Considerations of pool size in the manufacture of plasma derivatives

BACKGROUND

The pooling of human plasma from many donors for the purpose of manufacturing therapeutic proteins increases the risk of exposing recipients of these proteins to pathogens that may contaminate 1 or a few units included in the pool.

STUDY DESIGN AND METHODS

This risk is estimated for a range of manufacturing scales that would derive material from a varied number of donors and for a number of hypothetical infectious agents that may exist in the donor population over a wide range of prevalence. Risk is also calculated both for recipients of single doses of a plasma protein and for those who depend on long-term treatment with plasma derivatives.

RESULTS

Risk of exposure increases with pool size and the prevalence of the agent in question and accumulates with repeated treatments with material manufactured from different pools.

CONCLUSION

Reducing pool size would at best decrease this risk in proportion to the reduction in manufacturing scale. However, for individuals requiring repeated or continuous treatments, the risk of exposure to all but the rarest infectious agents would be only minimally affected, even by large reductions in manufacturing scale.

Effect of immune globulin on the prevention of experimental hepatitis C virus infection

The efficacy of postexposure prophylaxis for the prevention of hepatitis C virus (HCV) infection was studied in experimentally infected chimpanzees. Three chimpanzees were inoculated with HCV: Two were treated 1 h later with anti-HCV--negative intravenous immune globulin (IGIV) or hepatitis C immune globulin (HCIG), and a third animal was not treated. HCV infection was detected in all 3 animals within a few days of inoculation. Once passively transferred anti-HCV declined in the HCIG-treated animal, there was an increase of HCV antigen (Ag)--positive hepatocytes followed by reappearance of anti-HCV; HCV Ag disappeared concordant with the development of acute hepatitis. Acute hepatitis C developed in both the IGIV-treated and untreated chimpanzees, with peak liver enzyme activity on day 59, but was delayed in the HCIG-treated animal until day 146. Postexposure HCIG treatment markedly prolonged the incubation period of acute hepatitis C but did not prevent or delay HCV infection. IGIV had no effect on the course of HCV infection.

Detection and characterization of hepatitis C virus RNA in immune globulins

BACKGROUND

Hepatitis C virus (HCV) RNA was measured in immune globulins and its chemical and physical properties were characterized.

STUDY DESIGN AND METHODS

The study examined 69 immune globulin lots from 7 manufacturers, including 44 intravenous and 25 intramuscular immune globulin preparations. In addition, 8 experimental intravenous immune globulin preparations were investigated. Detection and quantitation of HCV RNA were achieved by reverse transcription and nested polymerase chain reaction at limiting dilution. A multi-antigen anti-HCV enzyme immunoassay was also used to test these immune globulins.

RESULTS

The highest level of HCV RNA was found in an experimental immune globulin lot derived from a plasma pool made up of 186 anti-c100-3-reactive units. HCV RNA was detected only in 1 of 7 manufacturers' experimental intravenous immune globulin preparations derived from a pool made up of 2887 anti-c100-3-negative units. It was also detected in commercial intravenous immune globulin lots prepared by the same manufacturer from source plasma, but not from recovered plasma. More than half of the commercial intramuscular immune globulin lots, including specific immune globulin products, were HCV RNA positive. All immune globulin products examined were reactive for anti-HCV. Certain similarities were found for HCV RNA present in an immune globulin product and plasma. Ethanol at 20 or 25 percent had no effect upon the buoyant density of HCV RNA.

CONCLUSION

Many immune globulin preparations contained HCV RNA, with levels depending upon both the type of starting plasma and the manufacturing process. Exposure to ethanol did not appear to affect the physical characteristics of HCV RNA.

Dimer formation in immunoglobulin preparations and speculations on the mechanism of action of intravenous immune globulin in autoimmune diseases

