Introduction
The prostate-specific membrane antigen (PSMA) receptor is abundantly expressed on the cell surface of nearly all prostate cancer (PCa) cells [
1]. Targeting the PSMA receptor using radiolabeled PSMA ligands proved a valuable strategy for both diagnostic imaging of PCa as well as treatment in the advanced setting. Despite its name, PSMA (or glutamate carboxypeptidase II [GCP II]) was also identified in various other tissues, such as the small intestine, kidney nephrons and salivary glands [
2,
3]. The active target of a PSMA receptor consists of two binding sites, namely the glutamate-sensing pocket and a lipophilic binding pocket (arene-binding site), and the affinity of ligands for the receptor is increased after addressing both binding sites [
4‐
6]. Folates (including polyglutamates, monosodium glutamate (MSG), folic acid and 5-methyltetrahydratefolate) contain a glutamate structure and also target the arene-binding site, and consequently, folates could act as a competitor by blocking the binding of the PSMA radioligands [
7‐
9]. For diagnostic imaging, the potential effect of folates on the biodistribution could impact decision making, while during radioligand therapy this could even affect treatment efficacy.
The putative effect of folate intake on the biodistribution of diagnostic PSMA ligands has been demonstrated in two small prospective studies [
10,
11]. Harsini et al
. and Armstrong et al
. found that MSG administration (12.7 g and 150 mg/kg for the respective studies) prior to gallium-68 (
68Ga) PSMA-11 or fluorine-18 (
18F) DCFPyL administration lowered tracer uptake in salivary glands and tumor lesions significantly [
10,
11]. Apart from having an undesirable effect on imaging accuracy, folate co-administration has also been suggested as a potential approach to intentionally reduce organ uptake, and subsequent toxicity, during PSMA-based radionuclide therapy [
12]. As a first attempt applying this approach, Sarnelli et al
. showed that administration of folic polyglutamate indeed significantly reduced lutetium-177 (
177Lu) PSMA-617 uptake in salivary glands compared to previous dosimetry evaluations, though effects on tumor uptake were not assessed [
13]. Although extrapolation from diagnostic to therapeutic results is challenging, diagnostic tumor uptake was reduced after folate administration, which indicated limited benefits for this approach in clinical practice due to potential decreased treatment efficacy [
10,
11]. However, high doses of MSG (12.7 g and 150 mg/kg) were administered to patients in these studies and lower folate doses could potentially achieve saturation on organ tissue without affecting tumor uptake. Therefore, it is crucial to evaluate the effect of different folate doses and timing of these doses on biodistribution, which is also important since folates are ingested in low quantities daily and are found in many vitamin supplements [
11,
14]. In addition, considering a tumor sink effect, a greater impact of folate administration on tumor uptake in patients with a high tumor burden might be expected [
11].
To study the effects of folate intake and timing of intake on PSMA radioligand distribution, a physiologically based pharmacokinetic (PBPK) modeling approach was used. Main advantages of this noninvasive approach are that predictions of folate effects can easily be extrapolated to different clinical scenarios (e.g., comparing patients with different tumor volumes) and subsequent prospective trials are either not required or can efficiently be informed by the results [
15]. Hence, the aim of this study was to predict the effect of different doses of folate administrations (representing both folate containing food intake, vitamin supplements and high-dose folic acid administration) prior to [
68Ga]Ga-PSMA-11 PET/CT on uptake in salivary glands, kidney and tumors using PBPK modeling. In addition, the effect of different timings of folate intake and increasing tumor volume on the impact of folate administration on [
68Ga]Ga-PSMA-11 biodistribution was determined. Gained information based on these simulations could then guide future trial design or decision making for the use of folates in clinical practice.
Discussion
PBPK model simulations of folate intake together with [68Ga]Ga-PSMA-11 (t = 0) showed clinically relevant decreases in salivary glands and kidney uptake only with high folate doses (5 and 10 mg), while effects of folate on tumor uptake were not clinically relevant for all simulated doses. In all cases, folate intake by means of folate containing food (150 µg) or vitamin supplements (400 µg) did not have a relevant impact on [68Ga]Ga-PSMA-11 accumulation in salivary glands, kidneys and tumors.
