Development and Biophysical Characterization of a Humanized FSH–Blocking Monoclonal Antibody Therapeutic Formulated at an Ultra–High Concentration

Highly concentrated antibody formulations are oftentimes required for subcutaneous, self-administered biologics. Here, we report the creation of a unique formulation for our first–in–class FSH–blocking humanized antibody, MS-Hu6, which we propose to move to the clinic for osteoporosis, obesity, and Alzheimer’s disease. The studies were carried out using our Good Laboratory Practice (GLP) platform, compliant with the Code of Federal Regulations (Title 21, Part 58). We first used protein thermal shift, size exclusion chromatography, and dynamic light scattering to examine MS-Hu6 concentrations between 1 and 100 mg/mL. We found that thermal, monomeric, and colloidal stability of formulated MS-Hu6 was maintained at a concentration of 100 mg/mL. The addition of the antioxidant L–methionine and chelating agent disodium EDTA improved the formulation’s long–term colloidal and thermal stability. Thermal stability was further confirmed by Nano differential scanning calorimetry (DSC). Physiochemical properties of formulated MS-Hu6, including viscosity, turbidity, and clarity, conformed with acceptable industry standards. That the structural integrity of MS-Hu6 in formulation was maintained was proven through Circular Dichroism (CD) and Fourier Transform Infrared (FTIR) spectroscopy. Three rapid freeze–thaw cycles at −80°C/25°C or −80°C/37°C further revealed excellent thermal and colloidal stability. Furthermore, formulated MS-Hu6, particularly its Fab domain, displayed thermal and monomeric storage stability for more than 90 days at 4°C and 25°C. Finally, the unfolding temperature (Tm) for formulated MS-Hu6 increased by >4.80°C upon binding to recombinant FSH, indicating highly specific ligand binding. Overall, we document the feasibility of developing a stable, manufacturable and transportable MS-Hu6 formulation at a ultra–high concentration at industry standards. The study should become a resource for developing biologic formulations in academic medical centers.


INTRODUCTION 51
Biotherapeutics have transformed the treatment of chronic inflammatory diseases and 52 cancers 1-3 , with monoclonal antibodies witnessing the most rapid growth because of their 53 specificity and adaptability 4-6 . As monoclonal antibodies, particularly blocking antibodies, are 54 generally of low affinity, and therefore must be administered at relatively high doses to elicit 55 higher (or more stable) indicate overall thermostability. To optimize the formulation, we 126 increased MS-Hu6 concentration from 0.2 to 100 mg/mL ( Figure 1A). T m values for the Fc and 127 Fab domains of MS-Hu6 in PBS at 1 mg/mL were 69.17±0.02 °C and 79.64±0.02 °C, 128 respectively. Increases T m values compared with MS-Hu6 in PBS were noted progressively up 129 to 10 and 20 mg/mL for the Fab and Fc regions of formulated MS-Hu6. However, at higher 130 concentrations (50 and 100 mg/ml), while T m values dropped somewhat for the MS-Hu6 131 formulation compared with MS-Hu6 in PBS, the difference (or ΔT m ) did not exceed 1 o C-which 132 provided assurance of thermostability even at 100 mg/mL. 133 We further used SEC to study monomeric stability of formulated MS-Hu6 based on 134 molecular size. The percentage of monomer (peak 2) and dimer (peak 1) areas were 135 monitored. Figure 1B shows that MS-Hu6 in PBS or in formulation exists mainly in its 136 monomeric form (>99%) with a dimer area of <1%; the latter is within acceptable industry 137 standards of <3-5%. 138 We examined colloidal stability using DLS, a well-recognized, quick, and sensitive 139 method to determine particle size and particle-size distribution. The method is based on light 140 scattering due to Brownian motion. While maintaining the original structure of proteins, DLS 141 measures the hydrodynamic diameter (rh) and polydispersity index (PDI). The samples were 142 diluted with respective buffers (PBS or Formulation Buffer) prior to DLS. Data were collected in 143 volume percent, with hydrodynamic radius (rh) presented for the prominent and smaller peaks 144 ( Figure 1C). Particle size data were reported in Z-average values. 145 The two MS-Hu6 (in PBS) samples displayed a prominent peak (volume % ~99%) with 146 an average rh of 4.96 nm, which fell within the industry standard (rh <10 nm). However, there 147 were two aggregated peaks with average volume percentages of <1%, but with an average rh of 148 91.4 and 2595 nm, respectively. The multiplicity of such peaks resulted in a high PDI, indicating 149 varying particle size ( Figure 1C). In contrast, MS-Hu6 formulations containing 1 and 50 mg/mL 150 MS-Hu6 displayed 99-100 volume % in a single peak, with average rh values of ~4 nm. In the 151 100 mg/mL formulation, however, a negligible peak (0.1% in sample 3) was associated with a 152 large rh of 2671 nm, whereas the other samples (1 and 2) displayed single peaks with an 153 average rh of 3.68 nm. This resulted in lower PDI values for 1 and 50 mg/L, with a similar PDI 154 for the 100 mg/ml formulation to that of MS-Hu6 in PBS. Nonetheless, the PDI values in all 155 cases were <1, which is the accepted cut-off. Large particles with a high rh values can form as 156 soluble reversible aggregates, which are well accepted in biotherapeutics development at levels 157 <10%. Thus, in all, the 100 mg/mL MS-Hu6 formulation showed an acceptable rh and PDI,158 suggesting that the monomeric nature of MS-Hu6 was maintained at higher concentrations. 159 160

