Preclinical Development and First-in-Human Imaging of the Integrin α_vβ₆ with [¹⁸F]α_vβ₆-Binding Peptide in Metastatic Carcinoma

A central tenet for the delivery of personalized therapy in the treatment of patients with cancer is the accurate evaluation of the extent of disease, which has primarily relied on a variety of radiologic studies (1, 2). [¹⁸F]FDG-PET/CT allows whole body imaging and has been incorporated into clinical practice for both initial evaluation of disease sites as well as a surrogate for response to therapy. Yet significant limitations exist for [¹⁸F]FDG-PET/CT including the lack of a target that is specific and biologically relevant to cancer cells (2–5). There is a significant and rapidly growing body of literature that has identified the integrin α_vβ₆ as one such target. This integrin is an epithelial-specific cell surface receptor that is usually undetectable in healthy adult epithelium but is significantly upregulated in a wide range of epithelial-derived cancers including breast, colon, non–small cell lung, oral squamous cell carcinoma, ovarian, and pancreas cancer (6–15). Furthermore, higher levels of expression have been correlated with invasiveness and clinically this has been correlated with metastatic progression and poor patient survival (10, 14–20). The preferential expression of this integrin in neoplastic tissues as well as the scope of malignancies overexpressing this integrin suggests that agents specifically targeting α_vβ₆ have significant potential in the evaluation, treatment, and management of multiple types of solid organ malignancies. Given the impact of molecular imaging in oncology, our group and others have chosen to develop α_vβ₆ specific PET molecular imaging agents (21–23).

The prototypical structure of α_vβ₆ integrin–binding agents has been based on the RG/TDLXXL motif (24), and these agents have undergone extensive preclinical evaluation. Several molecular imaging agents, including those identified using phage display (8) and cystine knot peptides (25), have been investigated. Our efforts have focused primarily on the 20-amino acid peptide A20FMDV2 that was first identified from the GH-loop of the VP1 protein of foot and mouth disease virus, which uses the α_vβ₆ integrin to mediate host cell infection (23). Building upon the information obtained using our first-generation peptide, [¹⁸F]FBA-A20FMDV2, and through subsequent structure–activity studies, derivatization through the addition of PEG modifiers, and in vitro and in vivo mouse studies, we developed [¹⁸F]α_vβ₆-BP.

Herein, we present the development of our lead candidate [¹⁸F]α_vβ₆-BP, including the in vitro testing, in vivo evaluation in mouse xenograft models of human cancer, and ultimately the first-in-human PET imaging in patients with a diagnosis of breast, colon, lung, or pancreas cancer. The data demonstrate the potential for molecular imaging of a broad spectrum of malignancies using [¹⁸F]α_vβ₆-BP.

The general reagents and consumables for the synthesis of the peptides used in the preclinical and clinical studies, and the high-performance liquid chromatography (HPLC) and analytic methods used, are described in the Supplementary Data.

Peptide synthesis and radiolabeling were performed on solid support and the radiotracer formulated in serum-free DMEM (for in vitro studies) or PBS (for in vivo studies; ref. 26). The cold standard [¹⁹F]α_vβ₆-BP was prepared by coupling 4-fluorobenzoic acid, pre-activated with 1-[bis(dimethylamino)methylene]-1H-1,2,3-triazolo[4,5-b]pyridinium 3-oxid hexafluorophosphate (HATU) and N,N-diisopropylethylamine (DIPEA), to the H₂N-peptidyl-resin; cleavage, purification, analysis, and storage followed procedures described previously (27, 28). Radiosynthesis of [¹⁸F]α_vβ₆-BP on solid support was done by coupling 4-[¹⁸F]fluorobenzoic acid (FBA; ref. 29), preactivated with HATU and DIPEA, to 5 mg of preswollen H₂N-peptidyl-resin in a 1-mL fritted syringe reactor; following cleavage with trifluoroacetic acid (TFA)/water/triisopropylsilane (TIPS) 95/2.5/2.5 (v/v/v; 2 × 0.5 mL; 15 minutes/each), [¹⁸F]α_vβ₆-BP was purified by HPLC and formulated (27).

Competitive binding ELISAs for integrins α_vβ₆ and α_vβ₃ were performed with [¹⁹F]α_vβ₆-BP as previously described and analyzed using Prism software (GraphPad Software; ref. 23).

