Pathologic Proteolytic Processing of N-Cadherin as a Marker of Human Fibrotic Disease

Abstract

Prior research has implicated the involvement of cell adhesion molecule N-cadherin in tissue fibrosis and remodeling. We hypothesize that anomalies in N-cadherin protein processing are involved in pathological fibrosis. Diseased tissues associated with fibrosis of the heart, lung, and liver were probed for the precursor form of N-cadherin, pro-N-cadherin (PNC), by immunohistochemistry and compared to healthy tissues. Myofibroblast cell lines were analyzed for cell surface pro-N-cadherin by flow cytometry and immunofluorescent microscopy. Soluble PNC products were immunoprecipitated from patient plasmas and an enzyme-linked immunoassay was developed for quantification. All fibrotic tissues examined show aberrant PNC localization. Cell surface PNC is expressed in myofibroblast cell lines isolated from cardiomyopathy and idiopathic pulmonary fibrosis but not on myofibroblasts isolated from healthy tissues. PNC is elevated in the plasma of patients with cardiomyopathy (p ≤ 0.0001), idiopathic pulmonary fibrosis (p ≤ 0.05), and nonalcoholic fatty liver disease with cirrhosis (p ≤ 0.05). Finally, we have humanized a murine antibody and demonstrate that it significantly inhibits migration of PNC expressing myofibroblasts. Collectively, the aberrant localization of PNC is observed in all fibrotic tissues examined in our study and our data suggest a role for cell surface PNC in the pathogenesis of fibrosis.

Keywords: N-cadherin; cirrhosis; fibronectin; fibrosis; heart failure; myofibroblasts; pro-N-cadherin; pulmonary fibrosis.

Figure 1

PNC is aberrantly expressed in fibrotic heart, lung, and liver tissues. (A) Representative images of stained tissues from explanted failed human hearts (n = 20, ischemic and non-ischemic etiology), lungs (n = 10, IPF etiology) and livers (n = 40, NAFLD-cirrhosis etiology) show positive expression and aberrant localization of PNC (Brown stain). Corresponding representative images of stained normal human heart (n = 24), lung (n = 18), and liver (n = 32) show lack of aberrantly expressed PNC at the tissue level. Perinuclear staining, consistent with normal N-cadherin processing can be seen in the healthy cardiac tissue. Scale bar = 100 µm. (B) Graphical representation of human samples analyzed in this study. Each column represents a single human sample, annotated by color to indicate the organ system of relevance to this study (top row), whether the patient had fibrosis (second row), the reported sex of the patient (third row), the age of the patient (fourth row) and the method by which the sample was analyzed in this study (fourth row). In cases where demographic data is unknown, unavailable, or unreported, a white bar with black hatch is shown. n = 257 total samples analyzed.

Figure 2

PNC is localized to the cell surface of myofibroblasts. Myofibroblasts from heart and lung tissues were stained and analyzed by flow cytometry using m-α-PNC mAb 10A10, excluding debris and dead cells via gating and 7AAD exclusion using Flowjo software. Myofibroblasts from fibrosis origins stain positive for cell surface PNC: (A) CF-DCM, cardiac myofibroblasts from dilated cardiomyopathy; Chi-Squared = 177.5; SE Dymax % Positive = 48.0. LL97A, IPF; Chi-Squared = 93.4; SE Dymax % Positive = 54.8, and LL29, IPF; Chi-Squared = 113.5; SE Dymax % Positive = 51.0. PNC was not detected on the surface of primary normal human cardiac myofibroblasts from healthy donor, NHCF; Chi-Squared = 3.47; SE Dymax % Positive = 8.54, primary normal human lung myofibroblasts from healthy donor, NHLF; Chi-Squared = 0; SE Dymax % Positive = 0, or immortalized CCD-16Lu lung myofibroblasts from healthy donor; Chi-Squared = 0; SE Dymax % Positive = 0. Mouse IgG1 isotype control (blue shaded) was compared to m-α-PNC mAb (unshaded). Results are representative of 3 independent experiments, n = 3. For all flow cytometry experiments, Chi-squared ≥ 4 is statistically significant. (B) Cell surface proteins were isolated from myofibroblasts and immunoblotted for N-cadherin, PNC, and Na,k-ATPase α-1 (ATP1A1) cell surface compartment loading control. PNC and N-Cadherin lysates were normalized to the ATP1A1 cell surface loading control for each sample and are reported as a relative value below each band.

