Discussion
The current study investigated the short-term impact of castration-induced AD on the uptake of metabolic PET tracers - [
18F]FDG, [
11C]choline and [
11C]acetate - in PCa cells
in vivo. The percentage changed uptake, based on SUV
max and SUV
meanTH, of each tracer was evaluated in control and AD-treated androgen-sensitive (LAPC-4) and androgen-independent (22Rv1) PCa xenografts in order to assess the impact of androgen on tracer uptake. In the clinical setting, SUV
max is usually the preferred value for evaluating therapy response after AD. Nevertheless, this approach quantifies only the highest tumour uptake localised in one voxel, which is not a true characterization of the metabolic status of the entire tumour. For this reason, we additionally looked at the effect of AD on the average tracer uptake in viable tumour cells. A threshold (SUV
meanTH) was set to exclude uptake values in necrotic tumour tissue, which was equally present in tumours of control and castrated animals as confirmed by histology. Moreover, since tumour sizes were relatively large and the main objective was to accurately quantify tumour uptake, FBP had minor drawbacks as compared to iterative reconstruction while avoiding possible positive bias due to iterative reconstruction [
19]. Therefore, all data were reconstructed with FBP using the same reconstruction protocol.
Regarding the image acquisition of [11C]choline and [11C]acetate, optimal time points following the injection of both tracers were chosen, taking into account background detection and tumoural uptake at a steady state. In the clinical setting, [11C]choline PET imaging is preferably acquired 5 min post injection. On the other hand, centres using [11C]acetate acquire images 30 min post injection. A high tumour uptake was observed very early after the injection of [11C]acetate. However, at that time, tracer uptake was also high in the liver and kidneys. Since background uptake of [11C]acetate diminished over time while tumour uptake remained rather stable, a static scan 30 min after tracer injection was the most optimal acquisition time point for tumour imaging with [11C]acetate.
Previously, the late-term effect of AD on the uptake of [
18F]FDG and [
11C]choline has been characterised in hormone-dependent and hormone-independent PCa xenograft models [
20]. The authors demonstrated the lower efficacy of [
11C]choline PET compared to that of [
18F]FDG PET for tumour detection and therapy response assessment
in vivo after surgical castration. Likewise, in this study, not only the trapping of [
11C]choline, but also of [
11C]acetate in the xenografted tumours was lower than that of [
18F]FDG; however, the clinical relevance of [
11C]choline and [
11C]acetate and their superiority to [
18F]FDG for PCa detection have been shown. In the current study, we focussed on evaluating the effect of intratumoural response to AD on PET imaging with both these tracers as well as with [
18F]FDG [
5]. Moreover, in order to minimise partial volume effects in imaging assessment, the xenografted tumours had to have a minimal volume of 150 mm
3. However, at this point, the proliferation rate of 22Rv1 tumours was rather fast, wherefore long observation periods after AD were not possible. Kukuk et al. evaluated the effect of AD on tracer uptake 2 and 3 weeks following castration [
20], which would reflect follow-up imaging of treated PCa in a clinical setting more precisely. However, Oyama et al. demonstrated at an early stage androgen-regulated effects on metabolic PET imaging in androgen-sensitive PCa [
18]. In this study, PCa xenografted tumours were androgen-deprived for 1 week by means of diethylstilbestrol (DES) administration. AD significantly diminished the uptake of [
18F]FDG, but not [
11C]acetate in androgen-sensitive PCa [
18]. Taken these findings into account and the fact that androgen-induced effects on gene expression were previously observed in preclinical models after several days [
21,
22], we investigated the short-term effect of AD, i.e. 5 days following the start of the treatment, on metabolic PET imaging of PCa xenografted tumours.
In the current study, AD was successful in all animals, but still, no significant change on tumour growth was observed in castrated animals. This observation can however be explained by the fact that AD alters gene expression - biological/molecular changes - rather fast in time, whereas its impact on tumour growth - morphological changes - is mostly a long-term effect. Since all metabolic tracers involved in the study are not specific, the effect of gene alteration on the uptake of the molecular probes might occur later in time. Therefore, an AD-induced inhibition of tumour proliferation, which could have been expected in especially the androgen-sensitive LAPC-4 model, was probably not detected in the present study. Also, we evaluated androgen-mediated effects by means of castration on the uptake mechanisms of metabolic tracers. Although surgical castration significantly reduces testicular androgen synthesis, adrenal androgen, which provides 10% of the total androgen production, remains present. Additionally, in tumour cells, bioconversion of the adrenal androgen androstanediol to dihydrotestosterone takes place of which the latter compound causes AR transactivation [
23]. Hence, surgical castration is not equivalent to medical treatment, i.e. AR antagonists, which reduces androgen activity locally and in turn also limits the effect of adrenal and tumoural androgen. Therefore, medical treatment could further extend the effect of AD on tumour proliferation and PET imaging.
