Molecular Imaging Probes Based on Matrix Metalloproteinase Inhibitors (MMPIs)

Matrix metalloproteinases (MMPs) are a family of zinc- and calcium-dependent endopeptidases which are secreted or anchored in the cell membrane and are capable of degrading the multiple components of the extracellular matrix (ECM). MMPs are frequently overexpressed or highly activated in numerous human diseases. Owing to the important role of MMPs in human diseases, many MMP inhibitors (MMPIs) have been developed as novel therapeutics, and some of them have entered clinical trials. However, so far, only one MMPI (doxycycline) has been approved by the FDA. Therefore, the evaluation of the activity of a specific subset of MMPs in human diseases using clinically relevant imaging techniques would be a powerful tool for the early diagnosis and assessment of the efficacy of therapy. In recent years, numerous MMPIs labeled imaging agents have emerged. This article begins by providing an overview of the MMP subfamily and its structure and function. The latest advances in the design of subtype selective MMPIs and their biological evaluation are then summarized. Subsequently, the potential use of MMPI-labeled diagnostic agents in clinical imaging techniques are discussed, including positron emission tomography (PET), single-photon emission computed tomography (SPECT) and optical imaging (OI). Finally, this article concludes with future perspectives and clinical utility.


Introduction
Molecular imaging is a rapidly emerging multidisciplinary field that can visualize physiological or pathological processes inside the body at the cellular or molecular level. Molecular imaging plays an important role in noninvasive earlier diagnosis, the accurate detection of diseases or dysfunctions, treatment follow-ups, personalized treatments, and it is also useful in drug development and discovery processes [1]. There are several modalities available for clinical imaging purposes, including positron emission tomography (PET), single photon emission computed tomography (SPECT), magnetic resonance imaging (MRI), X-ray computed tomography imaging (CT), optical imaging (OI) by either bioluminescence or fluorescence, and ultrasound (US). Each imaging modality has advantages and disadvantages in aspects, such as spatial resolution, penetration depth and sensitivity which are collected in Table 1 [2]. In recent years, many efforts have been made to improve the imaging instruments and advanced image reconstruction techniques to obtain high resolution images that disclose even tiny lesions and gather accurate quantification of pathological processes. Both academia and the pharmaceutical industry are making a great effort to develop novel molecular imaging probes. The ideal molecular imaging probe must bind to its target with enough affinity, have optimal circulation time inside the body, rapidly Hence, radio-/fluorophore-labeled small molecule MMPIs could be useful for the detection of activated MMPs by means of molecular imaging techniques, such as SPECT, PET and OI. In recent years, numerous labelled MMPIs imaging agents have emerged. This review summarizes the MMP-based imaging probes for diseases listed above. The labelled MMPIs imaging agents based on the specific disease application and on their structure have been organized. An intensive broad discussion has been made on the synthetic methods including the yield, purity, overall synthesis time and specific activity of the probes, in vitro MMP isoform selective binding affinity, in vitro imaging, in vivo biodistribution, metabolic stability assessment, and in vivo preclinical to human clinical imaging applications. This review specifically covers the development of small-molecule PET, SPECT and OI probes but does not include

