Ligation of the CD2 Co-Stimulatory Receptor Enhances IL-2 Production from ﬁrst-Generation Chimeric Antigen Receptor T Cells

SHORT COMMUNICATION Ligation of the CD2 co-stimulatory receptor enhances IL-2 production from ﬁrst-generation chimeric antigen receptor T cells

EJ Cheadle1,3, DG Rothwell2, JS Bridgeman1,4, VE Sheard1, RE Hawkins1 and DE Gilham1

T cells bearing chimeric antigen receptors (CARs) are broadly categorised into first- and second-generation receptors. Second-generation CARs contain a co-stimulatory signalling molecule and have been shown to secrete IL-2, undergo greater proliferation and to have enhanced persistence in vivo. However, we have shown that T cells bearing a first-generation CAR containing a CD19-targeting scFv (single-chain variable fragment) and the CD3z-signalling domain are able to produce IL-2 upon co-culture with CD19 þ B-cell lymphomas independent of CD28 activity. Here, we report that signalling through endogenous CD2 following ligation with its ligands, CD48 in mouse and CD58 in humans, drives IL-2 production by first- generation CD19-specific CAR. Moreover, the high levels of IL-2 produced by human T cells engrafted with a second-generation CD28-containing CAR during target-cell recognition are dependent to a degree upon CD2 receptor activity. These observations highlight the fact that the functional activity induced by T-cell-expressed CARs is dependent upon endogenous ‘natural’ receptor interactions. A deeper understanding of the role of these activities will serve to further refine the design of future CARs to either exploit or avoid these interactions.

Gene Therapy (2012) 19, 1114--1120; doi:10.1038/gt.2011.192; published online 1 December 2011 Keywords: chimeric antigen receptor; CD2; B-cell malignancies

INTRODUCTION driving CAR T-cell IL-2 production and to question whether these Adoptive immunotherapy as a treatment for cancer using T cells ‘natural’ T-cell--target cell interactions impacted upon the activity bearing a chimeric antigen receptor (CAR) is currently under of second-generation CAR T cells. clinical investigation in several sites worldwide.1 B-cell malignancies appear particularly amenable to this approach with significant anti-tumour efficacy seen in animal models.2 In their simplest RESULTS AND DISCUSSION form, CARs consist of an scFv (single-chain variable fragment) The interaction of the mouse T-cell surface-receptor CD2 with its linked to a T-cell-signalling molecule such as CD3z.3 Upon antigen ligand CD48 results in an increase in IL-2 and IFNg synthesis ligation, these first-generation CARs drive ‘signal one’-type through mRNA stabilisation.7 Mouse T cells bearing the CD19- signalling resulting in interferon gamma (IFNg) production and specific CD3z CAR (aCD19z) highly express CD2 (data not shown), target-cell cytotoxicity. However, signal one alone is thought to be whereas CD48 is also highly expressed on A20 Balb/c lymphoma sub-optimal as further co-stimulatory signals are required to drive cells gene-modified to express the human CD19 ligand full T-cell activation, survival and proliferation.4,5 To this end, (A20hCD19, Figure 1a). Blockade of the CD2--CD48 interaction second- and third-generation CARs containing the signalling between aCD19z T cells and A20hCD19 target cells by an anti- domains for receptors, such as CD28, 4--1BB and OX-40, have CD48 antibody significantly reduced the levels of antigen-driven been produced that can compensate for the lack of co-stimulation secretion of IFNg (Figure 1b) and IL-2 (Figure 1c). Experiments in provided by first-generation CARs. Interestingly, our previous which excess anti-CD48 was washed off the tumour cells before studies using mouse T cells engrafted with a CD19-specific CD3z the addition of aCD19z T-cells confirmed that CD48 expression on first-generation CAR produced IL-2 during co-culture with CD19 þ tumour cells and not T cells was responsible for co-stimulation B-cell lymphoma target cells,6 suggesting that these T cells were (data not shown). Transcript analysis further confirmed that the receiving co-stimulatory signals independent of the CAR, with the suppression in the level of cytokines also occurs at the mRNA level B-cell targets being the obvious source driving this co-stimulation during the CD2--CD48 blockade (Figure 1d) and that this effect. However, this IL-2 secretion was found to be independent suppression was rapid, occurring within 3 h. of the major ligands (CD80 and CD86) for the most studied To confirm the significance of the role of CD48, lentiviral- co-stimulatory receptor CD28.6 We therefore sought to identify mediated small hairpin RNA (shRNA) knockdown8 of CD48 the molecules present on B-cell lymphoma tumour cells that were expression in A20hCD19 lymphoma cells was used to generate a

