Article Figures & Data
Figures
- Figure 1. LLC cells exhibit a blunted response to IFNγ in vitro and in vivo compared with CMT167.
CMT167 or LLC cells were orthotopically injected into the left lung lobe of transgenic GFP-expressing C57BL/6J mice and were grown for either 2 (LLC) or 3 wk (CMT167). Tumor-bearing lung lobes were isolated and made into single-cell suspensions containing both GFP-positive (host cells) and GFP-negative (cancer cells). First, RNA was isolated from identical cancer cells grown in passage (in vitro condition). Second, RNA was isolated from recovered GFP-negative cancer cells (isolated via FACS-in vivo condition). Third, RNA was run for RNA-Seq from both conditions. The CMT167 condition had three experimental replicates per in vitro and in vivo conditions with five tumor-bearing lung lobes pooled per in vivo experimental replicate (15 mice used total). The LLC condition had five experimental replicates per in vitro and in vivo conditions with four tumor-bearing lung lobes pooled per in vivo experimental replicate (20 mice used total). (A) The “HM_IFNg Response” pathway containing 80 genes from the Broad Institute Molecular Signatures Database was used to generate a heat map of differentially expressed genes between the LLC and CMT167 in vitro and in vivo experimental conditions.
Source data are available for this figure.
Source Data for Figure 1[LSA-2019-00328_SDataF1.xlsx]
- Figure S1. LLC cells exhibit a blunted response to IFNγ in vitro and in vivo compared with CMT167.
LLC or CMT167 cells were orthotopically injected into the left lung lobes of mice and established primary tumors. After 2 (LLC) or 3 (CMT167) wk of tumor growth, the mice were euthanized and their tumor-bearing lung lobes were isolated and snap-frozen in liquid nitrogen. After first thaw, the lung lobes were homogenized, followed by RNA isolation from bulk tumor homogenate and qRT-PCR. (A) mRNA expression of Ifnγ relative to the housekeeping gene Gapdh from LLC and CMT167 tumor homogenate. (B, C) Conditioned media from cells plated at equal confluency was measured for (B) CXCL9 and (C) CXCL10 proteins via ELISA comparing LLC cells with CMT167 cells in vitro ±100 ng/ml IFNγ for 48 h. (D) Representative flow cytometry plots showing median fluorescence intensity (MFI) of LLC or CMT167 cells ±100 ng/ml IFNγ and ±1 μm ruxolitinib for 72 h. (E–H) Protein expression (MFI) as measured by flow cytometry of (E) PD-L1, (F) MHCII, (G) H2-D, and (H) H2-K in LLC or CMT167 cells in vitro. Statistics compare LLC and CMT167 cells treated with IFNγ alone or in combination with ruxolitinib. Error bars represent the mean of the data ± SEM after a t test (S1A) or a two-way ANOVA (S1B-C, E–H). (*P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001).
- Figure 2. In vitro treatment with IFNγ recapitulates RNA-Seq differences in the IFNγ signaling pathway in LLC and CMT167 cells.
(A) Immunoblots of LLC or CMT167 cells treated with ±10 ng/mL IFNγ in vitro for a time course ranging from 15 min to 24 h, showing p-STAT1 and total STAT1 expression compared with the housekeeping gene β-ACTIN. LLC or CMT167 cells were treated with ±100 ng/ml IFNγ in vitro for 24 h followed by isolation of RNA and qRT-PCR. (B–E) mRNA levels of (B) Cxcl9, (C) Cxcl10, (D) Cd274, and (E) Ciita are shown as absolute values (SQ values) normalized to the housekeeping gene Actb. Statistics compare all groups ± IFNγ. Error bars represent the mean of the data ± SEM after a two-way ANOVA (B–E) (*P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001).
Source data are available for this figure.
- Figure S2. Silencing Ifngr1 in the CMT167 line confers decreased response to IFNγ in vitro and in vivo.
