| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Basic Science Investigation |
1 Department of Radiology, Memorial Sloan-Kettering Cancer Center, New York, New York; and 2 Immunology Program, Department of Medicine, Memorial Sloan-Kettering Cancer Center, New York, New York
Correspondence: For correspondence or reprints contact: Vladimir Ponomarev, Molecular Imaging Laboratory, Department of Radiology, Memorial Sloan-Kettering Cancer Center, 1275 York Ave., Box 501, Z-2063, New York, NY 10021. E-mail: ponomarv{at}mskcc.org
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONReferences |
|---|
Key Words: PET • molecular imaging • adoptive immunotherapy • HSV1tk • 18F-FEAU
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONReferences |
|---|
Prostate-specific membrane antigen (PSMA) is an attractive target molecule for immunotherapy in patients with advanced prostate cancer. Gade et al. recently reported successful targeting of human PSMA (hPSMA)–positive tumors with human T lymphocytes grafted with anti-hPSMA CAR in a mouse model (13). However, despite high hPSMA-specific cytotoxic potential, only 50% of animals with orthotopic prostate tumors and 30% of mice with systemic tumors were cured. Continuous monitoring of the spatial distribution and dynamics of T-cell trafficking and persistence at tumor sites should help lead to an understanding of the impact of genetically modified T cells on therapy outcome. Because treatment results largely depend on multiple dynamic cell interactions, the monitoring of T-cell immunotherapy must be performed in a physiologic context. Noninvasive imaging of adoptively transferred T cells enables the study of T-cell homing, survival, and persistence as well as the assessment and prediction of therapeutic efficacy.
Several groups have demonstrated the feasibility of in vivo monitoring of adoptively transferred antigen-specific T cells transduced to express a herpes simplex virus type 1 thymidine kinase (HSV1tk) reporter gene in animal models. Murine splenocytes expressing HSV1sr39tk were shown to traffic to locally implanted subcutaneous tumors by PET with 9-4-18F-fluoro-3-(hydroxymethyl)butylguanine (14–16). Human HSV1tk-transduced T lymphocytes targeting subcutaneous Epstein–Barr virus–positive and Wilms' tumor antigen–positive xenografts were successfully imaged by PET with 2'-fluoro-2'-deoxy-1-β-D-arabinofuranosyl-5-124I-iodouracil (17) and 2'-18F-fluoro-2'-deoxy-1-β-D-arabinofuranosyl-5-ethyluracil (18F-FEAU) (18), respectively. There have been no published data on nuclear reporter gene imaging of T cells targeting systemic tumors as a more clinically relevant cancer model.
In this study, we used a systemic prostate carcinoma model for repetitive monitoring of hPSMA-targeted T cells by PET and bioluminescence imaging (BLI). We show that quantitative PET can be used not only for monitoring the spatial and temporal distributions of CAR-bearing T cells but also for predicting T-lymphocyte efficacy and therapy outcome. Moreover, this is the first study in which both therapeutic and nuclear reporter genes have been successfully cointroduced and coexpressed in human peripheral blood T lymphocytes and shown to be fully functional in transduced cells.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONReferences |
|---|
; Cz1) was kindly provided by Dr. Hinrich Abken (Universität zu Köln, Cologne, Germany) (22). The bicistronic SFG-PIP retroviral vector carries the hPSMA and puromycin resistance genes (21). Human CEA (hCEA) cDNA (kindly provided by Dr. Jeffrey Schlom, National Institutes of Health, Bethesda, MD) and a neomycin resistance gene were subcloned into the SFG-PIP retroviral vector, replacing the hPSMA and puromycin resistance genes, respectively. The resulting vector was named SFG-CIN.
