In the current issue of the European Journal of Nuclear Medicine and Molecular Imaging (EJNMMI), Yoo et al. [1] reported that [64Cu][Cu(ATSM-FITC)] positron emission tomography (PET) could accurately detect hydrogen sulfide (H2S) in brain pathophysiology, opening a new horizon for detecting neuroinflammation by mapping H2S level. In the below paragraphs, we first would like to introduce several targets and associated tracers used for detecting neuroinflammation and then the theranostic potential of [64Cu]Cu-ATSM in preclinical and clinical settings. At the end of the manuscript, we will highlight the findings and potential applications of [64Cu][Cu(ATSM-FITC)] reported by Yoo et al. By introducing the background and the most recent proceedings, readers may get a balanced understanding of the relevant information.

Brief introduction of neuroinflammation and associated tracers

Neuroinflammation is a key biological process in response to cell infection or injury that involves all the cells present within the central nervous system (CNS), including microglia, astrocytes, neurons, and macroglia [2,3,4]. Neuroinflammation often refers to the activation of the neuroimmune cells microglia and astrocytes. These glia cells provided pro-inflammatory and anti-inflammatory functionality and are involved in neurodegeneration, such as Alzheimer’s disease (AD), Parkinson’s disease (PD), and multiple sclerosis (MS), and psychiatric disorders, such as major depression disorder (MDD), schizophrenia and psychosis, and substance use. PET can visualize, characterize, and measure neuroinflammation in the brain by targeting different biomarkers from macrophages to angiogenesis [5].

The most common neuroinflammatory target is the 18 kDa translocator protein (TSPO), formerly known as peripheral benzodiazepine receptor (PBR) [6,7,8]. Several radiotracers were developed to quantify TSPO expression, such as the first-generation R-[11C]PK11195. However, R-[11C]PK11195 displayed low specific binding [9, 10]. Two second-generation tracers ([18F)DPA-714 and [11C]PBR28) showed about 1.5- and 5-folds higher affinity than R-[11C]PK11195 [11]. [18F]GE-180 is a so-called third-generation TSPO tracer with an even higher binding affinity, which provides a higher target-to-background ratio compared to these tracers [12,13,14]. With the development of these tracers, we can measure both activated microglia and astroglia reflecting by TSPO in AD, MS, and even depression.

There is increased TSPO expression in the frontotemporal cortex and slightly higher in the neocortex of patients with AD compared to that in healthy controls [4]. More recently, microglial activation was observed to propagate in Braak stages jointly as tau pathology by the [11C]PBR28 PET study [15]. Most PET studies also indicated elevated TSPO binding in the anterior cingulate cortex or prefrontal cortex in participants with major depressive disorders [3]. MS is the most common neuroinflammatory disease which caused nontraumatic disability. As shown in Fig. 1, MS displayed increased TSPO expression in white matter lesions identified with magnetic resonance imaging (MRI); it was observed in both relapsing–remitting MS and secondary progressive MS [12, 16,17,18]. These TSPO binding patterns in white matter lesions are indistinguishable on MRI, which suggests that TSPO PET can detect pathophysiological heterogeneity of MS to which MRI is not sensitive or specific [4]. Furthermore, non-lesional white matter and gray matter in patients with MS also showed greater TSPO binding than that in age-matched healthy controls.

Fig. 1
figure 1

Representative [18F]DPA-714 PET in a 28-year-old female patient with MS. T2-weighted MR (left) and TSPO PET (right) images were provided by the PET center, Huashan Hospital

Full size image

But there were several limitations for TSPO as a biomarker of neuroinflammation. Firstly, the expression of TSPO by astrocytes and in the vascular endothelium was not neglected other than its expression in the brain by microglia. Secondly, TSPO PET is sensitive to polymorphism (rs6971) in the TSPO gene, which generated high-affinity binders (two copies of the major allele), mixed-affinity binders (heterozygous allele), and low-affinity binders (two copies of the rare allele). It means different PET signals could be produced by individuals with the same TSPO density but different genotypes. Thirdly, the lack of a true reference region requires kinetic modeling through the use of the metabolite-corrected arterial input function for accurate measurement of its density. These shortcomings limited the application of TSPO PET, especially the quantification of TSPO. More importantly, TSPO PET was found to reflect the density of inflammatory cells rather than their activation in humans, because its expression in human myeloid cells is related to different phenomena. Therefore, TSPO PET can reflect activated microglia in rodents but not in humans, which means the interpretation of TSPO PET data requires revision [19]. Therefore, more neuroimmune imaging biomarkers are necessary to overcome the limitation of TSPO PET. Novel inflammatory targets, including P2X7, P2Y12, colony-stimulating factor 1 receptor (CSF1R), monoamine oxidase B (MAO-B), cyclooxygenase isoenzymes COX-1, and COX-2, were validated in human diseases.

