Eur. J. Immunol. 1992. 22: 165-173

Garth A. Wilbanks., Michele Mammolenti and J. Wayne Streilein Department of Microbiology and Immunology, and Department of Ophthalmology, University of Miami School of Medicine, Miami

TGF-P and ACAID

165

Studies on the induction of anterior chamber-associated immune deviation (ACAID) 111. Induction of ACAID depends upon intraocular transforming growth factor$* Delayed hypersensitivity (DH), the prototypical form of cell-mediated immune responsiveness, is mediated with the participation of considerable nonspecific inflammation which necessarily disrupts the anatomic integrity of involved and adjacent tissues. Damage of this type is of minor consequence to many visceral and cutaneous organs, but is of devastating consequence for organs such as the eye and the brain. At least in the case of the eye, the organ is remarkably adept at regulating the immune system’s ability to respond to intraocular antigens by selectively down-regulating both the induction and expression of delayed hypersensitivity while leaving other effector modalities intact. This ability of the eye to selectivity down-regulate systemic D H responses to intracamerally inoculated antigens is known as anterior chamber-associated immune deviation (ACAID) and is mediated in part by antigen-specific regulatory T cells. Recent work suggests that macrophages (MQ) that reside in the iris and ciliary body can migrate out of an antigen-bearing eye and activate regulatory Tcells within the spleen. In an effort to understand the mechanism by which intraocular MQ interact with antigen in the anterior chamber of the eye (AC) and subsequently induce splenic regulatory cells in ACAID, we have investigated what role, if any, the AC microenvironment itself plays in ACAID induction. The results reveal that CD4Y parenchymal iriskiliary cells secrete a soluble factor(s) locally and into the aqueous humor which endows resident, mature MQ with ACAIDinducing capabilities. Mice receiving infusions of these altered, antigen-pulsed MQ are incapable of mounting a significant D H response following immunization with antigen in adjuvant. Importantly, the ACAID-inducing effect is achieved when conventional, extraocular MQ are exposed in vitro to a soluble factor present in aqueous humor or culture SN from iris and ciliary body cells. Further investigations into the identity of this factor reveal it to be transforming growth factor+ (TGF-fI).The role of TGF-fI in the generation of ACAID, as well as the implications of these findings to an understanding of immunologic privilege in general, are discussed.

1 Introduction The understanding of mechanisms involved in peripheral T cell tolerance induction as it relates to autoimmunity, tumor growth and transplant rejection has been the target of intense experimental effort over the past decade. Despite the many examples of peripheral tolerance induction in adult individuals following introduction of antigen

[I 98551

*

This work was supported by USPHS EY-05678. Supported in part by Ophthalmology Specialty Training grant EY-07021-11.

Correspondence:J. Wayne Streilein, Department of Microbiology and Immunology, University of Miami School of Medicine. PO. Box 016960 (R-138), Miami, FL 33101, USA Abbreviations: AC: Anterior chamber of the eye ACAID: Anterior chamber-associated immune deviation DH: Delayed hypersensitivity IICB: Iris and ciliary body cells PEC: Peritoneal exudate cells TGF-P: Transforming growth factor$

0 VCH Verlagsgesellschaft mbH, D-6940 Weinheim, 1992

via i.v. [I],oral [2] or intracameral [3] routes, very little is known about the precise cellular mechanisms involved in initiating these processes. Over the past 15 years, this laboratory has studied the remarkable and unique systemic immune responses incited by antigens placed in the anterior chamber (AC) of the eye. Following the inoculation of antigens into the AC, specific T and B cells are activated systemically,but the resultant effector response is selectively deficient in delayed hypersensitivity (DH) and complement-fixing antibody production [3-61. The selective deficits of DH and complement-fixing antibody production, termed anterior chamber-associated immune deviation (ACAID), correlate with the splenic presence of CD8+ efferent T, lymphocyte populations, as well as, and CD4+ afferent T, lymphocyte poulatjons [6, 71. Viral, transplantation, tumor-specific, and soluble protein antigens injected into the AC have all been shown to induce ACAID [4,8-111. For example, AC inoculation of semiallogeneic cells impairs the subsequent capacity of rats to reject orthotopic skin grafts syngeneic with the inoculated cells [12, 131. Likewise, AC inoculation of minor histoincompatible tumor cells into the AC of mice results in progressive growth of the tumor in oculi (immune privilege) and in acceptance of orthotopic skin grafts syngeneic with 0014-~2980/92/0101-016S$3.S0 + .25/0

166

G. A. Wilbanks, M. Mammolenti and J. W. Streilein

the tumor [lo, 141. Moreover, T cell-mediated protection from autoimmune uveoretinitis can be achieved by AC inoculation of retinal autoantigens [ll]prior to experimental induction of the disease. Recent experimental evidence demonstrates the importance of the eye itself in the generation of the ACAID phenomenon. In particular, BM derived cells that reside in the murine iris and ciliary body (I/CB), and which express the mature MQ,marker, F4/80, appear to ingest intraocular antigen, escape the eye via the blood vasculature, and migrate to the spleen where regulatory T cells characteristically found in ACAID are generated [15, 161. While F4/80+ MQ,from the normal I/CB are capable of inducing ACAID, extraocular MQ, [peripheral blood monocytes, peritoneal exudate cells (PEC)] are not. However, extraocular MQ,populations can acquire ACAID-inducing properties following direct inoculation into the AC, a finding which implicates the microenvironment of the AC in modifying the antigen-presenting functions of MQ [16].

