(A) Expression of TARC (pg/ml ± SEM of triplicate tests, Y-axis) in BAL of individual naïve mice (#3001 and 3002) and tumor-bearing mice (#3507, 3512, and 3522) was assessed by ELISA. To prepare BAL, lungs were washed with 1 ml PBS. (B) Tumor CM, but not control medium (III), induces TARC expression in lungs of mice (I). Lungs were formalin fixed and paraffin embedded at day six after daily single i.p. injections with 0.5 ml tumor CM (I and II), or control medium (III), or PBS (IV). A representative staining result with anti-TARC Ab (I, III and IV) or control isotype-matched Ab (II) from three mice per group experiment is shown. (C) CCR4 on the surface of 4T1 tumor cells was stained with anti-CCR4 Ab (black line) and control Ab (grey line). (D) Unlike 4T1-PE and 4T1.2-PE cells, CCR4⁺4T1 tumor cells migrate to TARC (100 ng/ml), but not I-309 (100 ng/ml). Y– axis, Chemotactic Index ± SEM of triplicate experiment. *P<0.05.

(A) FoxP3⁺ cells (Tregs), but not NK cells, are significantly increased in the lungs of mice i.p. treated with tumor CM (as in Fig.1B). In contrast, both populations were increased in spleens of tumor CM –treated mice. Y- axis, change in absolute numbers of cells compared with PBS-treated group ±SEM of three mice per group experiment repeated twice. (B) Tregs presence in the lungs of NOD/SCID mice after transfer with BALB/C Tregs or splenocytes, or non-Tregs (Non-Tregs) or splenocytes from IL-10 deficient mice (IL-10 KO) at day 28 post tumor challenge. Shown, mean proportion ±SEM of CD4⁺ cells (left, Y-axis); and mean of lung metastasis ±SEM (right, Y-axis) of four mice per group experiment. (C) CCR4 expression on Tregs (left panel) and NK cells (right panel) is shown after staining with anti-CCR4 Ab (grey) and control Ab (open area), respectively. (D) Tregs shown in Fig.2D chemotax to TARC (100 ng/ml). They also migrate to extracts from lungs (Lung ext.) of mice treated with tumor CM (as in Fig.1B) utilizing CCR4. TARC was added in upper wells of chemotactic chamber to abrogate chemotaxis. Y– axis, Chemotactic Index ± SEM of triplicate experiment repeated three times. *P<0.05; **P<0.01; *** P<0.001.

Purified Tregs or non-Tregs were incubated with CFSE – labeled NK cells at indicated ratio prior staining with PI. Shown, proportion (%) of apoptotic NK cells (PI⁺CFSE⁺ ±SEM, A) and Tregs (PI⁺CFSE⁻ ±SEM, B) of triplicate experiments. (C) Effects of βGBP on NK cells (filled circles) and T cells (open circles). Proportion of apoptotic cells (%, PI⁺ ±SEM of triplicate experiment) is shown. Experiments were repeated at least three times. (D) Lung metastasis can be controlled by targeting CCR4⁺ cells. 4T1.2 tumor-bearing mice were i.v. treated with TARC-PE38 or TARC-Ag, or control TARC-Ag or mock (PBS). Shown, mean lung metastasis ±SEM. Experiments were reproduced three times. * P<0.05; **P<0.01; ***P<0.001.

Shown, mean lung metastasis foci ± SEM of four-five per group mice experiments of: (A) BALB/C mice that werechallenged with 4T1, 4T1.2, 4T1-PE and 4T1.2-PE cells; (B) 4T1.2 tumor-bearing BALB/C mice that were intratumorally treated with 10 μg of TARC-PE38 or control TARC-Ag; (C) 4T1.2 tumor-bearing BALB/C mice that were depleted of Tregs using anti-CD25 Ab. Control mice were treated with isotype-matched antibody (mock); and (D) 4T1.2 tumor-bearing NOD/SCID mice that were transferred with splenocytes or Tregs from naïve BALB/C mice, or injected with PBS (Mock). All experiments were reproduced at least three times. *P<0.05; **P<0.01; *** P<0.001.

(A) BALB/C splenocytes pretreated with TARC-Ag or mock (PBS) restore lung metastasis of 4T1.2 tumor cells in NOD/SCID mice, which is abrogated by one hour TARC-PE38 pre-treatment of the splenocytes prior the adoptive transfer (TARC-PE38). Control mice (No Splenocytes) were only challenged with 4T1.2 tumor. Shown, mean of lung metastasis ±SEM of five mice per group experiment repeated three times. (B) NK cells protect lungs of NOD/SCID mice from metastasis of 4T1.2 tumor cells. Tumor-bearing mice were i.v. treated with anti-asialo GM1. Control mice were injected with rabbit sera (IgG) or with PBS. Shown, mean of lung metastasis ±SEM of four mice per group experiment reproduced twice. (C) The numbers of NK cells and Tregs inversely correlate in tumor – bearing BALB/C mice. Tregs were depleted as in Fig.2c. Shown, mean proportion ±SEM (%) of Tregs and NK cels in peripheral blood (PB), lymph nodes (LN) and spleens (SP) of four mice per group experiment. (D) The transfer of BALB/C Tregs or BALB/C splenocytes significantly reduces NK cell numbers in PB of tumor – bearing NOD/SCID mice. Control mice were transferred with BALB/C non-Tregs or splenocytes from IL-10 deficient mice (IL-10 KO splenocytes). Cells were analyzed by FACS after staining for Pan-NK, CD27 and CD25^High CD4⁺ markers. *P<0.05; **P<0.01; ***P<0.001.

(A). The data are from nonparametric Mann Whitney analysis of previously untreated 73 breast cancer patients. Shown, T cell (CD3⁺) and NK cell counts in PB of patients with stage II and with the advanced disease (stage III and IV). (B) Schema that summarizes the finding. Primary breast cancer is heterogeneous, it consists of at least CCR4⁺ and CCR4⁻ cells (1). TARC is induced in lungs by the primary tumor growing in mammary pad (open arrows, 2). This allows CCR4⁺ tumors to leave the primary site (3) and metastasize to lungs (2). However, the cells are efficiently killed by NK cells (6), unless they migrate together with their protectors – CCR4⁺ Tregs (4). Tregs inactivate NK cells using βGBP (5). However, Tregs can be also killed by NK cells (7), if they are present at lower numbers. (C) Immunohistochemistry staining of mouse (I-III) and human (IV-VI) lungs with 4T1 tumors (II and III) and human breast cancer (V and VI). Shown, control Ab (III and VI) or anti-TARC Ab (II and V) stainings of tumor-bearing lungs or normal tumor-free lungs (I and IV) of mice and human, respectively.