Abstract
Short half-lives in circulation and poor transport across the blood–brain barrier limit the utility of cytokines and growth factors acting as receptor agonists. Here we show that surrogate receptor agonists with longer half-lives in circulation and enhanced transport rates across the blood–brain barrier can be generated by genetically inserting macrocyclic peptide pharmacophores into the structural loops of the fragment crystallizable (Fc) region of a human immunoglobulin. We used such ‘lasso-grafting’ approach, which preserves the expression levels of the Fc region and its affinity for the neonatal Fc receptor, to generate Fc-based protein scaffolds with macrocyclic peptides binding to the receptor tyrosine protein kinase Met. The Met agonists dimerized Met, inducing biological responses that were similar to those induced by its natural ligand. Moreover, lasso-grafting of the Fc region of the mouse anti-transferrin-receptor antibody with Met-binding macrocyclic peptides enhanced the accumulation of the resulting Met agonists in brain parenchyma in mice. Lasso-grafting may allow for designer protein therapeutics with enhanced stability and pharmacokinetics.
Similar content being viewed by others
Main
The clinical use of cytokines and growth factors as therapeutics has been approved by the US Food and Drug Administration1,2, and their potential application in emerging areas, such as neurogenesis and brain repair3,4, is a topic of intense research. However, their inherent structural properties render them challenging to engineer for improved physicochemical stability and pharmacokinetics, in particular, for better half-life or blood–brain barrier (BBB) penetrance. Existing methods that control the pharmacokinetics of protein therapeutics include poly(ethylene glycol) conjugation and fusion with the fragment crystallizable (Fc) region of immunoglobulin5,6,7 to extend their half-lives; and fusion with Fab of the anti-transferrin receptor (TfR) antibody to enhance their penetration across the BBB8,9,10. However, the extent to which these methods can improve the pharmacokinetics of the proteins without adversely affecting their bioactivities is dependent on the properties of each protein5,6,7,8. To overcome the inherent structural limitations of cytokines and growth factors, there has been progress in the development of surrogate agonists that are structurally unrelated to native ligands11,12. Despite these recent advances, there remains a large unmet need for more robust and versatile methods to design protein therapeutics with desired physicochemical stability and pharmacokinetics.
Macrocyclic peptides have emerged as a promising novel class of drug candidates boasting a number of desirable features, such as antibody-like binding affinity and specificity13,14, the ability to target unique chemical spaces15,16,17,18 and efficient discovery through both rational/computational design and in vitro display18,19,20,21. In general, macrocyclic peptides display greater affinities for targets than linear peptides due to their constrained cyclic structures. An intriguing possibility in extending the applicability of these macrocyclic peptides is to ‘engraft’ them onto protein scaffolds to allow for functional combinations of peptide and protein18,22,23,24,25,26. However, grafting de novo identified peptides to protein loops has been challenging due to the potential misfolding of both the grafted peptide and the host protein26.
The development of the RaPID (random non-standard peptides integrated discovery) system, which integrates messenger RNA display and genetic code reprogramming, has enabled the discovery of thioether-based cyclized macrocyclic peptides with exquisite binding specificity against target proteins16,17,18. In previous studies, we have shown that RaPID-derived pharmacophore sequences can be readily implanted onto surface-exposed loops of proteins and maintain both functions of the guest peptide and the host protein, which we termed ‘lasso-grafting’27,28,29. The observed exceptional grafting compatibility is probably due to the intrinsic property of the pharmacophore motifs to self-fold into target-binding conformation similar to the parental macrocycle, even in the context of unrelated loop structure of the scaffold proteins18,30.
In this study, we leverage the desirable scaffolding properties of the Fc fragment, its long half-life, versatility in combination with Fab5,6,31,32 and ease of production to show the feasibility and application of lasso-grafting. Using the Fc as a scaffold, we generated Met receptor agonists that are characterized by markedly improved half-life and BBB penetrance.
Results
Met agonists generated by lasso-grafting Fc
Met is a receptor tyrosine kinase activated upon dimerization triggered by its ligand, hepatocyte growth factor (HGF). Met-HGF is centrally involved in morphogenesis during development, wound repair and organ homoeostasis by stimulating cell growth, survival and migration33,34. Recombinant HGF protein displays therapeutic efficacy in preclinical models34,35 and in patients with specific diseases36,37. However, its short half-life and inability to be delivered beyond the BBB have long limited its potential in medical applications38,39. To address these shortcomings, we inserted by lasso-grafting the linearized pharmacophore sequence of aMD4 or aMD5, two macrocyclic peptides that bind to the ectodomain of Met40 into each of the eight loops of Fc (Fig. 1a). These two series of aMD4- or aMD5-grafted Fc fragments (hereafter Fc(aMD4) and Fc(aMD5), respectively), each displaying two Met binders within 31–42 Å of each other, have the potential to proximate two Met receptors on the cell surface. The expression levels of Fc(aMD4) and Fc(aMD5) in Expi293F cells were similar to that of the control Fc, with the exception of Fc(aMD5)T3 (Extended Data Fig. 1), suggesting that lasso-grafting generally preserved the high expression of Fc.
The activation of cellular Met by purified Fc(aMD4) or Fc(aMD5) was highly dependent on the grafting site (Fig. 1b,c), where B3 was the optimal site for aMD4 (Fig. 1b). Our analyses show that Fc(aMD4)B3 activated Met with an half maximal effective concentration (EC50) comparable to that of ungrafted aMD4-dimer peptide (EC50: 4.5 ± 0.2 nM vs 5.1 ± 0.6 nM, respectively), and a slightly reduced maximum activation relative to HGF-induced phosphorylation (74.8% ± 0.4% vs 101.3% ± 0.1%, respectively). In contrast, grafting aMD4 into other loops resulted in lower Met activation relative to aMD4-dimer peptide. In the case of aMD5, B1 was the best grafting site, where Fc(aMD5)B1 activated Met with a 3.7-fold higher EC50 compared with ungrafted aMD5-dimer peptide, and a slightly lower maximum activation (EC50: 16.6 ± 1.6 nM vs 4.5 ± 0.2 nM, respectively; maximum activation: 76.2% ± 1.4% vs 105.0% ± 4.2%, respectively; Fig. 1c). Grafting aMD5 into other loops either reduced or abolished Met activation. Notably, the agonistic activities of lasso-grafted Fc correlated well with their binding strengths to cell surface Met (Supplementary Fig. 1a) or Met ectodomain fragment (MetECD) (Extended Data Fig. 2 and Supplementary Fig. 1b), indicating that their efficacies are determined largely by their affinity to Met. The difference in the graft-site dependency of the two peptides is probably due to the steric effects imposed by the difference in their binding site on the Met ectodomain and their divergent spatial arrangements in the Fc structure (Supplementary Fig. 2).
To improve agonistic activity, we appended a Cys residue at each end of the aMD4 sequence to promote a tightly closed macrocyclic conformation via disulfide linkage (Fig. 2a and Supplementary Fig. 3). The disulfide-bonded derivatives aMD4 (hereafter prefixed with ‘ds’) at T1 and B3 showed greater maximum Met activation than the control (T1, P = 0.039; B3, P < 0.0001), while no effects were observed in the EC50 (Supplementary Fig. 3f) and dissociation constant (KD) values of MetECD binding (Extended Data Fig. 2). In contrast, dsB2 showed no improvement in agonistic activity, maximum activation or EC50 values (Supplementary Fig. 3f), suggesting that the reduced Met affinity after grafting into these two sites was not due to suboptimal conformation of the grafted peptide motif per se. This was further supported by the similar conformations of aMD4 or aMD5 when grafted at the B1 or B3 position (Supplementary Fig. 2).
