Role of ADAM10 and ADAM17 in Regulating CD137 Function

Human CD137 (4-1BB), a member of the TNF receptor family, and its ligand CD137L (4-1BBL), are expressed on immune cells and tumor cells. CD137/CD137L interaction mediates bidirectional cellular responses of potential relevance in inflammatory diseases, autoimmunity and oncology. A soluble form of CD137 exists, elevated levels of which have been reported in patients with rheumatoid arthritis and various malignancies. Soluble CD137 (sCD137) is considered to represent a splice variant of CD137. In this report, however, evidence is presented that A Disintegrin and Metalloproteinase (ADAM)10 and potentially also ADAM17 are centrally involved in its generation. Release of sCD137 by transfected cell lines and primary T cells was uniformly inhibitable by ADAM10 inhibition. The shedding function of ADAM10 can be blocked through inhibition of its interaction with surface exposed phosphatidylserine (PS), and this effectively inhibited sCD137 generation. The phospholipid scramblase Anoctamin-6 (ANO6) traffics PS to the outer membrane and thus modifies ADAM10 function. Overexpression of ANO6 increased stimulated shedding, and hyperactive ANO6 led to maximal constitutive shedding of CD137. sCD137 was functionally active and augmented T cell proliferation. Our findings shed new light on the regulation of CD137/CD137L immune responses with potential impact on immunotherapeutic approaches targeting CD137.


Introduction
CD137 (also known as 4-1BB and TNFRSF9) is a type I transmembrane protein of the TNFR family that is widely expressed on cells of both the innate and adaptive immune system. It was originally discovered in screens for inducible genes in activated T cells [1,2]. Numerous studies devoted to its functional characterization on activated T cells and natural killer cells have been published implicating decisive potential of CD137 as a target for immunotherapies [3,4]. Interaction with its ligand CD137L results in upregulation of antiapoptotic molecules, cytokine secretion, proliferation and enhanced effector function [5]. The type II transmembrane protein CD137L is constitutively expressed on several types of APCs such as mature dendritic cells and activated macrophages and transduces signals as well [6,7]. Rather remarkably, CD137L is also co-expressed with its receptor on activated T-lymphocytes [8][9][10].
Expression of CD137 and CD137L has also been reported for different types of cancer including lung cancer, leukemia and lymphoma [14][15][16]. The expression might be induced by activating K-ras mutations which are frequently observed, e.g., in pancreatic or colorectal cancer [17]. The biological significance of CD137 expression in cancer cells is presently unclear.

ADAM10 Is the Major Sheddase of CD137 in HT29 Cells
The human colon adenocarcinoma cell line HT29 was employed for the first analysis of sCD137 release. The cells were transfected with CD137 plasmid and release of sCD137 was assessed in the presence of the broad spectrum-metalloprotease inhibitor marimastat (MM), the preferential ADAM10 inhibitor GI254023X (GI), or GW280264 (GW), an inhibitor of both ADAM10 and ADAM17 ( Figure 1A). All three inhibitors led to a significant reduction in sCD137 in the supernatant. This was the first indication that metalloproteinases could cleave and release CD137 from cells, and that ADAM10 was the major sheddase involved.
ADAM10-and ADAM17-mediated shedding can be differentially activated dependent on stimuli. ADAM10 activity can be triggered by induction of calcium-influx, the membrane-active cationic peptide melittin can upregulate both ADAM10 and ADAM17 function, while the phorbol ester PMA is a classical selective stimulus of ADAM17 [44][45][46]. As shown in Figure 1B, enhanced shedding of CD137 was induced by ionomycin and melittin but not by PMA, indicating that ADAM10 represents the major sheddase that is also responsible for stimulated shedding.
CD137 transfection was controlled in parallel by immunoblot (Supplementary Figure S1A). Total amounts of CD137 can be found in Supplementary Figure S1B,C. reduction in sCD137 in the supernatant. This was the first indication that metalloproteinases could cleave and release CD137 from cells, and that ADAM10 was the major sheddase involved.
ADAM10-and ADAM17-mediated shedding can be differentially activated dependent on stimuli. ADAM10 activity can be triggered by induction of calcium-influx, the membrane-active cationic peptide melittin can upregulate both ADAM10 and ADAM17 function, while the phorbol ester PMA is a classical selective stimulus of ADAM17 [44][45][46]. As shown in Figure 1B, enhanced shedding of CD137 was induced by ionomycin and melittin but not by PMA, indicating that ADAM10 represents the major sheddase that is also responsible for stimulated shedding.
CD137 transfection was controlled in parallel by immunoblot (Supplementary Figure  S1A). Total amounts of CD137 can be found in Supplementary Figure S1B,C.

