T1/T2 Nanoscale Coordination Polymers as Novel Contrast Agents for MRI: a Preclinical Study for Brain Tumor

In the last years, extensive attention has been paid on designing and developing functional imaging contrast agents for providing accurate noninvasive evaluation of pathology in vivo. However, the issue of false-positives or ambiguous imaging and the lack of a robust strategy for simultaneous dual-mode imaging remain to be fully addressed. One effective strategy for improving it is to rationally design magnetic resonance imaging (MRI) contrast agents (CAs) with intrinsic T1/ T2 dual-mode imaging features. In this work, the development and characterization of one-pot synthesized nanostructured coordination polymers (NCPs) which exhibit dual mode T1/ T2 MRI contrast behavior is described. The resulting material comprises the combination of different paramagnetic ions (Fe3+, Gd3+, Mn2+) with selected organic ligands able to induce the polymerization process and nanostructure stabilization. Among them, the Fe-based NCPs showed the best features in terms of colloidal stability, low toxicity, and dual T1/ T2 MRI contrast performance overcoming the main drawbacks of reported CAs. The dual-mode CA capability was evaluated by different means: in vitro phantoms, ex vivo and in vivo MRI, using a preclinical model of murine glioblastoma. Interestingly, the in vivo MRI of Fe-NCPs show T1 and T2 high contrast potential, allowing simultaneous recording of positive and negative contrast images in a very short period of time while being safer for the mouse. Moreover, the biodistribution assays reveals the persistence of the nanoparticles in the tumor and subsequent gradual clearance denoting their biodegradability. After a comparative study with commercial CAs, the results suggest these nanoplatforms as promising candidates for the development of dual-mode MRI CAs with clear advantages.