The cell lines DX3puroβ6 (α_vβ₆ positive) and DX3puro (α_vβ₆ negative) were cultured at 37°C under 5% CO₂ in DMEM containing 10% heat-inactivated FBS and 1% PSG (26, 27). Expression levels of α_vβ₆ were confirmed by flow cytometry prior to the experiment. Binding of [¹⁸F]α_vβ₆-BP to and internalization into DX3puroβ6 and DX3puro cells were determined as described previously (26, 27). Cell-binding samples were counted on a Wizard 1470 γ-counter (PerkinElmer).

To evaluate serum stability, mouse serum (Sigma Aldrich) was combined with the [¹⁸F]α_vβ₆-BP, incubated at 37°C for various amounts of time, and, following precipitation of serum proteins with ethanol, analyzed by HPLC as described previously (26).

All animal procedures conformed to the Animal Welfare Act and were approved by the university's Institutional Animal Care and Use Committee. Female athymic nude mice (Charles River Laboratories) were inoculated subcutaneously with 3 × 10⁶ DX3puro and 3 × 10⁶ DX3puroβ6 cells on opposite flanks. Studies commenced once tumors reached a maximum diameter of approximately 0.5 cm (range, 43–225 mg).

The [¹⁸F]α_vβ₆-BP was injected into the tail vein of mice (imaging: 9.0–9.5 MBq/animal; biodistribution: 1.5–2.0 MBq/animal; autoradiography: 37 MBq/animal). For imaging, 2 mice/scan were placed side-by-side in a feet-first, prone position [n = 4 total; anesthesia: 1.5%–2.0% isoflurane); PET/CT scans (dynamic 4 × 15-minute PET emission scan starting 15 minutes postinjection (p.i.), single-frame 15-minute PET emission scans at 2 and 4 hours p.i.] were acquired using Inveon scanners (Inveon DPET scanner and Inveon SPECT/CT scanner, Siemens Medical Solutions) and analyzed as previously described using Inveon Research Workplace software (Siemens; ref. 26).

For biodistribution studies, mice were euthanized and dissected 1, 2, and 4 hours (n = 3/time point) following injection of [¹⁸F]α_vβ₆-BP; blocking studies were conducted by preadministration with [¹⁹F]FBA-PEG₂₈-A20FMDV2 (30 mg/kg, 10 mg/mL in saline, intravenously 10 minutes before the [¹⁸F]α_vβ₆-BP; ref. 30). Tissues were collected, rinsed, and radioactivity measured in a γ-counter; calibrated, decay-corrected radioactivity concentrations are expressed as percent of injected dose per gram of sample (% ID/g). For autoradiography of [¹⁸F]α_vβ₆-BP tumor uptake and correlation to integrin α_vβ₆-expression by IHC, one animal was injected with [¹⁸F]α_vβ₆-BP and tumor tissue collected (1 h p.i.), embedded in freezing medium (Tissue-Tek, Sakura), and sectioned (CM 1850 cryostat, Leica Biosystems). Autoradiography sections (20 μm) were exposed to a storage phosphor-screen overnight and the screen read on a STORM 860 phosphor imager at 50 μm resolution (GE Healthcare). For integrin α_vβ₆ IHC, adjacent sections (5 μm) were cut and fixed for staining (26).

IHC sections were stained using anti-integrin β₆ antibody (Calbiochem Clone 442.5C4, 1:200 dilution) on a DAKO link autostainer (Agilent). The same method was used for assessment of β₆ expression of human samples.

Quantitative data are reported as mean ± SD. Statistical analysis was done using a paired two-tailed Student t tests to evaluate statistical significance, where P < 0.05 was considered statistically significant.

The GLP synthesis of the precursor peptide H₂N-α_vβ₆-BP, and single-dose acute toxicity study of the [¹⁹F]α_vβ₆-BP are described in the Supplementary Methods.

[¹⁸F]α_vβ₆-BP was manufactured in compliance with current good manufacturing practice (cGMP) under the guidelines of USP Chapter <823> for administration to human subjects as a single intravenous dose under the auspices of a FDA-approved exploratory investigational new drug (eIND #124336).