Optimization of Antioxidant and Chelating Agent 161
Due to the presence of trace elements and to avoid catalytic oxidation during 162 manufacturing, disodium EDTA was tested as an excipient for formulated  Furthermore, given that there are two methionine residues in the Fc domain, namely Met 164 33(C H 3) and Met 209(C H 3), we opted also to use L-methionine to avoid methionine oxidation 165 during manufacturing and storage. We examined the effect of disodium-EDTA and L-166 methionine on the stability of the MS-Hu6 formulation using protein thermal shift and DLS. 167 Protein thermal shift showed that compared with MS-Hu6 (8.621 mg/mL) in PBS, formulated 168 MS-Hu6 (100 mg/mL) plus 1 mM EDTA and 1 mM L-methionine showed slight increased ΔT m 169 value for the Fc domain of 0.38 o C (Figure 2 A). This suggested that methionine oxidation and 170 catalytic oxidation were being prevented to enhance the thermostability of MS-Hu6. In contrast, 171 there was no significant T m shift in the Fab domain upon EDTA or L-methionine addition. 172 increase in absorbance, with a flattening at 100 mg/mL MS-Hu6 in formulation. The increase at 201 100 mg/mL is likely due to antibody crowding in solution and formation of reversible aggregates; 202 however, the latter phenomenon is not consequential in biopharmaceutical product 203

development. 204
Finally, there were also no differences in clarity at any wavelength except a minimal drop 205 with formulated MS-Hu6 at 100 mg/mL (Figure 3 C). In all, the percent transmittance for each 206 concentration was more than >99% at the three wavelengths, suggesting that the formulations 207 are clear between 0.2 and 50 mg/mL and slightly opalescent at 100 mg/mL. All concentrations 208 fall into the acceptable clarity categories of   48 . 209 210

