& Supramolecular Chemistry

Dendrimeric b-Cyclodextrin/GdIII Chelate SupramolecularHost–Guest Adducts as High-Relaxivity MRI Probes**

Jonathan Martinelli, Kalaivani Thangavel, Lorenzo Tei, and Mauro Botta*[a]

Abstract: We have synthesized a new macromolecular archi-tecture, (PAMAM)-CD8, which consists of eight b-cyclodextrinunits (b-CD) attached to a generation 1 poly(amidoamine)(PAMAM) dendrimer through a disulfide bond, which can becleaved under reducing conditions. This system shows a pro-nounced hosting capability towards GdIII chelates functional-ized with hydrophobic groups, thus leading to well-definedsupramolecular adducts. 1H NMR relaxometric investigationswere carried out to follow the formation of adducts withthree GdIII chelates based on the ligand architectures of 6-amino-6-methylperhydro-1,4-diazepinetetraacetic acid(AAZTA) or 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraace-tic acid (DOTA) suitably functionalized with benzyl or ada-mantyl (Ad) pendant groups. In particular, the ditopic com-

plex composed of two AAZTA chelating units connected toa central aromatic ring that bears an adamantyl groupshowed a strong affinity (ca. 106

m�1) for the CD units of the

dendrimer, which is two orders of magnitude higher thantoward human serum albumin (HSA). Remarkable relaxivityenhancements (i.e. , up to 71 % at 1 T and 25 8C) were ob-served upon the formation of the macromolecular host–guest adducts due to a decrease in the molecular tumblingrate and fast water-exchange. Reduction experiments andcompetition studies between the paramagnetic dendrimerand HSA were carried out by relaxometric techniques. Theresults show that the metal complexes are not displaced bythe protein, thus suggesting that this novel macromolecularprobe is potentially suitable for applications in vivo.

Introduction

Magnetic resonance imaging (MRI) enables the acquisition ofhigh-resolution three-dimensional images of the distribution ofwater in living organisms. It is widely understood that MRI, al-though providing images with excellent spatial resolution, isinherently characterized by a low sensitivity.[1, 2] The use ofa contrast agent (CA) that catalytically decreases the relaxationtimes of water protons coordinated to a paramagnetic metalcenter can reduce the overall acquisition time of the procedureand improve the diagnostic content of the image. The GdIII ionis particularly suitable for such agents due to the combinationof optimum chemical and magnetic properties.[1] However, cur-rent Gd-based commercial agents have limited contrast-en-hancement ability due to their low longitudinal relaxivity (r1),the net paramagnetic relaxation-rate enhancement of waterprotons per unit concentration of contrast agent (mm

�1 s�1).Such agents are, therefore, restricted to the visualization ofsites or organs where they can accumulate in sufficiently highconcentration, such as the bloodstream and kidneys. In thesearch for strategies for signal amplification, it was recognized

early that rotational dynamics is one of the main factors thatcontrols the efficiency of the low-molecular-weight GdIII che-lates. In fact, the relaxivity of a GdIII CA improves when increas-ing the molecular size and, hence, with slowing down of therotational tumbling rate (1/tR).[1–3] Therefore, over the past 20years, several approaches for the elongation of tR have beendeveloped by conjugating the paramagnetic complexes toa variety of macromolecular substrates.[3] For example, the re-versible formation of adducts with human serum albumin(HSA) represents one of the first strategies utilized to obtainhigh-relaxation enhancements for the GdIII complexes and toincrease their halftime in the vascular system. Besides the re-laxation enhancement, this macromolecular approach providesadditional advantages, such as increased local GdIII concentra-tion, longer retention in the biological environment, and thepossibility of multifunctional labeling that involves dyes, vec-tors, and drugs.

Among the broad collection of available macromolecularspecies (e.g. , polyaminoacids, polysaccharides, and proteins),dendrimers represent a promising and easily accessible alterna-tive. Relative to linear, cross-linked, or branched polymers, den-dritic architectures consist of nearly monodispersed, structure-controlled, and high-molecular-weight systems characterizedby an external multivalent surface formed by a well-definednumber of functional groups. In particular, poly(amidoamine)(PAMAM) dendrimers have been known since the 1980s andtheir early commercial availability has favored a wide range ofapplications.[4] The toxicity of dendrimers has been extensivelystudied because of the relevance of this issue for their biomed-

[a] Dr. J. Martinelli, Dr. K. Thangavel, Dr. L. Tei, Prof. M. BottaDipartimento di Scienze e Innovazione TecnologicaUniversit� del Piemonte Orientale “Amedeo Avogadro”Viale T. Michel 11, 15121 Alessandria (Italy)E-mail : mauro.botta@unipmn.it

[**] MRI = magnetic resonance imaging.

Supporting information for this article is available on the WWW underhttp ://dx.doi.org/10.1002/chem.201402418.

Chem. Eur. J. 2014, 20, 1 – 10 � 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim1 &&

These are not the final page numbers! ��

Full PaperDOI: 10.1002/chem.201402418

ical use,[5] and generally these compounds are more viablethan analogous linear polymers, probably because of the loweradherence to cell membranes, which is a consequence of theirglobular shape. This cytotoxicity can be further reduced byfunctionalization of the surface moieties.[6]

Dendrimer-based MRI CAs have been reported since 1990,when Lauterbur and co-workers exploited this new macromo-lecular class of compounds and observed some of the highestrelaxivity values reported at the time.[7] Since then, the proper-ties of these macromolecular CAs, prepared by conjugatingGdIII chelates to a dendrimeric architecture, have been exten-sively investigated and optimized.[8, 9] It is worth noting that,even if the Solomon-Bloembergen-Morgan theory of paramag-netic relaxation predicts a strong increase in the r1 values forthese systems because of their slow tumbling rate,[1, 10] the in-ternal rotational flexibility of the chelates and the slow rate ofexchange of the Gd-coordinated water molecule(s) often pre-vented the achievement of the theoretical r1 values.[11] A conse-quence of this behavior is the occurrence of a saturation forthe relaxivity that is observed for these systems after a succes-sion of few generations.[9] Nonetheless, because of the greatpotential of such compounds, several GdIII-loaded dendrimer-based agents have been developed and commercialized;among these systems, the so-called 24-GdIII–DTPA (DTPA = di-ethylenetriaminepentaacetic acid) cascade polymer[12] andGadomer 17[13] have shown to be particularly successful. Typi-cally, the relaxivity values of these Gd-based dendrimers at20 MHz and 25 8C cover the range 14–36 mm

�1 s�1 (per Gd ion)depending on the nature of the GdIII complex and the dendri-mer generation.