(A, B) In vitro mRNA expression of (A) Ifngr1 and (B) Ifngr2 in FPKM as assessed by RNA-Seq. (C) Immunoblots of IFNGR1 and IFNGR2 expression relative to β-ACTIN levels in LLC and CMT167 cells in vitro. (D–I) Full data sets for mRNA expression are shown for (D) Ifngr1, (E) Cxcl9, (F) Cxcl10, (G) Cd274, (H) Ciita, and (I) Socs1 after ±10 ng/ml IFNγ for 24 h. Statistics compare cells lines treated with IFNγ relative to the CMT-NT cell line. (J) PD-L1 protein expression (MFI) of CMT-NT versus CMT-sh68sc3 cells in vitro by flow cytometry after ±10 ng/ml IFNγ treatment for 72 h. Statistics compare all groups. (K, L) Representative images of H&E staining of tumors from the following experimental conditions: (K) CMT-NT IgG2a and CMT-NT PD-1 and (L) CMT-sh68sc3 IgG2a and CMT-sh68sc3 PD-1 (scale bars: 100 μm and magnified to 10 μm). Tumors are dense and stain dark purple relative to adjacent normal lung tissue. Red arrows mark lymphocyte nests. Error bars represent the mean of the data ± SEM after a t test (S2A, B) or a two-way ANOVA (D–J) (*P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001).
Source data are available for this figure.
- Figure 3. Silencing Ifngr1 in the CMT167 line confers decreased response to IFNγ in vitro and in vivo.
Two separate shRNAs targeting Ifngr1 (CMT-sh68 and CMT-sh69) and a nontargeting control vector (CMT-NT) were transduced into CMT167 cells expressing luciferase. CMT167 cells were then screened for functional and stable knockdown of Ifngr1 after 10 d of puromycin selection and subcloning of shRNA pools (CMT-sh68sc3 and CMT-sh69sc2). The cells were treated with ±10 ng/ml IFNγ for 24 h followed by isolation of RNA and qRT-PCR. (A) mRNA levels of (A) Ifngr1 was assessed for knockdown of shRNA subclones relative to the CMT-NT cell line. (B) Immunoblots showing p-STAT1, total STAT1, and β-ACTIN levels of the CMT-NT, CMT-sh68sc3, and CMT-sh69sc2 cell lines ± IFNγ after 15 min or 1 h in vitro. (C–F) mRNA expression of downstream IFNγ response genes (C) Cxcl9, (D) Cxcl10, (E) Cd274, and (F) Ciita are shown as absolute values (SQ values) normalized to the housekeeping gene Actb after ±10 ng/ml IFNγ treatment for 24 h. Statistics compare the CMT-NT cell line to other cell lines with IFNγ treatment. CMT-NT or CMT-sh68sc3 cells were orthotopically injected into the lungs of syngeneic mice, established for 7 d, then were treated with either an isotype control antibody (IgG2a) or an anti–PD-1 antibody for 3 wk followed by terminal euthanization at 4 wk post tumor cell injection. (G) Primary tumor volume was assessed using digital calipers. Error bars represent the mean of the data ± SEM after a two-way ANOVA (A, C–F) or a one-way ANOVA (G) (*P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001).
Source data are available for this figure.
- Figure 4. Silencing Socs1 in the LLC line confers increased response to IFNγ in vitro and in vivo.