ADAM10 and ADAM17 Can Release CD137 in HEK Cells
Studies addressing the individual roles of ADAM10/ADAM17 are aided by the availability of double-deficient (dKO) HEK293T cells [47]. Wild-type (WT) and dKO-cells were transfected with CD137 and shedding was analyzed in the presence of the different inhibitors. As shown in Figure 2A, release of sCD137 occurred constitutively in WT-cells, and shedding was significantly reduced in the presence of MM, the ADAM17/ADAM10 inhibitor GW, and the ADAM10 inhibitor GI. Constitutive CD137 release was conspicuously lower in double-deficient HEK cells compared with WT-cells and was not further decreased in the presence of inhibitors.
Stimulation experiments provided data that complemented and corroborated the findings. As shown in Figure 2B, shedding in WT-cells was significantly increased upon ionomycin (IO) stimulation and also enhanced upon exposure to melittin, whereas no responses were noted in the double-deficient cells. Finally, it was found that re-transfection of double knock-out cells with ADAM10 restored shedding capacity ( Figure 2C). Retransfection with ADAM17 had a similar effect ( Figure 2C). This was in line with earlier reports

ADAM10 and ADAM17 Can Release CD137 in HEK Cells
Studies addressing the individual roles of ADAM10/ADAM17 are aided by the availability of double-deficient (dKO) HEK293T cells [47]. Wild-type (WT) and dKO-cells were transfected with CD137 and shedding was analyzed in the presence of the different inhibitors. As shown in Figure 2A, release of sCD137 occurred constitutively in WT-cells, and shedding was significantly reduced in the presence of MM, the ADAM17/ADAM10 inhibitor GW, and the ADAM10 inhibitor GI. Constitutive CD137 release was conspicuously lower in double-deficient HEK cells compared with WT-cells and was not further decreased in the presence of inhibitors.
Stimulation experiments provided data that complemented and corroborated the findings. As shown in Figure 2B, shedding in WT-cells was significantly increased upon ionomycin (IO) stimulation and also enhanced upon exposure to melittin, whereas no responses were noted in the double-deficient cells. Finally, it was found that re-transfection of double knock-out cells with ADAM10 restored shedding capacity ( Figure 2C). Retransfection with ADAM17 had a similar effect ( Figure 2C). This was in line with earlier reports that, in the absence of one protease, ADAM10 and ADAM17 can replace each other in cleaving substrates such as TGF-.
CD137 transfection was controlled in parallel by immunoblot (Supplementary Figure S2A). Total amounts of CD137 can be found in Supplementary Figure S2B that, in the absence of one protease, ADAM10 and ADAM17 can replace each other in cleaving substrates such as TGF-.
CD137 transfection was controlled in parallel by immunoblot (Supplementary Figure  S2A). Total amounts of CD137 can be found in Supplementary Figure

Ionomycin-Induced Shedding of CD137 is PS Dependent
Recently, we have presented evidence that the interaction of ADAM10 and ADAM17 with externalized phosphatidylserine (PS) is required for the proteases to exert their sheddase function [34,35,48]. To discern whether stimulated shedding of CD137 was also PSdependent, competition experiments were conducted employing phosphorylserine (OPS, O-phospho-l-serine), the soluble headgroup of PS, or lactadherin (LA), which binds to surface-exposed PS [49,50] and thus blocks its accessibility to other PS-binders.

