PET Imaging Biomarkers of Anti-EGFR Immunotherapy in Esophageal Squamous Cell Carcinoma Models

Epidermal growth factor receptor (EGFR) is overexpressed and considered as a proper molecular target for diagnosis and targeted therapy of esophageal squamous cell carcinoma (ESCC). This study evaluated the usefulness of PET imaging biomarkers with 64Cu-PCTA-cetuximab and 18F-FDG-PET for anti-EGFR immunotherapy in ESCC models. In vivo EGFR status and glucose metabolism by cetuximab treatment were evaluated using 64Cu-PCTA-cetuximab and 18F-FDG-PET, respectively. Therapeutic responses with imaging biomarkers were confirmed by western blot and immunohistochemistry. TE-4 and TE-8 tumors were clearly visualized by 64Cu-PCTA-cetuximab, and EGFR expression on TE-8 tumors showed 2.6-fold higher uptake than TE-4. Tumor volumes were markedly reduced by cetuximab in TE-8 tumor (92.5 ± 5.9%), but TE-4 tumors were refractory to cetuximab treatment. The SUVs in 64Cu-PCTA-cetuximab and 18F-FDG-PET images were statistically significantly reduced by cetuximab treatment in TE-8 but not in TE-4. 64Cu-PCTA-cetuximab and 18F-FDG-PET images were well correlated with EGFR and pAkt levels. 64Cu-PCTA-cetuximab immuno-PET had a potential for determining EGFR level and monitoring therapeutic response by anti-EGFR therapy. 18F-FDG-PET was also attractive for monitoring efficacy of anti-EGFR therapy. In conclusion, PET imaging biomarkers may be useful for selecting patients that express target molecules and for monitoring therapeutic efficacy of EGFR-targeted therapy in ESCC patients.


Introduction
Esophageal cancer is the sixth leading cause of cancer-related mortality worldwide [1], and it is histologically classified into esophageal squamous cell carcinoma (ESCC) and esophageal adenocarcinoma. ESCC is the major histology in Asian countries, including Japan and China [2,3]. There has been a multidisciplinary approach in surgical techniques, chemotherapy, and radiotherapy treatments used for the condition, with 5-fluorouracil, platinum agents, and taxanes among commonly used agents. However, the outcome for patients with esophageal cancer remains poor [4][5][6][7]. Therefore, novel therapeutic strategies such as molecular-targeted therapy, including small molecule inhibitors of tyrosine kinases (TKIs) and monoclonal antibodies (mAbs), are needed [8,9].

Small Animal PET Imaging
We performed immunotherapy twice in TE4 and TE-8 models. In the first experiment (n = 6 or 7/group), 64 Cu-PCTA-cetuximab was injected 2 days before cetuximab treatment and imaged at 48 h postinjection. After immuno-PET imaging, isotype control or cetuximab were treated to tumor bearing mice. After the first week of immunotherapy, 64 Cu-PCTA-cetuximab was injected into mice, and immuno-PET images were acquired at 48 h. After treatment for 28 days, mice were sacrificed and tumors were excised, which were used for immunohistochemistry and western blot analysis. In the second immunotherapy treatment (n = 3 or 4/group), 18 F-FDG-PET imaging was performed before and after cetuximab or isotype control treatment for 3 weeks. 18 F-FDG-PET imaging was done once per week.
Immuno-PET imaging of tumor-bearing mice was performed using a small animal PET scanner (microPET R4, Concorde Microsystems, Knoxville, TN, USA). 64 Cu-PCTA-cetuximab (3.0 ± 0.3 MBq, n = 3/group) was injected intravenously into the mice, and static scans were acquired for 60 min at 48 h postinjection. The acquired 3D emission list-mode data were reconstructed for imaging using Fourier rebinning and ordered subsets expectation maximization reconstruction algorithm. Images were visualized using ASIPro display software. We studied no attenuation correction because attenuation for mice is fairly small, and there is very little change in the activity profile across the mice with attenuation correction scan [36]. Quantitative data were expressed as standardized uptake value (SUV), which is defined as tissue concentration (MBq/mL)/injected dose (MBq)/the body weight (g) [35]. The tumor uptake was evaluated from SUV images by 0.5 cm 3 volume of interest (VOI), which was manually drawn. 18 F-FDG-PET imaging also used the same instrument. 18 F-FDG (7.7 ± 0.6 MBq, n = 3/group) was injected intravenously 1 h prior to scan, and static scans were obtained for 20 min. PET images were analyzed and quantified using the abovementioned procedure.

