Manganese in PET imaging: Opportunities and challenges

Abstract

Several radionuclides of the transition metal manganese are known and accessible. Three of them, ⁵¹Mn, ^52mMn, and ^52gMn, are positron emitters that are potentially interesting for positron emission tomography (PET) applications and, thus, have caught the interest of the radiochemical/radiopharmaceutical and nuclear medicine communities. This mini‐review provides an overview of the production routes and physical properties of these radionuclides. For medical imaging, the focus is on the longer‐living ^52gMn and its application for the radiolabelling of molecules and other entities exhibiting long biological half‐lives, the imaging of manganese‐dependent biological processes, and the development of bimodal PET/magnetic resonance imaging (MRI) probes in combination with paramagnetic ^natMn as a contrast agent.

1. INTRODUCTION

Radioactive nuclides have been used in nuclear medicine for the assessment of functional processes since about a century.1 However, to be suitable for in vivo imaging applications, a radionuclide needs to meet specific physical demands: Its decay radiation should be in an energy range sufficiently high to escape a patient's body in detectable amounts but also low enough to allow an efficient measurement by available detectors. Suitable γ energies for this purpose are usually in the range of 50 to 600 keV for single photon emission computed tomography (SPECT) and 511 keV for positron emission tomography (PET). Furthermore, the physical half‐life (t_1/2) of the nuclide needs to be on the one hand long enough for a work‐up procedure, processing and suitable for the time scale needed to track the biological process of interest, but on the other hand short enough to result in an acceptable radiation burden for the investigated subject. Last but not least, practical radionuclides need to be accessible.

Since these requirements are not easily fulfilled, only a relatively small number of radionuclides have made their way into clinical practice. For scintigraphy and SPECT applications, nuclides like indium‐111 (¹¹¹In, t_1/2 = 2.8 d), iodine‐123 (¹²³I, t_1/2 = 13.2 h), and technetium‐99m (^99mTc, t_1/2 = 6.0 h) are nowadays routinely used, with ^99mTc being the working horse of nuclear medicine. For PET applications, an imaging technique based on the detection of annihilation radiation of positrons (β⁺), mainly short‐living nuclides produced by proton bombardment of appropriate targets in cyclotrons, are applied. Here, fluorine‐18 (¹⁸F, t_1/2 = 109.7 min) has become the most common nuclide because of its accessibility and excellent physical properties. Other important PET nuclides in this context are carbon‐11 (¹¹C, t_1/2 = 20.4 min), nitrogen‐13 (¹³N, t_1/2 = 10.0 min), and oxygen‐15 (¹⁵O, t_1/2 = 2.0 min). Furthermore, gallium‐68 (⁶⁸Ga, t_1/2 = 67.6 min) has increasingly found applications as a PET radiometal because of the introduction of ⁶⁸Ga generators and the establishment of somatostatin‐ and prostate‐specific membrane antigen (PSMA) tracers in the clinic.2 However, all these PET nuclides have a t_1/2 of 2 to 110 minutes. Therefore, they are not suitable to track compounds in vivo that exhibit a slow pharmacokinetic (several hours to days), such as antibodies, cells, nanoparticles, and liposomes.

The number of applications of antibodies radiolabelled with a PET radionuclide (immunoPET) has significantly increased over the last decade.3 Consequently, PET nuclides with a longer t_1/2 like zirconium‐89 (⁸⁹Zr, t_1/2 = 78.4 h) or copper‐64 (⁶⁴Cu, t_1/2 = 12.7 h) have been investigated to match an antibody's biological half‐life.4 As of recently, the PET isotope manganese‐52g (^52gMn) has been proposed as a suitable candidate for the combination with antibodies and proteins. It can be readily produced with a standard 16‐MeV cyclotron and has the potential to be used in bimodal PET/magnetic resonance (MR) systems as traceable nuclide in combination with a paramagnetic contrast agent based on nonradioactive manganese at the same time.

In this review, we give an overview of medically relevant β⁺ emitting manganese isotopes and their utility for applications in PET and PET/MR imaging to address questions of medical relevance. Although publications about the preclinical use of [⁵¹Mn]MnCl₂ are available, we mainly focus on the longer‐living ^52gMn compounds and their applications.