Ionomycin-Induced Shedding of CD137 Is PS Dependent
Recently, we have presented evidence that the interaction of ADAM10 and ADAM17 with externalized phosphatidylserine (PS) is required for the proteases to exert their sheddase function [34,35,48]. To discern whether stimulated shedding of CD137 was also PS-dependent, competition experiments were conducted employing phosphorylserine (OPS, O-phospho-l-serine), the soluble headgroup of PS, or lactadherin (LA), which binds to surface-exposed PS [49,50] and thus blocks its accessibility to other PS-binders.
HT29 and HEK cells were stimulated in the absence or presence of the inhibitors and released sCD137 was quantified. OPS dose-dependently reduced levels of sCD137 in the supernatants of both cell lines ( Figure 3A,B). In corroboration of this finding, lactadherin used at the entirely non-toxic concentration of 2 µM exerted maximal inhibition [35,36]. Total amounts of CD137 can be found in Supplementary Figure S3A,B. HT29 and HEK cells were stimulated in the absence or presence of the inhibitors and released sCD137 was quantified. OPS dose-dependently reduced levels of sCD137 in the supernatants of both cell lines ( Figure 3A,B). In corroboration of this finding, lactadherin used at the entirely non-toxic concentration of 2 µM exerted maximal inhibition [35,36]. Total amounts of CD137 can be found in Supplementary Figure 3A,B.

Anoctamin-6 Modulates CD137 Release
ANO6 activation through calcium-elevation results in rapid translocation of PS to the external membrane leaflet, which triggers ADAM10/17 sheddase function. Transfection experiments were next undertaken with ANO6 and with a point substitution mutant D408G that is uniquely hypersensitive to calcium [36]. This hypersensitivity leads to unrestrained activity and, in consequence, to permanent PS-exposure and continued ADAM10 activation in resting cells [36]. Figure 4A depicts the results of experiments with HT29 cells. Transfection with ANO6 did not primarily affect levels of constitutively released sCD137. However, the release was significantly increased relative to controls upon IO-stimulation. Very conspicuously, cells transfected with the hyperactive ANO6-mutant constitutively shed large amounts of CD137 in the absence of any stimulus. Figure 4B shows the data obtained with HEK cells. The findings followed the same pattern but were yet more pronounced. Stimulated shedding in ANO6-transfected cells was augmented threefold, and transfection of the hyperactive mutant led to constitute release of sCD137 to even higher levels. Total amounts of CD137 can be found in Supplementary Figure S4.

Anoctamin-6 Modulates CD137 Release
ANO6 activation through calcium-elevation results in rapid translocation of PS to the external membrane leaflet, which triggers ADAM10/17 sheddase function. Transfection experiments were next undertaken with ANO6 and with a point substitution mutant D408G that is uniquely hypersensitive to calcium [36]. This hypersensitivity leads to unrestrained activity and, in consequence, to permanent PS-exposure and continued ADAM10 activation in resting cells [36]. Figure 4A depicts the results of experiments with HT29 cells. Transfection with ANO6 did not primarily affect levels of constitutively released sCD137. However, the release was significantly increased relative to controls upon IO-stimulation. Very conspicuously, cells transfected with the hyperactive ANO6-mutant constitutively shed large amounts of CD137 in the absence of any stimulus. Figure 4B shows the data obtained with HEK cells. The findings followed the same pattern but were yet more pronounced. Stimulated shedding in ANO6-transfected cells was augmented threefold, and transfection of the hyperactive mutant led to constitute release of sCD137 to even higher levels. Total amounts of CD137 can be found in Supplementary Figure S4. WT cells were stimulated with ionomycin (IO; 1 µM) for 30 min. In addition, cells co-transfected with CD137 and hyperactive ANO6 were analyzed in the absence of any stimulus for 30 min. The relative amount of shedding products in the supernatant was determined in relation to total CD137 by ELISA. IO-induced shedding was significantly increased upon overexpression of ANO6. Shedding of CD137 was significantly increased upon overexpression of hyperactive ANO6 compared to mock-transfected non-stimulated cells. * indicates significant increase (* p < 0.05 (A) n = 5; (B) n = 3; ± s.e.m.). Data were analyzed by one-way analysis of variance and Holm-Sidak multiple comparison post hoc test.

