Safety and Efficacy of Immune Globulins

Investigators

Abstract

Our laboratory is involved in the following research efforts: 1. Comparison of immune globulin product antibody profiles against bacterial and viral pathogens, particularly those for which immune deficient people are at risk; 3. Determination of molecular integrity of products over time, and biochemical research on improved markers of stability, as well as long-term monitoring of specific products and standards; 4. Development and provision of standards and methods, e.g. testing and distribution of an interim standard for vaccinia immune globulin. 5. Research on mechanisms of infusion-related adverse reactions to immune globulin, using cell-culture based assays, and determining etiology of proinflammatory cytokine induction. 6. Development and testing of animal-derived anthrax immune globulin, to determine whether or not this type of preparation has potential as a therapeutic agent. 7. In vivo potency assay development for vaccinia immune globulins, including assays in immune deficient mice, and studies of ocular vaccinia. Cytokine Release as a Mechanism Mediating IGIV Adverse Reactions. Inflammatory immune responses with clinically relevant adverse consequences can be associated with TNF, IL-1, IL-6 and other cytokines. These cytokines are rapidly produced by bacteria and bacterial components, such as DNA and LPS, but can also be induced, under certain circumstances, by triggering of complement receptors. Manufacture of plasma derivatives involves steps which are not aseptic and which can lead to bacterial contamination. Intact bacteria are removed by sterile filtration, but this step does not remove LPS, bacterial DNA (bDNA), or other microbial constituents. Adverse events with IGIV administration include fever and hypotension; fever has been associated with elevated TNF levels in recipients, and hypotension is a known side effect of LPS. We have observed patterns of cytokine release by human monocytes that are similar after IGIV or microbial stimulation. The levels of bacterial DNA that were detectable in IGIV's were too low to effectively stimulate cytokine products, but studies using polymyxin B suggested that LPS was responsible for a portion of cytokine release. Multiple lots of IGIV from different manufacturers were assessed to determine whether they could stimulate TNF or IL-1 release from human mononuclear cells at concentrations that are present in humans after IGIV infusion. Many lots stimulated significant TNF and IL-1 production, most consistently observed when non-heat inactivated human serum was used in the culture media. Addition of polymyxin B, which neutralizes LPS, did not always abrogate cytokine release. Although no IGIV lots had LPS contamination > 0.91 EU/ml (the industry standard), dose response experiments using LPS showed that amounts lower than 0.91 EU/ml can stimulate cytokine release from monocytes. Bacterial DNA was detected in IGIV using PCR primers which recognized broadly conserved bacterial DNA sequences; the levels detected were lower than those needed to stimulate monocyte cytokines in vitro. However, the presence of bDNA suggests that other bacterial constituents, such as cell wall components of gram positive bacteria, could also be present, some of which may not be detected by current lot release standards such as LPS testing and rabbit pyrogen testing. Previous results show that in the presence of (complement-component-containing) non-heat inactivated serum, proinflammatory cytokine release is greater, suggesting that complement components interact with IGIV, to produce this effect. In the past year, we have changed the tissue culture system from individual donor monocytes, to consistent monocyte pools from ten donors, which are frozen in aliquots. This has enabled us to accomplish many experiments using the same monocytes. We have found that certain products are more likely to stimulate cytokine release than others, and confirmed that these effects cannot be completely attributed to LPS. Current work is focused upon increasing sensitivity of readout systems to stimulation by low levels of LPS and other microbial constituents. We have observed that surface interleukin-15 (IL-15) on monocytes appear to me a more sensitive method for detecting low levels of LPS, to the picogram range. Current work is focused upon determining what the lowest concentrations of various microbial components (LPS, CpG, lipoproteins, poly (I:C)) are, that can still increase expression of IL-15 on fresh or pooled human monocytes. Other proinflammatory cytokines are also be measured for comparison. IGIV's will be tested using the enhanced-sensitivity system. In order to discern the potential causative agent(s) of adverse reactions, the effect of toll-like receptor (TLR) blockade will be explored. By blocking various TLRs or determining signaling pathways, it may be possible to classify