Glucocorticosteroid Response-Modifying Factors Derived from Accessory Cells

GLUCOCORTICOSTEROID RESPONSE-MODIFYING FACTORS DERIVED FROM ACCESSORY CELLS * Robert I. Mishell, Linda M. Bradley,? Yu-hua Una Chen, Kenneth H. Grabstein,? and Stanley M. Shiigi? Department of Microbiology and Immunology University o f ,California Berkeley, California 94720

INTRODUCTION Glucocorticosteroid hormones exert strong immunosuppressive effects in vivo and in vitro.l-.LTheir mode of action in this respect is complex and poorly understood. Although it is well established that both T-cells and accessory cells (monocytes/macrophages) are targets of ~ t e r o i d s , ~much -~ remains to be learned about the effects of the hormones on specific cell subpopulations that participate in immune reactions. Some of the questions requiring further investigation are to identify the specific types of cells that are inhibited by steroids, to distinguish reversible effects from those that are irreversible, and to determine which cellular lesions occur as a result of direct hormone action and which are secondary to effects on other cells. A few years ago, we and others observed that cultured mouse spleen cells acquire resistance to the immunosuppressive effects of very high doses of steroids.2-" This phenomenon was shown to be due to the activation of accessory cells by bacterial adjuvants. Further experiments demonstrated that protection from steroid suppression occurs, at least in part, as a result of the secretion and action of cell-free mediators (monokines) . l o o l1 We are attempting now to understand in greater detail the cellular and molecular pathways by which adjuvant-activated accessory cells prevent steroid immunosuppression. Such studies may provide new insights concerning the action of steroids on the various subpopulations of cells that interact during the generation of immune responses. This report briefly summarizes our initial studies on acquired resistance to steroid immunosuppression, and presents recent experimental data on the biochemical characteristics of accessory cell factors and on the biology of protection. BACKGROUND The capacity of freshly isolated murine spleen cells to generate in vitro primary humoral immunity is inhibited by the inclusion of dexamethasone in the culture medium at concentrations of M or greater. However, spleen cells that are precultured overnight in the presence of various bacterial adjuvants become resistant to steroid immunosuppression even at high concentrations of *This work was supported by a grant from the Kroc Foundation, and National Institutes of Health Grants CA25056 and AI15482. t Supported by National Institutes of Health Grant CA09179.

433 0077-8923/79/0332-0433 $01.75/0 @ 1979, NYAS

434

Annals New York Academy of Sciences

the hormone (leeM ) . Further studies revealed that freshly isolated spleen cells become immediately resistant to high concentrations of steroids if they are cultured with adjuvant-activated accessory cells from the spleen or the resident peritoneal population." We subsequently found that resident peritoneal accessory cells secrete cell-free mediators which also prevent high-dose steroid immunosuppression when added to cultures of freshly isolated spleen l1 Other experiments showed that following overnight culture in steroids, antigenprimed T-helper cells lose their ability to restore the immune responses of T-depleted spleen cells. Taken together, these studies suggested that steroids suppress helper T-cell function but that mediators produced by adjuvantactivated accessory cells either block this effect or substitute for the helper T-cells in the induction of primary humoral immunity. To distinguish between these alternatives, experiments employing a two-step design were conducted. Briefly, spleen cells from antigen-primed mice were cultured initially in the presence or absence of dexamethasone and accessory cell factors. After two days of culture, the cells were harvested, washed and x-irradiated.$ They were then assayed for T-helper activity by determining their capacity to enable freshly isolated T-depleted spleen cells to generate humoral immunity during a second culture period. Experiments of this design showed that dexamethasone alone completely suppressed helper activity but that substantial T-helper activity was retained if the first step cultures contained accessory cell factors in addition to steroids." Since the factors were not present in the assay for helper activity, these data support the hypothesis that mediators obtained from adjuvant-activated accessory cells block steroid suppression of T-helper cells. For this reason, we have designated the factors as glucocorticosteroid-responsemodifying factors (GRMF) . Our previous experiments have delineated several aspects of the protection phenomenon. However, these experiments do not establish the mechanism of protection nor the cellular site of GRMF action. These and other questions can best be approached with well-defined molecular and cellular reagents. We have begun therefore to investigate the physical chemical properties of GRMF and to study the biology of protection in greater detail, using more homogenous cell populations. GLUCOCORTICOSTERIOD RESPONSE-MODIFYING FACTORS GRMF are defined as substances secreted by adjuvant activated accessory cells which protect mouse spleen cells from the immunosuppressive effects of glucocorticosteroid hormones. Their activity is conveniently detected in a single-step assay by comparing the humoral responses to sheep red blood cells of mouse spleen cells cultured without steroids (positive control), and others cultured with 1 0 - 6 M dexamethasone (steroid-suppressed control) to experimental groups containing dexamethasone and the test materials. The responses are measured with the direct heinolytic plaque forming assay. Positive control responses usually range between 1,000 and 20,000 plaque-forming cells/ culture when 4 X lo6 cells are cultured in 1 ml volumes for 5 days. Typically, steroid$ X-irradiation was used to eliminate potential antibody-forming cells from the precultured antigen-primed (T-helper cell) population. Presumably, suppressor T-cell functions were also reduced by this means."

