Silicone-Specific Blood Lymphocyte Response in Women ...eases, systematic immunologic studies ofwomenwith silicone gel implants are needed to assess any possible role of cell-mediated

CLINICAL AND DIAGNOSTIC LABORATORY IMMUNOLOGY, Nov. 1994, p. 689-695 Vol. 1, No. 61071-412X/94/$04.00+0Copyright © 1994, American Society for Microbiology

Silicone-Specific Blood Lymphocyte Response in Womenwith Silicone Breast Implants

EMMANUEL A. OJO-AMAIZE,1* VICTOR CONTE,2 HUN-CHI LIN,1 ROBERT F. BRUCKER, 2MELKON S. AGOPIAN,1 AND JAMES B. PETER'

Specialty Laboratories, Inc., Santa Monica, California 90404-3900,' andBalco Laboratories, Inc., Burlingame, Califomia 940102

Received 20 May 1994/Returned for modification 29 June 1994/Accepted 20 July 1994

A blinded cross-sectional study was carried out with 99 women, 44 of whom had silicone breast implants.Group I consisted of 55 healthy volunteer women without breast implants; group II comprised 13 volunteerwomen with breast implants or explants who felt healthy; group III comprised 21 volunteer women with breastimplants who had chronic fatigue, musculoskeletal symptoms, and skin disorders; and group IV comprised 10women who had their prostheses explanted but still presented with clinical symptoms similar to those of thewomen in group III. Proliferative responses of peripheral blood mononuclear cells from all 99 women weremeasured by [3H]thymidine uptake after exposure to SiO2, silicon, or silicone gel. The levels of proliferativeresponses were expressed as stimulation indices, which were obtained by dividing the counts per minute ofstimulated cells by the counts per minute of unstimulated cells. Abnormal responses to SiO2, silicon, or siliconegel were defined as a stimulation index of >2.8, >2.1, or >2.4, respectively. Abnormal responses were observedin 0%1 of group I, 15% of group 1, 29%o of group Ill, and 30%o of group IV (P < 0.0005 for group I versus groupsII and IV). Thirty-one percent of symptomatic women with silicone gel breast implants had elevated serumsilicon levels (>0.18 mg/liter); however, there was no significant correlation between abnormal cellularresponses and silicon levels in blood serum, type of implant, time since first implantation, prosthesisexplantation, number of implants, or report of implant leakage or rupture. Flow cytometric and cell depletionanalyses showed that the responding cells were CD4+ T cells, with no apparent contribution from the CD8+T-cell population. Our demonstration that silicon-specific T-cell responses are observed in twice as manysymptomatic as asymptomatic women exposed to silicone breast implants suggests that cell-mediatedimmunity plays a role in the development of abnormal immune reactions associated with silicone and providesa new, apparently specific screening blood test. Whether the activity observed in asymptomatic women ispredictive of symptom development is under prospective study.

The term silicone refers to a group of silicon-containingcompounds, which include fluids, gels, rubbers, sponges,foams, and resins (1). Although silicone was originally re-garded as being inert in the human body, its polymeric andhydrophobic characteristics and the presence of electrostaticcharges and organic side groups render silicone a potentiallyideal immunogen (32). Several reports suggest that siliconeproducts are associated with various complications that mayinvolve an immune reaction to silicone (4, 11, 16, 31). Approx-imately 1 to 2 million women in the United States have hadsilicone breast implants inserted for reconstruction or augmen-tation mammoplasty; some of these women are reported tohave developed a systemic autoimmune disease (19, 26, 29).

Presently, it is uncertain which complications have a cause-and-effect relationship and which represent coincidental find-ings (26). There is further confusion in distinguishing betweennonspecific local reactions and reactions that have an immu-nological basis (8). In view of the fact that silica mining isthought to predispose individuals to certain autoimmune dis-eases, systematic immunologic studies of women with siliconegel implants are needed to assess any possible role of cell-mediated immunity in the clinical complications of siliconebreast implants. Our studies reported herein were directed tothe search for silicon(e)-specific T cells in the circulation of

* Corresponding author. Mailing address: Specialty Laboratories,Inc., 2211 Michigan Ave., Santa Monica, CA 90404-3900. Phone: (310)828-6543, extension 310. Fax: (310) 828-6634.

women with silicone breast implants and to the possiblecorrelation of such reactivity with various complications asso-ciated with silicone medical devices.

