- Email: [email protected]
Treatment of Spontaneously Arising Retinoblastoma Tumors in Transgenic Mice with an Attenuated Herpes Simplex Virus Mutant
VIROLOGY
229, 283– 291 (1997) VY968414
ARTICLE NO.
SHORT COMMUNICATION Treatment of Spontaneously Arising Retinoblastoma Tumors in Transgenic Mice with an Attenuated Herpes Simplex Virus Mutant CURTIS R. BRANDT,*,†,1 PASCAL D. IMESCH,* NANCY L. ROBINSON,* NASREEN A. SYED,* SEEMA UNTAWALE,* SOESIAWATI R. DARJATMOKO,* RICHARD J. CHAPPELL,‡ PAUL HEINZELMAN,* and DANIEL M. ALBERT* *Department of Ophthalmology and Visual Sciences, †Department of Medical Microbiology and Immunology, and ‡Department of Biostatistics and Statistics, University of Wisconsin Medical School, Madison, Wisconsin 53706 Received July 26, 1996; returned to author for revision September 24, 1996; accepted December 17, 1996 The use of viruses to treat tumors has received renewed interest with the availability of genetically defined attenuated mutants. Herpes simplex virus (HSV) type 1 in particular has been shown to be effective for tumors of neuronal origin. However, the model systems used for these studies rely on the use of explanted tumor cells in immunodeficient animals. We have used a recently developed transgenic mouse model, wherein mice spontaneously develop retinoblastomas, to determine if a mutant HSV has a therapeutic effect against an endogenously arising tumor in an immunocompetent host. The injection of 1 1 10 6 PFU of the neuroattenuated HSV-1/HSV-2 recombinant RE6 into the vitreous of transgenic mice resulted in a significant inhibition of tumor growth compared to injection of medium alone (P Å 0.0063). Immunohistochemical analysis of viral antigen showed that viral replication was restricted to focal areas of the tumors and the retinal pigment epithelium. Viral growth was not significantly different in the eyes of transgene-positive and transgene-negative mice, suggesting that enhanced replication in tumor cells may not explain the effects. Tumor cells in the treated eyes were significantly less differentiated than those in the untreated eyes (P Å 0.04), suggesting that the virus may replicate better in certain cell types in the tumors. Although the injection of RE6 resulted in a difference in tumor size, the treatment did not result in the elimination of tumors in any of the mice. Improvements in the efficacy of tumor control are needed if this therapy is to be of use. q 1997 Academic Press
Traditionally, malignancy is treated with surgery, radiation, chemotherapy, or immunotherapy. Although these treatments are successful with some cancers, other types of cancer respond less well. Tumors of neuronal origin are among those resistant to traditional treatments. For example, high-grade malignant gliomas are the most common brain tumors in adults (1), and despite aggressive therapy with traditional methods, the 5-year survival rate is less than 5% (2– 5). One potential alternative therapy is the use of genetically modified herpes simplex virus (HSV). The advantages of HSV include its neurotropism and the existence of avirulent mutant viruses that grow to high titer in dividing cells, but show reduced replication in terminally differentiated cells (6– 10). Pioneering studies by Skinner et al. (11) showed that an HSV thymidine kinase (TK) mutant could reduce the rate of tumor development. More recent studies using a genetically engineered deletion mutant, dlspTK (12), showed a dose-dependent improvement in survival using a nude mouse– human gli-
oma explant model. Subsequent studies with other tumor cell lines (13), additional HSV mutants (14, 16– 21), or HSV mutants in combination with ganciclovir (15) have provided additional support for the use of HSV mutants for tumor therapy. These previous studies used model systems in which tumor cell lines were implanted into rodents. In addition, some of them also used immunodeficient nude (12– 16) or SCID mice (18). The immunodeficient models do not allow for the study of the effects of an active immune response to the attenuated HSV as a possible positive or negative cofactor in the treatment. Tumors arise spontaneously and usually in the presence of functional immunity, which can modulate direct cell killing by the virus and may enhance the response by recognition of new cell surface antigens. HSV encodes a gene, g134.5, that is required for complete neurovirulence and growth in a number of cell types including neurons (7, 8, 10, 22– 28). Successful treatment of neuronal tumors in animal models has been reported with several g134.5 mutants. Using a glioma cell implant model in nude mice, Markert (14) showed that treatment with RE6 and another g134.5 mutant, R3616, significantly prolonged survival and eliminated the tumors in 67% of
1 To whom correspondence and reprint requests should be addressed at University of Wisconsin Medical School, 6630 Medical Sciences Center, 1300 University Avenue, Madison, WI 53706-1532.
0042-6822/97 $25.00
283
AID VY 8414
/
6a29$$$421
01-31-97 20:35:39
Copyright q 1997 by Academic Press All rights of reproduction in any form reserved.
vira
AP: Virology
284
SHORT COMMUNICATION
the mice. Chambers et al. (18), using a SCID mouse model of human glioma, showed that treatment prolonged survival in a dose-dependent fashion. Finally, Randazzo et al. (17) showed that a g134.5 mutant, 1716, could increase survival time in an intracranial murine melanoma model system. These results suggest that g134.5 mutants have potential for therapy of neuronal tumors. A transgenic model for heritable retinoblastoma in which bilateral retinoblastomas arise spontaneously in the mice was recently described (29, 30). In this model, a transgene containing the SV40 large T-antigen (Tag) driven by the murine b-leutinizing hormone promoter (LHb-Tag mice) expresses Tag in the retina. The Tag binds the p105 Rb protein (29), resulting in functional inactivation of Rb and the development of ocular tumors that, in their histopathology and clinical course, closely resemble human retinoblastoma. The LHb-Tag mice thus represent a model in which to test the use of attenuated HSV mutants for the treatment of malignancy that more accurately represents the situation in patients. The RE6 strain of HSV was first described in 1983 as an avirulent HSV-1:HSV-2 intertypic recombinant (10). Subsequently, it was shown to be avirulent (PFU/LD50 ú 1 1 106) due to a defect in the g134.5 gene which resulted in severely reduced replication in neuronal tissues (10, 22). In this paper, we show that the attenuated HSV mutant, RE6, when injected intravitreally into LHb-Tag transgenic mice, results in a statistically significant reduction in retinoblastoma tumor size. This is the first report of the use of an HSV mutant to treat a spontaneously arising tumor in an immunocompetent animal and supports the possible use of this therapeutic modality for some cancers. Transgene-bearing mice were produced by mating LHb-Tag transgene-positive males with CB6F1/J (C57BL/ 6J BALB/cJ) females (The Jackson Laboratory, Bar Harbor, ME) and identifying transgene-positive offspring by PCR analysis of tail DNA using the 5* primer 5*-GACTTTGGAGGCTTCTGGGATGCAACTGAG-3* (SV40 Tag bases 4947– 4918) and the 3* primer GGCATTCCACCACTGCTCCCATTCATCAGT-5* (SV40 Tag bases 4527 – 4556), which resulted in a 421-bp product. High-titer stocks of the RE6 virus (10) (gift of Dr. John Yu, Massachusetts General Hospital, Boston, MA) were prepared in African green monkey kidney (Vero) cells as we described previously (6). The effect of treatment with the RE6 virus was tested initially in a pilot study using the right eyes from 16 mice. Eight of these were injected with 1 1 106 PFU of RE6 in 5 ml of medium, and 8 received medium only. Based on the results from the preliminary study, a second group of 36 mice was tested. In this second study, 17 eyes received virus and 19 eyes were injected with medium only. The left eye in these mice was not injected and served as an uninjected control. All mice were treated
