In Vivo Characterization of Two ¹⁸F-Labeled PDE10A PET Radioligands in Nonhuman Primate Brains

Abstract

Positron emission tomography with phosphodiesterase 10A (PDE10A) specific radioligands provides a noninvasive and quantitative imaging tool to access the expression of this enzyme in vivo under normal and diseased conditions. We recently reported two potent ¹⁸F-labeled PDE10A radioligands (¹⁸F-TZ19106B and ¹⁸F-TZ8110); initial evaluation in rats and nonhuman primates indicated stable metabolic profiles and excellent target-to-nontarget ratio (striatum/cerebellum) for both tracers. Herein, we focused on in vivo characterization of ¹⁸F-TZ19106B and ¹⁸F-TZ8110 to identify a suitable radioligand for imaging PDE10A in vivo. We directly compared microPET studies of these two radiotracers in adult male Macaca fascicularis nonhuman primates (NHPs). ¹⁸F-TZ19106B had higher striatal uptake and tracer retention in NHP brains than ¹⁸F-TZ8110, quantified by either standardized uptake values (SUVs) or nondisplaceable binding potential (BP_ND) estimated using reference-based modeling analysis. Blocking and displacement studies using the PDE10A inhibitor MP-10 indicated the binding of ¹⁸F-TZ19106B to PDE10A was specific and reversible. We also demonstrated sensitivity of ¹⁸F-TZ19106B binding to varying number of specific binding sites using escalating doses of MP-10 blockade (0.3, 0.5, 1.0, 1.5, and 2.0 mg/kg). Pretreatment with a dopamine D2-like receptor antagonist enhanced the striatal uptake of ¹⁸F-TZ19106B. Our results indicate that ¹⁸F-TZ19106B is a promising radioligand candidate for imaging PDE10A in vivo and it may be used to determine target engagement of PDE10A inhibitors and serve as a tool to evaluate the effect of novel antipsychotic therapies.

Graphical abstract

INTRODUCTION

Phosphodiesterase 10A (PDE10A) is a dual-substrate specific phosphodiesterase that mostly occurs on striatal medium spiny neurons (MSNs)¹ including direct (mostly dopaminergic (DA) D1-like receptor mediated) and indirect pathway neurons (mostly DA D2-like receptor mediated).^2–4 PDE10A is primarily membrane-bound and associated with postsynaptic densities on dendritic spines.^3,5 Under normal conditions, the activation of D1-like receptor in MSNs upregulates cyclic adenosine monophosphate (cAMP) production, whereas D2-like receptor activation reduces cAMP production.^2,6 Accordingly, excessive striatal DA release exerts differential effects on the regulation of PDE10A in the two striatal output pathways. PDE10A would shift toward a somatic, cytosolic localization in D1-like mediated MSNs caused by increased PDE10A, whereas it may shift to an axonal and dendritic localization in D2-like MSNs caused by disinhibition of PDE10A palmitoylation.⁷ Whether dopaminergic signaling modulates PDE10A function remains obscure but DA depletion alters PDE10A expression at both RNA and protein levels.⁸ PDE10A mRNA levels were decreased in ipsilateral striatal neurons 10 weeks after unilateral 6-hydroxydopamine lesion in a rat model of nigrostriatal injury. PDE10A protein levels and activity also decreased in striatal neurons and in striatopallidal and striatonigral projections. Thus, PDE10A levels respond to nigrostriatal injury and may provide a postsynaptic biomarker for striatal function.

The interaction between PDE10A and DA receptors was also investigated by applying PDE10A inhibitors. PDE10A inhibitors suppress D2-like receptors and concomitantly potentiate D1-like receptor-mediated neurotransmission, which reflect desirable antipsychotic effects.⁹ In direct pathway neurons, PDE10A inhibition by papaverine activates cAMP/protein kinase A (PKA) signaling, leading to potentiation of D1-like receptor signaling. In indirect pathway neurons, PDE10A inhibition by papaverine also activates cAMP/PKA signaling by simultaneously potentiating adenosine A2A receptor signaling and inhibiting D2-like receptor signaling. The balance of cAMP/PKA signaling between the direct and indirect pathways determines the output from basal ganglia.¹⁰ PDE10A inhibitors activate cAMP/PKA signaling in indirect and direct pathway neurons, but predominately affect indirect pathway neurons. Experimental support of this comes from an electrophysiology study showing PDE10A inhibition has a greater facilitatory effect on corticostriatal synaptic activity in indirect pathway neurons.¹¹ Biochemical action of PDE10A inhibitors resembles antipsychotic drugs that act primarily as D2-like receptor antagonists and increase dopamine- and cAMP-regulated neuronal phosphoprotein (DARPP-32) phosphorylation in indirect pathway neurons.¹² Thus, PDE10A radioligands have the potential to provide a metric of target engagement for novel antipsychotic therapies.