The N-succinimidyl 4-[¹⁸F]fluorobenzoate ([¹⁸F]SFB) was prepared on the GE TracerLab FX-FN one-pot automated synthesizer (GE Healthcare) from azeotropically dried [¹⁸F]fluoride in Kryptofix (10 mg/mL)/potassium carbonate (2 mg/mL) solution (acetonitrile/water 94/6 v/v) and 4-(ethoxycarbonyl)-N,N,N-trimethylbenzenaminium triflate (5 mg; 90°C, 10 minutes, acetonitrile), followed by saponification with tetraproplyammonium hydroxide (20 μL of 25% aq solution; 120°C, 1 minute, acetonitrile) and subsequent activation using O-(N-succinimidyl-N,N,N′,N′-tetramethyluronium tetrafluoroborate (TSTU; 12 mg; 90°C, 5 minutes, acetonitrile) to yield [¹⁸F]SFB. The crude [¹⁸F]SFB was purified by HPLC (Jupiter Proteo 10 μm 90 Å column, 250 × 10 mm; isocratic 10 mmol/L phosphoric acid/acetonitrile 67/33 v/v, 5 mL/minute). The HPLC fraction containing the [¹⁸F]SFB was trapped on a Sep-Pak C18 Plus cartridge, washed with water (5 mL), eluted with acetonitrile (1.5 mL), and dried by passing over a Sep-Pak Dry Sodium Sulfate Plus cartridge.

Following evaporation of the solvent (50°C, 10 minutes), the [¹⁸F]SFB was reconstituted with DMSO (245 μL) containing the precursor peptide H₂N-α_vβ₆-BP (0.5 mg) and DIPEA (5 μL). The coupling reaction was allowed to proceed for 30 minutes (15 minutes at ambient temperature, followed by 15 minutes at 50°C), quenched with phosphoric acid (10 mmol/L, 0.75 mL), and purified by semipreparative HPLC. The fraction containing pure [¹⁸F]α_vβ₆-BP was collected (≤3 mL) through a sterile 0.22 μm filter (SLGVM33RS; MilliporeSigma), formulated by dilution with saline (to a final volume of 20 mL) and dispensed through a sterile 0.22 μm filter (SLGVM33RS; MilliporeSigma) to yield [¹⁸F]α_vβ₆-BP in a ≤10% ethanolic v/v solution in saline.

Prior to injection, quality testing including visual clarity, filter integrity, pH, radionuclidic identity, purity and radiochemical identity via analytical HPLC, residual solvent analysis via gas chromatography, residual Kryptofix spot test, and pyrogenicity via limulus amebocyte lysate test (Charles River Laboratories) were performed. The full list of release criteria is given in Supplementary Table S1 in the Supplementary Data; they also included a postrelease 14-day sterility test using the direct inoculation of media method described in USP <71>.

This study protocol was approved by the UC Davis Institutional Review Board (FWA00004557), and written informed consent was obtained from all individual participants included in the study (ClinicalTrials.gov NCT03164486). All subjects enrolled in this study signed a consent form that provided their consent for the testing of the [¹⁸F]α_vβ₆-BP and the analysis of biopsy material for research purposes. The inclusion and exclusion criteria for the study are provided in the Supplementary Data. The studies were conducted following U.S. Common Rule.

Imaging was performed on a General Electric 690 time-of-flight PET scanner fitted with a 16-slice CT. Up to 370 MBq (326 ± 44 MBq, in up to 10 mL) of [¹⁸F]α_vβ₆-BP were injected (intravenously) over 1 minute. Immediately after the injection, vital signs (heart rate, pulse oximetry value, body temperature, EKG) were measured. Patients underwent PET/CT scanning from the apex of the skull to the proximal thighs with arms raised, using two minutes per bed position starting at 30 minutes, 1 hour, 2 hours, and 3 hours postinjection.

Data were reconstructed using the Ordered Subsets – Expectation Maximization (OSEM) algorithm, with 24 subsets and 2 iterations, and reviewed in the axial, coronal, and sagittal planes. A 6.4-mm filter was used transaxially and the manufacturer's “standard” filter was used axially. Image review was performed using GE Advantage Workstation software version 4.5. SUV_max values were obtained for lesions using manual sizing of rectangular ROI and standard 42% threshold.

Vital signs were checked at 15, 30, 45, 60, 75, 90, 105, 120, 135, 150, 165, and 180 minutes. In addition, vital signs and blood samples were obtained and checked 1 and 7 days after injection of [¹⁸F]α_vβ₆-BP to ensure no changes or abnormalities occurred. Preinjection and follow-up blood draws were assessed for complete metabolic panel including liver function tests.

Blood samples (1 mL) were taken throughout the imaging period (1, 3, 5, 10, 15, 30, 60, 90, 120, and 180 minutes), collected in heparinized tubes, and immediately stored on ice and processed.

