Tissue Samples and Immunohistochemistry

To ascertain the expression of sodium-glucose cotransporter 2 (SGLT2) in human lung cancers, we first compared SGLT2 staining across an array of normal lung tissue and pulmonary adenocarcinomas using immunohistochemistry (IHC). A tissue microarray with normal lung tissue was obtained from US Biomax (LC1201b, Rockville, MD, USA). Additionally, 44 tissue samples were obtained from patients at the University of Pennsylvania between 2023–2024 who underwent pulmonary wedge resection or lobectomy for lung cancer. The University of Pennsylvania Institutional Review Board approved this study, and all subjects provided written informed consent for use of their tissues.

IHC staining for SGLT2 was performed through a process of deparaffinization, rehydration, and washings in xylene, graded alcohols, and distilled water. The samples were then placed in 10 mM citrate buffer at pH 6 with subsequent microwave antigen retrieval procedure and then incubated with purified anti-SGLT2 monoclonal antibody (ab58298, Abcam, Cambridge, MA, USA) at 1:1000 dilution. The antigen–antibody reaction was visualized using the avidin–biotin-peroxidase complex and diaminobenzidine as the chromogen. The slides were counterstained with hematoxylin.

Technical adequacy of the staining was validated by normal kidney and lung parenchyma as known positive and negative controls, respectively. Scoring was performed by a board-certified pulmonary pathologist. Staining of the tumor cells was scored based on the intensity of the staining as negative (0), mild (1 +), moderate (2 +), and strong (3 +) in 4 discrete regions of interest. Samples were analyzed under 5 ×, 10 ×, and 20 × objectives. 3 + strong staining was readily visualized under the 5 × objective, 2 + moderate staining was visible at 10 ×, and 1 + weak staining required the 20 × objective to visualize staining. Stroma and other cellular components apart from tumor cells were not scored or analyzed. Scores in the 4 regions of interest were then averaged for a composite score.

Study Drug

To develop an SGLT2-targeted imaging agent, we hypothesized that a member of the SGLT2-inhibitor class of medications could be used as an effective ligand. SGLT2-inhibitors are an FDA-approved class of medication for type II diabetes mellitus that function as competitive antagonists highly specific for SGLT2 over other SGLTs. Multiple small molecule SGLT2 inhibitors were assessed for suitability for molecular imaging. Among these, dapagliflozin was selected as the small-molecule ligand as it has high specificity for SGLT2 over other isoforms of SGLTs (IC50 = 1.2 nM for SGLT2; 1400 nM for SGLT1), and was most amenable to conjugation with a fluorophore [17]. Dapagliflozin can be administered intravenously or orally and has a wide therapeutic range up to 500 mg per dose with minimal side effects in both diabetic and non-diabetic patients, offering an ample dosing range for clinical translation [21]. Dapagliflozin was manufactured in compliance with Good Manufacturing Practices (Astra-Zeneca, Hyderabad, India). The dapagliflozin molecule was modified to facilitate conjugation and was covalently bound to a modified indocyanine green (ICG) molecule, as described in Lansdell et al. [22]. The GlucoGlo solution was stored at − 20 °C in dimethyl sulfoxide. Before utilization, the frozen vials were thawed and diluted with methanol, phosphate-buffered saline (PBS) or culture media for the appropriate application.

Spectral Analysis

Absorption spectra of GlucoGlo were measured at 5 μM concentration in methanol with a near-infrared (NIR) JASCO V-700 spectrophotometer (JASCO, Easton, MD, USA) using a dye-free vehicle sample for correction. NIR emission spectra were recorded using the same solution with a QM8075-11-C fluorescence spectrophotometer (Horiba Scientific, Ontario, Canada). Depth of penetration was compared to other common NIR contrast agents including ICG and pafolacianine using a tissue phantom. The 1% Intralipid tissue phantom was prepared by diluting 20% Intralipid (Sigma-Aldrich, Burlington, MA, USA) with deionized water [23]. 10 μM aliquots of each dye were placed in a capillary tube at various depths of Intralipid solution and were imaged with 785 nm excitation and 820 nm emission wavelengths using the Pearl Imaging System (LI-COR Biosciences, Lincoln, NE, USA). Signal to background ratios (SBRs) and full width at half maximum (FWHM) measurements were performed using ImageJ (NIH; https://imagej.nih.gov/ij).

In Vitro SGLT2 Expression and GlucoGlo Binding

Several human non-small cell lung cancer (NSCLC) cell lines were obtained from American Type Culture Collection (Manassas, VA, USA), including H1299 (RRID:CVCL_B7N8), H1666 (RRID:CVCL_1485), and A549 (RRID:CVCL_A549) to evaluate GlucoGlo fluorescence in relation to SGLT2 expression in human cell lines. Cell lines were maintained in vitro using standard growth media (RPMI 1640 media supplemented with 10% fetal bovine serum (FBS), 1% glutamine, and 1% penicillin/streptomycin). Cell lines were regularly tested and maintained negative for Mycoplasma spp. and were cultured at 37 °C in 5% CO2 in a humidified incubator.

