Int. J. Immunopharmac., Vol. 1, pp. 273-284 Pergamon Press Ltd. 1979. Printed in Great Britain.

PRESENCE OF THYMOSIN-LIKE FACTORS IN HUMAN THYMIC EPITHELIUM CONDITIONED MEDIUM

LOUIS KATER,* ROBERT OOSTEROM,* JOHN McCLURE[" and ALLAN L. GOLDSTEINt

*Division of Immunopathology, Departments of Internal Medicine and Pathology, University Hospital Utrecht, The Netherlands and tDepartment of Biochemistry, The George Washington University, School of Medicine and Health

Sciences, Washington, D.C. 20037, U.S.A.

(Received 17 July 1979 and in final form 29 August 1979)

Abstract--The objective of the present study was to determine whether or not thymosin and conditioned medium from human thymus epithelial cultures (HTECM) contain Similar fractions, capable of inducing T-cell differentiation. Therefore we tested the ability of rabbit antisera to different thymosin fractions (thymosin fractions 5, 6 and al) of both bovine and human origin to block the stimulatory effect of HTECM on Con A and PHA response of mouse thymocytes. We also looked for reactivity of the antisera toward tissues of man, mouse and calf, and to tissue cultures of man and mouse. Anti-thymosin fraction 5 and 6, but not anti-thymosin-a 1, were found to inhibit the stimulatory effect of HTECM on both Con A and PHA responses. This cannot be attributed to cytostatic or cytotoxic effects of the antisera on thymocytes. No effect was seen when using antiserum to kidney fraction 5 or normal rabbit serum.

In both tissue sections and tissue cultures of the different species tested, anti-thymosin reacts only with thymus epithelial cells. The reactivity is blocked by neutralizing the antisera with thymosin fraction 5 but not by kidney fraction 5. The reactivity is also strongly diminished by neutralizing the antisera with HTECM but not with supernatants of conditioned media from non-thymic tissues.

The observations are highly suggestive for the presence of similar fractions in HTECM and thymosin, and that both are secreted by thymus epithelial cells.

A number of laboratories have described partially or fully purified thymic humoral factors which can replace thymus function in inducing T-lymphocyte maturat ion (van Bekkum, 1975; Low & Goldstein, 1978). These factors, isolated from thymus tissue, are thought to be secreted by thymic epithelial cells, although the actual site of production of these fac- tors has not been definitely established. These studies suggested that thymic humoral factors should be able to be produced in vitro by culturing thymic epithelial cells. The development of such in vitro systems may also solve the question of whether or not humoral factors isolated f rom thymic tissue are actually of epithelial or lymphocytic origin. Supernatants o f human thymic epithelium have proved to be capable of inducing E-rosette forming cells in fractionated bone marrow cells (Pyke & Gelfand, 1974; Willis- Carr, Ochs & Wedgewood, 1978). In another study, when spleen cells of adult thymectomized mice were incubated within a millipore chamber on human thymic monolayers, theta conversion of these cells was obtained (Papiernik, Nabarra & Bach, 1975). Similarly, the response of rat thymocytes to mito- genic and allogeneic stimulation could be increased with rat thymic epithelial culture supernatants (Kruisbeek, KrOse & Z~lstra, 1977; Kruisbeek,

Astaldi, G. C. B., Blankwater, Z~'lstra, Levert & Astaldi, A., 1978).

We have previously studied the effects of human thymic epithelial condit ioned medium (HTECM) on the mitogen responses of human and mouse lympho- cytes and the allogeneic stimulation of mouse thymo- cytes. H T E C M increased the responsiveness of human and mouse thymocytes to Con A and P H A (as measured by 3H-TdR incorporation) and the allo- geneic stimulation of mouse thymocytes while other conditioned media f rom non-thymic epithelial cul- tures as well as from non-epithelial origin, failed to do so (Oosterom, R., Kater & Oosterom, J . , 1979; Oosterom, R. & Kater, in press). No effect o f H T E C M on mitogen stimulation of human PBL was seen. The P H A response, but not the Con A and LPS responses, of mouse spleen cells was increased by H T E C M . The mitogen response of mouse lymph node cells was not affected by H T E C M . From these data the conclusion was drawn that the effect of H T E C M is not species specific and that the thymus, and to a lesser extent the spleen, contain cell popula- tions sensitive to H T E C M which are apparently not B-lymphocytes.