MATERIALS AND METHODS

Reagents. Lithium sulfate (Li2SO4), nickel sulfate hexahy-drate (NiSO4 -6H20), zirconyl chloride hydrate (ZrOCl2),mercuric chloride (HgCl2), chromic trioxide (CrO3), magne-sium sulfate (MgSO4), and silicon dioxide (Si02) were pur-chased from Sigma Chemical Company (St. Louis, Mo.). Otherreagents and sources included beryllium sulfate tetrahydrate(BeSO4 *4H20) (Aldrich Chemical Co. Inc., Milwaukee,Wis.), elemental silicon (Si) and silicone gel (Spex Industry,Edison, N.J., and Mentor Corporation, Santa Barbara, Calif.),RPMI 1640 medium and Hanks' balanced salt solution (HBSS)(Irvine Scientific, Santa Ana, Calif.), Ficoll-Paque in vitrolymphocyte isolation medium (Pharmacia, Piscataway, N.J.),penicillin-streptomycin mixture (GIBCO Laboratories, GrandIsland, N.Y.), and pooled human AB serum (Gemini Bioprod-ucts, Inc., Calabasas, Calif.). Fluorescence-activated cell sorterlysing buffer, phycoerythrin-labeled monoclonal antibody(MAb) Leu2a (CD8), and fluorescein isothiocyanate-labeledMAb Leu3a (CD4) were purchased from Becton Dickinson(San Jose, Calif.). An AIS MicroCellector for the selection ofT-cell subsets was obtained from Applied Immune Sciences,Inc. (Menlo Park, Calif.).

Blood donors. In a blinded cross-sectional study, peripheralblood was obtained by venipuncture from 55 healthy women

689

on August 23, 2020 by guest

http://cvi.asm.org/

Dow

nloaded from

690 OJO-AMAIZE ET AL.

TABLE 1. Characterization of 44 women with silicone gel breast implants

Group Clinical characteristics No. of No. with the following type of implant:subjects Silicone gel Lumena Memeb Unknown

I No implants, healthy 55II Implants or explants, well, asymptomatic 13 8 3 1 1III Implants, symptomatic (chronic fatigue, musculoskeletal 21 13 3 1 4

symptoms, autoimmune disease)IV Explants, symptomatic (chronic fatigue, musculoskeletal 10 7 3 0 0

symptoms, autoimmune disease)a Silicone inside and saline outside.b Polyurethane foam coated.

who did not have breast implants (group I) and from 44 womenwith silicone gel breast implants (groups II to IV). Informedconsent was obtained in accordance with institutional guide-lines. Of the 13 asymptomatic women, 4 had explanted theirbreast prostheses and the other 9 still had their implants inplace and felt well (group II). Group III consisted of 21 womenwith breast implants who had various symptoms, includingfatigue, fibromyalgia, insomnia, skin disorders, respiratorycomplaints, headache, joint pain, muscle cramps, arthritis,short-term memory problems, allergies, arrhythmia, and ane-mia. Group IV comprised 10 women who had their breastimplants removed because of overt connective tissue diseases.All of these women (groups II to IV) filled out questionnairesabout their medical history; they were referred by plasticsurgeons and primary-care physicians and through individualbreast implant support groups. Each woman provided detailedinformation with regard to date of placement of all implants,indications for implantation, type(s) of implant, reported overtleakage or rupture, and duration, types, and severity of symp-toms.Measurement of silicon levels in blood serum. Serum sam-

ples were analyzed by using an inductively coupled plasmaatomic emission spectrometer (Applied Research Laborato-ries, Dearborn, Mich.). The samples were introduced to theargon plasma by a peristaltic pump (Gilson Medical Electron-ics, Inc., Middleton, Wis.) and an auto sampler (Gilson).Details of the methodology and the instructions given tosubjects were essentially as previously described (9, 10, 20).

Preparation of lymphocytes. Peripheral blood mononuclearcells (PBMC) isolated from heparinized venous blood byFicoll-Hypaque gradient centrifugation (3) were washed threetimes with Hanks' balanced salt solution before determinationof viability by the trypan blue dye exclusion method. Cellsresuspended in complete culture medium (20% heat-inacti-vated pooled human AB serum, 2 mM L-glutamine, 2 mMN-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid [HEPES],and 1% penicillin-streptomycin [10,000 U/ml]) were adjustedto 2 x 106/ml.

