This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Wajchman, H. J. Articles by Dombrowski, K. E. Search for Related Content PubMed PubMed Citation Articles by Wajchman, H. J. Articles by Dombrowski, K. E.

Ex Vivo Expansion of CD8⁺CD56⁺ and CD8⁺CD56^- Natural Killer T Cells Specific for MUC1 Mucin

¹ VA Medical Center, Atlanta, Georgia; ² OED & Associates, Hermann, Missouri; and Departments of ³ Pathology, ⁴ Urology, and ⁵ Biochemistry and Microbiology & Immunology Emory University, Atlanta, Georgia

	ABSTRACT

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Prostate cancers express MUC1, but nearly all metastatic cells lack HLA class I molecules. Thus, a lymphocyte population that can sense its antigenic environment, while also able to react to stimuli of natural killer (NK) cells, may be a more versatile effector cell population for antitumor immune responses. Herein, we report that tumor-specific MUC1 peptide, interleukin 2, and interleukin 12 act synergistically to stimulate the ex vivo expansion of CD8⁺CD56^- T cells and CD8⁺CD56⁺ natural killer T (NKT) cells from the peripheral blood mononuclear cells of prostate cancer patients, as well as healthy male and female donors. Both the CD56⁺ NKT cells and CD56^- T cells lysed allogeneic mucin-bearing target cells, as well as NK target cells, but not lymphokine-activated killer target cells. However, the CD56⁺ NKT cells displayed a 2-fold greater cytolytic activity than the CD56^- T cells. The mucin-specific cytolytic activity and NK cytolytic activities for both lymphocyte populations were independent of HLA class I and CD1 molecules. The CD56^- T cells up-regulated CD56 with continued antigenic stimulation in the presence of interleukin 12, suggesting that CD8⁺CD56^- T cells are NKT cells. However, CD56⁺ NKT cells expand poorly to continued stimulation. All mucin-stimulated NKT cells exhibited the activated/memory CD45RO phenotype. The NKT cell lines express the /ß T-cell receptor (TCR). The TCR repertoire was limited and varied with cell line, but was not the V24Vß11 TCR typically associated with NKT cells. Whereas CD161 is generally considered a marker of NKT cells, the mucin-stimulated NKT cells did not express this marker. Thus, we have described two phenotypically distinct NKT types that do not display a biased TCR repertoire, but do display specificity for a tumor-specific peptide antigen (CTL-like activity), as well as HLA class I-deficient target cells (NK-like activity).

	INTRODUCTION

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Tumor cells express unique antigenic determinants, such as the core protein of MUC1 mucin, which render them targets for immune recognition and eradication. MUC1 mucin is a highly glycosylated transmembrane protein expressed by the glandular epithelial cells of the prostate and other organs (1 , 2) . The hallmark of MUC1 mucin is the 20-amino acid sequence PDTRPAPGSTAPPAHGVTSA, which is tandemly repeated 25–100 times per individual (3) . The role for this structural polymorphism in the function of mucin is not known. The oligosaccharide side chains of MUC1 mucin are O-linked to the Ser and Thr residues in the tandem-repeat domain (1) . The carbohydrate side chains form discontinuous patches masking much of the core protein (4) . However, malignant cells express a hypoglycosylated form of MUC1 mucin (1) . This hypoglycosylation results in the exposure of otherwise-masked antigenic determinants on the core protein. These unmasked epitopes are tumor-specific and provide a potential target for the immunotherapy of cancer.

Tumor cells also present a unique challenge to the adaptive immune response by evading immune detection. Diverse mechanisms that include inappropriate activation of antigen presenting cells (APCs; Ref. 5 ), antigen shedding, and down-regulation of HLA class I molecules have been described (5, 6, 7) . Under these circumstances, CTLs, which recognize antigen in context of HLA class I molecules, have distinct limitations in their ability to kill tumor cells.

Where the adaptive CTL immune response is limited, the innate immune response of natural killer (NK) cells can play a critical role in guarding the body against malignancies. The advantage that NK cells provide to immune surveillance is that they recognize and spontaneously lyse tumor cells without the need for prior sensitization or the involvement of HLA class I molecules (8) . Thus, the synergy between innate and adaptive immunity can play a key role in antitumor responses.

Even with both arms of the immune system in place, it is often the case that many tumors fail to elicit a strong in vivo antitumor immune response. Thus, targeting tumor-specific antigens by vaccination or the adoptive transfer of CTL and NK cells (9 , 10) is a promising means of specifically eradicating tumor cells to overcome a weak in vivo antitumor immune response.

CD8⁺ CTLs that are specific for the core protein of MUC1 mucin expressed by tumor cells have been expanded ex vivo from peripheral blood- and tumor-infiltrating lymphocytes (TILs) of patients with breast cancer (11 , 12) , ovarian cancer (12 , 13) , pancreatic cancer (14) , or MUC1-mucin expressing myeloma (15) . Although synthetic MUC1 peptides (12) , autologous APCs transfected with MUC1 mucin cDNA plus inhibitors of glycosylation to ensure exposure of tumor-specific epitopes (16 , 17) , and allogeneic tumor cells (15) have been used to stimulate the mucin-specific CTLs, each of these antigenic forms of mucin has yielded similar results. Namely, the mucin-stimulated CTLs express the ß T-cell receptor (TCR). In addition, although the cytolytic activity against MUC1-bearing target cells has been variable, and oftentimes weak, the mucin-stimulated CTLs are specific for the mucin core protein and generally recognize and lyse HLA-matched and unmatched MUC1-expressing target cell lines irrespective of tissue. These CTL cell lines generally lack NK cytolytic activity and, therefore, may be limited in their ability to kill tumor cells that fail to express HLA class I molecules.

