Coalition of Silicone Survivors
Bron: LANCET 1992
COSSkids is onderdeel van Coalition of Silicone Survivors - Boulder Collorado USA
Goldblum, Randall M. Pelley, Ronald P.
Antibodies to silicone elastomers and reactions to ventriculoperitoneal shunts.
Goldblum, Randall M. Pelley, Ronald P.
Lancet; 8/29/92, Vol. 340 Issue 8818, p510, 4p, 4 graphs
Silicone elastomers used to make medical implants and prostheses are generally believed to be biologically inert. However, we have seen two patients who showed severe, apparently immunemediated, reactions to ventriculoperitoneal (VP) shunts.
We used an enzyme-linked immunosorbent assay in which Silastic tubing served as the solid-phase antigen to test serum from the two patients, five other VP shunt patients without inflammatory reactions, and nine healthy adults. IgG binding to Silastic tubing was consistently higher in the two patients than in the healthy or patient controls. The IgG seemed to be binding specifically, since IgG Fab fragments also bound to the tubing, and preincubation of serum with Silastic or silylated proteins removed most of the activity.
These findings show that specific immune reactivity to elastomers of polydimethylsiloxane can develop in human beings.
Elastomers formed of cross-linked polydimethylsiloxane, commonly called silicone elastomers, are widely used to make medical implants and prostheses. Silicone elastomers are generally believed to be biologically inert, since tissue responses are usually limited to mild foreign-body reactions. However, over the past 10 years, there has been increasing suspicion that rare, severe inflammatory reactions to silicone elastomer implants have an immunological basis.( n1, n2) We report evidence that specific antibody reactivity to polydimethylsiloxane develops in some patients after repeated exposure to Silastic (Dow-Corning Wright, Midland, Michigan, USA) shunt tubing.
The first patient was a caucasian girl who had a low lumbar myelomeningocele repaired on the first day of life. A ventriculoperitoneal (VP) shunt was inserted at 19 days and revised at 4 months. When the patient was 6 years old an abdominal pseudocyst formed around the shunt. A year later a neck wound from a shunt revision became red, exuded clear fluid, and separated, exposing the underlying shunt tubing. This process recurred eleven times over the next 4 years. Similar reactions were seen at sites of silicone-coated sutures. Cultures from the wound sites, collected many times over 3 years, and four tissue samples failed to show any microorganisms, despite use of many special stains and electronmicroscopy. Histology of the inflamed tissue showed granulomatous inflammation with many lymphocytes, macrophages, giant cells, epithelioid cells, and occasional plasma cells. There was mild hypergammaglobulinaemia (IgA 3.24 g/1; IgG 15.9 g/1). Serum samples taken when the child was 9-11 years old were stored frozen until analysis. At the ninth shunt replacement (age 14 years)precautions were taken to cover the track with intact tissue; it has been well tolerated for longer than 3 years.
The second patient was a Latin American girl who developed hydrocephalus at 9 months. Computed tomography confirmed hydrocephalus and showed partial agenesis of the corpus callosum and a Dandy Walker deformity. A VP shunt was placed and revised when she was 5 1/2 years. 2 months later the shunt was extruded in the same way as in the first patient. A few months later an abdominal pseudocyst was noted. Local reactions developed along the shunt, but no microorganisms were found by culture or histology. Serum concentrations of IgG (15.4 g/1) and IgM (4.6 g/1) were moderately high and there was unexplained eosinophilia (1.2 x 10 9/1). The reactions gradually resolved over 10 days. Serum samples were taken at the time of admission to our hospital and a year later.
These intense inflammatory reactions surrounding recently implanted Silastic tubing, in the absence of infection, suggested an immunological reaction, possibly to the tubing. To investigate this possibility, we developed an assay to detect antibodies to Silastic tubing. We also collected samples from five patients with VP shunts (aged 5-37 years) who had had no clinically apparent reactions. Four of these patients had had malfunction of their shunts, requiring at least one surgical revision. The assay was a modified enzyme-linked immunosorbent assay (ELISA)( n3) with Silastic tubing as the solid-phase antigen. 1 ml volumes of serial ten-fold dilutions (1/10-1/1000) of serum in 0.05 mol/1 phosphate-buffered saline, pH 7.1, with 0.05% (by volume) Tween 20 (Sigma Chemical Co, St Louis, Missouri, USA) were incubated overnight at 20 Celsius in polystyrene test-tubes with 1 cm sections of surgical Silastic tubing, sliced linearly to allow full exposure of inner and outer surfaces. The pieces of tubing were washed three times with 3-4 ml buffer, then incubated for 4 h with rabbit antibody to human IgG conjugated with horseradish peroxidase (Dako, Carpinteria, California, USA), washed again, then transferred to clean polystyrene test tubes. 1 ml enzyme substrate (0.2 g/1 orthophenylenediamine dihydrochloride in citrate buffer with 0.25% hydrogen peroxide) was added to each tube and the enzyme reaction was allowed to continue for 20 min, then it was stopped by acidification. The optical density at 492 nm was measured (EIA reader, BioRad, Richmond, California). Each experiment contained one serum sample from patient 1 and a buffer control. Assays were run in duplicate. Optical densities for the buffer controls were always less than 7% of the maximum value for serum from patient 1.
