A highly potent CD73 biparatopic antibody blocks organization of the enzyme active site through dual mechanisms
2020; Elsevier BV; Volume: 295; Issue: 52 Linguagem: Inglês
10.1074/jbc.ra120.012395
ISSN1083-351X
AutoresJames E. Stefano, Dana M. Lord, Yanfeng Zhou, Julie Jaworski, Joern Hopke, Tara Travaline, Ningning Zhang, Karen Wong, Amanda Lennon, Timothy He, Eva Bric-Furlong, Cornishia Cherrie, Tristan Magnay, Élisabeth Rémy, William Brondyk, Huawei Qiu, Katarina Radošević,
Tópico(s)RNA Interference and Gene Delivery
ResumoThe dimeric ectonucleotidase CD73 catalyzes the hydrolysis of AMP at the cell surface to form adenosine, a potent suppressor of the immune response. Blocking CD73 activity in the tumor microenvironment can have a beneficial effect on tumor eradication and is a promising approach for cancer therapy. Biparatopic antibodies binding different regions of CD73 may be a means to antagonize its enzymatic activity. A panel of biparatopic antibodies representing the pairwise combination of 11 parental monoclonal antibodies against CD73 was generated by Fab-arm exchange. Nine variants vastly exceeded the potency of their parental antibodies with ≥90% inhibition of activity and subnanomolar EC50 values. Pairing the Fabs of parents with nonoverlapping epitopes was both sufficient and necessary whereas monovalent antibodies were poor inhibitors. Some parental antibodies yielded potent biparatopics with multiple partners, one of which (TB19) producing the most potent. The structure of the TB19 Fab with CD73 reveals that it blocks alignment of the N- and C-terminal CD73 domains necessary for catalysis. A separate structure of CD73 with a Fab (TB38) which complements TB19 in a particularly potent biparatopic shows its binding to a nonoverlapping site on the CD73 N-terminal domain. Structural modeling demonstrates a TB19/TB38 biparatopic antibody would be unable to bind the CD73 dimer in a bivalent manner, implicating crosslinking of separate CD73 dimers in its mechanism of action. This ability of a biparatopic antibody to both crosslink CD73 dimers and fix them in an inactive conformation thus represents a highly effective mechanism for the inhibition of CD73 activity. The dimeric ectonucleotidase CD73 catalyzes the hydrolysis of AMP at the cell surface to form adenosine, a potent suppressor of the immune response. Blocking CD73 activity in the tumor microenvironment can have a beneficial effect on tumor eradication and is a promising approach for cancer therapy. Biparatopic antibodies binding different regions of CD73 may be a means to antagonize its enzymatic activity. A panel of biparatopic antibodies representing the pairwise combination of 11 parental monoclonal antibodies against CD73 was generated by Fab-arm exchange. Nine variants vastly exceeded the potency of their parental antibodies with ≥90% inhibition of activity and subnanomolar EC50 values. Pairing the Fabs of parents with nonoverlapping epitopes was both sufficient and necessary whereas monovalent antibodies were poor inhibitors. Some parental antibodies yielded potent biparatopics with multiple partners, one of which (TB19) producing the most potent. The structure of the TB19 Fab with CD73 reveals that it blocks alignment of the N- and C-terminal CD73 domains necessary for catalysis. A separate structure of CD73 with a Fab (TB38) which complements TB19 in a particularly potent biparatopic shows its binding to a nonoverlapping site on the CD73 N-terminal domain. Structural modeling demonstrates a TB19/TB38 biparatopic antibody would be unable to bind the CD73 dimer in a bivalent manner, implicating crosslinking of separate CD73 dimers in its mechanism of action. This ability of a biparatopic antibody to both crosslink CD73 dimers and fix them in an inactive conformation thus represents a highly effective mechanism for the inhibition of CD73 activity. CD73 (ecto-5′-nucleotidase, NT5E) is a glycosylated 125-kDa homodimeric membrane-bound enzyme which dephosphorylates AMP in the extracellular milieu to adenosine (1Allard D. Allard B. Gaudreau P.O. Chrobak P. Stagg J. CD73-adenosine: A next-generation target in immuno-oncology.Immunotherapy. 