Gastrointestinal stromal tumor (GIST), the most common sarcoma, is characterized by KIT protein overexpression, and tumors are frequently driven by oncogenic KIT mutations. Targeted inhibition of KIT revolutionized GIST therapy and ushered in the era of precision medicine for the treatment of solid malignancies. Here, we present the first use of a KIT-specific DNA aptamer for targeted labeling of GIST. We found that an anti-KIT DNA aptamer bound cells in a KIT-dependent manner and was highly specific for GIST cell labeling in vitro. Functionally, the KIT aptamer bound extracellular KIT in a manner similar to KIT mAb staining, and was trafficked intracellularly in vitro. The KIT aptamer bound dissociated primary human GIST cells in a mutation agnostic manner such that tumors with KIT and PDGFRA mutations were labeled. In addition, the KIT aptamer specifically labeled intact human GIST tissue ex vivo, as well as peritoneal xenografts in mice with high sensitivity. These results represent the first use of an aptamer-based method for targeted detection of GIST in vitro and in vivo.
Gastrointestinal stromal tumor (GIST) is the most common sarcoma with approximately 3,000 new malignant cases in the United States annually (1). The treatment of GIST provided the first proof of principle for precision medicine in solid tumors as oncogenic-driver mutations in the KIT gene were identified and effectively targeted with a tyrosine kinase inhibitor (TKI), imatinib. KIT-mutated GIST represents approximately 60%–70% of disease and oncogenic KIT activation is associated with high levels of KIT expression (2, 3). However, non-KIT–mutant GIST also highly expresses KIT likely due to converging pathway activation (4). Currently, GIST diagnosis relies on analyzing tissue procured from either biopsy or surgical specimens. Although KIT-expressing GISTs are effectively diagnosed with IHC staining with anti-KIT antibodies, this approach requires ongoing hybridomas for production.
Aptamers are single-stranded oligonucleotide (DNA or RNA) ligands that are selected against specific targets (proteins and small molecules) through an in vitro iterative process called systematic evolution of ligands by exponential enrichment (SELEX; refs. 5, 6). Aptamers undergo intramolecular base pairing between complementary nucleotides and assume secondary, followed by tertiary structures. In turn, these can bind to their cognate targets with high affinity and specificity similar to antibodies. In contrast, aptamers demonstrate enhanced tissue penetration due to their small sizes and are nonimmunogenic as compared with antibodies. Aptamers are also amenable to a variety of chemical modifications and can be conjugated with compounds such as fluorophores or drugs. Thus, modified aptamers can be utilized for several applications including in vitro and in vivo imaging, as well as targeted drug delivery (7, 8).
Recently, Zhao and colleagues presented a ssDNA aptamer that was developed for use in an acute myeloid leukemia (AML) model that highly expresses KIT (9). The investigators utilized a hybrid SELEX method that involves sequential exposure of a library of ssDNA oligonucleotides to AML cells in vitro followed by enrichment with recombinant KIT. In their study, the final selected oligonucleotide (aptamer) was conjugated to methotrexate and shown to be cytotoxic to AML lines and patient-derived samples. In another study, Tanno and colleagues used the same aptamer to develop a miRNA-aptamer chimera to deliver miR-26a against chemotherapy-mediated myelosuppression (10). In both study studies, the anti-KIT aptamer was tested in the hematologic malignancies. However, the ability of an anti-KIT aptamer to bind and label solid tumor cells remains untested.
Here, we present a study of the utility of an anti-KIT aptamer in the detection of GIST. We find that the anti-KIT aptamer binds GIST cells in a KIT-dependent manner and is trafficked internally. Importantly, the aptamer binds to live and fixed primary human GIST cells, allowing for broad labeling applications. Finally, we found that the KIT aptamer conjugated to a fluorophore can detect GIST cells both ex vivo and in vivo. This study provides the first proof of principle that an anti-KIT DNA aptamer may be used for targeted detection for human GIST, irrespective of driver mutation status.