Release of Endogenously Expressed CD137 in T Cells
Expression of CD137 is best characterized in T cells, where it can readily be induced by their activation [1,51]. CD4 + and CD8 + T cells were activated with soluble CD3-and CD28-antibodies. Expression of CD137 was detected 6 h in CD8+ cells and 24 h in CD4 + cells following stimulation. Soluble CD137 was found to accumulate over time. Addition of the metalloprotease inhibitors marimastat, GI or GW to the media strikingly reduced levels of sCD137 in supernatants of both cell types ( Figure 5A,B). The reduction in sCD137 was accompanied by an increase in CD137 on the surface of the activated CD4 + and CD8 + cells ( Figure 5C,D). These findings show that ADAM10 is the principle effector of sCD137 release from human lymphocytes. WT cells were stimulated with ionomycin (IO; 1 µM) for 30 min. In addition, cells co-transfected with CD137 and hyperactive ANO6 were analyzed in the absence of any stimulus for 30 min. The relative amount of shedding products in the supernatant was determined in relation to total CD137 by ELISA. IO-induced shedding was significantly increased upon overexpression of ANO6. Shedding of CD137 was significantly increased upon overexpression of hyperactive ANO6 compared to mocktransfected non-stimulated cells. * indicates significant increase (* p < 0.05 (A) n = 5; (B) n = 3; ± s.e.m.). Data were analyzed by one-way analysis of variance and Holm-Sidak multiple comparison post hoc test.

Release of Endogenously Expressed CD137 in T Cells
Expression of CD137 is best characterized in T cells, where it can readily be induced by their activation [1,51]. CD4 + and CD8 + T cells were activated with soluble CD3-and CD28-antibodies. Expression of CD137 was detected 6 h in CD8+ cells and 24 h in CD4 + cells following stimulation. Soluble CD137 was found to accumulate over time. Addition of the metalloprotease inhibitors marimastat, GI or GW to the media strikingly reduced levels of sCD137 in supernatants of both cell types ( Figure 5A,B). The reduction in sCD137 was accompanied by an increase in CD137 on the surface of the activated CD4 + and CD8 + cells ( Figure 5C,D). These findings show that ADAM10 is the principle effector of sCD137 release from human lymphocytes.

Shed sCD137 Induces Proliferation of Activated T Cells
Cleavage of ADAM10 substrates always occurs close to the membrane in the stalk region of the respective substrate. Therefore, it can be assumed that shed sCD137 encompasses the extracellular domain and might retain the ability to bind to CD137 ligand.
Interestingly, T cells do not only express CD137 but also CD137L upon activation [8][9][10], (see also Supplementary Figure S5). Activation of CD137L through receptor binding might lead to pro-and anti-inflammatory effects as well [5]. Thus, we addressed the question how shed CD137 would affect T cell proliferation. According to our previous findings, huge amounts of sCD137 were released from transfected HEK cells in the presence of hyperactive ANO6 without any stimulus. Thus, we made use of the conditioned medium of CD137/hyperactive ANO6 co-transfected cells and mock-transfected cells or cells transfected with hyperactive ANO6 alone. When T cells were incubated in medium containing sCD137, T cell proliferation was significantly augmented ( Figure 6). These data indicate that shed sCD137 is functionally active and able to bind to CD137L expressing cells.

Shed sCD137 Induces Proliferation of Activated T Cells
Cleavage of ADAM10 substrates always occurs close to the membrane in the stalk region of the respective substrate. Therefore, it can be assumed that shed sCD137 encompasses the extracellular domain and might retain the ability to bind to CD137 ligand.
Interestingly, T cells do not only express CD137 but also CD137L upon activation [8][9][10], (see also Supplementary Figure S5). Activation of CD137L through receptor binding might lead to pro-and anti-inflammatory effects as well [5]. Thus, we addressed the question how shed CD137 would affect T cell proliferation. According to our previous findings, huge amounts of sCD137 were released from transfected HEK cells in the presence of hyperactive ANO6 without any stimulus. Thus, we made use of the conditioned medium of CD137/hyperactive ANO6 co-transfected cells and mock-transfected cells or cells transfected with hyperactive ANO6 alone. When T cells were incubated in medium containing sCD137, T cell proliferation was significantly augmented ( Figure 6). These data indicate that shed sCD137 is functionally active and able to bind to CD137L expressing cells.

