Photochemical &Photobiological Sciences

PAPER

Cite this: DOI: 10.1039/c3pp50169h

Received 3rd June 2013,Accepted 11th July 2013

DOI: 10.1039/c3pp50169h

www.rsc.org/pps

Citric acid–γ-cyclodextrin crosslinked oligomers ascarriers for doxorubicin delivery†

Resmi Anand,a Milo Malanga,b Ilse Manet,a Francesco Manoli,a Kata Tuza,b

Ahmet Aykaç,c C. Ladavière,d Eva Fenyvesi,b Antonio Vargas-Berenguel,c

Ruxandra Grefe and Sandra Monti*a

Two citric acid crosslinked γ-cyclodextrin oligomers (pγ-CyD) with a MW of 21–33 kDa and 10–15 γ-CyDunits per molecule were prepared by following green chemistry methods and were fully characterized.

The non-covalent association of doxorubicin (DOX) with these macromolecules was investigated in

neutral aqueous medium by means of circular dichroism (CD), UV-vis absorption and fluorescence. Global

analysis of multiwavelength spectroscopic CD and fluorescence titration data, taking into account the

DOX monomer–dimer equilibrium, evidenced the formation of 1 : 1 and 1 : 2 pγ-CyD unit–DOX com-

plexes. The binding constants are 1–2 orders of magnitude higher than those obtained for γ-CyD and

depend on the characteristics of the oligomer batch used. The concentration profiles of the species in

solution evidence the progressive monomerization of DOX with increasing oligomer concentration. Con-

focal fluorescence imaging and spectral imaging showed a similar drug distribution within the MCF-7 cell

line incubated with either DOX complexed to pγ-CyD or free DOX. In both cases DOX is taken up into the

cell nucleus without any degradation.

Introduction

Doxorubicin (DOX, Scheme 1) is one of the most powerfulanthracycline anticancer drugs, largely employed in the treat-ment of leukaemia and various solid tumors, in spite of draw-backs associated with poor water solubility, severe

cardiotoxicity and emerging multidrug resistance.1 A strategyto face these problems has been to develop suitable carriers tooptimize the administration of the drug. In this frame a nano-particle-based formulation of DOX (PEGylated liposomalDoxil) has been successfully introduced in the market.2

Cyclodextrins (CyDs) are water soluble, biocompatible cyclicpolysaccharides made of α-D-glucopyranose units joinedtogether by α(1–4) linkages and possess hydrophobic cavitiesthat can easily include lipophilic molecules. The natural com-pounds α-, β- and γ-CyD have six, seven, and eight sugar units,respectively. They have received a lot of attention as pharma-ceutical excipients for solubilisation, enhancement of stabilityand bioavailability of drugs.3,4 CyD derivatives have been inves-tigated as potential carriers for DOX from the 1990s to thepresent. They are able to improve the activity of DOX on bothsensitive and multidrug-resistant cancer cell lines5 and topromote DOX release in brain tissues.6 Saccharide-conjugatedCyDs7 and various CyD-containing nanoassemblies haveshown great potential to improve the pharmacological profileof DOX.4,8

Crosslinked β-CyD polymers, spontaneously forming nano-particles in aqueous solutions, have been recently prepared.9

Compared to the natural CyDs these systems dramaticallyenhance the apparent solubility of several guests. An epichloro-hydrin crosslinked β-CyD polymer (pβ-CyD) of high molecularweight10 has been investigated as a DOX complexing agent.11

One of the goals of that study was to examine the ability of

Scheme 1 Structure of Doxorubicin (DOX).

†Electronic supplementary information (ESI) available. See DOI: 10.1039/c3pp50169h

aInstitute for Organic Synthesis and Photoreactivity, CNR, via P. Gobetti 101, 40129

Bologna, Italy. E-mail: sandra.monti@isof.cnr.itbCycloLab, Cyclodextrin R&D Ltd., H1097 Budapest, HungarycDepartamento de Química y Física, Universidad de Almería, Carretera de

Sacramento s/n, 04120 Almería, SpaindEcole Polytechnique Universitaire Lyon 1, UMR CNRS 5223/IMP, Villeurbanne

Cedex, FranceeUMR 8612 CNRS Université Paris-Sud, Châtenay-Malabry, France

This journal is © The Royal Society of Chemistry and Owner Societies 2013 Photochem. Photobiol. Sci.

Publ

ishe

d on

12

July

201

3. D

ownl

oade

d by

Ohi

o L

ink

Off

ices

on

27/0

8/20

13 0

7:21

:57.

View Article OnlineView Journal

pβ-CyD to regulate the DOX self-association tendency inaqueous medium. Indeed drug aggregation is clearly detrimen-tal to the pharmacological action, because the DOX monomerplays a major antitumoral role whereas DOX self-aggregatedspecies (dimers) have no efficacy as such.12 Although theaverage binding constant of DOX to the β-CyD units of thepolymer was much lower than the drug dimerization con-stant,13 the nanoparticle frame of the polymeric host proved tobe able to disrupt the free DOX dimers present in solution,converting them into polymer-bound monomers. However, theaverage stability of the DOX complexes within the pβ-CyDnanoparticles is rather low, likely because the natural β-CyDcavity is not large enough to effectively include the DOX mole-cule. Among the natural CyDs, only γ-CyD (Scheme 2) formsinclusion complexes of fair stability with DOX.14,15 We haverecently shown the formation of several complexes, with molarratios of 1 : 1, 2 : 1, 1 : 2 and 2 : 2 γ-CyD–DOX stoichiometry,and have determined the stability constants and the absoluteCD, UV-vis absorption and fluorescence spectra of all of them.It has also emerged that γ-CyD cannot disrupt the drug dimerwhen this species is the predominant form of the drug in solu-tion.13 Considering our results on the interaction of DOX withpβ-CyD and γ-CyD we envisaged that a crosslinked γ-CyDnetwork might promote monomerization of DOX through its3D spatial organization as well as effective binding of the drugin the large cavity of the γ-CyD units. So, using green chemistrymethods, we synthesized and characterized water soluble citricacid–γ-CyD crosslinked oligomers (pγ-CyD, see Scheme 2) withan MW of 21–33 kDa and 10–15 γ-CyD units per molecule.Polymeric citric acid–γ-CyD networks were first prepared inwater-insoluble form mostly grafted to textiles,16 and were con-sidered for drug delivery in vascular prosthesis and boneimplants.17 Some publications reported that the solubility ofalbendazol and doxycyclin was improved by inclusion in citricacid–CyD-copolymers.18 A very recent study compared variouscitric acid–CyD polymers with a molecular weight of50–100 kDa with their corresponding monomers, showing thatthe binding constants for nine selected drugs improved up to5 times.19 The oligomers synthesized in this work are expectedto possess intermediate properties between the monomersand the polymers and may have several advantages withrespect to their higher MW polymeric analogues. Besideshigher solubility and lower viscosity, they can be more easilycleared from the body, after drug release, by renal excretion,whereas polymers with a higher molecular weight need todegrade before removal (the MW threshold for renal excretion

of polymeric drugs ranges from 30 to 50 kDa depending onthe polymer structure).20

We have studied the interaction of two different citric acid–γ-CyD oligomers with DOX in neutral TRIS buffer solution, thedrug concentration ranging from 1 × 10−5 to 2 × 10−4 M. Bymeans of CD, UV-Vis absorption and fluorescence spec-troscopy we have found that very stable 1 : 1 and 1 : 2 pγ-CyD_unit–DOX complexes are promptly formed upon addition ofthe oligomeric material at low concentration (<10−5 M) and aprogressive monomerization of the drug is produced on increas-ing the oligomer concentration. We have determined the stabi-lity constants and the individual spectroscopic properties of allthe complexes and compared them with the corresponding pro-perties of the γ-CyD–DOX and pβ-CyD–DOX analogues. We havealso examined the uptake and biodistribution of DOX, incu-bated with and without pγ-CyD, in the MCF-7 tumor cell line bytime-resolved confocal fluorescence microscopy. This studyhighlights the potential of these water soluble citric acid–γ-CyDoligomers as carriers for DOX delivery.

