We describe a platform for dual modality (radionuclide/optical) imaging of prostate cancer (PCa) based on targeting the prostate-specific membrane antigen (PSMA). An example provided demonstrates that after a single intravenous (IV) injection of tracer amounts (0.1 nmol) of imaging agent to a tumor-bearing mouse, both single photon emission computed tomography (SPECT) and near-infrared fluorescence (NIRF) imaging were capable of delineating tumor specifically. That such small injected amounts could identify tumor in vivo suggests that optical agents, as has long been known for radiopharmaceuticals, may obey the tracer principle, enabling more rapid clinical translation.
Prostate cancer will kill 32,000 men in the US this year. Two major issues that contribute to that mortality are: 1) the inability to predict which patients have indolent as opposed to aggressive disease, complicating the choice of appropriate therapy; and, 2) difficulty in assuring a clear surgical margin during prostatectomy. Further complicating management of PCa is the fact that there is no reliable way to image the disease either for the purpose of predicting which tumors may invade and metastasize or for identifying the number of metastatic sites during therapy. Molecular imaging, the non-invasive sensing of cellular and molecular processes in animal models of disease or in patients, can report not only on the anatomy of a tumor, but also on its underlying physiology – possibly predicting its future behavior. Unfortunately, unlike other cancers PCa does not image well with the standard molecular imaging agent, [18Ffluorodeoxyglucose, which can be detected in patients using positron emission tomography (PET). The integral membrane protein prostate-specific membrane antigen (PSMA) is becoming increasingly recognized as a viable target for imaging and therapy of prostate and other forms of cancer.[3–5] PSMA is significantly over-expressed in PCa and metastases, particularly with respect to the hormone-refractory form. PSMA may provide a negative prognostic indicator for PCa – enabling distinction of indolent from aggressive disease.
Dual-modality imaging agents that combine the sensitivity of radionuclide imaging, such as SPECT, with the high sensitivity and resolution of NIRF imaging would be useful in the preoperative and intraoperative realms, respectively.[8–10] For staging and preoperative assessment of tumor burden the radionuclide component could initially be used to map disseminated lesions or lesions within the prostate at sub-centimeter resolution. The optical mode could aid in image-guided surgery to delineate tumor margins precisely during and following resection, using a minimally invasive laparoscopic instrument equipped with a NIRF detection system.
Recently, we and others have demonstrated successful PSMA-targeted radiopharmaceutical imaging in experimental models of PCa using cysteine-glutamate or lysine-glutamate ureas.[11–15] With these agents the radionuclide (11C, 125I, 18F) is attached to the cysteine or lysine moiety via a small prosthetic group. For large molecular fragments, such as radiometal (99mTc, 68Ga) chelators, organic fluorescent molecules, and nanoparticles, we have determined that a linking moiety of at least 20Å between the large molecule and the lysine moiety is needed.[16–19] We hypothesized that conjugation of a dual (radionuclide and optical) modality imaging platform to a PSMA-targeting urea would enable sequential, dual modality imaging of experimental prostate tumors. Such a dual modality PSMA imaging agent could be used to address each of the two items confounding PCa management described above.
Here we report the first dual modality SPECT/NIRF imaging agent [111In]1 that demonstrates high and specific uptake in PSMA+ xenografts and pharmacokinetics suitable for targeting PSMA in vivo. The non-radiolabeled compound 1 is prepared conveniently by a combined solid phase and solution phase Fmoc-based peptide synthesis strategy as shown in Scheme 1. Commercially available Fmoc-Lysine(Boc)-Wang resin was coupled with Fmoc-Lysine(Dde)-OH and DOTA tris(t-Bu ester)-CO2H. The resin support was cleaved with a cocktail of TFA/CH2Cl2/Et3SiH/H2O to generate 2, which also resulted in simultaneous removal of O-t-Bu and N-Boc groups. Compound 3, Glu-urea-Lys-linker-NHS, was reacted with the free ε-amine of the lysine of 2 in the presence of DMSO and TEA to generate the urea-DOTA 4 in 60% yield. Removal of the Dde group using 2% hydrazine-hydrate in DMF afforded a free ε-amine that was subsequently conjugated with the commercially available N-hydroxysuccinimide ester of IRDye800CW to produce 1. Compound 1 was labeled with either non-radioactive 113/115In or radioactive 111In at pH 5.5. The radiochemical yield of [111In]1 was 60–70%. Compounds were purified by reverse phase HPLC and characterized by LCMS (Supporting Information). The specific radioactivity of [111In]1 was 46 – 325 GBq/μmol, and the radiochemical purity was ≥ 99%. The electronic spectra of 1 and non-radioactive [113/115In]1 exhibited an absorbance maximum at 776 nm with an extinction coefficient of 181,000 M−1cm−1. We did not observe any significant changes in the spectral properties between 1 and [113/115In]1. Upon excitation, both 1 and [113/115In]1 provided intense fluorescence with an emission maximum at 792 nm and a fluorescence lifetime of 420 and 483 psec, respectively, in aqueous solution with quantum yields of ~ 8.9% in PBS.