between 16 and 18 weeks of age and eyes were taken for analysis at 3– 5 weeks postinjection. Treatment at this time was chosen based on the growth characteristics of the tumor (data not shown). At 16 to 18 weeks of age, tumors have involved the full thickness of the retina and are beginning to extend into the vitreous. The analysis at 3 to 5 weeks was chosen to allow for viral clearance. For both studies, the left and right eyes were removed, fixed in 10% buffered formalin, embedded in paraffin, sectioned, and stained with hematoxylin and eosin. In the pilot study, three random sections of each eye were examined by a masked observer and the section with the largest tumor was chosen. The tumor area was calculated using a 2.5X objective and reticle. In the second study, single sections from each eye were cut through the pupil optic nerve axis, each section was digitized, the tumor was traced, and the tumor area in square millimeters was calculated. These methods have been published previously (31, 32). Each eye was counted as a single tumor. To confirm that these methods result in random sampling of the eyes, we serially sectioned the right eyes of four mice and measured the ratio of tumor area to total eye area in each section. The results showed that the ratio was the same from the optic nerve– pupil axis to at least 1 mm on each side (data not shown). At the optic nerve– pupil axis the ratios were 0.49 { 0.05, indicating that the size of the tumor in relation to the size of the eye is remarkably constant. These results suggest that we are taking random representative samples of the tumors using the method described for the second study. The data were analyzed in two strata corresponding to the two studies (33). As the stratification did not influence the estimates of treatment effect or significance, it was dropped in the final analysis for simplicity. Two additional comparisons were available only in the second group: the treated vs control eye contrast in the experimental animals, and the injected vs control eye contrast among the medium-only group. However, the major contrast between the eyes treated with RE6 in the experimental group and the injected eyes in the mediumonly group was available for all eyes. This yielded the P value of 0.0063 for treatment effect. Due to the presence of outliers in the form of a few very large tumors, tumor areas were subjected to the logarithmic transformation. The logarithmic scale induced approximate normality so that standard normaltheory confidence intervals and t tests were appropriately used. Thus, the means and confidence intervals were calculated on the log scale and exponentiated. Relative risks and their confidence intervals were computed from differencing means on the log scale and then exponentiating. The analysis was conducted using S-plus statistical software (33). All significance levels were obtained via t tests, either two-sample, when comparing eyes in
AID VY 8414
vira
/
6a29$$$422
01-31-97 20:35:39
AP: Virology
SHORT COMMUNICATION
285
FIG. 1. Effect of treatment on tumor size (A), and the ratio of treated to control eyes (B). These results represent the combined analysis of the data from the two studies.
different animals, or one-sample, when comparing injected and uninjected eyes in the same animal. As shown in Fig. 1A, the mean tumor area in the virustreated eyes was 1.97 mm2 (95% confidence interval 1.48 –2.61 mm2) compared to 2.54 mm2 (1.67– 3.85 mm2) for the uninjected eyes in the virus-treated group. The mean tumor area was 3.67 mm2 (2.59 – 5.18 mm2) for the eyes injected with medium only compared to 3.14 mm2 (2.23– 4.42 mm2) for the uninjected eyes in the medium control group. To assess the effect of the treatment we compared the means of the ratios of the treated and untreated eyes in the virus and medium-injected groups as well as between the virus-injected and medium-injected eyes
(Fig. 1B). The mean ratio in tumor size between the injected and the uninjected eyes in the medium-only group was 0.88 (95% confidence interval 0.61 – 1.28) and the difference was not significant (P Å 0.24). This result indicates that the injection of medium did not affect tumor size. The mean ratio in tumor size in the virus-treated group was 0.74 (0.56– 0.98). This difference was significant (P Å 0.019), indicating that the virus treatment resulted in smaller tumors. When we compared the eyes injected with virus to the eyes injected with medium, the mean ratio was 0.54 (0.35– 0.83) and the difference was highly significant (P Å 0.0063). It should be noted, however, that the 95% confidence intervals overlap in all groups. Representative low-magnification examples of
AID VY 8414
vira
/
6a29$$$422
01-31-97 20:35:39
AP: Virology
286
SHORT COMMUNICATION
sections through an untreated eye and a virus-treated eye are shown in Figs. 2A and 2B respectively. Morphological analysis of the tumors in the two groups was also carried out to determine if there were differences in the degree of involvement of the choroid, vitreous, iris – ciliary body angle, cornea, retinal pigment epithelium, and optic nerve. The criteria for this analysis have been described previously (31). A defining characteristic of retinoblastomas is the presence of Homer Wright rosettes, which represent attempts by the tumor cells to differentiate into photoreceptors (23, 24). These rosettes are characterized by a central luscent space that represents the attempt by the cells to form outer segments. The cell bodies surrounding the outer segments have a relatively constant shape and size with a high nuclear-to-cytoplasmic ratio. As shown in Table 1, the only differences noted were that the tumors in the RE6-treated eyes were less differentiated (P Å 0.03) and had fewer Homer Wright rosettes (P Å 0.03) than the control tumors. Figures 2C and 2D show representative intermediate-magnification views of a virus-injected (2C) tumor and an uninjected (2D) tumor. The virus-injected tumor shows an extensive area lacking the typical Homer Wright rosettes compared to the uninjected tumor, in which the majority of the cells are organized into rosettes. Higher magnification views in Figs. 2E and 2F show the organized structure of the rosettes in the uninjected tumors compared to the pleomorphic cells and disorganization in the virus-injected tumors. There was no difference in the amount of necrosis in the two groups at 3 to 5 weeks posttreatment. To test the hypothesis that the antitumor effect was due to the ability of the RE6 virus to grow better in tumor cells compared to normal eye tissue, we injected the right eyes of transgene-positive or transgene-negative littermates (five mice per group) with 1 1 10 6 PFU of RE6 and at various times postinfection (PI) we assayed the eyes for the amount of infectious virus using procedures we described previously (34). The results are shown in Fig. 3. On Day 1 PI, the RE6 titers were 2 1 104 and 1 1 10 3 for the LHb-Tag-negative and LHb-Tag-positive mice, respectively. On Day 2 PI, the titers were 10-fold higher in the LHb-Tag-negative mice. On Day 5 PI, there was only a 2-fold difference and by Day 7, infectious virus had essentially cleared in both groups of mice. Statistical analysis (t test) failed to reveal significant differences (P ú 0.05) in viral growth in the two groups.