2016; 8 (26808918): 145-16310.2217/imt.15.106Crossref PubMed Scopus (94) Google Scholar). Adenosine has potent immunosuppressive effects in the tumor microenvironment so CD73 has attracted wide interest as a target for cancer therapy (1Allard D. Allard B. Gaudreau P.O. Chrobak P. Stagg J. CD73-adenosine: A next-generation target in immuno-oncology.Immunotherapy. 2016; 8 (26808918): 145-16310.2217/imt.15.106Crossref PubMed Scopus (94) Google Scholar, 2Allard B. Longhi M.S. Robson S.C. Stagg J. The ectonucleotidases CD39 and CD73: Novel checkpoint inhibitor targets.Immunol. Rev. 2017; 276 (28258700): 121-14410.1111/imr.12528Crossref PubMed Scopus (490) Google Scholar, 3Allard D. Chrobak P. Allard B. Messaoudi N. Stagg J. Targeting the CD73-adenosine axis in immuno-oncology.Immunol. Lett. 2019; 205 (29758241): 31-3910.1016/j.imlet.2018.05.001Crossref PubMed Scopus (91) Google Scholar, 4Kats L. Darcy P.K. Beavis P.A. Bahreyni A. Rezaei M. Khazaei M. Fuiji H. Ferns G.A. Ryzhikov M. Avan A. Hassanian S.M. The potential role of adenosine signaling in the pathogenesis of melanoma.Int. J. Mol. Sci. 2018; 156 (30232037): 451-45710.1016/j.bcp.2018.09.018Google Scholar, 5Sek K. Mølck C. Stewart G.D. Targeting adenosine receptor signaling in cancer immunotherapy. 2018; 19 (30513816)383710.3390/ijms19123837Google Scholar, 6Yang J. Liao X. Yu J. Zhou P. Role of CD73 in disease: Promising prognostic indicator and therapeutic target.Curr. Med. Chem. 2018; 25 (29345574): 2260-227110.2174/0929867325666180117101114Crossref PubMed Scopus (18) Google Scholar). CD73 expression is associated with resistance to anti-HER2 therapy (7Turcotte M. Allard D. Mittal D. Bareche Y. Buisseret L. José V. Pommey S. Delisle V. Loi S. Joensuu H. Kellokumpu-Lehtinen P.L. Sotiriou C. Smyth M.J. Stagg J. CD73 promotes resistance to HER2/ErbB2 antibody therapy.Cancer Res. 2017; 77 (28855210): 5652-566310.1158/0008-5472.CAN-17-0707Crossref PubMed Scopus (73) Google Scholar), poor prognosis with reduced anti-tumor immune response in a variety of tumor types (1Allard D. Allard B. Gaudreau P.O. Chrobak P. Stagg J. CD73-adenosine: A next-generation target in immuno-oncology.Immunotherapy. 2016; 8 (26808918): 145-16310.2217/imt.15.106Crossref PubMed Scopus (94) Google Scholar), and the increased growth of tumor cells, migration, and invasion in vitro (8Zhi X. Chen S. Zhou P. Shao Z. Wang L. Ou Z. Yin L. RNA interference of ecto-5′-nucleotidase (CD73) inhibits human breast cancer cell growth and invasion.Clin. Exp. Metastasis. 2007; 24 (17587186): 439-44810.1007/s10585-007-9081-yCrossref PubMed Scopus (125) Google Scholar). A number of clinical studies are in progress with CD73-specific antibodies (9Siu L.L. Burris H. Le D.T. Hollebecque A. Steeghs N. Delord J.-P. Hilton J. Barnhart B. Sega E. Sanghavi K. Klippel A. Hedvat C. Hilt E. Donovan M. Gipson A. et al.Abstract CT180: Preliminary phase 1 profile of BMS-986179, an anti-CD73 antibody, in combination with nivolumab in patients with advanced solid tumors.Cancer Res. 2018; 78: CT18010.1158/1538-7445.AM2018-CT180Google Scholar, 10Overman M.J. LoRusso P. Strickler J.H. Patel S.P. Clarke S.J. Noonan A.M. Prasanna T. Amin M.A. Nemunaitis J.J. Desai J. O'Byrne K.J. George T.J. Englert J. She D. Cooper Z.A. et al.Safety, efficacy and pharmacodynamics (PD) of MEDI9447 (oleclumab) alone or in combination with durvalumab in advanced colorectal cancer (CRC) or pancreatic cancer (panc.).J. Clin. Oncol. 