Prostate-specific membrane antigen (PSMA) is a type II transmembrane protein highly expressed in almost all stages of prostate cancer (PCa) (> 90%) [1]. PSMA ligands for imaging and therapy are now available worldwide. Robust data suggested that PSMA-based radioligands carried the highest diagnostic value in the imaging of PCa. A meta-analysis including 4790 patients showed that PSMA PET improved the detection of metastases with biochemical recurrence (BCR), particularly at low pre-PET prostate-specific antigen (PSA) levels of > 0.2 ng/mL (33%) and 0.2–0.5 ng/mL (45%) [2]. PSMA-targeted radionuclide imaging is now significantly impacting clinical decision-making in about 54% of patients [3]. More importantly, PSMA PET is now a pivotal part of the management of PCa in international guidelines. The Food and Drug Administration has approved [68 Ga]Ga-PSMA-11 for PET imaging of patients with PCa in 2020 [4]. With great success in the imaging field, PSMA radioligand therapy is an ongoing area of great potential. In the past decade, several PSMA-targeted radiopharmaceuticals have been developed for the treatment of advanced PCa, enabling the delivery of radiation to both bone and soft tissue metastases [5]. Besides that, PSMA tracer uptake by other types of cancers has been increasingly reported [6, 7]. Several case reports and small-size clinical trials have reported the accumulation of PSMA-targeted agents in both hepatic lesions and extrahepatic metastases [8,9,10]. The possible mechanism is due to the overexpression of PSMA on the newly formed microvessels in non-prostate solid cancers [11,12,13,14] (Fig. 1a).

Fig. 1
figure 1

Prostate-specific membrane antigen (PSMA)-targeted theranostics of solid tumors. a Schematic diagram of PSMA-targeted radioligand binding to tumor-associated neovasculature. b Body diagram of all solid tumors currently linked to PSMA radioligand therapy

Full size image

We read with great interest the paper by Lu et al. recently published in the European Journal of Nuclear Medicine and Molecular Imaging [15] investigating [68 Ga]Ga-PSMA-617 PET imaging of hepatocellular carcinoma (HCC) by targeting tumor-associated endothelium. The preclinical study presented by Lu and colleagues is interesting as there was no in-depth basic research exploring molecular mechanisms mediating HCC uptake of [68 Ga]Ga-PSMA. Through in vitro and in vivo experiments, the authors demonstrated that uptake of [68 Ga]Ga-PSMA-617 in HCC was mediated by tumor-associated endothelium, which is of great significance for understanding the concentration of PSMA-targeted radioligands in non-prostate cancers. The authors first showed that the uptake of [68 Ga]Ga-PSMA-617 in HepG2 and HuH-7 cells was very low and could not be blocked by a PSMA inhibitor, suggesting the non-specificity of the PSMA-targeted probe binding to HepG2 and HuH-7 cells. Nevertheless, direct detection of the in vitro uptake of [68Ga]Ga-PSMA-617 by vascular endothelial cells is also very important. So far, the study of PSMA expression in vasculature has been impeded as cultured human umbilical vein endothelial cells (HUVEC) are PSMA negative and both cell-derived xenografts (CDX) and patient-derived xenografts (PDX) models are not known to express PSMA in their vasculature [16, 17]. In recent years, a growing body of research has shown that the incubation of endothelial HUVEC cells in tumor-conditioned media could significantly increase its PSMA expression [17,18,19]. Thus, in our opinion, it would be better to test the uptake of [68Ga]Ga-PSMA-617 in endothelial cells with or without HepG2 and HuH-7 conditioned medium.

