Lymphatic dysfunction in lupus contributes to cutaneous photosensitivity and lymph node B cell responses

Abstract

Patients with systemic lupus erythematosus (SLE) are photosensitive, developing skin inflammation with even ambient ultraviolet radiation (UVR), and this cutaneous photosensitivity can be associated with UVR-induced flares of systemic disease, which can involve increased autoantibodies and further end-organ injury. Mechanistic insight into the link between the skin responses and autoimmunity is limited. Signals from skin are transmitted directly to the immune system via lymphatic vessels, and here we show evidence for potentiation of UVR-induced lymphatic flow dysfunction in SLE patients and murine models. Improving lymphatic flow by manual lymphatic drainage (MLD) or with a transgenic model with increased lymphatic vessels reduces both cutaneous inflammation and lymph node B and T cell responses, and long-term MLD reduces splenomegaly and titers of a number of autoantibodies. Mechanistically, improved flow restrains B cell responses in part by stimulating a lymph node fibroblastic reticular cell-monocyte axis. Our results point to lymphatic modulation of lymph node stromal function as a link between photosensitive skin responses and autoimmunity and as a therapeutic target in lupus, provide insight into mechanisms by which the skin state regulates draining lymph node function, and suggest the possibility of MLD as an accessible and cost-effective adjunct to add to ongoing medical therapies for lupus and related diseases.

Keywords: Autoimmunity; Immunology; Inflammation; Lupus; Lymph; Vascular biology.

Figure 1. Patients with SLE and murine SLE models show evidence of cutaneous lymphatic dysfunction.

(A–C) Punch biopsies of forearm skin from healthy control (HC), positive APLs without SLE, and SLE subjects with or without APL were stained for CD31⁺PDPN⁺ lymphatic vessels. (A) Representative photomicrographs. Arrowheads point to lymphatic vessels. Original magnification ×40. (B) Lumenal area per vessel. Each symbol is an individual subject. (C) Number of lymphatic vessels per tissue area. (D–H) LPR mice and MRL controls were treated with UVR for 4 consecutive days, injected in ear pinna with Evans blue dye (EB) 1 day after final UVR dose, and ear harvested to assess EB content 1 day later, as in D. (E) EB retention and (F) ear swelling was quantified after UVR or (G and H) with no UVR. (I–K) Ears were examined by flow cytometry at 24 hours after final dose of UVR. (I) Monocyte, (J) TCR CD4⁺ T cell, and (K) TCRab⁺CD3⁺ CD4^–CD8^– DN T cell numbers. (L–O) B6 mice received IMQ on the right ear and were exposed to UVR, and EB retention after intradermal injection of left ear or footpad was compared with responses in vehicle-treated control mice, as in L. (M) EB retention in ear. (N) Ear swelling. (O) EB retention in footpad. (P and Q) Left ear was collected 24 hours after 1 or 4 doses of UVR. (P) Monocyte and (Q) TCR CD4⁺ T cell numbers. Each symbol represents 1 mouse; n = 2 to 21 per condition; data are from 2 (I–K), 3 (M and O), 4 (G and H), 6 (E and F), 10 (N), and 11 (P and Q) independent experiments. Normality was assessed using the Shapiro-Wilk test. If normal, unpaired t test was used. If data were not normal, Mann-Whitney U test was used. ***P < 0.001; **P < 0.01; *P < 0.05. Error bars represent SD.

Figure 2. Improving lymphatic flow reduces cutaneous photosensitive responses in SLE models.

(A–J) LPR mice and (K–P) B6-IMQ mice were treated with UVR and MLD targeting the left ear or were control handled. Left ear was then examined. (A and K) Experimental design. (B) Illustration of MLD technique. Please see Methods for details. (C and L) EB retention and (D, E, and M) ear thickness, absolute (D and M) or normalized (E). (F and N) TCR CD4⁺ T cell numbers, (G and O) normalized to controls. (H and P) Percentage of CD4⁺ T cells that express IFN-γ. (I) DN T cell numbers and (J) percentage that express IFN-γ. (Q–U) Flt4Cre^ERT2 PTEN^fl/fl mice treated with tamoxifen (LEC^PTEN) or without (LEC^WT) were treated with IMQ on right ear and UVR before left ear skin assessment, as in Q. (R) EB retention, (S) skin thickness, (T) TCR CD4⁺ T cell numbers, and (U) IFN-γ⁺ percentage. (V and W) Nontransgenic B6 mice were treated as described in Q. (V) EB retention and (W) ear thickness. Each symbol represents 1 mouse; n = 3 to 13 per condition; data are from 2 (H, J, L, V, and W), 3 (N, O, P, R, T, and U), 4 (F, G, I, and S), 6 (C), 7 (D and E), and 9 (M) independent experiments. Normality was assessed using the Shapiro-Wilk test. If normal, unpaired t test was used. If data were not normal, Mann-Whitney U test was used. ***P < 0.001; **P < 0.01; *P < 0.05. Error bars represent SD.

