Tumor-reactive immune cells protect against metastatic tumor and induce immunoediting of indolent but not quiescent tumor cells

Abstract

Two major barriers to cancer immunotherapy include tumor-induced immune suppression mediated by myeloid-derived suppressor cells and poor immunogenicity of the tumor-expressing self-antigens. To overcome these barriers, we reprogrammed tumor-immune cell cross-talk by combined use of decitabine and adoptive immunotherapy, containing tumor-sensitized T cells and CD25(+) NKT cells. Decitabine functioned to induce the expression of highly immunogenic cancer testis antigens in the tumor, while also reducing the frequency of myeloid-derived suppressor cells and the presence of CD25(+) NKT cells rendered T cells, resistant to remaining myeloid-derived suppressor cells. This combinatorial therapy significantly prolonged survival of animals bearing metastatic tumor cells. Adoptive immunotherapy also induced tumor immunoediting, resulting in tumor escape and associated disease-related mortality. To identify a tumor target that is incapable of escape from the immune response, we used dormant tumor cells. We used Adriamycin chemotherapy or radiation therapy, which simultaneously induce tumor cell death and tumor dormancy. Resultant dormant cells became refractory to additional doses of Adriamycin or radiation therapy, but they remained sensitive to tumor-reactive immune cells. Importantly, we discovered that dormant tumor cells contained indolent cells that expressed low levels of Ki67 and quiescent cells that were Ki67 negative. Whereas the former were prone to tumor immunoediting and escape, the latter did not demonstrate immunoediting. Our results suggest that immunotherapy could be highly effective against quiescent dormant tumor cells. The challenge is to develop combinatorial therapies that could establish a quiescent type of tumor dormancy, which would be the best target for immunotherapy.

Keywords: adoptive immunotherapy; breast cancer; dormancy; escape; relapse.

Figure 1.. AIT, with or without Dec, fails to induce regression-established mammary carcinoma.

(A and B) Animals were challenged with MMC (3 × 10⁶) i.d. in the mammary gland region; upon the tumor reaching 50–70 mm³, animals were conditioned with CYP (100 mg/kg). The following day, mice remained untreated (Control; n = 4) or received AIT (n = 4). (C) MMC tumor cells were cultured with (+Dec) or without Dec (Untreated; 3 mM) for 72 h; RNA was then extracted and converted to cDNA, followed by qRT-PCR, using primers specific for 6 murine CTAs. (D) Tumor-bearing (∼1000 mm³ i.d.) FVBN202 mice received 5 injections of Dec, 1/d (+Dec; 2.5 mg/kg; n = 1) or remained untreated (n = 1); the tumors were harvested 3 d later, and cDNA was generated to quantify CTA expression, which was normalized to GAPDH. AKAP4, A-kinase anchor protein 4; ESX1, ; MAGEA4, melanoma-associated antigen 4; MAGEB5, melanoma-associated antigen B5; SPA17, ESX1, Extraembryonic, spermatogenesis, homeobox 1; SPA17, Sperm Autoantigenic Protein 17. (E) Animals were challenged i.d. with MMC (3 × 10⁶) in the mammary gland region; after tumors reached 50–70 mm³, all animals were treated with Dec (every other day for 3 total injections; 2.5 mg/kg, i.p.; n = 3) or remained untreated (n = 3). Seven days later, mice were euthanized, and MDSCs were analyzed in the spleen. (F and G) Animals were challenged i.d. with MMC (3 × 10⁶) in the mammary gland region; after tumors reached 50–70 mm³, all animals were treated with Dec [every day for 3 total injections (×3); 2.5 mg/kg, i.p.]. Two days later, animals were conditioned with CYP (100 mg/kg, i.p.). The following day, mice remained untreated (Control; n = 3) or received AIT (n = 4), derived from a CTA⁺ tumor-bearing donor. (H) MMC cells remained untreated (MMC) or were treated with Dec (3 mM; 72 h) to induce CTA expression (CTA-MMC). Tumor cells were then cocultured with reprogrammed splenocytes (1:10) for 20 h. IFN-γ was detected in the supernatant by ELISA. Data represent means ± sem of duplicate wells.

Figure 2.. Combined use of Dec and AIT prolongs survival of animals and induces tumor immune editing.

