- Email: [email protected]
Exosomes and the MICA-NKG2D system in cancer
Blood Cells, Molecules, and Diseases 34 (2005) 206 – 213 www.elsevier.com/locate/ybcmd
Exosomes and the MICA-NKG2D system in cancer Aled ClaytonT, Zsuzsanna Tabi Department of Clinical Oncology and Palliative Medicine, Wales College of Medicine, Cardiff University, Velindre Cancer Centre, Whitchurch, Cardiff CF14 2TL, UK Submitted 9 March 2005; revised 15 March 2005 (Communicated by M. Lichtman, M.D., 15 March 2005)
Abstract Exosomes are nanometer sized vesicles, secreted by a diverse range of cell types, whose biological functions remain ambiguous. Several groups have demonstrated the potential of manipulating exosomes for activating cellular immune responses. The possibility that exosomes may inhibit immunological responses, however, has not been widely addressed. We have investigated if exosomes produced by tumor cells can inhibit immunological functions, through modulating expression of the NKG2D receptor by effector cells. Incubating tumor exosomes with fresh peripheral blood leukocytes resulted in a marked reduction in the proportion of NKG2D-positive CD3+CD8+ Cells, and CD3 cells by 48 h. This effect was dose dependent and was shown with exosomes from different tumor cells including breast cancer and mesothelioma. Analysis of tumor exosome-phenotype revealed positive expression of several NKG2D ligands, and antibody blocking experiments revealed the importance of such ligands in driving the reduction in the proportion of NKG2D-positive effector cells. The functional importance of the decrease in NKG2D-positive cells was addressed in vitro cytotoxicity assays. For example a CD8+ T cell line pre-incubated with tumor exosomes had significant decreased capacity to kill peptide-pulsed T2 target cells. These data highlight a role for tumor exosomes bearing NKG2D ligands as a mechanism contributing to cancer immune evasion. D 2005 Elsevier Inc. All rights reserved. Keywords: Exosome; MICA; NKG2D; Cancer
Introduction NKG2D is a homodimeric C-type lectin receptor, expressed by a range of immune cells, including NK, NKT, CD8 ah T cells, gy T cells and macrophages. Ligands for NKG2D comprise the polymorphic MHC class I related, stress-inducible proteins (including MICA and MICB) that exhibit highly restricted expression in vivo, and are generally only found upon infection and on some tumor cells. Additional ligands, discovered more recently, include the ULBP-1, -2, -3 and -4 [1] proteins, so named because of their binding to the cytomegalovirus encoded protein UL16, important for CMV immune evasion through preventing surface
T Corresponding author. E-mail address: [email protected] (A. Clayton). 1079-9796/$ - see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.bcmd.2005.03.003
expression of NKG2D ligands [2 –4]. It is likely that NKG2D is capable of binding yet more ligands, such as Letal [5] and others [6]. The true breadth of ligands for NKG2D is not yet fully appreciated. The activation of NK cells through NKG2D can overcome inhibitory signaling from self-recognition, and therefore directly trigger NK cytotoxic function, while NKG2D provides a costimulatory signal (in addition to signaling through TCR) in T cells. Thus tumor cells (and other cells under stress) expressing ligands for NKG2D are highly susceptible to killing by NK and T cells [7 –9]. Importantly, however, there is mounting evidence demonstrating that tumor cells may subvert this mode of targeting through shedding such ligands from the cell surface, through the activity of matrix metalloproteinases for example [10]. Such data are strengthened by observations that cancer patients have high levels of a
A. Clayton, Z. Tabi / Blood Cells, Molecules, and Diseases 34 (2005) 206 – 213
soluble form of MICA (sMICA) (and in some cases sMICB) in their serum, where such molecules are not detectable in healthy donor sera [11]. The effect of shed ligands results not only in lower levels of cell surface ligand expression, but there is also evidence that soluble ligands downregulate the expression of T cell NKG2D, and decrease the cytotoxicity of NK cells towards MICApositive targets [12]. In this report we have investigated the possibility that tumor cell-derived exosomes express ligands for NKG2D and are capable of modulating effector cell functions. Our data demonstrate that NKG2D is downregulated following exposure to tumor-derived exosomes, and that effector cytotoxic functions are impaired as a result. The production of NKG2D-ligand-bearing exosomes, therefore, may be a novel mechanism for tumor cell immune evasion, and further demonstrates the complexity of interactions possible between exosomes and the immune system.
207
Materials and methods Cell lines The breast carcinoma line T47d was a gift from Prof. R Nicholson, Tenovus, Cardiff University. Two mesothelioma cell lines (named patients 1 and 15) were generated in house from explants of pleural malignant mesothelioma. Jurkat and IB4 cells were from the MRC cooperative, Cardiff University, and were maintained in serum-free AIM-v medium. Other cells were maintained in RPMI1640 (with l-glutamine, pen/strep and HEPES), and 10% FBS (BRL Life Technologies). FBS was depleted of bovine exosomes by pelleting at 100,000 g for 2 h. Exosome purification Cell conditioned medium (48 – 72 h) was cleared of cells and particulates by serial centrifugation (100 g/5 min,
Fig. 1. A diagrammatic explanation of a 3-color flow cytometric method used for analyzing the proportion of NKG2D-positive cells within lymphocyte subsets. The selection of lymphocytes is performed using the FSC/SSC dot plot (region-1, R1). Lymphocytes which stain positive (R2) or negative (R3) for CD3 are identified using the FL-1/FCS plot. Finally, the proportion of NKG2D-positive T-lymphocytes (R1 and R2) are examined using an FL-3(CD8)/FL-2(NKG2D) plot, allowing the CD8+ and CD8 populations to be distinguished. Finally, analysis of non-T cells (R1 and R3) are examined using an FL-1/FL-2 plot. An example of NKG2D-low (a) and an NKG2D-high (b) population is shown.
