WO2024100251A1 - Formulation of a bevacizumab conjugate - Google Patents

Abstract

The present invention relates to a composition comprising a conjugate between bevacizumab and a dye and a buffer, said composition having a pH of 6.0 to 6.5 and a conductivity of at most 4.0 mS/cm.

Description

New International Patent Applciation claiming priority of EP 22206918.9 Applicant: SURGVISION GmbH Our Ref.: SUR17243PCT Formulation of a bevacizumab conjugate TECHNICAL FIELD OF THE INVENTION [001] The present invention relates to a composition comprising a conjugate between bevacizumab and a dye and a buffer, said composition having a pH of 6.0 to 6.5 and a conductivity of at most 4.0 mS/cm. BACKGROUND ART [002] Bevacizumab (Avastin, Roche AG) is a recombinant, humanized, IgG1 monoclonal antibody (mAb) with specific affinity for human vascular endothelial growth factor (VEGF). VEGF is commonly involved in tumor-induced angiogenesis and its expression is correlated with a poor response following chemoradiotherapy. Bevacizumab (Avastin) is supplied as a concentrate for solution for infusion in vials with a volume of 4 ml containing 100 mg Bevacizumab, or a volume of 16 ml containing 400 mg Bevacizumab. [003] RDye 800CW is a near-infrared organic fluorophore IRDye800CW (800CW-NHS ester), developed by LI-COR Biosciences (Lincoln, NE, USA). This fluorophore can be used for intra- operative imaging in tissue and has been shown in pre-clinical experiments to be safe and non- toxic. The near-infrared organic fluorophore IRDye800CW can be used for nearinfrared imaging in tissue. [004] Bevacizumab-800CW is a conjugation of the fluorescent dye IRDye 800CW to bevacizumab. The conjugation is performed by having bevacizumab diluted in phosphate buffered saline and mixed with IRDye 800CW NHS ester to perform the conjugation of the protein to the dye. Direct amide bonds are formed between the NHS ester and primary amines in the proteinbackbone. Lysine residues are the primary conjugation point for the dye. The calculated molecular weight of bevacizumab-800CW is approximately 152 kDa. The ratio of IRDye 800CW conjugated to bevacizumab is approximately 1.5 molecules dye per antibody. Conjugation to a targeted carrier such as an antibody enables localized fluorescence signal for applications such as fluorescence guided surgery. The product is composed of Bevacizumab- 800CW at a concentration of 1 mg/mL in formulation buffer. The formulation buffer consists of a 50 mM sodium phosphate buffered solution made isotonic with NaCl. The pH of the buffer is 7.0. The product is a green to blueish-green, clear solution. VEGF binding capacity after labeling is comp. [005] However, in order to be sufficiently stable for practical application and to maintain its biological activity, for protein based pharmaceuticals suitable formulations are required. Some formulations of Bevacizumab conjugates are known in the prior art. [006] Wu et al. (AAPS J., 2012 Jun;14(2):252-61) discloses a formulation comprising bevacizumab and the NIRF dye 800CW. Wu investigates the pharmacokinetics of NIRF-labelled bevacizumab in mice. Wu describes the conjugation of bevacizumab and 800CW. However, WU does not provide details on the exact composition of the formulation comprising NIRF-labelled bevacizumab. [007] Ter Weele et al. (Eur J Pharm Biopharm 2016 Jul;104:226-34) discloses formulations comprising bevacizumab and the NIRF dye 800CW. Ter Weele also discloses the process of conjugating 800CW to bevacizumab, whereby the conjugate was finally dissolved at a concentration of 1 mg/ml in 0.9% NaCl. [008] However, there is a need for further formulations of these bevacizumab conjugates. SUMMARY OF THE INVENTION [009] The present invention relates to a composition comprising a conjugate between bevacizumab and a dye and a buffer, said composition having a pH of 6.0 to 6.5 and a conductivity of at most 4.0 mS/cm. [0010] It has been found that bevacizumab conjugated to a buffer has a higher stability in the formulations of the present invention than in formulations of the prior art as discussed above. It has been found that low ionic strength, correlating to a conductivity of at most 4.0 mS/cm contributes to a higher stability of the inventive formulations. Moreover, it has been found that a pH value below 7, such as in the range of 6.0 to 6.5 also adds to a higher stabilization of the inventive formalization. DETAILED DESCRIPTION OF THE INVENTION Definitions [0011] It is noted that as used herein, the singular forms “a”, “an”, and “the”, include plural references unless the context clearly indicates otherwise. Thus, for example, reference to “a reagent” includes one or more of such different reagents and reference to “the method” includes reference to equivalent steps and methods known to those of ordinary skill in the art that could be modified or substituted for the methods described herein. [0012] Unless otherwise indicated, the term "at least" preceding a series of elements is to be understood to refer to every element in the series. Those skilled in the art will recognize, or be able to ascertain using no more than routine experimentation, many equivalents to the specific embodiments of the invention described herein. Such equivalents are intended to be encompassed by the present invention. [0013] The term "and/or" wherever used herein includes the meaning of "and", "or" and "all or any other combination of the elements connected by said term". [0014] The term “less than” or in turn “more than” does not include the concrete number. [0015] For example, less than 20 means less than the number indicated. Similarly, more than or greater than means more than or greater than the indicated number, f.e. more than 80 % means more than or greater than the indicated number of 80 %.^ [0016] Throughout this specification and the claims which follow, unless the context requires otherwise, the word “comprise”, and variations such as “comprises” and “comprising”, will be understood to imply the inclusion of a stated integer or step or group of integers or steps but not the exclusion of any other integer or step or group of integer or step. When used herein the term “comprising” can be substituted with the term “containing” or “including” or sometimes when used herein with the term “having”. When used herein “consisting of" excludes any element, step, or ingredient not specified. [0017] The term “including” means “including but not limited to”. “Including” and “including but not limited to” are used interchangeably.^ PEG means Polyethyleneglycol, CAS Nr.25322-68-3. Polyethlyeneglycole with an average molecular weight of 3350 g/mol is abbreviated as PEG 3350. [0018] The present invention relates to a composition comprising a conjugate between bevacizumab and a dye. [0019] Bevacizumab is a humanized monoclonal antibody (subtype IG1), with a Molecular weight of 149 kDa, which is directed against the Vascular Endothelial Growth Factor (VEGF) and thereby prevents the supply of tumors with new blood vessels. In the European Union, Bevacizumab is admitted for the treatment of six progressed cancer diseases, such as colon, lung, breast, kidney, ovarian and cervical cancer. Bevacizumab as such has the ATC-Code: L01XC07. [0020] Bevacizumab comprises a heavy chain SEQ ID No.1: EVQLVESGGGLVQPGGSLRLSCAASGYTFTNYGMNWVRQAPGKGLEWVGWINTYTGEPTY AADFKRRFTFSLDTSKSTAYLQMNSLRAEDTAVYYCAKYPHYYGSSHWYFDVWGQGTLVT VSSASTKGPSVFPLAPSSKSTSGGTAALGCLVKDYFPEPVTVSWNSGALTSGVHTFPAVL QSSGLYSLSSVVTVPSSSLGTQTYICNVNHKPSNTKVDKKVEPKSCDKTHTCPPCPAPEL LGGPSVFLFPPKPKDTLMISRTPEVTCVVVDVSHEDPEVKFNWYVDGVEVHNAKTKPREE QYNSTYRVVSVLTVLHQDWLNGKEYKCKVSNKALPAPIEKTISKAKGQPREPQVYTLPPS REEMTKNQVSLTCLVKGFYPSDIAVEWESNGQPENNYKTTPPVLDSDGSFFLYSKLTVDK SRWQQGNVFSCSVMHEALHNHYTQKSLSLSPGK and a light chain SEQ ID No.2: DIQMTQSPSSLSASVGDRVTITCSASQDISNYLNWYQQKPGKAPKVLIYFTSSLHSGVPS RFSGSGSGTDFTLTISSLQPEDFATYYCQQYSTVPWTFGQGTKVEIKRTVAAPSVFIFPP SDEQLKSGTASVVCLLNNFYPREAKVQWKVDNALQSGNSQESVTEQDSKDSTYSLSSTLT LSKADYEKHKVYACEVTHQGLSSPVTKSFNRGEC. [0021] The dye may be any dye suitable for labeling proteins, in particular antibodies known to the person skilled in the art. [0022] Preferably, the dye is an near infrared fluoerescence (NIRF) dye. More preferably selected from the group consisting of IRDye 800 CW, IRDye 800 RS, 800 CW PEG, ICG-Der- 02, IntegriSense680PEG, Cy5.5, Alexa Fluor 647, Alexa Fluor 680, most preferably IRDye 800 CW, IRDye 800 RS, 800 CW PEG, particular preferred IRDye 800 CW. [0023] Preferably, the conjugate is Bevacizumab-800 CW, as known in the art, see Ter Weele et al., Eur J Pharm Biopharm 2016 Jul;104:226-34. [0024] The conjugate may be present in the composition in a concentration of 0.5 to 2 mg/ml, preferably 0.7 to 1.5 mg/ml, more preferably 1 mg/ml. [0025] The composition further comprises a buffer. Preferably, the buffer is a succinate or phosphate buffer, more preferably succinate. Preferably, the buffer is present in the composition in a concentration of up to 60 mM, more preferably up to 40 mM, most preferably up to 20 mM [0026] In one embodiment the buffer is succinate, wherein said succinate is present in the composition in a concentration of up to 40 mM, preferably 20 mM. [0027] In one embodiment the buffer is phosphate, wherein said phosphate is present in the composition in a concentration of up to 60 mM, preferably 50 mM, more preferably 20 mM. [0028] The composition exhibts a conductivity of at most 4.0 mS/cm. In a further embodiment, the composition exhibts a conductivity of at most 3.0 mS/cm. In a further embodiment, the composition exhibts a conductivity of at most 2.6 mS/cm. [0029] The composition exhibits a pH of 6.0 to 6.5. [0030] The composition may further comprise trehalose or succrose. The trehalose or sucrose may be present in a concentration of up to 270 mM, preferably 270 mM. [0031] The composition may further comprise an amino acid, preferably alanine. [0032] The amino acid may be present in a concentration of up to 80 mM, preferably 60 mM more preferably 50 mM. [0033] The composition may further comprise an excipient selected from the group consisting of polysorbate 20, PEG 3350 and poloxamer 188. [0034] The polysorbate 20 may be present in 0.01 to 0.04, preferably 0.15 to 0.03, more preferably 0.02 % (v/v). [0035] The PEG 3350 may be present in 0.3 to 1.5, preferably 0.7 to 1.3, most preferably 1.0 % (v/v). [0036] The poloxamer 188 may be present in 0.2 to 0.08, preferably 0.03 to 0.07, more preferably 0.05 % (v/v). [0037] The composition may be lyophilized. General conditions for lyophilization are known in the art and exemplified in the examples. [0038] In one embodiment the formulation comprises 0.5 to 2 mg/ml, preferably 0.7 to 1.5 mg/ml, more preferably 1 mg/ml of Bevacizumab-800 CW, succinate in a concentration of up to 40 mM, preferably 20 mM, sucrose in a concentration of up to 270 mM, preferably 270 mM, and PEG 3350 in 0.3 to 1.5, preferably 0.7 to 1.3, most preferably 1.0 % (v/v). [0039] Further, the invention comprises the following items: 1. A composition comprising a conjugate between bevacizumab and a dye and a buffer, said composition having a pH of 6.0 to 6.5 and a conductivity of at most 4.0 mS/cm. 2. The composition of item 1, wherein said buffer is succinate. 3. The composition of any one of the preceding items, further comprising trehalose. 4. The composition of any one of the preceding items, further comprising an excipient selected from the group consisting of polysorbate 20, PEG 3350 and poloxamer 188. 5. The composition of any one of the preceding items, further comprising alanine. 6. The composition of any one of the preceding items, wherein said succinate has a concentration of up to 40 mM, preferably 20 mM. 7. The composition of any one of the preceding items, wherein said trehalose has a concentration up to 270 mM, preferably 270 mM. 8. The composition of any one of the preceding items, wherein said polysorbate 20 is present in 0.02 % (v/v). 9. The composition of any one of the preceding items, wherein said PEG 3350 is present in 1.0 % (v/v). 10. The composition of any one of the preceding items, wherein said poloxamer 188 is present in 0.05 % (v/v). 11. The composition of any one of the preceding items, wherein said dye is a NIRF dye. 12. The composition of item 11, wherein said NIRF dye is IRDye 800 CW. 13. The composition of any one of the preceding items, wherein said conjugate is through an NHS ester. 14. The composition of any one of the preceding items, wherein said bevacizumab has the VH amino acid sequence shown in SEQ ID NO: 1 and the VL amino acid sequence shown in SEQ ID NIO: 2. 15. The composition of any one of the preceding items, which is lyophilized. EXAMPLES OF THE INVENTION 1. Introduction [0040] The goal was to develop an alternative stable formulation to replace this formulation for Bevacizumab-800CW, preferably without particulates. 2. Materials and Methods 2.1 Materials [0041] The protein, equipment and reagents used for the study are listed in Tables 1-4. Table 1: Samples Received