IgG dimers occurring in therapeutic Ig preparations have been characterized as Id-anti-Id, in that the sites of interaction are localized to the distal tips of the Fab arms. The observation that such dimers are prevalent in Ig or Igi.v. prepared from large plasma pools, but absent in preparations from a single individual, supports the notion that the individual immune repertoire is small with respect to the species repertoire. A crude mathematical model that attempts to relate Id-dimer content to the number of donors is presented. This model suggests that Id-pairs may be much more prevalent in Ig than is reflected by the Id-dimer content, inasmuch as the concentrations of the individual Ids and anti-Ids may limit the equilibrium level of dimer. The model further suggests that the antibody diversity in Ig derived from thousands of donors may be representative of the species repertoire; hence, virtually all specificities, including anti-Id specificities, will be included. Because Ig is derived from normal healthy donors, it should be relatively free of pathogenic autoantibodies. However, there is no reason to suspect that anti-Ids to such autoantibodies would not occur, and indeed the presence of such anti-Ids has been demonstrated. Several mechanisms have been proposed by which such anti-Ids might ameliorate autoimmune disease. They may directly inhibit the binding of autoantibody to its target antigen, or they may target for destruction those cells expressing or secreting autoantibody. It may well be that anti-Ids play no role in the mechanism of action of Igi.v. in autoimmune disease. Such appears to have been demonstrated, for example, in the treatment of ITP. There is an obvious need for additional studies in order to elucidate the mechanism of action of Igi.v. in various autoimmune diseases. Experimental animal models of autoimmune disease, such as the mouse model for systemic lupus erythematosus (Mozes et al. 1993), might be very useful in this regard. Finally, it needs to be emphasized that the usefulness of high doses of Igi.v. in many autoimmune diseases remains to be established by controlled clinical trials. Because Igi.v. is a limited resource, and one which cannot be produced through biotechnological advances (at least in the foreseeable future), its widespread use should be restricted to the treatment of diseases for which efficacy has been demonstrated. To do otherwise might deprive appropriate patients of a valuable therapy.

The effect on the safety of intravenous immunoglobulin of testing plasma for antibody to hepatitis C

BACKGROUND

The safety of intravenous immunoglobulin (IGIV), manufactured from units testing negative for antibody to hepatitis C virus (anti-HCV), was investigated.

STUDY DESIGN AND METHODS

A study involving five chimpanzees was performed to determine whether the safety of IGIV would be compromised if units of plasma that reacted for anti-HCV were withheld from pools from which IGIV is manufactured. In the first phase of the experiment, two chimpanzees were infused with 25 mL per kg of unprocessed, pooled plasma from 2887 donors who did not react for anti-HCV in single-antigen (c100-3) enzyme-linked immunosorbent assays. In the second phase, each of three chimpanzees was infused with 1000 mg per kg of IGIV manufactured from the same plasma units. The immunoglobulin was made by seven United States-licensed manufacturers, each using its own approved method. Each chimpanzee received an equal dose of each manufacturer's IGIV.

RESULTS

The two chimpanzees that received anti-c100-3-nonreactive, unprocessed pooled plasma became infected with HCV. The three chimpanzees infused with IGIV did not show any evidence of infection with HCV 15 months after inoculation. Two of these animals were challenged with human non-A,non-B hepatitis-infectious plasma, and both subsequently showed evidence of HCV infection.

CONCLUSION

These studies demonstrate that, as determined by infectivity for chimpanzees, 1) the withholding of plasma units that react for anti-c100-3 from pools from which plasma products are manufactured does not render the source material noninfectious, and 2) the safety of IGIV manufactured from such plasma pools is not compromised by withholding the units that react for anti-c100-3.

Partitioning of hepatitis C virus during Cohn-Oncley fractionation of plasma

Because of concern about the safety of immune globulins with respect to transmission of hepatitis C, the partitioning of hepatitis C virus (HCV) during alcohol fractionation of a plasma pool prepared exclusively from anti-HCV-reactive donations was examined. Quantitation of HCV RNA was accomplished by nested polymerase chain reaction (PCR) at limiting dilutions. One PCR unit was arbitrarily defined as the minimum amount of HCV RNA from which an amplified product could be detected. The starting plasma pool contained 1.4 x 10(5) PCR units per mL. Most of the HCV RNA was found in cryoprecipitate and in Cohn fractions I and III, but it was also detected in fraction II, which is used for immunoglobulin G preparations. A 3.4-percent solution of IgG prepared from this fraction II contained 30 PCR units per mL. The fractionation process leading to immune globulin resulted in overall reduction in HCV RNA by a factor of 4.7 x 10(4). Although the presence of HCV RNA in the final product does not necessarily imply the presence of infectious virus, this work suggests that the safety of immune globulins with respect to HCV transmission is not due solely to the partitioning of HCV away from the immunoglobulin fraction.