Assessment of Structural Integrity 211
An alteration in the secondary structure or conformation of any protein indicates 212 inactivation or denaturation. We used Circular Dichroism (CD) spectroscopy to evaluate the 213 secondary structure of formulated MS-Hu6. As shown in Figure 4 A and through mathematical 214 derivations of ellipticity (millidegrees) and mean residue ellipticity, analysis in the far UV region 215 (190-200 nm), revealed that, as with other IgG molecules, the predominant secondary structure 216 in MS-Hu6 was regular β-sheets and unordered/random coils 49-51 . Importantly, there was no 217 alteration in any parameter in formulated MS-Hu6 versus MS-Hu6 in PBS (0.5 mg/mL). 218 However, since antibodies can be used for CD spectroscopy only at low concentrations, 219 in our case at 0.5 mg/mL. We utilized Fourier-transform infrared (FTIR) spectroscopy to further 220 study the structure of formulated MS-Hu6 at 50 and 100 mg/mL compared with MS-Hu6 in PBS. 221 The amide I band peak at 1637 cm -1 , a surrogate for intra-molecular β-sheets, and the random 222 coil structure peak at 1642-1657 cm -1 did not shift, confirming maintenance of the antibody's 223 native conformation in formulation (Figure 4 B). 224 . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted May 11, 2023. ;https://doi.org/10.1101https://doi.org/10. /2023 We finally employed nano differential scanning calorimetry (Nano DSC) to determine the 225 protein unfolding temperatures (T m s). MS-Hu6 at 50 and 100 mg/ml were diluted to 5 mg/ml in 226 formulation buffer. MS-Hu6 displayed multi-domain transitions, with the first transition 227 representing unfolding of C H 2 (Fc) domain (T m >71 °C), the second representing Fab unfolding 228 that formulated MS-Hu6 is stable under extreme stress and that there is minimal or no 249 aggregation or denaturation. 250 All rapid F/T samples were further evaluated by DLS to determine their colloidal stability. 251 For three cycles, the major volume peak (>99%) rh was found to be between 3 to 7 nm ( Figure  252 5 B was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted May 11, 2023. ;https://doi.org/10.1101https://doi.org/10. /2023 accelerated conditions at 25 o C for at least 90 days. These relatively stable T m s likely arise from 273 the stabilizing effects of the salt, stabilizer and sugar utilized in the optimized formulation. 274 We also assessed monomer loss by SEC. Samples stored at 4°C retained 99.36% of 275 monomer area at baseline (Figure 6  For samples stored at 25°C for 90 days, the average monomer area retained was 284 96.84±1.22% for three batches, and monomer loss was 3.16% on average (Figure 6 B). 285 However, two of the samples (batches 1 and 2) lost ~4% monomer within the acceptable limit of 286 <5%, while one sample retained 99.29% monomer. This increased monomer loss at 90 days is 287 likely due to the formation of soluble high molecular weight species (>150 kDa); these species 288 are not considered critical in biopharmaceutical development due to their reversible and 289 transient nature. In all, formulated MS-Hu6 at 4°C and 25°C for 90 days showed acceptable 290 stability and retained maximum monomeric area in accordance with set standards. 291

FSH Binding Studies 293
We evaluated formulated MS-Hu6 for its binding to human FSH. MS-Hu6 (1 µg/µL) in 294 formulation buffer or PBS was incubated with and without 10 µg/µL FSH, and its thermostability 295 was measured by protein thermal shift, as before. We expected that binding of FSH will 296 stabilize the Fab, but not the Fc domain of MS-Hu6, and hence result in a right-shift of the T m . 297 . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted May 11, 2023. ;https://doi.org/10.1101https://doi.org/10. /2023 We noted a ΔT m of 4.49±0.03 o C for MS-Hu6 in PBS and 4.80±0.07 o C for formulated MS-Hu6 298 ( Figure 7). There was little or no change in T m in the Fc domain. These data establish 299 conclusively the retained binding of formulated MS-Hu6 with its ligand, FSH. 300