The first PDE10A PET radiotracer, carbon-11 labeled papaverine was radiolabeled and evaluated in 2010.¹³ Due to the low retention of ¹¹C-papaverine in the brain, tremendous efforts have been paid to develop a more suitable radioligand for imaging PDE10A in vivo. Currently, there are several lead PDE10A radioligands (Figure 1); all of them showed potent binding affinity and good selectivity toward PDE10A. Several of these radiotracers already have been transferred into clinical settings, including 18F-JNJ42259152,^14,15 11C-IMA107,^{16–19 11}C-Lu AE92686,^{20,21 18}F-MNI-659,^22–25 and 11C-T-773.²⁶ MP-10 is the first generation of PDE10A compounds for treating schizophrenia, and was successfully radiolabeled with C-11 for in vivo evaluation. However, nonpolar radio-metabolites hampered its further application.^27,28 TZ1964B was developed and optimized based on MP-10²⁹ by changing its in vivo pathway of metabolism. We radiolabeled TZ1964B on the opposite side to avoid nonpolar radiometabolites that could cross the blood brain barrier,³⁰ and in vitro and in vivo characterization of 11C-TZ1964B in rodents and NHPs³¹ demonstrated that ¹¹C-TZ1964B is a suitable radioligand for imaging PDE10A.

Figure 1.

Chemical structures of leading PDE10A radioligands.

Encouraged by the promising result of ¹¹C-TZ1964B studies, our group recently explored and synthesized a series of PDE10A compounds for F-18 labeling, since F-18 labeled radioligands have several advantages over C-11 radiotracers in general, including longer half-life, better image resolution, and potential for delivery to off-site imaging facilities within a 3–4 h drive distance. Two new potent compounds were identified and radiolabeled: ¹⁸F-TZ19106B (named as ¹⁸F-20a), 3-(2-[¹⁸F]-fluoroethoxy)-2-((4-(1-methyl-4-(pyridin-4-yl)-1H-pyrazol-3-yl)phenoxy)methyl) quinolone, and 18F-TZ8110 (named as ¹⁸F-18d), 4-(2-[¹⁸F]Fluoroethoxy)-2-((4-(1-methyl-4-(pyridin-4-yl)-1H-pyrazol-3-yl)phenoxy)methyl) quinolone.³² Initial evaluation in rats and nonhuman primates indicated that both radiotracers have good stability in vivo and excellent target-to-nontarget ratio (striatum/cerebellum),³² although ¹⁸F-TZ19106B is even more stable than ¹⁸F-TZ8110. Therefore, the goal of the current study was to further investigate the in vivo binding properties of 18F-TZ19106B and 18F-TZ8110 through kinetic modeling, and identify a suitable radioligand for imaging PDE10A in vivo.

RESULTS AND DISCUSSION

MicroPET Baseline Scans in Nonhuman Primates

We previously reported high tracer accumulation in striatum and rapid clearance from nontarget brain regions of ¹⁸F-TZ19106B and ¹⁸F-TZ8110 in nonhuman primates.³² To quantitatively compare the imaging properties of the two PDE10A tracers, 3 h dynamic PET studies were performed in male adult macacca fascicularis. Tissue time activity curves showed that both tracers had high uptake in NHP striatum, while ¹⁸F-TZ19106B had much higher striatal retention than ¹⁸F-TZ8110. Tracer uptake of ¹⁸F-TZ19106B in NHP striatum reached the max SUV value (∼1.76) at 90–100 min post injection and decrease gradually (Figure 2a). In contrast, the peak (SUV value ∼ 0.58) appeared at 30–40 min post injection after ¹⁸F-TZ8110 injections, and declined relatively rapidly (Figure 2b). In addition, low tracer uptake in cerebellum was observed for both tracers, suggesting the feasibility of using cerebellum as the reference region for tracer kinetic modeling analysis.