The cellular responses induced by Fc(aMD4)B3 in EHMES-1 cells or in hepatocytes were highly comparable to those of HGF (Fig. 2 and Extended Data Fig. 3). Fc(aMD4)B3 at 1 nM induced phosphorylation of Met at Y1234/1235 and activated subsequent downstream signalling with phosphorylation of Akt and Erk1/2 at levels comparable to those induced by HGF (Fig. 2b and Supplementary Fig. 4). The levels of phosphorylation induced by Fc(aMD4)B3 and HGF at three different tyrosine residues of Met, as well as the induction kinetics of Met, Akt and Erk1/2 phosphorylation, were comparable (Extended Data Fig. 3). The selectivity of Fc(aMD4)B3 for Met was confirmed by a survey of tyrosine phosphorylation of 49 receptors (Fig. 2c and Supplementary Fig. 5). The changes in gene expression induced by Fc(aMD4)B3 or HGF were examined in primary human hepatocyte spheroid culture by RNA-seq (Fig. 2d–g). A scatterplot shows the correlation of gene expression changes induced by HGF and Fc(aMD4)B3 relative to untreated spheroids (Fig. 2e). Lastly, a heat map of differentially expressed genes (DEGs) relative to untreated spheroids indicates that both Fc(aMD4)B3 and HGF similarly altered the expression of more than 2,000 transcripts (Fig. 2f), with significant enrichment of gene ontology (GO) terms related to wound healing and liver functions (Fig. 2g).
Mechanism of Met dimerization by Fc(aMD4)B3
As peptides generated de novo could differentially bind to the target receptors in a way completely different from that of the native ligands, we could theoretically generate agonists that do not compete with native ligands or are capable of activating mutated receptors that lack ligand binding. To investigate this potential, we first mapped the putative Fc(aMD4)B3 binding site (shown in red in Fig. 3a) using chimeric MetECD fragments that variably fused human and mouse MetECD (Extended Data Fig. 4), exploiting the specificity of aMD4 for human Met40. The putative binding site is mapped to the lower face of the Sema domain of Met, which does not overlap with the HGF binding site41, hence explaining why aMD4 peptide does not compete with HGF-induced Met activation40.
We next examined the structure of the Met dimer induced by Fc(aMD4). Complexes between MetECD and Fc(aMD4) were stabilized by crosslinking using bis(sulfosuccinimidyl)suberate (BS3) and analysed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS–PAGE) (Fig. 3b and Extended Data Fig. 5a,b). The results indicated the formation of complexes between MetECD and Fc(aMD4) in 1:1 or 2:1 stoichiometry. The relative efficiency of various Fc(aMD4) in forming a 2:1 signalling complex was in the order of B3 > B1 > B2, which correlated well with their ability to activate cellular Met. The BS3-crosslinked 2:1 complex of MetECD and Fc(aMD4)B3 was purified by size exclusion chromatography (Extended Data Fig. 5c) and subjected to single-molecule examination by high-speed atomic force microscopy (HS-AFM)42, with Fc and MetECD as controls (Fig. 3c–f and Supplementary Videos 1–4). On HS-AFM, MetECD showed a globular Sema domain and Ig-like, plexins, transcription factors (IPT) stalk domains, with the relative positioning of each IPT domain being flexible (Fig. 3d and Supplementary Video 2). Fc(aMD4)B3 bound to the lower face of the Sema domain and bridged two Met molecules while having the IPT stalk domains sprayed away (Fig. 3e and Supplementary Video 3) or attached (Fig. 3f and Supplementary Video 4). The flexibility of IPT domains probably allows for the proper association of the intracellular kinase domains of two Met molecules, this association being essential for Met activation33,34. These results show that while Fc(aMD4)B3 binds to Met on a site different from that in HGF, its recruitment of two Met receptors in close proximity enables it to activate Met to the same extent as HGF.
FcRn binding, half-life and in vivo activity
The long half-lives of Fc and antibodies are mainly maintained through their binding to the FcRn43,44,45,46. As such, we have selectively lasso-grafted into loop locations that are distal from the FcRn binding site (Fig. 4a). Accordingly, our analysis of binding kinetics between Fc(aMD4) and immobilized FcRn showed that lasso-grafting into most loops did not affect the affinity of Fc to FcRn, although Fc(aMD4)T2 and M2 exhibited slightly higher KD values than control Fc (Fig. 4b and Extended Data Fig. 6). This suggests that lasso-grafting on Fc preserves the affinity of Fc to FcRn.
We next evaluated the serum half-life of Fc(aMD4)B3 in comparison to those of control Fc and HGF in wild-type mice after a single intravenous (i.v.) administration at equimolar amount per kg (Fig. 4c). The serum concentration of HGF decreased to below 0.01 nM within 1 h, which was unable to activate Met. In contrast, Fc(aMD4)B3 showed an extended half-life comparable to that of control Fc (Fig. 4c). To accurately determine the pharmacokinetics of our human Fc-based variants, we measured the serum half-lives of Fc(aMD4)B3 and control Fc in mFcRn−/− hFcRn+ mice47 (Fig. 4d). The serum half-life of Fc(aMD4)B3 in these mice was 49.4 h and comparable to that of control Fc (46.6 h). The serum concentration of Fc(aMD4)B3 remained above 1 nM, the minimum concentration for activating Met (Fig. 2a,b), for up to 200 h after a single i.v. administration at 0.74 mg kg−1 (Fig. 4d).
The in vivo activity of Fc(aMD4)B3 was next examined using chimeric mice transplanted with human hepatocytes with approximately 80% of hepatocytes being humanized (PXB-mice)48, as Fc(aMD4)B3 is incapable of activating murine Met40. We first confirmed the extended serum half-life of Fc(aMD4)B3 in PXB-mice after a single i.v. or subcutaneous (s.c.) injection at 5 mg kg−1 (Supplementary Fig. 6). Subsequently, the proliferation and gene expression of human hepatocytes 46 h after a single s.c. injection of Fc(aMD4)B3 or PBS in PXB-mice (Fig. 4e–j) were analysed. Fc(aMD4)B3 significantly increased replicative DNA synthesis of hepatocytes compared with those of control mice treated with PBS (3.73% ± 0.31% vs 0.37% ± 0.07%, respectively, P < 0.0001) (Fig. 4f,g). A heat map of DEGs between Fc(aMD4)B3-treated and PBS-treated livers indicated that Fc(aMD4)B3 altered the expression of more than 3,200 transcripts (Fig. 4h). Significantly enriched GO terms among them were primarily involved in mitotic DNA replication/cell cycle and metabolic processes (Fig. 4i,j). Taken together, these results show that lasso-grafting generated a Met-activating Fc that is bioactive in vivo, with a markedly improved half-life that is comparable to that of unmodified Fc.
Met agonists with BBB penetration
It has been reported that HGF-induced Met activation exerts beneficial neuroprotective effects in preclinical models of cerebral ischaemia, spinal cord injuries and neurological pathologies, such as Alzheimer’s disease, amyotrophic lateral sclerosis and multiple sclerosis35. In line with earlier studies, we confirmed that Met is expressed in neurons in mouse brains (Supplementary Fig. 7). To generate Met agonists capable of crossing the BBB, we applied lasso-grafting of aMD4 on the Fc of the anti-mouse TfR (mTfR) antibody49 in either single- or double-Fab configuration (sTfR(aMD4) and dTfR(aMD4), respectively) (Fig. 5a,b and Supplementary Fig. 8). Lasso-grafting did not significantly alter the previously reported9,10 affinity of these antibodies to mTfR, suggesting that Fab affinity is largely preserved (Fig. 5c). In contrast, both sTfR(aMD4) and dTfR(aMD4) showed reduced Met activation compared with Fc(aMD4), with 9.6- and 18.7-fold reduction in EC50 values (20.2 ± 3.4 nM vs 39.2 ± 22.3 nM vs 2.1 ± 0.2 nM, respectively) (Fig. 5d). This was possibly due to the presence of large Fab arm(s) on sTfR(aMD4) and dTfR(aMD4), which may have interfered with their binding and/or dimerization of Met on the cell surface.