The 30-minute blood sample was centrifuged (1,130 × g, 10 minutes; 4°C), proteins precipitated from the supernatant with an equal volume of ethanol, and removed by centrifugation (11,160 × g, 3 minutes; 4°C). The protein removal process was repeated with a half-volume of ethanol. The final supernatant was diluted with an equal volume water/TFA (0.05% v/v)/glacial acetic acid (2% v/v) for HPLC analysis.

At each collection time point a whole-blood aliquot (200 μL) was transferred to a γ-counter tube; cells were precipitated in the remaining sample by centrifugation (1,130 × g, 10 minutes; 4°C) and a plasma aliquot (200 μL) was transferred to a γ-counter tube. Samples were counted on a γ-counter and radioactivity concentrations calculated as decay-corrected percent of injected dose per gram of sample (dc % ID/g).

Nonradioactive [¹⁹F]α_vβ₆-BP was obtained in >98% purity after HPLC purification; [¹⁹F]α_vβ₆-BP: MS (MALDI) m/z = 4,919.9564 [M+H]⁺, calculated M (FBA-PEG-NAVPNLRGDLQVLAQRVARTPEG; C₂₁₈H₄₀₁FN₃₆O₈₆) = 4,918.8096, 4919.8174 [M+H]⁺. For preclinical studies, the radiotracer [¹⁸F]α_vβ₆-BP was prepared by solid-phase radiolabeling with [¹⁸F]FBA and obtained in 98.1 ± 1.0% radiochemical purity with a molar activity of 65 ± 25 GBq/μmol and a decay-corrected radiochemical yield of 6.6% ± 1.0% from [¹⁸F]FBA (n = 10). For clinical production, [¹⁸F]α_vβ₆-BP was prepared in solution with [¹⁸F]SFB in radiochemical purity ≥99% with molar activity ≥37 GBq/μmol, mass dose ≤50 μg/patient dose, and 15.1% ± 1.7% decay-corrected radiochemical yield from [¹⁸F]SFB (n = 18).

In vitro affinity and selectivity were evaluated by competitive ELISA against the biotinylated natural ligands and by cell-binding studies using the integrin α_vβ₆-expressing cell line DX3puroβ6 and the paired non-α_vβ₆-expressing DX3puro control; affinity (IC₅₀) of [¹⁹F]α_vβ₆-BP for integrin α_vβ₆ was 1.2 nmol/L, whereas no binding was observed to integrin α_vβ₃ at concentrations up to 10 μmol/L (Fig. 1A). Cell-binding studies with [¹⁸F]α_vβ₆-BP resulted in 72.5 ± 0.9% of total radioactivity bound to DX3puroβ6 cells at 1 hour, with 52.5% ± 1.8% of total radioactivity internalized (corresponding to 72.4% ± 2.5% of bound radioactivity; Fig. 1B); binding to and internalization in DX3puro cells was 3.5% ± 1.4% and 1.0% ± 0.1% of total radioactivity, respectively. Therefore, the ratios of α_vβ₆-expressing cell line compared with the nonexpressing cell line (DX3puroβ6/DX3puro) were 21/1 and 54/1, respectively, for total bound and internalized radioactivity (P = 2 × 10⁻¹⁰, P = 1 × 10⁻⁸). In mouse serum, [¹⁸F]α_vβ₆-BP remained 98.3% intact at 30 minutes at 37°C; with minimal metabolism at 60 and 120 minutes (93.3% and 88.9% intact, respectively).

Evaluation of [¹⁸F]α_vβ₆-BP in the paired DX3puroβ6/DX3puro tumor mouse model showed DX3puroβ6-tumor uptake of 2.51 ± 0.48% ID/g at 1 hour and good retention (2.64 ± 1.23% ID/g at 4 hours), while uptake and retention in the α_vβ₆-negative DX3puro control tumor were low (1 hour: 0.35 ± 0.14% ID/g; 4 hour: 0.21 ± 0.15% ID/g; Supplementary Table S2). This resulted in a steady increase in DX3puroβ6-to-DX3puro tumor ratio from 8.3 ± 4.2/1 at 1 hour to 18.3 ± 11.7/1 at 4 hours (Supplementary Table S2). When α_vβ₆-blocking peptide was preadministered 10 minutes before the radiotracer, DX3puroβ6 tumor uptake was reduced by 90% to 0.28 ± 0.07% ID/g (1 hour; P = 0.001), a level comparable with uptake in the DX3puro control tumor (0.26 ± 0.03% ID/g).

