Abstract
Δγ134.5 mutant herpes simplex type 1 viruses are under active clinical investigation as oncolytic therapy for cancer. Mitogen-activated protein kinase/extracellular signal-regulated kinase kinase (MEK) activity has been shown to suppress protein kinase R and thereby confer oncolytic susceptibility to some human tumors by R3616, a virus deleted for both copies of γ134.5. We report that systemic delivery of R3616 can selectively target and destroy human xenograft tumors that overexpress MEK activity compared with tumors that express lower MEK activity. These results suggest systemic delivery of R3616 may be effective in the treatment of some human tumors. [Cancer Res 2007;67(17):8301–6]
- HSV-1
- systemic
- oncolytic virotherapy
- γ34.5
- mitogen-activated protein kinase kinase
Introduction
Recent clinical trials and experimental studies in animals have shown that recombinant herpes simplex virus-1 (HSV-1) may be suitable for therapy of cancers that are resistant to current therapeutic regimens. A key unresolved question relevant to both design of suitable therapeutic viruses and clinical safety assessments is the ideal mode of viral delivery. The background that has led to the studies described in this report is as follows.
The most widely used mutants for assessment of clinical efficacy in the treatment of tumors lack both copies of the γ134.5 gene and are exemplified by the recombinant R3616 ( 1, 2). In brief, complementary viral RNAs accumulating in infected cells activate protein kinase R (PKR), which in turn phosphorylates the eukaryotic initiation factor-2α (eIF-2α; ref. 3). A consequence of the phosphorylation of eIF-2α is total shutoff of protein synthesis. In wild-type virus-infected cells, the product of the γ134.5 gene binds and recruits protein phosphatase 1α to dephosphorylate eIF-2α to enable protein synthesis to continue unabated ( 4). The basis for using Δγ134.5 mutants for cancer therapy is that they replicate in vitro in some tumor cell lines but fail to grow in normal tissues and are avirulent in experimental animal systems and in humans ( 2, 4). Phase I (safety) trials for potential oncolytic therapy have shown Δγ134.5 mutant viruses to be safe following escalating doses in cancer patients ( 5, 6). These clinical trials as well as the studies on human tumor xenografts in mice revealed that a fraction of tumors fail to support viral replication to a level sufficient to mediate tumor lysis.
Recent studies in our laboratory have shown that transduction of a cell line with a constitutively active mitogen-activated protein kinase (MAPK)/extracellular signal-regulated kinase (ERK) kinase (MEK) gene confers susceptibility to the R3616 mutant virus, whereas cells transduced with a dominant-negative MEK gene become more resistant to the recombinant virus ( 4). MEK is a key regulator in the MAPK pathway and is activated by MAPK kinase kinases (A-RAF, B-RAF, and C-RAF), which are downstream of RAS. MEK, in turn, phosphorylates its only known substrates, the MAPKs (ERK1 and ERK2). MEK is constitutively activated in a wide variety of tumors and functions to promote cell survival ( 1– 3) and to protect tumor cells from multiple apoptotic stimuli. Extensive analyses of the phenotype of the parent and transduced tumor cells exposed to the γ134.5 mutant virus indicate that, in cells transduced with the constitutively active MEK, PKR is not activated in contrast to cells transduced with the dominant-negative MEK.
In essence, we have developed two isogenic tumor cell lines differing in susceptibility to the Δγ134.5 mutant R3616. Availability of these lines enabled us to study the distribution and persistence of virus delivered by different routes. As expected, the virus replicates better and persists longer in the susceptible (high MEK activity) tumors in nude mice. This finding is confirmed in human tumor xenografts that differentially express MEK. The key finding, however, is that systemic viral administration to tumor-bearing mice can be as effective as intratumoral (i.t.) delivery with regard to tumor oncolysis.
Materials and Methods
Cells and tissue culture. HT-caMEK and HT-dnMEK are clonal cell lines constructed in our laboratory from the parental cell line HT1080, a human fibrosarcoma. The methods of transfection with genetic constructs pNC84 and pNC92, which express constitutively active and dominant-negative MEK, respectively, are reported elsewhere ( 4, 7). PC-3 (human prostate cancer) and Hep3B (human hepatoma) were originally obtained from the American Type Culture Collection. The above cell lines were grown in DMEM (Life Technologies/Invitrogen Corp.)-10% FCS (Intergen)-1% penicillin-streptomycin at 37°C and 7% CO2. HT-caMEK and HT-dnMEK were grown in medium supplemented with 500 μg/mL G418 (geneticin, Life Technologies).
