0021-972x/92/7403-0585$03.00/0 Journal of Clinical Endocrinology Copyright 0 1992 by The Endocrine

and Metabolism Society

Immunoglobulins Phospholipase-A2

Vol. 74, No. 3

Printed

in U.S.A.

from Graves’ Patients Stimulate in FRTLB Thyroid Cells*

ALFRED0 DI CERBO, MARIA DI GIROLAMOt, VITO DE FILIPPIS, AND DANIELA CORDA

VINCENZO

GUARDABASSO,

Laboratory of Cellular and Molecular Endocrinology (A.D.C., M.D.G., D.C.) and Biomathemathics and Applied Informatics Research Unit (V.G.), Istituto di Ricer& Farmacologiche “Mario Negri,” Consorzio Mario Negri Sud, 66030 Santa Maria Zmbaro, Chieti, Italy; and the Division of Endocrinology, “Casa Sollieuo della Sofferenza” General Hospital (A.D.C., V.D.F.), 71013 San Giouanni Rotondo, Foggia, Italy

ABSTRACT. The well documented ability of immunoglobulins G (IgGs) from Graves’ patients to stimulate CAMP production is believed to be involved in the pathophysiology of this disease. It is still under discussion whether other intracellular messengers known to regulate thyroid function might play a similar role. This study shows that phospholipase-A2, a-signal pathway unrelated to CAMP. is activated bv Graves’ IeGs. The IgGs from 67 patients with’active Graves’ disease, 8-patients with Graves’ disease in remission, 5 patients with idiopathic myxedema, 2 patients with Hashimoto’s thyroiditis, 57 patients with nonautoimmune thyroid disease, and 65 normal subjects were tested for their ability to stimulate phospholipase-A2 activity, as measured by arachidonic acid release from FRTL5 thyroid cells. The IgGs from patients with active Graves’ disease caused a significant increase in arachidonic acid release compared to those from normal subjects, patients with nonautoimmune thyroid diseases, and patients with Graves’ disease in remission (P

< 0.0001). The IgGs from active Graves’ patients were also able to increase CAMP accumulation in FRTL5 cells. This effect did not correlate with the ability of the same IgGs to induce arachidonic acid release, suggesting that Graves’ IgGs stimulate these two pathways by separate mechanisms. Moreover, a subgroup of IgGs that stimulated phospholipase-A2 did not increase the CAMP levels in FRTL5 cells. Our data suggest a novel mechanism of action of Graves’ IgGs, the activation of phospholipase-A2, well distinguishable from the known effect on CAMP accumulation. The assay we describe could be helpful in improving the diagnosis and therapy of Graves’ disease and in distinguishing it from nonautoimmune thyroid diseases. It also supplies the basis for a prospective subclassification of the Graves’ patients, which might become useful to clarify the pathophysiology of this disease. (J Clin Endocrinol Metab 74: 585-592, 1992)

G

pathways unrelated to CAMP might be triggered by Graves’ IgGs associated with the disease. For this purpose we have used a continuous line of rat thyroid cells (FRTL5), which has been widely employed to analyze the regulation of the thyroid (9-22). In addition to being well characterized, FRTL5 cells offer advantages of convenience and reproducibility for screening large numbers of sera and are already being used in clinical laboratory tests to measure CAMP accumulation induced by Graves’ IgGs (3, 17). Previous work in FRTL5 cells has shown that the regulation of iodide transport and iodination of thyroglobulin involves phospholipase-C and phospholipase-A2 and that these pathways are controlled by cqadrenergic and muscarinic receptors (12-14, 16, 18-21). The regulation of growth has also been characterized in FRTL5 cells and appears to be related to three transduction pathways: a CAMP-dependent mechanism activated by TSH, and two CAMP-independent mechanisms. The first involves stimulation of the insulin-like growth factor-1 (IGF-D-dependent kinase, while the second is mediated by the activation of phospholipase-A2 and the

RAVES’ disease has been associated with the presence of antithyroid immunoglobulins G (IgGs) (1). These antibodies affect thyroid function and growth by a mechanism that is still partially unclear. Some IgGs stimulate thyroid via the TSH receptor and are believed to act, like TSH, by increasing the cellular levels of CAMP, the main intracellular regulator of growth and differentiation in thyroid cells (l-5). Thyroid growth, however, can also be induced by IgGs that do not affect CAMP production, suggesting that other transduction mechanisms might account for their effects (6-8). The aim of this study was to examine which transduction Received March 22, 1991. Address requests for reprints to: A. Di Cerbo, Istituto di Ricerche Farmacologiche “Mario Negri,” Consorzio Mario Negri Sud, 66030 Santa Maria Imbaro, Chieti, Italy. *This work was supported in part by the Italian Association for Cancer Research (AIRC), Fidia S.p.A., and the Agenzia per la Promozione e lo Sviluppo de1 Mezzogiorno. t Recipient of a fellowship from the Centro di Formazione e Studi per il Mezzogiorno-Formez (Progetto Speciale “Ricerca Scientifica Applicata nel Mezzogiorno”). 585

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 12 November 2015. at 12:09 For personal use only. No other uses without permission. . All rights reserved.

