Fractional Laser Releases Tumor-Associated Antigens in Poorly Immunogenic Tumor and Induces Systemic Immunity

Currently ablative fractional photothermolysis (aFP) with CO2 laser is used for a wide variety of dermatological indications. This study presents and discusses the utility of aFP for treating oncological indications. We used a fractional CO2 laser and anti-PD-1 inhibitor to treat a tumor established unilaterally by the CT26 wild type (CT26WT) colon carcinoma cell line. Inoculated tumors grew significantly slower in aFP-treated groups (aFP and aFP + anti-PD-1 groups) and complete remission was observed in the aFP-treated groups. Flow cytometric analysis showed aFP treatment elicited an increase of CD3+, CD4+, CD8+ vand epitope specific CD8+ T cells. Moreover, the ratio of CD8+ T cells to Treg increased in the aFP-treated groups. Additionally, we established a bilateral CT26WT-inoculated mouse model, treating tumors on one-side and observing both tumors. Interestingly, tumors grew significantly slower in the aFP + anti-PD-1 groups and complete remission was observed for tumors on both aFP-treated and untreated sides. This study has demonstrated a potential role of aFP treatments in oncology.

Fractional photothermolysis (FP) can be characterized as laser-assisted treatment generating a pattern of microscopic treatment zones (MTZs) in biological tissue 1 . There are two FP modes, the non-ablative mode (nFP) and ablative (aFP) mode. nFP generates MTZs as small zones of thermally damaged tissue, whereas aFP additionally produces a central "hole" of physically-removed (ablated) tissue, surrounded by a small cuff of thermally-damaged tissue in MTZs 2,3 . In general, the width or diameter of the MTZs measures less than 0.5 mm. FP techniques typically expose only a small fraction of the tissue (often an areal fraction of approximately 5-27%), leaving the majority of tissue spared or unexposed 4 . A large range of dermatological indications, such as treatment of photodamaged skin, dyschromia, rhytides, and different kind of scars including acne, surgical and burn scars currently make use of FP 1,[5][6][7][8] . However, FP is not used for the treatment of tumors and no studies to date have investigated production of systemic effects using FP methods.
Mroz et al. reported laser irradiation of photodynamic therapy (PDT) to treat mice subcutaneously inoculated with CT26.CL25 colon carcinoma cells inducing local remission and antigen-specific immune response systemically, promoting regression of a remote and untreated tumor 9 . However CT26.CL25 cells are artificially transduced with the lacZ gene to stably express a tumor antigen (beta-gal). One limitation of the study is that the tumor model used is not a clinically relevant tumor, and instead represents an artificially-induced cancerous state in the subject mice. The phenomenon promoting regression of a remote and untreated tumor was not observed when they employed CT26 wild type (CT26WT) undifferentiated colon carcinoma cells, which are beta-gal negative parental carcinoma of CT26.CL25 cells. This is surprising because many reports showed PDT could induce the immune response against locally inoculated CT26WT colon carcinoma cells [10][11][12] . These carcinoma cells are a clone of the N-nitoroso-N-methylurethan (NMU)-induced grade IV carcinoma, which is rapidly-growing and readily-metastasizing 13 . Therefore we decided to challenge and treat these poorly immunogenic CT26WT colon carcinoma cells with aFP and anti-PD-1 inhibitor, which boosts the function of CD8+ T cells by blocking the PD-1/ PD-L1 pathway [14][15][16] , to confirm if aFP can be used for induction of anti-tumor immunity and promote Results aFP is significantly effective against local primary tumors. aFP laser irradiation was performed 8 days after tumor inoculation. The aFP significantly led to a reduction in tumor volume and growth rate of CT26WT tumors after the treatment (Fig. 1a). Tumors shrank in 6 mice of 11 mice in the aFP