2. RELEVANT PET RADIONUCLIDES OF MANGANESE FOR MEDICAL IMAGING

2.1. Physical properties

Table 1 provides an overview of all neutron‐deficient manganese isotopes with a t_1/2 useful for medical applications (>1 min). Manganese‐53 (⁵³Mn) and ‐54 (⁵⁴Mn) are excluded from the discussion below because of their overall poor physical decay characteristics, such as negligible β⁺ intensity and unfavourably long t_1/2. However, they have to be considered as possible side products/contaminants during the production of other manganese isotopes.

Table 1.

Positron emission tomography (PET) isotopes of manganese and their physical properties

Isotope	ß+, %	Eßmax, keV	Iγ, keVa	Half‐life
51Mn	97	2185.27	–	46.2 min
52mMn	97	2633.36	1434.07 (98.3%)	21.1 min
52gMn	29	575	744.23 (90.0%); 935.54 (94.5%); 1434.07 (100%)	5.6 d
53Mn	0/EC	–	X‐rays only	3.74 × 106 y
54Mn	5.7 × 10−7	1377	834.85 (99.9%)	312.2 d

^aMost abundant (intensity > 10%).

⁵¹Mn has a favourable β⁺ branching fraction and a t_1/2 (t_1/2 = 46 min) comparable with that of ⁶⁸Ga (t_1/2 = 68 min), which is suitable for the imaging of fast biological processes. However, the short t_1/2 of ⁵¹Mn leads to constraints regarding target separation and radiolabelling chemistry. The β⁺ energy is relatively high (E_ßmax = 2.19 MeV) in comparison with the “standard” PET nuclide ¹⁸F (E_ßmax = 0.6 MeV), which leads to an unfavourably long penetration range of the β⁺s in tissue and, thus, a deteriorated spatial resolution in the PET images.5

^52mMn has a shorter t_1/2 than ⁵¹Mn, comparable with that of the widely used ¹¹C (t_1/2 = 20 min), which renders its radiochemical handling even more challenging. Furthermore, the β⁺ energy of ^52mMn (2633 keV) is higher than that of ⁵¹Mn. This results in a mean ß⁺ range of 5.3 mm in tissue6 and, thus, in a correspondingly lower resolution of PET images. Additionally, ^52mMn decays partially via internal conversion to its ground state ^52gMn, leading to an increasing contamination with the longer‐living ^52gMn over time. A further drawback of this isotope with regard to medical applications is the presence of an additional prompt γ with relatively high energy (Table 1). Together with the β⁺ branching fraction of 93%, the arising dose rates and radiation protection concerns are unfavourable.

In contrast to the two Mn isotopes discussed above, ^52gMn has a convenient long t_1/2 (5.6 d), which is advantageous for target separations and chemical handling of the radionuclide. In addition, its t_1/2 is well suited for the investigation of slow biological processes, eg, the pharmacokinetics of antibodies. ^52gMn decays with a branching fraction for β⁺ of 29%, which is significantly lower than that of the previously mentioned Mn isotopes (Table 1). The β⁺ are emitted with a low maximum energy of E_ßmax = 0.6 MeV, which is among the lowest energies of all β⁺ emitting nuclides. This results in a comparatively low tissue penetration range of the β⁺s and, thus, better resolution of PET images. The main disadvantage of ^52gMn is the occurrence of three prompt γ rays with high energy and intensity (Table 1). These γ rays would significantly contribute to the radiation burden of patients and personnel of nuclear medicine departments. Furthermore, the prompt gammas cause erroneous signals in the PET detectors, which necessitates the implementation of prompt gamma correction techniques when using ^52gMn for PET imaging.7

2.2. Nuclide production

In theory, a large number of nuclear reactions can lead to the formation of each of the different manganese isotopes discussed in this review, including possible side products such as ⁵³Mn and ⁵⁴Mn (see IAEAs EXFOR database).8 However, many of the reported production routes have no practical relevance, and a comprehensive discussion of all possibilities is beyond the scope of this review. Instead, we focus on the most efficient ones, which includes the irradiation of solid chromium targets.