Introduction
Magnetic resonance imaging (MRI) is a powerful technique for gathering tomographic images of biological soft tissues in a non-invasive manner with a high spatial resolution and depth penetration. This radiation-free technique enables the coding of nuclear magnetization into 2D/3D images, being the most used tool in clinical diagnosis. The signal intensity of protonbased MRI ( 1 H-MRI) depends on a combination of factors: proton density, longitudinal (T 1 ) and transversal (T 2 ) relaxivity times and the cell microenvironment. 1,2 However, the intrinsic contrast provided by the combination of these factors and the biological changes due to a disease are not enough for obtaining accurate and sensitive diagnosis. For these reasons, the use of contrast agents (CAs) in MRI diagnosis is needed in order to improve image resolution due to their selective accumulation in the Region-Of-Interest (ROI). CAs can shorten longitudinal and transverse relaxation of protons respectively causing a positive enhancement (i.e. brighter image) in T 1 -weighted (T 1w ) MRI and negative enhancement (i.e. darker image) in T 2 -weighted (T 2w ) MRI in comparison with pre-contrast images. 3,4,5 Unfortunately, most of the T 1 and T 2 CAs clinically approved for MRI diagnosis, have associated several disadvantages. Usually, T 1 CAs are based on chelating complexes of paramagnetic gadolinium. 6 The presence of Gd 3+ has been related with some technical issues and health risks such as headache, nausea, dizziness and severe nephrogenic system fibrosis (NSF). 7,8,9,10,11 Furthermore, long-term deposition of Gd in the human brain due to their incomplete clearance from the organism, has been recently reported. 12,13 Apart from Gd, manganese (II) has been presented as alternative CA for T 1w images. However, the accumulative toxicity arising from Mn 2+ which induces neurological degeneration or oxidative stress in cells, might limit its clinical application. On the other hand, typical contrast agents in T 2W images are generally based on superparamagnetic iron oxide consisted mainly in i) direct conjugation of T 1 compounds (e.g., Gd or Mn-containing systems) and T 2 compounds (e.g., SPIONs). 21,[24][25][26] (ii) T 2 materials doped with T 1 contrast materials; 19,[27][28][29] (iii) magnetic nanoparticles with rational modulation of size and magnetization; 3,30 and (iv) integration of T 1 contrast materials with nonmagnetic porous materials. [31][32][33][34] However, these strategies are often involved in laborious procedures with multi-step synthesis. Furthermore, the possibility of quenching effect between T 1 and T 2 due to the interaction between both relaxation times requires a delicate control limiting their production and clinical applications. 23 Based on this rationale, nanoscale coordination polymers (NCPs) have appeared during the last years as good candidates for their use in medical applications. The interest of NCPs lies in their rational chemical design based on multiple combinations between metal ions and polydentate ligands. The potential multifunctionality of these nanosystems implies longer blood circulation times as compared to common chelates, less metal leaching and greater area between magnetically centers and the tissue due to the increased surface-to-volume ratio. Encouragingly, previous studies showed that biodegradable NCPs could provide T 1w properties and be combined with T 2w contrast agents to obtain nanosystems with dual-mode MRI properties. 35 Different approaches have been reported recently in order to synthesize DMCAs by Eu co-doping of a Gd NCP complex, 36 the use of porphyrin ligands for the chelation of Fe, 37 or polydopamine-based coordination complex and its affinity for Fe ions. 33 Despite the progress of DMCAs, some questions remain need to be addressed before the clinical application of DMCAs become a reality: i) facile synthesis procedures with possibilities for scale up, ii) chemically stable in biological environment by non-toxicity and nonaccumulation in the organism (avoiding the use of Gd and Mn ions) and iii) high efficiency in different tissues by good biodistribution (low uptake in liver and kidneys). 38 These key parameters have to be considered during the rational design of novel DMCAs overcoming the main drawbacks of current clinical approved CAs. Therefore, there is still need to overcome the disadvantages of single modality CAs by the preparation of robust DMCAs which are supposed to minimize the risks of ambiguity and improve the diagnostic sensitivity. Herein we demonstrate the intrinsic ability of NCPs complexes based on paramagnetic ions (Fe 3+ , Gd 3+ , Mn 2+ ) to generate in a facile, robust and reproducible one-pot synthesis biodegradable DMCAs nanoparticles. In a deeper study, we selected the Fe-based NCPs (Fe-NCP) nanosystem as a suitable candidate for preclinical studies, which overcomes the main issues of contrast agents: i) cheap one-pot procedure, ii) chemically stable, iii) non-toxic and biocompatible, iv) intrinsically T 1 and T 2 imaging capabilities in vitro and in vivo and v) good biodistribution with low uptake by liver and kidneys. This rational design provides a protocol for the synthesis of novel systems in successful way, which reduces the main limitations of other reported systems. To the best of our knowledge, this is the first report on the fabrication of Fe-NCP for T 1w /T 2w contrast enhancement for the diagnosis and visualization of glioblastoma in vivo allowing the imaging of cerebral lesions. Our work would help to the development of novel one-pot NCP-based DMCA systems as a valid strategy with potential applications in theranostic and clinical translation to T 1 /T 2 dual-mode MRI. Ultraviolet-visible spectroscopy (UV-Vis) was performed using a Cary 4000 UV-vis spectrometer (Agilent Technologies, Santa Clara, CA, USA) within range wavelengths from 200 to 800 nm and a 1 cm path length quartz cuvette (QS 10 mm). The baseline was corrected using a blank sample of pure solvent. All the measurements were taken under atmosphere conditions. Dynamic light scattering (DLS) measurements for obtaining size distribution and zeta potential (ζ-potential) of the nanostructures were performed using a Zetasizer Nano ZS 3600 (Malvern Instruments, U.K.). All analyzed samples were diluted in order to obtain a concentration of nanoparticles suitable for the experimental calculation of size dispersion. The data reported are values coming from the mean of measurements for each sample which were measured per quadruplicate. Inductive coupled plasma-mass spectroscopy measurements (ICP-MS) were obtained using an ICP-MS NexION 300X (Perkin Elmer, San Francisco, CA, USA). All samples were measured in argon atmosphere. X-ray photoelectron spectroscopy (XPS) measurements were performed with a Phoibos 150 analyzer (SPECS GmbH, Berlin, Germany) in ultra-high vacuum conditions (based pressure 10 -10 mbar). Monochromatic Al Kα was used as X-ray source (1486.6 eV). The electron energy analyzer was operated with pass energy of 50 eV. The analyzer was located perpendicular to the sample surface. The data was collected every eV with a dwell time of 0.5 s and treated with CasaXPS version 2.317PR1.1 /Casa Software LTD, Teignmouth, UK).  All the synthesis reactions were carried out during 30 minutes under room temperature and at atmospheric conditions. The obtained precipitates were centrifuged (7500 rpm, 5 min) and washed with water and ethanol. The precipitates were vacuum dried and the solvent was removed.    µl (0.4 mmol metal kg -1 ). In order to assess both T 1 and T 2 RCE, T 1w and T 2w images (TR/TE = 200/8.5 ms and 4200/12 ms, respectively) were acquired before and immediately after i.v.

