Synthesis and Characterization of a Theranostic Vascular Disrupting Agent for In Vivo MR Imaging

Published: March 16, 2011

r 2011 American Chemical Society 879 dx.doi.org/10.1021/bc100329t | Bioconjugate Chem. 2011, 22, 879–886

ARTICLE

pubs.acs.org/bc

Synthesis and Characterization of a Theranostic VascularDisrupting Agent for In Vivo MR ImagingTammy L. Kalber,*,† Nazila Kamaly,†,‡ Stephanie A. Higham,† John A. Pugh,§ Josephine Bunch,§,||

Cameron W. McLeod,§ Andrew D. Miller,‡ and Jimmy D. Bell†

†Metabolic and Molecular Imaging Group, MRC Clinical Sciences Centre, Imperial College London, Hammersmith Hospital,London W12 0NN, U.K.‡Imperial College Genetic Therapies Centre, Department of Chemistry, Flowers Building, Armstrong Road,Imperial College London, London SW7 2AZ, U.K.

§Centre for Analytical Sciences, Department of Chemistry, University of Sheffield, Brook Hill, Sheffield S3 7HF, U.K.

)School of Chemistry, University of Birmingham, Edgbaston, Birmingham B15 2TT, U.K.

’ INTRODUCTION

It is well established that tumor growth is dependent upon thegenerationof a functional vascular supply and that tumor vasculaturediffers significantly from that of vessels in normal tissues.1�4Normaltissues have well-ordered vessels with quiescent endothelial cells,whereas tumor vasculature is a chaotic network of tortuous, thin-walled, leaky vessels, with a high proportion of immature, proliferat-ing endothelial cells that rely on their tubulin cytoskeleton tomaintain shape.3,4 Consequently, tumor endothelium can be tar-geted in therapeutic approaches to cancer that leave normal bloodvessels unaffected.5 In particular, vascular disrupting agents (VDAs)invoke the selective collapse of tumor vasculature, leading to a loss intumor blood flow, secondary cell death, and tumor necrosis due toischemia.6 VDAs have largely consisted of tubulin-depolymerizingagents such as colchicine and its derivatives, and more recently, theflavonids such as 5,6-dimethylxanthenone-4-acetic acid (DMXAA).

Colchicine binds to β-tubulin, inhibiting tubulin polymeriza-tion, resulting in microtubule destabilization. Destabilization ofthe microtubule cytoskeleton induces cell shape changes, which inthe endothelial cells of tumors, results in cell detachment, exposureof the basal lamina, and vessel occlusion.7 However, colchicine istoo toxic for clinical use as it is only effective close to its maximumtolerated dose (MTD). Derivatives of colchicine have there-fore been developed, the best known being disodium combretas-tatin A-4 3-O-phosphate (CA-4-P) and ZD6126, both of whichinduce vascular occlusion and therefore antitumor effects in awide range of tumormodels at doses that are less than one-tenth ofthe MTD.8�11

Received: July 21, 2010Revised: February 17, 2011

ABSTRACT: Colchicine, a known tubulin binding agent andvascular disrupting agent, causes rapid vascular shut down andcentral necrosis in tumors. The binding of tubulin results in tubulindestabilization, with characteristic cell shape changes and inhibitionof cell division, and results in cell death. A gadolinium(III) labeledderivative of colchicine (Gd 3DOTA 3Colchicinic acid) was synthe-sized and characterized as a theranostic agent (enabling simulta-neous diagnostic/real time MRI contrast imaging). In vitro,Gd 3DOTA 3Colchicinic acid was shown to initiate cell changescharacteristic of tubulin-destabilization in both OVCAR-3 andIGROV-1 ovarian carcinoma cell lines in vitro over a period of 24h, while maintaining the qualities of the MR imaging tracer. In vivo,Gd 3DOTA 3Colchicinic acid (200 mg/kg) was shown to inducethe formation of central necrosis, which was confirmed ex vivo by histology, in OVCAR-3 subcutaneous tumor xenografts, whilesimultaneously acting as an imaging agent to promote a significant reduction in theMR relaxation time T1 (p < 0.05) of tumors 24 h post-administration.Morphological changeswithin the tumorwhich correspondedwith areas derived from the formation of central necrosiswerealso present on MR images that were not observed for the same colchicine derivate that was not complexed with gadolinium that alsopresented with central necrosis ex vivo. However, Gd 3DOTA 3Colchicinic acid accumulation in the liver, as shown by changes in liverT1 (p< 0.05), takes placewithin 2 h. The implication is thatGd 3DOTA 3Colchicinic acid distributes to tissues, including tumors, within 2 h,but enters tumor cells to lower T1 times and promotes cell death over a period of up to 24 h. As the biodistribution/pharmacokinetic andpharmacodynamics data provided here is similar to that of conventional colchicines derivatives, such combined data are a potentiallypowerful way to rapidly characterize the complete behavior of drug candidates in vivo.