Materials and Methods
Written informed consent was obtained for all study participants, including publication of clinical data. Patient tissue collection, acquisition of clinical data, and conducting experimental procedures on biological samples was approved by the UC San Diego Human Research Protections Program Institutional Review Board (IRB; Protocol #181755, La Jolla, CA). Pathologic diagnosis was made by an experienced pathologist based on light microscopic analysis of optimal cutting temperature (OCT) tissue sections and sections labeled with antibodies against KIT and DOG-1 (discovered in GIST-1), a membrane associated antigen that is expressed in GIST irrespective of KIT or PDGFRA mutations (11). All experiments were conducted with deidentified tissues in accordance with appropriate regulatory guidelines for use of human tissue.
ssDNA KIT aptamer sequence was obtained from the primary literature report (5′-GAGGCATACCAGCTTATTCAAGGGGCCGGGGCAAGGGGGGGGTACCGTGGTAGGACATAGTAAGTGCAATCTGCGAA-3′; ref. 9). Scrambled aptamer sequence was generated through a random oligonucleotide sequence generator and constrained to have equivalent free energy to the KIT aptamer sequence (5′-TGACGGGAGACTTAAAACGCAAGGGGTGCAGCTATCGCGG AGGCCAAGGGTTCAAGTCGACGGGTAGCTAGGTTGGA-3′; Oligo Calculator version 3.27, biotools.nubic.northwestern.edu). The aptamer and scrambled sequences were synthesized as an unmodified oligonucleotide, 5′-biotin modification, and 5′-Cy5.5 fluorophore modification (Integrated DNA Technologies) for various applications.
We obtained the GIST-T1 line containing a KIT exon 11 (heterozygous KIT p.V560_Y578del) mutation (12) from T. Taguchi (Kochi Medical School, Nankoku, Japan) and the GIST882 line containing a KIT exon 13 (homozygous KIT K642E) mutation (13) from S. Singer (Memorial Sloan Kettering Cancer Center, New York, NY). Pancreatic cancer cell lines, Panc-1 and MiaPaCa-2 were obtained from ATCC. The mouse ovarian surface epithelial cell line ID8 was provided by D. Schlaepfer (University of California, San Diego, La Jolla, CA), and the human ovarian cancer cell line SK-OV-3 was provided by D. Stupack (University of California, San Diego, La Jolla, CA). All cell lines were cultured as reported previously. GIST-T1, ID8, Panc-1, MiaPaCa-2, and SK-OV-3 were grown in DMEM with 10% FBS, 1% penicillin/streptomycin (Mediatech), and 2 mmol/L glutamine (Mediatech; refs. 12, 14–16). GIST882 were grown in RPMI with 20% FBS, 1% penicillin/streptomycin (Mediatech), and 2 mmol/L glutamine (Mediatech; ref. 13). The human mast cell line HMC 1.2 (obtained from I. Pass, Sanford Burnham Prebys Medical Discovery Institute, San Diego, CA) was cultured in Iscove's modified Dulbecco's Medium (Gibco) with 10% FBS, 1% penicillin/streptomycin, and 1.2 mmol/L 1-Thioglycerol (Sigma; ref. 17).
ID8 cells, a mouse ovarian cancer cell line, were used to establish a conditional KIT expression cell line using the Sleeping Beauty–based transposon system. Briefly, ID8 cells were cotransfected with plasmids containing the KIT gene (flanked by the inverted terminal repeats) and a transposase-encoding plasmid (18, 19). Stable transgenic cells were selected with hygromycin. KIT expression was controlled by a doxycycline-inducible promoter and transgenic cells were detectable by constitutive GFP expression.
In addition, a GIST-T1 with constitutive GFP expression was created for use in mouse imaging experiments. Briefly, the GIST-T1 line was transduced by lentivirus with GFP-expressing plasmid. Stable cell lines were created by puromycin selection.
Cells were harvested from monolayer cell cultures using trypsin 0.05% (HyClone) or Accutase (Sigma-Aldrich). Cells were washed in cold PBS with 1% FBS. 5′-biotynylated aptamers were incubated with streptavidin-phycoerythrin (SA-PE, ProZyme) for 10 minutes in the dark at room temperature. Cells were then treated with fluorophore-conjugated aptamer for 1 hour at room temperature. Cells were then washed with PBS buffer three times and resuspended in PBS for flow cytometry analysis (BD FACSCalibur). PE anti-human c-KIT antibody (Clone 104D2, BioLegend) was applied in 1:20 dilution for cell staining. FlowJo software was used to analyze the flow cytometry data.
