64Cu-DOTA-Anti-CTLA-4 mAb Enabled PET Visualization of CTLA-4 on the T-Cell Infiltrating Tumor Tissues

Cytotoxic T lymphocyte-associated antigen-4 (CTLA-4) targeted therapy by anti-CTLA-4 monoclonal antibody (mAb) is highly effective in cancer patients. However, it is extremely expensive and potentially produces autoimmune-related adverse effects. Therefore, the development of a method to evaluate CTLA-4 expression prior to CTLA-4-targeted therapy is expected to open doors to evidence-based and cost-efficient medical care and to avoid adverse effects brought about by ineffective therapy. In this study, we aimed to develop a molecular imaging probe for CTLA-4 visualization in tumor. First, we examined CTLA-4 expression in normal colon tissues, cultured CT26 cells, and CT26 tumor tissues from tumor-bearing BALB/c mice and BALB/c nude mice by reverse transcription polymerase chain reaction (RT-PCR) analysis and confirmed whether CTLA-4 is strongly expressed in CT26 tumor tissues. Second, we newly synthesized 64Cu-1,4,7,10-tetraazacyclododecane-N,N′,N″,N‴-tetraacetic acid-anti-mouse CTLA-4 mAb (64Cu-DOTA-anti-CTLA-4 mAb) and evaluated its usefulness in positron emission tomography (PET) and ex-vivo biodistribution analysis in CT26-bearing BALB/c mice. High CTLA-4 expression was confirmed in the CT26 tumor tissues of tumor-bearing BALB/c mice. However, CTLA-4 expression was extremely low in the cultured CT26 cells and the CT26 tumor tissues of tumor-bearing BALB/c nude mice. The results suggested that T cells were responsible for the high CTLA-4 expression. Furthermore, 64Cu-DOTA-anti-CTLA-4 mAb displayed significantly high accumulation in the CT26 tumor, thereby realizing non-invasive CTLA-4 visualization in the tumor. Together, the results indicate that 64Cu-DOTA-anti-CTLA-4 mAb would be useful for the evaluation of CTLA-4 expression in tumor.


Introduction
Cancer is a complex mixture of host and tumor cells. Whereas the human body has the ability to produce an anti-tumor immune response, cancers develop multiple strategies to evade the host immune system [1]. Cytotoxic T lymphocyte-associated antigen-4 (CTLA-4), also known as cluster of differentiation 152 (CD152), is one of the most important molecules that are involved in the downregulation of the immune system and the anti-tumor response. CTLA-4 is expressed predominantly on the surface of two major subsets of CD4 + T cells: regulatory T cells (Tregs) and activated CD4 + effector cells, and activated CD8 + effector T cells [2,3]. In addition, recent research showed that various tumor cells also express CTLA-4 [4].
CTLA-4 targeted therapy augments endogenous response to tumor cells, thereby leading to tumor cell death when utilized on its own or with other therapeutic interventions [3]. It is for this reason that CTLA-4 has attracted attention as a target molecule of cancer immunotherapy [5]. Fully human anti-CTLA-4 monoclo-nal antibodies (mAbs), ipilimumab and tremelimumab, were developed for the treatment of cancer patients. Ipilimumab is the first drug to demonstrate survival benefits in metastatic melanoma patients, and was approved by the US Food and Drug Administration (FDA) for the treatment of advanced melanoma in 2011. Pre-clinical and clinical trials of anti-CTLA-4 mAbs have been conducted for the treatment of other cancers, including colon, breast, lung, ovarian, and prostate cancers [3,6].
Although CTLA-4-targeted therapy is an attractive method for the treatment of various cancers, the therapy is beset by several problems. First, the enhanced T cell response by the CTLA-4 blockade frequently produces autoimmune-related adverse effects, such as rash, diarrhea, colitis, hepatitis, and hypophysitis [7,8]. A superagonist antibody for CD28 (TGN1412), which directly stimulates T cells, caused life-threatening inflammatory reactions in a London clinical trial [9]. Extreme precaution must be taken when CTLA-4-targeted antibodies are used for the treatment because CTLA-4 is an antagonist of CD28-ligand interactions [10]. Second, antibody drugs are extremely expensive. One treatment course of ipilimumab in the United States consists of four doses at US$30,000 per dose [2,11]. Clearly, there is an urgent need to develop a method to screen patients for sensitivity to the CTLA-4-targeted therapy, to eliminate adverse effects brought about by ineffective therapy and reduce unnecessary financial burden in non-sensitive patients. The identification of CTLA-4 expression in tumor prior to molecular targeted therapy would lead to evidence-based and cost-efficient medical care.
Biopsy is principally conducted to evaluate the expression of molecules of interest. However, it is an invasive and stressful procedure. Moreover, biopsy evaluates the expression of target molecules only in a localized region of the tumor. Thus, it is difficult to acquire information of a patient's sensitivity to a molecular targeted drug for tumors existing in whole body.
Molecular imaging can provide molecular information of the whole body in a noninvasive manner and be used for the determination of sensitivity to antibody drugs. Tumor imaging probes for human epidermal growth factor receptor 2 (HER2) [12][13][14], epidermal growth factor receptor (EGFR) [15][16][17][18], and vascular endothelial growth factor (VEGF) [19,20], which are the bevacizumab, respectively, have been developed. The expression of those molecules in tumor was detected with their respective probes by positron emission tomography (PET) or single photon emission computed tomography (SPECT). However, to our knowledge, a molecular imaging probe that targets CTLA-4 has yet to be developed.
In this study, we aimed to develop a molecular imaging probe for CTLA-4 visualization in tumor. First, CTLA-4 expression was examined in CT26 tumor tissues and cultured CT26 cells by reverse transcription polymerase chain reaction (RT-PCR) analysis. Second, we newly developed 64 Cu-1,4,7,10-tetraazacyclododecane-N,N9,N0,N--tetraacetic acid (DOTA)-anti-mouse CTLA-4 mAb by introducing DOTA groups to anti-mouse CTLA-4 mAb and subsequent radiolabeling with 64 Cu. The utility of 64 Cu-DOTA-anti-CTLA-4 mAb as an imaging probe was assessed by PET imaging and ex-vivo biodistribution analysis. We prepared tumor-bearing mice by syngeneic implantation of CT26 cells (mouse colon tumor cell line) to BALB/c mice for PET imaging. Immune-deprived mice bearing human tumor cell lines were not used because T cells might be responsible for the CTLA-4 expression in the tumor tissues.