Eur. J. Immunol. 1992. 22: 165-173

taining 1% normal mouse serum (NMS), 2 X M 2-ME, MEM nonessential amino acids (0.1 mM, Gibco Laboratories, Grand Island, NY), MEM sodium pyruvate (1 mM, Gibco) , L-glutamine (2 mM) , penicillin/streptomycin (100 U 100 pg/ml), and Hepes (5 mM, Gibco). Two hundred thousand I/CB cells in 2 ml of medium were placed in 35 mm tissue culture plates (Becton Dickinson, Lincoln Park, NJ) and incubated at 37 "C in a humidified, 5% C02 atmosphere for up to 2 months. SN was harvested approximately every 4 days and replaced with 2 ml of culture medium. Harvested SN was frozen at - 90 "C in siliconized vials for use at a later date.

+

2.3 PEC cells PEC were obtained from naive mice that received 2.5 ml of a 3% thioglycolate solution i.p. 4 days earlier.

2.4 ACAID induction protocol We have now examined the features of the AC that enable it to exert such a profound effect on MQ, antigen-presenting function in vivo. Our attention was drawn to two components of the AC, the transparent fluid filling the AC (aqueous humor) and cells of the I/CB which are largely responsible for aqueous humor production. Several investigators have demonstrated that both aqueous humor and cells of the I/CB can dramatically impair antigen-driven T cell proliferative responses, as well as IL 2 production in vitro [17-191. These effects are due, at least in part, to transforming growth factor-P (TGF-P) which has been found in aqueous humor [20,21] and I/CB culture SN [22]. With these findings in mind, we assessed the capacity of aqueous humor and I/CB culture SN to alter the antigenpresenting potential of F4/80+ extracular MQ. Our results indicate that the AC microenvironment endows MQ, with an ACAID-inducing potential, a phenomenon attributable, at least partially, to local TGF-6 production by non-BM-derived iridciliary body parenchymal cells.

2 Materials and methods 2.1 Animals Six- to twelve-week-old BALB/c mice (Andervont line) were obtained from the University of Miami mouse colony. All animals were treated according to the Association for Research in Vision and Ophtalmology (ARVO) resolution on the use of animals in research.

2.2 Preparation of iriskiliary cell cultures The method of prcparation has been described in previous publications [17]. Briefly, the iris and ciliary body were microdissected from the eyes of naive BALB/c mice and incubated at 37°C for 1 h in culture medium containing collagenase/dispase (1 mg/ml, Boehringer Mannheim Biochemicals, Indianapolis, IN). Following incubation, singlecell suspensions were obtained by repeated aspirations through an 18-gauge, followed by 21-gauge needle attached to a 5-mI syringe. After washing, the I/CB cells were resuspended in complete culture medium-RPMI 1640 con-

PEC were isolated from naive BALB/c mice and 2 x lo5 cells incubated in 150 pl of complete culture medium containing BSA (5 mg/ml, Sigma Chemical Co., St. Louis, MO) in a 96-well plate. Fifty microliters of PBS or I/CB culture SN was added to the cultures which were allowed to continue overnight. In some experiments neutralizing antibodies for TGF-Pl(1: 10, a kind gift of A. B. Roberts) and TGF-P2 (30 pg/50 p1 SN, R & D Systems, Minneapolis, MN) were added to the PEC-BSA-I/CB culture SN mixture prior to overnight culture. Neutralizing antibody titers used were based on in vitro studies utilizing the TGF-P-sensitive Mv 1 Lu mink lung epithelial cell line [23] American Type Culture Collection, CCL-64, Rockville, MD; data not shown). In other experiments, stock TGF-P1 and TGF-fi2 (R & D Systems) or PGE2 (Sigma) was added to PEC-BSA cultures in lieu of I/CB SN. Following incubation, the PEC were cooled for 30min at 4"C, dislodged from the culture plate by vigorous pipetting, washed twice, resuspended and infused i.v. (2 x lo3 cells/100 pI) into naive syngeneic recipients. Seven days later, all mice received an immunizing dose of BSA (50 mg) emulsified 1: 1 in CFA (0.5 mg Mycobacteriudml, Difco Laboratories, Detroit, MI) in a total volume of 50 ml injected S.C. into the nape of the neck. 2.5 Assessment of DH

DH was accomplished as previously described [24, 251. Seven days following S.C. immunization with BSA in CFA, mice received i.d. inoculations of BSA (400 mg/20 ml) into the right ear pinna. The ear swelling response was then assessed 24 and 48 h later using a micrometer (Mitutoyo 227-101, MTI Corporation, Paramus, NJ). 2.6 Cell separation Cell separation was accomplished using a panning technique developed by Wysocki and Sat0 [26] with minor alterations. Briefly, 100 x 20 mm tissue culture dishes (Becton Dickinson) were incubated with 10 ml of a Tris buffer-anti-F4/80 (F4/80 hyridoma, American Type Culture

TGF-p and ACAID

Eur. J. Immunol. 1992. 22: 165-173

Collection) or-anti-CD45 (PharMingen, San Diego, CA) solution for 90 min at 37 "C. Anti-F4/80 is a rat IgGzb mAb directed against a differentiation antigen restricted to mature mouse M@ and monocytes [27]. F4/80 does not bind to dendritic cells, lymphocytes, erythrocytes or granulocytes. After incubation, plates were washed thrice with 1% NMS solution and cells added at 1 x 106/ml in a 3-ml volume. Following a 90-min incubation at 4"C, adherent and nonadherent fractions were collected, washed and resuspended to the appropriate concentration prior to i.v. inoculation. Preliminary experiments using FCM analysis confirmed that optimal panning of I/CB cells with antiF4/80 occurred at a titer of 1: 50 yielding a 95% pure F4/80cell population in the anti-F4/80 nonadherent fraction and a 50% pure F4/80+ cell population in the anti-F4/80 adherent fraction (data not shown). Optimal panning with antiCD45 also occurred at a titer of 1 : SO, yielding a 70% pure CD45+ cell population in the anti-CD45-adherent fraction and an 88% pure CD45- cell population in the antiCD45-nonadherent fraction (data not shown). 2.7 Statistical analysis Ear swelling measurements were evaluated statistically using a two-tailed Student's t-test. All p values < 0.05 were deemed significant. All experiments were repeated at least twice.