2. Experimental2.1 Materials

Doxorubicin (DOX, Adriamycin), from ALEXIS Biochemicals,was used as received. Water was purified by passage through aMillipore MilliQ system. 0.01 M TRIS buffer at pH 7.4 was used.Two pγ-CyD samples (named batch 1 and batch 2 in the follow-ing) with different molecular weights and different γ-CyD con-tents were used (see below for their synthesis andcharacterization). Citric acid (≥99.5%) and sodium phosphatemonobasic dihydrate (≥99%) were from Sigma-Aldrich; γ-cyclo-dextrin and partially acetylated γ-CyD were products of CycloLab.

2.2 Synthesis and characterization of pγ-CyD oligomers

The synthetic procedure of Martel et al. was adopted usingγ-CyD and citric acid in 1 : 3 and 1 : 5 molar ratios, respect-ively.21 To the yellowish material obtained after heating at140 °C for 30 min, distilled water was added leading to insolu-ble gel-like structure immediately getting formed. The crudewas filtered to separate the insoluble fraction from the solubleone. The water phase was neutralized and then dialyzed over-night (12 kDa cellulosic membrane). The solution was finallydried under reduced pressure to yield a slightly yellow powder.

The product was analysed by FTIR and NMR recorded in aKBr disk using a Nicolet 205 FTIR and in D2O using a Varian

Scheme 2 Synthesis of citric acid–γ-CyD crosslinked oligomer (citric acid pγ-CyD) via the reaction of citric acid molecules with γ-CyD cyclic polysaccharides.

Paper Photochemical & Photobiological Sciences

Photochem. Photobiol. Sci. This journal is © The Royal Society of Chemistry and Owner Societies 2013

Publ

ishe

d on

12

July

201

3. D

ownl

oade

d by

Ohi

o L

ink

Off

ices

on

27/0

8/20

13 0

7:21

:57.

View Article Online

VXR-600 at 400 or 600 MHz. Size-exclusion chromatography(SEC), employing multiangle light scattering (MALS) andrefractive index (RI) detectors (SEC/MALS/RI, Wyatt Techno-logy, Santa Barbara, USA), of pγ-CyD oligomers providedweight average molecular weight values of ca. 21 000 (batch 2)and ca. 33 250 g mol−1 (batch 1) (Ð = 1.4 and 1.6, respectively)using dn/dc of dextran = 0.147 mL g−1 with TSK6000PW/TSK2500PW columns, and ammonium acetate–acetic acid pH4.5 as the eluent at a flow rate of 0.5 mL min−1.

2.3 Spectroscopic measurements

UV-Vis absorption spectra were recorded using a Perkin-ElmerLambda 650 spectrophotometer. Samples were measured inquartz cuvettes with 0.2 cm path length, with the pure solvent(TRIS buffer, 0.01 M, pH 7.4) as the reference. Circular dichro-ism (CD) spectra were obtained using a Jasco J-715 dichro-graph. The titration experiments were carried out at a constantDOX concentration and varying pγ-CyD concentration. Theellipticity signal was recorded for solutions in 0.2 cm quartzcuvettes accumulating three scans with a speed of 50 nmmin−1. The signal of the medium (TRIS buffer 0.01 M, pH 7.4,for DOX alone and pγ-CyD solutions of the suited concen-tration for the mixtures) was subtracted. Steady state fluo-rescence spectra of air-equilibrated solutions contained in acell with a triangular section were registered with excitation at530 nm (isosbestic region) on a SPEX Fluorolog 111 spectro-fluorimeter. The excitation beam was incident at 45° onto the“diagonal” cell surface and the emission was collected at rightangle; suitable filters were used to cut the excitation light. Thefluorescence spectra have been corrected for the wavelengthdependent response of the detection system. Fluorescencequantum yields were measured assuming a fluorescencequantum yield of 0.039 for a 1 × 10−5 M DOX solution inneutral buffer at 22 °C.13 Fluorescence lifetimes weremeasured in air-equilibrated solutions with a time correlatedsingle photon counting system (IBH Consultants Ltd). A nano-second LED source at 465 nm was used for excitation and theemission was collected at right angle at 590 nm.

The best complexation model and the association constantswere determined by multivariate global analysis of multiwave-length CD or fluorescence data, analyzing a set of spectracorresponding to different host–guest mixtures and using theprogram SPECFIT/32 (v. 3.0.40, TgK Scientific). The procedureis based on singular value decomposition (SVD) and non-linear regression modelling by the Levenberg–Marquardtmethod. The analysis also afforded the individual CD spectraof the complexes (see ESI† for more details).

For the pγ-CyD characterization, 1H, HSQC, DEPT NMR,13C-NMR and APT spectra were recorded using a VarianVXR-600 at 400 or 600 MHz. IR spectra were recorded in a KBrdisk using a Nicolet 205 FTIR.

2.4 Confocal fluorescence microscopy

Confocal fluorescence imaging was performed using aninverted Nikon A1 laser scanning confocal microscopeequipped with a CW argon ion laser for excitation at 457, 488

and 514 nm (Melles Griot, 40 mW), and a diode laser for exci-tation at 405 nm (LDH-D-C-405 of Picoquant GmbH Berlin,Germany) operating both in continuous mode (50 mW) andpulsed at 40 MHz (1.0 mW average power for pulse FWHM of70 ps). Confocal fluorescence imaging was carried out on thesamples at room temperature. The images were collected usinga Nikon PLAN APO VC 60× oil immersion objective with NA1.40. Images of 512 × 512 or 1024 × 1024 pixels have beenacquired applying a scan speed of 1 frame in 2–8 s and thepixel dimension of the xy plane falls in the range of0.1–0.4 μm. The hexagonal pinhole dimension was set to0.8–1.0 au corresponding to 25–38 μm and an optical thick-ness of 330–440 nm. Two dichroic mirrors reflecting 405, 488,541 and 640 nm or 457 and 514 nm were used. Bandpassfilters in front of the PMT selected fluorescence in the ranges425–475 nm, 500–550 nm and 570–620 nm. Spectral imagingwas done with a Nikon 32-PMT array detector with resolutionvarying from 6 to 10 nm per channel and a 20/80 beam splitterin the scan head instead of a dichroic mirror. For fluorescencelifetime imaging a time-correlated single photon counting(TCSPC) system of Picoquant GmbH Berlin was used excitingat 405 nm. Photons were detected in time tagged time resolved(TTTR) mode with two Single Photon Avalanche Diodes manu-factured by Micro Photon Devices (MPD), Bolzano, Italy. Fluo-rescence was filtered with the opportune fluorescenceSEMROCK bandpass filters 520/40 nm and 585/40 nm. ThePicoHarp 300 photon processor completes the TCSPC system.SymPhoTime v. 5.1 analysis software was used for image pro-cessing and lifetime fitting. A tail fit with multi-exponentialfunctions was performed to analyze fluorescence decays of theselected region of interest (ROI). The system allowed measure-ment of fluorescence lifetimes from 300 ps up to severalnanoseconds.

3. Results and discussion3.1 Synthesis and characterization of pγ-CyD

Two batches of pγ-CyDs were synthesized according to the pro-cedure shown in Scheme 2.21 The amount of the catalyst(NaH2PO4·2H2O) was kept constant in both experiments whilethe citric acid/γ-CyD molar ratio was set to 3 or 5 in order toobtain oligomers with different cyclodextrin contents andmolecular weights; the work-up was performed in a similarmanner for both procedures. Batch 1 with a molar ratio of5 has an average molecular weight of 33 kDa and a γ-CyDcontent of 60% wt; batch 2 with a molar ratio of 3 has anaverage molecular weight of 21 kDa and a γ-CyD content of40% wt. Completely self-consistent results were obtained fromnumerical analysis of titration experiments with both oligomerbatches when the host concentrations were expressed in termsof the average concentration of γ-CyD units. In the followingthe detailed spectroscopic characterization of batch 1 isdescribed.

3.1.1 pγ-CyD characterization with FT-IR spectroscopy. InFig. 1, the spectra of native γ-CyD, citric acid and pγ-CyDs are

Photochemical & Photobiological Sciences Paper

This journal is © The Royal Society of Chemistry and Owner Societies 2013 Photochem. Photobiol. Sci.

Publ

ishe

d on

12

July

201

3. D

ownl

oade

d by

Ohi

o L

ink

Off

ices

on

27/0

8/20

13 0

7:21

:57.