208
A. Clayton, Z. Tabi / Blood Cells, Molecules, and Diseases 34 (2005) 206 – 213
400 g/5 min, 2000 g/10 min at 4-C) and filtered (0.2 Am). Supernatant was underlayed with 30% Sucrose/D2O (as previously described [13,14] prior to ultracentrifugation at 100,000 g for 3 h. The collected sucrose cushion was diluted in excess PBS, and exosomes pelleted by a further ultracentrifugation step. Exosome quantity was determined by BCA protein assay (Pierce). Flow cytometry and antibodies Analysis of NKG2D expression by peripheral blood lymphocytes was performed by three-color flow cytometry (BD FACScan), using monoclonal antibodies against CD3-FITC (Dako), CD8-Cy5 (BD) and NKG2D-PE (R&D Systems). Analysis was performed on lymphocytes (gating with FSC/SSC) that were CD3 positive or CD3 negative and are expressed as % NKG2D-positive cells (schematically represented in Fig. 1). Antibodies for the known ligands of NKG2D, MICA, MICB, ULBP-1, -2 and -3 together with IgG2a and IgG2b isotype controls were from R&D Systems. Flow-cytometric analysis of ligand expression was performed on live cells (not treated with trypsin—because of the protease-sensitive
nature of the ligands) and on exosome-coated aldehyde– sulphate latex beads (from Interfacial Dynamics). CD8+ T cell cytotoxicity assays A CD8+ T cell line was generated by pulsing autologous dendritic cells with a cocktail of synthetic peptides, with sequences derived from published immunodominant viral epitopes (including influenza, Epstein – Barr virus and cytomegalovirus). Standard chromium release assay was performed with T2 target cells, loaded with Cr51 and pulsed with peptide cocktail (at a final concentration of 10 Ag/ml), at an effector: target ratio of 10:1.
Results The proportion of NKG2D-positive lymphocytes can be modulated by growth factors, tumor cells and tumor cell-derived exosomes In order to examine the effect of tumor exosomes on the number of NKG2D-positive lymphocytes we estab-
Fig. 2. Baseline studies of NKG2D expression by PBL of three healthy volunteers. Fresh PBL were incubated for 24 h T47d tumor cells (at 5:1 or 1:1 PBL: T47d ratio), in the presence or absence of IL-2 or IL-15. The absolute proportion of NKG2D-positive cells, for CD3+/CD8+, CD3+/CD8 and CD3 lymphocyte populations, is shown. The data demonstrate comparable levels of NKG2D-positive cells from each donor, and a tumor-cell-mediated decrease in this positivity, which overcomes the slight elevation due to cytokine treatment.
A. Clayton, Z. Tabi / Blood Cells, Molecules, and Diseases 34 (2005) 206 – 213
209
Fig. 3. Fresh PBL were pre-incubated for 48h with IL-15 (10 ng/ml), in order to increase NKG2D expression. T47d cells or T47d exosomes were subsequently added at the indicated doses, and following 48 h incubation, the proportion of NKG2D-positive lymphocytes was examined. Tumor cells and exosomes gave a dose-dependent reduction in the proportion of NKG2D lymphocytes, in each sub-population. Graph shows the proportion of NKG2D-positive cells as a percentage of IL-15-treated PBL (set at 100%, dotted line), mean + SD, of duplicates.
lished a 3-color flow cytometric protocol that allowed several lymphocyte sub-populations to be examined in the same sample. Fresh peripheral blood leukocytes (PBL)
were treated with various conditions, and live cells were harvested and stained with anti CD3 (FITC), anti CD8 (Cy-5) and anti NKG2D (PE). The gating scheme (shown
Fig. 4. Flow cytometric analysis of a panel of cancer cell lines, stained with irrelevant isotype-control antibody (filled histogram), or with antibody as indicated (unfilled histogram). The adhered cells (T47d and mesothelioma cell line 1 and 15) were harvested by washing the monolayer in 2 mM NaEDTA in PBS at 4-C for 5 min to avoid the use of protease. Live cells were washed in PBS, and labeled on ice with primary antibody (for 40 min), washed in PBS, followed staining with goat anti-mouse FITC (Fab). The panel demonstrates variable NKG2D ligand expression patterns across the different cell lines.
210
A. Clayton, Z. Tabi / Blood Cells, Molecules, and Diseases 34 (2005) 206 – 213
in Fig. 1) allowed the selection and analysis of three distinct lymphocyte subsets (CD3+/CD8+, CD3+/CD8 and CD3 cells as indicated). The proportion of NKG2Dpositive cells within each subset was calculated from the number of events in each quadrant. A representative dot plot for an NKG2D low (a) vs. high (b) population is shown. An example of typical % NKG2D-positive cells in each lymphocyte population is shown in Fig. 2. In this case, PBL from three healthy individuals was co-cultured for 24 h with breast cancer (T47d) cells, in the presence of IL-2 or IL-15 as indicated. Although the cytokines alone gave a small elevation in NKG2D positivity, the presence of tumor cells, particularly at a ratio of 1:1, were effective at reducing the proportion of NKG2D-
positive cells, even in the presence of cytokines. The general trend of change was seen in all three donors, and the absolute % values for each sub-population were comparable. We next examined whether exosomes purified from these tumor cells had the capacity to alter the proportion of NKG2D-positive PBL. In order to achieve a high level of initial NKG2D positivity, the PBL were treated with IL-15 for 48 h prior to the addition of tumor cells or exosomes. Following further 48h incubation, PBL were analyzed for NKG2D expression (Fig. 3). Tumor cells were effective at significantly reducing the proportion of NKG2D-positive lymphocytes. In addition, a dose-dependent inhibition was also observed following the addition of exosomes, and this was significant at the higher dose used (equal to 1 Ag
Fig. 5. Fresh PBL were incubated for 24 h with nothing, with TGFh1 (50 ng/ml), with T47d exosomes or with T47d cells at the indicated doses. Some exosomes were pre-incubated for 20 min with 5 Ag/ml monoclonal antibody against MICA or MICB. As a positive control for antibody blockade, some T47d cells were incubated with a cocktail of antibodies against NKG2D ligands (all at 5 Ag/ml) as indicated. The proportion of CD3+/CD8+ T cells that were positive for NKG2D was measured by flow cytometry as described. Graph shows levels relative to untreated controls dotted line (mean + SD of duplicate samples) (a). Exosomes from mesothelioma cells (b), the EVB-immortalized cell IB4 (c) or Jurkat cells were also tested for their capacity to inhibit NKG2D-positive CD8+ T cells in the same manner except for a longer (48 h) incubation time prior to analysis. Exosomes captured onto latex beads were stained with antibodies and analyzed by flow cytometry, gating on the single-bead population. Filled histograms represent staining with isotype control antibodies, while unfilled histograms show staining by antibodies as indicated (e). These data demonstrate a correlation between exosomal NKG2D ligand expression and a reduction in the proportion of NKG2D-positive CD8+ T cells.
