Co-Expression and Functional Interactions of Death Receptor 3 and E-Selectin in Clear Cell Renal Cell Carcinoma

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 ajp.amjpathol.org 63 64 65 66 67 68 69 70 71 72 73 74 Co-Expression and Functional Interactions of 75 76 Death Receptor 3 and E-Selectin in Clear Cell 77 78 79 80 81 Renal Cell Carcinoma Rafia S. Al-Lamki,* Jun Wang,* Jordan S. Pober,y and John R. Bradley*

Renal cell carcinoma (RCC) is the most common type of kidney cancer, with the clear cell subtype (ccRCC) representing approximately 70% of Q7 cases. 1 Prognosis for highgrade RCC is poor, and new approaches to treat this tumor are needed. In this study, we report a new autocrine/ paracrine signaling pathway in ccRCC cells involving death receptor 3 (DR3) and E-selectin that could be a target for new therapies.
DR3, also designated as TNFRSF25, TRAMP, LARD, WSL-1, Apo-3, or TR3, is one of eight death domain (DD)econtaining receptors of the tumor necrosis factor (TNF) superfamily. 2 On binding of its ligand, known as tumor necrosis factorelike 1A (TL1A also designated as TNFSF15) to DR3, signaling proteins, such the TNF receptor type 1-associated DD protein, TNF receptor associated factor 2 protein, and receptor-interacting serine/ threonine-protein kinase 1, are recruited to the cytosolic DD of the activated receptor, initiating new expression of proteins involved in immune responses and cell survival. 3e5 After some delay, the Fas associated via DD protein may also be recruited to this complex, initiating the extrinsic pathway of apoptosis. We have previously reported that DR3 is up-regulated on renal tubular epithelial cells (TECs) and vascular endothelial cells (ECs) in settings of renal injury and that binding of TL1A to DR3, which activates the same signaling pathways as TNF-a binding to TNF receptor type 1, can trigger cell injury. 6,7 However, others have reported effects of DR3 signaling that reduce acute kidney injury by antagonizing the proapoptotic signals induced by TNF. 8,9 TL1A mRNA is abundantly expressed in kidney  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60  61  62  63 tissue 7 and colon cancer. 10 Recent studies have reported DR3 expression by some tumors, but its role in tumorigenesis is controversial; it has been implicated as both an anticancer 11 and procancer molecule. 12,13 E-selectin (CD62E, also known as endothelial leukocyte adhesion molecule 1), is a type Ia transmembrane protein that contains lectin-like and epithelial growth factorelike domains, followed by short cysteine-rich repeats. 14 It is primarily expressed by activated but not resting postcapillary venular ECs, 15 but rare epithelial cell expression of E-selectin has been previously reported. 16,17 E-selectin is transcriptionally induced by cytokines, such as TNF or IL-1, mediated by NF-kB and a variant of activator protein 1. 18e20 These are the same signaling pathways activated by TL1A binding to DR3. In humans, E-selectin plays a critical and nonredundant role in tethering and rolling of leukocytes to ECs, allowing their subsequent extravasation into inflamed tissues. 21 Expression of E-selectin has been reported on some tumor vasculatures, 22e24 where it has been proposed as a target for metastasis of tumor cells bearing E-selectin ligands, 13,25e29 typically cell surface proteins or lipids bearing sialyl Lewis X (sLe X ) or sialyl Lewis A. 13,30e32 sLe X may be detected using a monoclonal antibody that binds to human skinehoming T cells where the antigenic determinant is designated as cutaneous lymphocyte antigen (CLA). sLe X expression increases from normal tissue to early-stage bladder or renal cancers to metastatic disease. 33,34 In RCC, sLe X correlates with conventional histopathologic parameters and serves as an indicator for prognosis. 35 Elevated levels of soluble serum E-selectin are associated with tumor grade, tumor stage, and metastases in breast cancer. 36 In addition to adhesion to ECs, one report has suggested that in colon carcinoma, E-selectin binds to and can serve as a functional ligand for DR3, triggering activation of p38 and ERK mitogen-activated protein kinases (MAPKs) and conferring metastatic and survival advantages. 13 Because DR3 can be induced on normal kidney TECs by inflammation, we hypothesized that it might also play a role in their malignant counterparts, namely, ccRCC cells. In this article, we report that both DR3 and, unexpectedly, Eselectin are expressed on ccRCC cells and that expression of both molecules increases with tumor grade. DR3 extracted from ccRCC tumors bears sLe X moieties and tumor cell Eselectin appears to be complexed with tumor cell DR3 in situ as assessed by proximity ligation assay. In ccRCC, organ culture exogenous E-selectin, like exogenous TL1A, acts as a DR3 ligand, leading to activation of NF-kB and MAPKs, increased expression of both DR3 and E-selectin, and increased cell cycle entry of ccRCC cells.