Immunohistochemistry
Following the third dose of isotype or cetuximab, the mice were sacrificed. The tumor tissues were excised, fixed in 4% paraformaldehyde, dehydrated, and embedded in paraffin. Subsequently, terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling (TUNEL) and phosphorylated Akt (pAkt) staining were carried out on tumor sections. In random 6 fields, TUNEL-positive nuclei per field were counted. The pAkt staining index (SI) was defined as the percentage of positive nuclei within the total number of nuclei.

Statistical Analysis
Quantitative data are represented as the mean ± SD, and statistical analysis was performed by one-way ANOVA or Student's t test using GraphPad Prism 5; p values of less than 0.05 were considered statistically significant.

Characterization of EGFR Expression in Esophageal Squamous Cell Carcinoma
ESCC TE-4 and TE-8 cell lines were examined for EGFR expression in RT-PCR, western blot, and flow cytometry in vitro. RT-PCR analysis revealed that EGFR mRNA were detectable in TE-4 and TE-8 cell lines (Figure 1a). The primers for EGFR and GAPDH gene sequence yielded amplification products of the expected size: 195 and 532 bp, respectively. Immunoblot was used to verify the EGFR expression level. EGFR and β-actin bands were detected in TE-4 and TE-8 cell lines (Figure 1b).
Flow cytometric analysis (Figure 1c) showed similar results as the western blot data. As determined by western blot and flow cytometry, the TE-8 cell line showed a relatively higher level of EGFR than the TE-4 cell line. TE-8 cells represented higher mean fluorescent intensity (MFI, 577.5) than TE-4 cells (MFI, 53.8).
Cells 2018, 7, x FOR PEER REVIEW  5 of 15   determined by western blot and flow cytometry, the TE-8 cell line showed a relatively higher level of  EGFR than the TE-4 cell line. TE-8 cells represented higher mean fluorescent intensity (MFI, 577.5) than TE-4 cells (MFI, 53.8). To determine the cell-surface EGFR expression using 64 Cu-PCTA-cetuximab, cell binding assay was performed ( Figure 2). A431 and U87-MG cell lines were used as a positive and a negative control in this study, respectively. Assuming that EGFR expression level of A431 was 100%, TE-8 cells (87.0 ± 1.8%) had higher EGFR expression, while TE-4 (18.7 ± 0.3%) cells had relatively low EGFR expression. These results indicated that binding of 64 Cu-PCTA-cetuximab represented the level of EGFR expression on the cells.

Cytotoxicity of Cetuximab in ESCC Cells
We examined the antiproliferative effect of cetuximab against high-EGFR-expressing cell line TE-8 and low-EGFR-expressing cell line TE-4. Cetuximab-induced cell growth inhibition was found in high-EGFR-expressing TE-8 cells, while it was minimal in low-EGFR-expressing TE-4 ( Figure 3). TE-8 cells showed 57.2 ± 3.9% (10 μg/mL) and 44.4 ± 7.5% (50 μg/mL) viability compared to control; however, TE-4 cells still kept more than 80% viability even with 50 μg/mL of cetuximab. To determine the cell-surface EGFR expression using 64 Cu-PCTA-cetuximab, cell binding assay was performed ( Figure 2). A431 and U87-MG cell lines were used as a positive and a negative control in this study, respectively. Assuming that EGFR expression level of A431 was 100%, TE-8 cells (87.0 ± 1.8%) had higher EGFR expression, while TE-4 (18.7 ± 0.3%) cells had relatively low EGFR expression. These results indicated that binding of 64 Cu-PCTA-cetuximab represented the level of EGFR expression on the cells.  To determine the cell-surface EGFR expression using 64 Cu-PCTA-cetuximab, cell binding assay was performed ( Figure 2). A431 and U87-MG cell lines were used as a positive and a negative control in this study, respectively. Assuming that EGFR expression level of A431 was 100%, TE-8 cells (87.0 ± 1.8%) had higher EGFR expression, while TE-4 (18.7 ± 0.3%) cells had relatively low EGFR expression. These results indicated that binding of 64 Cu-PCTA-cetuximab represented the level of EGFR expression on the cells.

Figure 2.
In vitro cell binding assay of 64 Cu-PCTA-cetuximab. A431 and U87-MG cells were used as positive and negative control, respectively. Data presented as the percentage (%) of relative cell-bound radioactivity to A431 cells.

Therapeutic Effects of Cetuximab in ESCC Tumors
Antitumor effects of cetuximab were assessed in TE-4 and TE-8 xenograft models (Figure 4). In the isotype group, the growth rate of TE-4 tumors was faster than that of TE-8 tumors. There was no difference in tumor growth between isotype and cetuximab treatment in TE-4 model (Figure 4a).

