Multimodal Imaging Probe for Melanoma Evaluating of PTK7 Expression by Sgc8-c Aptamer Recognition

Melanoma is one of the most aggressive and deadly skin cancers, and although histopathological criteria are used for its prognosis, biomarkers are necessary to identify the different evolution stages. The applications of molecular imaging include the in vivo diagnosis of cancer with probes that recognize the tumor-biomarkers specic expression allowing external images acquisitions and evaluations of the biological process in quali-quantitative ways. Aptamers are oligonucleotides that recognize targets with high anity and specicity presenting advantages that make them interesting molecular imaging probes. Sgc8-c (DNA-aptamer) selectively recognizes PTK7-receptor overexpressed in various types of tumors. Herein, Sgc8-c was evaluated, in two melanoma models, non-metastatic and metastatic, as molecular imaging probe for in vivo diagnostic. Firstly, two probes, radio- and uorescent-probe, were in vitro evaluated verifying the high specic PTK7 recognition and its internalization in tumor cells by the endosomal route. Secondly, in vivo proof of concept was performed in animal tumor models. Likewise, they have rapid clearance from blood exhibiting excellent target (tumor)/non-target organ ratios. Furthermore, optimal biodistribution was observed 24 hours after probes-injections accumulating almost exclusively in the tumor tissue. Sgc8-c is a potential tool for their specic use in the early detection of melanoma. induced in B16F10. Once melanomas palpable, the Sgc8-c-Alexa647 probe (25 µg, 2 nmol) was injected i.v. and at 0.5, 2, 24 and 48 h post- injection (n = 5 per time group), mice were sacriced. Ex vivo images of organs (liver, heart, lungs, spleen, kidneys, thyroid, muscule, bone, blood and tumor) were taken using the imaging equipment above mentioned, with the X-ray and uorescence model. This study was compared with the melanoma tumor model generated with B16F1 cells. The results were expressed in ROI and the tumor/blood and tumor/muscle ratios were calculated. A competition test was performed. One group of mice (n = 5) were rst i.v. injected with Sgc8-c-NH 2 in excess of 5 times more than the probe. After 30 min post-injection, the same mice were i.v. injected with the Sgc8-c-Alexa647 probe. After 2 h mice were sacriced, images taken and organs weighed. Similar evaluation was performed for the radiolabelled probe. Sgc8-c-NOTA-67Ga was i.v. biodistribution until 72 h post-injection. Over time, isourane perform the in vivo. Live were aquired for 0.5, 2, 24, 48 and 72 h after injection by X-rays and gamma modalities biodistributions acquired the levels radioactivity


Introduction
Melanoma is one of the most aggressive and deadly types of skin cancer 1,2 , with an annual increase in incidence during the last decade between 15-25 per 100,000 individuals 3 . Although histopathological criteria such as tumor thickness, mitotic rate, histologic subtype and ulceration 4,5 , are usually used for its prognosis, biomarkers are necessary to identify whether the primary melanoma has metastasized or even differentiate the stages of its evolution 6,7 . Malignant melanomas have been reported to have increased activity of protein tyrosine kinase 7 (PKT7) 8,9 . This membrane receptor is highly conserved in different species and is involved in signal transduction pathways that mediate cell growth, cell polarity, differentiation, and survival 10,11 . However, PTK7 may participate as a co-receptor and its protection by type 1 membrane metalloprotease is implicated in the progression of cancer 12,13 . PTK7 has also been shown to be a key regulator in the Wnt / β-Catenin or Wnt / planar cell polarity pathway, and correlates with aggressive clinicopathological characteristics in cancer 14,15 . Likewise, this receptor is overexpressed in different types of leukemia, colon, lung, prostate, breast, gastric tumors, and even metastases [16][17][18][19][20][21][22] .
Furthermore, it participates in the migration and endothelial invasion of tumor cells [23][24][25] . Through imaging, it is possible to identify the expression of tumor biomarkers, using molecular imaging, a non-invasive technique that manages to evaluate the strategies for in vivo administration of the tumor target 26,27 . Molecular imaging consists of in vivo visualization, characterization and measurement of biological processes at the cellular or molecular level 27,28 . Molecular imaging is a very useful tool for diagnosing cancer. The use of probes that have optimal imaging characteristics provides clinically essential information for this disease, which would allow a correct selection of the treatment to be followed and the monitoring of its effects 29 .
