A genome-wide CRISPR screen identifies WDFY3 as a regulator of macrophage efferocytosis

Abstract

Phagocytic clearance of dying cells, termed efferocytosis, is essential for maintaining tissue homeostasis, yet our understanding of efferocytosis regulation remains incomplete. Here we perform a FACS-based, genome-wide CRISPR knockout screen in primary mouse macrophages to search for novel regulators of efferocytosis. The results show that Wdfy3 knockout in macrophages specifically impairs uptake, but not binding, of apoptotic cells due to defective actin disassembly. Additionally, WDFY3 interacts with GABARAP, thus facilitating LC3 lipidation and subsequent lysosomal acidification to permit the degradation of apoptotic cell components. Mechanistically, while the C-terminus of WDFY3 is sufficient to rescue the impaired degradation induced by Wdfy3 knockout, full-length WDFY3 is required to reconstitute the uptake of apoptotic cells. Finally, WDFY3 is also required for efficient efferocytosis in vivo in mice and in vitro in primary human macrophages. This work thus expands our knowledge of the mechanisms of macrophage efferocytosis, as well as supports genome-wide CRISPR screen as a platform for interrogating complex functional phenotypes in primary macrophages.

Fig. 1. A pooled, fluorescence-activated cell sorting (FACS)-based genome-wide CRISPR knockout screen in primary mouse macrophages identified known and novel regulators of macrophage efferocytosis.

a Schematics of the CRISPR screen workflow. b Timeline of bone marrow isolation, lentiviral library transduction, puromycin selection, efferocytosis, and cell sorting. c Visualization of gating strategy for separation of non-eaters and efficient eaters. Successful separation was confirmed by fluorescent microscopy. d Volcano plot highlights the top-ranked screen hits that are known positive and negative regulators of macrophage efferocytosis. e Canonical pathways enriched in top-ranked positive regulators by Ingenuity Pathway Analysis (IPA). f Canonical pathways enriched in top-ranked negative regulators by IPA. g Validation of Wdfy3 as a positive regulator required for macrophage efferocytosis (n = 4 independent experiments). Data are presented as mean ± SEM. Two-sided P values were determined by a two-way ANOVA with Tukey’s multiple comparisons test in panel g. ACs, apoptotic cells; BMDMs, bone-marrow-derived macrophages.

Fig. 2. WDFY3 deficiency led to impaired uptake, as opposed to binding, of apoptotic cells (ACs) due to defective actin depolymerization.

a Schematics of breeding LysMCre mice with Wdfy3^fl/fl mice to obtain mice with myeloid-specific knockout of Wdfy3. b Validation of efficient knockout in BMDMs by western blot of WDFY3 (n = 4 biological replicates; the blot shown is a representative image of three independent experiments). c Cre^- and Cre⁺ BMDMs were incubated with Hoechst-labeled ACs at various AC: BMDM ratios of 3:1, 5:1, 10:1 respectively for 1 h and analyzed by flow cytometry (n = 3 biological replicates, each from the average of 2 technical replicates). d Cre^- and Cre⁺ BMDMs were incubated with PKH26-labeled ACs at various time points of 15 min, 30 min, and 60 min at a AC: BMDM ratio of 5:1 and analyzed by flow cytometry (n = 3 technical replicates). e Cre^- and Cre⁺ BMDMs were pretreated with cytochalasin D for 30 min to block polymerization and elongation of actin, thus testing the binding of ACs with BMDMs. The treated BMDMs were then incubated with TAMRA-stained apoptotic mouse thymocytes at 37 °C for 30 min and then extensively washed with DPBS to remove unbound ACs for imaging and quantification after fixation (n = 6 biological replicates). f Cre^- and Cre⁺ BMDMs were stained with CellTracker and incubated with ACs. Efferocytosis of ACs by BMDMs were observed using time-lapse confocal microscopy. The time required for phagosome formation was recorded and quantified (n  = 44 and 47 data points for Cre- and Cre+ respectively, each data point represents one BMDM with engulfed ACs, 4 biological replicates for each genotype). The arrows point to the BMDM engulfing an AC across the stages from phagocytic cup formation to phagosome closure (from left to right). g F-actin labeled by siR-actin in Cre^- and Cre⁺ BMDMs was quantified by flow cytometry (n = 4 biological replicates, each from the average of 3 technical replicates). h BMDMs were stained with CellTracker and siR-actin, then incubated with NuclearMask Blue-labeled apoptotic Jurkat cells for various time points (10 min, 20 min, 40 min, and 60 min). For each time point, unbound ACs were removed and BMDMs were fixed. BMDMs were imaged and the percentage of BMDMs with engulfed cargos surrounded by F-actin rings in all BMDMs with engulfed cargos was quantified (n = 4 biological replicates, data are representative of two independent experiments). Data are presented as mean ± SEM. Two-sided P values were determined by a two-way ANOVA with Tukey’s multiple comparisons test in (c, d, e, g, h), or by unpaired t test in panel f.

