key: cord-0807219-5of195vi
authors: Van de Walle, Inge; Silence, Karen; Budding, Kevin; Van de Ven, Liesbeth; Dijkxhoorn, Kim; de Zeeuw, Elisabeth; Yildiz, Cafer; Gabriels, Sofie; Percier, Jean-Michel; Wildemann, Johanna; Meeldijk, Jan; Simons, Peter J.; Boon, Louis; Cox, Linda; Holgate, Rob; Urbanus, Rolf; Otten, Henny G.; Leusen, Jeanette H.W.; Blanchetot, Christophe; de Haard, Hans; Hack, C. Erik; Boross, Peter
title: ARGX-117, a therapeutic complement inhibiting antibody targeting C2
date: 2020-09-11
journal: J Allergy Clin Immunol
DOI: 10.1016/j.jaci.2020.08.028
sha: 1eb352dfa895a91c043e83796054ad87702be8f5
doc_id: 807219
cord_uid: 5of195vi

Background Activation of the classical and lectin pathway of complement may contribute to tissue damage and organ dysfunction of antibody-mediated diseases and ischemia-reperfusion conditions. Complement factors are being considered as targets for therapeutic intervention Objective We here characterize ARGX-117, a humanized inhibitory monoclonal antibody against complement C2. Methods The mode-of-action and binding characteristics of ARGX-117 were investigated in detail. Furthermore, its efficacy was analyzed in in vitro complement cytotoxicity assays. Finally, a PK/PD study was conducted in cynomolgus monkeys. Results Through binding to the Sushi-2 domain of C2, ARGX-117 prevents the formation of the C3 proconvertase, and inhibits classical and lectin pathway activation upstream of C3 activation. As ARGX-117 does not inhibit the alternative pathway it is expected not to affect the antimicrobial activity of this complement pathway. ARGX-117 prevents complement-mediated cytotoxicity in in vitro models for autoimmune hemolytic anemia and antibody-mediated rejection of organ transplants. ARGX-117 exhibits pH- and calcium-dependent target binding and is Fc-engineered to increase affinity at acidic pH to FcRn, and to reduce effector functions. In cynomolgus monkeys, ARGX-117 dose-dependently reduces free C2 levels and classical pathway activity. A two-dose regimen of 80 and 20 mg/kg separated by a week, resulted in profound reduction of classical pathway activity lasting for at least 7 weeks. Conclusion ARGX-117 is a promising new complement inhibitor that is uniquely positioned to target both the classical and lectin pathway while leaving the alternative pathway intact.

Through binding to the Sushi-2 domain of C2, ARGX-117 prevents the formation of the C3 50 proconvertase, and inhibits classical and lectin pathway activation upstream of C3 activation. 51

As ARGX-117 does not inhibit the alternative pathway it is expected not to affect the 52 antimicrobial activity of this complement pathway. ARGX-117 prevents complement-53 mediated cytotoxicity in in vitro models for autoimmune hemolytic anemia and antibody-54 mediated rejection of organ transplants. ARGX-117 exhibits pH-and calcium-dependent 55 target binding and is Fc-engineered to increase affinity at acidic pH to FcRn, and to reduce 56 effector functions. In cynomolgus monkeys, ARGX-117 dose-dependently reduces free C2 57 levels and classical pathway activity. A two-dose regimen of 80 and 20 mg/kg separated by a 58

week, resulted in profound reduction of classical pathway activity lasting for at least 7 weeks. 59

Conclusion: 60 ARGX-117 is a promising new complement inhibitor that is uniquely positioned to target 61 both the classical and lectin pathway while leaving the alternative pathway intact. 62 J o u r n a l P r e -p r o o f Introduction 99

The complement system consists of over 20 soluble and membrane-bound proteins. The 100 system is activated via three pathways, the classical, lectin and alternative pathways (CP, LP  101 and AP, respectively), which converge at the level of C3 to activate a final common pathway 102

leading to the formation of the membrane attack complex (MAC). Whereas its role as an 103

innate defense system against micro-organisms is well established, complement as an 104

inflammatory mediator system contributing to organ dysfunction and tissue injury in human 105 disease has received increasing attention during the last decades 1,2 . Two inhibitors, 106 eculizumab (Soliris ® , Alexion), a monoclonal antibody (mAb) that blocks C5, and 107 recombinant (Ruconest ® , Pharming) and plasma-derived C1-inhibitor (Cinryze ® , Takeda, and 108