View Article Online

compared. An intensive absorption band appeared at1736 cm−1 which was not present in the spectra of the startingmaterials. Indeed the peak at 1736 cm−1 is a typical absorptionfor the carbonyl stretching of an ester moiety. The absorptionbands of ester groups indicated that the hydroxyl groups ofγ-CyD had reacted with the carboxyl groups of citric acid andthereby a three-dimensional network was formed.

3.1.2 pγ-CyD characterization with NMR spectroscopy. Thestructure of both oligomers was analyzed by NMR spectro-scopic techniques with DEPT and HSCQ experiments. A typical1H-NMR spectrum of citric acid cross-linked γ-CyD oligomers(see Fig. 2) shows two sets of signals between 3.25–5.5 ppmand 2.5–3.00 ppm indicating the presence of the γ-CyD andcitric acid moieties, respectively, in the oligomer structure.

The broad signal from 4.8 to 5.2 ppm corresponding to theCyD anomeric protons (8 protons per γ-CyD unit) in the1H-NMR spectrum confirmed the loss of the macrocycle C8-symmetry as a result of the partial acylation of it. Likewise, the13C NMR spectrum shows a broad signal at 101.5 ppmassigned to the anomeric carbons of a non-symmetrical γ-CyDring. The resonances of the citric acid methylene protonsappear as two broad signals between 2.5 and 3 ppm. As shownby the HSQC spectrum (Fig. 3), such methylene protons corre-late with a carbon signal at 43.2 ppm that corresponds to theresonance of the citric acid methylene carbons, as was con-firmed by DEPT experiments (Fig. 4). In addition to the unsub-stituted CyD C-6 signal at 60.1 ppm, further methylene carbonsignals appear in the DEPT spectrum at 64–65 ppm which

were assigned to the ester-C-6. A comparison of the 13C NMRspectra of the oligomers with the APT spectrum of a partiallyacetylated γ-CyD (Fig. 5) (aware of the fact that the peak at20 ppm belongs to the methyl moiety of the acetyl group) sup-ports the assignment of the ester-C-6 at more downfield reson-ance than that of the unsubstituted C-6. The NMR spectraprove that the CD units in the polymer are mainly substitutedat the C-6 position.

The citric acid/γ-CyD molar ratio in the products could thenbe estimated by integrating the peaks assigned to 4 protons ofthe citric acid methylene group divided by 8 anomeric protonsof γ-CyD units (Fig. 2). The CyD content in both oligomersamples was finally determined by dividing the molecularweight of the γ-CyD by the molecular weight of the repeatingunit calculated according to the molar citric acid/γ-CyD ratio.The values of 60% and 40% were obtained for batches 1 and 2,respectively. The degree of substitution (DS) calculated fromthese data is 4–5 and 10–11 for batches 1 and 2, respectively.

3.2 Association of DOX molecules with pγ-CyD

3.2.1 DOX–pγ-CyD association examined by means ofUV-Vis absorption and circular dichroism. The UV-visibleabsorption spectra of 1.6 × 10−4 M DOX alone and in the pres-ence of increasing concentration of pγ-CyD (batch1) in TRISbuffer (0.01 M at pH 7.4) at 22 °C are shown in Fig. 6a. pγ-CyDdoes not show any characteristic absorption band in the visibleregion (grey dashed lines in Fig. 6a). The rather intense tailobserved at high oligomer concentrations (pγ-CyD > 10−3 M)

Fig. 1 IR spectra; (a) γ-CyD, (b) citric acid and (c) pγ-CyD powder.

Paper Photochemical & Photobiological Sciences

Photochem. Photobiol. Sci. This journal is © The Royal Society of Chemistry and Owner Societies 2013

Publ

ishe

d on

12

July

201

3. D

ownl

oade

d by

Ohi

o L

ink

Off

ices

on

27/0

8/20

13 0

7:21

:57.

View Article Online

is likely to be due to a contribution of light scattering. Theabsorption maximum of DOX (black solid line in Fig. 6a) at480–500 nm has been assigned to the allowed 1A → 1Lb π−π*transition polarized along the long axis; the peak at 288 nmcorresponds to the allowed 1A → 1La π−π* transition polarizedalong the short axis, while the shoulder around 320–380 nm isrelevant to n−π* transitions of the three CvO groups in themolecule, partially allowed by electric dipole.22,23 In aqueoussolution, self-aggregation of DOX affects the shape of theabsorption spectrum and, in particular, decreases the appar-ent molar absorption coefficients of the visible band. At con-centrations higher than ca. 5 × 10−3 M, DOX tends to formhigh order aggregates. Below this limit a monomer–dimerequilibrium applies reasonably well.24 We have previouslydetermined a dimerization constant with log (Kd/M

−1) =4.8 ± 0.1 in buffer at pH 7.4 and 22 °C.13 At 1.6 × 10−4 M

concentration, DOX is ca. 81 mol% in dimeric form. Uponaddition of pγ-CyD (colored solid lines in Fig. 6a) at low con-centration (<4.8 × 10−5 M) the absorption band at 480 nmshows a decrease in intensity with a red shift in the maximum(∼10 nm, orange and cyan curves). An increase in the oligomerconcentration causes a progressive increment of the absorp-tion intensity accompanied by a blue shift of the whole band.Further insight into the association of DOX with pγ-CyD wasgained from circular dichroism analysis (Fig. 6b). The intrinsicchirality of the DOX molecule is due to several asymmetriccarbon centres. The ellipticity of a DOX solution (1.6 × 10−4 M)in neutral TRIS buffer exhibits a negative band at 293 nm, apositive band at 352 nm and a positive–negative splitting inthe visible band (positive component at ca. 460 nm andweaker negative component in the 510–540 nm region). Thelatter feature indicates the presence of dimers in solution.22

Fig. 2 1H-NMR spectrum of pγ-CyD. Red lines are for the integrals.

Fig. 3 Hetero Nuclear Single Quantum Coherence (HSCQ) NMR spectrum of pγ-CyD (300 MHz, 25 °C, D2O).

Photochemical & Photobiological Sciences Paper

This journal is © The Royal Society of Chemistry and Owner Societies 2013 Photochem. Photobiol. Sci.

Publ

ishe

d on

12

July

201

3. D

ownl

oade

d by

Ohi

o L

ink

Off

ices

on

27/0

8/20

13 0

7:21

:57.

View Article Online

Indeed no negative component in the visible region is presentat concentrations of DOX < 1.0 × 10−5 M, in agreement withthe predominance of the monomer. The ellipticity of the DOXsolution modifies upon titration with pγ-CyD in the rangefrom 6.1 × 10−6 M to 3.6 × 10−3 M (Fig. 6b). At low oligomerconcentration (orange and cyan curves) the spectral profileappears to be very different from that of DOX alone in both theUV and the visible region: the positive band at 460 nmincreases in intensity and shifts to the red of ca. 6 nm; below380 nm a negative band with a minimum at 310 nm appears,instead of the positive one at 352 nm observed in the absenceof the oligomer. At oligomer concentrations ≥ 4.8 × 10−5 M thepositive band in the visible region tends to decrease; furthershifting to the red and the dimer negative CD component at510–540 nm tends to disappear; below 380 nm a reshaping ofthe signal toward the initial profile occurs. Along the titrationtwo quasi-isoelliptic points are maintained at 493 nm and365 nm. These features reasonably indicate the formation ofDOX multimolecular aggregates associated with pγ-CyD at low

oligomer concentration and the establishment of an equili-brium of different DOX complexes (likely with lower stoichio-metries), at higher oligomer concentrations (see below).

