Cell culture

Mouse lung fibroblasts CCL-206, human lung fibroblasts IMR-90 and HEK293T were purchased from ATCC (CCL-206, CCL-186 and CRL-3216, respectively). The cells were cultured to 70% confluency in Dulbecco’s modified Eagle medium (DMEM), supplemented with 10% foetal bovine serum and 1% penicillin–streptomycin. To induce senescence, cells were treated with 50 μM etoposide (Sigma, E1383) for 48 h, washed three times with PBS and cultured for additional 5–7 days in DMEM. Replicative senescence (RIS) was induced by long-term passaging of the cells in tissue culture. The cells developed senescence phenotype after 35 population doublings. On the day of the experiment, the cells were detached using trypsin.

Lentivirus production and infection

Generation of lentiviruses and their infection of cells was performed as described previously44. Lentiviruses were generated by co-transfecting HEK293T cells with 4 μg of pLX401-INK4A (AddGene, 121919) and 2 μg each of pLP/VSVG, pLP1 and pLP2 plasmids using Lipofectamine 2000 (Invitrogen, 11668019). Growth medium was exchanged the following day, and lentivirus-containing supernatant was collected 48 h later. CCL-206, IMR-90 and HEK293T cells were infected with the indicated viruses for 12 h, washed three times with PBS and cultured for an additional 24 h in RPMI medium. For selection, the cells were re-seeded in fresh RPMI medium with 1 μg ml−1 of puromycin (Gibco, A1113803) for 4–5 days. Doxycycline (Sigma-Aldrich, D3072) was added to the medium at a concentration of 5 or 10 μg ml−1 for inducible expression of the plasmid system.

siRNA

Cells were transfected overnight with 50 nM of ON-TARGETplus SMARTpool small-interfering RNA (siRNA) targeting CDKN2A (L-011007-00-0005) or with non-targeting siRNA pool (D-001810-10-20) as a control (Dharmacon). At 24 h post-transfection, the remaining adherent cells were collected.

CDK4/6 inhibitors

Abemaciclib (Pubchem, LY2835219) and palbociclib (Sigma-Aldrich, PZ0383) were dissolved in DMSO (vehicle) to yield 10 mM stock solutions and stored at 80 °C. IMR-90 cells were treated with DMEM supplemented with either 1uM abemaciclib, palbociclib or equivalent amount of DMSO for 48 h.

Proteasome inhibition

IMR-90 cells were treated with DMEM supplemented with either 10 μM MG132 (Sigma-Aldrich, M7449) or equivalent amount of DMSO for 3 h.

Immunoblot and immunoprecipitation assay

Cells were incubated in the radio-immunoprecipitation assay lysis buffer containing protease inhibitor cocktail (1:100) (Sigma-Aldrich, P8340) and phosphatase inhibitor cocktail (1:100) (Sigma, p5726) for 20 min on ice. Lysates were spun down for 15 min at 13,000 rpm and 4 °C, and protein concentrations were determined with bicinchoninic acid assay (Thermo Scientific). Equal amounts of protein were resolved by sodium dodecyl sulfate polyacrylamide gel electrophoresis and immunoblotted using β-TrCP (Cell Signalling Technology, CST-4394S), p16 (Abcam: Ab108349, human; Ab211542, mouse), vinculin (Abcam, Ab129002), PD-L1 (Cell Signalling Technology, CST-13684, human; Abcam, Ab213480, mouse) and appropriate HRP-conjugated secondary antibody, and with Europa Component Library visualization.

For immunoprecipitations analysis, cells were lysed in HNTG buffer (0.05 M HEPES pH 7.5, 10% glycerol, 0.15 M NaCl, 1% Triton X-100, 0.001 M ethylenediaminetetraacetic acid (EDTA), 0.001 M EGTA, 0.01 M NaF and 0.025 M β-glycerol phosphate) supplemented with protease inhibitor cocktail (1:100) (Sigma-Aldrich, P8340) and phosphatase inhibitor cocktail (1:100) (Sigma-Aldrich, P5726). Exaactly 2,000 μg of total cell lysates were incubated with previously coated protein A/G agarose beads (Santa Cruz Biotechnology, 2003) with anti-PD-L1 antibody (3 μg ml−1) (Cell Signalling Technology, 13684) overnight at 4 °C with gentle rotation. The beads were thoroughly washed with HNTG buffer and eluted with 6× sodium dodecyl sulfate loading buffer by boiling at 95 °C for 10 min. Ubiquitination of PD-L1 was measured by immunoblotting with anti-ubiquitin antibody (Santa Cruz Biotechnology, 8017).

