Production of a Non-immunoglobulin Thyroid Stimulator by Human Lymphocytes during Mixed Culture with Human Thyroid Cells*

Human lymphocytes from normal individuals and from patients with Graves’ disease were cultured together with normal human thyroid cells in long term monolayer culture. Stimulation of thyroid cell CAMP content began on the 2nd day of co-culture and rapidly increased to reach a maximum after approximately 72 h. Cell-free culture medium obtained following the mixed culture of lymphocytes and thyroid cells contained thyroid stimulatory activity. Thyroid cells or lymphocytes cultured alone in autologous serum released little or no thyroid stimulating substance into the medium. Surprisingly, lymphocytic thyroid stimulator (LTS) was produced by lymphocytes from all 11 normal subjects tested (stimulation of target cell CAMP content to 271 f 16% SE. of basal values) as well as by lymphocytes from all five patients with Graves’ disease (324 f 44% S.E. of basal values). Human thyroid cells were not found to be specific activators of LTS production, in that such activation was observed with other human cell lines. In some, but not all, experiments, LTS stimulated CAMP content in target cells other than human thyroid, such as human fibroblast and dog thyroid cells. LTS was not precipitated by anti-human immunoglobulin antisera and was dialyzable and heat-stable. LTS bioactivity was extracted at pH 3.0 by 2:l chloroform:methanol and diethyl ether. Gel filtration on Sephadex G-25 and Sephadex G-10 showed that the molecular weight of LTS was 1000 or less, and LTS co-eluted with radiolabeled prostaglandins. On paper chromatography, LTS co-migrated together with prostaglandins, particularly of the E and F category. Comparison of the kinetics of LTS and prostaglandin E, stimulation of thyroid cell CAMP generation demonstrated close similarity in their time course of action. Indomethacin present during the mixed culture of human lymphocytes and human thyroid cells markedly reduced LTS generation, with a half-maximal inhibitory effect at approximately 10ey M. Chromatography of LTS on

Human lymphocytes from normal individuals and from patients with Graves' disease were cultured together with normal human thyroid cells in long term monolayer culture. Stimulation of thyroid cell CAMP content began on the 2nd day of co-culture and rapidly increased to reach a maximum after approximately 72 h. Cell-free culture medium obtained following the mixed culture of lymphocytes and thyroid cells contained thyroid stimulatory activity. Thyroid cells or lymphocytes cultured alone in autologous serum released little or no thyroid stimulating substance into the medium. Surprisingly, lymphocytic thyroid stimulator (LTS) was produced by lymphocytes from all 11 normal subjects tested (stimulation of target cell CAMP content to 271 f 16% SE. of basal values) as well as by lymphocytes from all five patients with Graves' disease (324 f 44% S.E. of basal values). Human thyroid cells were not found to be specific activators of LTS production, in that such activation was observed with other human cell lines. In some, but not all, experiments, LTS stimulated CAMP content in target cells other than human thyroid, such as human fibroblast and dog thyroid cells. LTS was not precipitated by anti-human immunoglobulin antisera and was dialyzable and heat-stable. LTS bioactivity was extracted at pH 3.0 by 2:l chloroform:methanol and diethyl ether. Gel filtration on Sephadex G-25 and Sephadex G-10 showed that the molecular weight of LTS was 1000 or less, and LTS co-eluted with radiolabeled prostaglandins.
On paper chromatography, LTS co-migrated together with prostaglandins, particularly of the E and F category. silicic acid columns provided further evidence for the prostaglandin nature of LTS and in addition, suggested that LTS is a combination of prostaglandin E and prostaglandin F. Since prostaglandins stimulate thyroid hormone secretion, it is possible that the hyperthyroidism of Graves' disease may be related to the local production of these substances by lymphocytes within the thyroid gland.
There is considerable evidence, accumulated over a period of nearly 20 years, that sera from many patients with the hyperthyroidism of Graves' disease contain thyroid-stimulating immunoglobulins (l-6). Because of the clinical association of Graves' disease with other autoimmune diseases (31, the lymphoid hyperplasia associated with Graves' disease (71, and the presence of thyroid stimulatory antibodies, there has been increasing interest in the role of the lymphocyte in Graves' disease. Several studies have demonstrated the production of a thyroid stimulator by lymphocytes from patients with Graves' disease cultured in vitro (8)(9)(10)(11)(12)(13). Neutralization of this stimulator-y activity by anti-human immunoglobulin antiserum further strengthened the concept of the abnormal production of a thyroid stimulatory immunoglobulin in Graves' disease (9,10,12).