In addition, aptamers have been used as a component of molecular imaging probes since they have the ability to bind, through non-binding covalent bonds, with high a nity and speci city for a molecular target 30,31 . They are oligonucleotides (ssDNAs or RNAs) that have a three-dimensional structure characterized by loops, stems or hairpins. Physicochemical properties, such as temperature and pH stability give an advantages to be functionalized, compared to antibodies 32,33 . In addition, their chemical synthetic process allows a low cost production and no batch to batch variability 34,35 . Furthermore, aptamers molecular weight (~ 15000Da) and charge provide rapid penetration into target tissues and elimination from the body 31,36 . Also, aptamers not immunogenic and non-toxic effects have been reported. Therefore, the characteristics of aptamers provide great advantages for their use in the development of new molecular imaging agents 37 .
Previously, we have modi ed the Sgc8-c aptamer to generate molecular imaging probes in the diagnosis of cancer [37][38][39][40] . The Sgc8-c aptamer is DNA (41 nt) and selectively recognizes the PKT7 receptor with a K d = 0.78 nM 10,41 . We have developed potential molecular imaging probes in different tumor models, through in vitro and in vivo evaluation. In this sense, radiolabelled probes (Sgc8-c-NOTA-67 Ga, Sgc8-c-DOTA-67 Ga, Sgc8-c-HYNIC-99m Tc) have been reported to have better tissue penetration and ability to accurately measure tissue, resulting in that allows quantitative images of the whole body to be obtained [37][38][39][40]. Likewise, the uorescent probe (Sgc8-c-Alexa647) allows the generation of optical images in the near infrared region with little interference, achieving optimal contrasts due to the molecules present in the tissues do not exhibit high absorption in that spectral region 37,38 . These features are very useful to perform, for example, guided surgeries in real time42,43. Even, using the appropriate imaging probes, it is possible to identify metastases 44 . Likewise, the sensitive and effective detection of PTK7 may represent a good strategy in the early diagnosis of melanoma. Due to this, the present work evaluates the potential of Sgc8-c probes in two different melanoma models, one being metastatic, in order to optimize biological control methods, both biodistribution and imaging.

Results And Discussion
PTK7 expression in B16F10 cells. Firstly, we evaluated the presence of PTK7 in metastatic melanoma B16F10 cells by ow cytometry and western blot. Previously, was evaluated in non-metastatic B16F1 melanoma cells 37 . Flow cytometry assays performed with a commercial anti-PTK7 antibody as probed reveled that approximately 40 % of the B16F10 cells expressed detectable levels of PTK7. Of note, near to 80 % of the positive control CCRF-CEM cells stained for PTK7, while the negative control, U87MG cells, showed null signal (Fig. 1A). Additionally, the Western blot studies con rmed the presence of PTK7 receptor on B16F10 cells (Fig. 1B, full-length gels are shown in Figure S1 in Supporting Information).
In vitro binding studies. Two different strategies were employed to analyze the ability of Sgc8-c probes to interact to B16F10: the use of the radiolabelled probe (Sgc8-c-NOTA-67 Ga) that can be measured by gamma counter system, and the use of the uorescent-probe (Sgc8-c-Alexa647), which allows analysis by ow cytometry and western blot.
The results showed that the Sgc8-c-NOTA-67 Ga probe binds to the B16F10 cell line. It was found that the binding percentage increased with time with signi cant differences at 4 h of incubation (Fig. 2). Regarding the blocking test of cells with the unlabeled aptamer, it was observed that the percentage of binding to the labeled probe decreases (compare 2 h of incubation and blocking incubation, p < 0.05, Fig. 2) indicating that there was a competition between both compounds con rming the probe interaction with PTK7.
For uorescent-ow cytometry assays, melanoma B16F10 cells were incubated with different concentrations of Sgc8-c-Alexa647. Figure 3 shows the percentage of positive cells and the speci c mean uorescence index (MFI). Results showed that the percentage of PTK7 positive cells is dependent on probe concentration, reaching a maximum of 80 % of B16F10 cells with approximately 0.5 µM. Likewise, the MFI indicates the amount of the probe that binds to PTK7, more precisely, the abundance of proteins at individual population cell level 45 . This suggests that the B16F10 cell line expresses signi cant high levels of the PTK7 receptor (Fig. 3).