Furthermore, [68Ga]Ga-PSMA-617 PET imaging in both HepG2/HuH-7 CDX and HCC PDX models showed obvious radioactive accumulation in the tumors and the uptake could be largely blocked by co-injection of the PSMA inhibitor. Subsequent immunohistochemical tests indeed revealed that PSMA expression was mainly localized in vascular endothelium in the xenografts. PSMA inhibitor ZJ-43 can effectively block the uptake by HepG2 and HuH-7 tumors in vivo, but not in in vitro settings. The authors supposed the difference in blocking results between the in vivo and in vitro experiments reflected PSMA expression in tumor-associated vasculature. These findings are important since the results support the view that PSMA-targeted imaging tracers and therapeutics may have much broader applicability by enabling the management of non-prostate tumors. The different animal tumors used in this study are certainly a strong point, especially the PDX models from HCC patients because they enabled the examination of tumor tissue in a naïve environment without significantly affecting the heterogeneity and stromal architecture of the neoplasms [20]. Before this, the mechanism for in vivo [68Ga]Ga-PSMA-617 uptake by HepG2 and HuH-7 tumors but not by cultured HepG2 and HuH-7 cells remains rather unclear. The non-prostate expression of PSMA has been reported exclusively within the neovasculature endothelial cells of non-prostate cancers but not on adjacent normal endothelium, suggesting that tumor-related factors may induce PSMA expression by neovessels [17,18,19]. The molecular mechanisms of HCC cells on the expression of PSMA by neovasculature endothelial cells needs to be further revealed in future research.

Beyond its well-documented expression in > 90% of PCa [1], PSMA is also found in the endothelial cells of tumor-associated neovasculature of most non-prostate solid cancers, especially those tumor entities that critically depend on angiogenesis, yet not in normal endothelium [12, 21]. Preclinical studies demonstrated that PSMA might be involved in cancer-related angiogenesis by participating in integrin signal transduction and degrading the extracellular matrix [22,23,24]. Similar to the introduction of PSMA-targeted theranostics in PCa, overexpression of PSMA by neovasculature and the important role of PSMA in tumor angiogenesis makes it a potential target for imaging and treatment of other cancers through PSMA-mediated delivery of chemotherapeutics or radiation agents [12].

A subset of patients with HCC, clear cell renal carcinoma, salivary gland cancer, glioblastoma, and thyroid cancer have shown sufficient PSMA tracer uptake [6, 10, 25,26,27,28], suggesting patients with these tumors might potentially benefit from PSMA-targeted radioligand therapy (Fig. 1b). In 2019, a patient with metastatic adenoid cystic carcinoma of the parotid received PSMA radioligand therapy (PSMA-RLT) was reported [29], for whom one cycle of 7.5 GBq of [177Lu]Lu-PSMA was given. Post-therapy whole-body SPECT imaging showed intense uptake in the metastases. Treatment was well tolerated with no side effects and some pain relief was reported. Kumar et al. reported a case of recurrent glioblastoma multiforme who received three cycles of 3.7 GBq of [177Lu]Lu-PSMA-617. Post-therapy magnetic resonance imaging showed a partial response with tumor shrinkage and importantly improvement of the quality of life [30]. This suggested that a sufficient radiation dose can be delivered despite PSMA expression being limited to the tumor vasculature. To date, most of the studies reporting PSMA-targeted imaging and radioligand therapy of non-prostate cancers are case series or case reports. Likely, positive incidental findings are much more reported than negative ones, known as publication bias. It is worth noting that, in contrast to radionuclide imaging, the effect of radionuclide therapy critically depends on the long-lasting tumor concentration of the radionuclide [31]. In non-prostate cancer tumors, PSMA is mainly expressed on the neovasculature, leading to a shortened tracer accumulation and rapid washout [10]. In other words, PSMA-RLT is not retained in the tumor for a longer time, resulting in a lower radiation dose to the tumor and less therapeutic effectiveness [13]. Large sample, multicenter, and prospective studies are expected to thoroughly evaluate the theranostic value of this promising target outside PCas. Besides that, considering the expression of PSMA in several benign tissue types such as the salivary glands, small intestine, and renal tubules, the potential side effects of PSMA-RLT on these tissues should be noticed and long-term data will have to be gathered in larger patient populations.

Agents targeting vascular endothelial-derived growth factor (VEGF) and VEGF receptors (VEGFR) have achieved remarkable success in treating various types of cancers [32, 33]. Meanwhile, novel anti-angiogenic therapies and tumor vascular normalization are actively explored for effective cancer treatment [34,35,36,37]. By targeting PSMA expression in tumor neovasculature, PSMA-targeted theranostics may find more broad applications in tumors beyond PCas.