Supplemental Material - de Melo et al. - Vol. 55, pp 1015-1026

Online Supplemental data for de Melo et al.: Figure Image files.

Files in this Data Supplement:

• de Melo et al., Supplemental Figure 1 - An overview of galectin-3 ligand distribution in normal human tissues and in selected types of tumors. Presence of galectin-3 ligands was evaluated using the chimeric probe Gal3/AP, further developed with Fast Red. Thus, red staining in the photomicrographs displayed here indicates presence of galectin-3 ligands within tissues and tumors. Lines designed as numbers represent a kind of neoplasia, i.e., line 1 is breast carcinoma; 2, prostate carcinoma; 3, pancreatic carcinoma; 4, squamous cell carcinoma; 5, melanoma; 6, neuroendocrine tumors; 7, neuroblastoma; 8, gliomas. Line 9 represents a control for the endogenous alkaline phosphatase activity. In these samples, Gal3/AP was not added. For lines 1 to 8, the first photomicrograph (A) illustrates a reaction in which Gal3/AP was incubated with tissues in the presence of 100 mM lactose. Photomicrographs represent respectively; (1A) normal breast tissue, (2A) normal prostatic tissue, (3A) normal pancreatic tissue, (4A) normal skin, (5A) normal retina, illustrating pigmented cells, (6A) neuroendocrine neoplasia, (7A) neuroblastoma and (8A) normal brain tissue (8A). Supplemental figures 2 to 9 explore in more detail the very same tissues and tumors.
• de Melo et al., Supplemental Figure 2 - Normal breast and breast carcinoma. (A) Negative control performed by incubating tissues with Gal3/AP in the presence of lactose. (B) In normal breast tissue galectin-3 ligands were expressed by ductal cells, inflammatory cells and endothelium, but not expressed by other stromal cells. Note that ductal cells have a luminal and apical pattern of expression, suggesting that galectin-3 ligands are secreted into lumen of the glands. This finding was also observed in well-differentiated breast carcinomas (example in C). In moderately to poorly differentiated carcinomas, galectin-3 expression tended to decrease, varying from moderate intensity to complete loss of staining. When observed, staining was cytoplasmic, with loss of apical or luminal polarity, as observed in well-differentiated carcinomas. As shown in D, it was frequent to observe loss of staining in tumor cells, whereas inflammatory cells stained intensely with Gal3/AP. Similar findings are depicted in photomicrographs E and F, with the exception that tumor cells had a diffuse cytoplasmatic expression of galectin-3 ligands in E.
• de Melo et al., Supplemental Figure 3 - Normal prostate and prostate carcinoma. (A) Negative control performed by incubating tissues with Gal3/AP in the presence of lactose. (B) In normal prostatic tissue, galectin-3 ligands were expressed by epithelial cells. Galectin-3 ligands were found within the cytoplasm and also associatated with apical membranes and secreted (intraluminal secretion). Well-differentiated carcinomas, as depicted in photomicrographs C and E, keep the same expression pattern shown in normal tissue. Note that in these cases, the stroma was negative and the inflammatory reaction was discrete, if present at all. As in the less differentiated cases depicted in photomicrographs D and F, galectin-3 ligands were found diffusely in the cytoplasm, when not lost. Intriguingly, loss of galectin-3 ligand expression in tumor cells was often accompanied by increased inflammatory infiltrate, which stained with gal3/AP (as shown in F).
• de Melo et al., Supplemental Figure 4 - Pancreatic Carcinoma. (A) Negative control performed by incubating tissues with Gal3/AP in the presence of lactose. (B) Normal pancreatic tissue showed a luminal staining pattern, with galectin-3 ligands present within the acinar secretion, stromal cells did not stain for galectin-3 ligands. (C) Well-differentiated papillary pancreatic carcinomas showed a strong cytoplasmatic staining, with apical distribution of galectin-3 ligands. Gal3/AP stained variably moderately to poorly differentiated pancreatic carcinomas (E and F), which showed both a cytoplasmic and diffuse pattern of expression of galectin-3 ligands. Inflammatory cells were strongly stained within the cytoplasm. As depicted in D, inflammation was inversely correlated with the presence of galectin-3 ligands in tumor cells.
• de Melo et al., Supplemental Figure 5 - Squamous cell carcinoma. (A) Negative control performed by incubating tissues with Gal3/AP in the presence of lactose. (B) In normal skin there was not expression of galectin-3 ligands by squamous cells. Diffuse and weak focal expression of galectin-3 ligands were observed in stromal and inflammatory cells. (C) Well-differentiated carcinomas expressed galectin-3 ligands in a diffuse cytoplasmic pattern, preferentially in the basal cell layer of the epidermis. In moderately to poorly differentiated carcinomas (D-F), the degree of expression of galectin-3 ligands increased, keeping the same pattern, i.e., diffuse cytoplasmic staining. Inflammatory reaction was variable (discrete to absent in D, intense in E and F). Noteworthy is the inverse correlation between tumor cell expression of galectin-3 ligands and the intensity of the peritumoral inflammatory infiltrate.
• de Melo et al., Supplemental Figure 6 - Melanoma. (A) Negative control performed by incubating tissues with Gal3/AP in the presence of lactose. The brown pigment represents melanin. (B) The normal retina showed a negative expression of galectin-3 ligands by melanocytes of the pigmented epithelium. In melanomas (here depicted as cutaneous melanomas), galectin-3 ligand expression was highly variable (C-F). C and D represent respectively a pigmented and an amelanotic melanoma with similar expression of galectin-3 ligands and weak to absent inflammatory infiltrate, whereas E and F depict melanomas which express virtually no galectin-3 ligands and were intensely infiltrated by strongly Gal3/AP reactive inflammatory cells.
• de Melo et al., Supplemental Figure 7 - Neuroendocrine tumors. (A) Negative control performed by incubating tissues with Gal3/AP in the presence of lactose. (B) Cells from the solid type neuroendocrine neoplasia showed a strong and diffuse expression of galectin-3 ligands, with scattered positive inflammatory cells within the tumor. (C and D) When tumors presented a secretory pattern, galectin-3 ligands were variably expressed and were either found in microacinar areas (C) or completely absent (D). Consistently with all the other tumors, as the solid type neuroendocrine neoplasias had an intense stromal reaction, tumor cell reactivity with Gal3/AP decreased markedly (E and F) followed by intense infiltration of strongly reactive inflammatory cells (F).
• de Melo et al., Supplemental Figure 8 - Neuroblastoma. (A) Negative control performed by incubating tissues with Gal3/AP in the presence of lactose. (B-F) The expression of galectin-3 ligands was correlated with neither type nor degree of differentiation of neuroblastomas. While the ganglionar component of neuroblastomas displayed galectin-3 ligands (B), the schwanomatous stroma did not (C). Poorly differentiated neuroblastoma, ganglioneuroma and ganglioneuroblastoma did not express galectin-3 ligands (D, E and F respectively). Scattered gal3/AP positive inflammatory cells were found in ganglioneuroblastomas (F).
• de Melo et al., Supplemental Figure 9 - Gliomas. (A) Negative control performed by incubating tissues with Gal3/AP in the presence of lactose. (B) A diffuse expression of galectin-3 ligands was found through the normal brain tissue. Note the intense reactivity of endothelial cells. In low-grade gliomas (C and D), galectin-3 expression was highly variable. Examples of high expression (in C) and no expression (D) of galectin-3 ligands in gliomas are depicted. In high-grade gliomas (E and F), tumor cells were either negative or weakly reactive to gal3/AP, whereas infiltrating inflammatory cells were strongly reactive.