To find out whether the IgG binding to Silastic tubing was specific, IgG was separated from the serum of patient 1, by means of a protein A column (BioRad), and cleaved into Fc and Fab fragments with papain immobilised on Sepharose beads (Pierce Chemicals, Rockford, Illinois).
Fab fragments were separated from intact IgG and Fc fragments, concentrated, and checked for purity by sodium dodecyl sulphate polyacrylamide gel electrophoresis. The binding of Fab fragments to Silastic tubing was measured by means of the rabbit anti-IgG or a mixture of anti-* (This character cannot be converted in ASCII text) and anti-Lambda conjugates (Dako) diluted 1/500 and 1/3000, respectively.
We investigated the specificity of IgG binding to Silastic tubing further by assaying serum samples previously exposed to tubing or methylsilexane-conjugated proteins. 1/10 dilutions of serum were exposed to 1 cm or 2 cm sections of Silastic tubing, as described above. The tubing was then removed and a fresh section of tubing was added to each absorbed serum. Both sets of Silastic tubing were processed to determine the amount of IgG bound.
Methylsiloxane-conjugated proteins were made by mixing 1 mg crystalline bovine serum albumin or ovalbumin (Sigma) with 1 ml pyridine in glass tubes previously treated ("siliconised") with a reactive oligodimethylsiloxane (Sigmacote, Sigma). The crystals were sonicated in a bath for 90 min at 56 Celsius and the mixture was left at 23 Celsius overnight to allow the pyridine m infiltrate the protein crystals. Varying amounts ( 10-1000,Mu l) of the silylating agents orthobis-(trimethylsilyl)-trifluoroacetamide (Pierce) or N-trimethylsilylimadazole (Supelco, Bellefonte, Pennsylvania) were incubated with the protein/pyridine suspension for 2 h at 56 Celsius with sonication. The mixture was evaporated to dryness under nitrogen, and the residue was suspended in 1 ml Tween-phosphate buffer. Control protein complexes were prepared in the same way, but no silylating agent was added; these samples were exposed to pyridine in "siliconised" glass tubes.
The oligomethylsiloxane-protein complexes were then used to absorb serum from patient 1. About 50 Mu g of the complex mixture was incubated with 10 Mu l serum for 15 min at 37 Celsius in a total volume of 1 ml. The absorbed serum was then used in the Silastic tubing ELISA.
In the ELISA for IgG binding to Silastic tubing, binding of serum IgG from patients 1 and 2 was easily shown at all three dilutions; serum from VP shunt patients without inflammatory reactions and from 9 healthy adults showed much lower or undetectable IgG binding (fig 1). To show that the IgG binding to Silastic was attributable to specific antibody, IgG and Fab fragments from patient 1 were compared. Binding of the Fab fragment (anti-light-chain conjugate) approached that of intact IgG (fig 2). Fig 3 shows the effect of incubating serum from patients 1 and 2 and a normal adult with tubing before assaying the serum for IgG antibodies to Silastic. In both patients, most of the IgG able to bind to Silastic tubing was removed by the preincubation, though the concentration of total IgG determined by rate nephelometry did not change. The amount of IgG able to bind to Silastic tubing was also reduced substantially (22-56%) by preincubation of the serum with protein that had been subjected to pyridine treatment in silicone-treated glassware (fig 4). Incubation of serum with proteins treated with increasing amounts of either silylating reagent resulted in further decreases in IgG binding. Proteins conjugated with a large molar excess (680-1400moles of trimethylsiloxane per mole of hydroxyl residues of protein) removed from the serum 72-81% of the binding activity for Silastic tubing (fig 4). The results for the two different proteins and two different silylating reagents were pooled, since they showed no significant differences by ANOVA.