Purinergic receptors were classified as P1 and P2 receptors. P2 receptors are further subdivided into P2X and P2Y receptors; seven P2X receptors and eight P2Y receptors were identified. Among them, P2X7 and P2Y12 are the most promising targets for imaging neuroinflammation. The P2X7 receptor is expressed on immune cells such as monocytes, macrophages, and microglia. However, the P2Y12 receptor is only expressed on microglia but not on peripheral macrophages in the brain [20]. Some P2X7 tracers were developed and evaluated in humans, such as [11C]GSK1482160, [11C]JNJ-54173717, and [18F]JNJ-64413739. Although P2X7 PET displayed excellent results in MS, the application in other diseases is still not successful. P2X7 PET demonstrated no difference between amyotrophic lateral sclerosis (ALS) and controls and between PD and controls [21, 22]. For the P2Y12 receptor, several radiotracers were developed, but currently, no tracer can cross the blood–brain barrier (BBB) [23]. Therefore, the development of the brain penetrating P2Y12R PET tracer is still urgent.

CSF1R is another promising neuroinflammatory biomarker which predominantly expressed by microglia in the CNS. Its expression in other cells, such as neurons, is limited. It was also reported to involve in neurodegeneration and psychiatric disorders. CSF1R was used as a therapeutic target in various autoimmune disorders and cancers. Mutations in the CSF1R gene causing an autosomal dominant disease of hereditary diffuse leukoencephalopathy with spheroids (HDLS) also emphasized its significance [24, 25]. However, suitable CSF1R radiotracers are not available now. [11C]CPPC and [11C]GW2580 are the most promising ones investigated in inflammatory animal models [26, 27]. An autoradiography study also confirmed the increased (75–99%) radiotracer binding in the AD brain by [11C]CPPC [27]. The major issue of [11C]CPPC is that the sensitivity is not adequate to detect the low density of CSF1R in the healthy brain.

MAO-B is located both in neurons and astrocytes and overexpressed under pathological conditions associated with astrocytosis. MAO-B is considered a good biomarker of astrocytes and displayed a significant role in the treatment and diagnosis of neurodegenerative and other brain disorders [28]. PET imaging of MAO-B could provide opportunities for the quantification of astrocytosis. The first generation of MAO-B tracer of L-[11C]deprenyl was originally developed in the 1980s [29]. Recently, a novel MAO-B tracer, [18F]SMBT-1, was developed from the tau tracer, [18F]THK-5351 [30, 31]. [18F]SMBT-1 displayed high binding affinity and selectivity to MAO-B and reversible kinetics. It is also valuable to measure astrocytosis in AD patients.

COX is essential in the synthesis process of pro-inflammatory prostanoids, and their release as inflammatory mediators is significant to neuroinflammation. COX-1 is mainly localized in microglia, contributing to pro-inflammatory responses, while COX-2 is also expressed in neurons. COX-2 is quickly and dramatically upregulated by inflammation. Therefore, COXs were used as the biomarkers for the treatment and diagnosis of chronic inflammatory diseases, such as pain, fever, arthritis, and neurodegeneration. Two selective tracers were validated in human volunteers, [11C]PS13 for COX-1 and [11C]MC1 for COX-2. [11C]PS13 displayed the highest uptake in the hippocampus and occipital cortex, followed by the pericentral cortex and adjacent neocortices in the brain [32]. Physiological COX-1 expression in these regions was confirmed by COX-1 mRNA expression in the healthy human brain. [11C]MC1 was investigated in patients with rheumatoid arthritis, and increased binding in the affected joints was observed. [11C]MC1 also found no detectable binding in the brain of healthy volunteers. However, patients with rheumatoid arthritis displayed elevated brain uptake, in an order of neocortex, subcortical gray matter, and cerebellum [33]. The ability of [11C]MC1 in the measurement of COX-2 expression was also validated in the rhesus monkey brain and periphery after lipopolysaccharide injection [33].