An unresolved question from the studies is to what extent thymic factors obtained either as whole thymic

273

274

extracts or as thymus epithelial conditioned medium contain identical fractions which are capable of in- ducing T-cell differentiation.

Ultrastructural and biochemical studies suggest that the source of the factor(s) are secretory granules in the cytoplasm of thymic epithelial cells (Clark, 1973) but also from other thymus components (Mandi& Giant, 1973). Antibodies to thymosin have been demonstrated by immunofluorescence to bind specifically to the vacuoles of thymic epithelial cells (Mandi& Giant, 1973; Teodorczyk & Potworowski, 1975; Potworowski, 1977; Oosterom, de Nooy-Van Heel, Kater, Goldstein & Pyke, 1976).

Fractionated bone marrow cells incubated with HTECM gave a significant increase in E-rosette forming cells (Pyke & Gelfand, 1974). This increase did not occur when cells were incubated in a mixture of both HTECM and anti-thymosin 5. Incubation with both components in different steps resulted in an increase of E-rosette forming cells (Kater & Pyke, unpublished observation).

In the present study, we measured the influence of antisera to thymosin fraction 5, thymosin fraction 6, and thymosin-al on the stimulatory effect of HTECM on the mitogen responses of mouse thymo- cytes. We also tested a variety of tissues and tissue cultures and lymphocytes of different species for re- activity with antisera to thymosin fractions.

EXPERIMENTAL PROCEDURES

Preparation o f antibodies to thymosin fractions and controls

The antisera to bovine and human materials were prepared in adult male and female New Zealand white rabbits. Generally, less refined protein fractions (fraction 5 preparations of thymus and kidney) were administered in large doses, 5 mg for each ad- ministration, in an uncoupled form. More refined peptide samples (thymosin fraction 6 and thymosin- al), coupled to immunogenic carrier proteins, were administered in small doses of 0.2-1.6 mg for each administration. The immune serum was passed through Millipore filters (0.45/am) and stored in the presence of sodium azide 0.05% (w/v) at - 20°C.

The preparation of antisera to fraction 5 of bovine thymus and kidney and human thymus was initiated by primary immunizing doses of 5 mg uncoupled protein in 2 ml of an emulsion (equal volumes of sterile saline and Freund's Complete Adjuvant (FCA) administered intradermally at 20-30 sites on the shaved backs of each rabbit). Boosting doses of 5 mg protein and FCA were administered subcutan- eously in the neck region once each week for 5 weeks.

L. KATER et aL

Maintenance doses of 5 mg protein in FCA/saline emulsion plus 0.5 cc Pertussis Vaccine (V-1034, Eli Lilly Co.) were given to each rabbit once each month for 5 months. Bleeding of the rabbits was done 10-12 days following each boosting dose of the antigen.

The antiserum to thymosin fraction 6 (G 95 L) was obtained from a rabbit which had been boosted for 2 years. The primary imm0nization with 5 mg antigen protein (uncoupled) and FCA, as well as all boosts, was administered subcutaneously in the neck region of each of six rabbits. Boosting doses of 5 mg thymo- sin fraction 6 were given every 2 weeks for 9 months (total of 95 mg protein administered to each rabbit). A rest period of 3 months was followed by initiation of boosts with 1 mg doses of fraction 6 coupled to hemocyanin (coupling by means of carbodiimides). Maintenance boosting was continued every 6 weeks for 1 year.

The purified peptide used as the antigen for the production of anti-thymosin-a s was obtained from bovine thymosin fraction 5 (Hooper, McDaniel, Thurman, Cohen, Schulof & Goldstein, 1975; Gold- stein et al., 1977). The antigen was coupled to bovine gamma globulin by means of reaction with carbodi- imide. For the primary immunization, preceded by two days with 0.5 cc Pertussis vaccine, 1.6 mg anti- gen, in coupled form, and 5 mg heat-killed M. tuber- culosis (H37RA Difco Laboratories, Detroit, MI) were suspended in 2 ml emulsion of sterile normal saline and FCA. The emulsion was administered sub- cutaneously in the neck region of each rabbit. Boosting doses of 1.5 mg thymosin-a~ in Freund's Incomplete Adjuvant were given each week for 6 weeks and then maintained every 6 weeks for 6 months. All antisera except anti-thymosin-a~ showed positivity to the respective antigen material by double immunodiffusion (Ouchterlony). The specificity of each antiserum was investigated by the utilization of direct hemagglutination (HA) and sepharose bead immunofluorescence tests (SIA). In addition, speci- ficity was determined by absorption of each anti- serum with the antigenic starting materials. Control tests were designed by absorption of antisera to thy- mosin fraction 5 and 6 with fraction 5 prepared from different tissues (i.e. kidney, liver, brain), with conditioned tissue culture media from thymus (HTECM), or control tissue cultures, and with sus- pensions of thoroughly washed human, calf, and mouse thymocytes.

