Reformulation and Thermal Stability of a Therapeutic Anti-Cocaine mAb

Abstract

We concentrated and reformulated the anti-cocaine mAb, h2E2, to reduce the amount of sucrose and histidine buffer infused with the mAb, to satisfy FDA maximum exposure levels for those components for use in clinical trials. After concentration of the original 20 mg/ml mAb, 4 reformulation buffers were evaluated for suitability. The concentration of histidine was reduced from 10 mM to 3 or 0 mM, and the concentration of sucrose reduced from 10% to 2, 4, or 6%. The approximately 100 mg/ml reformulated mAb samples were analyzed for oligomer formation, aggregation, concentration of the emulsifier polysorbate 80, and thermal stability. These reformulated mAb samples were also assessed for their stability at 40°C from 1 day to 12 weeks. As expected, long term thermal resistance to oligomer formation increased as a function of increasing sucrose concentration. Interestingly, unbuffered reformulated mAb displayed a less than or equal to tendency to form oligomers and aggregates, compared to the histidine buffered samples. Importantly, even after 12 weeks at 40°C, all the reformulated samples displayed little aggregation, and bound their antigen (cocaine) with identical affinities and thermodynamics, as measured by isothermal titration calorimetry (ITC). These ITC thermodynamic binding parameters are consistent with recently published values for the original formulation of this mAb. In all reformulated samples there was a slight decrease in the number of cocaine binding sites after 12 weeks at 40°C, likely due to the parallel small increase in soluble oligomeric antibody, suggesting that soluble oligomeric mAb may no longer bind cocaine with high affinity.

Graphical Abstract

Introduction

The Norman laboratory group developed a high affinity anti-cocaine IgG1 mAb, named h2E2^1-3 to attempt to fill an important unmet need, since there is currently no FDA approved drug or vaccine for the treatment of cocaine use disorders. This recombinant humanized monoclonal h2E2 antibody was previously produced in quantities needed for Phase 1 clinical testing by Catalent PharmaSolutions, Inc. It was formulated at a mAb concentration of 20 mg/ml in a stability-optimized formulation buffer containing 10 mM histidine, 10% sucrose and 0.01% PS 80, pH=6.0. However, due to larger than anticipated doses (and therefore volumes) of the h2E2 antibody required to be tested for safety by the FDA, the infusion volumes needed to satisfy the largest mAb doses would result in exceeding the maximum levels of sucrose and histidine that are approved by the FDA as safe for clinical usage. Therefore, the h2E2 mAb needs to be reformulated to decrease the amounts of these components.

Thus, we performed small scale, non GMP experiments, concentrating the mAb and testing several reformulation buffers related to the originally optimized formulation buffer to assess their suitability for h2E2 mAb reformulation. We evaluated this by measurements of mAb stability, cocaine binding, and tendency to form oligomers under thermally stressed (extended incubation at 40°C) conditions. We chose to use buffers containing 3 mM histidine, since that concentration would not exceed the FDA maximum levels for intravenous administration at the highest doses (and therefore largest mAb infusion volumes) of the reformulated 100 mg/ml mAb needed for Phase 1 safety trials. We also evaluated a reformulation solution lacking histidine and any other non-protein buffering component, since recent reports have shown that concentrated mAbs themselves have substantial buffering capacities, and therefore formulations containing no small molecule buffers maybe suitable for therapeutic antibodies⁴. We chose to vary the sucrose concentration up to the highest level allowable, assuming a 100 mg/ml reformulated mAb and the highest dosage required for the Phase 1 studies (i.e., up to 6% sucrose). Histidine and sucrose were both analyzed since they were included in the original (20 mg/ml mAb) optimized formulation buffer, and are both commonly used to stabilize mAbs⁵.

The main endpoints of our study are mAb stability and assessment of mAb oligomerization and aggregation after reformulation and concentration to 100 mg/ml, as well as the thermal stability of both the antibody and its ability to bind cocaine after extended incubation at 40°C. The findings reported in this study inform the choice of formulation buffer to be used for the larger scale and expensive GMP reformulation of sufficient quantities of this h2E2 mAb needed for Phase 1 trials. It is anticipated that the results and findings presented in this study may also be applicable and helpful for the formulation and analyses of other therapeutic mAbs and proteins under development.

