Detection of a Tyrosine Phosphatase LAR on Intestinal Epithelial Cells and Intraepithelial Lymphocytes in the Human Duodenum

Studies of tyrosine phosphorylation in the human duodenum have indicated that proliferating cells in the middle portion of the duodenal crypt were devoid of this feature, suggesting that tyrosine kinase activation is not a dominant factor in crypt cell proliferation, and that consequently tyrosine phosphatase activity may be a more critical factor in crypt cell development. We investigated the expression of the leukocyte common antigen-related receptor (LAR) family of tyrosine phosphatases. A flow cytometry system was used to examine cells from the surface, mid-portion, and lower part of the crypt. Individual cell populations were immunostained with anti-LAR antibodies using phycoerythrin-conjugated anti-CD3 to discriminate between epithelial cells (CD3−) and intraepithelial lymphocytes (CD3+). Epithelial cells expressed LAR throughout the crypt. Expression of LAR was maximal in the mid-portion of the crypt with lower expression at the top of the villi. Intraepithelial lymphocytes expressed low levels of LAR at the tips of the villi with stronger expression extending towards the base of the crypt. These findings were confirmed by immunohistochemistry on paraffin-fixed sections. Of note, peripheral blood lymphocytes expressed less LAR than IEL. These observations suggest the possibility that tyrosine phosphatase LAR may be of importance in the regulation of crypt cell proliferation. Moreover, as the extracellular domain of LAR has homology with adhesion molecules, the finding of this molecule on IEL could suggest a possible functional role in homing of this unique lymphocyte.


INTRODUCTION
Protein tyrosine phosphatases (PTPases) comprise a structurally diverse family of transmembrane and cytoplasmic-type enzymes that, together with protein tyrosine kinases (PTKases) control protein tyrosine phosphorylation [1]. Tyrosine phosphorylation is known to regulate such physiological events as cell-cell, and cellmatrix interactions and is associated with many cellular processes including activation, proliferation, differentiation, and migration [2,3,4]. The human transmembrane molecule leukocyte common antigen-related receptor (LAR) is part of a family of receptor protein tyrosine phosphatases, of which there are approximately 20 known members. It has structural features similar to the extracellular receptor region of the immunoglobulin superfamily of adhesion molecules, and the fibronectin type-III like repeats in the same region of the neural cell adhesion molecule 1 (N-CAM-1) [5,6]. The cytoplasmic segment of LAR consists of either one or two tandem phosphatase domains with a single transmembrane region [7]. Thus, LAR which is part of a distinct class of proteins might be involved in cell adhesive interactions as well as having tyrosine phosphorylation function [8].
The cytoplasmic domain of LAR is also similar to the leukocyte common antigen (LCA), CD45 (also known as T200 B220, and Ly-5), the prototype and first discovered receptor-linked protein tyrosine phosphatase [9]. CD45 is a high-molecular-weight glycoprotein, exclusively restricted to leukocytes, including intraepithelial lymphocytes (IEL) [1,10]. CD45 tyrosine phosphatase appears to be required for the coupling of T-cell antigen receptor signals to the phosphoinositol second messenger signal and may also participate in the rapid increase in tyrosine phosphorylation induced by activation of the T cell receptor [11,12]. It has been shown that both LCA and LAR have phosphatase activity in vitro [13,14].
The diversity in the extracellular region of LAR presumably reflects a diversity of ligands to which this enzyme responds. The identity of a ligand for CD45 is still not clear. It has been shown that heparin sulphate on bone marrow stromal cells serves as a ligand for CD45, while CD22 has been shown to be the molecule on B cells [15,16]. To date, little is known about the ligand for LAR although homotypic interactions have been reported. LAR may associate with the actin cytoskeleton through cadherins or catenin molecules [17]. It is also possible that LAR which is expressed in many tissues and cell types may act by modulating phosphorylation of the components of the signal pathways that regulate cell adhesion [18].
Until recently, protein phosphatases have played a low profile in the protein phosphorylation field. However, it is clear that protein phosphatases do not simply constitutively reverse the effects of protein kinases but rather play a specific role in cellular biology [19]. Studies using knock-out mice have demonstrated that intestinal crypts can develop in the absence of the epidermal growth factor (EGF), a tyrosine kinase receptor [20].
Furthermore, studies of tyrosine phosphorylation in the human duodenum have indicated that proliferating cells in the middle and lower portions of the duodenal crypt were devoid of tyrosine phosphorylation [21]. These data indicate that tyrosine kinase activation mediated by EGF is not necessarily a dominant factor in crypt cell proliferation and suggest that tyrosine phosphatase activity may be a more critical factor in crypt development. Lastly, it has been reported that cellular levels of tyrosine kinase activity is downregulated in colon cancer suggesting a possible correlation with colonic tumor growth, while structures similar to the extracellular domain of LAR have been reported in the colon-cancerassociated gene DCC [22,23].
Antibodies to the LAR receptor protein tyrosine phosphatase (RPTP) have made it possible to study the expression of LAR in different tissues and in single-cell suspensions [5,6]. We therefore investigated the expression of this family of molecules in the human duodenal crypt. A system was devised using two-colour flow cytometry to assay the sequential elution of intraepithelial lymphocytes (CD3 + ) and epithelial cells (CD3 − ) from the upper, middle, and lower regions of the human duodenal crypt. We determined the level of expression of LAR in IEL and epithelial cells in the duodenal crypt. In addition, we confirmed the specific location of LAR within the duodenal crypt by immunohistochemical staining on paraffin-fixed section.

