Keratinocyte expression of B7-1 in transgenic mice amplifies the ...

3 downloads 0 Views 1MB Size Report
would be more effective inducers of T-cell-mediated immune responses in skin, we prepared transgenic mice in which expression of the B7-1 costimulator was ...

Proc. Natl. Acad. Sci. USA Vol. 91, pp. 12780-12784, December 1994 Immunology

Keratinocyte expression of B7-1 in transgenic mice amplifies the primary immune response to cutaneous antigens IFOR R. WILLIAMS, RICHARD J. ORT, AND THOMAS S. KUPPER* Harvard Skin Diseases Research Center, Division of Dermatology, Brigham and Women's Hospital, Harvard Medical School, 75 Francis Street, Boston, MA 02115

Communicated by Aaron B. Lerner, August 29, 1994 (received for review May 12, 1994)

ABSTRACT Resting epidermal keratinocytes do not express B7-1 and other known CD28 counterligands with costimulatory activity. The absence of these costimulators on keratinocytes correlates with their ability to preferentially induce T-cell anergy instead of T-cell activation. To test the hypothesis that keratinocytes expressing a CD28 counterligand would be more effective inducers of T-cell-mediated immune responses in skin, we prepared transgenic mice in which expression of the B7-1 costimulator was targeted to basal keratinocytes by using the human K14 promoter. Keratinocytes from the K14/B7-1 transgenic line expressed high levels of surface B7-1. No spontaneous inflammatory changes were seen in transgenic skin, but epicutaneous application of contact sensitizers to these mice elicited a stronger primary ear swelling response than in controls. Sites of initial hapten application in transgenic mice also responded much more strongly to reapplication of hapten to a remote cutaneous site. Epidermal cell suspensions from transgenic mice contained normal numbers of Langerhans cells and dendritic epidermal T cells when analyzed by flow cytometry. Systemic treatment of the transgenic mice with interferon y induced high levels of class II major histocompatiblity complex expression on keratinocytes but was not sufficient to initiate an inflammatory response. We conclude that the constitutive expression of the B7-1 molecule in vivo on a nonprofessional antigen-presenting cell is not by itself sufficient to trigger inflammatory changes, but B7-1 expression amplifies the host immune responses after exposure to nonself antigens presented by B7-l-expressing cells.

function of class II-expressing keratinocytes at a site of inflammation is to promote resolution of the inflammatory process by actively down-regulating CD4+ T-cell function. A prediction based on this model is that if keratinocytes constitutively expressed a relevant costimulatory molecule at high levels, the outcome of T-cell exposure to class IIpositive keratinocytes should be amplification rather than suppression of the immune response. The CD28 molecule on T cells, by virtue of its capacity to interact with counterligands (B7-1, B7-2, and perhaps others) expressed by APCs, is a major receptor for delivery of costimulatory signals (5, 6). Engagement of CD28 on primary CD4+ T cells (7) or T helper cell clones from the Thl (8) or Th2 (9) subsets by B7-1 provides a second signal for T cells that promotes full activation and rescues them from anergy induction. Resting keratinocytes do not express CD28 counterligands detectable with either antibodies to B7-1 or the CTLA-4-immunoglobulin (Ig) and CD28-Ig fusion proteins (10). Thus, the absence of CD28 counterligands is one possible explanation of the impaired ability of keratinocytes to function as costimulators for CD4+ T-cell activation. We now report generation of a line of transgenic mice in which constitutive expression of B7-1, the prototype CD28 counterligand, has been targeted to the basal layer of epidermis. By using these mice, we have analyzed the effects of keratinocyte B7-1 expression on resident epidermal leukocyte populations and contact hypersensitivity (CHS) responses.

MATERIALS AND METHODS DNA Constructs and Transgenic Mice. A cDNA clone for mouse B7-1 (11) inserted into the EcoRI site of the vector pBluescript SK- (Stratagene) was the gift of Casey Weaver (University of Alabama, Birmingham). To express this cDNA in the basal epidermis of transgenic mice, the EcoRI insert was filled-in with the Klenow fragment of DNA polymerase I (Boehringer Mannheim), ligated to Bgl II linkers (New England Biolabs), and cloned into the unique BamHI site of the K14/human growth hormone (hGH) expression vector (12), provided by Elaine Fuchs (University of Chicago, Chicago), to generate the K14/B7-1 transgenic construct (Fig. 1). A 5.0-kb EcoRI fragment liberated from the K14/B7-1 plasmid by digestion with EcoRI was used for microinjection of inbred FVB/N embryos as described (13). DNA from 13 live-born mice was screened for the transgene by PCR with hGH-specific primers (13). A female transgenic founder was identified that had 50-100 copies of the transgene by Southern blot analysis.