The present consensus is that lymphocytes from normal subjects, in contrast to lymphocytes from patients with Graves' disease, do not elaborate a thyroid stimulator (8-10, 12, 13). Further evidence suggests that normal human thyroid tissue, but not other human tissues, activates lymphocytes to produce a thyroid stimulator (12). This reported difference between control lymphocytes and Graves' lymphocytes in their ability to produce a thyroid stimulator, in response to normal thyroid tissue antigen or PHA,' supports the observation of similar differences with respect to the generation of the lymphokines, migration inhibition factor (14-16), and cytotoxic factor (17). These studies suggest that patients with Graves' disease have an abnormal lymphocyte population, '  After centrifugation, the ether and chloroform phases were evaporated to dryness and the aqueous phases were lyophilized.
All fractions were then resuspended in medium containing 0.5 mM 3-isobutyl-l-methyl xanthine and measured for LTS bioactivity as described above.

Lymphocytic
Non-immunoglobulin Thyroid Stimulator complement had been destroyed during the heating procedure. It was subsequently found however that no significant difference in thyroid stimulator production was observed whether heated or unheated serum was present in the culture medium (Table I). Final proof for the lymphocytic origin of the thyroid stimulator was provided by the observation that the incubation of human lymphocytes together with a nonviable, thyroid cell particulate fraction did result in thyroid stimulator production (Table II). On this basis, the term lymphocytic thyroid stimulator was chosen to describe this factor.
Comparison of Time Course of Stimulatory Effect of TSH and LTS -Human thyroid cells were incubated for up to 2 h in medium containing either added TSH (200 microunits/ml) or LTS previously produced by the mixed culture of human lymphocytes and human thyroid cells for 72 h. The TSH was added to control medium incubated for the same period without either lymphocytes or thyroid cells. The concentration of TSH was chosen to produce a response of similar magnitude to that of LTS. Both TSH and LTS produced a maximal CAMP response at 10 to 15 min (Fig. 3). However, in contrast to the slow decline from maximum seen with TSH stimulation, Human lymphocytes from a normal subject were cultured for 3 days in the presence or absence of thyroid cells. The medium was then aspirated, centrifuged to remove cellular elements, and added to fresh dishes of target thyroid cells to measure thyroid cell CAMP stimulatory activity, as described under "Materials and Methods." 3-day co-culture  there was an initial rapid decline in CAMP values observed with LTS.
Comparison of Control Lymphocytes and Graves' Lymphocytes in Their Ability to Generate LTS-In separate experiments, lymphocytes from 11 normal individuals and 5 patients with active Graves' disease were incubated with normal human thyroid cells for approximately 72 h in medium containing 10% autologous serum. LTS generation was observed with all control lymphocytes (271 +-16% of basal CAMP values) as well as with all lymphocytes from patients with Graves' disease (324 + 44% S.E. of basal CAMP values) (Fig.   4). These values, as determined by the unpaired t test, are not significantly different from one another (P > 0.2). Lymphocytes from both groups cultured without thyroid cells did not produce LTS. As a further control, the incubation of thyroid cells without lymphocytes for 72 h did not result in thyroid stimulating activity in the medium (data not shown).
Ability of Different Cell Lines to Stimulate LTS Production by Normal Human Lymphocytes -Lymphocytes from subjects without known thyroid disease were incubated for 66 to 70 h in medium containing 10% autologous serum, either alone or in mixed culture, with a variety of cells of different species and tissue origin. The cell-free media were then tested for their ability to increase CAMP content in target human thyroid cells in monolayer culture (Table III). The two animal cell lines tested, rat hepatoma (HTC) and dog thyroid, did not stimulate LTS production. In contrast, human cell lines (WI-38 lung embryo libroblasts and HeLa cells) were potent stimulators of LTS generation. LTS production was also observed in a separate experiment using thyroid capsule fibroblasts (data not shown). The medium obtained after incubating WI-38 fibroblasts without lymphocytes did produce some stimulation of CAMP content in target human thyroid cells, but this stimulation was small relative to that seen when lymphocytes were also present. No such stimulatory effect was observed in separate experiments with medium obtained after a 3-day incubation of human thyroid cells without lymphocytes.
