Routes of Delivery for CpG and Anti-CD137 for the Treatment of Orthotopic Kidney Tumors in Mice

We have found previously that the tumor cell lines, Renca (a renal cancer) and MC38 (a colon tumor) which had been injected subcutaneously in mice, could be successfully treated with a combination therapy of an oligodeoxynucleotide (CpG1826) (injected intratumorally) and anti-CD137 antibody (injected intraperitoneally). Thus the combination treatment was expected to initiate a “danger” signal via TLR9 on immune cells, and the anti-CD137 was expected to further activate T cells. In the present study, we found that several other tumor types injected subcutaneously could also be successfully treated with this combination therapy. In addition, we wished to determine if the treatment could work as effectively in an orthotopic metastatic model, which is more physiologically relevant to cancer in humans. Renca was selected as we were familiar with injecting this orthotopically into the outer cortex of the kidney in mice, and it spontaneously metastasizes to lung and abdominal sites. We tested various routes of delivery of CpG combined with intraperitoneal delivery of anti-CD137. Orthotopic tumors were injected with CpG intratumorally, using ultrasound-guided delivery on multiple occasions, combined with anti-CD137 intraperitoneally. A reduction in primary tumor size was observed following intratumoral injection of CpG compared to other treatments. We found that there was a statistically significant increase in survival of mice with orthotopic Renca tumor following intratumoral injection of CpG. However, we determined that the most effective route of delivery of CpG was intravenous, which led to further significantly enhanced survival of mice when combined with anti-CD137 intraperitoneally, likely due to inhibition of metastatic disease. Our data supports future development of this combination therapy for cancer.


Introduction
CpGs are unmethylated oligodeoxynucleotides which mimic bacterial DNA. They bind to Toll-like receptor 9 (TLR9), which is expressed on murine monocytes, macrophages, plasmacytoid dendritic cells (DCs) and B cells. Ligation of TLR9 by CpG can initiate an inflammatory or ''danger'' signal from these cells, enabling immune cell recruitment and antigen presentation (reviewed in [1] and [2]). Anti-CD137 (aCD137) is an agonist antibody specific for CD137 expressed on activated T cells (more so on CD8+ than CD4+ T cells), DCs, natural killer cells (NKs), granulocytes and endothelial cells at inflammation sites. Ligation of CD137 on T cells leads to enhanced T cell proliferation, T cell cytotoxicity, prevents activation-induced cell death and prevents immune tolerance (reviewed in [3] and [4]). In a previous study we found that combined treatment of CpG1826 and aCD137 promoted tumor regression and increased survival in subcutaneous (SC) tumor models using MC38 (colon carcinoma) and Renca (renal cell carcinoma) in mice [5].
In the present study we expanded the testing of this therapy to further SC models: B16F10 (melanoma) and CT26 (colon carcinoma), finding CpG alone or in combination with aCD137 increased therapeutic efficacy. As it is increasingly clear in the literature [6][7][8][9] that orthotopic models give a more physiologically relevant indication of the efficacy of treatments both in mice and humans, we elected Renca to be injected orthotopically into the kidney of mice and treated with combined CpG1826 and aCD137. This produces not only primary renal tumors, but also metastases to other organs such as the lung and liver, which play a major role in disease progression and mortality. This model recapitulates sites of metastasis in human renal carcinoma [10]. Previously we found that orthotopic Renca tumor was minimally responsive to treatment with a combination of three antibodies (anti-DR5, anti-CD40 and anti-CD137) [11] injected intraperitoneally, and therefore we wished to determine if a different immunotherapy (combining CpG and anti-CD137) could have a better impact on metastatic orthotopic kidney cancer.
Initially, we tested SC injections (into the skin) of CpG1826 combined with IP aCD137. Subsequently, we used ultrasoundguided injection of CpG directly IT into kidney tumors. Ultrasound-guided injection of other therapies has been used for other tumors in mice, including those of bladder [12], prostate [13] and liver [14], but not to our knowledge has it been used to treat kidney tumors in mice. We also tested CpG injected intravenously combined with IP aCD137, and found this gave the greatest efficacy of the treatment against metastatic orthotopic kidney cancer.

Mouse tumor models
Ethics statement: This study was carried out in strict accordance with the recommendations of the Victorian Bureau of Animal Welfare, Department of Primary Industries, and the National Health and Medical Research Council's Australian code of practice for the care and use of animals for scientific purposes. The protocol was approved by the Peter MacCallum Cancer Centre Animal Experimentation Ethics Committee under Permit numbers E1274 and E1328. All efforts were made to minimize suffering. The individual shown in Figure S1 in this manuscript is the operator of the ultrasound equipment and has given written informed consent (as outlined in PLOS consent form) to appear in this figure.
