Photodynamic Therapy Plus Low-Dose Cyclophosphamide Generates Antitumor Immunity in a Mouse Model

Photodynamic therapy plus low-dose cyclophosphamide generates antitumor immunity in a mouse model

Ana P. Castano*†‡, Pawel Mroz*†, Mei X. Wu*†§, and Michael R. Hamblin*†§¶

*Wellman Center for Photomedicine, Massachusetts General Hospital, Boston, MA 02114; †Department of Dermatology, Harvard Medical School, Boston, MA 02115; and §Division of Health Sciences and Technology, Harvard–Massachusetts Institute of Technology, Cambridge, MA 02139

Edited by Mark Naylor, University of Oklahoma Health Sciences Center, Tulsa, OK, and accepted by the Editorial Board February 21, 2008 (received for review September 28, 2007) Photodynamic therapy (PDT) is a modality for the treatment of cidal effect include direct cytotoxicity to tumor cells (13), cancer involving excitation of nontoxic photosensitizers with shutting down of the tumor vasculature, thus starving the tumor harmless visible light-producing cytotoxic reactive oxygen species. of oxygen or nutrients (14), and the induction of a host immune PDT causes apoptosis and necrosis of tumor cells, destruction of the response (15). The precise mechanisms involved in the PDT- tumor blood supply, and activation of the immune system. The mediated induction of antitumor immunity are not yet com- objective of this study was to compare in an animal model of pletely understood, but advances are constantly being made (16). metastatic cancer PDT alone and PDT combined with low-dose It is known that after PDT there is an acute inflammatory cyclophosphamide (CY) a treatment that has been proposed to response in the treated area that rapidly causes a massive deplete regulatory T cells (T-regs) and increase the immune re- regulated invasion of neutrophils (17), mast cells, and mono- sponse to some tumors. We used J774 tumors (a highly metastatic cytes/macrophages (18) that may outnumber resident cancer reticulum cell sarcoma line) and PDT with benzoporphyrin deriv- cells. Most notable is a rapid accumulation of large numbers of ative monoacid ring A, verteporfin for injection (BPD; 1-mg/kg neutrophils, which have been shown to have a profound impact injected i.v. followed after 15 min by 150 J/cm2 of 690-nm light). CY on PDT-mediated destruction of tumors, and it has been shown (50 or 150 mg/kg i.p.) was injected 48 h before light delivery. PDT that depletion of neutrophils in tumor-bearing mice decreased alone led to tumor regressions and a survival advantage but no the PDT-mediated tumor cure rate (19). The acute inflamma- permanent cures were obtained. BPD–PDT in combination with tory response is thought to attract dendritic cells that can take low-dose CY (but not high-dose CY) led to 70% permanent cures. up necrotic and apoptotic tumor cells and migrate to lymph Low-dose CY alone gave no permanent cures but did provide a nodes (LNs) where they prime tumor-specific cytotoxic T cells survival advantage and was shown to reduce CD4؉FoxP3؉ T-regs (CTLs). The activity of CTLs is not limited to the original in lymph nodes, whereas high-dose CY reduced other lymphocyte PDT-treated site but can include disseminated and metastatic classes as well. Cured animals were rechallenged with J774 cells, lesions of the same cancer (20). and the tumors were rejected in 71% of mice. Cured mice had Reports dating from as early as 1981 (21, 22) have shown that tumor-specific T cells in spleens as determined by a 51Cr release under some conditions low-dose cyclophosphamide (CY) (a assay. We conclude that low-dose CY depletes T-regs and poten- traditional cytotoxic cancer drug that damages tumor DNA) can tiates BPD–PDT, leading to tumor cures and memory immunity. potentiate antitumor immunity in mouse models. Several mechanisms have been proposed to explain this observation, including immunosuppression ͉ cytotoxic T lymphocytes ͉ benzoporphyrin depletion of suppressor T cells (21), reduction of immunosup- derivative ͉ regulatory T cells ͉ tumor-associated antigen pressive cytokines (23), and antiangiogenesis (24). Since the explosion of interest in regulatory T cells (T-regs) expressing CD4, CD25, and the transcription factor Foxp3 (25, 26), it has espite a long-held realization that cancer leads to an im- become accepted that low-dose CY selectively depletes T-regs in paired immune response, and speculation that the immune D mice and thus increases both the priming and effector phases of system is capable of mounting an effective antitumor im- the antitumor immune response (27). mune response if the correct stimulation is applied, cancer Here, we report that PDT can be effectively combined with immunotherapy has not yet reached the status of a mainstream low-dose CY to cure a highly metastatic mouse tumor and therapy (1, 2). Many approaches have been suggested for re- produce tumor-specific CTLs and potent memory immunity. versing this ability of tumors to evade the immune system (3), including nonspecific biological response modifiers (4), cytokine Results therapies (5), cancer vaccine approaches (6), cell-mediated J774 Tumor. The J774 tumor is a BALB/c mouse reticulum cell immunotherapies (7), and gene therapy (8). Because it is well sarcoma (macrophage tumor) that is aggressive, invasive, and known that most deaths from cancer are caused by metastatic metastatic as shown by hematoxylin and eosin-stained histology tumor, the ideal cancer treatment would be one that not only (Fig. 1 A–D). It is always fatal to mice and difficult to cure; its effectively eradicated the primary tumor, but in addition, simul- immunogenicity is unknown. After sequential amputations after taneously activated the immune system to attack distant metastases. Although surgery and radiotherapy can often destroy primary tumors with high efficiency, at high doses they are Author contributions: A.P.C., P.M., M.X.W., and M.R.H. designed research; A.P.C. and P.M. MEDICAL SCIENCES thought to have an immunosuppressive effect, thus allowing performed research; M.X.W. contributed new reagents/analytic tools; A.P.C., P.M., and micrometastases to grow unchecked. M.R.H. analyzed data; and P.M. and M.R.H. wrote the paper. Photodynamic therapy (PDT) uses a photoactivated dye or The authors declare no conflict of interest. photosensitizer (PS) in combination with visible light that pro- This article is a PNAS Direct Submission. M.N. is a guest editor invited by the Editorial Board. duces reactive oxygen species and destroys tumor cells and ‡Present address: Renal Division, Brigham and Women’s Hospital, Department of Medicine, malignant tissue (9–11). PDT is approved for multiple indica- Harvard Medical School, Boston, MA 02115. tions in the United States and many other countries (12). ¶To whom correspondence should be addressed. E-mail: [email protected]. Mechanisms that have been shown to be involved in the tumori- © 2008 by The National Academy of Sciences of the USA