Functionalization of
injection of CA. Three T 1w and three T 2w images were acquired before the CA injection. Then, alternated T 1w -T 2w images were acquired continuously during 30 min after CA administration, resulting in a total of 15 frames for T 1w and 15 frames for T 2w images. In addition, individual T 1w and T 2w acquisitions were also performed at 2 and 24 h after CA administration. Only the slice with better defined contrast-enhanced region in the best study was used for measurements. For dual enhancement images, T 1w and T 2w images were processed with the algebra algorithm (MR signal T 1 /MR signal T 2 ) provided by Paravision 5.1.
Where S(i) is the absolute signal intensity of the "ipsilateral" region with respect to the contrast administration (visually shows contrast enhancement) and S(c) is the absolute signal intensity of the equivalent contralateral region.
For in vivo studies, two ROIs were manually selected in each slice: one corresponding to the tumor area and the second corresponding to contralateral normal brain tissue.  was used. Significant level was set to 0.05.

Synthesis and characterization of NCPs
The synthesis of NCPs containing three different paramagnetic ions widely used as CAs for MRI (Fe 3+ , Gd 3+ and Mn 2+ ) was achieved. The NCPs where synthesized using a previously published methodology 40  inducing low electrical repulsion between the nanoparticles (Table 1). With the aim to mimic the physiological environment and improve the colloidal stability of the NCPs, the nanoparticles were dispersed in a solution containing PBS and BSA (0.5 mM) at physiological conditions. 41,42 The results obtained demonstrate an improvement on the dispersions and colloidal stability of the NCPs in aqueous media, reducing the visible aggregation and showing hydrodynamic diameters close to the mean size values obtained by SEM (see Table 1 and Figure S1a

Biodegradability of NCPs
The size and the surface ξ-potential of the NCPs were studied at different pH in aqueous media (PBS buffer). In acidic media (pH < 4), the protonation of the carboxylic acid, and the consequent reduction of ξ-potential, induces the aggregation and precipitation of the nanoparticles. On the other hand, the lowest particle size value is obtained at pH = 9 in which a high percent of carboxylic groups are deprotonated and the surface charge is highly negative resulting in less aggregation due to electrostatic repulsion between nanoparticles ( Figure S6-S7).
For values below pH = 4 or above pH = 9, the decomposition of the polymeric material becomes   47 In basis of these results, the stability of Fe-NCP is enough for the MRI acquisition, without accumulation within the organism avoiding undesirable side effects.

Biocompatibility: cytotoxicity assays and reactive oxygen species (ROS) generation
The biocompatibility of NCPs using in vitro viability assays on HeLa (human cervix  Table 2 (graphical changes on r 1 and r 2 are shown in Figure S12a). The The MR behaviour of contrast agents relies on the relaxivity ratio (r 2 /r 1 ). If the material behaves as a T 2 contrast agent, r 2 /r 1 ≥ 10 and as T 1 contrast agent if r 2 /r 1 < 2. 55 Therefore, nanoparticles with an intermediate r 2 /r 1 ratio can be excellent candidates as T 1 /T 2 DMCAs. In our case, the in vitro phantom imaging studies are in agreement with this assumption for Fe-NCP, since the relaxivity ratio r 2 /r 1 is 2.1, indicating its potential use as DMCA. The r 1 value of Fe-NCP is comparable or even higher than those found for previously reported iron-based coordination polymers, 56,57 or molecular complexes, 58 which exhibit r 1 values up to 3.0 mM -1 s -1 .
Such good response is directly related to the Fe-NCP capacity to modify the relaxation times of the water protons in the surrounding medium when a magnetic field is applied. On the one hand, it has been reported that catechol-based ligands maximize second-sphere interactions with water molecules, and therefore to enhance T 1 , through hydrogen bonding with the oxygen atoms of the Fe-O-R linkages. 59 On the other hand, the coordination polymer probably contains water molecules strongly coordinated to the metal sites, as well as free water molecules. These labile water molecules are probably in exchange with bounded water molecules, diffusing through the polymeric matrix. 60 The mobility of the metal coordinated water in the first and second coordination spheres should induce an effect on the relaxation times of the water protons, resulting on the observed ratio r 2 /r 1 .
In the case of gadolinium-based nanoparticles (Gd-NCP and GdDTPA-NCP) although they perform satisfactorily as T 1 CAs, no significant effect as T 2 CA is detected and the r 2 values obtained (4.9 ± 0.2 mM -1 s -1 and 5.1 ± 0.3 mM -1 s -1 , respectively) are comparable to the commercial GdDTPA complex (4.6 mM -1 s -1 ) resulting in r 2 /r 1 ~ 1 ratio. The same trend is observed for Mn-NCP, which shows a typical behaviour of a T 1 contrast agent with r 2 /r 1 = 0.6 ratios exhibiting value comparable to the Mn-based commercial MnDPDP CA (r 1 = 3.1 ± 0.5 mM -1 s -1 , r 2 = 2.3 ± 0.1 mM -1 s -1 ).

