GOLD NANOROD PHOTOTHERMAL THERAPY INA GENETICALLY ENGINEERED MOUSE MODEL

OF SOFT TISSUE SARCOMA

KEVIN Y. LIN*,¶, ALEXANDER F. BAGLEY†,¶,ALEXIA Y. ZHANG‡, DANIEL L. KARL‡, SAM S. YOON‡

and SANGEETA N. BHATIA§,||

*Department of Chemical EngineeringMassachusetts Institute of Technology

77 Massachusetts Avenue, Cambridge, MA 02139, USA

†Division of Health Sciences and TechnologyMassachusetts Institute of Technology

77 Massachusetts Avenue, Cambridge, MA 02139, USA

‡Department of SurgeryMassachusetts General Hospital and Harvard Medical School

Boston, MA 02114, USA

§David H. Koch Institute for Integrative Cancer ResearchMassachusetts Institute of Technology

77 Massachusetts Avenue, Cambridge, MA 02139, USA||sbhatia@mit.edu

Received 11 March 2011Accepted 26 June 2011

Plasmonic nanomaterials are poised to impact the clinical management of cancer through theirability to convert externally applied energy into localized heat at sites of diseased tissue. However,characterization of plasmonic nanomaterials as cancer therapeutics has been limited to xenograftmodels, creating a need to extend these ¯ndings to more clinically relevant models of cancer. Here,we evaluate the method of photothermal ablation therapy in a genetically engineered mousemodel (GEMM) of sarcoma, which more accurately recapitulates the human disease in terms ofstructure and biology than subcutaneous xenograft models. Using polyethylene glycol (PEG)-coated gold nanorods (PEG-NRs), we quantitatively evaluate the ability of nanoparticles topenetrate and accumulate in sarcomas through passive targeting mechanisms. We demonstratethat PEG-NR�mediated photothermal heating results in signi¯cant delays in tumor growth withno progression in some instances. Lastly, by evaluating our photothermal ablation protocol in aGEMM, we observe o®-target heating e®ects that are not detectable in xenograft models andwhich may be of future clinical interest.

Keywords: Gold nanorods; photothermal ablation; genetically engineered mouse model.

¶These authors contributed equally to this work.||Corresponding author.

Nano LIFEVol. 1, Nos. 3 & 4 (2010) 277�287© World Scienti¯c Publishing CompanyDOI: 10.1142/S1793984410000262

277

1. Introduction

In the last decade, nanoparticles and nanomaterialshave been engineered for a wide array of biomedicalapplications. The clinical management of cancerstands to bene¯t greatly from nanoparticles that canmore directly and selectively target tumors fordiagnosis, imaging, and therapy. Noble metalnanomaterials are especially promising diagnostic,imaging, and therapeutic tools because they exhibitstrong optical absorption and scattering propertiesdue to an e®ect known as surface plasmon reson-ance. Gold nanoparticles are plasmonic materialswhich are characterized by facile synthesis andbioconjugation and low cytotoxicity1�8 and whichhave demonstrated potential as multimodal diag-nostic and therapeutic agents in vivo.7,9,10 As diag-nostic agents, gold nanoparticles enable imaging byoptical coherence tomography (OCT), photo-acoustic tomography, two-photon luminescence,X-ray computed tomography (CT), and surface-enhanced Raman spectroscopy.4,11�14 Therapeuti-cally, strategies employing gold nanoparticles haveharnessed their ability to selectively heat tumortissue through the localized conversion of light intothermal energy. By varying their geometrical prop-erties, gold nanoparticles, including nanoshells andnanorods, can be tuned to absorb speci¯c near-infrared (NIR) wavelengths, at which biologicaltissue exhibits relatively low extinction coe±cients.Gold nanoparticle�mediated photothermal ablationhas shown considerable e±cacy in the treatment ofcancer, leading to complete resolution of irradiatedtumor xenografts in some cases.6�8Gold nanoparticle-induced heating has also been combined withother therapeutic modalities to leverage synergiesbetween heat and radiation or chemotherapy,thereby sensitizing tumors to treatment.15,16

Finally, nanoparticle-induced heating has been usedas a photothermal trigger in heat-responsive drugdelivery systems.17,18