The expansion of tumor-specific CTLs ex vivo from peripheral blood mononuclear cells (PBMCs) and TILs is evidence that such T cells were present because of prior in vivo exposure to tumor (12) . In this communication, we show that PBMCs from patients with a MUC1-expressing adenocarcinoma of the prostate, as well as from men and women who had no history of cancer or immunosuppressive disorders (i.e., healthy), were stimulated with synthetic mucin peptides in the presence of interleukin 2 (IL-2) and interleukin 12 (IL-12), type 1 cytokines that direct strong Th1 cellular immune response. The resulting T-cell lines displayed cytolytic activity against both mucin-bearing and NK target cells and comprised CD3⁺CD8⁺ natural killer T (NKT) cells; no classical CD3^-CD56⁺ NK cells were present.

Two major subsets of NKT cells have been described (18) . The more widely studied subset is the NKT cell that is dependent on the presentation of the exogenous glycolipid antigen, -GalCer, through CD1d, a member of the family of nonpolymorphic, class I-like antigen-presenting molecules (18) . In humans, CD1d-dependent NKT cells generally display an invariant V24 TCR repertoire. Considerably less is known about the CD1d-independent NKT cells, but this subset expresses a nonbiased TCRß repertoire. The tumor-specific T cell lines that are described herein fulfill the criteria of this latter class of NKT cells. We believe that the observations presented in this communication are the first to describe human CD8⁺ NKT cells having a TCR repertoire that is not biased toward V24 and Vß11 and displays both tumor peptide antigen-specific and NK-like cytolytic activities that are independent of HLA class I and CD1 molecules for their cytolytic activity.

	MATERIALS AND METHODS

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Materials.
RPMI 1640, penicillin, streptomycin, L-glutamine, and 1 mM sodium pyruvate were purchased from Mediatech (Washington, DC). FCS was purchased from Hyclone (Logan, UT). Oxaloacetate/pyruvate/insulin was purchased from Sigma (St. Louis, MO). Recombinant human IL-12 was a generous gift from Genetics Institute (Cambridge, MA). Recombinant human IL-2 was obtained from the National Cancer Institute (Frederick, MD). The 40-amino acid MUC1-mtr₂ peptide {[PDTRPAPGSTAPPAHGVTSA]₂} was synthesized by the Emory University Microchemical Facility (Atlanta, GA). All other chemicals were of reagent grade or tissue-culture grade purity.

Cell Lines.
Tumor target cell lines, MCF-7 (breast), CaOV3 (ovary), and LN-CaP (prostate) were purchased from American Type Culture Collection (Rockville, MD). These cell lines were maintained in medium recommended by the American Type Culture Collection, which was supplemented with 10% FCS, oxaloacetate/pyruvate/insulin (Sigma), and 2 mM L-glutamine (Mediatech). K562, a NK target cell, and Raji, a lymphokine-activated killer (LAK) target cell, were purchased from American Type Culture Collection. Both K562 and Raji cells were cultured in RPMI 1640 containing 10% FCS and 2 mM L-glutamine. The EBV-transformed B cell line 32993 was cultured and maintained as described previously (19) . In addition, B cells from each donor were immortalized by transformation with EBV as described previously (19) for use as APCs. All cultures contained penicillin and streptomycin and were treated with fungazone (Mediatech) as needed.

Primary MUC1-Specific CTL and Long-Term CTL Cell Lines.
PBMCs were isolated from whole blood donated by men and women without any history of cancer or immunological disorders (i.e., healthy individuals) and nine men with diagnosed prostate cancer (CaP) after obtaining informed consent. PMBCs from two CaP patients were used to develop and characterize cell lines. Cell line CaP00–0705 was derived from a 44-year-old Caucasian male with clinical stage T₃N₀N₀ adenocarcinoma of the prostate, Gleason grade 3 + 4 = 7. The cancer involved both lobes of the prostate with predominance in the left lobe (80% of the biopsy cores). Serum prostate specific antigen was 37 ng/ml. Bone scintigraphy and computerized tomography of the abdomen and pelvis failed to document metastatic spread. There was a suggestion of increased uptake on tracer in the region of the right iliac vessels on ¹¹¹In prostatscintigraphy. However, subsequent bilateral pelvic lymph node dissection proved negative for metastases. This donor was otherwise healthy and was taking no medication. Patient was treated with total androgen blockade for three months followed by radical retropubic prostatectomy. Blood samples from this donor for this study were drawn during preoperative work up.

The CaP00–0720 cell line was derived from the PBMCs of a 61-year-old African-American male presenting with elevated prostate specific antigen who underwent radical retropubic prostatectomy with bilateral pelvic lymphadenectomy. Histological examination revealed clinical stage T₂ adenocarcinoma of the prostate, Gleason grade 2 + 3 = 5, involving 10% of the right lobe. The prostatic capsule, seminal vesicles, surgical margins, and lymph nodes were free of disease. Other medical problems included hyperlipidemia, obesity, hypertension, alcohol abuse, and degenerative joint disease. There was no history of immune disease or disorders. The patient took no medication. Blood from this donor was drawn for this study during preoperative work up.

