEK293T cells (right, Glosensor) after 2 h pre-treatment with GTAC or Nb-Fc, followed by 30 min stimulation with 1 nM RLN2. Data represent mean ± s.d. from n = 3 biological replicates. (f) Cell surface RXFP1 (left, anti-FLAG) and cell surface TfR1 (right, anti-v5) in Expi293 cells overexpressing FLAG-RXFP1, TfR1-v5, or both receptors, after 2 h treatment with 10 nM monovalent RXFP1 GTAC or Nb-Fc. Statistical comparisons via unpaired two-tailed t tests. Data represent mean ± s.d. from n = 3 biological replicates. (g) Cell surface RXFP1 (left, anti-FLAG) and cell surface TfR1 (right, anti-v5) in FLAG-RXFP1+ TfR1-v5+ Expi293 cells after 2 h treatment with 0.1 nM bivalent RXFP1 GTAC or Nb-Fc. Statistical comparisons via unpaired two-tailed t tests. Data represent mean ± s.d. from n = 3 biological replicates. (h) Cell surface GPCR (anti-FLAG), effector (anti-TfR1), and relative target/effector expression ratio in FLAG-AT1R+ TfR1+ Expi293 cells, FLAG-RXFP1+ TfR1+ Expi293 cells, and FLAG-CCR6+ Jurkat cells. Data represent mean ± standard deviation from n = 3 biological replicates. Statistical comparisons via unpaired two-tailed t tests. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Exact p-values are listed in SI Table 7. Diagrams in a, c and h created in BioRender; Zhou, X. https://biorender.com/hymki1y (2026).

Extended Data Fig. 7 Characterization of CCR6 GTAC variants.

(a) BLI measurements of intermediate affinity TfR1 Nb mutants TfR1INT1 and TfR1INT2. (b) Schematic of the split-luciferase assay shows cells incubated with GTAC-SmBit and Fc-H7-LgBit, where H7 is a TfR1-binding scFv non-competitive with the TfR1 Nb. (c) Split-luciferase assay reflects GTAC–TfR1 binding across GTAC affinity variants. Statistical comparisons performed via unpaired two-tailed t tests. Data represent mean ± s.d. from n = 3 biological replicates. (d) Cell surface levels of FLAG-CCR6 (left) and TfR1 (right) following 2 h or (e) 24 h treatment of CCR6+ Jurkat cells with GTAC variants via flow cytometry. Data represent mean ± s.d. from n = 3 biological replicates. (f) Transwell chemotaxis assay of CCR6+ Jurkat cells following 2 h GTAC pretreatment and 2 h migration. Data represent mean ± s.d. from n = 3 biological replicates. (g) Quantification of CCR6-GFP in Rab5-mCherry+, CCR6-GFP+, TfR1-BFP+ HeLa cells by confocal microscopy with 1 h, 4 h, or 24 h treatment with GTAC variants or CCL20. Data represent mean ± s.d. n of cells for analyses are summarized in SI Table 4. (h) Cell surface FLAG-CCR6 and TfR1 levels in CCR6+ PC9 cells with doxycycline-inducible TfR1 knockdown via flow cytometry. Cells were treated with doxycycline, followed by 24 h GTAC treatment. Data are presented as mean from n = 2 biological replicates. (i) Quantification of total CCR6-GFP via flow cytometry in CCR6-GFP+ PC9 cells following 18 h treatment of GTAC variants with or without 250 nM bafilomycin. Data represent mean ± s.d. from n = 3 biological replicates. (j) Flow cytometry of 10 nM CCR6 Fab-Fc cell surface binding to Lenti-X cells expressing FLAG-CCR6, FLAG-CCR7, or FLAG-CXCR2 after 10 min incubation on ice. Binding is reported as the ratio of antibody detected (Protein A) to receptor expression (FLAG). Statistical comparisons performed via unpaired two-tailed t tests. Data represent mean ± s.d. from n = 3 biological replicates. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Exact p-values are listed in SI Table 7. Diagram in b created in BioRender; Zhou, X. https://biorender.com/hymki1y (2026).

Extended Data Fig. 8 GTACs drive sustained receptor depletion through mechanisms distinct from native ligand signaling.

