CXCR4-Targeted Radiopharmaceuticals for the Imaging and Therapy of Malignant Tumors

C-X-C chemokine receptor type 4 (CXCR4), also known as fusin or CD184, is a 7-transmembrane helix G-protein-coupled receptor that is encoded by the CXCR4 gene. Involved in various physiological processes, CXCR4 could form an interaction with its endogenous partner, chemokine ligand 12 (CXCL12), which is also named SDF-1. In the past several decades, the CXCR4/CXCL12 couple has attracted a large amount of research interest due to its critical functions in the occurrence and development of refractory diseases, such as HIV infection, inflammatory diseases, and metastatic cancer, including breast cancer, gastric cancer, and non-small cell lung cancer. Furthermore, overexpression of CXCR4 in tumor tissues was shown to have a high correlation with tumor aggressiveness and elevated risks of metastasis and recurrence. The pivotal roles of CXCR4 have encouraged an effort around the world to investigate CXCR4-targeted imaging and therapeutics. In this review, we would like to summarize the implementation of CXCR4-targeted radiopharmaceuticals in the field of various kinds of carcinomas. The nomenclature, structure, properties, and functions of chemokines and chemokine receptors are briefly introduced. Radiopharmaceuticals that could target CXCR4 will be described in detail according to their structure, such as pentapeptide-based structures, heptapeptide-based structures, nonapeptide-based structures, etc. To make this review a comprehensive and informative article, we would also like to provide the predictive prospects for the CXCR4-targeted species in future clinical development.


Chemokines
The chemokines (chemotactic cytokines) are a series of small, mostly secreted proteins that consist of about 60 to 90 amino acids (8-10 kDa in mass) with an N terminal and a C terminal. At the N terminals of these proteins, there are two or four cysteine residues. In the biological environment, the primary function of chemokines is to induce cell migration [1].

Chemokine Receptor Chemokine
Involved in various physiological processes, CXCR4 could form an interaction with its endogenous partner, chemokine ligand 12 (CXCL12), which is also named SDF-1 [10]. In the past several decades, the CXCR4/CXCL12 couple has attracted a large amount of research interest due to its critical functions in the occurrence and development of refractory diseases, such as HIV infection, inflammatory diseases, and metastatic cancer, including breast cancer, gastric cancer, and non-small cell lung cancer [11]. The CXCR4-CXCL12 axis could promote angiogenesis and recruit myeloid bone marrow-derived cells to facilitate tumor recurrence and metastasis, thus mediating resistance to conventional as well as targeted therapies. Neutralization of CXCR4/CXCL12 chemotaxis using anti-CXCR4 antibodies, peptide antagonists, or small molecule antagonists could significantly reduce the metastasis.
The pivotal role of CXCR4 has encouraged researchers around the world to investigate CXCR4-targeted imaging and therapeutics [12]. Figure 1 illustrates the schematic diagram of CXCR4, which showed up as a 7-transmembrane helix protein. The CXCR4 crystal structure is present in the published literature [13].

Radiopharmaceuticals Based on CXCR4
The structures of CXCR4-based small molecule antagonists have been influenced by the structures of the corresponding peptide drugs to some extent. Specifically, a highly potent β-sheet-like 14-mer peptide (T140) was originally designed based on the structures of CXCR4 and CXCL12 [14]. A structure-activity relationship study of this 14-peptide showed that four of the amino acid residues (Arg2, Nal3, Tyr5, and Arg14) are indispensable for the activity of the 14-peptide.
Thus, after a series of structural optimizations, researchers obtained a pentapeptide compound, FC131, which is also a peptide compound with a high affinity for CXCR4 [15]. In this pentapeptide compound, any of the amino acid residues are very important, except for Arg2, which is less important. Taking this feature into consideration, CPCR4-2 has been further developed on the basis of FC131.
As for small molecules, a series of small molecule structures have been optimized according to the structure-activity relationship, such as pentapeptide-based antagonists, indole-based antagonists, tetrahydroquinoline-based antagonists, para-xylyl-enediamine-based antagonists, guanidine-based antagonists, quinoline-based antagonists, pyrimidine-based antagonists [13], benzenesulfonamide-based antagonists [16], etc. However, only a few of these small molecules were translated into the radiopharmaceutical field.
In 2009, Kiesewetter et al. [17] imply 64 Cu-AMD3100 ( Figure 2) for the first time in PET imaging of CXCR4-expressing tumors and proved that this radiotracer is useful in CXCR4-targeted imaging and therapies. They modified the synthetic route of AMD3100 in high yield and achieved 64 Cu-AMD3100 in high radiochemical yield and with high radiochemical purity. In the binding affinity assay, 64 Cu-AMD3100 exhibited an IC50 value of 62.7 µM towards Jurkat T-cells (CXCR4-positive). This value is much higher than

Radiopharmaceuticals Based on CXCR4
The structures of CXCR4-based small molecule antagonists have been influenced by the structures of the corresponding peptide drugs to some extent. Specifically, a highly potent β-sheet-like 14-mer peptide (T140) was originally designed based on the structures of CXCR4 and CXCL12 [14]. A structure-activity relationship study of this 14-peptide showed that four of the amino acid residues (Arg2, Nal3, Tyr5, and Arg14) are indispensable for the activity of the 14-peptide.
Thus, after a series of structural optimizations, researchers obtained a pentapeptide compound, FC131, which is also a peptide compound with a high affinity for CXCR4 [15]. In this pentapeptide compound, any of the amino acid residues are very important, except for Arg2, which is less important. Taking this feature into consideration, CPCR4-2 has been further developed on the basis of FC131.
As for small molecules, a series of small molecule structures have been optimized according to the structure-activity relationship, such as pentapeptide-based antagonists, indole-based antagonists, tetrahydroquinoline-based antagonists, para-xylyl-enediaminebased antagonists, guanidine-based antagonists, quinoline-based antagonists, pyrimidinebased antagonists [13], benzenesulfonamide-based antagonists [16], etc. However, only a few of these small molecules were translated into the radiopharmaceutical field.