A. Clayton, Z. Tabi / Blood Cells, Molecules, and Diseases 34 (2005) 206 – 213
exosomes per 50,000 PBL). The breast cancer cell, T47d and their exosomes are therefore potent inhibitors of lymphocyte NKG2D expression, even in the presence of the stimulating effects of IL-15. Expression of NKG2D ligands by tumor cells We next examined the possible expression of NKG2D ligands by T47d (and other) cancer cells by flow cytometry, using a panel of antibodies as indicated (Fig. 4). Adherent cells were harvested without trypsin, as this was found to cleave NKG2D ligands from the cell surface, giving false-negative results. T47d cells were negative for MICA, ULBP-1 and -3. There was very week staining for ULBP-2 and convincing positivity for MICB. In contrast, two mesothelioma cell lines, expressed a different repertoire of NKG2D ligands, with positive staining for MICA, MICB and ULBP2. The Tlymphoma line (Jurkat) was strongly positive for all ligands, except for ULBP-3, while EBV-transformed B cells were negative for all ligands. Tumor cells therefore express different patterns of NKG2D ligands. The correlation between NKG2D ligand expression and the capacity of these cells (or their exosomes) to modulate the expression of NKG2D by CD8+ T cells was examined (Fig. 5). T47d exosomes were effective at reducing NKG2D-positive cells (CD3+/CD8+ shown), and this was not affected by pre-incubating exosomes with anti-MICA antibody. In contrast, pre-incubation with anti-MICB was effective at preventing NKG2D suppression (Fig. 5a). Exosomes from other cellular sources were also tested for NKG2D-modulating ability. Two mesothelioma cell line exosomes were arguably more effective at NKG2D suppression than T47d exosomes (Fig. 5b), and this may reflect the wider repertoire of NKG2D ligands expressed by the mesothelioma cells. In contrast, exosomes of the EBV-immortalized B cell line, IB4, showed no inhibitory effect on NKG2D levels. Again this likely reflects the relative paucity of NKG2D ligands expressed by such cells (Fig. 5c). Because Jurkat cells expressed high levels of NKG2D ligands (Fig. 4), it was expected that such exosomes would be efficient at suppressing NKG2D expression. However, even at doses as high as 1 Ag of exosomes per 8000 PBL, there was no inhibition of NKG2D. In contrast Jurkat cells were as effective at NKG2D suppression as a mesothelioma cell line (Fig. 5d). To find a possible explanation for this observation, exosomes were immobilized onto beads, stained with antibodies and analyzed by flow cytometry for NKG2D ligand expression (Fig. 5e). Although Jurkat cells express very high levels of MHC Class I, the exosomes which they produce do not well represent this, yet other exosomal markers such as CD81 are strongly expressed. In addition Jurkat exosomes appeared negative for MICA and MICB, yet the cells were convincingly positive (Fig. 4). In contrast, T47d exosomes expressed a
211
phenotype identical to that of the parent cell, with low MHC class I, no MICA but positive staining for MICB. The findings to date, therefore, are consistent with the hypothesis that certain tumor cell exosomes express NKG2D ligands, which are capable of suppressing NKG2D-expression by lymphocytes. Exosome-mediated inhibition of T cell cytotoxicity We next questioned whether T47d exosomes had any functional effect on CD8+ T cells (i.e., other than downregulation of NKG2D). To do this, in vitro cytotoxicity assays were performed using a CD8+ T cell line-specific against several immunodominant viral epitopes, as an effector-cell population, and testing the capacity of this line to kill peptide-pulsed target cells following preexposure to T47d exosomes. The line was tested in standard chromium release assay for its ability to specifically kill T2 target cells in the absence or presence of peptide cocktail (0.8% and 26%
Fig. 6. The cytotoxic capacity of a CD8+ T cell line was tested in a standard chromium release assay, against T2-target cells that were non-pulsed or pulsed with a synthetic peptide cocktail at an effector: target ratio of 10:1 (a). The same T cell line was left untreated or treated for 48 h with two doses of T47d cells or T47d exosomes, prior to testing cytotoxicity against peptide-pulsed T2 cells (E:T of 10:1) (graph represents specific cytotoxicity as percentage of untreated controls (set at 100%; mean + SD of triplicate samples, *P < 0.05)). The data indicate that tumor exosomes (of mismatched MHC haplotype) can inhibit the cytotoxic capacity of T cells specific for viral epitopes.
212
A. Clayton, Z. Tabi / Blood Cells, Molecules, and Diseases 34 (2005) 206 – 213
specific lysis, respectively) (Fig. 6a). Pre-treatment of this effector cell population for 48 h with either T47d exosomes or with T47d Cells resulted in a drastic reduction in specific cytotoxicity, around 90% inhibition when cultured with T47d cells, and significant inhibition of killing following exposure to T47d exosomes (~65% inhibition, low dose and ~90% inhibition at high exosome dose) (Fig. 6b). These data demonstrate that tumor exosomes can significantly inhibit the cytotoxic function of CD8+ T cells.