(A) In vitro mRNA expression of Socs1 in FPKM as assessed by RNA-Seq in LLC and CMT167 cells. (B) Immunoblots of LLC or CMT167 cells treated with ±10 ng/ml IFNγ in vitro for a time course ranging from 15 min to 24 h, showing SOCS1 expression relative to the housekeeping gene β-ACTIN. Two separate shRNAs targeting Socs1 (LLC-sh20 and LLC-sh21) and a nontargeting control vector (LLC-NT) were transduced into LLC cells expressing luciferase. LLC cells were then screened for functional and stable knockdown of Socs1 after 10 d of puromycin selection. (C) Immunoblots showing p-STAT1, total STAT1, and β-ACTIN levels of the LLC-NT, LLC-sh20, and LLC-sh21 cell lines ± IFNγ at 15 min or 1 h in vitro. (D) mRNA expression of Socs1 via qRT-PCR shown as absolute values (SQ values) normalized to the housekeeping gene Actb. Statistics compare the LLC-NT line to the knockdown cell lines. (E) Immunoblot for SOCS1 relative to β-ACTIN protein levels in the LLC-NT, LLC-sh20, and LLC-sh21 cell lines in vitro at baseline. LLC-NT or LLC-sh21 cells were orthotopically injected into the lungs of syngeneic mice, established for 7 d, then were treated with either an isotype control antibody (IgG2a) or an anti–PD-1 antibody for 2 wk followed by terminal euthanization at 3 wk post tumor cell injection. (F) Primary tumor volume of lung tumors was assessed using digital calipers. LLC-sh21 cells were injected into the flanks of syngeneic mice, established for 7 d, then were treated as in (F). (G) Primary tumor volume of subcutaneous tumors (Flank) was also assessed with digital calipers. Error bars represent the mean of the data ± SEM after a t test (A, G) or a one-way ANOVA (D, F) (*P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001).
Source data are available for this figure.
- Figure S3. Silencing Socs1 in the LLC line confers increased response to IFNγ in vitro.
Two separate shRNAs targeting Socs1 (LLC-sh20 and LLC-sh21) and a nontargeting control vector (LLC-NT) were transduced into LLC cells expressing luciferase. LLC cells were then screened for functional and stable knockdown of Socs1 after 10 d of puromycin selection. (A) Immunoblots showing p-STAT1, total STAT1, and β-ACTIN levels of the LLC-NT, LLC-sh20, and LLC-sh21 cell lines ±100 ng/ml IFNγ after 15 min or 48 h in vitro. The cells were treated with ± 10 ng/ml IFNγ for 24 h followed by isolation of RNA and qRT-PCR. (B–G) mRNA levels of (B) Socs1, (C) Ifngr1, and downstream IFNγ response genes (D) Cxcl9, (E) Cxcl10, (F) Cd274, and (G) Ciita are shown as absolute values (SQ values) normalized to the housekeeping gene Actb from full data sets. Statistics compare the LLC-NT line with other cell lines with IFNγ treatment. Error bars represent the mean of the data ± SEM (*P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001) after a two-way ANOVA (B–G).
Source data are available for this figure.
- Figure S4. Silencing Socs1 in the LLC line confers increased response to IFNγ in vitro by protein.
LLC-NT or LLC-sh21 cells were treated ±100 ng/ml IFNγ for 48 h. (A, B) Conditioned media from cells plated at equal confluency was measured for (A) CXCL9 and (B) CXCL10 protein via ELISA comparing the LLC-NT line to LLC-sh21 cells in vitro. (C–F) Statistics compare all groups. Protein expression (MFI) by flow cytometry was assessed for (C) PD-L1, (D) MHCII, (E) H2-D, and (F) H2-K ±100 ng/ml IFNγ and ± 1 μM ruxolitinib treatment for 48 h. Statistics compare IFNγ treated cells alone with IFNγ or in combination with ruxolitinib. Error bars represent the mean of the data ± SEM (*P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001) after a two-way ANOVA (A–F).
- Figure S5. Socs1 KD tumors show alterations in multiple populations.