Injection of virus into immunocompetent animals might be expected to induce an inflammatory response which could contribute to the antitumor activity. We therefore compared the degree of inflammation in the virus-injected (treated) and untreated tumors at 3 weeks posttreatment using procedures described previously (31). The presence of inflammation was higher in the treated eyes (44%) compared to the control eyes (31%). However, the groups were not significantly different (P Å 0.43) (data not shown). When we characterized the inflammation with respect to cell type, by histological criteria, we found that neutrophils were the predominant infiltrating cell type; however, the comparisons between treated and untreated eyes failed to reveal significant differences with regard to the type of inflammatory cell (P ú 0.05) within the tumor (data not shown). To assess the distribution of virus in the eyes of the treated mice, the right eyes of 16- to 18-week-old transgene-bearing mice were injected intravitreally with 1 1 106 PFU of RE6. At 3 days PI, the eyes were removed, fixed in 10% buffered formalin, embedded in paraffin, and sectioned. Immunohistochemical detection of viral antigen was carried out using a commercially available kit (Vectastain ABC, Vector Laboratories, Burlingame, CA) and commercially available primary anti-HSV antibody (BO114, Dako Corp., Carpinteria, CA). Representative examples of the immunohistochemical staining are shown in Fig. 4. In tumor tissue, HSV antigen was detected only in focal areas (Fig. 4B) and did not appear to spread extensively, even though viral titers were highest at this time (Fig. 3). Viral antigen was also present in retinal pigment epithelial cells in focal areas (Fig. 4D). We found no evidence of immunohistochemical staining in any other ocular tissues. One possible mechanism for the selective effects of the attenuated HSV on tumors is that the mutant viruses grow better in rapidly dividing tumor cells than in normal tissues. For attenuated viruses lacking either viral thymidine kinase or ribonucleotide reductase, this hypothesis is consistent with the proposed role of these enzymes in viral replication. However, the in vivo growth of mutant viruses in tumor and normal tissue was not compared in most of the previous studies (12 – 21). In the one previous study using a g134.5 mutant where viral growth was tested in vivo, replication was more efficient and prolonged in tumor cells compared to normal brain (17). In contrast, we found that viral growth in the eyes of LHb-
FIG. 2. Tumor histopathology. (A) Representative example of an untreated (control) eye (original magnification, 123). (B) RE6-treated eye from the same animal. C, cornea; L, lens; T, tumor; N, optic nerve; R, retina (original magnification, 123). (C) Intermediate-magnification view of tumor from an untreated eye showing Homer Wright rosettes (original magnification, 1230), some of which are marked with asterisks. (D) Intermediatemagnification view of the tumor from an RE6-treated eye (original magnification, 1230). Note that some cells attempt to form rosettes but these are poorly organized. Note also the variable morphology of the cells. (E) High-magnification view of a tumor from an untreated eye showing Homer Wright rosettes with the central luscent space and surrounding cell bodies (original magnification, 1460). Note the uniformity in cell size and shape. (F) High-magnification view of a tumor from an eye treated with RE6 (original magnification, 1460). Note the lack of rosettes and the variability in cell size and shape.
AID VY 8414
/
6a29$$$422
01-31-97 20:35:39
vira
AP: Virology
SHORT COMMUNICATION
01-31-97 20:35:39 AID VY 8414
/
6a29$$8414
vira
287
AP: Virology
288
SHORT COMMUNICATION TABLE 1 Characteristics of Retinoblastoma Tumorsa Untreated (left eye)b
Treated (right eye) Number
%
Number
%
Pc
Number of eyes graded Tumor present Retina totally involved % of retina uninvolved Full thickness involvement
17d 17 5 58.13 ({33.58) 10
100 100 29
16 e 16 4 62.38 ({30.82) 10
94 100 16
1 1 1 0.63 0.67
Invasion of vitreous Invasion of ciliary body Invasion of iris Invasion of angle Invasion of cornea Invasion of choroid Optic nerve invasion Calcification % undifferentiated Necrosis present Flexner-Wintersteiner rosettes Homer Wright rosettes
7 4 2 2 0 7 2 12 34.69 ({28.37) 9 0 65.5 ({29.15)
41 24 12 12 — 41 12 71
4 4 3 4 1 5 2 9 14.38 ({20.32) 8 3 85.63 ({20.32)
25 16 19 25 6 32 13 56
Observation
Retinal involvement
Invasion of other ocular structures
Rosettes
59
53 —
63
16 19
0.43 0.71 0.33 0.33 0.33 0.75 1 0.5 0.03 1 — 0.03
a
All data from the second study, sections through pupil optic nerve axis. Uninjected paired eyes. c Unpaired t test. d For statistical analysis only the 16 paired eyes were included. e The intraocular contents of one eye were lost and could not be graded. b
Tag-positive mice was not significantly different than in the eyes of nontransgenic littermates (Fig. 3). Viral replication, as measured by staining for viral antigen, was restricted to focal areas of the tumors and to limited regions of retinal pigment epithelium. The magnitude of the antitumor effects seems surprising given the restricted distribution of viral antigen. However, these results are similar to those of Randazzo et al. (17) and suggest that the virus does not spread extensively in the
FIG. 3. Replication of RE6 virus in the eyes of transgene-positive (Rb/) and transgene-negative (Rb0) mice. Each data point is the mean titer from five eyes.
tumors. When considered together, these results suggest that the differences in tumor size between virus-injected and control mice are not related to viral replication. Additional mechanisms for the selective antitumor effects include the induction of an inflammatory response that kills tumor cells through bystander effects, or the induction of tumor-specific immunity as a consequence of viral infection. The observed inflammation in areas positive for viral antigens at Day 3 PI is consistent with possible bystander effects, but the data are not conclusive. The presence of similar degrees of inflammation in treated and untreated eyes in the same animal at 3 to 5 weeks PI makes it difficult to draw conclusions about the induction of tumor-specific immunity and further studies are needed to distinguish the effect of these different mechanisms. Several morphological criteria of the tumors were also analyzed. In addition to the significant difference in tumor size, we found that the tumors in the treated mice were less differentiated, having fewer Homer Wright rosettes than the untreated tumors. These results are tentative due to their marginal P values and the large number of parameters that were measured. One possible interpretation of this finding is that the virus preferentially infects or is more cytolytic in the more differentiated population of retinoblastoma cells in the tumor. This hypothesis would be consistent with previous results showing that cell type and cell state affect the growth of g134.5 mu-
AID VY 8414
vira
/
6a29$$$422
01-31-97 20:35:39
AP: Virology
FIG. 4. Immunohistochemical analysis of viral antigen distribution. Three days postinfection the eyes were collected, fixed in 10% formalin in phosphatebuffered saline, embedded in paraffin, sectioned, and stained using polyclonal anti HSV-1 antisera (Dako, Carpintera, CA), and an ABC vectastain kit (Vector Laboratories, Burlingame, CA). (A) Control (uninjected) eye. (B) RE6-injected eye. The arrows in B point to areas of positive staining for viral antigen. (C) Control eye showing retinal pigment epithelium. (D) RE6-injected eye showing viral antigen in RPE cells. The arrows in D point to RPE cells containing viral antigen. The bars represent 16 mm.
SHORT COMMUNICATION
AID VY 8414
/
6a29$$8414
01-31-97 20:35:39
vira
289
AP: Virology
290
SHORT COMMUNICATION
tants (28). Further studies are needed to determine if the RE6 virus preferentially replicates in more differentiated tumor cells and whether this is related to the g134.5 mutation in RE6. Although we have shown that intravitreal injection of RE6 results in a significant difference in retinoblastoma size, the 95% confidence intervals overlapped and we did not achieve complete regression of the tumor in any of the treated mice. This is in contrast to other reports with neuronal cell implants in which elimination of the tumors was achieved at least in a portion of the treated animals (14, 17, 21). The reason or reasons for our less robust results are not clear, but there are several possibilities. Several of these factors could also account for the variability in the results as reflected in the overlapping 95% confidence intervals. Our data suggest that viral growth, if it occurred at all, was not different between the transgene-bearing mice and transgene-negative mice, and lack of viral replication could have reduced the effectiveness of the treatment. Animal models utilizing tumor cell implants differ substantially from our retinoblastoma model in that tumors form in 2 to 3 weeks in the implant models compared to 16 to 18 weeks for the retinoblastomas. Thus, there is likely to be a substantial difference in the rate of cell division in the tumor cells between the two models that could account for the reduced viral growth in our model. Another possibility is that the mice mounted an immune response against infection that subsequently reduces the effectiveness of the virus. If this were true, we would expect that the inflammatory response in the virustreated eyes would have been different quantitatively, qualitatively, or both, and we found no evidence for this. It should be pointed out, however, that our analysis of immune responses was limited to histopathology and a more extensive study of the immune response to the virus and the tumor is needed. Tumor cell lines from these mice are not readily available as yet. Thus, direct determination of the ability of RE6 to grow in tumor cells was not tested. It is possible that retinoblastoma cells are resistant to the growth of g134.5 mutants in general or RE6 in particular. However, the growth of virus in cultured murine retinoblastoma cells may not reflect the in vivo situation and the direct measurement of replication in vivo may be more appropriate. Finally, the limited efficacy of the treatment in our model may be due to technical difficulties. At the time of treatment, the tumors are just extending into the vitreous. In these circumstances, the majority of the virus may be injected into the vitreous and not directly into the tumor. Thus, RE6 virus may have infected the tumor poorly or the majority of the virus may have failed to reach the tumor. The cell implant models have involved either direct injection into the tumor or preinfection of the cell with virus prior to implantation. Optimum results in our
model may require injections at later than 16 to 18 weeks when tumors are larger and direct intratumoral injection is technically less difficult. In summary, we have shown that treatment of spontaneously arising retinoblastoma tumors in a transgenic mouse model with a neuroattenuated HSV mutant (RE6) results in a significant difference in tumor size. This finding represents the first report of successful treatment of a tumor with a mutant HSV without the use of tumor explants. This model more closely mimics the situation in the treatment of human cancer. More importantly, the model will allow us to test viral treatment of tumors in situations where tumor alloantigens from explants are not a significant factor. The fact that these animals are immunocompetent will also allow us to examine the combined effects of the immune response and viral infection on tumor development.