Berinert ® , CSL/Behring), the main inhibitor of CP and LP, are currently approved for clinical 109 application 3,4 . However, there is no "one-size-fits-all" complement inhibitor, and both 110

inhibitors have their limitations. Eculizumab cannot address inflammatory or cytotoxic 111 reactions mediated by C3 activation products, which may explain its failure to prevent 112 extravascular hemolysis in paroxysmal nocturnal hemoglobinuria 5 . The C1-inhibitor is a poor 113

inhibitor of immune complex-bound C1 6 and has a relatively short half-life in plasma making 114 it less attractive for long-term prophylactic or chronic use 7 . Therefore, several other 115 complement factors are being considered as targets for therapeutic intervention 8,9 . 116 We hypothesized that inhibition of complement at the level of C2 is an attractive therapeutic 117 approach for several reasons. First, C2 is part of both CP and LP which are implicated in 118 diseases driven by auto-antibody formation or ischemia-reperfusion. Secondly, it leaves AP, 119

an important innate defense system, intact. Moreover, C2 deficiencies in humans are 120 associated with a lower prevalence of autoimmune diseases compared to C1 or C4 121 deficiencies, although predisposition to autoimmunity and systemic lupus is increased, in 122 particular in women. C2 deficiency also increases the risk for bacterial pyogenic infections 10 . 123

However, C2 deficiency is not associated with increased susceptibility to meningococcal 124

infections. Lastly, C2 in plasma is less abundant compared to other complement factors 11,12 . 125

It was previously shown that mAb-mediated inhibition of complement proteins in humans is 126 feasible 3 . However, owing to the high concentrations of most complement factors in plasma, 127

in combination with target-mediated clearance, high doses of mAbs are needed to fully block 128 complement. To overcome this, an antibody can be used that takes advantage of pH 129 differences between blood (pH 7.4) and endosome (pH 6.0) to dissociate the antigen from the 130 antibody-antigen complex in the endosome to be degraded into the lysosome, a so-called 131 sweeping antibody. Sweeping properties endow a mAb with a prolonged in vivo activity 13 . 132

Here we describe ARGX-117, an anti-human C2 mAb with pH and calcium dependent target 133 binding properties. This mAb was generated in mice, humanized, and formatted as a human 134

IgG1 antibody with mutations to knock out effector functions 14 Results of the surface plasmon resonance (SPR) experiments were analyzed using Biacore 198 T100 evaluation software. The inhibitory effect of ARGX-117 in the complement-mediated 199 phagocytosis assay and in the complement-dependent cytotoxicity assay was tested via one- with human C2, and screened for the production of anti-C2 antibodies with an ELISA using 207 solid-phase human C2. Inhibitory potency of anti-C2 antibodies was evaluated using an in-208

house ELISA in which fixation of C3 and C4 from fresh human serum to solid-phase-bound 209 aggIgG was measured. Antibodies that inhibited fixation of C3, but not that of C4, were 210 further analyzed. Among the nine inhibitory mAbs identified, anti-C2 mAb 5F2.4 was 211 selected for humanization using the human composite antibody technique (Abzena Ltd, 212

Cambridge, UK) as described 16 Figure 1A -C), whereas its affinity to FcRn at 217 acidic pH was increased compared to a control human wild-type IgG1 (Supplemental Figure  218 1D). Increased FcRn binding at low pH rescues ARGX-117 from lysosomal degradation and 219 enhances recycling into circulation, resulting in a prolonged half-life. 220

The affinity of the binding of ARGX-117 to C2 was quantitated with SPR (Supplemental 221 methods). The kD of the binding of Fab ARGX-117 to C2 was about 0.3 nM (Supplemental 222 Figure 2 , Supplemental Table 1 ). 223

Assessing cross-reactivity with C2 from several animal species in ELISA revealed that 224 ARGX-117 bound to both cynomolgus C2 and human C2. Moreover, it had a low affinity for 225 rat C2, but did not react with mouse, guinea pig, or rabbit C2 (data not shown). 226

Commercial complement assays, (see supplemental methods), were used to assess the 228 complement pathways inhibited by ARGX-117 in human serum. ARGX-117 potently 229

inhibited CP and LP (EC 50 = 30.5±4.5, and 93.4±10.4 µg/mL, respectively) in a 230 concentration-dependent manner ( Figures 1A-B) . Importantly, AP activation was not affected 231 by ARGX-117 ( Figure 1C ). The commercial assays measure complement activation by 232 assessing MAC fixation to the solid-phase. ARGX-117 also inhibited C3 fixation from human 233 serum to solid phase aggIgG, whereas it had no effect on C4 fixation ( Figure 1D ). 234