Thymic epithelial and control tissue cultures

Human thymic epithelial and control tissue cul- tures. Human thymuses were obtained from children undergoing cardiac surgery. Thymic epithelial cells were cultured in HEPES buffered RPMI 1640

Presence of Thymosin-like Factors in Human

(H-RPMI; GIBCO Biocult Ltd., Scotland) supple- mented with 2 mM L-glutamine, penicillin (100 tag/ml), streptomycin (100 tag/ml) and 2007o heat in- activated human AB serum. The detailed method for primary culturing of thymic epithelial cells from explants have been described (Oosterom et al., 1979). Human AB serum was used in the culture medium in order to circumvent the possibility that the antisera, directed toward fractions of calf thymus, would react with foetal calf serum components. Control cultures used were fibroblasts grown from a cervical biopsy (HFCM), cervical carcinoma, amnion epithelial cells and epithelial cells grown from sublabial mucosa (HLECM) and salivary glands (HSECM). The latter two epithelial cell cultures were chosen because they share a common ectodermal origin with thymus epi- thelium (Hamilton, Boyd & Mossman, 1972).

Mouse thymic epithelial cultures and control cul- tures. The experiments were performed with 6 -8 week old Swiss inbred mice killed by cervical disloca- tion. The thymuses were removed aseptically and dis- sected into small fragments after removal of capsular tissue. The tissue cultures were prepared in the same way as described for human tissues. Control cultures used were fibroblasts grown from a skin biopsy and 3T3 embryo fibroblasts, and epithelial cells grown from esophagus, mammary glands and skin.

Collecting o f conditioned media

Conditioned media from the human thymic epi- thelial cultures were collected from day 10 when thymocytes had disappeared from the cultures, until day 35 when fibroblasts started to overgrow the epi- thelial cells. Supernatants from the epithelial cultures of control tissues were collected in the same way. Conditioned media from fibroblasts were collected when the cells had covered more than 50% of the tissue culture flasks. The collected media were centri- fuged for 10 min at 1000g, filtered through Millipore filters (0.45 tam) and stored at - 20°C.

Mitogen stimulation o f mouse lymphocytes

Thymuses from 6-8 week old mice were dissected out under sterile conditions and minced in H-RPMI. The cells were counted in a haemocytometer. Via- bility was assessed by trypan blue dye exclusion: viability was always greater than 95°7o. The cell sus- pensions were adjusted to a 4 x 106/ml in RPMI with b icarbonate , ant ibiot ics and 1007o AB serum. Cultures were set up in round bottom microtiter plates (Dynateck, Ntlrtingen, FRG) and contained 50 tal mitogen solution and 50 tal diluted tissue culture condi t ioned medium or thymosin fract ion 5. Antiserum, non-immune serum, or medium was used to dilute tissue culture medium. In all experiments,

Thymic Epithelium Conditioned Medium 275

the dilution of conditioned medium used was the optimal dilution as had been previously determined and reported (Oosterom, 1979). Cell cultures were incubated for 2 days at 37°C in humidified air with 50/o CO 2. Mitogens used were Con A (79-003, Miles, Slough, UK), 2 tag/culture and P H A (HAl5 , Welcome, Beckenham, UK), 16 tag/culture. After 2 days of incubation, 1 taCi [methyl-3H]thymidine (5 Ci/mmole, Radiochemical Centre Amersham, UK) was added to each well. The cells were collected on Titertek glass-fiber filters, after 7 h, with an automatic culture harvester (Skatron, Lierbyen, Norway). The air-dried filters were placed in scintillation vials and 2.5 ml toluene scintillator was added. Radioactivity was measured with Nuclear Chicago Liquid Scintil lation counter (NC725). Results were expressed as the mean cpm of quad- ruplicate cultures+standard error (S.E.).

