TGF-β1 mediates pathologic changes of secondary lymphedema by promoting fibrosis and inflammation

Abstract

Background: Secondary lymphedema is a common complication of cancer treatment, and previous studies have shown that the expression of transforming growth factor-beta 1 (TGF-β1), a pro-fibrotic and anti-lymphangiogenic growth factor, is increased in this disease. Inhibition of TGF-β1 decreases the severity of the disease in mouse models; however, the mechanisms that regulate this improvement remain unknown.

Methods: Expression of TGF-β1 and extracellular matrix molecules (ECM) was assessed in biopsy specimens from patients with unilateral breast cancer-related lymphedema (BCRL). The effects of TGF-β1 inhibition using neutralizing antibodies or a topical formulation of pirfenidone (PFD) were analyzed in mouse models of lymphedema. We also assessed the direct effects of TGF-β1 on lymphatic endothelial cells (LECs) using transgenic mice that expressed a dominant-negative TGF-β receptor selectively on LECs (LEC^DN-RII ).

Results: The expression of TGF-β1 and ECM molecules is significantly increased in BCRL skin biopsies. Inhibition of TGF-β1 in mouse models of lymphedema using neutralizing antibodies or with topical PFD decreased ECM deposition, increased the formation of collateral lymphatics, and inhibited infiltration of T cells. In vitro studies showed that TGF-β1 in lymphedematous tissues increases fibroblast, lymphatic endothelial cell (LEC), and lymphatic smooth muscle cell stiffness. Knockdown of TGF-β1 responsiveness in LEC^DN-RII resulted in increased lymphangiogenesis and collateral lymphatic formation; however, ECM deposition and fibrosis persisted, and the severity of lymphedema was indistinguishable from controls.

Conclusions: Our results show that TGF-β1 is an essential regulator of ECM deposition in secondary lymphedema and that inhibition of this response is a promising means of treating lymphedema.

Keywords: TGF-β; fibrosis; inflammation; pirfenidone.

FIGURE 1

BCRL results in increased TGF‐β1 expression and signaling. (A) Representative IF localisation of TGF‐β1 (top) and pSMAD3 (bottom) in normal and lymphedematous (labelled LE) tissues. (B) Quantification of TGF‐β1 (top) and pSMAD3 (bottom) IF staining areas in tissue sections of patients with unilateral BCRL. Each circle represents an average of three HPF views per patient (N = 8). (C) mRNA expression of TGF‐β isoforms and TGF‐βRI comparing normal and lymphedematous limb of patients with unilateral BCRL. Each circle represents an individual patient (N = 14). (D) Representative Western blot of TGF‐β isoforms, pSMAD3 and tSMAD3 in normal and lymphedematous limbs of patients with unilateral BCRL. (E) Quantification of Western blots with relative changes comparing normal and lymphedematous limb of each patient. Each circle represents an average of two separate Western blots per patient (N = 8). BCRL, breast cancer‐related lymphedema; TGF‐β1, transforming growth factor‐beta 1; IF, immunofluorescence; LE, lymphedema; HPF, high‐power field; TGF‐βR‐I, transforming growth factor‐beta receptor I

FIGURE 2

BCRL increases fibrosis and ECM deposition. (A) Representative IF localisation of type 1 collagen (Col1; top) and CD26 (bottom) in normal and lymphedematous (labelled LE) tissues. (B) Quantification of type I collagen (top) and CD26 (bottom) IF staining areas in tissue sections of patients with unilateral BCRL. Each circle represents an average of three HPF views per patient (N = 8). (C) Left panel: Representative Western blot of ECM proteins in normal and lymphedematous limbs of patients with unilateral BCRL. Right panel: Quantification of ECM proteins. Each circle represents an average of two separate Western blots per patient (N = 8). (D) mRNA expression of ECM molecules comparing normal and lymphedematous limb of patients with unilateral BCRL. Each circle represents an individual patient (N = 12–14). BCRL, breast cancer‐related lymphedema; ECM, extracellular matrix molecules; IF, immunofluorescence; LE, lymphedema; HPF, high‐power field