Effect of LTS on CAMP Content in Target Cells from Different Tissues-LTS was produced by incubating human lymphocytes in medium containing 10% autologous serum with normal thyroid cells in monolayer culture for approximately 72 h. After removal of lymphocytes by centrifugation, the lymphocyte-human thyroid medium (L/HT) was tested for its ability to increase the CAMP content in a variety of target cells. Similar results were observed in separate experiments with lymphocytes from patients with Graves' disease (Table  IVA). In contrast to the stimulation of CAMP content in thyroid cells, no stimulation was seen in rat hepatoma cells and HeLa cells. In this representative experiment, only slight stimulation of CAMP content was seen in the WI-38 cells. However, much greater stimulation of WI-38 CAMP content was observed in two of three further experiments, one of which is shown in Table IVB. These data indicate that Cells were cultured for approximately 70 h, with or without lymphocytes from a normal subject, in 10% autologous (with respect to the lymphocytes) serum. The cell-free medium was then added for 15 min to fresh dishes of human thyroid cells, after which thyroid cellular CAMP was extracted and measured. stimulation of CAMP generation by the lymphocytic factor is not specific for the thyroid cell. Clear evidence that human thyroid cells were stimulated by L/HT and that the target cell CAMP responses observed were not those of "contaminating" fibroblasts was provided by the observation of the morphological transformation of polygonal thyroid cells into stellate forms (Fig. 5). A similar morphological effect occurs with TSH stimulation of thyroid cells in monolayer culture (22). Fusiform fibroblasts are not so affected.
Studies on Nature of Thyroid Stimulator Produced by Lymphocytes -Because it is presently accepted that the abnormal thyroid stimulator in Graves' disease is a 7 S immunoglobulin, attempts were made to neutralize LTS with specific anti-human immunoglobulin antisera. Anti-human IgM was tested because of the foregoing observation of LTS production by lymphocytes from normal individuals; i.e. it seemed possible that an IgM, rather than an IgG, would be produced over a 3-day period by lymphocytes not presensitized to thyroid antigen. Anti-human IgA was tested as control. Despite the formation, and removal by centrifugation, of substantial precipitates, no loss of LTS bioactivity was observed, indicating that LTS is not an immunoglobulin (Table V). Measured by radial immunodiffusion, the human IgG content in medium containing LTS was 3.4 mg/dl. After immunoprecipitation, no detectable human IgG was present (limit of sensitivity of the assay, 1 mg/dl).
Further studies indicated that LTS was dialyzable and heat-stable (Table VI), confirming the nonimmunoglobulin nature of the substance. Because CAMP itself is dialyzable and heat-stable, it was important to demonstrate that LTS  retarded on the column, eluting after the B chain of insulin (M,. = 3500) and vitamin Bi2 (M, = 1355). This confirmed the small nature of LTS and suggested a molecular weight for the factor of 1000 or less. Preliminary data indicating that LTS bioactivity was unaffected by prolonged treatment with pronase and was extractable at acid pH with chloroform (29) suggested that LTS was a lipid. The differential extractability of LTS bioactivity by lipid solvents was therefore examined further. At pH 3.0, LTS was extracted by 2:l chloroform:methanol as well as by diethyl ether (Table VII). At pH 7.4, more LTS bioactivity was extracted by diethyl ether than remained in the aqueous phase. These data suggested that LTS is an unsaturated fatty acid.
Since prostaglandins are unsaturated fatty acids which stimulate thyroid cell adenylate cyclase activity (30-32), it seemed possible that LTS might be a prostaglandin.
To test FIG. 6. Sephadex G-25 (fine) gel filtration of LTS. Culture medium obtained after the mixed culture of human lymphocytes and human thyroid cells was lyophilized and then extracted with 95% ethanol. The ethanol extract was evaporated, and the residue was resuspended in 50 mM NH,HCO,, pH 7.7, and applied to the column which was eluted with the same buffer. Fractions were lyophilized, resuspended in Leibovitz-15 medium, pH 7.4, with 0.5 mM 3-isobutyll-methyl xanthine, and tested for their ability to stimulate CAMP generation in target thyroid cells. to cause thyroid cell stimulation was applied to a Sephadex G-10 column. LTS bioactivity eluted in the same fractions as did the radioactive prostaglandin (Fig. 7), indicating that the molecular weight of LTS was similar to that of the prostaglandins.
Further evidence suggesting that LTS is a prostaglandin was obtained using paper chromatography.