LLC-NT or LLC-sh21 cells were orthotopically injected into the left lung lobes of mice and established primary tumors. After 2 wk of tumor growth with no treatment, mice were euthanized and their tumor-bearing lung lobes were isolated. Single-cell suspensions were made from tumor-bearing lung lobes or naive lungs. Naive biological replicates each contained lungs isolated from one mouse (three mice/three replicates total). LLC-NT and LLC-sh21 biological replicates each contained tumor-bearing lungs from three pooled mice (nine mice/three replicates total; three mice/replicate). Single-cell suspensions were then stained with a 39-antibody panel and analyzed on the Helios mass cytometer. Data show all viable single cells, subjected to the PhenoGraph algorithm. (A) Primary tumor volume was assessed by digital calipers. (B) PhenoGraph-based visualization on a tSNE plot, colored according to the expression of the cellular markers (parameters): CD3, CD4, CD8, CD11b, CD11c, CD19, CD44, CD45, CD64, CD90.2, Gr-1, MHC-II, and Siglec-F. (C) PhenoGraph-defined cellular distribution and clustering, as defined by tSNE1 and tSNE2, colored by cluster ID for all experimental groups and replicates. (D) Pie charts show the abundance of all clusters containing cancer cells (clusters #5 and #35) identified as CD45−and missing in naïve samples. (E) Cancer cell clusters in LLC-NT samples or LLC-sh21 samples colored according to Ki67, MHCI, MHCII, and PD-L1.
- Figure 5. Socs1 KD tumors show alterations in multiple populations.
LLC-NT or LLC-sh21 cells were orthotopically injected into the left lung lobes of mice and established primary tumors. After 2 wk of tumor growth with no treatment, the mice were euthanized and their tumor-bearing lung lobes were isolated. Single-cell suspensions were made from tumor-bearing lung lobes or naïve lungs. Naïve biological replicates each contained lungs isolated from one mouse (three mice/three replicates total). LLC-NT and LLC-sh21 biological replicates each contained tumor-bearing lungs from three pooled mice (nine mice/three replicates total; three mice/replicate). Single-cell suspensions were then stained with a 39-antibody panel and analyzed on the Helios mass cytometer. Data show all viable single cells, subjected to the PhenoGraph algorithm. (A) PhenoGraph-defined cellular distribution and clustering, as defined by tSNE1 and tSNE2, colored by phenotypic designation (legend provided in panel B) for all treatment conditions where all replicates per experimental condition are combined. (B) Pie charts show all 35 clusters colored by their phenotypic designations for all experimental conditions with numbers indicating which PhenoGraph-defined clusters were present in each phenotypic designation. (C–E) Clusters identified as statistically significant are shown as preferentially enriched in (C) both the LLC-NT and LLC-sh21 tumor samples, (D) LLC-NT samples alone, or (E) LLC-sh21 samples alone. Error bars represent the mean of the data ± SEM after a two-way ANOVA (C–E) (*P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001).
- Figure S6. PhenoGraph-based visualization on tSNE plots.
LLC-NT or LLC-sh21 cells were orthotopically injected into the left lung lobes of mice and established primary tumors. After 2 wk of tumor growth with no treatment, the mice were euthanized and their tumor-bearing lung lobes were isolated. Single-cell suspensions were made from tumor-bearing lung lobes or naive lungs. Naive biological replicates each contained lungs isolated from one mouse (three mice/three replicates total). LLC-NT and LLC-sh21 biological replicates each contained tumor-bearing lungs from three pooled mice (nine mice/three replicates total; three mice/replicate). Single-cell suspensions were then stained with a 39-antibody panel and analyzed on the Helios mass cytometer. Data show all viable single cells, subjected to the PhenoGraph algorithm. (A) PhenoGraph-based visualization on a tSNE plot, colored according to the expression of the cellular markers (parameters) not included in Fig S5: INOS, Tim3, CXCR3, IkBA, Arg1, NKp64, Ki67, CD103, PD-L1, IRF4, Foxp3, CD80/86, B-catenin, CD127, GITR, CD25, p-H2AX, CTLA-4, PD-1, Ly6A/E, CD117, Lag3, and ICOS.
- Figure 6. Socs1 KD tumors have an altered macrophage composition.