According to cytometry analysis, saturation concentration was reached at 0.3 µM of apatmer without achieving 100 % of the B16F10 cells. This phenomenon could be explained by different cell division and diferentiation stage in cell culture. In addition, a molecular cleavage phenomenon has been described upon Sgc8-PTK7 interaction 46 . If the cleaved PTK7 receptor had been excreted into the supernatant of the medium, as result of cleavage mechanism of the aptamer-PTK7 complex, it would be the cause of not reaching saturation when the process was analyzed by ow cytometry. Then, if the cleaved PTK7 receptor is excreted into the supernatant of the medium, as result of cleavage mechanism of the aptamer-PTK7 complex, its detection by ow cytometry is not possible. However, all experiments were performed in ice so the chances of cleavage are almost null. Besides, Western blot studies using supernatants rejected this hypothesis. The results indicated that both the antibody and the uorescent probe only recognize proteins from the cell pellet ( Fig. 1B) showing that there was no detectable molecular cleavage. Based on these results the hypothesis that the fast complex internalization does not allowing saturation due to the absence of a membrane. For this reason, uorescent confocal microscopy studies were performed to demonstrate the internalization of the receptor (see below).
In vivo binding studies. Previously to perform the in vivo biological studies we analyzed the ability of the Sgc-8-c-Alexa647 probe to recognize in vivo the PTK7 presence in target organs. For this, ux cytometry studies were performed on B16F10-tumors, liver, spleen, and bone marrows as negative controls. The results showed that the uorescent-probe marked high level of tumoral cells while in the rest of the nontarget organs the percentage of positive cells were very low (Table 1). Table 1 Percentage of positive cells for in vivo exposed organs to probe Sgc8-c-Alexa647. Confocal microscopy. Another mechanism propose for Sgc8-PTK7 interaction is complex internalization 47 . For this reason, uorescent confocal microscopy studies were performed to demonstrate the internalization of the receptor on the studied tumoral cells. To perform confocal microscopy assays, the B16F10 tumor cells were incubated with the Sgc8-c-Alexa647 probe during different times. The nuclei were marked with the Hoechst marker and the membrane with WGA-green. Clearly, since the rst time of the study it was evident that the Sgc8-c-Alexa647 probe was internalized by the cells (Fig. 4).
Once it was observed that Sgc8-c-Alexa647 is internalized, it was analyzed if the probe was also colocalized within the endosomes like other aptamers-probes 47,48 . For that, tumor cells were incubated with the uorescent probe for different times and subsequently analyzed the early endosomes with Rab5 and the nuclei with Hoechst. Sgc8-c-Alexa647 was observed to co-localize within endosomes in all cell lines ( Fig. 5 and Figure S2 in Supporting Information). Similarly, it was observed that the signal from the probe shows a tendency towards polarization. It has been seen that PTK7 is not found homogeneously distributed uniformly in the cell membrane, presenting a dynamic role in cell polarization 49 .
In vivo biological studies. In vivo studies using both Sgc8-c-Alexa647 and Sgc8-c-NOTA-67 Ga probes have been performed in murine melanoma models. To do this, once the induced tumors were palpable, probes were i.v. administered and imaging and biodistribution studies were performed at different times after injection. The results showed interesting characteristics regarding the uptake of the probes in the tumor (Fig. 6, and Figures S2 and S3 in Supporting Information). Rapid tissue penetration was visualized, with tumor retention (Fig. 6 and Figure S3 in Supporting Information). Using the uorescent probe, a tumor uptake of 32.9 % was observed at 2 h post-injection (pi), increasing signi cantly at 24 h (42.4 %) and at 48 h (50.3 %).