AID VY 8414
vira
/
6a29$$$422
01-31-97 20:35:39
ACKNOWLEDGMENTS These studies were supported in part by NIH Grants EY07736 (C.R.B.) and EY01917 (D.M.A.), by the University of Wisconsin Cancer Center Core Grant SP30 CA14520, by an unrestricted grant from Research to Prevent Blindness, Inc. to the Department of Ophthalmology and Visual Sciences, and the Wisconsin Lions. C.R.B. was a Research to Prevent Blindness, Inc. Robert E. McCormick Special Scholar during a portion of these studies. The authors thank Bernadette Bull for administrative assistance.
REFERENCES 1. Levin, A. L., Shelinji, G. E., and Gutin, P. H., In ‘‘Cancer: Principles and Practice of Oncology’’ (V. T. Vita, S. Hellman, and S. A. Rosenberg, Eds.), pp. 1557–1611. Lippincott, Philadelphia, 1989. 2. Mahaley, M. S., Jr., Mettlin, C., Natarajan, N., Laws, E. R., and Peace, B. B., J. Neurosurg. 71, 826– 836 (1989). 3. Salazar, D. M., Rubin, P., Feldstein, M. L., and Pizzutiello, R., Int. J. Radiat. Oncol. Biol. Phys. 5, 1733–1740 (1979). 4. Danmas-Duport, C., Scheithauer, B., O’Fallon, J., and Kelly, P., Cancer 62, 2152–2165 (1988). 5. Kim, T. S., Halliday, A. L., Hedley-Whyte, E. T., and Convery, K., J. Neurosurg. 74, 27 –37 (1991). 6. Brandt, C. R., Kintner, R. L., Pumfery, A. M., Visalli, R. J., and Grau, D. R., J. Gen. Virol. 72, 2043– 2049 (1991). 7. Chou, J., Kern, E., Whitley, R. J., and Roizman, B., Science 250, 1262–1266 (1990). 8. MacLean, A. R., Embed-Fareed, M., Robertson, L., Harland, J., and Brown, S. M., J. Gen. Virol. 72, 631–639 (1991). 9. Coen, D. M., Kosz-Vienchak, M., Jacobson, J. G., Lieb, D. A., Bogard, L. L., Schatter, P. A., Tyler, K. L., and Knipe, D. V., Proc. Natl. Acad. Sci. USA 86, 4736–4740 (1989). 10. Thompson, R. L., and Stevens, J. G., Virology 131, 171– 179 (1983). 11. Skinner, G. R. B., Cowan, M., Davies, J., Brookes, K., Billstrom, M., and Buchan, A., Br. J. Exp. Pathol. 69, 495– 504 (1988). 12. Martuza, R. L., Malick, A., Markert, J. M., Ruffner, K. L., and Coen, D. M., Science 252, 854– 856 (1991). 13. Markert, J. M., Coen, D. M., Malick, A., Mineta, T., and Martuza, R. L., J. Neurosurg. 77, 590–594 (1992). 14. Markert, J. M., Malick, A., Coen, D. M., and Martuza, R. L., Neurosurgery 32, 597– 603 (1993). 15. Mineta, T., Rabkin, S. D., and Martuza, R. L., Cancer Res. 54, 3963– 3966 (1994).
AP: Virology
SHORT COMMUNICATION
291
16. Mineta, T., Rabkin, S. D., Yazaki, T., Hunter, W. D., and Martuza, R. L., Nature Med. 1, 938– 943 (1995). 17. Randazzo, B. P., Kesari, S., Gesser, R. M., Alsop, D., Ford, J. C., Brown, S. M., MacLean, A., and Fraser, N. W., Virology 211, 94 – 101 (1995). 18. Chambers, R., Gillespie, G. Y., Soroceaun, L., Andreansky, S., Chatterjee, S., Chou, J., Roizman, B., and Whitley, R. J., Proc. Natl. Acad. Sci. USA 92, 1411– 1415 (1995). 19. Jia, W. W-G., McDermott, M., Goldie, J., Cynader, M., Tan, J., and Tufaro, F., J. Natl. Cancer Inst. 86, 1209– 1215 (1994). 20. Boviatsis, E. J., Scharf, J. M., Chase, M., Harrington, K., Kowall, N. W., Breakefield, X. O., and Chiocca, E. A., Gene Ther. 1, 323– 331 (1994). 21. Kaplitt, M. G., Tjuvajev, J. G., Leib, D. A., Berk, J., Pettigrew, K. D., Posner, J. B., Pfaff, D. W., Rabkin, S. D., and Blasberg, R. G., J. Neuro-Oncol. 19, 137– 147 (1994). 22. Thompson, R. L., Rogers, S. K., and Zerhusen, M. A., Virology 172, 435– 450 (1989). 23. Bolovan, C. A., Sawtell, N. M., and Thompson, R. L., J. Virol. 68, 48 – 55 (1994). 24. Javier, R. T., Thompson, R. L., and Stevens, J. G., J. Virol. 61, 1978– 1984 (1987).
25. MacLean, A., Robertson, L., McKay, E., and Brown, S. M., J. Gen. Virol. 72, 2305–2310 (1991). 26. Taha, M. Y., Brown, S. M., Clements, G. B., and Graham, D. I., J. Gen. Virol. 71, 1597–1601 (1990). 27. Whitley, R. J., Kern, E. R., Chatterjee, S., Chou, J., and Roizman, B., J. Clin. Invest. 91, 2837– 2843 (1993). 28. Brown, S. M., Harland, J., MacLean, A. R., Podlech, J., and Clements, J. R., J. Gen. Virol. 75, 2367– 2377 (1994). 29. Windle, J. J., Albert, D. M., O’Brien, J. M., Marcus, D. M., Disteche, C. M., Bernards, R., and Mellon, P. L., Nature 343, 665– 669 (1990). 30. O’Brien, J. M., Marcus, D. M., Bernards, R., Carpenter, J. L., Windle, J. J., Mellon, P., and Albert, D. M., Arch. Ophthalmology 108, 1145–1151 (1990). 31. Albert, D. M., Marcus, D. M., Gallo, J. P., and O’Brien, J. M., Invest. Ophthalmol. Vis. Sci. 33, 2354 –2364 (1992). 32. Backer, R. A., Chambers, J. M., and Wilks, A. R., ‘‘The New S Language.’’ Wadsworth’s Brooks/Cole, Pacific Grove, CA, 1988. 33. Pagano, M., and Gauvrean, K., ‘‘Principles of Biostatistics.’’ Duxbury, Belmont, CA, 1993. 34. Brandt, C. R., and Grau, D. R., Invest. Ophthalmol. Vis. Sci. 31, 2214– 2223 (1990).