The production of ⁵¹Mn was discussed by Klein et al including a survey of potential nuclear reactions as well as a short summary of previous results by other groups.9 They concluded that the ⁵⁰Cr(d,n)⁵¹Mn reaction previously investigated by Cogneau et al10 and the ^natCr(p,x)⁵¹Mn reactions are the most promising candidates for this task. The ⁵⁰Cr(d,n)⁵¹Mn reaction is based on the irradiation of isotopically enriched ⁵⁰Cr with 14 → 3 MeV deuterons, resulting in the desired isotope in good yields (Table 2). The alternative reaction ^natCr(p,xn)⁵¹Mn utilizes the high abundance of ⁵²Cr (83.8%) in natural chromium and uses the ⁵²Cr(p,2n)⁵¹Mn reaction. However, even if conducted with an optimized proton energy window, the coformation of the long‐living ⁵⁴Mn and ^52mMn as well as ^52gMn by competing (p,n) reactions cannot be avoided. It should be mentioned that the formation of ⁵⁴Mn can be circumvented by using highly enriched ⁵²Cr as target material. The potential formation of long‐living ⁵³Mn when using enriched ⁵²Cr targets has not been investigated so far.

Table 2.

Positron emission tomography (PET) isotopes of manganese and their production routes

Isotope	Reaction Channel	Energy Threshold, MeVa	Thick Target Yield, MBq/μAh
51Mn	52Cr(p,2n)51Mna	16.3	Not determined
	53Cr(p,3n)51Mna	24.4	Not determined
	50Cr(d,n)51Mnb	0	110b
52mMn	52Cr(p,n)52mMn	5.9	6910 ± 760
	53Cr(p,2n)52mMn	13.8
52gMn	52Cr(p,n)52gMn	5.5	13.7 ± 1.6
	53Cr(p,2n)52gMn	13.4

^aEnergy range 16.9 → 8.2 MeV protons for the nuclear reaction ^natCr(p,xn).

^bEnergy range 14 → 3 MeV deuterons for the nuclear reaction ⁵⁰Cr(d,n) with enriched target material.9, 11

One of the most promising routes for the production of ^52gMn and ^52mMn is represented by the irradiation of suitable chromium targets with 16 → 8 MeV protons (Table 2). Several studies demonstrated the feasibility of ⁵²Mn production by the ^natCr(p,xn)⁵²Mn reaction using 16 MeV cyclotrons.12, 13, 14 Another study presents cross‐sectional data for target beams up to 20 MeV.15 Because of similar energy thresholds, ^52mMn is always coproduced with ^52gMn when chromium targets are irradiated with protons. An isomerically pure production of these nuclides is therefore not possible by this approach. However, for the production of ^52gMn for imaging applications, this is not an issue. Because of the significant difference in t_1/2 of ^52gMn (21.1 min) and ^52mMn (5.6 d), the contamination can be removed by just simply letting the ^52mMn decay. The same applies also to potentially coproduced ⁵¹Mn. The only reported relevant impurity of ^52gMn mentioned in the literature is ⁵⁴Mn, which is produced in minor amounts from ⁵⁴Cr by a (p,n) reaction. Information about the potential long‐living impurity ⁵³Mn is scarcely discussed albeit considerable cross sections for the ⁵³Cr(p,n)⁵³Mn reaction have been published.16 However, the formation of both long‐living impurities (⁵³Mn and ⁵⁴Mn) can be avoided by irradiation of highly enriched ⁵²Cr and should therefore not impose any restriction for potential clinical applications of ^52gMn.

Alternatively to the production of Mn isotopes via proton irradiation, ^52mMn is available in high isotopic purity via a ⁵²Fe/^52mMn generator (t_1/2(⁵²Fe) = 8.28 h).17 However, because of the poor accessibility of the mother nuclide ⁵²Fe, this approach has not been fully explored yet.

2.3. Target separation

To separate ^52gMn isotopes from the chromium target material, several chromatographic ion exchange methods have been published.18, 19, 20 In most of them, the chromium target disc is first dissolved in an acidic medium and flushed over an ion exchange column. The radiomanganese is retained while chromium is washed out of the column. In a next step, the radiomanganese is eluted from the column with a different solvent composition. In the case of insufficient removal of chromium, the procedure has to be repeated. Different ion exchange resins, solvents, and solvent mixtures have been evaluated. For example, a recently published method required multiple subsequent column purifications resulting in a ^52gMn recovery below 70%.21 Thus, there is still room for future improvements. A further development is the combination of chemical and chromatographic separation techniques, resulting in higher purity of the desired ^52gMn.22 The different purification conditions are summarized in a review by Chaple et al,18 although none of them can be considered yet as “perfect” in terms of simplicity and recovery of the radionuclide.

