Sulfonamide derivative targeting carbonic anhydrase IX as a nuclear imaging probe for colorectal cancer detection in vivo.

Hypoxic microenvironment is a common situation in solid tumors. Carbonic anhydrase IX (CA9) is one of the reliable cellular biomarkers of hypoxia. The role of CA9 in colorectal cancer (CRC) remains to be clarified. CA9 inhibitor such as sulfonamides is known to block CA9 activation and reduce tumor growth consequently. Here, we aimed to investigate the CA9 expression in serum and tumor from different stages of CRC patients and utilize sulfonamide derivative with indium-111 labeling as a probe for CRC nuclear imaging detection in vivo. The serum CA9 was correlated with the tumor CA9 levels in different stages of CRC patients. Hypoxia increased cell viability and CA9 expression in colorectal cancer HCT-15 cells. Sulfonamide derivative 5-(2-aminoethyl)thiophene-2-sulfonamide (ATS) could bind with CA9 in vitro under hypoxia. Moreover, tumor tissues in HCT-15-induced xenograft mice possessed higher hypoxic fluorescence signal as compared with other organs. We also found that the radioisotope signal of indium-111 labeled ATS, which was utilized for CRC detection in HCT-15-induced xenograft mice, was markedly enhanced in tumors as compared with non-ATS control. Taken together, these findings suggest that CA9 is a potential hypoxic CRC biomarker and measurement of serum CA9 can be as a potential tool for diagnosing CA9 expressions in CRC clinical practice. The radioisotope-labeled sulfonamide derivative (ATS) may be useful to apply in CRC patients for nuclear medicine imaging.


INTRODUCTION
Colorectal cancer (CRC) is one of most common malignancy cancers worldwide [ 1 ]. Epidemiological studies have shown that the incidence and mortality of colorectal cancer is increasing over the past several decades [ 2 , 3 ]. The clinicopathological characteristics of CRC patients are well known, including the factors of age and tumor site [ 4 -6 ]. In addition, the higher mortality of CRC is due to the disease that is frequently diagnosed in the advanced stage without reliable diagnosis. Thus, accurate detection of cancerous lesions can provide more effective surgical and pharmacological therapies for decreasing in the mortality of CRC [ 7 ]. Colonoscopies are considered the most preferred tool for CRC screening in clinical; however, they are invasive and present a small but signifi cant risk for perforations and limitations in making an accurate diagnosis due to sampling errors or personal skill [ 8 , 9 ]. The development of high sensitivity and specifi city diagnostic methods is urgent and important.
Hypoxia (low oxygen tension) is a biologically important phenomenon with strong effects on cancer progression and tumor phenotype [ 10 ]. The hypoxic tumor cells rapidly outgrow their blood supply leading to portions of the tumor with regions where the oxygen concentration is markedly lower (typically Ϲ1% of overall oxygen content) than in healthy tissues, and accompany a decrease in extracellular pH (~ pH 6.5) in the tumor microenvironment [ 11 , 12 ]. In order to survive in the acidic microenvironment, these tumor cells have to maintain an intracellular pH near physiological levels (pH 7.2-7.4) [ 13 ]. Carbonic anhydrase IX (CA9), a zinc metallo-enzyme, is one of 15 human isoforms of carbonic anhydrase family, which catalyzes the reversible hydration of carbon dioxide to bicarbonate and a proton [ 14 , 15 ]. CA9 has been observed to be directly linked to an increase of hypoxia inducible factor (HIF)-1, and to correlate to cell survival, proliferation, migration, growth, adhesion, pH value, and cell-signaling pathways [ 16 , 17 ]. Previous studies have shown that CA9 is predominantly associated with and high expressions in many human tumor types, including breast [ 18 ], lung [ 19 ], kidney [ 20 ], cervix uteri [ 21 ], oral cavity [ 22 ], brain [ 23 ], pancreas [ 24 ], and gastric epithelium [ 25 ]; but its decreased expression may be associated with human gastric carcinogenesis [ 26 ]. Hence, more attention is focused on the CA9 that is being pursued as a biomarker for development of specifi c systemic therapies and diagnostic imaging probe [ 27 , 28 ].
Carbonic anhydrase inhibitors have been extensively studied and their inhibition mechanisms were wellestablished for carbonic anhydrase-correlated imaging and therapy [ 29 ]. Sulfonamide-based compounds (sulphonamides, sulphanilamides, sulphamates and their derivatives) are CA9 small molecular inhibitors, which inhibit CA9 by coordinating to the zinc ion within the active site with μM to nM Ki inhibition [ 30 , 31 ]. They are the most robust class of CA9 inhibitors due to their high affi nity, availability, and ease of chemical manipulation [ 32 ]. Fluorescent dye-conjugated sulfonamides have been developed and revealed the binding affi nities to CA9 [ 33 ]. The fl uorescent sulfonamides for measuring CA9 expressions have been used in renal and breast tumor xenografts [ 34 -36 ]. In addition, the near-infrared (NIR) fl uorescent labeled sulfonamide derivatives were synthesized for imaging hypoxia-induced CA9 expressions in tumor cells in vitro, ex vivo , and in vivo [ 37 ]. However, the specifi c effi ciency of sulfonamides on CA9-overexpressed CRC remains to be clarifi ed.
Nuclear imaging is known as a common clinicalused diagnostic tool in several cancers for providing a prominent proportion of information in surgical management, radiation planning, chemotherapeutic assessment, and follow-up evaluation of patients [ 38 ].
The nuclear imaging with cancer-specifi c probes such as monoclonal antibodies, peptides, nanoparticles and other specifi c small molecular compounds are potentiality for tumor diagnosis [ 39 -43 ]. Even though technetium-99m labeled sulfonamide derivative has been developed for CRC detection, the results were undesirable [ 44 ]. Therefore, in the present study, we aimed to investigate whether (1) serum CA9 (sCA9) levels are correlated with tumor tissue CA9 levels in clinical CRC patients and can as a CRC reliable biomarker; (2) CA9 can as a biomarker for hypoxic tumor diagnosis; (3) radiolabeled sulfonamide derivative can bind to CA9-overexpressed CRC tumors.