The HXY00–0103 and HXX98–0408 cell lines were derived from a 37-year-old male and 31-year-old female, respectively, with no history of cancer or immunosuppressive disorders. These donors were considered to be "healthy."

To generate cell lines, blood was collected from the healthy and CaP donors in vacutainer tubes containing heparin sulfate or ACD solution E tubes. Blood was diluted 1:2 with Dulbecco’s PBS (calcium and magnesium free), and PBMCs were isolated by Ficoll-Paque (Pharmacia, Piscataway, NJ) density gradient centrifugation. Purified PBMCs were washed three times with PBS, aliquoted, and stored at -70°C in Origen DMSO freeze medium (Igen International, Inc, Gaithersburg, MD). An aliquot of freshly isolated PBMCs was resuspended in RPMI 1640 containing 10% FCS and seeded at 1 x 10⁶ PBMCs/well in a Linbro 24-well plate (ICN, Costa Mesa, CA) with 1 x 10⁵ -irradiated (3000 rads) autologous PMBCs as APCs in a final volume of 1 ml of complete medium containing 10% FCS, 1% oxaloacetate/pyruvate/insulin, 2 mM L-glutamine, 50 µM 2-mercaptoethanol, 10–50 IU rIL-2, 25 µg of the 40-amino acid mucin peptide MUC1-mtr₂ and 2 ng of rIL-12. The cells were cultured in an atmosphere of 5% CO₂ in air at 37°C. For short-term experiments, cells were harvested at 6–8 days and used as effector cells in ⁵¹Cr-release cytotoxicity assays. To generate cell lines, cultures were fed every 3–5 days with complete medium [supplemented with either 10–50 units/ml human rIL-2 or 10–20% Lymphocult-T (Biotest, Denville, NJ)] and restimulated every 7–14 days with 25 µg of MUC1-mtr₂ peptide, irradiated autologous PBMCs, and 2 ng/ml human rIL-12. Alternatively, PBMCs were replaced with autologous, irradiated EBV-transformed B cells pulsed with the peptide. Lymphocytes were maintained at 60% confluence in wells. Debris and dead cells were removed by harvesting and pooling wells, followed by a Ficoll-Paque gradient. Viable cells were then returned to culture in 24-well plates with fresh complete medium at 1 x 10⁶ cells/well and restimulated as indicated above. Established cell lines were maintained in 24-well plates and expanded in 25-cm² or 75-cm² flasks (Falcon) with continued feeding and stimulation as before, but using 10 µg of the MUC1-mtr₂ peptide.

Cytotoxicity Assays.
Target cells were loaded with radioactivity by incubating 5 x 10⁶ cells with 100 µCi of Na₂[⁵¹CrO₄] (NEN, Boston, MA) in 200 µl of RPMI 1640 containing 10% FCS for 1 h at 37°C. Loaded cells were washed three times with RPMI 1640 containing 2% FCS and plated at 1 x 10⁴ cells/well in 96-well U-bottomed plates (Corning) containing effector cells at various concentrations to obtain the desired E:T ratios in a final volume of 200 µl. Effector cells were routinely >90% viable by trypan blue exclusion. All data points were performed in triplicate and are reported as the mean ± SD. Spontaneous release was determined by incubation of target cells in assay medium alone. Maximum release was determined by addition of 100 µl of 2% Triton X-100. Plates were incubated for 4 h in 5% CO₂ atmosphere at 37°C. Supernatants were harvested using a Skatron supernatant collection system (Skatron Instruments, Sterling, VA) and counted in a Packard Cobra II gamma counter. Percentage specific cytolytic activity was determined by the following equation: (experimental release - spontaneous release)/(maximum release - spontaneous release) x 100.

Flow Cytometric Analysis.
The phenotype of cell lines were determined using monoclonal antibodies (mAbs) conjugated with FITC, PE, PerCP, or APC [all mAbs were purchased from Caltag (Burlingame, CA) unless otherwise noted]. Cells were stained with mAbs directed against CD3-FITC (MEM-57, IgG-2a), CD4-FITC (S3.5, IgG-2a), CD8-FITC (3B5, IgG-2a), CD8-PerCP, CD56-PE, CD56-APC (NK1-nbl-1, IgG-1), CD45 RO-PE (UCHL1, IgG2a), /ß TCR-PE [BMA 031, IgG-2b; T10B9.1A-31, IgM (PharMingen)], / TCR-PE (B1.1, IgG1; PharMingen), V24-FITC, and CD161-PE. Briefly, 4 x 10⁵ cells were stained in "V"-bottom 96-well plates (Nunc) by incubating with the respective mAb for 30 min on ice. Plates were spun to collect cells and washed three times with PBS containing 3% FCS and 0.05% NaN₃. Labeled cells were analyzed by a FACScalibur (Becton Dickinson) using Cellquest and FloJo data analysis software.