Figure 2.

Time–activity curves (TACs) of microPET in NHP brains using two PDE10A radioligands ¹⁸F-TZ19106B, and ¹⁸F-TZ8110. (a) ¹⁸F-TZ19106B TACs in NHP striatum and cerebellum using averaged SUVs (n = 2 scans), striatum uptake reached the max SUV value (∼1.76) at 90–100 min postinjection and decreased gradually; (b) ¹⁸F-TZ8110 TACs in NHP striatum and cerebellum (n = 1 scan), the peak (SUV value ∼ 0.58) appeared at 30–40 min postinjection and declined relatively rapidly. Note: To minimize noise signal in the presentation, data in TAC graphics have been smoothed by “LOWESS” using GraphPad Prism 6.0 (GraphPad Software, Inc., San Diego, CA), while tracer kinetic analysis was based on original data without smoothing.

Further quantitative microPET analyses were carried out for trace kinetic comparison of the two PDE10A radioligands, using 0–120 min data for all scans. Reference based modeling methods have been validated for quantitative PET analysis of PDE10A radioligands in NHPs and humans.^20,22,28,31 Non-displaceable binding potential (BP_ND) is the typical measurement from reference tissue methods, which refers to the ratio at pseudoequilibrium of specifically bound radioligand to that of nondisplaceable radioligand in tissue.³³ R1 is a measure of radioligand delivery to tissue relative to reference region.³⁴ Parameters k₂, k₃, and k₄ are rate constants. k₂ (min⁻¹) represents the efflux of radiotracer through the blood-brain barrier (BBB) by diffusion. The rate constants k₃ (min⁻¹) and k₄ (min⁻¹) correspond to the transfer of radioligand between the compartment for nondisplaceable binding (C_ND) and the compartment for specific binding (C_B).³⁵ The radio of k₃/k₄ equals to BP_ND value. In the current study, both Logan graphic analysis (LoganREF) and reference tissue model (RTM) yielded similar striatal BP_ND with cerebellum as the reference (¹⁸F-TZ19106B: BP_ND-LoganREF = 4.36 ± 0.91 vs BP_ND-RTM = 4.70 ± 0.64; ¹⁸F-TZ8110: BP_ND-LoganREF = 2.02 ± 0.70 vs BP_ND-RTM = 2.15 ± 0.66, Table 1). BP_ND values from Logan REF strongly correlated with those using RTM (Figure 3). Moreover, RTM revealed similar R1 and k₂ values for both tracers (¹⁸F-TZ19106B: R1 = 0.85 ± 0.16, k₂ = 0.059 ± 0.008 min⁻¹; ¹⁸F-TZ8110: R1 = 0.89 ± 0.05, k₂ = 0.058 ± 0.004 min⁻¹, Table 1), while k₃ and k₄ values significantly differ between the two radiotracers. ¹⁸F-TZ19106B has higher k₃ value (0.052 ± 0.008 min⁻¹ vs 0.041 ± 0.009 min⁻¹) and lower k₄ value (0.011 ± 0.002 min⁻¹ vs 0.020 ± 0.003 min⁻¹) than ¹⁸F-TZ8110 (Table 1). These data suggest that the influx and efflux of both radioligands through BBB are similar; ¹⁸F-TZ19106B has faster transfer rate from C_ND to C_B (k₃) and slower transfer rate from C_B to C_ND (k4), compared to ¹⁸F-TZ8110. Accordingly, the difference in k₃ and k₄ values mainly contribute to higher BP_ND value of ¹⁸F-TZ19106B than ¹⁸F-TZ8110.

Table 1.

Kinetic Parameters of PET Brain Imaging in NHPs at Baseline Using ¹⁸F-TZ19106B and ¹⁸F-TZ8110

tracers	LoganREF	RTM
	BPND	R1	k2 (min−1)	k3 (min−1)	k4 (min−1)	BPND(k3/k4)
18F-TZ19106Ba	4.36 ± 0.91	0.85 ± 0.16	0.059 ± 0.008	0.052 ± 0.008	0.011 ± 0.002	4.70 ± 0.64
18F-TZ8110b	2.02 ± 0.70	0.89 ± 0.05	0.058 ± 0.004	0.041 ± 0.009	0.020 ± 0.003	2.15 ± 0.66

^an = 5 in 3 NHPs.