Ex vivo
Before performing in vivo studies, it is necessary to select the appropriate candidate which produces the best relative performance in a tissue-like environment. 61,62 Although in vitro T 1 and T 2 relaxivity times were calculated, the results obtained from them are not usually comparable in real tissue. Most of the in vitro strategies used to evaluate the CAs performance, are not able to reproduce the in vivo conditions that could modify their ability to generate contrast. [63][64][65] For these reasons, ex vivo studies were conducted essentially as reported previously. 62 Table 2 (graphical changes on T 1 and T 2 are shown in Figure S12b). Fe-NCP and Gd-NCP presents RCE T 1 comparable with the two commercial CAs GdDTPA and MnDPDP (with no significant differences). Mn-NCP produces poorer and non-reproducible results presumably due to the low stability of these nanoparticles in solution. Regarding RCE T 2 , Fe-NCP presented the best T 1w /T 2w result, although significance is only reached in comparison with commercial MnDPDP, probably due to high value dispersion. The dual potential of this CA is more clearly observed when the calculated ratio of RCE T 1 /T 2 is considered. Ideally, a DMCA should present high absolute values for RCE T 1 (signal increase) and low absolute values for RCE T 2 (signal decrease), resulting in a high RCE T 1 /T 2 ratio, as observed for Fe-NCP (RCE T 1 /T 2 = 7.7) ( Table 2). In this case, the ex vivo studies validate the enhanced performance described in vitro for Fe-NCP. Additionally, the low toxicity and high chemical stability make this iron-based complex the best candidate for its preclinical study.

Tolerability
Previous to in vivo assays, a tolerability study to determine the maximum safe dose for Fe-NCP complex was done. The study was carried out following a protocol adapted and previously reported. 66

MRI studies
For the in vivo CAs study, GL261 glioblastoma tumor-bearing mice were i.v. injected with 107.1 ± 6.4 µl containing the equivalent dose of 0.4 mmol metal kg -1 . The commercial GdDTPA CA was administered first, as reference for direct comparison (the complete description of the procedure and acquisition parameters is detailed in Experimental section). Representative T 1w (Figure 5a,b) and T 2w (Figure 5c,d) images were acquired pre-and post-CA injection.
In a second experiment, Fe-NCP was administered and T 1w and T 2w images acquired.
Consequently, dual enhancement images were obtained through a post-processing algebra algorithm application (T 1w /T 2w , Figure 5e,f). The RCE values (Table 3) were calculated by manual selection of two ROIs in each MRI slice: one corresponding to the tumor region and a second one to the contralateral normal brain tissue ( Figure 5). In this case, the RCE values give information about the contrast improvement before (normalized as 100%) and after CA injection.
Additionally, the basal RCE values for Fe-NCP are completely recovered after 24 h and 30 min post-injection for T 1 and T 2 , respectively. These values confirm the rapid biodegradability of Fe-NCP and its potential as DMCA for MRI (see Figure 5f where RCE T 1 /T 2 ratio is shown). It is worth noting that for Fe-NCP, the T 1max is observed at 9.4 ± 1.1 min, meanwhile for the commercial CA the maximum is achieved at 6.1 ± 1.1 min. This difference may be related to the greater retention and accumulation of the nanostructured material due to the enhanced permeability and retention effect (EPR) of nanoparticles in tumors as described in previous works. 65 In our case, the Fe-NCP nanoparticles are able to produce T 1 and T 2 RCE at reasonably short times in the preclinical glioblastoma model (between 3.95 and 10.72 min after administration for maximum T 1 /T 2 effect). This behaviour would be a clear advantage, allowing the acquisition of both data types (T 1w and T 2w ) in the same exploration, instead of performing two explorations with a large interlude. Thus, these nanoparticles present a clear interest for future studies, bearing a strong translational potential. This is not restricted to contrast enhancement, as they could be also used as nanocarriers for tumor-drug release and therapy in addition to diagnosis. 73,74

Biodistribution
To further verify the biocompatibility studies and ex vivo/in vivo properties as MRI CAs, the biodistribution of Fe-NCP nanoparticles was studied. The quantification of Fe 3+ levels within the different organs was determined by using ICP-MS. In Figure 6, the biodistribution of iron from Actually, the gradual increase concentration of iron in bladder is an indicative of a gradual biodegradation of the Fe-NCP since the size needed to pass through the kidneys and reach the bladder should be smaller than 8 nm. 75 The large uptake of the Fe-NCP observed in spleen indicates its important role in the nanoparticle pharmacokinetics based on their rapid clearance by the mononuclear phagocytic system (MPS) which is common for unmodified nanoparticles bigger than 40 nm. 76,77 Although the persistence of Fe-NCP in the spleen may arise initial concerns, no adverse effects were detected in the 30 day-long tolerability assays described before.

Supporting Information.
In vivo relative contrast enhancement (RCE %)