Because of their well-documented ability to produce and respond to multiple cytokines involved in inflammation and immune responses, keratinocytes are now properly regarded as active participants in the genesis of immune responses occurring in skin (1, 2). However, the precise role of keratinocyte antigen presentation in influencing T-cell responses to antigens encountered in epidermis is less clear. Keratinocytes displaying antigenic peptides in the binding site of class I major histocompatibility complex (MHC) molecules are potential antigen-presenting cells (APCs) for CD8+ class I-restricted T cells. The expression of class II MHC molecules by keratinocytes after in vivo activation with interferon 'y (IFN-y) (3) enables these activated keratinocytes to function additionally as APCs for class II-restricted CD4+ T cells. Previous studies of antigen presentation by class IIexpressing keratinocytes to CD4+ T cells have shown that keratinocytes, like other "nonprofessional" APCs, preferentially induce anergy rather than activation of CD4+ T cells both in vitro (3) and in vivo (4). The inability of keratinocytes to fully activate CD4+ T cells has generally been ascribed to their lack of expression of relevant costimulatory molecules for T-cell activation. A model was proposed (4) in which the

Abbreviations: APC, antigen-presenting cell; CHS, contact hypersensitivity; DETC, dendritic epidermal T cell; DNFB, 2,4dinitrofluorobenzene; FITC, fluorescein isothiocyanate; hGH, human growth hormone; IFN, interferon; LC, Langerhans cell; mAb, monoclonal antibody; MHC, major histocompatibility complex; Ig, immunoglobulin; TCR, T-cell receptor. *To whom reprint requests should be addressed.

The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact. 12780

Immunology: Wifiams et al. Human K14

Promoter/Enhancer 2100 bp

Mouse B7-1 cDNA 950 bp

Proc. Natl. Acad. Sci. USA 91 (1994)


Introns/ pA site (from hGH) 2150 bp _BEP!,e. .. ......................................:::.




11 I 1 kb



TAG (Stop)


1! :,~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~...


FIG. 1. K14/B7-1 transgenic construct with the promoter/ enhancer region of the human K14 gene (horizontal lines), a mouse B7-1 cDNA containing the start and stop codons (stippled box), and part of the hGH gene (solid boxes represent exons; open boxes represent four introns and the 5' and 3' untranslated sequence). E, EcoRI; B, BamHI; pA, polyadenylylation signal.

