Experimental model and treatment administration. (A) The 2-tumor site model. BALB/c mice were inoculated subcutaneously with 5 × 10⁶ A20 tumor cells at 2 different sites: right and left abdomen. Intratumoral injections of CpG were performed in the right tumor while the left tumor served for evaluation of the systemic antitumor immune response. (B) Schema of treatment. Treatment started 5 to 7 days after tumor inoculation when the tumors reached 0.5 cm to 0.7 cm in largest diameter. Mice then received daily injections of intratumoral CpG over 5 days (100 μg per injection, red arrows) and 2 intraperitoneal injections of mAbs used for T-cell modulation (anti-OX40, anti-CTLA4, anti-GITR, anti-FR4) on the first and last day of treatment (blue arrows). (C) CpG effect on the injected (right) and noninjected (left) tumors. In mice bearing 2 tumors, intratumoral injections of CpG induced regression of the injected (right) tumor, but failed to cure the distant (left) tumor, unless combined with effective T-cell modulation (*anti-OX40 + anti-CTLA4 mAbs were used in this example for T-cell modulation). All tumor curves presented in the following figures will show tumor growth at the left (non-CpG injected) site, which reflects the systemic antitumor effect of the immunotherapy.

T-cell modulation improves intratumoral CpG vaccination efficacy. (A) Single mAb therapy can improve the therapeutic efficacy of CpG vaccination. Using the 2-tumor model, mice were either left untreated or treated with CpG intratumorally with or without one mAb for T-cell modulation, as described in . A panel of 4 different antibodies was tested: anti-OX40 (400 μg/injection), anti-CTLA4 (100 μg/injection), anti-GITR (500 μg/injection), anti-FR4 (100 μg/injection). (B) Some double mAb therapy, but not all, can greatly enhance the therapeutic efficacy of CpG vaccination over single mAb therapy. Anti-OX40, anti-CTLA4, anti-GITR, and anti-FR4 were combined 2 by 2 in conjunction with CpG vaccination and tested for their ability to induce tumor regression. Similar to the previous experiment, mice bearing 2 tumors were either left untreated or treated with CpG intratumorally with or without injection of 2 different antibodies for T-cell modulation. Doses of antibodies were the same as previously described. To rule out a possible dose-effect, one additional group of mice received double dose (2x) of anti-OX40 mAb (800 μg/injection). All growth curves represent the size of the left (non-CpG–injected) tumor. Numbers indicate the ratio of tumor-free mice at day 100 (some groups experienced regression of the left [non-CpG–injected] tumor but eventually grew distant metastasic lymph nodes which are represented in the denominator shown).

CpG is required for therapeutic effect of anti-OX40 + anti-CTLA4 combination. Mice bearing tumors at 2 sites received either no treatment, CpG intratumorally alone at one site, anti-OX40 + anti-CTLA4 intraperitoneally (without CpG), or the combination of both (CpG + anti-OX40 + anti-CTLA4), according to the schedule described previously. Growth curves represent the size of the left (non-CpG–injected) tumor.

Mice cured with CpG + anti-OX40 + anti-CTLA4 combination are protected from rechallenge. Naive mice (●) or mice cured from A20 tumors with CpG + anti-OX40 + anti-CTLA4 given 100 days earlier (■) were (re)challenged with 2.5 × 10⁶ A20 tumors cells dorsally. Growth curves represent the average size of the tumor (n = 10 mice per group).

CpG + anti-OX40 + anti-CTLA4 induces antitumor IFN-γ–producing T cells and decreases the percentage of Treg cells. (A-F) Using the 2-tumor site model, mice bearing A20 tumors were either left untreated, treated with CpG alone, or treated with CpG + anti-OX40 + anti + CTLA4, using the same therapeutic protocol as previously described. Four days after the end of treatment, mice were killed and spleens and tumors were harvested separately and made into a single cell suspension. (A,B) Total splenocytes were cocultured in vitro in the presence of irradiated A20 tumor cells plus anti-CD28 mAb for 24 hours. CD4 and CD8 T cells were assayed for CD44 and intracellular IFN-γ expression. (A) Data from one representative mouse of each group shows IFN-γ expression according to CD44 expression. (B) Average percentage of CD4 and CD8 T cells expressing intracellular IFN-γ of each group (n = 3 mice per group). * indicates a statistically significant difference (P < .05). (C,D) Total cells from the tumor were cocultured in vitro in the presence of irradiated A20 tumor cells plus anti-CD28 mAb for 24 hours. Intracellular expression of IFN-γ for CD4 and CD8 T cells is shown for one representative mouse of each group (C) and as the average of each group (n = 3 mice per group; D). (E,F) Total cells from the tumor were directly analyzed for Treg subpopulation. The proportion of intracellular FoxP3 expressing CD4 T cells is shown for one representative mouse of each group (E) and as the average of each group (n = 3 mice per group; F).

CD4 and CD8 T cells are required for CpG + anti-OX40 + anti-CTLA4 therapeutic effect. (A) Using the 2-tumor site model, mice received either no treatment or CpG + anti-OX40 + anti-CTLA4, with or without CD4 or CD8 depletion. CpG (red arrows) and anti-OX40 + anti-CTLA4 mAb (blue arrows) were administered as previously. CD4 and CD8 depletion (green arrows) was performed by intraperitoneal injections of ascitic fluid containing 0.5 mg of either anti-CD4 mAb (GK1.5 hybridoma) or anti-CD8 mAb (2.43 hybridoma) on days −2, −1, 1, and 5 from start of treatment. Depletion of CD4 and CD8 T cells was confirmed by flow cytometry on peripheral blood. Growth curves represent the total tumor volume (with exception of the right, CpG-injected tumor) expressed as the sum of the surface of the left tumor and all superficial metastatic lymph nodes when present. Numbers indicate the ratio of tumor-free mice at day 100. (B) Alternatively, BALB/c CD8 KO mice bearing 2 tumors received either no treatment or CpG + anti-OX40 + anti-CTLA4 as previously described. Numbers indicate the ratio of tumor-free mice at day 100. (C) Cancer growth was strikingly different between CD4- and CD8-depleted mice after treatment: tumor cells rapidly disseminated to the lymph nodes (LNs) in CD8-depleted mice whereas they remained confined to the primary tumor site (T) in the CD4-depleted mice. Photographs show 3 representative mice from each group.