Mishell et al. : Hormone Response-Modifying Factors

435

containing controls generate fewer than 10 PFC/ culture, although occasionally responses in the range of 100-200 PFC occur. The responses of steroid-treated cultures containing optimal concentrations of GRMF are equivalent to or greater than those of the positive controls. Thus, there is usually a 100-fold or greater difference in the responses of protected and suppressed cultures. GRMF activity is dose-dependent, although usually it is not possible to demonstrate a truly linear dose response by simple titration. GRMF INDUCTION Normally, resident peritoneal cells do not secrete detectable levels of GRMF in vitro unless they are cultured with a bacterial adjuvant. (Occasionally GRMF are secreted by such cells without the addition of adjuvant, presumably reflecting in vivo activation prior to culture.) In most of our experiments, GRMF have been generated by culturing resident peritoneal cells ( 5 x10" cells/ml) for 3 days in medium supplemented with 3% fetal calf serum and 0.1 pglml lipopolysaccharide. Medium conditioned in this manner usually has sufficient GRMF activity to protect cultured mouse spleen cells completely from dexamethasone suppression when it is present at a concentration of 20-30%. The experiments defining the physical chemical properties of GRMF shown below are done with material prepared in this manner. However, we have also tested several other substances of bacterial origin for G R M F induction. This is illustrated by the three experiments shown in TABLE 1. While our experience with inducers other than LPS is limited, these experiments clearly show that substances with very diverse chemical structures are active as inducers of GRMF. PHYSICAL CHEMICAL PROPERTIES OF GRMF Initial characterization studies of GRMF containing conditioned media revealed that the activity is retained during dialysis, is stable to lyophilization, and is stable to thermal treatment of 6 5 ° C for 1 hour. More than 90% of the activity is lost, however, following treatment at 70" C for 1 hour. To further characterize GRMF, we have employed the molecular fractionation flow chart shown in FIGURE 1. A representative experiment analyzing GRMF activity in fractions separated on Sephadex (3-75 is shown in FIGURE 2. Two major peaks of activity corresponding to apparent molecular weights of 13,000 to 16,000 daltons and 23,000 to 28,000 daltons were detected. A similar fractionation experiment was conducted with Bio-Gel P-60. Again, two major peaks of activity were detected, although the apparent molecular weights were greater than those obtained with Sephadex (data not shown). The fractions obtained with the Bio-Gel P-60 separation were also tested for lymphocyte-activating factor (LAF) activity by Dr. Steven Mizel (NIDR, National Institutes of Health). He found two peaks of activity which corresponded with the fractions positive in the GRMF assay. Most of the LAF activity appeared in the smaller molecular weight peak whereas the larger molecular weight peak had more activity in the GRMF assay. The significance of these apparent quantitative discrepancies is unclear and will require further study. To determine whether GRMF activity is dependent on carbohydrate moieties, fractions corresponding to the two peaks of activity from a (3-75 chromato-

436

Annals New York Academy of Sciences TABLE1 INDUCTION OF GRMF BY BACTERIAL SUBSTANCES *

Conditioned Media

Experiment

Experiment

Experiment

393

394

397

Positive controls BG LPS (0.1 f i g ) PG (3 r g ) LTA (50 fig) PPD (10 ag) MDP (50 f i g )

2,970 70 13,000 7,030 1 1,400 3,470 370

2,850 20 5,800 9,000 8,820 5,330 3,590

13,350 360 17,900 20,600 25,200 14,200

(NT)