Therapeutic Effects of Cetuximab in ESCC Tumors
Antitumor effects of cetuximab were assessed in TE-4 and TE-8 xenograft models (Figure 4). In the isotype group, the growth rate of TE-4 tumors was faster than that of TE-8 tumors. There was no difference in tumor growth between isotype and cetuximab treatment in TE-4 model (Figure 4a). TE-8 tumor growth was inhibited after second administration of cetuximab, and the TE-8 tumor markedly regressed with cetuximab treatment (92.5 ± 5.9% tumor reduction, p < 0.001); however, the TE-8 tumor volume continuously increased with isotype treatment (Figure 4b). TE-8 tumor volume in the cetuximab treatment group showed a statistically significant difference after four days (p < 0.01). Cetuximab treatment was well tolerated in both TE-4 and TE-8 xenograft models, and no apparent body weight loss was observed ( Figure S1). . Cytotoxicity of cetuximab on TE-4 (low EGFR expression) and TE-8 (high EGFR expression) cells after five days of cetuximab treatment at each dose. The viable cell number was counted by ADAM cell counter. Data presented as the percentage (%) of viable cell number compared to the control.

Therapeutic Effects of Cetuximab in ESCC Tumors
Antitumor effects of cetuximab were assessed in TE-4 and TE-8 xenograft models (Figure 4). In the isotype group, the growth rate of TE-4 tumors was faster than that of TE-8 tumors. There was no difference in tumor growth between isotype and cetuximab treatment in TE-4 model (Figure 4a). TE-8 tumor growth was inhibited after second administration of cetuximab, and the TE-8 tumor markedly regressed with cetuximab treatment (92.5 ± 5.9% tumor reduction, p < 0.001); however, the TE-8 tumor volume continuously increased with isotype treatment (Figure 4b). TE-8 tumor volume in the cetuximab treatment group showed a statistically significant difference after four days (p < 0.01). Cetuximab treatment was well tolerated in both TE-4 and TE-8 xenograft models, and no apparent body weight loss was observed ( Figure S1).

Immuno-PET Imaging of Cetuximab-Induced Antitumor Activity
To evaluate the potential of 64 Cu-PCTA-cetuximab as an immuno-PET imaging agent for determining EGFR level, we performed immuno-PET imaging in TE-4 or TE-8 xenograft models. 64 Cu-PCTA-cetuximab immuno-PET images (n = 3) were obtained for each animal before treatment and after one week of treatment in TE-4 and TE-8 xenograft models. PET images clearly showed the uptake of 64 Cu-PCTA-cetuximab in TE-4 and TE-8 tumors at 48 h after injection (Figure 5a,b). The SUV of 64 Cu-cetuximab before treatment was 2.5-fold higher in TE-8 (4.6 ± 0.7) than TE-4 tumors (1.8 ± 0.3). Immuno-PET images were consistent with the data of flow cytometric assay and cell binding assay. Next, we assessed the tumor response to cetuximab depending on the level of EGFR expression. In TE-4 tumors, there was no statistically significant difference in SUV between pretreated group and post-treated group in isotype-treated (1.8 ± 0.4 vs. 1.7 ± 0.3, p = 0.5448) or cetuximab-treated mice (1.8 ± 0.4 vs. 1.7 ± 0.5, p = 0.8473) (Figure 5a,c). For TE-8 tumors, there was statistically significant SUV reduction (65.7%) in cetuximab-treated mice compared to isotype control (4.2 ± 0.7 vs. 1.4 ± 0.3, p < 0.001) (Figure 5b,d). The SUV of TE-8 tumors in isotype-treated mice seemed to trend downward when compared to the pretreated group; the trend was not significant (p = 0.0917). These data indicate that cetuximab-induced antitumor activity could be monitored by 64 Cu-PCTA-cetuximab, immuno-PET agent, which represents the relative level of EGFR expression in ESCC tumors.

Cetuximab-Induced Apoptosis and Signaling Blockade
TUNEL assay was performed to evaluate apoptosis induced by cetuximab treatment in TE-4 and TE-8 tumors. In both tumors, apoptosis was almost unobservable after isotype treatment. TUNEL positive apoptotic cells were more visualized in TE-4 tumor with cetuximab treatment compared to isotype treatment, but it was statistically insignificant (p = 0.1085). Apoptotic cells markedly increased in cetuximab-treated TE-8 tumors compared to isotype (p < 0.001), indicating that cetuximab treatment increased apoptosis in high-EGFR-expressing TE-8 tumor (Figure 7a,c). pAkt immunoreactivity was significantly reduced in TE-8 tumor with cetuximab treatment