Selection of T-lymphocyte subsets. The AIS MicroCellectorCell Culture Flask System was used to negatively select eitherCD4+ or CD8+ cells according to manufacturer's instructions.In the MicroCellector Cell Culture Flask System, MAbs are

permanently bound to the surface. When blood cells areadded, the immobilized ligands bind to surface antigens of thetargeted cells; cells not bearing the recognized antigens remainfree in suspension. A quantity of 4 x 107 PBMC was added toeither the AIS CD4 T-25 flask (MAb Leu3a bound; fordepletion of CD4+ cells and selection for CD8+ cells) or theCD8 T-25 flask (MAb Leu2a bound; for depletion of CD8+cells and selection for CD4+ cells). After incubation for 1 h atroom temperature (RT) on a flat nonvibrating surface, nonad-

herent cells were removed, washed twice in complete medium,counted, and analyzed by flow cytometry.Flow cytometric analysis ofT-lymphocyte subsets. Following

lysis of any remaining erythrocytes, P13MC were stained with amixture of phycoerythrin-labeled MAb CD8 and fluoresceinisothiocyanate-labeled MAb CD4. Stained cells were analyzedon a FACScan cytometer (Becton Dickinson). Lym.phocyteswere gated by forward and right-angle light scattering.

Antigens. Si02 was prepared in concentrated phosphoricacid (H3P04). A 2 mM stock solution was prepared and storedat RT until used for assay at final concentrations of 10, 1, and0.1 ,uM in complete culture medium.A stock solution of silicon dissolved in H20 at a concentra-

tion of 1 mg/ml was stored at RT until used for assay at finalconcentrations of 10, 1, and 0.1 ,ug/ml in complete culturemedium.A stock solution of 1 g of silicone gel dissolved in hexane to

yield 100 mg/ml was stored at RT until used for assay at finalconcentrations of 10, 1, and 0.1 jig/ml in complete culturemedium.

Lymphoproliferation. Lymphocytes (2 x 105/0.1 ml of com-plete culture medium) were dispensed in quadruplicate into96-well round-bottomed microtiter plates, and antigens (0.1 mlper well) were added in complete culture medium. Nonstimu-lated control wells contained 0.1 ml of cells and 0.1 ml ofcomplete culture medium. Cultures were maintained in ahumid incubator at 37°C in an atmosphere of 5% CO2 forperiods shown in preliminary experiments to be optimal (5and 7 days). At 4 h before harvest, the cultures in eachwell were pulsed with 1 ,uCi of tritiated thymidine (specificactivity, 719.5 mCi/mg; Dupont, Wilmington, Del.). Cells har-vested onto glass fiber filters (Packard, Downers Grove, Ill.)with a 96-well automatic cell harvester (TOMTEC, Hamden,Conn.) were counted directly on a Matrix 9600 direct betacounter (Packard). Data were expressed as the stimulationindex (SI) (counts per minute for stimulated wells/counts per

TABLE 2. Establishment of normal reference range fordetermination of degree of stimulation with three forms

of silicon in healthy women without breast implants

Type of No. of Mean SI + 3 SD at: Establishedsilicon subjects per SI cutoff

group 0.1 pLM 1.0 ,uM 10 ,uM value

Si02 40 2.4 2.9 3.1 2.8Silicona 15 2.9 2.0 1.4 2.1Silicone gela 15 3.0 1.8 2.3 2.4

a Elemental silicon and silicone gel were used in microgram concentrations(0.1, 1.0, and 10 ,ug/ml) on cells of the same 15 healthy women.

CLIN. DIAGN. LAB. IMMUNOL.

on August 23, 2020 by guest

http://cvi.asm.org/

Dow

nloaded from

T-CELL RESPONSE IN WOMEN WITH SILICONE BREAST IMPLANTS 691

TABLE 3. Silicon-specific T-cell proliferative response in womenwith silicone gel breast implants

No. of No. (%) with abnormalGroupa subjects T-cell response tosilicon(e)b

I 55 0(0)II 13 2 (15.3)CIII 21 6 (28.6)IV 10 3 (30.0)

a The groups are characterized in Table 1.b Individuals with SI values greater than the cutoff values shown in Table 2 at

any concentration, at any of the two optimal time points (5- or 7-day culture),and to any of the three forms of silicon.

c One of the two women with an abnormal cellular response to elementalsilicon had her breast implant in place, and the other had had hers explanted.Both were healthy at the time of study.

minute for unstimulated control cultures) ± standard error ofthe mean.

Determination of dose-response curve for response to stim-ulation with SiO2, silicon, or silicone gel. Viable cells, adjustedto 2 x 106/ml, were dispensed in 0.1-ml volumes (2 x 105 cellsper well) into microtiter wells and challenged with differentconcentrations of SiO2, silicon, or silicone gel in O.1-ml vol-umes. The cultures were incubated for 5 or 7 days. At 4 hbefore the end of the culture period, the cultures were pulsedwith 1 ,uCi of [3H]thymidine. Blastogenic responses weredetermined as described above.

