Overexpression of epidermal growth factor receptor (EGFR) is common in colorectal cancer. However, cetuximab as an EGFR-targeting drug is useful only for a subset of patients and currently no single predictor other than V-Ki-ras2 Kirsten rat sarcoma viral oncogene homolog (KRAS) mutation status has been established. In the present study, we investigated cetuximab accumulation in colorectal tumors and major organs using 111In-DOTA-cetuximab. We also evaluated the potential of positron emission tomography (PET) imaging of 64Cu-DOTA-cetuximab. Colorectal tumor xenografts with a different EGFR expression level and KRAS mutation status were subjected to in vivo biodistribution study and PET imaging at 48 h post-injection of radiolabeled cetuximab. The EGFR expression levels on colorectal tumors were determined by ex vivo immunoblotting and ELISA. We found that KRAS wild-type tumors had significantly higher 111In-DOTA-cetuximab accumulation than KRAS mutant tumors (P <0.001). Based on KRAS mutation status, a strong correlation was found between 111In-DOTA-cetuximab tumor uptake and EGFR expression level (KRAS wild type: r =0.988; KRAS mutant: r =0.829), and between 64Cu-DOTA-cetuximab tumor uptake with EGFR expression level (KRAS wild type: r =0.838; KRAS mutant: r =0.927). Significant correlation was also found between tumor uptake of 111In-DOTA-cetuximab and 64Cu-DOTA-cetuximab (r = 0.920). PET imaging with 64Cu-DOTA-cetuximab allowed clear visualization of tumors. Both radiolabeled cetuximab had effectively visualized cetuximab accumulation in colorectal tumors with a wide variety of EGFR expression levels and different KRAS mutation status as commonly encountered in the clinical setting. Our findings suggest that this radioimmunoimaging therefore can be clinically translated as an in vivo tool to predict cetuximab accumulation in colorectal cancer patients prior to cetuximab therapy. (Cancer Sci 2012; 103: 600–605)
In 2008, colorectal cancer was the third most commonly diagnosed cancer worldwide, the third most common in men and the second in women. Additionally, colorectal cancer was the fourth leading cause of death from cancer with an 8% mortality rate. By 2020, the number of newly diagnosed cases of colorectal cancer is expected to grow by approximately one-third.(1)
Approximately 25% and 19% of European and American patients, respectively, present with metastases at initial diagnosis, and 89% of them die within 5 years following the diagnosis.(2,3) Therefore, to enhance the efficacy of conventional chemotherapy for advanced or metastatic colorectal carcinoma, current guidelines recommend using targeted therapy.(2,4) One of these targeted drugs is cetuximab (Erbitux), which is a chimeric monoclonal antibody (mAb) specialized in targeting the extracellular domain of epidermal growth factor receptor (EGFR) and disrupting its downstream signaling by preventing EGFR ligands binding. Since its approval in 2004, cetuximab has been widely used in clinical practice as a single agent or with other modalities for treatment of metastatic or advanced colorectal carcinoma. However, some clinical trials revealed that cetuximab can be beneficial to only approximately 31.5% of patients.(5)
In the presence of EGFR ligands, EGFR has been shown to play an essential role in carcinogenesis, by initiating downstream signaling of several prosurvival cascades, such as the RAS and PI3K pathways. Activation of these pathways is believed to initiate cell proliferation, angiogenesis, invasion, metastasis and survival, making EGFR one of the most promising molecular targets in cancer therapy.(6) Although EGFR is overexpressed in 82% of colorectal cancer patients(7) and associated with a poor prognosis,(8) the EGFR level as determined by immunohistochemistry (IHC),(9) FISH(10) and quantitative PCR(11) failed to serve as an outcome predictor for cetuximab. Recently, some studies showed that the benefit of cetuximab is limited only to those with wild-type V-Ki-ras2 Kirsten rat sarcoma viral oncogene homolog (KRAS) and wild-type B-type Raf kinase gene (BRAF) tumors.(12,13) Additionally, KRAS mutation was claimed to be a negative predictor for EGFR inhibitor therapy.(2,4,14) However, a meta-analysis study found that 50–60% of patients with wild-type KRAS failed to respond to cetuximab.(15) Paradoxically, 9% of cetuximab responders were KRAS mutant patients, therefore patient selection criteria seems to be an issue for cetuximab therapy.(15)