IHC staining for integrin α_vβ₆ confirmed expression in DX3puroβ6 tumors, and the distribution of integrin α_vβ₆ corresponded to the distribution of [¹⁸F]α_vβ₆-BP as determined by autoradiography. In comparison, both staining for integrin α_vβ₆ and uptake of [¹⁸F]α_vβ₆-BP were at background levels for the DX3puro control tumors (Fig. 1D).

PET/CT imaging correlated with the biodistribution data, clearly visualizing the DX3puroβ6 tumor while showing no uptake in the DX3puro tumor (Fig. 1C). Imaging and biodistribution demonstrated rapid renal clearance as the major route of elimination, with minor hepatobiliary contribution (Fig. 1; Supplementary Table S2). Highest uptake was seen for kidneys at 1 hour (22.9 ± 8.9% ID/g), rapidly clearing to 7.9 ± 2.1% ID/g at 4 hours; all other organs showed <8% ID/g at 1 hour and <6% ID/g at 4 hours, with stomach, gallbladder, lung, and bladder being 5.6 ± 1.6, 4.5 ± 3.0, 2.3 ± 0.2, and 1.8 ± 1.8% ID/g, respectively, at 4 hours (Supplementary Table S2).

Patients with a prior diagnosis of breast, colon, lung, or pancreas cancer were recruited to the study (18 patients to date). The mean injected dose of [¹⁸F]α_vβ₆-BP was 326 ± 44 MBq (range: 220–364 MBq) with a mass dose ≤50 μg/patient dose. No noticeable changes in vital signs, on electrocardiograms, or in blood laboratory results were observed during the study, and no adverse side effects were noted at the 7-day follow-up after injection of the radiotracer.

[¹⁸F]α_vβ₆-BP was rapidly cleared from the blood stream (Supplementary Fig. S1); radio-HPLC of blood samples taken 30 minutes after injection showed no metabolite peaks and indicated that [¹⁸F]α_vβ₆-BP remained intact (Supplementary Fig. S2). PET images demonstrated significant uptake of [¹⁸F]α_vβ₆-BP in the primary and metastatic lesions, including subcentimeter lesions in adrenal glands, bone, brain, liver, and lung (Figs. 2–6; Supplementary Table S4). For example, for subject #1 the primary tumor to background ratio was 17.3:1 and right iliac wing lesion to background ratio was 67.5:1. For subject #2, the primary tumor to background ratio was 19.5:1 and the left iliac bone lesion to background ratio was 43.7:1. Concordant with the murine data, renal excretion of [¹⁸F]α_vβ₆-BP was observed with minimal uptake in normal bone, brain liver, or lung (i.e., major sites of metastasis). Evaluation of specific tissues demonstrated low to mild activity in the pituitary, lacrimal, and salivary glands where it decreased during the imaging timeframe, and in the thyroid where it appeared stable (Supplementary Table S3). High uptake was seen in the kidneys and was stable over the imaging timeframe whereas the bladder demonstrated low to moderate and variable activity depending on fluid and voiding status. The average radioactivity uptake in the entire stomach was moderate and steady throughout the imaging timeframe; however, uptake observed in the fundus overall minimally decreased over time. The average radioactivity uptake in the entire small bowel was moderate and increased over time. The ascending and descending colon demonstrated overall mild to moderate activity that increased over time to moderate activity. Initial impression of bowel activity and distribution pattern suggests uptake and secretion from the stomach and anterograde movement of activity over time in the bowel lumen. A summary of the SUV_max data at the 1 and 3-hour time points for the 5 patients presented here is listed in the Supplementary Table S3.

Subject 1 was a 53-year-old female diagnosed with stage IV adenocarcinoma of the lung. Uptake of [¹⁸F]α_vβ₆-BP was observed in both the primary lung lesion and a right iliac wing bone metastasis (Fig. 2A and B). At the 1-hour scan, a SUV_max of 5.2 and 13.5 was observed for the primary lung lesion and right iliac wing metastasis, respectively (Fig. 2; Supplementary Table S4). Analysis of a sample of a malignant pleural effusion demonstrated strong expression of the integrin α_vβ₆ (Fig. 2C a-b).