Ligand competition assay
Aptamer specificity was also tested through a competition experiment using the KIT ligand also known as stem cell factor (SCF). GIST-T1 cells were incubated with biotinylated-human-SCF (Arco Biosystems) conjugated to streptavidin-PE in the presence of unmodified anti-KIT or scrambled aptamer for 1 hour. Cells were washed thoroughly and resuspended in PBS for analysis by flow cytometry to quantify SCF-bound cells in each condition.
Confocal immunofluorescence microscopy
Cells were plated on glass-bottomed well slides and cultured to 50% confluency. Cells were washed thoroughly with PBS and then incubated with preconjugated aptamer-SA-PE (400 nmol/L) for 1 hour at room temperature. Cells were then fixed in 4% paraformaldehyde (Thermo Fisher Scientific), washed, and counter-stained with DAPI (1:50,000, Thermo Fisher Scientific).
Fluorescent dextran 10 was used for probing general fluid-phase endocytosis (macropinocytosis and micropinocytosis; ref. 20). We used CF488A Dextran (Biotium, Inc.) to study the localization of the anti-KIT aptamer on nonpermeabilized live cells. A total of 2 × 105 GIST-T1 cells were seeded in a 24-multiwell plate (Ibidi USA, Inc.) 24 hours before the treatment. Cells were treated with 10 kDa CF488A Dextran (3 nmol/L) and aptamer-PE conjugate (400 nmol/L) for 1 hour at 37°C. Cells were thoroughly washed with dPBS with MgCl2 and CaCl2 (Thermo Fisher Scientific) and fixed for 5 minutes with 1% paraformaldehyde (PFA). Fixed cells were then washed again and counter stained with DAPI (1:500; GeneTex, Inc.). Z-stacks at 63× magnification with LSM880 confocal (Zeiss) were acquired. The images were analyzed with Zen Software (Zeiss) to measure the localization of fluorescence signals. For each z stack, the maximum intensity projection was generated followed by fluorescence intensity plots to measure and calculate individual channel intensity.
OCT sections (7 μm) of primary human-resected GISTs were obtained under our IRB-approved protocol. Sections were washed thoroughly, labeled with PE-labeled aptamer for 1 hour at room temperature, and permeabilized with 70% ethanol. OCT sections were then counter-stained with DAPI. Fluorescence intensity was visualized with A1R confocal microscope (Nikon).
Cell lysate preparation and Western blotting
Cells were homogenized in RIPA Buffer (Thermo Fisher Scientific) containing protease and phosphatase inhibitors (catalog no. A32959, Thermo Fisher Scientific). Protein concentrations were determined. The lysates were then loaded and separated by SDS-PAGE before transfer to a nitrocellulose membrane (Bio-Rad). Membranes were incubated with primary anti-KIT antibody (1:1,000, Cell Signaling Technology). Secondary antibody horseradish peroxidase–conjugated anti-mouse IgG (1:5,000, Invitrogen) was added and antibody complexes were detected by the ECL system.
Primary tumor dissociation and single-cell suspension
Fresh tumor tissue was dissociated into single-cell suspensions using the gentleMACS Dissociator (Miltenyi Biotec) as described previously (21). Solid tissues were cut into 5-mm size pieces and were transferred to a gentleMACS C-Tube containing RPMI media and MACS human tumor dissociation enzyme cocktail (Miltenyi Biotec) according to the manufacturer's instructions for tough tumor tissue (h_Tumor_01). The sample was then passed through a 70-μm filter, and tumor cells were collected following centrifugation. Cells were then labeled using the flow cytometry approach described above.
Cell viability assay
Single-cell suspensions of tumor cells were seeded at 5,000 cells per well on a 96-well plate (Corning). The cells were grown for 48 hours and subsequently treated with 5, 10, or 20 μmol/L of unmodified KIT or scrambled aptamer for 24-, 48-, or 72-hour time points. Cell viability was analyzed by CellTiter-Glo Luminescent Assay (Promega) and luminescence measured on the Tecan Infinite 200 Microplate Reader (Tecan).