AID VY 8414
vira
/
6a29$$$423
01-31-97 20:35:39
AP: Virology
229, 283– 291 (1997) VY968414
ARTICLE NO.
SHORT COMMUNICATION Treatment of Spontaneously Arising Retinoblastoma Tumors in Transgenic Mice with an Attenuated Herpes Simplex Virus Mutant CURTIS R. BRANDT,*,†,1 PASCAL D. IMESCH,* NANCY L. ROBINSON,* NASREEN A. SYED,* SEEMA UNTAWALE,* SOESIAWATI R. DARJATMOKO,* RICHARD J. CHAPPELL,‡ PAUL HEINZELMAN,* and DANIEL M. ALBERT* *Department of Ophthalmology and Visual Sciences, †Department of Medical Microbiology and Immunology, and ‡Department of Biostatistics and Statistics, University of Wisconsin Medical School, Madison, Wisconsin 53706 Received July 26, 1996; returned to author for revision September 24, 1996; accepted December 17, 1996 The use of viruses to treat tumors has received renewed interest with the availability of genetically defined attenuated mutants. Herpes simplex virus (HSV) type 1 in particular has been shown to be effective for tumors of neuronal origin. However, the model systems used for these studies rely on the use of explanted tumor cells in immunodeficient animals. We have used a recently developed transgenic mouse model, wherein mice spontaneously develop retinoblastomas, to determine if a mutant HSV has a therapeutic effect against an endogenously arising tumor in an immunocompetent host. The injection of 1 1 10 6 PFU of the neuroattenuated HSV-1/HSV-2 recombinant RE6 into the vitreous of transgenic mice resulted in a significant inhibition of tumor growth compared to injection of medium alone (P Å 0.0063). Immunohistochemical analysis of viral antigen showed that viral replication was restricted to focal areas of the tumors and the retinal pigment epithelium. Viral growth was not significantly different in the eyes of transgene-positive and transgene-negative mice, suggesting that enhanced replication in tumor cells may not explain the effects. Tumor cells in the treated eyes were significantly less differentiated than those in the untreated eyes (P Å 0.04), suggesting that the virus may replicate better in certain cell types in the tumors. Although the injection of RE6 resulted in a difference in tumor size, the treatment did not result in the elimination of tumors in any of the mice. Improvements in the efficacy of tumor control are needed if this therapy is to be of use. q 1997 Academic Press
Traditionally, malignancy is treated with surgery, radiation, chemotherapy, or immunotherapy. Although these treatments are successful with some cancers, other types of cancer respond less well. Tumors of neuronal origin are among those resistant to traditional treatments. For example, high-grade malignant gliomas are the most common brain tumors in adults (1), and despite aggressive therapy with traditional methods, the 5-year survival rate is less than 5% (2– 5). One potential alternative therapy is the use of genetically modified herpes simplex virus (HSV). The advantages of HSV include its neurotropism and the existence of avirulent mutant viruses that grow to high titer in dividing cells, but show reduced replication in terminally differentiated cells (6– 10). Pioneering studies by Skinner et al. (11) showed that an HSV thymidine kinase (TK) mutant could reduce the rate of tumor development. More recent studies using a genetically engineered deletion mutant, dlspTK (12), showed a dose-dependent improvement in survival using a nude mouse– human gli-
oma explant model. Subsequent studies with other tumor cell lines (13), additional HSV mutants (14, 16– 21), or HSV mutants in combination with ganciclovir (15) have provided additional support for the use of HSV mutants for tumor therapy. These previous studies used model systems in which tumor cell lines were implanted into rodents. In addition, some of them also used immunodeficient nude (12– 16) or SCID mice (18). The immunodeficient models do not allow for the study of the effects of an active immune response to the attenuated HSV as a possible positive or negative cofactor in the treatment. Tumors arise spontaneously and usually in the presence of functional immunity, which can modulate direct cell killing by the virus and may enhance the response by recognition of new cell surface antigens. HSV encodes a gene, g134.5, that is required for complete neurovirulence and growth in a number of cell types including neurons (7, 8, 10, 22– 28). Successful treatment of neuronal tumors in animal models has been reported with several g134.5 mutants. Using a glioma cell implant model in nude mice, Markert (14) showed that treatment with RE6 and another g134.5 mutant, R3616, significantly prolonged survival and eliminated the tumors in 67% of
1 To whom correspondence and reprint requests should be addressed at University of Wisconsin Medical School, 6630 Medical Sciences Center, 1300 University Avenue, Madison, WI 53706-1532.
0042-6822/97 $25.00
283
AID VY 8414
/
6a29$$$421
01-31-97 20:35:39
Copyright q 1997 by Academic Press All rights of reproduction in any form reserved.
vira
AP: Virology
284
SHORT COMMUNICATION
the mice. Chambers et al. (18), using a SCID mouse model of human glioma, showed that treatment prolonged survival in a dose-dependent fashion. Finally, Randazzo et al. (17) showed that a g134.5 mutant, 1716, could increase survival time in an intracranial murine melanoma model system. These results suggest that g134.5 mutants have potential for therapy of neuronal tumors. A transgenic model for heritable retinoblastoma in which bilateral retinoblastomas arise spontaneously in the mice was recently described (29, 30). In this model, a transgene containing the SV40 large T-antigen (Tag) driven by the murine b-leutinizing hormone promoter (LHb-Tag mice) expresses Tag in the retina. The Tag binds the p105 Rb protein (29), resulting in functional inactivation of Rb and the development of ocular tumors that, in their histopathology and clinical course, closely resemble human retinoblastoma. The LHb-Tag mice thus represent a model in which to test the use of attenuated HSV mutants for the treatment of malignancy that more accurately represents the situation in patients. The RE6 strain of HSV was first described in 1983 as an avirulent HSV-1:HSV-2 intertypic recombinant (10). Subsequently, it was shown to be avirulent (PFU/LD50 ú 1 1 106) due to a defect in the g134.5 gene which resulted in severely reduced replication in neuronal tissues (10, 22). In this paper, we show that the attenuated HSV mutant, RE6, when injected intravitreally into LHb-Tag transgenic mice, results in a statistically significant reduction in retinoblastoma tumor size. This is the first report of the use of an HSV mutant to treat a spontaneously arising tumor in an immunocompetent animal and supports the possible use of this therapeutic modality for some cancers. Transgene-bearing mice were produced by mating LHb-Tag transgene-positive males with CB6F1/J (C57BL/ 6J BALB/cJ) females (The Jackson Laboratory, Bar Harbor, ME) and identifying transgene-positive offspring by PCR analysis of tail DNA using the 5* primer 5*-GACTTTGGAGGCTTCTGGGATGCAACTGAG-3* (SV40 Tag bases 4947– 4918) and the 3* primer GGCATTCCACCACTGCTCCCATTCATCAGT-5* (SV40 Tag bases 4527 – 4556), which resulted in a 421-bp product. High-titer stocks of the RE6 virus (10) (gift of Dr. John Yu, Massachusetts General Hospital, Boston, MA) were prepared in African green monkey kidney (Vero) cells as we described previously (6). The effect of treatment with the RE6 virus was tested initially in a pilot study using the right eyes from 16 mice. Eight of these were injected with 1 1 106 PFU of RE6 in 5 ml of medium, and 8 received medium only. Based on the results from the preliminary study, a second group of 36 mice was tested. In this second study, 17 eyes received virus and 19 eyes were injected with medium only. The left eye in these mice was not injected and served as an uninjected control. All mice were treated
between 16 and 18 weeks of age and eyes were taken for analysis at 3– 5 weeks postinjection. Treatment at this time was chosen based on the growth characteristics of the tumor (data not shown). At 16 to 18 weeks of age, tumors have involved the full thickness of the retina and are beginning to extend into the vitreous. The analysis at 3 to 5 weeks was chosen to allow for viral clearance. For both studies, the left and right eyes were removed, fixed in 10% buffered formalin, embedded in paraffin, sectioned, and stained with hematoxylin and eosin. In the pilot study, three random sections of each eye were examined by a masked observer and the section with the largest tumor was chosen. The tumor area was calculated using a 2.5X objective and reticle. In the second study, single sections from each eye were cut through the pupil optic nerve axis, each section was digitized, the tumor was traced, and the tumor area in square millimeters was calculated. These methods have been published previously (31, 32). Each eye was counted as a single tumor. To confirm that these methods result in random sampling of the eyes, we serially sectioned the right eyes of four mice and measured the ratio of tumor area to total eye area in each section. The results showed that the ratio was the same from the optic nerve– pupil axis to at least 1 mm on each side (data not shown). At the optic nerve– pupil axis the ratios were 0.49 { 0.05, indicating that the size of the tumor in relation to the size of the eye is remarkably constant. These results suggest that we are taking random representative samples of the tumors using the method described for the second study. The data were analyzed in two strata corresponding to the two studies (33). As the stratification did not influence the estimates of treatment effect or significance, it was dropped in the final analysis for simplicity. Two additional comparisons were available only in the second group: the treated vs control eye contrast in the experimental animals, and the injected vs control eye contrast among the medium-only group. However, the major contrast between the eyes treated with RE6 in the experimental group and the injected eyes in the mediumonly group was available for all eyes. This yielded the P value of 0.0063 for treatment effect. Due to the presence of outliers in the form of a few very large tumors, tumor areas were subjected to the logarithmic transformation. The logarithmic scale induced approximate normality so that standard normaltheory confidence intervals and t tests were appropriately used. Thus, the means and confidence intervals were calculated on the log scale and exponentiated. Relative risks and their confidence intervals were computed from differencing means on the log scale and then exponentiating. The analysis was conducted using S-plus statistical software (33). All significance levels were obtained via t tests, either two-sample, when comparing eyes in
AID VY 8414
vira
/
6a29$$$422
01-31-97 20:35:39
AP: Virology
SHORT COMMUNICATION
285
FIG. 1. Effect of treatment on tumor size (A), and the ratio of treated to control eyes (B). These results represent the combined analysis of the data from the two studies.