^bn = 6 in 3 NHPs.

Figure 3.

Correlation between BP_ND-LoganREF and BP_ND-RTM for ¹⁸F-TZ19106B and ¹⁸F-TZ8110 in NHP striatum. n = 5 scans in 3 NHPs for ¹⁸F-TZ19106B (a), n = 6 scans in 3 NHPs for ¹⁸F-TZ8110 (b). Data from 0 to 120 min post-tracer injection were used for BP_ND estimates. BP_ND-LoganREF values of both tracers strongly correlated with BP_ND-RTM values.

In addition, to establish the time dependence of BP_ND estimates, striatal BP_ND values were estimated by LoganREF and RTM using PET data acquiring from 0 to120 min recordings and 0 to 100- and 80 min truncation, and with exclusion of the first 40 min from the Logan linearization (Table 2). The result showed both ¹⁸F-TZ19106B and ¹⁸F-TZ8110 have good time stability of striatal BP_ND values, at least up to 80 min of the scan duration. For ¹⁸F-TZ19106B, BP_ND-LoganREF were 4.36 ± 0.91, 4.40 ± 0.89, 4.40 ± 0.92 using 120, 100, and 80 min data, and BP_ND-RTM were 4.70 ± 0.64, 4.76 ± 0.77, 4.83 ± 0.76, respectively. The striatal BP_ND estimates of 18F-TZ8110 using 120, 100, and 80 min data were 2.02 ± 0.70, 2.06 ± 0.67, 1.95 ± 0.68 (LoganREF), and 2.15 ± 0.66, 2.17 ± 0.72, 2.12 ± 0.81 (RTM).

Table 2.

Striatal BP_ND Estimates of ¹⁸F-TZ19106B and ¹⁸F-TZ8110 Using Different Scan Duration

radioligand	scan duration for analysisc	120 min	100 min	80 min
18F-TZ19106Ba	LoganREF	4.36 ± 0.91	4.40 ± 0.89	4.40 ± 0.92
	RTM	4.70 ± 0.64	4.76 ± 0.77	4.83 ± 0.76
18F-TZ8110b	LoganREF	2.02 ± 0.70	2.06 ± 0.67	1.95 ± 0.68
	RTM	2.15 ± 0.66	2.17 ± 0.72	2.12 ± 0.81

^an = 5 in 3 NHPs.

^bn = 6 in 3 NHPs.

^cLoganREF analysis uses data starting from 40 min post injection for both tracers, and RTM uses data starting from 0 min postinjection.

Blocking and Displacement Study of ¹⁸F-TZ19106B

Based on the baseline comparison data, ¹⁸F-TZ19106B showed higher striatal retention and favorable tracer kinetics than ¹⁸F-TZ8110. Therefore, we continued further pharmacological characterization of ¹⁸F-TZ19106B in nonhuman primates. As shown in Figure 4, pretreatment with a PDE10A selective inhibitor MP-10 significantly decreased striatal uptake of ¹⁸F-TZ19106B, suggesting the binding of ¹⁸F-TZ19106B is specific to PDE10A. Furthermore, we also demonstrated sensitivity of ¹⁸F-TZ19106B binding to varying number of available specific binding sites with escalating doses of MP-10 blockade (0.3, 0.5, 1.0, 1.5, and 2.0 mg/kg). The resulting BP_ND values were 2.93, 1.76, 1.06, 0.46, and 0.48 by Logan REF and 2.96, 1.62, 0.94, 0.52, and 0.45 by RTM (Table 3). Target (PDE10A in the striatum) occupancy levels then were calculated as the relative change in striatal BP_ND using either LoganREF or RTM; and it refers to the percentage of the enzyme PDE10A binding sites bound by the unlabeled drug molecule (MP-10). Accordingly, MP-10 pretreatment doses of 0.3, 0.5, 1.0, 1.5, and 2.0 mg/kg produced occupancy levels of 33%, 60%, 76%, 89%, 89% determined by Logan REF, and 35%, 64%, 79%, 89%, 90% by RTM (Table 3).

Figure 4.

Representative ¹⁸F-TZ19106B microPET images of NHP brain. High striatum uptake at baseline was blocked by pretreatments with 1.0 or 2.0 mg/kg MP-10, a specific PDE10A inhibitor. In contrast, pretreatment with (−)-eticlopride even increased the striatal uptake of ¹⁸F-TZ19106B. All PET images are summed from 0 to 120 min.

