Original Paper

Vox Sang 1995;68: 1 4

Martin Gonzhleza DominPo A. Muratweb Thermal Stability of Human ~

0

Immunoglobulins with Sorbitol Gerardo D. Fidelio a

a DeDartamento de Quimica Biolomca, - Fac. de Ciencias Quimicas y Laboratorios de Hemoderivados, Universidad Nacional de Cordoba, Argentina

A Critical Evaluation

................................................................................................. Abstract The effect of the additive sorbitol on the thermal stabilization of human IgG was investigated by differential scanning calorimetry and size exclusion chromatog- raphy. In the presence of 33% sorbitol, the temperature at which denaturation of IgG began (Ti) was increased from 52 to 65°C. Similarly, the temperature of the maximum heat capacity (T,,,,) was increased from 69 to 76°C. Sorbitol also de- creased dimer aggregation and the extent of oligomerization during heating com- pared with IgG dissolved in phosphate buffer. Sorbitol at 33% prevented massive protein denaturation but a 10-15% of oligomerization of high molecular weight aggregates with turbidity could not be avoided when heating for 10 hat 60°C. The use of sorbitol33% to avoid heat denaturation of human IgG during viral inacti- vation did not prevent protein aggregation or the appearance of turbidity. Conse- quently, further processing will be required to achieve a product suitable for phar- maceutical use. .....................

Introduction Materials and Methods

The World Health Organization (WHO) advises the use of pasteurization (60°C 10h in liquid) as one of the most appropriate methods of viral inactivation for plasma frac- tionated products [l]. The use of sorbitol and other polyols for virus inactivation of IgG solution for intravenous pro- duction has been reported [2-4]. Ideally, the heat treatment of liquid IgG allows safe virus inactivation and has several technical advantages compared with other processes.

According to the report from the WHO [l], the intrave- nous IgG shall be composed of not less than 90% of immu- noglobulins and not more than 10% of split products togeth- er with aggregates (MW higher than IgG trimer). In this work, we study the effect of sorbitol on the thermal stability of human IgG and the formation of aggregates with the peri- od of heating required for effective pasteurisation.

The molecular characteristics of the IgG preparations were deter- mined by size exclusion chromatography, using an FPLC System Pharmacia LCC 500-plus with a Superose-12 HR 10130 prepacked co- lumn. Cohn fraction I1 paste (30% protein w / ~ , >95% of IgG) was provided by the Laboratorios de Hemoderivados from Universidad Nacional de Cordoba, Argentina. The samples were prepared mixing 1.7 g of Cohn fraction 11 paste plus 3.3 g of solid sorbitol (Sigma Chemical Co., St. Louis, Mo., USA) and dissolved in a final volume of 10 ml of 100 mMphosphate buffer pH 6.8. The control was prepared in the same way without the additive. All IgG treatment was carried out at a final protein concentration of 5% w/v and the samples treated with additive, the final sorbitol concentration was 33% w/v. The addition of sorbitol does not change the IgG solubility. Protein concentration was determined by measuring the optical density at 280 nm, using an ex- tinction coefficient (E"") of 14. The volume of the samples for heating were 2 ml. The samples were heated at 60°C for 1.5 or 10 h with a com- puterized Haake Bath 3C. After heating, the samples were cooled down to room temperature. Then the samples were diluted with the

Received: March 15. 1994 0 I995 S. Karger AG. Basel Revised manuscript Departamento de Quimica Biologica-CIQUIBIC 0042-90071951068 I-0001

Accepted: July 2X, 1994

Dr. Gerardo D. Fidelio

received: July 26. I994 Fac. de Ciencias Quimicas $8.0010 Ciudad Universitaria Casilla de Correo 6 I 50 16-Cordoba (Argentina)

Table 1. Effect of sorbitol on heat-treated IgG IgG condition Native Heating 1.5 h Native Heating 1.5 h Heating 10 h

phosphate phosphate sorbitol33% sorbitol33% sorbitol33% buffer buffer

~

Polymer 2.6 32.2 2.3 6.5 11.2 Dimer 19.0 ND 18 ND ND Monomer 76.3 67.7 78 92.7 88.7 Split products 2.0 ND 1.5 ND ND

ND =Not detected. Results are expressed in percentages.

running buffer at an adequate concentration that provides an appropri- ate absorbance value and filtered in 0.22 pm Millipore filter before loading the column. The filtration process is done to avoid insoluble particles in order to preserve the column performance. This process does not alter the relative proportion of polymers, dimers and mono- mers of IgG. The running buffer was sodium phosphate 100 mM pH 6.8, filtered and degassed.

