360 | VOLUME 60 | NUMBER 2 | FEBRUARY 2007 www.neurosurgery-online.com

EXPERIMENTAL STUDIESSarah C. Jost, M.D.Department of Neurosurgery,Washington University,School of Medicine,St. Louis, Missouri

John E. Wanebo, M.D.Department of Neurosurgery,Uniformed Services University of theHealth Sciences National NavalMedical CenterBethesda, Maryland

Sheng-Kwei Song, Ph.D.Department of Radiology,Washington University,School of Medicine,St. Louis, Missouri

Michael R. Chicoine, M.D.Department of Neurosurgery,Washington University,School of Medicine,St. Louis, Missouri

Keith M. Rich, M.D.Department of Neurosurgery,Washington University,School of Medicine,St. Louis, Missouri

Thomas A. Woolsey, M.D.Department of Neurosurgery,Washington University,School of Medicine,St. Louis, Missouri

Jason S. Lewis, Ph.D.Department of Radiology andWashington University,School of Medicine,St. Louis, Missouri

Robert H. Mach, Ph.D.Department of Radiology andWashington University,School of Medicine,St. Louis, Missouri

Jinbin Xu, Ph.D.Department of Radiology,Washington University,School of Medicine,St. Louis, Missouri

Joel R. Garbow, Ph.D.Departments of Radiology andChemistry,Washington University,School of Medicine,St. Louis, Missouri

The Alvin J. Siteman Cancer Center(JSL, RHM, JRG)

Reprint requests:Joel R. Garbow, Ph.D.,Department of Radiology,Washington University,School of Medicine,4525 Scott Avenue,Campus Box 8227,St. Louis, MO 63110.Email: garbow@wustl.edu

Received, March 29, 2006.Accepted, September 18, 2006.

Glioblastoma (GBM) is the most commonprimary malignant neoplasm of theadult brain (62, 65). Treatment outcomes

are poor: median survivals are only 1 year (19).GBMs are highly invasive and tumor cells arefound in areas of the brain at significant dis-tances from the bulk tumor, making surgicalresection especially challenging (34, 54, 60). Thelimitations of current treatments make the devel-opment of new therapies imperative. Suchadvances depend, in part, on animal models thataccurately mimic human glioblastomas (28).

The evaluation of animal models requiresquantitative methods to assess the efficacy ofpossible therapeutic interventions. Small-animal imaging, including magnetic resonanceimaging (MRI), positron emission tomography(PET), and optical methods, can follow tumorprogression in vivo (39). Combination of thesemethods will allow quantification of intracra-nial tumor growth and monitoring tumor biol-ogy and will permit longitudinal assessmentof the effects of therapeutic interventions invivo. Complementing this approach are meth-

IN VIVO IMAGING IN A MURINE MODELOF GLIOBLASTOMA

OBJECTIVE: To use in vivo imaging methods in mice to quantify intracranial gliomagrowth, to correlate images and histopathological findings, to explore tumor markerspecificity, to assess effects on cortical function, and to monitor effects of chemother-apy.METHODS: Mice with DBT glioma cell tumors implanted intracranially were imagedserially with a 4.7-T small-animal magnetic resonance imaging (MRI) scanner. MRItumor volumes were measured and correlated with postmortem histological findings.Different nonspecific and specific positron emission tomography radiopharmaceuti-cals, [18F]2-fluoro-2-deoxy-D-glucose, [18F]3�-deoxy-3�-fluorothymidine, or [11C]RHM-I, a σ2-receptor ligand, were visualized with microPET (CTI-Concorde MicroSystemsLLC, Knoxville, TN). Intrinsic optical signals were imaged serially during contralateralwhisker stimulation to study the impact of tumor growth on cortical function. Othergroups of mice were imaged serially with MRI after one or two doses of the antimitoticN,N�-bis(2-chloroethyl)-N-nitrosourea (BCNU).RESULTS: MRI and histological tumor volumes were highly correlated (r 2 � 0.85).Significant binding of [11C]RHM-I was observed in growing tumors. Over time, tumorsreduced and displaced (P � 0.001) whisker-activated intrinsic optical signals but did notchange intrinsic optical signals in the contralateral hemisphere. Tumor growth wasdelayed 7 days after a single dose of BCNU and 18 days after two doses of BCNU.Mean tumor volume 15 days after DBT implantation was significantly smaller for treatedmice (1- and 2-dose BCNU) compared with controls (P � 0.0026).CONCLUSION: Mouse MRI, positron emission tomography, and optical imaging pro-vide quantitative and qualitative in vivo assessments of intracranial tumors that corre-late directly with tumor histological findings. The combined imaging approach pro-vides powerful multimodality assessments of tumor progression, effects on brain function,and responses to therapy.

KEY WORDS: Glioma, Magnetic resonance imaging, Optical imaging, Positron emission tomography,Therapeutic response

Neurosurgery 60:360–371, 2007 DOI: 10.1227/01.NEU.0000249264.80579.37 www.neurosurgery-online.com

blue dye, and counted with a hemocytometer. The cells wererecentrifuged and suspended in serum-free DMEM to a con-centration of 80,000 or 100,000 DBT cells/µl and cooled on icebefore injection.

Animal SurgerySeventy-four adult BALB/c mice weighing approximately

25 g were studied. All animal protocols were approved bythe Washington University Division of Comparative Medicineand met or exceeded American Associat ion for theAccreditation of Laboratory Animal Care International andNational Institutes of Health standards. The mice were anes-thetized with intraperitoneal ketamine (25 mg/kg ), xylazine(5 mg/kg), and acepromazine (2.5 mg/kg) before intracranialDBT cell implantation. After a midline scalp incision, a cran-iotomy was made over the cortex with a 1-mm cutting burr.Mice were secured in a stereotactic frame and 5 µl DBT tumorcell suspension were aspirated into a Hamilton syringeattached to the frame which was then passed into the brain.For the optical imaging studies, the whisker-barrel cortex(first somatosensory cortex) was targeted using cranium andvascular landmarks, and the syringe was advanced to 750 µmbelow the cortical surface. 2.0 � 105 tumor cells were injectedduring a 3-minute time period. For the BCNU treatment, PETexperiments, and some of the volumetric and MRI studies,2.5 � 105 tumor cells were injected into the striatum at a site2 mm lateral and 2 mm posterior to the bregma and 4 mmdeep to the cortical surface. After the needle was removed, thecraniotomy was sealed with bone wax and the scalp wasclosed with super glue or a metal staple that was laterremoved before MRI examination. After surgery, mice werehoused in nonbarrier animal facilities for the duration of theimaging experiments. Although animals with large braintumors are more likely to die under anesthesia than healthyanimals, mice were successfully and routinely imaged at mul-tiple time points throughout this study.

MRIImages were collected in an Oxford Instruments 4.7-T magnet

(33 cm, clear bore; Oxford, United Kingdom) equipped with 15-cminner diameter, actively shielded gradient coils (maximum gradi-ent, 18 G/cm; rise time, 100 µs). The magnet/gradients were inter-faced with a Varian INOVA console (Palo Alto, CA), and datawere collected using a 1.5-cm outer diameter surface coil (receive)and a 9-cm inner diameter Helmholtz coil (transmit). Before theimaging experiments, mice were anesthetized with isoflurane/O2

(4% [v/v]) and maintained on isoflurane/O2 (1.5% [v/v])throughout the experiments.

Initially, data were collected using T2- and diffusion-weighted, spin-echo, multislice imaging sequences with a1.5 � 1.5-cm2 coronal field of view (T2-weighted: repetitiontime [TR], 3 s; echo time [TE], 50 ms; slice thickness, 0.5 mm;diffusion-weighted imaging: TR, 1.5 s; TE, 42 ms; slice thick-ness, 1 mm). In later imaging studies, mice were injectedintraperitoneally with 500 µl Omniscan (Gadodiamide, GE

ods that assess function, including imaging intrinsic opticalsignals (IOSs).