The levels of sCA9 and tumor tissue of CRC patients
In order to evaluate CA9 as a CRC biomarker, we performed ELISA assay to observe the serum CA9 (sCA9) levels of CRC patients as compared to healthy volunteers. There were about 32% (8/25) of early stage and 44% (11/25) of late stage CRC patients displaying high-level sCA9 more than the average of sCA9 concentration in disease group, and about 76% (early stage:17/25, late stage: 21/25) of CRC patients revealing high-level sCA9 more than the average of sCA9 levels in control group ( Figure 1A ). The clinical characteristics of CRC patients and healthy volunteers were shown in Table 1 . To detect the correlation between sCA9 and tissue CA9 in the individual CRC patients, protein levels of CA9 in tumor tissues from patients with low-and high-level sCA9 were measured. As shown in Figure 1B , the protein expressions of CA9 in tumors were higher in patients with high-level sCA9 than in patients with low-level sCA9. Moreover, the CA9 expressions were also analyzed in tumor and non-tumor tissues from patients with high-level sCA9. As shown in Figure 1C , higher CA9 expressions were observed in tumor tissues than in non-tumor tissues. The sCA9 levels were correlated with tissue CA9 expressions in CRC patients ( p < 0.05, Figure 1D ), considering that detection of sCA9 may be used to refl ect the levels of CA9 in tumors as a reference during cancer therapy.

Histological and immunohistochemical analysis of clinical specimens of CRC
The histology of normal and cancer colon tissue were shown in Figure 2A -2C . The results exhibited strong staining of nuclei of (proliferating) epithelial cells in the cancer colon mucosa and stroma as compared with normal tissue. In order to confi rm the cancerous tumor and non-tumor tissue, methylene blue stain was performed to observe in CRC patients colon tissue ( Figure 2D -2F ).
To investigate tissue hypoxia condition in the different CRC stages, we analyzed HIF-1α, which have been proposed as a potential cell hypoxic biomarker for celland tissue-based detection by using immunohistochemical assay [ 45 ]. The immunohistochemical staining was conducted for CA9 and HIF-1α to verify their protein expressions within the normal and colon tumor tissues. HIF-1α exhibited strong staining in the colon cancerous tissues ( Figure 2H and 2I ). Within normal tissue, a majority of epithelial cells stained negative for HIF-1α ( Figure 2G ). Moreover, CA9 also exhibited strong staining within the tumor tissues ( Figure 2K and 2L ). Little to no staining was observed for CA9 in human normal colon tissues ( Figure 2J ). Interestingly, the expressions of CA9 and HIF-1α in late stage of CRC tissue were higher than in early stage of CRC tissue.