Materials and Methods
Cell culture CT26 was purchased from American Type Culture Collection and cultured in RPMI 1640 medium supplemented with 10% fetal bovine serum, 4 mM L-glutamine, 10 U/mL penicillin, and 10 mg/mL streptomycin at 37uC in a humidified atmosphere containing 5% CO 2 .

Preparation of subcutaneous tumor model mice
Female BALB/c and BALB/c (nu/nu) nude mice (4-6 weeks old) were purchased from CLEA Japan Inc. Tumor-bearing BALB/c and BALB/c nude mice were prepared by subcutaneously implanting CT26 cells (1-4610 6 cells). Investigations were initiated after receiving approval from the committee on animal experiments of Okayama University.

RT-PCR analysis
RNA extraction and cDNA synthesis were conducted by using the same methods as our previous report [21]. Total RNA was isolated from cultured cells and tissues with TRIZOL reagent (Life Technologies Co., Ltd.) and a PureLink RNA Mini Kit (Life Technologies Co., Ltd.). One microgram of total RNA was used as the template for single-strand cDNA synthesis with a Transcriptor First Strand cDNA Synthesis Kit (Roche Co., Ltd.). Analysis of mRNA expression levels was carried out with RT-PCR using TaKaRa Ex Taq (TaKaRa Co., Ltd.). The amplification of bactin is shown as internal control. Primer sequences are listed in Table S1. The amplicons were separated on agarose gel (AGAROSE I, Amresco, Inc.), stained with ethidium bromide, and visualized with a Benchtop 2UV Transilluminator (UVP, Inc.). 64 Cu-DOTA-anti-CTLA-4 mAb production Anti-mouse CTLA-4 mAb (200-500 mg) (R&D Systems, Inc.) was conjugated to DOTA-mono-N-hydroxysuccinimide ester (DOTA-mono-NHS ester; Macrocyclics, Inc.) in phosphatebuffered saline without calcium and magnesium (pH 7.5) (PBS (2)), by using a 100-fold molar excess of DOTA-mono-NHS ester. The mixture was stirred at room temperature (RT) for three hours to give the DOTA-anti-CTLA-4 antibody. The DOTA-anti-CTLA-4 antibody was purified with a PD-10 column (GE Healthcare Co., Ltd.) and an Amicon-Ultra 50 K device (Millipore Co., Ltd.). The DOTA-anti-CTLA-4 antibody was analyzed by size-exclusion high-performance liquid chromatography (SE-HPLC) using TSK-GEL Super SW3000 (Tosoh Co., Ltd.). The mobile phase of 10 mM PBS (2) containing 0.3 M NaCl was used and the flow rate was 0.35 mL/min. 64 Cu was produced by irradiating a 99.6% 64 Ni-enriched nickel target with 12 MeV protons using a cyclotron (CYPRIS-HM12, Sumitomo Heavy Industries, Ltd.). Then, 64 Cu was purified with a Muromac column (Muromachi Technos Co., Ltd.). The buffer solution of DOTA-anti-CTLA-4 mAb was replaced with 0.1 M acetate buffer (pH 6.5) three times by using an Amicon-Ultra 50 K device (Millipore Co., Ltd.). DOTA-anti-CTLA-4 mAb was radiolabeled with 64 Cu by incubating at 40uC for one hour. To remove excess 64 Cu, the buffer was replaced with 0.2 M glycine buffer by using the Amicon-Ultra 50 K device. Buffer of the purified antibody solution was replaced with PBS (2) by using the Amicon-Ultra 50 K device. The resultant solution was used for injection.
