C5b-9 Membrane Attack Complex Formation and Extracellular Vesicle Shedding in Barrett's Esophagus and Esophageal Adenocarcinoma

Abstract

The early complement components have emerged as mediators of pro-oncogenic inflammation, classically inferred to cause terminal complement activation, but there are limited data on the activity of terminal complement in cancer. We previously reported elevated serum and tissue C9, the terminal complement component, in esophageal adenocarcinoma (EAC) compared to the precursor condition Barrett's Esophagus (BE) and healthy controls. Here, we investigate the level and cellular fates of the terminal complement complex C5b-9, also known as the membrane attack complex. Punctate C5b-9 staining and diffuse C9 staining was detected in BE and EAC by multiplex immunohistofluorescence without corresponding increase of C9 mRNA transcript. Increased C9 and C5b-9 staining were observed in the sequence normal squamous epithelium, BE, low- and high-grade dysplasia, EAC. C5b-9 positive esophageal cells were morphologically intact, indicative of sublytic or complement-evasion mechanisms. To investigate this at a cellular level, we exposed non-dysplastic BE (BAR-T and CP-A), high-grade dysplastic BE (CP-B and CP-D) and EAC (FLO-1 and OE-33) cell lines to the same sublytic dose of immunopurified human C9 (3 µg/ml) in the presence of C9-depleted human serum. Cellular C5b-9 was visualized by immunofluorescence confocal microscopy. Shed C5b-9 in the form of extracellular vesicles (EV) was measured in collected conditioned medium using recently described microfluidic immunoassay with capture by a mixture of three tetraspanin antibodies (CD9/CD63/CD81) and detection by surface-enhanced Raman scattering (SERS) after EV labelling with C5b-9 or C9 antibody conjugated SERS nanotags. Following C9 exposure, all examined cell lines formed C5b-9, internalized C5b-9, and shed C5b-9⁺ and C9⁺ EVs, albeit at varying levels despite receiving the same C9 dose. In conclusion, these results confirm increased esophageal C5b-9 formation during EAC development and demonstrate capability and heterogeneity in C5b-9 formation and shedding in BE and EAC cell lines following sublytic C9 exposure. Future work may explore the molecular mechanisms and pathogenic implications of the shed C5b-9⁺ EV.

Keywords: complement system activation; exosome; extracellular vesicle (EV); microvesicle; terminal complement component (SC5b-9).

Figure 1

Representative immunohistofluorescence for C9 and C5b-9 in esophageal tissues during stages of esophageal adenocarcinoma development. Tissue microarrays of esophageal biopsies and esophageal adenocarcinoma (EAC) specimens were stained for C9 and C5b-9 using multiplex immunohistofluorescence, and then exported as chromogenic images for visualizations and scoring. Representatives images for the stages of EAC development are shown in panels (A–F), with the areas marked by black lined shapes: normal squamous epithelium (NSE), non-dysplastic Barrett’s esophagus (BE), low-grade dysplasia (LGD), high-grade dysplasia (HGD), HGD with intraepithelial carcinoma (HGD+IEC) and EAC. Cores are 1.5mm. The whole EAC section is cancerous. Staining intensity was scored for C9 (G) and C5b-9 (H) for each tissue phenotype on a 0-3 scale by a specialist pathologist.

Figure 2

Localization of C9 and C5b-9 in esophageal tissues during stages of esophageal adenocarcinoma development. High magnification images of the multiplex immunohistofluorescence staining for C9 (red) and C5b-9 (green), with the blue DAPI stain for nucleus. (A) NSE, (B) BE, (C) LGD, (D) HGD, (E) HGD+IEC. Arrows in (A, B) highlight areas of C5b-9 staining that are not associated with nuclei. Arrows in (C) show areas positive for C5b-9 that are associated with nuclei and are likely cellular: these cells are weakly positive for C9. Arrows in (D) highlight two areas with similar C9 deposition, but different C5b-9 formation: C9 positive cells are not always strongly positive for C5b-9. E demonstrates the strong epithelial C9 staining in cystically dilated glands. Scale bar is 50µm.