2018; 36412310.1200/JCO.2018.36.15_suppl.4123Crossref Google Scholar) and small molecule inhibitors (10Overman M.J. LoRusso P. Strickler J.H. Patel S.P. Clarke S.J. Noonan A.M. Prasanna T. Amin M.A. Nemunaitis J.J. Desai J. O'Byrne K.J. George T.J. Englert J. She D. Cooper Z.A. et al.Safety, efficacy and pharmacodynamics (PD) of MEDI9447 (oleclumab) alone or in combination with durvalumab in advanced colorectal cancer (CRC) or pancreatic cancer (panc.).J. Clin. Oncol. 2018; 36412310.1200/JCO.2018.36.15_suppl.4123Crossref Google Scholar, 11Seitz L. Jin L. Leleti M. Ashok D. Jeffrey J. Rieger A. Tiessen R.G. Arold G. Tan J.B.L. Powers J.P. Walters M.J. Karakunnel J. Safety, tolerability, and pharmacology of AB928, a novel dual adenosine receptor antagonist, in a randomized, phase 1 study in healthy volunteers.Invest. New Drugs. 2019; 37 (30569245): 711-72110.1007/s10637-018-0706-6Crossref PubMed Scopus (30) Google Scholar), alone or in combination with A2a adenosine receptor antagonists and antibodies to other targets, particularly the PD-1/PD-L1 axis (12Leone R.D. Emens L.A. Targeting adenosine for cancer immunotherapy.J. Immunother. Cancer. 2018; 6 (29914571): 5710.1186/s40425-018-0360-8Crossref PubMed Scopus (282) Google Scholar). MEDI9447 (oleclumab), a CD73-specific internalizing antibody with moderate inhibition of enzymatic activity, has shown some clinical efficacy as a monotherapy and in combination with the PD-L1 blocker durvalumab (13Hay C.M. Sult E. Huang Q. Mulgrew K. Fuhrmann S.R. McGlinchey K.A. Hammond S.A. Rothstein R. Rios-Doria J. Poon E. Holoweckyj N. Durham N.M. Leow C.C. Diedrich G. Damschroder M. et al.Targeting CD73 in the tumor microenvironment with MEDI9447.Oncoimmunology. 2016; 5 (27622077)e120887510.1080/2162402X.2016.1208875Crossref PubMed Scopus (188) Google Scholar). There are also indications that CD73 antibodies can exert effects independent of adenosine production. One study indicated that the enhancement of the immune response was mediated through FcγRIV engagement in mice (14Vijayan D. Barkauskas D.S. Stannard K. Sult E. Buonpane R. Takeda K. Teng M.W.L. Sachsenmeier K. Hay C. Smyth M.J. Selective activation of anti-CD73 mechanisms in control of primary tumors and metastases.Oncoimmunology. 2017; 6 (28638737)e131204410.1080/2162402X.2017.1312044Crossref PubMed Scopus (17) Google Scholar) and other work suggested a role for CD73 internalization at suppressing metastasis (10Overman M.J. LoRusso P. Strickler J.H. Patel S.P. Clarke S.J. Noonan A.M. Prasanna T. Amin M.A. Nemunaitis J.J. Desai J. O'Byrne K.J. George T.J. Englert J. She D. Cooper Z.A. et al.Safety, efficacy and pharmacodynamics (PD) of MEDI9447 (oleclumab) alone or in combination with durvalumab in advanced colorectal cancer (CRC) or pancreatic cancer (panc.).J. Clin. Oncol. 2018; 36412310.1200/JCO.2018.36.15_suppl.4123Crossref Google Scholar, 13Hay C.M. Sult E. Huang Q. Mulgrew K. Fuhrmann S.R. McGlinchey K.A. Hammond S.A. Rothstein R. Rios-Doria J. Poon E. Holoweckyj N. Durham N.M. Leow C.C. Diedrich G. Damschroder M. et al.Targeting CD73 in the tumor microenvironment with MEDI9447.Oncoimmunology. 2016; 5 (27622077)e120887510.1080/2162402X.2016.1208875Crossref PubMed Scopus (188) Google Scholar, 15Terp M.G. Olesen K.A. Arnspang E.C. Lund R.R. Lagerholm B.C. Ditzel H.J. Leth-Larsen R. Anti-human CD73 monoclonal antibody inhibits metastasis formation in human breast cancer by inducing clustering and internalization of CD73 expressed on the surface of cancer cells.J. Immunol. 