Viruses. R3616 is a recombinant virus derived from the prototype wild-type HSV-1(F) strain, deleted for both copies of the γ134.5 gene ( 8). R2636 is a recombinant virus carrying the luciferase gene driven by the glycoprotein C (gC) promoter (gC-luc) in place of the γ134.5 gene The derivation and properties of these viruses are described elsewhere ( 9).
In vivo tumor xenograft regrowth studies. Tumor xenografts in athymic nude mice were established by hind limb injection of 5 × 106 HT-caMEK, HT-dnMEK, Hep3B, or PC-3 tumor cells. At a mean volume of 115 to 150 mm3, the tumors were treated on days 0 and 5 by administration of R3616 via i.t. injection of 5 × 107 plaque-forming units (pfu) or i.p. injection of 106, 107, or 108 pfu of R3616 recombinant virus. Tumor xenografts were measured twice weekly with calipers. Tumor volume was calculated with the formula (l × w × h) / 2, derived from the formula for the volume of an ellipsoid (d3 / g; ref. 4). Tumor growth was measured at each time point by calculating the ratio of tumor volume (V) to initial tumor volume (V0).
Bioluminescence imaging. HT-dnMEK and HT-caMEK xenografts were established in the right hind limb of athymic nude mice by injection of 5 × 106 cells. At initial tumor volumes of 175 ± 60 mm3 for HT-caMEK and 131 ± 22 mm3 for HT-dnMEK, mice were injected with either i.t. (5 × 107 pfu) or i.p. (108 pfu) R2636. Animals were imaged on days 1, 3, 8, 12, and 22 following viral injection. Imaging was done on a charge-coupled device camera (Roper Scientific Photometrics). On days of imaging, animals were injected i.p. with 15 mg/kg body weight with d-luciferin (Biotium). After 5 min, animals were anesthetized with i.p. injection of ketamine (75 mg/kg) and xylazine (5 mg/kg) for imaging, which was done 10 min after injection of d-luciferin.
Quantification of bioluminescence imaging data. Pseudocolor images, with intensity ranging from low (blue) to high (red), were used to represent the relative intensity of light emission from animals treated with R2636. Grayscale images of the mice were superimposed on the pseudocolor images using MetaMorph image analysis software (Fryer Co.). Data for total photon flux were calculated using an area-under-the-curve analysis (MetaMorph).
Results
Comparison of i.t. and systemic delivery of genetically engineered virus on tumor xenografts derived by injection of isogenic tumor cells differing with respect to ectopically expressed MEK activity. In this series of experiments, we established tumor xenografts by injecting into the hind limb of athymic nude mice 5 × 106 HT-caMEK (expressing constitutively active MEK) or HT-dnMEK (expressing dominant-negative MEK) tumor cells. At a mean volume of 115 ± 13 mm3, the tumors were treated on days 0 and 5 by administration of R3616 via i.t. injection of 5 × 107 pfu or i.p. injection of 106, 107, or 108 pfu of R3616 recombinant virus. Tumor xenografts were measured twice weekly with calipers. Tumor volume was calculated with the formula (l × w × h) / 2, derived from the formula for the volume of an ellipsoid. Tumor growth was measured at each time point from day 0 to day 19 by calculating the ratio of tumor volume (V) to initial tumor volume (V0). The results of these experiments are shown in Fig. 1 . In the HT-caMEK xenografts ( Fig. 1A), i.p. treatment with 2 × 106, 2 × 107, or 2 × 108 pfu of R3616 resulted in a significant dose-dependent tumor response by 19 days (V/V0 of 9.1 ± 1.9, 7.3 ± 1.6, and 1.5 ± 0.6, respectively) compared with untreated HT-caMEK controls (V/V0 of 14.5 ± 1.7; P = 0.0221, 0.0371, and 0.0007, respectively). In HT-dnMEK xenografts ( Fig. 1B), no significant effect on tumor growth was seen by day 15 with i.p. administration of 2 × 106, 2 × 107, or 2 × 108 pfu of R3616 (V/V0 of 11.2 ± 1.9, 10.4 ± 1.6, and 9.6 ± 0.6, respectively) compared with untreated HT-dnMEK controls (V/V0 of 9.1 ± 3.1; P = 0.46, 0.35, 0.14, respectively). I.t. administration of 108 pfu of R3616 in HT-caMEK xenografts resulted in a significant antitumor effect with a V/V0 of 3.2 ± 1.1 by day 19 (P = 0.0020). I.t. administration of 108 pfu of R3616 in HT-dnMEK xenografts did not show a significant antitumor effect with V/V0 of 7.9 ± 1.1 by day 15 (P = 0.36). Thus, tumor xenografts genetically engineered to express constitutively active MEK were susceptible to oncolysis following systemic delivery by i.p. injection of R3616, whereas xenografts engineered to express dominant-negative MEK activity were resistant to R3616 oncolysis.
