1.

Introduction

Since its discovery in 1993, an increasing amount of effort has been invested in studying the pharmacology and physiopathological role of type 2 cannabinoid receptors ($\mathrm{C}{\mathrm{B}}_2\mathrm{R}$). Presently, both academia and the pharmaceutical industries consider this receptor as a major therapeutic target. Specifically, $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$ represents a promising target to treat inflammation,^1,2 pain,³ osteoporosis,⁴ autoimmune diseases,⁵ addiction,^6,7 psychiatric disorders,^8–10 diabetes,^11–13 cardiovascular disorders,¹⁴ cancers, and neurodegenerative diseases. For example, recent studies have shown that $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$ plays a fundamental role in tumorigenesis and that the $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$ agonists may represent promising venues to develop cancer treatments.^15,16 Specifically, some of the most exciting recent research carried out by several laboratories demonstrated that $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$ agonists potently inhibited viability, proliferation, adhesion, and migration of various cancer cells, such as breast,^17–19 prostate,^20,21 glioma,²² colon,²³ lung,²⁴ thyroid,²⁵ lymphoma,²⁶ skin,²⁷ pancreas,²⁸ and liver²⁹ cancers. These data were collected from both cellular systems and preclinical animal models. Additional evidence indicates that the $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$ agonists exhibit promising therapeutic value for treating neurodegenerative diseases. $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$ agonists attenuate Alzheimer’s disease pathogenesis by blocking $\beta$-amyloid peptide-induced activation of microglial cells³⁰ and play a neuroprotective role in Huntington’s disease and amyotrophic lateral sclerosis.³¹ The richness of $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$’s regulatory roles has rendered this receptor as an attractive target to study a variety of diseases and biological processes. However, the precise role of $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$ in the regulation of diseases remains unclear. The ability to specifically image $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$ would contribute to develop reliable $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$-based therapeutic approaches with a better understanding of the mechanism of $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$ action in these diseases.

Little has been done to target $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$ for imaging studies and therapeutic evaluations. The current imaging techniques to identify $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$ rely heavily on immunostaining; however, all currently available $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$ antibodies have significant nonspecific binding issues, leading to unreliable imaging results.^6,32 As such, the development of reliable contrast agents for $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$ imaging is critically needed in the field. Few laboratories have developed $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$ imaging agents for positron emission tomography (PET) imaging, and $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$ imaging using other modalities such as optical imaging has been virtually unexplored.³³ Although PET is a great imaging technique for clinical imaging and translational research due to high sensitivity and no limitation in tissue penetration, the spatial resolution is comparatively low.³⁴ Optical imaging is widely used for biomedical imaging due to its high sensitivity and resolution, as well as low instrument cost.³⁵ In clinical settings, optical imaging is a promising technology for intraoperative guidance^36–38 and optical biopsies.^39,40 Near-infrared (NIR) fluorescent dyes are typically used as the fluorophores for in vivo imaging applications because of the relatively deep tissue penetration and negligible autofluorescence in the NIR region (650 to 900 nm).⁴¹

In our previous study, we reported the first $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$-targeted NIR fluorescent probe, NIRmbc94, which was synthesized by coupling a conjugable pyrazole-based $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$ ligand, mbc94, with an NIR fluorescent dye, IRDye800CW.⁴² We selected the pyrazole structure to develop mbc94 because SR144528, a pyrazole molecule, is a well-known selective $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$ inverse agonist with subnanomolar binding affinity and well-characterized biology.⁴³ NIRmbc94 was successfully used to image $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$ in a transfected mouse malignant astrocytoma delayed brain tumor (DBT) cell line, ${\mathrm{CB}}_2$-mid DBT, that expresses $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$ at endogenous levels.⁴⁴ Later on, NIRmbc94 was also used to select specific $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$ ligands by high-throughput screening in the receptors’ native environment.⁴⁵ Recently, we reported the first in vivo optical imaging study using another NIR fluorescent probe based on mbc94, NIR760-mbc94, which has an NIR dye (NIR760) with easy synthesis, and high fluorescence quantum yield, stability, and molar extinction coefficient.⁴⁶