2013; 191 (24043904): 4165-417310.4049/jimmunol.1301274Crossref PubMed Scopus (96) Google Scholar). Nonetheless, adenosine levels in tumors can reach micromolar concentrations, so incomplete inhibition of CD73 activity may be a limiting factor for the efficacy of current CD73-targeting therapeutics (16Blay J. White T.D. Hoskin D.W. The extracellular fluid of solid carcinomas contains immunosuppressive concentrations of adenosine.Cancer Res. 1997; 57 (9205063): 2602-2605PubMed Google Scholar). Thus, the mechanism by which CD73 affects cancer progression may be complex, suggesting the need for very potent inhibition of enzymatic activity or a combination of mechanisms to achieve optimal efficacy. The CD73 monomer, with N- and C-terminal domains that are connected through a flexible α-helical linker, is expressed at the cell surface attached to C-terminal glycosylphosphatidylinositol anchor. In the physiological form two monomers associate through extensive noncovalent contacts between the C-terminal domains forming a dimer (17Heuts D.P. Weissenborn M.J. Olkhov R.V. Shaw A.M. Gummadova J. Levy C. Scrutton N.S. Crystal structure of a soluble form of human CD73 with ecto-5′-nucleotidase activity.Chembiochem. 2012; 13 (22997138): 2384-239110.1002/cbic.201200426Crossref PubMed Scopus (54) Google Scholar, 18Knapp K. Zebisch M. Pippel J. El-Tayeb A. Müller C.E. Sträter N. Crystal structure of the human ecto-5′-nucleotidase (CD73): Insights into the regulation of purinergic signaling.Structure. 2012; 20 (23142347): 2161-217310.1016/j.str.2012.10.001Abstract Full Text Full Text PDF PubMed Scopus (107) Google Scholar). The active site in each monomer of CD73 is comprised of substrate contact residues in both the N- and C-terminal domains in addition to zinc cofactors bound by the N-terminal domain (18Knapp K. Zebisch M. Pippel J. El-Tayeb A. Müller C.E. Sträter N. Crystal structure of the human ecto-5′-nucleotidase (CD73): Insights into the regulation of purinergic signaling.Structure. 2012; 20 (23142347): 2161-217310.1016/j.str.2012.10.001Abstract Full Text Full Text PDF PubMed Scopus (107) Google Scholar). Following binding of the AMP substrate to the C-terminal domain, the N-terminal domain and zinc cofactors align with the AMP in a "closed" CD73 conformation in which catalysis takes place to generate the adenosine product (19Knapp K. Zebisch M. Pippel J. El-Tayeb A. Müller C.E. Sträter N. Movie S1. Crystal structure of the human ecto-59-nucleotidase (CD73): Insights into the regulation of purinergic signaling. Morph illustrating the conformational change between the open and closed forms of ecto-5′-nucleotidase.Structure. 2012; 20 ([movie]) (23142347): 2161-217310.1016/j.str.2012.10.001Abstract Full Text Full Text PDF PubMed Scopus (133) Google Scholar). A large lateral rotation of the N-terminal domain to re-expose the substrate binding site in the "open" conformer then allows product release (18Knapp K. Zebisch M. Pippel J. El-Tayeb A. Müller C.E. Sträter N. Crystal structure of the human ecto-5′-nucleotidase (CD73): Insights into the regulation of purinergic signaling.Structure. 2012; 20 (23142347): 2161-217310.1016/j.str.2012.10.001Abstract Full Text Full Text PDF PubMed Scopus (107) Google Scholar). A limited solvent access to the active site in the closed conformer indicates that cycling between the two forms is required for substrate binding and product release, i.e. efficient enzymatic activity (18Knapp K. Zebisch M. Pippel J. El-Tayeb A. Müller C.E. Sträter N. Crystal structure of the human ecto-5′-nucleotidase (CD73): Insights into the regulation of purinergic signaling.Structure. 