3.2.2 DOX–pγ-CyD association examined by means of fluo-rescence. Fig. 7 shows the fluorescence spectra of 1 × 10−5 MDOX in the absence and presence of different concentrationsof pγ-CyD oligomers (batch 2, from 3.03 × 10−6 M up to9.3 × 10−4 M), with excitation at 530 nm. At this DOX concen-tration, in the absence of the oligomer, 42% of DOX is presentin dimeric form. According to a recent study of the timeresolved fluorescence of DOX on the ps–fs time scale, the DOXdimer has an ultrashort lifetime and a very low emissionquantum yield and does not practically contribute to thesteady state fluorescence intensity.25 Accordingly the fluo-rescence spectrum shows the characteristic features of theDOX monomer with vibronic bands at 560, 590 and 630 nm.26

Addition of pγ-CyD at the lowest concentration drasticallydecreases the fluorescence intensity (Fig. 7a). This can beexplained by the formation of multimolecular aggregatedspecies of DOX associated with pγ-CyD. Further increase in thepγ-CyD concentration leads to progressive increment of theDOX fluorescence intensity, reasonably associated with mono-merization of the drug with an increase of the overall emissionquantum yield. Inclusion in the pγ-CyD_unit and/or in the 3Dframe of the oligomer network can be hypothesized.

The decay of DOX fluorescence on the nanosecond timescale was monitored at 590 nm with excitation at 465 nm. In aDOX solution without oligomer the lifetime, assigned to theDOX monomer, was found to be 1.02 ns ( χ2 = 1.03). Additionof pγ-CyD at various concentrations tends to slow down a littlethe decay rate but a monoexponential decay law still applies.In the presence of 7 × 10−3 M pγ-CyD_unit the measured life-time is τ = 1.3 ns ( χ2 = 1.0), assigned to the complexedmonomer (see Table 2).

3.2.3 DOX–pγ-CyD association examined by mathematicalglobal analysis of spectroscopic data. Global analysis based onthe SPECFIT/32 program was performed on CD titration data

Fig. 4 Distortionless Enhancement by Polarization Transfer (DEPT) spectrum of pγ-CyD (300 MHz, 25 °C, D2O).

Fig. 5 Attached-Proton-Test (APT) spectrum of partially acetylated-γ-CyD(300 MHz, 25 °C, D2O).

Paper Photochemical & Photobiological Sciences

Photochem. Photobiol. Sci. This journal is © The Royal Society of Chemistry and Owner Societies 2013

Publ

ishe

d on

12

July

201

3. D

ownl

oade

d by

Ohi

o L

ink

Off

ices

on

27/0

8/20

13 0

7:21

:57.

View Article Online

of Fig. 6b (see ESI† for more information). The pγ-CyD titrantconcentration was expressed in terms of average concentrationof γ-CyD unit (indicated in the following as “pγ-CyD_unit”); weexcluded from analysis the spectra with a molar excess ofpγ-CyD_unit vs. DOX below 4. We included the DOX dimeriza-tion equilibrium with a fixed constant, log (Kd/M

−1) = 4.8, andfixed the absolute CD spectra of the free DOX monomer anddimer species, as determined in neutral aqueous buffer.13 Thebest fit corresponds to a complexation model with 1 : 1 and1 : 2 pγ-CyD_unit–DOX complexes in equilibrium withDOX monomer, DOX dimer and pγ-CyD_unit. The optimizedassociation constants are log (K11/M

−1) = 3.3 ± 0.2 and log(K12/M

−2) = 8.7 ± 0.3 (Durbin–Watson factor = 1.6). The individ-ual CD spectra of the complexes are reported in Fig. 8atogether with the spectra of the free DOX components. Theconcentration profiles of all the DOX species in solution arereported in Fig. 8b. In Fig. 9a the agreement between calcu-lated and experimental ellipticity at key wavelengths can beappreciated.

A confirmation of the binding model was obtained fromglobal analysis of the fluorescence titration data in Fig. 7a. Weassumed that emission comes exclusively from monomericDOX, either free or complexed, and again included the DOXdimerization equilibrium with a fixed constant log Kd = 4.8.13

In the explored concentration range the molar excess of pγ-Cy-D_unit over DOX is >3. The optimized association constantsare log (K11/M

−1) = 4.4 ± 0.1 and log (K12/M−2) = 11.1 ± 0.1 (DW

1.1). The corresponding individual emission amplitudes forthe free DOX monomer and the 1 : 1 pγ-CyD_unit–DOXcomplex are shown in Fig. 7b. The quantitative discrepancybetween binding constants from CD and fluorescence data isdiscussed in the following section. In Table 1 we collect theassociation constants of DOX species with various CyD hosts,from this work and the existing literature.

3.2.4 The DOX–pγ-CyD binding process and the spectro-scopic properties of the complexes. With pγ-CyD_unit/DOXmolar ratios <4 we evidenced a massive association of thehydrophobic drug with the oligomeric material to form

Fig. 7 Titration of a 1 × 10−5 M DOX solution in 0.01 M TRIS buffer, pH 7.4, with pγ-CyD in the concentration range 3 × 10−5 to 9.3 × 10−3 M (expressed in γ-CyDunits). (a) Corrected fluorescence spectra for excitation at 530 nm; (b) molar amplitude of emission for DOX monomeric species extracted by global analysis with log(K11/M

−1) = 4.4, log (K12/M−2) = 11.1; the 1 : 2 complex was assumed to be not emissive.

Fig. 6 Titration of a 1.6 × 10−4 M DOX solution in 0.01 M TRIS buffer, pH 7.4, with pγ-CyD from 8.8 × 10−5 M to 5.3 × 10−2 M (concentration expressed in γ-CyDunits). Cell path length 0.2 cm, T = 295 K. (a) UV-Vis absorption spectra of pγ-CyD solutions (grey dashed lines), a DOX solution (black solid line) and DOX–pγ-CyDmixtures (coloured solid lines); (b) CD spectra of a DOX solution (black solid line) and DOX–pγ-CyD mixtures (coloured solid lines, the same color code as in (a)). Thespectra of the mixtures were registered with the pγ-CyD solutions in TRIS buffer as the reference, whereas the spectra of the DOX alone and pγ-CyD alone sampleswere taken with TRIS buffer as the reference.

Photochemical & Photobiological Sciences Paper

This journal is © The Royal Society of Chemistry and Owner Societies 2013 Photochem. Photobiol. Sci.

Publ

ishe

d on

12

July

201

3. D

ownl

oade

d by

Ohi

o L

ink

Off

ices

on

27/0

8/20

13 0

7:21

:57.

View Article Online

multimolecular, possibly non-specific, aggregates character-ized by “distorted” CD spectra (Fig. 6b) and no or very weakfluorescence (Fig. 7a). We did not include such data in ourglobal analysis, which was limited to a titration range corres-ponding to higher pγ-CyD_unit/DOX molar ratios. A relatively

simple equilibrium model was found to apply fairly well, withthe formation of pγ-CyD_unit–DOX complexes of 1 : 1 and 1 : 2stoichiometry, in equilibrium with free DOX monomer anddimer. The concentration profiles of all DOX species in solu-tion presented in Fig. 8b indicate that DOX associates very

Fig. 8 (a) CD absolute spectra of pγ-CyD_unit–DOX complexes corresponding to log (K11/M−1) = 3.3 and log (K12/M

−2) = 8.7 and of free DOX species taken fromref. 11, and (b) concentration profiles of DOX species in solution. Inset: zoom on the pγ-CyD_unit concentration range below 0.01 M.

Fig. 9 (a) Comparison of experimental (symbols) and calculated (line) data at representative wavelengths, corresponding to log (K11/M−1) = 3.3 and log (K12/M

−2) =8.7. (b) Comparison of spectra of pγ-CyD_unit : DOX complexes (this work) with spectra of γ-CyD : DOX complexes taken from ref. 13.

Table 1 Association constants of DOX–CyD complexes in aqueous media at pH 7.4, at 22 °C, determined by global analysis of titration spectroscopic data with theSpecfit/32 program, and relevant literature data

DOX complexes Log K Units Technique Note

DOX2 (DOX dimerization) 4.8 ± 0.1 M−1 Abs Ref. 13pγ-CyD_unit–DOX 1 : 1 (3.3 and 3.4)a ± 0.2 M−1 CD This work

(4.0 and 4.4)a ± 0.1 FLpγ-CyD_unit–DOX 1 : 2 (8.5 and 8.7)a ± 0.3 M−2 CD This work

(10.4 and 11.1)a ± 0.1 FLpβ-CyD_unit–DOX 1 : 1 2.2 ± 0.1 M−1 CD Ref. 11β-CD–DOX 1 : 1 2.1 ± 0.1 M−1 Abs Ref. 15

2.3 ± 0.1 FLγ-CyD –DOX 1 : 1 2.7 ± 0.2 M−1 CD Ref. 13

2.3 ± 0.3 FLγ-CyD–DOX 1 : 2 7.80 ± 0.04 M−2 CD Ref. 13γ-CyD–DOX 2 : 1 4.4 ± 0.5 M−2 CD Ref. 13

4.9 ± 0.1 FL4.9 ± 0.4 Abs

γ-CyD–DOX 2 : 2 10.48 ± 0.21 M−3 CD Ref. 13

a Values relevant to the two batches (1 and 2, respectively) of the pγ-CyD host material.

