Cocke [50] found intracellular silicone in the surrounding tissue of an intact doublelumen prosthesis 120 h after its placement. In 1979, Vargas [54] also reported silicone shedding from the envelope of saline-filled silicone breast implants. Thus, it is now generally accepted that all types of silicone-gel implant bleed silicone through the intact envelope, both in vivo [21, 32, 50] and in vitro, with resulting granulomas [7, 55], lymphadenopathy and migration of free silicone to remote areas of the body through lymphatic or hematogenic pathways.

In 1964, Miyoshi et al. [56] reported on two patients who developed connective tissue-like disorders several years after augmentation mammoplasty by injection of paraffin and noted one complete resolution of the clinical symptoms after mastectomy to remove the foreign body. They were the first to name this disorder `human adjuvant disease', because it was considered the human counterpart of adjuvant arthritis found in rats after subcutaneous injection of Freund's complete adjuvant (dispersion of dried heat-killed tubercle bacilli in mineral oil). They defined six characteristics of the condition:

(1) Autoimmune disease-like symptoms which developed after the plastic surgery using foreign substances.

(2) Paraffin, silicone or related substances with possible adjuvant effects had been previously injected in the patient.

(3) Foreign body granulomata were observed histopathologically in the injected area.

(4) The presence of autoantibodies.

(5) There was no evidence of infection or malignancy in the operated area.

(6) Improvement occurred after the removal of the foreign substances. The first reports of an autoimmune connective tissue disease occurring in patients after augmentation mammoplasty with a silicone-gel-filled prosthesis came from van Nunen et al. [57] in Australia in 1982 and the following year from Baldwin et al. [20] from the US. However, it was Endo et al. [58] in 1984 and Varga et al. [59] and Varga and Jiminez [60] in 1989 who were the first to report human adjuvant disease in patients who had received saline breast implants. Since that time, a growing number of patients with diseases such as scleroderma SLE, Sjogren's syndrome, rheumatoid arthritis or atypical connective tissue disease have been described [10, 12, 26, 29, 37, 55-58,61-67]. Among those that have been described as atypical, there was one report of a life-threatening systemic illness that developed 24 h after implant removal [68] as well as an adult respiratory distress syndrome following augmentation by silicone injections [69].

It has been recognized that every normal individual makes autoantibodies, but only certain individuals produce pathogenic autoantibodies that may lead eventually to autoimmunity. Central to this process is the activation of self-reactive T-helper/inducer cells and it has been established that T-cells recognize a complex consisting of a major histocompatibility complex (MHC) restriction element and a peptide antigen fragment. After the processing of the antigen by the endosomes of the cell, some of these fragments become associated with class II molecules. Therefore, the ability of a specific antigen to bind with a MHC dictates an association between the immune response to a specific antigen [70, 71]. After the specific antigen is presented to the immune system, it binds with receptors on the surface of the B-lymphocyte. The antigen may also bind to an immunoglobulin-like globulin on the surface of the inducer T-lymphocyte and on the surface of a suppresser T-cell. The inducer T-lymphocyte may thus permit the B-lymphocyte to mature into a plasma cell and, thus, secrete antibody to the foreign antigen. A second T-lymphocyte carrying a specific receptor may then bind with the newly formed antibody [63]. This process will then permit the B-lymphocyte to become a plasma cell that will secrete an antigen and then bind the original B-lymphocyte. Therefore, B-lymphocyte activation is dependent on T-lymphocytes for subsequent autoantibody production [72, 73]. Peritoneal inflammatory granulonatosis, foamy conglomeration and, finally, plasmacytomagenesis in genetically susceptible mice (BALC/C.DBA/2-IHD1-PEP3) has been shown following intraperitoneal injection of silicone. In addition, experimental allergic encephalopathy (EAE) and experimental allergic neuropathy (EAN) have been extensively studied. Moreover, silicone gel and octamethylcyclotetrasiloxane (D4) have been shown to potentiate antibody production to bovine serum in A/J mice [74].