Classifications and Structures of Matrix Metalloproteinases
During the last 40 years, MMPs have become an attractive pharmacological target due to their relationship with several diseases, especially in cancer invasion and metastasis [12]. The main role of these zinc-dependent endopeptidases lies not only in their ability to remodel the extracellular matrix (ECM) by degrading the connective tissue, but also in their involvement in contact cell regulation [13]. The normal function of these enzymes is related to tissue homeostasis by maintaining the complex ECM and preserving its biochemical and mechanical properties. Therefore, a dysregulation in MMP activity could lead to both the development and cancer progression, as well as other diseases such as atherosclerosis, osteoarthritis, pulmonary and cardiovascular diseases [14]. The entire MMP genome includes 24 distinct genes encoding for more than 26 members. These have been classified based on the substrate type (collagenases, gelatinases, stromelysins, matrilysins, membrane-type MMPs and others) or by their structure [15]. Structurally distinct MMPs such as MMP-2, -7, -8, -9, -12 and -13 have been targeted with MMPIs labelled as diagnostic agents [6]. MMPs are synthesized and secreted as preproenzymes (zymogens) apart from MMP-8 and MMP-9 that are present in neutrophils and located inside granulomas. The activation process typically occurs extracellularly by proteolytic cleavage by serine proteinases or even by other MMPs. Commonly, MMPs are multi-domain proteins, which contain a signal sequence that targets the peptide for secretion; a propeptide responsible for the inactivation; a catalytic domain; a linker domain (also called a hinge region); and a hemopexin-like domain [16]. The propeptide is made up by the PRCGXPD motif that can inhibit the enzyme by a Cys-Zn 2+ chelation that occludes the active site. Apart from the auto inhibition by tissue inhibitors of matrix metalloproteinases (TIMPs) and the general plasma proteinase inhibitor, α2-macroglobulin can also regulate MMPs activity [17]. The MMP catalytic domain consists of a globular domain formed by approximately 170 amino acids where five central β-sheets are wrapped around four α-helixes. These metalloenzymes contain two zinc ions: One is part of the catalytic site and is coordinated by three histidines contained in the well conserved zinc-binding motif (HEXXHXXGXXH); the second exerts structural functions and its absence is linked to an autoinhibition allosteric process [18]. Furthermore, at least 2 or 3 calcium ions having structural functions are present depending on the 3D-crystal structure. Furthermore, the general structure presents five β-sheets (βI-βV) and three α-helixes (αA, αB and αC) within the N-terminal subdomain. The catalytic site is located in the αB (Figure 1). By linking αB and αC, a notable Ω-loop is found, presenting a variety of lengths and amino acid compositions depending on the type of MMP. The Ω-loop has become one of the paradigms in MMP drug discovery, as it is the least conserved fragment within the catalytic domain that can offer different ligand selectivity [19]. There are several hydrophobic pockets described in the catalytic domain of MMPs which are designated by a primed or unprimed S subsite. As described by Babine and Bender [20,21], the subsites on the left side of the Zn 2+ ion are designated without a prime S (S1, S2, S3..), while those present on the right side are primed (S1´, S2´, S3´. . . ) and the numbers refer to the residue numbers present on the substrate that occupy the pockets being the first ones cleaved by the MMP and 2 and 3 the adjacent. The S1' pocket is a highly hydrophobic cavity formed by the Ω-loop. Although this subsite is present in all MMPs, its volume and plasticity varies depending on the nature of the Ω-loop, giving rise to a MMP classification according to the size of the S1' pocket: Small (MMP-1, MMP-7, MMP-20, MMP-22), medium (MMP-2, MMP-8, MMP-9, MMP-12, MMP-14, MMP-16), and large (MMP-3, MMP-10, MMP-13) [22]. The S1' pocket is the major hydrophobic cavity near the active site and has attracted major attention in the design of selective MMP inhibitors because of its variation in size and depth [23]. Furthermore, another side pocket has been described, the S1'*, which extends beyond the S1' pocket, and its accessibility depends also on the position and conformation of the Ω-loop. The presence of this extra pocket has been only mentioned in some MMPs, such as MMP-8 and MMP-13 [24,25]. In addition to the S1' and S1'* pockets, the S1, S2, S3 and S2 , S3' subsites can also offer some degree of selectivity [21].