167

These animals were subsequently immunized against BSA in CFA S.C. Significantly reduced DH responses, compared to positive controls,were taken as evidence that a test factor was important in rendering M@ ACAID inducing. 3.2 Aqueous humor endows conventional M a with ACAID-inducing potential PEC were harvested from naive micc pretreated with 3% thioglycolate i.p. 4 days earlier. Some PEC were divided into FU8O-enriched and -depleted fractions by panning with anti-F4/80 antibody. Following panning, 2 X lo5 PEC were resuspended in a 100 p1 volume of fresh rabbit aqueous humor or PBS and allowed to incubate at 37 "C for 1 h. One hundred microliters of complete culture medium (2% NMS) containing BSA ( 5 mg/ml) was then added and the cultures allowed to incubate an additional 18 h. The following day, the cultures were harvested, washed and 2 X lo3 cells infused i.v. into the lateral tail vein of naive syngeneic recipients. This number was based on previous reports demonstrating that 2 x lo3 PEC contain a number of F4/80+ cells equal to what can be harvested from a single murine I/CB (1.2 x lo3 [17,27]). Group A (Fig. 1)received whole unpanned PEC; group B, F4/80-enriched PEC; and group C, F4/80-depleted PEC. Seven days later, all groups received an immunizing dose of BSA emulsified in CFA and 1week later, these same mice were challenged with soluble BSA (400 pg) in the right ear pinna. Ear swelling, (reflecting DH) was measured 24 and 48 h later.

3 Results 3.1 General experimental plan In the following experiments, F4/80+ M@ were harvested from the peritoneal cavity of thioglycolate-treated mice. The cells were then pulsed with antigen (BSA) in vitro in the presence of various eye-related factors suspected of being important in creating the so-called "ACAID-inducing" signal. After treatment, the cells were washed thoroughly and injected i.v. into naive syngeneic recipients.

Group:

(A)

(B)

All groups receiving infusions of BSA-pulsed PEC treated with PBS mounted vigorous D H responses following immunization with BSA in CFA (see Fig. 1, groups A, B and C). In contrast, groups receiving whole, unpanned PEC (group A) as well as groups receiving F4BO-enriched PEC (group B) pretreated with aqueous humor were unable to mount D H responses following immunization with BSA in CFA. As predicted by results of previous experiments demonstrating that F4/80+ I/CB cells are responsible for inducing the suppression of D H character-

(C)

Figure I . Rabbit aqueous humor endows PEC with ACAID-inducing capabilities. PEC were harvested from naive mice pretreated with 3% thioglycolate i.p. 4 days earlier. Some PEC were divided into F4/80-enriched and -depleted fractions by panning with anti-F4/80 antibody. Following panning, 2 X 105PEC were resuspended in fresh rabbit aqueous humor or PBS and allowed to incubate at 37°C for 1 h. Culture medium containing BSA (5 mg/ml) was then added and the cultures allowed to incubate overnight. The following day, cultures were harvested, washed and 2 X 10' cells infused i.v. into the lateral tail vein of naive syngeneic recipients (see text for details). Group A received whole unpanned PEC. Group B received F4BO-enriched PEC. Group C received F4/80-depleted PEC. Seven days later, all groups were immunized with BSA emulsified in CFA (50 pg). One week later, all groups received ear challenges with BSA. Bars represent the mean ear swelling measurements of mice 24 h post challenge with BSA. Asterisks indiate mean values significantly less than groups receiving PBS-treated PEC (p < 0.05).

Eur. J. Immunol. 1992. 22: 165-173

G . A. Wilbanks, M. Mammolenti and J. W. Streilein

168

istically seen in mice with ACAID [15], mice receiving F4BO-depleted PEC treated with aqueous humor had no evidence of DH suppression; all mice mounted vigorous DH responses following immunization (group C). Thus, a soluble factor appears to be present in aqueous humor that causes F4/80+, peritoneally derived, MQ populations to impair D H induction when injected into mice destined to be immunized with BSA in CFA.We will refer subsequently to this property as “ACAID-inducing”.

3.3 I/CB culture SN endow PEC with ACAID-inducing potential Having observed that a soluble factor present in normal aqueous humor is capable of altering MQ function resulting in suppressed DH responses, we next sought to determine the source and identity of this soluble, “Ma-modifying” factor. Because aqueous humor is produced primarily by ciliary body epithelium, we first tested whether SN from in vitro cultures of cell suspensions prepared from I/CB could endow M a with ACAID-inducing capabilities. PEC were harvested from naive mice pretreated with 3% thioglycolate. Some of the PEC were then divided into F4/80adherent and -nonadherent populations by panning with anti-F4/80 antibody. Following panning, cells were harvested, washed and resuspended at 2 x 105/150pl complete culture medium containing BSA (5 mg/ml). Thereafter 150 p1 of the cell suspension was added to several wells in a 96-well plate. PBS or 2-week-old I/CB culture SN (50 ml) was then added to each well and the cultures allowed to incubate for an additional 18-24 h. The following day, the cultures were harvested, washed and 2 x lo3 cells infused i.v. into the lateral tail vein of naive syngeneic recipients. Seven days later, all groups received an immunizing dose of BSA in CFA. One week later, the anti-BSA D H response was assessed following ear challenge with soluble BSA (400 Pg). As before, groups receiving infusions of BSA-pulsed PEC treated with PBS mounted vigorous D H responses follow-

150

ilture Tx: ] PBS 150 I/CB

ICulture sup. 100

. -4

a,

.C

v) L

m w

Group:

*

(A1

50

0

1 (Bl

150

ing immunization with BSA in CFA (see Fig. 2, groups A, B, and C). In contrast, groups receiving BSA-pulsed unselected PEC (group A) as well as F4BO-enriched PEC (group B) pretreated with I/CB culture SN were unable to mount DH responses following immunization with BSA in CFA. Group C confirms that PEC depleted of F4/80+ MQ fail to induce impaired D H reactivity. Thus, a soluble Ma-modifying factor, similar to that found in normal aqueous humor, is produced in vitro by cells harvested from the I/CB.