CHX chase assay

Twenty-four hours following transfection with siRNA, senescent IMR-90 cells were treated with 200 μM CHX (Sigma-Aldrich, C4859) for 3 h. Cells were lysed by incubation with 100 μl of radio-immunoprecipitation assay buffer (supplemented with PMSF and protease inhibitor cocktail) for 20 min on ice and protein concentrations were determined using bicinchoninic acid assay. A total of 10 μg of protein lysate was used to measure levels of PD-L1 by the enzyme-linked immunosorbent assay (ELISA), using the PD-L1/B7-H1 Quantikine ELISA Immunoassay kit (R&D, DB7H10), according to the manufacturer’s protocol. The optical density of each well was measured with the Infinite 200 plate reader (Tecan) at 450 nm with wavelength correction set at 540 nm. The experiment was performed twice and each sample was performed in duplicate.

Immunofluorescent staining of human tissue microarray

Formalin-fixed, paraffin-embedded sections of human lung tissue microarray (US Biomax, LC487) were incubated at 60 °C for 60 min, deparaffinized and incubated in acetone for 7 min at −20 °C, followed by subsequent incubation with 3% H2O2 for 15 min at room temperature to block endogenous peroxidase activity. Antigen retrieval was performed in a microwave (3 min at full power, 1,000 W, then 20 min at 20% of full power) in Tris-EDTA buffer (pH 9.0). The slide was blocked with 20% NHS and 0.5% Triton in PBS and primary antibodies were diluted in 2% NHS and 0.5% Triton in PBS (p16, 1:30—Abcam, Ab108349; PD-L1, 1:100—Abcam, Ab213524) in a multiplexed manner with the Opal reagents (Akoya Bioscience), each one overnight at 4 °C. Following overnight incubation with the first primary antibody, the slide was washed with PBS, incubated in 2% NHS in PBS with secondary antibody conjugated to HRP (1:100) for 90 min, washed again and incubated with Opal reagents for 15 min. The slide was then washed and microwaved (as described above), washed, stained with the next primary antibody and with DAPI at the end of the cycle, and mounted. We used the following staining sequence: p16 → PD-L1 → DAPI. Each antibody was validated separately, and then multiplexed immunofluorescence was optimized to confirm that the antibody signal was not lost or changed due to the multistep protocol. Slides were imaged with an Eclipse Ni-U microscope (Nikon), connected to a colour camera (DS-Ri1, Nikon, ×20), and DAPI, Cy3 and Cy5 cubes. Images were analysed using the Fiji v2.6.0 software. The QuePath v0.4.4 software was used for the identification and quantification of cells positive for the fluorescent signal of each marker.

Mice

Female C57BL/6 mice 10–14 weeks of age (young) or 24 months old (old) were used in all experiments. Mice were purchased from Harlan Laboratories. All mice were housed and maintained under specific pathogen-free conditions at the Weizmann Institute of Science in accordance with national animal care guidelines. The housing conditions were: 12-h dark/light cycle (lights on at 8:00), 22 °C temperature and 30–70% humidity. All procedures were performed in accordance with the protocols approved by the Weizmann Institute Animal Care and Use Committee (03320423-1, 06900820-2, 02720418-2, 05410621-3, 04000523-1 and 04040523-2).

LPS exposure and treatment

For chronic LPS exposure, mice were exposed to an aerosolized PBS alone or PBS containing Escherichia coli LPS (0.5 mg ml−1; Sigma-Aldrich, L2630) for 30 min, three times a week for 10 weeks, in a custom-built cylindrical chamber as described previously14. For short-term 5-day LPS exposure, mice were exposed as in chronic exposure, but only for 5 constitutive days. Mice were killed and lungs were collected 48 h after the last exposure.