group and 8 mice of 11 mice in the aFP + anti-PD-1 group, but such shrinkage occurred in only 1 mouse (9%) from the anti-PD-1 group or does not occur in the control group (Fig. 1b). The significance value for the difference between the survival curves are: control vs. anti-PD-1 (p < 0.05), control vs. aFP (p < 0.0001), control vs. aFP + anti-PD-1 (p < 0.0001), anti-PD-1 vs. aFP (p < 0.005) and anti-PD-1 vs. aFP + anti-PD-1 (p < 0.0005). aFP induces development of long term anti-tumor immunity. To assess the presence of any induced long term anti-tumor immunity, we performed a rechallenge experiment following the treatment. Six mice in the aFP group and eight mice in the aFP + anti-PD-1 group which survived for more than 90 days after the treatment, were subsequently inoculated subcutaneously with CT26WT cells in the contralateral (right) thigh. Eight age-matched naive mice were inoculated with the same number of CT26WT cells, respectively, in the right thigh as a control. The tumor on the naive mice in the control groups and one of the six mice in the aFP group progressed over time. Tumors on the rest of the survival mice in the aFP group and all survival mice in the aFP + anti-PD-1 group did not appear to progress (Fig. 1c). They remained tumor-free for at least another 60 days following the inoculation (Fig. 1c). The significance values for the differences between the survival curves are: control vs. aFP (p < 0.001), control vs. aFP + anti-PD-1 (p < 0.0001). aFP increases CD3+, CD4+ and CD8+ T cell numbers and develop epitope specific CD8+ lymphocytes in aFP-treated tumor. To investigate if the adaptive immune system was affected by aFP, we measured the number of CD3roup and all survival mice in, CD4roup and all survival mice in, and CD8+ T cells and Foxp3+ Tregs inside the tumor using flow cytometry 5 days after aFP treatment. We chose to measure on 5 days after aFP treatment because by this day the aFP-treated tumor in the aFP and aFP + anti-PD-1 groups showed significantly shrinkage. We found that CD3+, CD4+, CD8+ T cells increased per tumor weight (c) Kaplan-Meier survival curves of mice receiving the rechallenge test with CT26WT cells. These mice were each inoculated with 3.5 × 10 5 CT26WT cells subcutaneously in the contralateral (right) thigh against previous inoculation. The significance values for the difference between the survival curves are: control mice vs. survival mice (p < 0.005), control vs. aFP + anti-PD1: P < 0.0001. *The bars represent SD. P < 0.05, **P < 0.005, ***P < 0.0005, ****P < 0.0001.

Adaptive immunity is necessary to eradicate cancer cells after aFP. To investigate whether adaptive
immunity is necessary to eradicate cancer cells after aFP, we repeated the experiment with anti-CD8 depletion antibody to ablate CD8+ T cells in the mouse. As shown in Fig. 3, treatment of aFP in absence of CD8+ T cells failed to prevent tumor growth and eventual euthanasia, while 4 out of 6 mice in control group, with no CD8 depletion, survived. (P < 0.005; Fig. 3). This observation indicates that adaptive immunity is necessary to eradicate cancer cells after aFP.

aFP + anti-PD-1 inhibitor combination therapy induces development of systemic anti-tumor immunity.
To investigate whether aFP treatment could induce systemic anti-tumor immunity, we established a two tumor mouse model, which has a tumor on both hind legs, with the aFP-treated tumor on the left leg and observed the growth of both tumors. aFP laser irradiation was performed 7 days after tumor inoculation. The aFP significantly led to a reduction in tumor volume and growth rate of aFP-treated tumors after the treatment ( Fig. 4a and c). Regarding the untreated contralateral tumors, we found that the aFP + anti-PD-1 combination therapy led to a significant reduction in growth rate after the treatment ( Fig. 4b and c). Moreover the untreated contralateral tumors shrank in 2 of the 6 mice in the aFP + anti-PD-1 group completely, but such shrinkage did not occur in the rest of the