880 dx.doi.org/10.1021/bc100329t |Bioconjugate Chem. 2011, 22, 879–886

Bioconjugate Chemistry ARTICLE

The impact of VDAs on tumor tissue has been demonstratedhistologically in numerous tumor models, revealing central necrosiswithin 24 hof a single injectionwith a characteristic thin rimof viabletumor cells at the periphery.9,12 The rim is thought to be a conse-quence of these cells receiving nutrients/oxygen from nontumorvessels in adjacent normal tissue allowing these tumor regions toregrow. Therefore, VDA treatment is not expected to cause tumorshrinkage or significant growth delay as a single therapy. Accord-ingly, in the absence of tumor shrinkage, there is a need for noni-nvasive methods that can provide an assessment of tumor vascu-lature and pathophysiology in response to VDA therapy in vivo. Thehemodynamic tumor response to VDA therapy can be monitoredby dynamic contrast enhanced magnetic resonance imaging (MRI)using gadolinium(III) (Gd) chelates,13�17 as well as by intrinsic-susceptibility MRI, which is sensitive to changes in blood oxygena-tion.18,19 Otherwise, pathological changes have been observed byMR methodologies such as diffusion-weighted MRI20�22 or necro-sis-avid contrast agents capable of detecting tumor necrosis,23 aswellas methods to investigate apoptosis and cell death, such as cholinespectroscopy24 or targeted contrast agents.25 Since all these imagingmethodologies are responding to secondary changes at selectedtime points after VDA administration, then this suggests that VDAimaging itself could provide the means not only to reach a betterunderstanding of drug biodistribution/pharmacokinetics, routes ofclearance, etc., but also to simultaneously monitor pharmacody-namic phenotypic effects in real time as well as tumor necrosis inparticular.

Previously, colchicine has been conjugated with imagingagents26�31 including Gd-chelates,32 but mainly for applications invitro or in terminal in vivo studies. The aimof the study reported herewas to conjugate an MRI contrast agent to colchicine in order toproduce a novel theranostic VDA capable of remaining as afunctional VDA as well as acting as a competent contrast agentfor similtaneous diagnostic/real time MRI contrast imaging in vivo.Here, we describe the synthesis and characterization of such atheranostic VDA followed by a range of in vitro and in vivo imagingstudies over 24 h post-VDA administration, concluding withhistology and laser ablation inductively coupled mass spectrometry(LA-ICP-MS) to correlate Gd (III) distribution with associatedrelaxation time (T1) changes shown by in vivoMR.

’EXPERIMENTAL SECTION

Gerneral procedures. All chemicals were of analytical/highgrade and purchased from Sigma-Aldrich (Dorset, U.K.) or Macro-cyclics (Dallas, USA). Mass spectrometry (MS) were performedusing VG-070B, Joel SX-102, or Bruker Esquire 3000 electrospraymass spectrometry (ESI) instruments. 1H NMR spectra wererecorded on a 400 MHz Bruker Advance 400 spectrometer. AllMRI experiments were conducted on a 4.7 T Varian Direct DriveMRI scanner (Varian Inc., Palo Alto, CA, USA). All procedures onanimals were carried out in accordance with U.K. Home Officeregulations and the Guidance for the Operation of Animals(Scientific Procedures) Act (1986).Chemical Synthesis.Colchicinic acid (Scheme 1) was synthe-

sized according to the previously reportedmethod by Satpati et al.26

Tetraazacyclododecanetetraacetic acid (DOTA) in the form ofDOTA 3NHS ester (100 mg, 0.121 mmol) and colchicinic acid(41.55 mg, 0.121 mmol) were added to an air-evacuated round-bottom flask, and to this, dry dichloromethane (30mL) followed bytriethylamine (3 eq 0.363mmol) was added and stirred at 45 �C for12 h. The solvent was removed and the crude mixture purified byflash column chromatography usingMerck 0.040 to 0.063 mm, 230to 400 mesh silica gel, and the product eluted with dichloro-methane/methanol/acetic acid (6:3:1 v/v). The collected fractionswere reduced to yield a solid (87mg, 98%) DOTA 3Colchicinic acid(Scheme 1). Rf [dichloromethane/methanol/acetic acid, 6:3:1 v/v]0.11. 1H NMR (400 MHz, MeOD) 8.28 (s, br, 1H, �OH, H10),7.76 (s, 1H, H8), 7.46 (s, 1H, H11), 7.11 (d, 1H, H12), 6.92 (d, 1H,H4), 6.48 (s, br, 1H, �CNHCOCH2�), 4.34 (d, 2H, �NCO-CH2N�), 3.64 (m, 16H, 4� �NCH2CH2N�), 3.26 (m, 6H,3� �NCH2COOH), 3.12, (s, 3H, -OCH3, H1), 3.08 (m, 4H,�CH2CH2�, H5,6), 2.77 (s, 3H, �OCH3, H2), 2.63 (s, 3H,�OCH3, H3). MS (ESIþ) calculated for C35H47N5O12 m/z729.7740; found, 730.10 (M þ H)þ.A stoichiometric amount of gadolinium chloride (20.90 mg,

0.056 mmol) was added to DOTA 3Colchicinic acid (41.06 mg,0.056 mmol) and stirred in doubly distilled water (20 mL) at90 �C overnight. The water was freeze-dried to yield a powder(50 mg, quantitative yield): Gd 3DOTA 3Colchicinic acid(Figure 1). Rf [dichloromethane/methanol/acetic acid, 6:3:1

Scheme 1. Synthesis of Gd 3DOTA 3Colchicinic acida

a (i) 30% H2SO4, 100 �C, 5 h, 20%; (ii) DOTA-NHS-ester, DCM, 40 �C, 3 eq. Et3N, 12 h, 98%; (iii) 6H2O 3GdCl3, H2O, 90 �C, quantitative.



v/v] 0.18. MS (ESIþ) calculated for C35H44GdN5O12 m/z884.00; found, 885.23 (M þ H)þ. The gadolinium compoundwas characterized by ESI-MS, and the presence of free gadoli-nium ions was determined by the xylenol orange test previouslydescribed by Barge et al.33