During the 1992-1995 American College of Rheumatology meetings, numerous studies were presented that reported patients who had developed atypical rheumatic disease after silicone breast implant surgery [2, 46, 62, 73, 75-78]. Interestingly, Bridges et al. [76] reported 156 women with silicone breast implants who had developed atypical rheumatic disease. In their study, only 9% had tested positive for the rheumatoid factor and only 22% had tested positive for antinuclear antibodies. They concluded that women with silicone breast implants might develop atypical rheumatic diseases, which differ from the classical idiopathic disease. Furthermore, Love et al. [79] reached the same conclusion after investigating 13 patients who had developed myositis after receiving breast implants; they found that their clinical and immunogenic features differed from the classical idiopathic myositis (polymyositis-dermatomyositis). Moreover, Freundlich et al. [80] reported 24 patients with breast implants who had developed atypical Sjogren's-like syndrome with dry eyes and dry mouth, adenopathy and glandular mononuclear cell infiltrates in the absence of serological findings. In addition, Morse and Spera [73] reported a significant increase in B- and DR-cells, a decrease in CD2- and CD8-cells and an increase in the CD4:CD8 (T4:TB) ratio in 30 patients with breast implants when compared to controls. Some investigators, however, have reported novel antibodies in patients with breast implants; among them Tenenbaum et al. [81] identified an antibody that bound to large molecular weight proteins appearing as a bobble in over 70% of their patients with breast implants and that serum antibodies from healthy individuals failed to react to this protein, suggesting an atypical immune response in symptomatic breast implant recipients. Ostermeyer and Patten [51, 52, 82, 83] were the first to report adjuvant breast neurological disease, multiple sclerosis (MS)-like syndrome and motor neuron disease-like syndrome in silicone breast implant recipients. Some investigators have reported antimicrosomal thyroid antibodies in 30% of patients with silicone breast implants [72].

Vodjani et al. [34] reported abnormalities of the T-helper: T-suppresser ratio mechanism, increased autoimmunity and increased immune complexes in patients with breast implants when compared with healthy sex- and age-matched controls. Levine and Ilowite [84] reported 11 children with esophageal dysfunction who were breast-fed by mothers with breast implants. The same group also found increased nitrite and nitrate urinary excretion in those children breast-fed by mothers with implants which they thought was due to activated macrophages exposed to silicone [65].

The remission of some of the symptoms of silicone breast implant-associated disease after implant removal has also been reported. In fact, Brozema et al. [10] described a patient who presented with progressive scleroderma-like illness after silicone breast augmentation with dramatic recovery upon implant removal. Walsh et al. [68] reported on a patient with chylous effusions, peripheral edema and high antinuclear antibody titer whose symptoms resolved after implant removal. Gutierrez and Espinosa [85] also reported the reversal of progressive systemic sclerosis with severe hypertension in a woman after implant removal. Moreover, Kaiser et al. [16] reported on the remission of silicone-induced autoimmune disease after explantation, while Silver et al. [77] identified silicon in tissues involved by chronic inflammation and fibrosis such as implant capsules, synovium, skin and alveolar macrophages in three patients with connective tissue disease; all improved after implant removal.

There is convincing evidence that polyneuropathy (demyelinating) is associated with over 85% of our clinic patients complaining of symptoms after silicone gel implantation. This may be associated with a variety of autoimmune chemical phenomena. Less convincing, but an association of demyelination of the central nervous system and anti-thyroid antibodies may be found in approximately 30-35% of these patients with symptoms following implantation. Furthermore, local immune responses may be found in the capsule surrounding the silicone-gel implants and this includes activation of macrophages, B- and Tlymphocytes and selected T-cell receptor utilization and interleukin-2 antibodies. Moreover, lumen leukocyte antigen (HLA) typing has demonstrated that there is a significant HLA-DR53 positivity in those symptomatic patients with fibromyalgia associated with silicone-gel implants [86-89].

EAN and EAE have been extensively studied and each of the adjuvant-dependent antigens has been identified. In the EAN disease, which is produced in susceptible animals, Po glycoprotein, P2 lipid binding protein and myelin-associated globulin (MAG) have been etiologically associated with the development and progression of the disease which parallels that found in silicone breast implant adjuvant syndrome associated with peripheral neuropathy. The protein P1, also named myelin basic protein (MBP), is found only in the central nervous system and is responsible for the animal model of EAE. In the model of EAN, commonly produced in Lewis rats, a MHC II response regularly occurs which is mediated by T-cells and occurs in the following stages.

(1) Alteration of the blood-nerve barrier with the infiltration of T-cells within 72 h after the challenge.

(2) Migration of the inflammatory T-cells with the presence of edema, which it is associated with a decrease of nerve conduction. This occurs within 4-5 days following the induction of EAN.