Concerning the effect of androgen on PET imaging of PCa patients, AD using bicalutamide has been shown to significantly decrease the uptake of [
11C]choline in patients undergoing PET/CT for preoperative staging of PCa [
12]. Unlike that in patients, [
11C]choline uptake was not significantly reduced in AD-treated compared to untreated hormone-naive xenografts (LAPC-4) in this experiment. Because LAPC-4 cells are characterised by the expression of the wild-type AR, they are assumed to have a homeostatic AR signalling. A significant androgen-controlled effect on choline metabolism has further been observed in androgen-sensitive, wild-type AR-expressing PCa cell lines [
24,
25]. Based on these
in vitro observations, AD-induced alterations on metabolic imaging could have been expected in the current study. Moreover, Jadvar et al. evaluated AD-induced effects on [
11C]choline uptake in androgen-sensitive PCa xenografted tumours using autoradiography [
26]. Tumour uptake of [
11C]choline was determined 5, 10 and 20 min post injection. Results demonstrated a significantly higher tracer uptake in tumours of castrated animals at only the first two time points. Compared to our study, AD treatment was induced before androgen-sensitive tumour cells were injected subcutaneously and thus before solid tumours were developed. Although differences in [
11C]choline uptake in these tumours were assumed to result from androgen action, it is not obvious whether this is completely true. In fact, androgen-sensitive cells injected into an androgen-depleted environment are exposed to additional stress as compared to those cells injected into an androgen-rich environment (control animal). Since PSA expression was not monitored in this study and no baseline [
11C]choline uptake in the cells could be measured, it could be questioned whether androgen-sensitive cells grown under AD developed a truly androgen-sensitive solid tumour after certain weeks. Hence, this would hamper the interpretation of AD-induced effects on [
11C]choline uptake in androgen-sensitive PCa. Also, autoradiography quantifies the tracer uptake in several tissue slices of a certain organ, while PET offers a 3D image of a region of interest wherefore tracer internalisation in a specific tissue compartment of a living subject can be approached more precisely. The experimental set-up and imaging method used differed from those chosen in our current study, which is most likely the reason for different results obtained.
In patients with recurrent PCa, the development of biochemical failure during AD is observed over time. This stage of the disease - CRPC - is characterised by an increasingly androgen-independent proliferation [
27]. Nevertheless, androgen may still affect the expression of genes controlled by AR signalling and the progressive behaviour of tumours. Therefore, at this stage of the disease AD could diminish the uptake of metabolic PET tracers which in turn would suggest a favourable therapy response and delay the detection of recurrences. Hence, this study also examined androgen-induced effects on metabolic PET imaging of the androgen-independent yet androgen-responsive PCa xenograft model 22Rv1. AD decreased the uptake of [
18F]FDG and significantly of [
11C]choline
in vivo, while tumour histology in control and castrated animals was similar; however, [
11C]acetate PET imaging remained unaffected in both experimental groups. These observations are in line with our previous findings demonstrating an androgen-induced increased uptake of [
18F]FDG and [
11C]choline in 22Rv1 cells [
25]. Although [
11C]choline PET is currently used for the detection of recurrent PCa before and following AD treatment of PCa, it is not yet clear whether AD should be withdrawn on a regular basis before [
11C]choline PET is performed in androgen-resistant PCa. A recent study of Fuccio et al. demonstrated a diminishing effect of AD on the uptake of [
11C]choline in recurrent PCa which supports the retraction of AD therapy before follow-up imaging in order to increase the sensitivity of [
11C]choline PET [
13]. The current results support these findings, and it is noteworthy that caution may need to be taken for imaging recurrent PCa undergoing biochemical failure during or following AD.
AD treatment by means of surgical castration or AR antagonist both affect tumour viability, which can be detected by nuclear imaging techniques. However, in order to compare tracer internalisation in AD-treated versus non-treated animals accurately, tracer metabolism should remain as far as possible unaffected by AD. The biodistribution of [
18F]FDG and [
11C]acetate was studied in control and DES-treated animals [
18]. In healthy rats and tumour-bearing mice, AD did not alter tracer clearance from the blood. Also, the uptake of both tracers in background tissue compartments such as the liver, muscle and heart was not significantly affected [
18]. Therefore, the biodistribution profile of [
18F]FDG and [
11C]acetate is assumed to remain unchanged after AD. This would mean that tracer clearance should not affect PET imaging. Regarding [
18F]FDG PET, these data were further confirmed by Jadvar et al. [
28]. In a later study of this group, [
11C]choline biodistribution was studied in castrated and non-treated tumour-bearing mice using autoradiography. The authors concluded that AD did not change clearance of [
11C]choline [
26]. In our present study, we did not observe an altered clearance of [
18F]FDG, [
11C]choline and [
11C]acetate following castration, which supports these previously obtained results.
Besides therapy, micro-environmental factors play an important role in tumour behaviour and can affect the evaluation of metabolic PET imaging following the treatment of PCa. Necrosis and inflammation, of which the latter is a main therapeutic effect, interfere with therapy response assessment using PET. However, in this study, both micro-environmental factors were equally observed in tumour tissues of control and castrated animals. Taking this into account and also the observation that AD did not affect [11C]acetate uptake in androgen-sensitive and androgen-independent xenografted tumours, this tracer may be applicable for assessing tumour viability after AD treatment at all stages of PCa. The potential use of [11C]acetate PET as well as its advantage over the currently mostly employed [11C]choline PET for the evaluation of advanced PCa following AD should be considered. However, because the uptake of [11C]acetate was rather low and did not visualise proliferation as well as that of [18F]FDG and [11C]choline, we denote the significance of further preclinical and clinical evaluation in order to confirm the diagnostic efficacy of [11C]acetate PET for detecting the viability of PCa lesions before and after therapy.