Despite the controversies, EGFR as tumor-target therapy remains a focus of interest. While IHC and EGFR gene copy number has proven unreliable for tumor EGFR expression assessment,(16,17) radioimmunoimaging as noninvasive molecular imaging is expected to provide more comprehensive visualization of tumor-expressing EGFR.(18,19) Radioimmunoimaging of EGFR using cetuximab was studied in several tumor types.(18,20–25) It is well established that heterogeneity among affinity and distribution of EGFR exists in the tumor site, affecting cetuximab accumulation in tumors and thereby hampering the therapeutic effect.(26,27) Using positron emission tomography (PET), EGFR expression of the whole tumor has become possible to be quantified.(20) Moreover, radioimmunoimaging provides direct evaluation of the cetuximab accumulation level in tumors. Due to the fact that antibody accumulation depends not only on the EGFR expression level but also on tumor perfusion, vascularity and vascular permeability,(18) the cetuximab accumulation level in tumors is more predictive of the therapeutic response than merely the EGFR expression level. Assessment of the antibody accumulation level in tumors can be performed by immuno-PET.(18)
Copper-64 (64Cu) is an attractive radionuclide for PET imaging due to its appropriate half-life of 12.7 h and easy production.(28) PET imaging using 64Cu-labeled antibodies has been successfully performed in our laboratory.(29,30) Furthermore, PET imaging using 64Cu-labeled cetuximab has been used successfully to detect EGFR expression in prostate(31) and cervical cancer.(32) Therefore, the aim of the current study was to investigate the EGFR expression level and radiolabeled cetuximab accumulation in colorectal cancers and its correlation with KRAS mutation status, and evaluate the potential of PET imaging using 64Cu-labeled cetuximab.
Materials and Methods
Cetuximab was obtained from Merck KgaA (Darmstadt, Germany). Bifunctional ligand DOTA (2-(4-isothiocyanatobenzyl)-1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid or p-SCN-Bz-DOTA) was purchased from Macrocyclics (Dallas, TX, USA). Copper-64 (350–500 MBq) was produced on a biomedical cyclotron CYPRIS HM-18 (Sumitomo Heavy Industries Ltd, Tokyo, Japan) at Gunma University Hospital. Indium-111 (74 MBq/mL) was purchased from Nihon Medi-Physics (Nishinomiya, Japan).
Cell lines and xenografts. Human colorectal cancer cell lines HT-29, DLD-1, LoVo, SW480, SW620 and SW948 were obtained from ATCC (Manassas, VA, USA), and human colorectal cancer cell lines SNU-C4 and SNU-C5 were from Korean Cell Line Bank (Seoul, Korea). All cell lines were grown in their respective medium (McCoy’s 5A, Ham’s F12, RPMI 1640 or Leibovitz’s L-15, from Invitrogen [Carlsbad, CA, USA] or Wako [Osaka, Japan]) supplemented with 10% heat-inactivated FBS (Nichirei Bioscience, Tokyo, Japan) and 1% antibiotic (0.1 mg/mL penicillin and 100 U/mL streptomycin; Sigma-Aldrich, St Louis, MO, USA). All cell lines were grown as monolayers at 37°C in a humidified atmosphere comprising 5% CO2 and 95% air, except for SW480, SW620 and SW948, which had to be maintained in a CO2-free environment. KRAS mutation status analysis was performed using PCR-based direct sequencing as described previously.(33) Two cell lines (SW480 and SW620) had a homozygous mutation (G35T), whereas two cell lines (LoVo and DLD-1) had a heterozygous mutation (G38A). The four remaining cell lines (HT-29, SNU-C4, SNU-C5 and SW948) were wild type. Female athymic Balb/c nude mice were purchased from Japan CLEA (Tokyo, Japan). To make a colorectal tumor xenograft, 5 × 106 cells in 100 μL PBS suspension were injected subcutaneously in the dorsal flank of the mice and grown for 2–3 weeks (size approximately 200–300 mm3). The animal studies were performed in accordance with our institutional guidelines and were approved by the Local Animal Care Committee.