FIGURE 3

Neutralisation of TGF‐β1 decreases lymphedema and inflammation. (A) Left panel: Representative cross‐sectional H&E stain of a mouse tail treated with isotype control or TGF‐β1 neutralising antibody. Brackets show subcutaneous fibroadipose tissues; note decreased subcutaneous thickness in TGF‐β1 neutralising antibody‐treated mice. Right panel: Quantification of skin thickness in control or TGF‐β1 neutralising antibody‐treated mice. Each circle represents an average of three HPF views per animal (N = 4). (B) Left panel: Representative IF localisation of TGF‐β1 (top) and pSmad3 (bottom) in isotype and TGF‐β1 neutralising antibody‐treated mice. Right panel: Quantification of TGF‐β1 area and Smad3+ cells. Each circle represents an average of three HPF views per animal (N = 5). (C) mRNA expression of TGF‐β1 isoforms and downstream signaling pathways. Relative change to isotype control‐treated mice is shown. Each circle represents an individual animal (N = 5). Genes shown in the green‐ and red‐shaded zones represent upregulated (green) and downregulated (red) molecules. *p < .05, **p < .01. (D) mRNA expression of inflammatory mediators. Relative change to isotype control‐treated mice is shown. Each circle represents an individual animal (N = 5). Genes shown in the green‐ and red‐shaded zones represent upregulated (green) and downregulated (red) molecules. *p < .05, **p < .01. (E) Quantification of flow cytometry for leukocytes (CD45⁺), CD45⁺/CD4⁺ cells, Th1 (CD45⁺/CD4⁺/CXCR3⁺/CCR5⁺) cells, Th2 (CD45⁺/CD4⁺/CCR4⁺/CCR8⁺) cells, macrophages (CD45⁺/CD11b⁺/F480⁺) and neutrophils (CD45⁺/CD11b⁺/Ly‐6G⁺). Each circle represents an average of two flow cytometry runs for each animal (N = 5). (F) Quantification of CD4 T cells (CD45⁺/CD4⁺), macrophages (CD45⁺/CD11b⁺/F480⁺) and DCs (CD45⁺/CD11c⁺/CD86⁺) in draining lymph nodes and foot pad injection site after injection of LE lysate. Each circle represents an individual animal (N = 5). TGF‐β1, transforming growth factor‐beta 1; H&E haematoxylin and eosin; HPF, high‐power field; IF, immunofluorescence; DCs, dendritic cells; LE, lymphedema

FIGURE 4

TGF‐β1 increases fibroblast ECM expression and increases the stiffness of fibroblasts, LECs and LSMCs. (A) Left panel: Representative IF localisation of type I collagen (Col1) (top) and CD26 (bottom) in isotype and TGF‐β1 neutralising antibody‐treated mice. Right panel: Quantification of type I collagen (top) and CD26 (bottom) IF staining areas in tissue sections. Each circle represents an average of three HPF views per animal (N = 4). (B) Representative Western blot of ECM proteins in mouse tails of animals treated with isotype control or TGF‐β1 neutralising antibodies. (C) Quantification of Western blots of ECM proteins with relative changes comparing isotype control and TGF‐β1 neutralising antibody‐treated mice. Each circle represents an average of two separate Western blots per animal (N = 4). (D) Representative Western blot of ECM proteins in NIH3T3 fibroblasts stimulated with LE lysate containing isotype control or TGF‐β1 neutralising antibodies. (E) Quantification of Western blot of ECM proteins in NIH3T3 fibroblasts stimulated with LE lysate containing isotype control or TGF‐β1 neutralising antibodies. Each circle represents an average of four separate Western blots per group. (F) Proliferation of NIH3T3 fibroblasts 24 and 72 h after stimulation with normal media (labelled NM) or media supplemented with LE lysate containing isotype control or TGF‐β1 neutralising antibodies. (G) Left panel: Representative atomic force microscopy heatmaps of fibroblasts, LECs and LSMCs, stimulated with LE lysate containing isotype control or TGF‐β1 neutralising antibodies. Right panel: Quantification of Young's elastic modulus in fibroblasts, LECs and LSMCs stimulated with LE lysate containing isotype control or TGF‐β1 neutralising antibodies. TGF‐β1, transforming growth factor‐beta 1; ECM, extracellular matrix molecules; LECs, lymphatic endothelial cells; LSMCs, lymphatic smooth muscle cells; IF, immunofluorescence; HPF, high‐power field; LE, lymphedema

FIGURE 5

Inhibition of TGF‐β1 signaling in LECs increases lymphangiogenesis but does not improve lymphedema. (A) Left panel: Representative cross‐sectional H&E stain of WT or LEC^DN‐TBRII mouse tails. Right panel: Quantification of skin thickness in WT and LEC^DN‐TBRII mouse tails. Each circle represents an average of three HPF views per animal (N = 5). Bar = 500 μm. (B) Representative IF localisation of type I collagen (Col1) (top) and CD45 (bottom) in WT and LEC^DN‐TBRII mouse tails. (C) Quantification of type I collagen deposition in WT and LEC^DN‐TBRII mouse tails. Each circle represents an average of three HPF views per animal (N = 4). (D) Quantification of CD45⁺ cells in WT and LEC^DN‐TBRII mouse tails. Each circle represents an average of three HPF views per animal (N = 5). (E) Quantification of lymphatic vessels in WT and LEC^DN‐TBRII mouse tails. Each circle represents the average of three HPF views per animal (N = 6). (F) Quantification of bridging lymphatic channels in WT and LEC^DN‐TBRII mouse tails. Each circle represents an average of three HPF views per animal (N = 4). (G) Quantification of Ki67⁺ LECs in lymphatic vessels of WT and LEC^DN‐TBRII mouse tails. Each circle represents an average of three HPF views per animal (N = 5). (H) Lymphatic diameter in WT and LEC^DN‐TBRII mouse tails. Each circle represents the average of four lymphatic vessels per animal (N = 5). TGF‐β1, transforming growth factor‐beta 1; LECs, lymphatic endothelial cells; H&E haematoxylin and eosin; WT, wild‐type; HPF, high‐power field; IF, immunofluorescence