MMPs as Biomarkers in Cancer, Atherosclerosis, Osteoarthritis, Pulmonary and Cardiovascular Diseases
Although the knowledge that MMP overexpression has an influence in the risks and prognosis of several types of cancer and other diseases is not new, the application of MMPs as biomarkers is relatively recent. An ideal biomarker presents the capacity to be measured and evaluated as an indicator of the biological process of the target. Therefore, a biomarker should be accurate, non-invasive, sensitive, non-expensive and easy to perform [26]. The recently developed biomarkers are deficient in specificity and sensibility when they are applied in a single combination, so nowadays biomarker combination is one of the strategies to overcome these weaknesses [27].
Currently, MMPs are being explored as biomarkers in several fields such as diagnosis, monitoring and treatment efficacy in different diseases. In this line, their application in cancer, atherosclerosis, osteoarthritis, and pulmonary and cardiovascular diseases has been remarkable. This novel approach comes mainly because the overexpression of MMPs in these diseases is specific and elevated. Moreover, these enzymes are available in the extracellular media, which means the biomarker molecule does not need to cross the cell membrane. An extensive review made by Huang et al. [28] summarized how MMP-9 can be explored as a potential biomarker, and also how this MMP is overexpressed in several types of tumors such as colorectal carcinoma, breast, pancreatic, ovarian, cervical, osteosarcoma non-small cell lung cancer (NSCLC) and giant cell tumor of bone (GCTB). This makes MMP-9 a preferential candidate for the early detection of these types of tumors. The elevated MMP-9 expression in those tumors compared with healthy controls can be detected due to the recent development of experimental analytical techniques such as proteomic analysis, immunohistochemical studies, immunochemistry assays and tandem mass spectrometry [29][30][31]. The augmented levels of MMP-9 can be detected in plasma or blood samples, showing triple protein levels compared to healthy patients. The study carried out by Zajkowska et al. [29] showed that patients with cervical cancer in stages III and IV present MMP-9 plasma levels of 3344.44 ng/mL and 325.80 ng/mL, respectively, with almost three times more compared to healthy patients which present with normal levels of 166.00 ng/mL. De Chiara et al. [32] measured the MMP-9 levels in blood samples and found them to be triple (302.92 ng/mL) for NSCLC patients than in healthy controls (136.14 ng/mL). By tissue sampling, the overexpression of MMP-9 has also been detected by immunohistochemical assays using micro-RNA and labelled IgG in different types of cancer such as colorectal carcinoma, and breast cancer. Moreover, this overexpression has been linked to metastatic processes [30,31]. In the case of GCTB, the resection after surgery was the only available option to get the samples in order to measure the MMP-9 mRNA expression in peripheral tissue, confirming the overexpression in this type of cancer. These published diagnostic screenings in MMP-9 present as the main drawback the lack of tumor localization for blood samples and the inconvenience of the biopsy sampling. This gives an idea of the potential of MMP-9 as a biomarker.
Further, MMP-2, -7, -11, and -12, have been reported to be involved in cell growth and metastasis processes. In this line, MMP-12 overexpression has been linked to the grade and stage of esophageal squamous cell carcinoma in tissue sampling [33]. This protease has also been connected with the exposure to tobacco smoke and with gefitinib-resistant NSCLC that leads to a poor prognosis [34,35]. Related to its implication in the effects of the exposure to cigarette smoke, MMP-12 is a key enzyme in diseases such as chronic obstructive pulmonary disease (COPD) and has been proposed as an early biomarker due to its elevated levels in tissue sampling [36,37].
MMP-11 has promoted the occurrence and development of malignant tumors such as breast [38], gastric [39], colorectal (CRC) [40], and lung cancer [41]. Similarly, MMP-7 overexpression has been correlated with the incidence of gastric cancer. The growth/differentiation factor 15 overexpression was also linked with this type of cancer, showing almost more than ten times the protein level compared with healthy patients [42]. MMP-7 has been described to play critical roles in the development and progression of other malignant tumors including CRC [43,44] and triple-negative breast cancer [45], amongst others. In conclusion, MMP-7 has been designated as a tumor suppressor. MMP-2 has been involved in the migration and invasion of ovarian, colorectal, bladder and breast cancer, where the increased activity and expression of these enzymes have been described [46][47][48][49]. Additionally, a remarkable synergic effect between Cullin1 and MMP-2 in CRC prompt diagnosis has been proposed by Deng et al. An extensive study with 470 patients showed that protein levels were significantly upregulated in CRC tissues compared to the healthy ones [50]. However, the overexpression of MMP-2 has been linked with poor prognosis in patients with epithelial ovarian and lung cancers [51,52].