An important issue associated with predicted effects from a PBPK model is the dependency of predictions on the underlying input parameters. During the sensitivity analysis, the input parameters were varied with 10%, while uncertainty in some parameters might exceed 10%. Model predictions for folate effects will be highly dependent on the folate affinity for the PSMA receptor. Therefore, a more extreme range in
KD was taken into account in model simulations (1–10.8 nM). Unfortunately, the exact affinity of folic acid and 5-MTHF for the PSMA receptor is unknown, while prediction intervals (see Fig.
2) proved the importance of this input value for predictions of the folate effect. The prediction range especially becomes larger for the largest decreases in relative accumulation (i.e., stronger folate effects). This is explained by the fact that in the case of a low impact, the folate concentrations are too low to achieve any effect and predictions are thus barely dependent on the affinity for the PSMA receptor. Conversely, in cases where folate does impact [
68Ga]Ga-PSMA-11 accumulation (especially at higher folic acid doses) the impact of affinity for the PSMA receptor is more important and predictions are very reliant on affinity input values.
Final PBPK model predictions for [
68Ga]Ga-PSMA-11 showed accumulation plateaus in the organs of interest at ~ 20 min post-injection. Wen et al
. reported plateaus reached in liver at ~ 10 min post-injection, while for kidneys and salivary glands uptake increased up to (at least) 60 min post-injection [
35]. For kidneys, this could be due to an increase of urine content containing radioactivity over time, while in our predictions [
68Ga]Ga-PSMA-11 in urine was not part of simulated concentration–time profiles for the whole kidney organ. In this way, we were able to show the effect of folate administration on the PSMA receptor-mediated uptake. For salivary glands, the difference between reported times of the reached plateau and our predictions was less apparent. However, the salivary glands uptake mechanism was assumed similar to other organs (regarding receptor binding and internalization), while it has been suggested that this mechanism might be partly non-PSMA specific [
36,
37]. Afshar-Oromieh et al
. showed that for most patients uptake in salivary glands does not increase any further after 60 min post-injection [
38]. Therefore, the plateau reached in salivary glands in our predictions is probably only slightly earlier compared to reported uptake, suggesting that non-PSMA receptor-mediated uptake does not play a major role in salivary glands exposure.
To put the predictions of the folate effect on organ uptake into perspective, these were compared to published clinical results. In our study, co-administration of 10 mg folic acid (at
t = 0) resulted in a 36% (range 31–45%) decrease of [
68Ga]Ga-PSMA-11 uptake in salivary glands and 34% (range 27–45%) decrease in kidneys, which is rather comparable to decreases reported by Armstrong et al
. (46% and 52% decrease in SUV
mean for salivary glands and kidneys, respectively) and Harsini et al
. (26–42% and 28% decrease in SUL
mean for all glands and kidneys, respectively) [
10,
11]. Rousseau et al
. also showed a reduced uptake in salivary glands and kidney, without affecting tumor uptake, after MSG administration in a preclinical setting [
39]. In addition, Rousseau et al
. reported a clear dose-dependent effect in mice (MSG dose range 164–657 mg/kg), while our findings only implied a dose-dependent effect in organs up to a folate dose of 5 mg (for administration at
t = 0). Differences between mice and humans, for example in receptor expressions and renal folate clearance, could explain these different findings regarding dose dependency of the effects.