Immunoperoxidas Analysis. Frozen sections of ear skin punch biopsies were embedded in OCT (Miles) and frozen in cold isopentane. Five-micrometer cryostat sections were stained using the human CTLA4-Ig fusion protein (gift of Peter Linsley, Bristol-Myers Squibb, Seattle) as the primary reagent and the Vectastain Elite ABC kit (Vector Laboratories). Flow Cytometric Analysis of Epidermal Ceils and Cultured Keratinocytes. Epidermal cell suspensions were prepared from adult mouse ear skin by dispase and trypsin treatment (13). Cultures of neonatal transgenic keratinocytes were established in medium containing 0.05 mM Ca2+ (13). To measure B7-1 expression, cells were stained with the antiB7-1 monoclonal antibody (mAb) lGlO (PharMingen) or human CTLA-4-Ig. Secondary reagents used were fluorescein isothiocyanate (FITC)-labeled goat anti-rat Ig (Southern Biotechnology Associates) or FITC-labeled goat anti-human IgG Fc (Caltag, South San Francisco, CA). To quantitate epidermal cell subsets, epidermal cells were stained with FITC-30F11.1 (anti-CD45) and biotinylated KH116 (anti-IAq) (both mAbs from PharMingen) followed by streptavidinphycoerythrin (Immunotech, Westbrook, ME). In Vivo Treatment with IFN-y. To induce expression of class II MHC antigens on keratinocytes, recombinant mouse IFN- y(Genentech; specific activity, 9.8 x 106 units/mg) was administered to groups of two transgenic or control mice. The IFN-y was diluted in 0.5 ml of 0.15 M NaCl containing 0.1% bovine serum albumin and treatments were given daily for eight consecutive days by i.p. injection (14). Groups of mice not receiving IFN-y received daily injections of the diluent only. The mice were sacrificed 24 hr after the final dose of IFN-y, and epidermal cells from the pooled ears of both mice in the group were analyzed by flow cytometry. CHS Responses. B7-1 transgenic and control mice were sensitized to 2,4-dinitrofluorobenzene (DNFB) by application of 25 ,ul of 0.5% DNFB (Sigma) in 3:1 (vol/vol) acetone/ olive oil to shaved abdominal skin on two consecutive days. Five days after the first sensitizing dose, these mice were challenged on the right ear by applying 10 ,l of 0.2% DNFB in 3:1 acetone/olive oil to both the inner and outer surfaces of the ear. Sensitization and challenge of mice to oxazolone (Sigma) were done by applying 10 p1 of 1% oxazolone in ethanol to both surfaces of a single ear or 20 ,u to both rear footpads. Ear thickness was measured prior to challenge and at multiple times after challenge by using engineer's calipers (Dyer, Lancaster, PA). The results obtained with groups of three to five mice are expressed as the increase in ear thickness (in micrometers; mean + SEM). RESULTS K14/B7-1 Transgenic Mice Express B7-1 on the Surface of Basal Keratinocytes. Immunohistochemistry with human CTLA-4-Ig fusion protein was done to examine the cellular distribution of B7-1 in skin of the K14/B7-1 mice. Basal

A~ w..




FIG. 2. B7-1 expression is restricted to the basal epidermis in the skin of K14/B7-1 transgenic mice. Frozen sections of ear skin from an FVB control mouse (A) or a K14/B7-1 transgenic mouse (B) were stained for B7-1 with human CTLA-4-Ig. (xl100.)

keratinocytes in ear skin biopsies from transgenic mice stained strongly with CTLA-4-Ig (Fig. 2); no CTLA-4-Ig staining was detected in skin from controls. To measure more precisely the level of expression of B7-1 by basal keratinocytes, flow cytometric analysis of freshly isolated epidermal cell suspensions was done using both CTLA-4-Ig and a B7-1-specific mAb. The vast majority of epidermal cells in K14/B7-1 transgenic mice expressed high levels of surface B7-1 that could be detected by both reagents (Fig. 3 A and B). Cultured neonatal transgenic keratinocytes were also strongly positive for B7-1 (Fig. 3C). Constitutive Expression of B7-.1 Does Not Cause Spontaneous Cutanleous Inflammation. To determine whether constitutive expression of the B7-1 molecule on the surface of mouse keratinocytes was associated with inflammatory changes in skin or alterations in the leukocyte populations within epidermis, histological analysis of skin biopsies and two-color flow cytometric analysis of epidermal cells were done. Transgenic mice lacked spontaneous inflammatory changes in skin and epidermal maturation100was normal (data 10 B710 expresion 100 101s102icte to 101a 102emi 10 10e not 10Gshown). To probe for changes in the content of dendritic epidermal T cells (DETCs) and Langerhans cells (LCs), epidermal cells from FVB and K14/B7-1 transgenic mice kerainoctesin ersFluborescneinensfomtyaseicmc were stained with antibodies to CD45 and I-q and the percentages of CD45+, I-Aq~+ (LCs), and CD45+, I-Aq(DETCs) cells were determined. Fig. 4A shows that similar CD45/I-Aq contour plots were obtained using epidermal cells from control and B7-1 transgenic mice. Induction of Class II MHC Molecules on Transgenic KeratinIocyts Expressing B7-1 Does Not Provoke Cutaneous Inflammation. In spite of high constitutive B7 expression on basal keratinocytes in situ, inflammatory events were not observed. However, murine keratinocytes do not constitutively express MHC class II molecules. Without expression of MHC class II molecules to bind antigenic peptides, transgenic keratinocytes cannot provide a ligand for the T-cell 60

70 B l

Fluorescence ikL Ansity

FIG. 3. Keratinocytes from K14/B7-1 mice express high levels of surface B7-1 by flow cytometry. Epidermal cell suspensions prepared from K14/B7-1 mice were stained with CTLA-4-Ig (A) or mAb lGlO (B). Primary cultures of neonatal transgenic keratinocytes were also stained with CTLA-4-Ig (C). Overlays of the histograms obtained with the anti-B7-1 reagents (shaded histograms) and the secondary antibody alone (unshaded histograms) are shown.