(a) Multiplexed four-color live-cell confocal microscopy of HeLa cells expressing Rab11-mCherry, CCR6-GFP, and TfR1-BFP with 24 h treatment of 10 nM GTAC conjugated to Alexa Fluor 647 or 50 nM CCL20. (b) Pearson colocalization analysis of CCR6-GFP & Rab7-mCherry (late endosomes), CCR6-GFP and Rab11-mCherry (recycling endosomes), or (c) TfR1-BFP and Rab11-mCherry at different times with GTAC treatment. Data are presented as mean ± s.d. n of cells for analyses are summarized in SI Table 4. Statistical comparisons performed via unpaired two-tailed t-tests. (d) Live-cell confocal microscopy of FLAG-AT1R-GFP+ TfR1+ Expi293 cells after 30 min or 6 h treatment with 1 µM AngII, 25 nM AT1R GTAC, or 25 nM FLAG GTAC. Scale bar: 20 µm. (e) Quantification of intracellular AT1R, measured as the ratio of intracellular to total GFP fluorescence per cell. Statistical comparisons performed via unpaired two-tailed t tests. Data represent mean ± s.d. from n = 4 cells. (f) Confocal microscopy images showing RXFP1 localization in FLAG-RXFP1-GFP+ TfR1+ Hela cells after 6 h treatment with 50 nM bivalent RXFP1 GTAC, Nb-Fc, or untreated controls. Arrows indicate intracellular puncta. Scale bar: 20 μm. Images represent results of 2 independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Exact p-values are listed in SI Table 7. Diagram in e created in BioRender; Zhou, X. https://biorender.com/hymki1y (2026).

Extended Data Fig. 9 Intracellular trafficking of GTAC-mediated degradation of CCR6.

Multiplexed four-color live-cell imaging in FLAG-CCR6-GFP+, TfR1-BFP+ HeLa cells with 4 h treatment with 10 nM GTAC conjugated to Alexa Fluor 647 or CCL20. Colocalization is shown with mCherry-tagged endolysosomal markers including Rab5 (a), Rab7 (b), Rab11 (c), or Lamp1 (d), to indicate the early endosomes, late endosomes, recycling endosomes, or lysosomes, respectively. The TfR1HIGH GTAC is compared to the TfR1LOW GTAC. n of cells for each condition are summarized in SI Table 4.

Extended Data Fig. 10 Imaging of GTAC-mediated CCR6 trafficking with endogenous endolysosomal markers.

Three-color imaging of FLAG-CCR6-GFP+, TfR1-BFP+ HeLa cells with or without 4 h treatment with 10 nM of low-affinity GTAC. Cells were fixed, permeabilized, and stained with primary antibodies targeting Rab5 (a), Rab7 (b), Rab11 (c), or Lamp1 (d), in addition to secondary antibodies conjugated with Alexa Fluor 647. (e) Comparison of pearson colocalization coefficients of CCR6-GFP & Rab5, Rab7, Rab11, or Lamp1 in fixed FLAG-CCR6-GFP+, TfR1-BFP+ HeLa cells with endogenous endolysosomal markers with the same coefficients in live cells with overexpressed markers fused to mCherry, with or without 4 h GTAC treatment. n of cells for each analysis is summarized in SI Table 4.

Extended Data Fig. 11 AT1R GTAC suppresses AngII responses in hiPSC-cardiomyocytes.