In 2009, Kiesewetter et al. [17] imply 64 Cu-AMD3100 ( Figure 2) for the first time in PET imaging of CXCR4-expressing tumors and proved that this radiotracer is useful in CXCR4-targeted imaging and therapies. They modified the synthetic route of AMD3100 in high yield and achieved 64 Cu-AMD3100 in high radiochemical yield and with high radiochemical purity. In the binding affinity assay, 64 Cu-AMD3100 exhibited an IC 50 value of 62.7 µM towards Jurkat T-cells (CXCR4-positive). This value is much higher than the IC 50 of AMD3100, verifying that the incorporation of the Cu(II) ion enhances the binding of AMD3100 to CXCR4. In the biodistribution studies, 64 Cu-AMD3100 was observed accumulating in immune-related organs, such as the spleen (13%, 1 h post-injection), lymph nodes (10%, 1 h post-injection), and bone marrow (14%, 1 h post-injection). Other organs, such as the liver and kidney, also shared a large amount of the radiotracers [17]. the IC50 of AMD3100, verifying that the incorporation of the Cu(II) ion enhances the binding of AMD3100 to CXCR4. In the biodistribution studies, 64 Cu-AMD3100 was observed accumulating in immune-related organs, such as the spleen (13%, 1 h post-injection), lymph nodes (10%, 1 h post-injection), and bone marrow (14%, 1 h post-injection). Other organs, such as the liver and kidney, also shared a large amount of the radiotracers [17]. In 2010, Nimmagadda et al. evaluated the kinetics and biodistribution of [ 64 Cu]AMD3100 in subcutaneous brain tumor xenografts [18]. In a cell binding assay, they found [ 64 Cu]AMD3100 could bind specifically to glioblastoma cell lines (U87-stb-CXCR4) and breast cancer cell lines (DU4475) with high CXCR4 expression levels. In subcutaneous tumor xenografts, this specificity was also validated [18]. In a later work in 2011, Farber et al. evaluated the extensive dosimetry in mice and established the feasibility of this radiotracer in the human body [19].
In order to bind firmly with the CXCR4 receptor, AMD3100 adopts several binding modes with the cooperation of several residues of CXCR4. One of the possible geometries is contributed by ASP171, ASP262, and GLU288, with one cyclam of AMD3100 binding to ASP171 and another cyclam binding to ASP262 and GLU288. Another possible geometry is in the form of residues ASP262 and GLU288 binding to the bicyclam rings and residues PHE189 and TYR190 binding to the methylene linker. For more binding mode information about AMD3100 and AMD3465, which is beyond the scope of this review, readers could refer to the published elegant papers [20][21][22][23]. After transition metal complex formation, the binding modes switch from electrostatic bonds between the protonated cyclam primary amine groups and aspartate residue carboxylic acid groups (ASP171 and ASP262) to coordinate bonds [24].
Except for 64 Cu, other radioisotopes, such as 99m Tc, 67 Ga, and 62 Zn [25][26][27] were also used in the development of AMD3100-based radiopharmaceuticals, though they might possess certain shortcomings. For example, 99m Tc-based radiotracers are used in SPECT imaging rather than PET imaging, and [ 99m Tc]AMD3100 showed substantially reduced binding affinity toward the receptor compared with [ 64 Cu]AMD3100, which might be due to the deviations in the planar structure of the cyclam caused by the introduction of a relatively large metal ion [25]; the radiolabeling time of [ 67 Ga]AMD3100 is about 2 h, which is not suitable for the development of 68 Ga-based radiotracers; besides, 62 Zn is a seldom used isotope in labeling or imaging studies.
In 2009, Archibald et al. developed CB-Bicyclam, a cross-bridged analog of AMD3100 with a specific structure, to reinforce the interaction between the bicyclam and the aspartate residues on one surface of CXCR4 [28]. Crystal structure and computational studies confirmed the shortened and stronger interactions between the complexes and the carboxylates compared to unconstrained macrocycle complexes in AMD3100 [29]. In 2020, the same group labeled this small molecule with the radionuclide 64 Cu to form the mono-copper species 64 Cu-CuCB-bicyclam. Compared with AMD3100, CuCB-bicyclam reduced the up to six configurations to only one configuration when it formed the copper(II) complex. This rendered 64 Cu-CuCB-bicyclam with higher affinity and specific In 2010, Nimmagadda et al. evaluated the kinetics and biodistribution of [ 64 Cu]AMD3100 in subcutaneous brain tumor xenografts [18]. In a cell binding assay, they found [ 64 Cu]AMD3100 could bind specifically to glioblastoma cell lines (U87-stb-CXCR4) and breast cancer cell lines (DU4475) with high CXCR4 expression levels. In subcutaneous tumor xenografts, this specificity was also validated [18]. In a later work in 2011, Farber et al. evaluated the extensive dosimetry in mice and established the feasibility of this radiotracer in the human body [19].
In order to bind firmly with the CXCR4 receptor, AMD3100 adopts several binding modes with the cooperation of several residues of CXCR4. One of the possible geometries is contributed by ASP171, ASP262, and GLU288, with one cyclam of AMD3100 binding to ASP171 and another cyclam binding to ASP262 and GLU288. Another possible geometry is in the form of residues ASP262 and GLU288 binding to the bicyclam rings and residues PHE189 and TYR190 binding to the methylene linker. For more binding mode information about AMD3100 and AMD3465, which is beyond the scope of this review, readers could refer to the published elegant papers [20][21][22][23]. After transition metal complex formation, the binding modes switch from electrostatic bonds between the protonated cyclam primary amine groups and aspartate residue carboxylic acid groups (ASP171 and ASP262) to coordinate bonds [24].