The attachment of Gd chelates to a macromolecular scaffoldcan be achieved through covalent or noncovalent conjugation.The latter strategy involves reversible supramolecular interac-tions between a suitably functionalized guest and a macromo-lecular host by exploiting the receptor-induced magnetizationenhancement (RIME) effect.[1] The equilibrium between freeand bound forms is characterized by the specific associationconstant KA. The main advantage of exploiting noncovalent ad-ducts lies in the fact that the low-molecular-weight species,namely, the Gd-loaded guest, can be readily excreted in thefree form, thus reducing the risk of toxic effects. Several stud-ies have shown that the clearance time of dendrimer-basedCAs from the body depends on size and hydrophilicity and iscorrelated to their toxicity.[14]

Cyclodextrins (CDs) are known to form inclusion complexesin aqueous solution with a variety of compounds because hy-drophobic moieties can fit their cavities with a binding affinitythat depends on the size of the moiety and cavity.[15] The highbiocompatibility of CDs is demonstrated by a wide range ofpharmaceutical applications, for example, they are commonlyemployed for the solubilization of lipophilic compounds, thedelivery of diagnostic agents, and as artificial enzymes or drugcarriers. Poly-CDs and CDs attached to or incorporated intopolymeric scaffolds have been exploited for a number of bio-technological purposes, including MRI.[16] For the same reasons,dendrimers functionalized with CDs have also been report-ed.[17] Nanoparticles that bear several CD units on their surface

have been used in vivo as systems that can bind certain recep-tors and release drugs or active substrates.[18] Very recently,a macromolecular platform composed of eight b-CD unitsanchored to the narrower rim of a g-CD core has also been in-vestigated to exploit the different hosting abilities of b- and g-CDs for multicarrier purposes.[19] In addition, bioresponsive CDnanocapsules loaded with GdIII complexes have been reportedpreviously by our group.[20]

In recent years, particular interest in the design of new MRIcontrast agents has been related to the development of com-pounds able to target specific tissues or biological environ-ments, thus providing information on cellular function and me-tabolism, such as gene expression, enzyme activity, pH, or pO2.Responsive or “smart” agents are diagnostic probes with con-trast-enhancing properties that are sensitive to the biochemi-cal parameters that characterize the microenvironment inwhich they are distributed.[21] The responsiveness of suchagents is achieved by a significant change in r1 value mostlycaused by a variation in the molecular dynamics (tR) of theparamagnetic system or the hydration state (q) of the GdIII

complex. Accordingly, research has also focused on macromo-lecular degradable scaffolds.[22] In particular, the assessment ofthe redox state in tumor tissues is of great diagnostic value be-cause it is well known that altered redox homeostasis incancer cells plays a role in tumor progression and metastasisand chemo- and radioresistance.[23] Several MRI probes sensi-tive to the redox properties of hypoxic microenvironments, allbased on the thiol/disulfide redox couple, have been report-ed[24] because the reversible reduction of disulfide bonds intothiols is of great importance in biological systems.

Herein, we describe the synthesis and characterization ofa novel dendrimeric host that bears eight b-cyclodextrin units.The agent is redox-sensitive due to the presence of disulfidebridges between the CDs and the polymeric scaffold. It is alsoexpected that the degradation of the agent into biocompatiblefragments will favor its elimination from the body after admin-istration. The interaction between the macromolecular carrierand selected Gd complexes functionalized with hydrophobicmoieties was investigated and a comprehensive 1H relaxomet-ric study was performed to assess the degree of relaxation en-hancement obtainable.

Results and Discussion

Synthesis

The macromolecular architecture of host 1 consists of eight b-CD units attached to a generation-1 PAMAM dendrimerthrough a disulfide bond (see Scheme 1 for the synthesis).Briefly, b-CD was converted into its monothiol derivative 5 byfollowing a modification of reported procedures[25] through thecorresponding monotosyl b-CD 4.[26] In parallel, PAMAM wasfunctionalized with 3-(2-pyridyldithio)propionic acid[27] (PDPA;2) to introduce a linker containing a S�S bridge. Finally, thecoupling between 5 and the modified PAMAM 3 led to the de-sired host macromolecule PAMAM-CD8 (1), which was purifiedby dialysis. The identity of the product (and in particular the

Chem. Eur. J. 2014, 20, 1 – 10 www.chemeurj.org � 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim2&&

�� These are not the final page numbers!

Full Paper

exhaustive attachment of eight CD units) was confirmed byMALDI mass-spectrometric analysis.

Supramolecular chemistry and relaxometry

PAMAM-CD8 is expected to act as a host for functionalized Gdchelates that can bind to each CD unit independently, thus re-ducing their tumbling motion (longer tR values) and improvethe relaxivity. The relaxation enhancement by the supramolec-ular structure depends on several properties of the paramag-netic guest : number (q) of coordinated water molecules, theirrate of exchange (kex = 1/tM) with the bulk, and the degree oflocal rotational mobility of the complex in the bound form(tRL). The latter is correlated with the rotational flexibility of thelinker that connects the binding group with the metal chelateand with the compactness and stereochemical rigidity of thecomplex in the free form. In addition, the chemical nature ofthe pendant group also dictates the strength of the interac-tion.[15] For this reason, we compared three types of Gd che-lates (Figure 1) that differentiate in the geometry of the coordi-nation cage (6-amino-6-methylperhydro-1,4-diazepinetetraace-tic acid (AAZTA)- or 1,4,7,10-tetraazacyclododecane-1,4,7,10-tet-raacetic acid (DOTA)-like), in the nature of the binding moiety(phenyl or adamantyl (Ad) groups), or in the value of some keymolecular parameters (q and tM). The formation of host–guestadducts with PAMAM-CD8 were investigated for [(Gd-AAZTA)2-Ad] (Gd2-6),[28] [Gd-BzAAZTA] (Gd-7),[20] and [Gd-DOTAMA-Ad](Gd-8 ; DOTAMA = DOTA monoamide).[19] All three complexescan form inclusion adducts between their hydrophobic moiet-ies and the cavities of the CD units on the dendrimer. Further-more, their different characteristics enable a clear assessmentof the relationship between the molecular structure and boththe strength of the interaction and the relaxivity enhancement.Complexes Gd2-6 and Gd-7 are based on the AAZTA plat-form,[29] which is well known to provide Gd chelates with goodthermodynamic stability, optimal parameters for attaining high

relaxivities (q = 2, tM = ~90 ns), and pronounced inertnesstoward water displacement by mono- and bidentate anions ortoward transmetalation reactions.[30] In particular, investigationswere focused on Gd2-6, a ditopic complex, which we have re-cently reported,[28] that consists of two Gd-AAZTA-like unitslinked to a phenyl-ring core further functionalized with a pend-ant arm that bears a terminal adamantyl moiety. The relaxivityvalues reported for Gd2-6 at 25 and 37 8C and 20 MHz are 16.7and 13.7 mm

�1 s�1, respectively (per mm concentration of Gd).The dimer proved to have a high affinity toward the cavity ofthe b-CD unit, both in the free and polymeric forms, withKA values of about 104

m�1.