2012; 20 (23142347): 2161-217310.1016/j.str.2012.10.001Abstract Full Text Full Text PDF PubMed Scopus (107) Google Scholar). In our hands, obtaining potent inhibition of CD73 enzymatic activity (e.g. both a high percentage inhibition and a low EC50) with monospecific CD73 antibodies proved challenging, an experience apparently shared by others (20Geoghegan J.C. Diedrich G. Lu X. Rosenthal K. Sachsenmeier K.F. Wu H. Dall'Acqua W.F. Damschroder M.M. Inhibition of CD73 AMP hydrolysis by a therapeutic antibody with a dual, non-competitive mechanism of action.mAbs. 2016; 8 (26854859): 454-46710.1080/19420862.2016.1143182Crossref PubMed Scopus (69) Google Scholar, 21Perrot, I., Paturel, C., and Gauthier, L., inventors (2016) CD73 Blockade. WO2016055609 (A1).Google Scholar, 22Ditzel, H., and Gammelgaard, O. L., inventors (2017) Bispecific antibodies targeting human CD73. WO2017118613 (A1).Google Scholar). We decided to examine the possible advantage of a biparatopic antibody approach because of the potentially additive effect of combining antibody specificities. To identify biologically relevant mechanisms of inhibition, we assayed CD73 activity on cells using a highly sensitive LC-MS–based method (23McManus J. He T. Gavigan J.A. Marchand G. Vougier S. Bedel O. Ferrari P. Arrebola R. Gillespy T. Gregory R.C. Licht S. Cheng H. Zhang B. Deng G. A robust multiplex mass spectrometric assay for screening small-molecule inhibitors of CD73 with diverse inhibition modalities.SLAS Discov. 2018; 23 (29336194): 264-27310.1177/2472555217750386PubMed Google Scholar). Our results demonstrate that antibodies can exhibit potent CD73 inhibition when combined in biparatopic variants provided they bind nonoverlapping epitopes on CD73. As a result of this investigation, we discovered one antibody (TB19) that was able to synergize with half of the antibodies to form highly potent biparatopic variants. The structures of CD73 in the complexes with both TB19 and a partner TB38, which combine to form a particularly potent biparatopic, show CD73 in conformations not reported previously and provide further insights into the catalytic mechanism. Our analyses show that the activity of this potent biparatopic variant is provided by a dual mechanism of directly blocking formation of the catalytically active conformer in a complex stabilized by interactions with other CD73. We generated a panel of biparatopic antibodies against CD73 using Fab-arm exchange (cFAE) representing the pairwise combinations of 11 parental antibodies unrelated by sequence and previously showing >50% inhibition of CD73 activity in cell-based assays. Each Fab was expressed as a fusion with human IgG1 Fc containing either the F405L or K409R mutation, which destabilize the parental Fc and stabilize the Fc of the biparatopic duobody product (24Gramer M.J. van den Bremer E.T. van Kampen M.D. Kundu A. Kopfmann P. Etter E. Stinehelfer D. Long J. Lannom T. Noordergraaf E.H. Gerritsen J. Labrijn A.F. Schuurman J. van Berkel P.H. Parren P.W. Production of stable bispecific IgG1 by controlled Fab-arm exchange: Scalability from bench to large-scale manufacturing by application of standard approaches.mAbs. 2013; 5 (23995617): 962-97310.4161/mabs.26233Crossref PubMed Scopus (48) Google Scholar, 25Labrijn A.F. Meesters J.I. Priem P. de Jong R.N. van den Bremer E.T.J. van Kampen M.D. Gerritsen A.F. Schuurman J. Parren P.W.H.I. Controlled Fab-arm exchange for the generation of stable bispecific IgG1.Nat. Protoc. 2014; 9 (25255089): 2450-246310.1038/nprot.2014.169Crossref PubMed Scopus (75) Google Scholar, 26Labrijn A.F. Meesters J.I. de Goeij B.E. van den Bremer E.T. Neijssen J. van Kampen M.D. Strumane K. Verploegen S. Kundu A. Gramer M.J. van Berkel P.H. van de Winkel J.G. Schuurman J. Parren P.W. Efficient generation of stable bispecific IgG1 by controlled Fab-arm exchange.Proc. Natl. Acad. Sci. U. S. A. 2013; 110 (23479652): 5145-515010.1073/pnas.1220145110Crossref PubMed Scopus (220) Google Scholar). Parental antibodies were expressed in small-scale cultures, purified using protein A, and recombined by Fab arm exchange (26Labrijn A.F. Meesters J.I. de Goeij B.E. van den Bremer E.T. Neijssen J. van Kampen M.D. Strumane K. Verploegen