Immunohistochemical methods

The antisera were also tested for reactivity toward tissues, tissue cultures, and lymphocytes of different species. Human tissues tested consisted of sections of thymus, liver, kidney, pancreas and spleen, and of cultured thymus, fibroblasts, cervical carcinoma, labial epithelium, and salivary gland epithelium.

Mouse tissue tested included thymus, liver, kidney and spleen. Cultures of mouse thymus, fibroblasts, esophagus epithelial cells, mammillary cells, and skin were also tested.

Calf tissues examined were thymus, liver, kidney, pancreas, brain and spleen. Lymphoid cells used for the experiments were human and mouse thymocytes, human peripheral blood lymphocytes (PBL), mouse spleen cells, isolated as described for mitogen stimulation (Oosterom et aL, 1979).

Small tissue blocks were kept for 60 min in phos- phate buffered saline (PBS) pH 7.4 and subsequently fixed for 4 h in 10% formalin and then kept over- night in a 30070 sucrose solution. After short washing in PBS the tissue blocks were deep frozen in liquid nitrogen and stored at - 7 0 ° C until further proces- sing. The tissue was cut in a cryostat into 4 tam sections, mounted on slides and air dried. The sec- tions were first thoroughly rinsed in order to remove unbound plasma components from the tissue. The sections were then dried and incubated in a moist chamber with the antisera. All tissue sections of the different species were incubated with the various antisera to calf and human thymosin fractions and to calf kidney fraction 5.

The antisera were used in the indirect immuno- fluorescent technique (IF). The fluorescein (FITC) labelled conjugate was a horse anti-rabbit globulin serum (Central Laboratory of the Netherlands Red

276

Cross Blood Transfusion Service (CLB), Amster- dam, The Netherlands). Direct IF using this con- jugate served as a control. In referring to the human and bovine origin of the fractions against which the antisera were raised, human and bovine tissues were also studied in the IF using FITC labelled horse anti- human globulin serum (Central Laboratory of the Netherlands Red Cross Blood Transfusion Service, Amsterdam, The Netherlands) and an unlabelled rabbit-anti-bovine globulin serum (Dr. J. Gouds- waard, Division of Immunology, Veterinary Faculty, Utrecht, The Netherlands).

The same procedure was carried out on tissue cul- tures which were attached to the bottoms of culture flasks and fixed in cold alcohol/acetic acid mixture (95:5) for 10 min at - 20°C . The procedure for in- cubation of human or mouse thymocytes or mouse spleen cells was as follows: 106 cells were incubated with 0.1 ml antiserum diluted in Earle's Balanced Salt Solution with 0.1070 NaN3 for half an hour at 4°C, washed three times, then incubated with fluore- scein labelled antiserum and examined as viable pre- paration.

Blocking tests consisted of neutralizing antisera with fraction 5 of thymus, kidney, liver, brain, with tissue powder extracts, with thymocytes and with HTECM and control conditioned media.

The sections were examined with a Leitz Ortho- plan microscope, equipped with a xenon arc (XBO 75 Watt) for epi-illumination. For the narrow band ex- citation with blue light of long wave length (470-490 nm) we used a double band interference filter (TAL 480): the vertical illuminator (Leitz) was used in position 3, in which position an interfering dividing plate (~ H495), with 50°7o transmittance and 5007o re- flection at 495 nm and a barrier filter (K 495) were placed in the light path. A K 530 filter was employed as an extra barrier filter.

Additionally an immunohistoperoxidase method was used (Burns, Hambridge & Taylor, 1974) for which a peroxidase-labelled goat antiserum to rabbit IgG (Miles Laboratories, Ltd., Slough, UK) and counterstaining with hematoxylin was applied.

RESULTS Human and mouse thymic epithelial cultures

Large rounded cells with some fibroblasts were seen around the explants during the first days of cul- ture. Lymphocytes released by the explants were almost completely removed after two medium changes. At days 4 through 7, the rounded cells dis- appeared and outgrowth of polygonal cells were seen at the edge of the explants. These cells formed con- tinuous sheets with fibroblasts at the periphery of the outgrowth. Beginning with day 14, Hassall's cor-

L. KATER el at.

puscles could be seen at the outer part of the circular fields (Kater, 1973). Electron microscopic examina- tion showed desmosomes and tonofibrils, confirming the epithelial nature of the cells. Fibroblasts could never be completely removed and after 4-6 weeks fibroblasts had overgrown the epithelial cells. Out- growth of mouse thymic'explants showed the same pattern except that Hassall's corpuscles were rarely seen (Kater, 1975).