(a) Angiotensin II (AngII) increases beat-to-beat variation of the contraction waveform (single representative spot). Phase-warped, amplitude-normalized mechanical contraction waveforms for the 10 beats immediately before (blue) and 10 after (red) injection of Ang II. Thin lines: individual beats; bold line: across-beat mean; shaded band: ± 1 s.d. The shape variation (mean across-beat SD over the cycle) increases from 0.041 (Pre) to 0.073 (Post), +80%. (b) Mechanical beat shape variation before vs. after 1 µM AngII or DMSO injection. Each beat was phase-warped and amplitude-normalized; variation is the across-beat SD of the normalized waveform (lower = more reproducible shape). Values are normalized by each recording’s median pre-injection variation. Beating interval (c) and normalized mechanical beat-shape variation (d) in response to 1 µM Ang II injection across four drug treatments: an untreated cardiomyocyte culture (same data as “Untreated” in Fig. 5d and Extended Data Fig. 11b) and three pre-treated cultures: VHHA-based AT1R GTAC (same data as “70 nM GTAC” in Fig. 5d and Extended Data Fig. 11b), AT1R Nb-Fc, and TfR1 Nb-Fc. (e) Per-beat contraction amplitude (from onset to peak of the band-passed mechanical signal, a.u.) over the 10 beats immediately before vs. after injection of 1 µM Ang II or DMSO vehicle, on untreated hiPSC-cardiomyocytes. (f) Per-beat time-to-peak (TTP, s) over the 10 beats immediately before vs. after injection, on untreated cells. (g) Per-beat electromechanical delay (EMD) before vs. after injection. In contrast to beat interval and mechanical waveform variation, AngII did not significantly alter contraction amplitude, time-to-peak, or electromechanical delay under the conditions tested. Bar graphs: medians; error bars: interquartile ranges. Statistical comparisons performed via unpaired two-tailed t-tests. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Exact p-values are listed in SI Table 7.

Extended Data Fig. 12 Characterizing CCR6 GTAC binder cross-reactivity and performance in mice.

(a) Schematic of the CCR6-targeting GTAC with a competitive TfR1 scFv binder (H7). H7 scFv-Fc was included as a control. (b) Cell surface FLAG-CCR6 and TfR1 expression in CCR6+ Jurkat cells after 2 h treatment with the H7 GTAC or H7 scFv-Fc, measured via flow cytometry. Data represents mean ± s.d. from n = 3 biological replicates. (c) Western blot analysis of total CCR6 and TfR1 protein levels in CCR6+ Jurkat cells after 18 h H7 GTAC treatment. (d) Migration of hCCR6+ Jurkat T cells with varying doses of either hCCL20 or mCCL20 after 2 h. Data are presented as mean from n = 2 biological replicates. (e) BLI measurements demonstrate affinity of TfR1HIGH Nb and lower-affinity variant TfR1INT1 for binding to mouse TfR1. (f) Cell surface FLAG-hCCR6 and mTfR1 levels measured by flow cytometry on hCCR6+ MC38 cells after 2 h treatment of GTAC variants. Data are presented as mean ± s.d. from n = 3 biological replicates. (g) Hematoxylin and eosin staining and evaluation of epidermal thickness of mouse ear skin. Bar = 100 μm. Each dot represents an individual mouse. Statistical comparison performed via unpaired two-tailed t-test. Error bars indicate mean ± s.d. from n = 2 (Untreated), 3 (TfR1HIGH GTAC), or 4 (CCR6-Fab-Fc) ear skin samples. (h) Gating strategy for detecting neutrophils and inflammatory monocytes (left). Representative histograms of hCCR6 on neutrophils and inflammatory monocytes (right). (i) qPCR analysis of the mRNA levels of hCCR6, Ccl20, and Il23a in the whole ear skin 24 h after 6 applications of IMQ. Results are normalized to Gapdh expression. Statistical comparisons performed via two-tailed Mann-Whitney tests. Error bars indicate mean ± SEMs from n = 4 (Untreated, n = 3 for IL23a), 6 (TfR1HIGH GTAC), or 8 (CCR6-Fab-Fc) ear skin samples. Panels g-i: data shown are from a single experiment. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Exact p-values are listed in SI Table 7. Diagram in a created in BioRender; Zhou, X. https://biorender.com/hymki1y (2026).

Source data

Extended Data Fig. 13 Pharmacodynamic and pharmacokinetic analysis of CCR6 GTACs.