Immunology: Williams et al.

12782 A 1l4



6 5


a d

1cp 101 10°




.91 10'





B7-1 Tg



Proc. Natl. Acad Sci. USA 91




B7-1 Tg











4~ ~ ~ ~-

6 10

10I 101











-Fio.A4. Constitt4ive expression of B7-1 in mouse epilermiis-does (A) Epidermal cell suspensions from K14/B7-1 mice (B7-1 Tg) and controls were stained with FITC-anti-CD45 and biotinylated 'anti-I-Aq plus streptavidinphycoerythrin and analyzed by flow cytometry. Contour plots are labeled to show the percentage of cells in the upper right quadrant (LCs), lower right quadrant (DETCs),';and lower left quadrant (keratinocytes with a minor contribution of other CD45-' epidermal cells such as melanocytes). (B) K14/B7-1 mice and controls were treated with IFN- y(50,000 units/day) i.p. for eight consecutive days. One day after the last IFN-y injection, epidermal cell suspensions from the ear were stained for CD45 and I-Aq as described above. not induce epidertnal inflammation.

control and transgenic mice were very similar, indicating that coexpression of B7-1 and class II MHC by keratinocytes in the transgenic mice did not induce leukocyte infiltration of the epidermis after IFN-,y treatment. Expression of B7-1 by Transgenic Keratinocytes Alters the Primary Immune Response to Contact Sensitizers. To test the hypothesis that B7-1 expression by epidermal keratinocytes could amplify the host immune responses to contact sensitizers, groups of control and transgenic mice were treated on abdominal skin for two consecutive days with a conventional sensitizing dose of DNFB and challenged on the right ear 5 days after the first dose of DNFB. Both groups demonstrated vigorous ear swelling responses that peaked at 24 hr and then slowly abated (Fig. 5). The magnitude of the secondary CHS response in control mice sensitized with conventional doses of DNFB suggested that this might not be the optimal system for detecting enhancement of CHS in the B7-1 transgenic mice. Therefore, we turned our attention to the primary ear swelling response that peaks 5-6 days after application of contact sensitizing haptens to ear skin of naive mice (15). B7-1 transgenic mice had a much stronger primary response to oxazolone (P < 0.05 at day 6) than controls (Fig. 6). The ear swelling of the transgenic mice decreased after day 6 but remained greater than that of the control mice through day 12. Extended Persistence ofInflammatory Changes at Initial Site of Hapten Exposure in K14/B7-1 Mice. The stronger primary Primary (Right Ear)


Day 25 Challenge

Day 13


Challenge (Left Ear)

(Both Footpads)

Left Ear ---- Control 0

receptors (TCRs) of CD4+ a,B T cells. To determine whether

induction of widespread class II expression by keratinocytes already constitutively expressing B7-1 would result in emergence of inflammatory changes in skin, K14/B7-1 transgenic mice were treated daily for eight consecutive days with IFN-y i.p. (50,000 units/day). Twenty-four hours after the final IFN-y injection, epidermal cells were again stained with antibodies to CD45 and I-An (Fig. 4B). IFN-y treatment induced class II expression on 20-30% of the CD45- keratinocytes in both transgenic and control groups. However, the two-color contour map profiles for the IFN-ytreated


K14/B7-1 Tg



E 100 500

Right Ear



500 -Control -0-

K14/17-1 Tg











Time, days 0



Time, days FIG. 5. Kinetics ofthe secondary CHS response in K14/B7-1 mice sensitized with conventional doses ofDNFB is similar to nontransgenic controls. K14/B7-1 transgenic (Tg) niice (open circles) ornontransgenic controls (solid squares) were rendered DNFB-immune by painting abdominal skin with 25 plofO.5%DI4FB on two consecutive days. Five days after the initial dose, the mice were challenged on the right ear with 20 p1 of 0.2% DNFB. The increase in ear thickness relative to the prechallenge ear thickness was determined at multiple times after challenge and reported as the mean SEM.