Visualizing the in vivo progress of tumor growth has been alimiting factor in the characterization of many tumor models.Small-animal MRI has offered a potential solution to this chal-lenging problem (1, 5–7, 10, 44, 52, 55, 67, 78). LongitudinalMRI studies provide a practical approach for identifyingintracranial glioma tumors and following tumor progression insmall animals (1, 6, 28, 44, 57). Even when symptomatic withbrain tumors, mice generally tolerate serial MRI studies (6, 44).Recent work has shown that calculations of intracranial tumorvolumes from MRI correlate well with postmortem tumor vol-umes (16, 55). Success in imaging murine brain tumors usingclinical MRI scanners (7, 48, 67) and in imaging multiple micein a single experiment have also been reported (5, 78).

Although early in vivo studies of brain tumors in mice usedsubcutaneously implanted tumor cells (32, 37, 46), more recentstudies have used cultured tumor cells implanted directly inthe brain (1, 3, 13, 14, 44, 45, 51, 57). Here, we used implantedDBT cells, cells that are similar to human GBMs in their aggres-sive growth pattern, histopathological features, and immunos-taining (1, 15, 57, 76).

In this study, small-animal MRI scanning was used to char-acterize longitudinal DBT tumor growth quantitatively in mice.PET was used to assess biologically relevant markers of GBMs,including metabolism, mitoses, and tumor-specific σ2 recep-tors in vivo. IOS was used to evaluate the effects of GBM onsensory function of the brain in response to whisker stimula-tion. Finally, the efficacy of therapeutic interventions on gliomagrowth in mice with single and multiple doses of N,N�-bis(2-chloroethyl)-N-nitrosourea (BCNU) was measured.

MATERIALS AND METHODS

Mouse DBT Cell LineThe DBT glioblastoma cell line used in these experiments

was donated by Dr. Michael Lai, Department of Microbiology,University of Southern California. DBT cells frozen in culturemedium with 5% dimethyl sulfoxide were thawed, washedwith magnesium-free Hank’s balanced salt solution, and grownin Dulbecco’s modified Eagle’s medium (DMEM) with 15%heat-inactivated fetal calf serum, 0.2 mmol/L glutamine,50 mg/ml neomycin, 100 mg/ml penicillin, and 100 mg/mlstreptomycin. The cells were plated in T-75 culture flasks forincubation in a 5% CO2 humidified atmosphere at 37�C andused at low passages.

Preparation of Cells for InjectionFeeding medium was aspirated from the culture flasks and

3 ml of 0.5% trypsin-ethylenediamine tetra-acetic acid wereadded during a 1-minute time period. After the cells were sus-pended, trypsin was inactivated by adding 7 ml of DMEM withfetal calf serum. The suspended cells were centrifuged at250 � g for 10 minutes at 4�C and resuspended in 2 ml ofDMEM. An aliquot of cells was aspirated, stained with trypan

NEUROSURGERY VOLUME 60 | NUMBER 2 | FEBRUARY 2007 | 361

MONITORING BRAIN TUMORS IN VIVO WITH SMALL ANIMAL IMAGING

Healthcare, Little Chalfont, United Kingdom) contrast agent,diluted 1:10 in sterile saline, 15 minutes before being placed inthe magnet. T1-weighted, gradient-echo multislice images(coronal and horizontal views) were collected (TR, 0.125 s; TE,0.0025 s; field of view [coronal], 1.5 � 1.5 cm2; field of view[horizontal], 3 � 3 cm2; slice thickness, 0.5 mm). In addition,T2-weighted, multislice spin-echo coronal images were col-lected for each animal (TR, 1.5 s; TE, 0.05 s; field of view,1.5 � 1.5 cm2; thickness, 0.5 mm). T2-weighted images wereused when it was difficult to see tumor margins. Tumor vol-umes were measured by manual outlining tumors usingVarian’s ImageBrowser software (Varian NMR Instruments,Inc., Palo Alto, CA) or post hoc with the public domain programNIH Image (available at http://rsb.info.nih.gov/nih-image).

To compare data from different experimental groups, tumorvolumes were first normalized by dividing individual values ina particular case by the sum of all values for that case. Thesenormalized values were used to compute means and standarddeviations or ranges (for BCNU, �2). Computed values andvariations then were scaled by calculating the ratio of the finallargest values in a series to one (1) and multiplying all values inthe series by this scaling factor. None of these procedureschanges the exponents of the fitted curves.

PETImages were collected on either a microPET-Focus-220 or a

microPET-Focus-120 (CTI-Concorde MicroSystems LLC;Knoxville, TN) tomograph (63). Coregistration of the PET imageswas achieved in combination with a microCAT-II camera (ImtekInc., Knoxville, TN) for high-resolution computed tomographic(CT) images. Mice were anesthetized with 1 to 2% isofluranebefore scanning and were immobilized in the supine position ina custom cradle. Two mice were imaged side by side in the samebed position at all time points. Image registration betweenmicroCT and PET images was accomplished using a landmarkregistration technique and AMIRA image display software(AMIRA; TGS, Inc., San Diego, CA). The registration methodproceeds by rigid transformation of the microCT images fromlandmarks provided by fiducial markers directly attached to theanimal bed. This coregistration method allows for accurate reg-istration of CT and PET images to within �0.5 mm.

Analysis of PET ImagesMice were imaged at different time points during tumor

development with [18F]2-fluoro-2-deoxy-D-glucose ([18F]FDG)(42, 61) to monitor tumor metabolism, 18F-3�-deoxy-3�-fluo-rothymidine ([18F]FLT) to measure thymidine kinase-1 activ-ity (58, 59), and a 11C-labeled σ2 receptor ligand ([11C]RHM-I)(66) to evaluate proliferation. [18F]FDG was prepared usingthe Coincidence Technologies [18F]FDG synthesis module (GEHealthcare). [18F]FLT (26) and [11C]RHM-I were producedaccording to published protocols (66). Two groups of animalswere studied. The first group of animals (n � 8) was imagedwith [18F]FDG on a weekly basis over a 3-week period afterDBT cell injection. The second group (n � 4) was imaged dur-ing active tumor development with both [18F]FLT and [11C]-

RHM-I (24 h apart) then again the following week with only[11C]RHM-I. For [18F]FDG imaging, animals were fastedovernight, injected via a tail-vein catheter with approximately0.5 mCi of [18F]FDG, and imaged at 1 hour after injection(1 � 10 min frame). For [18F]FLT imaging, animals wereinjected with approximately 0.25 mCi [18F]FLT and imaged at1 hour after injection (1 � 10 min frame). For [11C]RHM-Iimaging, animals were injected with 0.4 to 0.5 mCi of the 11C-labeled compound and were imaged at 30 minutes and 1 hourafter injection (1 � 15 min frame). Following PET imagingprotocols, selected mice were imaged for coregistration onthe microCAT.

Ligand SpecificityThe tritiated compound [3H]RHM-1 was synthesized by

American Radiolabeled Chemicals, Inc. (St. Louis, MO) via O-alkylation of the corresponding phenol precursor (66); chem-ical purity was greater than 99% and the specific activity of theradioligand was 80 Ci/mmol. Membrane homogenates wereprepared from approximately 4-g DBT tumors that wereremoved from previously implanted mice and frozen immedi-ately on dry ice and stored at �80�C, thawed slowly on ice, andhomogenized at 4�C with a Potter-Elvehjem tissue grinder(Wheaton Science Products, Millville, NJ) at a concentration of1 g tissue/ml of 50 mmol/L Tris-HCl at pH 8.0. The crudemembrane homogenate was then transferred to a 50-ml cen-trifuge tube and resuspended to a concentration of 0.2 g of tis-sue/ml of 50 mmol/L Tris-HCl. Additional homogenizationwas accomplished using an Ultra-Turrax T8 polython homog-enizer (IKA Works, Inc., Wilmington, NC). The finalhomogenate then was centrifuged for 10 minutes at 1000 � g,the pellet was discarded, and the supernatant was mixed byvortexing. Aliquots were stored at �80�C until use. The proteinconcentration of the suspension was determined using the DCprotein assay (Bio-Rad, Hercules, CA) and averaged approxi-mately 10 mg protein/ml stock solution.