Discussion In this preliminary report, we describe for the first time a phenotypic alteration which occurs in several lymphocyte subsets in response to tumor exosomes, i.e., decreased NKG2D expression. Although the precise mechanism by which this alteration occurs requires a more thorough analysis, to date our data point to a mechanism involving exosomally expressed ligands of NKG2D, as the effect is abolished in the presence of blocking antibodies, and is consistent with the ligand-repertoire of the tumor cell exosome. The physiological importance of this observation is not yet clear, but parallels can be drawn with the shedding of plasma membrane NKG2D ligands by certain tumor cells as a strategy for evading NK cell and T cell mediated lysis. This phenomenon results not only in a loss of tumor cell surface markers that would otherwise initiate an immune response, but also act to sequester NKG2Dreceptors (either locally or systemically), leading to suboptimal or dysfunctional cellular immune responses [15] and/or impaired immune surveillance [11]. The relative potency of exosome-borne NKG2D ligands compared to their soluble counterparts (such as sMICA and sMICB) will be a key indicator of physiological importance that needs to be carefully scrutinized. It may be that the inclusion of such ligands into tumor exosomes is merely coincidental, thus representing one of many ‘‘housekeeping’’ proteins, with no particular exosomefunction per say. Alternatively, however, it may be that that tumor cells gain further advantage in eliminating NKG2D ligands by their inclusion into exosomes. Such exosomes may deliver a combination of signals, perhaps multiple NKG2D ligands, in addition to other signals through general or cell-targeted adhesion molecules for example. This aspect is currently being investigated in our laboratory. One concerning observation which has arisen from these studies is the apparent ability of tumor exosomes to significantly inhibit the cytotoxic ability of CD8+ T cells. Pre-incubating virus-specific T cells with T47d exosomes resulted in a significant loss of cytotoxic function. The underlying mechanisms are not currently known but may involve disruption of the cytotoxic machinery by signals emanating from NKG2D
ligation, or involve a lack of appropriate co-stimulatory signals, or indeed involve NKG2D independent mechanisms. Studies are currently underway to address these questions.
Acknowledgments This paper is based on a presentation at a Focused Workshop entitled Exosomes: Biological Significance, sponsored by The Leukemia and Lymphoma Society, in Montreal, Canada, May 20 – 21, 2005. The work was supported by the Leukaemia Research Fund and Cancer Research Wales.
References [1] N. Jan Chalupny, C. Sutherland, W. Lawrence, A. Rein-Weston, D. Cosman, ULBP4 is a novel ligand for human NKG2D, Biochem. Biophys. Res. Commun. 305 (2003 (May 23)) 129 – 135. [2] M. Lopez-Botet, A. Angulo, M. Guma, Natural killer cell receptors for major histocompatibility complex class I and related molecules in cytomegalovirus infection, Tissue Antigens 63 (2004) 195 – 203. [3] W. Cao, W. He, UL16 binding proteins, Immunobiology 209 (2004) 283 – 290. [4] S.A. Welte, C. Sinzger, S.Z. Lutz, H. Singh-Jasuja, K.L. Sampaio, U. Eknigk, H.G. Rammensee, A. Steinle, Selective intracellular retention of virally induced NKG2D ligands by the human cytomegalovirus UL16 glycoprotein, Eur. J. Immunol. 33 (2003) 194 – 203. [5] J.R. Conejo-Garcia, F. Benencia, M.C. Courreges, P.A. Gimotty, E. Khang, R.J. Buckanovich, K.A. Frauwirth, L. Zhang, D. Katsaros, C.B. Thompson, B. Levine, G. Coukos, Ovarian carcinoma expresses the NKG2D ligand Letal and promotes the survival and expansion of CD28 antitumor T cells, Cancer Res. 64 (2004) 2175 – 2182. [6] L. Bacon, R.A. Eagle, M. Meyer, N. Easom, N.T. Young, J. Trowsdale, Two human ULBP/RAET1 molecules with transmembrane regions are ligands for NKG2D, J. Immunol. 173 (2004) 1078 – 1084. [7] A. Diefenbach, A. Jamieson, S. Liu, N. Shastri, D. Raulet, Ligands for the murine NKG2D receptor: expression by tumor cells and activation of NK cells and macrophages, Nat. Immunol. 1 (2000 (Aug)) 119 – 126. [8] D. Cosman, J. Mu¨llberg, C. Sutherland, W. Chin, R. Armitage, W. Fanslow, M. Kubin, N. Chalupny, ULBPs, novel MHC class I-related molecules, bind to CMV glycoprotein UL16 and stimulate NK cytotoxicity through the NKG2D receptor, Immunity 14 (2001) 123 – 133. [9] A. Cerwenka, A. Bakker, T. McClanahan, J. Wagner, J. Wu, J. Phillips, L. Lanier, Retinoic acid early inducible genes define a ligand family for the activating NKG2D receptor in mice, Immunity 12 (2000 (Jun)) 721 – 727. [10] H. Salih, H. Rammensee, A. Steinle, Cutting edge: down-regulation of MICA on human tumors by proteolytic shedding, J. Immunol. 169 (2002) 4098 – 4102. [11] H.R. Salih, H. Antropius, F. Gieseke, S.Z. Lutz, L. Kanz, H.G. Rammensee, A. Steinle, Functional expression and release of ligands for the activating immunoreceptor NKG2D in leukemia, Blood 102 (2003) 1389 – 1396. [12] L. Raffaghello, I. Prigione, I. Airoldi, M. Camoriano, I. Levreri, C. Gambini, D. Pende, A. Steinle, S. Ferrone, V. Pistoia, Downregulation and/or release of NKG2D ligands as immune evasion strategy of human neuroblastoma, Neoplasia 6 (2004) 558 – 568. [13] F. Andre, N.E. Schartz, M. Movassagh, C. Flament, P. Pautier, P.
A. Clayton, Z. Tabi / Blood Cells, Molecules, and Diseases 34 (2005) 206 – 213 Morice, C. Pomel, C. Lhomme, B. Escudier, T. Le Chevalier, T. Tursz, S. Amigorena, G. Raposo, E. Angevin, L. Zitvogel, Malignant effusions and immunogenic tumour-derived exosomes, Lancet 360 (2002) 295 – 305. [14] H. Lamparski, A. Metha-Damani, J. Yao, S. Patel, D. Hsu, C. Ruegg, J. Le Pecq, Production and characterization of clinical grade
213
exosomes derived from dendritic cells, J. Immunol. Methods 270 (2002) 211 – 226. [15] L. Raffaghello, I. Prigione, I. Airoldi, M. Camoriano, I. Levreri, C. Gambini, D. Pende, A. Steinle, S. Ferrone, V. Pistoia, Downregulation and/or release of NKG2D ligands as immune evasion strategy of human neuroblastoma, Neoplasia 6 (2004) 558 – 568.