In Vitro. Phantom Studies. 0.5 mM of Gd 3DOTA 3Colchici-nic acid, DOTA 3Colchicinic acid, Gd 3DOTA (Dotarem, Guer-bet, France), or water was used to assess signal enhancement. Arange of 6 concentrations (0.03�0.5 mM) of Gd 3DOTA 3Col-chicinic acid or Gd 3DOTA was then used to calculate relaxivityfor each contrast agent. Phantomswere placed in a quadrature 1Hvolume coil and positioned into the MRI scanner. A spin�echosequence with the following parameters was used to assess T1: 10TRs ranging from 50 to 7000 ms; TE = 10 ms; FOV = 45 � 45mm; averages, 1; matrix size, 256� 128 and a 2.0 mm thickness;and bore temperature, ∼ 20 �C.Tissue Culture. Human ovarian carcinoma OVCAR-3 and

IGROV-1 cell lines were grown in T175 flasks (Fisher Scientific,Loughborough, UK) in Dulbecco’s modified Eagle's medium(DMEM) (Sigma), supplemented with 10% heat inactivated fetalcalf serum (GIBCO, Grand Island NY, USA) in an incubator at37 �C with 95% air and 5% CO2. Cells were grown to 100%confluence before being trypsinized, counted, and then either platedfor in vitro experiments or prepared for subcutaneous implantation.Toxicity Studies.Cells were plated at 4� 105 in 6well plates 24 h

prior to incubation with a range of concentrations (1 nM�1 mM)of either DOTA 3Colchicinic acid or colchicine dissolved in serumfree DMEM. Histological morphology of cells was recorded at 2 hintervals up to 8 h, and at 24 h using an Olympus IX71 microscope(Olympus, Middlesex, UK). The cells were harvested at the 24 htime point and the numbers counted using a hemocytometer.Gd 3DOTA 3 Colchicinic Acid Uptake. Cells were plated at 1.5

� 105 in T25 flasks 24 h prior to the incubation of 1 mL of 1 mMGd 3DOTA 3Colchicinic acid (the most efficient incubation dosefrom toxicity studies). Cells were then washed, harvested, andpelleted and resuspended in 1% agarose (Sigma, UK) at 2 hintervals up to 8 h, and at 24 h after incubation and scanned asdescribed for phantoms.In Vivo. MRI. OVCAR-3 cells (5 � 106/0.1 mL) were inocu-

lated into the flank of 6�8 weeks old Balb/c nude mice. Oncepalpable, tumors were measured in three orthogonal dimensionsevery 2 days up until treatment, then pre and 24 h post-treatment,using calipers. Tumor volumeswere estimated assuming an ellipsoidshape using the formula volume = length� width� depth� π/6.

When tumors reached approximately 7�10 mm3, mice wereanesthetized with an isoflurane/O2 mix and scanned as describedfor phantoms using the parameters; 5 TRs ranging from 400�5000ms; TE= 15ms; FOV=45� 45mm; averages, 1;matrix size, 256�128, 2.0 mm thickness; and 20 consecutive transverse slices coveringthe whole abdomen. Mice were removed from the magnet and a tailvein cannulated for the administration of 200 mg/kg (22.64 mmol)Gd 3DOTA 3Colchicinic acid (n = 4), 22.64 mmol DOTA 3Colchi-cinic acid (n = 4), 22.64 mmol Gd 3DOTA (n = 4), or 200 μL ofsaline (n= 4); 200mg/kgwas used as this is an optimal dose that hasbeen used for both CA-4-P and ZD6126 in mice. Mice were thenrescanned with the same parameters 2, 8, and 24 h post-injection.After the final scan mice were sacrificed, the tumor was excised andfrozen for LA-ICP-MS and hematoxylin and eosin (H& E) staining.The liver and kidneys were also taken for H & E only.MRI Data Analysis. A region of interest (ROI) was drawn

around the whole tumor excluding the skin on MR images usingImage J (National Institute of Health, USA).34 Mean signalintensities of the ROIs at different TR values were measured andused to calculate the MR transverse relaxation time T1 usingGraphpad Prism (San Diego, California, USA). Mean T1 valuesfor each cohort were derived and a two-tailed unpaired t testassuming equal variances performed at each time point todetermine significant difference, with a 5% level of statisticalsignificance. Circular ROIs were then drawn in the liver andkidney to assess clearance and the signal intensities converted toT1 as described above.LA-ICP-MS. Frozen tissue was mounted onto a cryostat chuck

using water/ice slush frozen to �20 �C and sections cut at 20 μmusing a cryotome. The laser ablation system (LSX-200, Cetac,266 nm Nd:YAG) was configured to perform multiple line rasters,generating 2D elemental distribution maps. A beam diameter of 50μm was used for ablating tumor sections (with energy 0.6 mJ), ascanning speed of 50 μm/s and laser frequency of 10 Hz. For ICP-MS (Agilent,HP4500 Series 100), the elementsGd157 andZn66 (anelement widespread in biochemistry) were monitored in a time-resolved mode, isotopes were selected on the basis of high-percentage abundance, and minimal isobaric and polyatomic inter-ferences. Tumor sample areas were in the range 130�190 mm2.Elemental maps were produced using the Graphis software package(Kylebank Software Ltd., Ayr, UK). Adjacent sections where takenfor H & E staining.Histological Assessment. Sections were incubated with hema-

toxylin solution for 1 min, and then in freshly prepared eosin Y(both Sigma) for 30 s. The sections were then dehydrated withalcohol and passed through histoclear (Fisher Scientific),mounted with DPX (VWR), and imaged using the OlympusIX71 microscope as described in the toxicity studies.

’RESULTS

Chemical Synthesis. The synthesis of Gd 3DOTA 3Colchici-nic acid was carried out in three steps as shown in Scheme 1. Thecontrol compound DOTA 3Colchicinic acid was obtained ingood yield and was characterized by 1H and MS spectroscopy.Complexation of Gd(III) within DOTA could not be confirmedby NMR due to the paramagnetic nature of Gd(III) and wastherefore confirmed by ESI-MS and the xylenol orange assay.Isotopic peaks of Gd(III) were highly visible in theMS trace, andthere were no significant free Gd(III) ions detectable aftercomplexation according to the xylenol orange assay, therebyconfirming the formation of Gd 3DOTA 3Colchicinic acid.