Figure 3

PNC is localized to cellular protrusions. Fixed, unpermeabilized myofibroblasts from fibrotic and healthy tissues were immunostained for m-α-PNC mAb 10A10 (red) and DAPI (blue). PNC is localized to cellular protrusions on pathological myofibroblasts DCM-CF, LL29, and LL97A (arrows). PNC is not expressed on the surface of NHCF, NHLF, or CCD-16Lu isolated from healthy donor.

Figure 4

Feasibility and development of a PNC ELISA. (A) Soluble PNC product was immunoprecipitated from LL29 conditioned media and patient plasma from pooled healthy donors, IPF patients, NAFLD-Cirrhosis patients and cardiomyopathy patients using m-α-PNC mAb 10A10 compared to mouse IgG1 isotype control (mIgG1) and immunoblotted. (B) Recombinant prodomain analyte was serially diluted in duplicate starting at 100 ng/mL in standard diluent and measured by ELISA to determine the range of the standard. (C) Linearity of endogenous analyte was measured. Plasma from patients with heart failure was serially diluted 1:3, 1:7, 1:15, 1:31 in standard diluent and sPNC was measured by ELISA in duplicate. All linear regression r² values were calculated to be within acceptable range (r² ≥ 0.99). (D,E) Healthy donor plasma was diluted in standard diluent to indicated dilutions (1:1, 1:3, 1:7, 1:15) then spiked with recombinant prodomain analyte at 10 ng/mL (D) and 5 ng/mL (E) and analyzed by ELISA. Recovery of analyte was quantified and compared to back calculation of the standard. All dilutions were within the consensus range of 20 percent ± the standard back calculations at concentrations 10 ng/mL and 5 ng/mL. (F) Plasma was assayed for sPNC from healthy controls (n = 26), patients with IPF (n = 9), NAFLD with cirrhosis (n = 12), and cardiomyopathy (n = 9). Ordinary one-way ANOVA analysis with Dunnett’s multiple comparisons test was performed to determine significance (* p ≤ 0.05, **** p ≤ 0.0001).

Figure 5

FN1 is a potential PNC binding partner. (A) Medium binding ELISA plates were coated with human fibronectin, collagen type I, or collagen type III, then blocked and incubated with his-tagged recombinant N-cadherin prodomain. After washing unbound prodomain, prodomain binding to the immobilized substrate was measured using a biotinylated his-tag specific monoclonal antibody and streptavidin-HRP for detection. Assay was performed in quadruplicate and is representative of at least 3 independent experiments. (B) Representative image of a cardiac myofibroblast isolated from failed cardiac explant tissue showing colocalization of PNC and FN1 immunostained for PNC (red), FN1 (green) and DAPI (blue). Yellow indicates PNC/FN1 colocalization (Merge). (C) Medium binding ELISA plates were coated with fibronectin, blocked, then incubated with either recombinant prodomain of N-cadherin or prodomain in combination with mouse IgG1 isotype control, human IgG4 isotype control, m-α-PNC mAb 10A10 or h-α-PNC mAb HC5LC4. Bound recombinant prodomain was detected using anti-his-tag monoclonal antibody in technical duplicates and representative of at least 3 independent experiments (n = 3). Ordinary one-way ANOVA analysis with Tukey’s multiple comparisons test was performed to determine significance (* p ≤ 0.05, ** p ≤ 0.01).

Figure 6

Cell surface PNC has a role in myofibroblast migration. Transwell permeable supports with a 6.5 mm polycarbonate membrane and 8 µm pores were coated with human fibronectin and used to separate the upper and lower chambers of a 24-well cell culture plate to measure migration of cells across the membrane (n = 4). Pathological myofibroblast migration is significantly reduced by h-α-PNC mAb (HC5LC4) and recombinant prodomain of N-cadherin (rPro) after 5 h; (A) DCM-CF, dilated cardiomyopathy myofibroblasts (C) LL29, IPF myofibroblasts. No significant effect on myofibroblasts isolated from healthy tissues was observed; (D) immortalized CCD-16Lu lung myofibroblasts from healthy donor (B) NHCF, normal human cardiac myofibroblasts. Two tailed T-test assuming Gaussian distribution analysis was performed to determine significance. (* p ≤ 0.05, ** p ≤ 0.01).

Figure 7

AUROC analysis of plasma PNC for specificity and sensitivity. Normal donor samples (n= 26) were compared to samples from cardiomyopathy (Red, n = 9), NAFLD-Cirrhosis (Green, n = 12) and IPF (Yellow, n = 9) patients. To determine tissue-agnostic AUROC, all fibrotic samples were combined (Black, n= 30) and compared to normal donor samples. AUROC = area under the receiver operating characteristics curve.