- Top of page
- MATERIALS AND METHODS
This study aims to investigate the pharmacokinetics of a recently developed radiotracer for imaging of the norepinephrine transporter (NET) in baboon brain, 123I-INER, using single photon emission computed tomography (SPECT). In addition, it also aims to determine NET occupancy by atomoxetine and reboxetine, two selective norepinephrine reuptake inhibitors, using 123I-INER in baboons. Baseline and preblocking studies with a high dose of atomoxetine (0.85 mg/kg) were conducted in three baboons using SPECT with 123I-INER administered as a bolus. Kinetic modeling analysis was investigated for different models, namely invasive and reference tissue models. Bolus plus constant infusion experiments with displacement at equilibrium using six different doses of atomoxetine (0.03–0.85 mg/kg) and four different doses of reboxetine (0.5–3.0 mg/kg) were carried out in several baboons to obtain occupancy measurements as a function of dose for the two NET selective drugs. Results showed that reference tissue models can be used to estimate binding potential values and occupancy measures of 123I-INER in different brain regions. In addition, the apparent volume of distribution was estimated by dividing concentration in tissue by the concentration in blood at 3 hours postinjection. After administration of atomoxetine or reboxetine, a dose-dependent occupancy was observed in brain regions known to contain high densities of NET. In conclusion, pharmacokinetic properties of 123I-INER were successfully described, and obtained results may be used to simplify future data acquisition and image processing. Dose-dependent NET occupancy for two selective norepinephrine reuptake inhibitors was successfully measured in vivo in baboon brain using SPECT and 123I-INER. Synapse, 2013. © 2012 Wiley Periodicals, Inc.
- Top of page
- MATERIALS AND METHODS
The norepinephrine transporter (NET), located in the outer membrane of presynaptic terminals, plays a critical role in the regulation of synaptic levels of norepinephrine (Ressler and Nemeroff,1999). In nonhuman primate brain, similar to man, structures known to be rich in NET include: the brainstem (particularly the locus coeruleus) and the midbrain. Conversely, lower NET densities can be found in the striatum and cerebellum, and a widespread pattern of NET is present throughout the cortex (Aston-Jones,2004; Donnan et al.,1991; Smith et al.,2006). Dysregulation of NET function has been associated with the pathophysiology of attention-deficit/hyperactivity disorder (ADHD), substance abuse, depression, post-traumatic stress and Alzheimer's disease (Brunello et al.,2002; Ding et al.,2003; Grandoso et al.,2004; Kiyono et al.,2004; Kung et al.,2004; Lakshmi et al.,2008; McConathy et al.,2004; Zeng et al.,2009). Consequently, in vivo imaging of NET provides the opportunity to further understand and follow disease progression. In addition, it has the potential to monitor drug efficacy and therefore, improve the drug discovery process.
Drugs that act as inhibitors of NET, such as, selective norepinephrine reuptake inhibitors, have shown to be effective treatments for various aforementioned disorders (Page,2003; Ressler and Nemeroff,1999; Seneca et al.,2006; Simpson and Plosker,2004; Takano et al.,2009; Wong et al.,2000). These findings have reinforced the link between dysregulation of NET and disease state. Atomoxetine and reboxetine are two selective norepinephrine reuptake inhibitors currently used for treatment of ADHD and depression, respectively (Page,2003; Seneca et al.,2006; Simpson and Plosker,2004). The in vitro affinity of atomoxetine and reboxetine for NET in membranes from MDCK cell lines transfected with human NETs is 5 nM and 11 nM, respectively (Simpson and Plosker,2004; Wong et al.,2000).