Cetuximab-Induced Apoptosis and Signaling Blockade
TUNEL assay was performed to evaluate apoptosis induced by cetuximab treatment in TE-4 and TE-8 tumors. In both tumors, apoptosis was almost unobservable after isotype treatment. TUNEL positive apoptotic cells were more visualized in TE-4 tumor with cetuximab treatment compared to isotype treatment, but it was statistically insignificant (p = 0.1085). Apoptotic cells markedly increased in cetuximab-treated TE-8 tumors compared to isotype (p < 0.001), indicating that cetuximab treatment increased apoptosis in high-EGFR-expressing TE-8 tumor (Figure 7a,c). pAkt immunoreactivity was significantly reduced in TE-8 tumor with cetuximab treatment compared to isotype (Figure 7b,d).
In TE-4 tumors, pAkt positivity slightly increased with cetuximab treatment compared to isotype; however, the difference was nonsignificant (p = 0.8614). pAkt immunohistochemical staining result was consistent with western blot analysis of excised tumors. Expression level of EGFR, pEGFR, and pAkt in cetuximab-treated TE-8 tumor markedly decreased compared to isotype-treated group. Exceptionally, pEGFR and pAkt expression level in TE-4 tumor slightly decreased and increased, respectively, with cetuximab treatment. pAkt immunohistochemical staining and western blot results suggested that cetuximab treatment inhibited the PI3K/Akt pathway on TE-8 cells.
Cells 2018, 7, x FOR PEER REVIEW 10 of 15 compared to isotype (Figure 7b,d). In TE-4 tumors, pAkt positivity slightly increased with cetuximab treatment compared to isotype; however, the difference was nonsignificant (p = 0.8614). pAkt immunohistochemical staining result was consistent with western blot analysis of excised tumors. Expression level of EGFR, pEGFR, and pAkt in cetuximab-treated TE-8 tumor markedly decreased compared to isotype-treated group. Exceptionally, pEGFR and pAkt expression level in TE-4 tumor slightly decreased and increased, respectively, with cetuximab treatment. pAkt immunohistochemical staining and western blot results suggested that cetuximab treatment inhibited the PI3K/Akt pathway on TE-8 cells.