Determination of kinetics of secondary response. Viablecells, adjusted to 2 x 106/ml, were dispensed in O.1-ml volumes(2 x 105 cells per well) into microtiter wells and challengedwith different concentrations of SiO2, silicon, or silicone gel in0.1-ml volumes. The cultures were incubated for variousperiods of time (1, 3, 5, 7, and 9 days). At 4 h before the endof each culture period, the cultures were pulsed with 1 ,uCi of[3H]thymidine. Proliferative responses, expressed as SIs, weredetermined as described above.

Evaluation of antigen specificity of silicon-reactive cells.The antigen specificity of silicon-reactive cells was evaluated bytesting a battery of related metal salts and comparing theresults with those for silicon, SiO2 and silicone gel. The salts,which were tested at three different final concentrations (0.1, 1,and 10 ,uM), included BeSO4, CrO3, Li2SO4, NiSO4, ZrOCl2,HgCl2, and MgSO4. SIs after 5 or 7 days of culture weremeasured as described above.

Establishment of SI cutoff values for determination ofabnormal response to Si02, silicon, or silicone gel. An abnor-mal or positive response is defined as a peak SI of >2.8 forSiO2, >2.1 for silicon, or >2.4 for silicone gel. These values arebased on the blood mean peak SI plus 3 standard deviations for40 healthy women (SiO2), 15 healthy women (silicon), and 15healthy women (silicone gel).

Statistical analysis. All values are means ± standard errorsof the means. Statistical comparisons were made with Stu-dent's t test.

RESULTS

Characterization of women with silicone breast implants.The 99 women who voluntarily participated in the blindedstudy were divided into four groups at the completion of thestudy. Group I consisted of 55 healthy women without breastimplants, group II comprised 13 healthy women, group IIIconsisted of 21 women with breast implants who had chronicfatigue and musculoskeletal symptoms, and group IV com-

8-

c-o

C

C-0cc

E._"co

6-

4-

I

n=55

IIn=13

mn=21

IVn=10

Group NumberFIG. 1. Silicone-specific T-cell response profile of women with

silicone gel breast implants. The highest response (SI) obtained foreach woman in each group at any of the three concentrations testedagainst any of the three forms of silicon is shown. Group I representshealthy controls, group II represents asymptomatic women with breastimplants or explants, group III represents symptomatic women withbreast implants, and group IV consists of symptomatic women whohave had their silicone breast prostheses explanted. All the women(groups I to IV) had normal responses to the T-cell mitogen phytohe-magglutinin (SI > 50). The woman with the highest SI value in groupIII is the same woman whose results were used to generate the resultsshown in Fig. 4. This woman is subject no. 4 in Table 4.

prised 10 women who had their prostheses explanted but stillpresented with clinical symptoms similar to those of thewomen in group III (Table 1).

Establishment of normal reference range for assessment ofcellular response to stimulation with three forms of silicon. Ofthe 55 healthy women without implants used to establish thenormal reference range, 40 were used for SiO2 and the other15 were used for silicon and for silicone gel. The mean SIs +3 standard deviations for each of three concentrations of eachantigen are shown in Table 2. Because the results did not differsignificantly, the average of the standard deviations at all threeconcentrations for each antigen was used as the cutoff value forthat antigen (Table 2).

Silicon-specific and T-cell mitogen proliferation response inwomen with silicone gel breast implants. Abnormal silicon-specific T-cell proliferative responses were observed in 0% ofgroup I, 15% of group II, 28.6% of group III, and 30% of

0

0

0~~~

0~~~

0~~~~~~~~~

* 0

.~~S

so*

*4 Jo 00. .a *|.r"t .L .-~8

VOL. 1, 1994

on August 23, 2020 by guest

http://cvi.asm.org/

Dow

nloaded from

692 OJO-AMAIZE ET AL.

4.

wCO)+1

02E,

3.