Hypoxia condition induced cells proliferation and CA9 overexpression in human CRC cells
We indicated that the hypoxia levels in CRC tumor tissues were higher than non-tumor tissues ( Figure 2 ). In order to estimate the effects of hypoxia, the cell viability were measured in colorectal tumor cells (HCT-15) and human normal colon cells (FHC). As shown in Figure 3A , the cellular viability was signifi cantly increased in HCT-15 under hypoxia condition for 24 and 48 hours as compared with normoxic condition. However, the cellular viability was markedly decreased in human normal colon cells (FHC) under hypoxia condition. To analyze the effect of hypoxia on cellular proliferation in HCT-15 cells, the cellular proliferated markers, Ki-67 and PCNA, were detected by immunoblotting assay. The protein patients with stage 1-2 and stage 3-4 detected using an ELISA assay. * P < 0.05, ** P < 0.01, *** P < 0.001. B. CA9 protein expression in CRC tissues responding the sCA9 levels. The tissues were chosen from four high and four low sCA9 levels of CRC patients for Western blotting. Protein levels were quantifi ed by densitometry and normalized by GAPDH levels. The data are showed as means ± SEM ( n = 4). ** P < 0.01, versus low sCA9 levels of CRC patients. C. The protein expressions of CA9 in tumors and non-tumors from the individual CRC patients with high sCA9 levels. Protein levels were quantifi ed by densitometry and normalized by GAPDH levels. The data are presented as means ± SEM ( n = 3). * P < 0.05, versus non-tumor. D. The sCA9 levels were correlated with tumor tissue CA9 protein expressions in CRC patients. * p < 0.001, versus sCA9. expressions of Ki-67 and PCNA were predominantly enhanced in hypoxia condition for 24 and 48 hours ( Figure 3B ). We further observed the expression of CA9 in HCT-15 and FHC cells under normoxic and hypoxic condition. As shown in Figure 3C , the expression of CA9 was remarkably up-regulated in HCT-15 under hypoxia, but there was no effect in FHC cells under normoxic or hypoxic condition. In addition, the expression of CA9 under hypoxia for 48 hours was higher than hypoxia for 24 hours.

Hypoxia and CA9 expression in HCT-15-induced xenograft mice
We next observed the tissue hypoxic imaging in HCT-15-induced xenografts in mice. The fl uorescence signals were markedly enhanced in the 4-week HCT-15-induced xenograft mice as compared with 1-week HCT-15-induced xenograft mice ( Figure 4A ). The distributions of hypoxia in several organs of 4-week HCT-15-induced xenograft mice were also estimated. Tumor tissue possessed higher fl uorescence signals as compared with other organs ( Figure 4B ). The protein expression of CA9 was also higher in tumor tissue than in other organs ( Figure 4C ). The CA9 levels were up-regulated predominantly in 4-week HCT-15-induced xenografts as compared with in 1-week HCT-15-induced xenografts, which was consistent with the results of fl uorescence imaging ( Figure 4D ).

Sulfonamide derivative interacted with CA9 protein and decreased HCT-15 cell viability under hypoxia
Sulfonamide derivative, 5-(2-aminoethyl)thiophene-2-sulfonamide (ATS) (2.5-100 μg/ml), decreased cell viability of HCT-15 cells under hypoxic condition ( Figure 5A ). However, there was no cytotoxicity in normoxic HCT-15 and other normal cells, including FHC, rat normal liver cells (Clone 9) and human normal kidney cells (HK2) ( Figure 5A ). Furthermore, the fl ow cytometry analysis was performed to measure the ATS specifi c function in hypoxic HCT-15 cells. As shown in Figure 5B , fl uorescein isothiocyanate (FITC)-labeled ATS (ATS-FITC) was markedly bound to hypoxic HCT-15 cells. The binding level of ATS-FITC was predominantly decreased in unlabeled ATS-pretreated HCT-15 ( Figure 5B ). To confi rm the binding ability of ATS-FITC in tumor and normal cells, immunofl uorescence staining was performed to analyze in HCT-15 and FHC cells under hypoxic and normoxic condition. The results revealed that fl uorescence imaging was prominently raised as compared to normoxic HCT-15 cells ( Figure 5C ). The fl uorescence signals were signifi cantly inhibited in unlabeled ATS-pretreated Abbreviations: BMI, body mass index; CEA, carcinoembryonic antigen.
HCT-15 cells. In addition, the fl uorescence imaging was also not detected in hypoxic and ATS-pretreated FHC cells. We further demonstrated that ATS could bind to HCT-15 cells via interacting with CA9 by using immunoprecipitation assay ( Figure 5D ).