3.4 F4/SO-depleted I/CB cell culture SN endow PEC with ACAID-inducing potential In an attempt to identify the cellular source of the MQ-modifying activity,we separated freshly isolated I/CB cells into distinct populations based on surface marker expression. We first chose to determine whether the Ma-modifying factor was produced by F4/80+ cells themselves. PEC were harvested from naive mice, washed and resuspended at 2 x 105/150p1 complete culture medium containing BSA (5 mg/ml). Thereafter 150 p1 of the cell suspension was added to several wells in a 96-well plate. As in the previously described experiment, 50 pl of PBS or 10-day-old I/CB culture SN was added to each well; however, prior to I/CB culture initiation, I/CB cells were purified into F4BO-enriched and F4BO-depleted populations, yielding SN from F4BO-enriched or -depleted I/CB cultures. SN from 10-day-old I/CB cultures were chosen because it was well after the time when cells stabilized in culture, yet before the time when the F4BO-enriched I/CB cell cultures attrited due to a lack of in vitro growth factor supplementation (= 2 weeks; data not shown). Following an 18-24 h incubation of the PEC-BSA-I/CB culture SN mixture, cultures were harvested, washed and 2 x lo3 PEC infused i.v. into the lateral tail vein of naive syngeneic recipients. Group A (Fig. 3) received PBS-treated PEC. Group B received PEC treated with F4/80-enriched I/CB culture SN. Group C received PEC treated with F4/80depleted I/CB culture SN. Seven days later, all groups were

4

100

50

(CI

Figure 2. I/CB bulk-culture SN endows PEC with ACAID-inducing potential. PEC harvested from thioglycolate stimulated mice were cultured with BSA (5 mglml) overnight in complete medium containing 25% PBS or 25% SN from 10-day-old I/CB cultures (see text for details). The following day, cultures were harvested, washed and 2 x lo3 cells infused i.v. into the lateral tail vein of naive syngeneic recipients. Group A received whole unpanned PEC. Group B received F4BO-enriched PEC. Group C received F4BO-depleted PEC. For further treatment see legend to Fig. 1.

Eur. J. Imr

E

nol. 1992, 22: 165-173

TGF-(3 and ACAID

169

150

E

0

r

X

3m

100

-L

m

2

50

t m h

W

0

Culture Tx: Group:

PBS

(A)

F4/80+I/CB

F4/80' IlCB

Culture SN

Culture SN (C)

( Bl

Figure 3. SN from F4/80-depleted, not F4/80-enriched, IlCB cultures endows PEC with ACAID-inducing capabilities. PEC, harvested from thioglycolate-stimulated mice, were cultured with BSA ( 5 mg/ml) overnight in complete medium containing 25% PBS or 25% SN from IlCB cultures enriched or depleted for F4/80+ cells (see text for details). The following day, cultures were harvested, washed and 2 X lo3 PEC infused i.v. into the lateral tail vein of naive syngeneic recipients. Group A received PBS-treated PEC. Group B received PEC treated with F4BO-enriched I/CB culture SN. Group C received PEC treated with F4BO-depleted I/CB culture SN. For further treatment see Fig. 1.

immunized with BSA in CFA followed a week later with a BSA ear challenge to assess D H responsiveness. Interestingly, only mice receiving F4BO-depleted I/CB culture SN (group C) were prevented from mounting significant D H responses following immunization. Mice receiving BSApulsed PEC treated with PBS (group A) or F4/80-enriched (group B) l/CB culture SN all mounted vigorous anti-BSA D H responses. These results suggest that, although F4/80+ MQ are clearly important in the induction of ACAID, an F4/80- cell population produces the soluble factor that endows M@ with ACAID-inducing properties. 3.5

CD45-depleted I/CB culture SN endow PEC with ACAID-inducing potential

Only approximately 2% of cells in a normal I/CB are BM derived [ 171. Because about 95% of these BM derived cells (CD45+) are F4/80+, the failure of F4/80+ I/CB cells to produce any MQ-modifying activity makes it unlikely that the MQ-modifying factor is produced by BM-derived cells. However, to confirm directly the possibility that CD45-, F4/80- I/CB cells are responsible for secreting the MQmodifying factor, we assessed the ability of CD45-depleted I/CB culture SN to transform PEC into ACAID-inducing cells. As previously described, PEC were harvested from naive mice, washed and resuspended at 2 x 105/150pl complete culture medium containing BSA (5 mg/ml). Thereafter 150 pl of the cell suspension was added t o several wells in a 96-well plate. PBS or 10-day-old SN (50 pl) from CD45-enriched or CD45-depleted I/CB cell cultures was added to each well. Following an 18-24 h incubation, cultures were harvested, washed and 2 x lo3 cells infused i.v. into naive syngeneic recipients. Group A (Fig. 4) received PBS-treated PEC. Group B received PEC treated with CD45-enriched I/CB culture SN. Group C received PEC treated with CD45-depleted I/CB culture SN. Seven days later, all groups received an immunizing dose of BSA emulsified in CFA and 1 week later, mice were

Tx: Group:

pBs

(4

T200+ T200Culture SNB Culture SN

(B)

(C)