Materials and Methods

Materials

The generation, production, and purification of the h2E2 anti-cocaine monoclonal antibody by the manufacturer, Catalent, was previously described³, and was formulated at 20 mg/ml mAb in 10 mM histidine, 10% sucrose and 0.01% PS 80, pH=6.0. 10 mM ligand stock solutions of cocaine in distilled water were made from cocaine hydrochloride solid (Research Triangle Institute, RTI batch 14201-12A) as described¹. Histidine (cat. H-8000) and histidine monohydrochloride (cat. H-8125) were purchased from Sigma-Aldrich. Sucrose was from Sigma (cat. S-9378), and polysorbate 80 (PS 80) surfactant was obtained from JT Baker (cat. 4117-02). 4,4'-dianilino-1,1'-binaphthyl-5,5'-disulfonic acid (Bis-ANS) used for assays of PS 80 concentration was purchased from Cayman Chemicals (cat. 21820). Sypro orange 5000X commercial dye stock was from Invitrogen/Life technologies and stored at −20°C. The solid 4-(4-(dimethylamino)styryl)-N-methylpyridinium iodide dye (4-Di-ASP iodide, DASPMI) was purchased from ThermoFisher Scientific (Invitrogen/Life Technologies, cat. D-288). A 20 mM DASPMI stock solution was made in dry DMSO and stored light protected at −20°C for use in DSF experiments. The Applied Biosystems 48 well RT PCR plates (cat. 4375816) and the MicroAmp plate sealing optical adhesive transparent film (cat. 4375928) used in the StepOne RT PCR instrument used to perform the DSF analyses were purchased from ThermoFisher Scientific. Sephacryl S-300 HR media for size exclusion chromatography analysis of soluble mAb oligomers was purchased from Pharmacia. Electrophoresis chemicals and SDS-PAGE reagents were purchased from BioRad. The SDS-PAGE gel protein standards were from SMOBIO (cat. PM1600). PBS buffer used for ITC experiments was diluted from PBS buffer concentrate (10X) from Cambrex (BioWhittaker, without calcium or magnesium, cat. 17-517Q). PCR tubes (200 μl, Fisher cat. 14230225) were used to incubate the reformulated mAb samples at 40°C in a Benchmark Model BSA200-5017 dry heating block (with a heated lid).

Methods

Concentration and reformulation of the h2E2 mAb

The originally formulated 20 mg/ml h2E2 mAb contains 0.01% PS 80, and is the only available mAb drug substance for this mAb. Since PS 80 surfactant cannot be readily eliminated by dialysis due to its large micellar size and small CMC , we concentrated this material as is to 120 mg/ml using Centriprep 30 centrifugal concentrators, knowing that some of the PS80 would be concentrated along with the mAb. The concentrated mAb was then subsequently sterilized using 0.22 μm Millex-GP filters, yielding a final mAb concentration of 120 mg/ml (measured by the absorbance at 280 nm). All UV A280 nm mAb concentration measurements were performed without dilution using an Implen N60 nanophotometer and an applied sample volume of 1.5 μl. For each reformulation buffer tested, approximately 2.7 ml of 120 mg/ml mAb was dialyzed 3 times vs approximately 330 ml of buffer each time, at 4°C, over a total time period of 22 hrs (3 hrs, 4 hrs, and 15 hrs dialysis at 4°C). 3-4 ml volumes of the reformulated mAb solutions were examined for particulates and turbidity in glass tubes using white light illumination from behind and underneath the sample glass tubes. In addition, 200 μl volumes of each 40°C heat-treated sample were examined for particulates and turbidity in clear, thin-walled PCR tubes, using white light illumination from behind the tubes.

Measurement of PS 80 concentrations in the reformulated, approximately 100 mg/ml mAb

We used the method employing the change in fluorescence of bis-ANS to quantitate PS 80, as described by Zheng et al⁶, with minor modifications, as described in Supporting Information.

Differential Scanning Fluorimetry (DSF) analysis of undiluted mAb in the 4 reformulation buffers

We employed both the Sypro orange dye and the DASPMI dye in these experiments. The samples were prepared for DSF analysis simply by adding a small volume of concentrated dye to each (0.5 μl to 99.5 μl sample), and loading 4, 20 μl aliquots of each sample in the PCR plate used for the DSF analysis, and scanning from 35° to 95°C using the slowest temperature ramp rate setting available in the instrument software (0.3), which corresponds to an actual temperature ramp rate of 0.45°C per minute for 20 μL samples.