Cell Line Preparation and Maintenance
NIH3T3 fibroblasts and retroviral producer cell line PG13 (American Type Culture Collection) were cultured in Dulbecco's modified Eagle medium with a high glucose concentration and 10% fetal calf serum and supplemented with 2 mM L-glutamine (Invitrogen). NIH3T3 B7.1 (CD80)–positive cells were transduced with bicistronic vector SFG-PIP or SFG-CIN. Fresh viral supernatants from PG13 cells were used for T-cell transductions. Murine prostate carcinoma cell line RM1 was cultivated in Dulbecco's modified Eagle medium with a high glucose concentration and supplemented with sodium pyruvate at a concentration of 110 mg/L, 10% fetal bovine serum (Hyclone), and L-glutamine at a concentration of 2 mmol/L. RM1 cells were transduced with hPSMA and the RFP/Rluc reporter gene. After transduction, the cells were sorted on the basis of RFP expression, and hPSMA-expressing cells were selected with puromycin at a concentration of 3 µg/mL.
T-Cell Transduction and Expansion
Peripheral blood (100 mL) was drawn from healthy donors who gave informed consent. At 48 h after stimulation with phytohemagglutinin at a concentration of 2 µg/mL, T cells were transduced with the Pz1- or Cz1-encoding retroviral vector as described previously (13) in the presence of interleukin 2 (20 U/mL). On day 7, CAR-transduced T cells were selectively expanded by cocultivation with NIH3T3 target cells expressing B7.1. At 48 h after restimulation on hPSMA-positive NIH3T3 target cells (for Pz1-positive T cells) or hCEA-positive NIH3T3 target cells (for Cz1-positive T cells), T lymphocytes were transduced with retroviral vector SFG-nesHSV1tk/GFP or SFG-nTGCBR. During T-cell expansion, the medium was supplemented with interleukin 15 at a concentration of 10 ng/mL every 3 d.
Flow Cytometry
Pz1 and Cz1 CAR expression was detected with phycoerythrin-conjugated cross-reactive goat antimouse IgG (Calbiochem) and phycoerythrin-conjugated F(ab')2 antihuman IgG1 (Southern Biotechnology), respectively. Antihuman antibodies against CD3, CD4, CD8, and CD80 (BD Biosciences) were used. hPSMA and hCEA were detected on tumor cells with antibody J591 (kindly provided by Dr. Neil Bander, Weill Medical College of Cornell University, New York, NY) and anti-hCEA antibody (anti-CD66 antibody; Pharmingen), respectively. Stained cells were processed with a FACScan (BD Biosciences) instrument and analyzed with Cell Quest software (BD Biosciences).
3H-FEAU Accumulation Assays In Vitro
Radiotracer accumulation assays were performed as described previously (23). The cells were seeded at a density of 2 x 106 cells per milliliter in 75-cm3 tissue culture flasks (Nunc). The cells were harvested after various periods of incubation (30, 60, and 120 min) with [methyl-3H]FEAU at 3.7 kBq/mL (1.48 TBq/mmol) (Moravek Biochemicals) and assayed for radioactivity concentrations with a TriCarb 1600 β-spectrometer (Packard) by standard 3H channel counting techniques. The data were expressed as a harvested cell-to-medium concentration ratio: ([disintegrations per minute] per gram of cells per minute)/([disintegrations per minute] per milliliter of medium per minute). This ratio is a measure of HSV1tk reporter gene activity and correlates with independent measures of gene expression.
Target-Specific Cell Lysis Assay
The cytolytic activity of genetically modified T cells was assessed in a 4-h chromium release assay on the day of injection (13). Specific 51Cr release was calculated according to the following formula: percentage of specific lysis = [(cpm of experimental release – cpm of spontaneous release)/(cpm of maximum release – cpm of spontaneous release)] x 100.
Xenograft Tumor Model for Adoptive Immunotherapy
Immunodeficient 6- to 8-wk-old SCID/beige mice (Taconic) were used. For establishment of a lung tumor model for prostate cancer, mice were challenged with 10,000 RM1 murine prostate carcinoma cells expressing hPSMA and the RFP/Rluc reporter gene by tail vein injection. Human T lymphocytes transduced with a chimeric receptor specific for either hPSMA (Pz1) or hCEA (Cz1) and cotransduced with a reporter vector were adoptively transferred into tumor-bearing mice. A total of 20 x 106 CAR-positive, GFP-positive T lymphocytes were injected via the tail vein on 2 consecutive days. Nontreated tumor-bearing mice were used as negative controls.