To date radiolabeled analogs of reboxetine have shown the highest potential for in vivo imaging of NET using either single photon emission computed tomography (SPECT) or positron emission tomography (Ding et al.,2003; Kanegawa et al.,2006; Seneca et al.,2006; Tamagnan et al.,2007; Wilson et al.,2003). We have previously reported the synthesis of a series of reboxetine analogs and found that the most potent analog corresponded to the iodinated 2S,3S-stereoisomer of reboxetine (Ki = 2.47 nM). Preliminary biological evaluation showed that the novel radiotracer, 123I-INER, presented a regional distribution consistent with known NET density in nonhuman primate brain, suggesting that 123I-INER was a promising imaging agent for NET in vivo using SPECT (Tamagnan et al.,2007). This study aims to characterize the in vivo pharmacokinetic properties of 123I-INER in nonhuman primate brain, including kinetic modeling analysis; and to measure the NET occupancy of atomoxetine and reboxetine in target regions, that is, the brainstem and midbrain, for six different atomoxetine doses and four different reboxetine doses.
- Top of page
- MATERIALS AND METHODS
Following i.v. bolus injection of the radiotracer, whole brain uptake peaked at 20 minutes for the baseline study at 2.5% of the injected radioactivity. Specific binding (regional uptake minus occipital uptake) reached a maximum at around 180 minutes. Marked accumulation of radioactivity was found at the level of the brainstem (i.e., locus coeruleus) and midbrain, which are known to be NET-rich regions (Figs. 1a and 2). Conversely, lower radioactivity accumulation was found in the cerebellum, caudate and occipital cortex. One hour post-injection, the parent fraction in venous and arterial plasma was 31.56 ± 0.81% and 25.39 ± 0.83% (mean ± SD, n = 3), respectively, reaching a value of less than 15% at 4 hours post-injection (Fig. 3).
Figure 1. Representative SPECT images showing the distribution of 123I-INER in baboon brain at baseline (a) and after preblocking with atomoxetine (0.85 mg/kg) (b), coregistered with MRI. Highest radioactive accumulation showed in red and lowest in green.
Download figure to PowerPoint
Figure 2. Representative time–activity curves of 123I-INER in multiple brain regions. Dashed lines and open symbols are after preblocking with high dose of atomoxetine (0.85 mg/kg).
Download figure to PowerPoint
Figure 3. Representative parent fraction profile in plasma over 4 hours after i.v. bolus injection of 123I-INER (top graph) and summary table of blood results obtained in three different baboons at selected time points (bottom table). Results presented as mean ± SD, n = 3. p.i. = postinjection.
Download figure to PowerPoint
Results from kinetic modeling showed that the two tissue compartmental model was the preferred model compared to the one tissue compartmental model to evaluate and quantify the pharmacokinetic properties of 123I-INER in baboon brain (lowest Akaike information criterion and Schwartz criterion values and highest model selection criterion values were obtained with the 2T model). Kinetic modeling of the baseline studies using the 2T compartmental analysis with either arterial or venous input functions showed that the highest VT and BPND values were found in the NET-rich regions, brainstem and midbrain, with lower VT and BPND values in the caudate, white matter, cerebellum, and occipital cortex (Table I). Pretreatment with a high dose of atomoxetine reduced the VT and BPND values in NET-rich regions to the level of non-target regions (Figs. 1b, 2 and Table I). The VT values obtained using venous input function correlated well with those obtained using arterial input function both at baseline and blockade conditions (r2 of 0.99 at baseline for all three baboons and r2 of 0.99 at blockade for baboon 2 and 3, no data available for baboon 1 at blockade conditions) but were underestimated compared to the VT values determined using arterial input function. Nonetheless, the BPND results derived using arterial input function correlated well with BPND values derived using venous input function and were close to the line of identity (Fig. 4). The BPND values for baseline and preblocking studies determined by SRTM were not statistically different to BPND values calculated using 2T model and arterial or venous input function (P > 0.5, ANOVA); and at baseline conditions the highest values were found in the brainstem and midbrain (SRTM mean BPNDs in the brainstem and midbrain of around 0.72 and 0.68, respectively). SRTM analysis also showed a reduction of brainstem and midbrain BPND following pretreatment with atomoxetine (Table I). The tissue to arterial or venous plasma activity concentration ratios (Ct/Cp) were plotted against time, and a transient equilibrium was found to happen around 3 hours postinjection. Representative Ct/Cp curves are shown in Figure 5, where it is possible to observe the transient equilibrium at 3 hours postinjection.