Table 3.

Binding Potential and Occupancy Levels at Baseline and in Pretreatment Studies of ¹⁸F-TZ19106B PET

	BPND-LoganREF	Occm, %a	BPND-RTM	Occm, %
baseline (avg) MP-10, mg/kg	4.36		4.54
2.0	0.48	89.00	0.45	89.99
1.5	0.46	89.40	0.52	88.61
1.0	1.06	75.76	0.94	79.36
0.5	1.76	59.69	1.62	64.27
0.3	2.93	32.81	2.96	34.86
(−)-eticlopride, 0.025 mg/kg	6.28	−44.08	6.09	−34.08

^aOcc^M = (BP_ND-baseline − BP_ND-pretreatment)/BP_ND-baseline × 100%.

Reversibility of radioligand binding is important to consider in the application of tracer kinetic models.³⁶ Thus, we administered unlabeled MP-10 2.0 mg/kg 40 min post injection of ¹⁸F-TZ19106B to determine reversibility of the binding of ¹⁸F-TZ19106B to PDE10A. As shown in Figure 5, striatal uptake declined tremendously after injection of MP-10, while the radiotracer uptake in cerebellum was not obviously impacted, indicating a specific and reversible binding of ¹⁸F-TZ19106B to PDE10A.

Figure 5.

Displacement study of ¹⁸F-TZ19106B in NHP brain. The striatal uptake of 18F-TZ19106B significantly decreased after i.v. injection of MP-10 at 40 min post tracer injection (purple solid circles), while the tracer uptake in the cerebellum was not impacted by the MP-10 injection (purple open circles).

Eticlopride Pretreatment Study Using ¹⁸F-TZ19106B

Emerging evidence suggests that PDE10A play a key role in regulation of striatal signaling which involve dopaminergic pathways and cAMP-dependent pathways. Reduced striatal PDE10A and dopaminergic D2 receptor levels were detected by 18F-MNI-659/¹¹C-raclopride PET or 3H-AMG-7980/³H-raclopride autoradiography in the same set of Huntington Disease (HD) model mice (zQ175 and R6/2).^37,38 A recent clinical study further revealed that D2 receptor and PDE10A availability, indicated by striatal uptake of ¹⁸F-MNI-659 and ¹¹C-raclopride, were 62 ± 12%, and 21 ± 33% of control values (p < 0.05) in stage 1 HD gene expansion carriers (HDGECs), and 72 ± 12% and 53 ± 22% of control values in in premanifest HDGECs.³⁹ In a longitudinal microPET study with ¹⁸F-JNJ42259152, repeated stimulation of the dopamine neuro-transmission by D-amphetamine significantly increased in vivo PDE10A binding in normal rat striatum, while chronic treatment with the selective D1 antagonist SCH23390 decreased PDE10A striatal binding.⁴⁰ In our study, acute pretreatment with (−)-eticlopride (0.025 mg/kg), a selective D2 antagonist, increased the striatal uptake of ¹⁸F-TZ19106B by 44% (BP_ND-LoganREF) or 34% (BP_ND-RTM). This funding was consistent with Dlaboga’s report which used quantitative immunoblot analysis to quantify PDE10A expression after treatment with haloperidol, a selective D2/3 receptor antagonist. A significant increase in PDE10A expression was detected after a 21-day treatment of rats with haloperidol.⁴¹ Based on our data and others’, the potential interaction between PDE10A pathway and dopaminergic signaling was summarized as shown in Figure 6. Intercellular cAMP levels in MSNs is modulated by adenylate cyclase and PDE10A. Adenylate cyclase, the enzyme that catalyzes the conversion of ATP to cAMP, is activated by D1 signaling and suppressed by D2 signaling. Concordantly, D2 inhibition induced by (−)-eticlopride removes suppression of adenylate cyclase, thereby increasing cAMP levels, which lead to compensatory upregulation of PDE10A to hydrolyze overexpressed cAMP. In agreement with our hypothesis, increased in vivo cAMP level causes upregulation of PDE10A BP_ND in rat striatum, measured by ¹⁸F-JNJ42259152 microPET.⁴² Since modulation of cAMP levels could have therapeutic benefits for a broad range of neuropsychiatric disorders,⁴³ PDE10A levels by PET measures may be used as a (indirect) functional readout of novel antipsychotic therapies enhancing cAMP levels.