FVBN202 mice were challenged with 1 × 10⁶ MMC cells i.v. Mice then remained untreated (Control; n = 4), received AIT (n = 7) on the same day as tumor challenge, received Dec (Dec; n = 4; 5 daily doses beginning on d 3 after tumor challenge), or received Dec and AIT (AIT + Dec; n = 6; AIT on the day of tumor challenge, followed by Dec beginning on d 3).

Figure 3.. AIT promotes immunoediting of lung metastatic lesions.

(A) Splenocytes were harvested from untreated mice (Control; n = 5) and AIT recipients (n = 3) and cultured in the presence of MMC cells (10:1) for 20 h. Supernatants were collected and subjected to IFN-γ ELISA. (B) Metastatic lesions in the lung of FVBN202 mice that remained untreated (Control; n = 3), AIT recipients (n = 6), and AIT + Dec recipients (n = 4) were harvested when mice became moribund. Tumor lesions were digested and analyzed. MFI of neu and frequency of neu loss (B) and MFI of PD-L1 (C) were then quantified using flow cytometry using MMC cell line (MMC) as an in vitro control. (D) MMC cells were cultured with IFN-γ or splenocytes of tumor-bearing mice, and PD-L1 was detected after 16–20 h (n = 2–3). (E) Spleens of FVBN202 mice bearing primary mammary carcinoma (n = 4) were harvested after tumors were ≥1000 mm³. PD-1 expression was then quantified on the splenocytes, pre- and postreprogramming. (F and G) Metastatic lesions in the lung of FVBN202 mice that remained untreated (Control) and AIT recipients were harvested when mice became moribund. (F) The frequency of CD8⁺ T cell infiltration metastatic lesion in the lung of control mice and the AIT group (n = 3) was determined on gated CD3⁺ cells. SSC-A, Side-scatter-area. (G) Expression of PD-1 was determined on CD3⁺CD8⁺ cells (Control, n = 1; AIT, n = 3) by gating on CD45⁺ viable leukocytes. (H) Spleens of FVBN202 mice that had received AIT (n = 4) or remained untreated (n = 3) and were i.v. challenged with MMC were analyzed by flow cytometry after tumor-bearing mice became moribund to quantify PD-1 expression. Data represent means ± sem.

Figure 4.. ADR treatment results in the emergence of indolent and quiescent tumor dormancy.

(A) MMC tumor cells were treated with 3 daily doses of ADR (1 µM for 2 h) and then remained untreated for 3 wk. The frequency of viable MMC cells was determined by quantifying FVS⁻ cells using flow cytometry. (B) At wk 1 and 3 post-treatment, adherent and viable tumor cells were counted by trypan blue exclusion. (C and D) MMC cells were treated for 3 consecutive d with ADR (1 µM, 2 h) or left untreated. Three weeks later, ADR-treated and untreated MMC cells were stimulated with IFN-γ (50 ng/ml) for 12–16 h to induce the expression of PD-L1. (C) Emergence of Ki67 was determined in control MMC cells (Untreated), as well as ADR-treated cells (+ADR) ± IFN-γ stimulation. (D) The expression of PD-L1/cell was calculated by dividing PD-L1 MFI by the frequency of Ki67⁻ cells in ADR-treated and untreated MMC cells ± IFN-γ stimulation. Data represent 3 independent experiments and means ± sem.

Figure 5.. Immunotherapy displays cytotoxic function against treatment refractory dormant tumor cells in vitro.

(A) MMC cells (n = 3) treated with ADR (1 µM, 2 h) or 2 Gy RT (RT-treated MMC) for 3 consecutive d and remained in culture for 8 d total to establish tumor cell dormancy in vitro. (B) On d 8, these dormant tumor cells were treated with a high-dose ADR (1 µM, 24 h; ADR-treated MMC + ADR^hi) or reprogrammed immune cells (ADR-treated MMC + immune cells; ADR-treated MMC + immune cells). Two days later, cells were stained with Annexin V/PI and analyzed by flow cytometry. Data represent 3 biologic repeats and means ± sem. (C) On d 8, these dormant tumor cells were treated with 18 Gy RT (RT-treated MMC + RT^hi) or reprogrammed immune cells (RT-treated MMC + immune cells). Two days later, cells were stained with Annexin V/PI and analyzed by flow cytometry. (D) MMC tumor cells or dormant MMC cells (RT-MMC, ADR-MMC) were cultured in the absence or presence of the reprogrammed immune cells in a 10:1 ratio for 24 h. Control immune cells were cultured alone (Medium). IFN-γ release was detected in the supernatant using ELISA. Data represent 2 biologic repeats and means ± sem.