1756 S. Miiller et al. Eur. J. Immunol. 1997.27: 1756-1761

Stefan Miiller', Myriam Jungo', Peter Aichele' and Christoph Mueller'

' Department of Pathology, University of Bern, Bern, Switzerland

* Institute of Experimental Immunology, University Hospital, Zurich, Switzerland

CD5- CD8a p intestinal intraepithelial lymphocytes (IEL) are induced to express CD5 upon antigen-specific activation: CD5- and CD5+ CD8ap IEL do not represent separate T cell lineages

We followed ap T cell receptor (TCR) usage in subsets of gut intraepithelial lymphocytes (IEL) in major histocompatibility complex class I-restricted ap TCR-transgenic (tg) mice. The proportion of tg ap TCR' CD8aP IEL is reduced compared with CD8' splenocytes of the same animal, particularly under con- ventional conditions of maintenance. Further fractionation of CD8aP IEL according to the expression level of surface CD5 revealed that in conventionally housed animals tg TCR'CDS-CD8aP IEL are as frequent as in specific pathogen-free (SPF) mice, whereas tg TCR' CD5'"' or, even more pronounced, tg TCR' CDSh' CD8aP IEL are greatly diminished when compared with mice kept urider SPF conditions. Upon antigen-specific stimulation of CD5- CD8aP IEL in v i m , CD5 surface expression is up-regulated on a large fraction of cells within 48 h. Up-regulation of CD5 surface expression is further enhanced by the pres- ence of the anti-aIEL monoclonal antibody 2E7. This clearly demonstrates that CD5- , and CD5' CD8aP IEL cannot be considered as separate T cell lineages.

1 Introduction

Despite an increasing number of reports on the biology of intestinal intraepithelial lymphocytes (IEL) during the past few years, IEL still remain a somehow enigmatic T cell compartment in terms of developmental pathways [l , 21. IEL are distinguished from conventional peripheral T cells by their unique cellular composition [3]: conven- tional peripheral T cell markers like Thy-1 and CD5 have been reported to be absent on extrathymically derived IEL [4, 51. The absence of CD5 surface expression was also observed on a fraction of CD8aP IEL [3] which, in con- trast to CD8aa IEL are generally believed to represent intrathymically differentiated T cells.

CD5 is associated with the T cell receptor (TCR)-CD3 complex on the T cell surface and its cytoplasmic tyrosine residues are rapidly phosphorylated upon T cell activation [6]. In the thymus, CD5 has been shown to be a compo- nent of a system that regulates TCR-mediated thymocyte maturation at the stage of CD4+ CD8' double-positive thymocytes, depending on the intensity of initial TCR- ligand affinity [7]. CD5 has also functional significance on

[I 165521

Received December 4, 1996; in revised form March 27, 1997; accepted April 18, 1997.

Correspondence: Christoph Mueller, Department of Pathology, University of Bern, Murtenstrasse 31, CH-3010 Bern, Switzerland Fax: +41-31-3 81 87 64; e-mail: muellerc@patho.unibe.ch

Abbreviations: IEL: Intestinal intraepithelial lymphocyte tg: Transgenic SPF Specific pathogen free GP: Glycoprotein gp33: Amino acids 33-41 of LCMV glycoprotein int: Intermedi- ate hi. High

Key words: Intraepithelial lymphocyte / T cell receptor / Trans- genic mouse / CD5 I Phenotype

mature peripheral T cells in that anti-CD5 mAb mediate co-stimulatory effects on TCR-mediated proliferation of peripheral T cells [8-lo].

Two reports on developmental T cell selection in IEL in an anti-male H-Y antigen-specific ap TCR-transgenic (tg) mouse model yielded, in part inconsistent, results on the selection of distinct IEL subsets. While Rocha et al. [ll] restricted the analysis to CD8aa (thymus-independent) IEL, Poussier et al. [12] also found variable usage of a pos- itively selected tg ap TCR within CD8aP IEL in terms of tg TCR expression levels. The composition of IEL subsets, however, is strongly dependent on antigenic stimulation by environment and food intake [13, 141. Thus, it is important to assess, whether, and to what extent, different conditions of maintenance [ i . e . specific pathogen free (SPF) vs. con- ventional] will affect the usage of a given tg TCR within distinct IEL populations.

