Calcium-Induced Differentiation of Human Colon Adenomas in Colonoid Culture: Calcium Alone versus Calcium with Additional Trace Elements

Abstract

Previous murine studies have demonstrated that dietary Aquamin, a calcium-rich, multi-mineral natural product, suppressed colon polyp formation and transition to invasive tumors more effectively than calcium alone when provided over the lifespan of the animals. In the current study, we compared calcium alone to Aquamin for modulation of growth and differentiation in human colon adenomas in colonoid culture. Colonoids established from normal colonic tissue were examined in parallel. Both calcium alone at 1.5 mmol/L and Aquamin (provided at 1.5 mmol/L calcium) fostered differentiation in the adenoma colonoid cultures as compared with control (calcium at 0.15 mmol/L). When Aquamin was provided at an amount delivering 0.15 mmol/L calcium, adenoma differentiation also occurred, but was not as complete. Characteristic of colonoids undergoing differentiation was a reduction in the number of small, highly proliferative buds and their replacement by fewer but larger buds with smoother surface. Proliferation marker (Ki67) expression was reduced and markers of differentiation (CK20 and occludin) were increased along with E-cadherin translocalization to the cell surface. Additional proteins associated with differentiation/growth control [including histone-1 family members, certain keratins, NF2 (merlin), olfactomedin-4 and metallothioneins] were altered as assessed by proteomics. Immunohistologic expression of NF2 was higher with Aquamin as compared with calcium at either concentration. These findings support the conclusions that (i) calcium (1.5 mmol/L) has the capacity to modulate growth and differentiation in large human colon adenomas and (ii) Aquamin delivering 0.15 mmol/L calcium has effects on proliferation and differentiation not observed when calcium is used alone at this concentration. Cancer Prev Res; 11(7); 413-28. ©2018 AACR.

Conflict of interest statement

Disclosure of Conflict of Interest: All the authors have no financial or personal conflict of interest.

©2018 American Association for Cancer Research.

Figures

Figure 1. Adenoma colonoid appearance in culture

Figure 1A: Phase-contrast and scanning electron microscopy. At the end of the incubation period, virtually all of the adenoma colonoids maintained in 0.15 mM calcium consisted of a core structure with multiple tiny buds on the surface as indicated by phase-contrast microscopy (Figure 1, panel a). Colonoids maintained in 1.5 mM calcium (Figure 1, panel c) or treated with Aquamin® to provide either 0.15 mM (Figure 1, panel b) or 1.5 mM calcium (Figure 1, panel d) had a smooth surface with few buds. Bar=200μm. Scanning electron microscopic images confirmed the presence of multiple tiny buds growing out from the colonoid core structure under low-calcium conditions (asterisk) (Figure 1, panel e) and the presence of fewer but larger buds in the colonoid maintained in 1.5 mM calcium (asterisk) (Figure 1, panel f). Bars=100μm. Figure 1B: Histology and transmission electron microscopy. At the end of the incubation period, colonoids were examined under light-microscopy after sectioning and staining with hematoxylin and eosin. Under control conditions (Figure 1, panel g), tiny crypts of approximately 8-20 cells in cross section were seen. In the presence of 1.5 mM calcium alone (Figure 1, panel i) or with Aquamin® providing 1.5 mM calcium (Figure 1, panel j), larger crypts made up of columnar epithelial cells surrounding a large, often irregular-in-shape lumen were seen. Goblet cells were apparent. Colonoids treated with Aquamin® providing 0.15 mM calcium contained a mix of tiny crypts and larger structures (Figure 1, panel h). Bar= 50μm. When examined by transmission electron microscopy, the tiny crypts in low-calcium medium were found to consist of cuboidal cells surrounding a tiny central lumen. No desmosomes or tight junctures were seen (Figure 1, panels k and l). Under high-calcium conditions, the large crypts were made up of columnar epithelial cells around a larger lumen. Numerous desmosomes (black arrows) along the lateral surface connected by visible intermediate filaments and a tight-juncture at the apical surface (white arrow) were seen (m and n). Bar=2μm in Figure 1, panel k and m; Bar=500nm in Figure 1, panel l and =100nm in Figure 1, panel n. Figure 1C (Left panel): Percentage of individual colonoids with few tiny buds and smooth surface (phase-contrast). Means and standard deviations based on 20 fields per condition with 10-20 colonoids per field at 200X. Asterisks indicate statistical significance from control at p<0.05 level. Figure 1C (Right panel): Number of buds per core structure (hematoxylin and eosin sections). Means and standard deviations based on 5-10 high power fields per condition with 10-20 colonoids per field at 200X. Asterisks indicate statistical significance from control at p<0.05 level.