* GRMF containing accessory cell-conditioned media were generated separately for each of three experiments (393, 394, and 397) according to the following protocol: Normal BALB/k peritoneal cells were depleted of T-lymphocytes by treatment with specific antiserum and complement. These cells were cultured 72 hours at 5 x l @ cells/ml in 4 ml volumes in the presence (or absence, background, BG) of the following bacterial substances. ( 1) LPS: lipopolysaccharide was isolated from Salmonella typhimurium LT2M1 by the method of Galanos et ul.; (2) PG: peptidoglycan was isolated from Staphylococcus uureus 52A2 by lysostaphin digestion and consists of peptidoglycan covalently bound with ribitol teichoic acid; 28 ( 3 ) LTA: lipoteichoic acid was isolated from Streptococcus faecafis by hot-phenol-water extraction, and was kindly provided by Dr. G. V. Shockman; (4) PPD: tuberculin purified protein derivative, lot number 89213, was the generous gift of Dr. A. Gray of Merck, Sharp and Dohme, West Point, PA; and ( 5 ) MDP: muramyl dipeptide or synthetic muramyl-L-alanine-D-isoglutamine, was kindly provided by Dr. L. B. Chedid. After 72 hours, the supernatants were harvested, stored at 4" C and subsequently tested for glucocorticosteroid response-modifying factors (GRMF) . Aliquots (0.3 ml) of each supernatant were added to duplicate cultures containing 4 x 10" BDF, spleen cells, 3 X 10"SRBC, and lo4 M dexamethasone. Results are expressed as direct anti-SRBC PFC/culture on day 5. Positive controls were cultured without dexamethasone and BG controls cultured with dexamethasone in medium conditioned by accessory cells without adjuvants. As an additional control, both LPS and PG were added to separate aliquots of unstimulated conditioned medium (BG) after the medium was harvested. These controls resulted in responses of less than 10% of those which occurred in the respective induced conditioned media. graph were separately pooled, concentrated and each was treated with 0.01 M metaperiodate. The data from this experiment are shown in TABLE 2. These results indicate that G R M F contain carbohydrate structures that are essential for their biological activity. Medium conditioned by adjuvant-activated accessory cells has also been analyzed with D E A E cellulose chromatography. All of the G R M F activity bound to DEAE, and most of it was eluted as a single peak with a linear NaCl gradient (data not shown). In addition to our studies o n medium conditioned by adjuvant-activated accessory cells, we have tested several other cellular supernatants for G R M F activity. A sample provided by Dr. David Wood (Merck Institute, Rahway, N J ) of partially purified B-cell-activating factor (BAF) prepared from cultured human monocytes was very active in the one-step G R M F assay. Similar strong G R M F activity was present in a sample of partially purified L A F prepared by Dr. Steven Mizel from the murine-monocyte cell line, P388D,. G R M F

Mishell et al. : Hormone Response-Modifying Factors

437

activity was also present in supernatants of several other murine monocytic cell lines. Some of the latter have been tested in the two-step assay for protection of T-helper function and were highly active in it as well. These studies on the chemical characterization of GRMF indicate that they are macromolecules, probably glycoproteins on the basis of heat lability and metaperiodate inactivation. Two major molecular species were found in the media conditioned by adjuvant-activated normal murine peritoneal cells. The relationship of the two molecular species is not clear, but both are active in the LAF assay as well as the GRMF assay. Moreover, preparations of LAF and BAF from other sources were active in protecting against steroid immunosuppression. These results are consistent with the hypothesis that the three activities are due to a single family of molecules. Several mediators are produced by adjuvant-activated accessory cells, including human pyrogenic factor,I3 murine macrophage T-replacing factor,l.' LAF,I&and BAF.16 These mediators all are similar in size to the smaller molecular species of G R M F and the HARVEST SUPERNATANTS ROM ADJUVANT ACTIVATED ACCESSORY CELLS; 10 JJG/ML LPS, MEM CONTAINING FCS, (lx CONCENTRATION)

-1

PRECIPITATE, 75%-AMMONIUM SULFATE

DIALYSE PRECIPITANT AGAINST CONCENTRATION)

0,15

SODIUM CHLORIDE,

(lox

L Y O P H I L I Z E AND REDISSOLVE I N WATER,

(1ooX

CONCEN-

TR AT I ON 1

CHROMATOGRAPHJ -SEPHADEX

6-75,

S T E R I L I Z E AND BIOASSAY FRACTIONS,

POOL PEAKS OF A C T I V I T Y AND TREAT POOLED FRACTIONS WITH

.01

METAPERIODATE,

DIALYSE TREATED POOLS,

BIOASSAY OF METAPERIODATE TREATED POOLS,

FIGURE1. Molecular characterization of GRMF. Scheme used for fractionation of accessory cell-conditionedmedium (see text).