Figure 1. In vitro cell studies of DOTA 3Colchicinic acid in OVCAR-3cells. A histology picture of OVCAR-3 cells, (A) without and (B) with1 mM 24 h after DOTA 3Colchicinic acid; (C) cell counts for bothcolchicine (white) and DOTA 3Colchicinic acid (gray) after 24 h atvarious concentrations. Magnification �100.



In Vitro. Phantoms showed that Gd 3DOTA 3Colchicinic acidachieved a 96% reduction in T1 compared to the same concen-tration of DOTA 3Colchicinic acid or water, and its relaxivity wascalculated at 3.83 mM�1s�1. This was equivalent to that of non-conjugated Gd 3DOTA, which was calculated at 3.66 mM�1s�1.The efficiency of DOTA 3Colchicinic acid for binding tubulinwas assessed in two ovarian carcinoma cell lines. Normally,endothelial cell lines such as HUVEC’s are used to assess thefunctionality of novel VDAs. However, the effects of tubulin-binding drugs are not specific to endothelial cells. Indeed, sometumor cell types are more sensitive to VDA tubulin depolymer-ization than endothelial cells.35 The extent to which direct tumorcell death has contributed to VDA therapy is likely dependent ontumor type. Therefore, DOTA 3Colchicinic acid was tested intumor cells to assess tubulin binding and tubulin destabilizationactivities in order to obtain a better idea of the amount of celldeath/antitumor effect that Gd 3DOTA 3Colchicinic acid mayprovide in vivo, in addition to VDA action.Cell counts showed that colchicine was effective at significantly

reducing cell numbers at all concentrations in both the OVCAR-3and IGROV-1 cell lines, whereas DOTA 3Colchicinic acid was onlyeffective in the millimolar range, causing reductions in cell numbersin both cell lines (Fgure 1C). These high concentration effects werethe same for the Gd(III) derivate (Gd 3DOTA 3Colchicinic acid)and correlated well with colchicine effects at the same highconcentrations. Arguably, these data suggest that Gd 3DOTA 3Colchicinic acid may be much less cytotoxic than colchicine to cellsin general but retains colchicine's drug-action mechanism in spite ofthe conjugated Gd 3DOTA chelate (Figure 1C). Light microscopydemonstrated the emergence of initial cell shape changes as early as2 h post-treatment with DOTA 3Colchicinic acid, followed byrounding of the cells, and then detachment from the flask walls at24 h (Figure 1A and B). MR images of cell pellets taken at specifictime points after incubation with Gd 3DOTA 3Colchicinic acid(1 mM) showed a 54% reduction in T1 at 2 h compared to thatin control cells, which decreased further reaching a maximum of

83% reduction at 6 h. After 6 h,T1 recovered slightly, but at no pointreturned to baseline over the 24 h period (data not shown). Thesedata are in accordance with the histology data and suggest that initialtubulin binding takes place in vitro within the first 2 h of incubation,followed by further cellular uptake and tubulin binding over 6 h.In Vivo.Mice showed no adverse effects throughout the study.

There was no significant difference in tumor size between cohorts,consistent with the fact that colchicine and its derivatives do nottypically cause tumor shrinkage or growth delay. Images of pre-treated tumors appeared to be homogeneous in presentation. At 24h, the tumor images showed evidence of morphological changesmainly evident in the central area of the tumor as areas of low signalintensity (dark areas) in 3 out of 4 of the Gd 3DOTA 3Colchicinic

Figure 2. A set of two contiguous T1 MRI images acquired from similaranatomic points from the samemouse bearing a subcutaneous OVCAR3tumor pre, 2, 8, and 24 h after 200 mg/kg DOTA 3Colchicinic acid orGd 3DOTA 3Colchicinic acid administration. Column 24 h (annotated)shows the enlarged 24 h tumor image, which includes annotated areas ofsignal change in the Gd 3DOTA 3Colchicinic acid treated tumors. Thewhite arrow denotes a phantom placed to correlate images.

Figure 3. T1 (ms) measured pre, 2, 8, and 24 h post-injection of 200mg/kg (22.64 mmol) Gd 3DOTA 3Colchicinic acid, 22.64 mmolDOTA 3Colchicinic acid, or 200 μL saline for (A) tumor, (B) liver,and (C) kidney. Data points( standard error of the mean (n = 4 in eachgroup). * = P < 0.05.



acid treated mice (Figure 2). These changes in signal intensity werenot observed at any time point, in any other cohort throughoutthe study.Changes in mean T1 are presented (Figure 3) for the tumor,

liver, and kidney. Gd 3DOTA exhibited similar values and trends assaline and DOTA 3Colchicinic acid controls; therefore, these datahave been omitted (from Figure 3). TumorT1 presented no changein mean T1 from the baseline for any cohort at 2 h. A slight butnonsignificant reduction of 3% inmeanT1 was observed at 8 h post-Gd 3DOTA 3Colchicinic acid administration, which decreasedfurther to become a significantly (p < 0.05) greater reduction of6% in mean T1 at 24 h post-administration in comparison with thatin saline, Gd 3DOTA, and DOTA 3Colchicinic acid control cohorts(Figure 3A). A significant (p < 0.05) reduction of 55% in meanT1 was also observed in the liver post-Gd 3DOTA 3Colchicinic acidadministration at 2 h in comparison to that in saline, Gd 3DOTA, orDOTA 3Colchicinic acid control cohorts. This reduction inmeanT1was sustained throughout, up until 24 h post-administration(Figure 3B). In the case of the kidney, no single significant changein T1 was observed in any cohort (Figure 3C).LA-ICP-MS. Elemental maps of Gd157 and Zn66 distribution in