Although tremendous targets were investigated as neuroinflammatory biomarkers, most of them targeted glia cells. The shortcomings of these targets and associated tracers also limited their applications. Therefore, discovery of new targets and development of new tracers are also necessary. Although traditionally known as a toxic gas, accumulating evidence supports that hydrogen sulfide (H2S) plays important roles in CNS diseases and traumatic brain injury [34,35,36,37]. Measurement of H2S levels in the brain could be a new strategy to monitor the CNS diseases.

Cu-ATSM in imaging cancers and neurodegenerative diseases

The copper II glyoxal bis(4-methyl-3-thiosemicarbazone) (Cu(II)-GTSM) (Fig. 2a) can deliver exogenously bound Cu directly into cells and activate pathways where Cu is a key cofactor. Copper II diacetyl-bis(4-methyl-3-thiosemicarbazone) (Cu(II)-ATSM) (Fig. 2b) is a biosimilar and only releases the exogenously bound Cu under hypoxic conditions. Both Cu-ATSM and ATSM are quickly cleared from the circulation with corresponding half-lives (T1/2) of 21.5 and 22.4 min, respectively [38]. The BBB is a highly selective gateway regulating the influx and efflux of molecules from the systemic circulation into the brain and in the reverse direction. Cu(II)-ATSM has enhanced permeability and can penetrate the BBB [39].

Fig. 2
figure 2

Chemical structures of a Cu(GTSM), b Cu( ATSM), and c Cu(ATSM-FITC)

Full size image

Both the properties and species of Cu affect the cellular uptake [40, 41]. Although the full mechanisms underlying the uptake of Cu-ATSM remain to be elucidated, the proposed mechanisms are shown in Fig. 3. One potential mechanism is the penetration of Cu(II)-ATSM into mitochondria and reduction of Cu(II) to Cu(I), leading to irreversible retention of metal or radiometal in the hypoxic cells [42, 43]. More recently, Yoshii et al. found that uptake of Cu-ATSM not only reflects hypoxia but also indicates the over-reduced intracellular states caused by mitochondrial dysfunction [44]. In normoxic cells, the Cu(I)-ATSM compound is rapidly re-oxidized into Cu(II)-ATSM by molecular oxygen and washed out from the cells [45, 46]. Another potential mechanism is the disassociation of Cu(II) from copper(II) complexes and the reduction of Cu(II) to Cu(I) by reductases in the circulation or tumor microenvironment. Hypoxic tumor cells have increased expression of copper transporter 1 (CTR-1) which mediates the transportation of Cu(I) into tumor cells [47].

Fig. 3
figure 3

Mechanisms mediating cellular uptake of retention of Cu-ATSM under normal and hypoxic conditions. Cu(II)-ATSM readily penetrates cells due to its low molecular weight, high membrane permeability, and low redox potential. In over-reduced cells such as when cells are under hypoxic conditions, the Cu(II) in Cu(II)-ATSM is reduced to Cu(I), which is released from the ATSM and trapped inside the cells. Another recently proposed mechanism accounting for the deposition of the tracers is the formation and deposition of CuS when there is an increased level of H2S in the CNS [1]