Figure 4. Comparative analysis between BPND values obtained from 2T compartment model using an arterial input function (x axis) and a venous input function (y axis) at baseline (a) and blockade conditions (b) (r2 = 0.99 for both graphs). Note close proximity of correlation to the line of identity. A total of 24 values were plotted for baseline study (3 baboons, each with 1 SPECT measurement × 8 brain regions) and 16 values were plotted for blockade study (2 baboons, each with 1 SPECT measurement × 8 brain regions).
Download figure to PowerPoint
Figure 5. Example of calculated Ct/Cp at baseline (a) and following preblocking with the highest dose of atomoxetine (0.85 mg/kg) (b) graphs. Note a reduction of Ct/Cp in NET-rich regions (brainstem and midbrain) to non-target levels following preblocking with atomoxetine. Ct/Cp was determined using venous blood samples results.
Download figure to PowerPoint
Table I. Kinetic analysis of bolus experiments using 2T compartment model with arterial or venous input function and SRTM modeling (baseline and atomoxetine). Note the reduction in the VT and BPND values in NET-rich regions when atomoxetine pretreatment was given in comparison with baseline measurements.
|VT arterial (ml/cc)||% COV||VT venous (ml/cc)||% COV||BPND arterial||BPND venous||BPND SRTM|
| Baseline||11.37 ± 1.48||5.55 ± 1.62||9.11 ± 2.29||5.21 ± 2.10||0.84 ± 0.08||0.84 ± 0.09||0.72 ± 0.01|
| Atomoxetine||6.15 ± 2.50a||4.30 ± 2.19a||5.95 ± 1.38||4.15 ± 1.88||0.02 ± 0.06||0.05 ± 0.11||0.13 ± 0.05|
| Baseline||10.78 ± 1.49||5.79 ± 0.88||8.52 ± 2.04||7.29 ± 2.02||0.74 ± 0.08||0.72 ± 0.08||0.68 ± 0.13|
| Atomoxetine||6.50 ± 2.59a||2.28 ± 0.81a||6.30 ± 1.50||3.09 ± 1.61||0.08 ± 0.05||0.11 ± 0.11||0.20 ± 0.04|
| Baseline||9.67 ± 2.36||8.58 ± 7.27||7.64 ± 2.65||7.20 ± 5.25||0.54 ± 0.20||0.51 ± 0.17||0.54 ± 0.20|
| Atomoxetine||7.36 ± 2.36a||11.91 ± 6.25a||7.00 ± 1.88||9.84 ± 9.63||0.24 ± 0.04||0.22 ± 0.03||0.21 ± 0.04|
| Baseline||8.59 ± 1.20||5.67 ± 1.15||6.78 ± 1.50||6.11 ± 2.34||0.39 ± 0.11||0.37 ± 0.10||0.36 ± 0.07|
|Atomoxetine||7.52 ± 1.95a||18.04 ± 17.99a||7.11 ± 2.00||19.92 ± 23.12||0.28 ± 0.13||0.23 ± 0.10||0.29 ± 0.15|
| Baseline||6.20 ± 0.84||1.62 ± 0.60||4.96 ± 1.24||1.85 ± 0.51||–||–||–|
| Atomoxetine||5.98 ± 2.12a||5.83 ± 1.24a||5.74 ± 1.51||5.96 ± 3.74||–|| || |
| Baseline||8.10 ± 1.32||3.14 ± 1.69||6.34 ± 1.56||2.10 ± 0.40||0.31 ± 0.12||0.28 ± 0.10||0.28 ± 0.04|
| Atomoxetine||6.33 ± 2.84a||5.04 ± 4.01a||6.07 ± 1.93||5.65 ± 5.77||0.04 ± 0.11||0.05 ± 0.11||0.08 ± 0.02|
| Baseline||8.21 ± 1.18||1.70 ± 0.26||6.51 ± 1.56||1.81 ± 0.13||0.32 ± 0.09||0.31 ± 0.09||0.32 ± 0.11|
| Atomoxetine||6.35 ± 1.54a||2.52 ± 0.01a||6.18 ± 1.37||2.59 ± 1.17||0.08 ± 0.12||0.09 ± 0.09||0.14 ± 0.12|