In the present study we attempted to address usage of a tg TCR in phenotypically distinct subpopulations of CD8aP IEL, defined by the level of CD5 surface expression, in mice kept under different conditions of maintenance. We further attempted to determine the lineage relationship of CD5-, CD5'"' and CDSh'CD8aP IEL to assess whether CD5-, and CD5' CD8aP IEL represent distinct T cell lin- eages. For this purpose we chose an aP TCR tg mouse model specific for the lymphocytic choriomeningitis virus (LCMV) glycoprotein (gp)-derived epitope gp33 [ 151.

2 Materials and methods

2.1 Normal and tg mice, tg TCR-specific peptide

C57BL/6 mice were obtained from the Institut fur Labor- tierkunde, Universitat Zurich, Switzerland. The genera- tion of anti-LCMV gp33-specific TCRaP (line 327) tg mice has been described [15, 161. gp33 peptide (sequence:

OO14-2980/97/0707- 1756$17.50 + .50/0 0 VCH Verlagsgesellschaft mbH, D-69451 Weinheim, 1997

Eur. J. Immunol. 1997.27: 1756-1761 CD5 up-regulation on CD5- CD8ap IEL 1757

KAVYNFATM) was a generous gift of Dr. H. P. Pircher and originally purchased from Neosystem Laboratoire (Strasbourg, France).

2.2 Monoclonal antibodies

Biotin- and fluorescence-conjugated mAb were purchased from either Pharmingen (San Diego, CA): B20.1 (anti-Va 22) [17], 53-5.8 (anti-CD8P) [18], and 2E7 (anti-aIEL) [19]; or Sigma (St. Louis, MO): 29B (anti-CD3) [20], H129.19 (anti-CD4) [21] and 53-6.7 (anti-CD8a) [22]; or were purified by protein G from supernatants of hybrid- omas originally obtained from ATCC and subsequently labeled according to standard protocols [23]: H57-597 (anti-TCRap) [24], GL3 (anti-TCRyG) [25], 53-7.3 (anti- CD5) [22], 53-6.7, F4/80 (specific for mature macro- phages) [26] and MU70 (anti-CDllb) [27]. Streptavidin- RPE/Cy5 (Southern Biotechnology Associates, Birming- ham, GB) or Streptavidin-Tricolor (Caltag Laboratories, San Francisco, CA) were used as second step reagents.

2.3 Single-cell preparations

2.3.1 IEL

IEL were isolated according to a method adapted from Lefrangois [28]. Briefly, IEL were dissociated from small intestinal pieces in Ca2+- and Mgf-free (CMF) HBSS con- taining 5 % horse serum (HS; Gibco-BRL, Life Technolo- gies, Gaithersburgh, MD), 1 mM dithiothreitol and 0.5 mM EDTA. Cells were serially passed through 70- and 40-pm nylon mesh cell strainers (Falcon, Becton Dickin- son, San Jose, CA) and incubated for 30 min at 37°C in 5 YO C02 before purification on a 44 %/67 YO Percoll (Phar- macia Biotech, Uppsala, Sweden) gradient (15 min, 800 x g at room temperature).

2.3.2 Splenocytes

Spleens were serially passed through 70- and 40-pm cell strainers, and washed once with HBSS + 5 'YO HS. Eryth- rocytes were lysed by osmotic shock.

2.4 Cell staining, cytofluorometry and cell sorting

Cells were resuspended at 106/100 pl HBSS containing 5 % HS and 0.05'30 NaN,; 0.25-0.5 pg primary mAb, or second-step reagent were used per lo6 cells. Incubation time was 15 rnin on ice for both the first and second step. Cells were analyzed on a FACScan or sorted on a FACS Vantage (Becton Dickinson). Analysis was carried out using Macintosh-based CellQuest software, and cell sort- ing was controlled by Pascal-based Lysys I1 software (Bec- ton Dickinson).

2.5 In vitro culture of IEL

Sorted IEL (2.5 x 104-4.5 x lo4) were added to 1 x 10'-2 x lo5 syngeneic bone marrow-derived macrophages per well in a U-bottom 96-well plate (Costar). Plates were cen- trifuged (3 min, 280 x g at room temperature), superna- tants were aspirated and 200 p1 IMDM containing 10% FCS and 10 pM gp33 per well were added; 6 U IL-2, 0.5 pg 2E7, or 6 U IL-2 plus 0.5 pg 2E7 were added to appropriate wells. Plates were incubated for 24 or 48 h at 37 "C in an atmosphere containing 5 % C02 . After incuba- tion, cells were detached with cell dissociation solution and harvested by vigorous pipetting for subsequent stain- ing. Statistical significance between two groups was as- sessed using the Student's t-test.

