Fig 1.
Myeloid cell-specific deletion of autophagy attenuates tumor metastasis.
(A) The Atg5 gene was specifically deleted in myeloid cells of LysM-Atg5-/- mice. The cells isolated from in vivo growing tumors (MC38) were prepared from wild-type (WT) or LysM-Atg5-/- mice, and CD11b+Ly6C+ monocytes, CD11b+F4/80+ macrophages, CD11b+Gr-1+ granulocytes, CD3ε+ T cells and B220+ B cells were isolated by flow cytometry. Expression of Atg5 was evaluated by RT-PCR. (B) Atg5flox/flox or LysM-Atg5 -/- mice were inoculated subcutaneously with MC38 or B16 cells. Tumor growth was measured on the indicated days. (C) B16 cells were injected via tail veins of Atg5flox/flox or LysM-Atg5 -/- mice. 28 days later, the mice were sacrificed and subjected to macroscopic analysis for metastatic lesions in the lung. The experiments were performed at least ten times with similar results. (D) MC38 cells were injected into the spleen of Atg5flox/flox or LysM-Atg5 -/- mice. Twenty-one days later, the mice were sacrificed and subjected to macroscopic analysis for metastatic lesions in the liver. Representative data (left) and statistical analysis of average numbers and areas (mm2) of metastatic lesions are shown (right). The tumor metastatic lesions are shown by white arrows. The experiments were performed at least ten times with similar results. (E) MC38 cells were injected into the spleens of Atg5flox/flox or LysM-Atg5 -/- mice. Twenty-one days later, the mice were sacrificed and subjected to macroscopic analysis for disseminated lesions in the peritoneal cavity. Representative data (left) and statistical analysis of average numbers and areas (mm2) of metastatic lesions are shown (right). The experiments were performed five times with similar results. *: p<0.05, ns: not significant.
Fig 2.
Myeloid cell-specific autophagy supports the chemoresistance and invasive properties of tumor cells through TGF-β1-dependent manner.
(A, B) MC38 colon cancer cells were cultured with supernatants from Atg5flox/flox or LysM-Atg5-/- TAM (MS) for 48 h in the presence of DMSO or the TGF-β receptor kinase inhibitor SB431542 (TGFβR-KI). (A) The indicated cells were then treated with Cisplatin (CDDP: 10 μM) for 24 h, and the cell viability was determined by flow cytometry with PI and Annexin V staining. (B) The indicated cells were evaluated for in vitro invasion by Matrigel assay. (C) MC38 tumor cells were cultured with supernatants from Atg5flox/flox or LysM-Atg5-/- TAM (MS) with or without SB431542 (TGFβR-KI) for 48 h. The cellular morphological appearance was analyzed by microscopy. (D, E) MC38 tumor cells were cultured with supernatants from Atg5flox/flox or LysM-Atg5-/- TAM (MS) with or without SB-431542 (TGFβR-KI) or recombinant TGF-β1 (100 ng/mL). The expression levels of N-cadherin (solid line) and vimentin (green) were analyzed by flow cytometry (D: left) and confocal microscopy (E), respectively. The statistical analysis was also performed for flow cytometry data (n = 3; D: right). *: p<0.05, ns: not significant.
Fig 3.
TGF-β upregulated by myeloid cell-specific autophagy is a major mediator in regulating the invasive properties of tumors.
(A)Production of TGF-β1, IL-6, IL-10, IL-12p40 and IFN-γ was assayed by ELISA in culture supernatants of CD11b+F4/80highGr-1low(F4/80HGr-1L) and CD11b+F4/80lowGr-1high (F4/80LGr-1H) TAM obtained from B16 lung metastasis lesions of Atg5flox/flox or LysM-Atg5-/- mice. (B) B16 melanoma cells were injected intravenously into Atg5flox/flox or LysM-Atg5-/- mice. Mice were then treated with anti-TGF-β1 mAb (1D11) or isotype control Ig twice per week. Twenty-one days later, the mice were sacrificed and subjected to macroscopic analysis for metastatic lesions in the lung. The tumor metastatic lesions are shown by white arrows. Representative data (top) and statistical analysis of average areas of metastatic lesions are shown (bottom). *: p<0.05, ns: not significant.
Fig 4.
Mechanism of TGF-βinduction by myeloid cell-specific autophagy.
(A)The mRNA levels of TGF-β1 in Atg5flox/flox or LysM-Atg5-/- TAM were quantified by RT-PCR. (B) The protein levels of pro-TGF-β in wild-type (WT), Atg5flox/flox or LysM-Atg5-/- TAM were evaluated by western blot. (C) Integrin-αv expression on Atg5flox/flox or LysM-Atg5-/- TAM was evaluated by flow cytometry. (D) Atg5flox/flox or LysM-Atg5-/- TAMs were pretreated with MG132 or DMSO for 6h, and pro-TGF-β protein levels were measured by western blot.
Fig 5.
Intratumoral immunomodulation by myeloid cell-specific autophagy.