Except for 64 Cu, other radioisotopes, such as 99m Tc, 67 Ga, and 62 Zn [25][26][27] were also used in the development of AMD3100-based radiopharmaceuticals, though they might possess certain shortcomings. For example, 99m Tc-based radiotracers are used in SPECT imaging rather than PET imaging, and [ 99m Tc]AMD3100 showed substantially reduced binding affinity toward the receptor compared with [ 64 Cu]AMD3100, which might be due to the deviations in the planar structure of the cyclam caused by the introduction of a relatively large metal ion [25]; the radiolabeling time of [ 67 Ga]AMD3100 is about 2 h, which is not suitable for the development of 68 Ga-based radiotracers; besides, 62 Zn is a seldom used isotope in labeling or imaging studies.
In 2009, Archibald et al. developed CB-Bicyclam, a cross-bridged analog of AMD3100 with a specific structure, to reinforce the interaction between the bicyclam and the aspartate residues on one surface of CXCR4 [28]. Crystal structure and computational studies confirmed the shortened and stronger interactions between the complexes and the carboxylates compared to unconstrained macrocycle complexes in AMD3100 [29]. In 2020, the same group labeled this small molecule with the radionuclide 64 Cu to form the mono-copper species 64 Cu-CuCB-bicyclam. Compared with AMD3100, CuCB-bicyclam reduced the up to six configurations to only one configuration when it formed the copper(II) complex. This rendered 64 Cu-CuCB-bicyclam with higher affinity and specific binding towards CXCR4expressing cells (U87.CXCR4 cells). Liver uptake could also be observed but could be blocked by Cu 2 CB-bicyclam [30].
Although both [ 64 Cu]AMD3465 and [ 64 Cu]AMD3100 exhibited significant tumor uptake, they both showed considerable uptake in other organs, such as the kidneys, liver, and spleen. The liver uptake has been hypothesized to be partly due to transchelation of 64 [33]. They evaluated the stability, binding property, and SPECT/CT performance of the complex. In a stability study, 99m Tc-AMD3465 could remain stable in saline and mouse serum for up to 4 h. In in vitro cellular studies, 99m Tc-AMD3465 exhibited radioactivity accumulation in the order CHO-CXCR4 > MCF-7 > CHO, which is in the same order as their CXCR4 expression level. In the biodistribution study, 99m Tc-AMD3465 showed a relative high tumor/background ratio and significant tumor uptake of 2.07 ± 0.39% ID/g, which could be blocked by AMD3465·6HBr to some extent. In SPECT/CT imaging studies, 99m Tc-AMD3465 exhibited a higher tumor uptake in CXCR4positive MCF-7 tumor xenografts compared with those in CXCR4-negative CHO tumors. The uptake could be blocked by the nonradioactive species AMD3465·6HBr. However, the tumor/muscle ratio of this radiotracer was also higher in MCF-7 tumors (1.4 and 3.9 at 30 and 60 min, respectively) than in CHO tumors (1.1 and 1.5 at 30 and 60 min, respectively) [33].
In 2014, Vries et al. developed a [ 11 C]Methyl-labelled CXCR4 antagonist based on AMD3465 [34]. N-[ 11 C]Methyl-AMD3465 was prepared within two steps with around 60% yield, and the total synthesis time is about 50 min. In the stability assay, more than 99% of N-[ 11 C]Methyl-AMD3465 remained intact after 2 h of incubation in human liver microsomes/rat plasma, showing the good stability of this radiotracer. In the binding affinity assay, N-Methyl-AMD3465 showed decreased binding affinity compared with AMD3465 but increased binding affinity compared with AMD3100. Furthermore, both biodistribution and PET studies demonstrated high and specific binding of N-[ 11 C]Methyl-AMD3465 in C6 tumors, whereas the accumulation of the radiotracer in other organs, such as the liver and spleen is still high [34].
Other radioisotopes, such as 76 Br and 131 I, were also labeled based on the AMD3465 and AMD3100 structures. The modification of the phenyl ring did not show an obvious change in binding affinity toward the CXCR4 binding target. Among the six studied radioligands, 76 Br-HZ270-1 ( Figure 3) exhibited the best performance for the imaging of CXCR4 expression in s.c.-located tumors rather than CNS-located tumors [35].
In 2014, Prof. Nimmagadda and his colleagues [36] developed a facile synthetic route to RAD1-24 and RAD1-52, which are cross-bridged analogs of cyclams, to evaluate their radiochemical properties. The author also tried to synthesize the side-bridged cyclam analogs. However, the trial was unsuccessful due to the instability of its exposure to oxygen peptides. In binding affinity studies, Cu(II)-coordinated compounds of RAD1-24 and RAD1-52 showed increased affinity compared to the parent cold compounds. This is largely due to the enhanced interactions between the configurationally restricted coordinated compound and the aspartate residue in the receptor binding pocket. Furthermore, high radiolabeling yields, high affinity, high tumor-to-background ratios, and prolonged target tissue residence were approved. [ 64 Cu]RAD1-24 and [ 64 Cu]RAD1-52 showed a higher uptake in CXCR4positive tumors than in control tumors. The author also tried to synthesize RAD1-39, a carboxylic acid analog, in order to improve the image contrast of the cross-bridged AMD3465 analogs. However, this molecule showed neither CXCR4-specific in vitro affinity nor in vivo uptake in the tumors. This may partly be due to steric hindrance or electrostatic repulsion caused by the introduction of the carboxyl group [36].