BLI
BLI was performed with a Xenogen IVIS 200 Imaging System (Xenogen). Mice were anesthetized with 2% isoflurane and injected with either coelenterazine (Biotium; 10 µg per animal) for RM1 RFP/Rluc-positive tumor cell imaging or D-luciferin (Xenogen; 1,500 µg per animal) for imaging of CBRluc-transduced T cells. Coelenterazine was used as the initial substrate for imaging of RLuc expression. Imaging of CBRluc expression with D-luciferin was performed 5 h later, when the Rluc signal was no longer detectable in the animals. Whole-body images were obtained and analyzed by previously described protocols (24) with Living Image 2.50 software (Xenogen). The time of acquisition varied from 1 to 3 min. The mean number of photons per square centimeter per second per steradian in the region of interest (ROI) (lungs) was determined.
Micro-CT Imaging
Micro-CT imaging was performed with a MicroCAT II CT scanner (ImTek Inc.) as described previously (13). Images were reconstructed 3 dimensionally with the Feldkamp cone-beam filtered backprojection algorithm provided by the manufacturer. Total CT acquisition time was 
10 min, and the spatial resolution was 250 µm. Visualization and analyses of CT images were performed with Amira software (Konrad-Zuse-Zentrum für Informationstechnik Berlin and Indeed Visual Concepts GmbH). Heart and blood vessel contrast was achieved via intravenous administration of Fenestra VC (Alerion Biomedical Inc.).
PET Imaging with 18F-FEAU
No-carrier-added 18F-FEAU was prepared as described previously (25). Mice were anesthetized with 2% isoflurane. PET was performed at 2 h after intravenous administration of 7.4 MBq (200 µCi) of 18F-FEAU per animal with a 3-dimensional (3D) small-animal R4 microPET scanner (Concorde Microsystems) (energy window of 350–750 keV and timing window of 6 ns) (26). The duration of the emission scans was 10 min, with 106 events per acquisition. The list-mode data were sorted into sinograms by Fourier rebinning and reconstructed with an ordered-subset expectation maximization iterative reconstruction algorithm without attenuation or scatter correction. Radioactivity concentrations in tumors (percentage injected dose per gram of tissue [%ID/g]) were measured from the PET images with ASIPro software (Concorde Microsystems) by use of a system calibration factor determined from an 18F-filled mouse-size phantom. Quantification of the specific signal above the background was performed by calculating the ratio of %ID/g in the ROI (lungs) to %ID/g in a background region (15). Micro-CT/small-animal PET fusion images were produced and analyzed with ASIPro software. Registration of the micro-CT and small-animal PET images was performed manually by use of a rigid transform derived by interactive rotation and translation of the 3D small-animal PET and micro-CT image sets with respect to one another.
Histologic Analysis
Histopathologic examination of the specimens was performed as described previously (27). For immunohistochemical analysis, lung tissues were preserved in 4% paraformaldehyde. For immunofluorescence analysis, freshly isolated tumor tissue was frozen in optimum cutting temperature compound (Sakura Finetek). The immunohistochemical reaction was carried out with paraffin-embedded specimens and the following monoclonal antibodies: anti-hPSMA 13d6 (kindly provided by Dr. P. Gregor, Memorial Sloan-Kettering Cancer Center [MSKCC]), anti-hCD8 (Dako), anti-GFP (Invitrogen), and secondary biotinylated horse antimouse IgG (BMK-2202; Vector). For immunofluorescence analysis, the specimens were incubated with mouse anti-hPSMA monoclonal antibody 7C12 (kindly provided by Dr. Polly Gregor) and anti-GFP rabbit polyclonal IgG (sc-8334; Santa Cruz Biotechnology) and then with secondary antimouse antibodies Alexa Fluor 555 and Alexa Fluor 488 (Invitrogen), respectively. Microscopy of live GFP and RFP fluorescence in selected animals was performed on the days of PET. For the visualization of nuclei, 4,6-diamidino-2-phenylindole was used. Immunofluorescence was assessed with a fluorescence microscope (Axioplan-2; Carl Zeiss Microimaging Inc.).