The nuclear imaging agent of ATS-DTPA-111 In for CRC tumor detection in vivo
Next, we tried to synthesize the ATS-conjugated diethylene triamine pentaacetic acid (DTPA)-111 In (ATS-DTPA-111 In) complex to diagnose the CRC in a xenograft mouse model. DTPA is known to specifi cally trap the indium-111 radioisotope. The construct of the sulfonamide-based nuclear imaging agent and synthesis fl ow chart were illustrated in Figure 6A . After synthesis and purifi cation, the molecular weight of ATS-DTPA compounds was analyzed to determine the successful rate of conjugations. DTPA was successfully conjugated with ATS determined by mass spectrometry assay ( Figure 6B ). To ensure the labeling effi cacy of indium-111 onto the ATS-DTPA, instant thin layer chromatography was performed to separate indium-111 and indium-111 labeled ATS-DTPA (ATS-DTPA-111 In) in 15 cm membrane electrophoresis and estimated the radioactive signals by using a gamma counter. The ATS-DTPA labeling effi cacy with indium-111 was about 98% ( Figure 6C ), demonstrating that the sulfonamide-based nuclear imaging agent was synthesized with stable conjugation in phosphate buffered saline. To measure the stability of ATS-DTPA, the concentrations of ATS-DTPA in various pH conditions (pH = 5.5, 7.4, and 8.5) were analyzed. The results exhibited that the ATS-DTPA levels were no difference for 3 days in different pH condition ( Figure 6D ). In addition, the stability of ATS-DTPA-111 In in human and fetal bovine serum was observed in a 144-hours period. The results displayed that the stability of ATS-DTPA-111 In was preserved more than 94% and 92% for 144 hours in human and fetal bovine serum, respectively ( Figure 6E ).
In order to detect the imaging effi ciency of ATS-DTPA-111 In, the CRC tumor imaging was measured in the 4-week HCT-15-induced xenograft mice. After 6 and 24 hours imaging agent injection, the radionuclide signals were signifi cantly increased in ATS-DTPA-111 In-treated mice as compared with control DTPA-111 In-treated mice ( Figure 7A ). Then, the 24 hours distributions of ATS-DTPA-111 In in several organs of 4-week HCT-15-induced xenografts were also estimated. Tumor tissue possessed higher ATS-DTPA-111 In signals as compared with DTPA- 111 In groups ( Figure 7B ). Moreover, the bio-distribution of ATS-DTPA-111 In and DTPA-111 In were also markedly difference in blood, kidney, spleen and muscle ( Figure 7B   tumor hypoxic tissues and consequently improve the imaging effi cacy.

Toxicity evaluation for ATS-DTPA
We next tried to do a preliminary toxicity evaluation for ATS-DTPA in mice after signal dose intravenous injection. There were no deaths for all experimental groups within 14 days observation. There were no differences in body weight change among four experimental groups ( Figure 8A ). The indicators of renal function (BUN and creatinine) and liver function (alanine aminotransferase and aspartate aminotransferase) were no difference between control and ATS-DTPA-treated groups ( Figure 8B ). Moreover, histological evaluation of liver and kidney from control and ATS-DTPA (50 mg/kg)-treated mice revealed no imprint of toxicological signature in these organs ( Figure 8C ).