Complex Gd-7 is a Gd-AAZTA derivative that features a pend-ant benzyl group that has a quite high value of relaxivity (r1 =

9.3 mm�1 s� at 20 MHz and 25 8C), which is consistent with the

presence of two water molecules (q = 2) in the first coordina-tion sphere of the metal ion.[20] The interactions between thephenyl rings and the b-CD cavity are typically characterized by

Scheme 1. Synthesis of 1. 1) Dicyclohexylcarbodiimide (DCC), 1-hydroxybenzotriazole (HOBt), N,N-dimethylformamide (DMF, RT, 15 h; 2) tosyl imidazole, NaOH,then NH4Cl, RT, 15 h; 3) thiourea, DMF, 75 8C, 2 days; 4) Na2S2O5, NaOH (1 m), RT, 30 m; 5) trichloroethylene, HCl, pH 1, RT, 15 h; 6) DMF, RT, 2 days. Py = pyri-dine.

Figure 1. Structure of the complexes Gd2-6, Gd-7, and Gd-8.

Chem. Eur. J. 2014, 20, 1 – 10 www.chemeurj.org � 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim3 &&

These are not the final page numbers! ��

Full Paper

KA constants on the order of 101–102m�1.[15] Hence, Gd2-6 and

Gd-7 have the hydration state and the kinetics of water ex-change in common. This comparison highlights the effect ofthe different selectivities of the binding groups and their differ-ent rotational dynamics. The dimeric complex is characterizedby a lower degree of local rotational motion (tRL) and thereforebetter motional coupling with the global reorientation of thesupramolecular structure (tRG), a factor that favors higher relax-ivity.

Unlike the other two systems, Gd-8 is a DOTA-monoamide(DOTAMA) derivative with a side chain that terminates with anadamantyl group and a r1 value at 20 MHz and 25 8C of6.3 mm

�1 s�1.[19] This significantly lower value mainly reflectsthe different hydration number of q = 1. Furthermore, in theGd-DOTAMA derivatives, the average lifetime of the boundwater molecule is quite long (tM>500 ns), thus representinga limit to relaxivity when the molecular motion is sloweddown (long tR values) ; for example, by following the formationof supramolecular adducts.[1, 2] Therefore, the comparison withGd2-6, with which Gd-8 has the nature of the binding group incommon, underlines the effect of nonoptimized molecular pa-rameters on relaxivity.

The formation of the inclusion compound between Gd2-6and PAMAM-CD8 was investigated by 1H NMR relaxometric ti-tration experiments at 25 8C and 20 MHz, namely, the so-called“E” and “M” titrations.[2, 31] The former is carried out by monitor-ing the change in the observed longitudinal relaxation rate(R1

obs = 1/T1) of a solution with a fixed concentration of the par-amagnetic probe while increasing the concentration of thehost. In the “M” experiment, a solution with a given concentra-tion of the host is titrated with increasing amounts of the Gdcomplex. The R1

obs value increases dramatically and almost line-arly with the addition of Gd2-6 up to a plateau for a guest/hostratio of about 8:1 (Figure 2). This behavior clearly indicates theremarkable strength of the interaction, the near equivalence ofthe binding sites (i.e. , eight b-CD units), and the pronouncedincrease in relaxivity of the paramagnetic probe in the boundform.

Consequently, the interaction can be analyzed in terms ofa simple 1:1 binding model that considers the presence ofeight equivalent and independent binding sites. The nonlinearfitting of the experimental data according to the paramagneticrelaxation enhancement (PRE)[31] equation provides the valuesof the association constant (KA = (2.3�0.1) � 106

m�1) and the

relaxivity for the bound ditopic complex (r1b = 28.5�

0.4 mm�1 s�1) per Gd ion. The relaxivity of the adduct repre-

sents an enhancement of over 70 % relative to the value of thefree complex, whereas the binding constant is approximatelytwo orders of magnitude stronger than that calculated for theinteraction of Gd2-6 with b-CD (1.0 � 104

m�1) and poly-b-CD

(1.3 � 104m�1).[28] Such a large KA value can be tentatively ac-

counted for with the occurrence of associative effects, oftenobserved in the case of multisite interactions that involve pro-teins.[32] These results are fully supported by the titration M,which shows a linear increase of the relaxation rate of thewater protons with the concentration of the added complexup to a 1/Gd2-6 ratio of approximately 1:8. The data can be re-

produced quite satisfactorily by using the KA and r1b values pre-

viously obtained (from the continuous line through the experi-mental points in Figure 2). Therefore, it can be concluded thatthe interaction of the ditopic complex with the functionalizeddendrimer results in the formation of a large assembly witha well-defined supramolecular architecture that contains16 Gd-AAZTA-like paramagnetic units. It is worth noting thatthe relaxivity of the whole system, that is, the value of r1 perdendrimer, is as high as 456 mm

�1 s�1. This very high value sug-gests a rather compact structure in which the ditopic metalcomplexes possess a limited rotational flexibility, probably dueto the high steric hindrance in the supramolecular adduct,which is reflected in a long “effective” rotational correlationtime (high coupling between global and local rotation).

The benefits associated with the molecular structure of Gd2-6 are clearly highlighted by the comparison with the corre-sponding data obtained for the monomeric complexes Gd-7and Gd-8. The benzyl group allows only a weak interaction ofGd-7 with the b-CD units of the dendrimer. In fact, whereasthe analysis of the titration experiments yields a r1

b value of17 mm

�1 s�1, which corresponds to a 82 % enhancement, theKA value is only 3 � 102

m�1 (Figure 3)

Therefore, saturation of the binding sites would be possibleonly in the presence of a large excess of the complex. The cor-responding data for Gd-8 (Figure 3) confirm the strength ofthe interaction between the adamantyl group and the b-CDcavity, with an association constant of 5.0 (�0.1) � 104

m�1.

However, the fitting provides a r1b value of only 10.8�

0.1 mm�1 s�1, limited by the low hydration state of the complex

(q = 1), the slow water exchange rate, and the presence ofa highly flexible linker between the adamantyl moiety and thechelate that results in a fast local rotation (short tRL). Then, thefunctionalized dimer combines the advantages of both mono-mers, that is, q = 2, fast exchange, and a very high affinity con-stant.

Figure 2. Top: plot of titration E ([Gd2-6] = 0.15 mm, pH 7, 20 MHz, 25 8C).Bottom: plot of titration M ([1] = 0.1 mm, pH 7, 20 MHz, 25 8C).

Chem. Eur. J. 2014, 20, 1 – 10 www.chemeurj.org � 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim4&&

�� These are not the final page numbers!