Lastly, we examined the BBB penetration of the three aMD4-grafted variants in wild-type mice after a single i.v. administration at therapeutically relevant doses of the proteins at equimolar amount per kg (Fig. 5e–j). Their concentrations in PBS-perfused brains or serum at 24 h post i.v. injection were determined by ELISA for human IgG Fc. Of the three variants, sTfR(aMD4) showed the highest percentage injected dose per gram brain (0.96% ± 0.07%), brain-to-serum ratio (8.3% ± 0.7%), as well as the highest brain concentration (10.7 ± 1.2 nM) (Fig. 5f–i), which is sufficient for Met activation (Fig. 5d). These results are in line with previous reports of low-affinity anti-TfR antibody accumulation in the brain9,10,50. In comparison, these values were significantly lower for dTfR(aMD4) and were nearly negligible in Fc(aMD4) (Fig. 5f–i). Concordant with the higher brain concentration determined by ELISA, sTfR(aMD4) showed prominent parenchymal staining co-localized with NeuN, GFAP and Iba1, indicating its distribution to neuron, astrocytes and microglia, respectively (Fig. 5j(middle right) and Supplementary Fig. 9). In comparison, dTfR(aMD4) showed less extensive neuronal staining and more endothelial staining (Fig. 5j(right)), in line with previous reports of high-affinity anti-TfR antibody accumulation in brain endothelial cells9,10. Lastly, Fc(aMD4) was undetectable in the brain parenchyma (Fig. 5j(middle left)). Further evaluation of detailed pharmacokinetic profiles, including time course, will be needed for future therapeutic evaluation of sTfR(aMD4) in preclinical models. Taken together, these observations showed that we have generated BBB-penetrating Met agonists by lasso-grafting of Met-binding macrocyclic peptide on the Fc of anti-mTfR antibody.
Discussion
The grafting of de novo identified peptides to improve their stability and bioavailability has been attempted on a variety of scaffolds23,24,25,26. These past studies revealed that successful grafting is highly dependent on the combination and compatibility between grafted peptides and scaffold proteins25,26. As such, the selection of scaffold protein is a key consideration, as it determines the success of peptide grafting while conferring desirable properties such as metabolic stability and cell permeability to the engineered products24,25,26. To cater for this broad spectrum of desirable features, disulfide-stabilized peptides and miniproteins have been engineered to act as specialized scaffolds for peptide grafting23,24,25,26. However, these specialized scaffolds are small non-human proteins with no particular bioactivities of their own. Although natural human proteins are desirable scaffolds for peptide grafting, there has been very limited attempts at grafting de novo identified peptides onto natural human scaffolds so far. A major reason behind this is that grafting synthetic de novo identified peptides into an independently folded domain of a natural scaffold would often result in the misfolding and inactivation of both entities. It is therefore remarkable that our data thus far indicate that the functional grafting of de novo identified macrocyclic peptides into the protein loops of natural scaffolds is far more feasible than previously expected27,28,29.
As a protein scaffold, the Fc is highly desirable because of its high expression, ease of manufacture, long half-life, and Fab compatibility5,6,31,32. In this study, we report that lasso grafting Fc with macrocyclic peptides yielded Met receptor agonists with extended half-life and BBB penetration. Our application of lasso-grafting on Fc has revealed three key strengths to this approach: Firstly, lasso-grafting preserved the overall biochemical characteristics of Fc in terms of expression level, FcRn affinity, and Fab compatibility. Secondly, lasso-grafted macrocyclic peptide largely preserved parental macrocyclic conformation even without disulfide linkage. Thirdly, there is a wide choice of suitable insertion positions in Fc. Our study further showed that the target affinity of grafted peptides is site-dependent. For example, the T-loops were less effective in our study, probably due to the presence of a N-terminus hinge region that could interfere with the binding of grafted peptides to the target protein, Met. Yet, these loops acted as suitable grafting positions for other receptor binders27,29. Notwithstanding this, the robust graft compatibility between macrocyclic peptides and Fc, in combination with a wide choice of insertion positions, renders lasso-grafting a particularly effective approach of Fc engineering.
An important implication of our findings is that lasso-grafting protein scaffolds with specific properties can serve as a framework for engineering designer receptor agonists with the desired properties based on the choice of protein scaffolds. As a demonstration of this, we generated BBB-permeable agonists by lasso grafting an anti-TfR antibody, thereby leveraging an established BBB-permeation technique8,9,10. At present, there are several efforts in developing anti-TfR Fab fusion antibodies with blocking properties, but only few that attempt to develop agonist antibodies51. Lasso-grafting could be used to design agonist BBB-permeating antibodies, in addition to antagonistic ones. Furthermore, one major drawback of current BBB-permeating antibodies is that, owing to their low brain-to-serum ratio and long half-life in circulation, systemic concentrations often exceed the desired range for extended periods. This could cause unwanted neo-immunogenicity or side effects, especially if the targeted molecule is functionally important in other tissues. With lasso-grafting, this could be avoided by grafting functional peptides directly into scaffold proteins with a short half-life, such as the anti-TfR Fab.
Contrary to existing methods that improve the pharmacokinetics of protein therapeutics by modifying the protein, lasso-grafting endows novel functions derived from small peptides to well-understood protein scaffolds27,28,29, hence avoiding potential pitfalls such as reduced biological activity, unfavourable protein characteristics for manufacturing or neo-immunogenicity arising from protein fusion or from poly(ethylene glycol)5,6,7. Although Fc serves as an excellent scaffold because of its immunosuppressive property52, the immunogenicity, safety and efficacy of lasso-grafted proteins await future evaluation as part of drug development. In the same vein, HGF has shown therapeutic efficacies in preclinical models of various diseases such as non-alcoholic steatohepatitis53,54 or neurological pathologies35,37. The therapeutic advantages of the lasso-grafted Met agonists generated in this study should likewise be further evaluated in disease models.
In view of the rapid advances in potent bioactive and graft-compatible macrocyclic peptides, lasso-grafting holds great promise for the generation of designer protein therapeutics with the desired physicochemical stability and pharmacokinetics, as well as for engineering multifunction proteins and antibodies.
Methods
Protein design, construction and purification
To design peptide-grafted Fc protein, Fc protein was structurally inspected to find an appropriate insertion point where the approximate distance between the two anchor point residues located within an exposed loop was less than 7 Å. For the construction of control or Met-binding macrocyclic peptide (aMD4 or aMD5)-grafted Fc, human IgG1 Fc (residues 104–330, Uni-Prot P01857) was used without any tags. The internal sequence of aMD4 or aMD5 appended with or without Cys and two spacer residues of Gly at both ends were inserted into these identified sites (T1 to B3). For the construction of anti-mouse TfR antibodies with aMD4ds insertion at the B3 site, the Fab fragment of a rat anti-mouse TfR monoclonal antibody (clone 8D3)49 was used. The VH to CH1 fragment of 8D3 was fused to the N terminus of Fc(aMD4) to yield heavy chain (HC) of dTfR(aMD4). The VL to CL fragment of 8D3 was used to yield light chain (LC). For the construction of sTfR(aMD4), a flag tag was added at the N terminus of Fc(aMD4). The amino acid sequences used for s/dTfR(aMD4) can be found in Supplementary Fig. 8. All expression constructs were made using pcDNA3.1-based backbone with appropriate signal peptide and the coding region of all expression constructs was verified by DNA sequencing. Protein expressions were performed using the Expi293 expression system (Thermo Fisher). Secreted proteins were purified on a HiTrap Protein A HP column (Cytiva) using elution with 0.1 M citrate acid-sodium citrate buffer (pH 3.5), followed by neutralization, dialysis against phosphate-buffered saline (PBS) and size exclusion chromatography on a Superdex 200 Increase 10/300GL column (Cytiva) equilibrated with PBS. SDS–PAGE analysis and purity of purified Fc variants used in the assay can be found in Supplementary Fig. 10. For the expression and purification of sTfR(aMD4), three DNAs (TfR(aMD4) HC, TfR(aMD4) LC and Fc(aMD4)) were used to co-transfect Expi293F cells, yielding the mixture of sTfR(aMD4), dTfR(aMD4) and Fc(aMD4). The sTfR(aMD4) was purified from the latter two components by sequential anti-Flag affinity purification and size exclusion chromatography on a Superdex 200 Increase 10/300GL column equilibrated with PBS.