JCE & M. 1992 Vol74.No3

DI CERBO ET AL.

586

subsequent formation of prostaglandin-E2 (10-12, 23). Prostaglandin-Ep has been shown to induce DNA synthesis in a CAMP-independent manner in FRTL5 cells (12). Furthermore, DNA synthesis induced by TSH and monoclonal antibodies against the TSH receptor is blocked in part by the cyclooxygenase inhibitor indomethacin (6). The above considerations attracted our attention to phospholipase-A2 as a likely target for the stimulating activity of IgGs from Graves’ patients. The results show that a subpopulation of Graves’ IgGs can be identified that induces arachidonic acid release without affecting CAMP, thus suggesting phospholipase-A2 as an alternative mechanism by which IgGs may alter thyroid cell function. These data suggest that it may be possible to identify clinically different subgroups of Graves’ patients on the basis of the biochemical effects of their IgGs.

Materials

and Methods

Rea&nts

Hormonesused in the tissue culture medium, Coon’smodified Ham’s F-12 medium, HEPES, EGTA, BSA, norepinephrine, calcium ionophore A23187, glucose, Tris (hydroxymethyl)aminomethane, and 3-isohutyl-l-methylxanthine (IBMX) were obtained from Sigma (St. Louis, MO). Tissue culture materials were obtained from Gibco (Grand Island, NY). Protein-A-Sepharose CL4B was purchased from Pharmacia LKB Biotechnology (Uppsala, Sweden), glycine from Bio-Rad Laboratories (Richmond, CA), and Minicon B15 concentrators from Amicon Division (W. R. Grace Co., Danvers, MA). [N-3H]Arachidonic acid and CAMP RIA kits were purchased from New England Nuclear Corp. (Boston, MA). Purified TSH was obtained from the National Hormone and Pituitary Program (University of Maryland). Patients

Sera were obtained from 67 patients with active Graves’ disease(9 malesand 58 females,aged 13-80 yr; 17 before any treatment, 2 with recurrent Graves’ diseaseafter subtotal thyroidectomy, 47 receiving methimazole,1 receivingpropylthiouracil treatment), 8 patients with Graves’ diseasein remission,2 patients with Hashimoto’s thyroiditis, 5 patients with idiopathic myxedema, 57 patients with nonautoimmune thyroid disease(including multinodular nontoxic goiter, autonomously functioning thyroid nodule,multinodular toxic goiter, and TSH thyrotoxicosis), and 65 normal subjects.Active Graves’ disease has been characterized by diffuse thyroid enlargement, high levelsof free thyroid hormones,and undetectablelevelsof TSH, with classicsymptomsand signsof thyrotoxicosis. Twenty-five patients had clinical evidence of Graves’ exophtalmopathy. Antithyroglobulin and antithyroid microsomalantibodieswere positive in 46 and 61 of Graves’ patients, respectively. Remissionof Graves’ diseasewasdefined asthe absenceof symptoms and signsof thyrotoxicosis and normal levels of free thyroid hormonesand TSH at least 1 yr after the interruption of the

therapy. TSH binding-inhibiting IgGs (TBII) were not present in the serafrom patients with uni- or multinodular goiter. Two patients with TSH thyrotoxicosis were suffering from a TSHand GH- plus TSH-secreting pituitary neoplasia,respectively. Normal subjectsincluded age- and sex-matchedblood donors. Their characteristics were similar to those of the patients with Graves’disease(10 menand 55 women,aged19-67yr). Subjects with familial history positive for autoimmune diseasesas well as a personal history of recent acute or chronic diseaseswere excluded from the study. Cells FRTL5 cells, a continuousline of functioning epithelial cells derived from normal Fisher rat thyroids, were cultured as previously describedin Coon’smodified Ham’s F-12 medium, 5% calf serum, and a six-hormone mixture [TSH, insulin, cortisol, transferrin, glycyl-L-hystidyl-L-lysine acetate, and somatostatin (6H medium)] (9, 15, 22). AtT-20 D 16/16 cells, a clone of tumoral mousepituitary cells, and RBL 2H3, cloned from rat tumoral mast cells, were kindly provided by A. Luini. Their characteristics and culture conditions have been previously described(24, 25). IgG purification

IgG purification was performed by a modification of the method of Readeret al. (26). Up to 2 mL serum were applied to protein-A-SepharoseCL4B columns (10 mM Tris-HCl, pH 7.4), and the bound IgGs were washed,then eluted with 0.1 M glycine-0.1 N NaCl (pH 2.9), concentrated on Minicon-B15 concentrators, and resuspendedin a saline solution to a final concentration of lo-15 mg/mL. Arachidonic

acid release

The arachidonic acid releasefrom monolayers of FRTL5 cells maintained in 6H medium was evaluated as previously reported (12,21). Labeling wascarried out overnight at 37 C in 1 mL 6H medium containing [3H]arachidonic acid (0.1 &i/ well). Cell washesand releaseassayswere performed with 1 mL Hanks’ Balanced Salt Solution (pH 7.4), containing fatty acid-freeBSA (2 mg/mL; HBSS-BSA), in the absenceof TSH. On the average, about 2000 cpm [3H]arachidonic acid were releasedin control wells. Normal and Graves’ IgGs were tested in duplicate at 10 and 100 pg/mL concentrations. Resultsare expressedas percent release us. that in the basal sample (HBSS-BSA only). Alternatively, the Graves’ patient IgGinduced releasewas normalized us.that of the IgGs of normal subjects.All determinations were performed in duplicate in a minimum of three different cell preparations. Coefficients of variation within the individual experiment and amongdifferent experimentswere 17% and 22%, respectively. CAMP assay