Tissue Samples
Experiments using human tissue were performed with informed consent of patients and approval of the local ethics committee and Cambridge University Hospitals Tissue Bank. RCC tissue obtained from radical nephrectomy specimens was immediately excised from tumors that grossly appeared to be ccRCC. This tumor classification was later verified by routine histologic assessment of paraffin-wax embedded samples. Noneclear cell histologic tumor types (eg, papillary, chromophobe, and collecting duct) 37 were excluded, and only ccRCC, graded according to the four-tiered Fuhrman nuclear grading system 38 and pathologically staged based on the TNM classification, 39,40 was used. Tissue samples from 40 patients were collected and scored as Fuhrman grade 1 (n Z 10), Fuhrman grade 2 (n Z 10), Fuhrman grade 3 (n Z 10), and Fuhrman grade 4 (n Z 10). In parallel, adjacent nonetumor kidney (NK) (n Z 40), categorized histologically as normal kidney cortex, was collected in sites remote from the tumor. All samples were fixed overnight at 4 C in 4% formaldehyde in 0.1 mol/L phosphate buffer (pH 7.6) and paraffin wax embedded for immunofluorescence or snap frozen in isopentane cooled in liquid nitrogen. Parallel unfixed fresh samples were processed for organ culture experiments. Sections 5 mmol/L thick were prepared for subsequent experiments, and hematoxylin and eosin (H&E) staining was performed for morphologic studies.

Kidney Organ Cultures
As previously described, 39 duplicate <1-mm 3 fragments of ccRCC (grade 1/2) and adjacent NK (n Z 5 per study group) were obtained immediately from surgically excised specimens. Fragments of tissue were placed in a Corning flat-bottomed, 96-well tissue culture plate (Appleton Woods Limited, Birmingham, UK) and immediately immersed in Medium 199 containing 10% heat inactivated fetal calf serum (TCS, Bucks, UK) and 2.2 mmol/L glutamine. Multiple randomized samples from each patient were used to obtain parallel-group comparisons and to assess the reliability and reproducibility of these assays. Tissue was left in media alone [untreated (UT) controls] or pretreated with rhTL1A (0.2 mg/mL) (R&D Systems, Oxford, UK) 39,41 and soluble E-selectin (sE-selectin) (5 mg/mL) (Tonbo Biosciences) for 3 hours at 37 C. In parallel, some organ cultures were pretreated with 5 mmol/L of the NF-kB inhibitor 42 BAY11-7082 for 1 hour at 37 C before treatment with TL1A or sE-selectin for 3 hours at 37 C. All cultures were harvested, half of the samples were fixed in 4% formaldehyde for 1.5 hours at 4 C and processed for paraffin wax embedding, and half were snap frozen and stored at À70 C for cryosectioning. H&E was performed formalin-fixed, paraffin-embedded (FFPE) sections for morphologic studies.