DISCUSSION 302
We report the development and biophysical characterization of a formulation containing 303 an ultra-high concentration MS-Hu6. We initially utilized protein thermal shift to find that 304 formulated MS-Hu6 (0.2 to 100 mg/mL) had comparable T m s when compared with MS-Hu6 in 305 PBS with ΔT m shifts of ≤1°C--this confirmed thermostability for both Fc and Fab domains 306 compatible with industry standards 37, 41, 60 . The improved thermostability is attributed to the 307 stabilizing effect of excipients, such as sucrose, NaCl, Tween 20, L-methionine and disodium 308 EDTA, each of which have distinct mechanisms of action. 309 To assess for monomeric stability and potential aggregation, we used SEC to find that 310 the monomeric peak was >99% at all MS-Hu6 concentrations-this suggested that formulated 311 MS-Hu6 was monomerically stabile at higher concentrations with minimal aggregation and no 312 fragmentation. Clinical-grade antibodies must display active protein loss of not more than 5% 313 and secondary species <2% in a freshly reconstituted therapeutic formulations 6, 58, 59 . We 314 complemented SEC with DLS to ascertain aggregate formation and colloidal stability. At each 315 MS-Hu6 concentration in formulation, the major peak volume was >99% with a rh of 3-6 nm, 316 and peak volumes for aggregated particles of <1%; this is consistent with that reported for 317 several clinical-grade monoclonal antibodies 43, 61 . Of note is that soluble reversible aggregates 318 up to 10 µm are accepted in biotherapeutics development. However, due to their potential 319 immunogenicity or adverse reactions, particulate matter or large aggregates may need to be 320 monitored under strict FDA guidelines 28, 62 . 321 . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted May 11, 2023. ; https://doi.org/10.1101/2023.05.11.540323 doi: bioRxiv preprint Methionine oxidation of Fc domains is a common post-translational modification that can 322 affect antibody bioactivity and potentially induce an immunogenic response. The interface of 323 the C H 2 and C H 3 domains of most IgG1 antibodies contains two conserved heavy chain (HC) 324 methionine residues. The oxidation of these methionine residues reduces thermal and colloidal 325 stability 63, 64 , and in doing so, impairs long-term storage stability (shelf-life). The anti-oxidant 326 L-methionine and chelating agent, disodium-EDTA are thus used to increase the shelf-life of 327 commercial IgG1 biotherapeutics. Free L-methionine protects methionine residues from 328 oxidation, while EDTA increases the effectiveness of anti-oxidants through its ability to complex 329 heavy metal 44, 65 . We found that the addition of 1 mM L-methionine and 1 mM disodium-EDTA 330 added together resulted in thermostability in the protein thermal shift assay, and colloidal 331 stability in DLS, with a major peak of 99.9% and rh 6.93 nm. 332 Viscosity is an essential parameter for any injectable dosage form, as highly viscous 333 formulations increase back pressure and pain at injection sites. This issue is magnified for 334 subcutaneously administered therapeutic antibodies at volumes <1.5 mL, making viscosity 335 control a key determinant of highly concentrated formulations. We found an expected increase 336 in viscosity with increasing concentrations of MS-Hu6, likely resulting from antibody crowding 47 , 337 compounded at high concentrations, by protein-protein interactions. However, the viscosity at 338 100 mg/mL MS-Hu6 was considerably less than the industry maximum of 50 cPs 66, 67 . The 339 latter is known to affect syringeability. However, viscosity can vary between 30 and 50 cPs 340 even at high antibody concentrations (>100 mg/mL). In recent years, viscosity control has been 341 extensively studied as a function of concentration, temperature, pH, ionic strength, and ion type 342 66 . 343 Turbidity and clarity are equally important parameters in biopharmaceutical 344 development. Increases in turbidity and reduced clarity could result from antibody overcrowding 345 or gelation and the formation of reversible aggregates due to weak protein-protein interactions, 346 . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted May 11, 2023. ; https://doi.org/10.1101/2023.05.11.540323 doi: bioRxiv preprint such as Fc-Fc, Fc-Fab, hydrogen bonding, and hydrophobic interactions. We found a small 347 increase in turbidity (absorbance) of 100 mg/mL formulated MS-Hu6 at 350 nm. Per European 348 Pharmacopoeia, this would be categorized as Ref II (slightly opalescent ≤6 NTU) [86,87]. In 349 the case of clarity, percent transmittance of formulated MS-Hu6 was >99% at all wavelengths, 350 indicative of clarity between 0.2 and 50 mg/mL, with a slight decrease at 100 mg/mL (350 nm). protein loss in a newly reconstituted biotherapeutic formulations should be <5%, with secondary 358 species at a maximum of 2% 6, 58, 59 . Intact secondary structure was confirmed by FTIR, in 359 which the amide I band peak at 1637 cm -1 (representing intra-molecular β-sheets) and the 360 random coil structure peak at 1642-1657 cm -1 were unaltered in formulated MS-Hu6 (100 361 mg/mL) compared with MS-Hu6 in PBS. The data rule out significant inter-molecular β-sheets 362 resulting from protein aggregation or clustering. For further confirmation, we used the nano 363 DSC to study multidomain transitions, as a surrogate for the conformation alterations of the Fab, 364 Fc (C H 2), and Fc (C H 3) domains. Comparison between 50 and 100 mg/mL MS-Hu6 revealed 365 almost overlapping peaks-further confirming retained conformational stability of the respective 366