Molecules 2023, 28, x FOR PEER REVIEW 6 of 22 radioligands, 76 Br-HZ270-1 ( Figure 3) exhibited the best performance for the imaging of CXCR4 expression in s.c.-located tumors rather than CNS-located tumors [35]. In 2014, Prof. Nimmagadda and his colleagues [36] developed a facile synthetic route to RAD1-24 and RAD1-52, which are cross-bridged analogs of cyclams, to evaluate their radiochemical properties. The author also tried to synthesize the side-bridged cyclam analogs. However, the trial was unsuccessful due to the instability of its exposure to oxygen peptides. In binding affinity studies, Cu(II)-coordinated compounds of RAD1-24 and RAD1-52 showed increased affinity compared to the parent cold compounds. This is largely due to the enhanced interactions between the configurationally restricted coordinated compound and the aspartate residue in the receptor binding pocket. Furthermore, high radiolabeling yields, high affinity, high tumor-to-background ratios, and prolonged target tissue residence were approved. [ 64 Cu]RAD1-24 and [ 64 Cu]RAD1-52 showed a higher uptake in CXCR4-positive tumors than in control tumors. The author also tried to synthesize RAD1-39, a carboxylic acid analog, in order to improve the image contrast of the cross-bridged AMD3465 analogs. However, this molecule showed neither CXCR4-specific in vitro affinity nor in vivo uptake in the tumors. This may partly be due to steric hindrance or electrostatic repulsion caused by the introduction of the carboxyl group [36].
In 2019, Aboagye et al. developed a radiotracer named [ 18 F]MCFB based on the structure of AMD3465 due to its superior binding affinity and selectivity toward CXCR4 [39]. To avoid defluorination of the 2-fluoropyridine and 4-fluoropyridine, fluorobenzene was taken into consideration as a design strategy. The logD octanol/PBS of [ 18 F]MCFB was −1.64 ± 0.06, which was moderate among the other reported radiotracers. In the binding assay, [ 19 F]MCFB showed an IC 50 value of 111.3 nM, which is comparable to that of AMD3465 (89.8 nM). In in vitro binding studies, both U2932 (higher CXCR4 expression) and SuDHL8 (lower CXCR4 expression) cell lines were used. The uptake of [ 18 F]MCFB in the U2932 cell line was higher than in the SuDHL8 cell line. However, the addition of AMD3465 will lead to partial inhibition of the binding, which means the existence of partial nonspecific binding. Even though the specific uptake was sensitive to the CXCR4 expression level, knockdown of CXCR4 in the MDA-MB-231 shCXCR4 cell line with shRNA decreased the [ 18 F]MCFB uptake, which is consistent with the decrease in the CXCR4 expression level. In PET studies, [ 18 F]MCFB showed almost two-fold higher uptake in the U2932 tumor than in the SuDHL8 tumor, which is consistent with the CXCR4 expression level. In biodistribution studies, tumor uptake in U2932 was still higher than SuDHL8. However, bone uptake was relatively low, showing a low possibility of defluorination [39].
The first 68 Ga-labeled CXCR4 imaging probe was published by Wester and his collaborators. In 2011, Wester et al. developed CPCR4-2 (Pentixafor), a small cyclic pentapeptidebased molecule, to chelate with 68 Ga 3+ through the interaction with the DOTA moiety ( Figure 4) [40]. In the binding assay, the indium complex of CPCR4-2 exhibited an affinity of 44 ± 4 nM towards Jurkat cells (CXCR4-positive), while the binding affinities of the gallium complex were 5 ± 1 nM, which is comparable to the unmodified pentapeptide FC131. In in vivo studies carried out in nude mice bearing human small cell lung cancer tumor xenografts, 68 Ga 3+ -labeled CPCR4-2 showed CXCR4-specific tumor uptake, fast renal excretion, and high tumor-to-muscle ratios.  In 2011, the same group carried out further pharmacologic studies of CPCR4-2. In the lipophilicity assay, 68 Ga-CPCR4-2 ( 68 Ga-Pentixafor) displayed enhanced hydrophilicity with a logPoctanol/PBS of 22.90 ± 0.08, which is much higher than that of 125 I-FC131. Competition binding studies and biodistribution studies showed that 68 Ga-CPCR4-2 possesses high and specific tumor accumulation and low uptake in the nontumor region, leading to high-contrast images of tumors in small-animal PET studies [41].
In order to evaluate the PET imaging property of 68   In 2011, the same group carried out further pharmacologic studies of CPCR4-2. In the lipophilicity assay, 68 Ga-CPCR4-2 ( 68 Ga-Pentixafor) displayed enhanced hydrophilicity with a logP octanol/PBS of 22.90 ± 0.08, which is much higher than that of 125 I-FC131. Competition binding studies and biodistribution studies showed that 68 Ga-CPCR4-2 possesses high and specific tumor accumulation and low uptake in the nontumor region, leading to high-contrast images of tumors in small-animal PET studies [41].
In order to evaluate the PET imaging property of 68 Ga-Pentixafor in patients with solid tumors, Vag et al. (2016) performed PET imaging experiments on 21 patients with histologically proven pancreatic cancer, laryngeal cancer, non-small cell lung cancer (NSCLC), prostate cancer, etc. [42]. Moreover, 18 F-FDG was also used in 10 out of 21 patients with a total of 27 lesions as a comparison. This comparison between 68 Ga-Pentixafor and 18 F-FDG demonstrates that 18 F-FDG is superior to 68 Ga-Pentixafor. For example, among the 27 lesions evaluated, only 19 of the 27 lesions could be detected with 68 Ga-Pentixafor, whereas 18 F-FDG could detect all 27 lesions. However, among all measured lesions, 18 F-FDG demonstrated significantly higher SUV max and T/B ratios compared with 68 Ga-Pentixafor [42].