www.pnas.org͞cgi͞doi͞10.1073͞pnas.0709256105 PNAS ͉ April 8, 2008 ͉ vol. 105 ͉ no. 14 ͉ 5495–5500 Downloaded by guest on October 4, 2021 Fig. 1. H&E histology of tissue sections removed from J774 tumor-bearing mice. (A) Tumor invading thigh muscle after s.c. implantation. (B) Metastatic tumor invading omental fat pad. (C) Metastatic nodule in lung. (D) Metastatic nodule in liver. (Scale bar: 100 ␮m.) (E) Survival in days of 10 mice after sequential daily amputation of the leg bearing s.c. J774 tumors in individual mice commencing 24 h after tumor injection.

the first day postinjection, 10 mice were evaluated for survival, treatment control tumor-bearing mice are shown in Fig. 2B. The and all mice were dead after 48 days (Fig. 1E). The mouse that median survival of the high-dose CY alone group was 30 days died first was the one who had the latest surgery (10 days after (not significantly different from control tumor-bearing mice), cell injection). Early surgery therefore predicted longer survival and the median survival of the high-dose CY plus PDT group consistent with J774 being a highly metastatic tumor that was 42 days, but these curves were not significantly different metastasizes as early as 24 h after injection. from each other or from control tumor-bearing mice. However, PDT combined with high-dose CY gave significantly shorter Ͻ PDT. All of the PDT-treated mice had a response typical of median survival than PDT alone (P 0.05). vascular PDT in their treated tumors with the development of a black eschar at 2–3 days post-PDT and subsequent partial or complete disappearance of the tumor with good local tissue healing.

Mouse Survival. Nine mice were treated with PDT alone and the survival curves of these mice compared with 10 control J774 tumor-bearing mice is shown in Fig. 2A. The median survival of the PDT group was 49 days, and the median survival of the control untreated mice was 35 days. These two survival curves were statistically different by log-rank test (P Ͻ 0.005), but nevertheless all PDT-treated mice died. Fifteen mice treated with the combination of both therapies [benzoporphyrin derivative monoacid ring A, verteporfin for injection (BPD)–PDT plus CY] had a dramatically different outcome from those treated with PDT alone. Fig. 2A shows that the combination group had a long-term (Ͼ90 days) survival rate of Ϸ66%. This survival curve was statistically different from curves in control (P Ͻ 0.0001) and BPD–PDT alone (P ϭ 0.0001) groups. This is a very important result when compared with no long-term survivals whatsoever in mice treated with BPD–PDT alone. It was important to study the effect of administering a single low dose of CY alone on mice bearing J774 tumors. Thirteen mice were therefore injected with CY (50 mg/kg i.p.) 7 days after the tumor was implanted. These mice had a median survival of 49 days, and the Kaplan–Meier survival curve is shown in Fig. 2A and was significantly different from the following groups: control (P Ͻ 0.0001), and BPD–PDT ϩ low-dose CY (P Ͻ 0.0001). There was no difference between the survival of the PDT alone and low-dose CY-alone groups. To discriminate between the antitumor effect of CY and its ϩ Fig. 2. Kaplan–Meier survival curves of groups of mice. (A) No treatment immune potentiating activity, we repeated the PDT CY and tumor, n ϭ 10; PDT alone, n ϭ 9; low-dose CY, n ϭ 10; low-dose CY ϩ PDT, n ϭ CY-alone groups of mice with a high dose (150 mg/kg) rather 15. (B) No treatment tumor, n ϭ 10; PDT alone, n ϭ 9; high-dose CY, n ϭ 10; than a low dose (50 mg/kg) of CY. The Kaplan–Meier survival high-dose CY ϩ PDT, n ϭ 10. Mice were killed when the primary tumor reached curves of these mice together with that of PDT alone and no 1.5 cm in diameter or their body weight dropped Ն15%.