Figure 3. Improving lymphatic flow reduces draining lymph node B and T cell responses in SLE models.

(A–Q) Left auricular lymph nodes of (A–I) LPR, (J–Q) B6-IMQ, and (R–X) LEC^PTEN-IMQ and LEC^WT- IMQ mice that were treated as in Figure 2A, K, and Q were examined. (A, J, and R) Lymph node cellularity. (B, K, and S) B cell, (C, L, and T) germinal center (GC) B cell, (D, M, and U) plasmablast, and (E, N, and V) TCR CD4⁺ T cell numbers. (F, O, and W) CD4⁺ T cell numbers normalized to control. Rel, relative. (G, P, and X) Percentage of CD4⁺ T cells that express IFN-γ. (H, Q, and Y) CD8⁺ and (I) DN T cell numbers. Each symbol represents 1 mouse; n = 4 to 17 per condition; data are from 2 (X), 3 (G, P, and Q), 4 (A–C, H–K, M, V, W, and Y), 5 (D, N, and O), 6 (E, F, L, R, and U), and 8 (S and T) independent experiments. Normality was assessed using the Shapiro-Wilk test. If normal, unpaired t test was used. If data were not normal, Mann-Whitney U test was used. ***P < 0.001; **P < 0.01; *P < 0.05. Error bars represent SD.

Figure 4. Improving lymphatic flow long term reduces systemic disease activity.

(A) Splenic weight of LEC^PTEN-IMQ and LEC^WT-IMQ mice treated with UVR for 4 days. (B) Splenic weight of LPR mice treated with UVR and MLD concurrently or control handled for 5 weeks. (C and D) Heatmap of normalized signal intensity (NSI) from autoantigen microarray panel for IgG of MRL mice, LPR mice treated with UVR, and LPR mice treated with both UVR and MLD for 4–5 weeks. (C) Each column represents 1 mouse and autoantibodies with significant differences between LPR UV and LPR UV+MLD are labeled in red. (D) Each column represents average NSI of all mouse serum samples. Each symbol represents 1 mouse; n = 1 to 12 per condition; data are from 1 (B), 2 (C and D), and 6 (A) independent experiments. Normality was assessed using the Shapiro-Wilk test. If normal, unpaired t test was used. If data were not normal, Mann-Whitney U test was used. **P < 0.01; *P < 0.05. Error bars represent SD.

Figure 5. Improving lymphatic flow increases lymph node FRC proliferation, FRC CCL2, monocyte ROS generation, and limits plasmablast numbers in a monocyte-dependent manner.

(A–Q) Left auricular lymph nodes of indicated mice exposed to 4 days of UVR were examined. (A) FRC expression of GFP in CCL2-GFP reporter mice treated with IMQ that received MLD or control handling with UVR. (B) CCL2 expression by FRCs in LEC^PTEN-IMQ and LEC^WT- IMQ mice. (A and B) Representative histograms (left) and graphs (right). MFI, geometric mean fluorescence intensity. (C–G) B6-IMQ mice received 4 days of +/-MLD with UVR. (C) FRC numbers. (D) Percentage of FRCs that express Ki-67, (E) normalized to control handled mice. (F) Ly6C^hi monocyte numbers, flow cytometry gating (left), and numbers (right). (G) Monocyte ROS measured using CM-H2DCFDA, representative histograms (left) and relative MFI of CM-H2DCFDA (right). (H–L) B6-IMQ mice were treated with anti-Gr-1 or isotype control at days –1, 0, and +2 of UVR and MLD treatments as shown in H. (I) Monocyte numbers and (J) normalized to isotype control. (K) Germinal center B cell and (I) plasmablast numbers. (M–Q) CCR2-DTR mice were treated with DT at days 0 and 2 of UVR and MLD treatments as shown in M. (N) Monocyte numbers and (O) normalized to control. (P) Germinal center B cell and (Q) plasmablast numbers. Each symbol represents 1 mouse; n = 6 to 19 per condition; data are from 2 (B), 3 (A and I–L), 5 (N and O), 6 (D–G, P, and Q), and 7 (C) independent experiments. Normality was assessed using the Shapiro-Wilk test. If normal, unpaired t test was used. If data were not normal, Mann-Whitney U test was used. ***P < 0.001; **P < 0.01; *P < 0.05. Error bars represent SD.