Quantification of phosphorylation level of Met, Akt and Erk1/2 in situ
EHMES-1 human mesothelioma cells were seeded at 8,000 cells per well in a 96-well black µClear-plate (Greiner Bio-One) and cultured in RPMI1640 medium supplemented with 10% fetal bovine serum (FBS) for 24 h. Cells were stimulated with each protein in RPMI1640 medium supplemented with 10% FBS for 10 min or indicated time, washed with ice-cold PBS, fixed with 4% paraformaldehyde in PBS for 30 min, washed with PBS and blocked with 5% goat serum and 0.02% Triton X-100 in PBS for 30 min. Phosphorylation of Met, Akt or Erk1/2 was detected using an anti-phospho-Met (Tyr1234/1235) antibody (1:1,000 dilution, D26, Cell Signaling Technologies), phosphorylated Akt (S473) (1:1,000 dilution, D9E, Cell Signaling Technology) or phosphorylated Erk1/2 (T202/Y204) (1:1,000 dilution, D13.14.4E, Cell Signaling Technology), and a horseradish peroxidase (HRP)-conjugated anti-rabbit goat antibody (1:1,000 dilution, Dako). Then, chemiluminescence developed with ImmunoStar LD reagent (Wako) was measured using an ARVO MX plate reader (Perkin Elmer).
Flow cytometry
The binding of peptide-grafted Fc proteins to Met-knockout CHO cells and Met-reconstituted CHO cells27 was detected using flow cytometry. Cells were detached from dishes by brief treatment with 0.25% trypsin and 1 mM ethylenediaminetetraacetic acid, plated at 200,000 cells per well and incubated with peptide-grafted Fc proteins diluted at 10 mg ml−1 in 100 ml Ham’s F-12 medium containing 5% FBS on ice for 1.5 h. After washing twice with ice-cold PBS, cells were incubated with Alexa Fluor 488-labelled goat anti-human IgG (1:400 dilution in 100 µl Ham’s F-12 medium containing 5% FBS, Thermo Fisher, A11013) on ice for 30 min, then analysed on an EC800 system (Sony). Gate on forward scatter vs side scatter was set to include all cell populations but excluding debris and dead cells as shown in Supplementary Fig. 11. The data were analysed with FlowJo software (Tomy Digital Biology).
Surface plasmon resonance analysis
The ectodomain fragment of human Met (MetECD, 1–931) was C-terminally fused with a biotin acceptor sequence (SSLRQILDSQKMEWRSNAGG) and co-expressed with a biotin ligase (BirA) in Expi293F cells to achieve BirA-mediated biosynthetic biotinylation and immobilized onto a Series S Biotin CAPture chip (Cytiva) at a surface density of ~210 resonance units (RU). The binding of Fc and peptide-grafted Fcs was evaluated by injecting the analyte solutions serially diluted using the running buffer (20 mM HEPES-NaOH (pH 7.5), 150 mM NaCl, 0.05% Tween 20). The runs were conducted in single cycle kinetics mode employing the following parameters: flow rate of 30 µl min−1, contact time of 120 s and dissociation time of 300 s. After each run, the surface was regenerated by injecting the regeneration buffer (6 M guanidine-HCl, 250 mM NaOH) until the response returned to the original baseline level. The binding curves were obtained by subtracting the RU of the reference cell (un-immobilized) from that of the measurement cell (immobilized with MetECD) and used to derive kinetic binding values. Data were obtained using a Biacore T200 instrument (Cytiva) at 25 °C, and the results were analysed using Biacore T200 evaluation software v3.2 (Cytiva). Biotinylated human FcRn/β2M (ACROBiosystems) was captured on the surface of streptavidin-coated chips (Cytiva) at 180~220 RU. Analytes were diluted in running buffer (50 mM phosphate (pH 6.0), 150 mM NaCl, 0.01% Tween 20) and injected at 30 µl min−1 for 1 min, followed by dissociation for 0.7 min. The binding curves were obtained by subtracting the RU of the reference cell (un-immobilized) from that of the measurement cell (immobilized with FcRn/2M) and following axis-zeroing, sensorgrams were fit locally to a 1:1 Langmuir binding model. Data were obtained using a Biacore 3000 instrument (Cytiva) at 25 °C, and the results were analysed using BIAevaluation software v4.1 (Cytiva).
Predicted structures of the CH3 domain of Fc with aMD4 or aMD5 insertion
Structures of the CH3 domain of Fc with aMD4 or aMD5 inserted in the B1 or B3 loop were predicted using ColabFold and AlphaFold2 in MMseqs255.
Trypsin digestion and liquid chromatography–mass spectrometry (LC–MS/MS)
Trypsin digestion and LC–MS/MS analysis were performed by KANEKA TECHNO RESEARCH. One hundred µg of Fc(aMD4ds)T1 or Fc(aMD4ds)B3 in PBS were treated with 100 mM ammonium bicarbonate (Sigma-Aldrich) containing 8 M urea (Fujifilm Wako Pure Chemicals). After desalting using an Amicon Ultra device (MWCO: 10 K, Millipore), 10 µg of each protein was digested with trypsin (Promega) for 12 h at 37 °C under non-reducing conditions. The reaction was stopped by the addition of formic acid. The samples were injected into a Nexera X2 UHPLC system (Shimadzu corporation) connected to a SCIEX TripleTOF 6600 system (AB SCIEX) equipped with an ESI ionization source. Peptides were chromatographically separated on a C18 reverse-phase column with linear-gradient of mobile phase A (water containing 0.1% formic acid) and mobile phase B (acetonitrile (Fujifilm Wako Pure Chemicals) containing 0.1% formic acid). MS and MS/MS data were collected using IDA, positive ion mode. The LC–MS raw data were processed using SCIEX BioPharmaView software. The following criteria were used to identify peptides: mass accuracy for the matched precursor must be within 5 ppm of mass error, and peptide matching score was set to at least 3 for peptide identification.
Crosslink and isolation of Fc(aMD4)/MetECD complex
MetECD (60 nM) and Fc(aMD4)B1/B2/B3 (20 nM) were mixed in 100 µl PBS with 0.01% Triton X-100 for 1 h. The samples were divided into two and treated with 1 mM BS3 (Thermo Fisher) or left untreated for 1 h, quenched by 50 mM Tris, subjected to SDS–PAGE (7.5% or 3–10% gel) and silver stained. For the HS-AFM observation, MetECD (2.6 µM) and Fc(aMD4)B3 (1.3 µM) were mixed in 100 µl PBS with 0.01% Triton X-100 for 1 h, treated with 1 mM BS3 for 1 h, quenched by 50 mM Tris and analysed by size exclusion chromatography on a Superose 6 Increase 10/300GL column (Cytiva) equilibrated with PBS. The separated fractions were analysed by SDS–PAGE and stained with Coomassie brilliant blue. A single fraction was subjected to HS-AFM analysis.
HS-AFM observations
HS-AFM measurements were operated in tapping mode. Cantilever deflection was detected with an optical beam deflection detector using an infrared (IR) laser (0.7 mW, 780 nm). The IR laser beam was focused onto the backside of a cantilever covered with gold film (Olympus, BL-AC10DS-A2) through a ×60 objective lens (CFI S Plan Fluor ELWD ×60; Nikon). Reflection of the IR laser from the cantilever was detected using a two-segmented PIN photodiode (MPR-1; Graviton). The free-oscillation amplitude was approximately 1 nm and the set-point amplitude was approximately 90% of the free amplitude for feedback control of HS-AFM. For the AFM probe, an amorphous carbon tip with a length of approximately 500 nm grown by electron beam deposition was used. For the AFM substrate, a mica surface treated with 0.01% 3-aminopropyl-triethoxysilane (Shin-Etsu Silicone) was used. All HS-AFM observations were performed in a buffer solution containing 50 mM Tris-HCl (pH 7.4) and 100 mM NaCl at room temperature.