5496 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0709256105 Castano et al. Downloaded by guest on October 4, 2021 Fig. 3. Radiochromium release assay for cell-mediated cytotoxicity carried out by incubating LN cells isolated from either cured mice that had rejected a rechallenge with J774 cells, J774 tumor-bearing mice, or naı¨vemice. Target cells were either J774 (specific) or EMT-6 (irrelevant) cells that had been loaded with radioactive chromate. The indicated E/T ratios were incubated together for 4 h, and specific 51Cr release was measured by scintillation counting. Values are means of duplicate determinations from three separate mice, and bars are SD. Significance was determined by ANOVA. P values are ***, Ͻ 0.001; **, Ͻ 0.01; *, Ͻ0.05.

Tumor Rechallenge. To test whether the total cures obtained with the combination of low-dose CY plus PDT were caused by activation of the host immune system, we rechallenged the mice Fig. 4. Quantiﬁcation of T-regs and splenocytes. (A) Percentages of CD4ϩFoxP3ϩ double positive LN cells from LNs extracted from mice with or with 1 million J774 tumor cells in the contralateral leg. The without J774 tumors 3 days after low-dose CY injection. (B) Numbers of rechallenge procedure showed the following rates of success: in splenocytes from whole spleens removed from mice as described above. the case of BPD-PDT ϩ CY five of seven mice rejected the Values are means of duplicate determinations from three separate mice, and tumor rechallenge (71% success). As a control for tumorigenic- bars are SD. Signiﬁcance was determined by ANOVA. ity of the J774 cells five of five naive mice injected with the same dose of the same cells grew tumors (zero of five rejections). These two groups were significantly different by Fisher’s exact to 22.7 Ϯ 1% (P Ͻ 0.01). J774 tumor-bearing mice injected with P Ͻ 50 mg/kg CY showed a marked decrease in double positive cells test ( 0.05). To test the tumor specificity of the rejection of Ϫ the tumor rechallenge the mice that rejected J774 cells were down to 5 Ϯ 1% (P Ͻ 10 6), whereas mice bearing J774 tumors challenged for a second time with a tumorigenic dose of EMT6 injected with 150 mg/kg had a significantly higher percentage of cells. This is a BALB/c mammary sarcoma line that is moderately double positive cells (8.1 Ϯ 0.3%) compared with low-dose CY immunogenic. Seven of seven mice (100%) that had previously (P Ͻ 0.05). As a measure of the total number of lymphocytes we rejected J774 cells grew tumors with EMT6 cells demonstrating also measured the total number of splenocytes isolated per that the induced immune response is specific to J774 tumors. mouse. The results given in Fig. 4B show that mice with J774 tumors demonstrated a consistent splenomegaly with a Ϸ50% CTL Chromium Release Assay. We carried out a radiochromium increase in the total number of splenocytes from 112 million to release cell-mediated cytotoxicity assay on splenocytes isolated 165 million (P Ͻ 0.005). After low-dose CY the mean number of from either cured mice that had rejected a rechallenge with J774 splenocytes decreased Ͼ60% compared with tumor-bearing Ϫ cells, J774 tumor-bearing mice, or naı¨ve mice and compared the mice to 66 million (P Ͻ 10 4), and after high-dose CY the cytolysis of specific target cells (J774) or irrelevant nontarget number decreased even further to 38 million (P Ͻ 0.001 vs. cells (EMT6). Fig. 3 shows that lymphocytes from cured mice low-dose CY). These results are consistent with low-dose CY that had rejected a rechallenge with J774 cells were able to lyse killing a major fraction of the overall splenocyte population, but ϩ ϩ J774 targets in a effector/target (E/T)-dependent manner with with some selectivity for CD4 FoxP3 T-regs, whereas the statistical significance at E/T ratios Ն10:1. There was some high-dose CY kills even more splenocytes and lymphocytes, thus suggestion of cytolytic activity exhibited by lymphocytes from making the relative proportion of T-regs appear to increase. tumor bearing mice at the highest E/T ratios (100:1). The absence of any cytolysis of J774 cells by lymphocytes from naı¨ve Discussion mice or by lymphocytes from cured mice against an irrelevant This study has demonstrated that PDT combined with low-dose target (EMT6 cells) demonstrates the tumor specificity of the CY produces a dramatic improvement in survival and significant observed cytolysis. numbers of cures in a highly metastatic mouse tumor model. No cures but some survival advantage were seen with either treat- Flow Cytometry for T-Regs. To elucidate the effects of treatment ment alone, whereas when PDT was combined with high-dose of the mice with either low-dose or high dose CY, we examined CY there was no additional benefit. There was a long-term

the proportion of T cells from mouse LN that expressed CD4 and memory immunity generated by the combination treatment of MEDICAL SCIENCES intracellular FoxP3 2 days after injection with either PBS, BPD–PDT and low-dose CY that allowed cured mice to reject a low-dose CY, or high-dose CY. Many reports have shown that rechallenge with a tumorigenic dose of J774 cells, and the the intracellular transcription factor FoxP3 is the most specific presence of lymphocytes that specifically caused cytolysis of measure of T-regs in mice (28). The mean fraction of target cells ex vivo was shown. CD4ϩFoxP3 double positive splenocytes gated on the CD4 The PS that we used, BPD after 15 min, is still localized in the population is shown in Fig. 4A. Mice with J774 tumors injected vasculature. It has been shown that ‘‘vascular’’-type PSs such as with PBS alone showed a small, but significant, increase in BPD have greater effectiveness if illumination is carried out at double positive cells compared with naı¨ve mice from 18 Ϯ 1% early time points after injection (29). Under these conditions