Control cultures Human fibroblasts grown from a cervical biopsy

were subcultured 17 times. Conditioned fibroblasts medium (HFCM) was used from subcultures 3 and later. Explants of labial mucosae and salivary glands showed outgrowth of epithelial cells from day 4, giving a monolayer after 2-3 weeks. The cultures were almost free of fibroblasts. After about 5 weeks the cells tended to lose contact with the culture flask.

In cultures from mouse esophagus, mammary glands, and skin, polygonal cells resembling epithelial cells were seen, and fibroblasts were present at all times.

Influence o f antibodies to thymosin fractions 5, 6 and ~ on the stimulatory effect o f conditioned medium on mitogen response o f mouse thymocytes

Mouse thymocytes cultured with mitogen and HTECM gave significantly higher 3H-TdR incor- poration than mitogen-stimulated cultures to which control medium was added, as described previously (P< 0.0025) (Oosterom el al., 1979).

The influence of anti-thymosin fraction 6 is shown in Fig. 1. A clear inhibition of the HTECM effect on Con A and PHA response was seen when anti- thymosin fraction 6 diluted 1:30 was used in the cultures, the effect being less pronounced with higher dilutions of the antiserum (Fig. 2). However, a smaller effect of anti-thymosin fraction 6 on mitogen responses in the controls was also found occasionally as demonstrated in the PHA effect (Fig. 1). The results observed with anti-thymosin fraction 6 and anti-thymosin fraction 5 were similar.

As indicated in Table 1 anti-thymosin-a~ shows no reactivity on the stimulatory effect of HTECM as measured in the Con A and PHA response. Similarly, antiserum to bovine kidney fraction 5 had no influ- ence in the HTCEM stimulatory effect on mitogen responses (Table 2). Normal rabbit serum had no effect on the stimulatory capacity of HTECM.

Reactivity o f antibodies to thymosin fractions with tissue components

Tissue cultures. With indirect immunofluorescence and immunoperoxidase techniques using rabbit anti- sera to bovine thymosin fraction 5, thymosin fraction

Presence of Thymosin-like Factors in Human Thymic Epithelium Conditioned Medium

6, and thymosin-a~, and to human thymosin fraction 5, reactivity was found only in the cytoplasm of normal human epithelial tissue culture cells in a fine granular pattern (Fig. 3) and human thymoma (epi- thelial) cells (Fig. 4). No staining was observed in other cells including fibroblasts, thymocytes or cells

277

of control tissue cultures of man (amnion epithelial cells, cervical carcinoma, sublabial mucosa and sali- vary glands) and mouse (skin fibroblasts, 3T3 mouse embryo fibroblasts, esophagus epithelial cells, mam- miilary cells and skin). The reactivity was not species specific in that similar results were found in tissue

100--

cpm x 10 = cpm x 10 =

....T..

80-

60-

40-

20-

Con A 30-

HTECM HLECM RPMI

i

25-

20-

15-

10-

5-

,X~ ANTI-THYMOSIN FR. 6 1:30

PHA

HTECM HLECM RPMI

Fig. 1. Inhibition by anti-thymosin fraction 6 of the effect of HTECM on mitogen stimulated mouse thymocytes. HTECM--human thymic epithelial conditioned medium diluted 1:18 HLECM--human labial epithelial conditioned

medium diluted 1 : 18.

c p m x 10 ~ cpmx 103

100- Con A 35-

80-

60-

) / ~ 30-

25-

20-

15-

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . t 10

5-

PHA

J¢ ~t z / I I I I I I I I | I

480 240 120 60 30 480 240 120 60 30

RECIPROCAL DILUTION OF ANTI-THYMOSIN FR. 6

Fig. 2. Effect of dilutions of anti-bovine thymosin 6 on the effect of HTECM • ~ or HLECM O--- <3 to mitogen stimulated mouse thymocytes measured by 3H-TdR incorporation. The figure shows the influence of various dilutions of

the antiserum.