different animals, or one-sample, when comparing injected and uninjected eyes in the same animal. As shown in Fig. 1A, the mean tumor area in the virustreated eyes was 1.97 mm2 (95% confidence interval 1.48 –2.61 mm2) compared to 2.54 mm2 (1.67– 3.85 mm2) for the uninjected eyes in the virus-treated group. The mean tumor area was 3.67 mm2 (2.59 – 5.18 mm2) for the eyes injected with medium only compared to 3.14 mm2 (2.23– 4.42 mm2) for the uninjected eyes in the medium control group. To assess the effect of the treatment we compared the means of the ratios of the treated and untreated eyes in the virus and medium-injected groups as well as between the virus-injected and medium-injected eyes
(Fig. 1B). The mean ratio in tumor size between the injected and the uninjected eyes in the medium-only group was 0.88 (95% confidence interval 0.61 – 1.28) and the difference was not significant (P Å 0.24). This result indicates that the injection of medium did not affect tumor size. The mean ratio in tumor size in the virus-treated group was 0.74 (0.56– 0.98). This difference was significant (P Å 0.019), indicating that the virus treatment resulted in smaller tumors. When we compared the eyes injected with virus to the eyes injected with medium, the mean ratio was 0.54 (0.35– 0.83) and the difference was highly significant (P Å 0.0063). It should be noted, however, that the 95% confidence intervals overlap in all groups. Representative low-magnification examples of
AID VY 8414
vira
/
6a29$$$422
01-31-97 20:35:39
AP: Virology
286
SHORT COMMUNICATION
sections through an untreated eye and a virus-treated eye are shown in Figs. 2A and 2B respectively. Morphological analysis of the tumors in the two groups was also carried out to determine if there were differences in the degree of involvement of the choroid, vitreous, iris – ciliary body angle, cornea, retinal pigment epithelium, and optic nerve. The criteria for this analysis have been described previously (31). A defining characteristic of retinoblastomas is the presence of Homer Wright rosettes, which represent attempts by the tumor cells to differentiate into photoreceptors (23, 24). These rosettes are characterized by a central luscent space that represents the attempt by the cells to form outer segments. The cell bodies surrounding the outer segments have a relatively constant shape and size with a high nuclear-to-cytoplasmic ratio. As shown in Table 1, the only differences noted were that the tumors in the RE6-treated eyes were less differentiated (P Å 0.03) and had fewer Homer Wright rosettes (P Å 0.03) than the control tumors. Figures 2C and 2D show representative intermediate-magnification views of a virus-injected (2C) tumor and an uninjected (2D) tumor. The virus-injected tumor shows an extensive area lacking the typical Homer Wright rosettes compared to the uninjected tumor, in which the majority of the cells are organized into rosettes. Higher magnification views in Figs. 2E and 2F show the organized structure of the rosettes in the uninjected tumors compared to the pleomorphic cells and disorganization in the virus-injected tumors. There was no difference in the amount of necrosis in the two groups at 3 to 5 weeks posttreatment. To test the hypothesis that the antitumor effect was due to the ability of the RE6 virus to grow better in tumor cells compared to normal eye tissue, we injected the right eyes of transgene-positive or transgene-negative littermates (five mice per group) with 1 1 10 6 PFU of RE6 and at various times postinfection (PI) we assayed the eyes for the amount of infectious virus using procedures we described previously (34). The results are shown in Fig. 3. On Day 1 PI, the RE6 titers were 2 1 104 and 1 1 10 3 for the LHb-Tag-negative and LHb-Tag-positive mice, respectively. On Day 2 PI, the titers were 10-fold higher in the LHb-Tag-negative mice. On Day 5 PI, there was only a 2-fold difference and by Day 7, infectious virus had essentially cleared in both groups of mice. Statistical analysis (t test) failed to reveal significant differences (P ú 0.05) in viral growth in the two groups.
Injection of virus into immunocompetent animals might be expected to induce an inflammatory response which could contribute to the antitumor activity. We therefore compared the degree of inflammation in the virus-injected (treated) and untreated tumors at 3 weeks posttreatment using procedures described previously (31). The presence of inflammation was higher in the treated eyes (44%) compared to the control eyes (31%). However, the groups were not significantly different (P Å 0.43) (data not shown). When we characterized the inflammation with respect to cell type, by histological criteria, we found that neutrophils were the predominant infiltrating cell type; however, the comparisons between treated and untreated eyes failed to reveal significant differences with regard to the type of inflammatory cell (P ú 0.05) within the tumor (data not shown). To assess the distribution of virus in the eyes of the treated mice, the right eyes of 16- to 18-week-old transgene-bearing mice were injected intravitreally with 1 1 106 PFU of RE6. At 3 days PI, the eyes were removed, fixed in 10% buffered formalin, embedded in paraffin, and sectioned. Immunohistochemical detection of viral antigen was carried out using a commercially available kit (Vectastain ABC, Vector Laboratories, Burlingame, CA) and commercially available primary anti-HSV antibody (BO114, Dako Corp., Carpinteria, CA). Representative examples of the immunohistochemical staining are shown in Fig. 4. In tumor tissue, HSV antigen was detected only in focal areas (Fig. 4B) and did not appear to spread extensively, even though viral titers were highest at this time (Fig. 3). Viral antigen was also present in retinal pigment epithelial cells in focal areas (Fig. 4D). We found no evidence of immunohistochemical staining in any other ocular tissues. One possible mechanism for the selective effects of the attenuated HSV on tumors is that the mutant viruses grow better in rapidly dividing tumor cells than in normal tissues. For attenuated viruses lacking either viral thymidine kinase or ribonucleotide reductase, this hypothesis is consistent with the proposed role of these enzymes in viral replication. However, the in vivo growth of mutant viruses in tumor and normal tissue was not compared in most of the previous studies (12 – 21). In the one previous study using a g134.5 mutant where viral growth was tested in vivo, replication was more efficient and prolonged in tumor cells compared to normal brain (17). In contrast, we found that viral growth in the eyes of LHb-
FIG. 2. Tumor histopathology. (A) Representative example of an untreated (control) eye (original magnification, 123). (B) RE6-treated eye from the same animal. C, cornea; L, lens; T, tumor; N, optic nerve; R, retina (original magnification, 123). (C) Intermediate-magnification view of tumor from an untreated eye showing Homer Wright rosettes (original magnification, 1230), some of which are marked with asterisks. (D) Intermediatemagnification view of the tumor from an RE6-treated eye (original magnification, 1230). Note that some cells attempt to form rosettes but these are poorly organized. Note also the variable morphology of the cells. (E) High-magnification view of a tumor from an untreated eye showing Homer Wright rosettes with the central luscent space and surrounding cell bodies (original magnification, 1460). Note the uniformity in cell size and shape. (F) High-magnification view of a tumor from an eye treated with RE6 (original magnification, 1460). Note the lack of rosettes and the variability in cell size and shape.