Castano et al. PNAS ͉ April 8, 2008 ͉ vol. 105 ͉ no. 14 ͉ 5497 Downloaded by guest on October 4, 2021 damage is caused to the tumor by shutdown of vasculature, loss However, considering the further decrease in lymphocytes of perfusion, and deprivation of the tumor of oxygen and caused by high-dose CY, we suggest that populations of CD4 nutrients normally supplied by the circulation. The cells in the T-helper (Th) cells and CD8 effector T cells necessary for tumor tissues undergo necrosis and apoptosis, and these pro- development of an effective antitumor immunity were depleted cesses increase the release of tumor antigen that can be recog- by high-dose CY, but not by low-dose CY, and that this differ- nized and presented to the immune system, leading to the ence explains the ineffectiveness of high-dose CY in combina- generation of a tumor-specific immune response. Other workers tion with PDT in inducing immunity. High-dose CY plus PDT who have studied the induction of antitumor immunity after actually reduced the survival achieved with PDT alone, suggest- PDT have used PSs with a longer drug–light interval (20, 30), and ing that there is some contribution of the host immune system to these might be less vascular in mode of action. No one to our the outcome of PDT alone, although not sufficient to lead to knowledge has compared vascular and cellular PDT regimens to cures. It is at present uncertain exactly what is the mechanism for see which is more efficient in inducing immune response against selective toxicity of low-dose CY toward T-regs. Reports have tumors. shown a switch from Th2 to Th1 cytokines after low-dose CY The J774 tumor model has not been studied much in relation (45), inhibition of proliferation in the more rapidly cycling T-regs to its immunogenicity or to its susceptibility to cancer immuno- (46), induction of apoptosis and functional inhibition in T-regs therapy approaches. Its highly metastatic nature does mean that (27), and inhibition of inducible nitric oxide synthase that a fully curative therapy has to involve immune stimulation, as controls T-reg numbers (47). even complete surgical removal as soon as 1 day after implan- PDT has mostly been thought of as a local therapy in which cell tation did not give a cure. Because these tumor cells are derived and tissue destruction happens primarily in the area that is from the monocyte/macrophage lineage they do express MHC exposed to light and the immediately surrounding areas. How- class II molecules in addition to MHC class I molecules, and ever, mounting evidence suggests that, as opposed to many therefore they might be able to prime the immune system to common cancer therapies that are largely immunosuppressive, recognize their own antigens. Our findings that PDT, in com- PDT can stimulate the immune system (16). In this respect PDT bination with low-dose CY, induces specific CTLs capable of ex is similar to adjuvant-enhanced laser immunotherapy that has vivo cytolysis suggests that effective tumor-rejection antigen(s) also been shown to be capable of destroying local tumors, while do exist in J774 cells but remain to be identified. The existence at the same time sensitizing the immune system to destroy distant of possible antigens in J774 cells capable of being recognized by metastases (48–51). The fact that we observed only long-term the immune system was reported by Fedoseyeva et al. (31) who cures with the low-dose CY combination treatment suggests that showed that J774 tumors growing in BALB/c mice generated in this tumor model (and perhaps in many others) there is some specific CD4 cells that recognized peptides derived from mutant host factor that is fighting against the immune stimulating effect p53 protein. of PDT. We suggest that this factor could be CD4ϩFoxP3ϩ T-regs protect the host from autoimmune disease by suppress- T-regs because of their susceptibility to low-dose CY and as ing self-reactive T cells and may also block antitumor immune shown by our flow cytometry results. The effect of low-dose CY responses mediated by tumor-specific T cells (26). The suppres- alone on the tumor was actually much more pronounced than the sive effect of naturally occurring T-regs against tumor-specific effect of high-dose CY alone (compare Fig. 2 A and B). This CD8 T cells has been demonstrated both in vitro (32) and in vivo finding suggests that the effects of CY on J774 tumor are caused (33, 34). Local depletion of CD4 T cells inside the tumor led to by an immunostimulatory effect rather than by the traditional eradication of well established tumors and development of cytotoxic role of CY. long-term antitumor memory, suggesting that suppression of Much work remains to be carried out to conclusively establish the antitumor immunity by T-regs occurs predominantly at the scope of the antimetastatic effect and antitumor immunity effect of tumor site and that local reversal of suppression, even late during the combination of low-dose CY and PDT. Does substitution of tumor development, can be an effective treatment (35). low-dose CY with an anti-CD25 antibody produce the same effect Whether there is a systemic increase of T-regs in cancer- in the combination therapy regimen? Does the combination of bearing mice is not yet clearly defined because conflicting data low-dose CY and PDT work equally well with other less vascular PS have been reported. In a colon carcinoma model, an expansion (compared with BPD) as the PDT agent? Does the combination of T-regs was observed only in the spleen but not in peripheral work with other poorly immunogenic mouse tumor models? When blood (36). In our results we found a small, but significant, further work on the role of T-regs in limiting the immune response increase in T-regs in tumor-bearing mice. Depletion of T-regs to human cancer and their role in the immune response after PDT together with other immunostimulatory approaches, for exam- has been carried out in patients, we believe serious consideration ple, CTLA-4 blockade, has also been tested (37). Because should be given to testing T-reg depletion therapies for patients who immune responses to malignant tumors often are weak and will receive PDT for cancer especially when it may have already ineffective, solely depleting T-regs might not always result in metastasized. tumor regression, therefore combining T-reg depletion with other immunologic interventions, for example, transfer of acti- Materials and Methods vated T cells (38) or dendritic cell-based vaccinations (39), have Cell Lines and Mouse Tumor Model. J774 cell line [murine BALB/c macrophage- been reported. Several molecules have been implicated in the like reticulum cell sarcoma (52)] was obtained from ATCC and cultured in RPMI suppressor mechanism underlying the immunoregulatory func- media 1640 containing Hepes, glutamine, 10% FCS, 100 units/ml penicillin, ϩ ϩ tion of CD4 FoxP3 T cells, including cell surface molecules and 100 ␮g/ml streptomycin. Cells were collected for injection by washing with such as CTLA-4 and glucocorticoid-induced tumor necrosis PBS without Ca2ϩ and Mg2ϩ and adding trypsin-EDTA to the plate for 5 min at factor receptor (40), and secreted molecules such as TGF-␤1 and 37°C. EMT-6 cells, a BALB/c breast sarcoma line (53), were used as an irrelevant IL-10 (41). target control for radiochromium release assays. All animal experiments were CY administration especially when carried out at low doses has approved by the Subcommittee on Research Animal Care of Massachusetts been suggested to have a specific effect in depleting T-regs General Hospital and were in compliance with National Institutes of Health guidelines. Male BALB/c mice (Charles River Labs), weighing 20–25 g, were (42–44) and has been used as an immunomodulator in several depilated (Nair; Carter-Wallace Inc.) on the right thigh, and the next day they studies with the intention to induce tumor regression and were injected s.c. with 106 J774 cells suspended in 100 ␮l of PBS. Tumors grew increase the concomitant immunity that can be suppressed by predictably in all mice and reached a size of 5- to 6-mm diameter in 8–10 days T-regs. In our case both low- and high-dose CY decreased the after injection at which time they were used for PDT. In some mice tumors number of total splenocytes and the proportion of T-regs. were allowed to progress for 4 weeks, and mice were killed for examination