Patients
Duodenal biopsy specimens were obtained endoscopically from patients giving informed consent. Four biopsy fragments were taken from each patient. These patients were receiving treatment for dyspeptic symptoms and were all found to have a normal small intestine on histological examination. The Ethics Committee of the Federated Dublin Voluntary Hospital approved protocols for the use of duodenal biopsy tissue.

Flow cytometric analysis of enterocyte LAR expression
LAR expression was measured by indirect immunofluorescence and flow cytometry. A mixed population of IEL and epithelial cells was removed from biopsy tissue by sequential elution over periods of 30, 60, 90, and 120 minutes, using a method previously described [24]. Briefly, specimens were incubated in a solution of 1 mM EDTA and 1 mM DTT in Ca 2+ and Mg 2+ free Hanks buffered saline solution. Cells were harvested, counted and adjusted to 10 6 cells/mL. They were incubated for 15 minutes in 2% (v/v) paraformaldehyde and washed twice in a solution of phosphated buffered saline containing 0.01% (v/v) bovine serum albumin, and 0.01% (v/v) sodium azide (PBS/BSA/NaN 3 ), henceforth used in all subsequent washes. Cells were incubated with an affinity purified polyclonal antibody to the E-subunit of the LAR molecule, rabbit anti-LAR (Transduction Laboratories, Lexington, Ky), for 30 minutes at a concentration of 0.1 µ g/mL. The antibody, though not tested for crossreactivity with other isoforms of LAR, should recognise any LAR proteins containing the E-subunit. A control polyclonal antibody, anti-keyhole limpet haemocyanin (KLH), (Sigma, St Louis, Mo), was also used and cells were incubated for 30 minutes with a control antibody at concentration 0.1 µ g/mL. The cells were then washed twice with PBS solution. Samples were reincubated with a fluorescein-conjugated goat anti-rabbit secondary antibody (Dakopatts, Glostrop, Denmark) for a further 30 minutes, and again washed twice with PBS solution. Normal rabbit serum was used to block nonspecific binding sites, for 10 minutes. Samples were again washed twice with PBS and finally incubated for 10 minutes with an anti-CD3 monoclonal antibody directly conjugated to phycoerythrin (Becton Dickinson, Mountain View, Calif). Two-colour flow cytometric analysis was performed using the Becton-Dickinson FACScan with Lysys 11 software. The CD3 + and CD3 − cell populations were identified by gating on forward scatter (FSC) against FL2/PE, (CD3 + ) showing histograms of FL1 positivity for LAR and KLH.