Exosomes and the MICA-NKG2D system in cancer Aled ClaytonT, Zsuzsanna Tabi Department of Clinical Oncology and Palliative Medicine, Wales College of Medicine, Cardiff University, Velindre Cancer Centre, Whitchurch, Cardiff CF14 2TL, UK Submitted 9 March 2005; revised 15 March 2005 (Communicated by M. Lichtman, M.D., 15 March 2005)
Abstract Exosomes are nanometer sized vesicles, secreted by a diverse range of cell types, whose biological functions remain ambiguous. Several groups have demonstrated the potential of manipulating exosomes for activating cellular immune responses. The possibility that exosomes may inhibit immunological responses, however, has not been widely addressed. We have investigated if exosomes produced by tumor cells can inhibit immunological functions, through modulating expression of the NKG2D receptor by effector cells. Incubating tumor exosomes with fresh peripheral blood leukocytes resulted in a marked reduction in the proportion of NKG2D-positive CD3+CD8+ Cells, and CD3 cells by 48 h. This effect was dose dependent and was shown with exosomes from different tumor cells including breast cancer and mesothelioma. Analysis of tumor exosome-phenotype revealed positive expression of several NKG2D ligands, and antibody blocking experiments revealed the importance of such ligands in driving the reduction in the proportion of NKG2D-positive effector cells. The functional importance of the decrease in NKG2D-positive cells was addressed in vitro cytotoxicity assays. For example a CD8+ T cell line pre-incubated with tumor exosomes had significant decreased capacity to kill peptide-pulsed T2 target cells. These data highlight a role for tumor exosomes bearing NKG2D ligands as a mechanism contributing to cancer immune evasion. D 2005 Elsevier Inc. All rights reserved. Keywords: Exosome; MICA; NKG2D; Cancer
Introduction NKG2D is a homodimeric C-type lectin receptor, expressed by a range of immune cells, including NK, NKT, CD8 ah T cells, gy T cells and macrophages. Ligands for NKG2D comprise the polymorphic MHC class I related, stress-inducible proteins (including MICA and MICB) that exhibit highly restricted expression in vivo, and are generally only found upon infection and on some tumor cells. Additional ligands, discovered more recently, include the ULBP-1, -2, -3 and -4 [1] proteins, so named because of their binding to the cytomegalovirus encoded protein UL16, important for CMV immune evasion through preventing surface
T Corresponding author. E-mail address: [email protected] (A. Clayton). 1079-9796/$ - see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.bcmd.2005.03.003
expression of NKG2D ligands [2 –4]. It is likely that NKG2D is capable of binding yet more ligands, such as Letal [5] and others [6]. The true breadth of ligands for NKG2D is not yet fully appreciated. The activation of NK cells through NKG2D can overcome inhibitory signaling from self-recognition, and therefore directly trigger NK cytotoxic function, while NKG2D provides a costimulatory signal (in addition to signaling through TCR) in T cells. Thus tumor cells (and other cells under stress) expressing ligands for NKG2D are highly susceptible to killing by NK and T cells [7 –9]. Importantly, however, there is mounting evidence demonstrating that tumor cells may subvert this mode of targeting through shedding such ligands from the cell surface, through the activity of matrix metalloproteinases for example [10]. Such data are strengthened by observations that cancer patients have high levels of a
A. Clayton, Z. Tabi / Blood Cells, Molecules, and Diseases 34 (2005) 206 – 213
soluble form of MICA (sMICA) (and in some cases sMICB) in their serum, where such molecules are not detectable in healthy donor sera [11]. The effect of shed ligands results not only in lower levels of cell surface ligand expression, but there is also evidence that soluble ligands downregulate the expression of T cell NKG2D, and decrease the cytotoxicity of NK cells towards MICApositive targets [12]. In this report we have investigated the possibility that tumor cell-derived exosomes express ligands for NKG2D and are capable of modulating effector cell functions. Our data demonstrate that NKG2D is downregulated following exposure to tumor-derived exosomes, and that effector cytotoxic functions are impaired as a result. The production of NKG2D-ligand-bearing exosomes, therefore, may be a novel mechanism for tumor cell immune evasion, and further demonstrates the complexity of interactions possible between exosomes and the immune system.
207
Materials and methods Cell lines The breast carcinoma line T47d was a gift from Prof. R Nicholson, Tenovus, Cardiff University. Two mesothelioma cell lines (named patients 1 and 15) were generated in house from explants of pleural malignant mesothelioma. Jurkat and IB4 cells were from the MRC cooperative, Cardiff University, and were maintained in serum-free AIM-v medium. Other cells were maintained in RPMI1640 (with l-glutamine, pen/strep and HEPES), and 10% FBS (BRL Life Technologies). FBS was depleted of bovine exosomes by pelleting at 100,000 g for 2 h. Exosome purification Cell conditioned medium (48 – 72 h) was cleared of cells and particulates by serial centrifugation (100 g/5 min,
Fig. 1. A diagrammatic explanation of a 3-color flow cytometric method used for analyzing the proportion of NKG2D-positive cells within lymphocyte subsets. The selection of lymphocytes is performed using the FSC/SSC dot plot (region-1, R1). Lymphocytes which stain positive (R2) or negative (R3) for CD3 are identified using the FL-1/FCS plot. Finally, the proportion of NKG2D-positive T-lymphocytes (R1 and R2) are examined using an FL-3(CD8)/FL-2(NKG2D) plot, allowing the CD8+ and CD8 populations to be distinguished. Finally, analysis of non-T cells (R1 and R3) are examined using an FL-1/FL-2 plot. An example of NKG2D-low (a) and an NKG2D-high (b) population is shown.