Figure 6.

Simplified circuitry cartoon showing the interaction between PDE10A pathway and dopaminergic signaling in medium spiny neurons (MSNs) of striatum. Intercellular cAMP levels in MSNs is modulated by adenylate cyclase and PDE10A. Adenylate cyclase, which catalyzes the conversion of ATP to cAMP, is activated by D1 signaling and suppressed by D2 signaling. Concordantly, D2 inhibition induced by (−)-eticlopride removes the suppression of adenylate cyclase, resulting in increased cAMP levels, which lead to compensatory upregulation of PDE10A to hydrolyze overexpressed cAMP.

In summary, quantitative comparison of ¹⁸F-TZ19106B and ¹⁸F-TZ8110 in in vivo imaging PDE10A were conducted in nonhuman primates via PET. The data showed that ¹⁸F-TZ19106B had higher striatal uptake and tracer retention in NHP brains than 18F-TZ8110. Reference-based modeling analysis revealed that both tracers showed good time stability of striatal BP_ND, while ¹⁸F-TZ19106B has higher BP_ND, larger k₃ value and smaller k₄ value compared to ¹⁸F-TZ8110. Thus ¹⁸F-TZ19106B was further characterized pharmacologically. Blocking and displacement studies with MP-10 suggest specific and reversible binding of ¹⁸F-TZ19106B to PDE10A. Pretreatment with (−)-eticlopride upregulated the striatal uptake of ¹⁸F-TZ19106B, indicating PET with a suitable PDE10A radioligand is not only useful for determining target engagement of PDE10A inhibitors, but also may serve as a tool to evaluate the effect of novel antipsychotic drugs. ¹⁸F-TZ19106B has high striatal retention over the 3h scan period in NHP brains, which may permit measurements for static scan acquisition with good consistency, but not for short-term dynamic scans post injection of radiotracers. The toxic evaluation of ¹⁸F-TZ19106B counterpart cold standard is in the process. We are planning to seek United States Food and Drug Administration (FDA) approval for human use of this radiotracer, then to perform translational clinical investigation in human subjects.

METHODS

Radiochemical Synthesis

The radiosyntheses of ¹⁸F-TZ19106B and 18F-TZ8110 were accomplished by a two-step procedure as reported recently by our group.³² First, ethylene ditosylate was reacted with ¹⁸F-KF/Kryptofix 2.2.2 in acetonitrile (ACN) and then purified on a reversed phase HPLC system to afford ¹⁸F-fluoroethyl tosylate with a 60–75% radiochemical yield. Second, nucleophilic substitution of corresponding phenol precursors with ¹⁸F-fluoroethyl tosylate in dimethyl sulfoxide (DMSO) followed by purifying on reversed phase HPLC afforded ¹⁸F-TZ19106B and ¹⁸F-TZ8110 in 65–72% and 30–35% radiochemical yield, respectively. The two-step radiosynthesis took 3 h, and both tracers have high specific activity > 74 GBq/μmol (decay corrected to EOS) and high radiochemical purity > 98%.

Animals

All animal experiments were performed following the Guidelines for the Care and Use of Research Animals under a research protocol approved by Washington University Institutional Animal Care and Use Committee. This work was conducted in nonhuman primate microPET facility at the Washington University School of Medicine in St. Louis.

PET imaging studies were carried out on adult male Macaca fascicularis, n = 5, weighing 7–8 kg. We collected five or six baseline scans in three NHPs for each radiotracer, and each scan lasted for 120 or 180 min. A same animal was used for pretreatment and displacement studies. Animals were prepared for microPET studies as previously reported.^30–32,44 1.5–2.5% isoflurane inhalation anesthesia was maintained throughout the microPET imaging sessions. A 20-gauge plastic catheter was inserted into a limb vein to permit hydration and injection of the radiotracer. For the same animal, the interval between two consecutive PET scans was at least 2 weeks.