A view of the human idiotypic repertoire. Electron microscopic and immunologic analyses of spontaneous idiotype-anti-idiotype dimers in pooled human IgG

It has previously been reported that up to 40% of the molecules of human IgG from pooled plasma (100,000 donations) spontaneously dimerize, whereas IgG prepared from a single donor contains only trace quantities of dimer. We have conducted immunoelectron microscopic analyses on such samples and have verified that the majority of dimers are composed of complexes in which two arms of each molecule are bound in a reciprocal fashion at or near the distal tips of their respective arms as previously seen in bona fide Id-anti-Id complexes. A significant role for Fc was ruled out by showing the formation of dimeric ring structures in purified F(ab')2 samples. Spontaneous dimerization was also observed in pooled bovine or mouse IgG, but not in that from single animals. Radiolabeled monomeric, single donor human IgG was used as a probe to investigate dimer formation; this material readily codimerized with IgG from other donors or from pooled plasma (either human or bovine), but did not dimerize with IgG from the same donor. Subclass analysis of multiple donor human IgG revealed that the most flexible subclass (IgG3) was overrepresented in the dimer fraction, whereas one of the less flexible subclasses (IgG2) was underrepresented. Although IgG2 could dimerize to some extent with IgG of other subclasses, it could not self-dimerize. These data suggest that structural constraints (hinge flexibility) may play a role in limiting dimerization. This suspicion was confirmed by showing that an additional 15.5% of the monomeric IgG fraction from a multidonor sample could dimerize after a chemically induced increase in hinge flexibility. Our results are interpreted to show that, although the number of functionally distinct Id produced by a species is immense, it is nontheless finite. Moreover, the Id repertoire of an individual is much smaller than that of the species. Pooling the IgG from a number of individuals increases the Id diversity, which increases the chance that any given Id-bearing molecular will encounter a complementary partner.

A view of the human idiotypic repertoire. Electron microscopic and immunologic analyses of spontaneous idiotype-anti-idiotype dimers in pooled human IgG

It has previously been reported that up to 40% of the molecules of human IgG from pooled plasma (100,000 donations) spontaneously dimerize, whereas IgG prepared from a single donor contains only trace quantities of dimer. We have conducted immunoelectron microscopic analyses on such samples and have verified that the majority of dimers are composed of complexes in which two arms of each molecule are bound in a reciprocal fashion at or near the distal tips of their respective arms as previously seen in bona fide Id-anti-Id complexes. A significant role for Fc was ruled out by showing the formation of dimeric ring structures in purified F(ab')2 samples. Spontaneous dimerization was also observed in pooled bovine or mouse IgG, but not in that from single animals. Radiolabeled monomeric, single donor human IgG was used as a probe to investigate dimer formation; this material readily codimerized with IgG from other donors or from pooled plasma (either human or bovine), but did not dimerize with IgG from the same donor. Subclass analysis of multiple donor human IgG revealed that the most flexible subclass (IgG3) was overrepresented in the dimer fraction, whereas one of the less flexible subclasses (IgG2) was underrepresented. Although IgG2 could dimerize to some extent with IgG of other subclasses, it could not self-dimerize. These data suggest that structural constraints (hinge flexibility) may play a role in limiting dimerization. This suspicion was confirmed by showing that an additional 15.5% of the monomeric IgG fraction from a multidonor sample could dimerize after a chemically induced increase in hinge flexibility. Our results are interpreted to show that, although the number of functionally distinct Id produced by a species is immense, it is nontheless finite. Moreover, the Id repertoire of an individual is much smaller than that of the species. Pooling the IgG from a number of individuals increases the Id diversity, which increases the chance that any given Id-bearing molecular will encounter a complementary partner.

Immunoglobulin G dimer: an idiotype-anti-idiotype complex

Immunoglobulin G (IgG) prepared from pooled human plasma contains variable amounts (up to 40%) of IgG dimer whereas IgG isolated from the plasma of a single individual is essentially monomeric. The amount of dimer increases with the number of donors contributing to the plasma pool from which the IgG is prepared. Dimerization is reversed by increasing the temp or decreasing the pH. Two intact antibody-combining sites (paratopes) are required for optimal dimer formation; the Fc region is unnecessary. These findings strongly suggest that IgG dimer consists of idiotype-anti-idiotype pairs formed between molecules of IgG from different individuals, and that a mechanism exists for suppressing idiotype-anti-idiotype formation in vivo.