Discussion
In the current study, therapeutic efficacy using cetuximab immunotherapy in ESCC model was associated with EGFR expression level measured by 64 Cu-cetuximab immuno-PET imaging. Volumetric reduction and metabolic 18 F-FDG uptake change were evaluated in high-EGFR-expressing TE-8 tumor model. Based on these results, we note that 64 Cu-PCTA-cetuximab immuno-PET imaging may be useful for evaluating the level of EGFR expression on ESCC tumors in vitro and in vivo and determining change of EGFR expression with cetuximab treatment. 18 F-FDG-PET was also useful for assessing the therapeutic response to cetuximab treatment in ESCC tumors. Immuno-PET can provide noninvasive, quantitative assessment of specific molecular targets, predict whether the patients are likely to benefit from targeted therapy, and also permit serial monitoring of therapeutic efficacy in the whole level [38].
The receptor expression level alone does not necessarily predict the therapeutic outcome due to lack of information about functional state of the receptor. Thus, pharmacodynamic (PD) biomarker is needed for early identification of responders by quantification of molecular and functional effects and prediction of therapeutic outcomes. Our data indicated that 64 Cu-PCTA-cetuximab uptake in TE-8 tumors was significantly reduced (65.9% SUV reduction) after one week of cetuximab treatment compared to baseline, suggesting 64 Cu-PCTA-cetuximab may be useful as an early-phase imaging biomarker for monitoring changes in EGFR expression level (Figure 5c,d). Decreased EGFR expression was also detected by western blot (Figure 7e). To our knowledge, this is the first report to validate the change in EGFR expression in ESCC tumor models with cetuximab treatment.
Cetuximab inhibits activation of intracellular signaling, including the Ras-Raf-MAPK and phosphoinositide 3-kinase (PI3K)/Akt pathway that are involved in cell growth, survival, and glucose transport [39]. Thus, we postulated that changes in 18 F-FDG uptake could indicate modulation of EGFR function. 18 F-FDG-PET showed 21% reduction in the SUV of TE-8 tumors after one week of cetuximab treatment. After three weeks of cetuximab treatment, 18 F-FDG uptake in TE-8 tumors decreased by 66.8% compared to the isotype treatment. These changes were correlated with the reduction in tumor volume (Figure 5b) and pAkt signal intensity (Figure 7e) of TE-8 tumor. These data are also in accordance with the results reported by Niu et al., who presented a significant decrease in 18 F-FDG uptake after five days of cetuximab treatment in SSC1 xenografts [40]. Berger et al. also evaluated metabolic response during cetuximab therapy in correlation with clinical response in metastatic colorectal cancer using 18 F-FDG-PET/CT [41]. On the other hand, preclinical studies using breast cancer cell line for anti-HER2 therapy have shown no significant difference in 18 F-FDG metabolism following treatment with Hsp90 inhibitor 17-AAG [42] and trastuzumab [43]. Targeted therapy may not always have any observable effect on 18 F-FDG uptake; however, this study shows the feasibility of 18 F-FDG-PET as an early PD biomarker of anti-HER1 therapy in ESCC xenograft models.
As shown in in vitro cytotoxicity assay (Figure 3), TE-4 tumor growth was not affected by cetuximab treatment in vivo ( Figure 2). The flexibility of tumors to use alternative receptors to activate downstream signaling pathways regulating cetuximab resistance has been reported [44][45][46]. Wheeler et al. reported that acquired resistance to cetuximab reflected dysregulation of EGFR internalization/degradation and subsequent EGFR-dependent activation of HER2 and HER3 in non-small cell lung cancer cell line. Similarly, Yonesaka et al. demonstrated that activation of HER2 signaling led to persistent extracellular signal-regulated kinase 1/2 signaling and consequently to cetuximab resistance. In western blot assay, TE-4 cell is HER2 expressing cell line (data not shown) and pAkt expression level in TE-4 tumor slightly increased with cetuximab treatment compared to isotype treatment (Figure 7b,e). Therefore, aberrant HER2 signaling could be used to activate a bypass signaling pathway in TE-4 tumors. However, the relationship between HER2 expression and cetuximab resistance mechanism need to be assessed in further studies.
Treatment with TKIs, such as gefitinib and lapatinib, has been shown in vitro and in vivo to inhibit downstream effector pathway and proliferation of EGFR-and/or HER2-overexpressing ESCC [47,48]. Thus, anti-HER2 targeted therapy is also an attractive approach to treat ESCC. Inhibition of HER2 may play a potential role in antiproliferation of tumor, and inhibition of EGFR and combination therapy with TKIs could be useful in EGFR-or HER2-overexpressing ESCC.
There were some limitations to the current study. First, because two ESCC cell line model was used, expanded preclinical and clinical research are required in various ESCC animal models and clinical patients with tumors expressing varying level of EGFR to validate the association between 64 Cu-PCTA-cetuximab accumulation in tumor lesion and prognosis after cetuximab immunotherapy and to establish PET signal cutoff value for target selection as imaging biomarker. Second, analysis of genetic mutation of EGFR in ESCC was absent. Although driver gene mutations have not been detected in ESCC, the somatic mutation rate in ESCC is relatively high compared to other solid tumors, and it may attenuate the therapeutic effect to EGFR-targeted therapy [49]. Third, the possibility of the low uptake of 64 Cu-PCTA-cetuximab in TE-8 model by cetuximab immunotherapy exists. However, there was no reduced uptake of 64 Cu-PCTA-cetuximab in TE-4 tumor with intermediate EGFR expression by cetuximab treatment (Figure 5a,c). Luo et al. [50] reported systemic pharmacokinetic data of cetuximab in preclinical EGFR-expressing tumor model and nude mice without tumor. The plasma half-life of cetuximab at 1 mg dose of intravenous injection was 39.6 h, T max was 3 h, and C max was 407.6 g/mL in nude mice without tumor. However, the plasma concentration of cetuximab at 1 mg dose of intraperitoneal injection was 153.6 g/mL at 6 h and 17.7 g/mL at 24 h in tumor-bearing mice. In case of tumor-bearing mice, plasma clearance was faster than nude mice without tumor. Our experiment was also performed in EGFR-expressing tumor. Therefore, the plasma clearance of i.v.-injected cetuximab for treatment would have shown similar patterns as the previous study. Furthermore, TE-4 tumor has moderate EGFR expression, which is easily blocked by treatment dose of cetuximab. However, in our experimental design, TE-4 tumor uptake of 64 Cu-PCTA-cetuximab was maintained in cetuximab-treated TE-4 tumor at a similar level as that before cetuximab treatment.
In conclusion, this study shows that 64 Cu-PCTA-cetuximab immuno-PET and 18 F-FDG-PET can be used as potential pharmacodynamic PET imaging biomarkers for the therapeutic response to cetuximab treatment in ESCC tumors. 64 Cu-PCTA-cetuximab immuno-PET imaging biomarker may be useful for selecting patients that express target molecules and monitoring therapeutic efficacy of molecular targeted therapy in clinical trial.