Approximately 350 µg of membrane homogenates werediluted with 50 mmol/L Tris-HCl buffer, pH 8.0, and incubatedwith [3H]RHM-1 in a total volume of 150 µl at 25�C in 96-wellpolypropylene plates (Fisher Scientific, Pittsburgh, PA). Theconcentrations of the radioligand ranged from 0.1 to 8 nM.After incubation for 60 minutes, the reactions were terminatedby the addition of 150 µl of cold wash buffer (10 mmol/L Tris-HCl, 150 mmol/L NaCl, pH 7.4, at 4�C) with a 96-channeltransfer pipette (Fisher Scientific), and the samples were har-vested and filtered rapidly to 96-well fiberglass filter plates(Millipore, Billerica, MA) presoaked with 100 µl 50 mmol/LTris-HCl buffer, pH 8.0, for 1 hour. Each filter was washed threetimes with 200 µl of ice-cold wash buffer. A Wallac 1450MicroBeta liquid scintillation counter (Perkin Elmer, Boston,MA) was used to quantitate the bound radioactivity (77).Nonspecific binding was determined from samples that con-tained 10 µM RHM-1. The equilibrium dissociation constant(Kd) and maximum number of binding sites (Bmax) were deter-mined by a linear regression analysis of the transformed datausing the method of Scatchard (53).

362 | VOLUME 60 | NUMBER 2 | FEBRUARY 2007 www.neurosurgery-online.com

JOST ET AL.

Optical Imaging of Functional ResponsesFunctional optical imaging experiments were performed at

two different time points at least 1 week apart in 23 animals(between 6 and 26 d after tumor injection). Details of the imagingsystem and analysis are provided elsewhere (20, 22). Illuminationwas provided by focused (Edmund Scientific, Tonawanda, NY)filtered light (520–560 nm) from a 150-W, 21-V EKE light bulb(Opti-Quip, Highland Mills, NY) with a regulated power supply.The voltage was stabilized by placing 10 0.29-F capacitors in par-allel with the power supply. The image of the brain surface wasmagnified �0.75 to �3 with a Nikon SMZ-U dissecting micro-scope (Nikon Corp., Melleville, NY) while focusing on surfacevessels through the cranium and captured with a charge-cou-pled device camera (72S; Hamamatsu Corp., Bridgewater, NJ) at30 Hz. These 640 � 480-pixel images were stored with a real-timecapture card (AG-5; Scion Corporation, Frederick, MD) in aPower Macintosh G3 Computer (Apple Computer, Inc.,Cupertino, CA) running NIH Image 1.62, optimized by modify-ing the source code. The IOS was averaged from 20 trials of 3 sec-onds of stimulation, and the prestimulation signal was subtractedto yield optical images. Both the normal and tumor-injected hemi-spheres of each animal were imaged during a single recordingsession.

IOS Response AssessmentIOS images from control and tumor-involved cortices were

rank-ordered in a blinded analysis. Color prints of all 78 IOSimages (46 from 23 tumor-injected hemispheres and 32 from 16uninjected hemispheres from a subset of these animals) wereprinted at the same magnification and coded. These were shuf-fled as a deck of playing cards and three members of the labo-ratory not participating in the experiment were asked to sortthe images according to color, active (hot) to inactive (cold). Theranks were assigned to each image, and these were convertedto percentile rankings (hottest, 100; coldest, 0) for each observer.The 3rd percentile values for each image were averaged. Thepercentile scores were then sorted by condition and the time ofimaging. The average ranks of the first control images werecompared with the second control images and the ranks of thefirst tumor images were compared with the second tumorimages by Student’s t test. The barrel fields and vascular pat-terns reconstructed from the histological findings were alignedcarefully to these images (22).

Histological Analysis and Comparisonof MRI and Histological Volume

Fourteen animals were sacrificed and their brains were har-vested and stained for correlation with MRI findings. In onesubset of mice, seven out of eight animals with tumor injectionsin the striatum survived and were imaged serially 5, 8, 12, and15 days later. After the last day of imaging, they were sacrificedwith sodium pentobarbital. Brains were removed from the miceand were placed in 4% paraformaldehyde with 30% sucrose for24 to 48 hours at 4�C. The specimens were removed and cut inthe coronal plane. Fixed specimens were stained with hema-

toxylin and eosin and were correlated with MRI data at similarlevels (Fig. 1). In a second subset of mice, seven animals withtumor cells injected into the barrel cortex were imaged by MRIand perfusion on Day 15 before they were sacrificed. Thesebrains were cut parallel to the pia and were stained forcytochrome oxidase (57). Histological tumor volumes fromdigital microscopic images were measured with NIH Image.Whisker barrels and related images of the cortical surface werereconstructed by standard techniques (57).

Evaluation of Therapeutic Response with MRITwenty-four animals were randomized into three groups of

eight after tumor injection. Seven surviving mice in the firstgroup were not treated (vide supra). Animals in the secondgroup were treated with a single intraperitoneal dose of BCNU(10 mg/kg) on postoperative Day 7, and those in the thirdgroup were treated with two intraperitoneal doses of BCNU(10 mg/kg) given on postinjection Days 7 and 14. All untreatedanimals (n � 7) and those with 1 dose of BCNU (n � 8) under-went MRI evaluation on postoperative Days 5, 8, 12, and 15.Of the eight animals treated with a second dose of BCNU, foursurvived to be imaged on postinjection Days 12, 15, 19, and 22.For two of these animals, data also were collected 26, 29, and 33days after implantation.

RESULTS

All 74 animals showed tumor growth by MRI. Hydro-cephalus, brain herniation (related to tumor location and size),and BCNU toxicity contributed to animal death. Furthermore,animals with large tumors occasionally died when anesthetizedfor MRI, PET, and optical imaging experiments.

The results reported are from 51 mice that were studied atleast twice by imaging or with a combination of several meth-ods, that is, MRI and histological analysis. Tumors were visu-alized consistently either with a T1-weighted MRI sequencewith gadolinium contrast agent (n � 19) or combined T2/DWIspin-echo protocols (n � 32). Tumor volumes measured fromterminal MRI scans correlated significantly with those from thesame animals measured from histological findings (Fig. 1)(r2 � 0.85, data not shown).

Tumor-related hemorrhage and hydrocephalus were clear inT2-weighted spin-echo images. Figure 2 shows images fromone of three animals displaying moderate hydrocephalusbefore significant tumor growth.

PET imaging experiments with either [18F]FDG or [18F]FLTfailed to show consistent, differential uptake in the region ofthe tumor. In eight mice studied with [18F]FDG 2 weeks afterinjection of DBT cells, [18F]FDG uptake in tumor was indistin-guishable from that in normal brain. Only five mice survived to3 weeks; it was possible to delineate the tumor in just one ofthese animals. In the [18F]FLT studies, although the PET imageswere consistent with accumulation of [18F]FLT in the tumor(two out of four mice), interpretation of the images was con-founded by the uptake of the tracer in bone (58).

NEUROSURGERY VOLUME 60 | NUMBER 2 | FEBRUARY 2007 | 363

MONITORING BRAIN TUMORS IN VIVO WITH SMALL ANIMAL IMAGING

364 | VOLUME 60 | NUMBER 2 | FEBRUARY 2007 www.neurosurgery-online.com

JOST ET AL.

B

A

C

FIGURE 1. A, coronal T2-weighted, spin-echo MRI scan of a mouse justbefore sacrifice 15 days after injecting DBT cells into the right hemisphere. Amass effect, bright tumor, and hemorrhage are evident. B, photomicrograph ofa coronal section showing stained tumor and hemorrhage corresponding to thearea indicated by the rectangle in the MRI scan in A. Tumor cell infiltrationat the margins and mass effect are obvious (hematoxylin and eosin). C, higherpower image of B showing the rectangular region outlined in A and B.

FIGURE 2. The same anatomic slice plane from contrast-enhanced, T1-weighted gradient-echo image (A) and T2-weighted spin-echo image (B) of amouse 8 days after injection of DBT cells. Hydrocephalus was more evident inT2-weighted spin-echo images and was often seen before tumor was detected.