Subject 2 was a 59-year-old female diagnosed with stage IV invasive mammary carcinoma. Uptake of [¹⁸F]α_vβ₆-BP was observed in the primary breast lesion as well as in metastases in multiple lymph nodes in the right axilla and mediastinum, and multiple osseous sites. At the 1-hour scan, a SUV_max of 3.9 was observed in the primary lesion and SUV_max ranged from 4.4 to 13.1 in the lymph nodes and metastases (Fig. 3; Supplementary Table S4). Tissue from the primary lesion and the left iliac osseous metastasis were available and demonstrated high levels of α_vβ₆ expression (Fig. 3C a-d).

Subject 3 was a 56-year-old female diagnosed with initial stage IIIB adenocarcinoma of the lung, subsequently metastatic. Uptake of [¹⁸F]α_vβ₆-BP was observed in the primary lung lesion in addition to multiple sites of metastases including brain, right adrenal gland, perifacial lymph node, and osseous sites. At the 1-hour scan, a SUV_max of 10.6 was observed in the consolidated mass in the lung and the SUV_max ranged from 1.0 to 7.3 in the lymph node, right adrenal, and osseous metastasis (Supplementary Tables S3 and S4. The [¹⁸F]α_vβ₆-BP PET images revealed numerous brain metastases, in the right and left parietal, left frontal and occipital lobes, with SUV_max of 2.7, 2.4, 2.4, and 1.0, respectively. Representative axial images are shown in Fig. 4, middle row); all regions demonstrating increased SUV_max, were clearly visualized against a brain background SUV_max of 0.1, and correlated with the known brain metastases visualized on MRI (Fig. 4, upper row).

Subject 4 was a 51-year-old female diagnosed with initial stage IV adenocarcinoma of the colon with metastases to liver, lungs, and abdominal lymph nodes at the time of diagnosis. Uptake of [¹⁸F]α_vβ₆-BP was observed in two sites in the left lobe of the liver (Fig. 5), bilateral lobes of the lungs, and lymph nodes of the thorax/low neck. At the 1-hour scan, a SUV_max of 2.3 was observed in an upper left hepatic lobe lesion (Fig. 5; Supplementary Table S4), 3.4 in a lower lateral left hepatic lobe lesion, a range of 0.6 to 0.9 for the right and left pulmonary lesions and a range of 1.4–1.7 for the thoracic and low neck lymph nodes (Supplementary Table S4).

Subject 5 was a 78-year-old male diagnosed with initial stage IV adenocarcinoma of the pancreas with metastases to the lung at time of diagnosis. Uptake of [¹⁸F]α_vβ₆-BP was observed in multiple small right lower lobe pulmonary densities (example depicted in Fig. 6). At the 1-hour scan, a SUV_max range of 0.6 to 1.4 (Supplementary Table S4) was seen in the pulmonary densities ranging in size from 0.6 to 0.8 cm.

There is increasing evidence of the important role of integrin α_vβ₆ in cancer, its suitability as a prognostic biomarker (linked to poor disease-free and overall survival), and the promise of an α_vβ₆-targeted clinical diagnostic tool for in vivo detection, patient stratification, as well as monitoring of response to treatment (31–33). The high contrast in signal gained due to aberrant, but specific expression of α_vβ₆ suggests that α_vβ₆-targeted agents have significant potential in the treatment and management of a multitude of malignancies. This has led to the development of several preclinical PET and single photon emission computed tomography (SPECT) radiotracers and, most recently, clinical reports in patients with head and neck squamous cell carcinoma and NSCLC (34).

The integrin α_vβ₆ is a receptor for fibronectin, tenascin, vitronectin, the latency associated peptide (LAP) of the TGFβ, and the VP1 coat protein of foot and mouth disease virus (FMDV). Our first generation α_vβ₆-targeting peptide was A20FMDV2, a 20 amino acid peptide (NAVPNLRGDLQVLAQKVART), from the sequence of the GH-loop of the VP1 protein from FMDV (35, 36). We initially demonstrated the promise of A20FMDV2 as an in vivo imaging agent for α_vβ₆ with the 4-[¹⁸F]fluorobenzoyl labeled [¹⁸F]FBA-A20FMDV2 (23). Although α_vβ₆-specific targeting in vivo was observed in murine xenograft models, rapid tumor wash out (0.66% ID/g at 1 hour, down to 0.06% ID/g at 4 hours) and poor in vivo stability precluded its further application. This motivated us to investigate a series of improvements (mainly through the size and locations of PEG modifiers and subsequent confirmation by structure–activity studies), culminating in [¹⁸F]α_vβ₆-BP as the lead compound for translation to a human clinical trial.