groups ( Fig. 4c and d). The significance value for the difference between the survival curves are: control vs. aFP + anti-PD-1 (p < 0.005), anti-PD-1 vs. aFP + anti-PD-1 (p < 0.005), aFP vs. aFP + anti-PD-1 (p < 0.01). aFP induces recruitment of CD3+, CD4+, CD8+ and epitope specific CD8+ T lymphocytes but increases infiltration of Treg into the untreated contralateral tumors in the two tumor mouse model. To investigate why the untreated contralateral tumor in the aFP group did not shrink even though epitope specific CD8+ T lymphocytes were developed in the aFP-treated tumor, we measured the number of CD3+, CD4+, CD8+, epitope specific CD8+ T cells, and Foxp3+ Tregs inside the untreated contralateral tumor using flow cytometry 12 days after aFP treatment. We chose to measure on 12 days after aFP treatment because on this day the untreated contralateral tumor in the aFP + anti-PD-1 group shrunk significantly (aFP-treated tumors in the two groups shrank within 5 days after aFP treatment). We found that CD4+, CD8+ and epitope specific CD8+ T cells increased per tumor weight in the aFP-treated groups (CD3/weight and CD8/weight: control vs. aFP, control vs. aFP + anti-PD-1 and anti-PD-1 vs. aFP + anti-PD-1 (p < 0.05); CD4/weight and epitope specific CD8+ T cells: control vs. aFP, control vs. aFP + anti-PD-1, anti-PD-1 vs. aFP and anti-PD-1 vs. aFP + anti-PD-1 (p < 0.05; Fig. 5a and c). Moreover the percentage of AH1 specific CD8+ T cells compared to total CD8+ T cells increased in the aFP-treated groups (control vs. aFP and control vs. aFP + anti-PD-1 (p < 0.05; Fig. 5c). However, we found that Tregs increased in the aFP-treated groups per tumor weight though there was no significant difference regarding Treg numbers in aFP-treated tumors in the all groups (Fig. 5a). Furthermore, there is no significant difference in the ratio of CD8+ T cells to Treg in the all groups (Fig. 5b).
Additionally to investigate the up-regulation of PD-L1 on tumor cells, which is mediating immune suppression to escape immune surveillance, we investigated the PD-L1 expression on the tumor cell using flow cytometry 5 days and 12 days after aFP treatment. We found that there were high percentage of PD-L1 positive cells in the tumor in the all groups 12 days after aFP treatment even though lower percentage of CT26WT cells were PD-L1 positive before inoculation (Supplementary Figure 1a). The percentage of PD-L1 positive cells increased as the tumor grew (Supplementary Figure 1b). These data suggest that aFP induces infiltration of epitope specific CD8+ T cells into the untreated contralateral tumor; however, the function of the cells was impaired by high percentage of PD-L1 positive cells, which increased as tumors grew. Therefore, unlike the aFP-treated tumor, the untreated contralateral tumor in the aFP group did not shrink.

Antigen specificity is necessary for systemic anti-tumor immunity.
To investigate whether the shrinkage of the untreated contralateral tumor in the aFP + anti-PD-1 group was due to antigen specific immunity induced by aFP or not, we performed experiments with two groups of mice that each had two mismatched tumors, CT26WT in left leg and 4T1 (murine mammary carcinoma) in right leg. One group was treated with only anti-PD-1 inhibitor and the other group was treated with aFP treatment in the left leg (CT26WT) and anti-PD-1 inhibitor. There were no effects on the size or growth rate of the contralateral untreated tumors and survival mice in either group (Fig. 6).

Discussion
This study demonstrates significant tumor size reduction and initiation of a systemic, tumor specific immune response induced by FP laser irradiation against a poorly immunogenic tumor 13 . A large range of clinical indications currently make use of FP, yet the ability of FP to induce tumor regression is unexpected, because FP treatments are generally applied to achieve localized tissue effects within the treatment area such as in the stimulation of wound healing and tissue regeneration.