Discussion
The TNFR-superfamily assumes a central position in the signaling network and functioning of the immune system. Many studies have been undertaken to unravel the role of the CD137/CD137L axis and its links to pathophysiological immune processes. An emerging topic relates to the function and relevance of membrane-anchored CD137 and its soluble counterpart sCD137. Elevated plasma levels of the latter have been found in patients with various malignancies as well as in patients with rheumatoid arthritis, where levels have been found to be associated with disease severity.
Current evidence indicates that sCD137 is a spliced variant of the receptor, whose expression is subject to regulation at the transcriptional level. Nevertheless, we deemed it worthwhile to investigate whether an alternative mechanism for sCD137 generation might exist that would allow for rapid, pliable modulation of its release from cells. From the beginning, similarities were noted between findings reported for sCD137 and soluble CD27 (sCD27), another member of the TNFR-superfamily. High levels of sCD27 have also been found in patients suffering from rheumatoid arthritis [52]. In vivo levels of sCD27 correlate with tumor load in patients with leukemia and lymphoma [52,53]. Huang et al. reported that sCD27 increased T cell activation and proliferation in vitro and in vivo [54]. The authors speculate that this activation might modulate tumor immunity.
sCD27 is known to be released by metalloproteinase-mediated cleavage of the membrane-anchored molecule, so the suspicion arose that the same might hold for sCD137. Application of metalloproteinase inhibitor has been reported to increase amounts of cellbound CD137 in murine lymphocytes [31], a finding that constitutes evidence for the notion of proteolytic shedding.
The collective results reported herein now leave little room for doubt that sCD137 is indeed generated mainly through proteolytic release by ADAM10. In the first set of experiments, constitutive and stimulated shedding was analyzed in CD137-transfected HEK and HT29 cells. In both cell lines, constitutive release of sCD137 was significantly diminished in the presence of a broad-spectrum metalloprotease inhibitor. Incubation with the preferential ADAM10 inhibitor GI led to a comparable reduction which indicated that