In tumor regrowth studies, systemic delivery of R3616 by i.p. injection results in oncolysis of xenografts dependent on tumor MEK activity. Tumor xenografts were established in the hind limbs of nude mice by injection of 5 × 106 cells per animal. Tumor volume was determined by direct caliper measurement. Once tumors reached a mean volume of 115 to 150 mm3, animals were treated on days 0 and 5 with 2 × 106, 2 × 107, or 2 × 108 pfu i.p. R3616 or 108 pfu i.t. R3616. Tumor growth was measured by calculating the ratio of tumor volume (V) to initial tumor volume (V0). A, HT-caMEK. I.p. treatment with 2 × 106, 2 × 107, or 2 × 108 pfu of R3616 resulted in a significant dose-dependent tumor response by 19 d (V/V0 of 9.1 ± 1.9, 7.3 ± 1.6, and 1.5 ± 0.6, respectively, with P = 0.0221, 0.0371, and 0.0007, respectively) compared with untreated HT-caMEK controls (V/V0 of 14.5 ± 1.7). I.t. administration of 108 pfu of R3616 resulted in a significant antitumor effect with a V/V0 of 3.2 ± 1.1 by day 19 (P = 0.0020). B, HT-dnMEK. No significant effect on tumor growth was seen by day 15 with i.p. administration of 2 × 106, 2 × 107, or 2 × 108 pfu of R3616 (V/V0 of 11.2 ± 1.9, 10.4 ± 1.6, and 9.6 ± 0.6, respectively, with P = 0.46, 0.35, and 0.14, respectively) or i.t. administration of 108 pfu of R3616 (V/V0 of 7.9 ± 1.1, P = 0.36) compared with untreated HT-dnMEK controls (V/V0 of 9.1 ± 3.1). C, Hep3B (high MEK activity). A dose-dependent effect was seen with i.p. administration of 2 × 106, 2 × 107, and 2 × 108 pfu of R3616, which resulted in V/V0 of 4.3 ± 1.0, 3.2 ± 0.5, and 1.4 ± 0.3 at 18 d, compared with untreated Hep3B controls, which reached a mean V/V0 of 6.1 ± 1 (P = 0.2050, 0.0858, and 0.0135, respectively). I.t. administration of 108 pfu of R3616 into Hep3B xenografts (C) resulted in a V/V0 of 1.1 ± 0.2 (P = 0.0130) by day 18. D, PC-3 (low MEK activity). No significant difference in tumor growth was observed between mice treated with i.p. doses of 2 × 106, 2 × 107, and 2 × 108 pfu of R3616 (P = 0.2327, 0.0882, and 0.2970, respectively) or i.t. dose of 108 pfu of R3616 (V/V0 of 8.9 ± 2.2, P = 0.102) compared with untreated control PC-3 xenografts by day 17.
In the second experiments, we established xenografts in hind limbs of athymic mice consisting of Hep3B, a human hepatoma cell line, and PC-3, a human prostate cancer cell line. As reported earlier, Hep3B expresses high MEK activity, whereas the PC-3 cells express almost no MEK activity ( 4). Hep3B and PC-3 xenografts were established in nude mice by hind limb injection of 5 × 106 cells per animal. Hep3B and PC-3 xenografts were grown to an average volume of 150 ± 4 mm3 and then treated on days 0 and 5 with either i.t. injection of 5 × 107 pfu of R3616 or i.p. injection of 106, 107, or 108 pfu of R3616. Hep3B xenografts ( Fig. 1C) showed a dose-dependent effect with i.p. administration of 2 × 106, 2 × 107, and 2 × 108 pfu of R3616, which resulted in V/V0 of 4.3 ± 1.0, 3.2 ± 0.5, and 1.4 ± 0.3 at 18 days compared with untreated Hep3B controls, which reached a mean V/V0 of 6.1 ± 1 (P = 0.2050, 0.0858, and 0.0135, respectively).