278 L. KATER e t al.

Table 1. Lack of influence of rabbit anti-thymosin-a I on the effect of HTECM on mitogen stimulated mouse thymocytes

Conditioned medium Antiserum Con A PHA

HTECM* 10989±760§ 2.07 II 2392±115 4.18 HTECM + Ra Thymosin-t~ t ~: i1593_+1242 2.39 2136-+203 3.93 HTECM + NRS 12156--+1221 2.40 1913_+179 4.40

HLECM t 5574±423 1.04 666±177 1.16 HLECM + Ra Thymosin-a I 5415+_287 1.12 740-+26 1.36 HLECM + NRS 5654±695 1.11 670±164 1.54

RPMI 5309-+712 572±64 RPMI + Ra Thymosin-a I 4831±611 543±80 RPMI + NRS 5064± 1035 434±76

* Human thymic epithelial conditioned medium diluted 1 : 18. t Human labial epithelial conditioned medium diluted 1 : t8. :~ Rabbit anti-bovine thymosin-a I 1:30. § Mean cpm~S.E. I! Factor by which the mitogen response is increased as compared with the corresponding cultures without conditioned medium.

Table 2. Lack of influence of anti-bovine kidney 5 on the effect of HTECM on mitogen stimulated mouse thymocytes

Conditioned medium Antiserum Con A PHA

HTECM* 41940±3772§ 2.34 II 11328±845 4.50 HTECM + Ra kidney 5~ 38639±849 2.33 12269±557 4.39 HTECM + NRS 34136-±2117 2.01 11558±965 3.95

H L E C M t 21908±1628 1.22 6851±415 2.72 HLECM + Ra kidney 5 17765± 1484 1.07 7791±372 2.79 HLECM + NRS 18766--+1109 1.10 9142_+364 3.12

RPMI 17858± 545 251 ~ 2 ! 2 RPM! + Ra kidney 5 16573± i 106 2792±349 RPMI + NRS 16907±606 2921 ±295

* Human thymic epithelial conditioned medium diluted 1 : 18. t Human labial epithelial conditioned medium diluted 1 : 18. :~ Rabbit anti-bovine kidney fraction 5 diluted 1:30. § Mean cpm±S.E. II Factor by which the mitogen response is increased as compared with the corresponding cultures without conditioned medium.

Fig. 3. lmmunohis tochemical reactivity of anti-bovine thymosin fraction 6 to human thymus epithelial tissue culture (arrows indicate well-defined epithelial cells with intense cytoplasmic fluorescence due to presence of ant i- thymosin anti-

bodies). Indirect immunofluorescence technique x 220.

Fig. 4. Immunohis tochemical reactivity of ant i- thymosin fraction 6 to cultured human thymoma cells (epithelial cells). Indirect immunofluorescence technique x 540.

279

Fig. 5. Immunohis tochemical reactivity of rabbit an t i -human thymosin fraction 5 on human thymus tissue. Staining is seen in the cytoplasm of epitheloid cells (some indicated by arrows) of the thymic medulla, lmmunoperoxidase technique

counterstained with hematoxylin x 250.

Fig. 6. lmmunohis tochemical reactivity of rabbit anti-bovine thymosin 6 after neutralization with bovine thymosin 5 on human thymus tissue. No staining is seen. lmmunoperoxidase technique counterstained with hematoxylin × 250.

280

Fig. 7. Immunohis tochemical reactivity of rabbit anti-bovine thymosin 6 after neutralization with bovine kidney fraction 5 on human thymus tissue. Staining is still seen in epithelial cells (see arrows) and Hassall 's body indicating specificity of

ant i- thymosin 6. Immunoperoxidase technique counterstained with hematoxylin x 250.

281

Presence of Thymosin-like Factors in Human

cultures of man and mouse or when using antisera to bovine or human thymosin fractions.

Absorption of anti-bovine thymosin fraction 5 and thymosin fraction 6 with bovine thymosin fraction 5 resulted in completely negative staining. After absorption of these antisera with bovine kidney 5, some staining remained, although in decreased in- tensity. Neutralization reactions with HTECM re- sulted in decrease of staining while blocking tests with conditioned media of control tissue cultures did not influence the reactions. No reactivity was found using anti-thymosin fraction 5, 6 or al toward iso- lated human or mouse thymocytes, human PBL and mouse lymphoid cells.

Tissue sections. The antisera were also tested on frozen sections of a variety of tissues from man, mouse and calf both by immunofluorescence and immunoperoxidase technique. Staining was seen in epithelial cells of both cortex and medulla of the thymus (Fig. 5). In the cortex isolated cells were found to react, while most reacting cells were seen in the medulla, often in close connection to each other. Some Hassall's corpuscles showed some weak posi- tivity in their outer rims. The pattern was finely granular and confined to the cytoplasm of epithelial cells. Nuclei and lymphocytes as well as intercellular spaces were negative.