TCR V and Vß Expression.
Total mRNA was prepared from 5 x 10⁶ NKT cells from the male donor using the Qiagen RNA miniprep kit according to the instructions of the manufacturer. cDNA was prepared by transcribing 2 µg of total mRNA using the TaqMan reverse transcriptase kit (PE Biosystems) using an oligodeoxythymidylic acid primer in a final reaction volume of 100 µl. Reaction conditions were 10 min at 25°C, 30 min at 48°C, and 5 min at 95°C. The TCR V and Vß repertoire was determined by relative real-time PCR using the Bio-Rad iCycler (Hercules, CA). Each PCR reaction contained 5 µl of a 1:5 dilution of the cDNA and 1 µM of the V or Vß chain primers with their respective C and Cß primers as previously described (20, 21, 22, 23, 24, 25, 26) in a volume of 12.5 µl plus an equal volume of SyberGreen 2x PCR MasterMix (PE Biosystems). cDNA was amplified for 40 cycles of 1 min at 93°C, 1 min at 55°C, and 1 min at 72°C. Immediately after the last cycle, a melt-curve was performed to detect the purity of the PCR products. Product size was confirmed by electrophoresis using a 1.5% agarose gel. The amount of each PCR product was quantitated by normalization against a 5-fold dilution series of the V1.5 product from Jurkat T cells.

	RESULTS

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Cytolytic Activity and Phenotype of Primary Cultures Stimulated with Tumor-Specific Antigen, IL-2, and IL-12.
Nonstimulated PBMCs from CaP patient CaP00–0705 displayed cytolytic activity against MUC1 mucin-expressing target cells ( 30% at an E:T ratio of 40:1) and NK target cells, K562 ( 50% at an E:T ratio of 40:1), but no lysis of the LAK target cell line Raji (Fig. 1A) . To test whether these levels of cytolytic activity were representative of nonstimulated PBMCs from CaP patients, PBMCs from eight additional CaP donors were assayed. These PBMCs displayed a mean cytolytic activity at an E:T ratio of 40:1 of 9.4% against MCF-7 (range 2.7–18.8%), 26.3% against K562 (range 2.5–50.4%), and 10% against Raji cells (range, 2.2–11.8%). Unstimulated PBMCs from the healthy male donor HXY00–0103 also displayed innate cytolytic activity against the NK target cell line, K562, with 80% specific target lysis at an E:T ratio of 40:1, which titered to 0% with halving dilutions (Fig. 1B) . The unstimulated PBMCs also displayed 40% specific cytolytic activity against MCF-7 cells at an E:T ratio of 40:1, which also titered to 0% lysis (Fig. 1B) . These cells did not display significant levels of cytolytic activity against the LAK target cell line, Raji (Fig. 1B) . Thus, nonstimulated PBMCs from both healthy and CaP donors displayed variable levels of endogenous cytotoxicity against MUC1-expressing and NK target cells, but little to no endogenous LAK activity.

View larger version (19K): [in this window] [in a new window]   Fig. 1. Interleukin 2 (IL-2) and interleukin 12 (IL-12) act synergistically in the presence of tumor-specific antigen to direct T cells away from lymphokine-activated killer activity. Peripheral blood mononuclear cells from prostate cancer donor CaP00–0705 and healthy male donor HXY00–0103 were stimulated in parallel in the absence (-) or presence (+) of the mucin peptide, interleukin 2, and interleukin 12, as indicated under each panel and described in detail in "Materials and Methods." Cytolytic activity was measured by 51Cr-release assay at initiation of the primary culture and seven days later against mucin-expressing MCF-7 target cells (), the natural killer target cell K562 (), and the lymphokine-activated killer target cell line Raji (). ND, not determined.

Because IL-12 stimulates Th1 responses and nonrestricted NK effector cells, it would be reasonable to expect that the NK-like cytolytic activity described in Fig. 1F could be accounted for by the presence of both CTL and NK cells. However, we observed that MUC1-expressing MCF-7 cells are not lysed by the human NK cell line NK3.3 at an E:T ratio of 40:1 (data not shown). Because MCF-7 is not a target for classical NK cells, it did not seem likely that the cytolytic activity against this target cell could be attributable to classical NK cells. Nonetheless, the phenotype of this primary culture was determined to rule-out the possibility that the lysis of K562 was attributable to classical NK cells. The 7-day primary culture stimulated with MUC1 peptide, IL-2, and IL-12 (Fig. 1F) contained >90% CD3⁺ T cells (not shown). Two-color flow cytometric analysis showed that 53% of the CD3⁺ T cells expressed CD4⁺ (Fig. 2A) , but <3% the cells were CD4⁺CD56⁺ (Fig. 2A) , a value that was not distinguishable from background staining with the isotype control (not shown). The culture also contained 43% CD8⁺ T cells, 7% of which coexpressed CD56 (Fig. 2B) , suggesting the presence of NKT cells. No significant numbers of cells displayed the CD3^-CD56⁺ phenotype characteristic of classical NK cells (not shown). Thus, the NK-like activity cannot be accounted for by classical NK cells, but possibly by "promiscuous" CTL or NKT cells, both of which would display cytolytic activity of CTL and NK cells.

View larger version (36K): [in this window] [in a new window]   Fig. 2. Phenotype of primary culture stimulated with mucin-peptide in the presence of interleukin 2 and interleukin 12. The primary culture described in Fig. 1 F was analyzed by two-color flow cytometry using anti-CD56-PE in combination with either anti-CD4-FITC or anti-CD8-FITC. Inset, values listed are the percentage of cells in each respective quadrant. Background isotype staining using IgG1-FITC and IgG2a-PE was <3%.