Paper Photochemical & Photobiological Sciences

Photochem. Photobiol. Sci. This journal is © The Royal Society of Chemistry and Owner Societies 2013

Publ

ishe

d on

12

July

201

3. D

ownl

oade

d by

Ohi

o L

ink

Off

ices

on

27/0

8/20

13 0

7:21

:57.

View Article Online

efficiently with pγ-CyD both as monomer (1 : 1 complex) andas dimer (1 : 2 complex). With increasing host concentrations,a progressive decrease of the 1 : 2 complex concentrationoccurs, the 1 : 1 complex becoming more and more favouredand finally becoming largely dominant (ca. 71% of DOX isassociated as monomer and ca. 29% as dimer with pγ-CyD 3.6× 10−3 M ([γ-CyD_unit] = 5.3 × 10−2 M)). The locally highγ-CyD_unit concentration and the oligomer 3D organizationlikely play a role in this effect, as already observed in the epi-chlorohydrin crosslinked β-CyD polymer (pβ-CyD), where thethree-dimensional nanoparticle structure proved to be able todisrupt the free DOX dimers present in solution convertingthem into polymer-bound monomers. Note that of the naturalCyDs, only γ-CyD forms inclusion complexes of significantstability with DOX. Consistently, a comparison of the associ-ation constants of the two polymeric hosts for monomericDOX (Table 1) shows that the binding ability of pγ-CyD islarger than that of the pβ-CyD polymer, reasonably due to theinclusion of the bulky aglycone nucleus of DOX in the largercavity. Further, the electrostatic interactions between the posi-tively charged DOX daunosamine moiety and the negativelycharged carboxyl groups of the citric acid crosslinker mayeffectively contribute to the complex stability in pγ-CyD. Aninformative comparison is also that with the association con-stants of the DOX complexes with natural γ-CyD. Clearly, theoligomeric structure considerably improves the binding abilityof the pγ-CyD_unit for both DOX monomer and dimer. Finally,for the pγ-CyD system it is worth comparing the associationconstants determined from CD and from fluorescence data.For each batch the values obtained with the latter techniqueare significantly higher. This result may be related to the lowerDOX concentration used in the fluorescence titration (10−5 Mvs. 1.6 × 10−4 M for CD experiments), resulting in a highermonomer/dimer molar ratio in solution (19/81 in CD vs. 58/42in fluorescence experiments). It may be reasonably hypoth-esized that the initial formation of multimolecular oligomer–DOX aggregates observed at low oligomer concentrationsmakes the two experimental conditions not completely equi-valent for the complexation equilibria to be further estab-lished. Unfortunately, the 1.6 × 10−4 M DOX concentrationcannot be employed in fluorescence experiments because oftoo high absorbance at the excitation wavelength.

The CD spectra of the pγ-CyD–DOX complexes are com-pared to those of the free species and the complexes withnatural γ-CyD13 in Fig. 8a and 9b, respectively. The CD of the1 : 2 pγ-CyD_unit–DOX complex is qualitatively very similar tothat of the 1 : 2 γ-CyD–DOX complex in the visible region andboth exhibit common structure features at 265, 287 and306 nm. The CD spectrum of the 1 : 1 pγ-CyD_unit–DOXcomplex is somewhat different from that of the DOX monomerin buffer and from that of the 1 : 1 γ-CyD–DOX complex,whereas it is practically identical to the CD spectrum of the2 : 1 γ-CyD–DOX complex. This suggests that DOX in the γ-CyDoligomer is effectively protected from contact with bulk buffer,similarly to the 2 : 1 γ-CyD–DOX species, which, on the basis ofMolecular Dynamics, was suggested to have both the aglycone

and the daunosamine moieties interacting with γ-CyD hostunits.13 It can be reasonably concluded that the 3D organiz-ation of the pγ-CyD oligomer provides a non-aqueous characterto the surroundings of most of the DOX units. The fluo-rescence molar amplitudes (Fig. 7b) give direct information onthe environment experienced by DOX in the 1 : 1 complex. The560 nm/595 nm intensity ratio probes the polarity of theenvironment of the emitting state.27 The value decreases con-siderably from buffer (0.62) to the 1 : 1 complex (0.44), indicat-ing that the excited state of the drug in the oligomer frameexperiences a polarity similar to that of alcoholic media. As forthe γ-CyD–DOX 1 : 1 complex, the proximity of the dihydroxy-anthraquinone chromophore to the hydroxyl groups of theγ-CyD rim is inferred. As regards the emissive ability, the areaunder the molar amplitude profiles in Fig. 7b is proportionalto the emission quantum yield of the free (Φf ) and complexed(Φb) DOX monomer. Free monomer and 1 : 1 complex appearto have quite similar emission quantum yields. Actually, usingthe value of Φ = 0.039 for a 1 × 10−5 M DOX solution in neutralbuffer at 22 °C (see the Materials and methods) and consider-ing (i) the fraction of excitation light absorbed by themonomer and the dimer in buffer 13 and the negligible contri-bution of the dimer to the steady state emission,25 we calcu-lated a value Φf = 0.058 for the free DOX monomer and Φb =0.054 for the 1 : 1 pγ-CyD_unit–DOX complex. In Table 2 thephotophysical parameters of DOX species in various CyDmedia are summarized. We observe that neither kr nor knr sub-stantially changes in any of the γ-CyD complexes when com-pared to buffer (kr ≈ 3–4 × 107 s−1 and knr ≈ 7–10 × 108 s−1).

3.3 Uptake and distribution of the DOX–pγ-CyD conjugatewithin tumour cells

Cancer cell uptake of DOX–pγ-CyD complexes has beenstudied with time-resolved fluorescence confocal microscopy.The autofluorescence of DOX was used for this scope. Fig. 10shows an intensity-based image for the MCF-7 cell line incu-bated for 1 hour with 10 μM DOX alone and with pγ-CyD(1.4 mg mL−1, 98% DOX is associated with the oligomer) inbuffer. In the images obtained by collecting fluorescencein the 500–550 nm range we can discern the entire cell, whilein the orange and red images we only observe the nucleus.The confocal spectra (Fig. 10) obtained in correspondence to

Table 2 Photophysical parameters

Samples τf/10−9 s Φf kr/10

7 s−1 knr/108 s−1 Note

1 × 10−5 M DOX inneutral buffer

1.0a 0.039 3.9 9.6 Ref. 13

γ-CyD–DOX1:1 complex

∼1.1a 0.032 ∼2.9 ∼8.8 Ref. 13

pβ-CyD_unit–DOX1 : 1 complex

∼1.5b 0.13 ∼8.7 ∼5.8 Ref. 11

pγ-CyD_unit–DOX1 : 1 complex

∼1.3a 0.054 ∼4.1 ∼7.3 This work

aMono-exponential decay. b Second lifetime component in abiexponential decay.

Photochemical & Photobiological Sciences Paper

This journal is © The Royal Society of Chemistry and Owner Societies 2013 Photochem. Photobiol. Sci.

Publ

ishe

d on

12

July

201

3. D

ownl

oade

d by

Ohi

o L

ink

Off

ices

on

27/0

8/20

13 0

7:21

:57.

View Article Online

ROI located in the nucleus clearly evidence the presence ofDOX in the nucleus, while in the cytoplasm autofluorescencemixes with DOX fluorescence. It can be concluded that underboth conditions DOX enters the cell nuclei, where the spectral

features of the drug emission are clearly recognizable (Fig. S1.A†).26 This fact indicates preservation of the DOX molecularstructure during both the uptake and the monitoring pro-cesses, irrespective of the incubation conditions.