Non-Reducing SDS-PAGE assessment of mAb aggregation state

The mAb samples were diluted to 1 mg/ml using the non-reducing sample buffer of Laemmli⁷, boiled for 3 minutes, and 5-10 μl was loaded onto the wells of a 7%, 1.5 mm thick Laemmli acrylamide gel. After running, the dye was stained for 1 hour using Quick Blue protein stain from IBI Scientific, washed several times with water for at least 2 hours at 22°C, and then photographed.

Soluble Oligomer assessment using Sephacryl S-300 sizing column

A 10 ml, 28 cm long Sephacryl S-300 column was equilibrated in PBS, and run at approximately 0.25 ml/min, monitoring absorbance at 280 nm (0.5 AUFS). Typically, 5 μl of ≈100 mg/ml mAb sample was diluted with 145 μl PBS and loaded onto the column. The strip chart recording of the chromatogram was scanned, and then the data digitized and integrated using Origin Pro 2022 software to quantitate the % oligomer formed in each sample.

Isothermal Calorimetry (ITC) assessment of binding affinity

All isothermal titration calorimetry (ITC) experiments were performed at 20°C using a MicroCal VP-ITC instrument, as described previously⁸. Briefly, twenty 14 μl injections of 0.1 mM cocaine in PBS were performed for each experiment, adding to the 1.4227 ml cell sample containing h2E2 mAb in PBS or PBS buffer only, all using 0.22 μm filtered PBS buffer. The data were corrected for background injection of cocaine by two methods. The first method subtracted a data set from an ITC experiment performed with only PBS buffer in the cell (i.e., no mAb present) from each mAb sample data set prior to fitting of the data, and the second method used an average of the last 5 injections of the mAb sample ITC data (after the mAb is saturated with cocaine) to subtract from all the integrated mAb injection ITC peaks, prior to fitting of the data. The second method was found to be more accurate and reproducible and was therefore used in this study. The data were analyzed and fit using the one binding site model incorporated into the Origin 7.0 software supplied with the instrument. Kd (dissociation constant) values were calculated by taking the reciprocal of the Origin best fitted ITC K_ASS (association constant) values, and ΔG values reported in the Table were calculated using the Origin fitted values of ΔH and program calculated values of ΔS, and the Gibbs free energy equation, ΔG = ΔH − TΔS (T = 293.15°K). Approximately 100 uL of each diluted protein sample that was used to load the ITC cell was saved used to determine the ITC cell mAb protein concentration of all samples, using the molar extinction coefficient at 280 nm of 219,500 M⁻¹cm⁻¹ for the h2E2 mAb, and a molecular weight of the mAb of 144,500 Da.

Results

The outline for the current study is shown in Figure 1. The originally formulated 20 mg/ml h2E2 mAb was concentrated using a 30 kDa molecular cutoff filter, yielding a final mAb concentration of 120 mg/ml. The 120 mg/ml mAb was then reformulated by extensive dialysis at 4°C into 4 reformulation buffers (delineated in Table 1) and sterilized using 0.22 μm filtration. The NaCl concentration used in each buffer was chosen to make all buffers close to isotonic. The pH of the dialysis solutions at the end of the 3rd dialysis was measured, and the samples collected and quantitated by 280 nm absorbance to yield the mAb protein concentrations reported in Table 1.

Figure 1. Flow chart outline description of this study.

Table 1.

Composition and characterization of the 4 reformulation buffers utilized for the h2E2 mAb.

Reformulation Buffer Name	[Histidine buffer] pH=6.0	[PS-80] (wt/vol)	[sucrose] (wt/vol)	[NaCl] (mM)	Measured buffer pH (after mAb dialysis)	[mAb] after dialysis vs reformulation buffer (mg/ml)
A	0 mM	0.01%	4%	85 mM	5.10	97.6
B	3 mM	0.01%	2%	110 mM	6.04	96.0
C	3 mM	0.01%	4%	85 mM	6.03	97.2
D	3 mM	0.01%	6%	60 mM	6.01	102.9

It was expected that the concentration of PS 80 would increase during the concentration of the mAb, and remain higher during the reformulation dialysis, since 0.01 % PS 80 is above the critical micellar concentration (CMC) of PS 80 (approximately 0.002% wt/vol), and the size of the PS 80 micelle is approximately 76 kDa⁹. Since the concentration of PS 80 is an important part of the reformulation buffer, we measured it in the 4 reformulated samples (see Supporting Information). The standard curve for this PS 80 assay is shown in Supplemental Figure 1. The PS 80 (wt/vol) concentrations of the 4 reformulated mAb samples were all close to 0.03% PS 80 (0.0289, 0.0303, 0.0307, and 0.0300% for reformulations in Buffer A-D, respectively), confirming that some of the PS 80 concentrated along with the mAb during the concentration process, and remained higher even after extensive reformulation dialysis versus 0.01% PS 80 (the original formulation buffer also contained 0.01% PS 80).