View larger version (19K): [in this window] [in a new window]   Fig. 3. Cytolytic activity of cell lines expanded from the peripheral blood mononuclear cells of a prostate cancer patient. The CaP00–0720 cell line was established by ex vivo stimulation of peripheral blood mononuclear cells using mucin peptide, interleukin 2, and interleukin 12 as described in "Materials and Methods." The cytolytic activity was measured 20 days after the fourth stimulation (A) and 13 days after the eighth stimulation (B) against MCF-7 cells (), K562 cells (), and Raji cells ().

View larger version (20K): [in this window] [in a new window]   Fig. 4. Mucin-stimulated T cell lines recognize mucin-bearing target cells. The F2 cell line from the healthy male donor (A) and the E8 cell line from the healthy female donor (B) were assayed for cytolytic activity 5 days poststimulation by 51Cr-release assay as described in "Materials and Methods" against the MUC1-expressing MCF-7 (), the MUC1 mucin-expressing ovarian adenocarcinoma cell line CaOV3 (), the NK target K562 (), the prostate cancer cell line LN-CaP (), and the lymphokine-activated killer target Raji ().

The observations that multiple mucin-expressing tumor cell lines derived from different tissues are lysed by the mucin-stimulated T cells support the idea that the cytolytic activity is directed against mucin and not tissue-specific determinants. Nevertheless, the mucin specificity of the cell lines was also assayed to further test the specificity of the cell lines for mucin, against the HLA-positive cell lines MS and MS-MUC1, the latter of which was stably transfected with MUC1 mucin. The mucin-expressing MS-MUC1 cell line, but not the non-mucin-expressing MS cell line, was lysed by the mucin-stimulated T cell line (Fig. 5A) . In addition, F2 (Fig. 5B) and E8 (Fig. 5C) cell lines lysed the allogeneic EBV-transformed B cell line 32993 pulsed with the 40-amino acid mucin peptide MUC1-mtr₂, but did not lyse the non-peptide-pulsed EBV-transformed B cells. Although the specific target lysis against the MUC1-expressing MS-MUC1 cell line and the MUC1-pulsed EBV-transformed B cells was lower than that against the tumor cell lines (e.g., MCF-7, Ca-OV3), a range of 10–30% specific target lysis by mucin-stimulated T cells against MUC1-transfected cell lines and target cell loaded with exogenous MUC1 peptide is not an uncommon observation (12 , 27, 28, 29, 30) . Taken together, the results presented above are consistent with the idea that the mucin-stimulated T-cell lines are specific for endogenously synthesized and exogenously loaded mucin.

View larger version (17K): [in this window] [in a new window]   Fig. 5. Mucin-stimulated T cell lines recognize mucin-bearing target cells. A, the cytolytic activity of the HXY00–0103 cell line from the male donor was determined by 51Cr-release assay 7 days poststimulation against the non-MUC1-expressing epithelial cell line, MS (), the MS cell line stably transfected and expressing MUC1, MS-MUC1 (), and Raji cells (). Cytolytic activity of the HXY00–0103 (B) and E8 (C) cell lines against allogeneic 32993 EBV-transformed B cells alone (-) and the 32993 B cells pulsed with MUC1 peptide (+) as described in "Materials and Methods."

Mucin-Specific T-Cell Lines Display the Phenotype of NKT Cells.
Our results have reproducibly shown that the T-cell lines established by stimulating PBMCs with mucin peptide, IL-2, and IL-12 display both mucin-specific and NK-like cytolytic activities. The primary culture displayed the phenotype of CD8⁺ CTL with a minor population of CD8⁺CD56⁺ NKT cells, but no classical CD3^-CD8^-CD56⁺ NK cells (Fig. 2) . Nevertheless, a detailed phenotypic analysis of the mucin-specific T cell lines described above (see Figs. 3 4 5 ) was performed to determine the identity of the cytolytic lymphocytes.

The E8 cell line generated from the PBMC of the healthy female donor is representative of our observations. This cell line was >99% CD3⁺ (Fig. 6A) , and nearly half expressed CD56 (Fig. 6B) . The lack of CD3^- cells suggested that classical CD3^-CD56⁺ NK cells were not present, but rather that the cell line contained CD3⁺CD56^- T cells and CD3⁺CD56⁺ NKT cells. This cell line did not contain CD4⁺CD56⁺ NKT cells (Fig. 6C) , but rather approximately 50% of the T cells were CD8⁺CD56⁺ NKT cells (Fig. 6D) . In addition, the CD8⁺ T cells expressed the /ßTCR (Fig. 6E) , but not the TCR (not shown), with staining ranging from high to low. Furthermore, the CD8⁺ T cells were virtually all 100% memory/effector cells based on the expression of CD45RO (Fig. 6F) .

View larger version (35K): [in this window] [in a new window]   Fig. 6. Phenotype of mucin-stimulated T-cell line. Flow cytometric analysis was performed on the HXX98–0408 cell line as described in "Materials and Methods" using the antibodies indicated in each panel. Inset, values listed are the percentage of cells in each respective quadrant. Background isotype staining using anti-IgG1-FITC and anti-IgG2a-PE was <1%. TCR, T-cell receptor.