domains. 367
Aggregation can occur at various stages of product development, including freeze-thaw, 368 manufacturing, filling, shipping, and storage 69 . This could destabilize and degrade the product, 369 and thereby reduce shelf-life. We thus carried out a rapid freeze-thaw (F/T) study at -370 80°C/25°C and -80°C/37°C and for three cycles to confirm formulation stability at extreme 371 . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted May 11, 2023. ;https://doi.org/10.1101https://doi.org/10. /2023 conditions. ΔT m s were found not to be significant (<1°C) for both Fc and Fab domains in all 372 stressed conditions, indicating that formulated MS-Hu6 was thermostable under extreme stress 373 conditions. The data also establish the absence of aggregation. For confirmation, the samples 374 were further evaluated by DLS for colloidal stability. All samples showed excellent colloidal 375 stability, with minimal reversible aggregation (<0.4%, acceptable limit <5-10) 55 . Moreover, the 376 PDI or rh under stressed conditions were broadly comparable for three F/T cycles, further 377 confirming colloidal stability. However, a small increase in PDI was noted at the third F/T cycle 378 at -80°C/37°C, likely attributable to reversible aggregates due to Fc-Fc, Fc-Fab, and air-liquid 379 interface interactions 47, 55-57 . For three cycles, these reversible aggregates remained in the 0.06 380 to 0.14% range, which is within acceptable limits (5-10%). This finding suggests that the ultra-381 high MS-Hu6 formulation retains its hydrodynamic radius at extreme storage conditions (-80 382 °C/37 °C for three cycles). 383 Accelerated stability refers to the extent to which a drug retains, within specified limits, 384 the near same properties and characteristics that it possessed at the time of manufacturing. It 385 is a surrogate for shelf life and determines optimum storage conditions. Both Fab and Fc 386 domains of formulated MS-Hu6 remained stable at 4°C for 90 days (ΔT m <1°C on thermal shift). 387 Moreover, at 25 °C, the Fc domain was stable for 60 days (ΔT m <1°C); however, after 90 days, 388 the T m fell slightly (>1°C), indicative of reduced stability. In contrast, the T m for of the Fab 389 domain did not change significantly (ΔT m -0.47°C). SEC was used to further confirm monomeric 390 stability after 90 days. At 4 and 25 °C, all three batches retained >99% and 96.8% monomer, 391 with loss of <1% and 3.16%, respectively (acceptable limit for active structure loss is 5%). No 392 fragmentation was noted in any samples at 4°C (90 days). However, monomer loss was noted 393 after 90 days at 25°C, due to the formation of soluble dimers, with molecular weights greater 394 than 150 kDa; the latter are not considered critical due to their reversible and transient nature. 395 Such dimerization is induced by antibody crowding in solution, weak interactions between Fc-396 . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted May 11, 2023. ;https://doi.org/10.1101https://doi.org/10. /2023 Fc, Fc-Fab, and Fab-Fab, and hydrophobic and air-liquid interactions 59, 70, 71 . These soluble 397 reversible aggregates are difficult to remove during antibody manufacturing and are therefore 398 accepted in biotherapeutics (5-10%). However, and importantly, no fragmentation (low 399 molecular weight species <150 kDa) was detected in any accelerated condition. Finally, MS-400 was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made  To provide information on colloidal stability of formulated MS-Hu6, DLS was carried out 451 to determine the size (hydrodynamic radius, rh) and homogeneity (polydispersity Index, PDI). 452 Briefly, 20 μL formulated MS-Hu6 was diluted into 1 mL using formulation buffer or PBS and 453 transferred to disposable 1 mL microcuvettes (Malvern Cat. ZEN0040). Using Zetasizer Nano-454 ZS 90 system (Malvern), the diluted sample was analyzed for 60 seconds at 90° scattering 455 angle at 25°C (5 cycles). The refractive index of the medium was set at 1.33 and dynamic 456 viscosities were measured. Data were reported as Z-average of rh and PDI. All experiments 457 were performed in duplicate, and representative particle size distribution (PSD) graphs are 458 was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted May 11, 2023. ;https://doi.org/10.1101https://doi.org/10. /2023 The protein thermal shift assay was routinely used to assess the thermostability of 462 formulated MS-Hu6. The assay utilizes Sypro Orange (Cat # 01288948, ThermoFisher), which 463 reports hydrophobic domains that are exposed during protein unfolding. Briefly, the 20 µL 464 Biotherapeutics are expected to experience extreme temperature conditions while being 479 transported or stored. Freeze-thaw cycle testing is a type of stability testing that determines the 480 robustness of an optimized formulation. Three batches of formulated MS-Hu6 (100 µL at 100 481 mg/mL) were subject to repeat freeze-thaws at -80°C/25°C or -80°C/37°C. Briefly, 100 µL 482 formulated MS-Hu6 was incubated at -80 °C for 3 hours (TSX series freezer, ThermoFisher). 483 Samples were thawed for 30 minutes at 25 °C in an Isotemp 250 thermostat (Fisher Scientific) 484 or at 37 °C in a water bath (PolyScience), and analyzed by dynamic light scattering or protein 485 . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made FTIR was carried out to confirm the structural integrity of formulated MS-Hu6 at high 500 concentration. Our Nicolet iS10 FTIR Spectrometer (ThermoFisher) is equipped with an 501 attenuated total reflectance (ATR) sampling accessory and a 45 o ZnSe crystal. Briefly, 10-20 502 µL blank reference buffer or protein samples were loaded onto the ATR sample holder, and 503 spectra were recorded from 500 cm -1 to 4500 cm -1 . After subtracting the reference spectra from 504 the sample spectra, smoothened second derivatives were generated using the Savitsky-Golay 505 7-point, 3 rd order polynomial, 2 nd derivative algorithm (Essential FTIR software, Operant). For 506 appropriate peak fitting, the 2 nd derivative peaks were multiplied by -1. The region of structural 507 interest, Amide I, was used to assess structural integrity. 508 509