Hepatocytes monolayer culture
Monolayers of primary human hepatocytes derived from chimeric mice with humanized liver were obtained from PhoenixBio and grown in collagen-coated 12-well plates. Hepatocytes were cultured in dHCGM medium consisting of Dulbecco’s modified Eagle’s medium (DMEM) with 10% FBS, 20 mM HEPES, 44 mM NaHCO3, antibiotics (100 U ml−1 penicillin G and 100 µg ml−1 streptomycin), 15 µg ml−1 l-proline, 0.25 µg ml−1 insulin, 50 nM dexamethasone, 5 ng ml−1 epidermal growth factor, 0.1 mM l-ascorbic acid, and 2% dimethyl sulfoxide (all supplements from PhoenixBio).
Hepatocyte spheroids
Cryopreserved human hepatocytes were obtained from Gibco. Cells were seeded into ultra-low-attachment 96-well plates (Nunclon Sphera, Thermo Fisher) at 1,500 viable cells per well and subsequently centrifuged at 200 × g for 2 min. Cells were seeded in 100 µl Williams E medium supplemented with 2 mM GlutaMAX, 100 U ml−1 penicillin, 100 µg ml−1 streptomycin, 4 µg ml−1 insulin, 1 µM dexamethasone, 15 mM HEPES (pH 7.4) and 10% FBS (all supplements from Oriental Yeast). From five days after seeding, when the spheroids were sufficiently compact, 50% of the medium was exchanged every 2 d with serum-free Williams E medium supplemented with 2 mM GlutaMAX, 100 U ml−1 penicillin, 100 µg ml−1 streptomycin, 6.25 µg ml−1 insulin, 6.25 µg ml−1 transferrin, 6.25 µg ml−1 selenous acid, 1.25 mg ml−1 BSA, 5.35 µg ml−1 linoleic acid, 100 nM dexamethasone and 15 mM HEPES (pH 7.4) (all supplements from Oriental Yeast). Spheroids were treated in a serum-free medium with 0.3 nM HGF or 30 nM Fc(aMD4)B3 on days 7 and 9. Forty-eight spheroids were collected for each group and subjected to RNA isolation at days 8 or 10.
Western blotting and phospho-RTK array
Monolayer hepatocytes were serum starved in DMEM with 0.2% BSA for 3 h, stimulated with HGF (0.01~1 nM), Fc(aMD4)B3 (1~100 nM), Fc (100 nM) or unstimulated (None) in DMEM with 10% serum for 10 min. EHMES-1 cells were serum starved in RPMI1640 with 0.2% BSA for 3 h, stimulated with HGF (0.04~1.1 nM), Fc(aMD4)B3 (1.3~33 nM) or unstimulated in RPMI1640 with 0.2% BSA for 10 min. These cells were washed with ice-cold PBS, and lysed in 400 µl of lysis buffer (50 mM Tris, 150 mM NaCl, 1% Triton X-100, 1% NP-40, 10% glycerol, 1 mM phenylmethylsulfonyl fluoride, 1 mM Na3VO4, 1 mM NaF, 1 mM ethylenediaminetetraacetic acid and protease inhibitors). The lysates were passed through a 27G needle and 0.45 µm filter, and centrifuged. Protein concentrations of lysates were measured by bicinchoninic acid assay (Thermo Fisher). The supernatants were subjected to western blotting for phosphorylated Erk1/2 (T202/Y204) (1:1,000 dilution, D13.14.4E, Cell Signaling Technology), Erk1/2 (1:1,000 dilution, 137F5, Cell Signaling Technology), phosphorylated Akt (S473) (1:1,000 dilution, D9E, Cell Signaling Technology), Akt (1:1,000 dilution, 11E7, Cell Signaling Technology) or GAPDH (1:1,000 dilution, 14C10, Cell Signaling Technology). The supernatants were subjected to immunoprecipitation with an anti-Met antibody (5 µg for 600 µl lysates, D-4, Santa Cruz Biotechnology) and western blotting for Met (1:1,000 dilution, D1C2, Cell Signaling Technology), phosphorylated Met (Y1234/Y1235) (1:1,000 dilution, D26, Cell Signaling Technology), phosphorylated Met (Y1003) (1:1,000 dilution, 13D11, Cell Signaling Technology) or phosphorylated Met (Y1349) (1:1,000 dilution, 130H2, Cell Signaling Technology). The blottings were followed by HRP-conjugated secondary antibody (Dako) and measurement of chemiluminescence developed with ImmunoStar LD reagent (Wako), using the image reader FUSION-SOLO.6S.EDGE (VIRVER). For the phospho-RTK array, hepatocytes were serum starved in DMEM with 0.2% BSA for 3 h, stimulated with HGF (1 nM), Fc(aMD4)B3 (100 nM), Fc (100 nM) diluted in DMEM with 10% serum for 10 min, washed with ice-cold PBS, lysed in 400 µl of lysis buffer 17 (R&D Systems) with protease inhibitor cocktail (Nacalai Tesque), passaged through a 0.45 µm filter and centrifuged. The lysates (1 mg) were analysed using the Human Phospho-RTK Array kit (R&D Systems).
RNA-seq
Total RNA of hepatocyte spheroids or RNAlater (Thermo Fisher)-treated livers from PXB-mice was prepared using the QIAzol lysis reagent (Qiagen) and shipped to the Bioengineering Lab (Kanagawa, Japan) for RNA quality control and RNA-seq on a DNBSEQ-G400 sequencer (MGI Tech). Complementary DNA libraries for RNA-seq were prepared using MGIEasy RNA directional library prep set (MGI Tech) and evaluated using AATI fragment analyzer (Advanced Analytical Technologies). The cDNA libraries were circularized using MGIEasy circularization kit (MGI Tech), prepared as DNA nano balls by DNBSEQ-G400RS high-throughput sequencing kit (MGI Tech), and followed by massively parallel sequencing (2×100 bp) on a DNBSEQ-G400 sequencer (MGI Tech). After excluding adaptor sequences using cutadapt (ver. 1.9.1) and sequences with low-quality score (<20) or less than 40 nucleotides using sickle (ver 1.33), reads were aligned to the reference human genome (GRCh38.p13) using the hisat2 software (v2.2.0). The alignment data were counted using featureCounts (ver. 2.0.0). Transcript abundances were measured in reads per kilobase of exon per million mapped reads (RPKM) and transcripts per million (TPM). DEG analysis was performed using DEGES in TCC (v1.18.0) and DESeq (v1.30.0) (FDR < 0.05 or <0.01, Benjamini–Hochberg, two-sided). All DEGs identified were subjected to cluster analysis using Gene Cluster 3.0. Significantly enriched GO terms were identified using GO Enrichment Analysis by PANTHER.
Half-life in serum
Half-life in serum up to 9 h was determined in wild-type C57BL/6 mice (female, 9 weeks, 18–21 g, obtained from Japan SLC) given a single tail-vein injection at 0.7 mg kg−1 for Fc, 0.74 mg kg−1 for Fc(aMD4)B3 and 1.0 mg kg−1 for human recombinant HGF (Kringle Pharma), diluted in 100 µl PBS. Half-life in serum up to 12 d was determined in mFcRn−/− hFcRn Tg 276 heterozygote mice (female, 9–12 weeks, 18–24 g) produced in-house by mating mFcRn−/− mice and hFcRn Tg 276 homozygote mice obtained from Jackson Laboratory. Mice were given a single tail-vein injection at 0.7 mg kg−1 for Fc or 0.74 mg kg−1 for Fc(aMD4)B3 diluted in 100 µl PBS. Half-life in serum up to 12 d was also determined in chimeric mice with humanized liver (PXB-mice, PhoenixBio, male, >12 weeks, 19–24 g) given a single tail-vein injection or subcutaneous injection at 5.0 mg kg−1 for Fc(aMD4)B3, diluted in 100 µl PBS. Blood was drawn from the submandibular vein using a Goldenrod animal lancet (Bio Research Center) or from the orbital plexus at each time point, processed to serum and stored at −80 °C until analysis. The serum concentrations of Fc and Fc(aMD4)B3 were determined using a human immunoglobulin recognition immunoassay (human IgG ELISA quantitation set, Bethyl Laboratories). The serum concentrations of HGF were determined using an in-house developed ELISA. All animal experimental procedures were conducted in accordance with the guidelines provided by the Proper Conduct of Animal Experiments (1 June 2006, Science Council of Japan). The procedures were approved by the Institutional Review Board of Kanazawa University or the Laboratory Animal Ethics Committee of PhoenixBio.