Discussion
The TNFR-superfamily assumes a central position in the signaling network and functioning of the immune system. Many studies have been undertaken to unravel the role of the CD137/CD137L axis and its links to pathophysiological immune processes. An emerging topic relates to the function and relevance of membrane-anchored CD137 and its soluble counterpart sCD137. Elevated plasma levels of the latter have been found in patients with various malignancies as well as in patients with rheumatoid arthritis, where levels have been found to be associated with disease severity.
Current evidence indicates that sCD137 is a spliced variant of the receptor, whose expression is subject to regulation at the transcriptional level. Nevertheless, we deemed it worthwhile to investigate whether an alternative mechanism for sCD137 generation might exist that would allow for rapid, pliable modulation of its release from cells. From the beginning, similarities were noted between findings reported for sCD137 and soluble CD27 (sCD27), another member of the TNFR-superfamily. High levels of sCD27 have also been found in patients suffering from rheumatoid arthritis [52]. In vivo levels of sCD27 correlate with tumor load in patients with leukemia and lymphoma [52,53]. Huang et al. reported that sCD27 increased T cell activation and proliferation in vitro and in vivo [54]. The authors speculate that this activation might modulate tumor immunity.
sCD27 is known to be released by metalloproteinase-mediated cleavage of the membraneanchored molecule, so the suspicion arose that the same might hold for sCD137. Application of metalloproteinase inhibitor has been reported to increase amounts of cell-bound CD137 in murine lymphocytes [31], a finding that constitutes evidence for the notion of proteolytic shedding.
The collective results reported herein now leave little room for doubt that sCD137 is indeed generated mainly through proteolytic release by ADAM10. In the first set of experiments, constitutive and stimulated shedding was analyzed in CD137-transfected HEK and HT29 cells. In both cell lines, constitutive release of sCD137 was significantly diminished in the presence of a broad-spectrum metalloprotease inhibitor. Incubation with the preferential ADAM10 inhibitor GI led to a comparable reduction which indicated that ADAM10 is the major metalloprotease involved in CD137 shedding. ADAM10-sheddase function is enhanced through induction of calcium-influx, which can be provoked by treatment of cells with calcium-ionophore and melittin. Both agents strongly augmented CD137 shedding. Employment of HEK cells deficient in ADAM10 and ADAM17 rounded up the findings. Constitutive release of sCD137 fell to low levels and neither ionomycin nor melittin stimulated shedding capacity, which was restored upon re-transfection of the cells with ADAM10. Interestingly, some substrates are cleaved by ADAM10 as well as ADAM17 depending on the cell type and stimulus. While ADAM17 is the major sheddase of TGF-α under normal conditions, ADAM10 releases this substrate when ADAM17 is absent [46]. Both, ADAM10 and ADAM17 are also involved in the shedding of the T cell immunoglobulin and mucin domain 3 (TIM3) protein from activated T cells [55]. Tim-3 is shed by ADAM10 and ADAM17 after PMA or ionomycin stimulation, respectively. In monocytes, ADAM10 and to a lesser extent ADAM17 were found to be responsible for LPS-induced down-regulation of Tim-3. In our study, re-transfection of ADAM17 also restored the release of CD137 in ADAM10/ADAM17 dKO cells. Thus, both proteases are in principle able to shed this protein.
Calcium-influx triggers scramblase activation resulting in exposure of PS at the membrane surface. According to the 3D structure of the extracellular domain [56] the ADAM10 active site is occluded by a short peptide loop located at the commencement of the stalk region. Attraction of the cationic PS-binding motif to surface exposed PS may thus serve to draw this loop out of the catalytic site enabling substrate access. This event is suppressed in the presence of the soluble phospholipid head group or through blockade of surface-exposed PS with lactadherin. Both, in HT29 and HEK cells, ionomycin-provoked enhancement of substrate cleavage was dose-dependently reduced by OPS and virtually abrogated by lactadherin.
Experiments followed in which cells were transfected with ANO6 to enhance expression of the major calcium-activated scramblase that translocates PS molecules from the inner to the outer membrane leaflet. Transfection was also undertaken with the hyperactive D408G substitution mutant. Overexpression of WT ANO6 did not alter constitutional shedding of CD137, but upon stimulation with calcium ionophore, however, ANO6-transfected cells shed significantly higher amounts of sCD137 into the supernatants. The behavior of cells transfected with the ANO6 hyperactive mutant was remarkable. Maximal shedding of sCD137 occurred constitutively in both HT29 and HEK cells in the absence of any stimulus. Transfection of CD137 encoding plasmid in the HT29 and HEK cell lines ruled out the contribution of alternative splicing in the generation of sCD137.
Currently, the major fraction of soluble CD137 is believed to represent the secreted splice form that is expressed by activated T cells [26]. Thus, a central question arose whether sCD137 release by these cells also involved ADAM10 or ADAM17. CD4+ and CD8+ T cells were activated and cultured in the presence or absence of ADAM inhibitors. sCD137 accumulated to high levels during the observation period of 96 h, particularly in supernatants of CD8+ lymphocytes, and release was effectively reduced by all inhibitors. These findings leave little room for doubt that generation of sCD137 in human lymphocytes also occurs mainly through the action of ADAM10. Additional experiments addressing the contribution of the spliced form in comparison to ADAM-derived soluble CD137 would be of high interest.
Rapid, regulated shedding of CD137 must therefore now be considered in the context of the function of the CD137/sCD137 axis with its multiple facets [5]. The soluble spliced form of CD137 is known to bind to its ligand. Studies are currently underway to examine the consequences, which can be expected to be complex. Interaction of cell-bound CD137 with CD137L elicits bidirectional signaling in participating cells. Tu et al. have reported that soluble CD137 also provokes reversed signaling. sCD137 was found to represent a chemotactic factor for monocytes and Jurkat T-cells that express the ligand CD137L [57].
An analogous finding was made in this study. Co-transfection of HEK cells with CD137 and hyperactive ANO6 led to vigorous constitutional shedding of sCD137 into cell supernatants. The conditioned media were found to provide a stimulus for activated lymphocytes that are known to express CD137L. The effect was not noted when HEK cells were transfected with hyperactive ANO6 alone.
This result was somewhat unexpected in the light of previous data of Eun et al. [58], who reported that CD137 ligand signaling to T cells limited T cell activation. However, their observation pertained only to murine T-cells activated with very low concentration of 0.1 µg/mL anti-CD3. The effect was much less apparent at higher concentrations of antibody. It will be of interest to analyze and compare the human and murine CD137/CD137L axis in the future. This will be important for interpretation of in vivo data obtained in animal models.
Cellular ATP-content was employed as the readout for activation as it is routinely used to assess T cell numbers [59][60][61]. ATP-increase may also derive from enhanced mitochondrial function. Interestingly, this has been described as a costimulatory effect of CD137 activation in T cells [62,63].
It is clear that the functional data presented herein are no more than preliminary and must be followed by in-depth analyses of underlying signaling pathways and cell biological events. The prime intent of this report is to draw general attention to the existing link between two major signaling networks of immune cells and malignant cells. Indeed, tumor biology will possibly become a particularly rewarding field for future research. Sophisticated ruses are employed by tumor cells for immune evasion. One is the upregulation and cell surface expression of molecules such as PD-L1 (programmed death ligand 1) and CTLA-4 (cytotoxic T-lymphocyte associated protein 4) whose interaction with the respective partners on immune cells have detrimental consequences.
Possibly, CD137/CD137L will eventually join the list of such molecules. CD137 is being detected on an ever-increasing number of tumors. It may be more than coincidence that overexpression of ADAM10 contributes to tumor development [43], and that hypoxic conditions induce increased expression of both the secreted CD137 spliced variant [64] and ADAM10. It is tempting to speculate, for example, that soluble CD137 functions as a decoy and competitor of CD137 to negatively impact on physiological functions of the CD137/CD137L axis. With the present identification of ADAM10 as the major generator of soluble CD137, a new link is provided that should prove useful to further research in the field.