FIG. 6. K14/B7-1 transgenic mice have an amplified ear swelling response after the primary epicutaneous application of oxazolone to ear skin. The right ears of hapten-naive K14/B7-1 transgenic (Tg) mice (open circles) or nontransgenic controls (solid squares) were treated with 20 ,u of 1% oxazolone. Thirteen days after the initial hapten painting of the right ear, the left ears of mice in both groups were painted with the same dose of oxazolone. Twenty-five days after initial hapten painting, the footpads of mice in both groups were treated with the same dose of oxazolone. The results are presented as the increase in ear thickness (mean ± SEM). Asterisks indicate times at which the differences between transgenic and control groups were statistically significant at the P < 0.05 level by Student's t test.

Immunology: Williams




FIG. 7. Persistent ear swelling of hapten-treated K14/B7-1 ear skin is associated with chronic inflammatory changes. Mice from the CHS experiment depicted in Fig. 6 were sacrificed 28 days after the initial application of oxazolone and sections of right ear from control (A) and transgenic (B) mice were stained with hematoxylin/eosin. (x80.)

immune response and delayed resolution ofthe CEIS reaction in the K14/B7-1 mice suggested that T-cell priming might be occurrng with greater efficiency after initial priiming on the ear. Responsiveness of mice to challenge with a contact sensitizer normally reaches a peak 3-5 days after. sensitization and then wanes to near background levels by 3 weeks after sensitization (16). We asked whether reexposure of K14/B7-1 mice to hapten during this phase normally characterized by waning responsiveness might reveal a greater residual response. Mice sensitized by a primary application of 1.0% oxazolone to the right ear were challenged 13 days later on the left ear with the same dose of oxazolone. The newly challenged ears and the originally sensitized ears both responded with increased swelling 24 hr after challenge in control and transgenic groups (Fig. 6). The secondary response of the originally sensitized ears of the transgenic group was much greater than in controls (P < 0.05 on days 14, 15, 16, 19, and 20) and substantially exceeded the peak level of the primary response. Thus, the persistent local inflammatory response at the original site of hapten application left this site primed to respond vigorously to a new exposure of the immune system to the same antigen at a different cutaneous site. The secondary response of the newly challenged left ears was of similar magnitude in transgenic and control mice at 24 hr, but the ear swelling resolved more slowly in the transgenic mice resulting in differences at all later time points. Mice that had been sequentially treated with oxazolone on the right and left ears were given a final hapten application to both footpads 25 days after their initial hapten exposure. Both the right and left ears of the K14/B7-1 mice responded with a new burst of ear swelling 24 hr later (Fig. 6); little if any further swelling was detected in the ears of the control mice. On "day 28 (3 days after footpad challenge), ear skin from the right ear of mice in both groups was biopsied. A dense mononuclear cell infiltrate was present in the transgenic ears (Fig. 7).

DISCUSSION Keratinocytes are usually classified as "nonprofessional" APCs because they are far less stimulatory for T cells than "professional" APCs such as dendritic cells, macrophages, and B cells. The absence of critical cell surface costimulatory molecules, such as B7-1, that are required for optimal T-cell activation contributes to the inability of resting keratinocytes to present antigen productively to T cells. Deficient costimulation also favors induction of anergy in T cells exposed to antigen presented by keratinocytes. Keratinocytes activated by certain stimuli, including phorbol esters and cytokines, have been shown to possess increased accessory cell function for T-cell responses (17, 18). However, even maximally