In vitro affinity and selectivity of [¹⁸F]α_vβ₆-BP for integrin α_vβ₆ were excellent and compared favorably with those reported by Altmann and colleagues for their SFTI-1 derivatives; however, it is noted that the experimental methods were somewhat different, and different cells lines with varying expression levels of the integrin α_vβ₆ were used (37). In vivo studies using the paired DX3puroβ6/DX3puro cell xenograft mouse model confirmed that [¹⁸F]α_vβ₆-BP met expectations for α_vβ₆ targeting and overall pharmacokinetics, with good tumor retention and kidney clearance. The tumor uptake values were similar to those noted by Altmann and colleagues for their SFITGv6 peptide which demonstrated good binding in mice bearing HNO97 xenografts (37).

Given these favorable in vitro and in vivo results of [¹⁸F]α_vβ₆-BP, we proceeded with our first-in-human microdose study to evaluate the safety and in vivo behavior of [¹⁸F]α_vβ₆-BP (exploratory Investigational New Drug [eIND] # 124336, NCT03164486). The evaluation of the overall pharmacokinetics revealed that [¹⁸F]α_vβ₆-BP was predominantly excreted renally, with moderate transient activity in the gastrointestinal tract and mild activity in the pituitary, lacrimal, salivary glands, and the thyroid. Minimal uptake was found in the four common sites of metastatic disease, namely bone, brain, liver, and lung.

Our observations align closely with those described by Keat and colleagues who recently reported a microdose study in four healthy subjects using our first generation compound [¹⁸F]FBA-A20FMDV2 (38). Similar uptake in the thyroid, salivary glands, kidney, and stomach was noted; however, they also observed additional significant uptake in the gallbladder, liver, and spleen. The lack of uptake of [¹⁸F]α_vβ₆-BP in these organs is due to the peptide modifications introduced; this improvement on biodistribution does provide lower background uptake and thus increases sensitivity for the detection of disease, as the liver is a frequent site of metastases. Recently, Altmann and colleagues reported the results of PET/CT scans performed in a compassionate use setting in 2 patients with [⁶⁸Ga]-DOTA-SFITGv6, a SFTI-1 derived peptide targeting α_vβ₆, which was identified using phage display (37). They describe a general biodistribution of their gallium-68 tracer similar to that of [¹⁸F]α_vβ₆-BP, revealing accumulation in the kidneys, with gastrointestinal secretion followed by an intraluminal transport to the terminal ileum and cecum. They concluded that these challenges could be overcome by the use of diuretics and laxatives, respectively.

Although [¹⁸F]FDG-PET/CT plays a crucial role in the staging of patients with a wide range of malignancies, its well-known limitations include uptake in inflamed or infected tissues and poor signal-to-noise ratio in the brain and liver (1, 2, 4, 5). This limits the detection of primary and metastatic disease in these organs and highlights the need for cancer-specific imaging agents such as [¹⁸F]α_vβ₆-BP. The integrin α_vβ₆ is often localized to the invasive front and infiltrating edges of tumors, plays a key role in invasion and metastases, and has been closely correlated with distant metastasis in breast, lung, and colon cancer as well as lymph node metastasis in pancreas cancer (31). Notably, the low nonspecific uptake of [¹⁸F]α_vβ₆-BP in normal bone, brain, liver, and lung suggest [¹⁸F]α_vβ₆-BP as a promising imaging agent to detect metastases. Uptake of [¹⁸F]α_vβ₆-BP was observed in several subcentimeter metastases in the lung of a patient with metastatic pancreatic adenocarcinoma (subject 5). This observation begins to address the issue of size threshold for detection of disease. In this respect, our data for [¹⁸F]α_vβ₆-BP indicate a very favorable performance characteristic for identification of small lesions that may be indeterminate on traditional thoracic CT imaging. The fact that pancreas cancer is one of the most lethal cancers in the United States, that the majority of patients present with metastatic disease, and that detection of metastases markedly alters treatment amplifies the need for accurate noninvasive imaging (39). Our preliminary data strongly demonstrate that [¹⁸F]α_vβ₆-BP holds tremendous promise to meet this need for accurate determination of the extent of disease.