Generation of tumor xenograft models
Tumor xenograft models were created using the GIST-T1 cell line. A total of 5 × 106 cells were injected subcutaneously into the right flank of 5- to 6-week-old nu/nu mice (The Jackson Laboratory) and allowed to grow until tumor volume was 100–200 mm3. Mice were then euthanized and tumors were harvested for ex vivo analysis.
GFP-labeled GIST-T1 cells were also used to create an intraperitoneal xenograft GIST model. Five- to 6-week-old nu/nu mice (The Jackson Laboratory) were injected with 5 × 106 cells intraperitoneally. Mice were monitored weekly for tumor growth by visual inspection and In Vivo Imaging System (IVIS) imaging as described below.
Ex vivo detection and animal imaging
Ex vivo imaging was performed on resected subcutaneous GIST-T1 xenografts. Tumor pieces were labeled with preconjugated aptamer-SA-PE (400 nmol/L) in PBS for 1 hour and then washed thoroughly. Tumor pieces were then imaged using the IVIS system (PerkinElmer). Imaging and quantification were performed with Living Image software (PerkinElmer).
Preparation for in vivo imaging was done by feeding with an alfalfa-free diet prior to the date of imaging. Anesthesia was induced with isoflurane (2.5% isoflurane, 3 L/minute, 5 minutes). Tumor growth was monitored by placing mice in the supine position and GFP fluorescence signal was detected using the IVIS system. After 1 month from intraperitoneal injection of cells, mice were anesthetized and injected intraperitoneally with aptamer (either KIT or scrambled). Mice were randomly assigned to treatment groups. The aptamer used for live imaging experiments were either directly conjugated to Cy5.5 or preconjugated with streptavidin-Cy5.5 (Rockland Immunochemicals) with a biotinylated aptamer. Two hours after injection, mice were euthanized, a laparotomy was performed, and in situ imaging was performed. Images were acquired using the GFP and Cy5.5 filters, which were normalized to background signal per the manufacturer's protocol.
Statistical analyses were performed using Prism GraphPad 7 (GraphPad Software). The investigators were not blinded to allocation during experiments or outcome assessment. Two-tailed unpaired Student t tests with Welch correction or one-way ANOVA for multiple comparisons when appropriate were used to determine statistical significance.
KIT aptamer specifically binds GIST cells
We first tested the specificity of the anti-KIT aptamer to bind cells in a KIT-dependent manner. A stable transgenic ID8 cell line was established using a transposon-based plasmid for the integration of c-KIT (human) gene into the mouse cell genome. The cell line had c-KIT gene expression under a doxycycline-inducible promoter, as verified by Western blot analysis (Fig. 1A). Next, KIT protein expression on the cell membrane was assayed by staining nonpermeabilized cells with fluorophore-conjugated anti-KIT aptamer followed by flow cytometry. An anti-KIT mAb was used as a positive control to compare the binding of the anti-KIT aptamer. Flow cytometry analysis demonstrated that both the anti-KIT aptamer (32.8% vs. 0.05%) and antibody (29.3% vs. 0.04%) bound KIT-expressing cells only in the presence or absence of doxycycline, respectively, indicating that binding of the aptamer is comparable with the monoclonal anti-KIT antibody (Fig. 1B).
We next confirmed high KIT protein expression in two GIST cell lines [GIST-T1 (KIT V560-Y579Δ5) and GIST882 (KIT K642E)] by Western blot analysis (Fig. 1C). Two pancreatic cancer cell lines, Panc1 and MiaPaca2 were utilized as negative controls and had undetectable KIT expression (Fig. 1C). The control scrambled and KIT aptamers were tested in a dose titration experiment by labeling 1 × 106 GIST-T1 cells to determine the optimal concentration for aptamer-target binding. The aptamer demonstrated binding in a dose-dependent manner, while the control scrambled sequence had little binding at any dose, further demonstrating the specificity of the anti-KIT aptamer against the KIT protein expressed on the cell surface (Fig. 1D). The aptamer binding was saturated at 400 nmol/L, and this concentration was used for all subsequent experiments.