Table 2: List of Reagents

Table 3: List of General Supplies

Table 4: List of Instruments and Specialized Supplies

2.2 Methods 2.2.1. Simulated shear stress [0042] Therapeutic proteins, such as Beva-800CW, are sensitive molecules held together by weak physical interactions and chemical bonds. The manufacturing of therapeutic proteins involves several processes that result in stress on the molecule in the form of shear, which may change its three-dimensional structure. The formulations were subjected to simulated-shear stress at room temperature to monitor the effect of this mechanical stress on Beva-800CW. Experiments were performed in 3 ml Reacti-Vials™ (Thermo Scientific) with matching triangular stir bars and stirred at 600 rpm for a pre-determined time. 2.2.25 cycles freeze-thaw [0043] Freeze-Thaw (FT) of a protein may have an impact on protein stability, which can be usually mitigated by proper selection of buffer systems and additives. To understand the impact of FT and to ensure adequate handling of Beva-800CW throughout the project, the effects of repeated (5 times) freezing and thawing were studied. Aliquots of the protein in formulation buffer were repeatedly frozen at -80°C and thawed at ambient temperatures for a maximum of 5 cycles before analysis. 2.2.3 Dialysis [0044] Dialysis promotes the exchange of aqueous buffer with the protein of interest via selective diffusion through a semi-permeable membrane with a known molecular weight cut off. Beva- 800CW was dialyzed at 2-8°C with slow stirring using Slide-A-Lyzer ® Dialysis Cassettes of 3-12ml capacity or 12-30ml capacity and 10 KD molecular weight cut off. The dialysis was carried out over 24 hours with two buffer exchanges after 2 and 4 hours. The final dialysis step took place overnight. 2.2.4 Lyophilization [0045] Lyophilization was used as a tool to develop and to select a formulation that could be lyophilized, if decided to do so. A generic lyophilization cycle was used and is given below: Table 5: Generic lyophilization cycle used for the formulation development efforts of Bevar- 800CW