Full size image

When labeled with Cu isotopes (60Cu [half-life, 0.395 h; β+-decay, 92.5%; electron capture, 7.5%], 62Cu [half-life, 0.16 h; β+-decay, 98%; electron capture, 2%], and 64Cu [half-life, 12.7 h; β+-decay, 17.4%; β−-decay, 38.5%; electron capture, 43%]), they can be used for tumor hypoxia imaging and blood perfusion visualization. In 1996, Fujibayashi et al. reported that [62Cu]Cu-ATSM had high brain and heart uptake but was quickly washed out from these tissues. Moreover, [62Cu]Cu-ATSM had significantly increased heart uptake under hypoxic conditions [48]. This was followed by a translational study in 2000 reporting the safety profiles, rapid clearance from circulation, and accumulation of the tracer in the tumors within minutes [49]. Meanwhile, Lewis et al. elucidated that uptake of Cu-ATSM was dependent on tissue oxygen pressure and thus validated the value in imaging hypoxia [50, 51]. In non-small cell lung cancers (NSCLCs), the uptake patterns of [62Cu]Cu-ATSM and 18F-FDG differed in squamous cell carcinoma (SCC) and adenocarcinomas [52]. Similarly, in a recent study including 30 patients with head-and-neck cancers, Okazawa and co-authors demonstrated that high accumulation of [62Cu]Cu-ATSM in the periphery and intense uptake of 18F-FDG in the center were seen in SCCs. In comparison, the uptake patterns of the two tracers in adenocarcinomas were homogeneous [53]. Ikawa et al. reported that PET imaging with [62Cu]Cu-ATSM could visualize the regional oxidative stress in a patient with mitochondrial myopathy, encephalopathy, lactic acidosis, and stroke-like episodes (MELAS) [54]. Oxidative stress and mitochondrial dysfunction may contribute to the pathogenesis of PD. The same group then elucidated the uptake patterns of [62Cu]Cu-ATSM in patients with PD, reporting a higher accumulation of [62Cu]Cu-ATSM in the striata of the PD patients than that in the controls [55]. In patients with cerebrovascular disease, the information of dynamic [62Cu]Cu-ATSM correlated well with that of 15O tracers in assessing cerebral blood flow and oxygen extraction fraction [56]. [62Cu]Cu-ATSM uptake was significantly higher in grade IV than in grade III gliomas and correlated well with HIF-1α expression [57]. In addition to probing hypoxia, Dehdashti et al. further reported that the tumor-to-muscle activity ratio of [60Cu]Cu-ATSM could predict the treatment responsiveness in NSCLCs and cervical cancers [58, 59].

Compared to 60Cu or 62Cu-labeled agents, 64Cu-labeled agents are longer-lived radiopharmaceuticals that will facilitate shipping to multiple centers for multi-center clinical trials. In particular, β− decay (38.5%) and Auger electron emission of 64Cu open the possibility of therapeutic applications with [64Cu]Cu-ATSM under proper conditions [60, 61]. [64Cu]Cu-ATSM largely accumulated around the outer rim of tumor masses where hypoxic but active tumor cells resided [62]. Granted as an Investigational New Drug by the US Food and Drug Administration, a head-to-head comparison of [60Cu]Cu-ATSM and [64Cu]Cu-ATSM showed a significant uptake correlation of the two tracers in 10 patients with uterine cervical cancers. It is notable that [64Cu]Cu-ATSM PET images had better tumor-to-background ratios than that of [60Cu]Cu-ATSM [63]. Moreover, among several radiotracers for imaging tumor hypoxia, [64Cu]Cu-ATSM has potential advantages over others in several aspects. [64Cu]Cu-ATSM may have enhanced BBB penetration over [18F]FAZA which is cleared from the urinary system [64, 65]. [64Cu]Cu-ATSM had higher uptake in the hypoxia tissues and more rapid washout in normoxic cells than [18F]fluoromisonidazole ([18F]FMISO), the latter requiring at least 2-h equilibrium before scanning is commenced [39, 45, 66]. However, Little et al. found that [18F]FMISO but not [64Cu]Cu-ATSM or [64Cu]Cu-ATSE identified hypoxic regions in acute ischemic stroke models [66], validating the negative results previously reported [67]. So far, nitroimidazole-derived tracers, such as [18F]FMISO and the sugar-coupled tracer [18F]FAZA, are still the prime contenders in imaging hypoxia [68].