(A) Atg5flox/flox or LysM-Atg5 -/- BMDM were exposed to irradiated EG7 cells for 2 h and then co-cultured with wild-type CD4+ or CD8+ T cells for 96 h. CD4+ T cells expressing IFN-γ, IL-4, IL-17 or Foxp3, as well as CD8+ T cells expressing granzyme-B (GZM-B) were assayed by flow cytometry. (B) B16-OVA cells were inoculated subcutaneously into LysM-Atg5-/- or Atg5flox/flox mice that had been adoptively transferred with OT-I cells. After tumors were established (~25 mm2), tumor-infiltrating lymphocytes were isolated from the tumor tissues 4 days after the final treatment of CDDP and analyzed for the expression of SIINFEKL/H-2Kb complex on TAM by flow cytometry. Representative dot plots (left) and statistical analysis (n = 4; right) are shown. (C) MC38 cells were injected into the spleens of Atg5flox/flox or LysM-Atg5-/- mice two days after intravenous administration of anti-CD8 Ab. Twenty-one days later, the mice were sacrificed and subjected to macroscopic analysis for metastatic lesions in the liver. Representative data (upper) and statistical analysis of average areas of metastatic lesions are shown (bottom). White arrows indicate metastatic tumor areas. *: p<0.05, ns: not significant.
Fig 6.
Myeloid cell autophagy suppresses antitumor immunity by TGF-β.
(A)Atg5flox/flox or LysM-Atg5-/- BMDM were exposed to irradiated EG7 thymoma cells for 2 h and then co-cultured with wild-type CD4+ or CD8+ T cells in the presence of anti-TGF-β1 neutralizing mAb (1D11) or control Ig for 96 h. CD4+ T cells expressing Foxp3 and CD8+ T cells expressing granzyme-B (GZM-B) were assayed by flow cytometry. (B) B16-OVA cells were inoculated intravenously into Atg5flox/flox or LysM-Atg5-/- mice, and TAM isolated from B16-OVA tumors in metastatic lungs were co-cultured with CD8+ T cells for 72 hr. The expression of OVA peptide (SIINFEKL) tetramer positive cells among the CD8+ T cells were analyzed by flow cytometry. Representative dot plots (top) and statistical analysis (n = 3; bottom) are shown. (C) MC38 cells were injected into the spleen of Atg5flox/flox or LysM-Atg5-/- mice two days after intravenous administration of anti-CD8 depleting mAb, anti-TGF-β1 neutralizing mAb, or both. Twenty-one days later, the mice were sacrificed and subjected to macroscopic analysis for metastatic lesions in the liver. Representative data (top) and statistical analysis of liver weight and average areas of metastatic lesions (mm2) are shown (bottom). The tumor metastatic lesions are shown by yellow arrows. *: p<0.05, ns: not significant.
Fig 7.
CSF-1 supports the survival and differentiation of immunosuppressive macrophages in tumors.
(A) Atg5flox/flox or LysM-Atg5-/- BMDM were exposed to serum starvation for 16 h, low oxygen (1%) for 16 h, or UV light (300 J/m2) for 10 sec after 16h of culture. The cell viability was evaluated by quantifying annexin-V/PI-positive populations. Representative dot plots (upper) or statistical analysis (bottom) are shown. (B) MC38 cells were treated with culture medium (CM) or 50% supernatants of Atg5flox/flox or LysM-Atg5-/- BMDM for 72h. The supernatants of BMDM-primed MC38 cells were then used to treat wild-type BMDM in the presence of TGFβ receptor kinase inhibitor (TGFβR-KI: 10μM) or DMSO for 72 h. The expression of CD11b and CD206 was evaluated by flow cytometry. NT: untreated BMDM. (C) The percentages of F4/80+CD11b+macrophages, CD11b+CD206high M2 macrophages in MC38 tumors from control Ig or anti-TGF-β-treated Atg5flox/flox or LysM-Atg5-/- mice were quantified by flow cytometry. (D) B16 cells were injected intravenously to Atg5flox/flox or LysM-Atg5-/- mice two days after intraperitoneal administration of GW2580 (50 mg/kg) or PBS twice per week. Twenty-one days later, the mice were sacrificed and subjected to macroscopic analysis for metastatic lesions in the lung. Representative data (top) and statistical analysis of areas of metastatic lesions (mm2) are shown (bottom). (E) The numbers of total or M2 macrophages in B16 tumors from control Ig or anti-GW2580-treated Atg5flox/flox or LysM-Atg5-/- mice were quantified by flow cytometry. *: p<0.05, ns: not significant.
Fig 8.
TAM-derived autophagy contributes to the invasive activities of cancer patient-derived tumor cells.
(A) Detection of LC3 in CD11bhighmyeloid cells from tumors (TAM) or peripheral bloods (PBM) of NSCLC patients (n = 5). Red arrows indicate the cells with punctate LC3 structures. (B) Protein levels of TGF-β1 in culture supernatants of CD11bhigh myeloid cells from tumors (TAM) or peripheral bloods (PBM) derived from NSCLC patients as measured by ELISA (n = 10). (C-E) A549 or primary NSCLC cancer cells (PLC-HU1) were cultured with supernatants of myeloid cells from tumors (TAM) or peripheral bloods (PBM) of NSCLC patients (n = 3) for 48 h in the presence of 3MA or TGFβR-KI. The indicated cells were evaluated for the expression of E-cadherin and N-cadherin by flow cytometry (C), the expression of vimentin by fluorescence microscopy (D) and the in vitro invasion by Matrigel assay (E). *: p<0.05, ns: not significant.
Fig 9.
Schema of mechanism by which myeloid cell-specific autophagy promotes tumorigenesis.
Myeloid cell-derived autophagy promotes tumor metastasis through TGF-β-mediated EMT, as well as activates pro-survival signals to protect myeloid cells from stress-prone tumor microenvironments in support with tumor-derived CSF-1. These dual cascade bestow the surviving tumor-infiltrating myeloid cells to further amplify pro-metastatic cascade.