A B

As an alternative, a microPET strategy to assess cell prolifer-ation in vivo using [11C]RHM-I, a 11C-σ2 receptor tracer, wasused in four mice. The efficacy of RHM-I as a probe for σ2

receptors was first established in DBT cells by standardScatchard methods (53). Direct saturation binding studies werecarried out using [3H]RHM-1 with membrane homogenates of

FIGURE 3. Scatchard analysis of [3H]RHM-1 binding to the σ2 receptors inmembrane homogenates from DBT cells. A, Scatchard plot showing saturationbinding indicating the specific bound, total bound, and nonspecific boundprotein. B, Scatchard plot showing Kd and Bmax values.

B

A

DBT mouse brain tumor xenografts (Fig. 3). The Kd and Bmax

values of the receptor-radioligand binding of [3H]RHM-1 were0.76 nM and 1190 fmol/mg protein. The high affinity and lownonspecific binding indicate RHM-1 is a valuable probe forevaluating the σ2 receptor status of solid DBT tumors. In vivouptake of the 11C-σ2 receptor tracer was seen in tumors in fourof four mice, and localization was confirmed by the coregistra-tion of the PET and CT findings. (Image fusion is accuratewithin �0.5 mm using the landmark fiducial method). The res-olution of the 11C PET images is limited by the positron rangeof the nuclide (38) and the microPET camera resolution(∼1.7 mm) (63). Although lower than the resolution of the cor-responding MRI scans, the microPET images in Figure 4 showthat the region of increased tracer uptake closely matches theregion identified as tumor in the MRI scan.

Serial MRI measurements characterized tumor growth qual-itatively and quantitatively (Fig. 5) in untreated (Fig. 6) andBCNU-treated mice (Fig. 7). For both untreated groups, tumorgrowth kinetics were consistent. Tumor volumes were calcu-lated after manual tracing of all image slices containing tumorin each mouse brain at each time point (e.g., Fig. 5E). Figure 6A

shows a plot of tumor volumes versus time for untreated miceinjected with DBT cells in the striatum (therapeutic controls;2.5 � 105 cells) and the somatosensory/barrel cortex (IOS con-trols; 2 � 105 cells). Corresponding tumor growth curves afternormalization for differences in tumor volumes are shown inFigure 6B. Growth-rate exponents for therapeutic and IOScontrol mice, computed from measurement of tumor volumeson postoperative Days 7 and 15, are 0.264 � 0.085 and0.223 � 0.023, respectively (mean � standard deviation;P � 0.36, not significant).

Whisker-stimulation-evoked IOS in the barrel cortex of theinjected and contralateral hemispheres (with and withoutglioma) of DBT-injected animals was recorded at two differenttimes. There was no statistical difference between the percentileranks of IOS taken on two occasions at least 1 week apart (54thpercentile � 30 vs. 49th percentile � 28; P � 0.58, not signifi-cant) in the uninjected hemispheres of 16 animals. However, inthe 23 animals injected with DBT cells that were imaged twice,there was significant reduction of IOS signal related to the bulktumor (64th percentile � 26 vs. 38th percentile � 26; P � 0.001;Fig. 8), as confirmed by postmortem histological examination.

FIGURE 5. Similar anatomical slicesfrom contrast-enhanced, T1-weightedmultislice MRI scans of an untreatedmouse collected 5 days (A), 8 days(B), 12 days (C), and 15 days (D)after injection of DBT cells showrapid tumor growth. The tumor in Dis hemorrhagic. E, reproduction of Dwith the tumor outlined manually.

A B

C D

E

FIGURE 4. A and C, two slices from contrast-enhanced, T1-weighted MRIscans. B and D, matched microPET/CT images of the cell proliferation marker[11C]RHM-I, a σ2 receptor ligand. In this mouse 15 days after injection ofDBT cells, localized tracer binding in microPET largely coincides with tumorextent seen with MRI.

A B

C D

NEUROSURGERY VOLUME 60 | NUMBER 2 | FEBRUARY 2007 | 365

MONITORING BRAIN TUMORS IN VIVO WITH SMALL ANIMAL IMAGING

366 | VOLUME 60 | NUMBER 2 | FEBRUARY 2007 www.neurosurgery-online.com

JOST ET AL.

The effects of BCNU in slowing tumor growth were obviousfrom comparison of serial MRI scans of animals with and with-out BCNU treatment. Tumor growth kinetics and mean tumorvolumes for four different groups of animals (two controls; twotreated) are shown in Figure 7. A single dose of BCNU delayedthe exponential phase of tumor growth by approximately7 days, and two doses of BCNU delayed exponential growth byapproximately 18 days. Tumors in all animals eventuallyreturned to growth rates approaching those of untreated con-trols. The difference in tumor volume at 15 days after injectionin controls versus the combined group of treated mice (one andtwo doses of BCNU) was statistically significant (P � 0.0026).

DISCUSSION

Progress in adjunctive treatment of human glioblastoma dur-ing the past 25 years has been slow (19). Clearly, better under-standing of tumor microenvironment and its impact on tumorbiology, brain function, and treatment effectiveness is central todeveloping new therapeutic strategies. Treatments that limittumor cell proliferation, restrict metastases, increase apoptosis,or specifically target tumor cells with novel agents are logicaltargets for study. Small-animal imaging increases the power ofsuch investigations (32, 37, 46). In the current study, we injecteda mouse tumor cell line intracranially, and tumor cells grew inthe microenvironment of the normal brain. The combination ofimaging techniques described here allows longitudinal, in vivoassessment of tumor growth, brain function, and tumor biology.The power of small-animal MRI scans for measuring the effectof therapeutic interventions has been previously demonstrated(12, 36, 52). The present study is novel because a combination ofimaging methods was used to evaluate gliomas in mice.

DBT cells have morphological features similar to humanglioblastoma cells, are motile, are glial fibrillary acidic proteinpositive, and develop multicellular colonies when suspendedin agar. They have been characterized in a previous in vivostudy evaluating rat C6 cells and DBT cells injected into barrelcortex of rats and mice, respectively (57). In the present work,as in previous studies, intracranial neoplasms developed in allanimals with invasion away from bulk tumor (not shown).Invasion follows arteries, the principal pathway for the dis-semination of glioma cells (4, 24, 43, 57). Although previousstudies have demonstrated correlation between MRI resultsand histological measures of tumor volume (16), such a corre-lation had not been established previously for the DBT glioblas-toma model. This study demonstrated a significant, direct cor-relation between MRI results and histological measurements oftumor volumes.

The potential for PET as an imaging tool to detect specificmolecular markers or receptors for brain tumors in vivo is sig-nificant because changes seen with anatomical imaging are notnecessarily tumor specific. The 11C-labeled σ2 tracer [11C]RHM-I is under development as a PET radiotracer to image prolifer-ation in solid tumors (40, 66, 75). Vilner and Bowen (68) andVilner et al. (69) reported a high density of σ2 receptors in abroad panel of human and rodent tumors, including human

FIGURE 7. Relative tumor volumes in the different control groups plotted forcomparison with data from the two different treatment paradigms. Error bars,� standard deviation, except for BCNU 2�, which indicate data ranges.

FIGURE 6. Tumor volumes measured in two different groups of mice 1 and 2weeks after injection of DBT cells. A, graph showing averages (� standarddeviation) demonstrating that tumors in the striatum are larger (therapeuticcontrols, 2.5 � 105 DBT cells injected) than those in somatosensory cortex(IOS controls, 2 � 105 DBT cells). B, graph showing that after normalization(see Materials and Methods), the data from the two groups correlate.

(U-138MG glioblastoma) and rat (C6 glioma) brain tumors,whereas Al-Nabulsi et al. (2) measured the density of σ2 recep-tors in 9L glioma cells to be approximately 300,000receptors/cell. In this study, microPET scans performed using[11C]RHM-I showed considerable uptake of tracer in the regionof the brain correlating with the location of the tumor in all fourmice studied. [11C]RHM-I delineates tumors without accumu-lation of the tracer in normal brain (as with [18F]FDG) or bone(as with [18F]FLT). The excellent qualitative agreement of theregions of tracer uptake (PET) and tumor growth (MRI) empha-sizes the power of combining these imaging methods to char-acterize brain tumors.