The radiochemical purity of 64 Cu-DOTA-antibodies in PBS (2) was confirmed by reversed phase radio-thin layer chromatography (TLC). This analysis was performed with a TLC aluminum sheet, RP-18 F254 S (Merck Chemicals Co., Ltd.) and methanol:water:acetic acid (4:1:1) was used as the mobile phase. TLC chromatograms were obtained by autoradiography (FLA-7000IR; GE Healthcare Co., Ltd.). 64 Cu-DOTA-isotype IgG 2A ( 64 Cu-DOTA-Control IgG) was produced in the same way as that for negative control by using rat IgG 2A isotype control (R&D Systems, Inc.).
PET imaging study 64 Cu-DOTA-anti-CTLA-4 mAb (4 mg, approximately 16 MBq) or 64 Cu-DOTA-Control IgG (4 mg, approximately 14 MBq) was intravenously administered to CT26-bearing BALB/c mice via the tail vein. Forty-eight hours after administration of the radiolabeled antibodies, probe uptake in the CT26-bearing mice was measured with a small-animal PET scanner (microPET Focus220; Siemens Medical Solutions Inc.). During PET imaging, the mice were anesthetized with 1.5% isoflurane and 1.5% N 2 O gas, and placed in the prone position. Emission data were acquired for 60 min. The acquired data were summed into sinograms and threedimensional images were reconstructed by maximum a posteriori (MAP). Coronal and sagittal images were displayed in 9186760 and 5506760 pixel formats, respectively, with a pixel size of 0.053 mm60.053 mm. The image intensity was expressed by standardized uptake value (SUV). SUV max was calculated by ASIPRO software package (Concorde Microsystems, Inc.).

Biodistribution study
Forty-eight hours after administration of 64 Cu-DOTA-anti-CTLA-4 mAb (4 mg, 1 MBq) or 64 Cu-DOTA-Control IgG (4 mg, 1 MBq), the animals were immediately sacrificed and the organs and blood were removed. The organs and blood were weighed and radioactivities were counted with a gamma counter (ARC-7001B, ALOKA Co., Ltd.). Decay-corrected uptake was expressed as the percentage of injected dose per gram and calculated as the ratio to blood or muscle for comparison of the accumulation abilities in the CT26 tumor between 64 Cu-DOTA-anti-CTLA-4 antibody and 64 Cu-DOTA-Control IgG.

Statistical analysis
SUV max data are expressed as means 6 standard deviation (SD) and other data are expressed as means 6 standard error of mean (SEM). Statistical significance was determined using the Student's t-test. The level of significance was taken as p,0.01. The tests were performed using GraphPad Prism software (GraphPad Software, Inc.).

CTLA-4 was strongly expressed in CT26 tumor tissues but not cultured CT26 cells
First, RT-PCR was carried out to examine CTLA-4 expression in CT26 tumor tissues and cultured CT26 cells, and the results were compared to those obtained with normal colon tissues (Fig. 1). CTLA-4 (amplicon length: 920 base pairs (bp)) was strongly expressed in CT26 tumor tissues compared with normal colon tissues.
On the other hand, CTLA-4 expression was extremely low in cultured CT26 cells. Moreover, Treg markers, such as forkhead box P3 (Foxp3) and folate receptor 4 (FR4) [23,24], were more strongly expressed in the tumor tissues than the normal colon tissues and the cultured CT26 cells. The expression of CD25 and CD69, which are molecules expressed on regulatory and activated T cells [23][24][25], was also increased in the tumor tissues compared to the normal colon tissues and the cultured CT26 cells. CD154, which is induced on T cells by T cell activation [26], was also more strongly expressed in the tumor tissues than the normal colon tissues and the cultured CT26 cells.