There are reported studies that utilize tumor associated antigen (TAA) properties of the tumor to show anti-tumor activity of cancer therapies 17 . However, the antigen expression on the cell in a normal state is not sufficient to stimulate an immune response to prevent tumor growth 9 . Therefore, in the present study, we employed aFP because we hypothesized it would produce thermal denaturation, potentially releasing TAA and enhancing anti-tumor immunity in order for the cytotoxic T lymphocytes to recognize and bind the immunodominant peptide epitope derived from the antigens on MHC class I as it simultaneously creates MTZ with intact interspersed tissue. The total MTZ/tumor volume ratio was adjusted to approximately 5% of inoculated tumors, thereby aiming to enhance inflammation while avoiding the direct killing of the tumor bulk. Mroz et al. successfully reported laser irradiation with PDT to treat mouse subcutaenous immunogenic CT26.CL25 colon carcinoma cells, which was artificially transduced with lacZ gene to express an immunodominant antigen (beta-gal) induced antigen-specific immune response 9 . Moreover, PDT induces the immune response even in poorly immunogenic CT26WT colon carcinoma cells, which are beta-gal negative parental carcinoma of CT26. CL25 cells [10][11][12] . Therefore, we were motivated to investigate whether aFP could induce anti-tumor immunity against clinically relevant mouse tumor models and found that aFP treatment against the poorly immunogenic tumors could induce anti-tumor immunity and produce improvement in long-term survival. The CT26WT colon carcinoma cells we employed in this study have TAA in a single peptide known as AH-1, a non-mutated nonamer derived from the envelope protein (gp70) of an endogenous ecotropic murine leukemia provirus 18 . Mroz et al. showed that PDT induced local remission of CT26.CL25 tumors as well as a systemic tumor-specific immune response derived from release of TAA (beta-gal) from the tumor by PDT inducing epitope specific CD8+ T cells 9 . In our present study the pentamer staining revealed increase of TAA (AH1) specific CD8+ T cells per tumor weight in the aFP-treated groups. The result suggests aFP can also release TAA and induce anti-tumor immunity. The effectiveness of aFP is associated with thermal protein denaturation 1 . Reports have shown that thermal denaturation induces gene expression 19,20 and up-regulates chemokines 21 . Furthermore, it leads to release aFP + anti-PD1: P < 0.005, aFP vs. aFP + anti-PD1: P < 0.01. The bars represent SD. *P < 0.05, **P < 0.01, ***P < 0.005. of damage-associated molecular patterns (DAMPs) 22 and induces innate immunity 23,24 . These facts suggest thermal denaturation induced by aFP initiated a strong immune response locally, resulting in induction of epitope specific CD8+ T cell response.
Interestingly, even though epitope specific CD8+ T cells infiltrated into both sides of the tumors in the aFP group in the two tumor mouse model, there were no surviving mice in the group. We assume the reason for this is the abundant presence of PD-L1+ tumor cells that can block systemic CD8+ T cell immunity mounted by aFP treatment in the untreated contralateral tumor cells 12 days after aFP treatment since it is indicated by the survival benefit when using anti-PD-1 inhibitor in the aFP + anti-PD-1 group.
In addition to blocking the PD-1/PD-L1 axis, PD1 inhibitor boosts clonal expansion of antigen specific CD8+ T cells 25 . Therefore, there were tendencies that AH1 epitope specific CD8+ T cells compared to total CD8+ T  cells increased more in the aFP + anti-PD-1 group than in the aFP group. Even though the PD-1/PD-L1 axis was blocked by anti-PD-1 inhibitor and the high frequency of the epitope specific CD8+ T cells was observed, complete remission of the untreated contralateral tumor was observed in fewer cases than the aFP-treated tumors in the aFP + anti-PD-1 group in the two tumor mouse model. We assume the reason is the ratio values of CD8+ T cells compared to Treg in the untreated contralateral tumors are lower than in the aFP-treated tumors.
In conclusion, we used aFP to induce anti-tumor immunity. Despite the small degree of thermal injury to the tumor, aFP induced a long-lasting anti-tumor immunity. Moreover, combining aFP therapy with an anti-PD-1 inhibitor further boosts the systemic immunity. These effects might be mediated by an increase in the number of tumor antigen-specific cytotoxic T cells, induced by aFP treatment. Thus, such a therapeutic strategy might achieve significant efficacy in poorly immunogenic cancers of colon, breast, lung and other organs.
Animals. Six-week-old female BALB/c mice (Charles River Laboratories, Boston, MA) were used for the study.
The care and handling of the animals was done in accordance with a protocol approved by the Subcommittee on Research Animal Care (IACUC) at Massachusetts General Hospital (MGH).