5498 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0709256105 Castano et al. Downloaded by guest on October 4, 2021 of organs by standard H&E histopathology. Low doses (50 mg/kg) or high make single-cell suspensions. Erythrocytes were lysed with red blood cell lysis doses (150 mg/kg body weight) of CY (Sigma) in sterile PBS was injected i.p. 2 buffer (Roche Diagnostics). Cells were washed three times with full culture days before the PDT treatment or after 7 days tumor growth in CY-alone medium, and viable trypan blue-excluding cells were counted on a hemocy- groups. In the literature of low-dose CY studies in mice and rats there have tometer, allowing the total number of splenocytes isolated per mouse to be been several doses used at both low doses and high doses. For low doses from calculated. Single-cell suspensions of splenocytes were washed and resus- 10 to 100 mg/kg have been used, and for high doses from 150 to 300 mg/kg pended in stain buffer (BD Pharmingen). Splenocytes were plated in six-well have been used (54). Our selection of 50 and 150 mg/kg as low and high dose, plates in culture medium for1htoallow adherent cells to attach. The respectively, was based on these reports. Mice were anesthetized with an i.p. supernatant with lymphocytes (nonadherent splenocytes) was collected, and injection of ketamine/xylazine mixture (90 mg/kg ketamine, 10 mg/kg xyla- cells were counted and used for further assays. Cytotoxicity was measured by 51 zine) before any invasive procedure (PDT or surgery). Cr release assay. Nonadherent splenocyte suspension (0.1 ml) as effector cells were dispensed to wells of U-bottom 96-well microtiter plates, with replicates of six wells for each E/T ratio. One million target cells (J774 or EMT6 as Amputation. Ten mice with s.c. J774 tumors in the midthigh had their legs irrelevant target) was labeled for 2 h with 100-␮Ci of 51Cr (as sodium chromate; sequentially amputated above the injection site on successive days starting at PerkinElmer Life and Analytical Sciences) and washed with culture medium 1 day after injection until day 10. This procedure was done under anesthesia. and then 10,000 target cells were mixed with effector CTLs at various E/T ratios With a scalpel, the femoral artery was isolated, sutured, and cut with VICRIL and incubated for4hat37°C, 5% CO . The amount of 51Cr released into the 7–0 synthetic absorbable suture (Ethicon Inc.), and the bone was cut after 2 supernatant by killed target cells was quantitated in a gamma scintillation hemostasis. The wound was closed in two planes: first, muscle was closed with plate reader (Wallac). The final percentage of specific lysis was calculated as VICRIL 7–0 suture, and the second plane was skin closed with black monofil- follows: [test 51Cr released Ϫ spontaneous 51Cr released] / [maximum 51Cr ament Nylon nonabsorbable surgical suture (Ethicon, Inc.). All of these pro- released Ϫ spontaneous 51Cr released]. The maximal release was obtained by cedures were done following aseptic techniques. incubation of target cells in 0.5% SDS.