MicroPET Data Acquisition

The microPET studies were done using a MicroPET Focus 220 scanner (Concorde/CTI/Siemens Microsystems, Knoxville, TN). Prior to each PET acquisition, a 45 min transmission scan for attenuation correction was done. Subsequently, 185–370 MBq of ¹⁸F-TZ19106B and ¹⁸F-TZ8110 was intravenously injected, and a 120 min dynamic (3× 1 min frames, 4× 2 min frames, 3× 3 min frames, and 20× 5 min frames) or a 180 min dynamic (3× 1 min frames, 4× 2 min frames, 3× 3 min frames, and 32× 5 min frames) PET scan was acquired. Pretreatment studies were performed by intravenous injection of either MP-10 (0.3, 0.5, 1.0, 1.5, and 2.0 mg/kg) or (−)-eticlopride (0.025 mg/kg) 5 min prior to the radiotracer injection. For the displacement study, 2.0 mg/kg MP-10 was intravenously injected at 40 min after injection of the radiotracer.

PET Image Processing

The PET/CT images were processed according to our published procedure.^31,44 Briefly, sinogram data were corrected for attenuation, random and scatter, and reconstructed using filtered back projection.⁴⁵ The final reconstructed resolution was 2.00 mm full width at half-maximum (fwhm) for all three dimensions (axial) at center of the field of view. The reconstructed PET images was coregistered with magnetization-prepared rapid gradient echo (MP-RAGE) MR images using Automated Image Registration (AIR),⁴⁶ and superimposed using Analyze 10.0 (AnalyzeDirect, Overland Park, KS). For quantitative analyses, a program VIDI that was coded by our PET imaging analysts and routinely used in our institute for PET data processing^47,48 was used. Regions of interest (ROIs) were manually drawn on MRI images for striatum and cerebellum, and transformed into PET images using the coregistration transformation matrix. The ROIs for each animal were identified on a baseline scan and kept fixed for all subsequent studies. Time-activity curves (TACs) were then obtained from the dynamic PET images. Activity measures were calculated and standardized to the body weight and the injected dose of radioactivity, and yield standardized uptake values (SUVs). To minimize noise signal in the presentation, data in TAC graphics have been smoothed by “LOWESS” using GraphPad Prism 6.0 (GraphPad Software, Inc., San Diego, CA), while tracer kinetic analysis was based on original data without smoothing.

Reference Based Tracer Kinetic Modeling

To quantitatively evaluate the kinetic property of each radioligand, LoganREF and RTM were carried out for kinetics modeling. LoganREF is a wide used reference-based graphical method.⁴⁹ LoganREF is able to provide highly correlated BP estimates compared with compartmental tracer kinetic models with blood input function; this method has simplified computation and is straightforward.⁵⁰ RTM is a compartmental modeling approach that used the reference tissue time activity curve as input.⁵¹ Compared to graphical analysis, the reference tissue model can extract more physiological information from measured tracer kinetics.⁵⁰ Furthermore, the use of higher order reference tissue models (such as RTM vs simplified reference tissue model (SRTM)), has been proposed to reduce the bias of BP or distribution volume ratio (DVR) of target tissue to reference tissue estimates.⁵⁰ Therefore, LoganREF and RTM were chosen for tracer kinetic modeling in the current study as an initial phase for evaluating the two radiotracers. The binding potential was defined as a ratio of specifically bound ligand to its free concentration. BP_ND was calculated from DVR using the term: BP_ND = DVR-1.³³ R1, k_2, k₃, and k₄ values were also obtained from the RTM model.^52,53 Target occupancy levels (Occ^M) were calculated as the relative change in striatal BP_ND using either LoganREF or RTM, as follow: Occ^M = (BP_ND-baseline − BP_ND-pretreatment)/BP_ND-baseline × 100%.

ABBREVIATIONS

AIR

Automated Image Registration

BBB

blood-brain barrier

BP

binding potential

BP_ND

nondisplaceable binding potential

cAMP

cyclic adenosine monophosphate

C_B

compartment for specific binding

C_ND

compartment for nondisplaceable binding

DA

dopaminergic

DARPP-32

dopamine- and cAMP-regulated neuronal phosphoprotein

DVR

distribution volume ratio

fwhm

full width at half-maximum

HD

Huntington disease

HDGECs

HD gene expansion carriers

LoganREF

Logan graphic analysis

MSNs

medium spiny neurons

NHPs

nonhuman primates

NLSF

nonlinear least-squares fitting

Occ^M

target occupancy levels

PDE10A

phosphodiesterase 10A

PKA

protein kinase A

ROIs

regions of interest

RTM

reference tissue model

SUVs

standardized uptake values

TACs

time–activity curves