The CAMP levelswere evaluatedby a commercialRIA. Cells in 96-well plates grown to 80% confluency were placed in 5H mediumfor 4-10 days. Purified IgGs (1 mg/mL) were addedin 100 PL of modified HBSS (250 mM sucrose,5.4 mM KCl, 1.3

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 12 November 2015. at 12:09 For personal use only. No other uses without permission. . All rights reserved.

GRAVES’

IgG

STIMULATION

OF

PHOSPHOLIPASE-A2

587

mM CaCI,, 0.4 mM MgS04, 0.5 mM MgCl?, 0.4 mM Na2HP04, 0.44 mM KH2POI, 5.5 mM glucose,and 10 mM HEPES) containing BSA (4 mg/ml) and IBMX (0.5 mM; pH 7.3). The

incubation was performed for 60 min at 37 C. The CAMP content, evaluated from the extracellular medium, was expressedas picomolesper well; alternatively, the results were normalized against basal CAMP production (HBSS-BSAIBMX

only)

or against

the

CAMP

values

obtained

when

the

cells were stimulated with normal IgGs. All determinations were performed in triplicate from at least three different cell preparations. Coefficients of variation within the individual experiment and among different experiments were 13% and 18%, respectively. Other assays TBII activity and serumTB, T4, and TSH were evaluated by commercial RRA and RIAs [Byk Sangtec (Dietzenbach, Germany), Amersham(Aylesbury, Buckinghamshire,United Kingdom), and Travenol (Cambridge,MA)]; antithyroglobulin and antithyroid microsomal antibodies were determined by the tanned red cell hemoagglutinationtechnique, usingcommercial kits (Fujirebio, Inc., Tokyo, Japan). Statistical analysis

FIG. 1. Phospholipase-A2

activity, as measured by arachidonic acid (AA) release (upper panel), and adenylyl cyclase activity, as measured by CAMP accumulation (lower panel), in FRTL5 cells. IgGs from normal subjects and patients with active Graves’ disease, Graves’ disease in remission, idiopathic mixedema, multinodular toxic goiter, multinodular nontoxic goiter, Hashimoto’s thyroiditis, and TSH thyrotoxicosis were added to the cells at final concentrations of 100 pg/ mL (AA) and 1 mg/mL (CAMP). In parentheses are indicated the numbers of subjects for each group. Two patients with autonomously functioning thyroid nodule have been included in the multinodular toxic goiter group. Results are expressed as a percentage of the basal arachidonic acid release (HBSS-BSA only). Individual data are the means of duplicate determinations obtained in three to eight experiments performed with different cell preparations. CAMP accumulation was evaluated by a commercial RIA; data are presented as picomoles per well. Individual data pointsrepresent the means of triplicate determinations obtained in two to four experiments performed with different cell preparations. The basal CAMP accumulation in FRTL5 cells was 0.38 + 0.04 pmol/well. The bottom and top edges of boxes superimposed on data are located at the sample 25th and 75th percentiles. The center horizontal thick line is drawn at the sample median. This plot provides a visual representation of central tendency and variability of data (see Materials and Methods). The IgGs from active Graves’ disease patients caused a significant increase in AA release and CAMP accumulation compared to those from the other groups (by one-way analysis of

The observed CAMP and arachidonic acid release for all groups was preliminarily analyzed to obtain parametric and ordinal descriptive statistics and tested for compliancewith the Gaussiannormal distribution. Upper and lower quartiles (75th and 25th percentiles) and medianswere usedto construct the plots of Fig. 1, where the data are plotted with boxes superimposed. The bottom and top edgesof a box are located at the sample25th and 75th percentiles. The center horizontal thick line is drawn at the samplemedian. This plot is a simplified version of the “box and whiskers” plot (27) and provides a visual representation of central tendency and variability of data. Due to significant departuresfrom normality within several groups and heterogeneity of variances, hypothesis testing was performed using nonparametric methods. The tests used were the Wilcoxon rank sum test for two-group comparisons, one-way analysisof variance appliedto ranked data (equivalent to Kruskal-Wallis test) for multiple group comparison, and Spearman’srank correlation. All statistical analyseswere performed using the SAS System (SAS Institute, Inc., Cary, NC) on a MicroVAX 3600 under VMS Operating System (Digital Equipment Co., Maynard, MA). Results Arachidonic acid release induced by normal and Graves’ IgGs We have previously reported that norepinephrine, carbachol, and TSH are able to release arachidonic acid from FRTL5 cells, and that this effect is related to variance on ranked data, P < 0.0001). The data from patients with Hashimoto’s disease and TSH thyrotoxicosis were excluded from analysis due to their small number.