Inactivation and partition of human T-cell lymphotrophic virus, type III, during ethanol fractionation of plasma

Because of concern about the safety of immune globulins prepared for injection, we studied the effects of ethanol fractionation of human plasma on human lymphotropic virus, type III, (HTLV-III) by spiking the products of various fractionation steps with HTLV-III. Tests of inactivation and removal indicated that the ratio of residual live virus in plasma fractions/live virus in starting plasma was about 1 X 10(-15) for precipitate II from which immune globulin for injection is manufactured. The results are reassuring regarding the potential safety of immune globulin.

Kinetics of activation and autoactivation of human factor XII

The kinetics of the enzymic reactions that participate in the contact activation system of human plasma were examined. These reactions are potentiated by dextran sulfate, a negatively charged solute that mimics many of the effects of glass or kaolin on this system. The reactions of reciprocal activation, consisting of activation of factor XII by kallikrein and of prekallikrein by activated factor XII, follow Michaelis-Menten kinetics; values of kcat and Km for each of these reactions were determined in the presence of dextran sulfate and in its absence. In the presence of dextran sulfate, the catalytic efficiency for factor XII activation was increased 11 000-fold, and that for prekallikrein was increased 70-fold. Autoactivation of factor XII in the presence of dextran sulfate also follows Michaelis-Menten kinetics with kcat = 0.033 s-1 and Km = 7.5 microM. This finding supports the concept that autoactivation is an enzymic process, initiated by traces of activated factor XII which are invariably present in factor XII preparations. At prekallikrein and factor XII levels equal to those in plasma, reciprocal activation is approximately 2000-fold more rapid than autoactivation. Thus, reciprocal activation is the predominant mode of factor XII activation in normal plasma.

Activation of factor XII by dextran sulfate: the basis for an assay of factor XII

A system was developed for studying the activation of factor XII (Hageman factor) in the presence of dextran sulfate (DS). Salient features of the system included low ionic strength (0.08), low concentration of factor XII (approximately 1/10,000 that in normal plasma), and an excess of exogenous prekallikrein (PK). In this system, factor XII was rapidly converted to the 80,000 molecular weight (mol wt) form of factor XIIa (alpha-factor-XIIa). Once formed, the factor XIIa converted PK to kallikrein at a rate that was proportional to the amount of factor XII originally present in the incubation mixture. This system was used to construct a simple sensitive assay for factor XII in plasma and other biologic samples. The kallikrein produced was measured spectrophotometrically with the chromogenic substrate (H-D-Pro-Phe-Arg-p-nitroanilide (S-2302). This assay was shown to be independent of the high molecular weight kininogen and the PK content of the sample being analyzed. The measurements obtained were consistent with fundamental enzymologic principles and, if desired, could be processed with a simple calculator program to achieve linear standard curves. When applied to the quantitation of factor XII in plasma, the assay yielded values in close agreement with those determined by coagulant assay or by radial immunodiffusion.

Plasma prekallikrein: quantitative determination by direct activation with Hageman factor fragment (beta-XIIa)

This report describes a plasma prekallikrein assay which, unlike methods that employ contact activation, is not affected by the factor XII or HMW kininogen content of the plasma analyzed. In this assay beta-XIIa, a potent fluid-phase activator of prekallikrein, is added to diluted plasma in the presence of 20% acetone (to inactivate kallikrein inhibitors) at 30 degrees C and the kallikrein generated is measured with the chromogenic substrate S-2302. Prekallikrein is fully activated under these conditions and the activity remains stable for at least 30 hr. The mean prekallikrein concentration in plasma samples from 24 healthy individuals was 1.50 +/- 0.35 (S.D.) S-2302 U/ml, corresponding to 20.3 +/- 4.7 micrograms/ml prekallikrein (the specific activity of highly purified human prekallikrein was determined to be 74 S-2302 U/mg). In contrast, the mean concentration in five plasma samples from patients deficient in HMW kininogen was 0.38 +/- 0.02 S-2302 U/ml. No activity was generated in prekallikrein-deficient plasma, and essentially normal levels (1.35 +/- 0.18 S-2302 U/ml) were measured in plasmas from three patients with factor XII deficiency. Plasma prekallikrein was also quantitated by radial immunodiffusion, which gave results similar to those obtained by functional assay with beta-XIIa. The determination of plasma prekallikrein by direct activation with beta-XIIa in the presence of acetone offers several advantages over the use of contact activators such as dextran sulfate. These advantages include complete inactivation of kallikrein inhibitors and total activation of prekallikrein (even in plasmas deficient in other contact factors) without simultaneous generation of plasmin.