208
A. Clayton, Z. Tabi / Blood Cells, Molecules, and Diseases 34 (2005) 206 – 213
400 g/5 min, 2000 g/10 min at 4-C) and filtered (0.2 Am). Supernatant was underlayed with 30% Sucrose/D2O (as previously described [13,14] prior to ultracentrifugation at 100,000 g for 3 h. The collected sucrose cushion was diluted in excess PBS, and exosomes pelleted by a further ultracentrifugation step. Exosome quantity was determined by BCA protein assay (Pierce). Flow cytometry and antibodies Analysis of NKG2D expression by peripheral blood lymphocytes was performed by three-color flow cytometry (BD FACScan), using monoclonal antibodies against CD3-FITC (Dako), CD8-Cy5 (BD) and NKG2D-PE (R&D Systems). Analysis was performed on lymphocytes (gating with FSC/SSC) that were CD3 positive or CD3 negative and are expressed as % NKG2D-positive cells (schematically represented in Fig. 1). Antibodies for the known ligands of NKG2D, MICA, MICB, ULBP-1, -2 and -3 together with IgG2a and IgG2b isotype controls were from R&D Systems. Flow-cytometric analysis of ligand expression was performed on live cells (not treated with trypsin—because of the protease-sensitive
nature of the ligands) and on exosome-coated aldehyde– sulphate latex beads (from Interfacial Dynamics). CD8+ T cell cytotoxicity assays A CD8+ T cell line was generated by pulsing autologous dendritic cells with a cocktail of synthetic peptides, with sequences derived from published immunodominant viral epitopes (including influenza, Epstein – Barr virus and cytomegalovirus). Standard chromium release assay was performed with T2 target cells, loaded with Cr51 and pulsed with peptide cocktail (at a final concentration of 10 Ag/ml), at an effector: target ratio of 10:1.
Results The proportion of NKG2D-positive lymphocytes can be modulated by growth factors, tumor cells and tumor cell-derived exosomes In order to examine the effect of tumor exosomes on the number of NKG2D-positive lymphocytes we estab-
Fig. 2. Baseline studies of NKG2D expression by PBL of three healthy volunteers. Fresh PBL were incubated for 24 h T47d tumor cells (at 5:1 or 1:1 PBL: T47d ratio), in the presence or absence of IL-2 or IL-15. The absolute proportion of NKG2D-positive cells, for CD3+/CD8+, CD3+/CD8 and CD3 lymphocyte populations, is shown. The data demonstrate comparable levels of NKG2D-positive cells from each donor, and a tumor-cell-mediated decrease in this positivity, which overcomes the slight elevation due to cytokine treatment.
A. Clayton, Z. Tabi / Blood Cells, Molecules, and Diseases 34 (2005) 206 – 213
209
Fig. 3. Fresh PBL were pre-incubated for 48h with IL-15 (10 ng/ml), in order to increase NKG2D expression. T47d cells or T47d exosomes were subsequently added at the indicated doses, and following 48 h incubation, the proportion of NKG2D-positive lymphocytes was examined. Tumor cells and exosomes gave a dose-dependent reduction in the proportion of NKG2D lymphocytes, in each sub-population. Graph shows the proportion of NKG2D-positive cells as a percentage of IL-15-treated PBL (set at 100%, dotted line), mean + SD, of duplicates.
lished a 3-color flow cytometric protocol that allowed several lymphocyte sub-populations to be examined in the same sample. Fresh peripheral blood leukocytes (PBL)
were treated with various conditions, and live cells were harvested and stained with anti CD3 (FITC), anti CD8 (Cy-5) and anti NKG2D (PE). The gating scheme (shown
Fig. 4. Flow cytometric analysis of a panel of cancer cell lines, stained with irrelevant isotype-control antibody (filled histogram), or with antibody as indicated (unfilled histogram). The adhered cells (T47d and mesothelioma cell line 1 and 15) were harvested by washing the monolayer in 2 mM NaEDTA in PBS at 4-C for 5 min to avoid the use of protease. Live cells were washed in PBS, and labeled on ice with primary antibody (for 40 min), washed in PBS, followed staining with goat anti-mouse FITC (Fab). The panel demonstrates variable NKG2D ligand expression patterns across the different cell lines.
210
A. Clayton, Z. Tabi / Blood Cells, Molecules, and Diseases 34 (2005) 206 – 213
in Fig. 1) allowed the selection and analysis of three distinct lymphocyte subsets (CD3+/CD8+, CD3+/CD8 and CD3 cells as indicated). The proportion of NKG2Dpositive cells within each subset was calculated from the number of events in each quadrant. A representative dot plot for an NKG2D low (a) vs. high (b) population is shown. An example of typical % NKG2D-positive cells in each lymphocyte population is shown in Fig. 2. In this case, PBL from three healthy individuals was co-cultured for 24 h with breast cancer (T47d) cells, in the presence of IL-2 or IL-15 as indicated. Although the cytokines alone gave a small elevation in NKG2D positivity, the presence of tumor cells, particularly at a ratio of 1:1, were effective at reducing the proportion of NKG2D-
positive cells, even in the presence of cytokines. The general trend of change was seen in all three donors, and the absolute % values for each sub-population were comparable. We next examined whether exosomes purified from these tumor cells had the capacity to alter the proportion of NKG2D-positive PBL. In order to achieve a high level of initial NKG2D positivity, the PBL were treated with IL-15 for 48 h prior to the addition of tumor cells or exosomes. Following further 48h incubation, PBL were analyzed for NKG2D expression (Fig. 3). Tumor cells were effective at significantly reducing the proportion of NKG2D-positive lymphocytes. In addition, a dose-dependent inhibition was also observed following the addition of exosomes, and this was significant at the higher dose used (equal to 1 Ag
Fig. 5. Fresh PBL were incubated for 24 h with nothing, with TGFh1 (50 ng/ml), with T47d exosomes or with T47d cells at the indicated doses. Some exosomes were pre-incubated for 20 min with 5 Ag/ml monoclonal antibody against MICA or MICB. As a positive control for antibody blockade, some T47d cells were incubated with a cocktail of antibodies against NKG2D ligands (all at 5 Ag/ml) as indicated. The proportion of CD3+/CD8+ T cells that were positive for NKG2D was measured by flow cytometry as described. Graph shows levels relative to untreated controls dotted line (mean + SD of duplicate samples) (a). Exosomes from mesothelioma cells (b), the EVB-immortalized cell IB4 (c) or Jurkat cells were also tested for their capacity to inhibit NKG2D-positive CD8+ T cells in the same manner except for a longer (48 h) incubation time prior to analysis. Exosomes captured onto latex beads were stained with antibodies and analyzed by flow cytometry, gating on the single-bead population. Filled histograms represent staining with isotype control antibodies, while unfilled histograms show staining by antibodies as indicated (e). These data demonstrate a correlation between exosomal NKG2D ligand expression and a reduction in the proportion of NKG2D-positive CD8+ T cells.