As SGLT2 expression has not yet been well characterized in common NSCLC cell lines, SGLT2 expression in H1299, H1666, and A549 was first quantified using flow cytometry with SGLT2 antibody staining to establish SGLT2 expression in positive and negative cell lines. Cells were cultured in 6-well plates with standard growth media for 24 h. Cells were then stained with monoclonal anti-SGLT2 antibody (ab58298, Abcam, Cambridge, MA, USA) for 1 h at 4 °C and washed three times with PBS to eliminate unbound ligand. Cells were stained with an Alexa Fluor 488-conjugated secondary antibody (ab150113, Abcam, Cambridge, MA, USA) for 15 min at room temperature, protected from light. Antibody binding was assessed using an LSR Fortessa X-20 flow cytometer (BD Biosciences, San Diego, CA, USA). Samples were analyzed using FlowJo software (Ashland, OR, USA).

Once SGLT2 expression had been quantified in each cell line, we then evaluated GlucoGlo binding and immunofluorescence in the same cell lines. Cells were cultured in poly-L-lysine coated 8-chamber slides (Thermo Fisher Scientific, Somerset, NJ) with standard growth media for 24 h, then incubated with 10 μM GlucoGlo for 1 h at room temperature and washed three times with PBS. Slides were mounted with ProLong Gold Antifade Reagent containing DAPI (Fisher Scientific, Waltham, MA, USA) and covered with a glass coverslip. GlucoGlo fluorescence was imaged using a Leica DM6 B fluorescence microscope (Leica Microsystems, Wetzlar, Germany). Relative immunofluorescence of GlucoGlo in the various cell lines was then compared to the relative SGLT2 expression in the cell lines as measured by antibody staining.

Dose Response Curves and Incubation Time

Dose–response studies were conducted with GlucoGlo concentrations ranging from 0.1 to 10 µM. Time-course studies were performed with incubation times ranging from 5 min to 2 h, maintaining the same staining protocol. To quantify the binding curve for GlucoGlo, H1299 cells were cultured for 24 h and then treated with varying concentrations of GlucoGlo ranging from 1 nM to 1 μM for 1 h at 37 °C and washed three times with PBS. Flow cytometric analysis was performed to quantify GlucoGlo fluorescence, using red laser excitation (640 nm) and fluorescence detection in the APC-Cy7 channel (720–840 nm). GlucoGlo binding was quantified by measuring the median fluorescence intensity of the treated cells. Analyses were completed in triplicate and analyzed using FlowJo software.

Competitive Inhibition

Competitive inhibition assays were performed to evaluate the specificity of GlucoGlo for the SGLT2 transporter using both unconjugated dapagliflozin and unconjugated glucose, SGLT2’s physiologic substrate. H1299 cells were seeded in the fashion previously described and treated with 10 μM GlucoGlo in the presence of excess unconjugated dapagliflozin (up to 1 mM) or free glucose (up to 300 mM). The same staining protocol was maintained, and GlucoGlo fluorescence was imaged using a Leica DM6 B fluorescence microscope and analyzed using Image J.

Small Animal Tumor Model and Imaging

Flank xenografts were established in female nude athymic mice (Taconic Biosciences, Germantown, NY, USA) by subcutaneously injecting 2 × 106 H1299 cells in 50 μL PBS and 50 μL Matrigel (Corning Life Sciences, Corning, NY, USA). Once the flank xenografts were palpable, GlucoGlo was administered via tail vein injection (0.05 mg/kg; n = 5). As a negative control, mice (n = 5) were pretreated with 5 mg/kg unconjugated dapagliflozin for 2 days prior to administration of GlucoGlo at 0.05 mg/kg. 48 h after GlucoGlo administration, mice were imaged with the Pearl Imaging System. SBRs were calculated by comparing mean fluorescence in tumors areas to that of the same area on the contralateral flank. The Animal Care and Use Committee of the University of Pennsylvania approved all animal study protocols, and experiments were conducted in compliance with the Guide for the Care and Use of Laboratory Animals.

Image Analysis and Statistics

Image analysis was conducted with ImageJ. After correcting for distant background fluorescence, ROI software was used to quantify GlucoGlo fluorescence and background fluorescence to calculate an SBR. Statistical analysis was performed using GraphPad Prism 8 (GraphPad Software, San Diego, CA, USA). Fisher’s exact tests and chi-square tests were used for categorical variables where applicable. Unpaired student t-tests and ANOVA were used to compare means, where applicable. P values < 0.05 were considered statistically significant.