PSMA inhibition constant (Ki) values for 1 and [113/115In]1 were determined using a fluorescence-based PSMA inhibition assay. Compounds 1 and [113/115In]1 had Ki values of 0.46 ± 0.05 nM and 1.24 ± 0.07 nM (Figure S4), respectively, identifying them as potent inhibitors of PSMA with IC50 values similar to related compounds employing a single radionuclide chelating or NIRF moiety.[17, 19] To determine selectivity and specificity in vitro, PSMA+ PC3-PIP and PSMA− PC3-flu cells were seeded into a 96 well plate containing growth medium and incubated with a serially diluted solution of [113/115In]1 at 4ºC and 37°C and then washed with PBS. We prefer the isogenic PIP vs flu comparison as the two cell lines are phenotypically identical, differing only in PSMA expression. Figure S5A shows the lower expression of PSMA within PIP vs. LNCaP, the traditional PSMA+ control tumor cell model. NIRF imaging of the 96 well plate showed a dose dependent, PSMA+ cell specific uptake of [113/115In]1 (Figure S6). Figure 1 shows the pharmacokinetic behavior of [111In]1 in vivo in severe-combined immune-deficient (SCID) mice bearing both PSMA+ PC3-PIP and PSMA− PC3- flu xenografts.. In this experiment 37 MBq of [111In]1 was administered IV and the animal was imaged repeatedly over a two day period. Intense radiotracer uptake was seen only in the PSMA+ PIP tumors and in the kidneys. Kidney uptake of the radiotracer is partially due to its route of excretion and to specific uptake from the expression of PSMA in mouse kidneys. To validate the in vivo imaging data, [111In]1 was also assessed for its pharmacokinetics ex vivo. Table 1 shows the percent injected dose per gram (%ID/g) of radiotracer in selected organs
Compound [111In]1 showed PSMA dependent binding in PSMA+ PC3-PIP xenografts with rapid accumulation at the tumor site that persisted throughout the course of the experiment. We observed tumor uptake values of 13.9 ± 1.9, 13.3 ± 2.4, 16.4 ± 3.7, 14.6 ± 1.3 %ID/g (SEM) at 1, 3, 5 and 24 h respectively. The blood, spleen, gastrointestinal tract, kidney and bladder displayed the highest uptake at 1 h. Steady clearance from the kidneys was demonstrated, from 104.4 ± 9.4 at 1 h to 48.1 ± 17.7 % ID/g at 24 h. The tumor uptake values of [111In]1 compare favorably with other low molecular weight PSMA imaging agents, including N-[N-[(S)-1,3-dicarboxypropyl]carbamoyl]-4-[18F]fluorobenzyl-L-cysteine, [18F]DCFBC, which has recently been administered to human subjects. PSMA specificity and selectivity of [111In]1 was further confirmed by effective blocking in PSMA+ PC3 PIP tumor and kidney (Figure S9) using a dose 50 mg/Kg of the potent, selective PSMA inhibitor N-[[[(1S)-1-Carboxy-3-methylbutyl]amino]carbonyl]-L-glutamic acid (ZJ43). A significant aspect of this work is the fact that we were able to image PSMA expressing PC3 PIP tumors by SPECT and NIRF sequentially, after using a single injection of [111In]1, which was of high specific radioactivity. In so doing we were able to visualize tumor with NIRF imaging after injection of only ~ 0.22 μg (0.1 nmol) of material, which is a tracer dose. For comparison, Table S1 shows clinical studies indicating similar, although somewhat higher, tracer doses of administered radiopharmaceutical. That finding simplifies translation of optical as well as dual modality agents to the clinic because increased safety margins for chemical toxicity are afforded by the low tracer dose. Significantly, since DOTA is a general chelating agent, 1 may also be used with other radiometals such as 68Ga, 64Cu, or 86Y, for PET/optical-based dual modality imaging. Other fluorophore/chelating agent/radiometal combinations can also be envisioned using this approach.
In conclusion, we have developed an efficient chemical and radiochemical synthesis of a dual modality, PSMA-binding compound, which provides sequential imaging by SPECT and NIRF at radiotracer levels. Concurrent, rather than sequential, imaging will be possible with this agent once a corresponding dual modality imaging device becomes available.