tumors against a saline controlwere generated 24hpost-Gd 3DOTA 3Colchicinic acid administration (Figure 4). LA-ICP-MS analysisindicated substantial amounts of Gd(III) present throughout thewhole tumor. Using Zn66 as an indication of viable tissue, it can beclearly seen that the largest accumulation of Gd(III) takes place incomplementary areas thatmaybenecrotic. This suggests thatGd(III)has accumulated in tumor zones that were originally well perfusedwith blood vessels but became necrotic due to the occlusion of bloodvessels post-VDA action. Our data indicate that theMRI-active VDAis still trapped within these apparently necrotic areas.Histology.H& E staining showed that all tumors 24 h post-Gd 3

DOTA 3Colchicinic acid or DOTA 3Colchicinic acid administration

possessed a large central core of necrosis with a rim of viable tumortissue at the tumor periphery. Therefore, the degree of necrosis wasnot dependent on Gd(III) incorporation (Figure 5B). In contrast,tumors 24 h post-saline and Gd 3DOTA administration possessedminimal necrosis and exhibited large areas of viable tumor tissuethroughout the tumor (Figure 5A). Otherwise, the liver, which wasalso observed to accumulate substantial amounts of Gd(III) post-Gd 3DOTA 3Colchicinic acid as evidenced by substantial reductionsin mean T1 values (Figure 3B), showed no significant differencesin liver morphology between animals treated with saline or anyagent used in this work (Figure 5C and D). The kidney too wassubstantially unaffected by Gd 3DOTA 3Colchicinic acid adminis-tration or the administration of any other control compounds(data not shown).

’DISCUSSION

In this study, we have synthesized and characterized a noveltheranostic tubulin binding agent capable of working as a vasculardisrupting agent that was capable of both reducing T1 as well askeeping its efficacy to bind to tubulin in vitro, resulting in cellshape changes and central tumor necrosis in vivo characteristic ofVDA treatment that could be followed with standard MRsequences longitudinally.

The conjugation of DOTA to colchicine resulted in DOTA 3Colchicinic acid, which was complexed with gadolinium toproduce the novel compound Gd 3DOTA 3Colchicinic acid. Invitro results showed that conjugation of Gd 3DOTA to colchicinedid not affect the relaxivity of the contrast agent as an MR probe.DOTA 3Colchicinic acid was also found to be a functional thera-peutic agent and was effective at causing characteristic VDAchanges in both ovarian cell lines in vitro at concentrations of themillimolar range, suggesting that cells were less sensitive to theeffects of cytotoxicity as afforded by unconjugated colchicine.

Figure 4. Laser ablation inductively coupled mass spectrometry data of Gd157 and Zn66 distribution in tumor sections of OVCAR3 tumors 24 h after200 mg/kg Gd 3DOTA 3Colchicinic acid or 200 μL of saline administration.



Colchicine is too toxic for clinical use and is effective only at orclose to its MTD, where direct cell cytotoxicity is the principalmechanism of action. A reduction in tumor cell cytotoxicity,therefore, suggests that DOTA 3Colchicinic acid has a lowerMTD and therefore a wider therapeutic window than colchicine,which would be well tolerated in animals.

DOTA 3Colchicinic acid was tested in ovarian carcinoma cellsinitially in vitro instead of endothelial cells, as rapidly dividing tumorcells are also susceptible to VDA therapy and would give a betterunderstanding about the secondary effects of VDA therapy ontumor cells that would be on top of the initial effect of vascular shutdown.35 Initial cell shape changes were observed at as early as 2 hafter treatment with DOTA 3Colchicinc acid, followed by cellrounding caused by the contraction of the tubulin cytoskeleton,decreasing intracellular adhesion, and subsequent cell loss due todetachment, similar towhat has been shown in endothelial cells afterdoses of either ZD6126 or combretastatin derivatives in vitro.9,36,37

Gd 3DOTA 3Colchicinic acid also caused T1 reductions in cells invitro at 2 h, which further decreased over 6 h. After 6 h, the T1recovery may be due to lower numbers of cells or changes in cellshape effecting water diffusion. However, as theT1 does not recovercompletely, this implies that binding is irreversible as with colchicineitself,38 resulting in long drug exposures that will therefore cause cellcytotoxicity, which would lead to increased tumor cell death.39

As expected for single dose VDA treatment, Gd 3DOTA 3Colchicinic acid did not cause tumor shrinkage or growth delay.Reductions in T1 were apparent at 8 h and became significant at