1Clinical and Experimental Immunotherapy Group, Department of Medical Oncology, School of Cancer and Enabling Sciences, The University of Manchester, Manchester Academic Health Science Centre, Manchester, UK and 2Clinical Immune and Molecular Monitoring Laboratory, Department of Medical Oncology, School of Cancer and Enabling Sciences, The University of Manchester, Manchester Academic Health Science Centre, Manchester, UK. Correspondence: Dr DE Gilham, Clinical and Experimental Immunotherapy Group, Department of Medical Oncology, Paterson Institute for Cancer Research, Wilmslow Road, Manchester M20 4BX, UK. E-mail: [email protected] 3Current address: Targeted Therapy Group, Department of Medical Oncology, School of Cancer and Enabling Sciences, The University of Manchester, Manchester, UK. 4Current address: T-cell Modulation Group, Cardiff University, UK. Received 3 June 2011; revised 10 October 2011; accepted 21 October 2011; published online 1 December 2011 IL-2 production driven by CD2 co-stimulation in CAR T cells EJ Cheadle et al 1115 A20 A20hCD19 Isotype CD19 CD48 A20hCD19 + IgG A20hCD19 + aCD48 18000 A20hCD19 ** 15000 12000 9000 A20hCD19LUC 6000

IFN gamma (pg/ml) 3000 0 A20hCD19CD48lo mock T-cells aCD19z

A20 A20hCD19 IFN gamma IL-2 500 A20hCD19 +IgG A20hCD19 + aCD48 ** 9 8 400 7 6 300 5 4 200 3 IL-2 (pg/ml) 2 100 1 *

relative expression RNA 0 0 aCD19z + aCD19z + aCD19z + aCD19z + mock T-cells aCD19z IgG 3hr aCD48 3hr IgG 24hr aCD48 24hr

100 Hamster IgG1 aCD48Ab Rat IgG2b aCD2Ab 7000 80 6000 60 * 5000 4000 40 ** 3000 120±2 81.5±3 150±6 59±2 IL-2 pg/ml 20 2000 1000 * % IL-2 compared to controls * * 0 0 * * A20hCD19 CD48 shRNA aCD48 Ab aCD19z CD4aCD19z CD8 aCD19z Figure 1. CD2/CD48 co-stimulation has a major role in IL-2 secretion from aCD19z T cells. (a) CD19 and CD48 expression levels determined by flow cytometry on mouse A20hCD19LUC, mouse A20hCD19CD48lo (A20hCD19 CD48lo) or isotype control. (b) IFNg release and (c) IL-2 release from mock transduced and aCD19z T cells during a 24-h co-culture with A20/A20hCD19 tumour in the presence or absence of blocking antibody to CD48 (data representative of three independent experiments). (d) IFNg and IL-2 mRNA transcript levels in aCD19z T cells following 3 or 24 h co-culture with A20hCD19 cells in the presence of anti-CD48Ab or isotype control. mRNA levels are expressed relative to those in aCD19z T cells with anti-CD48Ab. (e) IL-2 release from purified CD4 þ aCD19z T cells during a 24-h co-culture with A20hCD19 cells in the presence of anti-CD48Ab, anti-CD2Ab or relevant isotype controls. (f) IL-2 release from CD4 þ aCD19z T cells co-cultured for 24 h with A20hCD19CD48lo (CD48 shRNA) or A20hCD19pSicoRluc with anti-CD48Ab expressed as a percentage of IL-2 release compared with that from CD4 þ aCD19z T cells co-cultured with A20hCD19pSicoRluc tumour cells alone. Data are mean±s.e.m. of three independent experiments. *Po0.05, **Po0.001 (unpaired Student’s t-test). clone expressing a much reduced level of CD48 (mean that these results were not due to the targeting of a human fluorescence intensity of CD48 staining 2.5 for CD48lo versus 531 protein expressed on a mouse cell line (data not shown). These of A20hCD19 cells, Figure 1a). Consequently, when cultured with results strongly implicate CD2 as being the major receptor driving CD48loA20hCD19 tumour cells, aCD19z CAR T cells produced IL-2 production in aCD19z CAR T cells when targeting B-cell significantly reduced levels of IL-2 (49% of A20hCD19 control lymphoma. target cells, Figure 1e) despite equivalent levels of CD19 target As second-generation CAR’s incorporate co-stimulatory signal- expression (data not shown). These absolute levels were higher ling domains to drive the production of IL-2,9,10 we questioned than that achieved with antibody blockade, most likely due to the whether CD2 also impacted upon IL-2 production in these more remaining low level of CD48 expression on the target cells. advanced CARs when targeting tumours expressing CD48. The Furthermore, experiments using purified CAR T-cell subsets addition of CD28 to the CD19-targeting CAR (aCD19CD28z) did confirmed that CD4 þ CAR T cells were primarily responsible for not significantly increase the absolute level of IL-2 produced antigen-specific IL-2 production and this could be blocked by above that of aCD19z CAR T cells (Figure 2a), although there was a either anti-CD48 or anti-CD2 mabs (Figure 1f). Importantly, similar trend to higher IL-2 production. However, in contrast to aCD19z results were seen in preliminary experiments using mouse T cells T cells, blocking antibodies to CD48 on A20hCD19 tumour cells bearing a mouse CD19-specific CAR-targeting A20 cells, indicating or CD2 on T cells had no discernible effect upon IL-2 secretion