S. Kundu A. Gramer M.J. van Berkel P.H. van de Winkel J.G. Schuurman J. Parren P.W. Efficient generation of stable bispecific IgG1 by controlled Fab-arm exchange.Proc. Natl. Acad. Sci. U. S. A. 2013; 110 (23479652): 5145-515010.1073/pnas.1220145110Crossref PubMed Scopus (220) Google Scholar). Production of the desired products was verified by capillary isoelectric focusing (cIEF) (Fig. S1). Out of 121 (11 × 11) possible combinations, 88 biparatopic variants were generated that covered all possible combinations in at least one orientation. Eleven monospecific parental antibodies were also reconstructed as comparators by combining the parental F405L and K409R Fc variants to control for possible effect of the Fc mutations on antibody structure and function. In addition, 21 pairings were generated in both Fc orientations to control for possible positional effects of the mutations. Purified parental and biparatopic antibodies were tested for potency at 1 μg/ml on COR-L23 lung carcinoma cells expressing human CD73, and the product adenosine quantitated by a LC-MS–based assay (23McManus J. He T. Gavigan J.A. Marchand G. Vougier S. Bedel O. Ferrari P. Arrebola R. Gillespy T. Gregory R.C. Licht S. Cheng H. Zhang B. Deng G. A robust multiplex mass spectrometric assay for screening small-molecule inhibitors of CD73 with diverse inhibition modalities.SLAS Discov. 2018; 23 (29336194): 264-27310.1177/2472555217750386PubMed Google Scholar). The percentage of inhibition of CD73 enzymatic activity by the biparatopics at 1 ug/ml is shown in Fig. 1. Although the extent of inhibition varied widely, most of the biparatopic combinations exhibited higher potency than either parental antibody in the form of a duobody. A number of the parental antibodies yielded highly potent daughter biparatopic variants showing ≥90% inhibition when combined with more than one other antibody. Of these, TB19 and E3.2 formed the highest number of variants with ≥90% inhibition and several of the TB19 pairs, including those with E3.2, H19, TB38, or TC29, achieved ≥95% inhibition. The TB19 and E3.2 antibodies also combined with several other antibodies to achieve ≥80% inhibition. Although both these antibodies showed this promiscuous pairing capability, they were distinguished from each other by complementarity in their pairing patterns. No major differences in the extent of inhibition were observed between biparatopic variants tested in both Fc orientations (in total 16), indicating that the position of the duobody mutations in the Fc did not significantly influence the outcome. To assess whether both parental Fabs were necessary for potency, the parental antibodies were also crossed with an irrelevant antibody (AS30) to create monovalent variant IgGs with only a single Fab capable of interacting with CD73. All of these antibodies showed negligible potency, demonstrating that the Fabs from two cognate parentals must participate (Fig. 1). We compared the binding of several of the monovalent antibodies to the biparatopic variants of which they were a part to assess whether the higher potency of the biparatopic was because of additional interactions with CD73. Antibodies were bound to the support and binding to soluble CD73 dimer in solution determined using SPR. Antibody loading was reduced to the lowest feasible level to minimize individual CD73 dimers interacting with more than a single antibody on the support (see under "Experimental procedures"). However, 9 of the 11 biparatopics displayed biphasic dissociation kinetics (Fig. S8), although largely as a consequence of a minor fraction (≤15%) of a faster-dissociating component. In one case (H19/C16) the abundance of this component was larger and similar to that of the monovalent parent C16/AS30 (31% versus 38%), suggesting