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 12 November 2015. at 12:09 For personal use only. No other uses without permission. . All rights reserved.

588

DI CERBO

activation of phospholipase-A2; these ligands were used as positive controls and yielded results similar to those previously reported (data not shown) (12, 18, 19, 21). IgGs from patients with active Graves’ disease caused greater arachidonic acid release than those from normal subjects (41 of 67 samples were above the 75th percentile of the normal values; Fig. 1, upper panel). The relationship between Graves’ IgG concentration and arachidonic acid release is shown in Fig. 2. On the average, the increase was 145.4%. The difference was highly statistically significant (P < 0.0001). When IgGs from normal subjects were compared to HBSS, the median arachidonic acid release was inhibited by 23% (Fig. 1, upper panel). This inhibitory effect was not simply due to a protein effect, since up to 6 mg/mL BSA did not change the generation of arachidonic acid (data not shown). Figure 3 shows the frequency distributions of the release values induced by normal and Graves’ IgGs. Normal subjects appeared to have a bimodal pattern of distribution, with negative kurtosis (a relative flatness, with wider tails compared to the Gaussian distribution); however, the statistical normality testing did not allow rejection of the hypothesis of Gaussian distribution at the customary P < 0.05 level (P was 0.10). Arachidonic acid release induced by IgGs from nonautoimmune thyroid diseases

The arachidonic acid release induced by the IgGs of patients affected by multinodular nontoxic goiter, autonomously functioning thyroid nodule, multinodular toxic goiter, and TSH thyrotoxicosis was not significantly different from that of normal subjects (Fig. 1, upper panel).

ET AL.

JCE & M. 1992 Voll4.No3

Arachidonic

Acid release (% of basal)

FIG. 3. Frequency distribution patterns of arachidonic acid release induced by IgGs from patients with active Graves’ disease compared with those from normal subjects. The arachidonic acid release induced by the IgGs (100 rg/mL) was analyzed in the FRTLB cell assay, as reported in Fig. 1. Frequency histograms were constructed for data from 67 patients with active Graves’ disease (-) and 65 normal subjects (IV). The histogram for normal patients shows a bimodal pattern of distribution, with modes at about 70% and 100%. In addition, both histograms appear flatter and wider than the Gaussian distribution.

Arachidonic acid release induced by IgGs from Graves’ patients in remission

If a specific IgG plays a role in the pathophysiology of Graves’, it might be expected to decrease or disappear in patients that are no longer in the active phase. Indeed, IgGs from patients with Graves’ disease in remission had levels of arachidonic acid release within the range of normal subjects (Fig. 1, upper panel). IgGs from Graves’ patients at different stages of treatment tended to be less effective in stimulating arachidonic acid release than IgGs from untreated patients (43.5% vs. 51.3% above normal subject values, respectively). The difference was not statistically significant. Lack of a role for CAMP and cytosolic Ca2+in the arachidonic acid releaseinduced by IgGs

FIG. 2. Dose-response curve of arachidonic acid (AA) release induced by IgGs from Graves’ patients. The indicated concentrations of IgGs from 30 patients with active Graves’ disease were added to FRTL5 cells under the experimental conditions detailed in MoterMk and Methods. Results are expressed as a percentage of the basal arachidonic acid release (HBSS-BSA only) and are the mean + SE of duplicate determinations performed in 6 different cell preparations.

Graves’ IgGs induce CAMP accumulation in FRTL5 cells (see below). It might be argued that the stimulation of arachidonic acid release could be secondary to an increase in CAMP production. To rule out this possibility we have examined the effect on arachidonic acid release of agents known to increase CAMP. Cholera toxin, which stimulates adenylyl cyclase in FRTL5 cells (18), did not affect arachidonic acid release (100% and 104% of the basal values at 1 and 10 nM cholera toxin, respectively). Similarly, forskolin, which directly activates adenylyl cyclase, and 8-bromo-CAMP, a nonhydrolizable CAMP analog, were unable to stimulate arachidonic acid release (data not shown). The release of arachidonic acid might also be secondary to an increase in cytosolic Ca2+ (due to either the activation of phospholipase-C or the opening of calcium channels). Calcium might activate a phospholipase-A2

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 12 November 2015. at 12:09 For personal use only. No other uses without permission. . All rights reserved.

GRAVES’

IgG STIMULATION

and/or a phospholipase-C and, consequently, a diacylglycerol lipase acting on the product (diacylglycerol) of phospholipase-C (28). However, under the experimental conditions used to evaluate arachidonic acid release, Graves’ IgGs were not able to modify the cytosolic Ca2+ levels of FRTL5 cells (115 f 10 nM, as previously reported) (14, 20). This indicates that under the experimental conditions employed, Graves’ IgGs are unable to activate phospholipase-C and increase cytosolic Ca2+. The direct activation of phospholipase-A2 by Graves’ IgGs interacting with membrane proteins is, thus, the most likely mechanism leading to the release of arachidonic acid. Releaseof arachidonic acid by Graves’ IgGs is thyroid specific

To investigate whether the effect of Graves’ IgGs is general or restricted to the thyroid, we examined the effects of normal and Graves’ IgGs on the activity of phospholipase-A2 in cell lines of nonthyroidal origin. Table 1 shows that the IgGs of either normal or Graves’ serum did not affect arachidonic acid release in AtT-20 cells (originated from mouse pituitary cells) and RBL cells (originated from rat mast cells). In both lines, arachidonic acid could be generated by phospholipaseA2 stimulants, such as the Ca2+ ionophore or melittin (data not shown).