Quantification of NET occupancy using baseline and atomoxetine pretreatment studies and 2T compartmental analysis with arterial or venous input function showed that, at doses of 0.85 mg/kg, atomoxetine occupied around 95% and 80% of brainstem and midbrain transporters, respectively, as determined by the ΔBPND method (Table II). When the Lassen plot method was used to measure occupancy using the VT values calculated from the 2T compartment model analysis, the global occupancy of NETs by atomoxetine was around 75%. Applying the Lassen plot equation to derive VND for subsequent calculation of occupancy in individual brain regions showed that around 95% and 80% of brainstem and midbrain transporters were occupied by atomoxetine, respectively. By using the Lassen plot with VTapp, a global occupancy of around 75% was also determined and the NET occupancy in the brainstem and midbrain was found to be again around 95% and 80%, respectively (Table II). The NET occupancy in the brainstem and midbrain determined using the SRTM method was found to be around 85% and 75%, respectively. A representative Lassen plot used to derive the VNDs is shown in Figure 6. The comparison between VT and VTapp values for the occipital cortex determined using 2T modeling and tissue/blood ratios methods, respectively, and the VNDs values derived using the Lassen plot are shown in Table III. Statistical analysis showed no significant differences between these values for all investigated methods (P > 0.6, ANOVA).
Table II. NET occupancy in selected brain regions following pretreatment with a high dose of atomoxetine
|Brain regions||% Occupancy|
|Arterial blood||Venous blood||SRTM|
|ΔBPND indirect method||Lassen plot VTa||Lassen plot VTappa||ΔBPND indirect method||Lassen plot VTa||Lassen plot VTappa|
| Global occupancyb||–||–||–||–||76.4%||78.0%||–|
| Global occupancyb||–||72.0%||57.3%||–||71.2%||60.2%||–|
| Global occupancyb||–||81.8%||88.0%||–||72.9%||86.3%||–|
Table III. Summary of VND values obtained using the Lassen plot and occipital cortex VT and VTapp values obtained from 2T modelling and Ct/Cp method, respectively, using arterial or venous blood
| ||VT occipital baseline||VT occipital blockade||VND Lassen plot||VTapp baseline||VTapp blockade||VNDapp Lassen plot|
The displacement of 123I-INER by atomoxetine varied in a dose-dependent fashion (Fig. 7a), where the estimated ED50 values were 0.15 mg/kg in the brainstem (r2 = 0.61), 0.09 mg/kg in the midbrain (r2 = 0.71), and 0.10 mg/kg when using a mean of both regions (r2 = 0.75). The maximum occupancy was estimated to be 22% and 37% for the brainstem and midbrain, respectively (Fig. 7a), while the average occupancy of both NET-rich regions was estimated to be 29%. For reboxetine, again the measured occupancy was found to be dose-dependent with an ED50 of 2.33 mg/kg (r2 = 0.95) in the brainstem, of 0.44 mg/kg (r2 = 0.57) in the midbrain, and of 1.07 mg/kg using a mean of both regions (r2 = 0.94). The maximum occupancy determined from the fit was 104% and 56% for the brainstem and midbrain, respectively (Fig. 7b), while the average occupancy for both NET-rich regions was estimated to be 74%.