Fig. 3. WDFY3 deficiency led to defects in LC3-associated phagocytosis (LAP) and the degradation of engulfed ACs.

a Cre⁻ and Cre⁺ BMDMs were incubated with PKH26-labeled ACs for 1 h. After washing away the unengulfed ACs, BMDMs were placed back to the incubator for another 3 h. BMDMs were then fixed and imaged. The percentage of BMDMs showing non-fragmented PKH26 signals in the total number of PKH26⁺ BMDMs was quantified (n = 4 and 5 biological replicates for Cre− and Cre+ respectively, each from the average of 3 technical replicates). b Cre⁻ and Cre⁺ BMDMs were incubated with TAMRA-labeled ACs for 1 h. After washing away the unengulfed ACs, BMDMs were either collected for flow cytometry to quantify the MFI of TAMRA or placed back to the incubator for another 16 h and then collected for flow cytometry. The rate of degradation was calculated, as shown in the schematics (n = 8 and 9 biological replicates for Cre− and Cre+ respectively). c Cre⁻ and Cre⁺ BMDMs were incubated with ACs labeled by Hoechst, which stains DNA and is pH-insensitive, and pHrodo, which is pH-sensitive and shows fluorescent signal only under an acidified environment in the phagolysosome. The percentage of Hoechst⁺ BMDMs indicates uptake. The percentage of Hoechst⁺/pHrodo⁺ BMDMs in Hoechst⁺ BMDMs indicates acidification of the engulfed cargos (n = 8 biological replicates, each from the average of 2 technical replicates). d Schematics of known functional domains and binding partners of human WDFY3. e The interaction between WDFY3 and GABARAP was assessed by co-immunoprecipitation. Cre⁻ and Cre⁺ BMDM cell lysates were incubated with anti-GABARAP antibody and protein A/G agarose beads. Beads-bound proteins were detected with anti-WDFY3 antibodies (n = 3 independent experiments with similar results). HC refers to heavy-chain. f Cre⁻ and Cre⁺ BMDMs were incubated with ACs for 1 h. Unbound ACs were washed away and BMDMs were collected for measurement of LC3-II by western blot (n = 5 biological replicates. The image shows the representative blot. * denotes non-specific band). g BMDMs were incubated with Hoechst-labeled ACs to allow efferocytosis. After removal of unbound ACs, BMDMs were collected and treated with digitonin to remove non-membrane bound LC3, and then immunostained for LC3 that is lipidated and membrane-bound. LC3-II staining was then quantified by flow cytometry for BMDMs that had engulfed Hoechst-labeled ACs (n = 5 biological replicates). Data are presented as mean ± SEM in (b, c, f), or as median ± 95% CI in (a, g). Two-sided P values were determined by a two-way ANOVA with Tukey’s multiple comparisons test in (b, c, f), or by Mann–Whitney test in (a, g).

Fig. 4. The C-terminal WDFY3 is sufficient for regulating degradation yet the full-length WDFY3 is required for the uptake of ACs during efferocytosis.

a Schematics of lentiviral overexpression of C-terminal WDFY3 in BMDMs of Cre^- and Cre⁺ mice. b C-terminal WDFY3 did not restore uptake, yet partially rescued the defects in cargo acidification in Cre⁺ mice (n = 7 biological replicates, each from the average of 2 technical replicates). c C-terminal WDFY3 restored LC3-II levels in Cre⁺ mice as determined by western blot (n = 5 biological replicates. * indicates non-specific bands) and in (d) by flow cytometry (n = 3 biological replicates). e BMDMs from GFP-LC3 mice were transfected with tdTomato-fused C-terminal WDFY3_(2981-3526) plasmid via electroporation. BMDMs were fed with Hoechst-labeled ACs. Unengulfed ACs were washed away and BMDMs were imaged to visualize GFP-LC3 phagosome association, C-WDFY3 intracellular localization, and GFP-LC3/tdTomato-WDFY3 colocalization with and without AC engulfment (n = 21 cells from 5 biological replicates. Images are representatives of 5 independent experiments each using one GPF-LC3 mouse). Data are presented as mean ± SEM. Two-sided P values were determined by a two-way ANOVA with Tukey’s multiple comparisons test in (b–d) or by unpaired t test in (e).