Statistics
Data were analyzed with GraphPad Prism (version 4) software (GraphPad Software). A Mann–Whitney test was applied for comparing 2 independent groups. P values of <0.05 were considered to be statistically significant. Survival analysis was performed with the Kaplan–Meier method, and curves were compared in a log-rank test. Linear regression analysis was used to fit the quantitative small-animal PET data to days after tumor cell injection.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONReferences |
|---|
BLI of T-Cell Distribution and Persistence
RM1 tumor cells expressing hPSMA and RFP/Rluc were injected 7 d before T-cell adoptive transfer. Rluc expression was near background levels in both (Pz1-treated and Cz1-treated) groups on the day of T-cell injection. In vivo tracking of the adoptively transferred T cells, transduced with either Pz1 or Cz1 CAR and the nTGCBR reporter gene, was performed by BLI of CBRluc expression in T cells. Different substrates were used for the imaging of Rluc and CBRluc expression; therefore, we were able to image Rluc and CBRluc in the same animals on the same days without signal interference.
Dynamic BLI of T-cell localization and persistence and tumor progression in 2 representative animals from each group is shown in Figure 1. The highest signal intensity on the first day (+1) after T-cell injection was detected in the lungs, liver, and spleen in both groups. The bioluminescence signal intensity in the lungs gradually decreased by day +6 in both groups. However, the bioluminescence signal intensity in the lungs was significantly higher on days +1, +2, and +3 after T-cell injection in the Pz1-treated group than in the Cz1-treated group (P = 0.008, P = 0.031, and P = 0.008, respectively) (Fig. 2). We observed an increase in the bioluminescence signal intensity in the spleens in the Pz1-treated group on days +2 and +3 after T-cell injection, with a subsequent decline on day +6 (data not shown).
|
|
Quantitative Small-Animal PET Imaging of Genetically Modified T Lymphocytes
In another experiment, T cells transduced with either Pz1 or Cz1 CAR and the HSV1tk/GFP reporter gene were administered via tail vein injection 10 d after mice were challenged with hPSMA- and RFP/Rluc-positive RM1 cells. By that time, extensive tumors were detectable by BLI in all animals. A separate group of tumor-free animals was injected with Pz1- and HSV1tk/GFP-positive T cells. Small-animal PET imaging of 18F-FEAU accumulation was performed on the day of T-cell injection (day 0) and was repeated on days +3 and +6 thereafter. By calculating the ratio of the %ID/g for 18F-FEAU in the ROI (lungs) to the %ID/g in the region with background levels of radiotracer accumulation (muscle), we were able to avoid variability between images over time. The %ID/g values in the regions with background levels of tracer accumulation were comparable in all groups of animals at all time points (0.069 ± 0.032 [mean ± SD] %ID/g). T cells were detected diffusely in the lungs in both the Pz1-treated group and the Cz1-treated group on the day of T-lymphocyte injection (day 0) (Fig. 3A). On the day of T-cell injection, focus (lungs)-to-background ratios were comparable in both (Pz1-treated and Cz1-treated) groups (Fig. 3B). There was no 18F-FEAU accumulation in nontreated tumor-bearing mice (negative controls). In contrast, by day +3, we were able to detect distinct 18F-FEAU accumulation in the lungs only in tumor-bearing animals injected with hPSMA-specific T cells (Fig. 3A). The focus-to-background ratio was significantly higher in this group than in the group injected with hCEA-specific T lymphocytes (2.16 ± 0.32 vs. 1.16 ± 0.07; P = 0.024) (Fig. 3B). On day +6 after T-cell transfer, there was no 18F-FEAU accumulation in the lungs in either group. These results were confirmed in at least 2 independent experiments.