& 2012 Macmillan Publishers Limited Gene Therapy (2012) 1114 --1120 IL-2 production driven by CD2 co-stimulation in CAR T cells EJ Cheadle et al 1116

aCD19z aCD19CD28z CD4+ aCD19z CD4+ aCD19CD28z 12000 110 100 10000 90 8000 80 70 6000 60 50 4000 40 30 ** * 2000 20 % IL-2 relative to controls 10 0 0 (IL-2 (pg/ml) per 100000 CAR T-cells Unselected CD4+ CD8+ Isotype CD48 CD2 blockade blockade

aCD19z aCD19CD28 IgG aCD58 100 1800 90 1600 80 1400 70 1200 60 1000 50 800 40 * 30 600 20 400 Relative Cytolysis (%) 10 200 ** 0 0 Ham aCD48Rat aCD2Ham aCD48 Rat aCD2 IL-2 (pg/ml) per 100000 CAR T-cells Raji Nalm-6 IgG1 IgG2a IgG1 IgG2a

IgG aCD58 25000

20000

15000

10000

5000

IL-2 (pg/ml) per 100000 CAR T-cells Raji Nalm-6 Figure 2. Addition of CD28 co-stimulation to CAR bypasses CD2/CD48-mediated IL-2 production. (a) IL-2 release from mixed, CD4 þ and CD8 þ aCD19z and aCD1928z T cells upon 24-h co-culture with mouse A20hCD19 tumour (data are mean±s.e.m., n ¼ 3) (unselected and CD4 þ ), n ¼ 2 (CD8 þ ). (b) IL-2 release from CD4 þ aCD19z T cells or CD4 þ aCD19CD28z T cells co-cultured for 24 h with A20hCD19 cells in the presence of anti-CD48Ab or anti-CD2Ab expressed as a percentage of IL-2 release compared with that from CD4 þ aCD19z T cells co-cultured with A20hCD19 cells in the presence of isotype control (data are mean±s.e.m., n ¼ 3). *Po0.05, **Po0.001 (unpaired Student’s t-test). (c) Antibody blockade of CD48 or CD2 does not impact upon short-term cytotoxicity mediated by aCD19z or aCD19CD28z CAR T cells. Data are representative of three independent experiments. (d) IL-2 release from human aCD19z T cells co-cultured for 24 h with Raji or Nalm-6 cells in the presence of mouse IgG1 or anti-CD58Ab. Data are mean±s.e.m., n ¼ 2. *Po0.05, **Po0.001 (unpaired Student’s t-test). (e) IL-2 release from human aCD19CD28z T cells co-cultured for 24 h with Raji or Nalm-6 cells in the presence of mouse IgG1 or anti-CD58Ab. Data are mean±s.e.m., n ¼ 2.

from CD4 þ aCD19CD28z T cells (Figure 2b). Moreover, blocking domain into the CAR induced an enhanced level of IL-2 production CD2--CD48 ligation did not impact upon the ability of aCD19z as compared with aCD19z T --cells, which could be partially reduced or aCD19CD28z T cells to lyse CD19 þ tumour cells (Figure 2c). through CD58 blockade (Figure 2e). To extend these observations, No lysis was seen with control non-signalling aCD19 mtm T cells. aCD19z and aCD19CD28z CAR T cells from five donors were In humans, the ligand for CD2 is CD58, which is structurally co-cultured with Raji or Nalm-6 cell lines in the presence of anti- related to murine CD48 and is expressed on many haematopoietic CD58 or isotype control antibodies. CD58 blockade induced a cells including B cells.11 Antibody blockade of CD58 present on significant reduction in the frequency of CD4 þ IL-2 þ CAR T cells Raji tumour cells resulted in a 60% reduction of IL-2 production by when cultured with both cell lines irrespective of whether the human T cells engrafted with the aCD19z CAR after tumour-CAR CD28 domain was present (Figures 3a and b). Although the T-cell co-culture (Figure 2d). This reduction in IL-2 production was absolute frequency of CD8 þ IL-2 þ CAR T cells was much lower, even more pronounced when CD58 was blocked on the Nalm-6 the relative frequency of these cells was reduced in aCD19z and cell line (Figure 2d) potentially due to the lack of CD80/CD86 aCD19CD28z CAR T-cell populations when CD58 was blocked, co-stimulatory molecules expressed by this tumour line.12 In which reached statistical significance in the Nalm-6 cultures contrast to mouse T cells, the addition of the CD28-signalling (Figures 3c and d).