Figure 4. SN from CD45-depleted, not CD45-enriched, IlCB cultures endows PEC with ACAID-inducing capabilities. PEC, harvested from thioglycolate-stimulated mice were cultured with BSA (5 mg/ml) overnight in complete medium containing 25% PBS or 25% SN from I/CB cultures enriched or depleted for CD45-expressing cells (see text for details). The following day, cultures were harvested, washed and 2 X lo3 PEC infused i.v. into the lateral tail vein of naive syngeneic recipients. Group A received PBS-treated PEC. Group B received PEC treated with CD45enriched I/CB culture SN. Group C received PEC treated with CD45-depleted IlCB cultures SN. For further treatment see Fig. 1.

challenged with soluble BSA in the right ear pinna. Ear swelling, (reflecting DH) was measured 24 and 48 h later. As predicted, only mice receiving PEC treated with CD45-depleted I/CB culture SN (group C) were unable to mount significant anti-BSA D H responses following immunization. Mice receiving BSA-pulsed PEC treated with PBS (group A) or CD45-enriched I/CB culture SN (group B) all mounted effective anti-BSA DH responses. Thus, a non-BM-derived (CD45-) cell within the I/CB produces a soluble factor that endows MQ with ACAIDinducing properties.

3.6 Time-course assessment of Ma-modifying factor production by I/CB culture SN Exposure of cells to in vitro culture conditions may alter their function to such an extent that a cultured cell's functional properties may be vastly different from the original cell population from which it was derived.With this possibility in mind, we next assessed the effect of I/CB cell culture SN on PEC function over a 2-month time period (15-20 cell-pass cycles). As described in Sect. 2.2, fresh I/CB cells were placed in 35-mm tissue culture plates and incubated at 37 "C in a humidified, 5% COz atmosphere for up to 2 months. SN was harvested approximately every 4 days and frozen at - 90 "C. To assess the effect of in vitro culture conditions on the MQ-modifying capacity of I/CB culture supernatant, 50 pl of SN from 1-, 2- or 8-week-old I/CB cultures was added to PEC cultures (2 X lo5ce11/150 yl medium) containing BSA ( 5 mg/ml) and allowed to incubate overnight (18-24 h). One culture received 2 x lo5 F4/80-depleted I/CB cells to assess the ability of freshly isolated I/CB cells to render PEC ACAID-inducing. The following day, the cultures were harvested, washed and 2 X lo3 cells infused i.v. into the lateral tail vein of naive

170

G. A. Wilbanks, M. Mammolenti and J. W. Streilein

Eur. J. Immunol. 1992. 22: 165-173

Table 1. Ma-modifying factor production by IlCB cell cultures Time course Source of Mamodifying factor")

Recipients' DH resonsiveness (24 h ear swelling SEM)

a) The IlCB were isolated from naive, thioglycolate-stimulated BALBlc mice, rendered into a (wm) single-cell suspension, and cultured as described in Sect. 2.2. SN was harvested at PBC Control 120 lk 10 Fresh. F4BO-depleted IlCB cellb) 18k 2 various times and frozen for use at a later Culture SN from 1-week I/CB cultures 32k 3 date. PEC were isolated from naive BALBk mice and cultured with BSA (5 mglml) over33k s Culture SN from 2-week IlCB cultures 20k 7 night. PBS or IlCB culture SN was added to Culture SN from 8-week IlCB cultures the various culture groups as described in Sect. 2.4. Following incubation, the PEC were washed and 2 X 103cells infused i.v. into naive recipients, all of whom were immunized 1 week later with BSA emulsified in CFA. The anti-BSA ear swelling response was assessed 1 week thereafter. b) Fresh, F4l80-depleted I/CB cells were added directly to the PEC-BSA cultures in lieu of IlCB bulk culture SN to assess the immediate effects of freshly isolated IlCB cells on PEC function.

syngeneic recipients. Seven days later, all groups received an immunizing dose of BSA (50 yg) in CFA with anti-BSA DH reactivity being assessed in the usual manner 1 week thereafter. Results indicate that the Ma-modifying activity was present in all ages of I/CB SN tested. I/CB culture SN from 1-, 2- and 8-week-old cultures as well as freshly isolated F4/80- cultures all endow PEC with an ACAIDinducing capacity (Table 1). These results suggest that the Ma-modifying activity is associated with a relatively stable molecule which is produced by a fairly long-lived, nonBM-derived cell population normally resident in the I/CB and which is capable of surviving in vitro for at least 2 months.

3.7 TGF-fi renders PEC ACAID-inducing cells Having observed that a soluble factor, produced by CD45I/CB cells, possesses the ability to convert M a into ACAID-inducing cells, we wished to determine the identity of this factor. Recent work published by several laboratories suggests that TGF-P might be a critical cytokine in ocular immune privilege and in the ACAID-inducing process [20,21,28]. TGF-P is known to alter APC function

PEC were harvested from naive mice, washed and resuspended at 2 x 105/150p1 complete culture medium containing BSA ( 5 mg/ml). One hundred-fifty microliters of the cell suspension was added to several wells in a 96-well plate. Fifty microliters of distilled, filtered water, or varying amounts of PGEz M - ~ O - ~~ / 5 yl, 0 Sigma) or TGFPI/TGF-P~(1-50 ng/50 yl water, R & D Systems) was added to each well. Following an 18-24 h incubation, cultures were harvested, washed and 2 x lo3 cells infused i.v. into naive syngeneic recipients. Seven days later, all groups received an immunizing dose of BSA emulsified in CFA and 1week later, mice were challenged with soluble BSA in the right ear pinna. Ear swelling was measured 24 and 48 h later. Fig. 5 displays results which reveal that no concentration of PGE2 (ranging from lop7 to l O P 9 ~ )

"

"

PEC

Treatment:

profoundly [29], as well as to down-regulate Tcell proliferative responses in vitro [30]. Because the aqueous humor is known to contain significant amounts of TGF-0 [20, 211 and I/CB cell cultures are known to produceTGF-P in vitro [22], we wished to determine if TGF-P might endow M a with ACAID-inducing activity. PGE was also assessed for its Ma-modifying capacity because of its well known ability to suppress Tcell function [31].