Despite the therapeutic promise of gold nano-particles, translation to the clinic requires thattheir e®ectiveness be validated in physiologicallyrelevant models of human cancer. To date, deliveryof gold nanoparticles has been studied in subcu-taneous xenograft tumors, and the accumulationof nanoparticles within xenografts has largelyrelied on passive targeting mechanisms. However,because subcutaneous xenografts fail to recapitu-late important structural features of more clinically

representative genetically engineered mouse models(GEMMs), the degree to which passive accumu-lation plays a role in tumor targeting within geneticmodels remains to be determined.19�21 In particu-lar, there is increasing evidence that tumor inter-stitial °uid pressure is model dependent and sitespeci¯c; therefore, pressure variations betweensubcutaneous xenografts, orthotopic models, andGEMMs may directly in°uence the extravasationacross the vessel wall, the physical and biologicalentrapment of nanoparticles in the interstitium,and the intratumoral distribution of gold nano-particles.22 Di®erences in the anatomical locationand geometry of the tumor in di®erent mousemodels may further a®ect the ability to focus NIRlight, induce speci¯c versus o®-target heating, andlethally ablate the tumor. Demonstrating the pen-etration and accumulation of gold nanoparticles, aswell as their ability to induce site-speci¯c photo-thermal ablation in GEMMs, is therefore animportant step as these particles are evaluated fortheir utility in the clinic.

Here, we sought to directly compare the e®ec-tiveness of a polyethylene (PEG)-coated nanorods(PEG-NR) ablative therapeutic protocol in agenetically engineered mouse model to previousresults obtained in subcutaneous xenografts.6�8 Toassess the impact of PEG-NR ablative therapy, weselected a model of sarcoma harboring conditionalmutations in Kras and Trp53 and resemblinghuman sarcoma both in microscopic structure andby immunohistochemistry.23 Our lab has previouslydeveloped polyethylene glycol (PEG)-coated goldnanorods which exhibit high stability in circulation(t1=2 � 17 h) and a high optical absorption coe±-cient.7 Leveraging this technology, we show thatintravenously injected PEG-NRs passively accumu-late in both subcutaneous xenografts and the sar-coma GEMM. Additionally, NIR laser irradiationresults in rapid, sarcoma-speci¯c heating and abla-tion, leading to signi¯cant delays in tumor growth.We further show loss of extremity function due tonon-speci¯c heating of the sarcoma, an e®ect onlyobserved in the GEMM. Collectively, these studiesdemonstrate that passively targeted PEG-NRs arecapable of serving as highly absorbing antennas in aclinically relevant model of cancer, and suggest thatan optimal combination of PEG-NR delivery andappropriate ablation protocol could be translated toclinical utility.

278 K. Y. Lin et al.

2. Materials and Methods

2.1. Preparation of PEG-coatedgold NRs

Gold nanorods were prepared as described pre-viously.7 Brie°y,� 10� 40 nm cetyltrimethylammo-nium bromide (CTAB)-coated gold NRs withlongitudinal plasmon resonance at 810 (NanopartzInc.) were centrifuged at 16,000 rcf to concentrateand resuspended in 100�M of 5-kDa methyl-PEG-thiol (Layson Bio, Inc.). The solution of 5-kDamethyl-PEG-thiol and CTAB-coated gold NRs wasmixed and dialyzed overnight against ultrapurewater (18M� cm�1) using cellulose ester membranedialysis (Spectrapor). Dialyzed samples were washedand ¯ltered through 100-kDa ¯lters (Millipore) toremove excess polymer, resuspended in PBS, andstored at 4�C.

2.2. Generation of soft tissuesarcomas

The mouse model of sarcoma was described pre-viously by Kirsch et al.23 The 129 S4/SvJae mousestrain was bred and used for the generation oftransgenic sarcoma mice. These mice harbor thefollowing conditional mutations: LSL-KrasG12D=þand p53Fl=Fl. Soft tissue sarcomas were generated byintramuscular injection of 2:4� 107 pfu of adeno-virus (Ad-Cre) expressing Cre-recombinase (GeneVector Transfer Core, University of Iowa) in thelower extremity. All animal studies and procedureswere approved by the MIT and MGH InstitutionalAnimal Care and Use Committees.

2.3. Generation of HT-1080 xenografts

To generate subcutaneous xenograft models, nudemice (Jackson Labs) were injected bilaterally in thehind °anks with � 5� 106 HT-1080 cells suspendedin 200�L DMEM.