Gene Therapy (2012) 1114 --1120 & 2012 Macmillan Publishers Limited IL-2 production driven by CD2 co-stimulation in CAR T cells EJ Cheadle et al 1117 Raji Nalm-6

p=0.0456* p=0.0113* p=0.0460* p=0.0275* 20 20

15 15 T cells (%)

T cells (%) 10 10 + + lL-2 lL-2 + + 5 5 CD4 CD4 0 0

CD19z + Isotype CD19z + Isotype CD19z + anti-CD58 CD19z + anti-CD CD19CD28z + Isotype CD19CD28z + Isotype CD19CD28z + anti-CD58 CD19CD28z + anti-CD58

p=0.0086** p=0.0225* 8 NS NS 4

6 3 T cells (%) T cells (%)

4 + 2 + lL-2 lFN + 2 + 1 CD8 CD8 0 0

CD19z + Isotype CD19z + Isotype CD19z + anti-CD58 CD19z + anti-CD58 CD19CD28z + Isotype CD19CD28z + Isotype CD19CD28z + anti-CD58 CD19CD28z + anti-CD58

Raji Nalm-6 Raji Nalm-6 p=0.0383* NS p=0.0123* NS 20 15

15 10 T cells (%)

10 T cells (%) + + 5 lL-2 lFN + + 5 CD4 CD4 0 0

CD19CD2z + Isotype CD19CD2z + Isotype CD19CD2z + Isotype CD19CD2z + Isotype CD19CD2z + anti-CD58 CD19CD2z + anti-CD58 CD19CD2z + anti-CD58 CD19CD2z + anti-CD58 Figure 3. Blockade of CD58 reduces the relative frequency of IL-2-producing aCD19z and aCD19.CD28z CAR T cells. aCD19z or aCD19.CD28z CAR T cells were co-cultured for 16 h either in the presence of isotype or anti-CD58 (10 mgmlÀ1) with (a) Raji or (b) Nalm-6 B-cell lines with the relative frequency of CAR þ CD4 þ IL-2 þ T cells determined by intracellular flow cytometry. (c, d) The frequency of CD8 þ IL-2 þ CAR T cells was also identified in the same cultures. (e) aCD19.CD2z CAR T cells were cultured with Raji or Nalm-6 tumour cell lines in the presence of isotype control or anti-CD58 for 16 h and CAR þ CD4 þ IFNg þ T-cell frequency determined by intracellular flow cytometry. (f) The frequency of CAR þ CD4 þ IL-2 þ T cells was also identified by flow cytometry in the same assay. Each data point represents a single donor. Paired t-tests were used to determine significance.

Thus, the high levels of IL-2 produced as a result of aCD19CD28z signalling differences in mouse against human T cells or whether CAR activity in human T cells appears to incorporate a there is an intrinsic difference in the way that the human receptor contribution from CD2--CD58 interactions when targeting sequences function in mouse T cells remains unclear. CD58 þ tumour cells, whereas CD2 ligation in the mouse does CD58 is expressed on all cell types13 and at high levels on not appear to signiﬁcantly enhance IL-2 cytokine production aggressive diffuse large B-cell lymphoma,14 whereas it is also induced by the same receptor. Whether this is due to downstream frequently over expressed in B-cell acute lymphoblastic leukaemia.15