2.6 Autoradiography of sorted IEL

Mice were injected intraperitoneally with 5 pCi [3H]dTTP (Amersham Laboratories, Amersham, GB) 4 h before sacrifice. After isolation, IEL were sorted directly onto poly-L-lysine-coated slides (PolysineTM, Menzel Glaser, Braunschweig, Germany) according to their cell surface phenotype and fixed for 20 rnin in 4 % buffered parafor- maldehyde, washed sequentially in 3 x PBS, 1 x PBS and H20 . When dry, slides were dipped in NTB-2 emul- sion (Eastman-Kodak Company, New Haven, CT) at 43"C, dried and exposed for 2-4 weeks at 4°C in the dark. After exposure, slides were developed in D-19 developer, fixed in GBX fixer (Eastman-Kodak Company) and counterstained with nuclear fast red (0.05 YO in 5 % alumi- num sulfate).

2.3.3 Bone marrow-derived macrophages 3 Results

Bone marrow was collected from both femurs of a TCRafJ tg mouse and cultured in Iscove's modified Dulbecco's medium containing 5 % FCS and 5 YO HS with additional 20% of supernatant of the fibroblast cell line L929 as a source of macrophage (M)-CSF. After 10 days, >95 % of cultured cells were positive for the macrophage-specific markers recognized by mAb MU70 and F4/80. Bone marrow-derived macrophages were subsequently incub- ated for 15 min at 37 "C in 5 YO C 0 2 atmosphere in a petri dish with cell dissociation solution (Gibco-BRL). Mac- rophages were harvested by washing the petri dish and removing adherent cells with a cell scraper (Costar, Camb- ridge, MA).

3.1 Phenotypic analysis of IEL subsets and tg TCRaS usage in conventional vs. SPF TCRafi tg C57BL/6 mice

To assess the influence of maintenance conditions on IEL composition and tg TCR usage, we compared a series of conventionally raised and housed TCRaP tg C57BL/6 mice that were age-matched with animals kept under SPF conditions. Results of the phenotypic analysis are shown in Table 1. The ratio of TCRaP' to TCRyG' IEL differs dra- matically in the two groups of mice: we find a ratio of 1 : 2.1 in SPF vs. 2.2 : 1 in conventionally housed mice. Notably, even within the TCRaP' IEL significant alterations in the proportions of the subsets are observed. The increased fre- quency of TCRap+ IEL in conventionally housed mice

1758 S. Muller et al. Eur. J. Immunol. 1997.27: 1756-1761

Table 1. Frequencies of lymphocyte subsets assessed by two- and three-color FACS analysis in the spleen and the intestinal epithelium of 8-10-week-old SPF (n = 4), and age-matched conventionally housed (n = 3), TCRaP tg C57BU6 (H-2Db) mice, respectively (mean f SE)

Lymphocyte compartmenta' Spleen IEL

Subset SPF Conventional SPF Conventional

CD3' b, 39.4 k 1.2 43.8 f 3.2 85.7 f 3.8 90.0 f 2.8 TCRaP' '). dl 91.7 f 2.0 91.2 f 2.9 35.4 f 1.5 71.1 f 1.8 TCRyb' 8.1 f 0.8 4.9 f 1.0 73.5 f 2.2 32.4 f 1.2 CD4' ') 14.1 f 1.8 17.8 f 1.1 6.6 f 0.9 7.0 k 0.7 CD8aP' ') 79.6 k 2.5 71.3 f 1.9 21.0 f 2.7 59.6 k 0.9 TCRafV CD8aa' ') N D ND 11.4 f 1.0 10.7 f 1.1

a) The lymphocyte gate was determined on an FSC vs. SSC plot. b) In % of total lymphocytes. c) In % of total CD3' cells. d) The pan TCRaP-specific mAb H57-597 recognizes all, tg and endogenous ap TCR.

compared to SPF animals is due to a massive increase of the CD8aP' subpopulation, whereas relative frequencies of CD4' and TCRaP" CD8aa' IEL are comparable in both SPF and conventionally housed mice.