IF and IHC
The 5 mmol/L FFPE sections or cryosections of ccRCC, NK, and corresponding organ cultures were subjected to immunofluorescence (IF) and immunohistochemistry (IHC) as previously described. 41 Briefly, some sections were incubated with antibodies to DR3 and/or E-selectin or in combination with anti-CD31, anti-cytokeratin, antieCLA-1, anti-sLe X , antiphosphorylated p38(Thr180/Tyr182), anti-ERK(Tyr20) or anti-JNK (Thr183/Tyr185). To determine the presence of leukocytes, parallel sections of ccRCC and NK were incubated with anti-CD45 alone or with CD15, CD14, or CD68. The intensity of fluorescence was calculated as the corrected total cell fluorescence (CTCF) using ImageJ software version 1.53a (NIH, Bethesda, MD; http:// imagej.nih.gov/ij) to control for local background fluorescence. The following formula was used to calculate CTCF: integrated density À (area of selected cell Â mean fluorescence of background readings). Antibody-binding sites were visualized with fluorochrome-conjugated secondary antibodies (Alexa Fluor 488 or Alexa Fluor 568 or Alexa Fluro 405 ) plus Hoechst 333,342 for nuclei detection. The brightest fluorophores were used for low-abundant proteins to maximize sensitivity and with narrow emission spectra to avoid bleed-through. The indirect IF amplification technique was used, which permits simultaneous detection of two antibodies raised in the same species. 43 After IF, all sections were washed in phosphate-buffered saline, mounted in Vectashield mounting media, and imaged with a Leica SPE confocal laser scanning microscope (Leica Microsystem Ltd., Milton Keynes, UK). In addition, IHC was used on organ cultures treated with TL1A or sE-selectin with or without Bay 11-7089 to examine expression of NF-kBp65p-Ser276 and phosphorylated histone-H3 Ser10 (pH3-Ser10 ). Endogenous peroxidase was blocked using 30% H 2 O 2 in absolute methanol for 30 minutes at room temperature before incubation with blocking buffer (10% fetal calf serum in phosphate-buffered saline) for 1 hour in room temperature followed by 1:100 dilution of primary antibodies overnight at 4 C. After rinses in phosphate-buffered saline, sections were incubated with 1:200 dilution of secondary antibodies for 1 hour in room temperature and antibody binding sites visualized using 3,3 0 -diaminobenzidine tetrahydrochloride (Thermo Fisher Scientific) plus 0.01% H 2 O 2 and viewed using a Nikon Optiphot-2 microscope. To quantify the intensity of immunostaining, we imported images of 3,3 0 -

Immunoprecipitation and Immunoblotting
Total proteins were extracted from clinical samples of ccRCC grades 1 to 4 and adjacent NK with radioimmunoprecipitation assay lysis buffer containing protease and phosphatase inhibitors (Merck Life Sciences UK Ltd), and protein was quantified using a bicinchoninic acid kit (Merck). DR3 epitope was pulled down from tissue extracts using rabbit anti-DR3 antibody and with Magnetic Dynabeads-Protein G (catalog number 10007D; Fischer Scientific UK Ltd., Loughborough, UK) following the manufacturer's instructions. Briefly, DynabeadseProtein G was suspended in antibody binding and washing buffer containing anti-DR3 antibody and incubated for 10 minutes at room temperature. The mixture was then placed on the magnet and supernatant removed. The DynabeadseProtein G antibody complex was resuspended in binding and washing buffer and antigen-containing samples (typically 100 to 1000 mL) were added and incubated on a rotor for 10 minutes at room temperature. The mixture was then placed on the magnet and supernatant removed. The DynabeadseProtein G antibody antigen complex was rinsed in washing buffer and resuspended in elution buffer for 2 minutes at room temperature. The tube was placed on the magnet and the supernatant/sample transferred to a clean tube. Eluates were than resuspended in the SDS sample buffer and incubated for 10 minutes at 70 C and subjected to immunoblotting. Then 50 mg of total protein per sample was separated by SDS-PAGE. After transfer to a nitrocellulose membrane, proteins were probed with 1:1000 dilution mouse antieCLA-1 overnight at 4 C, and the signal was detected using Super Signal West Pico Chemiluminescent Substrate according to the manufacturer's instructions (Fisher Scientific UK Ltd.).
In Situ PLA FFPE sections of ccRCC and adjacent NK organ cultures treated with TL1A or sE-selectin or left UT in media alone (negative controls) were subjected to proximity ligation assay (PLA) as previously described. 44 In brief, sections were incubated with rabbit anti-DR3 and mouse antieEselectin or antieCLA-1 or anti-sLe X antibodies overnight at 4 C. This procedure was followed by a mixture of 1:5 PLA probes (anti-mouse and anti-rabbit antibodies attached with different oligonucleotides, which can ligate through the addition of two other oligonucleotides in ligation solution) for 1 hour. After rinses in 1Â wash buffer A (provided with the kit), sections were incubated in a ligation solution for 30 minutes at 37 C and then rinsed in 1Â wash buffer A and incubated in amplification solution in a dark humidity incubator for 100 minutes at 37 C. Slide were then washed in 1Â and 0.01Â wash buffer B (provided with the kit) and mounted in Vectamount mounting medium with DAPI. For negative controls, the primary antibody was replaced with isotype-specific serum. Sections were examined on confocal laser scanning microscope, and red fluorescence dots, indicative for protein localization, were counted using ImageJ and quantified with GraphPad Prism version 9.0 software (GraphPad, La Jolla, CA).