For the immune checkpoint blockade treatment mice received intravenous injection of 200 μg anti-PD-L1 (Ichorbio, ICH1086), 200 μg anti-PD1 (Ichorbio, ICH1091) or 200 μg isotype control IgG2b (Ichorbio, ICH2243).

In a short-term 5-day LPS exposure, mice were treated with immune checkpoint blockade on the second and fifth day of LPS inhalation. Old mice and mice undergoing chronic LPS exposure were treated with immune checkpoint blockade in five doses within 3 weeks, and the mice were euthanized 2 days after the final injection.

Bronchoalveolar lavage (BAL) fluid was collected from perfused lungs by double washing with 1 ml PBS through a tracheal catheter as previously described14.

Measurement of cytokines levels in plasma

Blood was taken from the mice through cardiac puncture. To obtain plasma, blood samples were diluted 1:1 with PBS containing 1 mM EDTA upon the collection and then centrifuged at 3,400g for 15 min at 4 °C. Plasma levels of cytokines were measured by Milliplex MAP Mouse High Sensitivity T Cell Panel (cat no. MHSTCMAG-70K; Millipore) on Luminex (MAGPIX) following the manufacturer’s instructions. All samples were assayed in duplicate and mean values analysed. BELYSA v1.2 software (Millipore) was applied for data analysis. Concentrations are reported in pg ml−1.

Epigenetic age predictions

Genomic DNA was isolated from whole blood using a Qiagen QIAamp DNA Mini and Blood Mini Kit (Qiagen) and DNA concentrations were measured with a NanoDrop 2000 spectrophotometer (Thermo Scientific). A total of 500 ng of genomic DNA was bisulfite converted with the Zymo Research Group EZ DNA Methylation Kit (Zymo Research). Pyrosequencing was performed with the PyroMark Q48 Autoprep system (Qiagen), and DNA methylation percentages were obtained for age-related CpGs in Aspa, Wnt3a, Prima1 and Hsf4 for the samples from young and old mice treated with anti-PD-L1 or matched isotype control. Primers, PCR conditions and targeted epigenetic age calculations were used, as described before45.

Tissue dissociation

To achieve single-cell suspension from the lung, mice were euthanized by administration of xylazine/ketamine and then perfused by injecting cold PBS via the right ventricle before lung dissection. Lung tissues were dissected from mice, cut into small fragments and suspended in 1.5 ml of DMEM/F12 (Invitrogen, 11330-032) containing elastase (3 U ml−1, Worthington, LS002279), collagenase type IV (1 mg ml−1, Thermo Scientific, 17104019) and DNase I (0.5 mg ml−1, Roche, 10104159001) and incubated at 37 °C for 20 min with frequent agitation. After dissociation procedure, cells were washed with an equal volume of DMEM/F12 supplemented with 10% foetal bovine serum and 1% penicillin–streptomycin (Thermo Scientific), filtered through a 100-μm cell strainer and centrifuged at 380g for 5 min at 4 °C. Pelleted cells were resuspended in red blood cell ACK lysis buffer (Gibco, A1049201), incubated for 2 min at 25 °C, centrifuged at 380g for 5 min at 4 °C and then resuspended in ice-cold fluorescence-activated cell sorting (FACS) buffer (PBS supplemented with 2 mM EDTA, pH 8 and 0.5% bovine serum albumin).

Flow cytometry

IMR-90 cells were stained with Zombie Aqua Viability fixable stain (423101) or Sytox Blue (Invitrogen, 34857) for evaluation of live/dead cells, followed by antibody Brilliant Violet 711-PD-L1 (329721) or isotype control (400353) staining (all from BioLegend).