2

1

0,0 .1 1 10 100 1000

SbC Concentraon (M)FIG. 2. Dose-dependent proliferative response of PBMC from women with silicone breast implants to stimulation with SiO2. PBMC from three

women with silicone breast implants were challenged with various concentrations of SiO2. SEM, standard error of the mean. Symbols: A, a womanwith an abnormal cellular response to SiO2 (SI > 2.8); O, a woman with a normal (intermediate) response to Si02; 0, a woman with a normal(low) response to Si02.

group IV (Table 3). The responses were significantly higher ingroups II, III, and IV than in group I (group I versus group II,P < 0.05; group I versus group III, P < 0.0005; and group Iversus group IV, P < 0.0000001). The average SI of eachindividual woman in each group is shown in Fig. 1. All thepatients responded normally (SI > 50) to the T-cell mitogenphytohemagglutinin.

Dose-dependent proliferative response of PBMC fromwomen with silicone breast implants to stimulation with Si02.PBMC from three women were used to establish the dose-response curve. In a prior experiment, one of these women hadan abnormal response to 1 ,uM SiO2, and the other two hadresponses below the cutoff value for SiO2. SiO2 concentrationsof 0.1 and 1.0 ,uM induced higher levels of responses thanconcentrations of >50 ,uM (Fig. 2). Similarly, concentrationsof 0.1, 1.0, and 10 ,ug/ml for silicon and silicone gel inducedoptimal levels of responses (data not shown). The data indicatethat responses to one or more forms of silicon at variousconcentrations can occur.

Kinetics of proliferative response of PBMC from a womanwith silicone breast implants to stimulation with SiO2, silicon,or silicone gel. PBMC from a woman who had previously beendetermined to have abnormal SIs for stimulation with Si02 andsilicone gel were used for the kinetics experiment. Followingstimulation with 0.1, 1.0, or 10 ,IM Si02 or with 0.1, 1.0, or 10j,g of silicon or silicone gel per ml, cells were cultured forvarious periods of time (1 to 9 days). Both forms of siliconinduced maximal responses between days 5 and 7 (Fig. 3), withthe greatest response being on day 7. These results are similarto those for the beryllium-induced T-cell response, whichshowed that either of two optimal time points is appropriatefor determining the response to beryllium (15, 18, 23). Anindividual with SI values greater than the cutoff values for eachof the antigens, at any of the three concentrations and at anyof the two optimal time points (5- or 7-day culture), is regardedas having an abnormal T-cell proliferative response to silicon.

Specificity of the silicon-reactive PBMC. PBMC from asymptomatic woman with silicone breast implants who waspreviously shown to have an abnormal response to stimulationwith Si02 were used to demonstrate that a metal-specific

immune response to Si02 could be documented. All of theantigen metals tested (except Si02 and silicone gel) failed toinduce significant responses in the silicone-sensitized woman(Fig. 4). Although the particular woman used for this experi-ment did not respond to elemental silicon, some of the womenin groups II, III, and IV responded to elemental silicon (Table4).

Effect of T-cell subset depletion. After depletion with theAIS MicroCellector Cell Culture Flask, two-color immunofluo-rescence and flow cytometry demonstrated that the silicon-reactive cells possess the CD4+ phenotype. No activity wasfound in the CD8+ population of T cells (Table 5).Serum silicon levels. Four of 13 (31%) symptomatic women

with silicone gel breast implants, 1 of 7 (14%) symptomaticwomen who had had their prostheses explanted, 0 of 4 (0%)asymptomatic women with implants, and 0 of 8 (0%) controlshad elevated levels of silicon in serum (>0.18 mg/liter).

DISCUSSION

In spite of the widespread use of silicone and relatedmaterials in humans, there have been very few systematicstudies on the immunological effects of silicone. Althoughsilicone is known to induce inflammatory responses (8, 16) andlymphadenopathy and giant-cell granulomas (7, 21), a specificcellular response to silicone has never been demonstrated bystandard immunological assays. On the other hand, specificT-cell immune responses to light metals were observed inseveral studies (13, 22, 23, 30). The present report shows, forthe first time, that silicon(e) can act as a specific sensitizingantigen in vivo, leading to a silicon(e)-specific immune re-sponse in vitro as measured by [3H]thymidine uptake by T cellsresponding to stimulation with either Si02, silicon, or siliconegel.Women with silicone breast implants are known to have

several types of autoantibodies against different self antigens(5, 25-27, 29, 32). Our data showing the involvement of CD4+cells in silicone-induced immune reactions suggest that at leastone of the mechanisms by which certain individuals withsilicone prostheses produce autoantibodies could be via the

CLIN. DIAGN. LAB. IMMUNOL.

on August 23, 2020 by guest

http://cvi.asm.org/

Dow

nloaded from

T-CELL RESPONSE IN WOMEN WITH SILICONE BREAST IMPLANTS

w

-H

cCo0C

Eco

9.

8.

7.