View larger version (32K): [in this window] [in a new window]   Fig. 7. Phenotype of mucin-stimulated T-cell lines from a healthy donor and a prostate cancer patient. Flow cytometric analysis was performed on the T-cell lines from the healthy male donor HXY00–0103 (A) and prostate cancer patient CaP00–0705 (B) using anti-CD8 and anti-CD56 as described in "Materials and Methods." Inset, values listed are the percentage of cells in each respective quadrant. Background isotype staining using anti-IgG2a-PE was <1%.

View larger version (56K): [in this window] [in a new window]   Fig. 8. Interleukin 12 (IL-12) contributes to the expression of CD56 on natural killer T cells. The mucin-stimulated T-cell line HXY00–0103 from the healthy male donor after 6 months in culture was stimulated for 1 week with MUC1 peptide and interleukin 2 in the absence (- IL-12, left column) or presence (+ IL-12, right column) of interleukin 12 as indicated. The phenotype of the cell cultures was measured 6 days poststimulation. Values listed are the percentage of cells in each respective quadrant. TCR, T-cell receptor.

View larger version (47K): [in this window] [in a new window]   Fig. 9. CD56- T cells having antigen-specific and natural killer-like cytolytic activities give rise to natural killer T cells. The phenotype (A) and cytolytic activity (B) of the HXY00–0103 MUC1-stimulated cell line were determined as described in "Materials and Methods." Cytolytic activity was measured by 51Cr-release assay using the MCF-7 (), K562 (), and Raji () cell lines. The CD56+ T cells were selected by magnetic bead separation using anti-CD56 (Miltenyi) according to the instructions of the manufacturer. The phenotype (C) and cytolytic activity (D) of the CD56+ T-cell subset was determined as described for A and B, respectively. Similarly, the phenotype (E) and cytolytic (F) activity of the CD56- T cells that did not bind to the anti-CD56 magnetic beads were also determined as described for A and B, respectively. G and H, the cytolytic activity of the CD56+ and CD56- populations, respectively, was measured by 51Cr-release assay using the MS (), MS-MUC1 (), and Raji () cell lines. I–K, the unseparated HXY00–0103 cell line and the CD56+ and CD56- populations were stimulated one time with MUC1 peptide, interleukin 2, and interleukin 12. Flow cytometric analysis was performed on the respective populations 14 days poststimulation by two-color flow cytometry using anti-CD8-APC and anti-CD56-PE. Values in each flow cytometric analysis plot indicate the percentage of cells in that quadrant.

View larger version (19K): [in this window] [in a new window]   Fig. 10. A, cytolytic activity of MUC1-specific natural killer T cells is independent of HLA class I molecules. The HXY00–0103 cell line was assayed for cytolytic activity by 51Cr-release assay as described in "Materials and Methods" against the MCF-7 cells () or MCF-7 cells preincubated with 10 µg/ml W6/32 (), 20 µg/ml W6/32 (), 10 µg/ml anti-IgG2a isotype monoclonal antibody (mAb; ). The cytolytic activity against K562 cells preincubated without () or with () 20 µg/ml W6/32 and Raji cells () was also measured. B, cytolytic activity of mucin-specific natural killer T-cell lines is independent of CD1 molecules. The HXX98–0408 cell line was assayed for cytolytic activity by 51Cr-release assay as described in "Materials and Methods" against MCF-7 cells alone () and MCF-7 cells preincubated for 1 h at 37°C with 10 µg/ml anti-isotype mAb of anti-isotype mAb (), anti-CD1a mAb (), anti-CD1b mAb (), anti-CD1c mAb (), or anti-CD1d mAb (). The negative control against Raji cells is also shown (). All target cells were preincubated for 1 h at 37°C with or without the antibodies as indicated.

MUC1-Stimulated NKT Cell Lines Do Not Express CD161 or the V24Vß11 TCR.
In addition to the expression of CD56, the expression of CD161 and the V24Vß11 TCR have also generally been regarded as markers that define these lymphocytes. However, <8% of the CD8⁺CD56⁺ NKT cells from the HXY00–0103 (Fig. 11A) , CaP00–0705 (Fig. 11B) , and HXX98–0408 (not shown) cell lines expressed CD161. This low level of CD161 expression was not considered significant because subsequent flow cytometric analysis of these cell lines did not display any staining with anti-CD161 (not shown). The expression of the V24 TCR on the mucin-specific ßTCR NKT cell lines was also tested. However, flow cytometric analysis using anti-V24 mAb showed that the HXY00–0103 (Fig. 11C) , CaP00–0705 (Fig. 11D) , and HXX98–0408 (not shown) cell lines did not express this V chain.

View larger version (38K): [in this window] [in a new window]   Fig. 11. NKT cells stimulated by MUC1 mucin peptide, interleukin 2 and interleukin 12 do not express CD161 or V24. Representative two-color flow cytometric analysis of the established natural killer T-cell lines from the healthy male donor HXY00–0103 (A) and the prostate cancer patient CaP00–0705 (B) using anti-CD8-PerCP and anti-V24-FITC. Representative two-color flow cytometric analysis was also performed on the HXY00–0103 (C) and CaP00–0705 (D) cell lines using anti-CD8-PerCP and anti-CD161-PE. Samples were prepared and analyzed as described in "Materials and Methods." Inset, values listed in each quadrant indicate the percentage of positive-staining cells.