Figure 7. Percent occupancy following i.v. bolus injection of increasing doses of atomoxetine (a) and reboxetine (b). A one site hyperbola function fits well the experimental values: atomoxetine curves r2 = 0.61, 0.71, and 0.75 for brainstem, midbrain, and combined region, respectively; and reboxetine curves r2 = 0.95, 0.57, and 0.94 for brainstem, midbrain, and combined region, respectively.
Download figure to PowerPoint
- Top of page
- MATERIALS AND METHODS
This study evaluated 123I-INER pharmacokinetic properties in nonhuman primate brain using SPECT. 123I-INER accumulation in vivo at baseline was found to be consistent with known NET distribution in baboon brain, which is in agreement with our preliminary evaluation (Tamagnan et al.,2007). Furthermore, the decrease in specific binding of 123I-INER following preblocking or displacement using NET selective drugs in different baboons confirms pharmacological selectivity in vivo of the radioligand binding and gives support to 123I-INER as a useful radiotracer for NET imaging. Obtained kinetic data suggest that the occipital cortex may be used for quantification of BPND and NET occupancy in baboon brain, demonstrating the lowest accumulation of radioactivity and the lowest VT of all regions. However, in two monkeys the VT values were lower at preblocking conditions compared with baseline measurements. These observations suggest that the occipital cortex might not be a true reference region, but it may be used as a pseudo-reference region for quantification of BPND values and NET occupancy using SPECT with 123I-INER. The use of a pseudo-reference region model, that does not require the reference region to be devoid of specific binding, has been previously described and can be used as a method for protocol simplification. Such approach uses a region with the lowest target expression as the reference and makes a correction for the level of binding in that region (Gunn et al.,2011). The choice of a reference region for imaging studies of NET in brain has been highly controversial, even though both the striatum (caudate and putamen) and the occipital cortex have been used as reference regions for quantification of NET radiotracer binding in vivo both in humans and nonhuman primates (Takano et al.,2009; Zeng et al.,2009). Our results indicate that the occipital cortex might be useful as a reference region for quantification of BPND and NET occupancy with some reservations, as corrections will likely have to be applied for the level of binding in that region using the pseudo reference tissue model. Future work with a larger sample and preblocking studies with different doses of atomoxetine will be important to better understand the potential bias in the occupancy results obtained using the occipital cortex as reference region and SPECT with 123I-INER.
Reference tissue methods, using the occipital cortex as reference region, yielded a BPND similar to the one determined using invasive quantification methods, with no statistical differences found among BPND values derived using these three different methods of quantification (arterial or venous derived BPND and SRTM derived BPND, P > 0.5, ANOVA) both at baseline and preblocking conditions. Thus, noninvasive reference tissue methods might be useful in estimating BPND and occupancy of NET in baboon brain using SPECT with 123I-INER, eliminating the need for arterial blood sampling. Blood analysis showed that 123I-INER metabolic profile was similar in arterial and venous blood. Further analysis found that VT obtained using the arterial or venous plasma input function correlated well both at baseline and blockade conditions. Despite the VT values derived using venous input function being underestimated compared to the VT values determined using arterial input function, partially due to a higher parent fraction in venous blood compared to arterial blood, BPND values determined using both arterial and venous input functions correlated strongly with a relationship close to the line of identity. This suggests that kinetic modeling quantification by 2T compartment model may be successfully achieved by using venous blood samples rather than more invasive arterial blood. Although a bias is present when using venous blood sampling in comparison with arterial blood sampling for determination of VT values, this bias is unlikely to contribute significant errors in occupancy determinations, because BPND can be adequately determined using venous input function. Thus, when measuring occupancy following drug administrations, the use of venous sampling might provide a simple and adequate quantification of 123I-INER pharmacokinetic parameters, and it is more practical and reduces the risk of complications associated with arterial sampling. A trade-off between accuracy and simplicity of image analysis is typically necessary when exploring simplified methods in comparison to invasive more accurate methods. In summary, BPND and, in particular, occupancy quantification of 123I-INER in baboon may be successfully achieved using noninvasive modeling or invasive modeling with venous input function. Both observations can be used to simplify image analysis of studies conducted using 123I-INER. To our knowledge, 123I-INER is the only useful NET SPECT radiotracer where venous input function and reference tissue models may be used for BPND and occupancy estimations, with the caveat that the use of a reference region for data quantification might require a correction for the level of binding in that region, as explained above. Results obtained here might be underestimating BPND as ketamine and isoflurane have been reported to increase the levels of norepinephrine in rats and to decrease brain uptake, respectively (Hildebrandt et al.,2008; Takano et al.,2009). Although ketamine effects are theoretically short-lived owing to a short half-life, in awake animals measured BPND values may be higher than the ones reported in this paper. Isoflurane is the most commonly used anesthetic when performing in vivo imaging, because it allows for an easier control of the amount of anesthetic administered over time comparing to injectable anesthetics. In addition, it also allows for better maintenance of cardiac function, good muscle relaxation and rapid animal recovery follow anesthesia. Prior studies have shown that isoflurane reduced the brain uptake of 18F-FDG and 11C-SCH23390, a dopamine D1 receptor (Hildebrandt et al.,2008), thus it is possible that in awake animals, radiotracer uptake may be higher than the one reported in this article.