The radiolabeled probe showed the same tendency as the uorescent probe to increase tumor uptake over time. At 2 h pi, a tumor uptake of 8.4 %ID/g was observed, increasing signi cantly at 24 h (28.8 %ID/g). Then, at 72 hours pi, tumor uptake decreased slightly (17.6 % ID/g), a non-signi cant signal comes from other organs, which allows us to observe a clear signal from the tumor for that time (Figs. 6, 9 and S3 in Supporting Information). Urinary elimination was observed, for both probes, at early time points. At 0.5 h after the injection of the uorescent probe, 3.2 % was eliminated, increasing to 42.5 % at 2 h pi ( Fig. 6 and Figure S3 in the Supporting Information) and at 0.5 h 7.1% ID was observed for the radioactive probe, which decreased at 2 h to 4.2 %ID (Fig. 6). These data were determined by the renal values that were due to the elimination of the probes or their metabolites (Figs. 6 and 8). However, it was possible to distinguish the high signals from both tumor and liver in the in vivo and ex vivo images, which were higher in the tumor at later time points (Figs. 8 and 9). In vivo blocking studies with Sgc8-c-NH 2 showed a statistically signi cant decrease of tumor signal for the uorescent probe ( Fig. 6 and Figure S3, Supporting information), con rming speci c probe training.
Differences in signal values between organs and tumor were evident in both tumoral models and with both probes, resulting in an optimal tumor/non-tumor organ ratio, having a signi cant increase over time (Fig. 7). Mainly at 48 h pi with radiolabeled probe in the metastatic model, generated by B16F10 cells, showed a tumor / muscle ratio of 35.5 ( Fig. 7B and Figure S4 for B16F1 model in Supporting Information). Likewise, it was observed that tumor uptake increased signi cantly with the time of postinjection in both models ( Figure S5 in Supporting Information). However, it was observed that the tumor uptake was slightly higher in the metastatic melanoma model, generated with the B16F10 cell line, than with the B16F1 non-metastatic melanoma model ( Fig. 6 and Figure S5 in Supporting Information). This difference could be explained by an increase in the expression of PTK7, since it has been seen that in metastatic melanomas, this receptor is one of the tyrosine kinases that participates in the positive regulation of the formation and function of the invadopodia 9 . Also, this difference in the distribution of the probes, which is also observed in the images obtained by confocal microscopy, could be affected by a structural heterogeneity in the tumor vasculature, since it has been seen that the models generated with the B16F10 line show a signi cant improvement in vascular density 50 .

Conclusions
The interesting results described herein together our previous studies 37,40 , showed that both probes developed with the Sgc8-c aptamer are potential tools for their speci c use in the early detection of melanoma. The results of in vitro studies were consistent with those obtained for biodistributions and the in vivo imaging. Obtained interesting characteristics related to the uptake of the probes in the tumor, with optimal tumor/non-target organs ratios. Even these methodologies applied here allowed us to detect differences in the expression of the tumor marker PTK7 in two different types of melanomas. However, it should be studied in depth whether this difference in the expression of PTK7 is really involved in signaling pathways that, consequently, grant greater metastatic power to cells.
The optimal tumor uptakes of the probes, in this new model of metastatic melanoma, make them promising tools to facilitate in vivo diagnosis and thus select an appropriate therapy. These probes could also be advantageous for developing intraoperative imaging devices, combined or not, the properties of both probes for use in guided surgeries; identifying and pointing out the tumor margins, helping in the surgical resection of tumors and even helpful in detecting metastases 42,51 . A deepened study with the metastatic melanoma model are currently in progress with the aim to improve the use of probes in early diagnosis that allows the selection of an e cient and personalized therapy, and even for monitoring after remission.
Our results support the potential role of the Sgc8-c-NOTA-67 Ga and Sgc8-c-Alexa647 as molecular imaging probes optimum to improve strategies in non-invasive molecular diagnosis in melanoma, as well as theranostic aproaches.

Methods
Synthesis and puri cation of Sgc8-c-NOTA-67Ga and Sgc8-c-ALEXA-647. The synthesis and puri cation of both probes were performed following previous reports from our laboratory 37,39,40 .