Immunohistochemistry
Paraffin-embedded sections were cut from duodenal biopsy tissue, showing clearly the upper, middle, and lower parts of the duodenal crypts. Pellets were prepared from a human umbilical vein endothelial cell (HU-VEC) line, and paraffin-embedded sections cut to use as a positive control for LAR staining. Sections were then mounted on APS (3-aminopropyltriethoxysilane) slides and left overnight at 55 • C. Sections were fixed in methanol/acetone for 10 minutes followed by dewaxing in xylene and alcohol. Endogenous peroxidase was blocked by incubating sections in a solution of 0.3% H 2 O 2 in methanol for 30 minutes. After washing in H 2 O followed by a Tris buffered saline (TBS) wash, the samples were microwaved at 800 W in 10 mM citrate buffer, pH 6, for 20 minutes as previously described [25]. Sections were blocked with 1.0% bovine serum albumin for 10 minutes. The LAR antibody was applied and incubated at 4 • C overnight, followed by two TBS washes. Sections were incubated in a biotinylated antirabbit second antibody (Amersham, Buckingham, UK) for 60 minutes at room temperature and, following two TBS washes, were incubated in ABC Elite solution (Vectastatin, Vector Laboratories, Burlington, Calif) for 30 minutes in a humidified chamber. Finally, sections were stained with diaminobenzidine tetrahydrochloride (DAB) solution for 7 minutes counterstained in haematoxylin (5 seconds), and mounted in DPX (distyrene tricresylephosphate xylene mountant for Microscopy, 360294H), BDH, Poole, England. Photographs were taken with a Nikon camera mounted on a Polyvar 2 Reichart Jung light microscope. (Figures 1a and 1b). There is less LAR expression detected in unfixed cells in comparison to that detected in fixed cells suggesting that the LAR molecule may be shed in unfixed cells. This is consistent with previous findings that LAR is shed [26,27]. Histograms of IEL CD3 + (positive) and epithelial CD3 − (negative) enterocyte cell populations, eluted after 30, 60, 90, and 120 minutes intervals following EDTA/DTT incubation show differential expression of LAR in IEL CD3 + cells (Figures 2a, 2c, 2e, and 2g) and in epithelial CD3 − cells (Figures 2b, 2d, 2f, and 2h). The 30-minute incubation with EDTA/DTT removes surface enterocytes from the tips of villi. This is an area where the cells are more differentiated and fewer IEL are extracted [21]. It is evident from Figure 2c that the phosphatase molecule LAR is significantly expressed in CD3 + intraepithelial lymphocytes, isolated after 60 minutes. This corresponds to the upper and middle portions of the crypt [21]. Cells in the middle and lower portions of the crypt are more likely to undergo proliferation. LAR is expressed in the CD3 − cell population at higher levels in the upper, middle, and lower portions of the crypts where cellular proliferation is more likely to occur (Figures 2b, 2d, 2f, and 2h) [28,29]. LAR is also expressed in peripheral blood lymphocytes ( Figure  3a), but expression is low in comparison with IEL ( Figure  3b).

FACScan profiles show LAR expression in unfixed and fixed HUVECs
Intraepithelial lymphocytes intercalate between the epithelial cells of the villi and crypts. Under high power (×400) they can be detected in the upper, middle and lower areas of the villi. LAR staining is present in IEL at the tips of the villi (Figures 4a and 4b), while there was strong staining evident in IEL at the base of the crypt (Figures 4a  and 4c). A black arrow indicates the IEL. LAR expression was also detected within the duodenal lamina propia lymphocytes. Closer examination of the pattern of expression of LAR staining within the cells suggests a predominantly surface pattern. This is confirmed by the fact that LAR was detected by surface staining and flow cytometry in nonpermeabilized cells. Finally, similar surface expression is seen in control HUVECs (Figure 4d). The negative control KLH staining is seen in duodenal biopsy tissue (Figure 4e).

