Accumulation of neoplastic traits prior to spontaneous in vitro transformation of rat cholangiocytes determines susceptibility to activated ErbB-2/Neu

Abstract

Cholangiocarcinoma, a severe form of biliary cancer, has a high mortality rate resulting partially from the advanced stage of disease at earliest diagnosis. A better understanding of the progressive molecular and cellular changes occurring during spontaneous cholangiocarcinogenesis is needed to identify potential biomarkers for diagnosis/prognosis or targets for novel therapeutics. Here, we show that with continued passage (p) in vitro, rat bile duct epithelial cells (BDEC) accumulated neoplastic characteristics that by mid-passage (p31-85) included alterations in morphology, increased growth rate, growth factor independence, decreased cell adhesion, loss of cholangiocyte markers expressed at low passage (p<30), and onset of aneuploidy. At high passage (p>85), BDEC cultures showed increasing numbers of cells expressing activated, tyrosine phosphorylated ErbB-2/Neu, a receptor tyrosine kinase previously reported to be at elevated levels in cholangiocarcinomas. Enrichment for high passage ErbB-2/Neu-positive cells yielded several anchorage-independent sub-lines with elevated levels of activated ErbB-2/Neu and increased expression of cyclooxygenase-2 (COX-2). When injected into immunodeficient beige/nude/xid mice, these sub-lines formed poorly differentiated cystic tumors strongly positive for rat cholangiocyte markers, a finding consistent with a previous report showing the susceptibility of high passage, non-tumorigenic BDEC to transformation by activated ErbB-2/Neu. Mid passage BDEC, in contrast, were resistant to the transforming activity of activated ErbB-2/Neu and remained anchorage dependent in vitro and non-tumorigenic in vivo following stable transfection. Based on these findings, we concluded that during progression to high passage, cultured BDEC undergo preneoplastic changes that enhance their susceptibility to transformation by ErbB-2/Neu. The ability to generate cells at different points in the process of spontaneous neoplastic transformation offers a valuable model system for identifying molecular features that determine whether over-expression of activated ErbB-2/Neu is necessary and sufficient to induce neoplastic conversion.

Fig. 1

BDEC undergo changes in morphology and growth with increasing passage. (A-C) Phase contrast micrographs show morphological changes in BDE cultures at low (A; p21), mid (B; p66) and high (C; p135) passage. Images were taken when cultures had reached 50-60% confluence. Low passage cultures formed tightly compacted colonies containing polygonal shaped cells (A). In contrast, high passage cells became more spindle shaped and less tightly packed with increased passage (C). (E,F) The number and distribution of cytoplasmic vesicles decreased with increasing passage number. In low passage BDEC (p21) cytoplasmic vesicles were arranged in a well-ordered array proximal to the cell surface membrane (E). In contrast, cytoplasmic vesicles in mid and high passage cells were decreased in number and less organized as compared to low passage cells (F). (D) Phase contrast shows there are no significant morphological differences between low passage BDEC (A) and low passage BDEC nucleofected with the constitutively activated transmembrane domain mutant of ErbB-2/Neu (ErbB-2/Neu-M). (G) Growth curves show that high (p99) passage BDEC display a 2.5-fold increase in growth rate relative to low (p29) passage cultures. High passage cells also showed resistance to the effects of serum starvation. Low passage cells underwent growth arrest under similar conditions. Scale bar represents 50 μm.

Fig. 2

Decreased E-cadherin expression in high passage cells correlates with decreased intercellular adhesion. (A) Immunoblot analysis of cell lysates from low (p29) and high (p99) passage BDEC showed reduced expression of E-cadherin in high passage cells. The anti-Hsp70 antibody was used to demonstrate equal loading of cell lysates. (B) Aggregation of single cell suspensions of low (p26) and high (p99) passage BDEC was determined at time points ranging from 0 to 90 minutes. Reduced expression of E-cadherin by high passage cells (A) closely correlated with decreased intercellular adhesion as evidenced by the formation of large cell aggregates in low but not high passage cultures.

Fig. 3

Indirect immunofluorescence shows loss of ductal markers with increasing passage. Comparison of low (p18) and high (p88) passage BDEC showed decreased expression of ductal markers OV-6, CK19 and BD.1. E-cadherin expression was also reduced in high passage cells as shown also in Fig. 2 by immunoblot analysis. Expression of the ductal marker OC.5 remained constant with increasing passage. Scale bar represents 50 μm.