On the other hand, MMPs have also been involved in the development and progression of atherosclerosis, which is correlated with cardiovascular diseases, such as a stroke and ischemic heart diseases, two of the top worldwide causes of death. The activity of these proteases promotes the loss of collagen, elastin, and other ECM proteins, inducing the necrotic core of atherosclerotic plaque, which finally could end in myocardial infarction or a stroke [10]. It has been shown that increased levels of MMP-9 are present in specific risk groups, like obese people that present a profile for early atherosclerosis development [53]. Furthermore, Rohde et al. demonstrated that MMP-9 serum levels were increased in patients with definite carotid atherosclerotic disease compared with normal patients; and also, that only MMP-9 levels were increased in contrast with other proposed biomarkers such as MMP-1 and MMP-3 [54]. A study carried out by Nilsson et al. [55] using a cohort of 1500 patients, and measuring the plasma levels of MMP-1, -3, -7, -10, and -12, demonstrated that MMP-7 and MMP-12 were elevated in type 2 diabetes, which is related to atherosclerosis and coronary events. Another meta-study carried out by Zhong et al. [56] in which the serum levels of MMP-1 and TIMP were evaluated in patients with atrial fibrillation (AF), demonstrated that elevated mRNA levels of MMP-1 and decreased circulating levels of TIMP2 were significantly associated with AF risk.
Osteoarthritis (OA) is also one of the diseases linked to MMP overexpression due to their implication in the breakdown of articular cartilage in synovial joint tissues. Many publications have linked OA to the overexpression of MMPs in the affected tissue, especially MMP-1, -2, -9 and -13 [57,58]. MMP-1 is proposed as an early OA biomarker due to its increased concentrations in the synovial fluid in the early stage of OA compared to decreased levels when OA continues progressing [59]. Remarkably, MMP-13 has raised interest in clinical studies, as it is overexpressed in the cartilage of patients with OA and is identified as a target protein in the signaling pathways implicated in the regulation of cartilage breakdown [60,61]. MMP-3 has also been proposed as an OA biomarker together with other pro-inflammatory enzymes due to its markedly up-regulation in this disease [62].  [66]. Preliminary in vitro data showed affinities towards MMP-2 and MMP-9 in the lower nanomolar range comparable to those of the parent compound CGS 27023A. In vivo biodistribution using [ 125 I]14 in CL57 B/6 mice for 65 min p.i. showed rapid blood and plasma clearance and low retention in normal tissues. This is indicative of the suitability of these radiotracers for imaging MMP activity in vivo. However, in vivo studies using animal models that represent diseases with known MMP overexpression (e.g., tumor: Lewis lung carcinoma bearing mice; atherosclerotic plaques: apolipoprotein E-deficient mice) are needed to support their observation. Wagner   The same group reported a fully automated robust and reproducible radiosynthesis for [ 18 F]15 with reasonable yields, whereas compound [ 18 F]17 was synthesized in a one-step radiosynthesis with moderate radiochemical yield. Despite the latter presenting a high lipophilicity (experimental log D = 1.85), the initial PET in vivo studies in WT mice (C57/BL6) showed neither unspecific accumulation of [ 18 F]17 in the heart or carotid regions, nor accumulation of radioactivity in the bones due to tracer decomposition. Compound [ 18 F]18 was obtained in its two enantiomerically pure (R)-and (S)-isomers, and both demonstrated to be potent fluorinated inhibitors of MMP-2 and MMP-9 [68]. Remarkably, the (S)-enantiomer was even more active than the (R)-isomer in contrast to CGS 25966. Unfortunately, (S)-[ 18 [69], possess increased hydrophilicity compared to CGS 25966A (clog D values ranging from 0.60 to 2.25) and nanomolar inhibition potencies for MMP-2, MMP-8, MMP-9, and MMP-13 (see Table 2) [70]. 19 showed an excellent serum stability in vitro, up to 120 min, and a rapid clearance through the hepatic and renal clearance routes without presenting unspecific binding to non-excretion organs. Although these results indicate that [ 18 F]19 is a promising MMP-targeted radiotracer, there is a lack of murine disease models characterized by the up-regulated levels of the activated MMPs to carry out PET/CT preclinical studies. Compound [ 18 F]20 exhibited excellent serum stability and rapid renal clearance. Despite no defluorination or nonspecific accumulation of [ 18 F]20 in non-excretory organs were observed in vivo, the biodistribution and metabolism studies in ICR (CD1) WT mice showed mainly polar metabolites of [ 18 F]20 [70].