Full Paper

These conclusions are very well supported by the measure-ment of relaxivity as a function of magnetic field strength. Theobtained data afford the so-called nuclear magnetic relaxationdispersion (NMRD) profiles, the shape and intensity of whichdepend strictly on the relaxometric properties of the paramag-netic systems and therefore on the value of the molecular pa-rameters. In the case of the ditopic complex, either free or in-cluded in the PAMAM-CD8 dendrimer (1-[Gd2-6]8), the profilesclearly show the predominant role of the rotational dynamicsin controlling the relaxivity. Following the formation of theadduct, we can observe not only an increase in the r1 valueover the entire range of proton Larmor frequencies, but alsothe formation of a pronounced peak at high fields, centeredaround 40-50 MHz. The formation of this peak is typical of sys-tems with a slow rate of molecular rotation, and the maximumtends to shift to lower frequencies when increasing thetR value. In the case of the adducts of 1 with Gd-7 and Gd-8,the NMRD profiles reflect the different properties of the che-lates (Figure 4). For 1-[Gd-7]8, the peak is less pronounced andthe relaxivity values at low and high fields are comparable.This consequence is likely to be due to a poorer motional cou-pling between the local and global rotations and a lower valueof tR associated with the lower molecular size of the adduct.On the other hand, in the case of 1-[Gd-8]8, we may note theabsence of a peak of relaxivity, which instead takes a nearlyconstant value between 10 and 70 MHz. Although the consid-erable flexibility of the linker between the chelate and the ada-mantyl group is responsible for a low value of the effectivetR value, the long lifetime of the coordinated water moleculelikely represents the dominant limiting factor.

HSA competition experiments

The high association constant found for the ditopic AAZTA de-rivative Gd2-6 toward PAMAM-CD8 is of particular interest in

view of potential utilization in vivo. Among the reasons thatmake the biological fluids much more complex than thesimple water or buffer solutions, the presence of a variety ofproteins plays a crucial role in the actual applicability of a diag-nostic probe or a drug because they can be inactivated bystrong interactions with these endogenous species. For thisreason and as a result of the fact that serum albumin is themost abundant protein in blood plasma, competition experi-ments with HSA were carried out (Figure 5).[33] A recent studydemonstrated the high affinity of Gd2-6 for HSA, with a KA valueof 1.2 � 104

m�1.[28] The longitudinal water-proton relaxation rate

R1obs for a 0.15 mm aqueous solution of Gd2-6 in the presence

of 0.7 mm HSA is 9.8 s�1 (37 8C, 30 MHz). Under identical experi-mental conditions, the R1

obs value for a 0.15 mm solution ofGd2-6 in the presence of 0.0375 mm PAMAM-CD8 is 7.2 s�1.Upon the addition of HSA to the latter solution up to a finalconcentration of 0.7 mm, the measured R1

obs value is 7.7 s�1.The relaxation enhancement is quite marginal and likely dueto a small increase in the viscosity of the solution, thus indicat-

Figure 3. Top: plot of titration E ([Gd-L] = 0.15 mm, pH 7, 20 MHz, 25 8C; Gd-7: empty circles, Gd-8 : black squares). Bottom: plot of titration M([1] = 0.1 mm, pH 7, 20 MHz, 25 8C).

Figure 4. 1H NMRD profiles of adducts 1-[Gd2-6]8 (black diamonds), 1-[Gd-7]8

(empty circles), and 1-[Gd-8]8 (gray squares) at pH 7 and 25 8C.

Figure 5. Top: comparison among the R1obs values (pH 7, 37 8C, 30 MHz) mea-

sured for different solutions of Gd2-6 (0.15 mm) in the presence or absenceof 1 (0.0375 mm) and HSA (0.7 mm). Bottom: 1H NMRD profiles that corre-spond to the solutions of the bar plot (Gd2-6 : empty triangles, 1 + Gd2-6 :gray circles, 1 + Gd2-6 + HSA: empty circles, Gd2-6 + HSA: black squares). Thegray area highlights the range of frequencies that correspond to the maxi-mum effect (30 MHz) and MRI experiments (40 MHz; see below).

Chem. Eur. J. 2014, 20, 1 – 10 www.chemeurj.org � 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim5 &&

These are not the final page numbers! ��

Full Paper

ing that the paramagnetic probes are not displaced by HSAfrom the dendrimeric structure and confirming the results ofthe titration experiments.

Further support of this conclusion comes from the NMRDprofiles, which clearly demonstrate that the GdIII complex isnot involved in significant interactions with HSA in the pres-ence of PAMAM-CD8. As additional experimental evidence, themixture Gd2-6/1/HSA was ultracentrifuged (molecular-weightcut-off (MWCO) = 30 kDa) to separate the dendrimer (Mw

�11 kDa) from HSA (Mw�66 kDa). The relaxation rates of thewater protons were measured for the two final phases. The so-lution containing HSA had a R1

obs value close to that of the cor-responding diamagnetic aqueous solution of the protein. Onthe other hand, a much higher value was measured in the por-tion that contained the dendrimer, in agreement with the pres-ence of Gd2-6.

Reduction experiments

In principle, the supramolecular adduct between Gd2-6 andPAMAM-CD8 is responsive to reducing environments becauseof the presence of disulfide bonds between the b-CD unitsand the PAMAM scaffold. This hypothesis was tested by moni-toring the changes in the R1 value for a solution of 1-[Gd2-6]8

following successive additions of the reducing agent TCEP.[34]

This reducing agent has the advantage of being able to oper-ate with efficiency at physiological pH. The cleavage of the S�S bridges would lead to disassembly of the supramolecular ar-chitecture and a decrease in the size of the paramagneticsystem from adduct 1-[Gd2-6]8 to the smaller inclusion com-plexes between Gd2-6 and monothiolated b-CD 5 (Scheme 2).Therefore, a lower R1

obs value is expected at the end of the ex-periment due to a decrease in the rotational correlation timetR.

It was not possible to perform reduction-kinetics investiga-tions by monitoring R1

obs versus time upon the addition ofTCEP because the cleavage occurs very rapidly and the reac-tion reaches completion within a few seconds. Instead, thechanges in R1

obs were plotted against the increasing amount ofTCEP added to a solution of 1-[Gd2-6]8, while keeping theirconcentrations constant. The relaxation rate decreases linearlyuntil it reaches a minimum value that corresponds to thecleavage of all the disulfide bonds (Figure 6). In terms of ther1 value, the change is quite large (ca. 23 %, from 30.8 to23.7 mm

�1 s�1). The NMRD profile of the final solution is fullyconsistent with that recorded for the inclusion adduct betweenGd2-6 and b-CD (Figure 6).

Imaging experiments

A phantom-imaging experiment was carried out to check if(and how) the improved relaxivity of the supramolecular ad-ducts translates into image contrast (Figure 7 and Table 1). Asolution of 1-[Gd2-6]8 (see (3) in Figure 7) was imaged in a MRIinstrument (1 T) together with solutions of Gd2-6 (see (1) inFigure 7) and the commercial contrast agent Prohance (see (6)in Figure 7). The experiment clearly shows the signal-intensityenhancement that accompanies the formation of the macro-molecular adduct. A solution of 1-[Gd2-6]8 in the presence ofHSA (see (4) in Figure 7) was also imaged and compared witha sample containing Gd2-6 and HSA (see (2) in Figure 7), thusshowing that the contrast effect of solutions 3 and 4 is defi-nitely similar and certainly lower than that of sample 2, inagreement with the relaxometric results. An increase of 71 % inthe R1

obs value was calculated due to the binding interaction ofGd2-6 with the functionalized PAMAM-CD8 dendrimer. A MRIphantom was also obtained at 1 T for a solution of 1-[Gd2-6]8

after addition of TCEP (see (5) in Figure 7), thus allowing

Scheme 2. Cleavage of the supramolecular adduct 1-[Gd2-6]8 by reduction of the disulfide bonds that connect the CD units to the PAMAM scaffold withtris(2-carboxyethyl)phosphine (TCEP).