In Situ Hybridization
Nonradioactive in situ hybridization was performed on 5mm FFPE sections of ccRCC and adjacent NK organ cultures as previously described. 6,41,45 Briefly, sections were incubated overnight at 37 C with hybridization solution that contained single-stranded antisense DNA oligonucleotide probes 5 0 end labeled with digoxigenin specific for human E-selectin 4 mg/mL (5 0 -ATTGTCCCCTAGCAAGGCAT-3 0 ) and for human DR3

Statistical Analysis
The mean number of tumor cells and normal TECs positive for NF-kBp65p-Ser276 and pH3-Ser10 and positive cells for CD45 and CD15, CD14, or CD68 were counted in 10 random high-power fields of view at Â40 magnification and divided by the total number of cells to generate the percentage of positive cells. The assays were repeated in three or more independent experiments, and results are expressed as means AE SEM. Statistical significance was assessed using an unpaired two-tailed t-test and analysis of variance using GraphPad Prism.

½F1 ½F1
, quantified in Figure 1C and Supplemental Figure S1  , AeF), and IF staining of similar sections using panleukocyte marker CD45 in combination with CD15, CD14, or CD68 showed a high number of infiltrating cells composed of granulocytes, monocytes, and tissue-associated macrophages in the tumor microenvironment (Supplemental Figure 2A, quantified in Supplemental Figure 2B). In addition, co-signal for DR3 and E-selectin was seen in approximately 35% of tumor cells in ccRCC grade 4 compared with <5% in cRCC grade 1 and a negligible level of co-signal detected in adjacent NK ( Figure 2, AeE, quantified in Figure 2F). These patterns of expression were similar across multiple different ccRCC samples of the same histologic grade, and a similar staining pattern was seen using different DR3 or E-selectin antibodies.