& 2012 Macmillan Publishers Limited Gene Therapy (2012) 1114 --1120 IL-2 production driven by CD2 co-stimulation in CAR T cells EJ Cheadle et al 1118 Together these observations suggest that CAR T cells targeting of Technology, USA) via Dr Tim Sommerville (Paterson Institute for Cancer these malignancies may benefit from CD2-based signalling Research, Manchester, UK). The lentiviral vectors pMD2.g21 and and enhanced-IL-2 production. As most clinical CAR T-cell pCMVDR8.91(ref. 22) was originally generated in the laboratory of Professor adoptive transfer protocols include in vivo IL-2 support, there Didier Trono (EPFL, Switzerland). may be a rationale to reduce the level of IL-2 given to patients where the tumour expresses CD58. On the other hand, many solid Cell culture tumours, such as gastric and colorectal carcinomas, have reduced/ All cell culture media were obtained from Invitrogen (Paisley, Scotland) absent CD58 expression and may benefit from the inclusion of the (ref. 23) 16 --18 unless stated otherwise. 293T (ATCC CRL-11268) and GP þ e86 cells CD2-signalling domain in CAR design. However, whereas were cultured in DMEM Glutamax media supplemented with 10% fetal mouse aCD19z T cells can effectively eradicate long-term systemic 5 calf serum and passaged by trypsinization. The A20/A20hCD19iGFP Balb/c B-cell lymphoma in syngeneic mice, these aCD19z T cells fail to 6 11 B-cell lymphoma line (ATCC TIB-208) and Raji (ATCC CCL-86) was grown in persist for extended periods suggesting that IL-2 induced by RPMI 1640 (BioWhittaker, Lonza, Verviers, Belgium) supplemented with CAR T cells through CD2 ligation is not sufficient to support 10% heat-inactivated fetal calf serum, 25 mM Hepes, 50 nM 2-mercapto- aCD19z T cells. Importantly, host B cells returned to normal ethanol and 2 mM L-glutamine, penicillin, streptomycin (complete RPMI). frequencies shortly after aCD19z T-cell depletion in this model.11 However, T cells bearing a CD28-containing CD19-specific CAR maintained a longer duration of B-cell depletion indicative of Generation of CARs and producer cell lines prolonged activity of the CAR T cells.19 This observation suggests The pMP71tCD34.2A.CD19z and pMP71tCD34.2A.CD19 mtm retroviral vectors that CD28 signalling from a CAR may provide greater survival were generated, as described previously.6,24 pMP71tCD34.2A.CD19CD28z signals than that of endogenous CD2 signalling in vivo. retroviral vectors were generated as follows. Truncated CD28(IEV) (bp 336-- 10 An obvious question raised by this work is why does 660, aa 114--220) sequences were generated by PCR, fused to human endogenous CD28 signalling not appear to have a role in the CD3z using overlapping primers and annealing PCR and sub-cloned aCD19z T-cell situation6 although CD2 does function? Recent into pMP71tCD34.2A.CD19/pMP71tCD34.2A.MFE as a NotI, HindIII frag- studies have confirmed that the CAR’s containing the CD3z ment. PG13/GP þ e86 producer cell lines were generated as previously 6 transmembrane domain strongly interacts with the endogenous described. TCR/CD3 complex20 and these interactions may affect endo- The aCD19.CD2z CAR was cloned by overlapping PCR. To this end, genous CD28 function. Moreover, although CD2 ligation in aCD19z CD2 was amplified from human complementary DNA using the follow- CAR T cells is active in terms of cytokine production, it is not clear ing primer pair: 50-GAGCTTTCCATGTAAATTTGTAGCCAGCTTCCTTCTG-30 whether CD2 is functioning equivalently to that achieved during and 50-TTAATTAGAGGAAGGGGACAATGAGTTTTCTGCTGC-30. A 471-bp frag- endogenous TCR binding of target antigen. Additionally, very little ment encoding the complete human CD2 transmembrane and cytoplasmic is known about the biochemical structure/cell surface location and domain, and nine extracellular amino acids, was then amplified from this structural interactions of receptors not using the CD3z transmem- product using a 50 primer that included a 50 NotI site (50-GCGGCCGCAG brane domain including the CD28-containing CAR used in these AGAAAGGTCTGGACATCTATCTCATCATTGGCATATGT-30), and a 30 reverse and other studies. Indeed, our own recent observations suggest primer that overlapped with the cytoplasmic domain of CD2 (50-CTCCTGCT that CARs incorporating a CD3z transmembrane domain with GAACTTCACATTAGAGGAAGGGGACAATGAGTTTTCTGCTGCCCCATGGGG-30). CD28-signalling domain have a far reduced co-stimulatory signal CD3z was amplified from rKat.om1.B18.CD3z.IRES.GFP using a 50 primer capacity as compared with receptors bearing the CD28 trans- that overlaps with the CD2 cytoplasmic domain (50-CATTGTCCCCTTCCT membrane domain (JSB, personal communication). Indeed, initial AATGTGAAGTTCAGCAGGAGCGCAGACGC-30) and a 30 reverse primer that testing of a CD19-specific CAR bearing the CD2 endodomain fused included a 30 XbaI site and stop codon (50-TTTGAGTGGTGAAATCTCTAGAT to CD3z failed to drive enhanced-IL-2 production when expressed TATTAGCGAGGGGGCAGGGCCTGC-30). The fragments were annealed and in human T cells and co-cultured with CD58 blocked Raji and the correctly annealed sequence amplified using the 50 CD2 and 30 CD3 Nalm-6 target cell lines (Figures 3e and f). As such, further detailed primers. The CAR was cloned into rKat.MFE23 using a NotI and blunted examination of the biochemical and structural aspects of T cells XbaI site. Finally, the MFE23 scFv sequence was replaced with the aĆD19 bearing CARs is required in order to fully understand the relative scFv via a ClaI/NotI digest from the existing aCD19.CD3z CAR sequence. frequency of T-cell activation mediated by the CAR and to understand the role that naturally expressed host proteins play in Isolation, transduction and ex vivo culture of murine T cells the anti-tumour response mediated by the CAR-expressing T-cell. Mouse T cells were generated as previously described.6