2.2.5 Appearance [0046] All samples were visually inspected for appearance using a 22W light source. The visual appearance was conducted to note the color, clarity, and presence of particulates. The visual appearance description was noted upon observing under a magnifying glass. [0047] Although digital photos were taken, they did not show the fine particles observed under the magnifying lens. 2.2.6 pH [0048] The pH of the sample was measured using a Thermo Scientific, Orion Star Model A 211 pH meter equipped with a Ross PerpHecT microelectrode, Model 8220BNWP. For buffer solutions preparation, a triode electrode was used to measure the pH (Thermo Scientific, US Gel-filled Ultra Triode Electrodes). The instruments were standardized using pH 4, 7 and 10 buffers traceable to NIST Standard Reference Solutions from Thermo Scientific. 2.2.7 UV-VIS Spectroscopy, Ratio of Fluorochrome/ Antibody Content and Antibody Content [0049] The sample was diluted in product buffer to an antibody concentration of approximately 0.3 – 0.7 mg/ml and measured in a wavelength range of 240 nm – 1000 nm. For determination of the UV-Vis spectrum, the wavelengths of the 3 peak maxima were evaluated. The ratio of fluorochrome/antibody content is calculated using following equation: The molar extinction coefficient for the antibody is 243340 M-1 cm-1 and the molar extinction coefficient of the fluorochrome is 240000 M-1 cm-1. The antibody content was calculated using following equation:

The specific extinction coefficient k of the antibody is 1.7 l/g (1 cm path length) according to the manufacturer’s information (Roche Diagnostics GmbH). V is the dilution factor of the sample. The content was corrected by the absorption of the fluorochrome at 280 nm which is 3.0 % of its absorption at 780 nm (0.03 x A780) according to the manufacturer (Manual IRDye 800CW Protein Labelling Kit High MW, LI-COR). All samples were centrifuged at 12,000 rpm for 5 min to remove any aggregates before absorbance measurements were taken. 2.2.8 Right Angle Light Scattering (RALS) [0050] Conformational changes in protein structure due to stress were detected by intrinsic fluorescence (IF) and right-angle light scattering (RALS). IF measures possible changes in the tryptophan (Trp) environment arising from stress-induced conformational changes. Depending on the polarity of its environment, the Trp fluorescence emission λmax can range from 305 to 355 nm (305 nm in a non-polar environment and 355 nm in a polar environment). Therefore, a ratio of intensities can yield information about the unfolding/structural response to external stimuli. [0051] Instrument settings for IF and RALS were established before the sample analysis. A 400 μl volume of sample was placed in a 500 μl quartz cuvette. Prior to changing the settings in the computer program, the samples were placed into an Aminco Bowman Series 2 Luminescence Spectrometer (Model AB2). [0052] Ambient RALS monitors macroscopic changes in an otherwise soluble molecule when transferring to an aggregated state (soluble and/or insoluble). RALS measures light scatter at a 90° angle. When a molecule unfolds and aggregates, the light scattering increases. On a scale of 0-10 (instrument scale of min/max intensity), the closer to 10, the more unfolded the molecule and the more aggregation is present. For Beva-800CW, excitation and emission were set at 320 nm and the RALS readings were taken between 400-475V. The RALS was measured at ambient conditions for physically or thermally stressed samples. [0053] For unstressed samples at T0, a temperature ramp of 20°C to 80°C at 2°C/min was applied with a data collection rate of every 30 seconds. Both RALS and IF measurements were carried out simultaneously during the temperature ramp, and the transition temperatures such as Tagg and Tm were obtained from their respective inflection points on the curve. The emission ratio at 320 nm/320 nm and 320 nm/360 nm was plotted for RALS and IF, respectively. The emission ratios were predetermined based on the guanidine denaturation experiment. The RALS voltage was set at 475V while the IF voltage was adjusted at the start of each iteration and continued with the pre-set voltage through-out that iteration. 2.2.9 MicroFlow Imaging (FlowCam) [0054] Subvisible particles in the samples were monitored by FlowCam (Fluid Imaging Technologies, Inc.). The instrument was focused using 10 μm polystyrene beads at 3000/mL (NIST standard). Samples were aspirated at 0.08 mL/min through a 100 μm x 1 mm flow cell, and images of the particles were taken with a 10X optics system. Particles per mL were calculated by operating Visual Spreadsheet software. In between samples, the flow cell was rinsed with hot deionized water, 70% 2-propanol, or placebo. [0055] NOTE: All samples were degassed under vacuum prior to analysis. However, variability could be seen in T7 and T14 thermally stressed samples with respect to particle count by FlowCam. Samples may require equilibration at room temperature prior to analysis. 2.2.10 Size-Exclusion Chromatography HPLC (SEC) [0056] The method parameters are listed in Table 6. All samples were centrifuged at 12,000 rpm for 5 minutes to remove any solids formed before injection. Table 6: Beva-800CW Size-Exclusion Chromatography Method Parameters

Table 7: SEC Representative Sample Analysis Sequence

NOTE: The method was modified to analyze the ratio of 775 nm to 280 nm of the main peak. 2.2.11 Ion-Exchange Chromatography HPLC (CIEX) The method parameters are listed in Table 8. All samples were centrifuged at 12,000 rpm for 5 min to remove any solids formed before injection. Table 8: Beva-800CW Ion-Exchange Chromatography Method Parameters