Preparation of beta-XIIa (Hageman factor fragment) from human plasma

beta-Factor XIIa (beta-XIIa, Mr approximately 30,000) was isolated from human plasma by a procedure which utilized, as an initial step, the adsorption of Factor XII to celite. Activation of Factor XII and subsequent release of beta-XIIa was brought about by the proteolytic action of co-adsorbed kallikrein. Two successive chromatographic procedures were then used to achieve a final purification of 4,420-fold over plasma and an overall field of 2.3 mg of beta-XIIa per liter.

A new preparation of modified immune serum globulin (human) suitable for intravenous administration. I. Standardization of the reduction and alkylation reaction

Immune serum globulin (ISG) prepared by Cohn cold alcohol fractionation of pooled human plasma was reduced with dithiothreitol (DTT) and alkylated with iodoacetamide and other alkylating agents. Our results show that there are a few labile interheavy chain disulfide bonds in ISG which react rapidly under mild, nondissociating conditions. The extent of disulfide cleavage is controlled primarily by the ratio of DTT to ISG until about 4-5 disulfide bonds have been reduced. We report detailed studies on the variables of ISG concentration, DTT to ISG ratio, pH, and time, leading to a chemically modified ISG that has a controlled and limited number of reduced and alkylated disulfide bonds.

A new preparation of modified immune serum globulin (human) suitable for intravenous administration. II. Functional characterization

The commercial immune serum globulin (ISG) product is suitable for use in passive immunization only by intramuscular administration; intravenous use results in vasomotor side reactions. We report the functional characterization of a chemically modified ISG (MISG) that is essentially free of nonspecific complement-binding activity and contains no new antigenic determinants. The half-life values for MISG given intravenously to rabbits, guinea pigs, and humans are shown not to be different from that of ISG. The MISG we describe is shown to fulfill the in vitro and in vivo criteria of stability, safety, function, and half-life, making it suitable for human intravenous administration.

Contact-activated factors: contaminants of immunoglobulins preparations with coagulant and vasoactive properties

The intramuscular or intravenous administration of ISG prepared from human plasma by ethanol fractionation can elicit such reactions as pain at the injection site, flushing, and even hypotension. Similar adverse reactions to plasma protein fraction, a volume expander also made by ethanol fractionation, have been associated with PKA (Hageman factor fragments) in the product. Twenty-five lots of commercial ISG were therefore analyzed for PKA and kallikrein, components of the contact activation system which could mediate such reactions through the generation of kinins in recipients. Kallikrein activity ranged from undetectable levels to > 60% of the total potential kallikrein activity in normal plasma. PKA, which was measured by its ability to catalyze the conversion of prekallikrein to kallikrein, ranged from 5% to 3950% of the activity in a reference plasma protein fraction that had caused hypotension. All but five lots increased vascular permeability in the guinea pig. The five lots which caused no increased were also the lowest in PKA and kallikrein activity. When ISG ws subjected to gel chromatography to separate the enzymic contaminants from immunoglobulin G, only the fractions containing PKA and/or kallikrein increased vascular permeability. Several lots of ISG shortened the nonactivated partial thromboplastin time of normal plasma fro 236 sec to 38 to 55 sec. During gel chromatography, coagulation activity was eluted in a position corresponding to a molecular weight of 150,000; it was inhibited by antibody to human factor XI. These data indicate that factor XIa is responsible for the coagulant activity observed and that PKA and/or kallikrein are potential mediators of vasoactive reactions to ISG.

Kinetics of activation of prekallikrein by prekallikrein activator

A 28 000 molecular weight activator of prekallikrein was isolated from human plasma, and the kinetics of its enzymic activity toward prekallikrein was investigated. The activation follows Michaelis-Menten kinetics with a kcat of approximately 3 S(-1); Km is strongly dependent upon the ionic strength. Under suitable conditions the activation obeys first-order kinetics, and the first-order rate constant may be used to quantitate prekallikrein activator activity. Thus a two-stage assay, in which the first step involves the activation of prekallikrein by the activator and the second step quantitates the kallikrein generated, was developed to allow the measurement of prekallikrein activator in biologic samples. The prekallikrein activator content of therapeutic protein solutions, as determined by means of this assay, correlated well with the hypotensive activity of these solutions as determined in an animal model.