It should be noted that, unlike in case of ^52gMn, these elaborate and lengthy separation techniques are not adequate to isolate the short‐living manganese PET isotopes ^52mMn and ⁵¹Mn, if produced via a solid chromium target. The development of a suitable liquid target enabling fast separation by solid phase extraction technology might provide a solution (a similar discussion is ongoing regarding the cyclotron production of ⁶⁸Ga).23

In summary, the transition metal manganese offers three isotopes interesting for potential PET applications. Two of those, namely, ⁵¹Mn and ^52mMn, are short living and thus, have the potential for the imaging of fast biological processes. In comparison, ⁵¹Mn is the better candidate because of its favourable decay characteristics (longer t_1/2, practically no additional γ‐rays), which are comparable with those of ⁶⁸Ga. ^52gMn, on the other hand, has a suitable half‐life for the PET imaging of slow biological processes including applications in immunoPET. However, the presence of several high‐energy γ rays with high intensity necessitates the use of suitable corrections in PET imaging and is demanding with regard to radiation protection.

3. MANGANESE‐52G IN PRECLINICAL RESEARCH

3.1. Manganese complexes—coordination behaviour and paramagnetism

Manganese is a first‐row transition metal and homologue of technetium and rhenium. A broad range of coordination compounds of manganese are known, most of them with the metal in the oxidation state +II and coordination numbers of 6 or 7.24 As a “hard” Lewis acid, the most stable manganese complexes are obtained with ligands coordinating via oxygen and nitrogen atoms. For example, manganese (II) forms stable complexes with DOTA 1 and DO3A (logK_ML = 19.89 and 19.30, respectively).25, 26, 27 Such complexes can be obtained at room temperature under mild reaction conditions and within short reaction times.25 Other studied chelators for the complexation of Mn (II) are EDTA 3 and its derivatives such as CDTA 4 (Figure 1).28

Figure 1.

Examples of important chelating agents for manganese (II)

Mn (II) in its octahedral high spin complexes owns five unpaired electrons, which results in a high paramagnetic moment. For this reason, manganese (II) complexes have been investigated as possible magnetic resonance imaging (MRI) contrast agents. For example, the dipyridoxyl diphosphate (DPDP) complex of manganese (II) (Teslascan, Figure 2) was used in the clinic for the diagnosis of liver lesions. However, the compound has meanwhile been withdrawn from the market because of its insufficient stability in vivo.29

Figure 2.

Mn‐DPDP (Teslascan), a manganese‐based T₁ contrast agent

3.2. Biological pathways of MnCl ₂ in vivo and neuronal connectivity imaging

Manganese is an essential element for all beings, for example, participating in numerous enzymatic processes as a cofactor.30 In the oxidation states +II and +III, it is transported in the blood bound to serum proteins31 before it accumulates in organs and tissue or is excreted.32 Each human body contains roughly 12 mg of manganese mostly stored in the bones, liver, and kidney.33 Notably, it is also known to cross the blood‐brain‐barrier.34 In higher doses, free manganese ions are known to cause a neurological disorder condition called manganism with psychiatric and, in later stages, Parkinson‐like symptoms.35

With regard to potential medical applications of manganese compounds, the pharmacokinetic profile of intravenously injected [^52gMn]MnCl₂ was investigated in mice.36 The highest accumulation of the radiometal was found in the liver, kidney, salivary glands, thyroid, and pancreas, whereas only low uptake of radioactivity was observed in the bones. These data may help to predict the fate of Mn (II) ions in case they get released from radiopharmaceuticals in vivo. Interestingly, the biodistribution pattern changed when [^52gMn]MnCl₂ was given by inhalation as a saline aerosol (eg, decreased uptake in the bones).36