The generation of fluid-phase activation products after complement activation can enhance 235 inflammation and infiltration of immune cells, for example in dense deposit disease and C3 236 glomerulopathies 17 . To assess inhibition of fluid-phase complement activation products by 237 ARGX-117, aggIgG were added to fresh human serum. Subsequently, the generation of C4b/c 238 and C3b/c was determined using specific ELISAs (see supplemental methods for details). Pre-239

incubation of serum with ARGX-117 had no effect on the generation of C4b/c as expected, 240

but dose-dependently inhibited C3b/c fluid-phase activation (Supplemental Figure 3A- as when heat-inactivated serum was used, indicating that phagocytosis was complement-257 dependent. Both C3 fixation and phagocytosis was dose-dependently inhibited by ARGX-258 117, but not by an anti-C5 mAb or an isotype control for ARGX-117 ( Figure 2C ). 259

Antibody-mediated rejection is a major concern in solid-organ transplantation necessitating 260

pre-transplantation screening of recipient for complement activating antibodies against donor 261

HLA antigens with the crossmatch test 19 . We modified this test by using human instead of 262 rabbit complement to evaluate the potency of ARGX-117 to block complement-dependent 263 cytotoxicity by anti-HLA antibodies (see supplemental methods). Human PBMCs were 264 sensitized with anti-HLA antibodies from hyperimmunized kidney transplant patients, 265

incubated with fresh human serum with or without ARGX-117, and assessed for cell lysis by 266 microscopy. In absence of ARGX-117 about 20% of the cells were lysed, which is less than 267 the lysis induced by rabbit complement due to protection of PBMCs against human 268 complement by membrane-bound complement regulatory proteins. ARGX-117 concentration-269 dependently inhibited lysis of anti-HLA sensitized PBMC ( Figure 2D ). 270

To identify the domain harboring the epitope for ARGX-117, domain swap mutants of C2 and 272 FB were generated. We took advantage of the fact that C2 and FB are homologous proteins 273 and have the same domain structure ( Figure 3A ); the small subunits (C2b and FBa) of both 274 proteins harbor three Sushi domains (Sushi 1-2-3, also known as the CCP or short consensus 275 repeat (SCR) domain), whereas the large subunits (C2a and FBb) contain a von Willebrand 276

Factor-A domain and a serine protease domain. ARGX-117 did not bind to full length FB and 277

to chimeric C2 containing the FB-derived Sushi-2 (S2) domain, whereas the binding was 278 restored to FB containing the S2 domain of C2 ( Figure 3B) . Thus, the epitope of ARGX-117 279

is located in the S2 domain of the C2b subunit of C2. 280

To investigate the effect of ARGX-117 on C2 cleavage in serum during activation, 281

C2-deficient serum was reconstituted with purified plasma derived C2, activated with aggIgG, 282 separated on SDS-PAGE, and immunoblotted with biotinylated ARGX-117 as detection 283 antibody (Supplemental methods, Figure 3C ) incubation with an anti-MASP2 mAb did not affect the increase of C2b and the decrease of 292 total C2 in serum by aggIgG, confirming the observed activation was mediated by the 293 classical pathway. Both an anti-C5 mAb, or isotype control did not inhibit total C2 cleavage 294 and showed increased C2b (lanes 8, 9 and 10) . 295

The inhibition of the cleavage of C2 into C2a and C2b during activation by ARGX-117 can 296 either be explained by i) steric hindrance of C1s cleavage of the peptidyl bond between C2a 297

and C2b or ii) preventing the interaction of C2 with C4b thereby avoiding exposure of the 298 peptidyl bond. ARGX-117 was not able to inhibit C2 cleavage by C1s in a purified system, 299

suggesting it interferes with the interaction of C2 with C4b. Note that in this purified system 300 EDTA does not inhibit C2 cleavage by C1s since the interaction of C1s with C2 is calcium 301 independent ( Figure 3D ). 302