In vitro biological studies Confocal microscopy. To perform confocal microscopy assays, 1 × 10 5 cells from the B16F10 and B16F1 tumor cell lines were grown on round glass coverslips (12 mm) inside 24-well plates. These cells were incubated with the Sgc8-c-Alexa647 probe (10 µg, 0.8 nmol) for different times (2, 4, and 16 h). Cells were washed with sterile PBS pH 7.4 and xed with 4 % paraformaldehyde. Subsequently, the coverslips were placed in a humid chamber, the cells were blocked with 2 % bovine serum albumin in PBS for 20 min at RT and then they were blocked for an additional 15 min, also at RT, with the same solution but adding Triton (0.3 %). Cells were incubated for 1 h at RT with the Hoechst 33342 nuclear marker (1:100, ImmunoChemistry Technologies, LLC) and with the WGA-green membrane marker (1:100, thermo sher scienti c, USA). They were washed three times with PBS pH 7.4 and three more times with mili Q water. The coverslips were mounted with ProLong® (thermo sher Scienti c, USA) and the images were taken in the confocal microscope LEICATCS-SP5-DMI6000 (HeNe laser, 10mW: 633 nm). To determine if the probe was internalized endosomically, we followed the same protocol as before, incubating for 0.5, 2 and 4 h and instead of using a membrane marker, the early endosomal marker Rab5 (rabbit, 1:100, C8B1 mAb 3547, Cell Signaling Technology, USA) was used. The secondary antibody anti-rabbit IgG, Alexa 488conjugated (goat, 1:500, ab 150077 Abcam, USA) was used.
In vivo biological studies (racks with ltered air) at 20 ± 2 o C with cycle of 14 hours of light and 10 hours of darkness. They were fed ad libitum to standard pellet diet and given water ad libitum and were used after a minimum of 3 days acclimation to the housing conditions. Animals were monitored daily, recording their behavior and the presence or absence of tumor. Tumor location and volume was recorded, and checked that they did not exceed a diameter of 5 mm. Iso urane was used for anesthesia and at the end of the experiments the animals were sacri ced by cervical dislocation.
Binding studies for uorescent-probe. For this assay 2.5 × 10 5 cells/100 mL of the B16F10 cell line were injected subcutaneously into the right ank of C57BL/6 mice. Once tumors were palpable (10-12 days), the Sgc8-c-Alexa647 probe (25 µg, 2 nmol) was injected intravenously (i.v.) through the tail vein. At 0.5, 2 and 24 h post injection, mice were sacri ced to obtain the tumor, liver, spleen and marrow derived from the femur. Organs and tissues were disaggregated by passing through a cell strainer 70 µm (BD Bioscience) and resuspended in sterile PBS pH 7.4. For each sample, 10000 events were detected using the same laser, detector, and equipment mentioned above. The test was done in quintupled. Data were analyzed using FACS Diva and FlowJo software.
Imaging and biodistribution. To generate the melanoma tumor model, murine cell line B16F10. Tumors were induced in the C57BL/6 mice as described above for B16F10. Once melanomas were palpable, the Sgc8-c-Alexa647 probe (25 µg, 2 nmol) was injected i.v. and at 0.5, 2, 24 and 48 h post-injection (n = 5 per time group), mice were sacri ced. Ex vivo images of organs (liver, heart, lungs, spleen, kidneys, thyroid, muscule, bone, blood and tumor) were taken using the imaging equipment above mentioned, with the Xray and uorescence model. This study was compared with the melanoma tumor model generated with B16F1 cells. The results were expressed in ROI and the tumor/blood and tumor/muscle ratios were calculated. A competition test was performed. One group of mice (n = 5) were rst i.v. injected with Sgc8c-NH 2 in excess of 5 times more than the probe. After 30 min post-injection, the same mice were i.v.
injected with the Sgc8-c-Alexa647 probe. After 2 h mice were sacri ced, images taken and organs weighed. Similar evaluation was performed for the radiolabelled probe. Sgc8-c-NOTA-67Ga was i.v. injected (~ 1850 kBq) and biodistribution of the probe was following until 72 h post-injection. Over time, the animals were anesthetized with iso urane to perform the images in vivo. Live images were aquired for 0.5, 2, 24, 48 and 72 h after injection by X-rays and gamma modalities in the imaging equipment. For the biodistributions animals were sacri ced, images were acquired and the levels of radioactivity in the tissue were measured using a gamma counter (PC-RIA MAS, Stratec). Radioactivity levels were expressed as percentage of injected dose per organ gram (%ID/g) and injected dose (%ID). Blocking assays were also performed, injecting mice rst with 0.5 nmol of Sgc8-c-NH 2 and after 30 min re-injected with Sgc8-c-NOTA-67Ga and at 2 h the imaging and biodistribution studies were performed.
Statistical analysis. Statistical analysis was performed using the Student's t-test and the p values of signi cance indicated in each gure.

Competing interests
The authors declare no competing interests.