Media

PGE210.~~4PGE,IO"M

PGE ,iO"M

H20

50 ng TGF-I3 10 ng TGF-I3 1 ng TGF-I3

Figure 5. TGF-P renders PEC ACAID inducing. As described in Fig. 2 , PEC were harvested from naive mice, washed and resuspended. Cells were cultured overnight in complete medium containing BSA (5 mglml).Various amounts of PGEz IV-~O-~ ~ / 5 pI),TGF-P, 0 and TGF-Pz (1-50 ngl50 PI) or water was added to each culture (see text for details). Following an 18-24 h incubation, cultures were harvested, washed and 2 x lo3 cells infused i.v. into naive syngeneic recipients. For further treatment see Fig. 1.

TGF-@and ACAID

Eur. J. Immunol. 1992. 22: 165-173

conferred ACAID-inducing potential upon PEC (Fig. 5 A ) . However, we found that PEC cultured with 1-10 ng TGF-fJ failed to induce conventional D H immune responses (Fig. 5 B), and induced a state of profound D H unresponsiveness. Paradoxically, PEC treated with 50 ng TGF-p induced an apparently normal D H response in recipients immunized with BSA emulsified in CFA. 3.8 Anti-TGF-P antibody abolished the effect of I/CB culture SN on PEC

171

presence of intracameral TGF- f3 plays an important role in the ACAID inductive process

4 Discussion Our studies into the nature of the AC microenvironment, a milieu which endows certain APC with ACAID-inducing properties, have helped to illuminate the interacting elements responsible for ocular immune privilege and ACAID. The data reveal that F4/80+ M a populations pulsed with antigen and exposed to factors from the anterior segment of the normal eye are capable of inducing a D H unresponsive state in naive mice subsequently immunized with BSA in adjuvant. Previous investigations indicated that this inability to mount D H responses is mediated by regulatory T lymphocytes in a spleen-dependent, antigen-specific manner [ 15,161. Our present findings reveal that ACAID-inducing properties of antigen-pulsed M@ are acquired if the M a are exposed in vitro to a soluble factor present in aqueous humor and/or I/CB culture SN. Moreover, this factor is produced by F4/80-, C D 4 Y parenchymal I/CB cells, and appears to be TGF-6, a constituent of normal aqueous humor and a secretory product of cells of the I/CB [20-221.

To confirm a role for TGF-f3 in ACAID induction, we next attempted t o block the Ma-modifying effect of I/CB culture SN with TGF-p-specific antibody. PEC were harvested from naive mice pretreated with 3% thioglycolate. After washing, cells were resuspended 2 X 105/150p1 complete medium containing BSA (5 mg/ml). One hundred-fifty microliters of the cell suspension was added to several wells in a 96-well plate. PBS or 2-week-old I/CB culture SN (50 PI), both of which were pretreated with anti-TGF-Bl and $2 neutralizing antibody 1 h earlier (see Sect. 2.4 for details), was added to each well and the cultures allowed to incubate for an additional 18-24 h.The following day, the cultures were harvested, washed and 2 x lo3 cells infused i.v. into the lateral tail vein of naive syngeneic recipients. Seven days later, all groups received When placed in the context of previous investigations into an immunizing dose of BSA in CFA and 1 week later, the ocular immune privilege and ACAID, these results make it anti-BSA D H response was assessed following ear chal- possible to postulate a sequence of events that explains lenge with soluble BSA. As depicted in Fig. 6, neutralizing ACAID induction at the cellular level. F4/80+, Ma-like antibody directed against TGF-f~ partially reversed the cells whose functional programs have been altered followability I/CB culture SN to render M@ ACAID inducing. ing exposure to the AC microenvironment capture soluble When considered along with data demonstrating that antigens, such as BSA, placed intracamerally. These TGF-P can change F4/80+ PEC into ACAID-inducing cells, ACAID-inducing M@ subsequently exit the eye via the and that TGF-(3 is present in normal I/CB culture SN (3-4 trabecular meshwork and canal of Schlemm into the blood ng/ml) [22], this result strongly supports the view that the vascular system, and journey to the spleen. At this site, suppressor regulatory T cells are activated, presumably by exposure to an antigen-specific signal associated with these eye-derived cells [6. 15-16,321. We have evidence that this Ma-borne signal is a non-native form of antigen, in that ACAID-inducing, F4/80+ cells in the peripheral blood 2 days after mice have received an AC inoculation of BSA do not express native BSA on their surface [lS].

Cell Treatment: Antibody Treatment:

I/CB SN

PBS

I/CB SN

Complete Media

Anti-TGF-O Anti-TGF-R

Complete Media

PBS

Figure 6. Anti-TGF-fi blocks the ability of I/CB culture SN to cndow PEC with ACAID-inducing capabilities. PEC were harvested in the usual manner, resuspended in complete culture medium, and cultured overnight in the presence of BSA, I/CB culture SN or PBS as described in Fig, 2. However, PBS and I/CB culture SN were pretreated with anti-TGF-fit and TGF-PZneutralizing antibody 1 h prior to culture addition (see Sect. 2.4 for details). Following overnight culture, cells were harvested, washed and 2 x lo3 cells infused i.v. into the lateral tail vein of naive syngeneic recipients. For further treatment see Fig. 1.