6.

5.

4

3

2

I

0 2 4 6 8Duration in Culture (days)

10

FIG. 3. Kinetics of secondary proliferative response of PBMC from a woman with a silicone breast implant to stimulation with SiO2, silicon,or silicone gel. PBMC were from a woman with a silicone breast implant who had an abnormal cellular response to Si02 (O), an abnormal responseto silicone gel (A), and a normal response to silicon (0). SEM, standard error of the mean. Results shown are peak SIs at any one of the followingSiO2 concentrations: 0.1, 1.0, or 10 ,uM.

amplification of T-cell help for autoreactive B cells. Thefinding that more symptomatic silicone implant-exposedwomen developed abnormal T-cell responses to silicone (com-pared with asymptomatic implant women) (Table 3) is consis-tent with the notion that autoimmune reactions prevalentamong this group of women may be associated with silicon

materials. The demonstration of 15% positivity in the asymp-tomatic group (group II) suggests that all asymptomaticwomen who are positive must be monitored over time and thatall women with silicone breast implants and individuals withother silicone prostheses should be tested for a hypersensitivityreaction to silicon(e). In support of this notion is the observa-

36

32

2 28

+1 24C')X 200

- 16C0

X 12EE 8n

4 m xxx .S

BeS: 4 CrO3 U2SO4 MgSO4 NiSO4 HgC12 ZrOC12 Silicon Sil gel SO2

FIG. 4. Specificity of the response to silicone. PBMC were from a symptomatic woman with a silicone breast implant who had an abnormalresponse to SiO2 and silicone gel but a normal response to silicon and the rest of the metal salts. This was the same woman with the highest SIvalue in group III, as shown in Fig. 1. SEM, standard error of the mean. Results are shown as peak SIs at any one of the three concentrations (0.1,1.0, and 10 p.M) and any one of the two time points (5- or 7-day culture).

VOL. 1, 1994 693

on August 23, 2020 by guest

http://cvi.asm.org/

Dow

nloaded from

694 OJO-AMAIZE ET AL.

TABLE 4. Responsiveness of 11 women with abnormal responseto different forms of silicon

SIb with:

Groupa Subject SiO2 at (>M) Silicon at Silicone gel atno. (Rg/ml): (Rg/ml):

0.1 1.0 10.0 0.1 1.0 10.0 0.1 1.0 10.0

II 1 2.72 3.6

III 3 4.3 3.1 2.94 33 9.6 7.7 4.85 2.26 2.9 4.57 2.88 7.4 4.3

IV 9 3.110 3.9 2.411 3.8

a The groups are characterized in Table 1.b The highest SI value obtained for each subject with either 5- or 7-day culture.

tion that only about 5% of all beryllium-sensitized individualsdevelop chronic beryllium disease, which is found only inberyllium-sensitized individuals (12, 15, 18, 23). In other words,beryllium sensitization is a necessary but not sufficient factorfor development of beryllium disease. This may also be true forsilicon. The fact that some implant-exposed individuals re-sponded to all three forms of silicon whereas others respondedonly to one, two, or none of the silicon forms may reflectdifferences in either the level of in vivo priming, type ofimplant, genetic susceptibility, active immunosuppression, ortolerance. The few implant patients with symptoms whoshowed abnormal T-cell responses to silicon materials were notclinically different from the symptomatic patients who showednormal responses. Whether this is related to the small numberof patients or to the genetic background of the individuals willbe addressed in future studies involving a larger number ofpatients.

Specific antibodies to silicone were observed in the sera ofonly a small proportion (1.7%) of 249 women with siliconebreast implants (14). The observation that cell-based reactivitywith silicone is more prevalent than antibody-based reactivityis consistent with the observation that for a related light metal,beryllium (2, 12, 15, 17, 18, 23, 24, 28), cell-mediated immunityas measured by T-cell proliferation is more common thanberyllium-specific antibodies (6). In addition, the demonstra-tion of antibody reactivity with metals is quite tedious becauseof the complexities of binding of light-metal antigens to

TABLE 5. Characterization of silicon-specific lymphocytesaccording to phenotype

% of cellsT-cell expressing T-cell Level of silicon-specific

population subset phenotypea: T-cell proliferativeresponse (SI)b ± SD

CD4 CD8

CD4 depleted 8.8 91.2 0.9 ± 0.2CD8 depleted 96.3 3.7 4.0 ± 0.6

a Percentages of cells expressing the designated antigens following depletionwith the AIS MicroCellector Cell Culture Flask System and subjection to flowcytometric analysis.