BrdU labelling in chimeric mice with humanized liver
Chimeric mice with humanized livers (PXB-mice, PhoenixBio) were generated from urokinase-type plasminogen activator-cDNA/severe combined immunodeficiency mice transplanted with human hepatocytes, with approximately 80% of hepatocytes being humanized48. PXB-mice (male, >12 weeks, 15–21 g) were given a single subcutaneous injection at 5 mg kg−1 for Fc(aMD4)B3 in 200 µl PBS or control PBS. Forty-four hours after administration of Fc(aMD4)B3 or control PBS, 5-bromo-2'-deoxyuridine (BrdU) (Sigma Chemical) was intraperitoneally injected into chimeric mice at a dose of 100 mg kg−1 for 2 h before killing. Livers isolated from PXB-mice were treated with RNAlater for RNA-seq or prepared for BrdU staining. Formalin-fixed paraffin sections of chimeric livers were incubated with a rat anti-BrdU antibody at 3 µg ml−1 (Abcam, ab6326), followed by an Alexa Fluor 488-conjugated anti-rat IgG goat polyclonal antibody at 1 µg ml−1 (Abcam, ab150157). Nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI, Thermo Fisher) at 300 nM. Fluorescent images were obtained using Keyence BZ-X810. The number of BrdU-positive nuclei was determined by counting at least 15,000 cells in 20 randomly selected visual fields in sections from the lateral left lobe and medial right lobe per animal. The procedures involving humanized liver tissues were approved by the Utilization of Human Tissue Ethical Committee of PhoenixBio. All animal experimental procedures were conducted in accordance with the guidelines provided by the Proper Conduct of Animal Experiments (1 June 2006, Science Council of Japan). The procedures were approved by the Laboratory Animal Ethics Committee of PhoenixBio.
Affinity to mTfR
MaxiSorp plates (Nunc) were coated with recombinant mouse TfR (2 µg ml−1 in PBS, 50 µl per well, R&D Systems) at 4 °C overnight, washed with 0.05% Tween 20 in PBS and blocked with 1% BSA in PBS. Plates were incubated with 1:3 serial titrations of dTfR(aMD4), sTfR(aMD4) or Fc(aMD4) diluted with 0.05% Tween 20 and 1% BSA in PBS at room temperature for 1 h, washed with 0.05% Tween 20 in PBS, incubated with HRP-conjugated anti-human IgG Fc antibody (0.5 µg ml−1, 100 µl per well, Bethyl Laboratories) at room temperature for 1 h and washed with 0.05% Tween 20 in PBS. Signals were generated by incubation with 3,3′,5,5′-tetramethylbenzidine substrate (Cell Signaling Technology) and absorbance was read using an ARVO MX plate reader (Perkin Elmer).
BBB penetration
Wild-type C57BL/6 mice (female, 7–8 weeks, 17–20 g, obtained from SLC) were given a single tail-vein injection at equimolar amount per kg: 26.5 mg kg−1 for dTfR(aMD4), 20.0 mg kg−1 for sTfR(aMD4) or 12.0 mg kg−1 for Fc(aMD4), diluted in 200 µl PBS. Twenty-four hours after injection, mice were anaesthetized and blood was drawn from the right ventricle and processed to serum. Brains were isolated from mice after perfusion with PBS at a rate of 2 ml min−1 for 8 min. Each half of the brain was homogenized in 1% NP-40 (Calbiochem) in PBS containing CompleteMini EDTA-free protease inhibitor cocktail tablets (Roche Diagnostics), rotated at 4 °C for 1 h, spun at 14,000 r.p.m. for 20 min, and supernatants collected. The serum and the brain lysates were subjected to antibody measurement using a human immunoglobulin recognition immunoassay (human IgG ELISA quantitation set, Bethyl Laboratories). To calculate the concentrations in the brain, ng g−1 brain was considered to be ng ml−1. Each half of the brain was embedded in optimal cutting temperature compound (Sakura Finetek Japan), frozen in liquid nitrogen and sectioned at 10 µm thickness. After fixation in 4% paraformaldehyde in PBS for 30 min, the sections were blocked in 5% BSA, 0.3% Triton X-100 in PBS and stained with a goat anti-human IgG Fc antibody (1:500 dilution, Bethyl Laboratories) and a mouse anti-mouse NeuN antibody (1:500 dilution, Millipore), a mouse anti-GFAP antibody (1:500 dilution, GA5, Cell Signaling Technologies) or a mouse anti-Iba1 antibody (1:500 dilution, GT10312, GeneTex) in 1% BSA and 0.1% Triton X-100 in PBS at 4 °C overnight. Anti-human IgG Fc and anti-NeuN were visualized with an Alexa Fluor 488-conjugated anti-goat secondary antibody or an Alexa Fluor 594-conjugated anti-mouse secondary antibody, respectively (1:200 dilution, Thermo Fisher). Fluorescent images of cortical regions were obtained with the Keyence BZ-X810. This study was approved by the Institutional Review Board of Kanazawa University and performed in accordance with the guidelines and regulations.
Statistical analysis
Graphpad Prism 6.0d was used for graphing and statistical analysis. Relevant statistical methods for individual experiments are detailed in figure legends.
Reporting summary
Further information on research design is available in the Nature Research Reporting Summary linked to this article.
Data availability
The main data supporting the findings of this study are available within the Article and its Supplementary Information. The RNA-seq data are available at the DDBJ Sequence Read Archive under accession numbers DRA014557 (hepatocyte spheroids) and DRA014558 (livers of PXB mice). The raw data generated during the study are available from the corresponding authors on reasonable request. Source data for the figures are provided with this paper.
References
Ray, A., Gulati, S. G. K. & Joshi, N. R. J. Cytokines and their role in health and disease: a brief overview. MOJ Immunol. 4, 00121 (2016).
Silva, A. C. & Sousa Lobo, J. M. Cytokines and growth factors. Adv. Biochem. Eng. Biotechnol. 171, 87–113 (2020).
Oliveira, S. L. et al. Functions of neurotrophins and growth factors. Cytometry A 83, 76–89 (2013).
Wang, S. et al. Therapeutic potential of a TrkB agonistic antibody for Alzheimer’s disease. Theranostics 10, 6854–6874 (2020).
Vargason, A. M., Anselmo, A. C. & Mitragotri, S. The evolution of commercial drug delivery technologies. Nat. Biomed. Eng. 5, 951–967 (2021).
Mitragotri, S., Burke, P. A. & Langer, R. Overcoming the challenges in administering biopharmaceuticals: formulation and delivery strategies. Nat. Rev. Drug Discov. 13, 655–722 (2014).
Ekladious, I., Colso, Y. L. & Grinstaff, M. W. Polymer-drug conjugate therapeutics: advances, insights and prospects. Nat. Rev. Drug Discov. 18, 273–294 (2019).
Terstappen, G. C., Meyer, A. H., Bell, R. D. & Zhang, W. Strategies for delivering therapeutics across the blood-brain barrier. Nat. Rev. Drug Discov. 20, 362–383 (2021).
Yu, Y. J. et al. Boosting brain uptake of a therapeutic antibody by reducing its affinity for a transcytosis target. Sci. Transl. Med. 3, 84ra44 (2011).
Niewoehner, J. et al. Increased brain penetration and potency of a therapeutic antibody using a monovalent molecular shuttle. Neuron 8, 49–60 (2014).
Janda, C. Y. et al. Surrogate Wnt agonists that phenocopy canonical Wnt and β-catenin signalling. Nature 545, 234–237 (2017).
Yen, M. et al. Facile discovery of surrogate cytokine agonists. Cell 185, 1414–1430.e19 (2022).
Valeur, E. et al. New modalities for challenging targets in drug discovery. Angew. Chem. Int. Ed. Engl. 56, 10294–10323 (2017).