Next, flow cytometry analysis was performed on the two KIT-mutant GIST cell lines. The geometric mean of the histogram plots from each treatment was divided by the geometric mean of the corresponding unstained cells to obtain the normalized geometric mean (GM) value. Two independent experiments were performed for each condition. We observed that the anti-KIT aptamer fluorescence intensity was significantly higher than scrambled control aptamer for GIST-T1 (normalized GM 48 vs. 3.1) and GIST882 (normalized GM 52.6 vs. 4.2). In contrast, the pancreatic cancer cell lines had KIT aptamer binding that was low magnitude and was comparable with the scrambled aptamer with Panc1 (normalized GM 3.1 vs. 2.5) and MiaPaCa2 (normalized GM 2.9 vs. 1.9, respectively). In both positive and negative cell lines, KIT antibody had comparable binding patterns with the KIT aptamer: (GIST-T1 normalized GM 48.0 vs. 65.2; GIST882 normalized GM 52.6 vs. 45.4; Panc1 normalized GM 3.1 vs. 1.2; and MiaPaCa2 GM 2.8 vs. 1; Fig. 1E and F). In addition, the anti-KIT aptamer reduced binding of PE-labeled KIT ligand (SCF) compared with a negative control (normalized GM 9.5 vs.57.9, respectively), indicating that aptamer binding resulted in inhibition of SCF binding to KIT. The control scrambled aptamer was not associated with similar inhibition of SCF binding (normalized GM 57.9 vs. 59.5, respectively; Fig. 1G and H). Taken together, these experiments demonstrate that this anti-KIT aptamer binds the KIT receptor in a specific manner.
KIT aptamer binds to other cancer cells expressing KIT
We also tested a KIT-mutant mast cell line, HMC 1.2 (KIT G560V and D816V; refs. 22, 23), and a human ovarian cancer cell line SK-OV-3 (KIT wild-type; cBioPortal for Cancer Genomics) for anti-KIT aptamer binding. Higher expression of KIT receptor protein was detected by Western blot analysis in the HMC 1.2 as compared with the SK-OV-3 cell line (Supplementary Fig. S1A). This difference was also detected in the flow cytometry data with higher KIT aptamer binding to the HMC 1.2 cells (normalized GM 94.2) as compared with the SK-OV-3 cells (normalized GM 28.3). The scrambled aptamer had lower background binding in both the HMC 1.2 and SK-OV-3 cells (normalized GM 1.9 and 6.8, respectively; Supplementary Fig. S1B and S1C). Taken together, the data demonstrates that the KIT aptamer can be a useful reagent for KIT detection in other cancers that overexpress this receptor.
KIT aptamer cellular localization
We next examined cellular localization of the anti-KIT aptamer in GIST-T1 and GIST882 cells by immunocytochemistry. Both cell lines demonstrated a similar pattern of anti-KIT aptamer localization at punctate foci consistent with either plasma membrane or intracellular aggregates (Fig. 2A and B). In contrast, the scrambled aptamer demonstrated significantly less fluorescence and was generally distributed in a diffuse pattern that may be due to nonspecific cell binding.
Cellular localization of the anti-KIT aptamer was further probed by costaining with fluorescent dextrans, which are readily internalized through fluid-phase endocytosis. Confocal immunofluorescence demonstrated that the anti-KIT aptamer colocalizes with the CF488A dextran. In addition, the analysis of the fluorescence intensity profiles of the signal for the aptamer and the dextran indicated colocalization of the anti-KIT aptamer with cellular vesicles. The scrambled aptamer did not colocalize with any structure in the cells. This data strongly suggests that the aptamer is internalized in GIST cells following KIT receptor binding and is retained within intracellular vesicles (Fig. 2C–E).
KIT aptamer colocalizes with KIT antibody
Next, we examined the binding and localization of the KIT aptamer as compared with KIT antibody. Cells were coincubated with both the anti-KIT antibody and anti-KIT DNA aptamer. Cells were then analyzed by flow cytometry (Fig. 3A and B). The fluorescence intensity of the KIT aptamer binding was diminished in the presence of KIT antibody (GIST-T1: GM 246.0 vs. 123.0; GIST882: GM 146.0 vs. 62.0, respectively). This observation suggests that there is either competitive or allosteric targeting of the KIT receptor. This was evident in both the GIST-T1 and GIST882 cell lines. This observation was further corroborated by immunocytochemical analysis. Coincubation of the anti-KIT aptamer and antibody with live cells resulted in colocalization of signals suggesting that both molecules bound similar aggregates of KIT receptors (Fig. 3C).