OF PHOSPHOLIPASE-AZ

589

Graves’ patients caused a pronounced increase in CAMP production, in confirmation of previous results (P < 0.0001; Fig. 1, lower panel). A significantly higher increase was found in active Graves’ patients compared to that in the other groups (P < 0.0001; Fig. 1, lowerpanel). Similar data have been previously reported (3-5). Interestingly, when active Graves’ IgGs were analyzed for their ability to stimulate both adenylyl cyclase and phospholipase-A2, no significant correlation was found between the two sets of data (Fig. 4, upper panel). A significant correlation (r, = 0.48; P < 0.0003) was found between TBII activity and adenylyl cyclase stimulation (Fig. 4, middle panel). Phospholipase-A2 activation did not significantly correlate with TBII activity, suggesting that the two assays, i.e. arachidonic acid release and CAMP accumulation, do not measure the same mechanisms of activation (Fig. 4, lower panel). Thyroid hormones, TSH, and TBII serum levels

The serum levels of thyroid hormones, TSH, and TBII were evaluated for all of the Graves’ patients and the normal subjects examined in this study. There was not a statistically significant correlation between the arachidonic acid release stimulated by the IgGs of Graves’ the patients and the serum levels of the thyroid hormones, TSH, and TBII (data not shown). Discussion

CAMP accumulation

The ability of IgGs from Graves’ patients to induce an increase in CAMP production has been previously reported by several laboratories (3-5, 7). We measured the CAMP accumulation induced by the same IgGs tested in the arachidonic acid release assay in order to compare their relative efficacy on these two pathways. IgGs from TABLE 1. Arachidonic acid release induced in FRTL5, RBL, and AtTPO cells by IgGs from normal subjects and patients with active Graves’ disease Cells FRTL5 RBL AtT20

Release (% of basal) Normal subjects 73.4 + 8.3 (72.5) 117.7 + 3.5 (114.5) 98.2 + 4.1 (99)

Active Graves’ disease 141.1 + 2.7 (141.2)” 117.2 + 1.6 (117) 99.3 + 4.6 (98)

IgGs from 10 randomly selected normal subjects and 11 patients with active Graves’ disease were added to the cells at a 100 rg/mL concentration. The Graves’ IgGs were randomly selected among those able to stimulate arachidonic acid release (see Fig. 3). Data are expressed as a percentage of the basal arachidonic acid release (HBSSBSA only). Results are the mean + SE of duplicate determinations performed with three different cell preparations. Medians are reported in parentheses. s Significantly different from the values obtained with IgGs from normal subjects (p < 0.0001, by Wilcoxon rank sum test).

The production of IgGs by B-lymphocytes infiltrating the thyroid is considered to be an essential link in the development of Graves’ thyrotoxicosis (29, 30). Some Graves’ IgGs are functionally similar to TSH, since they are able to stimulate CAMP production in several thyroid systems, including human thyroid slices and membranes (1, 2). The main finding in this paper is that IgGs from Graves’ patients can activate another important signalgenerating pathway, phospholipase-A2. PhospholipaseA2-stimulating IgGs are present in about 60% of the Graves’ sera, as measured by arachidonic acid release from FRTL5 cells, with an average increase of 45% over that in normal subjects. The stimulating activity was absent in patients with nonautoimmune thyroid disorders, Hashimoto’s thyroiditis, and idiopathic myxedema and in all Graves’ patients in remission. A small fraction (16.9%) of the IgGs from normal subjects had some stimulatory activity on phospholipase-A2. The stimulation was not related to sex or age (data not shown). A possibility is that the bimodal pattern of distribution of normal subjects could reflect a subclinical autoimmune thyroid disease in the individuals whose IgGs can activate phospholipase-A2. A follow-up of this population should be performed to further investigate this point.

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 12 November 2015. at 12:09 For personal use only. No other uses without permission. . All rights reserved.

DI CERBO

590

1

.

.

l

I

q

4. .

n .

.

l

.

32-

.

l-

.

/

.

l ..m--.m.-.

O60

AA mlraw

0

20

10

..

120

100

80

.

n

.

140

160

(K of basal)

30

40

60

60

TBll (‘A of Inhlbltlon)

s

140

I 5120

100

.

.

.

.

I

.

t

. . .

.

.

. . ,,..,.............. . ....‘.“’ ,,,...,,,............... .‘.’ :

. .

n .-

.

.

I

a, 2

m

: Zm .S, . ,....”. . ..._“...=1.... ~ . .. . .. .. ...!. .. .. . . .. .. . ,,,f. .. . . .. .. ...” . . .. . .. ..-a.

t

al

I l

0

.

.

70

80

l

l

.

.

.

% --

I

0

10

I.

20

I.

30

8.

40

I

50

I.