b Each cell population was stimulated with 1.0 ,ug of silicone gel for 7 days orleft unstimulated. Cells were obtained from one of the six women in group IIIwith an abnormal T-cell proliferative response to silicone gel.

polystyrene plates. In contrast to the small number (1.7%) ofsymptomatic women with silicone breast implants identified tobe silicone antibody positive by the silicone antibody test (14),we have identified 25% (11 of 44) of such women to be siliconehyperreactive by the T-cell proliferation assay (Table 3). Thisproportion must, of course, be considered an approximationuntil similar studies with larger numbers of subjects areconducted.Our data demonstrating that CD4+ T cells are the target

cells for silicon(e) are consistent with previous reports on theinvolvement of CD4+ T cells in the immune response againsta related light metal, beryllium (18, 24).

In conclusion, the silicon(e)-specific T-cell proliferation test,compared with a silicon-specific antibody test, is less cumber-some to perform, is more specific and sensitive, and permitsthe gathering of information on an individual's cellular abnor-mal reaction to either elemental silicon, SiO2, or silicone gel.

ACKNOWLEDGMENTS

We thank Raksha Inamdar, Jesusa Arevalo, and James Valente fortheir technical contribution, Rose G. Yesowitch for preparing themanuscript, Sharon Hunt Gerardo for her critical review, and BrianGoldman for preparing and reviewing the systemic health question-naires.

REFERENCES1. Ballantyne, D. L., T. D. Rees, and I. Seidman. 1965. Silicone fluid:

response to massive subcutaneous injections of dimethylpolysilox-ane fluid in animals. Plast. Reconstr. Surg. 36:330-338.

2. Bargon, J., H. Kronenberger, L. Bergman, R. Buhl, J. Meirsydow,and P. Mitrou. 1986. Lymphocyte transformation test in a group offoundry workers exposed to beryllium and non-exposed controls.Eur. J. Respir. Dis. 69(Suppl. 136):211-215.

3. Boyum, A. 1968. Separation of leukocytes from blood and bonemarrow. Scand. J. Clin. Lab. Invest. 21(Suppl. 97):77-89.

4. Breedveld, F. C., and D. E. Trentham. 1987. Progress in theunderstanding of inducible models of chronic arthritis. Rheum.Dis. Clin. N. Am. 13:531-544.

5. Claman, H. H., and A. D. Robertson. 1994. Antinuclear antibodiesand breast implants. West. J. Med. 160:225-228.

6. Clarke, S. M. 1991. A novel enzyme-linked immunosorbent assay(ELISA) for the detection of beryllium antibodies. J. Immunol.Methods 137:65-72.

7. Digby, J. M. 1982. Malignant lymphoma with intranodal siliconerubber particles following metacarpophalangeal joint replace-ments. Hand 14:326-328.

8. Endo, L. P., N. L. Edwards, S. Longley, L. C. Corman, and R. S.Panush. 1987. Silicone and rheumatic diseases. Semin. ArthritisRheum. 17:112-118.

9. Gitelman, H. J., and F. R. Alderman. 1990. Determination ofsilicon in biological samples using electrothermal atomic absorp-tion spectrometry. J. Anal. Spectrom. 5:687-689.

10. Gitelman, H. J., F. R. Alderman, and S. J. Perry. 1992. Siliconaccumulation in dialysis patients. Am. J. Kidney Dis. 19:140-143.

11. Gordon, M., and P. G. Bullough. 1982. Synovial and osseousinflammation in failed silicone rubber prosthesis: a report of sixcases. J. Bone Jt. Surg. 64A.:574-580.

12. Jones, W. W., and W. R. Williams. 1983. Value of berylliumlymphocyte transformation tests in chronic beryllium disease andin potentially exposed workers. Thorax 38:41-44.

13. Kapsenberg, M. L., E. A. Wierenga, F. E. M. Stiedema, A. M. B. C.Tiggelman, and J. D. Bos. 1992. TH1 lymphokine productionprofiles of nickel-specific CD4+ T-lymphocyte clones from nickelcontact allergic and non-allergic individuals. J. Invest. Dermatol.98:59-63.

14. Kossovsky, N., M. Zeidler, G. Chun, N. Papasian, A. Nguyen, S.Rajguru, J. Stassi, A. Gelman, and E. Sponsler. 1993. Surfacedependent antigens identified by high binding avidity of serumantibodies in a subpopulation of patients with breast prostheses. J.Appl. Biomater. 4:281-288.