AID VY 8414
/
6a29$$$422
01-31-97 20:35:39
vira
AP: Virology
SHORT COMMUNICATION
01-31-97 20:35:39 AID VY 8414
/
6a29$$8414
vira
287
AP: Virology
288
SHORT COMMUNICATION TABLE 1 Characteristics of Retinoblastoma Tumorsa Untreated (left eye)b
Treated (right eye) Number
%
Number
%
Pc
Number of eyes graded Tumor present Retina totally involved % of retina uninvolved Full thickness involvement
17d 17 5 58.13 ({33.58) 10
100 100 29
16 e 16 4 62.38 ({30.82) 10
94 100 16
1 1 1 0.63 0.67
Invasion of vitreous Invasion of ciliary body Invasion of iris Invasion of angle Invasion of cornea Invasion of choroid Optic nerve invasion Calcification % undifferentiated Necrosis present Flexner-Wintersteiner rosettes Homer Wright rosettes
7 4 2 2 0 7 2 12 34.69 ({28.37) 9 0 65.5 ({29.15)
41 24 12 12 — 41 12 71
4 4 3 4 1 5 2 9 14.38 ({20.32) 8 3 85.63 ({20.32)
25 16 19 25 6 32 13 56
Observation
Retinal involvement
Invasion of other ocular structures
Rosettes
59
53 —
63
16 19
0.43 0.71 0.33 0.33 0.33 0.75 1 0.5 0.03 1 — 0.03
a
All data from the second study, sections through pupil optic nerve axis. Uninjected paired eyes. c Unpaired t test. d For statistical analysis only the 16 paired eyes were included. e The intraocular contents of one eye were lost and could not be graded. b
Tag-positive mice was not significantly different than in the eyes of nontransgenic littermates (Fig. 3). Viral replication, as measured by staining for viral antigen, was restricted to focal areas of the tumors and to limited regions of retinal pigment epithelium. The magnitude of the antitumor effects seems surprising given the restricted distribution of viral antigen. However, these results are similar to those of Randazzo et al. (17) and suggest that the virus does not spread extensively in the
FIG. 3. Replication of RE6 virus in the eyes of transgene-positive (Rb/) and transgene-negative (Rb0) mice. Each data point is the mean titer from five eyes.
tumors. When considered together, these results suggest that the differences in tumor size between virus-injected and control mice are not related to viral replication. Additional mechanisms for the selective antitumor effects include the induction of an inflammatory response that kills tumor cells through bystander effects, or the induction of tumor-specific immunity as a consequence of viral infection. The observed inflammation in areas positive for viral antigens at Day 3 PI is consistent with possible bystander effects, but the data are not conclusive. The presence of similar degrees of inflammation in treated and untreated eyes in the same animal at 3 to 5 weeks PI makes it difficult to draw conclusions about the induction of tumor-specific immunity and further studies are needed to distinguish the effect of these different mechanisms. Several morphological criteria of the tumors were also analyzed. In addition to the significant difference in tumor size, we found that the tumors in the treated mice were less differentiated, having fewer Homer Wright rosettes than the untreated tumors. These results are tentative due to their marginal P values and the large number of parameters that were measured. One possible interpretation of this finding is that the virus preferentially infects or is more cytolytic in the more differentiated population of retinoblastoma cells in the tumor. This hypothesis would be consistent with previous results showing that cell type and cell state affect the growth of g134.5 mu-
AID VY 8414
vira
/
6a29$$$422
01-31-97 20:35:39
AP: Virology
FIG. 4. Immunohistochemical analysis of viral antigen distribution. Three days postinfection the eyes were collected, fixed in 10% formalin in phosphatebuffered saline, embedded in paraffin, sectioned, and stained using polyclonal anti HSV-1 antisera (Dako, Carpintera, CA), and an ABC vectastain kit (Vector Laboratories, Burlingame, CA). (A) Control (uninjected) eye. (B) RE6-injected eye. The arrows in B point to areas of positive staining for viral antigen. (C) Control eye showing retinal pigment epithelium. (D) RE6-injected eye showing viral antigen in RPE cells. The arrows in D point to RPE cells containing viral antigen. The bars represent 16 mm.
SHORT COMMUNICATION
AID VY 8414
/
6a29$$8414
01-31-97 20:35:39
vira
289
AP: Virology
290
SHORT COMMUNICATION
tants (28). Further studies are needed to determine if the RE6 virus preferentially replicates in more differentiated tumor cells and whether this is related to the g134.5 mutation in RE6. Although we have shown that intravitreal injection of RE6 results in a significant difference in retinoblastoma size, the 95% confidence intervals overlapped and we did not achieve complete regression of the tumor in any of the treated mice. This is in contrast to other reports with neuronal cell implants in which elimination of the tumors was achieved at least in a portion of the treated animals (14, 17, 21). The reason or reasons for our less robust results are not clear, but there are several possibilities. Several of these factors could also account for the variability in the results as reflected in the overlapping 95% confidence intervals. Our data suggest that viral growth, if it occurred at all, was not different between the transgene-bearing mice and transgene-negative mice, and lack of viral replication could have reduced the effectiveness of the treatment. Animal models utilizing tumor cell implants differ substantially from our retinoblastoma model in that tumors form in 2 to 3 weeks in the implant models compared to 16 to 18 weeks for the retinoblastomas. Thus, there is likely to be a substantial difference in the rate of cell division in the tumor cells between the two models that could account for the reduced viral growth in our model. Another possibility is that the mice mounted an immune response against infection that subsequently reduces the effectiveness of the virus. If this were true, we would expect that the inflammatory response in the virustreated eyes would have been different quantitatively, qualitatively, or both, and we found no evidence for this. It should be pointed out, however, that our analysis of immune responses was limited to histopathology and a more extensive study of the immune response to the virus and the tumor is needed. Tumor cell lines from these mice are not readily available as yet. Thus, direct determination of the ability of RE6 to grow in tumor cells was not tested. It is possible that retinoblastoma cells are resistant to the growth of g134.5 mutants in general or RE6 in particular. However, the growth of virus in cultured murine retinoblastoma cells may not reflect the in vivo situation and the direct measurement of replication in vivo may be more appropriate. Finally, the limited efficacy of the treatment in our model may be due to technical difficulties. At the time of treatment, the tumors are just extending into the vitreous. In these circumstances, the majority of the virus may be injected into the vitreous and not directly into the tumor. Thus, RE6 virus may have infected the tumor poorly or the majority of the virus may have failed to reach the tumor. The cell implant models have involved either direct injection into the tumor or preinfection of the cell with virus prior to implantation. Optimum results in our
model may require injections at later than 16 to 18 weeks when tumors are larger and direct intratumoral injection is technically less difficult. In summary, we have shown that treatment of spontaneously arising retinoblastoma tumors in a transgenic mouse model with a neuroattenuated HSV mutant (RE6) results in a significant difference in tumor size. This finding represents the first report of successful treatment of a tumor with a mutant HSV without the use of tumor explants. This model more closely mimics the situation in the treatment of human cancer. More importantly, the model will allow us to test viral treatment of tumors in situations where tumor alloantigens from explants are not a significant factor. The fact that these animals are immunocompetent will also allow us to examine the combined effects of the immune response and viral infection on tumor development.