DISCUSSION
In this study, we identifi ed that CA9 levels in serum are correlated with tumor tissue CA9 levels in clinical CRC patients and CA9 can as a biomarker for hypoxic tumor diagnosis. We also found that sCA9 and tumor CA9 levels in late stage of CRC patients were higher than in early stage of CRC patients. In addition, hypoxia could cause cell proliferation and increase CA9 expressions in cultured CRC cells. The sulfonamide derivative ATS directly and specifi cally reacted with CA9 in hypoxic CRC cells. We also provided a novel method that sulfonamide conjugated with DTPA followed 111 -In labeling. The ATS-DTPA-111 In was successfully captured by CRC tumor imaging in vivo . In addition, the serum biochemistry values and histology staining indicated that ATS-DTPA compounds were no toxic effects in blood circulation of animals. These fi ndings presume that ATS effectively binds to CA9 overexpressed CRC cells and ATS-DTPA-111 In, providing a new nuclear imaging reagent for CA9overxpressed CRC tumor detection.
The non-invasive NIR fl uorescence imaging is known as a clinical application for detection of human deep organs. However, the limitation of tissue penetration let it to be unrealistic. It has been suggested to require considerable improvement of sensitivity of the fl uorescence imager and quantum effi ciency of fl uorophores [ 45 ]. In addition, even though Akurathi et al. developed technetium-99m (Tc-99m)-labeled sulfonamide for CRC detection [ 46 ], the results were undesirable. They only collected 0-4 h drug biodistribution data, but not imaging data. The biodistribution of [99mTc]-sulfonamide in CA9 expressing tumor bearing mice at 0.5, 1, 2 and 4 h (i.p.) showed low tumor accumulation (< 0.1% ID). We presumed that the lower accumulated radiolabeled drug resulted in image collection with diffi culty. In the present study, the result of bio-distribution of In-111 -labeled sulfonamide showed tumor hypoxia accumulation more than 5% ID after 24 h i.v. injection ( Figure 7B ). Therefore, we consider that isotope labeling is better than NIR fl uorescent labeling for hypoxia CRC detection in clinical and In-111 is more benefi cial for sulfonamide labeling compared with Tc-99m. . The supernatant of lysates was incubated with or without biotinylated-CA9 polyclonal antibody (2 μg/ml) in the presence of streptavidin agarose beads at 4°C overnight. The fl uorescent signaling of immunoprecipitates was detected by using ELISA reader. In competitive group, The ATS (10 μg/ml) was pretreated for competitive inhibition assay. Data were presented as means ± SEM ( n = 3). * P < 0.05, versus group 2; # P < 0.05, versus group 3. www.impactjournals.com/oncotarget It has been recognized for several decades that there was a correlation between tumor hypoxia and resistance to chemotherapy and radiotherapy [ 47 -50 ]. Though surgery of CRC is the foundation of curative treatment in more than 90% of cases, chemotherapy is maintained after tumor resection due to the high risk of recurrence in advanced cancers [ 51 ]. In addition to a few CRC cases, radiotherapy is a major method for head-andneck cancers treatment in clinical [ 52 ]. Therefore, tumor hypoxia detection could provide physician to determine the optimal drug usage and decreased time redundancy in tumor treatment. CA9 is presumed as one of the most prominent markers of tumor hypoxia with potential to serve as a diagnostic biomarker, prognostic indicator as well as tumor therapeutic target in many cancers [ 53 , 54 ]. In the present study, we demonstrated that CA9 was overexpressed in hypoxic CRC tumor as compared with normoxia condition. CA9 may be considered to be as a biomarker for hypoxic CRC tumor diagnosis.
The colonoscopy and histological examination of biopsy specimens are considered the gold standard for diagnosis of CRC in clinical practice [ 55 ]. However, the poor resolution of endoscopy and sampling error in small CRC nodules and interpersonal variations may decrease the accuracy of tumor analysis. In addition, immunohistochemistry (IHC) is the most generally methods to estimate the protein status in cancerous samples. Even though IHC exhibited high standardization and semi-quantitative assay, it has numerous limitations, including technical operation and illustration, which have been found to impact on the reproducibility and accuracy of results. Moreover, the complicated and time-consuming  -111 In was estimated at human and fetal bovine serum for 144 hours at 37°C and measured by ITLC. The radioactivity of ATS-DTPA-111 In was preserved more than 94% and 92% at 144 hours in human and fetal bovine serum, respectively. Data are presented as means ± SEM ( n = 5). processes lead to undesirable screening in a lot of clinical samples. The high throughput serum biomarkers screening for CRC provides results acquired quickly and interpreted easily to doctor for deciding tumor therapeutic method. Currently, FDA has approved human serum detection kits to determine the status of CA9 expressions. High levels of CA9 in plasma and serum have been measured to be raised in the patients with breast, lung, gastric and cervical cancer [ 19 , 56 -58 ]. In contrast to tumor tissue CA9 protein analysis, serum CA9 level examination in CRC patients is relatively rare. In the present study, we also observed that the average serum CA9 levels were remarkably increased in CRC patients, which were correlated with tumor tissue CA9 expressions. In addition, the serum CA9 levels in early stage of CRC patients were signifi cantly higher than normal control group. These results implied that serum CA9 levels determined by ELISA assay might be a profi table and invasive tool for diagnosing CA9 expressions in CRC patients.
In the study for G250 monoclonal antibody, it has been displayed excellent specifi city for CA9 in clear cell renal cell carcinoma (CCRCC) analysis [ 59 ]. Moreover, The M75 was another highly specifi c antibody for CA9 [ 60 ], which recognized the extracellular proteoglycanlike domain of CA9 and widely used for western blotting, immunoprecipitation and IHC in human tumor tissues. Compared to CA9 specifi c monoclonal antibody, the CA9 specifi c small molecule-based probes provided some advantages for tumor imaging. Sulfonamides has been demonstrated to only bind to active CA9, indicating that the active site of CA9 is only available for sulfonamides binding in several cell lines during hypoxia [ 33 , 61 ]. Nevertheless, the transmembrane CA9 proteins are very stable (with a half-life of about 40 hours) for a relatively long time after reoxygenation. Therefore, sulfonamides can distinguish cells that are currently in hypoxic conditions from those that are previously hypoxic, while CA9 specifi c antibodies cannot do it. In addition, sulfonamides are relatively easy to make and low cost and low immune responses, which are highly acceptable for clinical application and might provide different imaging and prognostic information compared to monoclonal antibodies.
In the present study, we observed that ATS markedly decreased the viability of HCT-15 cells under hypoxia condition. Previous studies indicated that sulfonamides could inhibit the activity of CA9 and resulted in pH regulation dysfunction in tumor cells [ 33 , 34 , 62 ]. Said and colleagues have shown that the specifi c chemical CA9 inhibitors, acetazolamide and SU.D2, are leading to an inhibition of CA9 in protein and mRNA levels during tumorgenesis of glioblastoma [ 63 ]. In addition, we also found that ATS bound to hypoxic HCT-15 cells via interacting with CA9. Therefore, we presume that the sulfonamide derivative decreases the CRC cell viability due to the insuffi cient activity and CA9 expression and thus they are unable to maintain the acidic tumor environment.
Hypoxia leads to stabilization and transcriptional activation of HIF-1α, which is a transcriptional factor for regulating hypoxia-inducible-genes expressions, including CA9 gene [ 64 ]. A previous report exhibited a predominantly positive correlation between CA9 expressions and the levels of tumor hypoxia in cervical carcinoma [ 65 ]. In this study, we observed that the hypoxia level was remarkably enhanced in HCT-15 induced 4-week xenograft mice as compared with control. In addition, the CA9 expressions in tumors of HCT-15 induced 4-week xenograft mice were also higher than control group. Therefore, we believed that HCT-15 induced 4-week xenograft mice are desirable animal model for CA9 imaging. CA9 expressions are primarily associated with the areas of low pH and high rates of cell proliferation in normal tissue. In the human, CA9 levels are generally limited to a total exclusivity in the gastrointestinal tract epithelium, namely glandular gastric mucosa, epithelium of the gallbladder and cryptic enterocytes of duodenum, jejunum, and ileum [ 66 ]. In the adult mice, however, the highest levels of CA9 in the normal tissues are restricted to a few tissues, including stomach, colon, and brain [ 67 ]. In this study, the hypoxia distribution assay presented that stomach and colon revealed highly fl uorescent signaling in the HCT-15 xenograft mice. However, the CA9 levels in stomach and colon were lower than other organs of HCT-15 xenograft mice. Therefore, we considered that the un-expectable fl uorescent contaminants in the gastrointestinal tract might lead to the false signaling production under IVIS detection.
In conclusion, we demonstrated that CA9 is a target protein for CRC, and serum CA9 may be a potential tool for analyzing CA9 expressions in CRC. ATS can specifi cally bind to CA9-overexpressed CRC cells in vitro . To characterize in vivo imaging and assess the binding effi cacy, ATS-DTPA-111 In was performed in HCT-15 xenograft mice for measuring and analyzing CA9 imaging by using nanoSPECT/CT instrument. ATS-DTPA-111 In displayed higher tumor binding effi cacy in HCT-15-induced xenograft mice. Therefore, ATS-DTPA- 111 In may be an ideal nuclear imaging agent for CRC diagnosis. The CA9 bound sulfonamide-based nuclear imaging may has potential to improve the care and welfare of patients suspected of having CRC and could be used to monitor patients with a history of CRC after surgery or drug treatment.