A comparison of the tg TCRaP (VaWp8.1) usage, as- sessed by the expression of the Va2 chain, among TCRap' subsets isolated from SPF and conventionally housed mice, respectively, is shown in Fig. 1: the number of CD8ap IEL expressing the tg ap TCR is drastically decreased in conventionally housed compared to SPF mice (33.7 f 7.2 YO Va2' vs. 75.3 k 4.0 YO Va2+, respectively). On the other hand, tg TCR usage is comparable in SPF and conventionally housed mice for other TCRaP' sub- sets, including TCRaP' CD8aa" and CD4' IEL (Fig. 1). In TCRaP tg BALB/c mice with the non-selecting H-2d MHC haplotype, Va2 usage does not significantly differ from non-tg control mice in any T cell subset (data not shown).

3.2 Differential selection of the tg af3 TCR in CDS- and CDS' CD8af3 IEL

The comparative analysis of CD5 expression on CD8aP IEL, and CD8aP splenocytes revealed that according to their CD5 surface expression, three different populations

SPF conventional

CD4+ CDBas+

TcRap+ CD0aa+ CD4+

CD8a$+ TcRafP C D b +

0 25 50 75 100 0 25 50 75 100

% Va2 expressing cells in the respective T cell subset (mean f SE)

Spleen

IEL

Figure 1. Tg TCRaP usage in lymphocyte subsets analyzed by three-color FACS analysis in splenocytes and IEL of 8-10-weeks- old SPF (n = 4), and age-matched conventionally housed (n = 3) TCRaP tg C57BL/6 (H-2Db) mice. Bars indicate relative frequen- cies of tg (Va2') TCRaP expressing cells within respective subsets (mean f SE).

of CD8aP' IEL can be distinguished, i.e. CD8ap IEL lacking CD5 expression (CD5-), CD8aP IEL with inter- mediate (CD5'"') and high (CDSh') CD5 expression. The CDSh' subpopulation of CD8aP IEL express the CD5 mol- ecule at comparable levels as CD8' splenocytes (Fig. 2). While normal C57BL/6 mice generally lack large numbers of CD5-CD8aP IEL (data not shown), anti-LCMV- specific TCRaP tg mice of the selecting MHC haplotype regularly show substantial proportions of CD5- and CD5'"'CD8aP IEL (Figs. 2 and 4). In TCRaP tg mice of the selecting MHC haplotype, raised and housed under SPF conditions, CD5-, CD5'"' and CDShiCD8ap IEL show comparable high frequencies (71.9-94.2 YO) of Va2 expressing T cells (Fig. 3, upper panel). In TCRap tg C57BW6 mice kept under conventional conditions, again high frequencies of CD5- CD8aP IEL expressing Va2 (67.1 f 6.3%) are detected. However, Va2 usage is greatly diminished in CD5'"' (35.7 k 9.1 % Va2') and even more pronounced in CDSh'CD8aP IEL subsets (21.2 f 8.3 % Va2") (Fig. 3, lower panel).

When littermates of TCRap tg C57BL/6 mice kept under SPF conditions are transferred to a conventional animal facility 6 weeks prior to analysis, the following alterations

CDSc@+ T cells

I

t'.,' '. - log CDS

Figure 2. CD5 expression levels in CD8aP IEL (solid line) and splenocytes (dashed line), determined by two-color FACS analy- sis.

Eur. J. Immunol. 1997.27: 1756-1761

1 5 1

CD5 up-regulation on CD5- CD8afi IEL 1759

CD5int conventional CDS-

0 25 50 75 100

% Va2 expressing cells (mean k SE)

Figure 3. Usage of the tg afi TCR within distinct CD8aP IEL subsets, defined by the absence of CD5 (CD5-), CD5 intermedi- ate (int), and high (hi) expression, from SPF ( n = 4), and age- matched, conventionally (n = 3) raised and housed TCRaP tg C57BL/6 (H-2Db) mice. Bars indicate relative frequencies of Va2 expressing cells within respective CD8afi IEL subsets (mean f SE).