Next, the effect of ARGX-117 on the formation of the CP C3 proconvertase complex was 303 further studied using SPR. Biotinylated C4b was immobilized onto streptavidin-coated chips 304

and C2 was added to allow the formation of the C4bC2 complex on the chip surface ( Figure  305 3E, red line). The binding of C2 to C4b was not influenced by control mAb ( Figure 3E , black 306 line). However, pre-incubation of C2 with ARGX-117 in different molar ratios, completely 307

inhibited C2 binding to C4b ( Figure 3E , lines in shades of blue). Taken together these results 308

argue that ARGX-117 interferes with the binding of C2 via its Sushi 2 domain to C4b. 309

Most anti-complement antibodies suffer of a fast clearance in vivo due to target mediated 311 clearance. Surprisingly, a pilot PK study in cynomolgus monkeys using a variant of ARGX-312 117 showed long PK and sustained PD effect (data not shown). This observation triggered us 313 to investigate the binding properties of ARGX-117 to C2 at different pHs. ARGX-117 binds 314 several hundred-fold better to human C2 at physiological plasma pH (pH 7.4) than at lower 315 pH of endosomes (pH 6.0; half-maximal effective concentration EC 50 = 0.1 µg/mL at pH 7.4 316 versus EC 50 = 32.2 µg/mL at pH 6.0; Figure 4A ). Additionally, during setup of different 317 hemolysis assays, it was found that ARGX-117 was less potent in EDTA-treated serum 318 reconstituted with suboptimal concentrations of calcium, hinting that ARGX-117 would also 319 be affected by different calcium concentrations. Indeed, the binding to C2 is approximately 320 25-fold better at physiological (1.25 mM) plasma Ca 2+ concentrations (EC 50 = 0.1 µg/mL) 321 compared to lower (25 µM) Ca 2+ concentrations typically found in endosomes (EC 50 = 2.5 322 µg/mL; Figure 4B ). Therefore, it is expected that C2 dissociates from ARGX-117 in the 323 endosome to be degraded into the lysosome, whereas ARGX-117 will be recycled to the 324 circulation enabling binding of a new target 20 . This effect can be optimized by introducing 325 H433K and N434 mutations in the Fc part 21 , which improves the recycling of antibody in 326 vivo. The combination of half-life extension by Fc-engineering and pH-and calcium-327 dependent target binding in ARGX-117 translates into optimal "sweeping" properties. 328

In a single-dose study, groups of 4 animals (2m/2f) were administered with an intravenous 330 dose (bolus injection) of 0.5, 1.5, 5, 10, or 30 mg/kg of ARGX-117. Dose linearity was 331 observed for all dose groups when the log of the serum concentration of ARGX-117 was 332 plotted against time ( Figure 5A ). Using all these data together, an overall half-life of 333 ARGX-117 of approximately 2 weeks was calculated (Table 1) . Anti-drug antibodies (ADA) 334

were observed in 9 of the 20 animals across all dose groups (Supplemental Figure 4) . No 335 relation between apparent half-life and ADA formation was observed. 336

In a repeat-dose study, 3 cynomolgus monkeys received a loading dose of 80 mg/kg and a 337 second dose of 20 mg/kg one week later ( Figure 5A ). The apparent half-life of ARGX-117 338 after the second dose appeared to be longer than the mean half-life in the single dose group 339 (up to 21.6 days after the second dose), despite measurable ADA formation in all 3 animals. 340

The course of CP activity in time upon administration of ARGX-117 in the animals, was 341 measured with the Quidel assay ( Figure 5C) (Figure 3) . 366 ARGX-117 binds C2 with a high, sub-nanomolar affinity translating into effective inhibition 367 of CP and LP in various in vitro assays. Intrigued by the long PK and PD effect observed in 368 an exploratory cynomolgus monkey experiment (Supplemental Figure 5A /B), which is not the 369 case for many other complement targeting antibodies because of target-mediated clearance, 370

we decided to investigate the binding characteristics of ARGX-117. Importantly, this affinity 371 is optimal at physiological pH and free Ca 2+ concentration, and decreases by 100-to 1000-372 fold at pH 6.0 and a calcium concentration of 25 µM, which conditions resemble those in 373 endosomes 20,23 . These unique biochemical characteristics predict that ARGX-117 efficiently 374 captures C2 in the circulation, to release it upon endocytosis by endothelial cells in the 375 endosomes where C2 then is sorted to be degraded in the lysosomes (= sweeping antibody) 20 . 376