FIGURE 6

Topical PFD decreases the pathology of lymphedema. (A) Change in tail volume over time in mice treated with vehicle or PFD. Each circle is an average of duplicate measurements from each mouse (N = 9 animals per group). Results are presented as mean ± SEM (*p < .05, **p < .01). Statistical comparisons are between groups at the same time points. (B) Left panel: Representative cross‐sectional H&E stain of vehicle and PFD‐treated mouse tails. Brackets show subcutaneous fibroadipose tissue deposition. Right panel: Quantification of skin thickness in vehicle and PFD‐treated mouse tails. Each circle represents an average of three HPF views from 5 animals. Bar = 500 μm. (C) Quantification of type I collagen deposition in vehicle and PFD‐treated mouse tails. Each circle represents an average of three HPF views from 4 animals. (D) Quantification of TGF‐ β1 protein in tissues collected from vehicle and PFD‐treated mice. Each circle represents an average of duplicate ELISAs per animal (N = 9). (E) Quantification of the number of pSmad3+ cells in vehicle and PFD‐treated mouse tails. Each circle represents an average of cell counts in three HPF views from 7 animals. (F) mRNA expression of TGF‐β1 isoforms and downstream signaling pathways. Relative change to vehicle control‐treated mice is shown. Each circle represents an individual animal (N = 6). Genes shown in the green‐ and red‐shaded zones represent upregulated (green) and downregulated (red) molecules. *p < .05, **p < .01, ***p < .001. (G) mRNA expression of fibrosis and ECM genes. Relative change to vehicle control‐treated mice is shown. Each circle represents an individual animal (N = 6). Genes shown in the green‐ and red‐shaded zones represent upregulated (green) and downregulated (red) molecules. *p < .05, **p < .01, ***p < .001. (H) mRNA expression of inflammatory cytokines. Relative change to vehicle control‐treated mice is shown. Each circle represents an individual animal (N = 6). Genes shown in the green‐ and red‐shaded zones represent upregulated (green) and downregulated (red) molecules. *p < .05, **p < .01, ***p < .001. (I) Quantification of flow cytometry for leukocytes (CD45⁺), macrophages (CD45⁺/CD11b⁺/F480⁺), CD45⁺/CD4⁺ cells, Th1 (CD45⁺/CD4⁺/CXCR3⁺/CCR5⁺) cells and Th2 (CD45⁺/CD4⁺/CCR4⁺/CCR8⁺) cells. Each circle represents an average of two flow cytometry runs per animal (N = 4). (J) Left panel: Representative IF localisation of CD4 (red) and LYVE‐1 (green) in vehicle and PFD‐treated mouse tails. Right panel: Quantification of perilymphatic CD4⁺ cells. Each circle represents an average of three HPF views from 9 animals. PFD, pirfenidone; SEM, standard error of the mean; H&E haematoxylin and eosin; HPF, high‐power field; TGF‐β1, transforming growth factor‐beta 1; ELISA, enzyme‐linked immunosorbent assay; ECM, extracellular matrix molecules; IF, immunofluorescence

FIGURE 7

Topical PFD improves lymphatic function. (A) Representative heat map of Tc uptake in the sacral lymph nodes of vehicle and PFD‐treated mouse tails. (B) Left panel: Quantification of decay adjusted Tc uptake in the sacral lymph nodes of vehicle and PFD‐treated mouse tails. The rate of Tc uptake (centre panel) and peak nodal uptake (right panel) are also shown. Each circle represents an average of three measurements per animal at each time point (N = 7 animals/group). (C) Experimental plan for PLND and analysis of lymphatic pumping/dermal backflow. (D) Representative ICG image of vehicle and PFD‐treated mice. Note increased collateral vessel formation and decreased dermal backflow (white fluorescence) in PFD‐treated animals. (E) Representative plot showing changes in light intensity in collecting lymphatic vessels of vehicle and PFD‐treated mice. (F) Left panel: Quantification of collecting lymphatic pumping (packets/min) in vehicle and PFD‐treated mice. Right panel: Quantification of dermal backflow. Each circle represents an average of two measurements per animal (N = 5). PFD, pirfenidone; PLND, popliteal lymph node dissections; ICG, indocyanine green