in the composition of the IEL compartment are observed in comparison to the littermates continuously kept under SPF conditions: (i) an increased proportion of CD5'"' and CDSh' vs. CD5- CD8aP IEL, and (ii) decreased frequen- cies of Va2 expressing cells in CD5'"', and even more pro- nounced, in CDSh' CD8aP IEL (Fig. 4). Tg a@ TCR usage in CD5- CD8aP IEL is only slightly decreased when mice transferred to conventional conditions of maintenance are compared with mice kept in the SPF facility: CD5- CD8aP IEL of mice transferred to the conventional facility still express Va2 at high frequencies of 68.2 f 3.8%. How- ever, CD5'"' and particularly CDSh' CD8aP IEL show sig- nificantly decreased frequencies of Va2 expressing cells (57.3 f 5.0%, and 35.8 + 3.1%, respectively), com- pared to their littermates still kept in the SPF facility (80.4 k 3.2% and 84.1 k 4.3%, respectively).

In absolute numbers, from the small bowel of four SPF animals, 3.77 (+ 0.26) x 105 CD5-CD8aP IEL and 2.57 (f 0.52) X lo5 CD5'CD8aS IEL are isolated, whereas lower numbers

t=Oh

CD5

of CD5-CD8aP IEL

CD5- CDSint CDShi CD5- CD5int CDShi

SPF SPF -> conventional

CD8ap IEL subset

Figure 4 . Tg TCRap usage in CD8afi IEL subsets, defined by the absence of CD5 (CD5-), CD5 intermediate (int) or CD5 high (hi) expression isolated from TCRaP tg C57BU6 (H-2Db) mice, kept under SPF conditions; and littermates, transferred from SPF to a conventional facility 6 weeks before analysis, respectively. Stacked bars indicate frequencies (in % of total CD3' IEL) of endogenous (Va2-, open bar) and tg afi TCR (Va2*, hatched bar) expressing cells in the CD5-, CD5'"', and CDSh' CD8ap IEL subsets. Bars represent means of four mice per group.

[2.56 (+ 0.32) x lo'] and higher numbers of CD5' CD8aP IEL [4.37 (+ 0.87) X 1051 are isolated from the small bowel of three conventionally housed ap TCR tg C57BL/6 mice.

3.3 CD5 up-regulation on purified CD5- CD8afl IEL in vitro

t = m t-48h

To test, whether CD5- and CDYCD8aP IEL represent distinct T cell lineages, or whether CD5+ CD8aP IEL may originate from CD5- CD8aP IEL purified CD5- CD8aP IEL from TCRaP tg mice were primed with specific anti-

- CD5

8P33

gp33 + IL-2

gp33 + 2E7

gP33 + IL2 + 2E7

Figure 5. Up-regulation of CD5 sur- face expression on FACS-sorted CD5- CD8afi IEL in vitro. Sorted cells were cultured with bone marrow-derived macrophages in the presence of the tg TCRap-specific peptide gp33 (first row), gp33 and IL-2 (second row), gp33 and activating anti-aIEL integrin mAb 2E7 (third row) or gp33, IL-2 and 2E7 (fourth row). CD5 surface expression on cultured IEL was analyzed after 24 h (left column) and 48 h (right column). Percentages indicate frequencies of CD5- (left marker), CD5 intermediate (middle marker) and CD5 high (right marker) expressing cells. Fluorescence intensities are shown in logarithmic scale. Similar results were obtained in five independent experiments.

S. Miiller et al.

CD8ap IEL subset

Figure 6. In vivo ['HH]dTTF' incorporation by CD5 , and CD5', CD8ap IEL, respectively. TCR tg C57BU6 mice ( n = 3) received 50 pCi ['HldTTP intraperitoneally 4 h before sacrifice and IEL isolation (mean k SE).

gen presented by autologous bone marrow-derived mac- rophages in vitro. As shown in Fig. 5, after a 24-h culture in vitro a gradual up-regulation of CD5 surface expression in the starting population of CD5- CD8aP IEL is clearly visible, particularly when the cells are additionally stimu- lated with anti-aIELP7 integrin antibody. In contrast to the activating anti-aIEL mAb (clone 2E7), IL-2 seems to exert only minor effects in propagation of CD5 up- regulation in vitro. After 48 h of culture in the presence of 2E7 mAb, about 70-80 YO of the initial CD5- CD8aP IEL express CD5 at intermediate and almost 10 YO even at high levels, whereas less than 25 YO remain C D Y . When puri-