Introducing histidine residues into the complementarity determining regions to increase the 377 pH dependent antigen binding furthermore was not effective, suggesting that the lead 378 antibody already had maximal sweeping capacity (data not shown). Furthermore, H433K and 379 N434F mutations ("NHance") in its Fc-region will rescue ARGX-117 from lysosomal 380 degradation as these mutations increase the affinity to FcRn at the acidic pH of the 381 endosomes 14 thereby favouring recycling of the antibody into circulation where it may capture 382 additional C2 molecules. Indeed, comparison of PK and PD of the C2 antibody with NHance 383 mutations demonstrated a longer serum half-life and longer acting effect as compared to the 384 version without NHance mutations (Supplemental Figure 5C /D). 385

Consistent with the in vitro findings, a single administration of ARGX-117 to cynomolgus 386 monkeys at doses ≥1.5 mg/kg induced a temporary dose-dependent inhibition of CP as 387 evidenced by decreasing CH50 activity ( Figure 5) . A dose-linear PK profile was observed in 388 all ARGX-117 dose groups (0.5 to 30 mg/kg single dose), with an apparent half-life of 389 clearance of 2 weeks. This relatively long half-life strongly suggests the clearance of 390 ARGX-117 is independent of its target C2, and in this respect seems different from anti-C1s 391 sutimlimab 24 or C1q mAbs 25 . After a loading dose of ARGX-117 of 80 mg/kg followed by a 392 maintenance dose of 20 mg/kg at day 8 (80 + 20 mg/kg group), free C2 levels were low to 393 undetectable for up to 7 weeks, and CP was completely inhibited for 7 weeks or longer 394 ( Figure 5 ). The long half-life and the prolonged complement inhibition in cynomolgus 395 monkeys support such a recycling mechanism by ARGX-117. Notably, the in vivo potency of 396 ARGX-117 in humans may be underestimated as inhibition of CP in cynomolgus serum using 397 several in vitro complement assays requires ~2-fold more ARGX-117 than comparable 398 inhibition of CP in human serum (data not shown). 399

Various complement effector functions (cell lysis, opsonization, release of fluid phase 400 activation products) may contribute to a different extent to complement-mediated pathologies. 401

Importantly we show that ARGX-117 can inhibit multiple complement effector functions. It 402 prevents fixation of C3 and TCC to solid-phase aggregated IgG (Figure 1 ), reduces fluid 403 phase activation of C3 (Supplemental Figure 2) and prevents complement-dependent 404 cytotoxicity ( Figure 3) . Importantly, ARGX-117 prevents C3 activation via CP or LP. 405

Therefore, ARGX-117 has therapeutic potential in conditions such as C3 mediated-406 extravascular hemolysis in autoimmune hemolytic anemia (Figure 2 ). It may also be useful to 407

inhibit CP-or LP-dependent generation of C3a after C3 activation which is believed to 408 contribute to graft rejection after organ transplantation by enhancing T cell responses 26 Human serum was supplemented with ARGX-117 at the indicated concentrations. A) 548

Inhibition of CP-mediated complement activation was tested using the Quidel CH50 assay. B) 549

Inhibition of LP-mediated complement activation was tested using the Wieslab MP320 assay. 550 C) No inhibitory effect of ARGX-117 on AP-mediated complement activation was observed 551

using the Wieslab AP330 assay.D) Both C4 and C3 fixation were used as read-out in a solid-552 phase aggregated IgG-mediated complement activation assay which was developed in-house. 553

A concentration dependent inhibitory effect of ARGX-117 was observed on C3 fixation, 554

whereas fixation of C4, which acts upstream of C2, was unaffected. Data represent mean and 555 SEM of 3 (A, B,) or 2 (C, D) independent experiments. Anti-C5 mAb was tested in the assay 556

for AP as a positive control. See for further details M&M. 557 Inhibition of complement-dependent lysis of PBMC sensitized with anti-HLA antibodies. 565

Rabbit complement was used as positive control and 10 mM EDTA as negative control, 566

percentage lysis was calculated to controls. ARGX-117 concentration-dependent inhibits 567 complement-mediated cell lysis (compared to 20% serum condition, one-way ANOVA, 568 followed by Dunnett's multiple comparisons test, ****<0.0001, ***<0.001, **<0.001). 569

Results represent mean and SEM of three (A, B and C) and two (D) independent experiments. 570 ARGX-117 on the solid-phase was incubated with the swap mutants, which were detected 573 with biotinylated anti-FLAG antibody. Note that ARGX-117 binds to S2 domain of C2b. C) 574