We sought to determine if usable DSF stability data could be obtained on the reformulated samples directly (with no dilution) to assess the relative stability of the mAb in the reformulation buffers. This might have been problematic, since the mAb concentration (≈100 mg/ml = 692 μM) is much higher than that typically employed for DSF (1-5 μM¹⁰), and the elevated PS 80 detergent concentration can adversely affect DSF data quality and results^11,12 by PS 80 binding the dye directly or by affecting the binding of the dye to the exposed hydrophobic patches on the thermally unfolding mAb. For these DSF experiments, we employed both the most commonly used Sypro orange dye, as well as the DASPMI dye, which we previously showed only bound to the Fab portion of the mAb, and thus can be used as a reporter for cocaine binding to the mAb¹⁰. Therefore, the DSF mAb samples were prepared simply by adding a small volume (0.5% vol/vol) of concentrated dye to each reformulated mAb solution, and analyzing 4 aliquots of each reformulated sample, as shown in Figure 2. For each dye, the top traces are the dye fluorescence, and the bottom traces are the first derivative of the dye fluorescence. Derivative melting temperatures (TmDs), indicative of the stability of the part(s) of the mAb to which the dye binds, are the peaks, either positive (for the DSPMI dye) or negative (for the Sypro orange dye). These TmD temperatures are plotted as a function of sucrose concentration in the reformulation buffers in Figure 3.

Figure 2. Differential scanning fluorimetry (DSF) data for the undiluted (≈100 mg/ml), reformulated h2E2 mAbs in their respective reformulation buffers, using either a 1:200 dilution of commercial stock Sypro orange dye, or 100 μM of DASPMI dye (quadruplicate aliquot data are shown for all samples).

For both dyes, the upper panel shows the raw fluorescence data, while the lower panel shows the first derivative of that fluorescence data, which is used to determine the derivative thermal melting temperatures (TmDs) that are plotted in Figure 3. The data in this figure and other figures is color coded as follows; black = Buffer A, red = Buffer B, green = Buffer C, and blue = Buffer D.

Figure 3. The effect of the concentration of sucrose on the DSF melting temperature (TmD) of the mAb in the 4 reformulation buffers.

Straight line fits of the DSF TmD data from the reformulation buffers containing 3 mM histidine, pH=6.0 are shown in the figure for both DSF dyes utilized, and the data for the reformulation buffer that contains no histidine buffer is also included, and annotated by arrows. Error bars are the standard deviations for the quadruplicate aliquot wells analyzed. Note that the presence or absence of histidine buffer in the samples containing 4% sucrose made no significant difference for either dye utilized.

The starting, originally formulated 20 mg/ml mAb, the concentrated 120 mg/ml mAb, and the 4 reformulated mAbs (approximately 100 mg/ml) as a function of time of incubation at 40°C were analyzed by non-reducing SDS-PAGE. The major monomeric mAb migrates under these conditions as a band of ≈180 kDa apparent molecular weight. There are some lower molecular weight bands, which are likely to be mAb missing one or more of the 4 constituent mAb components (i.e., 2 heavy chains and 2 light chains) which are often seen in small amounts in commercially produced mAbs. We previously confirmed by mass spectral analysis that one of these h2E2 lower molecular weight mAb bands consists of only 2 heavy chains¹³, migrating under these gel conditions at approximately 120 kDa. However, we cannot exclude the possibility that these other bands could be proteolytic fragments. Regardless, for the purposes of the current study, we are interested in bands with apparent sizes greater than the monomeric 180 kDa mAb band. As can be seen in Panel A of Figure 4, intentional overloading of the gel with 5 and 10 μg of mAb allows visualization of 2 aggregated species in all samples which were concentrated from the original 20 mg/ml formulation, with 2 such aggregation bands being present, one that does not enter the gel, and one that does enter the gel. Panel B in Figure 4 compares the 4 reformulated mAbs after long incubations at 40°C (4, 8, and 12 weeks), with only a small increase in staining of these aggregated mAb bands being observed as a function of time at 40°C.

