破骨细胞与硬脑膜转移中的淋巴引流障碍

Fig. 1 | Distinct skull erosion and impaired MLVs in DM. a, Cranium CT from patients with meningioma (n = 14 patients), breast cancer BM (BRCA BM, n = 10 patients), breast cancer LM (BRCA LM, n = 7 patients) and breast cancer DM (BRCA DM, n = 8 patients). Skull thickening and thinning are indicated by green and pink arrowheads, respectively. b, Ratios of local thickened, preserved or local thinned skull in areas related to those in a and Extended Data Fig. 1d. c, Screening strategy for DM. d, 3D reconstruction and two-dimensional (2D) sections of CT scans showing cranium from 4T1-DM mice. Pink arrowheads indicate the same site of skull bone thinning in both 3D and 2D images. e, Statistics of skull thickness related to images in d and Extended Data Fig. 1g; n = 5 mice per group. L, left; R, right; Rem., remote to lesion; Adj., adjacent to lesion. f, Dura from non-tumour and 4T1-DM mice. Stained for LYVE1, CD31 and 4′,6-diamidino-2-phenylindole (DAPI). tdTom, tdTomato+ tumour cells. Dilations, blebs and local injuries were marked with white, magenta and yellow rectangles or arrowheads, respectively, and the boxed areas are shown magnified on the right of the images (Zoom). g–i, Statistics of diameters (g), blebs (h) and local injuries (i); n = 5 mice per group. j, Experimental design of MLV drainage evaluation. k, Whole-mount images of dCLNs from j. Blue, DAPI; BF, bright field; green, tracers; Red, tdTomato+ tumour cells; n = 5 mice per group. OVA, OVA-AF488. l, Statistics from k. Data were normalized to non-tumour controls; n = 5 mice per group. Data represent means ± s.e.m. Statistics were performed using one-way analysis of variance (ANOVA) with Tukey’s multiple comparison test (e,g,h,i) or two-sided unpaired t-test (l). NS, not significant. Scale bars, 100 μm (f, large images), 25 μm (f, inset) and 2 mm (k).

Fig. 2 | scRNA-seq analysis reveals T cell exhaustion in DM. a, Experimental set-up of scRNA-seq of dura maters from non-tumour, 4T1-LM and 4T1-DM mice; n = 3 samples per group (each sample contains mixture of dura maters from six mice). b, Uniform manifold approximation and projection (UMAP) visualization of colour-coded clusters based on manually annotated cell types and cells aggregated from samples in 4T1-DM mice. c, Proportions of immune cell populations of dura maters from scRNA-seq data of non-tumour (Extended Data Fig. 3a), 4T1-LM (Extended Data Fig. 3c) or 4T1-DM mice (from b). d, Dot plot depicting the expression of marker genes across all clusters from scRNA- seq data of DM cells in b. e, UMAP of re-clustered T+NK cells from dura maters of non-tumour (Extended Data Fig. 3a), 4T1-LM (Extended Data Fig. 3c) and 4T1-DM (from b) mice. f, Proportions of T+NK cell subsets from e. g, UMAP of exhaustion score (based on Pdcd1, Ctla4, Tigit, Lag3 and Havcr2) in T cells of non-tumour, 4T1-LM and 4T1-DM mice. h,j, Flow cytometry of CD8+ (h) and CD4+ (j) T cells in non-tumour dura and dura maters from 4T1-DM mice, showing the PD1- and TIM3-expressing cells. PE, phycoerythrin; BV421, brilliant violet (BV) 421. i,k, Statistics of data from h (i) and j (k); n = 5 mice per group, from three independent experiments. l,n, Flow cytometry of CD8+ (l) and CD4+ (n) T cells in non-tumour dura and dura maters from 4T1-DM mice, showing TNF- and IFNγ- expressing cells. m,o, Statistics of data from l (m) and n (o); n = 5 mice per group, from three independent experiments. Data represent means ± s.e.m. Two-sided unpaired t-test (i,k,m,o).