Chem. Eur. J. 2014, 20, 1 – 10 www.chemeurj.org � 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim6&&

�� These are not the final page numbers!

Full Paper

a visual comparison of the contrast with the same mixture inthe absence of the reducing agent; therefore, a decrease insignal intensity (contrast effect) is clearly visible, which is in fullagreement with the relaxometric results.

Conclusion

A generation-1 PAMAM dendrimer has been conjugated toeight b-cyclodextrin units through disulfide bridges can assem-ble suitably modified GdIII chelates based on noncovalent“host–guest” inclusion complexation. The supramolecularsystem was characterized by large T1 ionic relaxivities, a funda-mental requisite for an enhanced MRI contrast effect. In partic-ular, the best results were obtained for the ditopic Gd2-6 com-plex, which combines several favorable properties, such as thepresence of two inner-sphere water molecules with a sufficient-ly rapid exchange rate, a restricted degree of local rotation,and a strong affinity for b-CD. Upon formation of the “host–guest” complex with 1, the relaxivity of Gd2-6 increases by ap-proximately 70 %, thus reaching the remarkable value of60 mm

�1 s�1 per dimer (480 mm�1 s�1 per adduct) at 1 T and

25 8C. The noncovalent interactions are so strong that the para-magnetic probes are not displaced by serum albumin from thesupramolecular structure, which augurs well for future applica-tions in vivo in preclinical studies. Finally, the biodegradabilityof the supramolecular architecture under reducing conditionsis relevant for its viability because the possibility of degrada-tion into smaller fragments allows for safer excretion from thebiological environment. In addition, the b-CD cavities can in-clude other types of guests, such as fluorescent dyes, vectors,or drugs. Such versatility makes 1 well suitable for developingmultimodal imaging or theranostic systems.[35]

Experimental Section

Materials and equipment

All the chemicals were purchased from Sigma–Aldrich Co. or AlfaAesar Co. and were used without further purification. 1H and13C NMR spectra were recorded on JEOL Eclipse Plus 400 andBruker Avance III spectrometers operating at 9.4 and 11.7 T, respec-tively. Chemical shifts are reported relative to TMS and were refer-enced to the residual-proton solvent resonances. Electrospray ioni-zation mass spectra (MS(ESI)) were recorded on a SQD 3100 MassDetector (Waters) in the positive- or negative-ion mode withformic acid in methanol (1 %, v/v) as the carrier solvent. Matrix-as-sisted laser-desorption/ionization time-of-flight (MALDI-TOF) massspectra were recorded on a Voyager DE-PRO MALDI-TOF massspectrometer (Applied Biosystems, Foster City, CA, USA) equippedwith a nitrogen laser (l= 337 nm). A 3,5-dimethoxycinnamic acidmatrix was obtained from LaserBio Labs (Sophia Antipolis Cedex,France) and prepared as a solution in acetonitrile/ultrapure H2O/tri-fluoroacetic acid (TFA) 50:50:0.1 (10 mg mL�1). IR spectra were re-corded in the range n= 4000–400 cm�1 at a resolution of 4 cm�1

on a Bruker Equinox 55 spectrometer.

Synthesis of PAMAM-G1-PDP (3)

PDPA[27] (296 mg, 1.37 mmol) and HOBt (185 mg, 1.37 mmol) inDMF (2 mL) were added to a solution in MeOH of PAMAM-G1(20 %, 1 mL, 0.11 mmol). After bubbling N2 through the reactionmixture for 10 min, DCC (284 mg, 1.37 mmol) was added, and thesuspension was stirred at room temperature for 15 h. The mixturewas evaporated under vacuum and the residue was suspended inMeOH and dialyzed against MeOH (MWCO = 2 kDa). The final solu-

Figure 6. Top: R1obs values (pH 7, 25 8C, 40 MHz) measured upon the addition

of TCEP to a solution of Gd2-6 (0.15 mm) and 1 (0.038 mm). Bottom:1H NMRD profiles at pH 7 and 25 8C that correspond to solutions of Gd2-6 +b-CD (empty circles), 1 + Gd2-6 + TCEP (gray squares), 1 + Gd2-6 (black di-amonds; [Gd2-6] = 0.15 mm, [1] = 0.038 mm, [TCEP] = 1.25 mm, [b-CD] = 10 mm).

Figure 7. T1-weighted multislice spin-echo phantom images (1 T, TR/TE/NEX = 250/8, 9/10) for solutions of 1) Gd2-6, 2) Gd2-6 + HSA, 3) Gd2-6 + 1,4) Gd2-6 + 1 + HSA, 5) Gd2-6 + 1 + TCEP, and 6) Prohance ([Gd2-6] = 40 mm

([Gd] = 80 mm), [1] = 5 mm, [HSA] = 80 mm, [TCEP] = 40 mm, [Prohan-ce] = 80 mm).

Table 1. Relaxometric data relative to the MR images shown in Figure 7and described in the text.

Sample Content Signal intensity [a.u.] R1obs [s�1] r1 [mm

�1 s�1]

1 Gd2-6 1523 1.67 16.22 Gd2-6 + HSA 2832 3.80 42.83 Gd2-6 + 1 2189 2.59 27.74 Gd2-6 + 1 + HSA 2265 2.84 30.85 Gd2-6 + 1 + TCEP 1758 1.97 20.06 Prohance 794 0.77 4.8

Chem. Eur. J. 2014, 20, 1 – 10 www.chemeurj.org � 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim7 &&

These are not the final page numbers! ��

Full Paper

tion was evaporated under vacuum to yield 3 as a white solid(269 mg, 81 %). 1H NMR (400 MHz, 25 8C, CD3OD): d= 8.38 (m, 8 H),7.76 (m, 16 H), 7.18 (m, 8 H), 3.3–3.2 (m, 40 H), 3.04 (m, 16 H), 2.9–2.7 (m, 32 H), 2.61 (m, 16 H), 2.55 (m, 4 H), 2.4–2.3 ppm (m, 24 H);13C NMR (100 MHz, 25 8C, CD3OD): d= 173.7, 172.2, 159.8, 149.2,137.9, 121.2, 119.8, 52.1, 49.8, 38.9, 38.8, 37.4, 35.0, 34.1, 33.4 ppm;IR (KBr disk): n= 3325, 2928, 2850, 1626, 1574, 1535, 1435, 1311,1243, 1114, 1088, 1044, 892, 760, 641 cm�1; MS(ESI): m/z : 1504 [M +2H+/2]+ , 1003 [M + 3H + /3]+ , 753 [M + 4H+/4]+ .