Fluorescent Probes for Cancer
Faust et al. developed a hydroxamate based Cyanine5.5 labelled fluorescence imaging probe for MMPs [71]. Starting from CGS 27023A and CGS 25966A, they synthesized compound 21 after a 11 steps synthetic route. Before conjugating Cy5.5, the parent compound showed complete inhibition of MMP-2/-9 activities in vitro, both for purified MMP-2/-9 (active and pro-form) and MMP-2/-9 containing cell culture supernatants. However, as expected, the fluorescent probe 21 lost the nanomolar potency of its unlabeled precursor. In situ zymography studies were conducted cryostat sections of A-673 (rhabdomyosarcoma), HT1080 (fibrosarcoma), and BT-20 (breast cancer) xenograft (CD-1 nude mice), showing high, weak and low MMP-2 and -9 activities, respectively. The MMP positive tumor cryostat sections specifically showed a higher signal with compound 21 while MMP negative tumors were not targeted. Thus, compound 21 (Scheme 3) may be used for a sensitive imaging of MMP activity in various tumours. This non-peptidic low-molecular-weight fluorescent probe may be clinically translated for endoscopic or surgical procedures.

MMP Inhibitor Imaging Probes with Carboxylic Acid as Zing Binding Group (ZBG)
Furumoto et al. [73] designed and developed 18 25 showed tumor-specific accumulation of radioactivity, with higher accumulation in the bone and intestine. However, the increase of the uptake in the bone was time-dependent, suggesting the accumulation of free 18 F − anion, released from the terminal alkyl chain of [ 18 F]25. These results suggest that the prodrug strategy could be useful, however the release of free [ 18 F − ] should be avoided.
Structurally related compound [ 11 C]26 has been studied as a potential PET cancer tracer [75]. The biodistribution and micro-PET studies in breast cancer animal models, such as MCF-7 transfected with IL-1α implanted athymic mice, and MDA-MB-435 implanted athymic mice confirmed that the localization of [ 11 C]26 in the tumor was mediated by non-specific processes ( Figure 2).  The in vitro zymography and enzyme assays showed IC 50 in the nanomolar range on gelatinases and high selectivity for MMP-2 (see Table 2). The in vivo biodistribution in NMRI white mice and in A549 lung carcinoma bearing athymic mice [78] [84]. The corresponding non-labeled compound showed some selectivity towards MMP-2 and -9 (IC 50 MMP-2 = 7 nM; MMP-9 = 2 nM). The same group reported the structurally related [ 18 F]43 radiotracer [85]. The unlabeled analogous showed nanomolar MMP inhibition potencies with higher selectivity for MMP-9 against MMP-2 (MMP-2, IC 50 = 23 nM; MMP-9, IC 50 = 7 nM). Preliminary biodistribution studies indicated that there was no tissue specific accumulation in WT C57/BL6 mice and demonstrated favorable pharmacokinetic behavior corresponding to the physical half-life of the positron emitter. Small-animal PET images of the coronal whole-body showed that the tracer uptake in non-target organs was low over all the time points and showed rapid clearance through the renal and hepatobiliary route. Thus, these barbiturate-based MMP-targeted radiotracers may be applicable in noninvasive in vivo imaging of MMP-2 and MMP-9-associated diseases by means of PET.

Breyholz et al. described a barbiturate-based MMPI radiotracer [ 125 I]42 (Scheme 7) for the noninvasive in vivo visualization of activated MMPs using SPECT or PET imaging modalities
A second generation of radio fluorinated barbiturate-based MMP tracers containing a triazole linker was described (Scheme 8) [86]. The in vivo biodistribution studies of [ 18 F]44 in WT mice showed a high level of radioactivity in the liver, and a slower liver clearance and faster kidney clearance than the more hydrophobic [ 18 F]43. Hence, the alteration of the chemical structure by introducing a mini-PEG and a triazole unit in [ 18 F]44 led to a favorable pharmacokinetic behavior for in vivo PET-imaging. However, preclinical PET/CT studies in disease models with known MMP up-regulation (e.g., tumor, Lewis lung carcinoma bearing mice; atherosclerotic plaques, apolipoprotein E-deficient mice) should be assessed [87] 47. These compounds presented high inhibitory potency (IC 50 (MMP-2) = 29 nM, IC 50 (MMP-9) = 1.3 nM) [88].
In 2008, Faust et al. reported Cy5.5 labeled barbiturate based MMPIs as fluorescence photoprobes for imaging MMPs [89]. They designed a convergent 10 step synthetic route to obtain non-hydroxamate MMP inhibitor 48, which showed high affinity (IC 50 value: 48 nM for MMP-2) measured by a fluorogenic assay using commercially available MMP-substrates and the purified enzyme. Fluorescence microscopy studies using highly (A-673, rhabdomyosarcoma) and moderate (HT-1080, fibrosarcoma) MMP-2 expressing cell lines showed efficient binding of 48 to the MMP-2 positive cells while no binding to MMP-2 negative cells (MCF-7, breast cancer cells). Pre-dosing experiments using 50 µM of the unlabeled barbiturate successfully blocked the binding of the Cy 5.5-labeled ligand 48 and showed specific uptake of the tracer. Therefore Cy 5.5 labeled tracer 48 (Scheme 9) may be a promising candidate for sensitive MMP detection in vivo.