Calorimetric studies were performed with a MicroCal MC-2D scanning calorimeter with digital acquisition data software provided by the manufacturer. The sample was degassed before being injected into calorimeter cell (cell volume 1.2388 ml). The reference cell was filled with 100 mM phosphate buffer pH 6.8 or bufferisorbitol 33% Both cells were under Nz pressure (2 atm.). The normalized scan rate was 55.6"Cih and the IgG working concentration was 0.026 mM, ade- quate for DSC studies. The scans showed are representative of at least three determinations and the errors in the determination of the temper- ature initiation ofthe unfolding process (T, ) and in the temperature at maximal heat capacity (T,,,,,,), were within + _ I T .

Results and Discussion

Using molecular exclusion chromatography allows one to determine the relative fraction of aggregates in the sam- ple before and after heating. The aggregation process was analyzed by analytical chromatography in the Superose-12 gel filtration column (fig. I ) and the quantitative evaluation is shown in table 1. The native IgG from fraction I1 of Cohn in phosphate buffer presents a characteristic profile of elu- tion (fig. la). The amount of oligomers and aggregates of high molecular weight increases with heating time (1.5 h at 60"C, fig. I b). The sorbitol 33% w/v induces a decrease in dimer content and significantly reduces the amount of oli- gomers when the IgG is heated at 60°C (compare fig. la and c, table 1). The decrease in dimer content in the presence of the polyols is a temperature dependent effect since the addi- tion of sorbitol per se without heating does not modify the profile obtained with buffer alone (table 1). Heating the sample during 10 h increases the amount of aggregates near

the maximum levels recommended by WHO [I] and the sample shows a slight but definite turbidity (fig. Id, table 1).

The influence of sorbitol on the thermotropic properties of IgG was examined by differential scanning calorimetry (DSC). The heat capacity curve in the absence and in the presence of sorbitol is shown in figure 2. Naturally, IgG is unstable against heat and the unfolding (heat denaturation) begins near 52°C with a T,,, at 69"C, then a considerable exothermic peak is observed due to the aggregation of the unfolded molecules. Sorbitol induces stabilization and shifts T, from 52 to 65°C and T,,, from 69 to 76°C (fig. 2). The thermogram of figure 2 for total human IgG exhibits three main transitions, in agreement with the reported un- folding for rabbit IgG [7]. For these cases, the shift in the temperature of the initiation (T,) of the cooperative unfold- ing process is more important than the temperature of half completion (T,,,) or the temperature of maximal heat capac- ity of the thermogram (T,,J.

The mechanism by which sorbitol induces thermal stabi- lization is due to an unfavorable interaction ofthe cosolvent with the protein surface. The effect brought about by the polyol is a preferential hydration on the protein surface with the exclusion of the additive molecules from the cosolvent- protein interface. As the folded conformation has less area exposed to solvent than the unfolded conformation, the con- sequence of the unfavorable interaction with the additive is the stabilization of the native structure of the protein [6]. The aggregation of IgG is an irreversible process [5] and therefore the effect of sorbitol as a stabilizing agent is con- siderable since the additive not only disaggregates dimers on heating but also is preventing the massive aggregation after the complete unfolding.