EGFR expression analysis. Tumor tissues from xenografts were lysed with PMSF-added T-PER tissue protein extraction buffer (Pierce Biotechnology, Inc., Rockford, IL, USA) and homogenized with Polytron (Kinematica Inc., Bohemia, NY, USA). The protein concentration was determined using DC Protein assay kit (Bio-Rad Laboratories, Hercules, CA, USA). Proteins (50 μg) were fractionated on 5–20% gradient SDS-PAGE gel (Atto, Tokyo, Japan) and blotted onto PVDF membrane (Millipore, Billerica, MA, USA) by electrotransfer. After blocking for 1 h, the blots were then incubated overnight at 4°C with primary antibodies directed against EGFR (#2232; Cell Signaling, Beverly, MA, USA) or anti β-actin (clone AC-15; Sigma-Aldrich) as a loading control. Bound antibodies were visualized using HRP-linked secondary antibodies (sc-2030; Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA) or anti-mouse IgG (KPL Inc., Gaithersburg, MD, USA), developed with Amersham ECL detection reagents (GE Healthcare UK Ltd, Buckinghamshire, UK). Film was scanned using positive-film mode. Densitometry analysis to obtain an adjusted EGFR band intensity was made using ImageJ software (version 1.43, developed by Wayne Rasband; The National Institute of Health, Bethesda, MD, USA). Quantitative confirmation of the EGFR expression level was made by performing ELISA using Human EGFR ELISA Kit (Boster Biological Technology, Wuhan, China) following the manufacturer’s instructions. The result was reported as the EGFR concentration in protein.
Preparation of DOTA-cetuximab. Cetuximab (2 mg/mL) was washed with borate-buffered saline (0.1 M, pH 8.5) in Vivaspin (Sartorius Stedim Biotech, Aubagne, France) and then concentrated. DOTA was dissolved in dimethyl formamide, added into concentrated cetuximab (in an equilibrium 5:1 to cetuximab) and incubated overnight at 37°C. The conjugate was then purified with centrifugation in Vivaspin. The concentration of the resulting DOTA-cetuximab conjugate was measured using a Nanodrop spectrophotometer (Thermo Scientific, Wilmington, DE, USA).
Radiolabeling of DOTA-cetuximab with 64Cu and 111In. Copper-64 was provided in a dry state as 64CuCl2 (300–500 MBq). Purification of DOTA-cetuximab was performed using Bio-spin 6 Tris column (Bio-Rad). Acetate buffer (0.25 M, pH 5.5) was mixed with DOTA-cetuximab (500 μg) before dissolving 64CuCl2. Incubation of this compound lasted 1 h in 40°C. Purification to complex-free 64Cu was carried out using PD-10 desalting column (GE Healthcare Biosciences AB, Uppsala, Sweden) with the help of 100 mM EDTA (5 μL). Indium-111 was provided as 111InCl3 in a liquid state. The same procedure was performed to obtain 111In-DOTA-cetuximab. Specific activity of the final product was measured using a CRC-15R dose calibrator (Capintec, Ramsey, NJ, USA) and the labeling yield was checked using the TLC method. After purification, the radiochemical purity of 64Cu-DOTA-cetuximab and 111In-DOTA-cetuximab was more than 99%. The specific activity of the final product per milligram cetuximab ranged 200–250 MBq for 64Cu and 2–3 MBq for 111In.