|
Detailed Analysis of Small-Animal PET/Micro-CT Fusion Imaging
Micro-CT imaging with vasculature contrast enhancement revealed an established nodular pattern of tumor progression. A detailed analysis of small-animal PET/micro-CT coregistration was performed by estimating the levels of 18F-FEAU accumulation in the regions of intrathoracic tumors, large vessels, and lung parenchyma. The images of one representative animal from each tumor-bearing group on day +3 are shown in Figure 4. Higher levels of accumulation of 18F-FEAU were found by small-animal PET imaging at the sites corresponding to the tumor foci on micro-CT images for the Pz1-treated animal (Fig. 4A, boxes 1, 2, 4, and 5) but not in the lung parenchyma or vessels (Fig. 4A, boxes 3 and 6). To address whether T-cell accumulation was specific, we analyzed small-animal PET/micro-CT data from Cz1-treated and nontreated animals; no differences in the levels of 18F-FEAU accumulation in various thoracic regions were found in these animals (Figs. 4B and 4C). A quantitative analysis of 18F-FEAU accumulation is shown in Table 1.
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONReferences |
|---|
The immunotherapeutic efficiency of genetically modified T lymphocytes was recently shown in several animal tumor models (8,13,28–30). Gade et al. demonstrated successful eradication of hPSMA-positive RM1 murine prostate carcinoma tumors in lungs with human T lymphocytes transduced with anti-hPSMA CAR in a mouse model (13). It is noteworthy that RM1 is a very aggressive tumor, which forms nodules primarily in the lungs and later in the peripheral lymph nodes and which leads to the death of animals within 2–3 wk after tail vein inoculation of 5,000–10,000 tumor cells. In our initial experiments, we made numerous attempts to image Pz1-grafted T cells injected immediately after hPSMA-positive RM1 tumor cell administration, but we were not able to confirm the prolonged persistence of T cells in the lungs by small-animal PET. A short period of time between tumor cell inoculation and T-cell injection could result in the persistence of a small portion of injected T cells that engage tumor targets. In the experiments reported here, we explored the feasibility of imaging hPSMA-specific T cells targeting well-established hPSMA-positive lung tumors. The results of these experiments showed that in this systemic model, T cells persisted at tumor sites in an antigen-specific manner and had a substantial impact on animal survival.
The introduction of optical (28,30,31) or radionuclide-based (14,17) reporter genes into tumor and immune cells permits new insights into the biology of tumor–immune cell interactions with excellent temporal and spatial characteristics. Our study is the first to combine both therapeutic (CAR) and PET components through the use of genetic modifications of primary human T lymphocytes. We showed that the cotransduction of CAR-expressing T cells with a nuclear reporter gene did not impair CAR-mediated cytotoxicity and allowed for reliable PET of T lymphocytes during the course of adoptive immunotherapy. This approach potentially can be translated into clinical practice to permit the noninvasive monitoring of adoptive cell therapies with genetically modified T lymphocytes.
In our initial experiments, we used CBRluc to track the migration and persistence of genetically modified human T lymphocytes grafted with anti-hPSMA or anti-hCEA CAR. Tumor progression was monitored by Rluc BLI in the same animals. We were able to demonstrate a sustained Pz1-positive T-cell bioluminescence signal in lung tumors that was significantly higher on days +1–3 after T-cell injection than it was in the Cz1-treated control animals. We observed a transient accumulation of T cells in the spleen on days +1–3 and a subsequent decline by day +6 in the Pz1-treated animals. This latter observation may imply a possible redistribution of human T cells in a xenogeneic environment, the mechanisms for which are yet to be elucidated.