438

Annals New York Academy of Sciences

a W

IW

I

FRACTION NUMBER FIGURE2. Sephadex G-75 fractionation of GRMF. 3 ml of 1 0 0 ~concentrated medium conditioned by adjuvant-activated accessory cells was chromatographed on a 2 x 1 5 0 cm column of Sephadex (3-75 superfine (Pharmacia Fine Chemicals, Uppsala, Sweden) in buffered saline (0.15 M NaCI, 0.01 M phosphate, pH 7.4). Three ml fraations were collected and measured for optical density at 278 nanometers (0-0 ) . Glucocorticosteroid response-modifying factor activity of individual fractions was determined in 2 separate experiments. Aliquots of each fraction were added to 4x108 spleen cells cultured with SRBC and 10." M dexamethasone. Cultures were assayed on day 5. Results are expressed as % of control cultures which received no steroids. In one experiment (A-A), 0.3 ml aliquots of the fractions were tested (positive control, 22,400 PFCIculture, dexamethasone control, 240 PFC/culture). In the second experiment (0 O), 0.2 ml aliquots were tested (positive control, 2,070 PFC/culture; dexamethasone control, < 10 PFC/cullure).

-

activities may therefore be functions of identical or similar molecules. However, although the mediators appear to resemble one another, information about their molecular characteristics is limited and several molecular species with similar chemical properties may exist. Further purification and testing of the reagents in each of the biological assays may reveal more heterogeneity than has been established thus far.

BIOLOGYOF ACQUIRED RESISTANCE As summarized above, our previous studies showed that adjuvant-activated accessory cells and/or mediators produced by them (GRMF) protect primed helper T-cells from steroid immunosuppression. We have extended these studies

Mishell et al. : Hormone Response-Modifying Factors

439

to determine whether the helper activity of unprimed T-cells as well as other types of T-cell activities are sensitive to steroids and whether they can be protected from steroids by activated accessory cells and/or GRMF. For many of these studies, we have used the unprimed, immunologically mature thymocyte population as a source of functional T-cells. This population is isolated on the basis of differential binding with peanut agglutinin (PNA) . 2 1 Immature cortical thymocytes, which account for 90-95% of the thymocyte population, have surface glycoproteins with exposed galactose residues to which PNA specifically binds. The galactose residues are masked by sialic acid on immunologically mature medullary thymocytes. Therefore, in the presence of PNA, the sialic acid negative (SA-) immature thymocytes agglutinate. They are readily separated by 1 x g sedimentation from the nonagglutinated sialic acid positive (SA+) mature thymocytes.

SUPPRESSION OF HELPERACTIVITY OF SA+ THYMOCYTES BY STEROIDS The experimental design for investigating the steroid sensitivity of SA+ 3. SA+ thymocytes were initially cultured for thymocytes is outlined in FIGURE a total of 48 hours. During the 24-48 hour period, lo-' M dexamethasone was added to some of the cultures. At 48 hours, the cells were harvested, washed, and added to cultures of freshly prepared T-depleted normal spleen cells to assess their T-helper activity. The results of three experiments are summarized in TABLE 3. The SA+ thymocytes completely restored the responses of the T-depleted spleen cells. However, if the SA+ thymocyte cultures contained dexamethasone during the 24 to 48 hour period, T-helper activity was severely inhibited. These results show that the SA+ thymocytes exhibit T-helper activity in vitro and that the helper activity is sensitive to steroids.

TABLE2 EFFECTS OF METAPERIODATE ON GRMF * Peak 1

Peak 2

-~

Volume

Untreated

Metaperiodate

Untreated

Metaperiodate

0.3 ml 0.2 mi 0.1 rnl

9,450 5,370 1,550

630 70