Marimastat Based MMPIs Labelled Imaging Probes
Overall, a series of labelled Marimastat derivatives for in vivo PET imaging of MMP in cancer [90,91] have been described. First, FITC conjugated Marimastat (49) (Scheme 10) was reported to image MDA-MB-231 breast cancer cells stably transfected with MMP-14 by fluorescence. Syngeneic 67NR murine mammary carcinoma or 67NR/CMV luciferase derived primary tumor in Balb/C mice were used for in vivo PET imaging. After the injection of [ 18 F]50 to tumor-bearing mice and PET image acquisition, a low but detectable and specific uptake in the primary tumor was observed. However, in mice injected with higher concentration of [ 18 F]50 or control-ArBF 3 lacking the Marimastat moiety, no tumor localization was observed, with clearance only to the liver, bladder, and submaxillary salivary gland. As a specificity control, tumor-bearing mice were blocked with an unlabeled marimastat prior to the tracer injection. The pre-blocked mice had clearly reduced uptake with similarly sized tumors. The time-activity curve analysis revealed that [ 18 F]50 accumulated in the tumor after 60 min. Overall, the PET imaging and the time-activity curve analyses showed specific 18 F-labeling of the tumor as well as the bladder, liver, stomach, and gut. The resulting radiotracer [ 18 F]51 was evaluated in human bronchial epithelium 16HBE cells and breast cancer MCF-7 cells showing rather low binding to these cell lines. Specificity studies showed that cellular binding was reduced by 36.6% and 27.5% in MCF-7 and 16HBE cells, respectively, after co-incubation with the unlabeled counterpart.
The in vivo kinetics of [ 18 F]51 were examined in a HT1080 fibrosarcoma tumor bearing mice. In microPET scans, HT1080 tumors exhibited a low and homogeneous uptake of the tracer, suggesting that tracer binding was not only to membrane-bound disintegrin and metalloproteinase (ADAMs) but also to the extracellular MMPs. The tumors of mice injected with [ 18 F]51 showed specific uptake confirmed by blocking studies. The ex vivo biodistribution showed a rapid excretion through the kidneys and the liver. The metabolite assays indicated that the parent tracer represented 23.2 ± 7.3% (n = 2) of the total radioactivity in plasma after one and half hours after treatment. The autoradiography of a tumor slice confirmed the regular uptake on the tumor. A high kidney uptake was observed in the microPET/CT images. The change of the tumor to plasma ratio was statistically significant in contrast to the change of the tumor to muscle ratio. Therefore, the binding of [ 18 F]51 in the HT1080 xenograft mice was target mediated. Compound [ 18 F]51 may be suitable for the visualization/quantification of diseases overexpressing simultaneously MMPs and ADAMs.

Thiirane Based MMPIs Labelled Imaging Probes
Wagner et al. reported 18 F-fluorine labeled radiotracers based on SB-3CT, a slow-binding and mechanism-based MMP-2 and -9 inhibitor [93]. Inspired by the selectivity and activity of SB-3CT, they synthesized non-radiolabeled analogues and evaluated their MMP inhibition profile showing K i values ranging from 20.9 nM to 815 nM. These values are comparable to SB-3CT but they were non-selective inhibitors, with mixed kinetics and attractive residence times for MMP-2. However, this compound was unstable in human and mouse serum. The in vivo biodistribution studies revealed a rapid hepatobiliary elimination of the thiirane [ 18 F]52 (Scheme 12) and its metabolites in C57BL/6 mice and therefore, it is not suitable for PET imaging and further in vivo evaluations in MMP associated mouse models of disease.  for autoradiography and PET studies in apoE −/− mice, respectively [95].

CGS27023A Based MMPIs Labelled Imaging Probes
The cold compound 53 yielded nanomolar IC 50 values for MMP inhibition (298 and 153 nM for MMP-2 and MMP-9, respectively) making it suitable for radioligand synthesis for in vivo imaging of these MMPs activities. Upon pretreatment with non-labeled compound 53, compound [ 123 I]53 was able to specifically image MMP activity in vivo in the MMP-rich vascular lesions that develop after carotid artery ligation and a cholesterol-rich diet in apolipoprotein E-deficient mice, using WT C57/BL6 mice as the control. Additionally, compound [ 123 I]53 presented no uptake in thoracic cavity and the brain and was mainly excreted through the renal and hepatic pathways.
Compound [ 125 I]54 presented no specific uptake in the ligated carotid arteries and the results did not improve upon pretreatment with the non-labeled compound. Compound [ 124 I]55 presented a specific uptake in the ligated carotid arteries which was significantly higher in mice injected with [ 124 I]55 alone than in mice pretreated with an excess of unlabeled MMPI (Figure 4).
Although additional studies are needed to test the potential of this approach as a novel noninvasive clinical diagnostic tool for the management of human MMP-related cardiovascular diseases, these results prove that imaging of MMP activity in vivo is feasible using radiolabeled MMPIs in vascular lesions.