T Cell Isolation and FACS Analysis
Peripheral blood mononuclear cells (PBMC) were isolated from leukocyte concentrates of healthy adult blood donors obtained from the Institute of Transfusion Medicine, University Hospital Schleswig-Holstein (Kiel, Germany) by Ficoll density gradient centrifugation, Sigma Aldrich (St. Louis, MO, USA).

Expression Vectors and Transfection
The expression vector for human CD137-GFP was purchased from OriGene, (Rockville, MD, USA. The expression vector for murine ADAM17 was from Gillian Murphy (Cambridge, UK). The expression vector for murine ADAM10 was described before. ANO6-WT plasmid (EX-I1781-M02) was purchased from GeneCopoeia, (Rockville, MD, USA. The hyperactive D408G ANO6 mutant was generated by point mutation of WT ANO6 as described in Veit et al. [36]. HEK293T, HT29 and ADAM10/ADAM17 double-deficient HEK293T cells were transfected using Turbofect Transfection Reagent, Thermo Fisher Scientific (Waltham, MA, USA), according to the manufacturer's instructions. Then, 24 h after transfection of expression vectors, cell medium was replaced by fresh DMEM. Transfection efficiency of CD137 was always controlled in parallel by Simple Western analyses (Supplementary Figures S1 and S2).

Proliferation Assay
To investigate the effect of shed CD137 on T cell proliferation CD4 + and CD8 + T cells were incubated with conditioned medium of transfected HEK293T cells. The latter were transfected with mock vectors, hyperactive ANO6 or with CD137 and hyperactive ANO6. Then, 24 h after transfection, medium was exchanged. Supernatants were collected after 24 h and precleared of cells, debris and larger vesicles by centrifugation. Primary T cells were pre-activated with anti-CD3/CD18 antibodies for 24 h and seeded into 96 well plates in 50 µL medium. Then, conditioned medium of HEK cells was added 1:1. After 72 h the cell number was determined using Cell Titer Glo Luminescent Cell Viability Assay, Promega (Fitchburg WI, USA), according to the manufacturer's instructions.

Automated Western
Cells were lysed in lysis buffer (5 mMTris-HCl (pH 7.5), 1 mM EGTA, 250 mM saccharose, 1% Triton X-100) supplemented with complete inhibitor cocktail (Roche Applied Science, Penzberg, Germany) and 10 mM 1,10-phenanthroline monohydrate. Automated Western was performed with equal concentrations of protein per sample using WES™ (Protein Simple, San Jose, CA, USA) according to the manufacturer's instructions. Detection of CD137 in HT29 and HEK cells via Simple Western is shown in Supplementary Figures S1 and S2 using anti-tGFP antibodies.

Statistical Analysis
All values for the ectodomain shedding assays are expressed as means ± standard error of the mean. The standard error values indicate the variation between mean values obtained from at least three independent experiments. Statistics were generated using one-way or two-way analysis of variance (one-way/two-way ANOVA) and multiple comparison post hoc test as designated. p values < 0.05 were considered statistically significant (either indicated with * or #).