[64Cu]Cu-ATSM is currently in clinical trials for imaging tumor hypoxia in locally advanced rectum cancer (NCT03951337) and bulky tumors (NCT04875871). Automated cyclotron production and synthesis of [64Cu]Cu-ATSM with a commercially available synthesis module (GE Tracerlab™ FX2 N) has been reported by Liu et al. [69]. Multiple administration of [64Cu]Cu-ATSM may result in liver toxicity [70, 71]. Administration of penicillamine, a heavy metal chelator, could reduce radiation absorption doses in critical organs such as the liver and small intestine [72]. Different formulations of [64Cu]Cu-ATSM have been developed and validated by two research groups for diagnostic and theranostic applications, respectively [38, 63]. The potential risk associated with the chemical impurities from [64Cu]Cu-ATSM degradation is negligible, even in a therapeutic dose of [64Cu]Cu-ATSM [73]. PET imaging with [64Cu]Cu-ATSM may provide clinically relevant information about tumor oxygenation (hypoxia) and value in predicting therapeutic responses and survival in patients with solid tumors [74]. While most of the clinical trials are planned to image tumor hypoxia and predict the therapeutic responses of [64Cu]Cu-ATSM, there are also clinical studies evaluating the safety profiles and preliminary efficacies of Cu(II)ATSM in patients with PD or amyotrophic lateral sclerosis/motor neuron disease. The information on several finished and ongoing clinical trials is summarized in Table 1. Meanwhile, [64Cu]Cu-ATSM is fully exploited as a theranostic agent in Japan. Therefore, uncovering mechanisms other than the well-recognized radiobiological effects may help understand the therapeutic effects [75, 76].

Table 1 Registered clinical trials using [64Cu]Cu-ATSM or Cu(II)ATSM
Full size table

[64Cu]Cu-ATSM-FITC is an emerging agent sensing hydrogen sulfide

As mentioned above, accumulating evidence indicates that H2S plays important roles in the pathogenesis of a range of brain tissues [77]. A recent study elucidated that Cu(II)-ATSM, but not Cu(II)-GTSM, enhanced P-glycoprotein (P-gp) expression and may further contribute to the clearance of amyloid beta (Aβ) from the brain [78]. How to noninvasively and reliably detect endogenous level H2S remains a challenge. In the long run to develop probes for sensing H2S levels, both radioactive and non-radioactive probes have been developed [79]. Radioactive probes have advantages over fluorescent probes in terms of detection sensitivity, penetration ability, and high consistency. In the recent work published in the EJNMMI [1], Yoo et al. exquisitely developed [64Cu]Cu-ATSM-FITC (Fig. 2c) based on the bis(thiosemicarbazone) backbone. The tracer penetrated BBB and accumulated exceptionally (> 9%ID/g) high in the mice brain, due to the lipophilicity of the radiotracer with a D7.4 value of ~ 1.70. [64Cu]Cu-ATSM-FITC reacts with H2S instantly to immobilize gaseous H2S into an insoluble copper sulfide ([64Cu]CuS) precipitate. While [64Cu]Cu-ATSM-FITC has no fluorescence signal due to fluorescence quenching by Cu(II), the released ATSM-FITC ligand emits fluorescence. Elegant in vitro cell studies and in vivo PET imaging studies validated that uptake of [64Cu]Cu-ATSM-FITC was H2S-dependent. Furthermore, the authors reported that [64Cu]Cu-ATSM-FITC had higher uptake in lipopolysaccharide-induced neuroinflammation models.

The preclinical promise can potentially be translated into clinical reality to detect brain diseases [80,81,82,83]. Initial studies have shown the diffuse uptake of [62Cu]Cu-ATSM in healthy controls and neurological diseases [54, 55]. Since [64Cu]Cu-ATSM can delineate ischemic and hypoxic changes [84] as well as over-reduced intracellular states caused by mitochondrial dysfunction [44, 54], the net signal of [64Cu]Cu-ATSM or [64Cu]Cu-ATSM-FITC contributed by H2S should be carefully interpreted. Sequential use of [64Cu]Cu-ATSM (or [64Cu]Cu-ATSM-FITC) and tracers reflecting hypoxia or blood flow may better address clinical challenges. Furthermore, the role of CTR-1 in mediating the uptake of 64Cu-labeled radioligands should be carefully examined. From a diagnostic perspective, understanding mechanisms accounting for [64Cu]Cu-ATSM uptake is essential to fully interpret the imaging findings and find the application scenarios. Visualization of H2S levels in neurological diseases is attractive. As the authors mentioned, [64Cu]Cu-ATSM-FITC may help diagnose various diseases such as traumatic brain injury, stroke, encephalitis, and neurodegenerative diseases (e.g., PD and AD) [1]. Besides that, [64Cu]Cu-ATSM-FITC and other similar probes can also be used to detect cardiovascular diseases such as myocardium infarct and atherosclerosis [79, 85]. Further translation studies are needed to fill the gaps and address the clinical value.