In PC-3 xenografts ( Fig. 1D), there was no significant difference between i.p. doses of 2 × 106, 2 × 107, and 2 × 108 pfu of R3616 (P = 0.2327, 0.0882, 0.2970, respectively) and untreated control PC-3 xenografts by day 17. I.t. administration of 108 pfu of R3616 into Hep3B xenografts ( Fig. 1C) resulted in a V/V0 of 1.1 ± 0.2 (P = 0.0130) by day 18. In PC-3 xenografts, i.t. administration of 108 pfu of R3616 did not result in a significant antitumor effect with a V/V0 of 8.9 ± 2.2 (P = 0.102; Fig. 1D). These results show that tumor regrowth studies with natively high (Hep3B) and low (PC-3) MEK activity tumors are similar to the results obtained with tumors genetically engineered to express constitutively active or dominant-negative MEK activity.
Luciferase imaging shows increased viral replication, which localizes to HT-caMEK tumors, compared with attenuated viral replication in HT-dnMEK tumors. R2636 is a γ134.5-deficient virus constructed from the R3616 backbone that expresses the firefly luciferase gene under the control of the late HSV-1 gC promoter. Using R2636, we did in vivo imaging of viral replication. Detectable luciferase expression in tissues represents active viral replication because gC-driven expression marks the expression of late viral structural genes. Hind limb xenografts were established in nude mice by the injection of 5 × 106 cells of the fibrosarcoma cell lines HT-caMEK or HT-dnMEK. At initial tumor volumes of 175 ± 60 mm3 for HT-caMEK and 131 ± 22 mm3 for HT-dnMEK, mice were injected with either i.t. (5 × 107 pfu) or i.p. (108 pfu) R2636. Animals were imaged on days 1, 3, 8, 12, and 22 following viral injection.
In HT-caMEK xenografts that received i.t. injections ( Fig. 2A ), an increase in luminescence remained localized to the hind limb tumor only. In HT-dnMEK xenografts injected i.t., luminescence reached a plateau early in the study and showed much lower activity (approximately 4- to 16-fold lower from day 3 to day 22) than their HT-caMEK counterparts injected i.t. ( Fig. 2B). HT-caMEK tumor-bearing ( Fig. 2C) mice that received i.p. R2636 showed an increase in luminescence in the abdominal cavity (in the liver and spleen) on day 1 that disappeared by day 3 and remained absent up to the conclusion of the study at day 22, whereas a steady increase in luminescence was observed in the hind limb bearing xenografted tumors. HT-dnMEK tumor-bearing mice treated by i.p. R2636 ( Fig. 2D) showed a similar increase in luminescence in the abdominal cavity, in the liver and spleen, on days 1 and 3, which abated by day 8 and remained absent up to the conclusion of the study on day 22 with no localization to the hind limb xenografts. Luminescence was measured and relative intensity was quantified as total photon flux ( Fig. 3 ). HT-dnMEK tumors treated with either i.t. or i.p. R2636 failed to show significantly increased luminescence above the baseline luminescence measured in untreated HT-dnMEK control tumors.
In vivo luciferase imaging of R2636 replication shows that HT-caMEK tumors permit increasing viral replication and HT-dnMEK tumors restrict viral replication. I.p. administration of R2636 in HT-caMEK tumor-bearing mice allows viral localization to the hind limb xenograft and subsequent replication. Tumor xenografts were established as described previously. Mice were injected with i.t. (5 × 107 pfu) or i.p. (108 pfu) R2636. On days 1, 3, 8, 12, and 22 following R2636 treatment, imaging of luciferase activity was done on a charge-coupled device camera 15 min following i.p. injection of 15 mg/kg body weight with d-luciferin. A, HT-caMEK, i.t. B, HT-dnMEK, i.t. C, HT-caMEK, i.p. D, HT-dnMEK, i.p.