Fig. 10 Confocal images of MCF-7 cells incubated for 1 h with (A) free DOX (10 μM) and (B) DOX (10 μM) + pγ-CyD (1.4 mg mL−1); (top left) overlay of the imagescollecting fluorescence in the ranges 500–550, 570–620 and 665–735. (C) Fluorescence spectra of DOX and pγ-CyD loaded DOX in the MCF-7 cell line for ROI in thenucleus and free DOX in water, resolution of 6 nm per channel, DC2 used reflecting 488 nm. (D) and (E) show 4-exponential fitted decay and frequency histogramsof the four lifetimes for the entire image in A and B.

Paper Photochemical & Photobiological Sciences

Photochem. Photobiol. Sci. This journal is © The Royal Society of Chemistry and Owner Societies 2013

Publ

ishe

d on

12

July

201

3. D

ownl

oade

d by

Ohi

o L

ink

Off

ices

on

27/0

8/20

13 0

7:21

:57.

View Article Online

For the sake of comparison we registered the fluorescencespectrum of a DOX aqueous solution on the confocal system.Even though there seems to be a more pronounced vibrationalstructure for DOX in the cellular context the differences are toosmall to draw conclusions on the location of DOX in thenucleus.

We also performed time-resolved fluorescence imagingexciting the DOX and DOX/oligomer incubated cells at405 nm, the only pulsed laser source available on the system.We have to consider that under these conditions we cannotavoid excitation of intrinsic cell fluorophores. Time-resolvedfluorescence detection was performed in the 565–605 nmrange corresponding to the DOX maximum intensity in thefluorescence spectrum. Careful analysis started from the fit ofthe fluorescence decay of a ROI located in the nucleus with abiexponential function and yielded in all cases two lifetimes of1.0 (25%) and 2.4 ns (75%) with good χ2 values of ca. 1.0. Wethen calculated the histogram for the complete image and afour-exponential function was necessary to fit the decay withgood chi-square values fixing the 1.0 and 2.4 ns lifetimes. Twomore lifetimes of less importance were found, one short of ca.250 ps and one longer both with negligible weight in thenucleus. The 1 ns lifetime is the same as that of DOX in waterand can be ascribed to DOX in aqueous environments,whereas the longer lifetime of 2.4 ns is more difficult toassign. The known mode of interaction of the anthracyclineswith DNA is intercalation of the dihydroxyanthraquinonemoiety between two base pairs, with location of the sugarresidue in the minor groove.28 This binding mode is associatedwith a hypochromic effect,26 a decrease of the emissionability29 and a drastic shortening of the emission lifetime to afew ps.30 So the longer-lived emission component of 2.4 nscannot be attributed to DOX intercalated on DNA. We suggestthat it may pertain to the drug located in a hydrophobicpocket of a nuclear protein, where the lifetime is expected tobe longer than in water, according to the data reported inTable 2 for DOX incorporated into pβ-CyD nanoparticles. Thefluorescence confocal imaging experiments have given verysimilar results for two types of DOX inoculation in the cancercells. This may suggest that DOX does no longer reside in theoligomer once inside the cell. Not being fluorescent it is notpossible to ascertain whether the oligomer has entered the celland, if it has, where it localizes.

4. Conclusions

Self-association of DOX compromises its therapeutic activityand thus represents an unfavourable factor in clinical appli-cations. We recently reported on epichlorohydrin–β-CyD copo-lymer based nanoparticles with 15 nm diameter promotingDOX monomerization in aqueous solution. However, theapparent association constants of DOX with this pβ-CyDsystem were quite low. We thus oriented toward CyD networksendowed with a suitable architecture for more efficient inter-action with DOX and focused in particular on highly water

soluble and biodegradable citric acid–γ-CyD polymers(pγ-CyDs), which were prepared according to a recently pub-lished procedure. Two oligomers with an MW of 21–33 kDaand 10–15 γ-CyD units per molecule were synthesized in a wellreproducible way and fully characterized by spectroscopicmethods. The CyD content was calculated from the NMRsignals. Compared to the pβ-CyD system the pγ-CyD oligomerswere able to associate DOX with much higher binding con-stants, promoting monomerization of the drug as well. Likely,due to the low molecular weight, the γ-CyD-based materialsdid not form nanoparticles as the epichlorohydrin β-CyD co-polymer. Nevertheless entrapment of DOX as monomer wasfavoured by the 3D frame of the pγ-CyD macromolecule, wherethe high local concentration of γ-CyD host units and the pres-ence of nanocavities in the sponge-like crosslinked network,surrounded by negatively charged carboxylate groups, mayconcur to disrupt the DOX dicationic dimer.

Incubation with both uncomplexed and pγ-CyD-complexedDOX resulted in effective DOX uptake in the nuclei of theMCF-7 breast tumor cell line with complete preservation of thedrug. FLIM data indicate that in addition to DNA, also nuclearproteins may provide suitable binding sites. The presentresults suggest that γ-CyD-based oligomeric systems mighthold great potential in the delivery of anthracycline anticancerdrugs for enhanced therapeutic efficacy. In particular they mayhave several advantages with respect to higher MW polymericsystems in terms of higher solubility, lower viscosity and morefavourable pharmacokinetics.

Acknowledgements

We thank Junta de Andalucía (grant FQM-6903) and the FP7PEOPLE-ITN project no. 237962-CYCLON for funding theresearch, and A. Maksimenko for helpful input for in vitrostudies. I.M. and S.M. acknowledge the CARISBO foundationfor the generous contribution to the acquisition of confocalfluorescence microscope images with FLIM detection.

References

1 D. Dal Ben, M. Palumbo, G. Zagotto, G. Capranico andS. Moro, DNA topoisomerase II structures and anthra-cycline activity: insights into ternary complex formation,Curr. Pharm. Des., 2007, 13, 2766–2780; G. Minotti,P. Menna, E. Salvatorelli, G. Cairo and L. Gianni, Anthra-cyclines: Molecular advances and pharmacologic develop-ments in antitumor activity and cardiotoxicity, Pharmacol.Rev., 2004, 56, 185–229.

2 A. Gabizon, H. Shmeeda and Y. Barenholz, Pharmaco-kinetics of pegylated liposomal doxorubicin - Review ofanimal and human studies, Clin. Pharmacokinet., 2003, 42,419–436; A. A. Gabizon, Pegylated liposomal doxorubicin:Metamorphosis of an old drug into a new form of chemo-therapy, Cancer Invest., 2001, 19, 424–436.

Photochemical & Photobiological Sciences Paper

This journal is © The Royal Society of Chemistry and Owner Societies 2013 Photochem. Photobiol. Sci.

Publ

ishe

d on

12

July

201

3. D

ownl

oade

d by

Ohi

o L

ink

Off

ices

on

27/0

8/20

13 0

7:21

:57.

View Article Online

3 M. E. Davis and M. E. Brewster, Cyclodextrin-based phar-maceutics: Past, present and future, Nat. Rev. Drug Discov-ery, 2004, 3, 1023–1035; V. Villari, A. Mazzaglia, R. Darcy,C. M. O’Driscoll and N. Micali, Nanostructures of CationicAmphiphilic Cyclodextrin Complexes with DNA, Biomacro-molecules, 2013, 14, 811–817.

4 F. van de Manakker, T. Vermonden, C. F. van Nostrumand W. E. Hennink, Cyclodextrin-Based PolymericMaterials: Synthesis, Properties, and Pharmaceutical/Bio-medical Applications, Biomacromolecules, 2009, 10, 3157–3175.

5 P. Y. Grosse, F. Bressolle and F. Pinguet, Methyl-beta-cyclo-dextrin in HL-60 parental and multidrug-resistant cancercell lines: effect on the cytotoxic activity and intracellularaccumulation of doxorubicin, Cancer Chemother. Pharma-col., 1997, 40, 489–494; P. Y. Grosse, F. Bressolle andF. Pinguet, Antiproliferative effect of methyl-beta-cyclodex-trin in vitro and in human tumour xenografted athymicnude mice, Br. J. Cancer, 1998, 78, 1165–1169; P. Y. Grosse,F. Bressolle and F. Pinguet, In vitro modulation of doxorubi-cin and docetaxel antitumoral activity by methyl-beta-cyclo-dextrin, Eur. J. Cancer, 1998, 34, 168–174; P. Y. Grosse,F. Bressolle, P. Vago, J. Simony-Lafontaine, M. Radal andF. Pinguet, Tumor cell membrane as a potential target formethyl-beta-cyclodextrin, Anticancer Res., 1998, 18, 379–384; A. Al-Omar, S. Abdou, L. De Robertis, A. Marsura andC. Finance, Complexation study and anticellular activityenhancement by doxorubicin-cyclodextrin complexes on amultidrug-resistant adenocarcinoma cell line, Bioorg. Med.Chem. Lett., 1999, 9, 1115–1120.