AID VY 8414
vira
/
6a29$$$422
01-31-97 20:35:39
ACKNOWLEDGMENTS These studies were supported in part by NIH Grants EY07736 (C.R.B.) and EY01917 (D.M.A.), by the University of Wisconsin Cancer Center Core Grant SP30 CA14520, by an unrestricted grant from Research to Prevent Blindness, Inc. to the Department of Ophthalmology and Visual Sciences, and the Wisconsin Lions. C.R.B. was a Research to Prevent Blindness, Inc. Robert E. McCormick Special Scholar during a portion of these studies. The authors thank Bernadette Bull for administrative assistance.
REFERENCES 1. Levin, A. L., Shelinji, G. E., and Gutin, P. H., In ‘‘Cancer: Principles and Practice of Oncology’’ (V. T. Vita, S. Hellman, and S. A. Rosenberg, Eds.), pp. 1557–1611. Lippincott, Philadelphia, 1989. 2. Mahaley, M. S., Jr., Mettlin, C., Natarajan, N., Laws, E. R., and Peace, B. B., J. Neurosurg. 71, 826– 836 (1989). 3. Salazar, D. M., Rubin, P., Feldstein, M. L., and Pizzutiello, R., Int. J. Radiat. Oncol. Biol. Phys. 5, 1733–1740 (1979). 4. Danmas-Duport, C., Scheithauer, B., O’Fallon, J., and Kelly, P., Cancer 62, 2152–2165 (1988). 5. Kim, T. S., Halliday, A. L., Hedley-Whyte, E. T., and Convery, K., J. Neurosurg. 74, 27 –37 (1991). 6. Brandt, C. R., Kintner, R. L., Pumfery, A. M., Visalli, R. J., and Grau, D. R., J. Gen. Virol. 72, 2043– 2049 (1991). 7. Chou, J., Kern, E., Whitley, R. J., and Roizman, B., Science 250, 1262–1266 (1990). 8. MacLean, A. R., Embed-Fareed, M., Robertson, L., Harland, J., and Brown, S. M., J. Gen. Virol. 72, 631–639 (1991). 9. Coen, D. M., Kosz-Vienchak, M., Jacobson, J. G., Lieb, D. A., Bogard, L. L., Schatter, P. A., Tyler, K. L., and Knipe, D. V., Proc. Natl. Acad. Sci. USA 86, 4736–4740 (1989). 10. Thompson, R. L., and Stevens, J. G., Virology 131, 171– 179 (1983). 11. Skinner, G. R. B., Cowan, M., Davies, J., Brookes, K., Billstrom, M., and Buchan, A., Br. J. Exp. Pathol. 69, 495– 504 (1988). 12. Martuza, R. L., Malick, A., Markert, J. M., Ruffner, K. L., and Coen, D. M., Science 252, 854– 856 (1991). 13. Markert, J. M., Coen, D. M., Malick, A., Mineta, T., and Martuza, R. L., J. Neurosurg. 77, 590–594 (1992). 14. Markert, J. M., Malick, A., Coen, D. M., and Martuza, R. L., Neurosurgery 32, 597– 603 (1993). 15. Mineta, T., Rabkin, S. D., and Martuza, R. L., Cancer Res. 54, 3963– 3966 (1994).
AP: Virology
SHORT COMMUNICATION
291
16. Mineta, T., Rabkin, S. D., Yazaki, T., Hunter, W. D., and Martuza, R. L., Nature Med. 1, 938– 943 (1995). 17. Randazzo, B. P., Kesari, S., Gesser, R. M., Alsop, D., Ford, J. C., Brown, S. M., MacLean, A., and Fraser, N. W., Virology 211, 94 – 101 (1995). 18. Chambers, R., Gillespie, G. Y., Soroceaun, L., Andreansky, S., Chatterjee, S., Chou, J., Roizman, B., and Whitley, R. J., Proc. Natl. Acad. Sci. USA 92, 1411– 1415 (1995). 19. Jia, W. W-G., McDermott, M., Goldie, J., Cynader, M., Tan, J., and Tufaro, F., J. Natl. Cancer Inst. 86, 1209– 1215 (1994). 20. Boviatsis, E. J., Scharf, J. M., Chase, M., Harrington, K., Kowall, N. W., Breakefield, X. O., and Chiocca, E. A., Gene Ther. 1, 323– 331 (1994). 21. Kaplitt, M. G., Tjuvajev, J. G., Leib, D. A., Berk, J., Pettigrew, K. D., Posner, J. B., Pfaff, D. W., Rabkin, S. D., and Blasberg, R. G., J. Neuro-Oncol. 19, 137– 147 (1994). 22. Thompson, R. L., Rogers, S. K., and Zerhusen, M. A., Virology 172, 435– 450 (1989). 23. Bolovan, C. A., Sawtell, N. M., and Thompson, R. L., J. Virol. 68, 48 – 55 (1994). 24. Javier, R. T., Thompson, R. L., and Stevens, J. G., J. Virol. 61, 1978– 1984 (1987).
25. MacLean, A., Robertson, L., McKay, E., and Brown, S. M., J. Gen. Virol. 72, 2305–2310 (1991). 26. Taha, M. Y., Brown, S. M., Clements, G. B., and Graham, D. I., J. Gen. Virol. 71, 1597–1601 (1990). 27. Whitley, R. J., Kern, E. R., Chatterjee, S., Chou, J., and Roizman, B., J. Clin. Invest. 91, 2837– 2843 (1993). 28. Brown, S. M., Harland, J., MacLean, A. R., Podlech, J., and Clements, J. R., J. Gen. Virol. 75, 2367– 2377 (1994). 29. Windle, J. J., Albert, D. M., O’Brien, J. M., Marcus, D. M., Disteche, C. M., Bernards, R., and Mellon, P. L., Nature 343, 665– 669 (1990). 30. O’Brien, J. M., Marcus, D. M., Bernards, R., Carpenter, J. L., Windle, J. J., Mellon, P., and Albert, D. M., Arch. Ophthalmology 108, 1145–1151 (1990). 31. Albert, D. M., Marcus, D. M., Gallo, J. P., and O’Brien, J. M., Invest. Ophthalmol. Vis. Sci. 33, 2354 –2364 (1992). 32. Backer, R. A., Chambers, J. M., and Wilks, A. R., ‘‘The New S Language.’’ Wadsworth’s Brooks/Cole, Pacific Grove, CA, 1988. 33. Pagano, M., and Gauvrean, K., ‘‘Principles of Biostatistics.’’ Duxbury, Belmont, CA, 1993. 34. Brandt, C. R., and Grau, D. R., Invest. Ophthalmol. Vis. Sci. 31, 2214– 2223 (1990).
AID VY 8414
vira
/
6a29$$$423
01-31-97 20:35:39
AP: Virology