A. Clayton, Z. Tabi / Blood Cells, Molecules, and Diseases 34 (2005) 206 – 213
exosomes per 50,000 PBL). The breast cancer cell, T47d and their exosomes are therefore potent inhibitors of lymphocyte NKG2D expression, even in the presence of the stimulating effects of IL-15. Expression of NKG2D ligands by tumor cells We next examined the possible expression of NKG2D ligands by T47d (and other) cancer cells by flow cytometry, using a panel of antibodies as indicated (Fig. 4). Adherent cells were harvested without trypsin, as this was found to cleave NKG2D ligands from the cell surface, giving false-negative results. T47d cells were negative for MICA, ULBP-1 and -3. There was very week staining for ULBP-2 and convincing positivity for MICB. In contrast, two mesothelioma cell lines, expressed a different repertoire of NKG2D ligands, with positive staining for MICA, MICB and ULBP2. The Tlymphoma line (Jurkat) was strongly positive for all ligands, except for ULBP-3, while EBV-transformed B cells were negative for all ligands. Tumor cells therefore express different patterns of NKG2D ligands. The correlation between NKG2D ligand expression and the capacity of these cells (or their exosomes) to modulate the expression of NKG2D by CD8+ T cells was examined (Fig. 5). T47d exosomes were effective at reducing NKG2D-positive cells (CD3+/CD8+ shown), and this was not affected by pre-incubating exosomes with anti-MICA antibody. In contrast, pre-incubation with anti-MICB was effective at preventing NKG2D suppression (Fig. 5a). Exosomes from other cellular sources were also tested for NKG2D-modulating ability. Two mesothelioma cell line exosomes were arguably more effective at NKG2D suppression than T47d exosomes (Fig. 5b), and this may reflect the wider repertoire of NKG2D ligands expressed by the mesothelioma cells. In contrast, exosomes of the EBV-immortalized B cell line, IB4, showed no inhibitory effect on NKG2D levels. Again this likely reflects the relative paucity of NKG2D ligands expressed by such cells (Fig. 5c). Because Jurkat cells expressed high levels of NKG2D ligands (Fig. 4), it was expected that such exosomes would be efficient at suppressing NKG2D expression. However, even at doses as high as 1 Ag of exosomes per 8000 PBL, there was no inhibition of NKG2D. In contrast Jurkat cells were as effective at NKG2D suppression as a mesothelioma cell line (Fig. 5d). To find a possible explanation for this observation, exosomes were immobilized onto beads, stained with antibodies and analyzed by flow cytometry for NKG2D ligand expression (Fig. 5e). Although Jurkat cells express very high levels of MHC Class I, the exosomes which they produce do not well represent this, yet other exosomal markers such as CD81 are strongly expressed. In addition Jurkat exosomes appeared negative for MICA and MICB, yet the cells were convincingly positive (Fig. 4). In contrast, T47d exosomes expressed a
211
phenotype identical to that of the parent cell, with low MHC class I, no MICA but positive staining for MICB. The findings to date, therefore, are consistent with the hypothesis that certain tumor cell exosomes express NKG2D ligands, which are capable of suppressing NKG2D-expression by lymphocytes. Exosome-mediated inhibition of T cell cytotoxicity We next questioned whether T47d exosomes had any functional effect on CD8+ T cells (i.e., other than downregulation of NKG2D). To do this, in vitro cytotoxicity assays were performed using a CD8+ T cell line-specific against several immunodominant viral epitopes, as an effector-cell population, and testing the capacity of this line to kill peptide-pulsed target cells following preexposure to T47d exosomes. The line was tested in standard chromium release assay for its ability to specifically kill T2 target cells in the absence or presence of peptide cocktail (0.8% and 26%
Fig. 6. The cytotoxic capacity of a CD8+ T cell line was tested in a standard chromium release assay, against T2-target cells that were non-pulsed or pulsed with a synthetic peptide cocktail at an effector: target ratio of 10:1 (a). The same T cell line was left untreated or treated for 48 h with two doses of T47d cells or T47d exosomes, prior to testing cytotoxicity against peptide-pulsed T2 cells (E:T of 10:1) (graph represents specific cytotoxicity as percentage of untreated controls (set at 100%; mean + SD of triplicate samples, *P < 0.05)). The data indicate that tumor exosomes (of mismatched MHC haplotype) can inhibit the cytotoxic capacity of T cells specific for viral epitopes.
212
A. Clayton, Z. Tabi / Blood Cells, Molecules, and Diseases 34 (2005) 206 – 213
specific lysis, respectively) (Fig. 6a). Pre-treatment of this effector cell population for 48 h with either T47d exosomes or with T47d Cells resulted in a drastic reduction in specific cytotoxicity, around 90% inhibition when cultured with T47d cells, and significant inhibition of killing following exposure to T47d exosomes (~65% inhibition, low dose and ~90% inhibition at high exosome dose) (Fig. 6b). These data demonstrate that tumor exosomes can significantly inhibit the cytotoxic function of CD8+ T cells.