2.4. Silver enhancement staining

Para±n-embedded tissue sections were dewaxedand rinsed with double-distilled water for up to30 sec. Silver enhancement was performed using theSilver Enhancer Kit for Membranes (Cytodiagnos-tics). Equal volumes of Silver Enhancer Solution Aand Silver Enhancer Solution B were mixed, and50�100�L was added to tissue sections, ensuring

that the entire tissue was covered. Several dilutionsof SolutionA/Solution Bmixtures in double-distilledwater were tested. To determine the optimalenhancement time, representative 20X ¯elds wereimaged every 10min for up to 80min (NikonECLIPSE Ti). After incubation in silver enhance-ment solution, tissues were rinsed well with double-distilled water for up to 60 sec. Samples werecounterstained with hematoxylin, dehydrated, andmounted using standard protocols.

Quanti¯cation of silver deposits was performedusing ImageJ software. To generate particle counts,the Analyze Particles command was applied to eachcontrast-enhanced, 8-bit, thresholded image. Thesame thresholding values were applied to all imagesincluded in the analysis.

2.5. ICP-MS

Samples for inductively coupled plasma mass spec-trometry, or ICP-MS analysis, were frozen, lyophi-lized, and dissolved in aqua regia, prepared byadding 100�L of HNO3 to 300�L of 37% HCl for72 h to dissolve gold particles. Then, samples werediluted to 10mL with 9.6mL of 2% HNO3 andanalyzed via ICP-MS against standards (Thermo-Scienti¯c Finnigan ELEMENT2). Control salineand organ samples with exogenously added GNRswere used for calibration.

2.6. In vivo photothermal therapy

Approximately 90 days after injection of Ad-Cre,mice bearing sarcomas between 150�200mm3 wererandomized into one of three groups: PEG-NRþNIR, PEG-NR only, and NIR only (n ¼ 4�5mice per group). Mice were then injected throughthe tail vein with PEG-NRs in PBS (40mg Au/kg).After allowing 72 h for vascular clearance of PEG-NRs, the mice were anaesthetized with ketamine/xylazine and sarcomas were irradiated with a NIRlaser (�0.5W/cm2, 810 nm, �1-cm beam diameter).Prior to irradiation, the area around the tumor wasshaved to remove excess hair. To monitor surfacetemperature during irradiation, an infrared ther-mographic camera (FLIR, Thermacam S60) wasused. To assess tumor growth following treatment,tumors were measured every two to three days usingdigital calipers. Mice were euthanized when tumorsexceeded 15mm in any single dimension.

Gold Nanorod Photothermal Therapy in a Genetically Engineered Mouse Model 279

2.7. Immunohistochemistry

Immunohistochemistry was performed as describedpreviously.24 Brie°y, immunostaining was per-formed on formalin-¯xed para±n embedded sectionsfollowing antigen retrieval (10mM citrate bu®er(pH � 6:0) at 95�C for 20min; 22�C for 20min).Primary antibodies rat anti-mouse CD31 (1:100,Pharmingen, San Jose, CA, USA) and mouse anti-PCNA (1:500, PC-10; Santa Cruz Biotechnology,Santa Cruz, CA, USA) were applied to tissue sec-tions for 1 h at room temperature. Secondary anti-bodies were applied for 30min at room temperature:biotinylated rabbit anti-rat IgG (1:1000, VectorLaboratories, Burlingame, CA, USA) or biotiny-lated horse anti-mouse IgG (1:1000, Vector Lab-oratories, Burlingame, CA, USA). Sections werethen incubated for 30min with ABC reagent (Vec-tor Laboratories, Burlingame, CA, USA), rinsedwith PBS-T, and incubated in DAB chromagenreagent (Vector Laboratories, Burlingame, CA, USA).The sections were rinsed under running tap water for5min and counterstained in Mayer's Hematoxylin(Sigma, St. Louis, MO USA), dehydrated, andmounted using Permount (Fisher Scienti¯c, Pitts-burgh, PA, USA). Standard hematoxylin and eosin(H&E) staining was performed on tissue sections.