To study the uptake and retention of manganese with regard to possible neuroimaging techniques, manganese‐enhanced magnetic imaging (MEMRI) as well as PET imaging was utilized by Brunnquell et al.37 For this purpose, different doses of nonradioactive MnCl₂ were administered intravenously to rats. The uptake in different parts of the brain was studied using quantitative MEMRI at different time points postinjection (24 h to 14 d). Brain uptake studies with nca and ca [^52gMn]MnCl₂ were performed via gamma counting of resected brain areas. However, the authors concluded that the brain uptake of [^52gMn]Mn²⁺ with intact blood‐brain barrier was too low for neuroimaging applications, especially when using the carrier added radiotracer. In addition, the authors stated that [^52gMn]MnCl₂, although not suitable in this case, still holds promise as a radiotracer for specific uptake in the salivary glands and pancreas. Another publication evaluated ^52gMn for neuroimaging by stereotactic injection into the rat brain.38 The neuronal pathways between rat brain regions could be imaged successfully. Through application of different doses of nca [^52gMn]MnCl₂ (30 kBq‐170 MBq), the radiotoxicity was also studied, revealing that a low dose of 20 kBq is sufficient for imaging, while no histological and behavioural noxious effects occurred.

Saar et al studied the biodistribution of ⁵¹Mn and ^52gMn in different organs.39 Further, they investigated the neuronal olfactory pathway in monkeys and rodents using nasally administrated ⁵¹Mn and ^52gMn. It was shown that an administration of [^52gMn]MnCl₂ allows for a tracing of neuronal connections and that ^52gMn is able to enter excitable cells in a similar manner as nonradioactive manganese ions in MEMRI.

3.3. Manganese‐52g for immunoPET, cell labelling, and ß‐cell mass monitoring

The use of radiolabelled antibodies for immunoPET has become an important tool in nuclear oncology.3, 40 Radiolabelled antibodies offer the advantage of high affinity and specificity. On the other hand, antibodies exhibit slow pharmacokinetics and the time until satisfying tumour uptake and/or tumour‐to‐background ratios are reached is often too long for other short‐living PET radionuclides such as ⁶⁸Ga, ¹⁸F, or even ⁶⁴Cu. Therefore, the longer‐living ⁸⁹Zr (t_1/2 = 78.4 h) is currently predominantly used for the labelling of clinically relevant antibodies.40 However, ^52gMn also offers a suitable t_1/2 for immunoPET applications while displaying a lower β⁺ energy and a higher β⁺ intensity than ⁸⁹Zr (Table 3). Thus, ^52gMn may represent a promising, alternative radiometal for applications in immunoPET.41

Table 3.

Nuclides used for radiolabelling of antibodies and their physical properties

Nuclide	ß+, %	Eßmax, keV	Iγ, keVa	Half‐life, h
64Cu	17.9	653	–	12.7
89Zr	22.8	902	909 (99.0%)	78.4
52gMn	29	575	744.23 (90.0%); 935.54 (94.5%); 1434.07 (100%)	134.4

^aMost abundant (intensity > 10%).

The first and so far only in vivo study with a ^52gMn‐labelled antibody was published in 2015 by Graves et al.21 For this study, the chelator DOTA was conjugated to the AT1‐targeting antibody TRC105 and tested in radiolabelled form for the PET imaging of a breast cancer xenograft mouse model. The chelator DOTA allowed for the ^52gMn‐labelling at room temperature (see above), which is a necessary requirement to avoid denaturation of the protein during radiolabelling. Despite a slower blood clearance, the in vivo biodistribution and PET imaging yielded comparable results to a similar ⁸⁹Zr‐labelled antibody conjugate (Figure 3). The research group used ^52gMn instead of the more common ⁸⁹Zr to demonstrate the opportunity of PET imaging at late time points postinjection of the radiotracer and the possibility for triple coincidence PET measurements,42 a new imaging technology for which ^52gMn has suitable physical properties.

Figure 3.

Serial maximum intensity projection (MIP) positron emission tomography (PET) images of mice injected with ⁵²Mn‐DOTA‐TRC105 and ⁵²MnCl₂. Significant thyroid accumulation in the ⁵²MnCl₂ images contrasting the lack of uptake in the ⁵²Mn‐DOTA‐TRC105 images indicates highly stable DOTA chelation of ⁵²Mn²⁺ even at late time points. H, heart; K, kidneys; L, liver; T, tumour; Th, thyroid. Reprinted with permission from Graves et al21