MATERIALS AND METHODS Isolation, transduction and ex vivo culture of human T cells General reagents Human T cells were generated as previously described.25 Hamster anti-mouse CD3e (clone145-2C11) and hamster anti-mouse CD28 (clone 37.51) were obtained from BD-Pharmingen (Cowley, UK). Human Generation of CD48 shRNA A20hCD19 cell line Interleukin-2 (Proleukin) was obtained from Novartis (Horsham, UK) and We prepared CD48shRNA lentiviral vectors by cloning palindromic double- recombinant murine IL-7 from R&D Systems (Abingdon, UK). FITC, PE and stranded oligonucleotide sequences of CD48shRNA into HpaI and XhoI PE-Cy5-conjugated monoclonal antibodies to human CD2, CD19, human sites of a pSicoR-cherry red vector following instructions from the CD34, mouse CD2, CD4, CD8, CD19 and CD48, were obtained from laboratory of Dr Taylor Jacks (Massachusetts Institute of Technology).8,26 BD-Pharmingen. PE-conjugated anti-human CD58 was obtained from Lentivirus was produced by transient transfection of 293T cells in 150 mm3 Serotech and all chemical reagents from Sigma-Aldrich (Poole, UK) unless dishes by calcium chloride precipitation.27 Vectors were added in the ratio otherwise stated. Matched antibody pairs to mouse IFNg and IL-2 were 20 mg pSicoR, 13 mg pCMVDR8.91, 7 mg pMD2.g. Lentiviral particles were obtained from BD-Pharmingen (R4-6A2 and JES6-1A12/5H4, respectively). puriﬁed by ultracentrifugation at 71935 g, 4 1C, 2 h (Sorvall Ultra pro 80). Matched antibody pairs to human IL-2 were obtained from R&D Systems A20hCD19iGFP cells were transduced with pSicoR-luciferase or pSicoR- (MAB602 and BAF202). Blocking antibody to mouse CD48 (HM48-1) and CD48 by two rounds of spinfection at 2500 rpm for 3 h. Cells were sorted

mouse IgG1 (G235-2356) were obtained from BD-Pharmingen, blocking 5--7 days later for cherry-red expression on a FACSVantage and analysed antibody to mouse CD2 (RM2-5) and Rat IgG2b (eB149/10H5) from for CD48 expression by flow cytometry. Cells transduced with the shRNA eBioscience (Hatfield, UK) and blocking antibody to human CD58 from vectors with the 19mer 145 (50-GGACCATATAAACGTATCA-30) and 655 R&D Systems (Novartis). The shRNA vectors pSicoR cherry red-luc and (50-GGAGTATGTTGGACTGCAA-30) sequences were selected as giving the pScioR cherry red were a gift from Dr Taylor Jacks (Massachusetts Institute highest level of CD48 knockdown by flow cytometry. A20hCD19iGFP cells

Gene Therapy (2012) 1114 --1120 & 2012 Macmillan Publishers Limited IL-2 production driven by CD2 co-stimulation in CAR T cells EJ Cheadle et al 1119 that went through two rounds of transduction with the CD48shRNA 655 in the Molecular Biology Core Facility and the flow cytometry service. We also thank vector and one transduction with the CD48shRNA 145 vector were sorted the Kay Kendall Leukaemia Fund, Cancer Research UK and the FP6 programme (FACSVantage) for CD48-negative cells and selected as the line with the ‘ATTACK’ for funding this work. EJC was funded by the Kay Kendall Leukaemia Fund. lowest possible CD48 expression levels. A20hCD19iGFP pSicoRluc cells DGR, JSB, REH and DEG were funded by Cancer Research UK. were generated by two rounds of transduction with the pSicoR cherry red-luc vector and cell sorting for cherry-red expression. REFERENCES T-cell/tumour cell co-cultures 1 Morgan RA, Dudley ME, Rosenberg SA. Adoptive cell therapy: genetic 5 5 A total of 10 tumour cells were cultured with 10 murine T cells (30--40% modification to redirect effector cell specificity. Cancer J 2010; 16: 336 --341. CAR þ ) for 24 h at 37 1C in U-bottomed 96-well plates and supernatant 2 Park JH, Brentjens RJ. Adoptive immunotherapy for B-cell malignancies with analysed for IFNg and IL-2 by enzyme-linked immunosorbent assay. The autologous chimeric antigen receptor modified tumor targeted T cells. Discov limit of detection was 31.25 pg mlÀ1. In some instances, A20hCD19 cells Med 2010; 9: 277 --288. À1 3 Bridgeman JS, Hawkins RE, Hombach AA, Abken H, Gilham DE. Building better were pre-treated for 30 min with 10 mgml Hamster IgG1/anti CD48Ab chimeric antigen receptors for adoptive T cell therapy. Curr Gene Ther 2010; 10: that was not removed during the 24-h co-culture or mouse T cells were 77 --90. pre-treated for 30 min with 10 mg/ml Rat IgG2b/anti-CD2Ab. At the end of 4 Hombach A, Sent D, Schneider C, Heuser C, Koch D, Pohl C et al. T-cell activation the culture period, cells were re-suspended in phosphate-buffered saline/ by recombinant receptors: CD28 costimulation is required for interleukin 2 1% fetal calf serum containing anti-mouse CD19PE and incubated on secretion and receptor-mediated T-cell proliferation but does not affect receptor- ice for 30 min. 10 000 CountBright counting beads (Molecular Probes, mediated target cell lysis. Cancer Res 2001; 61: 1976 --1982. Invitrogen, Eugene, OR, USA) were added and samples acquired on a 5 Sadelain M, Brentjens R, Riviere I. The promise and potential pitfalls of chimeric FACscan or FacsCalibur. The number of remaining CD19 þ cells was antigen receptors. Curr Opin Immunol 2009; 21: 215 --223. calculated using the formula: (number of Ag-positive cell events/number of 6 Cheadle EJ, Hawkins RE, Batha H, Rothwell DG, Ashton G, Gilham DE. Eradication bead events) Â number of beads added. Relative cytotoxicity was deter- of established B-cell lymphoma by CD19-specific murine T cells is dependent on host lymphopenic environment and can be mediated by CD4+ and CD8+ T cells. mined by taking the number of tumour cells remaining in signalling-active J Immunother 2009; 32: 207 --218. CAR T-cell cultures divided by the number of tumour cells in the relevant 7 Musgrave BL, Watson CL, Haeryfar SM, Barnes CA, Hoskin DW. CD2-CD48 aCD19 mtm CAR culture expressed as a percentage to determine relative interactions promote interleukin-2 and interferon-gamma synthesis by stabilizing cytolysis. cytokine mRNA. Cell Immunol 2004; 229:1--12. Human T-cell/tumour (Raji, Nalm-6) co-cultures were performed in 8 Ventura A, Meissner A, Dillon CP, McManus M, Sharp PA, Van Parijs L et al. Cre-lox- a similar manner except that tumour cells were pre-treated with mouse regulated conditional RNA interference from transgenes. Proc Natl Acad Sci USA