Wild type female BALB/c and C57BL/6 mice were purchased from the Walter and Eliza Hall Institute of Medical Research (Bundoora, Australia). At 8-12 weeks of age, mice were injected subcutaneously (SC) with either 5610 5 B16F10, 1610 6 MC38, 5610 5 Renca or 1610 6 CT26 cells in 100 ml saline. When tumors averaged 48 mm 2 (for B16F10), 33-35 mm 2 (for MC38), 38 mm 2 (for Renca) and 20-33 mm 2 (for CT26) in size (day 7 for B16F10, day 10 for MC38, day 9 for Renca and day 6-7 for CT26), mice were randomized and assigned into groups of comparable ranges of tumor sizes, with 5-9 mice/group, and treated as described below. Tumors were measured (longest and shortest dimensions) every 2-3 days and tumor area calculated. Mice were culled when tumors reached 200 mm 2 in size or at the first signs of stress.
For intrakidney tumor experiments, BALB/c female mice were injected with 1610 5 Renca cells in 10 ml PBS intrakidney (IK) into the outer cortex of the kidney. On day 12 after tumor injection, mice were randomly assigned into 3 groups: not treated, treated with 10 ml PBS IT plus aCD137 antibody intraperitoneally (IP) or treated with CpG1826 (CpG) IT plus aCD137 antibody IP, as detailed below. As negative controls, some groups received isotype control rat IgG2a (MAC4, produced in-house from European Collection of Cell Cultures hybridoma) instead of anti-CD137 antibody, and/or saline instead of CpG.

Treatment
Treatment consisted of injections with CpG (CpG1826 = 59-TCCATGACGTTCCTGACGTT-3 from Coley Pharmaceuticals, Canada) at the dose of 50 mg/50 ml for IT injection of SC tumors, 50 mg/10 ml for IT injection of IK tumors, or 50 mg/200 ml for IV injection, starting on the first day of treatment specified above and then repeated every 2-3 days for a total of 2-4 injections. aCD137 antibody (in-house generated using caprylic acid precipitation) was injected IP, starting on the first day of treatment, at the dose of 100 mg/200 ml. aCD137 injections were repeated every 3-4 days for a total of 3-6 injections. Isotype control antibody (MAC4) was given at the same dose and rate as aCD137 to appropriate control groups of mice.
The method of injecting IK tumors intratumorally is described as follows. Imaging of the kidney tumor was performed using a 40 MHz single-element mechanical transducer with focal length 6 mm; axial resolution 40 mm; lateral resolution 80 mm (Vevo 770 high-resolution small-animal ultrasound system, Visualsonics, Toronto, Ontario, Canada). The fur of the mice was locally shaved over the kidney area then further removed using a depilatory cream (Veet, Reckitt Benckiser Group, Slough, Berkshire, England). Mice were then injected with ketamine hydrochloride (80 mg/kg, Ketalar, Hospira Australia, Mulgrave, VIC, Australia) plus xylazine hydrochloride (10 mg/kg, Ilium xylazil-20, Troy Laboratories, Smithfield, NSW, Australia) and once anesthetized, placed on a temperature-controlled stage of the integrated rail system (Visualsonics). This setting allowed adjustment of the mouse's positioning while maintaining the 30 gauge injection needle mounted on a Hamilton syringe (Hamilton Company, Reno, Nevada, USA) within the transducer-imaging plane. The therapeutic was then delivered to the tumor by moving the tumor into the scan plane using the XYZ controls on the mouse stage.

Cell depletions
In one experiment, mice were depleted of CD8 + T cells, by IP injection of anti-CD8 monoclonal antibody (clone 53-6-72, produced in-house from hybridoma) at the dose of 200 mg/ 200 mL PBS on day 21 of tumor injection, then 100 mg/200 mL PBS on days 0 and 1, and every 3 to 4 days thereafter after start of treatment. Control monoclonal antibody (MAC4), was given to the control group at the same dose and rate. NK cells were depleted in another group by injection of anti-asialo GM1 (rabbit immunoglobulin, Wako Pure Chemical Industries Ltd, Osaka, Japan): 1 mL was diluted 1/10 in PBS and 200 mL injected on the day of start of treatment, then every 3 to 4 days thereafter.