KIT aptamer has minimal effect on cell viability
Treatment with anti-KIT aptamer inhibits SCF binding to GIST cells (Fig. 1G and H). In turn, this may inhibit KIT receptor-mediated cell signaling pathways and cell viability. We tested this in GIST-T1 and GIST882 cells by assessing cell viability following treatment with unmodified scrambled or KIT aptamer. Cells were treated at different concentrations (5, 10, and 20 μmol/L), and analyzed for cell viability at 24-, 48-, and 72-hour time points. GIST-T1 cells had a decrease in cell viability after treatment with the aptamer compared with the scrambled control [5 μmol/L at 24 hours: 90.7% (SD, 2.0), 48 hours: 87.6% (SD, 0.7), and 72 hours: 73.9% (SD, 5.9); P < 0.001]. Although there as an observed time-dependent effect, the change in cell viability did not differ between the 5, 10, and 20 μmol/L doses (Fig. 4A). In the GIST882 cell line, there was no significant change in cell viability over time [5 μmol/L at 24 hours: 94.2% (SD, 8.9), 48 hours: 101.8% (SD, 2.1), and 72 hours: 98.5% (SD, 3.6); P = 0.88; Fig. 4B]. Similar values were observed for all doses tested. Although there was a difference between the sensitivity of the GIST-T1 and GIST882 cell line, the anti-KIT aptamer generally had minimal or no intrinsic cytotoxic effect on GIST cells. In addition, the aptamer concentrations tested here were several log-fold higher than the concentration used for cell binding experiments (400 nmol/L). This strongly suggests that the aptamer has negligible adverse cytotoxic effect when utilized for in vitro or in vivo applications (24).
KIT aptamer binds primary human GIST cells
Next, we tested the capability of the anti-KIT aptamer to bind primary human GIST cells. We utilized several tumor samples encompassing a range of primary tumor locations and mutational profiles (Fig. 5A). KIT aptamer bound cells with several log-fold higher affinity than the scrambled control in all samples tested (GIST#1: GM 35.8 vs. 8.1; GIST#2: GM 407.0 vs. 99.0; GIST#3: GM 239.0 vs. 12.4, respectively). Moreover, we compared the aptamer staining with KIT antibody staining. Fluorescence intensity of the aptamer was higher than antibody in GIST #2 (GM 407.0 vs. 173.0), similar in GIST #3 (GM 239.0 vs. 313.0), and lower in GIST#1 (GM 35.8 vs. 114, respectively; Fig. 5B and C).
Next, we tested the ability of the KIT aptamer to label cells from tissue specimens on glass-mounted slides. Two resected GISTs were labeled with either KIT or scrambled aptamer and visualized by confocal microscopy (Fig. 5D). KIT aptamer binding demonstrated diffuse binding with some perinuclear enhancement. The staining pattern likely differs from prior immunocytochemistry as these samples were permeabilized at the time of labeling. The staining pattern was similar for a GIST with a KIT exon 11 mutation (GIST #4) and one with a PDGFRA mutation (GIST #5). In both cases, scrambled aptamer signal was faintly detectable. This indicates a potential application of the anti-KIT aptamer as a reagent for the pathologic identification of primary patient GISTs by aptamer-based staining.
KIT aptamer binds GIST tumors ex vivo and in vivo
Finally, we tested the capacity of the KIT aptamer to bind GIST xenografts either ex vivo or in vivo. First, we harvested GIST-T1 xenografts grown subcutaneously in nu/nu mice. Tumors were fragmented and labeled with either anti-KIT or scrambled aptamer. Retention of aptamer signal was assessed using the IVIS system. Representative images from three xenograft tumors are shown (Fig. 6A). KIT aptamer had a higher proportion of PE-positive tumors (11/13, 84.6%) compared with scrambled aptamer (2/10, 20%). Moreover, the median fluorescence intensity for the scrambled aptamer was significantly lower than for the KIT aptamer (5.6 × 107 vs. 7.5 × 107 counts; P = 0.01; Fig. 6B).