heterogeneity of the C16 monoclonal used for producing both. TA9/AS30 showed a similar heterogeneity (29% lower stability) that was not reflected in the biparatopic daughter TA9/H7. The kd values and their abundances are presented in Table S2 and the half-times for dissociation compared with the monovalent parentals are shown in Fig. S10. In 8 of 11 cases, the kd of the principal dissociation component was within 2.2-fold of the monovalent parent having the highest stability. In contrast, the kds for the monovalent parentals differed by an average of 15-fold (range 1.5- to 73-fold, median 6.2) suggesting in these cases CD73 is bound by a single parental Fab arm on the immobilized antibody. However, in three cases (E3.2/TB19, CL25/TB19, and H19/TB19) the interaction with the biparatopic was significantly more stable than with either monovalent parental (5.4-, 8.8-, and 26-fold, respectively) suggesting the presence of additional contacts with the biparatopic. Biphasic kinetics of association were also apparent from the limited period provided for binding manifested by a rapid increase in RU immediately following injection followed by a significant decline in rate after 100s. Projection of the RU expected at early times from the rate after 100s assuming pseudo first-order kinetics showed a residual consistent with a fast component binding with first-order kinetics which contributed a significant fraction to the RU (30–49%). A reiterative process to fit both components yielded a combined fit within ± 0.2 of the observed RU over 90% of the course of binding (Fig. S8). Similar to dissociation, the ka value for each of the two components was within 3-fold of a monovalent parent (2.04 ± 1.4-fold, range 1.02–2.71) in contrast to an average ∼6-fold difference between them (5.9 ± 2.1, Table S2), suggesting they reflect the independent binding of CD73 by each parental Fab arm. The ka using a Langmuir 1:1 model was within 30% of the average of the two components (Table S2). Because each component ka could not be unequivocally assigned either one for dissociation, the affinities of the biparatopic and monovalent parents for CD73 were compared using an average KD value combining the kd of the principal dissociation component with the ka based on Langmuir 1:1 binding. As for dissociation, the apparent affinity of the biparatopic variants (KD) was similar to those of the more affine monovalent parental antibodies, suggesting the interaction of the biparatopics could be largely attributed to binding of a single Fab arm. In two cases also seen by comparing complex stabilities (CL25/TB19 and TB19/H19) the biparatopic variant showed a significant increase over that of either monovalent parental (26- and 69-fold, respectively). This increase was specific to those combinations because the parents (TB19, H19, CL25) did not produce a similar enhancement with other partners. Because these increases required two cognate arms, we infer that this reflects the interaction of both arms of these two biparatopic variants with CD73, similar to the conclusion arrived at from the dissociation kinetics. However, in the majority of cases the affinity for CD73 was not increased by the addition of a second cognate Fab arm despite its being necessary for potency, suggesting interaction of the biparatopic antibody with an additional CD73 is required for potent inhibition on cells. To further evaluate the benefit of combining the parental antibodies in biparatopic format we determined the EC50 and maximum inhibition at saturating antibody concentrations for the most active biparatopics along with their parental mAbs, either alone or in a mixture on COR-L23 cells (Table 1 and Fig. S2). In agreement with the results in Fig. 1, each biparatopic was more potent than either of their two parental