Macrocyclic Hydroxamate Based MMPIs Labelled Imaging Probes
The activation of MMPs after myocardial infarction (MI) contributes to adverse left ventricular (LV) remodeling. This remodeling process gives rise to LV dysfunction and progressive heart failure. The radiolabeled compounds [ 111    The micro-SPECT/CT imaging studies with analogous [ 99m Tc]58 radiotracer demonstrated a 5-fold increase of myocardial uptake. Additionally, tracer [ 99m Tc]58 showed specific uptake in atherosclerotic lesions in apolipoprotein E (ApoE −/− ) mice deficient in the low-density-lipoprotein receptor (LDLR −/− ) ( Figure 6). After treating abdominal aortic aneurysm (AAA) mouse models with [ 99m Tc]59, the SPECT/CT images showed that [ 99m Tc]59 displayed a maximum tracer uptake (Figure 7). Although further validation is required, [ 99m Tc]59-based molecular imaging might improve patient AAA management, as well as other disorders associated with MMP activity dysregulation.   [ 123 I]62 presented a high selectivity for MMP-2/-9 over MMP-1. The biodistribution and autoradiography studies of [ 123 I]62 showed an uptake in atheroprone mice, which is a clear indicator of the suitability of this compound for SPECT imaging. However, metabolite detection studies should be carried out.

MMPIs Labelled Imaging Agents for Experimental Autoimmune Encephalomyelitis (EAE) and Multiple Sclerosis (MS)
MMP-2 and MMP-9 are important for the induction of neuroinflammatory symptoms in experimental autoimmune encephalomyelitis (EAE) in a multiple sclerosis (MS) mouse model [103].
Compound 21 and [ 18 F]15, which were mentioned before as cancer probes, were also studied for EAE and MS. The injection of 21 in EAE mice showed an uptake at higher severities. Blocking studies demonstrated its specific uptake in EAE tissues (Figure 8).
They synthesized [ 18 F]15 using the method previously described by Wagner et al. [104] and used it for translational PET imaging in MS patients. A total of five patients with known or suspected relapsing-remitting MS (RRMS) presenting acute neurological symptoms participated in a [ 18 F]15-PET study ( Figure 9). All patients with the up-regulation of activated MMP-9 in cerebrospinal fluid samples exhibited a higher uptake of [ 18 F]15 into the brain (Figure 9).  (4 nmol) in two score 3 EAE mice, the brains were sliced into thick sections (200 mm) and imaged by FRI; color-encoded near infrared fluorescence and white-light images were captured and matched (left images). The same brain slices were then analyzed by in situ zymography for detection of activated gelatinases and subsequently stained with pan-laminin to define vessel borders and with CD45 antibodies to visualize inflammatory infiltrates (right images). Scale bars, 100 mm. (C,D) FRI of thick EAE brain sections (200 mm) at 5 h after intravenous injection of (4 nmol). Overlay of color-encoded near-infrared fluorescence image with significant MMP activity (red) projected onto the white-light image. The same brain slice was subsequently immunofluorescently stained with CD45 antibody; blue signal. 4 ,6-diamidino-2-phenylindole (DAPI) staining of nuclei. Two separate brains are shown in (C,D). Scale bars, 100 mm. Reprinted with permission from [103] Copyright © 2016, American Association for the Advancement of Science.

MMPIs Labelled Imaging Agents for Rheumatoid Arthritis (RA), and Osteo Arthritis (OA)
MMP-13 is overexpressed in the cartilage of patients with OA. The unlabeled Compound 63 (Scheme 15) is a picomolar MMP-13 inhibitor (IC 50 = 0.07 nM) with high selectivity towards other MMPs and ADAMs [105]. Interestingly, [ 18 F]63 exhibited excellent serum stability, and favorable in vivo biodistribution, which may be useful for PET imaging in RA and OA.  66 tracer-associated activity from the liver was detected, which lead the authors to assume that the tracer somehow accumulates in the liver, due to the metabolism. They concluded that radiolabeled selective MMP-13 inhibitors-based PET tracers could enable earlier and more specific diagnosis of several MMP-13 related diseases, such as OA and RA.