Animal tumor model. Mice were inoculated unilaterally at the left leg or bilaterally with 3.5 × 10 5 CT26WT subcutaneously into the depilated thigh, after being anesthetized through intraperitoneal injection of a cocktail of ketamine (90 mg/kg) and xylazine (10 mg/kg). Anti-PD-1 blocking antibodies (29 F.1A12; BioXCell,West Lebanon, NH) were administered intraperitoneally at a dose of 200 µg per mouse on days 8, 10, 12, 14, and 16 in the one tumor mouse model, and days 7, 9, 11, 13, and 15 in the two tumor mouse model after tumor cell inoculation (Fig. 7). Anti-CD8 depletion antibodies (2.43; BioXCell,West Lebanon, NH) were administered intraperitoneally at a dose of 200 µg per mouse every 3 days from one day before tumor inoculation to removal of mice as endpoint. The tumor volume was determined 3 times per week by measuring the longest dimension and orthogonal dimension of the tumor with vernier calipers. Tumor volumes were calculated according to the formula volume = 4π/3 × [(a + b)/4] 3 , where a and b represent the long and short axis lengths, respectively. If the tumor volume exceeded 700 mm 3 or showed severe ulceration, the mouse was removed from the study as endpoint.
Fractional CO 2 laser irradiation. Ablative fractional laser exposure was performed on the tumor site at day 7 or 8 after tumor cell inoculation (Fig. 7). At that time, the tumors have reached a typical diameter of about 5-6 mm. Exposures were performed with an Ultrapulse Encore CO 2 laser (Lumenis Inc, Yokneam, Israel). A single aFP treatment was performed within a treatment area of 6 × 6 mm, with a pulse energy of 100 mJ, at a nominal density of 5% (i.e., 5% within the treatment area surface was irradiated), and a pulse repetition frequency of 300 Hz. No skin cooling was applied and the anesthesia was performed by intraperitoneal injection of a cocktail of ketamine (90 mg/kg) and xylazine (10 mg/kg).
Rechallenge. Mice surviving 90 days after tumor inoculation were rechallenged with matched tumor cells (3.5 × 10 5 CT26WT) subcutaneously at the contralateral (right) leg from previously aFP-treated side. Age-matched naive mice were inoculated with the same number of the cells in the right leg as the controls. The inoculated mice were monitored for another 60 days to confirm tumorigenesis.

Flow cytometry analysis.
To confirm whether CD3+ and CD8+ lymphocyte, and regulatory T cell (Treg) numbers were affected by aFP, and the PD-L1 expressed on the tumor cells, flow cytometry analysis was performed. For leukocyte isolation from tumor, fresh CT26WT tumors were dissociated mechanically filtering through 70 μm strainer on 6 well culture plate with DNaseI (10 μg/ml, Roche; Nutley, NJ) and collagenase (10 mg/ml, Life technologies) and incubated 60 minutes at 37 °C. The dissociated cells were positively and negatively selected for CD45 surface expression using CD45 MicroBeads and magnetic-activated cell sorting (MACS) system (Miltenyi Biotec Inc., Auburn, CA). Then the positively selected cells were stained with antibodies CD3-FITC (145-2C11), CD4-APC-eFluor780 (RM4-5), CD8a-PE/CY7 (53-6.7) for 30 minutes at 4 °C after they were blocked at 4 °C for 15 minutes using CD16/CD32 antibody (93) and negatively selected cells were stained PD-L1-PerCPeFluor 710 (MIH5) and CD45-PE (30-F11; all from eBioscience; Santa Clare, CA) for 30 minutes at 4 °C. Following staining for surface markers, the negatively selected cells were fixed in 2% paraformaldehyde at room temperature for 20 minutes and the positively selected cells were fixed and permeabilized by Foxp3/ Transcription Factor Staining Buffer Set (eBioscience) according to the manufacturer's instructions at room temperature for 30 minutes. After the permeabilization the cells were stained with Foxp3-APC antibody (FJK-16s; eBioscience; Santa Clare, CA) at 4 °C over night. The next day the stained cells were analyzed on FACSCanto (BD Bioscience, Franklin Lakes, NJ).

Statistics.
All experiments were repeated at least once. All statics analyses were performed with GraphPad Prism 7.0 (GraphPad Software). All values are expressed as the mean ± SD. Flow cytometric results were compared with a two-tailed Mann-Whitney test. Tumor growth curves were compared with a two-tailed Wilcoxon rank-sum test. Survival analysis was performed using the Kaplan-Meier method and a log-rank test. Values of P < 0.05 were considered statistically significant.