antibodies, which showed only partial inhibition up to 10 nm. EC50 values for all the biparatopics were in the range of 0.2–0.8 nm. In most cases, the mixtures of parental antibodies yielded similar maximal inhibition as the biparatopics, but in half of the tested combinations, the biparatopic variant in addition showed a lower EC50. In the most striking case (TB19/TC29), the biparatopic showed an EC50 at least 40-fold lower than the antibody mixture despite the similar apparent affinity for the TC29 monovalent parent and the biparatopic for CD73 (Fig. 2). Strikingly, the very high affinities obtained for a number of the biparatopics based on kds derived from the principal dissociation component seen by SPR were not replicated in low EC50s, suggesting some interactions with CD73 in solution may not be fully accessible with CD73 on the cell surface, possibly because of proximity to the membrane or differences in conformational states. In only a single case (CL25/TA10) was the mixture more potent (∼4-fold), indicating that interactions with CD73 provided by that mixture could not be replicated with the biparatopic antibody.Table 1Potency of biparatopic antibodies and parental mixtures against CD73 on COR-L23 cells. EC50 and maximum extents of inhibition are based on nonlinear regression analysis (see "Experimental procedures")BiparatopicParental MixParentalsEC50 (nm)Max. inhibitionamaximum inhibition.EC50 (nm)Max. inhibitionamaximum inhibition.TB19/TB380.777100%0.841106%bextrapolated value.H19/TB190.38298%0.62998%E3.2/TB190.44397%0.81198%CL25/TB190.61997%0.636109%bextrapolated value.H19/E3.20.22496%0.28399%TB19/TC290.26495%13.0137% bextrapolated value.H7/TB190.27095%0.54195%F1.2/E3.20.30593%0.25697%H19/C160.23993%0.86377%CL25/TA100.26691%0.07395%TA9/H70.22966%0.65880%a maximum inhibition.b extrapolated value. Open table in a new tab Epitopes of the parental antibodies with the highest number of highly potent combinations (TB19, E3.2, TB38, H19, and E3.2) were binned using biolayer interferometry (Fig. 3A). In this approach, monovalent IgG antibodies were used to coat the solid support for capturing CD73 mixed with competitor Fabs. The result of interrogating a subset of the parental antibodies is shown in Fig. 3B and Fig. S9. Higher values indicate capture of CD73 bound by the challenge Fab with no/low competition for binding (i.e. that the Fab binds to a CD73 epitope not overlapping with that of the coated antibody) whereas lower values reflect blocking of the epitope by the Fab for capture by the immobilized antibody. Allocation of the antibodies to different epitope bins based on these results is shown in Fig. 3C. One of the bins contained TB38, H19, and the mostly overlapping TC29, all of which showed susceptibility to each of the Fabs except TB19. However, these three also showed differences in their susceptibilities to competition by different Fabs. For example, the capture of CD73 by a monovalent TB38 IgG was more susceptible to competition by H19 Fab than the capture either by TC29 or H19, whereas TC29 was distinguished from the other two by its partial resistance to competition by the F1.2 Fab, which was unique among all of the antibodies. Although the bins were in most cases clearly delineated, intermediate levels of inhibition were also observed in several cases (H19+H19, TC29+H19, TC29+F1.2, TB19+H19, F1.2+H19, and F1.2+TB19), possibly reflecting partially overlapping epitopes (27Abdiche Y.N. Yeung A.Y. Ni I. Stone D. Miles A. Morishige W. Rossi A. Strop P. Antibodies targeting closely adjacent or minimally overlapping epitopes can displace one another.PLoS One. 2017; 12 (28060885)e016953510.1371/journal.pone.0169535Crossref PubMed Scopus (26) Google Scholar) and/or significant differences in affinity. E3.2 could not
Referência(s)