Figure 4. Non-reducing SDS-PAGE analysis of the originally formulated h2E2 mAb, as well as the concentrated mAb, and the subsequently reformulated mAb at various times of incubation at 40°C.

7% acrylamide gels were used, loading either 5 or 10 ug of each mAb sample per well, after boiling the samples in non-reducing SDS-PAGE sample buffer for 3 minutes. Panel A is data obtained using the originally formulated mAb, the concentrated (120 mg/ml) mAb, and reformulated samples (≈100 mg/ml) after 2 weeks at 40°C. Panel B shows results obtained using reformulated mAbs after 4, 8, and 12 weeks incubation at 40°C. The dashed lines were added for ease of interpretation of the data and are superimposed on the images of a single gel shown in each panel. The gels were intentionally overloaded to allow visualization of the oligomer and aggregated mAb high molecular weight bands.

Although we did not measure viscosity, we noted little, if any increase in viscosity upon concentration of the mAb from 20 mg/ml to a maximum of 120 mg/ml. In addition, it should be noted that no particulates or turbidity were visible in any of the reformulated mAb samples even after 12 weeks at 40°C. However, we quantitated the amount of soluble oligomer formed in these samples by size exclusion chromatography for all the 40°C time points for all 4 of the reformulation buffers. Example raw chromatograms for the 40°C, 12 week time point samples are shown in Supplemental Figure 2. The % of oligomeric mAb present in all these samples was determined by integration of peaks recorded on these sizing column chromatograms, and the results are shown in Figure 5. From this data it is evident that the buffer in which the most soluble oligomer formation occurs is Buffer B, and that Buffers A and D show substantially less, and essentially equivalent, amounts of oligomer formation.

Figure 5. The 40°C time course of the formation of soluble oligomers from h2E2 mAb reformulated in the 4 reformulation buffers, as assessed by Sephacryl S-300 size exclusion chromatography in PBS buffer.

The data is color coded as follows; black = Buffer A, red = Buffer B, green = Buffer C, and blue = Buffer D. Example chromatograms showing the raw data used to derive the % oligomer data shown in the figure for the 12 week (84 day) samples are shown in Supplemental Figure 2.

Isothermal titration calorimetry analysis was employed to determine if there was a change in the reformulated mAbs in terms of cocaine affinity, number of binding sites, or binding thermodynamics after incubation at 40°C for 12 weeks. Unlike the other techniques employed in this study, ITC can reveal if there are any changes in either the number of functional cocaine binding sites, or in the thermodynamic parameters driving the high affinity binding of cocaine to the thermally stressed h2E2 mAb. Example data for the 120 mg/ml unheated (control) mAb which was used to generate the 4 reformulation samples is shown in Figure 6. These results are very similar to recently published ITC data using the original h2E2 mAb 20 mg/ml formulation (i.e., N=0.785 (cocaine bound per mAb Fab), K_ASS = 1.33 e8 M⁻¹, ΔH = −15400 cal/mol, ΔS= −15.4 cal/mol/deg, ΔG = −10881 cal/mol, Kd = 8.1 nM⁸). The unheated 120 mg/ml “control” ITC parameter determinations, as well as those determined for the 4 reformulated mAbs after incubation at 40°C for 12 weeks are given in Table 2. All of the thermodynamic parameters and calculated binding affinities are unchanged from the concentrated control (unheated) mAb, but there is a small (approximately 5-11%) decrease in the number of cocaine binding sites after incubation at 40°C for 12 weeks.

Figure 6. Representative ITC data for the binding of cocaine to concentrated (120 mg/ml) control h2E2 mAb at 20°C in PBS buffer, pH=7.4.

The injections corrected for the cocaine ligand blank ITC data are shown in the upper panel, and the fitting of the resultant titration curve using a single binding site model is shown in the lower panel. The concentrations of cocaine ligand in the syringe and h2E2 mAb in the sample cell are given in the top panel, and the fitted parameters describing the binding for cocaine to the control h2E2 mAb are shown in the bottom panel.

Table 2.

Summary of cocaine binding - Isothermal Titration Calorimetry (ITC) thermodynamic data (at 20°C) for the mAb samples heated at 40°C for 12 weeks (84 days) in the 4 reformulation buffers.