Results also showed a transient equilibrium was reached at around 3 hours postinjection, and VTapp could be estimated at that time point. When using the Lassen plot, the determined occupancy using VT values and VTapp values yield similar values in NET-rich regions, suggesting the that VTapp values may be used for occupancy measurement using the Lassen plot method. The Lassen plot was applied to obtain the global occupancy in brain following a preblock with atomoxetine, as previously described in the literature (Cunningham et al.,2010). The Lassen plot is a method typically used for quantification of global occupancy when no suitable reference region is available. In this study, the Lassen plot was also used to obtain VND, which was consequently used to calculate individual brain regions NET occupancy. For regions with high density of NETs, such as brainstem and midbrain, the occupancies determined using the Lassen plot methods were similar to the occupancies determined from the change of BPND derived from the 2T compartment models.
As previously mentioned, atomoxetine has been clinically used for treatment of ADHD (Michelson et al.,2001; Simpson and Plosker,2004; Witcher et al.,2003). Previous clinical trials have found that doses of 1.2 mg/kg were as effective as 1.8 mg/kg and the effect of 0.5 mg/kg was not significantly different from that of placebo (Michelson et al.,2001). The half-life of atomoxetine in plasma was found to be around 3 hours, although studies have found longer elimination times, as function of severe or reduce metabolizers (Simpson and Plosker,2004; Witcher et al.,2003). Our results from the displacement studies using multiple doses of atomoxetine showed that NET occupancy by atomoxetine is dose dependent with an estimated ED50 of 0.15 and 0.09 mg/kg for brainstem and midbrain, respectively. This suggests that for doses around 0.10 mg/kg half of the transporters are occupied by atomoxetine. The occupancy results plateau at around 30%, corresponding to atomoxtetine doses of around 0.85 mg/kg. In contrast, using the preblocking protocol, 0.85 mg/kg of atomoxetine was able to block 80% of NETs in baboon brain. This considerable difference in the measured occupancy of the displacement studies compared to the blocking study is due to technical difficulties in reaching equilibrium prior to displacement, as well as, insufficient data following displacement that makes difficult quantification of percent specific binding displaced. This suggests that percent specific binding displaced was overall underestimated and consequently, maximum occupancy might also be underestimated. Another important difference between the displacement and blocking experiments rely on the method used for drug administration. For displacement studies, the drugs were administrated as a bolus injection, while for preblocking experiments atomoxetine was given as a bolus followed by constant infusion throughout the study duration. This difference in the method of administration will result in different plasma availability levels of the injected drugs, further explaining the differences in results obtained using displacement study design compared to preblocking experiments. The measured occupancy of 80% for atomoxetine (0.85 mg/kg) in NET rich structures with SPECT and 123I-INER do not seem to correlate with clinical efficacy of atomoxetine previously described in clinical trials, where only for doses above 1.2 mg/kg clinical significance was reached. Such gap between obtained NET occupancy results and previously described efficacy of atomoxetine may be explained by the fact that our data only evaluate NET occupancy at peak time and not at trough levels; furthermore, clinical trials are typically designed to evaluate the effects of chronic dosing rather than single dose effects. Indeed, Takano and coworkers,2009, also found similar findings using (S,S)-[18F]FMeNER-D2 for evaluation of NET occupancy by atomoxetine. Implications of these observations remain to be elusive, but it was suggested by Takamo et al. (2009) that a sustained high NET occupancy at trough level could lead to different clinical outcomes.