D H is an important armament used by the immune system in the defense against bacterial pathogens and is known to play a protective role against intracellular pathogens, as well as in tumor and transplant rejection. However, a price must be paid for the protection afforded by D H because this immune effector modality recruits intense nonspecific inflammation which causes innocent-bystander tissue injury. It has been proposed that this price is unacceptably high in organ systems whose functional integrity is absolutely dependent on an inflammation-free environment and a precisely maintained microanatomy. Organs such as the eye, brain and even recently transplanted organs are exquisitely sensitive to surprisingly small inflammatory insults. In this context, the eye appears to possess a remarkable ability to regulate the immune system's response to intraocular antigens by causing the downregulation of both the induction and expression of delayed hypersensitivity. Although a number of potent inhibitory molecules are known to be produced by tissues lining the AC [20-221, TGF-6 appears to be primarily responsible for ACAID induction, as evidenced by our ability to block the

172

G. A. Wilbanks, M. Mammolenti and J. W. Streilein

M@-modifying ability of I/CB culture SN with TGFP-specific neutralizing antibody (Fig. 6). Moreover, the ability of in vitro exposure toTGF-P to alter M a function, such that when the cells are placed in vivo they induce regultory T cells capable of suppressing D H responsiveness, portends the possible application of this technology in the management or prevention of immune responses in which DH is believed to play a pathologic role. Examples include transplant rejection [33, 341, uveoretinitis 1111, various arthritides [2, 351, several autoimmune demylinating diseases 12, 361 and the response to Leishmania infection [37, 381. This possibility is particularly attractive when one considers that the CD8+ regulatory T lymphocytes found in mice with ACAID can selectively impair activated DH effector lymphocytes, a valuable asset for any immunotherapeutic regimen targeted at ongoing cellmediated DH responses [6, 71.

Eur. J. Immunol. 1992. 22: 165-173

are F4/80+ MWmonocytes. These data lead to the suggestion that eye-derived M a carry the major burden of ACAID induction in vivo. In summary, we believe these investigations to shed new light on the phenomena of ocular immune privilege and ACAID-both of which are characterized by an atypical immune outcome when antigen is placed in the AC of the eye. It is interesting that the ACAID-inducing, Ma-borne signal is dependent on the effects of factors produced by non-BM-derived parenchymal cells in the local I/CB tissue. Ocular cells are thus largely responsible for creating a unique intraocular microenvironment which, in turn, strongly influences the character of subsequent local and systemic immune responses to intraocular antigens. Although the relevance of these investigations to other forms of immune privilege and tolerance remains to be determined, microenvironmental factors produced by tissue-specificparenchymal cells will almost certainly prove to play major roles in the unique forms of regulation that characterize tissue and organ-restricted immune responses in vivo.

Although a number of independent investigators have demonstrated the presence of inhibitory M a populations, relatively little is known about the mechanism by which these cells lead toT5induction 139-431. Ishikura et al. have noted that exposure of specific immunogenic M a hybridomas to IFN-y-containing SN or rIFN-y endows the cells with the capacity to induce effector T, [44]. Although the mechanism of T, induction was not elucidated, blocking experiments utilizing anti-I-J idiotype suggest the mechanism to depend on Ts-I-J interaction with an, as of yet undefined, APC surface molecule. Our future experiments will examine any similarities between these M a hyridomas and the fresh M a we harvest from the I/CB.

5 References

We are fascinated by the differential susceptibility of APC to the “ACAID-ogenic”effects of the AC.Williamson et al., using a tumor model of ACAID (in which P815 mastocytoma cells alone, injected into the AC of BALB/c mice induced ACAID), demonstrated that co-injection of P815 cells and epidermal Langerhans cells into the AC provoked vigorous DH immune reactivity (i.e. prevented ACAID), whereas co-injection of P815 cells with spleen cells or even with LPS-treated B-cells into the AC did not prevent ACAID [45]. Related to this observation, Demidem et al., demonstrated that fresh Langerhans cells pretreated with TGF-P could stimulate allogeneic T cells whereas cultured Langerhans cells, as well as peripheral blood monocytes pretreated with TGF-P were no longer allostimulatory [29, 461. The resistance of fresh Langerhans cells to the “ACAID-ogenic”effects of aqueous humor/TGF-P may be related t o the extraordinary power of Langerhans cells to process and present exogenous antigens. For example, McKinney and Streilein [47] have used as few as 10 i.v. injected allogeneic Langerhans cells to prime cytotoxic Tcells in vivo. Other recent studies have revealed Langerhans cells to be virtually the only APC capable of inducing a primary immune response to soluble protein antigens in vitro [48]. Interestingly, when Williamson et al. inoculated APC other than fresh Langerhans cells (splenic cells, B lymphocytes, or A20 lymphoma cells) into the AC along with allogeneic tumor cells, the recipient mice developed ACAID [45], suggesting that B lymphocytes might also be susceptible to the Ma-modifying effects of TGF-P. However, we have previously demonstrated that following the AC inoculation of BSA, the only cells recoverable from the peripheral blood of mice that possess ACAID-inducing capabilities when transferred directly into naive recipients