Table 9: CIEX Representative Sample Analysis Sequence

2.2.11 HPLC system suitability of Beva-800CW [0057] System suitability working reference standard was prepared using the Beva-800CW received from SurgVision at 1mg/mL. To 600 μl Eppendorf tubes, 100µl aliquots were added and 100 of said tubes were stored at 2-8°C. A fresh vial was removed from 2-8°C chambers, before each SEC/CIEX-HPLC sequence, and was used as a system suitability/ working reference standard. The percent RSD of the system suitability was calculated for each sequence. A percent RSD ≤ 2% was considered acceptable. 2.2.12 SDS - Capillary Electrophoresis (SDS-CE), non-reduced and reduced [0058] SDS-CE was performed using a bare fused silica with 50 μm inner diameter and a length of 30 cm. A sample amount corresponding to 250 μg antibody was desalted with water for injection (WFI) in centrifugal filter units with a cut-off of 10 kDa. The SDS buffer (100 mM Tris/HCl pH 9.0 / 1% SDS) spiked with a 10kDa molecular weight standard was added to each sample. For the non-reduced samples, iodoacetamide was added and the protein was denatured by incubating for 10 minutes at 70 °C. For the reduced samples, β-Mercaptoethanol was added and the protein denatured by incubating at 20 minutes at 70 °C. Electrophoresis was conducted at maximum of 30 kV and 300 μA and the electropherogram was recorded at 214 nm. [0059] Calculation of relative retention time: For each run, the relative retention time was calculated by dividing the sample retention time by the retention time of the 10 kDa molecular weight standard. 2.2.13 SDS gel electrophoresis (SDS-PAGE) [0060] The samples were run by SDS-PAGE and the method parameters are listed in Table 10 below. Table 10: Method parameters for Non-Reducing SDS-PAGE

3. Overview over experiments [0061] Bevacizumab-800CW is comprised of bevacizumab conjugated to IRDye 800CW. The ratio of IRDye 800CW conjugated to bevacizumab is approximately 1.5 molecules dye per antibody. The product is a green to blueish-green, clear solution. The goal was to develop a stable formulation for Bevacizumab-800CW at 1mg/ml with low subvisible particles and no aggregation for intravenous administration. The research project has been carried out SEC- HPLC, CIEXHPLC, UV-Vis spectrophotometry, and SDS-CE (non-reduced and reduced) methods.. Since the UV-Vis method 780/280 molar ratio calculation included the dye bound antibody as well as the free dye, the SEC analysis was modified to include the calculation of the ratio between the antibody monomer AUC at 775 nm over the antibody monomer AUC at 280 nm (iterations 1-4). This modification allowed for the analysis of the dye bound specifically to the antibody, and in the absence of free dye In addition, the samples were run by non-reduced SDS-PAGE and scanned using the Licor scanner to monitor for the reduction in fluorescence intensity which might indicate a loss of bound dye. A pre-formulation assessment of the Beva- 800CW was performed. A series of experiments designed to understand the protein’s strengths and weaknesses when exposed to a wide variety of environmental and physical stress has been planned. The assessment was performed on Beva-800CW formulation containing 50 mM Na-Phosphate and 100 mM NaCl at pH 7.0. Testing of Beva-800CW led to the following conclusions: • Beva-800CW undergoes thermally induced conformational changes resulting in transitions at ~60°C based on IF and RALS. Therefore, during the formulation development program, thermal stress for formulation screening was performed at 50°C. • Based on the emission spectra shift in the presence of guanidine hydrochloride used as a denaturing agent. IF measurements of Beva-800CW were performed using the emission ratio of 320/360. • Beva-800CW is extremely sensitive to simulated shear stress and to a lesser extent sensitive to five cycles of freeze-thaw. This was evidenced by lower recoveries in SEC and CIEX HPLC, lower molar ratio by UV-Vis and increased particles by FlowCam. Pre-formulation and formulation development experiments followed a rationally designed program and included the systematic screening/evaluation of buffers, stabilizers, and surfactants. Excipients were first screened individually and then in combination to identify a final formulation via short- term accelerated stability testing and simulated-shear stress. The formulation development allowed for identifying the lead formulation at 1 mg/ml Beva 800CW. The screening process also utilized lyophilization as a tool and it was demonstrated that the lyophilized formulations of Beva- 800CW had higher stability for up to 14 days at 50°C as compared to the corresponding liquid formulations. Table 11: Summary results of all iterations

0 l l l l l - t t

[0062] Based on the excipient screening program, the lead formulation was identified as 1 mg/ml Beva-800CW in 20 mM succinate, 270 mM sucrose, and 0.02% tween 20 at pH 6.0. This formulation protects Beva-800CW against thermal and shear sensitivity and meets the criteria of low number of visible particles required for manufacturability. In addition, the formulation is able to be lyophilized. Lyophilization improved the thermal stability of the formulation leading to higher protein recovery and dye/protein ratio when compared to liquid formulations. 4. Iteration 1: Influence of pH and buffer. [0063] The buffer screening was carried out with eleven buffers (Table 12). Beva-800CW was received in 50 mM Na-Phosphate and 100 mM NaCl at pH 7.0 and was designated as the current formulation (CF). The current formulation (CF) was studied as the reference. The screening study was initiated with dialysis of Beva-800CW in the respective buffer, followed by a concentration step using the Amicon Ultra-15 Centrifugal Filter, (Ultracel 10K) to approximately 1 mg/ml for the protein in the Gly-Gly buffers (pH 7.5 and 8). All samples were filtered using a 0.2 µm cellulose acetate filter. The protein recovery after dialysis is given in Table 13 and the concentration after centrifugal filter concentration and filtration is given in Table 14. The concentrations were measured by the Nanodrop UV/Vis spectrophotometer. The protein recovery was >90% for all samples. [0064] Aliquots of 1.4 ml were dispensed into 3 ml glass vials, closed with a gray rubber stopper and aluminum crimped for the study. The container closure is listed in Table 3. A specialized vial (ReactiVial) was used for the simulated shear stress study. [0065] Aliquots of the formulations were subjected to stress conditions such as freeze-thaw (F/T) at - 80°C to ambient temperatures for five cycles, simulated shear stressed at 600 rpm stirring at room temperature for 3 hours, and incubation at 50°C for up to fourteen days. The sample aliquots were designated as T0, T0 FT, T0 Shear, T7/50°C or T14/50°C. T0 represented the control without any stress, and all stress samples were compared with the control (T0) to evaluate the effect of stress. [0066] NOTE: Due to lack of specificity of the UV-Vis spectrophotometric analysis of the 780/280 (molar ratio to the dye bound/associated to the antibody vs unbound/dissociated dye), the SEC chromatograms at 775 nm were integrated in addition to the 280 nm chromatogram (iterations 1-4). The SEC chromatogram (775 nm) was used to integrate the _{dye bound antibody peak} ^{, which is well separated by the SEC} _{column. The AUC ratio of the} antibody monomer peak (MP) at 775/280 was calculated. Table 12: Buffer screening iteration 1: list of formulations

Table 13: Buffer screening: dialysis concentration, iteration 1 buffer and pH screen

NOTE: CF was not dialyzed Table 14: Buffer screening: final concentration after concentration and filtration, iteration 1 buffer and pH screen

NOTE: CF was not dialyzed [0067] A SEC analysis at 280 nm for total percent of HMW T0 samples has been carried out. The results showed, see table 15, that the succinate and histidine buffers showed a lower total percent HMW when compared to phosphate and glycylglycine. Table 15: SEC 280 nm analysis of Beva-800CW at 280 nm for T0, iteration 1 buffer and pH screen

[0068] A FlowCam analysis for T0 samples has been carried out, see table 16. At T0, the lowest total particle counts were seen with formulations containing succinate pH 5.0, 6.5, glycylglycine pH 7.5 and histidine pH 5.5 and 6.0. The formulation containing phosphate showed relatively higher particle counts at all pH values compared to the succinate, glycylglycine and histidine buffers. Table 16: Flow imaging particle analysis for T0, iteration 1 buffer and pH screen