Unabsorbed antisera to thymosin fractions, how- ever, also reacted with liver, kidney, spleen, brain, and pancreas, while antibody to kidney fraction 5 reacted with thymus. This was in accordance with the results in HA and SIA. After absorption of anti- thymosin fraction 5 and fraction 6 with thymosin fraction 5 staining was abolished in all tissues tested (Fig. 6). After absorption of the serum with kidney fraction 5 staining was left in the thymus, although in decreased intensity, but not in control sections of other tissues (Fig. 7). The reactivity was not species specific since similar patterns of reactivity could be found on cryostat sections of thymus in man, mouse and calf by using rabbit anti-bovine thymosin or rabbit anti-human thymosin. The intensity of the staining in using the anti-bovine fractions was strongest on bovine thymus and that of anti-human thymosin fraction 5 on human thymus tissue. This is in accordance with the observations using HA and SIA. The intensity of staining with anti-thymosin-a Z was weak in comparison with anti-thymosin fraction 5 and 6, and no reactivity of Hassall's corpuscles could be found.

DISCUSSION

In studying the effects of human thymus epithelial conditioned medium (HTECM) on mitogen res-

Thymic Epithelium Conditioned Medium 283

ponses of human and mouse lymphocytes, we have concluded that HTECM increases the responsiveness of human and mouse thymocytes. Control condi- tioned media did not show any significant effect on human thymocytes but a slight effect is seen on mouse thymocytes (Oosterom et al., 1979). The phenomenon may be due to a xenogeneic effect, possibly due to shedding of cell surface antigens.

Macrophages probably do not contribute to the stimulatory effect of HTECM on thymocytes be- cause macrophages were not evident in the cultures and would contaminate equally cultures of the thymus and the control tissues which were examined. Anti-thymosin fraction 5 and 6 had a clear inhibitory influence on the effect of HTECM and in the PHA and Con A responses of mouse thymocytes. Anti- thymosin a 1 did not inhibit in these experiments. This observation may fit with the lack of effect of thymo- sin a~ observed on in vitro proliferative responses (Low & Goldstein, 1978).

lmmunohistochemical studies using anti-thymosin fraction 6 on tissue cultures showed only localization in thymus epithelial cells. Electron microscopy con- firmed the epithelial nature of the cultured cells, while macrophages were not found in routine exam- ination. In frozen sections using anti-thymosin frac- tion 6 absorbed with kidney fraction 5 staining of the thymus epithelial cells was still observed, while staining of the thymus by anti-thymosin fraction 6 was completely neutralized by thymosin fraction 5. Staining by anti-thymosin fraction 5 was strongly diminished by HTECM, but absorption by control conditioned media did not influence the staining.

The observations reported here show that antisera directed to thymosin fraction 5 and thymosin frac- tion 6 can inhibit the stimulatory effect of HTECM on the mitogen responses of thymocytes. The anti- sera show reactivity with thymus epithelial cells. These observations support the view that the bio- logical activity present in the supernatant is secreted by thymus epithelial cells. The data presented here also are highly suggestive for the presence of similar peptide components in both HTECM and thymosin.

Acknowledgements--These studies were supported in part by the Hartford Foundation, Inc., The National Cancer Institute grant number NIH CA 25017, and Hoffman- LaRoche, Inc.

We are indebted to Jonet Oosterom, Felie Schmitz du Moulin, Renan Atasoy, and Jan Geertzema for their excel- lent technical assistance, to Mr. E. E. W. Dumernit for preparation of the photographs, and to Barbara Smith for typing the manuscript. We would also like to express our appreciation to the cardio-surgery teams of the Antonius Hospital and the University Children Hospital in Utrecht for their attention in providing us with thymus tissue. The

284 L. KATER et al.

electron microscopy studies were performed by Dr. L. The expert technical assistance of Mr. James Oliver and Rademakers. Rabbit anti-bovine globulin serum was kindly Ms. Janelle Hatcher in the preparation of anti-thymosin provided by Dr. J. Goudswaard (Div. Immunology, antisera is acknowledged. Veterinary Faculty, Utrecht, The Netherlands).