Far-UV Circular Dichroism (CD) Spectroscopy 510
. CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted May 11, 2023. ; https://doi.org/10.1101/2023.05.11.540323 doi: bioRxiv preprint Far-UV circular dichroism (CD) spectroscopy was used to assess secondary structure 511 stability and integrity of formulated MS-Hu6. For this, 0.5 mg/mL (100 µL) of antibody samples 512 were transferred to a cuvette. CD spectra were collected at 25 °C in a 10-mm path length with a 513 50-nm/min scan speed and a 0.45-sec response time using a Chirascan CD spectrometer 514 (Applied Photophysics). A bandwidth of 2 nm was used to average five scans from 180 to 250 515 nm at 25 °C. Built-in functions were used to subtract the reference spectra from the respective 516 In addition to protein thermal shift assays, DSC was also performed to evaluate the 528 thermal stability of formulated MS-Hu6 at high concentrations. Briefly, samples were diluted 529 with formulation buffer to 5 mg/mL, and 500 µL of buffer (reference) or diluted sample were 530 loaded into 24 K gold cylindrical capillary cells in a Model-602000 Nano DSC (Waters), 531 equilibrated for 10 minutes, and heated from 25 °C to 90 °C at a rate of 1° C/min at a constant 532 pressure of 3 atm. To obtain the precise unfolding temperature, the specific heat capacity (Cp) 533 of the references were subtracted from the respective protein samples followed by data fitting 534 . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted May 11, 2023. ;https://doi.org/10.1101https://doi.org/10. /2023 using the three-transition model (NanoAnalyze Software). The respective thermograms for 535 formulated MS-Hu6 is reported. 536 537