Statistical analysis
Statistical significance in the experiments compared survival of mice between the various groups and was determined by the Wilcoxon log-rank (Peto-Prentice) test, using StatsDirect software (version 3.0.86 beta, StatsDirect Ltd, Altrincham, Cheshire, UK). Statistical significance comparing tumor sizes between groups was determined using the Mann-Whitney test and the unpaired T test, using StatsDirect software. Statistical analysis of lung metastases burden and ascites incidence was determined using Fishers' Exact Test, with StatsDirect software.

Results
Treatment with CpG alone, or in combination with aCD137, enhanced the survival of mice bearing various subcutaneous tumors C57BL/6 mice injected SC with either B16F10 or MC38 tumor cells, and BALB/c mice injected SC with either Renca or CT26 were treated with a combination of CpG IT and aCD137 IP.  Figure 1C) and ,0.0001 ( Figure 1E), for MC38: P = 0.002 ( Figure 1D) and 0.0039 ( Figure 1F)). In the Renca, CT26 and MC38 models, injection of CpG alone was also able to increase the survival of mice compared to non-treated mice ( Figure 1B, C, D, E). In addition, administration of aCD137 alone was also able to increase mouse survival of mice bearing CT26 or MC38 tumors ( Figure 1E and F). Except for the MC38 model ( Figure 1D and F), statistical significance was not quite achieved for groups receiving combined therapy with CpG+ aCD137 compared to monotherapy with CpG or aCD137 alone. However, given the trend of mice to survive longer after combined treatment in the current study, and the statistically significant improvement in survival afforded by combination therapy in our previous studies [5], we performed the remainder of the current study in Renca using the combination of CpG+aCD137. However, in the experiment shown in Figure 1B using the Renca model, we cannot claim that the combined therapy with CpG and aCD137 was more effective than monotherapy with CpG or aCD137 alone. Nevertheless, the use of the combination therapy allowed us to determine the feasibility of treating Renca using the model and schedules described below.
All these tumors were injected SC, but only B16F10 originated from cancer of this organ (melanoma). Both MC38 and CT26 are of bowel cancer origin and Renca is of renal cancer origin. Therefore we wished to determine if the significant survival seen with all these tumors injected SC would apply to the same tumors located orthotopically, a more realistic test of therapy efficacy. As we were familiar with Renca injected into the kidney producing renal cell cancer in mice, we decided to proceed with this model. Administration of CpG by various routes in combination with aCD137 IP enhances the survival of mice bearing orthotopic Renca tumors Renca was injected orthotopically into the outer cortex of the kidney. Orthotopic injection of Renca produces a renal tumor, which frequently metastasizes to the lung and occasionally to the liver, diaphragm and peritoneum. Orthotopic Renca is thus a very much more difficult mouse model in which to achieve survival with therapies compared with Renca SC which does not metastasize at all. In previous mouse studies and human clinical trials by other investigators, CpG has been used as a monotherapy or combined with other therapies for treating solid and hematological cancers (including melanoma, renal cell cancer, breast cancer and lymphomas), and has been injected by several routes, including SC, IT and intravenously (IV) [16][17][18][19][20]. In some of these studies intratumoral injection of CpG alone has been sufficient to induce regression or slow the growth of tumors. Given this, we initially tried injecting CpG into the skin (SC) with IP aCD137 to determine if this would affect the survival of the mice. Figure 2 shows that this led to no long term survivors, either as single therapeutic agents, or in combination, although all were statistically significant compared with not treated mice (P = 0.0038 for CpG only, P = 0.0008 for aCD137 only and P = 0.0006 for combined treatment).