REFERENCES

BURNS, I., HAMaRIDGE, M. & TAYLOR, C. R. (1974). Intracellular immunoglobulins. J. clin. Pathol. 27, 548-557. CLARK, S. L. (1973). The intrathymic environment. In Contemporary Topics in lmmunobiology, Vol. 2 (eds. Davies, A.

J. S. & Carter, R. L.) pp. 77-99. Plenum Press, New York. GOLDSTEIN, A. L., LOW, T. L. K., McADOO, M., McCLURE, J., THURMAN, G. B., ROSSIO, J., LAI, C-Y., CHEN(;, D.,

WANG, S-S., HARVEY, C., RAMEL, A. H. & MEIENHOFER, J. (1977). Thymosin-al: isolation and sequence analysis of an immunologically active thymic polypeptide. Proc. hath. Acad. Sci. U.S.A. 74, 725 729.

HAMILTON, W. J., BOYD, J. D. & MOSSMAN, H. W. (1972). In Human Embryology: Prenatal Development o f Form and Function, p. 162, Heifer, Cambridge.

HOOPER, J. A., McDANIEt., M. C., THURMAN, G. B., COHEN, G. H., SCHULOF, R. S. & GOLDSTEIN, A. L. (1975). Purifica- tion and properties of Bovine thymosin. Ann. N. Y. Acad. Sci. 249, 125-144.

KATER, L. (1973). A note on Hassall's Corpuscles. In Contemporary Topics in Immunobiology, Vol. 2 (Eds. DAVIES, A. J. S. and CARTER, R. L.) pp. 101-109, Plenum Press, New York.

KATER, L. (1975). Impairment of the thymus function. In The Biological Activity o f Thymic Hormones, (Ed. VAN BEKKUM, D. W.) p. 61, Kooyker Scientific, Rotterdam.

KRUISBEEK, A. M., KROSE, T. C. J. M. & Z~?I.STRA, J. J. (1977). Increase in T cell mitogen responsiveness in rat thymo- cytes by thymic epithelial culture supernatant. Eur. J. lmmun. 7, 375 381.

KRU~SBEEK, A. M., ASTAI.DI, G. C. B., BLANKWATER, M. J., Z~?LSTRA, J. J., LEVERT, L. A., & ASTAIDI, A. (1978). The in vitro effect of a thymic epithelial culture supernatant on mixed lymphocyte reactivity and intracellular c AMP levels of thymocytes and on antibody production to SRBC by Nu/Nu spleen cells. Cell. lmmun. 35, 13,1-147.

Low, T. L. K. & GOLDSTEIN, A. L. (1978). Structure and function of thymosin and other thymic factors. In The Year of Hematology (eds. Silber, R., Lobue, J. and Gordan, A. S.), Plenum Press, New York.

MANDI, B. & GI.ANT, T. (1973). Thymosin producing cells of the thymus, Nature, New Biol. 246, 25. OOSTEROM, R., DE NOOY-VAN HEEL, A., KATER, L., GOLDSTEIN, A. L. & PYKE, K. W. (1976). Detection of thymosin by

means of in vitro thymus systems. 17th Dutch Federation Meeting, p. 315. OOSTEROM, R., KATER, L. & OOSTEROM, J. (1979). Effects of human thymic epithelial conditioned medium on mitogen

responsiveness of human and mouse lymphocytes. Clin. Immun. and Immunopath. 12,460-470. OOSTEROM, R. & KATER, L. Ann. N. Y. Acad. Sci., in press. PAPIERNIK, M., NABARRA, B. 8z BACH, J. F. (1975), In vitro culture of functional human thymic epithelium. Clin. exp.

lmmun. 19, 281-287. POTWOROWSKI, E. F. (1977). The thymic microenvironment: localization of a biologically active insoluble fraction. Clin.

exp. Immun. 30, 305-308. PYKE, K. W. & GEl FAND, E. W. (1974). Morphological and functional maturation of human thymic epithelium in

culture. Nature, Lond. 251,421. TEODORCZYK, J. & POTWOROWSKI, E. F. (1975). Cellular localization and antigenic species specificity of thymic factors.

Nature, Lond. 158, 617 619. VAN BEKKUM, D. W. (ed.) (1975). The Biological Activity o f Thymic Hormones. Kookyer, Rotterdam. WILI.IS-CARR, J. l., OCHS, H. D. 8~ WEDGEWOOD, R. J. (1978). Induction of T-lymphocyte differentiation by thymic

epithelial cell monolayers. Clin. lmmun, lmmunopath. 10, 315-324.