[0069] The samples were analyzed by SEC-HPLC, see table 17. Upon five-freeze-thaws, the samples formulated as CF and the formulations containing phosphate buffers pH 7.5 and 8 showed an increase in the percent HMW, whereas there were no increases seen with the other formulations. Table 17: SEC 280 nm analysis of Beva-800CW at 280 nm for T0 FT, iteration 1 buffer and pH screen

[0070] A FlowCam analysis for T0 FT samples and T0 has been carried out, see table 18. There appeared to be a large increase in particle number following freeze-thaw for all formulations, with the largest seen with the formulations containing phosphate. The formulations containing succinate showed the least increase in particles. Table 18: : Flow imaging particle analysis for T0_FT, iteration 1 buffer and pH screen e 6.0 2230 843 92 8

0 2488 58354 58419 12448

919 [0071] The Effect of different pH buffers on RALS at 475V for T0 and T0 FT samples has been measured. The formulations containing succinate showed the least increase in scattering from T0 upon five freeze-thaws compared to the other formulations. Table 19: RALS analysis for T0_FT, iteration 1 buffer and pH screen

[0072] : UV-Vis analysis for antibody recovery (A) and 780/280 molar ratio (B) for T0 FTsamples and T0 has been carried out, see table 20. Analysis of antibody concentration showed that the formulations containing succinate had the least reduction in recovery relative to T0 upon freeze- thaw. Table 20: Antibody concentration for T0 FT, iteration 1 buffer and pH screen

[0073] A FlowCam analysis for T0 FT samples and T0 has been carried out, see table 21. There appeared to be a large increase in particle number following freeze-thaw for all formulations, with the largest seen with the formulations containing phosphate. The formulations containing succinate showed the least increase in particles. Table 21: Flow imaging particle analysis T7/50°C and T14/50°C, iteration 1 buffer and pH screen

[0074] The effect of different pH buffers on RALS at 475V for T0 and T0 FT samples has been measured, see table 22. The formulations containing succinate showed the least increase in scatteringfrom T0 upon five freeze-thaws compared to the other formulations.

[0075] An UV-Vis analysis in respect to antibody recovery (A) and 780/280 molar ratio (B) for T0 FT samples and T0 has been carried out, see table 23. Analysis of antibody concentration showed that the formulations containing succinate had the least reduction in recovery relative to T0 upon freeze-thaw. Table 23: Antibody concentration for T0 FT, iteration 1 buffer and pH screen

[0076] A SEC analysis at 775 nm % MP recovery at T7/50°C and T14/50°C relative to T0 has been carried out. Based on the above results, that at 775 nm, for the T7/50°C and T14/50°C samples, the highest main peak purity, main peak recovery, and 775/280 ratio for the main peak was seen with the formulation containing succinate buffer at pH 6.0. Overall, the protein recovery was reduced in all buffers after time and temperature dependent stress. Table 24: SEC analysis at 775 nm T7/50°C and T14/50°C, iteration 1 buffer and pH screen

[0077] A FlowCam analysis at T7/50°C, T14/50°C and T0 has been carried out. Increase in particles of different samples. The data showed that at T7/50°C the formulations containing succinate at pH 6.0 and 6.5, as well as the formulations containing phosphate at pH 7.0 and 7.5 had the smallest increase in particles relative to T0. Table 25: Flow imaging particle analysis T7/50°C and T14/50°C, iteration 1 buffer and pH screen

. . . . . . . . . . . T

[0078] Based on molar ratio recovery, the formulation containing succinate showed relatively higher recovery after time and temperature-dependent stress compared to the other formulations. See table 26. Table 26: UV-Vis antibody concentration and 780/280 molar ratio T7/50°C and T14/50°C, iteration 1 buffer and pH screen

[0079] In summary, based on the overall results, in respect to stability, samples comprising succinate showed the most promising results. [0080] Overall, the succinate buffer at pH 6.0 showed the highest stability to physical and thermal (time and temperature) stress when compared to the other buffers. 5. Iteration 2: Measurements of ionic strength and in respect to influence of ionic strength. [0081] The goal of iteration 2 was to use the most promising formulation from iteration 1 as the base formulation to screen for the stabilizers of Beva-800CW, from class of amino acids and salts (at different ionic strengths). The Beva-800CW formulation in presence of ionic strength and/or amino acids with maximum stability to physical (freeze-thaw and simulated shear) and thermal (time and temperature) stress were identified and thus, classified as the leading excipients from this iteration. Preparation of formulations [0082] The lead formulation from iteration 1 (succinate buffer at pH 6.0), was used to assess the effect of ionic strength and the use of amino acids as potential stabilizers. The formulations used in the study are given in Table 109. The Beva-800CW was received in 50 mM NaPhosphate and 100 mM NaCl at pH 7.0 and designated as the current formulation (CF). Beva- 800CW in the current formulation (CF) as an undialyzed sample (as received) was dialyzed into CF and was used as the reference. The screening study was initiated with dialysis of Beva-800CW in CF at pH 7.0 and in the succinate buffer at pH 6.0. Aliquots of the dialyzed samples were spiked with salt or amino acids according to Table 109 and the pH was readjusted to 6.0. All samples were filtered using a 0.2 µm PVDF filter. The protein recovery after dialysis is given in Table 110. The concentrations were measured by Nanodrop UV/Vis spectrophotometer. The protein recovery was >90% for all samples. Aliquots of 1.5-2 ml were dispensed into 5 ml glass vials, closed with a gray rubber stopper and aluminum crimped for the study. The container closure system is listed in Table 3. A specialized vial (ReactiVial) was used for the simulated shear stress study. Formulation aliquots were subjected to stress conditions such as freeze-thaw (F/T) at -80°C and at ambient temperatures for five cycles, simulated shear stress at 600 rpm stirring at room temperature for 8 hours and incubation at 50°C up to fourteen days. The samples were designated as T0, T0 FT, T0 Shear, T7/50°C or T14/50°C. T0 represented the control in the absence of any stress and all stress samples were compared to the control. Table 27: Screening for ionic strength and amino acid stabilizers, iteration 2: List of formulations

[0083] The recovery of the monomer peak at 775 nm appeared lower with increasing NaCl concentration, see table 28. Among the amino acids, the formulation containing arginine showed the lowest 775 nm monomer peak recovery. Table 28: SEC-HPLC analysis at 775 nm for monomer recovery

[0084] Total particles distribution has been determined by FlowCam for T7/50°C and T14/50°C, see table 29. Overall, there was an increase in the number of sub-visible particles across allformulations relative to CF. With increased salt concentration, there was a larger increase in particle counts. The formulation containing alanine showed the lowest particle count at T7/50°C. At T14/50°C, no clear trend was observed for formulations containing amino acids. Table 29: Total particles by FlowCam for T7/50°C and T14/50°C, iteration 2 ionic strength and amino acid screening Suc/ 50m NaCl 222 115 15 8 0 360 2538 6 2846