View larger version (24K): [in this window] [in a new window]   Fig. 12. T-cell receptor (TCR) V and Vß expression of mucin-specific natural killer T-cell lines. Relative real-time reverse-transcription PCR was performed using mRNA from the mucin-stimulated natural killer T-cell lines to test for the expression of the V (A) and Vß (B) TCR repertoire for the HXY00–1003 cell line and the V (C) and Vß (D) TCR repertoire for the CaP00–0705 cell line. Dotted lines, relative expression levels of V TCR <2 and Vß <4 were considered to be background based on melt-curve analysis, which was confirmed to be primer-dimers by agarose gel electrophoresis.

	DISCUSSION

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

NKT cells that express a nonbiased TCRß repertoire have also been reported (31) . Although the target specificity of these NKT cells is not known, they are activated independent of CD1 molecules. Herein, we described the ex vivo expansion of CD8⁺CD56⁺CD161^- NKT cells from the PBMCs of CaP patients and healthy male and female donors. The V and Vß TCR repertoire of the NKT cell lines varied among individuals, but none of the cell lines expressed V24 or Vß11. The CD8⁺CD56⁺CD161^- NKT cells lysed target cells that express MUC1 mucin or were pulsed with mucin peptide regardless of the tissue of origin. These results suggest that the NKT cells are specific for MUC1 mucin. In addition, the mucin-stimulated NKT cells lysed NK target cells to the same extent as mucin-bearing target cells, but the mechanism by which the NK target cells are recognized is not yet known. However, unlike the V24⁺Vß11⁺ NKT cells that often display high levels of LAK activity, the mucin-specific V24^-Vß11^- NKT cells displayed little to no LAK activity. This suggests that the peptide antigen might play a key role in directing antigen specificity of NKT cells and away from nonspecific LAK activity.

In addition to the CD8⁺CD56⁺ NKT cells described here, CD8⁺CD56^- mucin-specific T cells were also expanded. Like the CD8⁺CD56⁺ NKT cells, the CD8⁺CD56^- T cells also had mucin-specific and NK-like cytolytic activities, but no LAK activity. However, the cytolytic activity of the CD8⁺CD56^- T cells was 2-fold lower than that of the CD8⁺CD56⁺ NKT cells. Although CTLs have been reported to display "promiscuous" cytolytic activity against both MCF-7 and K562 (44) , we concluded that the CD8⁺CD56^- T cells described in this communication were a unique population of NKT cells based, in part, on their cytolytic activity profile. Moreover, CD56 was up-regulated on the CD56^- population to yield the CD8⁺CD56⁺ NKT cell phenotype, as well as the enhanced cytolytic activity in response to IL-12. The role that CD56 plays in the enhanced cytolytic activity is being investigated. Thus, these results suggest that mucin peptide, in combination with IL-2 and IL-12, resulted in the ex vivo expansion of CD8⁺CD56⁺ and CD8⁺CD56^- NKT cells, the latter of which gives rise to the CD8⁺CD56⁺ NKT cell phenotype.

Mucin-specific CD8⁺V24^-Vß11^- NKT cells were generated from cancer patients, as well as healthy men and women. These observations suggest that parameters such as gender and in vivo exposure to tumor may not have a large impact on the ex vivo expansion of these cells for adoptive immunotherapy of cancer. In addition, the generation of lymphocytes having antimucin cytolytic activity also argues against the assumption that an individual need to be presensitized by the tumor in vivo for successful ex vivo expansion (12) . Cell lines generated from the CaP patients also had nearly 50% higher levels of CD8⁺CD56⁺ NKT cells than cell lines generated from healthy individuals. The reason for this difference is not known. However, we observed that the CD56⁺ NKT cell population responded poorly to continued antigenic stimulation, whereas the CD56^- NKT cells and cell lines containing both CD56⁺ and CD56^- NKT cells were able to be maintained in long-term culture. This may suggest that CD56⁺ NKT cells are terminal and that a steady state between CD56⁺ and CD56^- NKT cells is maintained by the up-regulation of CD56 on the CD56^- NKT cells. Although multiple mechanisms contribute to a weak in vivo antitumor immune response, we speculate that cancer patients can generate an in vivo antitumor NKT cell response, but these cells may not adequately proliferate in response to the tumor burden.

The mucin-specific NKT cells described herein share similarities with previously reported mucin-specific CTLs (12, 13, 14, 15 , 29 , 45 , 46) , but they also display characteristics that also distinguish them as a unique T-cell population. Although the CD8⁺CD56^- NKT cells phenotype described gives the impression of a mucin-specific CTL, CD56 was shown to be up-regulated. Thus, we do not believe that the expression of this NK marker is alone sufficient to distinguish between an NKT cell and a CTL. However, we believe that a more appropriate definition of an NKT cell should be based on criteria of effector function. Namely, although the CD56⁺ and CD56^- NKT cells and CTLs display mucin-specific cytolytic activity against non-HLA matched allogeneic target cells, the NKT cells described here displayed strong NK-like cytolytic activity, whereas the previously reported mucin-specific CD8⁺ CTLs were devoid of NK-like cytolytic activity. This difference alone suggests that the lymphocytes reported herein are distinct from previously reported mucin-specific CTLs. Other differences include the observations that the cytolytic activity of the CD8⁺ CTLs was inhibited by the W6/32 anti-HLA class I mAb, whereas the cytolytic activity of the NKT cells reported herein was not inhibited by the antibody. This suggests that, although the CD8⁺ CTLs are not restricted to a single HLA haplotype, TCR-HLA interactions play an important role in target recognition. By contrast, the cytolytic activity of the mucin-specific NKT cells was independent, not only of HLA class I molecules, but also the CD1 family of antigen-presenting molecules. The mechanism of target cell recognition by the mucin-specific NKT cells is not yet known, but may involve intercellular adhesion molecules, similar to the HLA-unrestricted, mucin-specific CD8⁺ CTLs (27) . Work is in progress to determine the extent to which intercellular adhesion molecules play in the antigen-specific and NK-like cytolytic activities of NKT cells.