DISCUSSION
In this study we have developed a two-colour flow cytometric method for the detection of the tyrosine phosphatase LAR on epithelial cells and IEL eluted from endoscopically obtained duodenal biopsy tissue. We have shown that LAR is expressed in CD3 negative epithelial cells in the upper, middle, and lower regions of the duodenal crypts. It has been shown that the middle and lower regions of the crypts are areas of cellular proliferation and that low levels of tyrosine phosphorylation have been detected in the epithelial cells in these areas [21,28]. Although there is no doubt that EGF can cause intestinal epithelial cellular proliferation, there is emerging evidence that this is not its only function within the gastrointestinal tract [30]. It is also reported that crypt villi can grow in the absence of EGF receptor, a tyrosine kinase receptor molecule [20]. In addition, LAR gene products were detected to be abundantly expressed in rat oesophagus, skin, and small intestine, all of which are areas of continuously renewable epithelia [31]. These findings suggest that molecules other than tyrosine kinases may have a greater role to play in crypt cell proliferation. Tyrosine phosphatases, which cleave phosphate groups from tyrosine residues, may be the relevant molecules involved in crypt epithelial proliferation. The LAR family of tyrosine phosphatases may play an active role [32].
We have also shown LAR to be expressed in CD3 positive IEL eluted at 60 minutes in EDTA/DTT. These are a highly specialised subpopulation of T lymphocytes restricted to the mucosal epithelia. They differ from T cells in other lymphoid organs in that they are predominantly CD3+CD8+. The integrin αEβ7 is highly expressed in these cells and it has been hypothesised that IEL possess specific receptors enabling interaction with local ligands and controlling homing to and migration through basement membrane [33]. It is thought that adhesion molecules may play a major role, but the mechanism by which IEL rapidly migrate to sites of inflammation is not fully understood. The expression of LAR in a significant subpopulation of IEL suggests a role in the homing of this unique population to distinct sites in the gut mucosa. The extracellular domain of LAR has homology with adhesion molecules, for example N-CAM, suggesting a possible role in IEL enterocyte interactions [17]. As LAR interacts homotypically, LAR-LAR interactions between IEL and epithelial cells could represent one mechanism by which IEL epithelial cell interactions occur. Further definition of the phenotype of LAR positive IEL subpopulations should provide important insight into the mechanisms of IEL migration.
It has been shown that LAR is expressed in the cell surface as a complex of two noncovalently associated subunits. LAR E-subunit contain the cell adhesion molecule receptor region, and the LAR P-subunit contain a short segment of extracellular, transmembrane, and cytoplasmic domains. The LAR cell adhesion molecule E-subunit is shed during cell growth, suggesting that receptor shedding may be an important mechanism for regulating PT-Pase function [26,27]. The apparent loss of LAR expression in the CD3 positive cell population at the tips of the villi could reflect simply loss of the E-subunit. Hence, the shedding of LAR E-subunit might reduce the responsiveness of LAR to a cognate ligand. The LAR (PTPase) and LAR interacting protein 1 (LIP.1) have been shown to colocalise at focal adhesions suggesting that LAR and LIP.1 may regulate the disassembly of focal adhesions and thus regulate cell-cell, cell matrix interactions and migration, by facilitating the recruitment of LAR to cytoskeletal targets. They may also regulate LAR extracellular matrix interactions, retaining LAR at the cell surface [8].
These findings suggest the possibility that tyrosine phosphatase activity may be of significant importance in the regulation of crypt cell proliferation. Secondly, as the extra cellular domain of LAR has homology with adhesion molecules, the finding of this molecule on IEL could suggest a possible functional role in migration and homing of this unique lymphocyte subpopulation.