PDT. BPD was a generous gift from QLT Inc. (Vancouver). The lyophilized powder Flow Cytometry for T-Regs. On the third day after CY administration mice were was reconstituted with sterile 5% dextrose injection (Abbott Laboratories). Mice killed, and inguinal LNs and spleens were removed from naı¨ve mice and were anesthetized as above, and BPD at 2 mg/kg was injected in 0.1 ml in the tumor-bearing mice with and without CY. LN in 1.5-ml microcentrifuge tubes lateral tail vein with an insulin syringe (30-gauge needle). Tails were warmed in containing 0.5 ml of ice-cold PBS were homogenized with a pellet pestle motor water (50°C) to facilitate the procedure. PDT was carried out at 15 min after (Kontes Glass). Spleens were manually homogenized in a ground glass ho- injection of BPD. Illumination was carried out by using a 1-W solid state diode mogenizer in 5 ml of PBS. Homogenates were passed through a 70-␮m mesh laser emitting light at 690 nm (ϩ/Ϫ nm) (B&W Tek Inc.) for activation of BPD. The nylon cell strainer (BD Falcon) to make single-cell suspensions. Splenocytes laser was coupled into a 400-␮m fiber via a SMA connector, and light from the were counted by trypan blue exclusion on a hemocytometer after red blood distal end of the fiber was focused into a uniform spot with an objective lens (no. cell lysis. T-regs were stained with a mouse regulatory T cell staining kit (no. 774317; Olympus). The spot had a diameter of 1.2 cm and was positioned so that 88-811540; eBioscience). A total of 106 LN cells were resuspended in 1 ml of the entire tumor and a surrounding 2–3 mm of normal tissue were exposed to fixation/permeabilization solution, incubated on ice for 40 min, washed, light. A total fluence of 150 J/cm2 was delivered at a fluence rate of 100 mW/cm2 treated with Fc receptor blocking antibody, and incubated with FITC-labeled as measured with a power meter (model DMM 199 with 201 standard head; anti-CD4 and phycoerythrin-Cy5-labeled anti-FoxP3 antibodies for1honice. Coherent). At the completion of the illumination mice were allowed to recover in After washing, LN cells were analyzed by flow cytometry (FACScalibur; BD an animal warmer until they resumed their normal activity. Biosciences) together with the relevant isotype controls. Thirty thousand total events were counted. FACS scattergrams were analyzed by first gating for size ϩ Follow-Up of Mice. Mice were examined three times a week. They were and CD4 expression on FL1 vs. FSC dot plot and next by replotting CD4 cells weighed, and the orthogonal tumor dimensions (a and b) were measured on FL1 vs. FL3 dot plot to assess the percentage of CD4–FoxP3 double positive with vernier calipers. The tumor volume was calculated according to the cells that was multiplied by FL1 mean. formula, volume ϭ 4␲/3 [(a ϩ b)/4]3. Tumors were treated when they were Ϯ 6 mm in diameter (volume Ϸ110 mm3). Mice were killed when either the Statistics. All values are expressed as SD. Survival analysis was performed by primary tumor reached a diameter of 1.5 cm or they had lost 15% or more using the Kaplan–Meier method. Survival curves were compared, and differ- of their body weight. Mice were considered cured when they were healthy ences in survival were tested for significance by using a log rank test in the and tumor-free for 90 days and their body weight remained normal. Mice computer program GraphPad Prism. Differences between means of chromium that were cured were rechallenged with tumor (injection of 1 million J774 release and FACS analysis were tested for significance by single-factor ANOVA in Microsoft Excel. The tumor rejection frequency was analyzed by Fisher’s cells in the contralateral leg) together with naı¨vemice. Mice that rejected exact test. P Ͻ 0.05 was considered significant. the J774 rechallenge were subsequently challenged with 1 million EMT6 cells. ACKNOWLEDGMENTS. We thank John Demirs for help with histopathology and QLT Inc. for the generous gift of BPD. This work was funded by National CTL Chromium Release Assay. Whole spleens from tumor-immune mice killed Institutes of Health Grant R01-CA/AI838801 (to M.R.H.). A.P.C. was supported 2 days after another J774 rechallenge and from J774 tumor-bearing mice and by Department of Defense Congressionally Directed Medical Research Pro- naı¨vemice as controls were forced through coarse (70-mesh) steel screens to gram Breast Cancer Research Grant W81XWH-04-1-0676.