(3) CD4 (T)-cells predominate with the production of cytokines, which in turn increase the cell adhesion molecules by endothelial cells.

(4) Finally, there is an accumulation of macrophages, T-cells and polymorphonuclear leukocytes which, when activated, release free oxygen, hydroxyl radicals, proteases and lipases. The damage appears to be oxidative damage, while the protein and lipid enzymes are produced in order to digest the damaged cell debris. These changes have been observed in patients with sural nerve biopsies.

In EAN the peripheral nerve myelin is a complex structure that is synthesized and maintained by Schwann cells. The chemical composition of peripheral nerve myelin is largely lipid with a small percentage of proteins. The major protein is Po glycoprotein (50%) and this protein is not detected in the central nervous system. Sequencing of the amino acids in mammals shows 219 amino acids organized into three structural domains: an extracellular domain containing a single glycosylation site, a hydrophobic transmembrane domain and a basic cytoplasmic domain. It is this protein that is thought to play a major role in stabilizing the compaction of the extracellular apposition of the myelin membrane in the peripheral nervous system. EAN is a cell-mediated process, induced passively in experimental animals by lymphocytes but not by serum, although there is recent evidence that serum may induce demyelination in peripheral nerves. The Po glycoprotein readily produces EAN; however, MBP and P2 protein may induce EAN as well. Po protein and P2 proteins of peripheral nerves may initiate a neurotogenic T-cell response in experimental animals and produce similar demyelinating neuropathy. There is a naturally occurring syndrome of demyelinating neuropathy in humans and this is known as Guillian-Barre-Strohl-Landry syndrome. There is considerable evidence that this syndrome, which follows prodromal infections, is associated with increased levels of complement-fixing, anti-peripheral, nerve myelin glycoprotein antibodies. Certain patients with demyelinating neuropathy develop an immunoglobulin (Ig) M monoclonal antibody response to peripheral nerve myelin antigens. In approximately 60-70% of these patients, these antibodies (M proteins) attach to a carbohydrate determinant shared by MAG, Po and three glycoprotein components of peripheral nerve myelin. The crucial event in the pathogenesis of demyelinating neuropathy is the peripheral activation and expansion of a neuritogenic T-cell response (possibly Po) which then induces blood-nerve barrier dysfunction. The antibody may then cross into the peripheral nervous system and act synergistically with the T-cell response to enhance clinical disease. The balance between the intensity of the initial inflammatory T-cell response and the antibody concentration may then determine the clinical course of the disease.

EAE is produced in experimental animals by inoculation of the brain, spinal cord extracts and MBP along with Freund's adjuvant. MBP represents approximately 30% of the protein in the central nervous system myelin and has a molecular weight of approximately 18 500 and is composed of 169 amino acid residues. Sequencing of the amino acids in MBP has been determined in many species. Moreover, MS appears to be the human counterpart to EAE and has been closely associated with HLA (B7 and Dw2). Demyelinating plaques develop around blood vessels and neuropathologically parallel the course of the human disease. The myelin proteins of peripheral nerve Po and P2 will not induce the demyelinating lesions in the central nervous system of experimental animals.

Immune responses alter neural and immune functions and, in turn, neural and endocrine functions alter immune function. Many regulatory peptides and their receptors are known to be expressed by both the brain and the immune system. The central nervous system itself can be involved in immune reactions, whether arising from within the brain or in response to peripheral immune stimuli. Activated immunocompetent lymphocytes and macrophages can penetrate the blood-brain barrier and take up residence in the brain, where they secrete their full repertoire of cytokines and other inflammatory mediators, such as leukotrienes and prostaglandins. Microglia, which are embryologically and functionally related to peripheral macrophages and astrocytes, are, like macrophages and monocytes, activated by toxins, antigens and products of cell injury arising within the brain or reaching the brain from the periphery. These cells, in turn, secrete cytokines and inflammatory mediators. Furthermore, the endothelial and smooth muscle cells of cerebral blood vessels secrete cytokines such as interleukin-1 and interleukin-6 in response to circulating antigens and toxins. Moreover, the activation of cytokines in the central nervous system can lead to profound changes in neural function, ranging from mild behavioral disturbances to anorexia, drowsiness, increased slow-wave sleep, dementia coma and the destruction of neurons. None of these changes may be detectable by routine medical technologies, including magnetic resonance imaging (MRI) of the brain.