[18F]FDG PET has been used widely in the study of braintumors and was, therefore, important to evaluate in the presentstudy. Clinical applications of PET imaging to human glioblas-toma therapy principally use [18F]FDG PET to distinguishrecurrent tumor from radiation necrosis, a distinction that isdifficult to make with CT or MRI examination. However,[18F]FDG PET lacks specificity for GBM (or any other tumortype). Increased uptake is confounded by other processes suchas seizures (70) and the high resting metabolism of normal

brain. We did not observe a significant increase in [18F]FDGuptake in the regions of mouse brain with DBT tumors, exceptin one mouse at a late stage of tumor growth. Radiolabelednucleosides, such as [18F]FLT, which measure the salvage path-way of DNA synthesis, are under investigation as potentialmarkers for cell proliferation (47, 58, 59). In this study, althoughit was possible to see accumulation of [18F]FLT in regions con-taining tumor in two of four mice, the interpretation of the PETimage was confounded by [18F]FLT uptake in the cranium.

Although definitive studies documenting that σ2 receptorbinding correlates quantitatively with histological measures ofcell proliferation (e.g., Ki-67 or bromodeoxyuridine labelingstudies) should be carried out, the present encouraging resultssuggest that microPET imaging with σ2 tracers will have useful-ness for in vivo monitoring of specific tumor markers staticallyand in response to treatment paradigms. Because σ2 receptorsalso are localized in brain regions, such as the hippocampus,caution is appropriate in attributing all nervous system labelingwith this radiopharmaceutical to proliferation (8, 27).

Serial MRI measurements provide accurate, in vivo assess-ment of tumor growth. In this study, brain tumor growth was

FIGURE 8. IOSs from a control and DBT-injected hemisphere. A, surface ofthe intact left hemisphere (projected as the right for comparison) of the mousebrain through the cranium. The scale bar applies to all panels. The compassindicates medial (m) and anterior (a) in this and in panels C and E. B, aftercontralateral whisker stimulation, a vigorous response (warmer colors indicatebigger changes) is evoked. Postmortem histological analysis localized the sig-nal over stimulated barrels in layer IV. Selected barrels (A1, C3, and E5) arelabeled; the red lines indicate the principal branches of the MCA. The orien-tation is the same as in A. C, surface of the right hemisphere 7 days after DBTcells were injected (asterisk). The arrowhead indicates the proximal MCA.D, whisker stimulation does not activate as much cortex as vigorously as in

controls. The projected size of the tumor (purple) was estimated from growthrates (see Fig. 6). The DBT cell injection site is indicated (asterisk). The ori-entation is the same as in C. E, surface of the right hemisphere 26 days afterinjecting DBT cells. Brain swelling is evident. The center of the tumor out-lined postmortem is marked (asterisk). The arrowhead points to the proxi-mal MCA. F, whisker activation of the brain is reduced and confined to somewhisker barrels posterolateral to the tumor. On postmortem histological analy-sis, the pattern of the whisker barrels was distorted by the tumor (asterisk,dashed line). The orientation is the same as in E. MCA, a lateral branch ofthe middle cerebral artery.

NEUROSURGERY VOLUME 60 | NUMBER 2 | FEBRUARY 2007 | 367

MONITORING BRAIN TUMORS IN VIVO WITH SMALL ANIMAL IMAGING

measured for two different groups of untreated, control mice.The difference in number of cells injected resulted in differenttumor volumes in these two sets of animals. However, normal-ized tumor volume versus time curves for these two sets ofcontrol mice were virtually superimposed, indicating nearlyidentical growth rates in untreated animals that are independ-ent of injection site.

The reduction of optical signal elicited by cortical activationwithin tumor-involved functional cortex is consistent with neu-ronal dysfunction in patients with tumors in eloquent cortex(17). In these instances, it is critical to understand the mecha-nisms involved in loss of function so that therapies can bedeveloped that promote recovery of brain function while elim-inating the tumor. Progressive loss of whisker-stimulated IOSwas observed in this study as gliomas enlarged. We filteredlight at 520 to 560 nm and primarily evaluated hemoglobinconcentration related to changes in local vascular volume andhematocrit produced by blood flow. Several factors may con-tribute to IOS changes with glioma growth. These have not yetbeen delineated but clearly include local redistribution of bloodflow from brain to tumor; compression of neural tissues lead-ing to embarrassed metabolism; functional disconnection fromother cortex, thalamus, and descending targets; systemicresponses to tumors with paraneoplastic effects; and, poten-tially, compounds released by tumors that affect nervous func-tion (18, 49).

In models of stroke, IOS has been used to identify specificbrain regions functionally to guide studies designed to deter-mine the extent of pathological characteristics and their impacton brain function (9, 71–73). The relatively late appearance ofsymptoms (e.g., dysfunction, seizures) with many brain tumorssuggests compensatory neuroplasticity. Like the studies of thestroke model, IOS can guide studies to determine the impact oftumors on brain structure and function. The ability to assess thefunction of the whisker barrel cortex in vivo with optical imag-ing (20, 21, 33) and functional MRI (79) makes this model idealfor evaluating alterations in neurological function in responseto tumor growth and invasion. Optical imaging of rat andhuman gliomas using contrast agents has demonstrated prom-ise in identifying tumor margins for resection (29, 30).

The effects of BCNU treatment on tumor growth were eval-uated. BCNU is highly lipophilic and freely crosses the blood-brain barrier to reach high concentrations in cerebrospinalfluid (23, 74). It is a nitrosourea compound that nonspecifi-cally inhibits deoxyribonucleic acid, ribonucleic acid, and pro-tein synthesis. BCNU is an established cytotoxic agent thathas been used in the treatment of glioblastoma in humans.BCNU at standard human doses (5–10 mg/kg) did not showa significant effect in earlier small-animal research (32, 35, 37),but higher doses have been reported to be effective (46). Serialimaging was a sensitive measure of the therapeutic impact ofa 10-mg/kg dose of BCNU in our mouse GBM model. A sin-gle BCNU dose of 10 mg/kg, administered 1 week aftertumor implantation, significantly slowed tumor growth(Fig. 7). This indicates that MRI and other imaging methods,combined with an intracranial model, offer a more sensitive

and accurate measure of therapeutic efficacy than previousstandard models.

The aim of this proof of principle study was to demonstratethat these imaging methods can be used with relative efficiencyto test the effects of agents known to limit tumor growth andfunction. Future studies will evaluate the effects of more cur-rent and investigational therapeutic interventions. With thesecollective imaging tools, it should be possible to detect earlyeffects and to study mechanisms of novel interventions, todetermine dose-response relationships, and to quantitate theimpact of possible therapies on tumor growth with better sta-tistical power and using fewer animals.

The same technologies used to assess tumor growth, func-tion, and biology in vivo in a simple DBT cell model can beapplied to animal models that more directly reduplicate thegenotypic and histological features of human brain tumors.These methods can be used to assess the efficacy of novel ther-apeutic agents and the impact of current treatments on biologyand brain function and to test novel imaging methods that offerdirect translation to humans. Combining imaging methodsoffers a strategy to characterize human tumors that powerfullyaugments current genetic and biological studies.

In humans, MRI is used predominately to evaluate tumorvolume and localization of function, or both, before surgicaland other therapeutic interventions. Investigational studiessuggest, however, that newer MRI technologies, including dif-fusion methods (11, 31, 50) and dynamic contrast enhancement(10, 25, 41, 64), will provide important insights into the molec-ular aspects of neuro-oncology. Imaging techniques that focusnot just on the quantitative measurement of tumor volume,but also on the delineation of tumor and tissue biology and thein vivo effects of therapeutic intervention, offer great promise.Such research will facilitate longitudinal in vivo assessment oftumor biology, tumor growth, and their interactions with thenervous system in vivo. These applications will be directlydependent on the integration of MRI, PET, and optical imagingtechnologies described in this article. Twenty-five years ago,Shapiro et al. (56) suggested that the subcutaneous andintracranial implantation of brain tumor cells into nude micewould contribute to revolutionizing the field of neuro-oncology. We think that the development of small-animalmolecular and functional imaging techniques will lead to thefulfillment of Shapiro et al.’s vision, ultimately to the benefit ofpatients with glioblastoma.