Adoptive immunotherapy using tumor-specific CD8⁺ CTLs provides an avenue for the eradication of tumor having minimal to no nonspecific damage of normal cells. However, the effectiveness of CTLs may be limited because many tumors, including mucin-expressing adenocarcinomas, evade immune detection through mechanisms that include antigen shedding and down-regulation of HLA class I molecules (5 , 6) . Because NK cells also infiltrate tumor and cause regression (8 , 47 , 48) , we speculate that the dual cytolytic activities of the effector/memory NKT cells described herein that allow the cells to sense their antigenic environment as a CTL, while also having the ability to react to stimuli of NK cells when loss of HLA molecules occur, may make these lymphocytes a more versatile cell type for the active immunity against cancer and future prevention against relapse.

	ACKNOWLEDGMENTS

 
We thank Dr. Judith A. Kapp and the members of her laboratory for their discussions on this work, Drs. Olivera Finn and John McKolanis for the MS and MS-MUC1 cell lines, and Dr. Jacki Kornbluth for the NK3.3 cell line. We also thank the Genetics Institute for a gift of human IL-12.

	FOOTNOTES

 
Grant support: This work was supported by a Merit Review award from the Department of Veterans Affairs (K. E. Dombrowski).

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

Note: H. J. Wajchman is currently at AARoN Immunotechnologies Laboratory, Atlanta, Georgia.

Requests for reprints: Kenneth E. Dombrowski, VA Medical Center, 1670 Clairmont Road, Decatur, GA 30033. Phone: (404) 327-4981; E-mail: ken.dombrowski{at}med.va.gov

Received 11/14/02. Revised 10/15/03. Accepted 11/19/03.

	REFERENCES

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 

1. Hanisch F-G., Muller S. MUC1: the polymorphic appearance of a human mucin. Glycobiology, 10: 439-449, 2000.[Free Full Text]
2. Zhang S., Zhang H., Cordon-Cardo C., Ragupathi G., Livingston P. Selection of tumor antigens as targets for immune attack using immunohistochemistry: protein antigens. Clin. Cancer Res., 4: 2669-2676, 1998.[Abstract]
3. Debailleul V., Laine A., Huet G., Mathon P., d’Hooghe M., Aubert J., Porchet N. Human mucin genes MUC2, MUC3, MUC4, MUC5AC, MUC5B, and MUC6 express stable and extremely large mRNAs and exhibit a variable length polymorphism. J. Biol. Chem., 273: 881-890, 1998.[Abstract/Free Full Text]
4. Russell B., Moddeman W., Birkbeck J., Wright S., Millington D., Stevens R., Dombrowski K. Surface structure of human mucin using X-ray photoelectron spectroscopy. Biospectroscopy, 4: 257-266, 1998.[CrossRef][Medline]
5. Blades R., Keating P., McWilliam L., George N., Stern P. Loss of HLA class I expression in prostate cancer: implications for immunotherapy. Urology, 46: 681-687, 1995.[CrossRef][Medline]
6. Ferrone S., Marincola F. Loss of HLA class I antigens by melanoma cells: molecular mechanisms, functional significance and clinical relevance. Immunol. Today, 16: 487-494, 1995.[CrossRef][Medline]
7. Agrawal B., Krantz M., Reddish M., Longenecker B. Cancer-associated MUC1 mucin inhibits human T-cell proliferation, which is reversible by IL-2. Nat. Med., 4: 43-49, 1998.[CrossRef][Medline]
8. Whiteside T., Herberman R. The role of natural killer cells in immune surveillance of cancer. Curr. Opin. Immunol., 7: 704-710, 1995.[CrossRef][Medline]
9. Agrawal B., Gendler S., Longenecker B. M. The biological role of mucins in cellular interactions and immune regulation: prospects for cancer immunotherapy. Mol. Med. Today, : 397-403, 1998.
10. Chattophadhyay U. Tumour immunotherapy: developments and strategies. Immunol. Today, 20: 480-482, 1999.[CrossRef][Medline]
11. Finn O. Immunity to epithelial tumors and mucin-based vaccine design. Proc. Am. Assoc. Cancer Res., 36: 675-680, 1995.
12. Wright S., Kilinski L., Talib S., Lowe K., Burnside J., Wu J-Y., Dolby N., Dombrowski K., Lebkowski J., Philip R. Cytotoxic T-lymphocytes (CTL) from humans with adenocarcinomas stimulated by native MUC1 mucin and a mucin peptide mutated at a glycosylation site. J. Immunother., 23: 2-10, 2000.
13. Ioannides C., Fisk B., Jerome K., Irimura T., Wharton J., Finn O. Cytotoxic T cells from ovarian malignant tumors can recognize polymorphic epithelial mucin core peptides. J. Immunol., 151: 3693-3703, 1993.