In this study, we developed a novel $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$-targeted NIR fluorescent probe based on a quinolone structure. Quinolone-based molecules were recently reported as highly selective $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$ ligands with binding affinities as high as 0.2 nM.⁴⁷ To develop the quinolone-based $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$ probe (NIR760-Q), we synthesized a novel conjugable $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$ ligand based on the quinolone structure, followed by bioconjugation with the NIR760 dye. To demonstrate the translational potential of this new $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$ probe, NIR760-Q was used to image $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$ in Jurkat human acute T-lymphoblastic leukemia cells that naturally overexpress $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$. To our best knowledge, this is the first $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$-targeted cellular imaging of human cancer cells that naturally overexpress the target receptor.

2.

Methods and Materials

3.

Results and Discussion

The design of NIR760-Q has the following considerations: (1) The quinolone structure was chosen as the targeting moiety because certain quinolone molecules, such as 4-quinolone-3-carboxamide, were recently reported to have high $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$ selectivity and binding affinity ($K_d$ as high as 0.2 nM);⁴⁷ (2) A terminal amino group was introduced to the targeting moiety to allow for universal conjugation with various signaling moieties, such as fluorescent dyes (for optical imaging) and metal chelators (for PET, single photon emission-computed tomography, and magnetic resonance imaging); (3) NIR760 was selected as the fluorophore for NIR fluorescence imaging, which has high fluorescence quantum yield, stability, and molar extinction coefficient and can be synthesized with only two steps.⁴⁶ In addition, the four sulfonate groups on NIR760 introduce high hydrophilicity to the imaging probe; (4) A di(ethylene glycol) linker between the targeting moiety and NIR760 dye provides ether oxygens as hydrogen bond acceptors that can potentially facilitate the binding.⁵⁰ To synthesize NIR760-Q (Fig. 1), we first prepared quinolone compound 3 through 7-steps reactions using previously reported methods.^48,49 The following click reaction between compound 3 and 11-azido-3,6,9-trioxaundecan-1-amine provided the $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$ ligand 4 with a terminal amine group, which was conjugated with fluorescent dye NIR760 to yield the desired probe, NIR760-Q.

Fig. 1

Synthesis of NIR760-Q.

Upon the synthesis of NIR760-Q, absorption and emission spectra were collected and compared with NIR760 as shown in Fig. 2. NIR760 has intense NIR absorption and emission in water with peaks at 760 and 781 nm, respectively. After conjugation with $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$ ligand 4, the dye showed redshifted absorption and emission spectra. The maximum absorption of NIR760-Q is located at 768 nm with a high molar extinction coefficient of $2.5\times {10}^5{\mathrm{M}}^{-1}\mathrm{c}{\mathrm{m}}^{-1}$, which is comparable with that of NIR760 ($\epsilon =2.4\times {10}^5{\mathrm{M}}^{-1}\mathrm{c}{\mathrm{m}}^{-1}$).⁴⁶ NIR760-Q also exhibits the intense fluorescence centered at 787 nm with a high quantum yield (16.5% in water), which is roughly 60 times higher than that of indocyanine green (ICG) ($\mathrm{\Phi }=0.28\%$ in water⁵¹). ICG is the only Food and Drug Administration-approved NIR fluorescent dye and typically serves as the gold standard for NIR fluorescence imaging.

Fig. 2

The absorption and emission spectra of NIR760 (solid) and NIR760-Q (dash) in ${\mathrm{H}}_2\mathrm{O}$ at a concentration of 1.0 μM. (${\lambda }_{\mathrm{ex}}=720\mathrm{nm}$).