REFERENCES

1. Adzamli K, Yablonskiy DA, Chicoine MR, Won EK, Galen KP, Zahner MC,Woolsey TA, Ackerman JJ: Albumin-binding MR blood pool agents as MRIcontrast agents in an intracranial mouse glioma model. Magn Reson Med49:586–590, 2003.

2. Al-Nabulsi I, Mach RH, Wang LM, Wallen CA, Keng PC, Sten K, Childers SR,Wheeler KT: Effect of ploidy, recruitment, environmental factors, and tamox-ifen treatment on the expression of sigma-2 receptors in proliferating and qui-escent tumour cells. Br J Cancer 81:925–933, 1999.

3. Benedetti S, Bruzzone MG, Pollo B, DiMeco F, Magrassi L, Pirola B, CireneiN, Colombo MP, Finocchiaro G: Eradication of rat malignant gliomas by

368 | VOLUME 60 | NUMBER 2 | FEBRUARY 2007 www.neurosurgery-online.com

JOST ET AL.

retroviral mediated in vivo delivery of the interleukin 4 gene. Cancer Res59:645–652, 1999.

4. Bernstein JJ, Woodard CA: Glioblastoma cells do not intravasate blood ves-sels. Neurosurgery 36:124–132, 1995.

5. Bock NA, Konyer NB, Henkelman RM: Multiple-mouse MRI. Magn ResonMed 49:158–167, 2003.

6. Bock NA, Zadeh G, Davidson LM, Qian B, Sled JG, Guha A, HenkelmanRM: High-resolution longitudinal screening with magnetic resonance imag-ing on a murine brain cancer model. Neoplasia 5:546–554, 2003.

7. Brockmann MA, Ulmer S, Leppert J, Nadrowitz R, Wuestenberg R, Nolte I,Petersen D, Groden C, Giese A, Gottschalk S: Analysis of mouse brain usinga clinical 1.5T scanner and standard small loop surface coil. Brain Res1068:138–142, 2006.

8. Cao J, Kulkarni SS, Husbands SM, Bowen WD, Williams W, Kopajtic T, KatzJL, George C, Newman AH: Dual probes for the dopamine transporter andsigma1 receptors: Novel piperazinyl alkyl-bis(4’-fluorophenyl)amine ana-logues as potential cocaine-abuse therapeutic agents. J Med Chem46:2589–2598, 2003.

9. Carmichael ST, Wei L, Rovainen CM, Woolsey TA: New patterns of intracor-tical projections after focal cortical stroke. Neurobiol Dis 8:910–922, 2001.

10. Cha S, Johnson G, Wadghiri YZ, Jin O, Babb J, Zagzag D, Turnbull DH:Dynamic, contrast-enhanced perfusion MRI in mouse gliomas: Correlationwith histopathology. Magn Reson Med 49:848–855, 2003.

11. Chenevert TL, McKeever PE, Ross BD: Monitoring early response of experi-mental brain tumors to therapy using diffusion magnetic resonance imaging.Clin Cancer Res 3:1457–1466, 1997.

12. Chenevert TL, Stegman LD, Taylor JM, Robertson PL, Greenberg HS,Rehemtulla A, Ross BD: Diffusion magnetic resonance imaging: An earlysurrogate marker of therapeutic efficacy in brain tumors. J Natl Cancer Inst92:2029–2036, 2000.

13. Chicoine MR, Silbergeld DL: Assessment of brain tumor cell motility in vivoand in vitro. J Neurosurg 82:615–622, 1995.

14. Chicoine MR, Silbergeld DL: Invading C6 glioma cells maintain tumorgeni-city. J Neurosurg 83:665–671, 1995.

15. Chicoine MR, Won EK, Zahner M: Intratumoral injection of lipopolysaccha-ride causes regression of subcutaneously implanted mouse glioblastoma mul-tiforme. Neurosurgery 48:607–615, 2001.

16. Cornelissen B, Kersemans V, Jans L, Staelens J, Oltenfreiter R, Thonissen T,Achten E, Slegers G: Comparison between 1 T MRI and non-MRI based vol-umetry in inoculated tumours in mice. Br J Radiol 78:338–342, 2005.

17. Cushing H: Surgery of the head, in Keen WW (ed): Surgery: Its Principles andPractice. Philadelphia, W.B. Saunders, 1919, pp 17–276.

18. Darnell RB, Posner JB: Paraneoplastic syndromes involving the nervous sys-tem. N Engl J Med 349:1543–1554, 2003.

19. Davis FG, Freels S, Grutsch J, Barlas S, Brem S: Survival rates in patientswith primary malignant brain tumors stratified by patient age and tumor his-tological type: An analysis based on Surveillance, Epidemiology, and EndResults (SEER) data, 1973–1991. J Neurosurg 88:1–10, 1998.

20. Dowling JL, Henegar MM, Liu D, Rovainen CM, Woolsey TA: Rapid opticalimaging of whisker responses in the rat barrel cortex. J Neurosci Methods66:113–122, 1996.

21. Erinjeri J, Lui D, Woolsey TA: Timing of intrinsic optical and pial vessel diam-eter changes during activation of mouse barrel cortex. Soc Neurosci Abs28:632, 1998.

22. Erinjeri JP, Woolsey TA: Spatial integration of vascular changes with neuralactivity in mouse cortex. J Cereb Blood Flow Metab 22:353–360, 2002.

23. Fewer D, Wilson CB, Boldrey EB, Enot KJ, Powell MR: The chemotherapy ofbrain tumors. Clinical experience with carmustine (BCNU) and vincristine.JAMA 222:549–552, 1972.

24. Giese A, Westphal M: Glioma invasion in the central nervous system.Neurosurgery 39:235–252, 1996.

25. Gossmann A, Helbich TH, Kuriyama N, Ostrowitzki S, Roberts TP, ShamesDM, van Bruggen N, Wendland MF, Israel MA, Brasch RC: Dynamic contrast-enhanced magnetic resonance imaging as a surrogate marker of tumorresponse to anti-angiogenic therapy in a xenograft model of glioblastomamultiforme. J Magn Reson Imaging 15:233–240, 2002.

26. Grierson JR, Shields AF: Radiosynthesis of 3�-deoxy-3�-[18F]fluorothymidine:[18F]FLT for imaging of cellular proliferation in vivo. Nuc Med Biol27:143–156, 2000.

27. Gundlach AL, Largent BL, Snyder SH: Autoradiographic localization ofsigma receptor binding sites in guinea pig and rat central nervous systemwith (+)3h-3-(3-hydroxyphenyl)-n-(1-propyl)piperidine. J Neurosci6:1757–1770, 1986.

28. Gutmann DH, Baker SJ, Giovannini M, Garbow J, Weiss W: Mouse models ofhuman cancer consortium symposium on nervous system tumors. CancerRes 63:3001–3004, 2003.

29. Haglund MM, Berger MS, Hochman DW: Enhanced optical imaging ofhuman gliomas and tumor margins. Neurosurgery 38:308–317, 1996.

30. Haglund MM, Hochman DW, Spence AM, Berger MS: Enhanced opticalimaging of rat gliomas and tumor margins. Neurosurgery 35:930–941, 1994.

31. Hall DE, Moffat BA, Stojanovska J, Johnson TD, Li Z, Hamstra DA,Rehemtulla A, Chenevert TL, Carter J, Pietronigo D, Ross BD: Therapeuticefficacy of DTI-015 using diffusion magnetic resonance imaging as an earlysurrogate marker. Clin Cancer Res 10:7852–7859, 2004.

32. Hanley ML, Elion GB, Colvin OM, Modrich PL, Keir S, Adams DJ, Bigner DD,Friedman HS: Therapeutic efficacy of vinorelbine against pediatric and adultcentral nervous system tumors. Cancer Chemother Pharmacol 42:479–482,1998.