6 S. Tilloy, V. Monnaert, L. Fenart, H. Bricout, R. Cecchelliand E. Monflier, Methylated beta-cyclodextrin as P-gpmodulators for deliverance of doxorubicin across an invitro model of blood-brain barrier, Bioorg. Med. Chem. Lett.,2006, 16, 2154–2157; V. Monnaert, D. Betbeder, L. Fenart,H. Bricout, A. M. Lenfant, C. Landry, R. Cecchelli,E. Monflier and S. Tilloy, Effects of gamma- and hydroxy-propyl-gamma-cyclodextrins on the transport of doxorubi-cin across an in vitro model of blood-brain barrier,J. Pharmacol. Exp. Ther., 2004, 311, 1115–1120.

7 K. Hattori, A. Kenmoku, T. Mizuguchi, D. Ikeda,M. Mizuno and T. Inazu, Saccharide-branched cyclodex-trins as targeting drug carriers, J. Inclusion Phenom.Macrocyclic Chem., 2006, 56, 9–16; Y. Oda, N. Kobayashi,T. Yamanoi, K. Katsuraya, K. Takahashi and K. Hattori,Beta-cyclodextrin conjugates with glucose moietiesdesigned as drug carriers: Their syntheses, evaluationsusing concanavalin A and doxorubicin, and structuralanalyses by NMR spectroscopy, Med. Chem., 2008, 4,244–255; Y. Oda, H. Yanagisawa, M. Maruyama,K. Hattori and T. Yamanoi, Design, synthesis and evalu-ation ofD-galactose-beta-cyclodextrin conjugates as drug-carrying molecules, Bioorg. Med. Chem., 2008, 16, 8830–8840; G. J. L. Bernardes, R. Kikkeri, M. Maglinao,P. Laurino, M. Collot, S. Y. Hong, B. Lepenies andP. H. Seeberger, Design, synthesis and biological

evaluation of carbohydrate-functionalized cyclodextrinsand liposomes for hepatocyte-specific targeting, Org.Biomol. Chem., 2010, 8, 4987–4996.

8 W. Zhu, Y. Li, L. Liu, Y. Chen and F. Xi, Supramolecularhydrogels as a universal scaffold for stepwise deliveringDox and Dox/cisplatin loaded block copolymer micelles,Int. J. Pharm., 2012, 437, 11–19; Q.-D. Hu, H. Fan, Y. Ping,W.-Q. Liang, G.-P. Tang and J. Li, Cationic supramolecularnanoparticles for co-delivery of gene and anticancer drug,Chem. Commun., 2011, 47, 5572–5574; L. Y. Qiu, R. J. Wang,C. Zheng, Y. Jin and L. Q. Jin, Beta-cyclodextrin-centeredstar-shaped amphiphilic polymers for doxorubicin delivery,Nanomedicine, 2010, 5, 193–208; H. Kim, S. Kim, C. Park,H. Lee, H. J. Park and C. Kim, Glutathione-Induced Intra-cellular Release of Guests from Mesoporous Silica Nano-containers with Cyclodextrin Gatekeepers, Adv. Mater.,2010, 22, 4280–4283; E. S. Gil, J. S. Li, H. N. Xiao andT. L. Lowe, Quaternary Ammonium beta-CyclodextrinNanoparticles for Enhancing Doxorubicin Permeabilityacross the In Vitro Blood-Brain Barrier, Biomacromolecules,2009, 10, 505–516; Y. Hagiwara, H. Arima, F. Hirayama andK. Uekama, Prolonged retention of doxorubicin in tumorcells by encapsulation of gamma-cyclodextrin complex inpegylated liposomes, J. Inclusion Phenom. MacrocyclicChem., 2006, 56, 65–68; H. Arima, Y. Hagiwara, F. Hirayamaand K. Uekama, Enhancement of antitumor effect of doxo-rubicin by its complexation with gamma-cyclodextrin inpegylated liposomes, J. Drug Targeting, 2006, 14, 225–232;T. Liu, X. J. Li, Y. F. Qian, X. L. Hu and S. Y. Liu, Multifunc-tional pH-Disintegrable micellar nanoparticles of asymme-trically functionalized beta-cyclodextrin-Based starcopolymer covalently conjugated with doxorubicin andDOTA-Gd moieties, Biomaterials, 2012, 33, 2521–2531;N. Lin and A. Dufresne, Supramolecular Hydrogels from InSitu Host-Guest Inclusion between Chemically ModifiedCellulose Nanocrystals and Cyclodextrin, Biomacromole-cules, 2013, 14, 871–880.

9 S. Daoud-Mahammed, P. Couvreur, K. Bouchemal,M. Cheron, G. Lebas, C. Amiel and R. Gref, Cyclodextrinand Polysaccharide-Based Nanogels: Entrapment of TwoHydrophobic Molecules, Benzophenone and Tamoxifen,Biomacromolecules, 2009, 10, 547–554; S. Daoud-Mahammed, J. L. Grossiord, T. Bergua, C. Amiel,P. Couvreur and R. Gref, Self-assembling cyclodextrinbased hydrogels for the sustained delivery of hydrophobicdrugs, J. Biomed. Mater. Res., Part A, 2008, 86A, 736–748;S. Daoud-Mahammed, P. Couvreur, C. Amiel, M. Besnard,M. Appel and R. Gref, Original tamoxifen-loaded gels con-taining cyclodextrins: in situ self-assembling systems forcancer treatment, J. Drug Deliv. Sci. Technol., 2004, 14,51–55; S. Daoud-Mahammed, C. Ringard-Lefebvre,N. Razzouq, V. Rosilio, B. Gillet, P. Couvreur, C. Amiel andR. Gref, Spontaneous association of hydrophobized dextranand poly-beta-cyclodextrin into nanoassemblies. Formationand interaction with a hydrophobic drug, J. Colloid InterfaceSci., 2007, 307, 83–93; S. Daoud-Mahammed, P. Couvreur

Paper Photochemical & Photobiological Sciences

Photochem. Photobiol. Sci. This journal is © The Royal Society of Chemistry and Owner Societies 2013

Publ

ishe

d on

12

July

201

3. D

ownl

oade

d by

Ohi

o L

ink

Off

ices

on

27/0

8/20

13 0

7:21

:57.

View Article Online

and R. Gref, Novel self-assembling nanogels: stability andlyophilisation studies, Int. J. Pharm., 2007, 332, 185–191.

10 E. Renard, A. Deratani, G. Volet and B. Sebille, Preparationand characterization of water soluble high molecularweight beta-cyclodextrin-epichlorohydrin polymers, Eur.Polym. J., 1997, 33, 49–57; R. Gref, C. Amiel, K. Molinard,S. Daoud-Mahammed, B. Sebille, B. Gillet, J. C. Beloeil,C. Ringard, V. Rosilio, J. Poupaert and P. Couvreur, Newself-assembled nanogels based on host-guest interactions:Characterization and drug loading, J. Controlled Release,2006, 111, 316–324.

11 R. Anand, F. Manoli, I. Manet, S. Daoud-Mahammed,V. Agostoni, R. Gref and S. Monti, Beta-cyclodextrinpolymer nanoparticles as carriers for doxorubicin and arte-misinin: a spectroscopic and photophysical study, Photo-chem. Photobiol. Sci., 2012, 11, 1285–1292.

12 T. Nakanishi, S. Fukushima, K. Okamoto, M. Suzuki,Y. Matsumura, M. Yokoyama, T. Okano, Y. Sakurai andK. Kataoka, Development of the polymer micelle carriersystem for doxorubicin, J. Controlled Release, 2001, 74, 295–302.