Dougherty, P. G., Qian, Z. & Pei, D. Macrocycles as protein-protein interaction inhibitors. Biochem. J. 474, 1109–1125 (2017).
Villar, E. A. et al. How proteins bind macrocycles. Nat. Chem. Biol. 10, 723–731 (2014).
Otero-Ramirez, M. E., Passioura, T. & Suga, H. Structural features and binding modes of thioether-cyclized peptide ligands. Biomedicines 6, 116 (2018).
Sakai, K. et al. Macrocyclic peptide-based inhibition and imaging of hepatocyte growth factor. Nat. Chem. Biol. 15, 598–606 (2019).
Huang, Y., Margarete, M. & Suga, H. RNA display methods for the discovery of bioactive macrocycles. Chem. Rev. 119, 10360–10391 (2019).
Kale, S. S. et al. Cyclization of peptides with two chemical bridges affords large scaffold diversities. Nat. Chem. 10, 715–723 (2018).
Sohrabi, C., Foster, A. & Tavassoli, A. Methods for generating and screening libraries of genetically encoded cyclic peptides in drug discovery. Nat. Rev. Chem. 4, 90–101 (2020).
Wang, W., Khojasteh, S. C. & Su, D. Biosynthetic strategies for macrocyclic peptides. Molecules 26, 3338 (2021).
Traxlmayr, M. W. et al. Integrin binding human antibody constant domains—probing the C-terminal structural loops for grafting the RGD motif. J. Biotechnol. 155, 193–202 (2011).
Zoller, F. et al. Combination of phage display and molecular grafting generates highly specific tumor-targeting miniproteins. Angew. Chem. Int. Ed. Engl. 51, 13136–13139 (2012).
Ji, Y. et al. In vivo activation of the p53 tumor suppressor pathway by an engineered cyclotide. J. Am. Chem. Soc. 135, 11623–11633 (2013).
Wang, C. K. & Craik, D. J. Designing macrocyclic disulfide-rich peptides for biotechnological applications. Nat. Chem. Biol. 14, 417–427 (2018).
Wang, C. K. & Craik, D. J. Linking molecular evolution to molecular grafting. J. Biol. Chem. 296, 100425 (2021).
Mihara, E. et al. Lasso-grafting of macrocyclic peptide pharmacophores yields multi-functional proteins. Nat. Commun. 12, 1543 (2021).
Komatsu, Y. et al. De novo peptide grafting to a self-assembling nanocapsule yields a hepatocyte growth factor receptor agonist. iScience 24, 103302 (2021).
Sugano-Nakamura, N. et al. De novo Fc-based receptor dimerizers differentially modulate PlexinB1 function. Structure 30, 1367–462 (2022).
McAllister, T. E., Coleman, O. D., Roper, G. & Kawamura, A. Structural diversity in de novo cyclic peptide ligands from genetically encoded library technologies. Pept. Sci. 113, e24204 (2021).
Lobner, E., Traxlmayr, M. W., Obinger, C. & Hasenhindl, C. Engineered IgG1-Fc – one fragment to bind them all. Immunol. Rev. 270, 113–131 (2016).
Kariolis, M. S. et al. Brain delivery of therapeutic proteins using an Fc fragment blood-brain barrier transport vehicle in mice and monkeys. Sci. Transl. Med. 12, eaay1359 (2020).
Trusolino, L., Bertotti, A. & Comoglio, P. M. MET signalling: principles and functions in development, organ regeneration and cancer. Nat. Rev. Mol. Cell Biol. 11, 834–848 (2010).
Sakai, K., Aoki, S. & Matsumoto, K. Hepatocyte growth factor and Met in drug discovery. J. Biochem. 157, 271–284 (2015).
Desole, C. et al. HGF and MET: from brain development to neurological disorders. Front. Cell Dev. Biol. 9, 683609 (2021).
Hirano, S. et al. A phase I/II exploratory clinical trial for intracordal injection of recombinant hepatocyte growth factor for vocal fold scar and sulcus. J. Tissue Eng. Regen. Med. 12, 1031–1038 (2018).
Nagoshi, N. et al. Phase I/II study of intrathecal administration of recombinant human hepatocyte growth factor in patients with acute spinal cord injury: a double-blind, randomized clinical trial of safety and efficacy. J. Neurotrauma 37, 1752–1758 (2020).
Ido, A. et al. Safety and pharmacokinetics of recombinant human hepatocyte growth factor (rh-HGF) in patients with fulminant hepatitis: a phase I/II clinical trial, following preclinical studies to ensure safety. J. Transl. Med. 9, 55 (2011).
Sugiura, T. et al. Pharmacokinetic modeling of hepatocyte growth factor in experimental animals and humans. J. Pharm. Sci. 102, 237–249 (2013).
Ito, K. et al. Artificial human Met agonists based on macrocycle scaffolds. Nat. Commun. 6, 6373 (2015).
Stamos, J. et al. Crystal structure of the HGF beta-chain in complex with the Sema domain of the Met receptor. EMBO J. 23, 2325–2335 (2004).
Ando, T., Uchihashi, T. & Scheuring, S. Filming biomolecular processes by high-speed atomic force microscopy. Chem. Rev. 114, 3120–3188 (2014).
Roopenian, D. C. & Akilesh, S. FcRn: the neonatal Fc receptor comes of age. Nat. Rev. Immunol. 7, 715–725 (2007).
Sockolosky, J. T. & Szoka, F. C. The neonatal Fc receptor, FcRn, as a target for drug delivery and therapy. Adv. Drug Deliv. Rev. 91, 109–124 (2015).
Oganesyan, V. et al. Structural insights into neonatal Fc receptor-based recycling mechanisms. J. Biol. Chem. 289, 7812–7824 (2014).
Datta-Mannan, A., Witcher, D. R., Tang, Y., Watkins, J. & Wroblewski, V. J. Monoclonal antibody clearance. Impact of modulating the interaction of IgG with the neonatal Fc receptor. J. Biol. Chem. 28, 1709–1717 (2007).
Petkova, S. B. et al. Enhanced half-life of genetically engineered human IgG1 antibodies in a humanized FcRn mouse model: potential application in humorally mediated autoimmune disease. Int. Immunol. 18, 1759–1769 (2006).
Tateno, C. et al. Generation of novel chimeric mice with humanized livers by using hemizygous cDNA-uPA/SCID mice. PLoS ONE 10, e0142145 (2015).
Boado, R. J., Zhang, Y., Wang, Y. & Pardridge, W. M. Engineering and expression of a chimeric transferrin receptor monoclonal antibody for blood-brain barrier delivery in the mouse. Biotechnol. Bioeng. 102, 1251–1258 (2009).
Chang, H. Y. et al. Brain pharmacokinetics of anti-transferrin receptor antibody affinity variants in rats determined using microdialysis. MAbs 13, 1874121 (2021).
Clarke, E. et al. A Single Domain Shark Antibody Targeting the Transferrin Receptor 1 Delivers a TrkB Agonist Antibody to the Brain and Provides Full Neuroprotection in a Mouse Model of Parkinson’s Disease. Pharmaceutics 14, 1335 (2022).
De Groot, A. S. et al. Activation of natural regulatory T cells by IgG Fc-derived peptide ‘Tregitopes’. Blood 112, 3303–3311 (2006).
Li, N. et al. Therapeutic effect of HGF on NASH mice through HGF/c-Met and JAK2-STAT3 signalling pathway. Ann. Hepatol. 17, 501–510 (2018).
Yang, Y. M. et al. Interventional potential of recombinant feline hepatocyte growth factor in a mouse model of non-alcoholic steatohepatitis. Front. Endocrinol. 9, 378 (2018).
Mirdita, M. et al. ColabFold - making protein folding accessible to all. Nat Methods 19, 679–682 (2022).