Proc. Natl. Acad. Sci. USA 91 (1994)


activated keratinocytes are substantially less efficient at

providing accessory function on a "per cell" basis when

compared to professional APCs such as cultured LCs. It has been proposed that these negative effects of keratinocyte antigen presentation on host T-cell immunity play an important immunoregulatory role in limiting immune reactivity to antigens encountered in epidermis (4). Since keratinocytes are the most plentiful cell in epidermis, it could be argued that the cumulative effect of multiple nonproductive or anergy-inducing T-cell-keratinocyte interactions would be the net inhibition of ongoing immune responses in epidermis. It has been difficult to test such hypotheses in vivo. We now report the generation" of a transgenic mouse line in which basal epidermal keratinocytes constitutively express high levels of surface B7-1. The generation of this transgenic line permits us to test these hypotheses, at least in part. The major cutaneous findings in these mice are the absence of any secondary effects of B7-1 overexpression in the unperturbed epidermal microenvironment and a selective enhancement of CH$ tespoo$es. The sabence of spontaneois inflammatory changes in the skin of the .K14/B74i transgenic mice- is lilety to be a reflection of the limited access of cells expressing CD28 to epidermis. Among leukocytes, CD28 expression has been detected on T cells and natural killer cells (5, 19). The only lymphocytes normally resident in mouse epidermis are DETC, yO T cells that share an identical TCR and apparently recognize a keratinocyte-derived self antigen (20). Many T cells in the 'y6 lineage, like their a8 counterparts, express CD28 and utilize it as a receptor for costimulatory signals (21). However, intraepithelial y6 T cells from the intestine (22, 23) and the vagina (24) express little if any CD28. To determine whether transgenic keratinocyte B7-1 might bind to CD28 on DETCs in situ, resulting in enhanced signaling upon stimulation of the DETC TCR by locally available antigen, we stained epidermal cells from normal FVB mice with the anti-CD28 mAb 37.51 (25). No staining with antiCD28 mAb was detected in epidermal cell suspensions containing 3% CD3+ DETCs (data not shown). The lack of CD28 expression on DETCs in situ provides a straightforward explanation of why transgenic B7-1 on keratinocytes does not influence DETC number or function. A cutaneous phenotype was elicited in the K14/B7-1 mice by application of contact sensitizers to epidermis. The primary response to haptens, which peaked 6 days after epicutaneous application on ear skin, was significantly enhanced in B7-1 transgenic mice. However, not all parameters of the CHS response were uniformly enhanced in these transgenic mice. The peak ear swelling during the secondary response of transgenic mice sensitized with "conventional" doses (26) of haptens was not significantly greater than in nontransgenic mice, although delayed resolution of the swelling response was consistently noted. We propose that the increased primary ear -swelling response of K14/B7-1 transgenic mice beginning at 6 days (Fig. 6) is the result of local costimulation of effector T lymphocytes after their interactions with B7-1-expressing keratinocytes, leading to increased T-cell cytokine production, further recruitment of leukocytes, and delayed resolution of the inflammatory process. Keratinocytes expressing B7-1 and haptenated peptide-MHC complexes can simultaneously provide two distinct signals to hapten-specific T cells: (i) ligation of the TCR by peptide-MHC complexes on the keratinocyte and (ii) engagement of CD28 by keratinocyte B7-L.-Peptide-MHC complexes formed from both class I and class II MHC molecules are likely to be involved in this direct costimulation pathway, since several studies have shown that both CD4+,'and.CD8+' effector cells mediate CHS responses (27, 28). While direct keratinocyte costimulation of CD8+ effector cells can hypothetically begin immediately after


Immunology: Wflliams et al.