60

I

70

TBII (% of inhibition)

FIG. 4. Correlations of the results of CAMP production, arachidonic acid (AA) release, and TBII activity in 52 patients with active Graves’ disease. CAMP values are expressed as picomoles per well. TBII activity is expressed as percent inhibition of [“‘I]TSH binding on thyroid membranes. AA release is expressed as a percentage of the basal release (HBSS-BSA only). Each point represents the mean of triplicate determinations obtained in at least two separate experiments. The dotted regression lines, obtained from linear regression analysis, are shown for descriptive purposes only. Correlation was tested by Spearman’s nonparametric rank correlation. A significant correlation was found between CAMP production and TBII activity (middle panel; r. = 0.48; P c 0.0003), but not between CAMP production and AA release or AA release and TBII activity (upper ancl lower panels, respectively; r. = 0.02, P = NS; and rs = 0.2, P = NS).

Most IgGs from normal subjects, however, slightly but significantly inhibited basal arachidonic acid release. It should be noted that normal IgGs have also been reported to inhibit CAMP accumulation in human thyroid cells (31). Likewise, they reduce by about 50% the basal CAMP accumulation in FRTL5 cells, as reported in this paper. The meaning of these effects is not clear. It could be speculated that thyroid cells have Fc receptors exerting an inhibitory influence on both adenylyl cyclase and

ET

AL.

JCE & M.

1992

Vol74.No3

phospholipase-A2. All of the IgGs available were also tested in the CAMP assay. Fifty-two of 67 (77.6%) Graves’ IgGs stimulated CAMP production. When we tried to correlate the ability to stimulate phospholipase-A2 with the stimulation of CAMP accumulation by the same IgGs, we found that 1) no correlation existed between the two pathways; and 2) the IgG activity existed in all possible combinations with respect to the two assays, namely some IgGs induced high arachidonic acid release and high CAMP production, while other IgGs yielded high arachidonic acid release and low CAMP production or vice uersa. Finally, some IgGs exclusively stimulated only one of the two pathways. This reproduces a behavior already reported with regard to the TBII assay and the stimulation of CAMP accumulation (32-34). A possible explanation is that two different IgGs, one able to activate the arachidonic acid cascade and the other able to generate CAMP, may be present in Graves’ patients. The relative amounts of the two IgGs would be the variable determining the different behaviors described above (high CAMP vs. low arachidonic acid, and vice uersa). The interest of this observation is that it suggests that a subset of Graves’ patients cannot be detected by the classical CAMP test, but only by the arachidonic acid release assay. Another implication of potential clinical interest is that different subtypes of Graves’ disease might be defined according to the activity of the Graves’ IgGs on CAMP accumulation and arachidonic acid release. Future work might also establish whether these biochemical subtypes could correlate with different manifestations of Graves’ disease. We are now employing the arachidonic acid release to follow several cases of Graves’ from diagnosis to remission in order to characterize whether the results may relate to the various stages of the disease. It appears that arachidonic acid release, similar to what has been previously reported for CAMP production, is significantly lower in Graves’ patients in remission than in those in the active stage, suggesting that this assay could be of diagnostic value to estimate the clinical course of the disease. It may be asked whether the stimulation of phospholipase-A2 might play a role in the regulation of thyroid function. Preliminary results from our laboratory show that some Graves’ IgGs that stimulate arachidonic acid release and do not affect CAMP levels are also able to stimulate thymidine incorporation in FRTL5 cells, indicating that this pathway might indeed be involved in the regulation of thyroid cell proliferation (data not shown). By what mechanism is arachidonic acid release induced by Graves’ IgGs? Throughout the paper it has been assumed that phospholipase-A2 is the enzyme responsible for this effect. Arachidonic acid can be gener-

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 12 November 2015. at 12:09 For personal use only. No other uses without permission. . All rights reserved.

GRAVES’

IgG STIMULATION

ated through two main pathways, one involving phospholipase-A2, and the other involving phospholipase-C, with consequent production of inositol trisphosphate, which releases cytosolic Ca2+, and of diacylglycerol, a substrate of diacylglycerol lipase, which finally liberates arachidonic acid. The observation that the release of arachidonic acid is not accompanied by an increase in cytosolic Ca2+ indicates that phospholipase-C is not involved (see Results). Moreover, diacylglycerol lipase has been shown to be inactive in FRTL5 cells (35). Thus, phospholipase-A2 activation is likely to account for the observed stimulation of arachidonic acid release. It is not clear which membrane protein might mediate the activation of phospholipase-A2 by IgGs. The lack of correlation between arachidonic acid release and antithyroglobulin or antimicrosomal antibodies tends to exclude that these antigens could be involved in stimulation of the enzyme. A possible target is the low affinity portion of the TSH receptor, which might be responsible for the ability of TSH to activate the arachidonic acid cascade at micromolar concentrations (19). Other candidates might be the muscarinic and adrenergic receptors, which can stimulate phospholipase-A2, although in this case the existence of thyroid-specific receptors should be postulated to explain the effect of Graves’ IgGs on the thyroidal tissue only. The possibility that the arachidonic acid release induced by Graves’ disease IgGs could be secondary to the CAMP production elicited by the IgGs is excluded by several findings. One is the existence of IgGs (19.4% in our series) able to stimulate arachidonic acid release without affecting CAMP accumulation. The second is that arachidonic acid release is not affected by agents known to potently stimulate CAMP production in FRTL5 cells, such as cholera toxin and forskolin (see Results).