Quantified luciferase activity from HT-caMEK and HT-dnMEK tumor-bearing mice treated with 5 × 107 pfu i.t. R2636 or 108 pfu i.p. R2636. Using image analysis software to process images generated from R2636-treated mice bearing HT-caMEK and HT-dnMEK xenografts, luminescence was quantified as total photon flux, calculated using an area-under-the-curve analysis. The baseline luminescence in the untreated HT-caMEK tumors was 1.8 × 105 ± 5.9 × 103 photons. In HT-caMEK tumors injected i.t. with 5 × 107 pfu of R2636, the measured photon activity was 1.8 × 106 ± 6.6 × 105, 1.1 × 107 ± 3.9 × 106, 2.7 × 106 ± 1.2 × 106, 4.3 × 106 ± 3.1 × 106, and 1.6 × 107 ± 6.7 × 106 on days 1, 3, 8, 12, and 22, respectively (P = 0.042, 0.0208, 0.0726, 0.2149, and 0.0477, respectively, with reference to baseline luminescence in untreated control mice bearing HT-caMEK tumors). HT-caMEK xenografts treated with 108 pfu of i.p. R2636 resulted in measured photon emission of 6.6 × 105 ± 1.1 × 105, 2.4 × 106 ± 1.1 × 106, 8.4 × 106 ± 2.7 × 106, 1.1 × 107 ± 5.0 × 106, and 4.8 × 107 ± 2.1 × 107 on days 1, 3, 8, 12, and 22, respectively (P = 0.0019, 0.064, 0.0163, 0.0557, and 0.0499, respectively, with reference to untreated control tumor-bearing mice). In untreated control mice bearing HT-dnMEK tumors, baseline luminescence was 9.9 × 104 ± 1.3 × 104 photons. HT-dnMEK xenografts injected i.t. with 5 × 107 pfu R2636 resulted in measured photon activity of 4.0 × 106 ± 1.6 × 106, 6.8 × 105 ± 2.3 × 105, 6.9 × 105 ± 5.0 × 105, 9.4 × 105 ± 7.9 × 105, and 3.2 × 106 ± 2.8 × 106 on days 1, 3, 8, 12, and 22, respectively. HT-dnMEK xenografts treated with 108 pfu i.p. R2636 resulted in measured photon activity of 5.0 × 105 ± 1.4 × 105, 2.6 × 105 ± 7.3 × 104, 2.0 × 105 ± 1.5 × 105, 4.2 × 104 ± 4.1 × 103, and 4.4 × 104 ± 1.9 × 103 on days 1, 3, 8, 12, and 22, respectively.
To study i.t. distribution of R3616 in HT-caMEK tumors following i.t. or i.p. injection, xenografts were harvested 5 days after treatment with either 5 × 107 pfu of i.t. R3616 or 108 pfu of i.p. R3616. Immunohistochemistry for HSV-1 antigen in HT-caMEK xenografts injected i.t. showed viral replication along the needle tract ( Fig. 4A ). In contrast, HT-caMEK xenografts treated by i.p. injection showed a more diffuse pattern of viral distribution with multiple foci of viral replication throughout the tumors ( Fig. 4B). No HSV-1 antigens were detected by immunohistochemistry in HT-dnMEK xenografts 5 days following i.t. injection of 5 × 107 pfu of R3616 or i.p. injection of 108 pfu of R3616 (data not shown). To examine recovery of R3616 from HT-caMEK tumors following treatment with either i.t. or i.p. R3616, HT-caMEK xenografts were harvested 5 days after treatment with either i.t. 5 × 107 pfu or i.p. 108 pfu of R3616. Viral titers from homogenized samples were determined by standard plaque formation assays on Vero cell monolayers. I.t. administration of 5 × 107 pfu of R3616 yielded a titer of 4 × 105 ± 1 × 105 pfu. I.p. administration of 108 pfu of R3616 yielded a comparable titer of 2 × 105 ± 1 × 105 pfu ( Fig. 5 ). No detectable levels of R3616 were recovered from HT-caMEK xenografts treated with either i.p. 107 or 106 pfu at day 5.