Acquisition of the tissues and sera of CRC
Clinical tissues and sera were acquired from Buddhist Tzu Chi General Hospital and Cheng Hsin General Hospital, Taiwan. The study protocol was approved by regulatory authorities and Institutional Review Boards of both institutions. Signed and informed written consent were obtained from CRC patients and healthy volunteers. The primary tumors (T) and adjacent non-tumor (NT) tissues from the individuals of CRC were captured followed by surgery. Tumor histopathology, grade, and stage were assigned according to the rules of American Joint Commission on Cancer Staging (AJCCS) system.

Cell lines
Human colorectal carcinoma cells (HCT-15), human normal colorectal cells (FHC), rat liver normal cells (Clone 9), and human proximal tubule cells (HK2) were purchased from the American Type Culture Collection (ATCC) and maintained in ATCC's recommended culture medium with 10% of fetal bovine serum. All cells were incubated at 37°C and 5% CO 2 .

Animals
The male BALB/c nude mice and male BALB/c littermates (7-8 weeks of age, weighing 21-23 g) were acquired from BioLASCO Taiwan Co., Ltd. These mice were selected and housed on a 12-h light cycle at 22°C and given food (mouse standard diet) or water ad libitum . All animal studies were approved by the institutive ethical review committee and were carried out in accordance with government regulations and NIH guidelines on the care and welfare of laboratory animals.

Human serum CA9 assay
The serum CA9 levels in the clinical serum samples (50 colorectal cancer patients and 30 control participants) were analyzed by ELISA system (eBioscience, USA). Measurements were performed strictly according to the manufacturer's protocol and quality control was ensured. Samples were run in duplicate and serial dilution was performed in order to fall within the detection limits of the assay (0-300 pg/ml).