This was further observed in the patient with stage IV invasive mammary carcinoma (subject 2) where uptake of [¹⁸F]α_vβ₆-BP was observed in the primary breast lesion as well as metastases in multiple lymph nodes. The uptake in the primary lesion and the left iliac osseous metastasis was subsequently confirmed by IHC to correlate with α_vβ₆ expression. Numerous reports have demonstrated that tumors originating from the breast express high levels of α_vβ₆ and suggests that there may be potential for broad utility in staging of breast cancer. Here, [¹⁸F]α_vβ₆-BP also demonstrated great promise to guide the staging and treatment of these patients as the identification of occult metastatic disease significantly alters the intent of therapy.

In subject 3, [¹⁸F]α_vβ₆-BP clearly identified known brain metastases as small as 0.5 cm with low background uptake in the brain parenchyma. These images demonstrated that [¹⁸F]α_vβ₆-BP enters the brain, targets metastases, and suggest that this radiotracer could provide combined staging of the brain and body within a single imaging study. To our knowledge, this case is the first to demonstrate PET imaging of brain metastases using an α_vβ₆ targeting peptide and suggests an exciting opportunity for high-resolution systemic staging and patient selection for novel therapeutics using [¹⁸F]α_vβ₆-BP PET/CT.

In conclusion, [¹⁸F]α_vβ₆-BP demonstrated high affinity and selectivity for integrin α_vβ₆ in vitro and in vivo, along with significantly improved pharmacokinetics over previous generations. [¹⁸F]α_vβ₆-BP can be prepared reliably for clinical imaging and is well tolerated by patients. Significant uptake of [¹⁸F]α_vβ₆-BP in both the primary lesion and in metastases in a variety of cancers was observed, and uptake correlated with α_vβ₆ expression as shown by IHC for all available specimen samples. Although preliminary, these results are very promising, are in concurrence with the preclinical observations, and demonstrate the ability to successfully image lesions well under 1 cm. Furthermore, a variety of sites of metastases including lung, liver, bone, and brain were imaged as a result of the favorable pharmacokinetics coupled with the low background affinity in these common sites of metastasis. The clinical impact of this first-in-human study is immediate for a broad spectrum of malignancies. This approach facilitates precision medicine, with α_vβ₆-directed imaging allowing the early detection and monitoring of response to therapy, thus enabling prediction of outcome and optimized treatment regimes.

S. Hausner is a co-inventor of intellectual property related to [¹⁸F]α_vβ₆-BP. M.E. Daly reports receiving commercial research grants from EMD Serono, and is a consultant/advisory board member for Boston Scientific. J.L. Sutcliffe is an employee of and holds ownership interest (including patents) in Luminance Biosciences, Inc., and is a co-inventor of intellectual property related to [¹⁸F]α_vβ₆-BP. No potential conflicts of interest were disclosed by the other authors.

Conception and design: S.H. Hausner, R.J. Bold, C.C. Foster, J.L. Sutcliffe

Development of methodology: S.H. Hausner, R.J. Bold, R.A. Davis, C.C. Foster, J.L. Sutcliffe

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): S.H. Hausner, R.J. Bold, L.Y. Cheuy, H.K. Chew, M.E. Daly, C.C. Foster, E.J. Kim, J.L. Sutcliffe

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): S.H. Hausner, R.J. Bold, L.Y. Cheuy, C.C. Foster, J.L. Sutcliffe

Writing, review, and/or revision of the manuscript: S.H. Hausner, R.J. Bold, H.K. Chew, M.E. Daly, C.C. Foster, E.J. Kim, J.L. Sutcliffe

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): C.C. Foster, J.L. Sutcliffe

Study supervision: C.C. Foster, J.L. Sutcliffe

Other (characterization and quality control analysis of compounds and final radiopharmaceutical product): R.A. Davis

The authors acknowledge the nuclear medicine staff Denise Caudle, Heather Hunt, and Kristen McBride for taking care of the patients and performing the PET/CT imaging, Dr. Abimbola Olusanya for consenting patients enrolled in this study, Regina Gandour-Edwards for performing the IHC staining, Jennifer Fung, Charles Smith, and David Kukis of the Center for Molecular and Genomic Imaging as well as David Boucher for support of the preclinical studies. Last but not least, we thank all of our patient volunteers for selflessly dedicating their time. This work was funded in part through Department of Energy, Office of Science, award #DE-SC0008385 NIH R01CA211554-01 and UC Davis Research in Science and Engineering grant.

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.