Fig. 5. Mice with myeloid Wdfy3 knockout show impaired efferocytosis in vivo.

a Schematics of experimental design for in vivo thymus efferocytosis assay. b Thymus weight. c Total number of cells per thymus. d Percentage of F4/80⁺ macrophages in the thymus determined by flow cytometry. e Percentage of Annexin V⁺ ACs per thymus determined by flow cytometry. A higher percentage implies impaired efferocytic clearance. f Thymic sections were stained with TUNEL for ACs, and CD68 for macrophages. The ratio of macrophage-associated TUNEL⁺ cells vs. free TUNEL⁺ cells was quantified and summarized. The white and yellow squares highlight the macrophage-associated and free TUNEL⁺ cells, respectively (n = 5 biological replicates). g Schematics of experimental design for in vivo peritoneal macrophage efferocytosis assay. h Peritoneal exudates were stained for F4/80 and the percentage of TAMRA⁺ peritoneal macrophages was determined by flow cytometry (n = 5 biological replicates). n = 9 Cre^- and 9 Cre⁺ biological replicates for PBS group, n = 10 Cre⁻ and 14 Cre⁺ biological replicates for Dexamethasone group in (b, c, e). n = 8 Cre⁻ and 7 Cre⁺ biological replicates for PBS group, n = 9 Cre^- and 12 Cre⁺ biological replicates for Dexamethasone group in (d). Data are presented as mean ± SEM in (b–e), as median ± 95% CI in (f) and (h). Two-sided P values were determined by a two-way ANOVA with Tukey’s multiple comparisons test in (b–e), or by Mann–Whitney test in (f) and (h).

Fig. 6. WDFY3 regulates efferocytosis in human macrophages.

a Schematics of human monocyte differentiation to macrophages (HMDMs) and knockdown of WDFY3 with Lipofectamine RNAiMAX-mediated transfection of siRNAs targeting WDFY3, or non-targeting siRNAs as the control. b Validation of knockdown efficiency at mRNA level by qRT-PCR (n = 4 independent experiments, each from the average of 3 technical replicates). c Validation of knockdown efficiency at protein level by western blot (n = 2 biological replicates, data are representative of 3 independent experiments). d Efferocytosis of apoptotic Jurkat cells labeled by both Hoechst and pHrodo. The percentage of HMDMs with Hoechst-labeled ACs (indicating uptake), and the percentage of Hoechst⁺/pHrodo⁺ HMDMs in Hoechst⁺ HMDMs (indicating acidification upon uptake) were quantified by flow cytometry. Both uptake and acidification of ACs were impaired in HMDMs with siRNA-mediated WDFY3 knockdown (n = 4 independent experiments, each from the average of 2 technical replicates). e Fragmentation of engulfed ACs was assessed 3 h after washing away the unengulfed ACs. The percentage of HMDMs with non-fragmented PKH26 staining in all PKH26⁺ HMDMs was determined (n = 4 independent experiments). f Flow cytometry-based degradation assay was performed in HMDMs with procedures as described in Fig. 3b (n = 5 independent experiments). g RNA-seq was performed for HMDMs either unstimulated (M0) or treated with LPS and IFNγ for 18-20 h (M1-like). The expression of WDFY3 was visualized (n = 48 biological replicates). The box shows Q1, median, and Q3; the whiskers show 1.5 x interquartile range, though the lower whiskers in this plot only extend to the minimum as the minimum values are greater than the values corresponding to the lower whiskers. There is one outlier in M0 with FPKM 7.404. There are two outliers in M1-like with FPKMs 5.943 and 4.345. Data are presented as median ± 95% CI. Two-sided P values were determined by Mann–Whitney test.

Fig. 7. Schematic figure summarizing how WDFY3 regulates macrophage efferocytosis.

WDFY3 is discovered as a new regulator of efferocytosis by macrophages. WDFY3 deficiency in macrophages specifically impaired uptake, not binding, of apoptotic cells due to defective actin depolymerization, thus phagosome formation. WDFY3 directly interacts with GABARAP, one of the seven members of the LC3/GABARAP protein family, to facilitate LC3 lipidation and the subsequent phagosome-lysosome fusion and degradation of the engulfed AC components.