CTLA-4 and T cell marker expression was low in CT26 tumor tissues from tumor-bearing BALB/c nude mice
From the results of Fig. 1, we assumed that T cells were involved in CTLA-4 expression in the CT26 tumor tissues from the tumor-bearing BALB/c mice, and CT26 tumor tissues from the tumor-bearing BALB/c nude mice did not express CTLA-4 due to a marked decrease of T cells in those mice. Thus, we prepared two subcutaneous tumor models by syngeneic subcutaneous transplantation of CT26 into normal BALB/c mice or BALB/c nude mice, and compared CTLA-4 and T cell marker   (Fig. 2).
RT-PCR showed that CTLA-4 expression was dramatically decreased in the CT26 tumor tissues from the tumor-bearing BALB/c nude mice, compared with those from the tumor-bearing BALB/c mice. Furthermore, we confirmed that the expression of CD4 and CD8 as well as Foxp3, FR4, CD69, CD154, and CD25 was markedly decreased in the tumor tissues from the tumorbearing BALB/c nude mice.

DOTA-conjugated antibody probe was synthesized
DOTA chelators were conjugated to each mAb, as shown in Figure 3A. The chromatograms of all the mAbs showed a single peak. In addition, the retention times of anti-CTLA-4 mAb, DOTA-anti-CTLA-4 mAb, Control IgG, and DOTA-Control IgG were 9.42, 9.36, 10.57, and 10.33, respectively (Fig. 3B). The retention times of DOTA-conjugated antibodies were slightly shorter than those of the original antibodies, suggesting that anti-CTLA-4 mAb or Control IgG conjugated to DOTA and was well purified. Furthermore, MALDI-TOF-MS analysis was carried out to measure the average number of DOTA chelators that were conjugated to anti-CTLA-4 mAb or Control IgG ( Table 1) The results were consistent with those of the ex-vivo biodistribution study (Fig. 5). 64 Cu-DOTA-anti-CTLA-4 mAb showed significantly higher accumulation in the CT26 tumors than 64 Cu-DOTA-Control IgG (7.4960.32%ID/g vs. 5.8460.38%ID/g, p,0.01). Moreover, 64 Cu-DOTA-anti-CTLA-4 mAb showed higher tumor-to-blood and tumor-to-muscle ratios than 64 Cu-DOTA-Control IgG (tumor-to-blood ratio: 0.5860.03 vs. 0.4060.02, p,0.001; tumor-to-muscle ratio: 8.4860.63 and 5.3160.35, p,0.01).
In addition, CTLA-4 protein expression in the CT26 tumor was confirmed by immunohistochemical staining (Fig. S1A). CTLA-4 was weakly expressed in the normal tissues surrounding the tumor (Fig. S1B).