Biodistribution study. Xenografts of colorectal cell lines were intravenously injected with 50 kBq 111In-DOTA-cetuximab (protein dose: 20 μg) via the tail vein. Mice were killed 48 h after injection by decapitation. Blood, major organs and tumors were collected and weighed, and the radioactivity was measured in an automated γ-counter ARC-7001 (Aloka Co. Ltd, Tokyo, Japan). The distribution was expressed as percentage of injected dose/g (%ID/g) of tissue for all samples.
PET study. Xenografts of colorectal cell lines were intravenously injected with 20 MBq 64Cu-DOTA-cetuximab (protein dose: 100 μg) via the tail vein. PET images were taken 48 h after injection. After anesthetization by inhalation of isoflurane, whole body imaging was performed with a small-animal PET scanner (Inveon, Siemens, IL, USA). The energy window was set between 350 and 650 keV. The imaging data were reconstructed using a 2-D ordered-subsets expectation maximization algorithm. No correction was performed for attenuation and scattering. Quantification of the PET tracer uptake was done using the AMIDE (A Medical Image Data Examiner) software program (version 1.0.0, developed by Andreas M. Leoning; Crump Institute for Molecular Imaging, UCLA School of Medicine, CA, USA). Uptake of 64Cu-DOTA-cetuximab in organs was expressed as standardized uptake values (SUV), which was defined as the counts per second per pixel in a region of interest (ROI) encompassing the entire tumor divided by the total counts per second per pixel in the mouse.
Statistical analysis. Comparison of means was performed using an independent sample t-test and Mann–Whitney U-test, and P-values <0.01 were considered statistically significant. Correlation was considered as very strong when the correlation coefficient (r) was more than 0.8.
EGFR expression in colorectal cancer cell lines. Various levels of EGFR expression were observed in the eight cell lines (Fig. 1). Measurements of densitometry analysis were normalized to SW620 (1.00) to obtain a semi-quantitative EGFR expression level. Quantitative confirmation of EGFR expression obtained from ELISA was in agreement with the result of densitometry analysis and showed that SNU-C5 and LoVo had high EGFR expression, whereas EGFR expression for SW948 and SW620 was very low.
Biodistribution of 111In-DOTA-cetuximab. Table 1 summarizes 111In-DOTA-cetuximab uptakes in major organs and tumors 48 h after injection between KRAS wild-type tumor xenografts and KRAS mutant tumor xenografts. Compared to background accumulation, such as in muscle, cetuximab had high accumulation in tumors (Fig. 2). Accumulation in normal organs was similar in both the KRAS wild-type and KRAS mutant group. In the KRAS wild-type tumors group, SNU-C5 had the highest tumor accumulation of cetuximab (28.8 ± 2.9 %ID/g) and SW948 had the lowest (13.2 ± 7.3 %ID/g). In the KRAS mutant tumors group, LoVo accumulated the highest level of cetuximab (15.4 ± 1.2 %ID/g) while SW620 had the lowest (5.7 ± 3.1 %ID/g). KRAS wild-type tumors had a significantly higher accumulation of cetuximab than the KRAS mutant tumors (P <0.001).
Table 1. Comparison of the biodistribution of 111In-DOTA-cetuximab in colorectal tumor xenografts†
Mean uptake (%ID/g) ± SD
KRAS wild type
†n =37; 4–5 mice each cell line (KRAS wild type: SW948, HT-29, SNU-C4, SNU-C5, n =18; KRAS mutant: SW620, SW480, DLD-1, LoVo, n =19).