774 ¹⁵ 0 9021 Suc/ 50m NaCl 28670 6 12683 4 9087 192 46 42286

5 Messurement of Ionic strength [0085] Several formulations, which have previously shown to stabilize BEVA800, were prepared and electrical conductivity (EC) and ionic strength (IS) were assessed. Stock solutions of Succinate 20 mM (Merck Millipore, cod. 1.00682), Phosphate 50 mM and Phosphate 20 mM (Sodium dihydrogen phosphate monohydrate, Merck Millipore, cod. 1.06345; Disodium hydrogen phosphate, Sigma Aldrich, cod. 1.0658) were prepared fresh. 32 solutions were arranged from stock in aliquots of 25 mL each. The excipients Trehalose 270 mM (Sigma Aldrich, cod. T0167), Tween20 (Sigma-Aldrich, cod. P7949), PEG3350 (Sigma-Aldrich, cod. 1546547-USP), Poloxamer 188 (Sigma-Aldrich, cod. P2164009) and Human Serum Albumin (Sigma-Aldrich, cod. A9511) were added to the buffer solutions according to the experimental plan. Each solution was corrected for the pH value as listed below. In addition, a solution of Beva800 1 mg/mL in 50 mM Phosphate buffer + 100 mM NaCl pH 7 (BEVA800) was also evaluated for comparative purposes. [0086] EC was measured by using VWR® CO_3100L conductivity meter. The protocol included a preliminary conditioning step of the sensor (at least 30 minutes in deionized water), useful to identify any instrumental drift, and a calibration/verification step to assess the cell constant of the conductivity measuring cell using 0.01 mol/L KCl standards (Mettler Toledo, cod.51302049 and 30111140). Data was acquired at room temperature (23°C ca.). [0087] IS was determined from EC using an empirical method based on a linear relationship between electrical conductivity and ionic strength (Snoeyink, V.L. and D. Jenkins, Water Chemistry, John Wiley & Sons, New York, 1980), expressed by the formula: [0088] EC (µS/cm) = 6.2104 × I (mol/L), equivalent to I (mol/L) = 1.610-5 × EC (µS/cm) Table 30: Experimental data

Table 31: Furher measurements related to ionic strength

[0089] Conclusions from iteration 2 ionic strength and amino acid screening • Formulations containing NaCl showed reduced stability to physical and thermal (time and temperature) stress with increasing concentrations of NaCl • With and without stress, the formulation containing the amino acid arginine showed a lower 775/280 ratio by SEC and a lower 780/280 ratio by UV-Vis, and the formulation containing aspartic acid had a lower recovery and a higher HMW content. • Overall, the formulation containing the amino acid alanine showed the greatest stability to physical and thermal stress based on the cumulative scoring. Therefore, from iteration 2, the lead formulation was selected as 20mM succinate and 50mM alanine at pH 6.0. 6. Iteration 3: Excipient screening for carbohydrates and surfactants [0090] The goal of iteration 3 was to use the lead formulation from iteration 2 as the base formulation to screen for stability of Beva-800CW in the presence of carbohydrates and surfactants. Formulations containing alanine (lead from iteration 2) with and/or without a carbohydrate or surfactant were also included in the study. The Beva-800CW formulation from iteration 3 containing carbohydrate and surfactant providing maximum stability to physical (freeze-thawand simulated shear) and thermal (time and temperature) stress, was identified as the lead formulation from iteration 3. Here, lyophilization was used as a screening tool for formulation development in the event that there is a need to lyophilize Beva-800CW. Preparation of formulations [0091] The lead formulation from iteration 1, succinate buffer at pH 6.0, was used to assess the effect of carbohydrates and surfactants as stabilizers on liquid and lyophilized formulations.The formulations used in the study are given in Table 33. The Beva-800CW was received in 50 mM Na-Phosphate and 100 mM NaCl at pH 7.0 and was designated as the current formulation (CF). Beva-800CW dialyzed in CF was studied as the reference (CF in CF). The screening study was initiated with dialysis of Beva-800CW in CF at pH 7.0 or the succinate buffer at pH 6.0. Aliquots of the dialyzed sample were spiked with the excipient according to Table 33 and the pH was re-adjusted to 6.0. All samples were filtered using a 0.2 µm PVDF filter.. The concentrations were measured by Nanodrop UV/Vis spectrophotometer. The protein recovery was >90% for all samples. Aliquots of 1.5-2 ml were dispensed into 5 ml glass vials, stoppered with a gray rubber stopper and aluminum crimp topped for the study as liquid formulations. A specialized vial (ReactiVial) was used for the simulated shear stress study. Aliquots of 1.5 ml in 5 ml glass vials were also lyophilized. A generic lyophilization cycle was used as shown in the table below. Table 32: Lyophilization cycle, iteration 3 carbohydrate and surfactant screening

[0092] Following lyophilization, the samples were stoppered with a gray rubber stopper and aluminum crimp topped for the study as lyophilized formulations. The container closure is listed in Table 3. Liquid formulation aliquots were subjected to stress conditions such as freeze-thaw (F/T) at -80°C for five cycles, simulated shear stressed at 600 rpm stirring at room temperature for 8 hours, and incubated at 50°C up to fourteen days. The samples were designated as T0 Liq, T0 FT, T0 Shear, T7 Liq/50°C, and T14 Liq/50°C. T0 represented the control without any stress and all stress samples were compared to the control to evaluate the effect of stress. Similarly, the lyophilized formulations were studied after incubation at 50°C up to fourteen days. The samples were designated as T0 Lyo, T7 Lyo/50°C, and T14 Lyo/50°C. Prior to analysis, the lyophilized formulations were reconstituted in WFI. The reconstitution volume was determined by taking the vial weight filled with the sample, before and after lyophilization. The reduction in weight was taken as the reconstitution volume with water for injection (WFI, Table 2). For reconstitution, the appropriate volume of WFI was added slowly to the cake along the inner wall of the vial, the vial was gently swirled while keeping flat on a table, and the time taken for complete dissolution of the cake was recorded. Table 33: List of formulations, iteration 3 carbohydrate and surfactant screening

. [0093] A FlowCam analysis for total particle distribution of T0 Lyo and T0 Liq has been carried out, see table 34. Amongst the carbohydrates evaluated, the formulation containing sucrose had arelatively lower number of sub-visible particles. The formulations containing surfactants had lower number of particles. Formulations containing tween 20, tween 80 and PEG 3350 had a lower particle count than with poloxamer 188. A sharp increase in sub-visible particles was seen with the CF and the formulation containing mannitol relative to T0 Liquid after lyophilization. The particle count remained low for the formulations containing surfactants.. Table 34: FlowCam analysis for total particles of T0 FT samples relative to T0 Liq, iteration 3 carbohydrate and surfactant screening

[0094] A UV-Vis analysis in respect to antibody recovery has been carried out. Following freeze-thaw, there was a decrease in antibody recovery and molar ratio for formulations without surfactants, whereas the formulations containing surfactants showed no reduction in recovery and molar ratio. The formulation containing sucrose showed a higher antibody recovery relative to those containing trehalose or mannitol. Table 35: Antibody concentration and molar ratio, iteration 3 T0 FT