Reboxetine has been used for treatment of depression over the last years, where doses typically between 4.0 and 10.0 mg per day have been prescribed to patients (Page,2003; Wong et al.,2000). Reboxetine is rapidly absorbed, with peak plasma concentrations being reached within 2 hours and an elimination half-life of approximately 13 hours (Page,2003). Studies on its therapeutic efficacy are contradictory and controversial. Recently published reviews showed its efficacy using doses ranging between 4.0 and 10.0 mg/kg was similar to placebo (Cipriani et al.,2009; Eyding et al.,2010), however, others have found significant improvements in reboxetine treated patients compared to placebo (Ferguson et al.,2002; Montgomery et al.,2003). Our results showed that reboxetine occupies NET in a dose-dependent manner, where an estimated ED50 of 2.33 and 0.44 mg/kg was found for the brainstem and midbrain, respectively. Maximum occupancy of NETs by reboxetine was found to be 104% and 57% for the brainstem and midbrain, respectively. This difference is likely to be due to the occupancies at 3.0 mg/kg driving the fit. When combining the brainstem and midbrain, the measured occupancy at 3.0 mg/kg of reboxetine was 54%, allowing a more reliable estimation of the maximum occupancy found to be 74%. Implications of our results remain elusive and whether high NET occupancy is required for the treatment of depression is unknown, but high doses of reboxetine seem to be needed to reach maximum occupancy in the brain. Further evaluation of NET occupancy by reboxetine in vivo may provide more insight on its pharmacokinetic profile and give the opportunity to improve dosing protocols and understand currently contradictory observations.
The ED50 values determined for reboxetine were higher than the estimated ED50 values for atomoxetine. As mentioned previously, using membranes from MDCK cell lines transfected with human NETs, the Ki of reboxetine was determined to be 11 nM, while the Ki of atomoxetine was determined to be 5 nM (Bymaster et al.,2002). The ED50 values determined in vivo using SPECT with 123I-INER are in agreement with in vitro observations, showing atomoxetine to have a higher affinity for the NET than reboxetine.
In conclusion, the data reported here suggest that the occipital cortex may be used as a reference region for studies using SPECT with 123I-INER; however, this would need to be confirmed in future studies using a larger animal sample and additional studies at moderate occupancy to evaluate potential bias. This study showed that noninvasive reference tissue models using the occipital cortex as reference region were able to measure 123I-INER BPND and occupancy values in vivo using SPECT; however, the occipital cortex might not be a true reference region, which will mean that corrections for the level of binding in that region will likely have to be applied. In addition, a reasonable estimation of occupancy can be made using VT determined by dividing the concentration in tissue by the concentration in plasma from blood samples at 3 hours post injection (i.e., VTapp), even though VTapp values were biased in comparison to VT values determined using 2T modeling. Both findings may simplify considerably data acquisition and image analysis, providing a feasible and easy platform for future nonhuman primates and, potentially, human studies. Furthermore, this study showed that atomoxetine and reboxetine occupied NET in a dose-dependent manner using 123I-INER and SPECT in nonhuman primate brain. The use of molecular imaging techniques, such as the ones presented in this study, could improve development of novel drugs or advance knowledge on currently developed drugs for treatment of ADHD, depression and other disorders associated with the NET system.