8.13 ± 2.53
8.73 ± 2.31
16.57 ± 10.11
11.80 ± 8.11
7.86 ± 2.18
7.31 ± 1.70
1.63 ± 0.30
1.77 ± 0.37
8.12 ± 5.29
5.71 ± 4.93
1.66 ± 0.08
1.51 ± 0.17
4.75 ± 0.88
4.64 ± 1.07
2.58 ± 0.22
2.47 ± 0.38
1.22 ± 0.33
1.16 ± 0.13
1.14 ± 0.13
0.95 ± 0.15
20.21 ± 7.15
10.16 ± 4.90
Correlation of cetuximab tumor uptake with EGFR expression level and KRAS mutation status. Figure 3 shows the correlation analysis of tumor 111In-DOTA-cetuximab uptake with adjusted EGFR band density and with EGFR concentration based on the KRAS mutation status. Linear relationships with a positive slope existed in both graphs indicating a positive correlation between variables. Tumor uptake strongly correlated with the adjusted EGFR band density in KRAS wild-type tumors (r =0.988) and in the KRAS mutant tumors (r =0.829). The same tendency was also found on the correlation between tumor uptake and EGFR concentration in both the KRAS wild-type and KRAS mutant groups (r =0.945 and r =0.944, respectively).
PET imaging and biodistribution of 64Cu-DOTA-cetuximab. To visualize non-invasively the whole body distribution and tumor-targeting efficiency of the cetuximab, we performed small-animal PET imaging at 48 h after injection of 64Cu-DOTA-cetuximab on colorectal cancer xenografts. As shown in Figure 4, 64Cu-DOTA-cetuximab was highly accumulated in tumors with high EGFR expression. Quantification of the PET images revealed that tumor uptake (SUV) in KRAS wild-type SW948, HT-29, SNU-C4 and SNU-C5 was 5.98, 6.13, 7.37 and 7.20, respectively, and KRAS mutant SW620, SW480, DLD-1 and LoVo was 4.52, 4.27, 5.40 and 6.71, respectively. A significant correlation was observed between tumor accumulation of 111In-DOTA-cetuximab (%ID/g) and 64Cu-DOTA-cetuximab (SUV) within all tumor models of colorectal cancer (Fig. 5, r =0.920). Tumor uptake of 64Cu-DOTA-cetuximab was correlated with the EGFR expression level in both KRAS wild-type (r =0.838) and KRAS mutant (r =0.927) tumors.
Colorectal cancer ranks second in terms of KRAS mutations among cancers with EGFR overexpression.(34) Resistance to anti-EGFR therapy particularly in KRAS mutant patients has led to intense investigations over the past few years.(1,12,15) In the current guidelines, KRAS mutation is considered as a negative predictor for cetuximab therapy; however, the reason behind the finding that KRAS wild-type patients and KRAS mutant patients represent 59% of non-responders and 9% of responders, respectively, remains a challenge.(2,4,15) To the best of our knowledge, this is the first PET study targeting EGFR expression with reference to the KRAS mutation status for colorectal cancers.
In the present study, based on the KRAS mutation status, a strong linear correlation was found between EGFR expression levels and cetuximab accumulation in colorectal tumors with various EGFR expression levels and KRAS mutation status. Furthermore, visualization of cetuximab accumulation in colorectal tumors using radiolabeled cetuximab PET imaging was in agreement with the radioactivity count of cetuximab accumulation in the biodistribution study. Although the uptake of cetuximab in KRAS mutant tumors was lower than KRAS wild-type tumors, their different accumulation levels were detectable and quantifiable in PET images, including the SW620 tumor, which is commonly used as a negative control for EGFR expression by immunohistochemistry and other assays.(35,36) A study involving colorectal cancer patients suggests that the oncogenic KRAS alters the EGFR localization in tumor cells.(14) This might explain the lower level of cetuximab uptake in KRAS mutant tumors.
Therefore, our finding might provide some clues regarding how some KRAS mutant patients could benefit from cetuximab therapy. Moreover, as cetuximab is expected to mediate antibody-dependent cellular cytotoxicity,(37) it would be imperative to show the therapeutic effect against KRAS mutant tumors, particularly in patients with high cetuximab accumulation. As a result, imaging with radiolabeled cetuximab could be an important outcome-predictor for cetuximab treatment.