Mono(6-deoxy-6-mercapto)-b-cyclodextrin (5)

A modification of reported procedures was used.[25] Mono(6-deoxy-6-para-toluensulfonyl)-b-cyclodextrin[26] (4 ; 3.00 g, 2.33 mmol) andthiourea (1.77 g, 23.3 mmol) were dissolved in DMF (120 mL). Thesolution was stirred at 75 8C for 2 days. After cooling at room tem-perature, the mixture was added to Et2O (500 mL) and stirred for10 min. The precipitate was filtered and washed with Et2O(100 mL), suspended in acetone (120 mL), and heated to reflux for2 h. The suspension was cooled to room temperature and filtered.The collected white solid was dried under vacuum and added por-tionwise to a solution of Na2S2O5 (120 mg) in 1 m NaOH (100 mL).After stirring for 30 min, the solution was acidified to pH 3 withconcentrated HCl. Trichloroethylene (5.3 mL) was added to the re-action mixture, and the resulting suspension was sonicated for10 min. A white solid residue was collected by filtration and driedunder vacuum to yield 5 (2.52 g, 94 %), which was confirmed bycomparison with reported characterization data.[25]

Synthesis of PAMAM-CD8 (1)

Compounds 3 (53 mg, 0.018 mmol) and 5 (245 mg, 0.21 mmol)were dissolved in DMF (6 mL) and stirred for 2 days. After evapora-tion under vacuum, the residue was suspended in H2O and dia-lyzed against H2O (MWCO = 2 kDa). The final suspension was fil-tered, and the filtrate was freeze-dried to yield 1 as a white solid(132 mg, 64 %). 1H NMR (400 MHz, 25 8C, D2O): d= 5.1–4.9 (m, 56 H),4.0–3.8 (m, 104 H), 3.8–3.7 (m, 112 H), 3.7–3.6 (m, 48 H), 3.6–3.5 (m,32 H), 3.4–3.2 (m, 64 H), 3.0–2.8 (m, 56 H), 2.6–2.5 (m, 28 H), 2.5–2.3 ppm (m, 24 H); 13C NMR (100 MHz, 25 8C, CD3OD): d= 173.7 (m),101.8, 84.6, 81.1, 80.5, 73.0, 71.7 (m), 70.9 (m), 60.2, 51.3, 48.8 (m),38.7, 36.1, 34.9, 32.9, 32.3 ppm; IR (KBr disk): n= 3354, 2922, 1640,1550, 1409, 1364, 1154, 1079, 1030, 941, 854, 754, 703, 579 cm�1;MALDI-TOF MS: m/z : 11330, 11372.

Relaxometric measurements

The water-proton longitudinal relaxation rates as a function of themagnetic-field strength were measured in non-deuterated aqueoussolutions on a fast field-cycling Stelar SmarTracer relaxometer(Stelar s.r.l. , Mede (PV), Italy) over a continuum of magnetic-fieldstrengths from 0.00024 to 0.25 T (corresponding to 0.01–10 MHzproton Larmor frequencies). The relaxometer operates under com-puter control with an absolute uncertainty in 1/T1 of �1 %. Addi-tional longitudinal and transverse relaxation data in the range 15–70 MHz were obtained on a Stelar Relaxometer connected toa Bruker WP80 NMR electromagnet adapted to variable-field meas-urements (15–80 MHz proton Larmor frequency). The exact con-centration of GdIII ions was determined by measurement of thebulk magnetic-susceptibility shifts of a tBuOH signal or by using in-ductively coupled plasma mass spectrometry (ICP-MS; Element-2,Thermo-Finnigan, Rodano (MI), Italy). Sample digestion was per-formed with concentrated HNO3 (70 %, 2 mL) with microwave heat-ing at 160 8C for 20 min (Milestone MicroSYNTH Microwave lab sta-

tion equipped with an optical-fiber temperature control and HPR-1000/6M six-position high-pressure reactor; Bergamo, Italy). The 1HT1 relaxation times were acquired by the standard inversion-recov-ery method with a typical pulse width (908) of 3.5 ms and 16 experi-ments of 4 scans. The 1H T2 relaxation times were acquired by theCarr-Purcell-Meiboom-Gill (CPMG) sequence method with a typicalpulse width (908) of 3.5 ms, 64 experiments, and a typical echo timeof 100–500 ms. The temperature was controlled with a Stelar VTC-91 airflow heater equipped with a calibrated copper-constantanthermocouple (uncertainty of �0.1 8C).

Titrations E

A solution of 1 (0.30 mm) and Gd2-6, Gd-7, or Gd-8 (0.15 mm) wasprepared by mixing volumes of the respective stock solutions anddiluting with H2O when necessary. An aliquot was taken from thissolution, which was mixed again with H2O and the stock solutionof Gd2-6, Gd-7, or Gd-8 to keep the concentration of the complexat 0.15 mm. The R1

obs value for the resulting solution was mea-sured, and this procedure was repeated several times until anR1

obs value very close to that measured for a 0.15 mm solution ofjust the complex was reached. The curve obtained by plotting R1

obs

versus the molar concentration of 1 was fitted according to thePRE theory.[31]

Titrations M

A solution of 1 (0.10 mm) and Gd2-6, Gd-7, or Gd-8 (1.00 mm) wasprepared by mixing volumes of the respective stock solutions anddiluting with H2O when necessary. An aliquot was taken from thissolution, which was mixed again with H2O and the stock solutionof 1 to keep the concentration of the dendrimer at 0.10 mm. TheR1

obs value for the resulting solution was measured, and this proce-dure was repeated several times until an R1

obs value very close tothat measured for the matrix (i.e. , a 0.10 mm aqueous solution of1) was reached. The curve obtained by plotting R1

obs versus themolar concentration of Gd2-6, Gd-7, or Gd-8 was fitted by usingappropriate equations.

Cleavage experiments

TCEP (final concentration = 1.25 mm, from a 0.50 mm stock solutionat pH 7) was added portionwise to an aqueous solution of Gd2-6(0.15 mm) and 1 (0.0375 mm) at pH 7. The reaction mixture wasstirred at 25 8C for 5 min before the measurement. The R1 datawere collected at 40 MHz and 25 8C.

MR-phantom imaging

MR images at 1 T were acquired on an Aspect MRI System (AspectMagnet Technologies Ltd. , Netanya, Israel), which consisted of anNdFeB magnet, equipped with a solenoid coils (inner diameter =35 mm). The images were obtained by using a standard T1-weight-ed spin-echo sequence with the following parameters: TR/TE/NEX250/8.9/10, FOV = 30 mm, matrix = 128 � 128, slice thickness =5 mm.

Acknowledgements

Gd-BzAAZTA (Gd-7) and Gd-DOTAMA-Ad (Gd-8) were kindlyprovided by Prof. Giovanni B. Giovenzana (Universit� del Pie-monte Orientale) and Dr. Carla Carrera (Universit� di Torino), re-spectively. We thank Dr. Francesco Marsano (Universit� del Pie-

Chem. Eur. J. 2014, 20, 1 – 10 www.chemeurj.org � 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim8&&

�� These are not the final page numbers!