Intratumoral delivery of CpG into SC tumors has been previously demonstrated to be more effective than systemic or remote injection of CpG in some cases [16,21], and given the success of IT injection in SC tumors ( Figure 1) we wished to determine if this could be replicated in the orthotopic metastatic tumor model. We therefore tested injecting the CpG intratumorally into kidney tumors combined with IP aCD137 to determine if this treatment regimen could induce an enhanced degree of mouse survival similar to that observed with subcutaneous Renca tumors. The therapeutic injection of kidney tumors was ultrasound-guided as per the method described above. Figure S1A and B depict the set-up of mouse, syringe and needle and ultrasound imaging equipment. To validate the accuracy of ultrasound-guided IT injection Figure S1C shows a kidney tumor filled with a dye (Trypan blue) which was performed during development of this technique. In experiments where CpG was injected IT, it was possible to see the solution being injected into   Table 1. Summary of necropsies performed on individual mice on day of death or at the end of experiment for the two experiments shown in Figure S3. the tumor, as shown in the montage of time-lapse photography ( Figure S2 and Movie S1). Permeation of tumors by CpG solution was evident by an observed reduction in density of the tumor during injection of the solution. Injection of material into each renal tumor was displayed on-screen in all ultrasound guided experiments, and permeation of CpG into tumors was readily and reproducibly observed. Figure 3 shows the pooled results of two experiments in which mice bearing kidney tumors were treated with CpG IT, performed using this methodology, in combination with aCD137 IP. For direct comparison, a group was treated with CpG IV plus aCD137 IP within the same experiment. In Figure 3, three groups were treated, either with PBS IT plus aCD137 IP or CpG IT plus aCD137 IP or CpG IV plus aCD137, and compared with a not treated group. The data for this figure was derived from two separate experiments, which are shown individually in Figure S3. The majority of mice which had to be culled in all groups had large kidney tumors, but mice were also culled because of ascites in the peritoneum, or due to large numbers of tumor metastases in the lungs, or large abdominal tumors ( Table 1). There was a significant increase in survival of both CpG groups (P = 0.0138 for IT injected group, and P = ,0.0001 for the IV injected group, Figure 3) compared with the no treatment group. In addition, there was a statistical significant difference between these two treatment groups (P = 0.0057) with the CpG IV treated group surviving longer than the CpG IT treated group. There was no significant difference in survival with aCD137 alone group compared with the not treated group (P = 0.3654).
Measurements of kidney tumors during ultrasound-guided procedures revealed that by day 19 tumor growth was inhibited significantly with intratumoral injection of CpG combined with intraperitoneal aCD137 when compared to not treated ( Figure  S4A), but was not significantly inhibited when compared with PBS injected intratumorally combined with aCD137 ( Figure S4B,C). All mice died with large renal tumors and/or metastatic burden. The primary tumor sizes of mice in each group on the day of death were compared within each experiment, shown in Table 1, and subjected to statistical analysis ( Table 2). Using the unpaired single-sided T test, tumor size following IT CpG delivery combined with aCD137 was statistically significantly smaller compared with tumors following IV delivered CpG plus aCD137 in Experiment 1 (P = 0.0296). Analysis of tumor size in Experiment 2 revealed that tumors in mice receiving CpG IT plus aCD137 were significantly smaller than those receiving PBS IT plus aCD137 (P = 0.032 T test). There was no statistically significant difference between primary tumor size at death in mice in other treatment groups. Taken together, these experiments suggest that CpG delivered IT could reduce the size of primary kidney tumors.
The majority of kidney-tumor bearing mice developed lung metastases ( Table 1), resulting in extensive lung burden at the time of death with the number of metastases often being too numerous to count (TNTC). When the proportion of mice with lung metastases TNTC was compared between mice receiving systemic or intratumoral CpG, no statistically significant difference was observed (Table 3). Similarly, when the proportion of mice developing ascites was compared to those without ascites, no statistically significant difference between mice in any treatment group was observed ( Table 4). In addition, no statistically significant difference was observed in the number of abdominal metastases between mice receiving any treatment ( Table 5).
For the combination therapy, comparison of the mice injected SC with CpG in Figure 2 with the mice injected IT with CpG in Figure 3 showed no statistical difference (P = 0.0887). CpG Table 1. Cont. injected alone was also not statistically different from the combination therapy with CpG injected IT (P = 0.4563).
Our main aim in this study was to determine the feasibility of ultrasound-guided administration of therapy for kidney tumors. We combined the aCD137 and CpG reagents since this combination has been demonstrated to be effective against some subcutaneous tumors. However, in the present study, we did not determine the contributions of each reagent individually, and it remains possible that there was no contribution to therapy from aCD137 in the kidney tumor setting.
To ascertain the mechanism of the enhanced effect of systemic administration of CpG we performed an experiment in which various immune subsets were depleted. Three IV injections of CpG together with IP aCD137 led to significantly longer survival (P = 0.002) of mice compared to not treated mice (Figure 4). Previously Uno et al [22] showed that CD8 and IFNc were necessary for a combination therapy three monoclonal antibodies, including anti-CD137. Thus we explored depletion of either CD8 + T cells or NK cells in our model. Results (Figure 4) revealed that CD8 + T cells were crucial for this effect (P = 0.1144 for CD8 depletion showing no statistical difference compared with not treated group, whereas P = 0.0043 for NK cell depletion showing that these cells were not needed for the therapy to work effectively). While CD4 + helper T cells can also be important in some cases for anti-tumor responses, it was not possible to determine their contribution in this model since anti-CD4 administration also leads to depletion of CD4 + regulatory T cells (Treg), which have been previously shown to play a role in responses against Renca tumors [23].