1 Germain, R. N . and Benacerraf, B., Scand. J. Immunol. 1982. 13: 1. 2 Thompson, H. S. G . and Staines, N. A , , Immunol. Today 1990. 11: 396. 3 Streilein, J. W., FASEB J. 1987. I : 19. 4 Mizuno, K . , Clark, A . F. and Streilein, J. W., Invest. Ophthalmol. Vis. Sci. 1989. 30: 1112. 5 Wilbanks, G. A. and Streilein, J. W., Immunology 1990. 71: 566. 6 Wilbanks, G. A . and Streilein, J. W., Immunology 1990. 71: 383. 7 Streilein, J. W. and Niederkorn, J. Y., J. Immunol. 1985. 134: 1381. 8 Whittum, J. A., Niederkorn, J.Y., McCulley, J. l? and Streilein, J. W., Curr. Eye Res. 1983. 2: 691. 9 Niederkorn, J. Y. and Streilein, J. W., J. Immunol. 1983. 131: 2670. 10 Niederkorn, J.Y., Streilein, J. W. and Shadduck, J. A . , Invest. Ophthalmol. Vis. Sci. 1980. 20: 355. 11 Mizuno, K., Clark, A. F. and Streilein, J. W., Invest. Ophthalmol. Vis. Sci. 1989. 30: 772. 12 Kaplan, H. J. and Streilein, J. W., J. Immunol. 1978. 120: 689. 13 Subba Rao, D. S.V.and Grogan, J. B., Transplantation 1977.24: 377. 14 Niederkorn, J.Y. and Streilein, J. W., Transplantation 1982. 33: 573. 15 Wilbanks, G. A. and Streilein, J. W., J. Immunol. 1991. 146: 2610. 16 Wilbanks, G. A., Mammolenti, M. and Streilein, J. W., J. Immunol. 1991. 146: 3018. 17 Williamson, J. S. l?, Bradley, D. and Streilein, J.W., Immunology 1989. 67: 96. 18 Kaiser, C. J., Ksander, B. R . and Streilein, J.W., Reg. Immunol. 1989. 2: 42. 19 Streilein, J. W. and Bradley, D., Invest. Ophthalmol. Vis. Sci. 1991. 32: 2700.

The authors thank Dr. Scott Cousins for aqueous humor samples and Dr. Bruce Ksander for helpful discussions.

Received August 28, 1991.

Eur. J. Immunol. 1992. 22: 165-173 20 Granstein, R . , Staszewsk, R., Knisely, T. L., Zeira, E., Nazareno, R. M. Latino and Albert, D. M., J. Immunol. 1990. 144: 3021. 21 Cousins, S.W., McCabe, M. M., Danielpour, D. and Streilein, J. W., Invest. Ophthalmol. ViA. Sci. 1991. 32: 2201. 22 Helbig, H., Kittredge, K. L., Coca-Prados, M., Davis, J., Palestine, A. G. and Nussenblatt, R. B., Gruefes Arch. Clin. Exp. Ophthulmol. 1991, in press. 23 Tucker, R. F., Shipley, G. D. Moses, H. L. and Holley, R. W., Science 1984. 226: 705. 24 Wilbanks, G. A . and Streilein, J. W., Reg. Immunol. 1989. 2: 390. 25 Williamson. J. P. and Streilein, J. W., Reg. Immunol. 1988. 1: 15. 26 Wysocki, L. J. and Sato,V. L., Proc. Nutl. Acad. Sci. U S A 1978. 75: 2844. 27 Austyn, J. M. and Gordon, S., Eur. J. lmmunol. 1981. 11: 805. 28 Cousins, S. W. and Streilein, J. W. in Usui, M., Ohno, S. and Aoki, K. (Eds.), Oculur lmmunology Today, Elsevier Science Publishers, Amsterdam 1990. 29 Demidem, A.,Taylor, J. R . , Grammer, S. F. and Streilein, J.W., J. Invest. Dermatol. 1991. 96: 401. 30 Kehrl, J. H.,Wakefield, L. M., Roberts, A. B., Jakowlew, S., Alvarez-Mon, M., Derynck, R. Sporn, M. B. and Fauci, A. S., J. Exp. Med. 1986. 163: 1037. 31 Goodwin, J. S. and Webb, D. R . , J. Clin. Immunol. 1983. 3: 295. 32 Streilein, J.W. and Niederkorn, J.Y., J. Exp. Med. 1981. 153: 1058. 33 Larsen, C. €!, Morris, l? J., Austyn, J. M., J. Exp. Med. 1990. 171: 307.

TGF-B and ACAID

173

34 Faustman, D. L., Steinman, R. M., Gebel, H. M., Hauptfeld, V., Davie, J. M. and Lacy, l? E., Proc. Nutl. Acad. Sci. USA 1984. 81: 3864. 35 Thompson, H. S. G. andstaines, N . A., Clin. Exp. Immunol. 1985. 64: 581. 36 Higgins, l? J. and Weiner, H. L., J. Immunol. 1988. 140: 440. 37 Mosmann,T. R. and Moore, R.W., Immunol. Today 1991.12: 49. 38 Finkelman, F. D., Pearce, E . J., Urban, Jr., J. E and Sher, A., Immunol. Today 1991. 12: 62. 39 Hillinger, S. M. and Herzig, G. €!, J. Clin. Invest. 1978. 61: 1620. 40 Goodwin, T. S., DeHoratius, R., Israel, H., Peake, G. T. and Messner, R. €!, Ann. Intern. Med. 1979. 90: 169. 41 Markenson, J. W., Locksin, M. €?, Joachim, C. and Winfield, J. B., Proc. Soc. Exp. Biol. Med. 1978. 158: 5. 42 Ju, S. T. and Dorf, M. E., J. Immunol. 1985. 134: 3722. 43 Hausman, €? B., Kawasaki, H . , O H a ra , Jr., R. M., Minami, M., Sherr, D. H. and Dorf, M. E., J. Immunol. 1986. 137: 3717. 44 Ishikura, H., Jayaraman, S., Kuchroo,V., Diamond, B., Saito, S. and Dorf, M. E., J. Immunol. 1989. 143: 414. 45 Williamson, J. S. €! and Streilein, J.W. ,Transplantation 1989.47: 519. 46 Streilein, J. W., Grammer, S. EYoshikawa, T., Demidem, A. and Vermeer, M., Immunol. Rev. 1990. 117: 159. 47 McKinney, E. C. and Streilein, J. W., J. Immunol. 1989. 143: 1560. 48 Hauser, C. and Katz, S. I., Proc. Natl. Acad. Sci. U S A 1988.85: 5625.