Fig. 4

Karyotypic analysis of BDEC showed rapid escalation in the degree of aneuploidy with increasing passage. Low passage BDEC (BDE4 (p14) and BDE1.1 (p23)) showed a near normal diploid karyotype with chromosome counts ranging from 42-44. Chromosomal instability was evident by mid passage (p37) as 26% of the spreads were diploid, 63% of spreads had 85 chromosomes and 11% of the spreads had 64-65 chromosomes. In contrast, high passage anchorage-independent clones had acquired a triploid karyotype with 100% of metaphase spreads showing 64-65 chromosomes. A total of 20 metaphase spreads were prepared for each passage.

Fig. 5

ErbB-2/Neu expression is increased in high passage BDEC. (A) Indirect immunofluorescence demonstrated that ErbB-2/Neu expression in high passage BDEC (p99) was upregulated relative to low passage (p19) cells. (B) Quantitative analysis of ErbB-2/Neu (Ab-4) membrane staining using the two-tailed Student's T test with 99% confidence intervals demonstrated that the mean fluorescence of low and high passage cells was significantly different (p ≥0.0001). The mean values were: BDEC p99, 29.85± 0.926, BDEC p19, 20.03 ±0.208. Scale bar represents 25 μm.

Fig. 6

High passage anchorage-independent BDEC were tumorigenic when injected into immunodeficient beige/nude/xid mice. (A, B) Tumors produced in beige/nude/xid mice from two soft agar (SA) anchorage-independent sublines, SA 3.1 (p130) and SA 10.3 (p123), are shown as representative examples of tumorigenicity observed at one week following injection. (C) When mid-passage BDEC (p78) were injected into immunodeficient mice, only small nodules were detected after four weeks.

Fig. 7

Cloned high passage anchorage-independent BDEC sub-lines form desmoplastic ductal cholangiocarcinomas when injected into immunodeficient mice. (A) Histological characterization by hematoxilyn and eosin staining demonstrates clearly defined ducts throughout the tissue, marked by arrows. (B-F) Phenotypic characterization of tumors by indirect immunofluorescence. (B) Double-labeling for the ductal marker OC.5 (green) and phosphorylated histone 3 (red) demonstrated actively proliferating ducts containing BDEC. (C-F) Ductal cells in tumors were strongly positive for plasma membrane ErbB-2, (C; Neu C-18), phosphorylated ErbB-2 (E; p-neu Y1248) and nuclear COX-2 (F). The inset in panel F shows a magnification of the region outlined in the rectangle. (D) The DAPI image that corresponds to Panel C defined a distinct morphology of ductal cells relative to surrounding epithelial tissue. Scale bars represent 50 μm.

Fig. 8

High passage anchorage-independent BDEC express activated ErbB-2/Neu and COX-2 proteins. (A) FACS analysis for ErbB-2 (Neu F-11) expression in low (p20), mid (p66), and anchorage-independent BDEC sub-lines. 13% of low passage BDEC (p20) expressed detectable levels of ErbB-2/Neu protein. In contrast, 80-97% of high passage, tumorigenic sub-lines expressed ErbB-2/Neu showing that these tumorigenic sub-lines expressed a 6-24-fold increase in ErbB-2/Neu expression relative to low passage cultures. Mid passage cells (p66) had a 3-fold increase in both the percentage of cells expressing ErbB-2 and fluorescence intensity as compared to low passage BDEC. (B) FACS analysis for COX-2 expression in low, mid, and high passage BDEC. While 57% of low passage BDEC (p20) expressed COX-2, the percentage of cell expressing COX-2 in tumorigenic sub-lines was between 70-99%, although this increase in positive cells did not always correlate with an increase in expression levels. Mid-passage BDEC also displayed an increase in the number of cells positive for COX-2. (C) Cell lysates from low passage BDEC (p35) and anchorage-independent sub-lines were analyzed by immunoblot for ErbB-2/Neu (C-18) and phosphorylated ErbB-2/Neu (p-Neu Tyr 1248). Activated, phosphorylated ErbB-2/Neu expression was elevated in tumorigenic sublines (lanes 2-6) as compared to low passage BDEC (lane 1) and unselected, nontumorigenic high passage anchorage-independent BDEC (SA.NT; lane 7). Hsp70 expression demonstrates equal loading between lanes.

Fig. 9

Expression of activated ErbB-2 is not sufficient to induce transformation of low or mid pass BDEC. (A) RT-PCR shows that transfected low passage BDEC (p22) express activated ErbB-2/Neu transcript (lane 4). ErbB-2/Neu expression was not detected in non-nucleofected low pass BDEC (lane 1), high pass BDEC (lane 2) or cultured hepatocytes (lane 3). Indirect immunofluorescence confirmed protein expression of activated ErbB-2/Neu (Neu C-18-G) in nucleofected low pass BDEC (C). ErbB-2/Neu expression was not detected in control non-nucleofected low passage BDEC (B). OC.5 expression, shown as a control, was not affected by nucleofection (D, E). Scale bar represents 50 μm.