IgG1/aCD58Ab where appropriate. 2004; 101: 10380 --10385. 9 Krause A, Guo HF, Latouche JB, Tan C, Cheung NK, Sadelain M. Antigen- dependent CD28 signaling selectively enhances survival and proliferation in Intracellular flow cytometry genetically modified activated human primary T lymphocytes. J Exp Med 1998; A total of 5 Â 105 tumour cells were co-cultured with 5 Â 105 CAR or 188: 619 --626. control T cells in 1 ml of complete media supplemented with 1 mlof 10 Maher J, Brentjens RJ, Gunset G, Rivie`re I, Sadelain M. Human T-lymphocyte Golgistop (BD Biosciences, Oxford, UK) per well of a 24-well plate. After cytotoxicity and proliferation directed by a single chimeric TCRzeta /CD28 16 h, cells were collected, blocked with Fc blocking reagent (Miltenyi receptor. Nat Biotechnol 2002; 20: 70 --75. Biotec, Bergisch Gladbach, Germany) and stained with CD4-FITC and CD34- 11 Kato K, Koyanagi M, Okada H, Takanashi T, Wong YW, Williams AF et al. CD48 is a counter-receptor for mouse CD2 and is involved in T cell activation. J Exp Med Pe-Cy5 mabs (BD Biosciences) for 20 min. After washing, intracellular 1992; 176: 1241 --1249. staining was performed using Cytofix/Cytoperm kit (as directed by the 12 Brentjens RJ, Santos E, Nikhamin Y, Yeh R, Matsushita M, La Perle K et al. manufacturer’s instructions) using anti-IFNg and anti-IL-2 (BD Biosciences). Genetically targeted T cells eradicate systemic acute lymphoblastic leukemia Samples were analysed using a BD FACs Calliber and FlowJo (Treestar, xenografts. Clin Cancer Res 2007; 13: 5426 --5435. Ashland, OR, USA) software. In order to identify cytokine-producing T cells, 13 Krensky AM, Sanchez-Madrid F, Robbins E, Nagy JA, Springer TA, Burakoff SJ. The lymphocytes were gated on a FSC-H v SSC-H plot, CAR T cells identified by functional significance, distribution, and structure of LFA-1, LFA-2, and LFA-3: cell CD34 expression and cytokine-producing cells identified by comparison surface antigens associated with CTL-target interactions. J Immunol 1983; 131: with T cells cultured in the absence of target cells. 611 --616. 14 Jacob MC, Agrawal S, Chaperot L, Giroux C, Gressin R, Le Marc’Hadour F et al. Quantification of cellular adhesion molecules on malignant B cells from non- Real-time PCR Hodgkin’s lymphoma. Leukemia 1999; 13: 1428 --1433. A total of 5 Â 105 A20hCD19iGFP tumour cells were co-cultured with 15 Lee RV, Braylan RC, Rimsza LM. CD58 expression decreases as nonmalignant B 5 Â 105 aCD19z murine T cells (38% CAR þ ) in 24-well plates for 3 or 24 h in cells mature in bone marrow and is frequently overexpressed in adult and À1 þ pediatric precursor B-cell acute lymphoblastic leukemia. Am J Clin Pathol 2005; the presence of 10 mgml Hamster IgG1 or anti-CD48Ab. CAR T cells were purified using CD34 microbeads (Miltenyi Biotech) and RNA isolated 123: 119 --124. 16 Diederichsen AC, Stenholm AC, Kronborg O, Fenger C, Jensenius JC, Zeuthen J using the QIAGEN RNeasy kit as per the manufacturer’s instructions et al. Flow cytometric investigation of immune-response-related surface (Qiagen, Crawley, UK) and up to 1 mg of total RNA used to synthesize molecules on human colorectal cancers. Int J Cancer 1998; 79: 283 --287. complementary DNA using the Superscipt III first-strand complementary 17 Koretz K, Schlag P, Moller P. Sporadic loss of leucocyte-function-associated DNA synthesis kit (Invitrogen). Real-time PCR for mouse IL-2 was antigen-3 (LFA-3) in colorectal carcinomas. Virchows Arch A Pathol Anat performed on an ABI 7900 Prism Instrument using the mouse IL-2 or Histopathol 1991; 419: 389 --394. mouse b-actin TaqMan Gene Expression Assays (Applied Biosystems, 18 Mayer B, Lorenz C, Babic R, Jauch KW, Schildberg FW, Funke I et al. Expression of Warrington, UK) as per the manufacturer’s instructions. leukocyte cell adhesion molecules on gastric carcinomas: possible involvement of LFA-3 expression in the development of distant metastases. Int J Cancer 1995; 64: 415 --423. CONFLICT OF INTEREST 19 Kochenderfer JN, Yu Z, Frasheri D, Restifo NP, Rosenberg SA. Adoptive transfer of syngeneic T cells transduced with a chimeric antigen receptor that recognizes The authors declare no conflict of interest. murine CD19 can eradicate lymphoma and normal B cells. Blood 2010; 116: 3875 --3886. 20 Bridgeman JS, Hawkins RE, Bagley S, Blaylock M, Holland M, Gilham DE. ACKNOWLEDGEMENTS The optimal antigen response of chimeric antigen receptors harboring the We thank Professor Doug Fearon for providing the 1D3 hybridoma, and Dr Taylor CD3zeta transmembrane domain is dependent upon incorporation of the Jacks and Dr Tim Sommerville for the pSicoR vectors. We would also like to thank receptor into the endogenous TCR/CD3 complex. J Immunol 2010; 184: the service units at the Paterson Institute for Cancer Research, in particular, the staff 6938 --6949.