3. Results

3.1. Sarcoma treatmentand experimental schedule

Sarcoma-bearing mice were injected with eitherPEG-NRs or PBS control and subjected to the pho-tothermal ablation protocol [Fig. 1(a)]. Followingirradiation,mice in the treatment trial were regularlymonitored for tumor burden. A second group of micewas sacri¯ced either 24 or 72 h after ablation to assessthe short-term histopathological e®ects of ablation.We veri¯ed the uniform structure of PEG-NRs bytransmission electron microscopy [Fig. 1(b)] anddemonstrated the capacity of PEG-NRs to speci¯-cally and signi¯cantly heat sarcomas exposed to NIRirradiation [Figs. 1(c) and 1(d)]. A representativephotograph of the sarcoma [Fig. 1(e)] provides per-spective on the challenges of locally speci¯c ablationand the potential source for non-speci¯c heatinge®ects [Figs. 1(d) and 1(e), Supp. Fig. 1].

3.2. Comparison of PEG-NRaccumulation in GEMM andxenograft models of sarcoma

To con¯rm the presence of PEG-NR accumulationin the tumor interstitium of both the GEMM and

Fig. 1. Schematic of GNR heating with NIR laser irradiation in a GEMM of sarcoma. (a) Timeline of sarcoma generation andphotothermal ablation procedure. (b) TEM image of PEG-NRs. Scale bar represents 50 nm. ((c), (d), (e)) Bright ¯eld (c) and IRthermographic images of mice with NIR irradiation only (d) and PEG-NRþNIR irradiation (e).

280 K. Y. Lin et al.

xenograft, we performed silver enhancement stain-ing on tumor sections to visualize PEG-NR micro-distribution within the tumors. Both the GEMMsand HT-1080 xenografts displayed comparablePEG-NR microdistributions, with PEG-NRsappearing throughout the tumor tissue [Figs. 2(a)and 2(b)], while control samples displayed little tono detectable background staining [Figs. 2(c) and2(d)]. Quanti¯cation of silver-enhanced PEG-NRsin histological sections revealed a greater number ofparticles accumulating in the HT-1080 xenograftscompared to the GEMM; further, both the HT-1080xenografts and GEMMs exhibited signi¯cantaccumulation of particles relative to uninjectedtumor controls [Fig. 2(e)]. Additionally, we did notdetect signi¯cant silver deposits in surrounding tis-sues, including skeletal muscle, indicating thatPEG-NRs accumulated preferentially in sarcomas(Supp. Fig. 2). ICP-MS con¯rmed the presence ofgold in the GEMMs (6.60 %ID/g) in amounts

comparable to those seen in the HT-1080 xenografts(11.32%ID/g) [Fig. 2(f)]. These results directlycon¯rm that PEG-NRs are able to penetrate andaccumulate in sarcoma GEMMs in amounts su±-cient for photothermal ablation protocols.

3.3. Immunohistochemistry of tumorsfollowing photothermal ablation

Examination of H&E-stained para±n sectionsrevealed regions of gross necrosis at 24 and 72 hpost-irradiation. These sections were characterizedby loss of tumor architecture, cellularly hypodenseregions, lymphocytic in¯ltrates, and irregularnuclear staining patterns (Fig. 3 and Supp. Fig. 3).These regions likely correspond to the portions ofthe tumor receiving the majority of photothermalenergy during the ablation procedure.

To assess cellular proliferation, we immunohis-tochemically stained the sarcoma sections for

(a) (b) (c)

(d) (e) (f)

Fig. 2. Accumulation of PEG-NRs in GEMM and subcutaneous xenograft. Representative silver enhancement staining of PEG-NRs in a GEMM of sarcoma (a) and HT-1080 xenograft (b). Representative silver enhancement staining in an uninjected GEMM ofsarcoma (c) and HT-1080 xenograft (d). Scale bar represents 50�m. (e) Quanti¯cation of silver enhanced spots. Contrast-enhancedimages captured at 20� magni¯cation were analyzed on ImageJ software (n ¼ 9 per group). (f) ICP-MS quanti¯cation of PEG-NRdeposition in GEMM of sarcoma and HT-1080 xenografts 72 h after intravenous administration.

Gold Nanorod Photothermal Therapy in a Genetically Engineered Mouse Model 281

proliferating cell nuclear antigen (PCNA). Consist-ent with the H&E results, we observed wide regionsof absent PCNA staining throughout the tumors at24 and 72 h following ablation (Fig. 3). Interest-ingly, we identi¯ed isolated regions located exclu-sively at sites of the tumor most distal from the skinsurface that stained positively for PCNA. Theseareas of viable tissue account for anywhere between0% to approximately 20% of the area of some tissuesections (Supp. Fig. 3), suggesting that some por-tions of the tumor did not receive su±cient photo-thermal energy during the ablation procedure.