Lung single-cell suspension was stained with anti-mouse CD16/32 (eBioscience, 14-0161-82) to block Fc receptors before labelling with fluorescent antibodies against cell-surface epitopes. For samples that were used for p16 intracellular staining, we used the following antibodies for extracellular staining: Brilliant Violet 605-CD45 (103140), FITC-CD11c (117306) and Brilliant Violet 421-SiglecF (155509) purchased from BioLegend. We used two clones of PD-L1 antibody (either Brilliant Violet 785-PD-L1, 124331, or PE-Cy5-PD-L1, 124344, with both cloning 10F.9G2, and PE-PD-L1, 153611, which clones MIH6) purchased from BioLegend, which yielded similar results. Then cells were fixed with 90% methanol for 10 min at 4 °C. All centrifugation steps after fixation were done at 850g for 5 min at 4 °C. For intracellular staining, cells were stained with p16 antibody (Abcam, Ab54210) conjugated to Alexa Fluor 647 fluorophore (Thermo Scientific, A20186). Cells were stained with Zombie Aqua Viability fixable stain for evaluation of live/dead cells. For characterization of immune subsets in BAL, we used the following antibodies: pacific blue-CD69 (104523), Brilliant Violet 605-ICOS (313537), Brilliant Violet 785-NK1.1 (108749), PerCP-CD19 (115531), FITC-CD3 (100204), PE-CD25 (102007), PE-Dazzle 595-TIGIT (142109), PE-Cy5-CD8 (100709), PE-Cy7-CTLA4 (106313), APC-LAG3 (125209), Spark Nir 685-CD4 (100475), Alexa Fluor 700-CD44 (103025), APC/Cy7-PD1 (135223) and APC Fire810-CD45 (103173). All antibodies were purchased from BioLegend and diluted 1:100 in FACS buffer before staining. Cell populations were recorded using LSR-II new (BD Biosciences) or Aurora (Cytec) and analysed using FlowJo v10 software (BD Biosciences) and Prism v7 software.

For imaging, flow cytometry cells were stained with FITC-CD45 (BioLegend, 103107), Brilliant Violet 786-PD-L1 (BioLegend, 124331) and Ax647-p16 (Abcam, Ab54210, conjugated to Alexa Fluor 647 fluorophore from Thermo Scientific, A20186). Before acquisition, cells were stained with DAPI and filtered through a 100 µm membrane. All antibodies were diluted 1:100 in FACS buffer before staining. Cells were acquired using ImageStreamX mark II (Amnis, part of EMD Milipore Merck) and image data analysis was performed using IDEAS v6.2 software as described in previously6.

Mass cytometry

All antibodies used in the study, their corresponding clone, provider and catalogue number are listed in Supplementary Table 1. Antibodies were obtained in protein-free buffer. Custom metal-conjugated antibodies were generated using MaxPAR antibody labelling kits (Fluidigm) or the MIBItag Conjugation Kit (IONpath) according to the manufacturer’s instructions. After metal conjugation, the concentration of each antibody was determined with a Nanodrop (Thermo Scientific) and adjusted to 0.5 mg ml−1 with Antibody Stabilizer PBS (CANDOR Bioscience, 131050) for long-term storage at 4 °C. Lung single-cell suspension was washed once in 1 ml of cell staining buffer (CSB) (Fluidigm, 201068). To ensure homogeneous staining, 4 × 106 cells from each sample were used. For viability staining, cells were incubated with 1.25 μM Cell-ID Cisplatin (Fluidigm, 201064) for 3 min before quenching with CSB. Before antibody staining, cells were incubated for 10 min at 4 °C with anti-mouse CD16/32 (Invitrogen, 14-0161-82) to block Fc receptors. Cells were stained with the epithelial or immune-centric antibodies for 45 min at 4 °C. An antibody cocktail of extracellular markers was prepared as a master mix and 50 μl of the cocktail was added to the samples resuspended in 50 μl of CSB. Cells were washed twice with CSB and permeabilized with fixation/permeabilization buffer (eBioscience, 88-8824-00). Then, samples were washed twice with CSB, incubated with 5% goat serum (Sigma-Aldrich, G-9023) and resuspended in 50 μl of CSB before the addition of 50 μl of cocktail of intracellular antibodies. For DNA-based detection, cells were stained with 125 nM Cell-ID Intercalator-Ir (Fluidigm, 201192 A) in PBS with 1.6% paraformaldehyde (Electron Microscopy Sciences, 15700) overnight at 4 °C. Cells were then washed once in CSB and twice in Maxpar Water (Fluidigm, 201069). For mass cytometry acquisition, samples were diluted to 3 × 105 cells ml−1 in Maxpar Water containing 1:10 EQ Four Element Calibration Beads (Fluidigm, 201078) and filtered through a 35-μm filtermesh tube (Falcon). For acquisition CyTOF Helios system (Fluidigm) was used and samples were acquired at the rate of 200 events s−1. Data were collected as.fcs files. Data were normalized and concatented when necessary, via the CyTOF software v7.0 (Fluidigm). Then, the Cytobank platform (Beckman Coulter) was used to gate out the normalization beads according to the 140Ce channel. Next, several gates were applied to gate out live cells for further analysis. First, live single cells were gated using the cisplatin 195Pt, iridium DNA label in 193Ir, followed by the event length, and the Gaussian parameters of width, centre, offset and residual channels. To normalize data, a hyperbolic arcsine transformation (with a scale factor of five) was first applied. FlowSOM k-NN clustering and two-dimensional viSNE projections were calculated using Cytobank v9.0 software. Subsequently, mass cytometry data were analysed in Matematica (v14.0) and all custom-generated code is available in the AlonLabWIS git.

