Wild type mesenchymal cells contribute to the lung pathology of lymphangioleiomyomatosis

Abstract

Lymphangioleiomyomatosis (LAM) is a rare disease leading to lungs cysts and progressive respiratory failure. Cells of unknown origin accumulate in the lungs forming nodules and eventually resulting in lung cysts. These LAM cells are described as clonal with bi-allelic mutations in TSC-2 resulting in constitutive mTOR activation. However LAM nodules are heterogeneous structures containing cells of different phenotypes; we investigated whether recruited wild type cells were also present alongside mutation bearing cells. Cells were isolated from LAM lung tissue, cultured and characterised using microscopy, immunocytochemistry and western blotting. Fibroblast-like cells were identified in lung tissue using immunohistochemical markers. Fibroblast chemotaxis toward LAM cells was examined using migration assays and 3D cell culture. Fibroblast-like cells were obtained from LAM lungs: these cells had fibroblast-like morphology, actin stress fibres, full length tuberin protein and suppressible ribosomal protein S6 activity suggesting functional TSC-1/2 protein. Fibroblast Activation Protein, Fibroblast Specific Protein/S100A4 and Fibroblast Surface Protein all stained subsets of cells within LAM nodules from multiple donors. In a mouse model of LAM, tuberin positive host derived cells were also present within lung nodules of xenografted TSC-2 null cells. In vitro, LAM 621-101 cells and fibroblasts formed spontaneous aggregates over three days in 3D co-cultures. Fibroblast chemotaxis was enhanced two fold by LAM 621-101 conditioned medium (p=0.05), which was partially dependent upon LAM cell derived CXCL12. Further, LAM cell conditioned medium also halved fibroblast apoptosis under serum free conditions (p=0.03). Our findings suggest that LAM nodules contain a significant population of fibroblast-like cells. Analogous to cancer associated fibroblasts, these cells may provide a permissive environment for LAM cell growth and contribute to the lung pathology of LAM lung disease.

Fig 1. Characteristics of wild-type fibroblast-like cells obtained from LAM lungs.

(a) Explant culture of lung tissue from patients with LAM yields a fibroblast-like population with spindle shaped morphology under phase contrast microscopy (phase) and immunoreactivity to anti-alpha Smooth Muscle Actin, and fibroblast markers FSP, S100A4 and TE-7 (these cells are also DAPI stained for clarity). (b) Western blot of lysates of fibroblast like-cells from two donors with LAM compared with TSC-2 mutation bearing 621–101 cells. Unlike 621–101 cells fibroblast-like cells have full length tuberin, and suppression of phospho-S6 in the absence of serum.

Fig 2. LAM nodules express fibroblast markers.

LAM nodules react with antibodies against alpha-Smooth Muscle Actin, Fibroblast Surface Protein (FSP), Fibroblast Activation Protein (FAP) and S100A4. Left panels are x20 magnification and right are inset area at x40 taken from serial sections of a representative donor. Fibroblast markers display different expression patterns within LAM tissue, with anti-FSP reacting with 30–70% of cells within a nodule, anti-FAP detecting the majority of cells and S100A4 detecting only small nests of cells within nodules.

Fig 3. FSP is expressed by variable numbers of cells within LAM nodules.

There is heterogeneity between different LAM lung donors in the number of anti-FSP immunoreactive cells in LAM nodules. Panels (a) and (b) at x10 magnification, (c) at x20, (d) at x40. (e) and (f) show anti-tuberin reactive and non-reactive cells (arrowed) within nodules from two donors, x20.

Fig 4. Fibroblast markers and melanoma markers are expressed in different cell populations within LAM nodules.

LAM tissue subjected to immunofluorescence using melanoma marker antibodies PNL2 (green) and anti-gp100 (red) (the merged image shows cells which are positive for both markers and appear yellow/orange), and anti-gp100 (red) and the fibroblast marker anti-FSP (green), which do not overlap. Main images x10 magnification.

Fig 5. A mouse model of LAM also contains wild-type cells in lung nodules.

NCRNU-M athymic nude mice received tail-vein injections of cells obtained from TSC-2^-/- mouse kidney epithelial tumours. Mice develop lung nodules and air space enlargement (a). Lung tissue was stained with an anti-tuberin antibody which shows a large number of tuberin positive (host) cells within nodules admixed with tuberin null (tumour) cells (b) examples highlighted with arrows, suggesting host cells have migrated toward the TSC2^-/- cells. The absence of phosphorylation of ribosomal protein S6 in many cells is also an indicator of functional tuberin activity in these cells.

Fig 6. LAM 621–101 cells and fibroblasts form spontaneous aggregates in 3D co-cultures which are CXCR4/CXCL12 dependent.

(a) 621–10 cells (red) and fibroblasts (green) are co-cultured in a three dimensional extra-cellular matrix gel and imaged by confocal microscopy. After three days cells spontaneously form clumps with fibroblasts appearing to surround 621–101 cells. (b) Addition of 200μg/ml AMD3100 results in fewer and smaller 621–101 cell—fibroblast aggregates.

Fig 7. LAM cell medium is chemotactic for fibroblasts, and this is partially dependent on CXCR4.

(a) 621–101 cells secrete CXCL12 protein; this is partially sensitive to rapamycin. (b) The presence of 621–101 cell conditioned medium (CM) increases fibroblast migration, which is partially blocked by AMD3100. (c) Fibroblasts migrate towards CXCL12 in a modified Boyden chamber (Transwell) assay; this is inhibited by the CXCR4 receptor antagonist AMD3100 at 100μg/ml (*p = 0.05). (d) Fibroblasts migrate towards 621–101 cells in a bidirectional migration assay. (i) Removal of the fence leaves a gap between the peripheral fibroblasts (green) and the central reservoir of (red) 621–101 cells (dotted), (ii) Fibroblasts migrate towards the central reservoir (arrowed), (iii) fibroblasts migrate into the central area (arrowed). (e) Addition of 200μg/ml AMD3100 reduces migration of fibroblasts in response to 621–101 cells by 56%. In the absence of 621–101 cells in the central reservoir migration is reduced by 93%.

Fig 8. Fibroblasts protect 621–101 cells against apoptosis.

(a) Culture of either 621–101 cells or fibroblasts on glass in serum free conditions results in cell death. Co-culture of the two cell types under the same conditions results in formation of viable cell colonies. (b) Under serum-free conditions, 621–101 cells have a basal rate of apoptosis of 8% assessed by TUNEL positive nuclear staining (arrowed). TUNEL positivity is reduced by half in the presence of fibroblast conditioned medium (*p = 0.032).