CLA-1 and E-Selectin Binding Carbohydrate Moiety sLe X Are Highly Expressed in Tumor Cells and Are Associated with DR3 in ccRCC
Epitopes recognized by antieCLA-1 antibodies and other Eselectinebinding carbohydrate moiety sLe X -bearing molecules are highly expressed in tumor cells and are associated with DR3 in ccRCC. A relationship among CLA-1 molecules, sLe X determinants, and E-selectin has been previously reported. 46 E-selectin binds to proteins that bear sLe X oligosaccharides. 13 We, therefore, examined whether CLA-1 is associated with DR3 in ccRCC by immunoprecipitating DR3 from extracts of ccRCC and adjacent NK, separating the
DR3 þ ccRCC Cells Are Responsive to TL1A We previously reported that TL1A, the principal ligand for DR3, is up-regulated in TECs and vascular ECs in kidney inflammation and injury and that signaling through DR3 can activate NF-kB signaling in organ culture of human kidney. 7 Moreover, various MAPKs, including ERKs, JNKs, and p38 MAPK, contribute to tumor cell growth and survival, and phosphorylation of JNK is increased in ccRCC. 47 We assessed whether DR3 in ccRCC cells is functional using a human organ culture model established in our laboratory. 7 Organ cultures were treated with TL1A or left untreated and analyzed for the presence of phosphorylated p65 subunit of NF-kB (NF-kBp65 Ser276 ), indicative of induction of gene transcription, for expression of various activated MAPKs, namely, p38, ERK, and JNK, using antibodies specific for the activated phosphorylated forms of these enzymes and for induction of cell-cycle entry comparing the expression of pH3 -Ser10 , an indicator of cell proliferation. The effect of treatment on morphology was determined on H&E-stained sections from all the cultures. Compared with UT controls, both the treatments induced distortion of the cell architecture, some areas of fibrosis, and thickened wall in some vessels (Supplemental Figure S4 621  622  623  624  625  626  627  628  629  630  631  632  633  634  635  636  637  638  639  640  641  642  643  644  645  646  647  648  649  650  651  652  653  654  655  656  657  658  659  660  661  662  663  664  665  666  667  668  669  670  671  672  673  674  675  676  677  678  679  680  681  682  683   684  685  686  687  688  689  690  691  692  693  694  695  696  697  698  699  700  701  702  703  704  705  706  707  708  709  710  711  712  713  714  715  716  717  718  719  720  721  722  723  724  725  726  727  728  729  730  731  732  733  734  735  736  737  738  739  740  741  742  743  744 42% AE 0.2%) and in vascular ECs in ccRCC compared with UT controls (approximately 4.5% AE 0.6%) ( Figure 4 ½F4 ½F4 A). TL1A also induced NF-kBp65 Ser276 expression in NK, but this was less frequent by approximately 2.2-fold (approximately 19% AE 0.8%), with staining mainly confined to normal TECs and vascular ECs ( Figure 4B). Tissue pretreated with Bay11-7082, an inhibitor of the enzyme IKKb, which is required for canonical NF-kB signaling, before the addition of TL1A showed weakened green fluorescence nuclei staining of p65 in both study groups with pronounced effects in ccRCC, consistent with canonical signaling (quantified in Figure 4C). Interestingly, TL1A also induced a statistically significant increase in the level of activation of MAPKs with much pronounced signal for p38 and phosphorylated JNK than for phosphorylated ERK P < 0.001 versus P < 0.05) in ccRCC compared with NK (by approximately threefold) and UT controls and staining mainly localized to tumor cells, vascular ECs, and infiltrating cells ( Figure 4D, quantified as CTCF in Figure 4E). Tumor cells (approximately 35.1% AE 0.4%) and vascular ECs in treated ccRCC also showed an increased expression of nuclear pH3 -Ser10 compared with UT controls (approximately 4.7% AE 0.1%), and treated cultures of NK showed a much reduced effect (by approximately twofold) with signal detected mainly in normal TECs ( Figure 4F, quantified in Figure 4G). Cumulatively, these data demonstrate that DR3 in ccRCC tumor cells is functional, can activate NF-kB and MAPKs, and promotes tumor cell cycle entry, indicative of mitogenesis.

TL1A Signaling Induces the Expression of E-Selectin in ccRCC Cells
The expression of E-selectin on ccRCC cells was unexpected and could arise from adsorption of shed E-selectin from ECs or de novo synthesis by ccRCC cells. To address this, we analyzed the effect of TL1A on regulation of Eselectin mRNA and protein expression using organ cultures of low-grade ccRCC (grades 1 and 2). Signal for E-selectin protein and mRNA was seen in a few scattered tumor cells in UT controls, whereas TL1A-treated cultures showed an enhanced signal in tumor cells (approximately 26% AE 0.6%), vascular ECs, and infiltrating leukocytes (
Having established that DR3 signaling events activated by TL1A in tumor cells result in the induction of both DR3 and E-selectin expression and that induced E-selectin interacts with DR3, we then analyzed downstream cellular responses to determine whether sE-selectin can function as a ligand for DR3 in ccRCC. For this, low-grade ccRCC and adjacent NK cultures were treated with sE-selectin and left UT and examined for the presence of NF-kBp65 Ser276 , activation of MAPKs, and induction of cell cycle entry. Nuclear NF-kBp65 Ser276 expression was induced in tumor cells (approximately 22.82% AE 0.5%) and in vascular ECs