Based on the structure of pentixafor, Poschenrieder et al. [61] developed the other two pentapeptide-based structures, NOTA-pentixather and NODA-NCS-pentixather. Both molecules were synthesized and labeled with Al[ nat F], while only [ 18 F]AlF-NOTA-pentixather was evaluated in vitro and in vivo. The logP octanol/water value of [ 18 F]AlF-NOTA-pentixather was −1.4, which is a little higher than [ 68 Ga]pentixafor (−2.9). This could be due to the one less carboxylate group in NOTA compared with DOTA. As a result, [ 18 F]AlF-NOTApentixather showed increased accumulation in the gall bladder and intestines. In the binding affinity assay, [ nat F]AlF-NOTA-pentixather showed 1.4-fold higher CXCR4 affinity compared with [ nat Ga]pentixafor. Both [ 18 F]AlF-NOTA-pentixather and [ 68 Ga]pentixafor were evaluated in a biodistribution study conducted in Daudi xenograft-bearing mice. In accordance with the hydrophilic trend, [ 18 F]AlF-NOTA-pentixather showed delayed blood clearance. Furthermore, relatively high bone activity levels were observed, which was attributed to the defluorination phenomenon. However, both high CXCR4-specific in vivo uptake and high contrast in PET imaging were observed, and thus this molecule once again demonstrates the excellent properties of pentapeptide (FC131)-based radiotracers [61].
[ 68   With the development of [ 68 Ga]Pentixafor and [ 68 Ga]Pentixather and their clinical achievements, researchers tend to modify the fine chemical structures of these two tracers to further improve the ligand-receptor interaction. In 2020, Schottelius et al. [63] replaced the AMBA-linker in [ 68 Ga Lu]DOTA-r-a-ABA-iodoCPCR4 showed more than 4-fold total cellular uptake than [ 177 Lu]Pentixather, which is identical with their affinity results. In the lipophilicity assay, due to the cationic nature of the r-a-ABA structure, the introduction of the r-a-ABA linker led to a generally reduced lipophilicity of [ 68 Ga/ 177 Lu]DOTA-r-a-ABA-CPCR4 and [ 68 Ga/ 177 Lu]DOTA-r-a-ABA-iodoCPCR4 compared with the reference ligands [ 68   With the development of [ 68 Ga]Pentixafor and [ 68 Ga]Pentixather and their clinical achievements, researchers tend to modify the fine chemical structures of these two tracers to further improve the ligand-receptor interaction. In 2020, Schottelius et al. [63] replaced the AMBA-linker in [ 68 Ga Ga] for PET imaging and [ 177 Lu] for therapeutic application, were selected for further evaluation. In the internalization study, both [ 177 Lu]DOTA-r-a-ABA-CPCR4 and [ 177 Lu]DOTA-r-a-ABA-iodoCPCR4 showed more than 4-fold total cellular uptake than [ 177 Lu]Pentixather, which is identical with their affinity results. In the lipophilicity assay, due to the cationic nature of the r-a-ABA structure, the introduction of the r-a-ABA linker led to a generally reduced lipophilicity of [ 68 Ga/ 177 Lu]DOTA-r-a-ABA-CPCR4 and [ 68 Ga/ 177 Lu]DOTA-r-a-ABA-iodoCPCR4 compared with the reference ligands [ 68 Ga Ga]Pentixafor, which might be due to the enhanced background signal [63].
In 2020, Ferro-Flores and Jiménez-Mancilla et al. [64] developed 99m Tc and 177 Lu labeled pentapeptide-based CXCR4 targeted radiotracer pairs, 99m Tc-CXCR4-L and 177 Lu-CXCR4-L, for theranostic purposes. In molecular docking calculations, the theoretical affinity of HYNIC-CXCR4-L (−10.9 kcal/mol) was comparable with that of CVX15 (−9.2 kcal/mol, cyclopeptide co-crystallized with CXCR4 monomer downloaded from the RCSB Protein Data Bank). In the study in cancer cells, the uptake of 177 Lu-CXCR4-L in DU-4475 and C6 cells (CXCR4-positive) was significantly higher than that of 99m Tc-CXCR4-L in the same cells. However, in the aspect of internalization, 177 Lu-CXCR4-L showed a lower value than 99m Tc-CXCR4-L, which might be due to the chemical effect of the DOTA structure in 177 Lu-CXCR4-L. In biodistribution studies, 99m Tc-CXCR4-L showed 2.5% ID/g after 3 h post-injection, whereas 177 Lu-CXCR4-L showed 1.5% ID/g after 96 h post-injection. Micro-SPECT/CT images of the same animal injected with 99m Tc-CXCR4-L at day 0 and 177 Lu-CXCR4-L at day 2 clearly showed the tracer uptake in the CU-4475 and the C6 tumors, demonstrating the possibility of this radiotracer pair as a theranostic pair [64]. Further, nine patients with evidence (MRI) of brain tumors were screened with SPECT after 99m Tc-CXCR4-L injection, and seven of them were diagnosed as grade II oligodendroglioma, grade IV glioblastoma, grade IV gliosarcoma, metastasis, or diffuse astrocytoma. The other two negative SPECT patients were diagnosed with reactive gliosis, confirmed with immunohistochemistry [65].
Due to its superior binding property toward CXCR4, the pentapeptide moiety (CPCR4) was introduced into other imaging methods, such as NIR (near-infrared) fluorescence imaging. In 2022, Quante et al. [66] developed MK007, in which the CPCR4 structure was conjugated to the sulfo-Cy5 moiety with the help of a linker. The lipophilicity was determined to be −1.83 ± 0.02 using [ 125 I]MK007. This new fluorescence probe was determined to be a superior probe for NIR fluorescence imaging [66].
Complexation with Ga 3+ , Lu 3+ , and Y 3+ did not change the IC 50 value of FRM001 too much. As a comparison, FC131, AMD3465, and AMD3100 showed almost 10 times higher IC 50 values. In the internalization assay, FRM001 was labeled with 67 Ga to determine the internalization activity, and the majority of [ 67 Ga]FRM001 was found to remain at the cell membrane rather than be internalized. A biodistribution study of [ 67 Ga]FRM001 was performed in CCRF-CEM tumor-bearing mice, and [ 67 Ga]FRM001 showed a high tumor-toblood ratio of 59 at 4 h post-injection. Similar to other CXCR4-targeted radiotracers, the hepatic accumulation was still high to some extent. However, co-injection with AMD3100 will reduce this accumulation. A similar phenomenon was observed in the PET image acquired at 1 h post-injection [69].