Next, we examined KIT aptamer binding in an intraperitoneal model of GIST. GIST-T1-GFP cells were used for this model to visualize intraperitoneal tumor burden. One month after intraperitoneal injection, mice underwent intraperitoneal injection of either KIT-Cy5.5 or scrambled-Cy5.5 aptamer. Binding was assessed after 2 hours by IVIS (Fig. 6C). Accuracy of tumor binding was assessed by the ratio of Cy5.5 signal to GFP signal for each individual tumor, thus acting as internal controls. Forty percent (4/10) of the scrambled aptamer group had Cy5.5 signal detected above background fluorescence, while 85% (11/13) of the KIT aptamer group had Cy5.5 signal detected above background fluorescence (Fig. 6D). In addition, the median signal intensity of scrambled aptamer overlapping with GIST-T1-GFP signal was significantly lower than that of the KIT aptamer signal (−2.1 × 106 vs. 8.6 × 107 counts; P = 0.002; Fig. 6E). Collectively, these results suggest that scrambled aptamer binding was nonselective and comparable with background signal, while the KIT aptamer had a high rate of tumor detection and strong signal intensity at the multifocal sites of disease.
Gastrointestinal stromal tumor is a disease with well characterized diagnostic markers, including KIT, that also serve as a cognate drug target. Despite widespread use of targeted KIT inhibitors for clinical treatment of GIST, a targeted diagnostic probe is not clinically available. Here, we present the first use of a KIT-specific DNA aptamer for targeted labeling of GIST. We found that the KIT aptamer bound cells in a KIT-dependent manner and is specific for GIST cells, but not cancer cell lines lacking KIT expression. The KIT aptamer appears to bind the extracellular domain of KIT and is trafficked intracellularly. The aptamer also colocalizes with KIT antibody. Importantly, the KIT aptamer binds primary human GIST cells in a driver mutation–independent manner, suggesting that it has the potential for broad applications. Finally, the KIT aptamer specifically labels GIST tissue ex vivo and in vivo. These results represent the first use of an anti-KIT DNA aptamer-based method for targeted detection of GIST.
Although GIST is often associated with KIT mutations, there are a variety of other genes that are implicated in GIST tumorigenesis, including PDGFRA, SDHx subunits, RAS pathway (KRAS, NF1, and BRAF), and gene fusions such as ETV6-NTRK3 and FGFR1 fusions (25). However, even non-KIT–mutant GIST frequently has KIT surface expression. This fact has been leveraged for clinical GIST diagnosis through IHC (4). In this study, we found that this KIT aptamer efficiently labeled GIST cells that harbor KIT exon 11, KIT exon 13, and PDGFRA mutations. In addition, the aptamer was effective at labeling both live primary human GIST cells and glass-mounted fixed, permeabilized cells. These scenarios underscore the potential applications that a KIT aptamer may be utilized for in GIST diagnosis.
But it is important to consider that as compared with immortalized cell lines, primary human tumors are markedly heterogeneous. We found that primary human GISTs had differential labeling with both the KIT aptamer and the KIT antibody control. We found that the aptamer had higher, lower, or equivalent fluorescence intensity as antibody. However, these differences reflect the inherent variations related to tumor heterogeneity and the expected limitations with any single diagnostic approach.
There have been several studies that have tested the capacity of KIT–antibody conjugates to perform a variety of tasks based on KIT-specific targeting. Prior work by our group has demonstrated the feasibility of a near-IR–conjugated KIT antibody to be used for intra-abdominal imaging of GISTs (26). Another group developed a radiolabeled KIT antibody for imaging in a mouse model (27). Several groups have utilized KIT antibodies as a targeting system for delivery of chemotherapeutic agents or other cytotoxic approaches (28, 29). Others have utilized directly KIT-blocking antibodies to abrogate KIT signaling (30, 31). During preparation of this article, Shraim and colleagues reported the development and characterization of 2′ fluoro-pyrimidine–modified RNA aptamers against the KIT receptor kinase domains (KIT wild-type, KIT D816V, and KIT D816H). Interestingly, one of the selected aptamers, V15, specifically inhibited the in vitro kinase activity of mutant KIT D816V with an IC50 value that was 9-fold lower than sunitinib (32). Here, we present the first usage of a DNA oligonucleotide–based approach for targeted detection of KIT-expressing solid tumors because aptamers have several advantages to antibody-based approaches for cell targeting. First, aptamers are far easier to synthesize than antibodies, which require hybridoma maintenance and complex synthesis protocols. Moreover, aptamers are dynamic molecules that can be readily conjugated to a variety of molecules making them well suited for a variety of clinical applications if the target specificity is high. Finally, aptamers are small molecules that may have superior tissue distribution and are less likely to promote immunogenicity compared with antibodies. As demonstrated in the primary study by Zhao and colleagues and reinforced by our findings, this KIT aptamer is an excellent candidate for further development into a clinical diagnostic tool (9).