24 h after treatment with Gd 3DOTA.Colchcinic acid. As therewas no change in T1 for the saline or DOTA 3Colchicinic acidcohorts, the decrease in T1 relaxation is purely due to Gd 3DOTA 3Colchicinic acid acting as a contrast agent rather thaneffects caused by therapeutic effect, as these were not apparent inthe nongadolinium derivative, which presented necrosis scoressimilar to that of Gd 3DOTA.colchicinic acid by histology. AlthoughGd 3DOTA also contains gadolinium, the nature of the moleculemeans that it has a short half-life and rapid clearance; therefore, thefact that there was no change in T1 for this cohort suggests washoutwithin the initial 2 h period. Morphological changes correspondingto necrosis were detected within the tumor at 24 h onMR images inthe Gd 3DOTA.Colchcicinic acid cohort only. Darkening of areaswithin the originally homogeneous tumor presentation were de-tected mainly within the center of the tumor and correlated wellwith the degree of necrosis by histology. LA-ICP-MS data revealedhigh concentrations of gadolinium throughout the tumor. UsingZn66 as a marker of viable tissue shows that areas assumed to benecrotic have high degrees of gadolinium present and thereforemayrepresent previously well vascularised areas where Gd 3DOTA 3Colchicinic acid is still bound within the occluded vessels. Theincreased contrast between what is believed as the necrotic core andviable rim on images for the Gd 3DOTA 3Colchicinic acid cohort ismost likely due to perfusion in the presence of bound Gd 3DOTA 3Colchicinic acid and is enough to enhance the signal intensity onT1images, whereas in areas of necrosis and increased paramagneticsubstances due to therapy, such as increased permeability, plateletaggregation, coagulation, and hemorrhagic necrosis, it causes dar-kening on the image and therefore increases contrast between thetwo areas.40 The liver also showed signal attenuation due toGd 3DOTA 3Colchicinic acid accumulation, whereas no changeswere observed in kidneyT1. As histology did not show any structuralabnormalities and therefore no cyctoxicity to normal tissue in eitherorgan, this may indicate the route of excretion as being via the liverand bile, which has also been shown for ZD6126 when tagged withcarbon 14.41,42

The therapeutic effect of Gd 3DOTA 3Colchicinic acid istherefore comparable to changes induced by both CA-4-P andZD6126 by MR where CA-4-P studies have shown early changesin signal with recovery to baseline at 24 h in both animal modelsand humans at clinically relevant doses,17 whereas ZD6126studies show no evidence of tumor recovery at 24 h16,43 or upto 96 h after administration,43 suggesting as with the in vitro datathat the Gd 3DOTA.Colchicine compound binds irreversibly in asimilar manner to ZD6126.38

As well as carbon 14, colchicine has previously been con-jugated to a wide range of imaging tracers, including Gd 3DOTA,using the same demethylation of the methoxy group adjacent to thecarbonyl.26�32 These studies are mainly in vitro or if in vivo haverequired animal sacrifice at each end point, such as the radiolabeledstudies,26,30,31,41,42 and therefore do not provide longitudinal in vivodata. Studies that retained the toxicity of colchicine such as studiesby Clark et al. evaluated the binding and subsequent cell changes invitro but could not be used in vivo.27 In other studies that reduced thetoxicity of colchicines, although showing increased uptake andretention in vitro32 and in vivo,26 were unclear as to whether therewas any significant therapeutic response.26

In conclusion, we have successfully synthesized and character-ized a novel derivative of colchicine that is capable of functioningas a competent imaging tracer and as a therapeutic tubulin-binding agent in vitro and in vivo. The utilization of this hasallowed for in vivo imaging of central necrosis as a therapeutic

Figure 5. Hematoxylin and eosin stained sections of OVCAR-3 tumors(A,B) and normal liver parenchyma (C,D) 24 h after the administrationof (A and C) saline and (B) 200 mg/kg Gd 3DOTA 3Colchicinic acid.Magnification is �100.



response to VDA therapy in tumors in real time and has beenvalidated by both histology and LA-ICP-MS. This has given abetter understanding of the timing of necrosis and has strength-ened the requirement for smart probes to be incorporated intothe experimental design for in vivo applications.

’AUTHOR INFORMATION

Corresponding Author*Tel:þ44 208 383 3497. Fax:þ44 208 383 1511. E-mail: [email protected].

’ACKNOWLEDGMENT

Funding for T.K. was provided by the MRC. The GTC issupported by EPSRC grant support and is also grateful for recentsupport from ImuThes Limited.

’REFERENCES

(1) Vaupel, P., Kallinowski, F., and Okunieff, P. (1989) Blood flow,oxygen and nutrient supply, and metabolic microenvironment of humantumors: a review. Cancer Res. 49, 6449–6465.(2) Denekamp, J. (1993) Review article: angiogenesis, neovascular

proliferation and vascular pathophysiology as targets for cancer therapy.Br. J. Radiol. 66, 181–196.(3) Carmeliet, P., and Jain, R. K. (2000) Angiogenesis in cancer and

other diseases. Nature 407, 249–257.(4) Bergers, G., and Benjamin, L. E. (2003) Tumorigenesis and the

angiogenic switch. Nat. Rev. Cancer 3, 401–410.(5) Siemann, D. W., Chaplin, D. J., and Horsman, M. R. (2004)

Vascular-targeting therapies for treatment of malignant disease. Cancer100, 2491–2499.(6) Siemann, D. W., Bibby, M. C., Dark, G. G., Dicker, A. P., Eskens,

F. A., Horsman, M. R., Marme, D., and Lorusso, P. M. (2005)Differentiation and definition of vascular-targeted therapies. Clin. CancerRes. 11, 416–420.(7) Micheletti, G., Poli, M., Borsotti, P., Martinelli, M., Imberti, B.,

Taraboletti, G., and Giavazzi, R. (2003) Vascular-targeting activity ofZD6126, a novel tubulin-binding agent. Cancer Res. 63, 1534–1537.(8) Horsman, M. R., Ehrnrooth, E., Ladekarl, M., and Overgaard, J.