Another potential use for longer‐living radiometals in nuclear medicine is the labelling and tracking of cells and liposomes in vivo. In 2018, Gawne et al demonstrated the suitability of [^52gMn]Mn (oxinate)₂ to label different cell types as well as DOXIL/CAEXYL liposomes.43 It was shown that the efficiency of the radiolabelling was comparable with analogous ⁸⁹Zr‐labelling, although the method was limited by cell efflux of [^52gMn]Mn²⁺. In vivo studies of ^52gMn‐labelled liposomes in mice showed sufficient stability of the conjugate for up to 24 hours as long as the compound remained in the bloodstream. After the liposomes entered cells and tissues, uptake of radioactivity in the kidneys, salivary glands, and pancreas was detected, indicating decomposition and release of free [^52gMn]Mn²⁺ ions. The authors concluded that their method was not suitable for in vivo tracking of cells but could serve as a model to study the biological fate of ^52gMn once delivered inside of cells in vivo.

A different approach towards radiolabelled liposomes using ^52gMn was published by Jensen et al.44 In this comparative study, ^52gMn(II) and ⁶⁴Cu(II) were prepared as their DOTA chelates and used for both internal loading and surface labelling of liposomes. In vivo biodistribution studies with the ^52gMn‐labelled liposome preparations revealed that liposomes with internal ^52gMn‐loading had a longer plasma half‐life than their surface labelled counterparts. The authors concluded that the reduced blood plasma half‐live of the surface modified liposomes could result from insufficient in vivo stability of the radiometal‐DOTA chelates.44

Mn⁺² was shown to be taken up significantly by the pancreas31 (see also above), possibly by mimicking Ca²⁺ in pancreatic metabolic pathways. This feature was used by Hernandez et al to monitor ß‐cell mass with [^52gMn]MnCl₂ by ex vivo and in vivo imaging of ß‐cell metabolism in type 1 and type 2 diabetes mouse models.45 Previous work on this topic using MRI and nonradioactive manganese ions as contrast agents showed that the utility of this method was limited by the toxicity of free Mn²⁺ ions. Using the radiotracer [^52gMn]MnCl₂ and PET imaging solved the toxicity issues. It was shown that the uptake of [^52gMn]Mn²⁺ strongly depends on the activity of ß‐cell voltage‐dependent Ca²⁺ channels and that the uptake of the radiometal correlates with Ca²⁺ uptake. The authors concluded further that because of the rapid uptake mechanisms, the use of the shorter‐living ⁵¹Mn might be a good alternative. This was later confirmed by a subsequent study using [⁵¹Mn]MnCl₂.46

3.4. Relaxivity and stability—manganese for PET/MRI

In the past years, complexes of paramagnetic Mn (II) have been discussed as potential alternatives to the well‐established, but in some cases, disputed gadolinium‐based contrast agents (eg, Gadopentetat‐Dimeglumine, Magnevist).47, 48 For details on Mn‐based MRI contrast agents, the reader is referred to an excellent review on the topic.49 In addition, the availability of PET nuclides of manganese offers the possibility of isotopically radiolabelled manganese MR contrast agents for use in bimodal PET/MR imaging. Hybrid imaging by PET/MRI has lately received considerable attention in the nuclear medicine and radiology because it combines the high sensitivity PET with the high resolution of MRI.50 The concept is attractive, however, hampered by the different sensitivities of the modalities: PET allows for the application of very low concentrations of the radioactive substance (10⁻⁹‐10⁻¹²M) for achieving excellent contrast. In comparison, paramagnetic contrast agents for MRI are applied in millimolar concentrations (one dose of Gd‐DTPA, Gadovist: 0.1 mmol/kg body weight).51 The contradicting requirements of the two modalities for contrast agents can be met by using mixtures of a PET tracers with the respective nonradioactive analogous MRI contrast agent (carrier added radiotracers). Because the PET tracer (^52gMn) and the MR contrast agent (^natMn) are structurally identical, they exhibit equal biological properties.52 Employment of such mixtures enables, eg, the quantification of the MR contrast agent in areas of low uptake by PET. Furthermore, the concept is particularly attractive for applications to “smart MR contrast agents”.53, 54

The ability of a coordination compound to enhance longitudinal MR contrast is described by the value “relaxivity” r₁. The relaxivity of paramagnetic coordination compounds is influenced not only by the number of unpaired electrons of the central metal ion but also by its internal rotation, size, and most importantly, magnetic influence on directly bound and surrounding water molecules.55

Since a water molecule as an additional ligand inside the coordination sphere of the metal is required for contrast enhancement, Mn (II) complexes with hexa‐ or lower dentate ligand systems and at least one inner sphere water molecule are potential candidates for MRI contrast agents.60 The tendency of the chelator DOTA 1 to form octadentate complexes with manganese of excellent stability (see above) has its downside in this context: Because of the lack of an inner sphere water molecule in [^natMn (DOTA)], there is no MR contrast (Table 4). Therefore, complexes of Mn (II) with DOTA 1 (and DO3A)28 have good properties for PET imaging but are not suitable for PET/MR applications.