As LY2510924 is a potent peptide antagonist of CXCR4, in 2019, Bénard and Lin et al. [70] developed BL01 based on the structure of this cyclic nonapeptide. Radionuclides 68 Ga and 177 Lu were used to form complexes with BL01, and assays such as affinity and biodistribution were conducted on the tracers. In binding affinity assays, the IC 50 values of LY2510924, Ga-BL01, and Lu-BL01 are 27.8 ± 7.4, 21.2 ± 15.9, and 7.1 ± 1.7 nM, respectively. These data demonstrate that the introduction of the DOTA moiety and complex formation, especially the Lu 3+ -based complex formation, contribute to the increased binding affinity. The logD values of Ga-BL01 and Li-BL01 are −3.36 ± 0.09 and −3.34 ± 0.10, respectively, demonstrating the dominant renal clearance pathway. In the PET/CT study, [ 68 Ga]Ga-BL01 uptake could be observed in tumors, the liver, kidneys, and bladder. The tumor region could be clearly observed, and preinjection of LY2510924 could significantly reduce tumor uptake (Figure 7). In a biodistribution study, preinjection of LY2510924 could significantly reduce the uptake of [ 68 Ga]Ga-BL01 and [ 177 Lu]Lu-BL01 in tumors. For [ 177 Lu]Lu-BL01 in 1 h post-injection, tissue uptake was 12.95 ± 1.27% ID/g (lung), 11.55 ± 1.78% ID/g (spleen), and 14.00 ± 1.12% ID/g (tumor), respectively. At 4 h post-injection, the tumor-to-blood and tumor-to-muscle ratios of [ 177 Lu]Lu-BL01 were 92.9 ± 24.7 and 105 ± 24.8, respectively. At 24 h post-injection, these ratios increased to 229 ± 32.9 and 131 ± 27.7, respectively. At 24 h and 72 h post-injection, tumor uptake of [ 177 Lu]Lu-BL01 was 10.09 ± 1.41 and 3.62 ± 0.68% ID/g, respectively [70]. value of FRM001 too much. As a comparison, FC131, AMD3465, and AMD3100 showed almost 10 times higher IC50 values. In the internalization assay, FRM001 was labeled with 67 Ga to determine the internalization activity, and the majority of [ 67  showed a high tumor-to-blood ratio of 59 at 4 h post-injection. Similar to other CXCR4-targeted radiotracers, the hepatic accumulation was still high to some extent. However, co-injection with AMD3100 will reduce this accumulation. A similar phenomenon was observed in the PET image acquired at 1 h post-injection [69]. As LY2510924 is a potent peptide antagonist of CXCR4, in 2019, Bénard and Lin et al. [70] developed BL01 based on the structure of this cyclic nonapeptide. Radionuclides 68 Ga and 177 Lu were used to form complexes with BL01, and assays such as affinity and biodistribution were conducted on the tracers. In binding affinity assays, the IC50 values of LY2510924, Ga-BL01, and Lu-BL01 are 27.8 ± 7.4, 21.2 ± 15.9, and 7.1 ± 1.7 nM, respectively. These data demonstrate that the introduction of the DOTA moiety and complex formation, especially the Lu 3+ -based complex formation, contribute to the increased binding affinity. The logD values of Ga-BL01 and Li-BL01 are −3.36 ± 0.09 and −3.34 ± 0.10, respectively, demonstrating the dominant renal clearance pathway. In the PET/CT study, [ 68 Ga]Ga-BL01 uptake could be observed in tumors, the liver, kidneys, and bladder. The tumor region could be clearly observed, and preinjection of LY2510924 could significantly reduce tumor uptake (Figure 7). In a biodistribution study, preinjection of LY2510924 could significantly reduce the uptake of [ 68       In 2021, Jin et al. [72] developed an LY2510924-based radiopharmaceutical, [ 64 Cu]NOTA-CP01, which is conjugated with the NOTA moiety and labeled with 64 Cu. [ 64 Cu]NOTA-CP01 was stable in saline and FBS within 12 h of incubation. In partition coefficient evaluation, [ 64 Cu]NOTA-CP01 showed a logP value of −3.44 ± 0.12, which means a relatively high hydrophilic property. Competitive binding studies showed that the binding of [ 64 Cu]NOTA-CP01 to CXCR4 was specific, and the calculated IC 50 value was 1.61 ± 0.96 nM. In a micro-PET/CT imaging study and biodistribution study, [ 64 Cu]NOTA-CP01 was injected into EC109 tumor-bearing mice (CXCR4-positive). Among the PET images captured during 0.5-24 h, the image at 6 h was the best due to the uptake of [ 64 Cu]NOTA-CP01 in the EC109 tumor. At the 6 h timepoint, the tumor/blood and tumor/muscle ratios of the radiotracer are 4.79 ± 0.06 and 15.44 ± 2.94, respectively. However, the liver uptake of [ 64 Cu]NOTA-CP01 was still high, though the blood clearance was fast due to the hydrophilic property of the molecule [72].
In 2016, to overcome the high liver uptake of 64 Cu-AMD3100, Denat et al. [73] developed three radiotracers, AMD3100-DOTA, AMD3100-NODAGA, and AMD3100-ph-NODAGA, based on the AMD3100 moiety and DOTA/NODAGA chelators, using PEG 3 as a linker (Figure 9). The PEG 3 linker was introduced to reduce the high lipophilicity of 64 Cu-AMD3100, and the DOTA/NODAGA moiety was introduced to avoid the metal release phenomenon. However, to avoid radiolabeling of cyclam with 68 Ga, Ni 2+ ions were used as blocking reagents because the Ni 2+ /cyclam complex was stable even in strong acidic solutions and the only method to remove Ni 2+ ions from cyclam was to form cyanide at high temperature. However, the liver uptake of [ 64 Cu]NOTA-CP01 was still high, though the blood clearance was fast due to the hydrophilic property of the molecule [72].