Radioimmunoimaging of cetuximab in other tumor types showed diverse results regarding the EGFR expression and its correlation with tumor uptake.(18,20,22–25,32) Therefore, generally, PET imaging might not be able to predict the response of cetuximab therapy for a variety of tumors because it is only descriptive of EGFR reachability to cetuximab. However, our finding indicates that 64Cu-DOTA-cetuximab PET is a useful imaging tool to describe cetuximab accumulation in colorectal tumor, especially in KRAS mutant tumors. In the present study, we selected colorectal cancer lines with a wider spectrum of EGFR expression levels and different KRAS mutation status. Therefore, our subjects were much closer to the clinical setting compared with previous studies. In the clinical setting, assessment of the EGFR expression level for patients with colorectal cancer remains critical because some patients whose tumors have no EGFR expression detected using IHC, or any increase in EGFR copy number detected using FISH might benefit from cetuximab; on the contrary, patients with high EGFR staining do not necessarily respond to cetuximab.(7,9,10) Actually, IHC is limited by its technical factor, which can not overcome the heterogeneity among EGFR affinity and distribution, making it not representative for describing the EGFR expression level in the whole entity of tumors.(16) FISH is also limited by excluding EGFR protein expression in the cell membrane.(17) Furthermore, the KRAS mutation, which is believed to be used as an outcome predictor for cetuximab therapy with high specificity (0.91), is still insensitive (0.48).(15) Additionally, tumor vascularity and the microenvironment influence the reachability of EGFR as a target.(18,25) These factors could be overcome by the use of radioimmunoimaging, which is able to comprehensively visualize the tumors. Therefore, assessment of the tumor’s capability to accumulate cetuximab, as determined by 64Cu-DOTA-cetuximab PET, plays a superior role in patients’ selection criteria compared with any other assay that depicts EGFR expression in tumors merely anatomically.
A clinical study successfully performed PET imaging of 64Cu-labeled monoclonal antibody for the detection of colorectal tumors, without complication up to 24 months after imaging.(38) For clinical translation, the present study signifies that cetuximab remains accumulated in sufficient quantities in both KRAS wild-type and KRAS mutant colorectal tumors. Therefore, 64Cu-DOTA-cetuximab might serve as a tool for non-invasive assessment of cetuximab accumulation in tumors before initiation of cetuximab therapy and during cetuximab therapy. Our findings also hypothesize that cetuximab-based radioimmunotherapy could be a reasonable means for treating colorectal tumors regardless of their KRAS mutation status, because our results suggest that even in KRAS mutant tumors cetuximab accumulatation was in relation to the amount of EGFR expression. Considering the fact that approximately 10% of KRAS mutant patients have responded to cetuximab therapy,(15) quantitative assessment of cetuximab accumulation in tumors with 64Cu-DOTA-cetuximab imaging might be clinically important as a non-invasive technique for patient selection.
PET imagings in tumors expressing EGFR were also successfully performed using 89Zr and 86Y.(21,24,25) Zirconium-89 is an attractive positron emitter because of its long half-life (78.4 h) and PET imaging using 89Zr-labeled antibodies is also effectively performed in clinical practice.(39) Hence, we are in no doubt that in the near future, immuno-PET will function as a non-invasive tool for the assessment of antibody accumulation in tumors prior to therapy.
In conclusion, this study has effectively visualized cetuximab accumulation in colorectal tumors with a wide variety of EGFR expression level and different KRAS mutation status such as that commonly encountered in the clinical setting. This radioimmunoimaging as a consequence can be clinically translated into an in vivo tool to predict cetuximab accumulation in both KRAS wild-type and KRAS mutant tumors in colorectal cancer patients prior to cetuximab therapy.
The authors are grateful to Merck KGaA for kindly providing cetuximab.