hapten application (because keratinocytes constitutively express class I MHC), costimulation of CD4+ effector cells by the same mechanism cannot begin until keratinocytes are induced to express MHC class II molecules. This relatively late event occurs 3 days after epicutaneous hapten application in sensitized mice (15, 29) and 5 days after skin painting of naive mice (15). K14/B7-1 transgenic keratinocytes also have the potential to augment T-cell activation in the skin by "transcostimulation" (30-32) of T cells interacting with other APCs (e.g., LCs). In vitro transfection studies have shown that B7-1 will costimulate a T-cell response when present on a bystander cell distinct from the APC providing a TCR-ligating stimulus to T cells (30-32), provided the B7-1/CD28 interaction occurs shortly after TCR stimulation (30, 32). While this form of B7-1-mediated costimulation is less efficient than the combined delivery of both signals by a single cell (31), the preponderance of keratinocytes in epidermis may compensate for this inefficiency. Trans-costimulation by B7-1expressing keratinocytes will lead to earlier costimulation of CD4+ effector cells and additional costimulation of CD8+ effector cells during CHS responses. The most dramatic difference in the immune response of the K14/B7-1 mice to contact sensitizers was the ability of a secondary challenge of sensitized transgenic mice at a different cutaneous site to rekindle the smoldering inflammatory response at the original site of hapten painting, leading to a vigorous new round of ear swelling that was substantially greater than the response of nontransgenic mice. This response is an example of the "recall flare reaction" that can occur at local sites (e.g., skin and eye) of prior antigen injection after systemic reexposure to the same antigen (33-35). We postulate that the exaggerated recall flare reaction in the B7-1 transgenic mice reflects the persistent stimulation of residual hapten-specific T cells at the primary site of hapten application. Our analysis of the K14/B7-1 mice chronically treated with IFN-ydemonstrates the importance of exogenous rather than endogenous antigens in initiating a series of cellular interactions that eventually elicits a differential inflammatory response in the transgenic mice. Keratinocyte expression of I-Aq was efficiently induced by IFN-y in these mice, but the concurrent expression of class II MHC and B7-1 alone did not stimulate spontaneous inflammatory changes. We interpret this result as indicating that putative T-cell tolerance to skin autoantigens cannot be overcome simply by presentation of these antigens on keratinocytes coexpressing class II and B7-1. Furthermore, it highlights the necessary role of other mediators in the affected microenvironment, including chemoattractant molecules and adhesion molecules, to instigate a full-fledged inflammatory response. The in vivo effects of constitutive expression of the B7-1 costimulator on nonprofessional APCs have also been recently examined in transgenic mice expressing B7-1 transgenes in pancreatic 8 cells (36, 37). Both groups found that B7-1 expression on pancreatic f8 cells, as an isolated event, did not elicit spontaneous inflammation of pancreatic islets (36, 37). However, just as in our K14/B7-1 mice, additional manipulations of the mice that led to either presentation of an exogenous antigen by pancreatic 8 cells or enhanced presentation of l3-cell autoantigens triggered inflammation within islets that progressed to 3-cell destruction and autoimmune diabetes (37, 38). Thus, all of these B7-1 transgenic mice illustrate the same principle: expression of the B7-1 costimulator by nonprofessional APCs that normally lack this molecule will not stimulate T-cell-dependent immune responses unless antigenic MHC-peptide complexes and nontolerant T cells reactive with these complexes are also provided in the same microenvironment.