33. Hodge CJ Jr, Stevens RT, Newman H, Merola J, Chu C: Identification of func-tioning cortex using cortical optical imaging. Neurosurgery 41:1137–1145,1997.

34. Kelly PJ, Daumas-Duport C, Kispert DB, Kall BA, Scheithauer BW, Illig JJ:Image based stereotaxic serial biopsies in untreated intracranial glial neo-plasms. J Neurosurg 66:865–874, 1987.

35. Khil MS, Kolozsvary A, Apple M, Kim JH: Increased tumor cures using com-bined radiosurgery and BCNU in the treatment of 9l glioma in the rat brain.Int J Radiat Oncol Biol Phys 47:511–516, 2000.

36. Kim B, Chenevert TL, Ross BD: Growth kinetics and treatment response ofthe intracerebral rat 9L brain tumor model: A quantitative in vivo study usingmagnetic resonance imaging. Clin Cancer Res 1:643–650, 1995.

37. Kreklau EL, Pollok KE, Bailey BJ, Liu N, Hartwell JR, Williams DA, EricksonLC: Hematopoietic expression of O(6)-methylguanine DNA methyltransferase-P140K allows intensive treatment of human glioma xenografts with combina-tion O(6)-benzylguanine and 1, 3-bis-(2-chloroethyl)-1-nitrosourea. MolCancer Ther 2:1321–1329, 2003.

38. Laforest R, Rowland DJ, Welch MJ: MicroPET imaging with nonconventionalisotopes. IEEE Trans Nucl Sci 49:2119–2126, 2002.

39. Lewis JS, Achilefu S, Garbow JR, Laforest R, Welch MJ: Small animal imaging:Current technologies and perspectives for oncological imaging. Eur J Cancer38:2173–2188, 2002.

40. Mach RH, Smith CR, Al-Nabulsi I, Whirrett BR, Childer SR, Wheeler KT:Sigma 2 receptors as potential biomarkers of proliferation in breast cancer.Cancer Res 57:156–161, 1997.

41. Maxwell RJ, Wilson J, Prise VE, Vojnovic B, Rustin GJ, Lodge MA, Tozer GM:Evaluation of the anti-vascular effects of combretastatin in rodent tumours bydynamic contrast enhanced MRI. NMR Biomed 15:89–98, 2002.

42. Nabi HA, Zubeldia JM: Clinical applications of 18F-FDG in oncology. J NuclMed Technol 30:3–9, 2002.

43. Nagano N, Sasaki H, Aoyagi M, Hirakawa K: Invasion of experimental ratbrain tumor: Early morphological changes following microinjections of C6glioma cells. Acta Neuropathol 86:117–125, 1993.

44. Nelson AL, Algon SA, Munasinghe J, Graves O, Goumnerova L, Burstein D,Pomeroy SL, Kim JY: Magnetic resonance imaging of patched heterozygousand xenografted mouse brain tumors. J Neurooncol 62:259–267, 2003.

45. Peiper RO: Defined human cellular systems in the study of glioma develop-ment. Front Biosci 8:S19–S27, 2003.

46. Plowman J, Waud WR, Koutsoukos AD, Rubinstein LV, Moore TD, GreverMR: Preclinical antitumor activity of temozolomide in mice: Efficacy againsthuman brain tumor xenografts and synergism with 1, 3-bis(2-chloroethyl)-1-nitrosourea. Cancer Res 54:3793–3799, 1994.

47. Ponde DE, Oyama N, Welch MJ, McCarthy TM: Synthesis of fluoro thymidine([F-18]FLT as radiotracer of tumors. Abstr Pap Am Chem Soc 223:081–NuclPart 2, 2002.

NEUROSURGERY VOLUME 60 | NUMBER 2 | FEBRUARY 2007 | 369

MONITORING BRAIN TUMORS IN VIVO WITH SMALL ANIMAL IMAGING

Mechanisms for Neuromodulation and Neuroprotection. Ann Arbor, NPP Books,1992, pp 341–353.

69. Vilner BJ, John CS, Bowen WD: Sigma-1 and sigma-2 receptors are expressedin a wide variety of human and rodent tumour cell lines. Cancer Res55:408–413, 1995.

70. Weber W, Bartenstein P, Gross MW, Kinzel D, Daschner H, Feldmann HJ,Reidel G, Ziegler SI, Lumenta C, Molls M, Schwaiger M: Fluorine-18-FDGPET and iodine-123-IMT SPECT in the evaluation of brain tumors. J NuclMed 38:802–808, 1997.

71. Wei L, Erinjeri J, Zeng JS, Rovainen C, Woolsey TA: Correlated optical imag-ing, electrical activity and 2DG reveal functional plasticity after focal ischemiain rat whisker barrel cortex. Soc Neurosci Abstr 26:1033, 2000.

72. Wei L, Erinjeri JP, Rovainen CM, Woolsey TA: Functional recovery after localischemia revealed by online optical imaging. Soc Neurosci Abstr 23:1578, 1999.

73. Wei L, Rovainen CM, Woolsey TA: Ministrokes in rat barrel cortex. Stroke26:1459–1462, 1995.

74. Weingart J, Brem H: Biology and therapy of glial tumors. Curr Opin NeurolNeurosurg 5:808–812, 1992.

75. Wheeler KT, Wang LM, Wallen CA, Childers SR, Cline JM, Keng PC, MachRH: Sigma-2 receptors as a biomarker of proliferation in solid tumours. Br JCancer 82:1223–1232, 2000.

76. Won EK, Zahner MC, Chicoine MR, Grant EA, Gore P: Analysis of the anti-tumoral mechanisms of lipopolysaccharide against glioblastoma multiforme.Anticancer Drugs 14:457–466, 2003.

77. Xu J, Tu Z, Jones LA, Vangveravong S, Wheeler KT, Mach RH: [(3)H]N-[4-(3,4-dihydro-6, 7-dimethoxyisoquinolin-2(1H)-yl)butyl]-2-methoxy-5-methyl-benzamide: A novel sigma-2 receptor probe. Eur J Pharmacol 525:8–17, 2005.

78. Xu S, Gade TP, Matei C, Zakian K, Alfieri AA, Hu X, Holland EC,Soghomonian S, Tjuvajev J, Ballon D, Koutcher JA: In vivo multiple-mouseimaging at 1.5 T. Magn Reson Med 49:551–557, 2003.

79. Yang X, Hyder F, Shulman RG: Functional MRI BOLD signals coincides withelectrical activity in the rat. Magn Reson Med 38:874–877, 1997.

AcknowledgmentsSupported by National Institutes of Health grants NS28781, NS49048, and

CA102869; a National Cancer Institute/National Institutes of Health SmallAnimal Imaging Resource Program (SAIRP) grant (R24 CA83060); the Alvin J.Siteman Cancer Center at Washington University in St. Louis, a National CancerInstitute Comprehensive Cancer Center (P30 CA91842); an award from theSpastic Paralysis Foundation of the Illinois-Eastern Iowa District of the KiwanisInternational; and, in part, by generous donations of the family and friends ofDeborah Forbush. We thank Michael Zahner for preparation of the DBT cells;Joseph P. Erinjeri, M.D., Ph.D., for IOS technical innovations; Scott P. Leary,M.D., for histological reconstructions; Kathryn Diekmann, Brad Miller, Ph.D.,and John Christensen for evaluating IOS images; Douglas J. Rowland, Ph.D., andthe technical support of John Engelbach, Nicole Fettig, and Jerrel Rutlin in theWashington University small-animal PET laboratory; and Zhude Tu, Ph.D., andDatta Ponde, Ph.D., for synthesizing [11C]RHM-I and [18F]FLT.

COMMENTS

This study describes serial multimodality in vivo imaging ofglioblastoma-bearing mice by magnetic resonance imaging (MRI),

positron-emission tomography (PET), and optical imaging. The authorsconclude that tumor volumetric measurements calculated from MRIscans correlate closely with tumor volumes measured by histologicalsectioning, that significant binding of 11C-RHM-I occurs in growingtumors, and that intrinsic optical signals induced by whisker stimula-tion are reduced and displaced over time by a growing tumor.