In 2016, to overcome the high liver uptake of 64 Cu-AMD3100, Denat et al. [73] developed three radiotracers, AMD3100-DOTA, AMD3100-NODAGA, and AMD3100-ph-NODAGA, based on the AMD3100 moiety and DOTA/NODAGA chelators, using PEG3 as a linker (Figure 9). The PEG3 linker was introduced to reduce the high lipophilicity of 64 Cu-AMD3100, and the DOTA/NODAGA moiety was introduced to avoid the metal release phenomenon. However, to avoid radiolabeling of cyclam with 68 Ga, Ni 2+ ions were used as blocking reagents because the Ni 2+ /cyclam complex was stable even in strong acidic solutions and the only method to remove Ni 2+ ions from cyclam was to form cyanide at high temperature. The IC50 values of nat Ga-AMD3100-DOTA, nat Ga-AMD3100-NODAGA, and nat Ga-AMD3100-ph-NODAGA are 516, 1485, and 121 nM, respectively. The affinities of these complexes are all lower than those of AMD3100, which are 14 nM. A flow cytometry assay confirmed this tendency. Binding and internalization assays demonstrated that 68 Ga-AMD3100-ph-NODAGA possesses higher total cell uptake (1.77 The IC 50 values of nat Ga-AMD3100-DOTA, nat Ga-AMD3100-NODAGA, and nat Ga-AMD3100-ph-NODAGA are 516, 1485, and 121 nM, respectively. The affinities of these complexes are all lower than those of AMD3100, which are 14 nM. A flow cytometry assay confirmed this tendency. Binding and internalization assays demonstrated that 68 Ga-AMD3100-ph-NODAGA possesses higher total cell uptake (1.77 ± 0.10%) compared with 68 Ga-AMD3100-DOTA (0.61 ± 0.22%), indicating that 68 Ga-AMD3100-ph-NODAGA might be suitable for the biodistribution study and PET imaging. Unfortunately, in H69 xenograft (CXCR4 positive) bearing nude mice, 68 Ga-AMD3100-ph-NODAGA showed lower accumulation of the radioactivity in the tumor than 64 Cu-AMD3100, with comparable accumulation in immune-related organs [73].
In 2020, Shim et al. [76] developed a benzenesulfonamide-based radiotracer (compound 5, herein referred to as [ 18 F]benzenesulfonamide-based molecule-1) according to the optimizing result of the Schrödinger Suite. The optimization process has considered each atom's contribution to the entire benzenesulfonamide-based molecule developed by the author previously. As a result, the negatively contributing moiety was replaced to introduce the radionuclide F-18. The formed [ 18 F]benzenesulfonamide-based molecule-1 is shown in Figure 10. In an in vitro binding assay, it showed an IC 50 of 6.9 nM to block TN14003, which is a CXCR4-targeted peptide. At the same condition, AMD3100 showed an IC 50 of 66 nM. In in vivo imaging studies, [ 18 F]benzenesulfonamide-based molecule-1 showed significantly higher radioactivity in the lesion of paw edema, which demonstrates its ability to visualize γ-carrageenan-induced inflammation. Furthermore, [ 18 F]benzenesulfonamidebased molecule-1 exhibited preferential accumulation in the lesion of orthotopic xenograft SCCHN (4.00 ± 0.28% ID/g) and metastatic tumors arising in the lung (1.66 ± 0.14% ID/g), though the uptake in bone marrow was also specific [76].
In 2020, Shim et al. [76] developed a benzenesulfonamide-based radiotracer (compound 5, herein referred to as [ 18 F]benzenesulfonamide-based molecule-1) according to the optimizing result of the Schrödinger Suite. The optimization process has considered each atom's contribution to the entire benzenesulfonamide-based molecule developed by the author previously. As a result, the negatively contributing moiety was replaced to introduce the radionuclide F-18. The formed [ 18 F]benzenesulfonamide-based molecule-1 is shown in Figure 10. In an in vitro binding assay, it showed an IC50 of 6.9 nM to block TN14003, which is a CXCR4-targeted peptide. At the same condition, AMD3100 showed an IC50 of 66 nM. In in vivo imaging studies, [ 18 F]benzenesulfonamide-based molecule-1 showed significantly higher radioactivity in the lesion of paw edema, which demonstrates its ability to visualize γ-carrageenan-induced inflammation. Furthermore, [ 18 F]benzenesulfonamide-based molecule-1 exhibited preferential accumulation in the lesion of orthotopic xenograft SCCHN (4.00 ± 0.28% ID/g) and metastatic tumors arising in the lung (1.66 ± 0.14% ID/g), though the uptake in bone marrow was also specific [76].

Current Clinical and Marketing Information [77]
At present, there are about 76 potential small molecular and peptide drugs (including chemical drugs and radiopharmaceuticals) targeting the CXCR4 preclinical and clinical stages, but only one drug (AMD3100, Plerixafor, MOZOBIL) has been approved for marketing. The number of related drugs filed in China is about 36. In the field of CXCR4, there are about 200 clinical studies in progress worldwide, of which less than 10 are in China.
In the aspect of chemical drugs, the small molecular drugs entering clinical trials on the Chinese mainland are those based on the structure of AMD3100 (Plerixafor), which is generally used for the treatment of non-Hodgkin's lymphoma or multiple myeloma. Phase III clinical trials of AMD3100 have been completed on the Chinese mainland

Current Clinical and Marketing Information [77]
At present, there are about 76 potential small molecular and peptide drugs (including chemical drugs and radiopharmaceuticals) targeting the CXCR4 preclinical and clinical stages, but only one drug (AMD3100, Plerixafor, MOZOBIL) has been approved for marketing. The number of related drugs filed in China is about 36. In the field of CXCR4, there are about 200 clinical studies in progress worldwide, of which less than 10 are in China.