Sample	N (# sites per mAb Fab)	Kass (M−1)	ΔH (cal/mol)	ΔS (cal/mol/deg)	ΔG (cal/mol)	Kd (nM)
120 mg/ml Control h2E2 mAb	0.816 (100%)	9.99e7	−14000	−11.2	−10700	10.0
Buffer A : 40°C, 12 week h2E2 mAb	0.773 (94.7%)	9.01e7	−14200	−12.0	−10700	11.1
Buffer B : 40°C, 12 week h2E2 mAb	0.727 (89.1%)	7.42e7	−14300	−12.8	−10500	13.5
Buffer C : 40°C, 12 week h2E2 mAb	0.744 (91.2%)	8.33e7	−14000	−11.4	−10600	12.0
Buffer D : 40°C, 12 week h2E2 mAb	0.744 (91.2%)	8.85e7	−13900	−11.0	−10700	11.3

Discussion

Formulation of therapeutic proteins and mAbs is very important. The buffer and pH used, the use of stabilizing excipients such are sucrose and arginine, and the use of non-toxic surfactants/emulsifiers like PS 20 and PS 80 are all important formulation buffer considerations. The choice of the proper concentrations of all these components further complicates the design of formulation buffers. In the case of the anti-cocaine h2E2 mAb we have developed, all these components are important not just for maintaining the stability and monomeric nature of the mAb, but also must be considered for their effect on the physiology of the person to which they will be administered. This is because to be effective, the h2E2 mAb must essentially bind all the cocaine which the user chooses to self-administer, requiring a large amount of anti-cocaine mAb to be present in the blood during times when self-administration is likely. This means that large amounts of this mAb must be administered, which requires large volumes (up to hundreds of ml) of concentrated mAb to be delivered intravenously, magnifying any possible physiological effects of buffers, excipients, and emulsifiers present in the formulated anti-cocaine mAb. The original formulation of this h2E2 mAb, like most therapeutic mAb formulations, was designed principally to maximize the stability and shelf-life of the mAb. For the h2E2 mAb, it was also critical to minimize the concentrations of the histidine buffer and sucrose excipient to allow for the needed large volumes to be infused to bind all cocaine. Large volumes are needed even after concentration of the mAb to 100 mg/ml. The 6 most commonly used buffers and excipients for high concentration therapeutic mAbs which have been approved by the FDA are polysorbate, histidine, sucrose, arginine, sodium chloride, and methionine, in that order (see Figure 2 in ¹⁴. Our reformulations contained 4 of these 6 components. We did not include arginine or methionine, since the effective concentrations of those excipients would result in exceeding the Maximum Daily Exposure (MDE) doses that have been approved to date by the FDA for total infusion at the larger volumes needed to contain the higher doses of the anti-cocaine mAb to be tested clinically.

Consideration of all this has necessitated the reformulation of this h2E2 mAb from its original formulation of 20 mg/ml mAb in 10 mM histidine, 10% sucrose and 0.01% PS 80, pH=6.0. Increasing the concentration of the mAb and decreasing the concentrations of the buffer and sucrose in the formulated mAb are necessary to avoid exceeding the maximum approved amounts of these components allowable by the FDA at the highest doses of mAb to be tested.

For an effective cocaine use disorder therapeutic, the h2E2 mAb should be present while any patient with cocaine use disorder is at risk of relapse. Because humanized mAbs typically have a half-life of approximately 3-4 weeks, it is anticipated that approximately monthly dosing would maintain sufficient levels of h2E2. The maximum infused h2E2 mAb therapeutic dose that is currently proposed is 120 mg/kg (8,400 mg/70 kg person). Intravenous doses up to 1,200 mg/kg of h2E2 mAb appear to have no observable toxic effects in rats (data shared with the FDA). The human dose of 8,400 mg mAb/70 kg person has the capacity to bind 0.5 mg/kg of cocaine base. Because the volume of distribution (Vd) of mAbs is low, approximately 0.1 – 0.2 L/kg, compared to cocaine, the plasma concentration of h2E2 mAb will be high, enhancing its effectiveness to lower the Vd of cocaine and sequestering cocaine in the peripheral circulation and away from the brain, thereby antagonizing the central effects of cocaine¹⁵. As the sequestered plasma cocaine is mostly bound to the h2E2 mAb, the plasma cocaine is pharmacologically inert, as evidenced by the mAb antagonizing the cardiovascular effects of cocaine in rats¹⁶. Since the effective dose of h2E2 mAb is relatively high, the mAb formulation needs to have a high concentration of h2E2 mAb (100 mg/ml) to avoid very large infusion volumes that would result in higher doses of histidine and sucrose than those currently approved by the FDA. Specifically, the highest Maximum Daily Exposure (MDE) doses that are currently approved by the FDA for intravenous delivery are 63 mg, 7,560 mg, and 900 mg for histidine, sucrose, and polysorbate 80, respectively (see the FDA database for excipients: Inactive Ingredient Search for Approved Drug Products (fda.gov) at the web address: www.accessdata.fda.gov/scripts/cder/iig/index.cfm.) At a mAb concentration of 100 mg/ml, this translates to a maximum infusion volume of 84 ml per 70 kg person of the 100 mg/ml mAb formulated in 3 mM histidine, 4% wt/vol sucrose and 0.03% wt/vol polysorbate 80. Thus, the 84 ml infusion volume would contain 39 mg histidine, 3360 mg sucrose, and 25.2 mg polysorbate 80, all well below the previously approved FDA Maximum Daily Exposure (MDE) levels. Whether the maximum dose of 84 ml of a 100 mg/ml mAb is well tolerated by humans is to be determined. However, assuming 4L plasma volume in human adults, the total h2E2 mAb at the highest dose would be a fraction of the total IgG content, i.e., 8.4 g h2E2 mAb compared to approximately 40 g of IgG present in 4L of plasma, assuming the average concentration of 10 mg/ml IgG in human plasma.