Proc. Natl. Acad. Sci. USA 91 (1994) We thank D. Daley, T. Karaoli, and L. Jiang for expert technical assistance; R. Barnhill for suggestions on immunohistochemistry; and J. W. Streilein for helpful comments regarding mechanisms of the recall flare reaction. We thank C. Weaver and E. Fuchs for plasmids and P. Linsley for CTLA-4-Ig. This work was supported by a Dermatology Foundation Career Development Award (I.R.W.), a Thomas B. Fitzpatrick Research Award from the KAO Foundation of Japan (I.R.W.), and National Institutes of Health Grants AI-25082 and AR-40124 (T.S.K.). This project utilized core facilities funded in part by the Harvard Skin Disease Research Center (National Institutes of Health Grant AR-42869). 1. Kupper, T. S. (1990) J. Clin. Invest. 86, 1783-1789. 2. Nickoloff, B. J. & Turka, L. A. (1994) Am. J. Pathol. 143, 325-331. 3. Gaspari, A. A., Jenkins, M. K. & Katz, S. I. (1988) J. Immunol. 141, 2216-2220. 4. Gaspari, A. A. & Katz, S. I. (1991) J. Immunol. 147, 4155-4161. 5. Linsley, P. S. & Ledbetter, J. A. (1993) Annu. Rev. Immunol. 11, 191-212. 6. Lenschow, D. J. & Bluestone, J. A. (1993) Curr. Opin. Immunol. 5, 747-752. 7. Azuma, M., Cayabyab, M., Buck, D., Phillips, J. H. & Lanier, L. L. (1992) J. Exp. Med. 175, 353-360. 8. Harding, F. A., McArthur, J. G., Gross, J. A., Raulet, D. H. & Allison, J. P. (1992) Nature (London) 356, 607-609. 9. McArthur, J. G. & Raulet, D. H. (1993) J. Exp. Med. 178, 16451653. 10. Nickoloff, B. J., Mitra, R. S., Lee, K., Turka, L. A., Green, J., Thompson, C. & Shimizu, Y. (1993) Am. J. Pathol. 142, 1029-1040. 11. Wang, R., Murphy, K. M., Loh, D. Y., Weaver, C. & Russell, J. H. (1993) J. Immunol. 150, 3832-3842. 12. Vassar, R. & Fuchs, E. (1991) Genes Dev. 5, 714-727. 13. Williams, I. R. & Kupper, T. S. (1994) Proc. Natl. Acad. Sci. USA 91, 9710-9714. 14. Gaspari, A. A. & Katz, S. I. (1988) J. Immunol. 140, 2956-2963. 15. Stringer, C. P., Hicks, R. & Botham, P. A. (1991) Br. J. Dermatol. 125, 521-528. 16. Brown, W. R. & Shivji, G. M. (1991) Acta Derm. Venereol. 71, 44-47. 17. Simon, J. C., Cruz, P. D., Jr., Bergstresser, P. R., Davis, L. S. & Tigelaar, R. E. (1991) J. Immunol. 146, 476-484. 18. Nickoloff, B. J., Mitra, R. S., Green, J., Zheng, X.-G., Shimizu, Y., Thompson, C. & Turka, L. A. (1993) J. Immunol. 150, 2148-2159. 19. Nandi, D., Gross, J. A. & Allison, J. P. (1994) J. Immunol. 152, 3361-3369. 20. Havran, W. L., Chien, Y.-H. & Allison, J. P. (1991) Science 252, 1430-1432. 21. Sperling, A. I., Linsley, P. S., Barrett, T. A. & Bluestone, J. A. (1993) J. Immunol. 151, 6043-6050. 22. Gramzinski, R. A., Adams, E., Gross, J. E., Goodman, T. G., Allison, J. P. & Lefrancois, L. (1993) Int. Immunol. 5, 145-153. 23. Ohteki, T. & MacDonald, H. R. (1993) Eur. J. Immunol. 23, 1251-1255. 24. Nandi, D. & Allison, J. P. (1991) J. Immunol. 147, 1773-1778. 25. Gross, J. A., Callas, E. & Allison, J. P. (1992) J. Immunol. 149,

380-388. 26. Kurimoto, I. & Streilein, J. W. (1993) J. Invest. Dermatol. 101, 132-136. 27. Gocinski, B. L. & Tigelaar, R. E. (1990) J. Immunol. 144, 41214128. 28. Kalish, R. S. & Johnson, K. L. (1990) J. Immunol. 145, 3706-3713. 29. Aiba, S., Aizawa, H., Obata, M. & Tagami, H. (1984) Br. J. Dermatol. 111, 507-516. 30. Sansom, D. M., Wilson, A., Boshell, M., Lewis, J. & Hall, N. D. (1993) Immunology 80, 242-247. 31. Liu, Y. & Janeway, C. A., Jr. (1992) Proc. Natl. Acad. Sci. USA 89, 3845-3849. 32. Ding, L. & Shevach, E. M. (1994) Eur. J. Immunol. 24, 859-866. 33. Rapaport, F. T. & Converse, J. M. (1957) Ann. N. Y. Acad. Sci. 64, 836-841. 34. Seegal, D. & Seegal, B. C. (1931) J. Exp. Med. 54, 249-263. 35. Jansen, L. H. & Bleumink, E. (1970) Br. J. Dermatol. 83, 48-55. 36. Guerder, S., Picarella, D. E., Linsley, P. S. & Flavell, R. A. (1994) Proc. Natl. Acad. Sci. USA 91, 5138-5142. 37. Harlan, D. M., Hengartner, H., Huang, M. L., Kang, Y.-H., Abe, R., Moreadith, R. W., Pircher, H., Gray, G. S., Ohashi, P. S., Freeman, G. J., Nadler, L. M., June, C. H. & Aichele, P. (1994) Proc. Natl. Acad. Sci. USA 91, 3137-3141. 38. Guerder, S., Meyerhoff, J. & Flavell, R. (1994) Immunity 1, 155-166.

Suggest Documents