Flow cytometry cell sorting

Cell populations were sorted using BD FACS Aria Fusion flow cytometer (BD Biosciences). Before sorting, all samples were filtered through a 70 mm nylon mesh. Populations that were sorted were epithelial cells (EpCam+CD31−CD45−Ter119−), AMs (CD45+CD11c+SiglecF+p16 high/low) and CD8a+ T cells (CD45+SiglecF−CD3+CD8a+). Sytox Blue (Invitrogen, 34857) or Aqua Zombie was used for viability staining. A range of 5,000–10,000 live cells were sorted into a low-bind eppendorf tube containing 50 µl of lysis/binding buffer (Invitrogen). Immediately after sorting, samples were spun down, snap frozen and stored at −80 °C until further processing.

To sort out lung epithelium, cells were stained with following antibodies: Brilliant Violet BV605-CD31 (102427), PE-CD45 (103106) and Alexa Fluor 488-EpCam (118210) all purchased from BioLegend and eFluor450-TER-119 (eBioscience, 48-5921-82). To sort out AMs, cells were stained with Brilliant Violet 605-CD45 (103140), FITC-CD11c (117306), Brilliant Violet 421-SiglecF (155509) and Brilliant Violet 786-PD-L1 (124331) all from BioLegend and p16 (Abcam, Ab54210) conjugated to Alexa Fluor 647 fluorophore (Thermo Scientific, A20186).

To sort out CD8a+ T cells, cells were stained with Brilliant Violet 605-CD45 (103140), Brilliant Violet 421-SiglecF (155509), FITC-CD3 (100204) and APC-CD8a (100711). All antibodies were purchased from BioLegend and diluted 1:100 in FACS buffer before staining.

Preparation of libraries for RNA-seq

A total of 1 × 104 cells of lung epithelium (EpCam+CD31−CD45−) were sorted into 50 μl of lysis/binding buffer (Life Technologies). mRNA was captured with 15 μl of Dynabeads oligo(dT) (Life Technologies), washed and eluted at 70 °C with 6.1 μl of 10 mM Tris-Cl (pH 7.5). Complementary DNA libraries were prepared from pooled samples of the same cell type (10,000 cells per sample) according to a bulk variation of massively parallel RNA single-cell sequencing (MARS-seq)46 and were sequenced on Illumina NextSeq 500 (Illumina). Intracellular staining and sequencing (INs-seq) libraries were prepared as previously described27, followed by bulk MARS-seq.