Finally, to understand how photothermal abla-tion in°uences the distribution of blood vessels inthe sarcoma, we measured expression of CD31, anendothelial cell marker commonly used to identifyblood vessels. Necrotic regions exhibited di®use,non-speci¯c staining patterns bearing no resem-blance to the blood vessels observed in untreatedcontrol tissue (Supp. Fig. 4). Thus, in regions oftumor receiving su±cient irradiative heat to inducenecrosis, the tumor-associated vasculature is alsoablated.

3.4. Therapeutic assessmentof photothermal ablationin a transgenic sarcoma model

To test the ability of a single dose of PEG-NRs tosigni¯cantly delay tumor growth following one ses-sion of near-IR irradiation, genetically modi¯ed129 S4/SvJae mice-bearing-induced soft tissue sar-comas were injected with PEG-NRs. Mice were ran-domized to one of three cohorts (PEG-NRsþNIR,PEG-NRs only, NIR only). After plasma clearance ofPEG-NRs, tumors in the PEG-NRsþNIR and NIRonly groups were irradiated for 5min (�0.5W/cm2,810 nm) and tumor volume was measured over time[Fig. 4(a)]. Mice receiving the PEG-NRsþNIRtreatment experienced signi¯cant tumor growthdelay and extended lifespan compared to controltreatment groups. During the post-ablation obser-vation period, all mice in the PEG-NR�only andNIRonly groups developed substantial tumor burdensrequiring euthanization by day 16, whereas micereceiving PEG-NRsþNIR survived until day 30 orlonger [Fig. 4(b)]. At the end of the study period, the

Fig. 3. Immunohistochemical analysis of short-term heating e®ects on GEMM of sarcoma. (a) H&E and PCNA staining of sarcoma72 h after heating versus an unheated control. Scale bar represents 100�m. Contrast-enhanced images captured at 10� magni¯-cation. (b) Schematic of sarcoma cross-section following heating. (c) Photograph of PEG-NR þ NIR treated sarcoma prior toexcision, 72 h post-heating.

282 K. Y. Lin et al.

two remaining mice had tumor volumes < 150mm3

(Supp. Fig. 5). Notably, four of the ¯vemice receivingPEG-NRþNIR lost lower extremity function afterthe ablation procedure. Additionally, in a repeattherapeutic trial, we observed delayed progression ofsarcoma consistent with our initial results.

4. Discussion

Surgery, chemotherapy, and radiation haveremained the standard of care in cancer therapy fordecades. Novel approaches harnessing the uniqueproperties of nanomaterials have been proposed toimprove cancer therapy in the future. To date,numerous therapeutic protocols, involving a varietyof nanomaterials, have been described in xenograftmouse models. While xenograft models o®er astraightforward and reliable method for quantitat-ively assessing tumor burden, studies in morephysiologic tumor microenvironments, such as thosethat occur in GEMMs, may be more accurate pre-dictors of therapeutic e±cacy.25 For these reasons,GEMMs are gaining favor in the pharmaceuticalindustry and will almost certainly ¯nd utility inassessing the clinical validity of nanomaterial-basedtherapeutic protocols as well.

This study represents the ¯rst demonstration of atherapeutic, nanoparticle-mediated thermal abla-tion protocol in a GEMM. We have demonstratedthat PEG-NRs accumulate in the sarcomas at levelscomparable to those in subcutaneous HT-1080

xenografts. Deposition of PEG-NRs is su±cient tofacilitate photothermal ablation of the sarcomasupon localized administration of near-infraredirradiation.

Although numerous lines of evidence support thenotion that passively targeted nanotherapeutics canpenetrate and accumulate in human tumors, thisphenomenon has not been well characterized in aclinical setting.26,27 Routinely used xenograft modelsfail to capture important characteristics of humandisease that may impact the e®ectiveness of passivetargeting mechanisms.19 This uncertainty is signi¯-cant as many nanotherapeutic approaches rely onpassive targeting to varying degrees in order for thetherapeutic payload to accumulate at the site ofdisease. Our demonstration of untargeted PEG-NRaccumulation in a GEMM of sarcoma provides evi-dence that passive targeting is indeed su±cient forPEG-NRs to accumulate in a physiologic tumormicroenvironment.