[0095] A SEC-HPLC analysis for percent total recovery at 280 nm has been carried out, see table 36. Formulations containing surfactants showed a lower decrease in total recovery at 280 nm. Among formulations containing different carbohydrates, the sucrose formulation showed better recovery. Table 36: SEC-HPLC analysis of Beva-800CW at 280 nm for T0 Shear, iteration 3 carbohydrate and surfactant screening

[0096] A CIEX-HPLC analysis for percent total recovery of T0 Shear samples relative to T0 Liq has been carried out, see table 37. [0097] The formulation containing sucrose showed better recovery relative to formulations containing trehalose or mannitol. Except for poloxamer 188 (lower recovery), the other surfactants showed similar recovery following simulated shear. Formulations containing surfactants showed better recovery compared to those without surfactants. Table 37: Analysis by CIEX-HPLC FLD detector for T0 Shear, iteration 3 carbohydrate and surfactant screening

[0098] A UV-Vis analysis for antibody recovery has been carried out. Following simulated shear, there was a reduction in antibody recovery and molar ratio across all formulations. The decrease in recovery and molar ratio was less for the formulations containing surfactants. The formulation containing sucrose among carbohydrates and formulations containing tween 20 among surfactants, showed relatively higher recoveryand molar ratio. Table 38 : Antibody concentration and molar ratio for T0 Shear, iteration 3 carbohydrate and surfactant screening

[0099] Total particle distribution has been determined by FlowCam for thermally stressed liquid samples, see table 39. Overall, there was an increase in sub-visible particles across all formulations after being subjected to thermal stress. Formulation containing Suc-Ala and Tween 20 showed the least change over time and temperature. Table 39: Flow imaging particle analysis for T0 Liquid, T7 and T14 Liquid/50°C, iteration 3 carbohydrate and surfactant screening Suc/ reh/ olo 553

3 84 338 056 710 131 114 143 1 8

06 [00100] A SEC-HPLC has been carried out in order to determine the % of total HMW at 280 nm for T9 and T16 Lyo/50°C, see table 40. All formulations showed lower increase in percent HMW relative to CF, except for mannitol (consistent with visible precipitate). No differences were observed in formulations containing different carbohydrates. Formulations containing PEG and poloxamer showed lower percent HMW compared to tween 20 and 80. Table 40: SEC-HPLC analysis of Beva-800CW at 280 nm for T9 Lyo/50°C, iteration 3 carbohydrate and surfactant screening.

[00101] An analysis of total particles by FlowCam for thermally stressed samples, has been carried out. Overall, formulations containing surfactants showed a low number of sub- visible particles; lower than the corresponding liquid formulations. Table 41: Flow imaging particle analysis for T0 Lyo and T9 Lyo/50°C, iteration 3 carbohydrate and surfactant screening uc/ reh/ olo 153 38 8 0 0 199 161 8 0 0 0 169 61 15 8 0 0

84 [00102] Conclusions from iteration 3 carbohydrate and surfactant screening • Based on analytical characterization, the addition of surfactants to the formulations improved the stability of Beva-800CW to physical and thermal stress. Surfactants also decreased sub- visible particle counts by FlowCam and visual observations for particles (barely visible particles). • Lyophilization improved the stability of all formulations to thermal (time and temperature) stress as observed by a higher monomer recovery and 775/280 ratio by SEC and 780/280 molar ratio by UV-Vis. • Overall, from iteration 3, formulations containing both the carbohydrate, sucrose, and the surfactants, tween 20 and PEG, appeared to show higher stability to physical and thermal stress and were selected as the lead formulations. 7. Iteration 4: Final screening of lead buffer and excipients [00103] The goal of iteration 4 was to execute a screen leading to the identification of a combination of buffers and excipients from the previous iterations and to perform a final screen of formulations that provides the highest stability for Beva-800CW. The screen included the lead excipients from iterations 2 and 3 along with the two leading buffers from iteration 1, 20 mM succinate and 20mM phosphate buffers as the base formulation. The formulation which gave the Beva-800CW molecule maximum stability to physical (freeze-thaw and simulated shear) and thermal (time and temperature) stress was identified as the lead formulation from iteration 4. Similar to iteration 3, lyophilization was used as a screening tool for formulation development in the event lyophilization of Beva-800CW is needed. Preparation of formulations [00104] The lead buffers and excipients from iterations 1-3 were studied in iteration 4 (the final iteration). Here, in addition to the succinate buffer at pH 6.0, the phosphate buffer at pH 6.0 was also included. The formulations used in the study are given in Table 20. The Beva- 800CW was received in 50 mM Na-Phosphate and 100 mM NaCl at pH 7.0 and was designated as the current formulation (CF). Beva-800CW was dialyzed into CF and was used as the reference formulation (CF in CF). The screening study was initiated with dialysis of Beva-800CW into CF at pH 7.0 or the succinate/phosphate buffer pH 6.0 with a carbohydrate and the amino acid, alanine. Aliquots of the dialyzed sample were spiked with surfactant (Tween 20 or PEG 3350) according to Table 20 and the pH was verified. All samples were filtered using a 0.2 µm PVDF filter (Table 3).. Protein concentrations were measured by the Nanodrop UV/Vis spectrophotometer. The protein recovery was >90% for all samples. For liquid formulations, aliquots of 2 ml were dispensed into 5 ml glass vials, stoppered with a gray rubber stopper and aluminum crimp topped for the study. A specialized vial (ReactiVial) was used for the simulated shear stress study. Aliquots of 2 ml in 5 ml glass vials were also lyophilized. Here, lyophilization was used as ascreening tool for formulation development in the event lyophilization of Beva- 800CW is needed. A generic lyophilization cycle was used and is shown in the Table 43. Following lyophilization, the samples were stoppered with a gray rubber stopper and aluminum crimp topped for the study as lyophilized formulations. Aliquots of the liquid formulation were subjected to stress conditions which included: five cycles of freeze-thaw (F/T) at -80°C to ambient temperature, simulated shear stress with stirring at 600 rpm, room temperature for 8 hours, and thermal stress by incubating at 50°C for up to fourteen days. The samples were designated as T0 Liq, T0 FT, T0 Shear, T7 Liq/50°C or T14 Liq/50°C. T0 represented the control in the absence of stress and results from the testing of all stress samples were compared to the control values enabling an evaluation of the effects of stress. [00105] Similarly, the lyophilized formulations were studied after incubation at 50°C up to fourteen days. The samples were designated as T0 Lyo, T7 Lyo/50°C or T14 Lyo/50°C. Prior to analysis, the lyophilized formulations were reconstituted in WFI. The reconstitution volume was determined by taking the weight of the sample filled vial before and after lyophilization. The reduction in weight was taken as the reconstitution volume with water for injection. For reconstitution, the appropriate volume of WFI was added slowly to the cake along the inner wall of the vial, the vial was gently swirled while keeping flat on a table, and the time taken for complete dissolution of the cake was recorded. Table 42: Screening for carbohydrate and surfactants as stabilizers, iteration 4: List of formulations