Table 4.

Contrast agents based on Gd and Mn and their physicochemical properties

Contrast Agent	Longitudinal Relaxivity r1, 37°C, 20 MHz [mmol−1s−1]	Stability Constant logKML a	Ref
Gd‐DOTA (Dotarem)	3.83	24.7	56, 57, 58
Gd‐DTPA (Magnevist)	4.02	22.5	25
Mn‐DPDP (Teslascan)	2.80	11.6	59
Mn‐CDTA	2.75	14.3	28
Mn‐PyC3A	2.10	14.1	60
Mn‐DOTA	–	19.9	60
Mn‐DO3A	1.30	19.4	27

^aK_ML = Equilibrium stability constant: [ML]/[M]^*[L] with [ML] complex concentration; [M] metal ion concentration; [L] ligand concentration, in equilibrium.

While the combination of Mn (II) with DOTA 1 provides stable complexes but no contrast enhancement, the use of other chelators (eg, EDTA 3) results in complexes with appropriate relaxivity but insufficient stability. Therefore, the quest for new chelators that fulfil both requirements is still ongoing. The best compromise between sufficient stability and contrast enhancement described so far was achieved with chelators on the basis of the 1,2‐trans‐cyclohexyldiaminocarboxylate (CDTA 4) scaffold, for example, PyC3A 5 (Figure 4 with structure 4‐6).56 Gale et al reported in 2015 that the Mn (II) complex of PyC3A proved to be of sufficient thermodynamic stability (logK_MnL = 14.14) and suitable for contrast‐enhanced MR angiography.60 This successful proof‐of‐concept study represents an important step towards the development of new alternatives of Gd‐based MR contrast agents.

Figure 4.

Successful chelating agents for Mn based on CDTA 4 for magnetic resonance imaging (MRI) using manganese

In 2016, Vanasschen et al published the first isotopically labelled bimodal PET/MRI agent based on mixtures of radioactive and nonradioactive manganese.61 The stability of the investigated [^52g/55Mn]Mn‐CDTA complex in human blood serum (HBS) was studied, as well as its radiolabelling reaction using nca and ca [^52g/55Mn]Mn⁺². The radiolabelling could be performed successfully (radiochemical yield > 99%) at room temperature within 30 minutes. The stability study in HBS showed a dissociation half‐life of [^52g/55Mn]Mn‐CDTA of 12 hours with most of the released [^52gMn]Mn²⁺ being bound to larger serum proteins.61

Inspired by earlier work of Zhu et al,62 a rigid dendrimeric scaffold with three isotopically ^52g/55Mn‐labelled complexes has been recently reported.63 The investigated contrast agent ^natMn‐Tris‐CDTA‐1,3,5‐tris‐triazolobenzene [^natMn][Mn₃(TTB‐(CDTA)₃] 6a exhibited a dramatically increased overall T₁ relaxivity in comparison with monomeric Mn‐CDTA, obviously the result of the presence of multiple paramagnetic centres. Also, the relaxivity of each paramagnetic centre could be increased by 144% because of the rigidity and restricted internal rotation of the molecule in comparison with Mn‐CDTA.28, 63 Labelled with isotopic mixtures of ^natMn and ^52gMn, a bimodal PET/MR contrast agent based on manganese with high relaxivity was obtained. Through functionalization of the chelator, a bifunctionalized chelating agent ((EtVal‐TTB‐(CDTA)₃ 6b was synthesized to be used for further derivatization and potentially for bioconjugations.