Acknowledgements
This work was supported by: World Premier International Research Center Initiative (WPI), MEXT, Japan; a Grant-in-Aid for JSPS Scientific Research (C) (20K06553) to K.S.; a Grant-in-Aid for JSPS Scientific Research (B) (19H03499) and Challenging Research (21K18250) to K.M.; Program for Basic and Clinical Research on Hepatitis from the Japan Agency for Medical Research and Development (AMED) (JP21fk0210087) to K.S.; AMED Platform Project for Supporting Drug Discovery and Life Science Research (Basis for Supporting Innovative Drug Discovery and Life Science Research) to H.S. (JP19am0101090) and J.T. (JP19am0101075); and a Grant-in-Aid for JSPS Scientific Research (B) (21H01771) to M.S. This work was performed under the Cooperative Research Program of the Institute for Protein Research, Osaka University (CR15-05), and an Extramural Collaborative Research Grant from the Cancer Research Institute (Kanazawa University). We thank Dolphin Co. Ltd for the English language review.
Author information
Authors and Affiliations
Contributions
K.S., H. Suga, J.T. and K.M. conceived and designed the study. K.S. performed expression and purification of proteins, cellular assays, SPR for FcRn, crosslink, RNA-seq, serum half-life experiments and mTfR-binding assay. N.S-N. performed flow cytometry and SPR for Met. E.M. performed a preliminary assessment of Fc grafts and their activities. S.W. performed protein expression and purification, and the mapping of the binding site via the chimeric Met experiments. N.M.R.-C. performed BBB penetration experiments. I.S. performed ds-variants. H. Sato performed immunohistochemistry and contributed to animal experiments. R.I. contributed to animal experiments. D.C.-C.V. contributed to animal experiments and critically revised the manuscript. C.Y. and C.T. performed BrdU labelling in PXB mice. M.S. performed HS-AFM. The manuscript was written by K.S. All authors discussed the results and commented on the manuscript.
Corresponding authors
Ethics declarations
Competing interests
H. Suga and J.T. are co-founders and shareholders of MiraBiologics Inc. E.M., K.S. and K.M. are also shareholders of the same company. The other authors declare no competing interests.
Peer review
Peer review information
Nature Biomedical Engineering thanks Richard Siegel, Birgit Schoeberl, Manuel Sanchez-Felix 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
Extended Data Fig. 1 Expression of peptide-grafted Fc variants.
Conditioned medium of Expi293F suspension culture expressing control Fc or Fc variants were pulled down using protein A beads and analyzed by SDS-PAGE under non-reducing conditions and stained with Coomassie brilliant blue. The image for Fc(aMD4) was reproduced from our previous report27.
Extended Data Fig. 2 Binding of Fc or Fc(aMD4) to human MetECD.
Binding sensorgrams of Fc or Fc(aMD4) to human MetECD as determined by surface plasmon resonance. Measured curves are shown in red. Fitted curves are shown in black. RU, resonance units; SPR, surface plasmon resonance.
Extended Data Fig. 3 Met tyrosine phosphorylations (a) and induction kinetics (b, c) were comparable between Fc(aMD4)B3 and HGF.
a, Phosphorylation levels of Met tyrosine residues in EHMES-1 cells treated with HGF or Fc(aMD4)B3 for 10 min. The lysates were analyzed by Western blotting. b-c, EHMES-1 cells were treated with HGF (b) or Fc(aMD4)B3 (c). Activation of Met, Akt, or Erk1/2 at each time point after stimulation was quantified in situ with anti-phosphorylated Met (Tyr1234/1235), anti-phosphorylated Akt (S473), or anti-phosphorylated Erk1/2 (T202/Y204) antibody, respectively. The results are shown as the mean ± S.E.M. of fold-induction relative to non-stimulated controls (n = 4, independent experiments).
Extended Data Fig. 4 Mapping of Fc(aMD4)B3-binding sites on human MetECD.
a, The association between Fc(aMD4)B3 and human MetECD, mouse MetECD, or chimeric MetECD that variably fused human and mouse was evaluated by pull-down assay. PA-tagged MetECD was precipitated using anti-PA tag antibody conjugated beads. Consistent with our previous report that aMD4 peptide binds to human but not mouse MetECD40, Fc(aMD4)B3 bound to human MetECD (hWT) but not to mouse MetECD (mWT). Replacement of the PSI domain and the IPT domains to mouse sequence (variants A–E) preserved the binding to Fc(aMD4)B3, indicating the PSI domain and the IPT domains were dispensable for Fc(aMD4)B3 binding. In contrast, human-specific residues in blades 5–7 of the Sema domain were required for Fc(aMD4)B3 binding (variant F). The numbers in the Sema domain indicate each blade. b, Human-specific residues in blade 6 (B6-1 and B6-2 in d) were indispensable for Fc(aMD4)B3 binding. c, Human-specific residues in blade 5 (B5-1 and B5-2 in d) were also indispensable or contributing for Fc(aMD4)B3 binding, respectively. d, Amino acid sequence alignment of human and mouse Met Sema domains. Residues that are different between human and mouse are indicated in bold. The red bars (B5-1, B6-1, and B6-2) indicate the residues indispensable for Fc(aMD4)B3 binding. The orange bar (B5-2) indicates the residues that contribute to Fc(aMD4)B3 binding.
Extended Data Fig. 5 Analysis of BS3-stabilized 2:1 complex of MetECD and Fc(aMD4)B3.
a, Molecular weight analyzed by SDS-PAGE indicated 2:1 and 1:1 complex of MetECD and Fc(aMD4)B3. SDS-PAGE analysis was performed using 7.5% gels (blue, from Fig. 3b) and 3-10% gels (orange, from b). b, SDS-PAGE analysis of Met dimer formation by Fc(aMD4) in vitro. The complexes between MetECD and Fc(aMD4) or control Fc were crosslinked by BS3 and analyzed by SDS-PAGE (3-10% gel) under non-reducing conditions and silver stained. c, MetECD and Fc(aMD4)B3 were crosslinked by BS3 and subjected to size exclusion chromatography (SEC). The results of non-reducing SDS-PAGE of the SEC fractions with Coomassie brilliant blue staining are shown. The fraction indicated in red was subjected to HS-AFM as shown in Fig. 3e, f.
Extended Data Fig. 6 Binding of Fc or Fc(aMD4) to FcRn/β2 M for Fig. 4b.
a, Binding sensorgrams of Fc to FcRn/β2 M at pH 6.0 and 7.4. Ass, addition of Fc; Diss, buffer wash. b, Binding sensorgrams of Fc or Fc(aMD4) to FcRn/β2 M at pH 6.0. Sensorgrams show duplicates of three concentrations of analytes (250, 500, 1,000 nM). RU, resonance units; SPR, surface plasmon resonance.
Supplementary information
Supplementary Information
Supplementary figures.
Supplementary Video 1
HS-AFM analysis of Fc(aMD4)B3.
Supplementary Video 2
HS-AFM analysis of MetECD.
Supplementary Video 3
HS-AFM analysis of a 2:1 complex of MetECD and Fc(aMD4)B3 with the IPT stalk domains sprayed away.
Supplementary Video 4
HS-AFM analysis of a 2:1 complex of MetECD and Fc(aMD4)B3 with the IPT stalk domains attached.
Supplementary Dataset 1
Source data for Supplementary Fig. 3.
Supplementary Dataset 2
Source data for Supplementary Fig. 6.
Source data
Source Data For Fig. 1
Source data.
Source Data For Fig. 2
Source data.
Source Data For Fig. 4
Source data.
Source Data For Fig. 5
Source data.
Source Data For Extended Data Fig. 3
Source data.
Source Data For Extended Data Fig. 5
Source data.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, 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 license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license 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 license, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Sakai, K., Sugano-Nakamura, N., Mihara, E. et al. Designing receptor agonists with enhanced pharmacokinetics by grafting macrocyclic peptides into fragment crystallizable regions. Nat. Biomed. Eng 7, 164–176 (2023). https://doi.org/10.1038/s41551-022-00955-6
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1038/s41551-022-00955-6
This article is cited by
-
Considerations for the genotoxicity assessment of middle size peptide drugs containing non-canonical amino acid residues
Genes and Environment (2023)
-
Lasso-grafted designer cytokines
Nature Biomedical Engineering (2022)