In vivo lineage tracing defines the role of acinar-to-ductal transdifferentiation in inflammatory ductal metaplasia

Abstract

Background & aims: Chronic injury results in regeneration of normal pancreatic tissue and formation of a metaplasia of ductal phenotype. Metaplastic ductal lesions are seen in pancreatitis as well as in specimens of pancreatic cancer and are thought to represent a condition with increased risk of neoplasia. Acinar-to-ductal transdifferentiation is thought to be the source of this metaplasia. This has been suggested for flat duct-like lesions called tubular complexes and for lesions exhibiting a mucinous metaplasia. However, available studies are based on interpretation of static data rather than on direct evidence. Transdifferentiation from acinar to ductal cells has never been confirmed in the adult pancreas.

Methods: Here, we use Cre-loxP-based genetic lineage tracing in vivo to investigate whether transdifferentiation of acinar cells contributes to regeneration and metaplasia in pancreatitis.

Results: The results show that transdifferentiation does not play a role in regeneration of normal tissue. Acinar cells are regenerated by preexisting acinar cells and not from other cell types. Three different types of metaplastic ductal lesions are observed and analyzed. Whereas the majority of metaplastic lesions are not of acinar origin, acinar-to-ductal transdifferentiation is identified in a minority of mucinous metaplastic lesions.

Conclusions: Here, we provide direct evidence that acinar-to-ductal transdifferentiation occurs in the adult pancreas in vivo. However, it accounts for only a minority of metaplastic lesions.

Figure 1

Tagging in the ElastaseCreER™;Z/AP animal. Identification of HPAP by BCIP/NBT (blue) (A,C) and immunofluorescence (B,D). (A, B) HPAP is identified on the luminal membrane of acinar cells (arrowheads) and infrequently in islets (arrows) but not in the ductal epithelium (stars), including terminal ducts (arrowheads) and centroacinar cells (arrows) (C,D).

Figure 2

New acinar cells are formed from preexisting acinar cells. (A) Stable percentage of tagged acinar cells independent of the duration of pancreatitis. (B) BrdU incorporation (brown). (C,D) Identification of HPAP activity, blue (C) and immunofluorescence (D). In CP acinar cell density is decreased but the percentage of HPAP⁺ acinar cells remains stable. HPAP is not identified in ducts (arrows in C,D). (E) BrdU incorporation, 2 hrs (red). In the normal pancreas (left panel), proliferative activity is identified in single differentiated acinar cells (arrowhead) expressing amylase (green). In response to inflammation (right panel), proliferation is frequently identified in acinar cells (arrowheads).

Figure 3

TC1 are acinar cell-derived. (A) Quantitative analysis of lineage tracing. Expected and observed numbers of HPAP⁺ events. (B) Normal acinar cells and TC1 (arrows) are labeled with HPAP (blue). Unlabeled duct-like cells in TC1 (arrowheads). (C,D) TC1 express both HPAP (green) and amylase (red) (arrow in C) like acini (arrowheads). With loss of cytoplasm, amylase expression is downregulated but HPAP persists (arrows in D). (E,F) Double staining for HPAP (green) and CK (red) reveals that TC1 contain cells that are, like terminal ducts (star in E), strongly CK⁺ but never tagged with HPAP. A clear border (arrows) between CK⁺ cells and acinar-derived cells is maintained both in lesions with beginning luminar dilation (E) and with wide lumen (F).

Figure 4

TC2 are not of acinar origin. (A) Quantitative analysis of lineage tracing. Comparison (*) of the number of expected HPAP⁺ events assuming acinar origin and the number of actually observed HPAP⁺ cells of ductal morphology. (B) HPAP activity (blue) is detected in acinar but not in duct-like cells within fields of TC and disappears with progressive loss of acinar cells. (C) The apoptosis marker cleaved caspase-3 (brown) is identified in acinar cells (arrowheads), but not in duct-like cells within fields of TC2. Fields without acinar cells are devoid of cleaved caspase-3 expression. (D) The epithelial lining of TC2 is not tagged with HPAP (green), but expresses the ductal marker CK (red). Note tagging of adjacent acinar cells (arrowheads). (E) Proliferation (BrdU-red) is frequently identified in CK⁺ (green) duct-like cells and occasionally in the stroma (arrowheads). (F) Hes-1 expression (brown). In controls, Hes-1 is identified in centroacinar cells (arrowheads), located near the acinar lumen (arrow) and in ductal cells (arrowhead, inset). In chronic injury, Hes-1 is almost always identified in centroacinar cells (arrows). Terminal ductal cells frequently express Hes-1 (arrowheads). TC2 mirror this expression pattern. Acini remain consistently negative for Hes-1.

Figure 5

Acinar-to-ductal transdifferentiation accounts for a minority of MML. Separate analysis for lesions of mixed (A-C) and pure ductal phenotype (D-F). In most lesions mucinous metaplastic cells are not tagged with HPAP (B,E); In a minority of mixed and pure lesions HPAP⁺ ductal cells can be identified (C,F). Mixed phenotype: (A) Quantitative analysis reveals that HPAP is identified in cells of acinar morphology with a frequency similar to that expected for acinar origin. A minority (5.8%) of mixed lesions additionally contain HPAP⁺ mucinous cells of ductal morphology. (B) Mixed lesions, in which mucinous duct-like cells are not of acinar origin. HPAP/Alcian blue stains show that only acinar cells are tagged (blue), whereas mucinous metaplastic cells (turquoise) are untagged. Note that the lesion has formed in the terminal ductal to acinar region, whereas the draining duct (arrows) exhibits normal morphology. Serial double immunofluorescence identifies HPAP (green) in amylase⁺ cells (red, upper panel), but not in mucinous metaplastic cells, which are positive for both CK (red, mid panel) and gastric mucins (red, lower panel). Secreted amylase can be identified at the luminar membrane of the metaplastic ductal lining (arrowheads). (C) Mixed lesions with evidence of acinar-to-ductal transdifferentiation. Serial stains show expression of both HPAP (blue) and Alcian blue-positive mucins (turquoise) in cells of ductal morphology. Serial double immunofluorescence shows co-expression of HPAP (green) and amylase (red) in acinar cells (arrowhead), and confirm HPAP expression in duct-like cells without amylase expression (arrows, upper panel). These HPAP⁺ cells instead express both CK (red, middle panel) and gastric mucins (red, lower panel), suggesting transdifferentiation. Pure ductal phenotype: (D) In lesions of pure ductal phenotype HPAP⁺ duct-like cells are observed with a frequency (4.9%) similar to mixed lesions. (E) Most lesions (turquoise) are HPAP^-. Double immunofluorescence confirms expression of HPAP (green, upper panel) in acini, but absence of HPAP in metaplastic lesions, which instead express gastric mucins (red, middle panel). There is no overlap between HPAP and mucin expression (lower panel). (F) Pure ductal lesions with evidence of acinar-to-ductal transdifferentiation: Serial stains show expression of both HPAP (blue) and Alcian blue-positive mucins (turquoise) in cells of ductal morphology. Double immunofluorescence of the same lesion confirms co-expression of HPAP (green) and the ductal marker CK (red).

Figure 6

(A) MML exhibit a mosaic expression pattern of Hes-1 (brown). (B) Expression of Shh (brown) is frequently identified in MML. (C) Ki-67 expression (brown) and (D) incorporation of BrdU (green) demonstrate proliferative activity. (E) Apoptosis-cleaved caspase-3 (brown) is infrequently identified in MML.