instead of CD5- CD8aP IEL are cultured in the same way as described above, the proportions of CD5'"' and CDShi expressing CD8aP IEL do not change significantly during the 48 h of culture irrespective of addi- tional stimulation with IL-2 and/or 2E7 mAb (data not shown). Sorted CD5- and CD5'"'/hi CD8aP IEL were both >96% pure at the beginning of the in vitro culture. Recovery of IEL after the 48-h culture period and staining of the cells did not differ significantly (p > 0.4) between CD5-, and CD5'CD8aP IEL cultures in four indepen- dent experiments (26.1 k 5.4%, and 35.9 f 10.0%, respectively). When a H-2d-restricted, irrelevant peptide is added to the cultures instead of gp33, CD5 is not up- regulated on IEL and eventually, both purified C D Y , and CD5' CD8aP IEL are eliminated during a 48-h culture in vitro (data not shown).

fied c ~ ~ i n t l h i

3.4 Proliferation of CD5- and CD5+ CD8uf3 IEL in vivo

In vivo proliferative activity was measured to further sub- stantiate whether CD5- CD8aP IEL represent predomi- nantly naive T cells, not (yet) specifically stimulated, and whether CD5' CD8aP IEL are more mature T cells that have already been specifically stimulated and, at least in part, clonally expanded as a result of recent antigen- specific activation. Injection of [3H]d'ITP into convention- ally housed TCRaP tg H-2Db mice ( n = 3) and isolation of IEL 4 h later with subsequent cell sorting and autoradi- ography revealed that CD5' CD8aP IEL show prolifera-

Eur. J. Immunol. 1997.27: 1756-1761

tive activity (1.88 k 0.26 YO positive cells in autoradiogra- phy), whereas only 0.27 f 0.17% of CD5-CD8aP IEL incorporated [3H]dTTP in vivo (Fig. 6).

4 Discussion

Studies on the usage of a given positively selected tg TCR by IEL in the gut are hampered by the fact that the effects of positive and negative selection of extrathymically and intrathymically derived T cells are largely disguised by the clonal expansion of antigen-reactive T cells at this site of T cell effector function. This difficulty is further substanti- ated by the lack of distinct cell lineage markers, resulting in yet poorly defined lineage relationships between differ- ent IEL subsets.

Therefore, we attempted to follow the usage of a positively selected tg TCR in mice kept under SPF, and conventional conditions of maintenance, respectively. The dramatic effects of housing conditions on the cellular composition of IEL have been reported for the relative frequency of TCRaP, and TCRys IEL [13] and are also clearly visible in our analysis of distinct CD8aP JEL subsets. The propor- tion of CD5' CD8aB IEL increases almost threefold upon transfer of SPF animals to conventional housing conditions (Fig. 4). In parallel, a reduced usage of the tg TCR is observed in CD5' CD8aP IEL. These two observations may be attributed to a selective expansion and/or enhanced recruitment of CD5' CD8aP IEL with an en- dogenous TCR, triggered by antigenic peptides present in conventional animals, or the expansion of both CD5- and CDS' CD8aP IEL with an endogenous surface TCR, with concomitant up-regulation of CD5 on CD5- CD8aP IEL. To discriminate between these possibilities, we analyzed expression of CD5 on CD8aP IEL upon antigen-specific stimulation. Addition of the anti-aIEL mAb [19] to these cultures should mimic the environment of IEL since the ligand of aIEL, E-cadherin, is present on intestinal epi- thelial cells [29, 301. The results obtained are remarkable in two respects: first, they indicate the importance of the aIEL-ligand interaction which can even substitute for the requirement of exogenous 1L-2 for the induction of surface CD5 expression upon antigen-specific activation, and sec- ond, they clearly demonstrate that C D Y , and CD5' CD8aP IEL do not represent mutually exclusive T cell lineages and that, therefore, presence or absence of CD5 expression cannot be considered a marker for intra- thymically, and extrathymically derived T cells, respec- tively. Furthermore, stimulation-dependent regulation of CD5 surface expression on CD8aP IEL is also supported by the observed down-regulation of CD5 from high to intermediate levels on freshly isolated CDShi CD8aP IEL during a 20-h "Cr-release assay (manuscript in prepara- tion). The increase in the frequency of CD5'"' or CDShi CD8ap IEL within 48 h after antigenic stimulation cannot be attributed to selective expansion of contaminat- ing CD5' CD8aP IEL in the starting population because: (i) CD5 expression gradually increases during antigen- specific stimulation, (ii) the number of CD8aP IEL re- covered after in vitro culture is comparable when CD5-, and ~~5 in l /h i CD8aP IEL are used as a starting population, and (iii), after 24 and 48 h of culture in vitro, both initially CD5- and CD5i"'/hi CD8aP IEL had incorporated ['H]d?TP at comparable low frequencies (data not