Regarding the relevance of the above data in human thyroid, there is indeed a possibility that phospholipaseA2 activation by IgGs could play a role in the pathophysiology of Graves’ disease. Prostaglandin-E2, an arachidonic acid metabolite, has been shown to be generated and to have mitogenic activity in FRTL5 cells (12). However, there is no information on the modulation of phospholipase-A2 by Graves’ IgGs in human thyroid, nor is it known whether this enzyme is involved in the regulation of the human system. This is being presently investigated in our laboratory. In conclusion, we show that Graves’ IgGs activate arachidonic acid release in FRTL5 cells. This effect might be relevant to the pathophysiology of the disease. The arachidonic acid release assay described in this study is a technically simple test that, in combination with the CAMP test, should be of help in differentiating autoimmune from nonautoimmune hyperthyroid disorders and should provide a further means to define the stage of the

OF PHOSPHOLIPASE-A2

591

disease and/or predict its relapse. Furthermore, it is possible that clinical subgroups of Graves’ patients will be identified according to the biochemical effect of their IgGs. Acknowledgments Drs. A. Balsam0 and G. Gallone assisted in the care of some patients at the Mauriziano Hospital (Turin, Italy) and provided the related clinical information included in this article. We thank Drs. S. M. Aloj and A. Luini for critical reading of the manuscript. We are indebted to L. Mongardi for performing the TBII assay, to C. Di Filippo for assistance in data handling, and especially to R. M. Marfisi for assistance with the use of SAS for the statistical analysis. We acknowledge the gift of purified TSH from the National Hormone and Pituitary Program. Dr. F. S. Ambesi-Impiombato and Interthyr Research Foundation, Inc., supplied the FRTL5 cells used in this study.

References 1. Burman KD, Baker Jr JR. Immune mechanisms in Graves’ disease. Endocr Rev. 1985;6:183-232. 2. Rees Smith B, McLachlan SM, Furmaniak J. Autoantibodies to the thyrotropin receptor. Endocr Rev. 1988;9:106-21. 3. Vitti P, Rotella CM, Valente WA, et al. Characterization of the optimal stimulatory effects of Graves’ monoclonal and serum immunoglobulin G on adenosine 3’,5’-monophosphate production in FRTL5 thyroid cells: a potential clinical assay. J Clin Endocrinol Metab. 1983;57:782-9. 4. Zakarija M, McKenzie JM. The spectrum and significance of autoantibodies reacting with the thyrotropin receptor. Endocrinol Metab Clin North Am. 1987;16:343-63. 5. Rapoport B, Greenspan FS, Filetti S, Pepitone M. Clinical experience with a human thyroid cell bioassay for thyroid-stimulating immunoglobulin. J Clin Endocrinol Metab. 1984;58:332-8. 6. Kohn LD, Aloj SM, Tombaccini D, et al. The thyrotropin receptor. In: Litwack G, ed. Biochemical action of hormones. New York: Academic Press; 1985;12:457-512. 7. Valente WA, Vitti P, Rotella CM, et al. Antibodies that promote thyroid growth. A distinct population of thyroid-stimulating autoantibodies. N Engl J Med. 1983;309:1028-34. 8. Drexhage HA, Bottazzo GF, Doniach D, Bitensky L, Chayen J. Evidence for thyroid-growth-stimulating immunoglobulins in some goitrous thyroid diseases. Lancet. 1980;2:287-92. 9. Valente A. Vitti P. Kohn LD. et al. The relationshiu of arowth and adenylatecyclase activity in cultured thyroid cells:.separate bioeffects of thyrotropin. Endocrinology. 1983;112:71-9. 10. Tramontano D, Moses AC, Veneziani BM, Ingbar SH. Adenosine 3’,5’-monophosphate mediates both the mitogenic effect of thyrotropin and its ability to amplify the response to insulin-like growth factor I in FRTL5 cells. Endocrinology. 1988;122:127-32. 11. Tramontano D. Cushine GW. Moses AC. Inzbar SH. Insulin-like growth factor-lstimulat& the’growth of rat thyroid cells in culture and synergizes the stimulation of DNA synthesis induced by TSH and Graves’-IgG. Endocrinology. 1986;119:940-2. 12. Burch RM, Luini A, Mais DE, et al. cYi-Adrenergic stimulation of arachidonic acid release and metabolism in a rat thyroid cell line. J Biol Chem. 1986;261:11236-41. 13. Weiss SJ, Philp NJ, Grollman EF. Effect of thyrotropin on iodide efflux in FRTL5 cells mediated by Ca’+. Endocrinology. 1984;114:1108-13. 14. Corda D, Marcocci C, Kohn LD, Axelrod J, Luini A. Association of the changes in cytosolic Ca2+ and iodide efflux induced by thyrotropin and by the stimulation of cul-adrenergic receptors in cultured rat thvroid cells. J Biol Chem. 1985:260:923-6. 15. Ambesi-Impiombato FS, Perrild H, eds. FRTL5 today. Amsterdam:

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 12 November 2015. at 12:09 For personal use only. No other uses without permission. . All rights reserved.