The authors address an important topic. New research into tumormolecular pathways is providing a plethora of new drugs which mustbe assessed in animal models before the initiation of human clinical tri-als. The traditional method of assessing chemotherapeutic efficacy inanimal models has been by euthanasia and examination of histologicalsections. The advent of noninvasive imaging methods that permit serialimaging of individual animals over time has the potential of providing

48. Poptani H, Duvvuri U, Miller CG, Mancuso A, Chargundla S, Fraser NW,Glickson JD, Leigh JS, Reddy R: T1rho imaging of murine brain tumors at 4T. Acad Radiol 8:42–47, 2001.

49. Posner JB: Immunology of paraneoplastic syndromes: Overview. Ann N YAcad Sci 998:178–186, 2003.

50. Ross BD, Moffat BA, Lawrence TS, Mukherji SK, Gebarski SS, Quint DJ,Johnson TD, Junck L, Robertson PL, Muraszko KM, Dong Q, Meyer CR,Bland PH, McConville P, Geng H, Rehemtulla A, Chenevert TL: Evaluation ofcancer therapy using diffusion magnetic resonance imaging. Mol CancerTher 2:581–587, 2003.

51. Rubin JB, Kung AL, Klein RS, Chan JA, Sun Y, Schmidt K, Kieran MW, LusterAD, Segal RA: A small-molecule antagonist of CXCR4 inhibits intracranialgrowth of primary brain tumors. Proc Natl Acad Sci U S A 100:13513–13518,2003.

52. Rustamzadeh E, Hall WA, Todhunter DA, Low WC, Liu H, Panoskaltsis-Mortari A, Vallera DA: Intracranial therapy of glioblastoma with the fusionprotein DTIL13 in immunodeficient mice. Int J Cancer 118:2594–2601, 2006.

53. Scatchard G: The attractions of proteins for small molecules and ions. AnnN Y Acad Sci 51:662–672, 1949.

54. Scherer HJ: The forms of growth in gliomas and their practical significance.Brain 63:1–35, 1940.

55. Schmidt KF, Ziu M, Schmidt NO, Vaghasia P, Cargioli TG, Doshi S, Albert MS,Black PM, Carroll RS, Sun Y: Volume reconstruction techniques improve thecorrelation between histological and in vivo tumor volume measurements inmouse models of human gliomas. J Neurooncol 68:207–215, 2004.

56. Shapiro WR, Basler GA, Chernik NL, Posner JB: Human brain tumor trans-plantation into nude mice. J Natl Cancer Inst 62:447–453, 1979.

57. Sherburn EW, Wanebo JE, Kim P, Song SK, Chicoine MR, Woolsey TA:Glilomas in rodent whisker barrel cortex: A new tumor model. J Neurosurg91:814–821, 1999.

58. Shields AF, Grierson JR, Dohmen BM, Machulla HJ, Stayanoff JC, Lawhorn-Crews JM, Obradovich JE, Muzik O, Mangner TJ: Imaging proliferation invivo with [F-18]FLT and positron emission tomography. Nat Med4:1334–1336, 1998.

59. Shields AF, Grierson JR, Stayanoff JC, Lawhorn-Crews JM, Obradovich J,Muzik O, Mangner T: [F-18] FLT can be used to image cell proliferation invivo. J Nucl Med 39:228, 1998.

60. Silbergeld DL, Chicoine MR: Isolation and characterization of human malignantglioma cells from histologically normal brain. J Neurosurg 86:525–531, 1997.

61. Som P, Atkins HL, Bandoypadhyay D, Fowler JS, MacGregor RR, Matsui K,Oster ZH, Sacker DF, Shiue CY, Turner H, Wan CN, Wolf AP, Zabinski SV: Afluorinated glucose analog, 2-fluoro-2-deoxy-D-glucose (F-18): Nontoxictracer for rapid tumor detection. J Nucl Med 21:670–675, 1980.

62. Surawicz TS, Davis F, Freels S, Laws ER Jr, Menck HR: Brain tumor survival:Results from the National Cancer Data Base. J Neurooncol 40:151–160, 1998.

63. Tai YC, Ruangma A, Rowland DJ, Siegel S, Newport DF Chow PL, Laforest R:Performance evaluation of the microPET Focus: A third-generation microPETscanner dedicated to animal imaging. J Nucl Med 46:455–463, 2005.

64. Tofts PS, Brix G, Buckley DL, Evelhoch JL, Henderson E, Knopp MV, LarssonHB, Lee TY, Mayr NA, Parker GJ, Port RE, Taylor J, Weisskoff RM: Estimatingkinetic parameters from dynamic contrast-enhanced T1-weighted MRI of adiffusable tracer: Standardized quantities and symbols. J Magn ResonImaging 10:223–232, 1999.

65. Tooth HH: Some observations on the growth and survival period of intracra-nial tumours, based on the records of 500 cases, with special reference to thepathology of the gliomata. Brain 35:61–108, 1912.

66. Tu Z, Dence CS, Ponde DE, Jones L, Wheeler KT, Welch MJ, Mach MH:Carbon-11 labeled sigma-2 receptor ligands for imaging breast cancer. NuclMed Biol 32:423–430, 2005.

67. van Furth WR, Laughlin S, Taylor MD, Salhia B, Mainprize M, Henkelman M,Cusimano MD, Ackerley C, Rutka JT: Imaging of murine brain tumors usinga 1.5 tesla clinical MRI scanner. Can J Neurol Sci 30:326–332, 2003.

68. Vilner BJ, Bowen WD: Characterization of sigma-like binding properties ofNB41A3, S-20Y, and N1E-115 neuroblastoma, C6 glioma, and NG 108-15neuroblastoma-glioma hybrid cells: Further evidence for sigma-2 receptors, inKamenka JM, Domino EF (eds): Multiple Sigma and PCP Receptor Ligands:

370 | VOLUME 60 | NUMBER 2 | FEBRUARY 2007 www.neurosurgery-online.com

JOST ET AL.

a far more efficient method of drug development and testing than tra-ditional methods can provide.

The primary weakness of this report is that it describes little that isnew. Many groups have already reported serial MRI and PET scanningof tumor-bearing mice, both as a tumor grows and as it undergoestreatment; many groups have also reported volumetric measurementsof these tumors in conjunction with MRI. The whisker stimulation/optical imaging paradigm reported in this study has already beendescribed by the authors previously, and it is not clear what it offersover the other imaging methods used in this report. The authors’ tumormodel, which involves implantation of tumor cells, is also not state-of-the-art; genetically engineered mouse models offer a far richer environ-ment in which to develop and test therapeutic agents and to investigatebasic questions about tumor biology than the approach described here.

Elizabeth BullittAllan FriedmanDurham, North Carolina

In vivo imaging of small animal tumor models continues to be animportant relative limitation in studying the behavior of tumors

with multidimensional relevancy; principally, the longitudinal assess-

ment of tumor progression, effect on normal brain function, andtumor regression in response to treatment. Previous work has beenlimited with respect to these multiple factors of tumor biology. Inthis work, Jost et al. have taken the next logical step and made apotentially significant contribution to the advancement of neuro-oncology by integrating a multidimensional approach. Tumor pro-gression was assessed by serial MRI measurements as well as PETimaging specifically targeting cell proliferation by using the Û2-receptor ligand 11C-RHM-I. Likewise, intrinsic optical signal meas-urements were serially imaged to assess the effect of tumor growth onbrain function. The overall results are very promising, even thoughthe MRI technique used was very basic and the images at the rela-tively low field strength of 4.7 T were only acceptable. This workpoints to further investigations at significantly higher field strengths(9.4 T studies have been safely performed in humans), particularlywith MRI techniques being studied in vivo in humans, such as diffu-sion tensor imaging and susceptibility-weighted techniques, includ-ing MRI perfusion and permeability imaging.

Paul E. KimNeuroradiologistLos Angeles, California

MONITORING BRAIN TUMORS IN VIVO WITH SMALL ANIMAL IMAGING

Joe Louis (right) defends his heavyweight title against Tami Mauriello at Yankee Stadium, September 18, 1946. Louisdefeated Mauriello with a knockout two minutes and nine seconds into the first round.