1. Antonia S, Mule JJ, Weber JS (2004) Current developments of immunotherapy in the 11. Castano AP, Demidova TN, Hamblin MR (2005) Mechanisms in photodynamic therapy: clinic. Curr Opin Immunol 16:130–136. Part two–cellular signaling, cell metabolism, and modes of cell death. Photodiagn 2. Emens LA, Jaffee EM (2003) Cancer vaccines: An old idea comes of age. Cancer Biol Ther Photodyn Ther 2:1–23. 2:S161–S168. 12. Dougherty TJ (2002) An update on photodynamic therapy applications. J Clin Laser 3. Frost P, Bonavida B (2000) Circumvention of tumor cell escape following speciﬁc Med Surg 20:3–7. immunotherapy. Cancer Biother Radiopharm 15:141–152. 13. Henderson BW, et al. (1985) Tumor destruction and kinetics of tumor cell death in 4. Gupta S, Kanodia AK (2002) Biological response modiﬁers in cancer therapy. Natl Med two experimental mouse tumors following photodynamic therapy. Cancer Res J India 15:202–207. 45:572–576. 5. Parton M, Gore M, Eisen T (2006) Role of cytokine therapy in 2006 and beyond for 14. Abels C (2004) Targeting of the vascular system of solid tumors by photodynamic metastatic renal cell cancer. J Clin Oncol 24:5584–5592. therapy (PDT). Photochem Photobiol Sci 3:765–771. 6. Durrant LG, Ramage JM (2005) Development of cancer vaccines to activate cytotoxic T 15. Korbelik M (1996) Induction of tumor immunity by photodynamic therapy. J Clin Laser

lymphocytes. Exp Opin Biol Ther 5:555–563. Med Surg 14:329–334. MEDICAL SCIENCES 7. Schirrmacher V (2005) T cell-mediated immunotherapy of metastases: State of the art 16. Castano AP, Mroz P, Hamblin MR (2006) Photodynamic therapy and antitumor immu- in 2005. Exp Opin Biol Ther 5:1051–1068. nity. Nat Rev Cancer 6:535–545. 8. Qian C, Prieto J (2004) Gene therapy of cancer: Induction of antitumor immunity. Cell 17. Gollnick SO, et al. (2003) Role of cytokines in photodynamic therapy-induced local and Mol Immunol 1:105–111. systemic inﬂammation. Br J Cancer 88:1772–1779. 9. Castano AP, Demidova TN, Hamblin MR (2004) Mechanisms in photodynamic therapy: Part 18. Cecic I, Parkins CS, Korbelik M (2001) Induction of systemic neutrophil response in mice by one–photosensitizers, photochemistry and cellular localization. Photodiagn Photodyn Ther photodynamic therapy of solid tumors. Photochem Photobiol 74:712–720. 1:279–293. 19. Korbelik M, et al. (1996) The role of host lymphoid populations in the response of 10. Castano AP, Demidova TN, Hamblin MR (2005) Mechanisms in photodynamic therapy: mouse EMT6 tumor to photodynamic therapy. Cancer Res 56:5647–5652. Part three-photosensitizer pharmacokinetics, biodistribution, tumor localization, and 20. Korbelik M, Dougherty GJ (1999) Photodynamic therapy-mediated immune response modes of tumor destruction. Photodiagn Photodyn Ther 2:91–106. against subcutaneous mouse tumors. Cancer Res 59:1941–1946.