& 2012 Macmillan Publishers Limited Gene Therapy (2012) 1114 --1120 IL-2 production driven by CD2 co-stimulation in CAR T cells EJ Cheadle et al 1120 21 Naldini L, Blo¨mer U, Gallay P, Ory D, Mulligan R, Gage FH et al. In vivo gene antitumor activity of CD19-speciﬁc engineered T cells. J Immunol 2010; 184: delivery and stable transduction of nondividing cells by a lentiviral vector. Science 1885 --1896. 1996; 272: 263 --267. 25 Cheadle EJ, Gilham DE, Hawkins RE. The combination of cyclophosphamide and 22 Zufferey R, Nagy D, Mandel RJ, Naldini L, Trono D. Multiply attenuated human T cells genetically engineered to target CD19 can eradicate established B- lentiviral vector achieves efﬁcient gene delivery in vivo. Nat Biotechnol 1997; cell lymphoma. Br J Haematol 2008; 142:65--68. 15: 871 --875. 26 Jacks T. The pSico and pSicoR Information Page. Massachusetts Institute of 23 Markowitz D, Goff S, Bank A. A safe packaging line for gene transfer: Technology, http://web.mit.edu/jacks-lab/protocols/pSico.html. separating viral genes on two different plasmids. J Virol 1988; 62: 27 Gilham DE, O’Neil A, Hughes C, Guest RD, Kirillova N, Lehane M et al. Primary 1120 --1124. polyclonal human T lymphocytes targeted to carcino-embryonic antigens and 24 Cheadle EJ, Hawkins RE, Batha H, O’Neill AL, Dovedi SJ, Gilham DE. Natural neural cell adhesion molecule tumor antigens by CD3zeta-based chimeric expression of the CD19 antigen impacts the long-term engraftment but not immune receptors. J Immunother 2002; 25: 139 --151.

Gene Therapy (2012) 1114 --1120 & 2012 Macmillan Publishers Limited