Understanding the pharmacological properties of a metabolic PET tracer in prostate cancer

Abstract

Generally, solid tumors (>400 mm(3)) are inherently acidic, with more aggressive growth producing greater acidity. If the acidity could be targeted as a biomarker, it would provide a means to gauge the pace of tumor growth and degree of invasiveness, as well as providing a basis for predicting responses to pH-dependent chemotherapies. We have developed a (64)Cu pH (low) insertion peptide (pHLIP) for targeting, imaging, and quantifying acidic tumors by PET, and our findings reveal utility in assessing prostate tumors. The new pHLIP version limits indiscriminate healthy tissue binding, and we demonstrate its targeting of extracellular acidification in three different prostate cancer models, each with different vascularization and acid-extruding protein carbonic anhydrase IX (CAIX) expression. We then describe the tumor distribution of this radiotracer ex vivo, in association with blood perfusion and known biomarkers of acidity, such as hypoxia, lactate dehydrogenase A, and CAIX. We find that the probe reveals metabolic variations between and within tumors, and discriminates between necrotic and living tumor areas.

Fig. 1.

⁶⁸Ga-DOTA–labeled pHLIP variants. In vitro binding studies (n = 3) display higher binding of ⁶⁸Ga-DOTA-WT and ⁶⁸Ga-DOTA-Var7 variants as the pH of the incubation medium is decreased. Note that the opposite was observed with the control peptide, K-WT.

Fig. 2.

In vivo pharmacokinetic optimization studies in prostate tumor xenografts. (A) Tissue distribution of ⁶⁸Ga-DOTA–labeled WT and Var7 demonstrates the superiority of Var7 in terms of tumor uptake at 4 h p.i. in PC3-wt tumor implants. (B) In the same tumor model, ⁶⁴Cu-NOTA-Var7(D) displays faster clearance and less nonspecific binding, particularly in hepatobiliary, intestinal, and renal tissues, in contrast to ⁶⁴Cu-DOTA-Var7. Tumor uptake of both probes was comparable at 24 h p.i. The PET images (C and D) of mice bearing dual LNCaP and PC3-wt tumors demonstrate the advantages of ⁶⁴Cu-NOTA-Var7(D) compared with ⁶⁴Cu-DOTA-Var7. The tumor uptake displayed in the tables is the mean ± SD of all acquired images (n > 3; i.e., with and without pHe measurements).

Fig. 3.

pH-dependent interaction of Cu-NOTA-Var7(D) with the lipid membrane bilayer. Cold Cu-NOTA-Var7(D) was studied for the presence of the three basic states of pHLIP: State I is the peptide in solution at pH 8, state II is the peptide in the presence of 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) liposomes at pH 8, and state III is the folding and insertion of the peptide with POPC liposomes from pH 8–3.6. The states were monitored by changes of the steady-state CD (A) and tryptophan fluorescence spectroscopy at λ_ex = 295 nm (B). (C) Changes in the intrinsic fluorescence were monitored as a function of pH wherein a pK_a of ∼5.9 was obtained. λ_max, maximum wavelength.

Fig. 4.

In vivo pHe measurements. (A) Representative ¹H-decoupled ³¹P MR spectrum of a PC3-CAIX tumor after 3-APP injection, where 3-APP, phosphoethanolamine (PE), phosphocholine (PC), P_i, phosphocreatine (PCr), NDP/NTPs. (B) pHe measurements show PC3-wt as the most acidic (6.93 ± 0.03; P = 0.035) compared with the PC3-CAIX (7.07 ± 0.06) and LNCaP (7.07 ± 0.04) xenografts. (C) Intracellular acidity of PC3-wt (6.94 ± 0.07; P = 0.012) is higher than that of the other prostate implants. (D) Positive proton fluxes are observed for PC3-CAIX (0.33 ± 0.13, n = 5; P = 0.018) and LNCaP (0.27 ± 0.10, n = 5; P = 0.020) tumors, but not for PC3-wt tumors (−0.010 ± 0.055, n = 5) (D). a.u., absorbance units.

Fig. 5.

pHLIP-PET shows a direct association with extracellular acidity. (A) ⁶⁴Cu-NOTA-Var7(D) tumor uptake is shown to increase as the pHe of the PC3-wt tumor model decreases. (B) Plot of the pHe against radiotracer tumor uptake (24 h p.i.) from all three tumor models demonstrates a notable threshold where higher probe accretion (>3%ID/g) correlates to an acidic extracellular space with pHe <6.9. Because of tumor heterogeneity even within the same xenograft model, each measured tumor is treated as a separate data point.

Fig. 6.

Histology and autoradiography. ⁶⁴Cu-NOTA-Var7(D) autoradiography (24 h p.i.) and correlative histology from 10-μm adjacent sections obtained from PC3-CAIX (A), LNCaP (B), and PC3-wt (C) tumors. The distributions of ⁶⁴Cu-NOTA-Var7(D) (white), pimonidazole (hypoxia, green), and Hoechst 33342 (perfusion, blue), as well as the expression of LDH-A (red), are shown. White arrows indicate discordance between pimonidazole uptake and LDH-A expression, whereas red arrows mark skin accumulation of the imaging probe.