LLC-NT or LLC-sh21 cells were orthotopically injected into the left lung lobes of mice and established primary tumors. After 2 wk of tumor growth with no treatment, mice were euthanized and their tumor-bearing lung lobes were isolated. Single-cell suspensions were made from tumor-bearing lung lobes or naïve lungs. Naïve biological replicates each contained lungs isolated from one mouse (three mice/three replicates total). LLC-NT and LLC-sh21 biological replicates each contained tumor-bearing lungs from three pooled mice (nine mice/three replicates total; three mice/replicate). Single-cell suspensions were then stained with a 39-antibody panel and analyzed on the Helios mass cytometer. Data show all viable single cells, subjected to the PhenoGraph algorithm. (A, B) Pie charts show the relative frequency of cell clusters containing CD64+ events (a pan macrophage marker) in (A) LLC-NT or (B) LLC-sh21 tumors. Clusters are colored according to PD-L1 expression. (C–E) Pie charts are quantified based on (C) PD-L1 lo (cluster #2), (D) PD-L1 mid (cluster #7), or (E) PD-L1 high (clusters #3 and #4) expression. LLC-sh21 cells were orthotopically injected into the lungs of either WT or PD-L1−/− mice, established for 7 d, then were treated with either an isotype control antibody (IgG2a) or an anti–PD-1 antibody for 2 wk followed by terminal euthanization at 3 wk post tumor cell injection. (F) Primary tumor volume was assessed via digital calipers. Error bars represent the mean of the data ± SEM after a t test (C–E) or one-way ANOVA (F) (*P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001).
- Figure 7. Socs1 KD tumors exhibit a more T cell–inflamed phenotype and increased CD8+ T cell activation.
LLC-NT or LLC-sh21 cells were orthotopically injected into the left lung lobes of mice and established primary tumors. After 2 wk of tumor growth with no treatment, the mice were euthanized and their tumor-bearing lung lobes were isolated for either flow cytometry or FFPE and T cell staining by immunofluorescence. (A, B) Representative images of T cell staining from (A) an LLC-NT tumor or (B) an LLC-sh21 tumor (scale bars: 100 μm), showing CD8+ T cell staining. DAPI is shown in blue, CD3 in green, CD8 in red, and a merge of all channels in yellow. (C, D) Quantification of CD3+ T cells and (D) CD8+ T cells per high-power field (HPF) in LLC-NT versus LLC-sh21 tumors. There were six tumors from LLC-NT mice, and six tumors from LLC-sh21 mice. T cell numbers per HPF were averaged over four experiments (six random fields per tumor × four staining experiments = 24 fields averaged/tumor in total). Quantification of T cells was performed by two blinded observers (BB & AN). For flow cytometry, single-cell suspensions were made from tumor-bearing lung lobes. There were three experimental replicates × three tumor-bearing pooled lung lobes, for a total of nine lung lobes per the experimental conditions of “LLC-NT” or “LLC-sh21”. (E, F) For the “T Cell Phenotypic Panel,” single-cell suspensions were assessed for the following: (E) the percentage of PD-1–expressing CD8+ T cells or (F) double-positive PD-1/CD69–expressing CD8+ T cells gated as a percentage of all CD8+ T cells. For the “T Cell Stimulation Panel,” single-cell suspensions were stimulated with brefeldin A, monensin, and a cellular stimulation cocktail (PMA/Ionomycin) for 5 h to determine intracellular cytokine production at the time of harvest. Cell suspensions were either unstimulated (treated with brefeldin and monensin alone) or stimulated (treated with brefeldin, monensin, and PMA/Ionomycin). (G, H) Single-cell suspensions were assessed for the following: (G) the percentage of single-positive IFNγ-expressing CD8+ T cells or (H) double-positive IFNγ/TNFα–expressing CD8+ T cells gated as a percentage of all CD8+ T cells. Error bars represent the mean of the data ± SEM after a t test (C–F) or a two-way ANOVA (G–H) (*P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001).
- Figure S7. Socs1 KD tumors exhibit a more T-cell–inflamed phenotype and increased T-cell activation.