Figure 2. Proliferation marker expression

At the end of the 30-day incubation period, adenoma colonoids were examined after immunostaining. Figure 2A: Ki67 expression by immunohistology. The tiny crypts (regardless of condition) were almost universally-positive for Ki67 (Figure 2, panels a-d). The large crypts contained a mix of Ki67-positive and Ki67-negative cells. Bar=50μm. Figure 2B: Percentage of Ki67-positive nuclei. Means and standard deviations based on 15-25 individual crypts per condition. Asterisks indicate statistical significance from control at pInsert: A markup image exemplifying color coding of Ki67 stained nuclei in all four conditions (nuclear algorithm v9) is shown. Figure 2C: Confocal immunofluorescence microscopy. The majority of Ki67- staining was in the tiny crypts in either 0.15 mM calcium (Figure 2, panel e) or 1.5 mM calcium (Figure 2, panel f). Where the tiny crypts were still visibly attached to the core structure (DAPI counterstain), staining was in outer-facing cells. Bar=100μm.

Figure 3. Differentiation and apoptosis marker expression

At the end of the 30-day incubation period, adenoma colonoids were examined after immunostaining. Figure 3A. Immunohistology. At the end of the 30-day incubation period, adenoma colonoids were examined under light-microscopy after immunostaining for CK20 (Figure 3, panels a-d), E-cadherin (Figure 3, panels e-h), occludin (Figure 3, panels i-l) and cleaved caspase-3 (Figure 3, panels m-p). CK20 was primarily diffuse and intracellular under low-calcium conditions, though a few cells were strongly positive (Figure 3, panel a). E-cadherin was diffuse and intracellular (Figure 3, panel e) while occludin was barely detectable and limited to the luminal surface (Figure 3, panel i). Cleaved caspase-3 was not detected (Figure 3, panel m). In the large crypts of the other three conditions, intense staining for CK20 was observed in many cells, though negative areas remained (Figure 3, panels b-d). With E-cadherin, both cell surface and intracellular staining were observed. The most intense surface staining was in cells at the lumen (Figure 3, panels f-h). With occludin, staining was present along the lateral surface between cells and at the luminal surface (Figure 3, panels j-l). Cleaved caspase-3 was not detected in colonoids maintained in 0.15 mM Aquamin® (Figure 3, panel n) but was seen in a few cells in the crypt wall (arrows) and in debris present in the lumens of crypts maintained in 1.5 mM calcium (Figure 3, panel o) or with Aquamin® (1.5 mM) (Figure 3, panel p). Bar=50μm. Figure 3B: Positivity (measured using Positive Pixel Value v9). Means and standard deviations based on 25-75 individual crypts per condition. Asterisks indicate statistical significance from control at p<0.05 level.

Figure 4. Proteomic profile in adenoma colonoids

At the end of the 30-day incubation period, lysates were prepared for proteomic analysis. Figure 4A: Venn plots showing proteins altered (increased or decreased) by an average of 1.8-fold or greater across all three adenomas in response to each intervention and the overlap between pairs of interventions. The scatterplots demonstrate quantitative relationships between individual proteins altered by each pair of interventions. Figure 4B. Overlap in proteins altered (increased or decreased) with each intervention by 1.8-fold or greater in each of the three adenomas. Figure 4C. NF2 (merlin): Values in the upper bar graph represent fold-change under each condition relative to control. Values are means and standard deviations based on the proteomic assessment of the three adenomas. Asterisks indicate statistical significance from control at p® at both 0.15 and 1.5mM but little response to calcium. Two asterisks indicate statistical significance relative to calcium at 1.5 mM and control. Insert: NF2 (merlin) stained colonoids. Bar=50μm. Figure 4D. Metallothionines (MT1E and MT1H). Values represent fold-change and are means and standard deviations based on the proteomic assessment of the three adenomas. Both proteins were down-regulated in colonoids exposed to Aquamin® (at 0.15 and 1.5mM) and calcium at 1.5mM. Asterisks indicate statistical significance from control at p<0.05 level.

Figure 5. Colonoids derived from histologically-normal colon tissue

Phase-contrast (Figure 5, panels a-c). Colonoids maintained under all conditions consisted of a mix of thin-walled, translucent “cystic” structures and budding structures that resembled those seen in the adenoma cultures. Bar=200μm. Histology (Figure 5, panels d-f). At the histological level, crypts consisted of a single layer of epithelium surrounding a central lumen after staining with hematoxylin and eosin. Immunohistology (Figure 5, panels g-o). Immunostaining revealed a mix of Ki67-positive and -negative cells. Virtually all were strongly positive for CK20 but there was little or no staining for cleaved caspase-3. Bars=50μm.