We were not able to correlate T-cell bioluminescence signal intensity in each animal with tumor progression and overall survival under these conditions. We found that despite reliable optical imaging of tumor progression, BLI failed to assess small differences in T-cell dynamics between individual animals. Two-dimensional imaging does not allow for precise quantitative analysis of T-cell persistence because of a high degree of variation in bioluminescence signal intensity that substantially depends on lymphocyte localization (such as depth or distribution in a tissue volume), tissue attenuation, and nonequivalent animal body positioning during different imaging sessions. Nevertheless, BLI still can be used as a cost-effective and high-throughput imaging method for the longitudinal assessment of T-cell distributions in vivo.
In contrast to photon emission, 
-arrays are not substantially influenced by tissue thickness, localization of cells in the ROI, or the position of the animal. To circumvent this limitation of BLI, we took advantage of the HSV1tk nuclear reporter gene for in vivo labeling of T cells with 18F-FEAU and quantitative PET. This thymidine-based, short-lived (18F half-life, 2 h) radiotracer is suitable for high-contrast repetitive PET because of its fast clearance and low background accumulation in normal, nontransduced tissues (25,32). According to the studies of Zanzonico et al., HSV1tk-transduced T cells can be labeled to at least 1,300 cGy without any loss of target cell cytotoxicity or specificity (33); this value is approximately 17 times greater than the total absorbed dose estimate based on cell-level dosimetry for the 18F-FEAU study performed (Pat Zanzonico, written communication, 2006). Therefore, HSV1tk reporter gene imaging with 18F-FEAU can be safely incorporated into immunotherapy protocols with CAR for patients with disseminated prostate cancer.
Our experiments were mainly focused on quantitative small-animal PET imaging of T-cell localization and persistence. Small-animal PET data indicated that in our model, only antigen-specific lymphocytes could persist for up to 3 d at tumor-positive sites, a finding that was confirmed by histologic analysis (Supplemental Fig. 4). 18F-FEAU accumulation on day +3 was negatively correlated with tumor bioluminescence signal intensities on day +6 in each animal in the Pz1-treated group (Supplemental Fig. 3). A nearly linear relationship was observed between 18F-FEAU accumulation on day +3 and overall survival in individual animals in the Pz1-treated group (Fig. 3C). These findings confirmed that nuclear imaging techniques allow for quantitative measurements of genetically modified T-cell localization and persistence at tumor sites over time and can be used for the prediction of tumor response and therapy outcome.
We did not anticipate any long-term remissions in our studies, because the treatment was applied to animals with a large tumor burden. It is important that the first generation of CAR (without costimulatory motifs) was used in this study. The rapid decline in PET signal intensities can be explained by the lack of costimulation and exogenous cytokine support required for prolonged T-cell survival (8,34). Therefore, we observed only a delay in tumor progression and not complete tumor rejection.
Modern radiologic examination requires precise colocalization of anatomic imaging and functional imaging. PET/CT is routine clinical practice in many medical centers. We showed the feasibility of multimodality imaging for 3D colocalization of tumor sites and reporter gene–expressing, CAR-positive T lymphocytes in our systemic tumor model. Detailed analysis showed that the localization of CAR-positive tumor-specific T cells corresponded to the areas of established tumors but not to normal lung tissue or vessel structures. PET allowed for precise quantitation of T-cell distributions at tumor sites. This key feature makes PET the method of choice for the noninvasive monitoring of adoptive immunotherapy with tumor-targeted genetically modified T lymphocytes in cancer patients.
|
CONCLUSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONReferences |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
|
References |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONReferences |
|---|
Related articles in JNM:
This article has been cited by other articles:
|
|
 |
|
  V. Ponomarev Nuclear Imaging of Cancer Cell Therapies J. Nucl. Med., July 1, 2009; 50(7): 1013 - 1016. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| JOURNAL OF NUCLEAR MEDICINE TECHNOLOGY | THE JOURNAL OF NUCLEAR MEDICINE |