(1998) The effect of combretastatin A-4 disodium phosphate in a C3Hmouse mammary carcinoma and a variety of murine spontaneoustumors. Int. J. Radiat. Oncol. Biol. Phys. 42, 895–898.(9) Blakey, D. C., Westwood, F. R., Walker, M., Hughes, G. D.,

Davis, P. D., Ashton, S. E., and Ryan, A. J. (2002) Antitumor activity ofthe novel vascular targeting agent ZD6126 in a panel of tumor models.Clin. Cancer Res. 8, 1974–1983.(10) Tozer, G. M., Prise, V. E., Wilson, J., Locke, R. J., Vojnovic, B.,

Stratford, M. R., Dennis, M. F., and Chaplin, D. J. (1999) Combretas-tatin A-4 phosphate as a tumor vascular-targeting agent: early effects intumors and normal tissues. Cancer Res. 59, 1626–1634.(11) Siemann, D. W., and Horsman, M. R. (2002) Enhancement of

radiation therapy by vascular targeting agents. Curr. Opin. Invest. Drugs3, 1660–1665.(12) Dark, G. G., Hill, S. A., Prise, V. E., Tozer, G. M., Pettit, G. R.,

and Chaplin, D. J. (1997) Combretastatin A-4, an agent that displayspotent and selective toxicity toward tumor vasculature. Cancer Res.57, 1829–1834.(13) Dowlati, A., Robertson, K., Cooney,M., Petros,W. P., Stratford,

M., Jesberger, J., Rafie, N., Overmoyer, B., Makkar, V., Stambler, B.,Taylor, A., Waas, J., Lewin, J. S., McCrae, K. R., and Remick, S. C. (2002)A phase I pharmacokinetic and translational study of the novel vasculartargeting agent combretastatin a-4 phosphate on a single-dose intrave-nous schedule in patients with advanced cancer. Cancer Res. 62,3408–3416.

(14) Beauregard, D. A., Hill, S. A., Chaplin, D. J., and Brindle, K. M.(2001) The susceptibility of tumors to the antivascular drug combre-tastatin A4 phosphate correlates with vascular permeability. Cancer Res.61, 6811–6815.

(15) Maxwell, R. J., Wilson, J., Prise, V. E., Vojnovic, B., Rustin, G. J.,Lodge, M. A., and Tozer, G. M. (2002) Evaluation of the anti-vasculareffects of combretastatin in rodent tumours by dynamic contrastenhanced MRI. NMR Biomed. 15, 89–98.

(16) Evelhoch, J. L., LoRusso, P. M., He, Z., DelProposto, Z., Polin,L., Corbett, T. H., Langmuir, P., Wheeler, C., Stone, A., Leadbetter, J.,Ryan, A. J., Blakey, D. C., and Waterton, J. C. (2004) Magneticresonance imaging measurements of the response of murine and humantumors to the vascular-targeting agent ZD6126. Clin. Cancer Res.10, 3650–3657.

(17) Galbraith, S. M., Maxwell, R. J., Lodge, M. A., Tozer, G. M.,Wilson, J., Taylor, N. J., Stirling, J. J., Sena, L., Padhani, A. R., and Rustin,G. J. (2003) Combretastatin A4 phosphate has tumor antivascularactivity in rat and man as demonstrated by dynamic magnetic resonanceimaging. J. Clin. Oncol. 21, 2831–2842.

(18) Robinson, S. P., McIntyre, D. J., Checkley, D., Tessier, J. J.,Howe, F. A., Griffiths, J. R., Ashton, S. E., Ryan, A. J., Blakey, D. C., andWaterton, J. C. (2003) Tumour dose response to the antivascular agentZD6126 assessed by magnetic resonance imaging. Br. J. Cancer 88,1592–1597.

(19) Robinson, S. P., Kalber, T. L., Howe, F. A., McIntyre, D. J.,Griffiths, J. R., Blakey, D. C., Whittaker, L., Ryan, A. J., and Waterton,J. C. (2005) Acute tumor response to ZD6126 assessed by intrinsicsusceptibility magnetic resonance imaging. Neoplasia 7, 466–474.

(20) Thoeny, H. C., De Keyzer, F., Chen, F., Vandecaveye, V.,Verbeken, E. K., Ahmed, B., Sun, X., Ni, Y., Bosmans, H., Hermans, R.,van Oosterom, A., Marchal, G., and Landuyt, W. (2005) Diffusion-weighted magnetic resonance imaging allows noninvasive in vivo mon-itoring of the effects of combretastatin a-4 phosphate after repeatedadministration. Neoplasia 7, 779–787.

(21) Thoeny, H. C., De Keyzer, F., Chen, F., Ni, Y., Landuyt, W.,Verbeken, E. K., Bosmans, H., Marchal, G., and Hermans, R. (2005)Diffusion-weighted MR imaging in monitoring the effect of a vasculartargeting agent on rhabdomyosarcoma in rats. Radiology 234, 756–764.

(22) Wang, H., Sun, X., Chen, F., De Keyzer, F., Yu, J., Landuyt, W.,Vandecaveye, V., Peeters, R., Bosmans, H., Hermans, R., Marchal, G.,and Ni, Y. (2009) Treatment of rodent liver tumor with combretastatina4 phosphate: noninvasive therapeutic evaluation using multiparametricmagnetic resonance imaging in correlation with microangiography andhistology. Invest. Radiol. 44, 44–53.

(23) Ni, Y., Bormans, G., Chen, F., Verbruggen, A., and Marchal, G.(2005) Necrosis avid contrast agents: functional similarity versusstructural diversity. Invest. Radiol. 40, 526–535.

(24) Madhu, B., Waterton, J. C., Griffiths, J. R., Ryan, A. J., andRobinson, S. P. (2006) The response of RIF-1 fibrosarcomas to thevascular-disrupting agent ZD6126 assessed by in vivo and ex vivo 1Hmagnetic resonance spectroscopy. Neoplasia 8, 560–567.

(25) Zhao, M., Beauregard, D. A., Loizou, L., Davletov, B., andBrindle, K. M. (2001) Non-invasive detection of apoptosis usingmagnetic resonance imaging and a targeted contrast agent. Nat. Med.7, 1241–1244.