In the aspect of chemical drugs, the small molecular drugs entering clinical trials on the Chinese mainland are those based on the structure of AMD3100 (Plerixafor), which is generally used for the treatment of non-Hodgkin's lymphoma or multiple myeloma. Phase III clinical trials of AMD3100 have been completed on the Chinese mainland (registration date: July 2014, for the treatment of non-Hodgkin's lymphoma), sponsored by Genzyme Corporation (Cambridge, MA, USA), Patheon UK Ltd. (Wiltshire, UK), and Labcorp drug development (Beijing, China) Co., Ltd., at the Peking University People's Hospital. While the study for the treatment of multiple myeloma with AMD3100 has entered Phase IV in China at the First Affiliated Hospital of Soochow University, with a registration date of July 2021.
Internationally, small molecules and peptides that entered clinical trials were also dominated by the structure of AMD3100 (Plerixafor) in the US, EU (European Union), Australia, Japan, and many other countries and regions.
In addition, a large number of other drugs, such as AMD3465, TN14003, and GSK812397, are also in preclinical studies. LY2510924 is in clinical phase I for metastatic pancreatic cancer, metastatic rectal cancer, and advanced solid tumors; Burixafor is in clinical phase II for the treatment of non-Hodgkin's lymphoma, multiple myeloma, acute myeloid leukemia, Hodgkin's disease, haematological neoplasms, etc.; Motixafortide is in registration period for the treatment of multiple myeloma, haematological neoplasms, etc.; Balixafortide is in clinical phase III for the treatment of recurrent metastatic breast cancer, multiple myeloma, metastatic breast cancer, acute myocardial infarction, myocardial infarction, HIV infection, acute myeloid leukaemia, HER2-negative breast cancer, etc.
Small molecules and peptides targeting CXCR4 are currently limited to Plerixafor injection in international markets, which was first marketed Compared with the field of chemical drugs, in the field of radiopharmaceuticals, there are few types of drugs that have reached the clinical trial stage.
Radiopharmaceuticals that target CXCR4 are mainly 68 Ga-Pentixafor, which has entered Phase II clinical trials in the US and EU with PentixaPharm GmbH as the sponsor and has entered Phase I clinical trials in China with several sponsors, including First Affiliated Hospital of Fujian Medical University, Peking Union Medical College Hospital, and Zhongnan Hospital of Wuhan University. Furthermore, this drug has also entered the clinical phase in Australia, with Royal Brisbane and Women's Hospital as the sponsor. The drug is mainly used for PET imaging and can be used to diagnose hematological tumors, secondary CNS lymphomas, multiple myeloma, lymphomas, primary CNS lymphomas, etc.
There are currently no drugs targeting CXCR4, either internationally or on the Chinese mainland, that are available on the market in the radiopharmaceutical field.
MOZOBIL's worldwide sales figures show an annual average sale of approximately 200 million euros over the last five years, with year-on-year growth [77].
Considering the paramount role of CXCR4 in the human body, we anticipate that with the development of CXCR4-targeting radiopharmaceuticals in preclinical research and clinical trials, CXCR4-targeting radiopharmaceuticals will have a prosperous future.

Perspective Development
CXCR4-based radiopharmaceuticals are likely to have two main development directions in the future: On the one hand, compared to CXCR4-based chemotherapeutics, CXCR4-based radiopharmaceuticals are currently receiving attention within only a few molecular structures, e.g., AMD3100, Pentixafor, Pentixather, etc. In recent years, LY2510924, a peptide that was radionuclide-labeled, has also undergone certain preclinical studies. However, there are still a large number of chemical drug structures that have reached the clinical study stage but have not been radionuclide-labeled for further adequate study, and there are also still potential molecular structures that have not yet been designed and molecular docking simulated. The successful translation of these potential radiopharmaceutical precursors will not only enable a more comprehensive range of CXCR4-based radiopharmaceuticals, which will help the drugs be effective against a wider range of cancer types, but will also hopefully address the shortcomings of existing CXCR4-based radiopharmaceuticals with high hepatotoxicity. Therefore, radiopharmaceutical practitioners should not only pay attention to the progress of the development and marketing of popular radiopharmaceutical precursors, such as AMD3100 and Pentixafor, for generic drug development but also pay attention to and try to develop other novel structures for innovative drug research.
On the other hand, the rise of multimodal technologies has made it possible to combine the technical means of PET and SPECT, on which radiopharmaceuticals are based, with other technical means, such as photothermal therapy and chemotherapy [66,78,79]. Such research has helped to deepen our understanding of the underlying mechanisms and intermolecular interactions and has also helped to promote the creation of new instruments and assays. However, such technologies are currently largely confined to universities and research institutes, with less attention paid to them by for-profit companies.
The application prospects of CXCR4-targeted radionuclide therapy are very promising. According to the existing literature, in the early detection of multiple myeloma, the radiotracer 68 Ga-Pentixafor showed a higher positive rate than the commonly used radiopharmaceutical 18 F-FDG. In addition, in the aspect of radionuclide labeling, it is difficult to modify the structure of 18 F-FDG to realize the application of radiotherapy. In contrast, 68 Ga-Pentixafor could realize radiotherapy through structural modification ( 177 Lu-Pentixather). On the other hand, in terms of metabolism in the body, polypeptide drugs are more beneficial to radiotherapy than small molecules such as 18 F-FDG. In general, CXCR4-targeted radiopharmaceutical research is conducive to the development of integrated diagnosis and treatment. In addition, CXCR4 is not limited to the diagnosis and treatment of multiple myeloma. Since it is overexpressed on the surface of more than 23 types of human cancer cells, including non-Hodgkin lymphoma, multiple myeloma, chronic lymphocytic leukemia, and acute myeloid leukemia, it will also make progress in the detection and treatment of other types of tumors.

Conclusions
Overall, the development of CXCR4-based radiopharmaceuticals is still on the rise and requires the concerted efforts of researchers around the world, as well as the collaboration of staff from companies, universities, and research institutes.