Accelerated Stability 538
Protein thermal shift assays and size exclusion chromatography were used to examine 539 the stability of formulated MS-Hu6 at 4 °C (long term storage) or 25 °C (accelerated storage). 540 Three batches for each condition were evaluated. Samples were stored in 0.5 mL Eppendorf 541 tubes and collected at 0, 15, 30 and 90 days. Protein thermal shift assays were performed to 542 assess thermal stability. To determine monomeric stability, size exclusion chromatography was 543 carried out on samples that have been stored for 90 days at 4° C and 25° C. Experiments were 544 run in triplicate and representative chromatographs are reported. 545

FSH Binding Assay 547
To confirm binding, formulated MS-Hu6 or MS-Hu6 in PBS (1 µg/µL) was incubated with 548 or without human FSH (10 µg/µL) at room temperature for 30 min. Protein thermal shift assays 549 using Sypro Orange were performed as above. Experiments were carried out in triplicate. 550 551 . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted May 11, 2023. ;https://doi.org/10.1101https://doi.org/10. /2023  was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted May 11, 2023. ;https://doi.org/10.1101https://doi.org/10. /2023 pending patent application relating to the ultra-high formulation of MS-Hu6. These patents are 577 owned by Icahn School of Medicine at Mount Sinai (ISMMS), and the inventors and co-578 inventors. would be recipients of royalties, per institutional policy. M.Z. also consults for Rani 579 Pharmaceuticals, and several financial platforms, including Gerson Lehman Group and 580 Guidepoint, on drugs for osteoporosis and genetic bone diseases. 581

DATA AVAILABILITY 583
All data generated or analyzed during this study are included in the manuscript. Source 584 Data files have been provided for Figures 1-7. 585 586 . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted May 11, 2023. ; https://doi.org/10.1101/2023.05.11.540323 doi: bioRxiv preprint  Formulations were also evaluated for turbidity (opalescence) at 350 nm. Low absorbance (AU, 610 absorbance unit) was noted at lower concentrations; however, there was a concentration-611 dependent increase in absorbance at 100 mg/mL formulated MS -Hu6 (B). Formulated MS-Hu6 612 . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted May 11, 2023. ;https://doi.org/10.1101https://doi.org/10. /2023 evaluated for clarity at wavelengths of 350, 540 and 595 nm demonstrated that the formulation 613 is clear and transparent even at higher concentrations (100 mg/mL) that correlated well with the 614 turbidity data (C). β-sheets and unordered/random coils (A). Secondary structure was also confirmed at higher 620 formulation concentrations (50 and 100 mg/mL) using Fourier-transform infrared (FTIR) 621 spectroscopy. The amide I band peak at 1637 cm -1 (intra-molecular β-sheets) and the random 622 coil (1642-1657 cm -1 ) did not shift, confirming maintenance of the native conformation in 623 formulation (B). Thermostability was further confirmed using nano differential scanning 624 calorimetry (Nano DSC). MS-Hu6 concentrations of 50 and 100 mg/ml had comparable T m s, 625 indicating that the overall structure is conformationally and thermally stable at high 626 dynamic light scattering (DLS) for colloidal stability. The major peak (Peak 1) hydrodynamic 635 radius (rh) was found to be between 3 and 7 nm at both storage conditions (-80/25 °C and -636 80°C/37 °C). PDI-polydispersity index. The data suggest no or minimal reversible 637 aggregation (<0.5%; acceptable limit:5-10%). 638 . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted May 11, 2023. ;https://doi.org/10.1101https://doi.org/10. /2023  was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted May 11, 2023. ;https://doi.org/10.1101https://doi.org/10. /2023 . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted May 11, 2023. ;https://doi.org/10.1101https://doi.org/10. /2023 870 . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted May 11, 2023. ;https://doi.org/10.1101https://doi.org/10. /2023    . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

A. Thermostability (Protein Thermal Shift)
. CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted May 11, 2023. ; https://doi.org/10.1101/2023.05.11.540323 doi: bioRxiv preprint

A. Viscosity
. CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted May 11, 2023. ; https://doi.org/10.1101/2023.05.11.540323 doi: bioRxiv preprint   . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made A. Thermostability (Protein Thermal Shift)

Figure 5
. CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made  Thermostability (Protein Thermal Shift) . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted May 11, 2023. ;https://doi.org/10.1101https://doi.org/10. /2023