We concentrated the mAb and designed 4 reformulation buffers that include these same components as the original formulation buffer, but decreased and varied their concentrations in ranges that would not exceed FDA maximum approved levels at the highest anticipated doses of the h2E2 mAb. We chose the maximum concentrations of histidine and sucrose to be consistent with the maximum allowable administered amounts of these components. We also evaluated a formulation containing no small molecule buffer, since it was shown that high concentrations of mAb can exhibit substantial “self-buffering” capacities on their own⁴, and, as reported in a recent review of formulations of commercially available antibodies¹⁷, there is one FDA approved product that does not contain a small molecule buffer.

We also considered the maximum allowable PS 80 surfactant/emulsifier, but the above calculations showed that we would not exceed the PS 80 limit even if all of the PS 80 was concentrated with the mAb during concentration and reformulation. However, it is important to know the concentration of PS 80 in the reformulated product, and we showed that a simple PS 80 assay based on changes in bis-ANS fluorescence was suitable for our purposes. All four reformulated ≈100 mg/ml mAb solutions contained approximately 0.03% PS 80, which was an increase from the 0.01% PS 80 in the original 20 mg/ml mAb formulation, suggesting that about half of the PS 80 concentrated along with the mAb, which is consistent with the CMC and micellar size of PS 80 and the 30 kDa molecular weight cutoff of the concentrators used. This increased concentration of the PS80 should prove advantageous, since the higher PS 80 concentration should better inhibit the adsorption and aggregation of the mAb.

There have been relatively few attempts at characterizing the biophysical and stability properties of concentrated mAbs directly in their formulation buffers¹⁸. It was of interest to determine if we could assess the thermal stability of the reformulated h2E2 mAb as is, without diluting it. Therefore, we used differential scanning fluorimetry (DSF) and two fluorescent dyes, Sypro orange (a solvatochromic dye, the fluorescence of which increases when it binds to hydrophobic patches of the mAb exposed by thermal unfolding) and DASPMI (a rotor dye, the fluorescence of which is sensitive to its rotational freedom that is decreased when the dye binds mAb sites exposed upon thermal unfolding¹⁹). We previously showed that the DASPMI dye binds only to the Fab portion of the h2E2 mAb, and thus can be used to assess the binding of cocaine and cocaine metabolites to this mAb¹⁰. The DSF results shown in Figures 2 and 3 demonstrate that usable DSF data can indeed be generated using the undiluted reformulated mAbs, and that, as expected, the thermal stability of the mAb increases with the % sucrose in the reformation solution. The two dyes give rise to very different looking first derivative DSF curves, and different TmD values. This is most likely due to the Sypro dye binding to several unfolded patches located in multiple regions of the mAb, whereas the DASPMI dye is primarily binding to only one region – the Fab fragment near the cocaine binding site¹⁰.

We used non-reducing SDS-PAGE to assess aggregate formation in the mAb samples (see Figure 4). Large amounts (5 and 10 μg) of mAb samples were loaded on these gels to allow visualization of minor larger molecular weight oligomers and aggregated mAb. Some aggregates appeared after concentration of the mAb to 120 mg/ml, and this small amount of aggregated mAb persists in the reformulated samples, with a very small increase in high molecular weight aggregates observed with increasing time of incubation at 40°C. Importantly, the amount of aggregated mAb was small even after 12 weeks at 40°C, indicating that all 4 reformulation buffers did a good job inhibiting thermal aggregation.