Immunofluorescence staining of mTfR1 in (a) spleen, (b) or brain, sections from hCCR6+ mice treated with 10 mg/kg TfR1HIGH GTAC or CCR6 Fab-Fc. (c) Quantification of mTfR1 immunofluorescence in spleen and brain sections. Each dot represents an individual tissue section. Data are presented as mean ± s.d. from n = 5 individual tissue sections. Statistical comparisons performed via Mann-Whitney test. (d) Intracellular iron levels in MDA-MB-231 cells after 72 h treatment with or without 10 nM antibody or 100 μM Ammonium iron (II) sulfate. Statistical comparisons performed via unpaired two-tailed t-tests. Data are presented as mean from n = 2 biological replicates. (e) Western blots of plasma levels of CCR6 GTACs and CCR6 Fab-Fc across time in mice injected intraperitoneally with 10 mg/kg antibody. n = 3 mice per group. Exact p-values are listed in SI Table 7.

Source data

Supplementary information

Supplementary Figures (download PDF )

This file contains Supplementary Figs. 1–10.

Reporting Summary (download PDF )

Supplementary Table 1 (download XLSX )

AT1R depletion and antagonism using GTAC designs explained in Supplementary Fig. 3. Potency and efficacy of GTAC-mediated depletion of cell surface AT1R expression in FLAG-AT1R+ TfR1+ Expi293 cells after 2 h treatment with GTACs, measured by flow cytometry (rows 1–7); IC90 values of GTAC-mediated depletion of cell surface AT1R or total AT1R levels in FLAG-AT1R-GFP+ TfR1+ Expi293 cells after 4 h GTAC treatment, measured by flow cytometry (row 8). Data represent curve fits from n = 3 biological replicates. Potency and efficacy of GTAC-mediated NFAT inhibition in FLAG-AT1R+ TfR1+ NFAT-Luc+ Expi293 cells after 2 h pre-treatment with GTACs, followed by 4 h stimulation with 20 nM AngII (rows 1–7). Data represent curve fits from n = 2 biological replicates.

Supplementary Table 2 (download XLSX )

RXFP1 depletion and antagonism using GTAC designs explained in Supplementary Fig. 4. Potency and efficacy of GTAC-mediated depletion of cell surface RXFP1 expression in FLAG-RXFP1+ TfR1+ Expi293 after 2 h treatment. Data represent curve fits from n = 2 biological replicates, except for constructs 1–4, 11, 13, 14 and 18, which are from n = 3 biological replicates. Potency and efficacy of GTAC-mediated cAMP inhibition in FLAG-RXFP1+ TfR1+ GloSensor+ HEK293T, after 2 h pre-treatment with GTAC and 30 min stimulation with 1 nM RLN2. Data represent curve fits from n = 3 biological replicates.

Supplementary Table 3 (download XLSX )

CCR6 depletion by GTACs using GTAC designs explained in Supplementary Fig. 5. Potency and efficacy of GTAC-mediated depletion of cell surface CCR6 expression after 2 h treatment. Flow cytometry (anti-FLAG staining) with FLAG-CCR6+ Jurkat cells. Data represent curve fits from n = 2 biological replicates, except for constructs 1, 2, 3, 4 and 11, which are from n = 3 biological replicates.

Supplementary Table 4 (download XLSX )

Total number of CCR6-GFP+ and TfR1-BFP+ HeLa cells imaged and analysed across different endolysosomal markers and treatment conditions.

Supplementary Table 5 (download XLSX )

Pearson colocalization coefficients of CCR6-GFP+ and TfR1-BFP+ HeLa cells imaged and analysed across different endolysosomal markers and treatment conditions.

Supplementary Table 6 (download XLSX )

qPCR primer sequences. Primers for MMP9, MMP23B, NOTCH1, NOTCH3 and DVL1 for qPCR in Extended Data Fig. 2l. Primers for mGapdh, mIl1β, mCxcl2, mCcl20, mIl17a, mIl23a and hCCR6 for qPCR in Fig. 5m and Extended Data Fig. 12i.

Peer Review File (download PDF )

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Rhee, K., Shue, L., Adachi, A. et al. GPCR antagonism via rewiring of receptor trafficking and degradation. Nature (2026). https://doi.org/10.1038/s41586-026-11088-0

Download citation

  • Received: 07 July 2025

  • Accepted: 26 August 2026

  • Published: 30 September 2026

  • Version of record: 30 September 2026

  • DOI: https://doi.org/10.1038/s41586-026-11088-0

来源:Nature 综合主刊 · nature.com