The results showed that even when sorbitol seems to be an effective stabilizing agent, the long-term heating process required by pasteurisation is not sufficient to prevent unde- sired unfolding and aggregation (fig. 1 and table I ) . Immu-

Gonzalez/Murature/Fidelio Effect of Sorbitol on IgG Heat-Treatment

Fig.1. Elution profile of IgG from size exclusion chromatography on Superose-12 column. The final IgG concentration during treatment was 5% w/v. The corresponding elution volumes ofstandard is shown in a IgG without additive in phosphate buffer. b IgG without additive heated 1.5 h at 60°C. c IgG with sorbitol33% heated 1.5 h at 60°C. d IgG with sorbitol33% heated 10 hours at 60°C.

Fig.2. Thermogram for IgG in the ab- sence of additive (dashed), and in the pres- ence of sorbitol 330/0 (filled). IgG concentra- tion 0.026 mM.

noglobulin solution pasteurised in this manner will require further processing to prepare a pharmaceutically acceptable product. Our results are in disagreement with those reported by Uemura et al. [3]. In their report the authors stated that human IgG treated by sorbitol33Yo resulted in a low level of

IgG oligomers ( ~ 0 . 2 % ) . However, from this paper is not clear if this value is after further PEG precipitation of IgG heated or not.

Pasteurization only involves the following steps: (1) di- rect dissolution of Cohn fraction I1 paste with the additive,

3

(2) one-step pasteurization and (3) removal by ultrafiltra- tion and diafiltration of the additives and dilution of the IgG with the final adequate intravenous solution. From an eco- nomical, pharmaceutical and medical point of view, the fi- nal liquid product has also some advantages: (1) ready to use without dilution of a lyophilized IgG powder avoiding nurs- ing handling, (2) any macroscopic aspect of precipitation or turbidity in the liquid IgG preparation can be easily exam- ined and discarded during the quality control storage and (3) a more simple packaging, and finally (4) cheaper proc- essing.

The ideal additive (or a mixture of combined additives) to prevent a complete heat denaturation of human IgG dur- ing the viral inactivation in liquid state is still under lab-

oratory investigation. Our experience, either on laboratory or pilot scale, using sorbitol33% as unique additive was not satisfactory as a single method to obtain intravenous IgG.

Acknowledgments

This work was supported by a specific grant from SECyTECOR from Provincia de Cordoba, and from Laboratorios de Hemoderivados from Universidad Nacional de Cordoba, Argentina. M.G. is a fellow from CONICOR and G.D.F is member of the Carreer of Investigator from CONICET, Argentina. The authors are grateful to Dr. J. Zarzur and his collaborators for technical assistance and critical reading ofthe manuscript.

................................................................................................................................................... References

4

1 Suomela H: Viral inactivation of blood and blood products. WHO LBS 1992; 92-95.

2 Welch AG, Cuthbcrtson B. McIntosh RV. Foster PR: Non A. non B hepatitis from intravenous immunoglobulin. Lancet 1983;ii:1198-1199. Uemura U, Uriyu K, Hirao U, Takechi K. Ishika- wa H, NakajimaT, Kagitani Y, Yokoyama K, Fu- nakoshi S, Nishida M, Yabushita S, Furuta K, Hamamoto Y, Tochicura T, Yamamoto M: Inac- tivation and elimination of viruses during the fractionation of an intravenous immunoglobu- lin preparation: Liquid heat treatment and poly- ethylene glycol fractionation. Vox Sang 1989; 56: 155-1 61.

3

4 Hamalainen E, Suomela H, Ukkonen P: Virus inactivation during intravenous immunoglobu- lin production. Vox Sang 1992:63:6-1 I . Rosenqvist E, Jossang T, Feder J : Thermal prop- erties of human IgG. Mol lmmunol 1987;24: 459-501. Timasheff SN, Arakawa T: Stabilization ofpro- tein structure by solvents; in Creighton TE (ed): Protein Structure. A Practical Approach. Ox- ford, IRL Press, pp 331-345. Tischenko VM, Zajlalov VP. Medgyesi GA. Potekhin SA, Privalov PL: A thermodynamic study of cooperative structures in rabbit immu- noglobulin G. Eur J Biochem 1987; 126:s I7- 52 I

5

6

7

GonzaledMuratureiFidelio Effect of Sorbitol on IgG Heat-Treatment