The Nestin progenitor lineage is the compartment of origin for pancreatic intraepithelial neoplasia

Abstract

To determine the cell compartment in which initial oncogenic mutations occur in pancreatic ductal adenocarcinoma (PDAC), we generated a mouse model in which endogenous expression of mutated Kras (Kras(G12D)) was initially directed to a population of pancreatic exocrine progenitors characterized by the expression of Nestin. Targeting of oncogenic Kras to such a restricted cell compartment was sufficient for the formation of pancreatic intraepithelial neoplasias (PanINs), putative precursors to PDAC. PanINs appeared with the same grade and frequency as observed when Kras(G12D) was targeted to the whole pancreas by a Pdx1-driven Cre recombinase strategy. Thus, the Nestin cell lineage is highly responsive to Kras oncogenic activation and may represent the elusive progenitor population in which PDAC arises.

Fig. 1.

Nestin cell lineage characterization and Nestin expression in the pancreas. (A–C) LacZ staining showing Nestin cell lineage tracing in Nestin-Cre/R26R mice. (A and B) Three- to four-month-old pancreas. Strong but heterogeneous staining is seen the in the acinar cells in some pancreatic segments, whereas weaker staining is seen in other segments; endothelial cells within blood vessels exhibit LacZ staining (B, open arrows), indicating that these cells also derive from Nestin-expressing progenitor cells, in contrast to the endocrine cells, which are always devoid of LacZ staining (B, star). (C) Three-month-old brain. All of the cells express LacZ. (D) GFP expression in the pancreas of a 2-month-old Nestin-GFP mouse is observed in discrete acinar (arrows) and endothelial/mesenchymal cells (arrowheads) throughout the pancreatic parenchyma; some endothelial cells (F Inset) within the islets are GFP-positive, whereas the endocrine cells are always devoid of a GFP signal. (E and F) Hes-1 immunolabeling on 3-month-old Nestin-Cre/R26R LacZ-stained pancreas. Note that the Hes-1-labeled cells (light brown, open arrows) are surrounded by a white ring, showing that they are not LacZ-stained. (G and H) Serial sections of a LacZ-stained acinus and its associated unstained ductule (open arrows), with (H) D. biflorus agglutinin staining revealing that the lectin binds to the duct cells and the centroacinar cells that are clearly devoid of LacZ staining. (Magnification: A, ×200; B, ×400; C, ×50; D, ×400; D Inset, ×200; E–H, ×600.)

Fig. 2.

Conditional activation of Kras in the Nestin cell lineage is sufficient for generating mPanINs. (A) Breeding strategy of LSL-Kras^G12D mice with Nestin-Cre mice. The Cre recombinase expressed under the control of Nestin-regulatory elements will mediate the recombination/excision of the LSL specifically in the Nestin-expressing cells. Subsequently, Kras will be conditionally activated only in the Nestin cell lineage. (B) RT-PCR/RFLP was performed on a 4-month-old pancreas and brain to confirm Kras^G12D expression based on a Kras^G12D allele-specific HindIII site. PCR-amplified cDNA was undigested (−) or digested (+) with HindIII, generating the mutant amplicon that migrates as the lower band in relation to the wild-type product. Lanes 1 and 2 are Nestin-Cre/LSL-Kras^G12D (KxN) brain cDNAs, lanes 3 and 4 are wild-type brain cDNAs, lanes 5 and 6 are wild-type pancreas cDNAs, and lanes 7 and 8 are KxN pancreas cDNAs. There are similar levels of wild-type and mutated Kras products in the brain, whereas there is a preponderance of the mutated allele in the pancreas in 5 of 5 mutant mice. (C) mPanINs in Nestin-Cre/LSL-Kras^G12D compound mutant mice. (Ci and Cii) Two-month-old pancreas. (Ci) Focus of PanINs1-A. (Cii) Normal ducts (arrowheads) can be observed next to a PanIN1-B (12). (Ciii–Cvi) Four-month-old pancreas. (Ciii) Isolated PanINs (arrows) can be observed, but the acini and the islets (star) are not affected. (Civ and Cv) PanIN1-B, where the transition between normal (arrowheads) and abnormal ductal epithelium is clearly observed. (Cvi) multiple PanIN foci are observed. (Magnification: Ci, Cii, Civ, and Cv, ×400; Ciii and Cvi, ×200.)

Fig. 3.

Characterization of PanINs in 4-month-old Nestin-Cre/LSL-Kras^G12D pancreas. (A) CK19 staining confirms the epithelial origin of the PanINs. (B) Alcian blue staining reveals the abundant mucin content of the lesions as confirmed by strong Muc5a immunoreactivity (C), a high level of Ki67 observed in PanIN 1-A and surrounding inflammatory cells (arrowheads) but no staining observed in adjacent normal acini (D), Cox2 expression limited to the abnormal areas of the ductal epithelium (arrows) (E), but Gli3 (F) staining the nuclei of all of the cells in the PanINs as well as in the adjacent inflammatory (solid arrowheads) and some acinar (open arrowheads) cells. (Magnification: ×400.)

Fig. 4.

Kras^G12D targeting to the Nestin cell lineage generates a more specific model for PanINs formation. (A) Less acinar–ductal metaplasia observed at 6 months in the pancreas of Nestin-Cre/LSL-Kras^G12D mutants as shown by comparison of CK19 staining between Nestin-Cre/LSL-Kras^G12D (A1) and Pdx1-Cre/LSL-Kras^G12D (A2) pancreas. Note the abundance of acinar-like structures that stain for CK19 in the Pdx1-Cre/LSL-Kras^G12D pancreas, next to normal ducts (arrowhead). (B) Cell lineages determination model. During pancreas development, multiple progenitor populations with progressively more restricted differentiation potential are generated. We propose that the progenitor population characterized by Nestin expression (in gray) specifically gives rise to exocrine pancreas; that in the adult pancreas, the Nestin-expressing cells function as adult progenitor cells; and that this cell population could be the cell of origin for PDAC.