Full Paper

monte Orientale) for the acquisition of the MALDI mass spec-tra. Phantom-imaging experiments were conducted by Dr.Eliana Gianolio at the Molecular Biotechnology Center (Univer-sit� di Torino). Financial support of Regione Piemonte (Italy)through the Nano-IGT and PIIMDMT Projects is gratefully ac-knowledged.

Keywords: cyclodextrin units · magnetic resonance imaging ·relaxivity · responsive systems · supramolecular chemistry

[1] P. Caravan, J. J. Ellison, T. J. McMurry, R. B. Lauffer, Chem. Rev. 1999, 99,2293 – 2352; S. Aime, M. Botta, E. Terreno, Adv. Inorg. Chem. 2005, 57,173.

[2] A. Merbach, L. Helm, �. T�th, The Chemistry of Contrast Agents in Medi-cal Magnetic Resonance Imaging, 2nd ed., John Wiley & Sons, New York,2013.

[3] M. Botta, L. Tei, Eur. J. Inorg. Chem. 2012, 1945 – 1960; A. J. L. Villaraza,A. Bumb, M. W. Brechbiel, Chem. Rev. 2010, 110, 2921 – 2959.

[4] D. A. Tomalia, Aldrichimica Acta 2004, 37, 39 – 57; D. A. Tomalia, H. Baker,J. Dewald, M. Hall, G. Kallos, S. Martin, J. Roeck, J. Ryder, P. Smith, Mac-romolecules 1986, 19, 2466 – 2468; D. A. Tomalia, R. Esfand, Chem. Ind.1997, 416 – 420.

[5] U. Boas, P. M. H. Heegaard, Chem. Soc. Rev. 2004, 33, 43 – 63; R. Duncan,L. Izzo, Adv. Drug Delivery Rev. 2005, 57, 2215 – 2237.

[6] N. Bourne, L. R. Stanberry, E. R. Kern, G. Holan, B. Matthews, D. I. Bern-stein, Antimicrob. Agents Chemother. 2000, 44, 2471 – 2474; J. C. Roberts,M. K. Bhalgat, R. T. Zera, J. Biomed. Mater. Res. 1996, 30, 53 – 65.

[7] E. C. Wiener, M. W. Brechbiel, H. Brothers, R. L. Magin, O. A. Gansow,D. A. Tomalia, P. C. Lauterbur, Magn. Reson. Med. 1994, 31, 1 – 8.

[8] S. D. Swanson, J. F. Kukowska-Latallo, A. K. Patri, C. Chen, S. Ge, Z. Cao,A. Kotlyar, A. T. East, J. R. Baker, Int. J. Nanomed. 2008, 3, 201 – 210; S.Langereis, A. Dirksen, T. M. Hackeng, M. H. P. van Genderen, E. W. Meijer,New J. Chem. 2007, 31, 1152 – 1160; A. R. Menjoge, R. M. Kannan, D. A.Tomalia, Drug Discovery Today 2010, 15, 171 – 185; H. Kobayashi, S. Ka-wamoto, S. K. Jo, H. L. Bryant, M. W. Brechbiel, R. A. Star, BioconjugateChem. 2003, 14, 388 – 394; H. Kobayashi, N. Sato, A. Hiraga, T. Saga, Y.Nakamoto, H. Ueda, J. Konishi, K. Togashi, M. W. Brechbiel, Magn. Reson.Med. 2001, 45, 454 – 460; C. Kojima, B. Turkbey, M. Ogawa, M. Bernardo,C. A. S. Regino, L. H. Bryant, Jr. , P. L. Choyke, K. Kono, H. Kobayashi,Nanomedicine 2011, 7, 1001 – 1008; J. Rudovsky, M. Botta, P. Hermann,K. I. Hardcastle, I. Lukes, S. Aime, Bioconjugate Chem. 2006, 17, 975 –987; M. M. Ali, M. Woods, P. Caravan, A. C. L. Opina, M. Spiller, J. C. Fet-tinger, A. D. Sherry, Chem. Eur. J. 2008, 14, 7250 – 7258; Z. J�szber�nyi, L.Moriggi, P. Schmidt, C. Weidensteiner, R. Kneuer, A. E. Merbach, L. Helm,E. Toth, J. Biol. Inorg. Chem. 2007, 12, 406 – 420; G. Gugliotta, M. Botta,L. Tei, Org. Biomol. Chem. 2010, 8, 4569 – 4574.

[9] L. H. Bryant, Jr. , M. W. Brechbiel, C. Wu, J. W. Bulte, V. Herynek, J. A.Frank, J. Mater. Chem. J. Magn. Reson. Imaging 1999, 9, 348 – 352.

[10] N. Bloembergen, J. Chem. Phys. 1957, 27, 572 – 573; N. Bloembergen,E. M. Purcell, R. V. Pound, Phys. Rev. 1948, 73, 679 – 712; I. Solomon,Phys. Rev. 1955, 99, 559 – 565.

[11] G. M. Nicolle, E. Toth, H. Schmitt-Willich, B. Raduchel, A. E. Merbach,Chem. Eur. J. 2002, 8, 1040 – 1048.

[12] G. Adam, J. Neuerburg, E. Spuntrup, A. Muhler, K. Scherer, R. W. Gunth-er, J. Mater. Chem. J. Magn. Reson. Imaging 1994, 4, 462 – 466.

[13] Q. Dong, D. R. Hurst, H. J. Weinmann, T. L. Chenevert, F. J. Londy, M. R.Prince, Invest. Radiol. 1998, 33, 699 – 708.

[14] F. N. Franano, W. B. Edwards, M. J. Welch, M. W. Brechbiel, O. A. Gansow,J. R. Duncan, Magn. Reson. Imaging 1995, 13, 201 – 214.

[15] M. V. Rekharsky, Y. Inoue, Chem. Rev. 1998, 98, 1875 – 1917.[16] E. Battistini, E. Gianolio, R. Gref, P. Couvreur, S. Fuzerova, M. Othman, S.

Aime, B. Badet, P. Durand, Chem. Eur. J. 2008, 14, 4551 – 4561; C. W. Lim,B. J. Ravoo, D. N. Reinhoudt, Chem. Commun. 2005, 5627 – 5629; G.Giammona, G. Cavallaro, L. Maniscalco, E. F. Craparo, G. Pitarresi, Eur.Polym. J. 2006, 42, 2715 – 2729; E. Gianolio, R. Napolitano, F. Fedeli, F.Arena, S. Aime, Chem. Commun. 2009, 6044 – 6046.