13 R. Anand, S. Ottani, F. Manoli, I. Manet and S. Monti, Aclose-up on doxorubicin binding to gamma-cyclodextrin:an elucidating spectroscopic, photophysical and confor-mational study, RSC Adv., 2012, 2, 2346–2357.

14 O. Bekers, J. H. Beijnen, M. Otagiri, A. Bult andW. J. M. Underberg, Inclusion complexation of Doxorubicinand Daunorubicin with Cyclodextrins, J. Pharm. Biomed.Anal., 1990, 8, 671–674; O. Bekers, J. H. Beijnen, B. J. Vis,A. Suenaga, M. Otagiri, A. Bult and W. J. M. Underberg,Effect of Cyclodextrin Complexation on the Chemical-Stabi-lity of Doxorubicin and Daunorubicin in Aqueous-Solu-tions, Int. J. Pharm., 1991, 72, 123–130; O. Bekers,J. J. Kettenesvandenbosch, S. P. Vanhelden, D. Seijkens,J. H. Beijnen, A. Bult and W. J. M. Underberg, InclusionComplex-Formation of Anthracycline Antibiotics withCyclodextrins – a Proton Nuclear-Magnetic-Resonance andMolecular Modeling Study, J. Inclusion Phenom. Mol. Recog-nit. Chem., 1991, 11, 185–193.

15 N. Husain, T. T. Ndou, A. M. Delapena and I. M. Warner,Complexation of Doxorubicin with Beta-Cyclodextrins andGamma-Cyclodextrins, Appl. Spectrosc., 1992, 46, 652–658.

16 B. Martel, M. Weltrowski, D. Ruffin and M. Morcellet, Poly-carboxylic acids as crosslinking agents for grafting cyclo-dextrins onto cotton and wool fabrics: Study of the processparameters, J. Appl. Polym. Sci., 2002, 83, 1449–1456.

17 T. H. H. Thi, F. Chai, S. Lepretre, N. Blanchemain,B. Martel, F. Siepmann, H. F. Hildebrand, J. Siepmann andM. P. Flament, Bone implants modified with cyclodextrin:Study of drug release in bulk fluid and into agarose gel,Int. J. Pharm., 2010, 400, 74–85; N. Blanchemain,Y. Karrout, N. Tabary, C. Neut, M. Bria, J. Siepmann,H. F. Hildebrand and B. Martel, Methyl-beta-cyclodextrinmodified vascular prosthesis: Influence of the modificationlevel on the drug delivery properties in different media,Acta Biomater., 2011, 7, 304–314.

18 S. Joudieh, P. Bon, B. Martel, M. Skiba and M. Lahiani-Skiba, Cyclodextrin Polymers as Efficient Solubilizers ofAlbendazole: Complexation and Physico-Chemical Charac-terization, J. Nanosci. Nanotechnol., 2009, 9, 132–140;Y. Bakkour, G. Vermeersch, M. Morcellet, F. Boschin,B. Martel and N. Azaroual, Formation of cyclodextrininclusion complexes with doxycyclin-hyclate: NMR investi-gation of their characterisation and stability, J. InclusionPhenom. Macrocyclic Chem., 2006, 54, 109–114.

19 C. Danel, N. Azaroual, C. Chavaria, P. Odou, B. Martel andC. Vaccher, Comparative study of the complex formingability and enantioselectivity of cyclodextrin polymers byCE and H-1 NMR, Carbohydr. Polym., 2013, 92, 2282–2292.

20 J. R. E. Fraser, T. C. Laurent, H. Pertoft and E. Baxter,Plasma-Clearance, Tissue Distribution and Metabolism ofHyaluronic-Acid Injected Intravenously in the Rabbit,Biochem. J., 1981, 200, 415–424; M. I. ul-Haq, B. F. L. Lai,R. Chapanian and J. N. Kizhakkedathu, Influence of archi-tecture of high molecular weight linear and branched poly-glycerols on their biocompatibility and biodistribution,Biomaterials, 2013, 33, 9135–9147; T. Etrych, V. Subr,J. Strohalm, M. Sirova, B. Rihova and K. Ulbrich, HPMAcopolymer-doxorubicin conjugates: The effects of mole-cular weight and architecture on biodistribution andin vivo activity, J. Controlled Release, 2012, 164, 346–354;M. E. Fox, F. C. Szoka and J. M. J. Frèchet, Soluble PolymerCarriers for the Treatment of Cancer: The Importance ofMolecular Architecture, Acc. Chem. Res., 2009, 42, 1141–1151.

21 B. Martel, D. Ruffin, M. Weltrowski, Y. Lekchiri andM. Morcellet, Water-soluble polvmers and gels from thepolycondensation between cyclodextrins and poly(car-boxylic acid)s: A study of the preparation parameters,J. Appl. Polym. Sci., 2005, 97, 433–442.

22 M. M. L. Fiallo, H. Tayeb, A. Suarato and A. Garnier-Suil-lerot, Circular dichroism studies on anthracycline anti-tumor compounds. Relationship between the molecularstructure and the spectroscopic data, J. Pharm. Sci., 1998,87, 967–975.

23 B. Samori, A. Rossi, I. D. Pellerano, G. Marconi,L. Valentini, B. Gioia and A. Vigevani, Interactionsbetween Drugs and Nucleic-Acids 0.1. Dichroic Studies ofDoxorubicin, Daunorubicin, and Their Basic Chromo-phore, Quinizarin, J. Chem. Soc., Perkin Trans. 2, 1987,1419–1426.

24 P. Agrawal, S. K. Barthwal and R. Barthwal, Studies on self-aggregation of anthracycline drugs by restrained moleculardynamics approach using nuclear magnetic resonancespectroscopy supported by absorption, fluorescence,diffusion ordered spectroscopy and mass spectrometry,Eur. J. Med. Chem., 2009, 44, 1437–1451.

25 P. Changenet-Barret, T. Gustavsson, D. Markovitsi, I. Manetand S. Monti, Unravelling molecular mechanisms in thefluorescence spectra of Doxorubicin in aqueous solution byfemtosecond fluorescence spectroscopy, Phys. Chem. Chem.Phys., 2013, 15, 2937–2944.

Photochemical & Photobiological Sciences Paper

This journal is © The Royal Society of Chemistry and Owner Societies 2013 Photochem. Photobiol. Sci.

Publ

ishe

d on

12

July

201

3. D

ownl

oade

d by

Ohi

o L

ink

Off

ices

on

27/0

8/20

13 0

7:21

:57.

View Article Online

26 L. Angeloni, G. Smulevich and M. P. Marzocchi, Absorp-tion, Fluorescence and Resonance Raman-Spectra of Adria-mycin and Its Complex with DNA, Spectrochim. Acta, Part A,1982, 38, 213–217.

27 K. K. Karukstis, E. H. Z. Thompson, J. A. Whiles andR. J. Rosenfeld, Deciphering the fluorescence signature ofdaunomycin and doxorubicin, Biophys. Chem., 1998, 73,249–263.

28 A. H. J. Wang, G. Ughetto, G. J. Quigley and A. Rich, Inter-actions between an Anthracycline Antibiotic and DNA –

Molecular-Structure of Daunomycin Complexed to

D(Cpgptpapcpg) at 1.2-a Resolution, Biochemistry, 1987, 26,1152–1163.

29 V. Rizzo, C. Battistini, A. Vigevani, N. Sacchi, G. Razzano,F. Arcamone, A. Garbesi, F. Colonna, M. L. Capobianco andL. Tondelli, Association of anthracyclines and synthetichexanucleotides. Structural factors influencing sequencespecificity, J. Mol. Recognit., 1989, 2, 132–141.

30 X. G. Qu, C. Z. Wan, H. C. Becker, D. P. Zhong andA. H. Zewail, The anticancer drug–DNA complex: Femto-second primary dynamics for anthracycline antibiotics func-tion, Proc. Natl. Acad. Sci. U. S. A., 2001, 98, 14212–14217.

Paper Photochemical & Photobiological Sciences

Photochem. Photobiol. Sci. This journal is © The Royal Society of Chemistry and Owner Societies 2013

Publ

ishe

d on

12

July

201

3. D

ownl

oade

d by

Ohi

o L

ink

Off

ices

on

27/0

8/20

13 0

7:21

:57.

View Article Online