(26) Satpati, D., Korde, A., Pandey, U., Dami, P., Banerjee, S., andVenkatesh, M. (2006) Synthesis and evaluation of 90Y-DOTA-Colchi-cine conjugate in murine fibrosarcoma model. J. Labelled Compd.Radiopharm. 49, 951–958.

(27) Clark, J. I., and Garland, D. (1978) Fluorescein colchicine.Synthesis, purification, and biological activity. J. Cell Biol. 76, 619–627.

(28) Bagnato, J. D., Eilers, A. L., Horton, R. A., and Grissom, C. B.(2004) Synthesis and characterization of a cobalamin-colchicine con-jugate as a novel tumor-targeted cytotoxin. J. Org. Chem. 69, 8987–8996.

(29) van Rossenberg, S.M., Sliedregt-Bol, K.M., Koning, G., van denElst, H., van Berkel, T. J., van Boom, J. H., van der Marel, G. A., andBiessen, E. A. (2003) Improvement of hepatocyte-specific gene expres-sion by a targeted colchicine prodrug. ChemBioChem 4, 633–639.



(30) Zareneyrizi, F., Yang, D. J., Oh, C. S., Ilgan, S., Yu, D. F., Tansey,W., Liu, C. W., Kim, E. E., and Podoloff, D. A. (1999) Synthesis of[99mTc]ethylenedicysteine-colchicine for evaluation of antiangiogeniceffect. Anticancer Drugs 10, 685–692.(31) Korde, A., Satpati, D., Mathur, A., Mallia, M., Banerjee, S.,

Kothari, K., Sarma, H. D., Choudhari, P., and Venkatesh, M. (2006)99mTc-labeling of colchicine using [99mTc(CO)3(H2O)3]þ and[99mTc triple bond N]2þ core for the preparation of potentialtumor-targeting agents. Bioorg. Med. Chem. 14, 793–799.(32) Efthimiadou, E. K., Katsarou, M. E., Fardis, M., Zikos, C., Pitsinos,

E. N., Kazantzis, A., Leondiadis, L., Sagnou, M., and Vourloumis, D. (2008)Synthesis and characterization of novel natural product-Gd(III) MRIcontrast agent conjugates. Bioorg. Med. Chem. Lett. 18, 6058–6061.(33) Barge, A., Cravotto, G., Gianolio, E., and Fedeli, F. (2006) How

to determine free Gd and free ligand in solution of Gd chelates. Atechnical note. Contrast Media Mol. Imaging 1, 184–188.(34) Rasband, W. S. (1997�2008) ImageJ, U.S. National Institutes

of Health, Bethesda, MD, http://rsb.info.nih.gov/ij/.(35) Ahmed, B., Van Eijk, L. I., Bouma-Ter Steege, J. C., Van Der

Schaft, D. W., Van Esch, A. M., Joosten-Achjanie, S. R., Lambin, P.,Landuyt, W., and Griffioen, A. W. (2003) Vascular targeting effect ofcombretastatin A-4 phosphate dominates the inherent angiogenesisinhibitory activity. Int. J. Cancer 105, 20–25.(36) Davis, P. D., Dougherty, G. J., Blakey, D. C., Galbraith, S. M.,

Tozer, G. M., Holder, A. L., Naylor, M. A., Nolan, J., Stratford, M. R.,Chaplin, D. J., and Hill, S. A. (2002) ZD6126: a novel vascular-targetingagent that causes selective destruction of tumor vasculature. Cancer Res.62, 7247–7253.(37) Galbraith, S. M., Chaplin, D. J., Lee, F., Stratford, M. R., Locke,

R. J., Vojnovic, B., and Tozer, G. M. (2001) Effects of combretastatin A4phosphate on endothelial cell morphology in vitro and relationship totumour vascular targeting activity in vivo. Anticancer Res. 21, 93–102.(38) Lin, C. M., Ho, H. H., Pettit, G. R., and Hamel, E. (1989)

Antimitotic natural products combretastatin A-4 and combretastatinA-2: studies on the mechanism of their inhibition of the binding ofcolchicine to tubulin. Biochemistry 28, 6984–6991.(39) McGown, A. T., and Fox, B. W. (1989) Structural and

biochemical comparison of the anti-mitotic agents colchicine, combre-tastatin A4 and amphethinile. Anti-Cancer Drug Des. 3, 249–254.(40) Tozer, G. M., Kanthou, C., and Baguley, B. C. (2005) Disrupt-

ing tumour blood vessels. Nat. Rev. Cancer 5, 423–435.(41) Partridge, E. A., D’Souza, R. A., Lenz, E. M., Smith, S. M.,

Clarkson-Jones, J., and Roberts, D. W. (2008) Disposition and meta-bolism of the colchicine derivative [14C]-ZD6126 in rat and dog.Xenobiotica 38, 399–421.(42) D’Souza, R. A., Partridge, E. A., Roberts, D. W., Ashton, S.,

Ryan, A., Patterson, A. B., Wilson, Z., and Thurrell, C. C. (2007)Distribution of radioactivity and metabolite profiling in tumour andplasma following intravenous administration of a colchicine derivative(14C-ZD6126) to tumour-bearing mice. Xenobiotica 37, 328–340.(43) McIntyre, D. J., Robinson, S. P., Howe, F. A., Griffiths, J. R.,

Ryan, A. J., Blakey, D. C., Peers, I. S., and Waterton, J. C. (2004) Singledose of the antivascular agent, ZD6126 (N-acetylcolchinol-O-phos-phate), reduces perfusion for at least 96 h in the GH3 prolactinomarat tumor model. Neoplasia 6, 150–157.

Synthesis and Characterization of a Theranostic Vascular Disrupting Agent for In Vivo MR Imaging

Documents

Transcript of Synthesis and Characterization of a Theranostic Vascular Disrupting Agent for In Vivo MR Imaging