the contaminating agent(s), since TLR usage is specific for particular microbial constituents. This work will allow identification and characterization of agents in IGIV which may mediate adverse events, and could lead to development of better predictors of adverse events. Overall, these studies will add to understanding of how infusion-related IGIV adverse events are mediated, and may lead to manufacturing improvements which can abrogate such effects from all immune globulins. Characterization of Anti-Vaccinia Antibodies in Licensed Immune Globulins Intravenous (Human). Our laboratory has used the SCID mouse lethality model to demonstrate neutralization of vaccinia virus with VIG and VIGIV. We have shown that 5 mg of VIGIV (but not VIG) can neutralize 106 vaccinia organisms (Wyeth strain), when co-incubated with virus, and injected into SCID mice. In experiments using commercially licensed IGIV products, we have shown that all IGIV's tested to date can delay vaccinia-mediated mortality in SCID mice, demonstrating that these products have some anti-vaccinia activity. We have run Western blots to characterize the vaccinia protein binding present in VIG products and IGIV's. This data indicates the presence of multiple antibody specificities in both types of product, although the specific banding pattern differs in some respects between IGIV and VIG's. IGIV-mediated disease delay was unaffected by dialysis of the starting material, indicating that artifacts due to excipients are unlikely. We have recently compared IGIV's (high and low titer) to VIGIV (in diminishing doses), in SCID mice infected with 105 PFU of virus. The results suggested good efficacy of IGIV's against this low dose viral challenge. In pre- and early post-exposure prophylaxis studies, which are similar to possible scenarios in people, we have shown that VIG's and IGIV's demonstrate efficacy. Current experiments are focused upon dose ranging, and timing of VIG and IGIV dosing, including treatment of pox-expressing mice. Planned studies include determination of protection in mice that have immune deficits similar to those present in human populations, e.g. HIV (CD4-deficient), and primary immune deficiency (B cell deficient mice). Scarification and protection experiments will be performed in these models of human immune deficiency, with IGIV and VIG's. The function and efficacy of anti-vaccinia antibodies may be different in IGIV compared to VIG products, because the donors were vaccinated in the distant past. Any differences in antibody subclass or affinity between anti-vaccinia antibodies in IGIV and VIG products would have potential clinical consequences. Such findings would also inform efforts to produce more potent VIG product. Pending funding, we plan to study antibody subclass and affinity in high-titer IGIV and VIGIV. Antibody subclasses have been enriched (using protein A sepharose columns) from selected high-titer IGIV lots, as well as VIGIV. Our laboratory is experienced in subclass isolation. Each subclass will be tested for neutralizing ability by PRN assays, B-gal assays, and in vivo neutralization in SCID mice. Antibody affinity and off-rates are undergoing characterization using the Biacore. A Biacore chip has been coated with vaccinia virus, and studies are beginning to optimize a system that can be used with polyclonal, complex antibody mixtures. Positive controls, including monoclonal antibody preparations, are now being tested against vaccinia epitopes. Ongoing experiments are focused upon characterizing conditions (e.g. concentration of antigen on chip, flow rates, optimal serum/IGIV concentrations for detection of specific binding). Interaction curves (refractive units) will be generated and compared. It is hoped that the reaction will follow "pseudo first order" kinetics, enabling calculation of accurate affinity constants. These results will may enable better predictive characterization of vaccinia antibodies in VIG's and IGIV's, and the methods will be broadly useful for other specific immune globulins. The results would also be applied to stability monitoring of specific VIG's over time. Potency of VIG Products: Development of in vivo models to assess VIG potency. We have recently observed that VIG preparations which appear to have equivalent PRN titers, have different ability to neutralize vaccinia in vivo in SCID mice, suggesting that PRN assays as currently performed are unable to measure an important marker(s) of efficacy. Vaccinia viruses occur in two infectious forms. The intracellular mature form (IMV) remains inside cells until lysis, and accounts for the majority of virus. The extracellular enveloped form buds from live cells, is surrounded by an extra host-derived lipid envelope, and contains host and virus proteins that are absent from IMV. Since the 1970's, it has been recognized that distinct subsets of antibodies neutralize each form; furthermore that antibodies