Histology and immunohistochemistry
The samples were sectioned at 10 μm with a cryostat (HM525, Microm, Germany), and sections were thawmounted on gelatin-coated glass slides. For histology staining, these 10-μm-thick tissue section slides were stained with hematoxylin and eosin for morphology detection. The methylene blue stain was performed to detect the cancerous lesions. For immunohistochemistry, immunostaining was performed by immunoperoxidase secondary detection system (Merck Millipore, Germany) and the manufacturer's instruction with a modifi cation. Briefl y, sections were incubated for 1 h at 4°C with blocking reagent. The primary antibodies for CA9 (#SAB5300133, Sigma-Aldrich) and HIF-1α (#H6536, Sigma-Aldrich) were used. After washing in PBS, the sections were incubated for 30 min at room temperature with a secondary antibody reagent. The sections were washed again in PBS and incubated for 15 min at room temperature with streptavidin HRP reagent. After a fi nal wash in PBS, the sections were reacted with freshly prepared DAB Chromogen Reagent for 15 min. Sections were counterstained with hematoxylin counter stain solution (Sigma-Aldrich, USA) and coverslipped with Malinol (Muto Pure Chemicals, Tokyo, Japan). Images of the sections were obtained using a biological microscope Olympus DP70 (Olympus, Tokyo, Japan) and DP controller digital imaging software for research (Olympus) under 400 or 1000 magnifi cation.

Cell viability assay
WST-8 assay (Sigma-Aldrich, USA) was performed for cell viability measurement. Cells were seeded in 96-well plates at a density of 1 × 10 4 cells/well and cultured in 37°C and 5% CO 2 for 24 hours. After preincubation, each well, which containing about 80% confl uent cells were treated with ATS (1-500 μg/ml) for 24 and 48 hours under normoxic and hypoxic condition and then 10 μl of WST-8 solution was added to each well. After 2 hours-incubation at 37°C, the samples were analyzed by using a microplate reader (Bio-Rad, USA).

Flow cytometry analysis
HCT-15 and Clone 9 cells were seeded at 2 × 10 6 cells/ml in 10 cm dish, which were incubated under normoxic or hypoxic condition for 24 hours. In competitive group, HCT-15 cells were pre-treated with ATS (10 μg/ml) for 1 hour and incubated in hypoxic condition. After 24-hour incubation, medium was removed and washed with PBS for three times, and then incubated with FITC-labeled ATS (ATS-FITC, 10 μg/ml) for 3 hours. Cells were washed and harvested for fl ow cytometric analysis by FACSCalibur Flow Cytometer (BD Bioscience, USA).

Immunofl uorescence staining
The ATS-FITC binding assay was analyzed at HCT-15 cells by immunofl uorescence staining. Briefl y, cells were seeded in 8-well Millicell EZ slide (Merck Millipore, Germany) with cultured medium for 16 hours at 37°C and 5% CO 2 . Then, each well (2 × 10 3 cells) was incubated 37°C and 5% CO 2 for 24 hours under normoxic and hypoxic condition. ATS (10 μg/ml) was a competitor for competitive inhibition assay. Then, cells were washed by PBS and then fi xed using 4% paraformaldehyde for 15 min at room temperature. Fixed-cells were incubated with SuperBlock Blocking Buffers (ThermoFisher Scientifi c, USA) for 1 hour at room temperature. After blocking reaction, cells were incubated with fl uorescent labeled-ATS (10 μg/ml) for 4 hours at room temperature. DAPI was used to stain the cell nuclei at a concentration of 0.2 μg/ml for 10 minutes. After cells washing, the images were digitally acquired with a fl uorescence microscope (BX53, Olympus, Japan), coupled to a CCD camera. Digital images were taken using Image-Pro Plus software (Media Cybernetics, USA). Each staining was repeated 3 times in 3 different cultures.

HCT-15 xenograft mouse model and in vivo hypoxia imaging and CA9 distribution assay
HCT-15 cells (3 × 10 6 ) were subcutaneous inoculated into the right fl ank of nude mice ( n = 5 per each group). Tumors were established for one week and four weeks. For in vivo hypoxic imaging assay, the hypoxiadetected agent labeled with fl uorescent (HypoxiSense www.impactjournals.com/oncotarget 680, PerkinElmer, USA) was used. The commercial hypoxia-detection agent is a CA9 specifi c fl uorescent in vivo imaging agent that can be used to image CA9 overexpression in tumors in response to regional tumor hypoxia. The agent was intravenously injected into tail vein of HCT-15 xenograft mice. After injection for 24 hours, in vivo imaging system (IVIS, PerkinElmer, USA), a fl uorescent detection system, were performed to capture the whole animal body images for analyzing hypoxic distribution.