We recently reported a $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$-targeted NIR fluorescent probe, NIR760-mbc94, which was successfully used to image $\mathrm{C}{\mathrm{B}}_2\mathrm{R}+$ cells and tumors.⁴⁶ NIR760-Q and NIR760-mbc94 share the same fluorescent dye, NIR760, while the targeting molecules are based on quinolone and pyrazole structures, respectively. Compared to NIR760-mbc94, NIR760-Q has similar maximum absorption (768 versus 766 nm) and emission (787 versus 785 nm) wavelengths, but higher molar extinction coefficient ($\epsilon =2.5\times {10}^5{\mathrm{M}}^{-1}\mathrm{c}{\mathrm{m}}^{-1}$ versus $1.44\times {10}^5{\mathrm{M}}^{-1}\mathrm{c}{\mathrm{m}}^{-1}$ in water) and quantum yield ($\mathrm{\Phi }=16.5\%$ versus 15.2% in water). The enhanced absorption and emission near 800 nm, and excellent water solubility render NIR760-Q as an outstanding fluorescent probe for biomedical imaging.

We selected Jurkat human acute T-lymphoblastic leukemia cells to test the binding affinity and in vitro imaging potential of NIR760-Q. In our previous $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$-targeted imaging studies,^42,52 we used a transfected mouse malignant astrocytoma DBT cell line, ${\mathrm{CB}}_2$-mid DBT, that expresses $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$ at endogenous levels.⁴⁴ Along with ${\mathrm{CB}}_2$-mid DBT cells, we used WT DBT cells as the $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$-control cells. Although CB2-DBT and WT-DBT cells are an excellent pair for studying $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$ targeting specificity, these cells are mouse cell lines and $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$ is not naturally expressed. To demonstrate the translational potential of the developed $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$-targeted probe, we used Jurkat cells that have been reported to naturally express $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$.²⁶ Before Jurkat cells were used for imaging, we verified their $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$ expression using RT-PCR. RT-PCR was performed using primers for the human ${\mathrm{CB}}_2$ receptor. As shown in Fig. 3, a $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$ positive band at 1415 base pairs was detected, indicating positive mRNA expression of $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$ in Jurkat cells.

Fig. 3

$\mathrm{C}{\mathrm{B}}_2\mathrm{R}$ mRNA expression in Jurkat cells was assessed by RT-PCR. The band at 1415 base pairs (bp) corresponds to $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$.

To determine NIR760-Q’s binding affinity to $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$ in living Jurkat cells, we used an in vitro saturation binding assay to measure the equilibrium dissociation constant and the maximum specific binding. A large excess of 4Q3C was added to a parallel set of cells to saturate receptor binding sites and account for nonspecific binding. Figure 4 shows a representative saturation binding curve. NIR760-Q binds to $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$ with a $K_d$ of $75.51\pm 27.97$ nM and $B_{\max }$ of $440.9\pm 71.75\mathrm{pmol}/\mathrm{mg}$.

Fig. 4

In vitro $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$ saturation binding assay of NIR760-Q using intact cells. (a) Total binding data (solid) were represented as relative fluorescent units (RFU) on the $y$-axis as a function of NIR760-Q concentration ($x$-axis) in the absence of 4Q3C (blocking agent). Nonspecific binding data (dash) were represented as RFU as a function of NIR760-Q concentration in the presence of 4Q3C. (b) Specific binding data were represented as RFU ($y$-axis) as a function of NIR760-Q concentration ($x$-axis). Data in $y$-axis were obtained by subtracting the nonspecific binding data from the total binding data. The dissociation constant ($K_d$) and receptor density ($B_{\max }$) were estimated from the nonlinear fitting of the specific binding versus the concentration of NIR760-Q using Prism software. Each data point represents the $\text{mean}\pm \mathrm{SEM}$ based on triplicate samples.

The $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$ targeting specificity and imaging potential of NIR760-Q was evaluated using Jurkat cells that were incubated with 5 μM of NIR760-Q or NIR760. For blocking studies, Jurkat cells were treated with 5 μM of NIR760-Q and 10 μM of 4Q3C. We observed strong fluorescence signal from cells incubated with NIR760-Q, which primarily localized in the cytoplasm [Fig. 5(a)]. In contrast, no significant fluorescence signal was seen from cells incubated with the same concentration of free dye (NIR760) control. Moreover, challenged cells treated with both NIR760-Q and 4Q3C showed lower fluorescence signal than unchallenged cells. These results indicate specific binding of NIR760-Q to the target receptor.