Fig. 3 | Targeting OCs rescues MLV drainage and decreases dural T cell exhaustion. a, OC score (Ctsk, Acp5, Nfatc1, Mmp9 and Atp6v0d2) of Fig. 2b. b, TRAP staining of coronal head sections. Statistics of TRAP+ cells per field; n = 5 mice per group. c, Experimental set-up of d–n. d, Dura from c. The boxed areas in the left column are shown enlarged in the central and right columns. Dilations, blebs and local injuries are marked with white, magenta and yellow rectangles or arrowheads, respectively. e, Statistics of LYVE1+ area from d; n = 5 mice per group. f–h, Statistics of diameter (f), bleb (g) and local injuries (h) from d; n = 5 mice per group. i, The dCLNs from mice in c. Stained as in Fig. 1k. j, Statistics of tracer signals from i. Data were normalized to non-tumour mice; n = 5 mice per group. k,l, Statistics of PD1+TIM3+ in dural CD8+ (k) and CD4+ (l) T cells of c; n = 5 mice per group, from three independent experiments. Related to Extended Data Fig. 8f. m, Left, tumour growth from mice in c; right, BLI images; n = 5 mice per group, from three independent experiments. sr, steradian. n, Survival of mice in c; n = 10 mice per group, from two independent experiments. o, Experimental set-up of p and q. Ab, antibody; IVIS, in vivo imaging system. p, Left, tumour growth in o; right, BLI image, n = 5 mice per group, from three independent experiments. q, Survival of mice in o; n = 10 per group, from two independent experiments. Data represent means ± s.e.m. Two-sided unpaired t-test (e–h,k–m), one-way ANOVA with Tukey’s multiple comparison test (b,j,p) and log-rank test (n,q). Scale bars, 100 μm (b), 500 μm (d, left column), 200 μm (d, middle column) and 2 mm (i).

Fig. 4 | Dynamic tracing reveals two distinctive and inter-related subsets of OCs in DM. a, TRAP staining (red) of coronal head section in 4T1-DM model, from three independent experiments. Top, DM section; middle and bottom, expansion of areas boxed in grey and green in top panel, showing OCs and ITCs, respectively. b, Images showing 3D reconstruction and 2D sections of skull– dural interface from LLC-DM-CopGFP CtskCre:Ai9 mice. SHG, second harmonic generation. c, Statistics of cell volume of tdTomato+ OCs and ITCs from b; n = 5 mice per group. d, Intravital two-photon microscopy (2PM) images (left, baseline; right panels, time-lapse series) of tumour lesion showing tdTomato+ OCs in the LLC-DM-CopGFP model of CtskCre:Ai9 mice. e, Experimental set-up related to f and g. CtskCre:Ai9 mice were injected intravenously with 5-FAM- ZOL at 1, 3, 6, 12, 24 and 72 h before being euthanized. DM sections were obtained on day 12 after LLC-DM inoculation for immunofluorescence staining. f, Skull–dural interface from e. Dashed line, skull edge. g, Curve chart showing the variation of proportion of 5-FAM-ZOL+ cells in tdTomato+ OCs/ITCs or PanCK+ tumour cells from f. n = 5 mice. h, Experimental set-up for i. i, Intravital time-lapse 2-photon microscopy images (left, baseline; right panels, time- lapse series) of tumour lesion showing tdTomato+ OCs in the LLC-DM model of CtskCre:Ai9 mice. Experiments were performed 1 h following 5-FAM-ZOL injection. j, UMAP of colour-coded re-clustered OCs extracted from Fig. 2b. k,l, GO biological processes enriched from up-regulated genes in OC subsets, cut-off = log2 fold change > 1. Data represent means ± s.e.m. Two-sided unpaired t-test (c) and two-sided hypergeometric distribution with Benjamini–Hochberg- adjusted P (k,l). Scale bars, 200 μm (a, top panel), 25 μm (a, lower panels), 50 μm (b,d (left panel)), 10 μm (d (enlarged right panels),i), 20 μm (f).

Fig. 5 | ANG2 triggers OC-related MLV impairments and T cell exhaustion. a, Normalized expression of genes related to lymphatic remodelling in control OCs (GSE147174) and two OC subsets in DM. Ctrl, control. b, Immunofluorescent images from the LLC-DM-CopGFP model of CtskCre:Ai9 mice. Dashed line, skull edge. c, Proportions of ANG2+ cells in tdTomato+ OCs or ITCs from b; n = 5 mice per group. d, Experimental set-up for e–o. 4T1-DM mice were treated with ZOL, or ZOL+ANG2 since tumour inoculation. e, Dura from d. The boxed areas in the left column are shown enlarged in the central and right columns. Dilations, blebs and local injuries are labelled with white, magenta and yellow rectangles or arrowheads, respectively. f, Statistics of LYVE1+ region area from e; n = 5 per group. g–i, Statistics of diameter (g), bleb (h) and local injuries (i) from e; n = 5 mice per group. j,k, Statistics of normalized 70 kDa dextran (j) and OVA (k) signals; n = 5 mice per group. Related to Extended Data Fig. 11g–i. l,m, Statistics of PD1+TIM3+ of dural CD8+ (l) and CD4+ (m) T cells from d; n = 5 mice per group, from three independent experiments. Related to Extended Data Fig. 13h,i. n, Left, tumour growth in d; right, BLI images; n = 5 mice per group, from three independent experiments. o, Survival of mice in d; n = 12 mice per group, from two independent experiments. p, Schematic diagram. SAS, subarachnoid space. Data represent means ± s.e.m. Two-sided unpaired t-test (f–n) and log-rank test (o). Scale bars, 40 μm (b, left panel), 10 μm (b, magnified images in right panels), 400 μm (e, left panels) and 100 μm (e, magnified images in right panels).