to EEV as well as IMV are required for active and passive immune protection. The PRN, by nature of the viral preparation used, measures antibodies against IMV, but not EEV, and this could be one possible cause of the discrepancy between the in vivo and in vitro neutralization data. Other possibilities include effects of complement fixation, and antibody-dependent cellular cytotoxicity induction by VIG preparations, which are also not measured in vitro, but may differ among immune globulin preparations. The SCID mouse model, while useful in preliminary experiments, has several drawbacks, both practical and theoretical. All experiments to date have been performed by mixing antibodies and virus in vitro, followed by i.p. injection of the combined preparations. While this demonstrates differences among products (which are not detected by PRN or HPLC), the model has been criticized because it does not simulate the likely human clinical situation. Indications for VIG in humans have been prophylaxis against complications (in eczema, immunodeficiency, pregnancy), and treatment of disseminated infection after vaccination (progressive vaccinia and eczema vaccinatum) (CDC Advisory Committee on Immunization Practices, Vaccinia (Smallpox) Vaccine Recommendations of the Advisory Committee on Immunization Practices, 2001 at www.cdc.gov/mmwr/preview/mmwrhtml/rr5010a1.htm; Goldstein et. al. Pediatrics 55 (3): 342-7, 1975). It has also been suggested that smallpox attack rates among primary contacts can be diminished by VIG. Thus more relevant models would use different routes of exposure (e.g. intradermal, mucosal), would measure in vivo neutralization of virus by VIG, and would compare the ability of products to provide prophylaxis and treat extensive vaccinia infection. Furthermore the lethality assay in SCID mice is prolonged (28 days minimum), and requires extensive monitoring twice daily, 7 days/week. A new, rapid, relevant animal model could be used to develop optimal in vitro assays, or combinations of assays; to compare products; and to determine product characteristics that are required for efficacy. We have begun to develop in vivo models of widespread vaccinia infection, using relevant exposure methods, in mice. Eczema-prone mouse strains are being infected with vaccinia, and monitored clinically for skin lesions. Scarification studies in normal mice have not lead to pox formation (even using the Wyeth strain), so we are focusing upon other mouse models that may simulate human cutaneously acquired infection that has disseminated. Immunodeficient mice (SCID, and IL-15 KO) will be injected with doses of vaccinia (Wyeth strain, and Dryvax, 103 to 106 PFU) i.d. (flank), and by scarification, to determine dose-response to virus, and time to lethality. IL-15 KO mice are susceptible to severe vaccinia infection, due to lack of NK cells and inability to generate CTL. SCID mice lack B and T cells, and are thus unable to generate either cellular or humoral immunity, although they retain NK function. A dose of virus and optimal mouse model will be selected using results of preliminary experiments, and mice will receive prophylaxis or treatment (in separate experiments) with VIG preparations (5- 50 mg/mouse) i.v. and/or i.p. The outcome for these experiments is lethality, and no consistent surrogate marker has been validated that could permit earlier completion of experiments. The use of VIG's for keratitis (corneal inflammation) by vaccinia is controversial because of a report that VIG worsens corneal scarring in rabbits with ocular vaccinia infection. However, no such evidence has been generated in humans; furthermore ocular vaccinia may involve other structures in the orbit, and the eyelids, that could respond well to VIG. Using a beta-galactosidase-expressing vaccinia virus, we are developing a mouse model of vaccinia keratitis. Early results are promising in that infection can be established. We are working to optimize the infection system. This will be followed by detailed studies of whether VIG's alter infection clearance, and whether they result in long-term corneal damage. VIG products may be required for emergency use in case of mass vaccination or a smallpox attack. It is important to maintain current data on stability and potency of these products. VIG (Baxter 1994, the current IND product available for use), and VIGIV products (all lots) are retained by our group, and stored at the recommended storage temperature. We periodically assess each preparation by HPLC (which indicates fragmentation, and aggregation), and by B-galactosidase assay, using our new frozen CBER working standard as a control preparation. As assays evolve and become more specifically correlated with protection, we will perform additional in vivo and in vitro evaluations as they are deemed relevant. This project incorporates FY2002 projects 1Z01BQ004018-08, 1Z01BQ004024-01, and 1Z01BQ004025-01.