CLIN. DIAGN. LAB. IMMUNOL.

on August 23, 2020 by guest

http://cvi.asm.org/

Dow

nloaded from

T-CELL RESPONSE IN WOMEN WITH SILICONE BREAST IMPLANTS 695

15. Kreiss, K., F. Miller, L. S. Newman, E. A. Ojo-Amaize, M. D.Rossman, and C. Saltini. 1994. Chronic beryllium disease-fromthe workplace to cellular immunology, molecular immunogenetics,and back. Clin. Immunol. Immunopathol. 71:123-129.

16. Nakamura, A., Y. Kawasaki, K. Takada, Y. Aida, Y. Kurokama, S.Kojima, H. Shintani, M. Matsui, T. Nohmi, A. Matsuoka, T.Sofuni, M. Kurihara, and N. Miyata. 1992. Difference in tumorincidence and other tissue responses to polyetherurethanes andpolydimethylsiloxane in long-term subcutaneous implantation intorats. J. Biomed. Mater. Res. 26:631-650.

17. Ojo-Amaize, E. A., M. S. Agopian, T. N. Markham, and J. B. Peter.1992. Primary sensitization and restimulation of human lympho-cytes with beryllium in vitro. J. Allergy Clin. Immunol. 89:203.(Abstract.)

18. Ojo-Amaize, E. A., M. S. Agopian, and J. B. Peter. 1994. Novel invitro method for identification of individuals at risk for berylliumhypersensitivity. Clin. Diagn. Lab. Immunol. 1:164-171.

19. Press, R I., C. L. Peebles, Y. Kumagai, R. L. Ochs, and E. M. Tan.1992. Antinuclear autoantibodies in women with silicone breastimplants. Lancet 340:1304-1307.

20. Roberts, N. B., and P. Williams. 1990. Silicon measurement inserum and urine by direct current plasma emission spectrometry.Clin. Chem. 36:1460-1465.

21. Rogers, L. A., J. A. Longtime, M. B. Garnick, and G. S. Pinkus.1988. Silicone lymphadenopathy in a long distance runner: com-plication of a silastic prosthesis. Hum. Pathol. 19:1237-1239.

22. Romagnoli, P., G. A. Spinas, and F. Sinigagla. 1992. Gold-specificT cells in rheumatoid arthritis patients treated with gold. J. Clin.Invest. 89:254-258.

23. Rossman, M. D., J. A. Kern, J. A. Elias, M. R. Cullen, P. E.

Epstein, 0. P. Pruess, T. N. Markham, and R. P. Daniele. 1988.Proliferative response of bronchoalveolar lymphocytes to beryl-lium: a test for chronic beryllium disease. Ann. Intern. Med. 108:687-693.

24. Saltini, C., K. Winestock, M. Kirby, P. Pinkston, and R G.Crystal. 1989. Maintenance of alveolitis in patients with chronicberyllium disease by beryllium-specific helper T cells. N. Engl. J.Med. 320:1103-1109.

25. Sergott, T. J., J. P. Limoli, C. M. Baldwin, and D. R Laub. 1986.Human adjuvant disease, possible autoimmune disease after sili-cone implantation: a review of the literature, case studies, andspeculation for the future. Plast. Reconstr. Surg. 78:104-114.

26. Shons, A. R, and W. Schubert. 1992. Silicone breast implants andimmune disease. Ann. Plast. Surg. 28:491-501.

27. Spierra, H. 1988. Scleroderma after silicone augmentation mam-moplasty. JAMA 260:236-238.

28. Stokes, R F., and M. D. Rossman. 1991. Blood cell proliferationresponse to beryllium: analysis by receiver-operating characteris-tics. J. Occup. Med. 33:23-28.

29. Varga, J., R. Chumacher, and S. A. Jimenez. 1989. Systemicsclerosis after augmentation mammoplasty with silicone implants.Ann. Intern. Med. 111:377-383.

30. Winchurch, R A. 1988. Activation of thymocyte responses tointerleukin-1 by zinc. Clin. Immunol. Immunopathol. 47:174-180.

31. Yoshida, K. 1973. Post mammoplasty disorder as an adjuvantdisease of men. Shikoku Acta Med. 29:318-332.

32. Yoshida, S. H., C. C. Chang, S. S. Teber, and M. E. Gershwin.1993. Silicon and silicone: theoretical and clinical implications ofbreast implants. Regul. Toxicol. Pharmacol. 17:3-18.

VOL. 1, 1994

on August 23, 2020 by guest

http://cvi.asm.org/

Dow

nloaded from