LLC-NT or LLC-sh21 cells were orthotopically injected into the left lung lobes of mice and established primary tumors. After 2 wk of tumor growth with no treatment, the mice were euthanized and their tumor-bearing lung lobes were isolated for either flow cytometry or FFPE and T-cell staining by immunofluorescence. (A, B) Representative images of T-cell staining from an (A) an LLC-NT tumor or (B) an LLC-sh21 tumor (scale bars: 100 μm) showing CD4+ T-cell staining. DAPI is shown in blue, CD3 in green, CD4 in red, and a Merge of all channels in yellow. (C) Quantification of CD4+ T cells and per high-power field (HPF) in LLC-NT versus LLC-sh21 tumors. (D) Gating strategy for flow cytometry (Figs 7E–H and S7E–H). Gates were determined based on isotype controls (for IFNγ and TNFα) or single stain controls for Aqua Viability, CD45, MHCII, CD8, or CD4. (E, F) Percentage of PD-1 expressing CD4+ T cells or (F) double-positive PD-1/CD69 expressing CD4+ T cells gated as a percentage of all CD4+ T cells. (G, H) Percentage of single-positive IFNγ-expressing CD4+ T cells or (H) double-positive IFNγ/TNFα–expressing CD4+ T cells gated as a percentage of all CD4+ T cells. Error bars represent the mean of the data ± SEM after a t test (C, E, F) or a two-way ANOVA (G, H) (*P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001).
- Figure 8. Socs1 KD tumors have elevated levels of Cxcl9.
LLC-NT or LLC-sh21 cells were orthotopically injected into the left lung lobe of transgenic GFP-expressing C57BL/6J mice and were grown for 3 wk with no treatment. Tumor-bearing lung lobes were isolated and made into single-cell suspensions containing both GFP-positive (host cells) and GFP-negative (cancer cells). First, RNA was isolated from identical cancer cells grown in passage (in vitro condition). Second, RNA was isolated from recovered GFP-negative cancer cells (isolated via FACS-in vivo condition). Third, RNA was run for RNA-Seq from both conditions. Both the LLC-NT and LLC-sh21 conditions had three experimental replicates per in vitro and in vivo conditions with five tumor-bearing lung lobes pooled per in vivo experimental replicate (15 mice used total). (A) In vitro and in vivo mRNA expression of (A) Cxcl9 in FPKM as assessed by RNA-Seq. The RNAScope system was used for in situ hybridization. (B) Example of positive tumor staining with a murine Cxcl9 probe and its adjacent H&E stain (scale bars: 10 μm). Red arrows denote the tumor edge, tumor middle, adjacent normal lung, or positive Cxcl9 staining. (C) Staining controls of (C) tumor tissue stained with a negative control probe, dapB, or adjacent normal lung stained with a murine Cxcl9 probe that both show negative staining (scale bars: 100 μm). (D, E) Four separate tumors from either the LLC-NT or (E) LLC-sh21 experimental conditions stained with a probe targeting murine Cxcl9 (scale bars: 100 μm). Dark brown regions represent positive Cxcl9 staining. Error bars represent the mean of the data ± SEM after a one-way ANOVA (A). (*P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001).
Source data are available for this figure.
Supplementary Materials
Source Data for Figure 2[LSA-2019-00328_SDataF2.pdf]
Source Data for Figure S2[LSA-2019-00328_SDataFS2.pdf]
Source Data for Figure 3[LSA-2019-00328_SDataF3.pdf]
Source Data for Figure 4[LSA-2019-00328_SDataF4.pdf]
Source Data for Figure S3[LSA-2019-00328_SDataFS3.pdf]
Source Data for Figure 8[LSA-2019-00328_SDataF8.xlsx]
In this Issue
Subjects
Related Articles
Cited By...
- Adaptive immunity is required for durable responses to alectinib in murine models of EML4-ALK lung cancer
- Characterization of neoantigen-specific T cells in cancer resistant to immune checkpoint therapies
- Overcoming resistance to STING agonist therapy to incite durable protective antitumor immunity
- The FOXC2 Transcription Factor Promotes Melanoma Outgrowth and Regulates Expression of Genes Associated With Drug Resistance and Interferon Responsiveness