The considerable heterogeneity of human cancersis an important factor for PEG-NR-based therapies.It is not anticipated that a \one-size-¯ts-all"therapy will be identi¯ed, but rather that thera-peutic protocols must be tailored to each type ofcancer. Features including grade, stage, tissue oforigin, anatomical location, and metastatic pro-gression will in°uence if, and to what extent, PEG-NR-based therapies will have a place in theclinic.23,28,29 For example, because of constraintsrelated to depth of NIR irradiation into tissues,

(a) (b)

Fig. 4. Photothermal ablation of GEMM of sarcoma using PEG-NRs. (a) Volumetric changes in tumor volume are plotted overtime (n ¼ 4�5). Error bars represent standard deviations. PEG-NR is statistically di®erent from NIR and PEG-NR from day 7onwards with p < 0:0001 based on analysis of variance. (b) Survival of mice over time (n ¼ 4�5).

Gold Nanorod Photothermal Therapy in a Genetically Engineered Mouse Model 283

tumors that develop close to the skin surface, suchas some soft tissue sarcomas, would be better can-didates for PEG-NR-based therapies than tumorsthat develop at sites more distant from the skinsurface. Interestingly, in our GEMM, a portion ofthe sarcoma developed closer to the surface whilethe remainder developed deeper within surroundingskeletal muscle. In contrast to HT-1080 xenograftmodels which replicate super¯cial truncal tumors,our GEMM replicates a deep extremity sarcoma,which represents the majority of human sarcomas.We propose that this spatial heterogeneity con-tributed to the di®erential success in ablativetherapy that we observed. Additionally, tumors thatare identi¯ed at earlier stages may be better candi-dates for PEG-NR therapies than tumors identi¯edat late stages, such as pancreatic adenocarcinoma,30

because of the increased challenge of optimallyirradiating multiple metastatic lesions.

Notably, a signi¯cant proportion of mice toundergo photothermal ablation lost function inthe irradiated limb following treatment. Because thesarcomas invade surrounding skeletal muscle, thethermal energy generated within the sarcomadamaged the skeletal muscle as well as the sciatic andfemoral nerves in the leg.We believe this observationis important in two respects. First, it suggests thatimprovements are required in the irradiation proto-col to provide more precise thermal ablation at cel-lular resolution. Pulsed laser sources represent onepossibility for improving the localization of photo-thermal ablation. Second, this unwanted side e®ectof the therapy could only be detected and fullyappreciated in a GEMM, in which the sarcoma'sanatomical location closely mimics that of humanextremity soft tissue sarcomas. While a subcu-taneous xenograft model can provide important in-formation such as tumor burden in response totreatment, only in a GEMM can these realisticclinical consequences be detected and assessed.

5. Conclusion

The ability of mouse models to accurately recapitu-late human cancers and yield clinically insightfulresults is critical to the continuing development ofnanotechnology for therapy and diagnostics. Ourwork highlights the therapeutic e±cacy and potentialchallenges of PEG-NR�mediated photothermalablation therapy in a GEMM of sarcoma. We an-ticipate that our ¯ndings will enable the development

of improved therapeutic protocols as PEG-NRs andother nanotherapeutics advance closer to clinical use.

Acknowledgments

This project was supported by the MIT-MGH Col-laborative Initiative in Translational CancerResearch and the NIH (R01 CA124427). S. B. is anHHMI investigator. A.F.B. acknowledges supportfrom the NIH/Medical Scientist Training Program.The authors would like to acknowledge Dr. AliceChen for her helpful suggestions on the manuscript.The authors also thank Matt Kinsella for assistancewith ICP-MS, Prof.Warren Chan for assistance withsilver enhancement staining, Prof. Yoel Fink for useof the IR camera, andNickiWatson for TEM images.Histology support from the Koch Institute HistologyCore Facility was provided by Michael Brown.