Western blot analysis of C2-depleted (C2-DPL) serum reconstituted with serum purified C2 575 and activated with aggIgG (800µg/ml 

C2a C2b SP S1 S2 S3 VWFA Serine protease FBb FBa SP S1 S2 S3 VWFA Serine protease S1 S2 S3 VWFA Serine protease S1 S2 S3 VWFA Serine protease S1 S2 S3 VWFA Serine protease S1 S2 S3 VWFA Serine protease S1 S2 S3 VWFA Serine protease S1 S2 S3 VWFA Serine protease S1 S2 S3 VWFA Serine protease S1 S2 S3 VWFA Serine protease S1 S2 S3 VWFA Serine protease S1 S2 S3 VWFA Serine protease C2-S1 

Complement and its receptors: new insights into human disease

New insights into the immune 451 functions of complement

Discovery and development of 453 the complement inhibitor eculizumab for the treatment of paroxysmal nocturnal 454 hemoglobinuria

Replacement therapy in hereditary angioedema: successful treatment of acute episodes 457 of angioedema with partly purified C1 inhibitor

Complement 459 fraction 3 binding on erythrocytes as additional mechanism of disease in paroxysmal 460 nocturnal hemoglobinuria patients treated by eculizumab

Activator-bound C1 is less susceptible to inactivation by C1 462 inhibition than is fluid-phase C1

464 Population pharmacokinetics of plasma-derived C1 esterase inhibitor concentrate used to treat 465 acute hereditary angioedema attacks

The renaissance of complement 467 therapeutics

Clinical promise of next-generation complement 469 therapeutics

Systemic Lupus Erythematosus and Deficiencies of Early 471 Components of the Complement Classical Pathway

Quantification 473 of human complement C2 protein using an automated turbidimetric immunoassay

Complement determinations in human disease

Long 478 lasting neutralization of C5 by SKY59, a novel recycling antibody, is a potential therapy for 479 complement-mediated diseases

Effector function 481 activities of a panel of mutants of a broadly neutralizing antibody against human 482 immunodeficiency virus type 1

Divergent activities of an 484 engineered antibody in murine and human systems have implications for therapeutic 485 antibodies

Circumventing immunogenicity in the development of 487 therapeutic antibodies

Generation of multiple fluid

C3b:plasma protein complexes during complement activation: possible implications in C3 490 glomerulopathies

Role of Complement in Autoimmune Hemolytic Anemia

Sensitized renal transplant recipients: current protocols and 494 future directions

Sweeping antibody as a novel therapeutic antibody 496 modality capable of eliminating soluble antigens from circulation

Divergent activities of an 499 engineered antibody in murine and human systems have implications for therapeutic 500 antibodies

High Risk for Invasive 502

Meningococcal Disease Among Patients Receiving Eculizumab (Soliris) Despite Receipt of 503

Calcium-505 dependent antigen binding as a novel modality for antibody recycling by endosomal antigen 506 dissociation

A 508 randomized, first-in-human, healthy volunteer trial of sutimlimab, a humanized antibody for 509 the specific inhibition of the classical complement pathway

Antibody-mediated organ-allograft rejection

Nonclinical 514 Development of ANX005: A humanized anti-C1q antibody for treatment of autoimmune and 515 neurodegenerative diseases

The first case 518 of COVID-19 treated with the complement C3 inhibitor AMY-101

Eculizumab 522 treatment in patients with COVID-19: preliminary results from real life ASL Napoli 2 Nord 523 experience

Complement 530 associated microvascular injury and thrombosis in the pathogenesis of severe COVID-19 531 infection: A report of five cases

This work was supported by a grant by LSH Impuls Programme from The Life Science and

(AUC), per group of the PK/PD single dose study in cynomolgus monkeys and in 537 cynomolgus monkeys doses with 80mg/kg (day 0) followed by dose of 20mg/kg (day 7). 538Average ± SD. Upon administration of ARGX-117 at doses of ≥1.5 mg/kg, an immediate 539 decrease in free C2 levels was observed, which then gradually returned to pre-dose levels in 540 time ( Figure 5B ). In animals that received 80 and 20 mg/kg, free C2 levels stayed low for up 541 to 7 weeks, and only at the latest timepoints free C2 levels increased again. Notably, at the 542 latest timepoints, ADA formation had started in all animals. a AUC 0-last TD last sampling at day 543 58, b AUC 0-last TD last sampling at day 54, c Half-life and C max of ARGX-117 were determined 544 after the second dosing. 545