Discussion In this preliminary report, we describe for the first time a phenotypic alteration which occurs in several lymphocyte subsets in response to tumor exosomes, i.e., decreased NKG2D expression. Although the precise mechanism by which this alteration occurs requires a more thorough analysis, to date our data point to a mechanism involving exosomally expressed ligands of NKG2D, as the effect is abolished in the presence of blocking antibodies, and is consistent with the ligand-repertoire of the tumor cell exosome. The physiological importance of this observation is not yet clear, but parallels can be drawn with the shedding of plasma membrane NKG2D ligands by certain tumor cells as a strategy for evading NK cell and T cell mediated lysis. This phenomenon results not only in a loss of tumor cell surface markers that would otherwise initiate an immune response, but also act to sequester NKG2Dreceptors (either locally or systemically), leading to suboptimal or dysfunctional cellular immune responses [15] and/or impaired immune surveillance [11]. The relative potency of exosome-borne NKG2D ligands compared to their soluble counterparts (such as sMICA and sMICB) will be a key indicator of physiological importance that needs to be carefully scrutinized. It may be that the inclusion of such ligands into tumor exosomes is merely coincidental, thus representing one of many ‘‘housekeeping’’ proteins, with no particular exosomefunction per say. Alternatively, however, it may be that that tumor cells gain further advantage in eliminating NKG2D ligands by their inclusion into exosomes. Such exosomes may deliver a combination of signals, perhaps multiple NKG2D ligands, in addition to other signals through general or cell-targeted adhesion molecules for example. This aspect is currently being investigated in our laboratory. One concerning observation which has arisen from these studies is the apparent ability of tumor exosomes to significantly inhibit the cytotoxic ability of CD8+ T cells. Pre-incubating virus-specific T cells with T47d exosomes resulted in a significant loss of cytotoxic function. The underlying mechanisms are not currently known but may involve disruption of the cytotoxic machinery by signals emanating from NKG2D
ligation, or involve a lack of appropriate co-stimulatory signals, or indeed involve NKG2D independent mechanisms. Studies are currently underway to address these questions.
Acknowledgments This paper is based on a presentation at a Focused Workshop entitled Exosomes: Biological Significance, sponsored by The Leukemia and Lymphoma Society, in Montreal, Canada, May 20 – 21, 2005. The work was supported by the Leukaemia Research Fund and Cancer Research Wales.
References [1] N. Jan Chalupny, C. Sutherland, W. Lawrence, A. Rein-Weston, D. Cosman, ULBP4 is a novel ligand for human NKG2D, Biochem. Biophys. Res. Commun. 305 (2003 (May 23)) 129 – 135. [2] M. Lopez-Botet, A. Angulo, M. Guma, Natural killer cell receptors for major histocompatibility complex class I and related molecules in cytomegalovirus infection, Tissue Antigens 63 (2004) 195 – 203. [3] W. Cao, W. He, UL16 binding proteins, Immunobiology 209 (2004) 283 – 290. [4] S.A. Welte, C. Sinzger, S.Z. Lutz, H. Singh-Jasuja, K.L. Sampaio, U. Eknigk, H.G. Rammensee, A. Steinle, Selective intracellular retention of virally induced NKG2D ligands by the human cytomegalovirus UL16 glycoprotein, Eur. J. Immunol. 33 (2003) 194 – 203. [5] J.R. Conejo-Garcia, F. Benencia, M.C. Courreges, P.A. Gimotty, E. Khang, R.J. Buckanovich, K.A. Frauwirth, L. Zhang, D. Katsaros, C.B. Thompson, B. Levine, G. Coukos, Ovarian carcinoma expresses the NKG2D ligand Letal and promotes the survival and expansion of CD28 antitumor T cells, Cancer Res. 64 (2004) 2175 – 2182. [6] L. Bacon, R.A. Eagle, M. Meyer, N. Easom, N.T. Young, J. Trowsdale, Two human ULBP/RAET1 molecules with transmembrane regions are ligands for NKG2D, J. Immunol. 173 (2004) 1078 – 1084. [7] A. Diefenbach, A. Jamieson, S. Liu, N. Shastri, D. Raulet, Ligands for the murine NKG2D receptor: expression by tumor cells and activation of NK cells and macrophages, Nat. Immunol. 1 (2000 (Aug)) 119 – 126. [8] D. Cosman, J. Mu¨llberg, C. Sutherland, W. Chin, R. Armitage, W. Fanslow, M. Kubin, N. Chalupny, ULBPs, novel MHC class I-related molecules, bind to CMV glycoprotein UL16 and stimulate NK cytotoxicity through the NKG2D receptor, Immunity 14 (2001) 123 – 133. [9] A. Cerwenka, A. Bakker, T. McClanahan, J. Wagner, J. Wu, J. Phillips, L. Lanier, Retinoic acid early inducible genes define a ligand family for the activating NKG2D receptor in mice, Immunity 12 (2000 (Jun)) 721 – 727. [10] H. Salih, H. Rammensee, A. Steinle, Cutting edge: down-regulation of MICA on human tumors by proteolytic shedding, J. Immunol. 169 (2002) 4098 – 4102. [11] H.R. Salih, H. Antropius, F. Gieseke, S.Z. Lutz, L. Kanz, H.G. Rammensee, A. Steinle, Functional expression and release of ligands for the activating immunoreceptor NKG2D in leukemia, Blood 102 (2003) 1389 – 1396. [12] L. Raffaghello, I. Prigione, I. Airoldi, M. Camoriano, I. Levreri, C. Gambini, D. Pende, A. Steinle, S. Ferrone, V. Pistoia, Downregulation and/or release of NKG2D ligands as immune evasion strategy of human neuroblastoma, Neoplasia 6 (2004) 558 – 568. [13] F. Andre, N.E. Schartz, M. Movassagh, C. Flament, P. Pautier, P.
A. Clayton, Z. Tabi / Blood Cells, Molecules, and Diseases 34 (2005) 206 – 213 Morice, C. Pomel, C. Lhomme, B. Escudier, T. Le Chevalier, T. Tursz, S. Amigorena, G. Raposo, E. Angevin, L. Zitvogel, Malignant effusions and immunogenic tumour-derived exosomes, Lancet 360 (2002) 295 – 305. [14] H. Lamparski, A. Metha-Damani, J. Yao, S. Patel, D. Hsu, C. Ruegg, J. Le Pecq, Production and characterization of clinical grade
213
exosomes derived from dendritic cells, J. Immunol. Methods 270 (2002) 211 – 226. [15] L. Raffaghello, I. Prigione, I. Airoldi, M. Camoriano, I. Levreri, C. Gambini, D. Pende, A. Steinle, S. Ferrone, V. Pistoia, Downregulation and/or release of NKG2D ligands as immune evasion strategy of human neuroblastoma, Neoplasia 6 (2004) 558 – 568.