Castano et al. PNAS ͉ April 8, 2008 ͉ vol. 105 ͉ no. 14 ͉ 5499 Downloaded by guest on October 4, 2021 21. Ray PK, Raychaudhuri S (1981) Low-dose cyclophosphamide inhibition of transplant- 39. Prasad SJ, et al. (2005) Dendritic cells loaded with stressed tumor cells elicit long-lasting able fibrosarcoma growth by augmentation of the host immune response. J Natl protective tumor immunity in mice depleted of CD4ϩCD25ϩ regulatory T cells. JIm- Cancer Inst 67:1341–1345. munol 174:90–98. 22. North RJ (1982) Cyclophosphamide-facilitated adoptive immunotherapy of an established 40. Shimizu J, et al. (2002) Stimulation of CD25(ϩ)CD4(ϩ) regulatory T cells through GITR tumor depends on elimination of tumor-induced suppressor T cells. J Exp Med 155:1063–1074. breaks immunological self-tolerance. Nat Immunol 3:135–142. 23. Matar P, et al. (2000) Mechanism of antimetastatic immunopotentiation by low-dose 41. Fontenot JD, Rudensky AY (2004) Molecular aspects of regulatory T cell development. cyclophosphamide. Eur J Cancer 36:1060–1066. Semin Immunol 16:73–80. 24. Zhang Q, Kang X, Zhao W (2006) Antiangiogenic effect of low-dose cyclophosphamide 42. Awwad M, North RJ (1988) Cyclophosphamide (Cy)-facilitated adoptive immunother- combined with ginsenoside Rg3 on Lewis lung carcinoma. Biochem Biophys Res apy of a Cy-resistant tumor. Evidence that Cy permits the expression of adoptive T Commun 342:824–828. cellmediated immunity by removing suppressor T cells rather than by reducing tumor 25. Yamaguchi T, Sakaguchi S (2006) Regulatory T cells in immune surveillance and burden. Immunology 65:87–92. treatment of cancer. Semin Cancer Biol 16:115–123. 43. Awwad M, North RJ (1989) Cyclophosphamide-induced immunologically mediated 26. Beyer M, Schultze JL (2006) Regulatory T cells in cancer. Blood 108:804–811. 27. Lutsiak ME, et al. (2005) Inhibition of CD4(ϩ)25ϩ T regulatory cell function implicated regression of a cyclophosphamide-resistant murine tumor: A consequence of elimi- ϩ in enhanced immune response by low-dose cyclophosphamide. Blood 105:2862–2868. nating precursor L3T4 suppressor T cells. Cancer Res 49:1649–1654. ϩ ϩ 28. Ziegler SF (2006) FOXP3: Of mice and men. Annu Rev Immunol 24:209–226. 44. Ghiringhelli F, et al. (2004) CD4 CD25 regulatory T cells suppress tumor immunity 29. Chen B, et al. (2005) Combining vascular and cellular targeting regimens enhances the but are sensitive to cyclophosphamide which allows immunotherapy of established efficacy of photodynamic therapy. Int J Radiat Oncol Biol Phys 61:1216–1226. tumors to be curative. EurJ Immunol 34:336–344. 30. Gollnick SO, Owczarczak B, Maier P (2006) Photodynamic therapy and anttumor 45. Matar P, et al. (2002) Th2/Th1 switch induced by a single low dose of cyclophosphamide immunity. Lasers Surg Med 38:509–515. in a rat metastatic lymphoma model. Cancer Immunol Immunother 50:588–596. 31. Fedoseyeva EV, et al. (2000) CD4ϩ T cell responses to self- and mutated p53 determi- 46. Ercolini AM, et al. (2005) Recruitment of latent pools of high-avidity CD8(ϩ) T cells to nants during tumorigenesis in mice. J Immunol 164:5641–5651. the antitumor immune response. J Exp Med 201:1591–1602. 32. Shevach EM, et al. (2003) Control of T cell activation by CD4ϩCD25ϩ suppressor T cells. 47. Loeffler M, Kruger JA, Reisfeld RA (2005) Immunostimulatory effects of low-dose Novartis Found Symp 252:24–36; discussion 36–44, 106–114. cyclophosphamide are controlled by inducible nitric oxide synthase. Cancer Res 33. Seshadri M, et al. (2005) Tumor vascular response to photodynamic therapy and the 65:5027–5030. antivascular agent 5,6-dimethylxanthenone-4-acetic acid: Implications for combina- 48. Chen WR, et al. (2005) Enhancement of laser cancer treatment by a chitosan-derived tion therapy. Clin Cancer Res 11:4241–4250. immunoadjuvant. Photochem Photobiol 81:190–195. 34. Turk MJ, et al. (2004) Concomitant tumor immunity to a poorly immunogenic mela- 49. Chen WR, et al. (2002) Effect of different components of laser immunotherapy in noma is prevented by regulatory T cells. J Exp Med 200:771–782. treatment of metastatic tumors in rats. Cancer Res 62:4295–4299. 35. Yu P, et al. (2005) Intratumor depletion of CD4ϩ cells unmasks tumor immunogenicity 50. Chen WR, et al. (2001) Antitumor immunity induced by laser immunotherapy and its leading to the rejection of late-stage tumors. J Exp Med 201:779–791. adoptive transfer. Cancer Res 61:459–461. 36. Liu JY, et al. (2005) The changes of CD4ϩCD25ϩ/CD4ϩ proportion in spleen of 51. Naylor MF, et al. (2006) In situ photoimmunotherapy: A tumor-directed treatment for tumor-bearing BALB/c mice. J Transl Med, 10.1186/1479-5876-3-5. 37. Sutmuller RP, et al. (2001) Synergism of cytotoxic T lymphocyte-associated antigen 4 melanoma. Br J Dermatol 155:1287–1292. blockade and depletion of CD25(ϩ) regulatory T cells in antitumor therapy reveals 52. Ralph P, Prichard J, Cohn M (1975) Reticulum cell sarcoma: An effector cell in antibody- alternative pathways for suppression of autoreactive cytotoxic T lymphocyte re- dependent cell-mediated immunity. J Immunol 114:898–905. sponses. J Exp Med 194:823–832. 53. Siemann DW, et al. (1981) Cell subpopulations dispersed from solid tumors and 38. Tanaka H, et al. (2002) Depletion of CD4ϩ CD25ϩ regulatory cells augments the separated by centrifugal elutriation. Br J Cancer 44:100–108. generation of specific immune T cells in tumor-draining lymph nodes. J Immunother 54. Motoyoshi Y, et al. (2006) Different mechanisms for antitumor effects of low- and 25:207–217. high-dose cyclophosphamide. Oncol Rep 16:141–146.

5500 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0709256105 Castano et al. Downloaded by guest on October 4, 2021