References

1. R. Elghanian, J. J. Storho®, R. C. Mucic, R. L.Letsinger and C. A. Mirkin, Science 277, 1078(1997).

2. D. S. Grubisha, R. J. Lipert, H.-Y. Park, J. Driskelland M. D. Porter, Anal. Chem. 75, 5936 (2003).

3. J. B. Jackson, S. L. Westcott, L. R. Hirsch, J. L.West and N. J. Halas, Appl. Phys. Lett. 82, 257(2003).

4. X. Qian et al., Nat. Biotech. 26, 83 (2008).5. L. R. Hirsch et al., Proc. Nat. Acad. Sci. USA 100,

13549 (2003).6. D. P. O'Neal, R. H. Leon, J. H. Naomi, J. D. Payne

and L. W. Jennifer, Cancer Lett. 209, 171 (2004).7. G. von Maltzahn et al., Cancer Res. 69, 3892 (2009).8. B. E. Dickerson et al., Cancer Lett. 269, 57 (2008).9. A. M. Gobin et al., Nano Lett. 7, 1929 (2007).

10. G. von Maltzahn et al., Adv. Mater. 21, 3175(2009).

11. J. Chen et al., Nano Lett. 5, 473 (2005).12. Y. Wang et al., Nano Lett. 4, 1689 (2004).13. H. Wang et al., Proc. Natl. Acad. Sci. USA 102,

15752 (2005).14. D. Kim, S. Park, J. H. Lee, Y. Y. Jeong and S. Jon,

J. Am. Chem. Soc. 129, 7661 (2007).15. J.-H. Park et al., Proc. Natl. Acad. Sci. 107, 981

(2010).16. R. L. Atkinson et al., Sci. Transl. Med. 2, 55ra79

(2010).17. J.-H. Park et al., Adv. Mater. 22, 880 (2010).18. M. S. Yavuz et al., Nat. Mater. 8, 935 (2009).19. O. J. Becher, E. C. Holland, E. A. Sausville and

A. M. Burger, Cancer Res. 66, 3355 (2006).

284 K. Y. Lin et al.

20. E. A. Sausville, A. M. Burger, O. J. Becher and E. C.Holland, Cancer Res. 66, 3351 (2006).

21. K. K. Frese and D. A. Tuveson, Nat. Rev. Cancer 7,654 (2007).

22. C. Brekken, O. Bruland and C. de Lange Davies,Anticancer Res. 20, 3503 (2000).

23. D. G. Kirsch et al., Nat. Med. 13, 992 (2007).24. S. S. Yoon et al., Int. J. Radiation Oncol. Biol. Phys.

74, 1207 (2009).

25. M. Singh et al., Nat. Biotech. 28, 585 (2010).26. J. Fang, H. Nakamura and H. Maeda, Adv. Drug

Deliv. Rev. 63, 136 (2011).27. M. E. Davis, Z. Chen and D. M. Shin, Nat. Rev. Drug

Discov. 7, 771 (2008).28. M. Nogawa et al., Cancer Lett. 217, 243 (2005).29. J. K. Mito et al., PLoS One 4, e8075 (2009).30. M. Hidalgo, New Eng. J. Med. 362, 1605

(2010).

Supplementary Information

Fig. 1. Temperature pro¯le of GEMM of sarcoma heating with and without PEG-NR (n ¼ 4 or 5). Error bars represent standarddeviations. Temperatures are statistically di®erent at each time point from 30�300 s with p < 0:005 based on Student's t-test.

Fig. 2. Silver enhancement staining with H&E counterstaining in a sarcoma tissue section from a mouse injected with PEG-NRreveals silver deposits present in the sarcoma but absent in adjacent skeletal muscle. Silver deposits are indicated by arrowheads.Images were acquired at 4� magni¯cation.

Gold Nanorod Photothermal Therapy in a Genetically Engineered Mouse Model 285

Fig. 3. Immunohistochemical analysis of short-term heating e®ects on GEMM of sarcoma. Scale bar represents 50�m. Contrast-enhanced images captured at 20� magni¯cation. Arrow indicates the direction of NIR irradiation.

286 K. Y. Lin et al.

Fig. 5. Photothermal destruction of transgenic sarcoma using PEG-NRs. Volumetric changes in tumor volume are plotted overtime (n ¼ 4 or 5). Error bars represent standard deviations.

Fig. 4. Immunohistochemical analysis of short-term heating e®ects on GEMM of sarcoma. CD31 staining of tumor vessels at 72 hafter heating and in an unheated control. Scale bar represents 50�m. Contrast-enhanced images captured at 20� magni¯cation.

Gold Nanorod Photothermal Therapy in a Genetically Engineered Mouse Model 287