Eur. J. Immunol. 1997.27: 1756-1761 CD5 up-regulation on CD5- CD8aP IEL 1761

shown). The results of our in vitro experiments using IL-2 andor 2E7 suggest that intestinal CD8aP T cells may also require local stimuli for activation, as has been previously described for TCRaP+ CD8aa IEL [12, 311. Increased functional activity of IEL upon interaction with the stimu- lating anti-aIEL mAb HML-1 has been previously reported for human IEL [32].

Our data on CD5 up-regulation on previously CD5- CD8aP IEL in vitro can be interpreted in line with the finding that CD5+, but not CD5- CD8aP IEL expressing endogenous TCR, selectively expand upon transfer of mice from SPF to conventional conditions: we suggest that CD5- CD8aP IEL expressing endogenous TCR may con- tain T cells specific for MHC class I-restricted antigens in the intestine of conventional mice. Such activated CD5- CD8aP IEL up-regulate CD5 surface expression upon clonal expansion leading to the increased proportion of CD5+CD8aP IEL expressing an endogenous TCR as shown in Fig. 4. A transition of CD5- to CD5+ in v i vo rather than an outgrowth of CD5+CD8aP IEL alone is also suggested by the predominance of CD5- CD8aP IEL over CD5+CD8aP IEL in SPF animals, whereas in con- ventionally housed animals CD5' CD8aP IEL predomi- nate and total numbers of CD8aP IEL do not differ signifi- cantly between these two groups ( p > 0.6). Although we cannot formally rule out the possibility that CD5+ CD8a P T cells bearing an endogenous TCR are recruited from other sites, the reported absence of T cell traffic from peripheral lymphoid organs to the intestinal epithelium in immune competent mice [31] argues against an important contribution by this pathway. The observed sevenfold higher proliferation rate in CD5+ CD8ap IEL compared to CDYCD8aP IEL in vivo, is also compatible with our hypothesis of an activation induced transition from CD5- to CD5+ CD8aP IEL concomitant with proliferative expansion.

In summary, in addition to IEL expressing a CD8aa homodimer, a subset of CD8aP IEL is also negative for CD5 surface expression. In this subset, absence of CD5 expression reflects merely a functional state rather than a lineage distinct from CD5' CD8aP IEL. The absence of CD5 may affect TCR signaling requirements similar to TCRaP' CD8aa IEL, which are refractory to stimulation in vitro with specific antigen and IL-2 [12]. The capability, or inability, of a CD5- T cell to up-regulate CD5 upon acti- vation might be a striking difference between generally considered thymus-derived CD8aP and putatively extra- thymically differentiating CD8aa IEL.

The authors would like to thank Hanspeter Pircher for providing transgenic mice and peptide, and Hans Hengartner and Jean A . Laissue for helpful discussions and support. This work was sup- ported by grant 31 -43495.95 of the Swiss National Science Founda- tion to Christoph Mueller.

5 References

1 Lefrangois, L. and Puddington, L., Irnrnunol. Today 1995. 16:

2 Poussier, P. and Julius, M., J. Exp. Med. 1994. 180: 1185. 3 LefranFois, L., J . Immunol. 1991. 147: 1746. 4 Ferguson, A. and Parrot, D. M. V., Clin. Exp. Immunol. 1972.

16.

12: 477.

5 Mayrhofer, G., Blood 1980. 55: 532. 6 Davies, A. A,, Ley, S. C. and Crumpton, M.J., Proc. Natl.

Acad. Sci. USA 1992. 69: 6368. 7 Tarakhovsky, A,, Kanner, S. B., Hombach, J. , Ledbetter, J.

A., Miiller, W., Killeen, N. and Rajewsky, K., Science 1995. 269: 535.

8 Ledbetter, J. A., Martin, P. J., Spooner, C. E., Wofsby, D., Tsu, T. T., Beatty, P. G. and Gladstone, P., J . Irnmunol. 1985. 135: 2331.