Figure 3

Time course of C5b-9 puncta formation in Barrett’s esophagus and esophageal adenocarcinoma cell lines after exposure to immunopurified C9. Cell lines representing non-dysplastic BE (BAR-T, CP-A), high-grade dysplasia (CP-B, CP-D) and EAC (FLO-1 and OE-33) were incubated in serum-free medium for 4 hours before being incubated in medium containing C9-depleted human serum for 30 minutes (A), or medium containing C9-depleted human serum plus 3 µg/ml purified human C9 for 0, 30 or 75 minutes (B). Coverslips were removed and stained with C5b-9 antibody and imaged by confocal microscopy. Maximum intensity projection of C5b-9 signal at 0, 30 and 75 min in all cell lines, cell nuclei are represented by blue (DAPI) and C5b-9 by green. Images are all at the same magnification. All images ( Figure S4 ) were used to quantify the average number of puncta per cell (C, D) and the area positive for C5b-9 per cell (E, F) in untreated or cells incubated for 30 minutes in C9 depleted human serum (C, E), or with C9 for 0, 30 or 75 minutes (D, F). *P<.05. **P<0.01. ***P<0.001. ****P<0.0001.

Figure 4

Cell proliferation and death of esophagus and esophageal adenocarcinoma cell lines with and without C9 exposure. Cultured CP-B (A, B), CP-D (C, D), FLO-1 (E, F) and OE33 (G, H) cells were seeded into 96 well plates with and without C9 present (3µg/ml). Incucyte time-lapse videos were used to capture cell proliferation (B, D, F, H) and death (A, C, E, G) (by propidium iodide staining). Two way ANOVA with Sidak’s multiple comparison test was performed, and showed no significant effect of C9 on either proliferation or cell death (p values for C9 effect shown in figure). FLO-1 cells reach 100% confluence by 60hr, indicated by vertical line. No other cell lines reached confluence during the experiment.

Figure 5

Subcellular localization of C5b-9 in Barrett’s esophagus and esophageal adenocarcinoma cell lines after exposure to purified C9 in culture. After 75 minutes of C9 exposure of CP-B (A, B) or FLO-1 (C, D) cells as described for Figure 3 , co-staining of C5b-9 (green) was conducted with wheat germ agglutinin (WGA, red) to visualize the plasma membrane. The nucleus was stained with DAPI (blue). C5b-9 was observed at the cell surface (A, C), as well as intracellularly (B, D). These images represent one slice of Z-stack images acquired by confocal microscopy, adjoining Z-stack images are shown in Figure S4 . A separate experiment with 120 minutes of C9 exposure was used for co-staining with the lysosome marker LAMP-1 (red) in CP-B (E) and FLO-1 (F). Yellow color in the combined images indicates colocalization between the red and green signals.

Figure 6

Barrett’s esophagus and esophageal adenocarcinoma cell lines release extracellular vesicles containing C9 and C5b-9 after exposure to C9. BAR-T, CP-B and FLO-1 cells were exposed to C9-depleted serum plus 3 μg/ml C9 for 0 or 30 min as described in Figure 3 . The medium were collected and ab EV-capture SERS immunoassay performed to analyze C9 and C5b-9 on EV. (A–D) Schematic of EV-SERS assay: (A) Conditioned cell culture medium is added to (B) anti-CD9/CD63/CD81 antibody functionalized microelectrodes to capture C9⁺ and C5b-9⁺ EVs. (C) Captured EVs are detected with SERS nanotag-labelled anti-C9 or anti-C5b-9 antibody. In (B, C), the application of an alternating current electric field on the microelectrodes induces a nanoscopic fluid flow (black arrow) that stimulates collisions of EVs with antibody-functionalized microelectrode surface. (D) SERS mapping reveals the expression levels of C9 and C5b-9 on EVs. (E, F) SERS assay measured extracellular vesicles that are positive for C9 (E) or C5b9 (F), n=3. Controls include conditioned medium from cells exposed to C9 depleted serum (DEPL) for 30 minutes without addition of C9, as well as unconditioned medium (KSFM and DMEM) with depleted serum measured by 2 way ANOVA and Tukey’s multiple comparison test (*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001).