592

DI CERBO

Elsevier; 1989. 16. Marcocci C, Luini A, Santisteban P, Grollman EF. Norepinephrine and thyrotropin stimulation of iodide efflux in FRTL5 thyroid cells involves metabolites of arachidonic acid and is associated with the iodination of thyroglobulin. Endocrinology. 1987;120:1127-33. 17. Vitti P, Chiovato L, Lopez G, et al. Measurement of TSAb directly in serum using FRTL-5 cells. J Endocrinol Invest. 1988;11:313-7: 18. Corda D. Bizzarri C. Di Girolamo M. Valitutti S. Luini A. G Protein-linked receptors in the thyroid. Adv Exp Med Biol. 1989;261:245-69. 19. Corda D, Iacovelli L, Di Girolamo M. Coupling of the cul-adrenergic and thyrotropin receptors to second messenger systems in thyroid cells. Role of G-proteins. In: Maggi M, Johnston CA, eds. Horizons in endocrinology. New York: Raven Press; 1988.169-80. 20. Bizzarri C, Di Girolamo M, D’Orazio MC, Corda D. Evidence that a guanine nucleotide-binding protein linked to a muscarinic receptor inhibits directly phospholipase C. Proc Nat1 Acad Sci USA. 1990;87:4889-93. 21. Di Girolamo M, D’Arcangelo D, Bizzarri C, Corda D. Muscarinic regulation of phospholipase A2 and iodide fluxes in FRTL-5 thyroid cells. Acta Endocrinol (Conenhl. 1991:125:192-200. 22. Ambesi-Impiombato FS, Parks LAM, Coon HG. Culture of hormone-dependent functional epithelial cells from rat thyroids. Proc Nat1 Acad Sci USA. 198&77:3455-g. 23. Jin S, Hornicek FJ, Neylan D, Zakarija M, McKenzie JM. Evidence that adenosine 3’,5’-monophosphate mediates stimulation of thyroid growth in FRTLB cells. Endocrinology. 1986;119:802-10. 24. Luini A, Lewis D, Guild S, Corda D, Axelrod J. Hormone secretagogues increase cytosolic calcium by increasing CAMP in corticotropin-secreting cells. Proc Nat1 Acad Sci USA. 1985;82:8034-8. 25. White KN, Metzger H. Translocation of protein kinase C in rat basophilic leukemic cells induced by phorbol ester or by aggregation of IgE receptors. J Immunol. 1988;141:942-7. 26. Reader SCJ, Davison B, Beardwell C, Ratcliffe JG, Robertson WR.

ET AL.

27. 28. 29. 30. 31.

32.

33.

34.

35.

JCE & M .1992 Vol74.No3

Protein-A purified human imunoglobulins: a comparison of thyroid stimulating and thyrotrophin receptor binding activities in thyrotoxicosis. Clin Endocrinol (Oxfj. 1986;25:441-51. Tukey JW. Exploratory data analysis. Reading: Addison-Wesley; 1977. Bell RL, Kennerly DA, Stanford N, Majerus PW. Diglyceride lipase: a pathway for arachidonate release from human platelets. Proc Nat1 Acad Sci USA. 1979;76:3238-41. McLachlan SM, Pegg CAS, Atherton MC, Middleton SL, Clark F, Rees Smith B. TSH receptor antibody synthesis by thyroid lymphocytes. Clin Endocrinol (Oxf). 1986;24:223-30. Kendall-Taylor P, Knox AJ, Steel NR, Atkinson S. Evidence that thyroid-stimulating antibody is produced in the thyroid gland. Lancet. 1984;1:654-6. Rapoport B, Filetti S, Takai N, Seto P, Halverson G. Studies on the cyclic AMP response to thyroid stimulating immunoglobulins (TSI) and thyrotropin (TSH) in human thyroid cell monolayers. Metabolism. 1982:31:1159-67. Sugenoya A, Kidd A, Row VV, Volpi R. Correlation between thyrotropin-displacing activity and human thyroid-stimulating activity by immunoglobulins from patients with Graves’ disease and other thyroid disorders. J Clin Endocrinol Metab. 1979;48:398402. Borges M, Ingbar JC, Endo K, et al. A new method for assessing the thyrotropin binding inhibitorv activitv in the immunoelobulins and whole serum of patients with Graves’ disease. J Clin Endocrino1 Metab. 1982;54:552-8. Hensen J, Kotulla P, Finke R, et al. Methodological aspects and clinical results of an assay for thyroid-stimulating antibodies: correlation with thyrotropin binding-inhibiting antibodies. J Clin Endocrinol Metab..1984;58:980-7. Burch RM, Luini A, Axelrod J. Phospholipase A2 and phospholipase C are activated by distinct GTP-binding proteins in response to cYi-adrenergic stimulation in FRTLB thyroid cells. Proc Nat1 Acad Sci USA. 1986;83:7201-5.

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 12 November 2015. at 12:09 For personal use only. No other uses without permission. . All rights reserved.