Fig. 5

NIR760-Q specifically binds to $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$ in Jurkat cells. Jurkat cells were incubated for 30 min with 5 μM of NIR760-Q or free NIR760, with or without 10 μM of blocking agent 4Q3C. Cells were then washed three times with serum free medium before imaging. (a) Fluorescence imaging using a Zeiss Axio Observer fluorescent microscopy with ApoTome 2 imaging system. From left to right: ICG filter (red), ICG filter (red) + DAPI filter (blue) merged and differential interference contrast (DIC). Scale bar: 20 μm. (b) Quantitative fluorescent signal was measured with a Synergy H4 Hybrid Multimode Microplate Reader. Wild type DBT-cells were used as a $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$ negative control group for comparison. Each data point represents the $\text{mean}\pm \mathrm{SEM}$ based on triplicate samples. (***$p<0.001$, ****$p<0.0001$).

To evaluate the targeting specificity of NIR760-Q in a quantitative manner, we used a multiwell plate reader system to perform Jurkat cell binding assays. Similar to the fluorescence imaging study described above, Jurkat cells were divided into three groups: (1) cells incubated with 5 μM of NIR760-Q; (2) cells incubated with 5 μM of free dye; (3) cells incubated with 5 μM of NIR760-Q and 10 μM of 4Q3C. We also used DBT-cells as the $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$-cell line for comparison. As shown in Fig. 5(b), Jurkat cells treated with NIR760-Q showed a 2.8-fold higher fluorescence signal than those treated with NIR760 ($10469.00\pm 495.26$ versus $3755.67\pm 311.81$, $p<0.0001$). Treatment with 4Q3C reduced the uptake of NIR760-Q in Jurkat cells by 40% (from $10469.00\pm 495.26$ to $6274.33\pm 350.55$, $p=0.0001$). In addition, $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$-WT DBT cells treated with NIR760-Q showed 40% less fluorescence signal than Jurkat cells ($6297.67\pm 250.68$ versus $10469.00\pm 495.26$, $p=0.0001$). These results indicate specific binding of NIR760-Q to the target receptor and are consistent with our previous cellular imaging studies using NIR760-mbc94, whose uptake was also blocked by 40% when challenged with a $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$ ligand.⁴⁶ The partial inhibition indicates nonspecific binding of NIR760-Q, which may be due to its net negative surface charge. In a recent study, Choi et al.⁵³ reported that replacing the negatively charged NIR fluorescent dyes with a zwitterionic dye that has no net surface charge significantly reduced the nonspecific binding of NIR fluorescent imaging probes. With a net surface charge of $-4$, NIR760-Q is likely to have nonspecific binding. We noticed that the cell uptake of NIR760-Q in DBT-WT cells is higher than that of NIR760 in Jurkat cells. Although both uptakes are due to nonspecific binding, the higher lipophilicity of NIR760-Q than NIR760 could have caused the enhanced level of nonspecific binding. Future study will involve developing zwitterionic $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$ probes with reduced nonspecific binding to allow for enhanced imaging contrast at the target site. In addition, we plan to develop a xenograft mouse model using Jurkat cells and evaluate the potential of NIR760-Q and new zwitterionic $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$ probes in vivo.

4.

Summary

In summary, we have developed a novel $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$-targeted NIR probe NIR760-Q, which demonstrated specific binding in Jurkat human cells. Compared to our previously reported $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$ probe, NIR760-mbc94, NIR760-Q has enhanced absorption and emission and comparable binding affinity and specificity. We also report the first $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$-targeted optical imaging of human tumor cells that naturally express $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$. The combined data indicate that NIR760-Q is a promising imaging probe for $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$-targeted imaging with the potential of translational studies. Such an imaging tool may have great value in elucidating the regulatory role of $\mathrm{C}{\mathrm{B}}_2\mathrm{R}$ in various diseases.