[17] K. Matsuoka, M. Terabatake, Y. Saito, C. Hagihara, Y. Esumi, D. Terunuma,H. Kuzuhara, Bull. Chem. Soc. Jpn. 1998, 71, 2709 – 2713; C. Kojima, Y.Toi, A. Harada, K. Kono, Bioconjugate Chem. 2008, 19, 2280 – 2284; J.Suh, S. S. Hah, S. H. Lee, Bioorg. Chem. 1997, 25, 63 – 75; H. Arima, F.Kihara, F. Hirayama, K. Uekama, Bioconjugate Chem. 2001, 12, 476 – 484;F. Kihara, H. Arima, T. Tsutsumi, F. Hirayama, K. Uekama, BioconjugateChem. 2002, 13, 1211 – 1219; F. Kihara, H. Arima, T. Tsutsumi, F. Hiraya-ma, K. Uekama, Bioconjugate Chem. 2003, 14, 342 – 350; T. Tsutsumi, F.Hirayama, K. Uekama, H. Arima, J. Controlled Release 2007, 119, 349 –359.

[18] M. E. Davis, J. E. Zuckerman, C. H. J. Choi, D. Seligson, A. Tolcher, C. A.Alabi, Y. Yen, J. D. Heidel, A. Ribas, Nature 2010, 464, 1067 – 1140.

[19] A. Barge, M. Caporaso, G. Cravotto, K. Martina, P. Tosco, S. Aime, C. Car-rera, E. Gianolio, G. Pariani, D. Corpillo, Chem. Eur. J. 2013, 19, 12086.

[20] J. Martinelli, M. Fekete, L. Tei, M. Botta, Chem. Commun. 2011, 47, 3144 –3146.

[21] C. S. Bonnet, E. Toth, Future Med. Chem. 2010, 2, 367 – 384; L. M.De Leon-Rodriguez, A. J. M. Lubag, C. R. Malloy, G. V. Martinez, R. J. Gil-lies, A. D. Sherry, Acc. Chem. Res. 2009, 42, 948 – 957; J. L. Major, T. J.Meade, Acc. Chem. Res. 2009, 42, 893 – 903; E. L. Que, C. J. Chang, Chem.Soc. Rev. 2010, 39, 51 – 60; B. Yoo, M. D. Pagel, Front. Biosci. 2008, 13,1733 – 1752.

[22] M. Calder�n, M. A. Quadir, M. Strumia, R. Haag, Biochimie 2010, 92,1242 – 1251; E. Fleige, M. A. Quadir, R. Haag, Adv. Drug Delivery Rev.2012, 64, 866 – 884; M. S. Shim, Y. J. Kwon, Adv. Drug Delivery Rev. 2012,64, 1046 – 1058; M. Ye, Y. Qian, J. Tang, H. Hu, M. Sui, Y. Shen, J. Control.Rel. 2013, 169, 239 – 245.

[23] J. E. Biaglow, M. E. Varnes, E. P. Clark, E. R. Epp, Radiat. Res. 1983, 95,437 – 455; S. E. Moriarty-Craige, D. P. Jones, Annu. Rev. Nutr. 2004, 24,481 – 509; W.-S. Wu, Cancer Metastasis Rev. 2007, 25, 695 – 705.

[24] C. Carrera, G. Digilio, S. Baroni, D. Burgio, S. Consol, F. Fedeli, D. Longo,A. Mortillaro, S. Aime, Dalton Trans. 2007, 4980 – 4987; G. Digilio, V. Cat-anzaro, F. Fedeli, E. Gianolio, V. Menchise, R. Napolitano, C. Gringeria, S.Aime, Chem. Commun. 2009, 893 – 895; N. Raghunand, G. P. Guntle, V.Gokhale, G. S. Nichol, E. A. Mash, B. Jagadish, J. Med. Chem. 2010, 53,6747 – 6757; N. Raghunand, B. Jagadish, T. P. Trouard, J. P. Galons, R. J.Gillies, E. A. Mash, Magn. Reson. Med. 2006, 55, 1272 – 1280; R. Cheng, F.Feng, F. Meng, C. Deng, J. Feijen, Z. Zhong, J. Controlled Release 2011,152, 2 – 12.

[25] K. Fujita, T. Ueda, T. Imoto, I. Tabushi, N. Toh, T. Koga, Bioorg. Chem.1982, 11, 72 – 84; G. Nelles, M. Weisser, R. Back, P. Wohlfart, G. Wenz, S.MittlerNeher, J. Am. Chem. Soc. 1996, 118, 5039 – 5046.

[26] F. Trotta, K. Martina, B. Robaldo, A. Barge, G. Cravotto, J. InclusionPhenom. Macrocyclic Chem. 2007, 57, 3 – 7.

[27] G. Digilio, V. Menchise, E. Gianolio, V. Catanzaro, C. Carrera, R. Napolita-no, F. Fedeli, S. Aime, J. Med. Chem. 2010, 53, 4877 – 4890.

[28] G. Gambino, S. De Pinto, L. Tei, C. Cassino, F. Arena, E. Gianolio, M.Botta, J. Biol. Inorg. Chem. 2014, 19, 133.

[29] S. Aime, L. Calabi, C. Cavallotti, E. Gianolio, G. B. Giovenzana, P. Losi, A.Maiocchi, G. Palmisano, M. Sisti, Inorg. Chem. 2004, 43, 7588 – 7590; G.Gugliotta, M. Botta, G. B. Giovenzana, L. Tei, Bioorg. Med. Chem. Lett.2009, 19, 3442 – 3444.

[30] Z. Baranyai, F. Uggeri, G. B. Giovenzana, A. Benyei, E. Bruecher, S. Aime,Chem. Eur. J. 2009, 15, 1696 – 1705.

[31] S. Aime, M. Botta, M. Fasano, S. Geninatti, S. Crich, E. Terreno, J. Biol.Inorg. Chem. 1996, 1, 312.

[32] M. Mammen, S. K. Choi, G. M. Whitesides, Angew. Chem. Int. Ed. 1998,37, 2754 – 2794.

[33] T. J. Peters, All About Albumin: Biochemistry Genetics, and Medical Appli-cations, Academic Press, San Diego, 1995.

[34] J. A. Burns, J. C. Butler, J. Moran, G. M. Whitesides, J. Org. Chem. 1991,56, 2648 – 2650.

[35] Y. Liu, N. Zhang, Biomaterials 2012, 33, 5363 – 5375.

Received: February 28, 2014

Published online on && &&, 0000

Chem. Eur. J. 2014, 20, 1 – 10 www.chemeurj.org � 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim9 &&

These are not the final page numbers! ��

Full Paper

FULL PAPER

& Supramolecular Chemistry

J. Martinelli, K. Thangavel, L. Tei,M. Botta*

&& –&&

Dendrimeric b-Cyclodextrin/GdIII

Chelate Supramolecular Host–GuestAdducts as High-Relaxivity MRI Probes

A leading edge : A new dendrimeric ar-chitecture bears eight b-cyclodextrinunits at the periphery linked througha reducible disulfide bond (see figure).The hosting capability of this dendrimertoward three different GdIII chelatesfunctionalized with hydrophobic groupswas investigated by 1H NMR relaxome-try. The stability of these systems underphysiological conditions and the cleav-age of disulfide bridges in a reducingenvironment were also tested.

Chem. Eur. J. 2014, 20, 1 – 10 www.chemeurj.org � 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim10&&

�� These are not the final page numbers!

Full Paper