Discussion
CT26 is a N-nitroso-N-methylurethane-induced, undifferentiated colon carcinoma cell line and recent cancer immunotherapy studies have shown that CTLA-4 blockade reduced CT26 colon tumor size and was effective in CT26 tumor models [27,28]. Therefore, in this study, we used the CT26 cell line to prepare subcutaneous tumor models for PET imaging.
On the other hand, we found that CTLA-4 was not expressed in the cultured CT26 cells, although it was strongly expressed in the CT26 tumor tissues. Contardi et al. reported some human tumor cell lines that expressed CTLA-4 [4]. On the other hand, CTLA-4 is expressed also on CD25 + (and/or) Foxp3 + (and/or) FR4 + CD4 + Tregs, activated CD4 + effector T cells, and activated CD8 + effector T cells [2,3,24]. In addition, flow cytometry analysis by Valzasina et al. revealed that almost all CD4 + T cells in the CT26 tumor tissues expressed CD25 [35]. From those reports, we hypothesized that CTLA-4 expression in the CT26 tumor tissues regulated by T cells. To prove our hypothesis, we examined the expression of several T cell markers in normal colon tissues, CT26 tumor tissues, and cultured CT26 cells. RT-PCR analysis showed  that Treg markers and T cell activation markers were strongly expressed in the CT26 tumor tissues but not the normal colon tissues or cultured CT26 cells. Therefore, we assumed that the T cells were responsible for CTLA-4 expression in the CT26 tumor tissues. Then, we used RT-PCR to compare the expression of CTLA-4 and T cell markers in the CT26 tumor tissues of tumorbearing BALB/c mice with those of tumor-bearing BALB/c nude mice in hopes of elucidating the relationship between CTLA-4 expression in the CT26 tumor tissues and T cells. BALB/c nude mice are thymus-deficient and thus, the number of T cells is greatly reduced in those mice. Interestingly, we found that the expression of CTLA-4 as well as T cell markers was quite low in the CT26 tumor tissues from the BALB/c nude mice. The results indicated that T cells were responsible for the CTLA-4 expression.
Second, we developed a molecular imaging probe that targets CTLA-4 and examined its utility in mice bearing CTLA-4expressing CT26 tumor. The anti-CTLA-4 mAb for the imaging probe synthesis was made by using recombinant mouse CTLA-4 representing an extracellular domain of mouse CTLA-4 (Ala36-Phe161) as the immunogen. We selected this mAb for CTLA-4 imaging probe synthesis because clinically used anti-CTLA-4 mAb (ipilimumab) also recognizes the extracellular domain of CTLA-4 and a mAb probe that recognizes the extracellular domain of CTLA-4 is suited for the prediction of the efficacy and drug disposition of ipilimumab.
For CTLA-4 imaging, we conjugated anti-CTLA-4 mAb to DOTA. DOTA-conjugated mAb was prepared by reacting the nucleophilic amino group in the amino acid residue (particularly in lysine) of mAb with the electrophilic DOTA-mono-NHS ester. The binding activity of antibodies to CTLA-4 may be reduced, particularly when DOTA conjugates to lysine residues critical for mAb binding to CTLA-4 [36]. Therefore, it is necessary to check in advance the binding activities of the antibodies. We examined the binding activity of DOTA-conjugated anti-CTLA-4 mAb and compared it with that of control IgG. ELISA confirmed that the binding activity of DOTA-anti-CTLA-4 mAb was preserved for use in CTLA-4 imaging although a slight reduction (86.362.8%) was observed relative to the binding activity of the original anti-CTLA-4 mAb. The binding activity of DOTA-Control IgG was extremely low (0.360.1%) compared with that of DOTA-anti-CTLA-4 mAb. The results indicate the successful preparation of metal-chelator-conjugated anti-CTLA-4 mAb having CTLA-4 binding activity.
The choice of the positron emitter is an important factor for successful PET imaging. In our study, 64 Cu was used for labeling mAbs. 64 Cu decay generates positron emissions applicable to PET and the half-life of 64 Cu (T 1/2 = 12.7 h) is sufficiently long for imaging up to 24 to 48 h after administration to accommodate the mAb localization time. Therefore, 64 Cu has been used for the development of mAb-based radiopharmaceuticals [13,15,19]. In our preliminary biodistribution analyses, tumor uptake on 64 Cu-DOTA-anti-CTLA-4 mAb was not significantly higher than that on 64 Cu-DOTA-Control IgG at 24 h after administration of the probes (data was not shown). Therefore, we chose later time point (48 h) for PET imaging and ex-vivo biodistribution analysis.
The utility of 64 Cu-DOTA-anti-CTLA-4 mAb was examined by PET and ex-vivo biodistribution analysis. We were able to visualize CTLA-4-positive tumor by PET with 64 Cu-DOTA-anti-CTLA-4 mAb as the probe. Ex-vivo biodistribution analysis also revealed that 64 Cu-DOTA-anti-CTLA-4 mAb showed significant accumulation in the CT26 tumor compared to 64 Cu-DOTA-Control IgG. The results indicated that 64 Cu-DOTA-anti-CTLA-4 mAb is useful for the noninvasive imaging of CTLA-4 expression in tumor. Furthermore, we quantified PET images and calculated 64 Cu-DOTA-anti-CTLA-4 mAb/ 64 Cu-DOTA-Control IgG ratio in the CT26. The mean ratio of SUV max values in CT26 was 1.29 in the PET image and that in CT26 was 1.28 in the ex-vivo biodistribution analysis; further, the quantitative value of PET was similar to the ex-vivo biodistribution data. Thus, quantitativity was ensured in our PET experiment. In addition, although further investigation is needed, 64 Cu-DOTA-anti-CTLA-4 mAb could be used in the diagnosis of other types of tumor invaded by T cells, regardless of CTLA-4 expression in the tumor cells.
In conclusion, we have developed 64 Cu-DOTA-anti-CTLA-4 mAb and evaluated its potential as a new radiotracer for the noninvasive evaluation of CTLA-4 expression in tumor. Our results demonstrated that 64 Cu-DOTA-anti-CTLA-4 mAb visualized CTLA-4 expression in CT26 tumor in a noninvasive manner. Therefore, 64 Cu-DOTA-anti-CTLA-4 mAb is useful for evaluating CTLA-4 expression in the tumor. The evaluation of CTLA-4 expression in tumors using 64 Cu-DOTA-anti-CTLA-4 mAb would enable selection of patients sensitive to CTLA-4-targeted therapy, thereby eliminating the adverse effects brought about by ineffective therapy and reducing unnecessary financial burden in non-sensitive patients.