Studies On The Immune Recognition Pattern Of ULBPs, The Functions Of ULBP Splice Variants And The ULBP-related Immune Escape Mechanism

NKG2D is a C-type lectin-like homodimer receptor, which is expressed on NK, NKT, CD8+αβT cells andγδT cells and recognizes a group of ligands that represent distant relatives of MHC class I molecules including MICA/B in human and RAE-1-α-ε, H60, and MULT-1 in mouse. ULBPs are recently identified ligand family for NKG2D in human. They are frequently expressed by tumor cells and stimulate anti-tumor immune responses mediated by NKG2D or TCRγδ1 pathway.The fist issue of the study was aimed at discerning whether ULBPs could induce expansion of human ovarian epithelial carcinoma (OEC)-or leukemia-derivedγδT cells in vitro, and identifying whether ULBP-reactiveγδT cells were cytotoxic towards these malignant tumor cells. In addition, we also performed the analysis of CDR3 region of ULBP-reactiveγδT cells and the study of the relationship between TCRγδand NKG2D during the recognition and activation process of ULBP-reactingγδT cells, which would advance our knowledge about the recognition pattern ofγδT cells. Because of the MHC-unrestricted recognition pattern,γδT cells might directly recognize ULBPs expressed on tumor cells. Such pattern will make it possible for developing a novel immune therapy method.First, ULBP gene family were cloned from various tumor cell lines by RT-PCR. ULBP1, RAET1G/G2 were cloned from K562 human chronic myeloid leukemia cell line. ULBP2 and ULBP3 were cloned from Hela human cervix carcinoma cell line, while ULBP4 (also termed as RAET1E) was cloned from HO-8910 human ovarian carcinoma cell line. The recombinant ectodomains of ULBPs were expressed as a six histidine, C-terminal fusion protein in prokaryocyte and purified successfully. Meanwhile, stable transfection of ULBPs into CHO Chinese hamster ovary cells was completed successfully with SuperFect Transfection Reagent. Also, we expressed the ectodomain of ULBP3 in CHO cells and purified the recombinant protein successfully.Second, prokaryotic expressed ectodomains of ULBP4 and RAET1G2 were used as antigens to prepare monoclonal antibodies (mAb) by means of the conventional B lymphocyte hybridoma technique. Positive clones were identified by various ELISA, immunofluorescence assay, flow cytometry and western blot.Finally, we made the primary study for the immune recognition against ULBP3. Originally, because the peripheral bloods of malignant tumor patients or tumor tissues were unobtainable, we attempted to use the peripheral bloods of healthy individuals as reagents and unexpectedly found that ULBP3 not only stimulated NK cells (54.7% VS 9.9%), butγδT cells (15.3% VS 2.7%) to proliferate. Given the lower efficiency of proliferation, it would need lots of PBMC and several times of flow cytometry sorting for the functional studies. Therefore, we directly analyzed the sequences in CDR3 region of ULBP3-reactiveγδT cells instead and found that there was one dominant sequence inδ1 subtype (78.57%, 33/42); while the frequency of this sequence was only 10.71% (3/28) in thatδ1 subtype derived from the anti-TCRγδmAb-stimulatedγδT cells of the same individual. Sequence analysis indicated that it contains completed Vδ1, D3 and Jδ1 germline gene fragments, just like the T22-specific TCRδchain in mouse. However, there were no obvious characteristics about ULBP3-reactiveγ9 chain, suggesting that the interaction between TCRγ9δ1 and ULBP3 maybe independent ofγ9 chain. Although subsequent TCR gene transfection study failed because of the lower expression of TCRγδ1, the data will be helpful for finally and successfully establishing a gene transfection platform for confirming a certain antigen specific TCRγδand the research of the recognition pattern ofγδT cells against other members of ULBPs.The second issue of the study was to identify the functions of ULBP splice variants. Originally, we planned to search for the possible shortened ULBP4 splice variant without transmembrane region for confirming the immune escape hypothesis of ULBPs at RNA level by direct RT-PCR. However, we accidentally cloned three splice variants of ULBP4 instead: ULBP4-Ⅰ, ULBP4-Ⅱand ULBP4-Ⅲ, in HepG2 liver carcinoma cells, WISH human amnion cells and Hep-2 larynx carcinoma cells by RT-PCR, respectively, although their frequencies were very low. All of them are type 1 membrane-spanning molecule as parental ULBP4. Also, RAET1G3, another splice variant of RAET1G2 was identified in K562 leukemia cells by the same method. BLAST analysis shows that ULBP4-Ⅱis the same as RAE-l-like transcript 4 (GenBank AY054974) that also presents in U-373 MG glioblastoma cell line and some colon cancer cell lines. Moreover, ULBP4-Ⅱtranscripts could be detected in CNE2 human nasopharyngeal carcinoma cell line as well. Of interesting, CNE2 cells contained all of the identified splice variants. Nevertheless, with respect to any of these four splice variants, it should be noted that its frequency was very low, no matter in any tumor cells positive for these splice variant transcripts, as the positive transformed DH5 a cell clones containing any splice variant were no more than 5% in all of its individual positive transformed DH5αcell clones. The result was unlike RAET1G2, whose frequency was almost 50%. To test whether these scarce variants are also functional, we cloned them into eukaryotic expression vectors and showed that ectopic expressions of ULBP4 splice variants including ULBP4-Ⅰ, ULBP4-Ⅱand ULBP4-Ⅲled to the recognition by human NKG2D. In addition, free soluble forms of ULBP4 splice variants (theirα1+α2 ectodomains) together with RAET1G3 could down-regulate the expression of NKG2D on NK-92 cells. Finally, RAET1G3 could also stimulate NK-92 cells to secrete IFN-γ. Taken together, identification of these functional splice variants will advance our knowledge about the overall functions of ULBP gene family, since it's important to determine whether they are redundant in function or exhibit specificity.The third issue is whether ULBP4, that is RAET1E, can mediate tumor immune escape and what the immune escape mechanism is. In present study, we first screened many kinds of tumor cells for their RAET1E transcript expressions by RT-PCR at first. Consequently, we cloned and sequenced the shortened form of RAET1E gene, termed as RAET1E2, which could be detected in all tumor cells positive for RAETlE mRNA and whose protein expression was also confirmed by western blotting. Pre-incubating NK-92 cells with culture supernatants of tumor cell lines positive for RAET1E2 including MGC-803, HO-8910 and HepG2 cells or of RAETlE2 gene-transfected COS-7 cells, decreased the expression of NKG2D on NK-92 cells. Furthermore, recombinant RAET1E2-Histag fusion protein also induced the down-regulation of NKG2D expression on NK-92 cells. As a result, reduced NKG2D surface levels were paralleled by a marked reduction in NKG2D-mediated NK cytotoxicities to MGC-803, HepG2 and K562 tumor cells. Taken together, our data provide the first evidence for an immune escape mechanism of tumors via selective splicing of an ULBP RNA to generate a free soluble ULBP protein, RAET1E2, which might impair NKG2D-mediated immunosurveillance.The final issue is "one economical and simple method to amplify Vγ9Vδ2 T Cells". Establishment and improvement of the method will make it convenient for further researches.