MMPIs Labelled Imaging Agents for Chronic Obstructive Pulmonary Disease (COPD) and Lung Inflammation
The synthesis and application of [ 18 F]45 in cancer was described by Elsinga and colleagues as described before [92]. In 2015, they studied its utility in COPD [107]. Dynamic microPET studies in air or cigarette smoke (CS) exposed BALB/c mice showed a 2-fold increased the uptake in mice exposed to CS compared to air exposed mice ( Figure 10). The in vivo and ex vivo PET confirmed a 2-fold increased accumulation in the CS exposed lungs compared to normal mice ( Figure 11).  A macrocyclic hydroxamate based radiotracer [ 99m Tc]59 was mentioned before for its utility in early vascular imaging and AAA [109,110]. Further, in vivo and ex vivo studies of [ 99m Tc]59 in murine models of COPD demonstrated its higher uptake in lung tissue.
Sadeghi et al. [111] developed fluorescent probes (69-73) (Scheme 17) to image active MMP-12 in inflammation and aneurysm. Among these probes, 70 (K i = 6.1 nM) and 71 (K i = 3.4 nM) presented the higher potency and selectivity towards MMP-12. Animals injected with 71 showed the highest signal in previously implanted sterile polyvinyl material (sponges) harvested for 60 min and a specific aneurysmal carotid artery when compared to a non-specific (scramble) control probe in WT mice. Mukai and col. [112] described 74 (Scheme 18), a dibenzofuran sulfonamide-based MMPI. Interestingly, due to the specific interactions with the S1' pocket, compound 74 exhibited more than a 150-fold activity for MMP-12 over MMP-8 and -13. Moreover, it prevented lung inflammation induced by recombinant human MMP-12 in C57BL/6 mice [113]. Inspired by these results, Mukai and colleagues designed and synthesized novel MMP-12-targeted dibenzofuran radiolabelled compounds with a variety of linker structures such as carbamate, amide, and sulfonamide. Among these, the best results were obtained for the carbamate linker present in compound [ 125 I]75 (IC 50 = 8.5 nM). Wagner and colleagues designed and synthesized a series of dibenzofuran unlabeled sulfonamides 76-78 [36]. These compounds revealed excellent inhibitory activities (IC 50 values: 0.0004−0.19 nM) and high selectivity (≥25 fold) for MMP-12 and to date, they are the most potent MMP-12 inhibitors described in the literature.
Preliminary biodistribution and PET studies in C57BL/6 mice showed that the corresponding radiotracers [ 18 F]76-78 exhibited no specific uptake in non-excretion organs, indicating them as promising MMP-12 based PET imaging agents.

Summary and Future Perspectives
In summary, in last two decades, several fluorescent-labelled or radio-labelled MMPIs have been developed as imaging agents for different diseases. Remarkably, pilot-clinical studies of [ 18 F]15 in MS patients described by Wagner et al. form an excellent basis for the feasibility of MMPI probes toward clinical trials [104]. A combined in vitro and in vivo evaluation revealed that molecular imaging probes based on MMPIs could have potential as validated imaging agents, but further research is necessary to reach clinical application. Although biodistribution studies of certain MMPI radiotracers in wild type mice have been described, further extensive SPECT/PET preclinical studies with proper controls in murine disease models (at least 5 mice per group), metabolite formation, stability studies, and toxicity studies are essential to confirm their clinical utility.
The clinical translation of MMPs for PET or SPECT imaging as well as other prognostic and diagnostic biomarkers, requires time and is very costly. Therefore, it is crucial to select the candidates that have the highest possibilities of success to be translated into the clinic, based not only in the abovementioned preclinical models and pharmacokinetic studies, but also on economic perspectives.

Conclusions
The overexpression of different MMPs has been associated with a number of relevant diseases such as cancer, atherosclerosis, stroke, arthritis, cardiovascular diseases, periodontal disease, respiratory tract disorders, glomerulonephritis, abdominal aortic aneurysm expansion, inflammatory bowel disease, neurodegeneration, chronic obstructive pulmonary disease, multiple sclerosis and liver fibrosis. In recent years, many MMPIs have been developed as novel therapeutics for these disorders, with some of them having entered clinical trials. More recently, numerous MMPIs labeled imaging agents have emerged as powerful tools for the early diagnosis and assessment of the efficacy of currently approved therapies. The structure-based design of highly selective MMPIs broadens the possibilities of succeeding in the development of clinically useful imaging tools. This knowledge, gathered by a wide number of research groups, would undoubtedly allow the validation of specific MMPs as relevant biomarkers in different diseases.
This review summarizes the main achievements in the field of MMPIs imaging probes and constitutes an important contribution to support further research in this area. Our research group has designed, in the last years, highly potent and selective inhibitors of MMP-2, while avoiding other MMPs, including the highly homologous gelatinase MMP-9 [114][115][116][117][118]. The structure based drug design on MMP-13 carried out by the authors has also provided very interesting and highly selective inhibitors [119]. All these compounds constitute an important starting point for the development of imaging probes. Based on this experience, the authors are currently outlining the design strategies to obtain 18 F marked analogues for PET and fluorochrome derivatives for fluorescent detection.
As highlighted in this review, a combination of medicinal chemistry and chemical biology tools has become essential for the detection of diseases in the early stages remarkably improving the treatments. MMPs have demonstrated to be relevant biomarkers for several of the most prevalent diseases. Although many contributions have been reported to date, a deeper knowledge of the involvement of these metalloenzymes will afford, in the near future, small molecule candidates as molecular imaging probes which can be useful in the clinical practice.