. [00106] A Total particle analysis of T0 liquid and T0 lyo by FlowCam has been carried out. For liquid and lyohpilized samples at T0, with the exception of the formulation phosphate/trehalose/PEG, all other formulations containing surfactants showed a lower total of sub-visible particles as measured by FlowCam. There was a sharp increase in the number. of sub-visible particles following lyophilization of formulations not containing surfactants. This increase is consistent with visual observations. Table 44: Flow imaging particle analysis for T0 Liq and T0 lyo, iteration 4 final screening Phos Treh PEG 164 1 2_{58 23 0} 0 1922 433 4 3_{56 48 0} 0 4738

[00107] A RALS analysis of T0 Liquid and Lyo samples at 475V has been carried out. The RALS data showed no apparent differences between liquid and lyophilized formulations at T0 with the exception of CF and phosphate/trehalose, which showed an increase in scattering following lyophilization. Table 45: RALS analysis, iteration 4 T0 Liq and lyo at 475V

[00108] A FlowCam analysis of T0 FT samples relative to T0 Liq has been carried out. Except for CF, succinate/trehalose, and succinate/trehalose/PEG, all otherformulations, showed no increase in total sub-visible particles by flow imaging particleanalysis. Table 46: Flow imaging particle analysis for T0 Liq and T0 FT, iteration 4 final screening Pho s /Treh /PEG 1641 258 23 0 0 1922 176 10 0 10 0

196 [00109] A total particle analysis of T0 liquid and T0 lyo by FlowCam has been carried out. For liquid and lyohpilized samples at T0, with the exception of the formulation phosphate/trehalose/PEG, all other formulations containing surfactants showed a lower total of sub-visible particles as measured by FlowCam. There was a sharp increase in the number of sub-visible particles following lyophilization of formulations not containing surfactants.This increase is consistent with visual observations. Table 47: Total particle analysis of T0 liquid and T0 lyo by FlowCam, iteration 4 final screening

[00110] A FlowCam analysis of T0 FT samples relative to T0 Liq has been carried out, see table 48. Conclusion: Except for CF, succinate/trehalose, and succinate/trehalose/PEG, all other formulations, showed no increase in total sub-visible particles by flow imaging particle analysis. Table 48: Flow imaging particle analysis for T0 Liq and T0 FT, iteration 4 final screening

[00111] The total distribution of particles by FlowCam for T7 and T14 Liquid/50°C relative to T0 Liquid has been determined. Overall, there was an increase in the number of sub-visible particles across all formulations following thermal stress, except for CF. Amongst the other formulations, formulations containing phosphate had lower number of total sub-visible particles measured by FlowCam. Table 49: Flow imaging particle analysis for T0 Liq and T0 FT, iteration 4 final screening hos reh EG 41 258 23 0 0 22 176 10 0 10 0

196 [00112] The total distribution of particles by FlowCam for T7 and T14 Liquid/50°C relative to T0 Liquid has been determined. Overall, there was an increase in the number of sub-visible particles across all formulations following thermal stress, except for CF. Amongst the other formulations, formulations containing phosphate had lower number of total sub-visible particles measured by FlowCam. Table 50: Flow imaging particle analysis for T0 Liquid, T7 and T14 Liquid/50°C, iteration 4 final screening hos/ reh/ EG3 50 1 2 1 1 0

[00113] The total distribution of particles by FlowCam for T7 and T14 lyo/ 50°C relative to T0 Lyo (top) has been determined. The total sub-visible particle count for lyophilized formulations is much lower when compared to the corresponding liquid formulations, with the exception of formulations without surfactants. Comparing T0 with T7 data, the formulations containingsuccinate/alanine/tween 20, succinate/sucrose/tween 20, succinate/sucrose/PEG 3350,succinate/trehalsoe/PEG 3350, phosphate/sucrose/tween 20, and phosphate/trehalose/tween 20, showed lower sub-visible particle counts relative to the other formulations. Table 51: Flow imaging particle analysis for T0 Lyo, T7 and T14 Lyo/50°C, iteration 4 final screening os/ h/ G3 0

[00114] The percent of LMW of samples has been analyzed by non-reducing SDS-CE for T0 and T14 Liquid samples. For liquid formulations, succinate/sucrose/PEG 3350, phosphate/sucrose/PEG 3350, and phosphate/trehalose/PEG 3350 formulations showed the smallest increase in percent LMW following time and temperature (thermal) stress. Table 52: % AUC of samples analyzed by non-reducing SDS-CE for T0 and T14 Liquid samples, iteration 4 final screening

T0 Liquid %AUC

[00115] The percent of LMW of samples has been analyzed by reduced SDS-CE for T0 and T14 Liquid samples. All liquid formulations showed a smaller increase in percent LMW when compared to CF after time and temperature (thermal)stress. The lowest was seen in the sample containing succinate/sucrose/tween 20. However, there were only slight differences between formulations. Moreover, The percent HMW of samples has been analyzed by reduced SDS-CE for T0 and T14 Liquid samples. All liquid formulations showed a smaller increase in percent HMW when compared to CF following time and temperature (thermal) stress. The differences between each formulation were small with the succinate/trehalose/PEG 3350 formulation showing the least percent HMW.

Table 53 : % AUC of samples analyzed by reduced SDS-CE for T0 and T14 Liquid samples,

[00116] The lead formulation from iteration 4 was the succinate/sucrose/PEG 3350 formulation.

Claims

CLAIMS 1. A composition comprising a conjugate between bevacizumab and a dye, said composition having a pH of 6.0 to 7.0 and a conductivity of at most 4.0 mS/cm, in a mixture consisting of a buffer with a concentration up to 40 mM and at least one excipient selected from a nonionic surfactant, trehalose, a polyethylene glycol (PEG) and alanine or mixtures thereof. 2. The composition of claim 1, wherein said buffer has a concentration of 20 mM. 3.. The composition of claims 1 or 2, wherein said buffer is succinate. 4. The composition of claims 1 or 2 wherein said buffer is phosphate. 5. The composition of any of the preceding claims wherein the nonionic surfactant is selected from polysorbate 20 and poloxamer 188 and the polyethylene glycol (PEG) is PEG 3350. 6. The composition of any one of the preceding claims, wherein said trehalose has a concentration up to 270 mM, preferably 270 mM. 7. The composition of claim 5, wherein said polysorbate 20 is present in 0.02 % (v/v). 8. The composition of claim 5, wherein said PEG 3350 is present in 1.0 % (v/v). 9. The composition of claim 5, wherein said poloxamer 188 is present in 0.05 % (v/v). 10. The composition of any one of the preceding claims, further comprising albumin, preferably present in 0.1 % w/v. 11. The composition of any one of the preceding claims, wherein said dye is a NIRF dye. 12. The composition of claim 11, wherein said NIRF dye is IRDye 800 CW. 13. The composition of any one of the preceding claims, wherein said conjugate is through an NHS ester. 14. The composition of any one of the preceding claims, wherein said bevacizumab has the VH amino acid sequence shown in SEQ ID NO: 1 and the VL amino acid sequence shown in SEQ ID NIO: 2. 15. The composition of any one of the preceding claims, which is lyophilized.