Discussion
We previously reported that DR3 is induced in renal TECs and ECs in kidney injury associated with transplant rejection. 6 We extended our studies to examine the expression of DR3 in a clinical cohort of excised ccRCC tumors because these cells are the malignant counterpart of tubular epithelium. We found that DR3 is expressed in tumor cells and vascular ECs in ccRCC, and its expression increases with tumor grade. In endothelium, signaling through DR3 results in expression of E-selectin, thought to be a specific response of this cell type. Surprisingly DR3 þ tumor cells also show expression of E-selectin that increases with tumor grade. We found that DR3 on ccRCC cells also contains sLe X and that   1365  1366  1367  1368  1369  1370  1371  1372  1373  1374  1375  1376  1377  1378  1379  1380  1381  1382  1383  1384  1385  1386  1387  1388  1389  1390  1391  1392  1393  1394  1395  1396  1397  1398  1399  1400  1401  1402  1403  1404  1405  1406  1407  1408  1409  1410  1411  1412  1413  1414  1415  1416  1417  1418  1419  1420  1421  1422  1423  1424  1425  1426  1427   1428  1429  1430  1431  1432  1433  1434  1435  1436  1437  1438  1439  1440  1441  1442  1443  1444  1445  1446  1447  1448  1449  1450  1451  1452  1453  1454  1455  1456  1457  1458  1459  1460  1461  1462  1463  1464  1465  1466  1467  1468  1469  1470  1471  1472  1473  1474  1475  1476  1477  1478  1479  1480  1481  1482  1483  1484  1485  1486  1487  1488 sE-selectineinduced responses appear less pronounced than those induced by TL1A. This difference may be ascribed to an artifact of differences in the concentrations of ligands use or in affinity of sE-selectin and TL1A for DR3 or that many sLe X groups on DR3 are already engaged by endogenous Eselectin and unavailable for binding and hence DR3 clustering. TL1A and sE-selectin induced similar effects in organ cultures of NK but to a much lower extent. Currently, little is known about the role(s) of DR3 in cancer, particularly its potential as an antitumor target. Increased levels of DR3 expression have been reported in various cancer cell lines and tumors compared with normal counterparts. 13,48e50 DR3 mRNA encoding the membrane and soluble receptors has been reported in colorectal cancer, 50 and DR3 transcript expression levels are increased in high-grade breast tumors. 48 As noted above, a potential cancer migratory role for DR3 has also been previously highlighted, 48 and DR3 has been reported to be a new receptor for E-selectin that confers metastatic and survival advantages to colon cancer via MAPKs 13 and phosphatidylinositol 3-kinase/NF-kB axis. 51 Our observation that DR3 in tumor cells is induced by TL1A and sE-selectin suggests that it is a functional receptor in ccRCC with the ability to activate downstream pathways that promote survival/growth signals. Interestingly, DR3 is labeled as a death receptor, but its activation by sE-selectin does not induce apoptosis in colon cancer cells, except when ERK is inhibited. 13 A previous study by Wen et al 52 reported that the binding of TL1A to DR3 activated the ERK pathway, whereas JNK and apoptosis-inhibiting protein c-IAP2 prevented DR3-mediated apoptosis in TF-1 cells. Because DR3 uses the same signaling apparatus as TNFR1 for both protein induction and apoptosis initiation, it is also possible that TNF-inducible antiapoptotic proteins, such as cellular FADD-like IL-1b converting enzymeeinhibitory protein, are induced and then play a protective role.
Accumulating evidence suggests that E-selectin facilitates metastasis in various cancers. 28,29,49,53e55 The mechanism entails adhesion cascade governed by cell-cell interactions between circulating tumor cells and vascular ECs through a sequential affinity interaction between adhesion molecules and the counterreceptor ligand. 13 We document E-selectin expression in tumor cells in ccRCC, which is unexpected. However, epithelial cell expression of E-selectin has been previously reported in human colonic epithelium. 24 The factors that result in preferential expression of E-selectin in ccRCC cells and its pathogenic importance are at present unclear. Although the major role of E-selectin expression on ccRCC tumor cells is likely to be autocrine/paracrine signaling, it also could function to retain leukocytes within the tumor microenvironment.