Temporary and long-term implantation of various devices made from silicones has become common medical therapy. Reactions to these foreign materials are usually restricted to mild fibrosis,( n4) but immune mechanisms were proposed for some cases of obstruction of VP shunts, when inflammatory cells were detected in the tubing lumen.( n5, n6) Other evidence from patients and from experiments in animals suggests that the silicones may not be immunologically inert( n2, n3, n7-n10) and may elicit inflammatory reactions.
The two patients described here probably represent unusual complications of VP shunt placement, but similarities in their histories suggest common mechanisms. Their VP shunts were well tolerated at first, but after surgical revision of the abdominal pseudocysts and intense subcutaneous reactions developed at the sites of silicone implants and silicone-coated sutures, which could not be attributed to infections.
All serum samples from both patients contained IgG that bound to the tubing in greater quantities than did IgG from normal adults or other VP shunt patients without inflammatory reactions.
The bound immunoglobulin seemed to be specific antibodies, since the Fab fragment of IgG also bound to the tubing and most of the binding IgG was removed by preincubation with similar tubing or siloxane-conjugated proteins. However, we cannot attribute the local inflammatory reactions to these antibodies. The granulomas observed in patient 1 are more consistent with T-cell-mediated immune lesions than with antibody-mediated reactions.( n11) We have not yet found a form of polydimethylsiloxane adequate for in-vitro testing of cellular immunity in these patients, though the oligomethylsilane-protein complexes we used may be useful. The nature of the silicone antigen that elicits the immune response is not known.
Because therapeutic use of polydimethylsiloxane is widespread, the frequency of immune responses to these materials and their relation to adverse reactions to silicone implants should be studied further. Better understanding of the mechanisms by which patients become sensitised to polydimethylsiloxane could facilitate the selection of patients for implantation procedures, and aid in the development of new synthetic polymers that reduce the risk of adverse reactions to implantation of important medical devices.
We thank Dr Michael Wasserman for help in the development and testing of the immunoassay and Dr Richard Weiner and Dr Steven Blackwell for clinical management of the patients.
(n1.) Heggers JP, Kossovsky N, Parson RW, Robson MC, Pelley RC, Talmage RJ. Biocompatibility of silicone implants. Ann Plastic Surg 1983; 11: 38-45.
(n2.) Kossovsky N, Heggers JP, Robson MC. Experimental demonstration of the immunogenicity of silicone-protein complexes. J Biomed Mater Res 1987, 21: 1125-33.
(n3.) Bowman BH, Schneider L, Barnett DR, Kurosky A, Goldblum RM. Novel urinary fragments from human basement membrane collagen. J Biol Chem 1980; 255: 9484-89.
(n4.) Frisch EE. Technology of silicones in biomedical applications. In: Rubin LR, ed. Biomaterials in reconstructive surgery. New York: CV Mosby, 1983: 73-90.
(n5.) Peimer CA, Medige J, Eckert BS, Wright JR, Howard CS. Reactive synovitis after silicone asthroplasty. J Hand Surg 1986; 11A: 624-38.
(n6.) Kircher T. Silicone lymphadenopathy. Hum Pathol 1980; 11: 240-44.
(n7.) Varga J, Schumacher HR, Jimenez SA. Systemic sclerosis after augmentation mammoplasty with silicone implants. Ann Intern Med 1989, 111: 377-83.
(n8.) Baldwin CM, Kaplan EN. Silicone-induced human adjuvant disease? Ann Plastic Surge 1983; 10: 270-73.
(n9.) Gower DJ, Lewis JC, Kelly DL. Sterile shunt malfunction: a scanning electron microscopic perspective. J Neurosurg 1984; 61: 1079-84.
(n10.) Snow RB, Kossovsky N. Hypersensitivity reaction associated with sterile ventriculoperitoneal shunt malfunction. Surge Neurol 1989; 31: 209-14.
(n11.) Warren KS. A functional classification of granulomatous inflammation. Ann NY Acad Sci 1976; 278: 7-18.
ADDRESSES: Departments of Pediatrics (R. M. Goldblum, MD, A. A. O'Donell, MD, Prof D. Pyron, MT), Human Biological Chemistry and Genetics (R. M. Goldblum), Pharmacology and Toxicology (R. P. Pelley, MD), Family Medicine (A. A. O'Donell), Surgery and Microbiology (Prof J. P. Heggers, PhD), University of Texas Medical Branch, Galveston, Texas 77550, USA. Correspondence to Dr Randall M. Goldblum.