9 Ceuppens, J. L. and Baroja, M. L., J . Imrnunol. 1986. 137: 1816.

10 McAteer, M. J., Lagarde, A. C., Georgiou, H. M. and Bell- grau, D., Eur. J . Irnmunol. 1988. 18: 1111.

11 Rocha, B., von Boehmer, H. and Guy-Grand, D., Proc. Natl. Acad. Sci. USA 1992. 89: 5336.

12 Poussier, P., Teh, H. S. and Julius, M., J . Exp. Med. 1993. 178: 1947.

13 Bandeira, A., Mota-Santos, T., Itohara, S., Degermann, S., Heusser, C., Tonegawa, S. and Coutino, A., J . Exp. Med. 1990. 172: 239.

14 Poussier, P. and Julius, M., Annu. Rev. Immunol. 1994. 12: 521.

15 Pircher, H.-P., Biirki, K., Lang, R., Hengartner, H. and Zin- kernagel, R. M., Nature 1989. 342: 559.

16 Brandle, D., Miiller, C., Riilicke, T., Hengartner, H. and Pir- cher, H.-P., Proc. Natl. Acad. Sci. USA 1992. 89: 9529.

17 Pircher, H.-P., Rebai, N., Groettrup, M., GrCgoire, C., Spei- ser, D. E., Patt Happ, M., Palmer, E., Zinkernagel, R. M., Hengartner, H. and Malissen, B., Eur. J . Immunol. 1992. 22: 399.

18 Ledbetter, J. A., Rouse, R., Micklem, S. and Henenberg, L. A., J. Exp. Med. 1980. IS2: 280.

19 Lefrangois, L., Barrett, T. A., Havran, W. L. and Puddington, L., Eur. J . Immunol. 1994.24: 635.

20 Portole, P., Rojo, J., Golby, A., Bonneville, M., Gromkowski, S., Greenbaum, L., Janeway, C. A. Jr., Murphy, D. B. and Bottomly, K., J . Immunol. 1989. 142: 4169.

21 Pierres, A., Naquet, P., Van Agthoven, A., Bekkhoucha, F., Denizot, F., Mishal, Z., Schmitt-Verhulst, A.-M. and Pierres, M., J. Immunol. 1984.132: 2775.

22 Ledbetter, J. A. and Henenberg, L. A., Irnrnunol. Rev. 1979. 47: 63.

23 Holmes, K., Fowlkes, B. J., Schmid, I. and Giorgi, J. V., in Coligan, J. E., Kruisbeek, A. M., Marguelies, D. H., She- vach, E. M. and Strober, w. (Eds.), Current Protocols in Immunology, Wiley, New York 1996, p. 5.3.5.

24 Kubo, R. B., Kappler, J. W., Marrack, P. and Pigeon, M., J . Imrnunol. 1989. 142: 2736.

25 Goodman, T. and Lefrangois, L., J. Exp. Med. 1989. 170: 1569.

26 Austyn, J. M. and Gordon, S., Eur. J . Irnrnunol. 1981.11: 805. 27 Springer, T., Galfre, G., Secher, D. S. and Milstein, C., Eur. J .

Imrnunol. 1979. 9: 301. 28 LefranFois, L., in Coligan, J. E., Kruisbeek, A. M., Margue-

lies, D. H., Shevach, E. M. and Strober, W. (Eds.), Current Protocols in Immunology, Vol. 1, Wiley, New York 1993, p. 3.19.1.

29 Lefrangois, L., Barrett, T. A., Havran, W. L. and Puddington, L., Eur. J . Imrnunol. 1994.24: 635.

30 Cepek, K. L., Shaw, S. K., Parker, C. M., Russell, G. J. , Mor- row, J. S., Rimm, D. L. and Brenner, M. B., Nature 1994.372: 190.

31 Karecla, P. I., Bowden, S. J., Green, S. J. and Kilshaw, P. J., Eur. J . Immunol. 1995. 25: 852.

32 Poussier, P., Edouard, P., Lee, C., Binnie, M. and Julius, M., J . Exp. Med. 1992.176: 187.

33 Sarnacki, S., Begue, B., Buc, H., Le Deist, F. and Cerf- Bensussan, N., Eur. J . Imrnunol. 1992.22: 2887.