It should be noted that other approaches for the development of PET/MRI imaging agents have also been reported. For example, Notni et al combined [^68/69Ga]Ga and ^natGd in a scaffold containing the chelators TRAP and DOTA52; Frullano et al studied ^natGd‐DOTA complexes with a pendant ¹⁸F atom as PET reporter.53 In addition, iron particles spiked with radionuclides have been studied by different groups as potential PET/MRI imaging agents.64

4. SUMMARY AND CONCLUSION

There are three isotopes of manganese that are interesting for PET applications. The most promising among them is ^52gMn due to its low β⁺ energy and suitable t_1/2. ⁵²Mn can be readily produced using a small 16‐MeV cyclotron and separated from the target material by effective but time‐consuming ion exchange chromatography. For clinical applications, ^52gMn with a half‐life of 5.5 days is an interesting candidate for the radiolabelling of molecules with slow pharmacokinetics, for example, antibodies for immunoPET. In this context, ^52gMn outperforms the current standard for immunoPET ⁸⁹Zr not only in terms of t_1/2 but also in terms of β⁺ energy, the latter resulting in better resolution of PET images for ^52gMn. On the other hand, a drawback of ^52gMn in comparison with ⁸⁹Zr is the occurrence of high‐energy γs that will increase the dose rates for the patients, as well as the lack of commercial sources, which currently restricts its use in radiopharmaceutical research.

Because ^natMn (II) is paramagnetic and PET isotopes of the metal are available, isotopically radiolabelled manganese‐based PET/MRI contrast agents are within reach. The main challenge is the identification of suitable chelators, which provide manganese complexes of sufficient stability and relaxivity. Different approaches towards the development of such manganese complexes have been reported, but issues regarding the in vivo stability remain to be addressed.

In summary, manganese is a versatile transition metal with available isotopes well suited for MR, PET, and PET/MR imaging. Because of the currently limited availability, only a relatively small number of publications describe the production and use of manganese PET radionuclides. However, this may change in the future as in particular ^52gMn has high potential to become a new emerging radiometal for PET and PET/MR applications in nuclear medicine.

Biographies

Marie Brandt studied Chemistry at the Universities of Bonn and Cologne, Germany. She received her PhD in Radiochemistry from the Institute of Neuroscience and Medicine INM‐5: Nuclear Chemistry at Forschungszentrum Jülich, Germany in 2017. In fall 2017, she joined the LBI Applied Diagnostics / Department of Biomedical Imaging and Image Guided Therapy, Division of Nuclear Medicine at the Medical University of Vienna, Austria as postdoctoral fellow in Prof. Mindt's program line.

Jens Cardinale studied Chemistry at the University of Cologne, Germany. He received his PhD in Radiochemistry from the Institute of Neuroscience and Medicine INM‐5: Nuclear Chemistry at Forschungszentrum Jülich, Germany in 2012. From 2013 – 2017, he joined the research group of Prof. Kopka at German Cancer Research Center in Heidelberg, Germany. In spring 2017, he became Senior Scientist in Prof. Mindt's group at the LBI Applied Diagnostics / Departmentof Biomedical Imaging and Image Guided Therapy, Division of Nuclear Medicine at the Medical University of Vienna, Austria.

Ivo Rausch has studied technical Physics at the UT Vienna and holds a PhD in Medical Physics. Today, he is a post doc researcher in the Quantitative Imaging and Medical Physics group at the Center of Medical Physics and Biomedical Engineering, Medical University of Vienna. His main research interests relate to quantitative hybrid imaging in nuclear medicine with a focus on PET/CT and PET/MRI imaging.

Thomas L. Mindt is the Head of the Program Line “Imaging Biomarkers“ at the Ludwig Boltzmann Institute Applied Diagnostics. He obtained a chemical engineering degree from the University of Applied Sciences (Switzerland) and Ph.D. in organic chemistry from Brown University (USA). Upon return to Europe,he worked as a senior scientist in Radiopharmaceutical Sciences at the ETH in Zurich (Switzerland). In 2009, he accepted a call of the University of Basel (Switzerland) as an Assistant Professor in Radiopharmaceutical Chemistry and in 2015, he was promoted to Honorary Professor in the same discipline. In 2016, he moved to Vienna as a co‐founder of the new Ludwig Boltzmann Institute. Since 2018 he is a private lecturer (P.D.) at the Faculty of Chemistry of the University of Vienna.

Brandt M, Cardinale J, Rausch I, Mindt TL. Manganese in PET imaging: Opportunities and challenges. J Label Compd Radiopharm. 2019;62:541‐551. 10.1002/jlcr.3754