Immunohistochemistry of HT-caMEK tumor for HSV-1 antigen 5 d following R3616 treatment shows a different pattern of viral spread with i.t. versus i.p. injection. HT-caMEK xenografts were harvested 5 d following i.t. (5 × 107 pfu) or i.p. (108 pfu) injection of R3616. Tumors were formalin fixed, paraffin embedded, and probed with anti-HSV-1 antibody. A, i.t. injection (low and high power) shows viral spread outward from the needle tract. B, i.p. injection shows a more diffuse pattern with multiple foci of replication.
Viral recovery from HT-caMEK tumors 5 d following i.t. injection with 5 × 107 pfu R3616 or 108 pfu R3616 is comparable. HT-caMEK xenografts were harvested 5 d after treatment with either i.t. 5 × 107 pfu or i.p. 108 pfu of R3616. Viral titers from homogenized samples were determined by standard plaque formation assays on Vero cell monolayers.
Discussion
We used systemic delivery of R3616 based on the observations of Smith et al. ( 4) who reported that MEK activity suppresses PKR following tumor cell infection with R3616 and thereby increases viral recovery from tumors injected with the virus. The salient features of our report with regard to the systemic administration of HSV-1 are as follows. (a) R3616 shows greater oncolytic activity in xenografted flank tumors with high levels of active MEK compared with tumors that express lower levels of active MEK. This finding holds true in human tumors genetically engineered to express constitutively active MEK as well as tumors that natively express high MEK activity. (b) The superior oncolytic effects of R3616 in high MEK activity tumors are corroborated by in vivo imaging studies with R2636, a Δγ134.5 mutant based on the R3616 backbone in which the late viral promoter for gC drives luciferase expression. In vivo imaging with R2636 shows that systemic administration permits Δγ134.5 mutant virus localization to constitutively active MEK tumors with subsequent intratumoral viral replication. In contrast, in dominant-negative MEK xenografts, R2636 replication is diminished and systemic administration of R2636 does not lead to persistent intratumoral viral replication. (c) Although equal amounts of virus were recovered from caMEK-expressing tumors 5 days following i.p. administration compared with i.t. administration, the kinetics of viral proliferation differed as reflected by quantified bioluminescence imaging.
Applications of HSV-1 oncolytic therapy have principally used local injection of virus directly into the tumor. For this reason, HSV-1 vectors have been clinically tested primarily in malignant gliomas, which remain confined to the central nervous system. In the context of developing HSV-1 as a broader anticancer agent, it would be extremely valuable to be able to administer HSV-1 systemically (i.v. or i.p.) to effectively treat disseminated metastases in addition to the primary tumor. Metastatic disease is responsible for the vast majority of cancer deaths, often in spite of control of the primary tumor. A variety of human tumor types, such as melanomas, sarcomas, and carcinomas of the colon, ovary, liver, breast, esophageal, stomach, pancreas, and lung, have been reported to overexpress MEK activity ( 10, 11). Our results suggest that systemic delivery of R3616 may effectively treat metastases from these tumors. In addition, systemic delivery and selective localization of R3616 to tumors may permit detection of clinically occult metastatic disease as well as serial evaluation of tumor response to oncolysis by positron emission tomography in the clinical setting ( 12). Further, assays of MEK activation and potentially other kinases in tumors might allow for individualized targeted therapy with R3616 or similar viruses.
The limitations of our study include the fact that only a limited number of tumors were used in a flank model not a formal model of metastasis. Additionally, the study of human xenografts required the use of immunosuppressed murine hosts and it is possible that immune activity against HSV-1 in immunocompetent hosts could limit the application of R3616 in humans. Notably, Wong et al. ( 13) have shown oncolytic efficacy of systemically delivered HSV mutants deleted for a single copy of γ134.5 in immunocompetent murine hosts bearing syngeneic squamous cell carcinoma tumors. It is also noteworthy that anti-HSV-1 immune activity has not been reported to limit the use of Δγ134.5 mutants in human trials to date. In spite of these limitations, our results suggest that Δγ134.5 mutant viruses hold promise in the treatment of disseminated metastatic disease.
Acknowledgments
Grant support: NIH grant CA7193309 (R.R. Weichselbaum) and Virginia and D.K. Ludwig Foundation for Cancer Research (R.R. Weichselbaum).
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Footnotes
-
Note: J. Veerapong and K.A. Bickenbach contributed equally to this work.
- Received April 27, 2007.
- Accepted June 14, 2007.
- ©2007 American Association for Cancer Research.
References
Volume 67, Issue 17