After preliminary experiments, this system was chosen over thin layer chromatography because of the practical advantages of being able to apply relatively large volumes of sample and the ease of eluting strips for the subsequent bioassay of LTS. In an acid solvents system (butanol:acetic acid:water:pyridine), LTS bioactivity migrated very close to the solvent front (Fig. 81, as did the bioactivities of a number of different prostaglandins. Of the prostaglandins tested, the migration of PGE, appeared to be closest to that of LTS, however in separate experiments PGF,, behaved similarly (data not shown). As judged by their migration in the same system, a number of other dialyzable substances, either proven or potential thyroid stimulators, were excluded from being LTS. These included ATP, adenosine, CAMP, ADP, histamine, acetylcholine, epineph-  (butanol:ethanol:2 M ammonia, 10:3:10). Paper strips l-cm-wide were eluted and assayed for radioactivity and LTS bioactivity. Bioactivity was measured as described in Fig. 8. rine, and isoproterenol (Fig. 8). In an alkaline solvent system (butanol:ethanol:2 N ammonia), LTS migrated similarly to [3H]PGE, (Fig. 9). The co-migration of LTS and prostaglandins in two separate solvent systems therefore provided further evidence that LTS is a prostaglandin or a closely related substance.
In order to compare further the biological properties of LTS and prostaglandins, the time course of stimulation of thyroid cell CAMP content by LTS and PGE, was examined. Measured simultaneously, the kinetics of LTS and PGE, were similar, with maximal stimulation occurring at 10 to 15-min (Fig. 10).
Since indomethacin is a powerful inhibitor of prostaglandin synthesis, the presence of this agent during the period of mixed culture of human lymphocytes and human thyroid cells would be expected to inhibit the generation of LTS if LTS were indeed a prostaglandin.
This was found to be the case. Thus;indomethacin inhibited LTS with a half-maximal inhibitory effect at approximately 10m9 M (Fig. 11). Total inhibition of LTS generation was achieved at an indomethacin concentration of 10m7 M. These data strongly suggested that LTS is a prostaglandin.
Finally, evidence supporting the prostaglandin-like nature of LTS was obtained by silicic acid chromatography. This  11. Effect of indomethacin on LTS generation. Human lymphocytes and human thyroid cells were co-cultured for 70 h, after which the medium was aspirated and the lymphocytes were removed by centrifugation. After the addition of 3-isobutyl-l-methyl xanthine to a final concentration of 0.5 mM, the cell-free medium was assayed for its ability to stimulate CAMP generation in target thyroid cells.  (Table VIII). Although PGA and PGB were previously shown to have biological activity in our system (Fig. 8), no LTS bioactivity was present in the first fraction corresponding to PGA and PGB.

DISCUSSION
The present study describes the production of a thermostable, dialyzable stimulator of thyroid cell CAMP content as a result of the co-culture of peripheral human lymphocytes and human thyroid cells in monolayer.
It is shown that the lymphocyte is the site of production of the stimulator. Thus, (a) thyroid cells cultured alone do not produce the stimulator, (b) heat-inactivated lymphocytes co-cultured with living thyroid cells do not generate the stimulator, and (c) viable lymphocytes co-cultured with a nonviable thyroid cell particulate fraction do produce the stimulator.
We have therefore termed this factor lymphocyte thyroid stimulator (LTS). Strong evidence is provided that LTS is a mixture of prostaglandins E and F, primarily the former.

Lymphocytic
Non-immunoglobulin Thyroid Stimulator 639 Prostaglandins are ubiquitous and their production has been described in many tissues. With regard to lymphoid cells, Ferraris and DeRubertis (33) have demonstrated that stimulation of mouse spleen cells by staphylococcal enterotoxin B and mitogens results in prostaglandin E generation. Similarly, the intravenous injection of sheep erythrocytes into mice is followed by PGF,, synthesis within the spleen (34). To our knowledge, however, the present report is the first to indicate that antigenic stimulation can induce prostaglandin production by human peripheral lymphocytes.
There is considerable evidence that prostaglandins play a role in modulating lymphocytic function. Thus, prostaglandins have been shown to suppress the lymphocytic response to antigen stimulation with the subsequent inhibition of lymphocyte-mediated cytotoxicity (34)(35)(36)(37)(38). In addition to this inhibitory effect, prostaglandins may also be important mediators of the inflammatory process (39). Our data, together with those of others (33,341, suggest that prostaglandins may, in part, mediate the effects of lymphocytes on target cells in the cell-mediated immunity process. The present findings may therefore prove to be of significance in a variety of immune and inflammatory processes such as rheumatoid arthritis. Our data differ in a number of respects from previous data obtained following the exposure of human lymphocytes to thyroid tissue antigen (12). Thus, in the present study, (a) normal lymphocytes as well as lymphocytes from patients with Graves' disease released a thyroid stimulating factor (LTS) into the culture medium and (b) the antigen necessary to induce the production of LTS was present not only in human thyroid cells but also in a number of other human cells. It must be emphasized, however, that these data cannot readily be compared because the stimulatory factor described by others has been shown to be an immunoglobulin (9, 10, 12), whereas the stimulatory substance demonstrated in our system is a prostaglandin.