RNA-seq analysis

Raw data were processed with the User-friendly Transcriptome Analysis Pipeline47. Only reads with unique mapping to the 3′ of RefSeq annotated genes (mm10, NCBI Mus musculus annotation release 109) were analysed. For differential gene expression analysis, we used DESeq2 (ref. 48), following standard workflow, to analyse RNA-seq count data derived from lung epithelial cells, comparing LPS to PBS. Genes with <30 unique molecular indentifiers across samples were pre-filtered. Differentially expressed genes were selected to have fold change > 1.25 and Benjamini–Hochberg-adjusted P < 0.05. For GSEA, we used DESeq2 (ref. 38) to derive gene fold changes for LPS versus PBS epithelial cells and for p16+ versus p16− macrophages, controlling for treatment (LPS/PBS) as a covariant. We then applied GSEA to the ranked fold changes. We used Fast Gene Set Enrichment Analysis (‘fgsea’) library49 implemented in R to test for enrichment of gene sets (no. genes >10) from the mouse C5 v5p2 Gene Ontology collection of the Molecular Signature Database50.

Real-time PCR analysis

mRNA was extracted from 5,000 CD8a+ T cells sorted into 50 µl of lysis/binding buffer (Invitrogen) and captured using Dynabeads oligo(dT) (Invitrogen) kit according to the manufacturer’s protocols. For lung tissue total RNA was extracted using Qiagen kit. For quantitative PCR analysis, mRNA was reverse transcribed using SuperScript II (Invitrogen, 11904018) and cDNA was diluted 1:10 for quantitative PCR analysis performed using the SYBR Green system. The relative gene expression was determined using the ΔΔCt method and normalization to Actb. We used four biological replicates for each condition. One-tailed t-tests were used to perform statistical analysis.

The following primers were used: mouse Actb—forward, 5′-GGAGGGGGTTGAGGTGTT-3′, reverse,

5′- TGTGCACTTTTATTGGTCTCAAG-3′; Ifng—forward,

5′- TGAACGCTACACACTGCATCTTGG-3′, reverse, 5′-CGACTCCTTTTCCGCTTCCTGAG-3′; p16—forward, 5′-TTGGGCGGGCACTGAATCTC-3′, reverse,

5′-AGTCTGTCTGCAGCGGACTC-3′; p19—forward, 5′-GCCGCACCGGAATCCT-3′; reverse,

5′- TTGAGCAGAAGAGCTGCTACGT-3′; p21—forward, 5′-GACAAGAGGCCCAGTACTTC-3′; reverse, 5′-GCTTGGAGTGATAGAAATCTGTC-3′; Il-1b—forward,

5′-GGAGAACCAAGCAACGACAAAATA-3′;reverse, 5′-TGGGGAACTCTGCAGACTCAAAC-3′;

Tnf—forward, 5′-CCACGCTCTTCTGTCTACTG-3′; reverse, 5′-GATGAGAGGGAGGCCATTTG-3′; Il-6—forward, 5′-TAGTCCTTCCTACCCCAATTTCC-3′; reverse,

5′-TTGGTCCTTAGCCACTCCTTC-3′; CXCL10—forward,

5′-CCATCAGCACCATGAACC-3′; reverse, 5′-TCCGGATTCAGACATCTC-3′; and HPRT—forward,

5′-TGACACTGGCAAAACAATGCA-3′; reverse, 5′-GGTCCTTTTCACCAGCAAGCT-3′.

Statistics and reproducibility

For mice experiments, no statistical method was used to pre-determine sample sizes. In each experiment, the number of animals was chosen to have sufficient statistical power on the basis of the literature and experience6,7,15. For cell culture experiments, the sample size was determined to be at least n = 3 independent biological repeats, while in each experiment every sample had three technical repeats. Data are presented as means ± standard error of the mean (s.e.m.), unless otherwise noted. Comparisons between two groups were performed by an unpaired two-tailed Student’s t-test, unless otherwise noted. Comparisons between the three groups were performed by one-way analysis of variance (ANOVA). A chi-squared test was performed for RNA-seq analysis of differentially expressed genes. For consistency in comparisons, significance in all figures is denoted as follows: *P < 0.05, **P < 0.01, ***P < 0.001 and ****P < 0.0001.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.