The FDA approval of pegaptanib (pegylated anti-VEGF aptamer) marked the introduction of clinical use of aptamer-based therapeutics (33). Imaging and localization studies of the KIT aptamer suggest cellular internalization, which raises the possibility of theranostic applications. For example, conjugation to clinically relevant TKIs (i.e., imatinib, sunitinib, or regorafenib) may improve drug delivery permitting dose reduction and increased tolerability of these drugs. Cytotoxicity assays suggest that the KIT aptamer itself has minimally intrinsic cytotoxicity. However, conjugation to a cytotoxic molecule may enable a KIT–aptamer conjugate to serve as a targeted treatment for GIST. In addition, there are several diagnostic clinical roles for the KIT aptamer. Whole body imaging with a KIT aptamer probe could provide tumor-specific diagnosis, as well as monitoring of treatment response. The KIT aptamer also could potentially be adapted for in vitro use as a liquid biopsy surveillance tool to detect GIST cells in patients at high risk for disease recurrence.
In conclusion, for the first time we report aptamer labeling of human GIST cells, including primary human tumor cells. KIT aptamer labeling was equivalent or superior to the anti-KIT antibody and bound a similar distribution to KIT molecules in vitro and in vivo. Taken together, these studies provide proof of principle for investigating the utility of anti-KIT aptamers for developing novel GIST diagnostics.
Disclosure of Potential Conflicts of Interest
J.K. Sicklick is a consultant (paid consultant) at Loxo Oncology and Deciphera, reports receiving a commercial research grant from Foundation Medicine, other commercial research support from Amgen, and has received speakers bureau honoraria from Roche. No potential conflicts of interest were disclosed by the other authors.
Conception and design: S. Banerjee, J.K. Sicklick, P. Ray
Development of methodology: S. Banerjee, M. Gilardi, J.S. Shankara Narayanan, R.R. White, J.K. Sicklick, P. Ray
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): S. Banerjee, M. Gilardi, J.S. Shankara Narayanan, P. Ray
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): S. Banerjee, M. Gilardi, J.S. Shankara Narayanan, R.R. White, J.K. Sicklick, P. Ray
Writing, review, and/or revision of the manuscript: S. Banerjee, M. Gilardi, J.S. Shankara Narayanan, J.K. Sicklick, P. Ray
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): S. Banerjee, H. Yoon, M. Yebra, C.-M. Tang, P. Ray
Study supervision: S. Banerjee, J.K. Sicklick, P. Ray
We appreciate funding support from the Surgical Society of the Alimentary Tract Mentored Research Award (to S. Banerjee) and NIH T32 CA121938 Cancer Therapeutics (CT2) Training Fellowship (to S. Banerjee). In addition, we appreciate funding support from Hope for a Cure Foundation, The Life Raft Group, Kristen Ann Carr Fund, Lighting the Path Forward for GIST Cancer Research, The David Foundation, NIH K08 CA168999, NIH R01 CA226803, and FDA R01 FD006334 (all to J.K. Sicklick). Finally, this work was supported by UC San Diego Health Sciences Research Grant (to J.K. Sicklick and P. Ray). We thank Biorepository and Tissue technology shared resource for biospecimen collection. Biorepository and Tissue technology shared resource is supported by CCSG Grant P30CA23100. We thank the Cancer Center Microscopy core for providing microscopes and imaging systems. The Cancer Center Microscopy core is supported by the UCSD Specialized Cancer Center Support Grant (NCI) P30 2P30CA023100-28. The microscope LSM880 is supported by NIH S10OD021831 grant to the La Jolla Institute's imaging core facility. We would also like to thank the UCSD Academic Senate Health Sciences Research Grant awarded to P. Ray and J.K. Sicklick for funding the project.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.