Discussion
Previous investigations using a combination therapy of intratumoral CpG (30 mg, three times/week for 3 weeks) with 100 mg aCD137 (once weekly for 3 weeks) were performed by Smith et al [24] in BALB/c-neuT mice bearing SC TUBO tumors. The combined treatment resulted in survival of two of a total of 9 animals from two experiments, but none in a total of 10 mice from a further two experiments. In another study, SC CpG (20 mg) with IP aCD137 (100 mg) therapy (one injection of each only) was used against SC B16 melanoma in mice, with and without combination with Trp2 peptides [25]. Tumor growth with just CpG and aCD137 at this dose was less effective than when combined with Trp2 peptides.
We demonstrate in the present study that the therapy of four injections of 50 mg CpG injected IT combined with three IP injections of 100 mg aCD137 resulted in significant survival of mice injected with a variety of tumors injected SC, including B16F10 (melanoma tumor), Renca (renal carcinoma), and CT26 and MC38 (both colon cancers). However when one of these tumors, Renca, was injected orthotopically, neither injecting CpG subcutaneously or into the actual renal tumor itself, in combination with aCD137 IP, promoted survival of mice as substantial as that seen in mice injected with Renca SC. A difference between effectiveness of some therapies in SC tumors compared with orthotopic tumors has been shown previously with other mouse tumor models [26][27][28][29][30] and by us using a different immunotherapy injected IP into mice bearing orthotopic Renca tumors [11,31]. Therefore recommendations have been made to test therapies in orthotopic models [6,7,29,30,[32][33][34] which are thought to be   [27,28,31,35]. Ultrasound imaging has been used in other models to treat tumor [12][13][14], but not to intratumorally treat renal tumors, to our knowledge. Several studies have used injection of orthotopic kidney tumors with therapeutic agents, but this has involved surgery on the mice to reveal the kidney and tumor and then injecting the tumor directly (but only limited to one injection timepoint) [36][37][38][39]. The method we have shown here demonstrates that ultrasound-guided injection of therapeutic agents is fast, minimally invasive and therefore suitable for repeated administrations and high throughput. Ultrasound imaging ease of use and sensitivity allows for a high resolution visualization of the injection of CpG medication into the tumor. This technique may be of use in other future therapies involving IT injection of kidney tumors.
We found that intratumoral injection of CpG reduced the size of primary kidney tumors compared to other treatments. In addition, we found that optimal treatment of metastatic orthotopic kidney cancers with the combined therapy of CpG plus aCD137 consisted of IV injections of CpG combined with IP injections of aCD137. This regimen statistically enhanced survival of mice, and was found to be the best route of delivery for CpG for this particular therapeutic combination. Since higher local concentrations are assumed to occur in IT delivery, this result was unexpected and future work would need to be done to explore why this is so. One potential explanation relates to the metastatic nature of the kidney tumor model, which differs from previous subcutaneous models used to compare local and systemic delivery of CpG [40,41]. In our tumor model, Renca grows as a primary kidney tumor and spontaneously metastasizes to lung, liver and other abdominal sites. Indeed, in this model, metastasis plays a major role in disease progression and mortality. Thus, systemic delivery of CpG may have reduced the burden of metastases to a greater extent than intratumoral delivery of CpG. This is supported by the observation that intratumoral CpG-injected mice died earlier than systemic CpG-treated mice despite a reduction in primary tumor size mediated by intratumoral CpG. Systemic CpG may have impacted directly on metastases, or may have mediated its effects more efficiently than IT injection through lymphoid compartments including the lymph nodes. Previous studies have been reported in which delivery of CpG to relevant lymphoid compartments is more effective than subcutaneous injection at inducing antibody and cell-mediated responses to model antigens in a vaccine setting [42].
It is also possible that local injection of CpG into kidney tumors is less effective due to the inherent nature of the kidney tumor microenvironment, which has recently been demonstrated to be more immunosuppressive than subcutaneous tumors [31]. It may also be that different routes are appropriate for different combinations of therapy.
This study provides support for the feasibility of ultrasoundguided injection of therapeutic material into kidney tumors. However, increases in mouse survival were only of the order of several days and very few mice survived long term. This is in contrast to many previous observations using CpG and aCD137 in the subcutaneous setting where survival of a large proportion of mice was extended for several weeks, with frequent long term  survivors observed [5]. The current study highlights the importance of using more physiologically relevant tumor models, involving both local and metastatic disease, when investigating therapies using immune-based agents.