In order to quantitatively assess the effects of the different reformulation buffers on the rate of formation of soluble mAb oligomers, the 4 reformulated mAb solutions were analyzed as a function of incubation time at 40°C by size exclusion chromatography, as shown in Figure 5. Representative example raw data for the samples incubated for 12 weeks at 40°C are shown in Supplemental Figure 2. The results indicate that the least formation of soluble oligomers is seen in Buffers A and D. Buffer A contains 4% sucrose and no histidine buffer, while buffer D contains 6% sucrose and 3 mM histidine buffer, pH=6.0. It should be noted that the measured pH of Buffer A after dialysis was 5.10, while the measured pH of Buffer D after dialysis was 6.01. Importantly, the unbuffered 100 mg/ml mAb sample (Buffer A) was equal to or better than the buffered mAb formulations in terms of limiting the amount of soluble oligomer formed during thermal stressing at 40°C.

It is also very important to measure the effect of the different reformulations on the therapeutic function of the h2E2 mAb, i.e., the ability to bind cocaine with high affinity. To assess the affinity as well as the number of binding sites and the thermodynamic parameters involved in the binding of cocaine to the mAb, isothermal titration calorimetry (ITC) of the mAb after dilution into PBS buffer was performed, and these results were compared with each other and previous results obtained with the original 20 mg/ml formulation of the mAb, all in PBS buffer. Representative results for the concentrated (but not heated) mAb are shown in Figure 6, and these results are very similar to previously published results using the originally formulated mAb⁸. Importantly, even after incubation at 40°C for 12 weeks, the affinity and thermodynamics of the binding of cocaine to all four of the reformulated mAbs were not altered (Table 2). However, the N value (number of cocaine molecules bound per mAb Fab) was slightly decreased in the heated samples. This decrease in N correlated fairly well with the amount of soluble mAb oligomer formed, which is shown in Figure 5. Specifically, comparing the % control N cocaine binding site values vs the % monomeric mAb remaining after 12 weeks at 40°C : Buffer A, 94.7% vs 92.1%; Buffer B, 89.1% vs 88.1%; Buffer C, 91.2% vs 89.8%; and Buffer D, 91.2% vs 92.0%. This close correlation suggests that the soluble oligomeric form of the mAb does not bind cocaine with high affinity.

There are other considerations that factor into the choice of a reformulation buffer for this h2E2 mAb. For example, there may be some potential “polishing” mAb purification steps needed after reformulation that could be affected by a lack of a small molecule buffer. In addition, the highest concentration of sucrose evaluated here (6%) will result in a near approach to the maximum allowable infusion amount of sucrose at the highest mAb dose to be tested. Also, as of the review published in 2021¹⁷, there was only 1 FDA approved commercial antibody that contains no small molecule buffer (i.e., citrate-free Humira), which may cause hesitancy to use a formulation containing no small molecule buffer.

Highlights.

Formulated anti-cocaine h2E2 mAb was reformulated to reduce histidine and sucrose

A formulation lacking a small molecule buffer was evaluated and proved effective

Reformulated mAbs were evaluated by DSF and non-reducing SDS-PAGE

Reformulated mAbs were thermally stressed for 1 day to 12 weeks at 40°C

Stressed mAb evaluated for oligomer and aggregate formation as a function of time

Stressed mAb was evaluated for antigen (cocaine) binding thermodynamics by ITC

Abbreviations

mAb

monoclonal antibody

h2E2

humanized anti-cocaine monoclonal antibody

Fab

fragment antigen-binding

PS 80

polysorbate 80

Bis-ANS

4,4'-dianilino-1,1'-binaphthyl-5,5'-disulfonic acid

CMC

critical micellar concentration

DSF

differential scanning calorimetry

DASPMI

(4-(4-(dimethylamino)styryl)-N-methylpyridinium iodide

SEC

size exclusion chromatography

NR SDS-PAGE

non-reducing sodium dodecyl sulfate polyacrylamide gel electrophoresis

ITC

isothermal titration calorimetry

PBS

phosphate buffered saline

K_ASS

binding association constant

Kd

binding dissociation constant

(GMP)

good manufacturing practice