Although our studies were directed at investigating the pathogenesis of Graves' disease, it is unclear from the present data whether LTS plays a role in the hyperthyroidism of Graves' disease, and if so, how this role is related to the thyroid stimulating antibodies which are of fundamental importance in the 'disease (40). Williams et al. (41) have suggested the possibility of a nonimmunoglobulin thyroid stimulator in Graves' disease on the basis of studies in which dexamethasone produced a clinical improvement in hyperthyroidism faster than could be explained by the disappearance from the circulation of stimulatory immunoglobulins.
A direct effect of glucocorticoids on the thyroid is unlikely in view of the evidence that glucocorticoids do not impair the thyroid response to TSH stimulation (41-45). Since certain prostaglandins are well known stimulators of thyroid tissue in vitro (30-32), and of thyroid hormone secretion in uiuo (461, and since lymphocytic filtration is a hi&pathological feature of the thyroid in Graves' disease (471, one consideration is that prostaglandins, produced locally within the thyroid gland, may play a role in the hyperthyroidism of Graves' disease. Prostaglandins may produce local vascular dilatation, and it is therefore also possible that these agents may be responsible for the hypervascularity of the thyroid gland which is a feature of Graves' disease. Our observation of an in vitro lack of difference between control lymphocytes and Graves' lymphocytes in their ability to produce LTS may be evidence against a role for LTS in Graves' disease. On the other hand, there is increasing evidence that lymphocytes from healthy individuals have the potential to respond to self-antigens and that the expression of autoimmune disease is prevented by both cellular and humoral mechanisms of lymphocyte suppression (48-50). Further, there is evidence that, in uitro, the normal immunological checks and balances that exist in uiuo may be disturbed, with the consequent loss of tolerance of normal lymphocytes for normal tissue antigen (51, 52). Another alternative is that cells in tissue culture may expose normally cryptic tissue antigens which are then recognized by the lymphocytes (53). Finally, it is possible that in our in vitro system humoral suppressive factors may have been diluted in that 10% autologous serum was used.
A second line of evidence against a role for LTS in the hyperthyroidism of Graves' disease may be the fact that LTS stimulated human fibroblast CAMP accumulation in some of the present experiments and is therefore not a unique thyroid stimulator.
However, a lack of thyroid specificity does not exclude a pathogenetic role for prostaglandins in Graves' disease. Thus, since the thyroid is infiltrated with lymphocytes in this disease, thyroid stimulation may be produced by virtue of a high local concentration within the gland of a nonspecific stimulator, i.e. the specificity of the stimulation may be determined by the site of lymphocytic infiltration rather than by the specificity of the stimulatory substance.
Third, LTS may not be of importance in the pathogenesis of the hyperthyroidism of Graves' disease because the cellular antigen responsible for the induction of LTS was not unique to human thyroid cells but was also common to a variety of human cells. It is of interest that previous studies have demonstrated that lymphocytes cultured together with retroorbital flbroblasts stimulate mucopolysaccharide synthesis by the latter, a process mimicked by CAMP and abolished by corticosteroids (54,55). In addition, there are recent data demonstrating that PGE, stimulates fibroblast mucopolysaccharide production (56). These results obtained in a different system are therefore quite similar to ours and raise the possibility that we are observing the same phenomenon. Since Werner et al. (51) have demonstrated by immunofluorescent techniques that the immune reaction in Graves' disease appears to be limited to the thyroid tissue stroma and follicular basement membrane, it cannot be excluded that the inflltration of thyroid tissue by lymphocytes in Graves' disease represents an immune response directed against thyroid connective tissue cells rather than the thyroid cells themselves, and the thyroid cells are stimulated as innocent bystanders. These data may in the future also provide a clue as to the relationship between the thyroidal and extrathyroidal manifestations of Graves' disease in that these may all involve the same pathogenetic mechanism (prostaglandin generation) with the different clinical features depending upon the site and extent of lymphocytic infiltration.
For the present, however, any relationship between LTS and Graves' disease remains tenuous and speculative.