Predicted effects of high doses folic acid on the normal organs and tumors should be perceived with some considerations. The larger predicted decrease in organ uptake after a 10 mg dose compared to 5 mg at time points 4 h and 12 h prior to [
68Ga]Ga-PSMA-11 could be due to an underestimation of renal clearance of folate. The fraction of the folate dose excreted renally increases with increased dosing [
40], but the PBPK model was not evaluated for 10 mg folate administrations due to lack of patient data. Therefore, the extrapolation to 10 mg in our predictions should be interpreted with some caution. Still, results for 10 mg at
t = 0 showed that no clear additional effect on organ uptake was predicted compared to 5 mg, which reflected maximum folate effects (probably due to full occupancy of the receptors). However, at the same time point, a decrease in tumor uptake was observed after 10 mg compared to 5 mg folate, which implied that tumor uptake will further reduce with higher folate doses. PCa, and especially metastatic lesions, show a clear overexpression of PSMA receptors compared to healthy human tissues. Higher folate doses will eventually also induce total receptor saturation in tumor lesions, and hence, reduced [
68Ga]Ga-PSMA-11 uptake in the tumors as a result of competitive binding. This would also clarify differences in our findings for tumor uptake compared to previously published results [
10,
11], where tumor uptake of PSMA ligands was significantly reduced after MSG intake. Armstrong et al
. reported a decrease of 38% in SUV
mean after 150 mg/kg MSG, while Harsini et al
. observed a 29% decrease in SUL
mean after 21.7 g MSG [
10,
11]. Extrapolating our model simulations to a folic acid intake of 36.5 g (comparable to MSG doses in the prospective studies), tumor uptake indeed decreased even further with a relative difference of 15% (range 2–42%). Therefore, based on our predictions as well as previously published patient studies, caution is warranted with dose selection in case of using folate administration to reduce organ uptake and doses > 10 mg possibly could negatively affect tumor uptake. On the other hand, low doses (< 400 µg) are not expected to affect [
68Ga]Ga-PSMA-11 biodistribution at all.
The timing of folate intake also impacts the effect on biodistribution, as the decrease in organ uptake seemed more profound in case of 10 mg 4 h prior to [
68Ga]Ga-PSMA-11 (59% decrease for salivary glands and 46% decrease for kidney) compared to co-administration (
t = 0). As already discussed, predictions with 10 mg folic acid should be interpret with caution, but still these findings could be explained by a delayed maximum plasma concentration of folates (especially the metabolite 5-MTHF) after oral ingestion of folic acid. However, this increased folate effect seemed diminished with administration 12 h prior to [
68Ga]Ga-PSMA-11, due to the relatively high clearance of the folate metabolites from the systemic circulation [
41].
Radioligand accumulation is even more crucial when considering PSMA-based radionuclide therapy, as accumulation in salivary glands, bone marrow and kidneys is known to induce dose-dependent toxicities and insufficient tumor uptake can lead to a reduced therapy efficacy. Unfortunately, direct translation of our simulations to predict the effects of folate intake on [
177Lu]Lu-PSMA remains challenging. The main difference between those ligands is the total administered peptide amount, which is an almost 50-fold higher during therapy. Thus, a probable assumption is that also larger folate doses are needed to compete with PSMA ligands for the PSMA receptor during therapy. Another approach to reduce organ uptake by competing for the PSMA receptor could be to add a cold PSMA ligand during the administration of the radiolabeled PSMA ligand. A dose of 5 mg folic acid (0.0103 mmol) would then be comparable to, for example, administration of 9.75 mg PSMA-11 [
42], which is somewhat higher compared to the highest cold PSMA-11 mass that was suggested by Kalidindi et al
. (5.30 mg) [
43]. However, one major difference that needs to be considered is that PSMA-11 probably has an increased affinity to the PSMA receptor compared to folates, resulting in a more profound effect, and thus, more prominent reductions in organ and tumor uptake. In case one would design a prospective trial with folate intake prior to [
68Ga]Ga-PSMA-11, we would recommend a rather low dose of 5 or 10 mg folic acid (at
t = 0). Minor or negligible effects on tumor uptake are expected, while organ uptake is probably reduced in a clinically relevant manner. Potential dose extrapolations could be performed while examining and evaluating effects on organ and tumor uptake, since a decrease in tumor uptake might be expected.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.