DTPA labeled-sulfonamide synthesis and measurement
For preparation of diethylene triamine pentaacetic acid (DTPA) labeled-sulfonamide (ATS-DTPA), ATS and DTPA dianhydride (w/w 1:50) was incubated at 25°C with bicarbonate buffer (pH 8.5) for 4 hours. Then, the conjugated-compounds were purifi ed by high pressure liquid chromatography with G-10 column. For molecular weight assay, liquid chromatography-mass spectrometry (LTQ Orbitrap XL™ Hybrid Ion Trap-Orbitrap, Thermo Scientifi c, USA) was performed for ATS analysis. And matrix-assisted laser desorption/ionization time-offl ight mass spectrometry (Ultrafl exIII TOF/TOF, Bruker Daltonics GmbH, Germany) was performed for DTPA and ATS-DTPA detection. Briefl y, the compounds were crystallized with an equal volume of freshly prepared α-cyano-4-hydroxycinnamic acid matrix (10 mg/ml in 50% acetonitrile/0.5% TFA) for mass spectrometry analysis. The calibration standards were used as external calibrator with mass accuracy within 100 ppm. The spectra were processed using FlexAnalysis™ 3.0 software (Bruker, Germany). The ATS-DTPA conjugation effi ciency with indium-111 was evaluated from the radiolabeling yields of ATS-DTPA-111 In determined by instant thin layer chromatography (ITLC) (AR-2000 radio-TLC Imaging Scanner, Bioscan, France) on silica gel impregnated glass fi ber sheets (PALL corporation, USA) using PBS buffer, pH 7.4, as the mobile phase.

The stability assay of ATS-DTPA and ATS-DTPA-111 In in vitro
To investigate the stability of ATS-DTPA, the levels of ATS-DTPA were measured to observe the variation of molecular weight based on pH of the solvent. Briefl y, the freeze-dried ATS-DTPA was dissolved in aqueous solvent with different pH: 5.5, 7.4, and 8.5 at 37°C. Then, the alteration of ATS-DTPA levels in an aqueous solution over time (up to 4 days) was estimated. Samples (200 μl) were collected from the stock solution and analyzed by using HPLC. For ATS-DTPA-111 In stability assay, the radiolabeled-compounds were incubated in human and bovine serum for 144 hour. Then, the radiolabeling yields of ATS-DTPA-111 In were analyzed by ITLC every 24 hour. Each assay was repeated 3 times in 3 different containers.

In vivo CA9 nuclear imaging
In vivo CA9 nuclear imaging was measured on male BALB/c nude mice, subcutaneous inoculated with HCT-15 cells (2 × 10 6 cells). Tumors were established for four weeks. For CA9 nuclear imaging assay, all mice received indium-111 labeled with or without ATS conjugation via intravenous injection with 2 mCi of indium-111 . NanoSPECT/CT (Mediso Medical Imaging Systems, USA) was performed to capture the imaging for detecting CA9 distribution at 6 and 24 hours. The radioactive intensity was estimated against the baseline (i.e., the radioactive intensity of muscle derived from a leg without the tumor embedded).

Ex vivo bio-distribution study
For tissue bio-distribution assay, 2 × 10 6 HCT-15 cells were injected subcutaneously into the right hind limb. Tumor size and the state of health of animals were checked regularly. Approximately 4 weeks after inoculation, all animals were performed isofl uorane anesthesia. Then, mice received a single tail vein injection of ATS-DTPA-111 In or DTPA-111 In (0.3-0.4 MBq [8][9][10][11]). At 24 hours post-injection, mice were sacrifi ced and then removed the organs and blotted dry. All samples were weighed and the radioactivity was analyzed using a gamma counter (1470 WIZARD, PerkinElmer, USA) along with known standards. Data points were corrected for radioactive decay. The percent injected dose per gram of tissue (%ID/g) was calculated.

Toxicity evaluation for ATS-DTPA
Forty male BALB/c mice were divided into 4 groups, each consisting with 10 animals. Group I was control group and received only vehicle (PBS). Group II, III, and IV were treated with ATS-DTPA at doses of 0.5, 5, 50 mg/kg, respectively. All groups were intravenously administrated with a signal dose. The body weight changes % of all group were determined daily for 14 days. Fourteen days after injection, blood was collected from retro-orbital sinus of anesthetized mice and serum was separated for determination of blood urea nitrogen (BUN), serum creatinine, alanine aminotransferase (ALT), and aspartate aminotransferase (AST). Mice were sacrifi ced and the kidneys and livers were isolated. The kidneys and livers were immersed in buffered 4% paraformaldehyde solution for 48 h to obtain histological sections.

Statistical analysis
All data analysis was performed by ANOVA followed by post hoc analysis with Bonferroni's test.
Data were presented as means ± SEM. Differences between means were considered signifi cation when p value < 0.05. Statistical analysis was performed by using GraphPad Prism V5.01 software (GraphPad Software, Inc. USA).