NF1 loss of function as an alternative initiating event in pancreatic ductal adenocarcinoma

Abstract

A long-standing question in the pancreatic ductal adenocarcinoma (PDAC) field has been whether alternative genetic alterations could substitute for oncogenic KRAS mutations in initiating malignancy. Here, we report that Neurofibromin1 (NF1) inactivation can bypass the requirement of mutant KRAS for PDAC pathogenesis. An in-depth analysis of PDAC databases reveals various genetic alterations in the NF1 locus, including nonsense mutations, which occur predominantly in tumors with wild-type KRAS. Genetic experiments demonstrate that NF1 ablation culminates in acinar-to-ductal metaplasia, an early step in PDAC. Furthermore, NF1 haploinsufficiency results in a dramatic acceleration of Kras^G12D-driven PDAC. Finally, we show an association between NF1 and p53 that is orchestrated by PML, and mosaic analysis with double markers demonstrates that concomitant inactivation of NF1 and Trp53 is sufficient to trigger full-blown PDAC. Together, these findings open up an exploratory framework for apprehending the mechanistic paradigms of PDAC with normal KRAS, for which no effective therapy is available.

Keywords: CP: Cancer; alternative genetic drivers in pancreatic ductal adenocarcinoma; mosaic analysis with double markers; mutant KRAS; pancreatic ductal adenocarcinoma initiation; tumor-suppressor gene NF1; tumor-suppressor gene PML; tumor-suppressor gene TP53.

Figure 1.. NF1 genetic alterations in PDAC

(A) Oncoplot showing genetic coding alterations in NF1 and KRAS in human PDAC samples found in several public datasets, as described in STAR Methods. (B) Distribution of PDAC nonsense mutations identified in NF1. The recurrent mutations are highlighted in blue. (C) Numbers of overlapping and independent PDAC mutations in NF1 and KRAS of samples shown in (A). (D and E) NF1 and/or cytokeratin-19 (CK19) expression was analyzed by immunohistochemistry or co-immunofluorescence using three different human TMAs. Representative pictures are shown. Scale bars: (D) 25 and (E) 50 μm. (F) NF1 expression was analyzed by qRT-PCR using pancreatic tissues from control (2 months) and KC mice with progressive PanIN (2 months for PanIN1, 4 months for PanIN2, 6 months for PanIN3) and PDAC (12 months) lesions (n = 6; 3 females and 3 males). Data are expressed as the mean ± SEM.

Figure 2.. NF1 ablation leads to ADM

(A) Schematic representation of genetically engineered mouse models (GEMMs) of PDAC used in the experiments. (B) Immunoblot analysis of PDAC-relevant proteins in pancreatic extracts from NF1^KO mice and their wild-type littermates (n = 3; 2 females and 1 male). (C) Analysis of pancreatic tissues from control or NF1^KO mice at different ages (3, 6, 9, or 12 months) by hematoxylin and eosin (H&E) or immunohistochemistry using anti-CK19 antibody. Representative pictures are shown (n = 6; 3 females and 3 males). Scale bars: 50 μm. (D) Two-month-old control or NF1^KO mice were treated with vehicle or caerulein, and pancreas histology was analyzed by H&E or immunohistochemistry using anti-Sox9 antibody. Representative pictures are shown (n = 6; 3 females and 3 males). Scale bars: 50 μm.

Figure 3.. NF1 haploinsufficiency accelerates Kras^G12D-driven PDAC

(A) Pancreatic tissues from 6-week-old mice highlighted in the schematic (left) were analyzed by H&E or immunohistochemistry using antibodies to CK19 and Muc5AC (right). Representative pictures are shown (n = 6; 3 females and 3 males). Scale bars: 50 μm. (B) Kaplan-Meier analysis of survival of mice with the indicated genotypes (n = 12–21; 7–10 females and 5–11 males). (C) Pancreatic extracts from 6-week-old mice with the indicated genotypes were analyzed for pERK or pAKT. A representative immunoblot is shown on the left. Right: the graphs show the quantification of six samples from each genotype (n = 6; 3 females and 3 males). (D) PDAC growth in 2-month-old mice (3 females and 3 males) with the indicated genotypes highlighted by the scheme (left) was analyzed by IVIS bioluminescence. Representative luciferase images are shown (middle). Right: the graph shows the minimum and maximum of luciferase counts recorded in six experiments.

Figure 4.. NF1 restricts PDAC growth via a cell-autonomous mechanism

(A–C) Isogenic HPDE cell lines stably expressing various combinations of HA-Kras^G12D and gRNA targeting NF1 were analyzed for HA-Kras^G12D, NF1, pERK, ERK, pAKT, AKT, and β-actin expression by immunoblotting (A). Cells were passaged 12 times, and representative pictures of cells or spheres at different passages are shown (B). Cell proliferation was assessed either by cell counting at different time periods or by MTT after 96 h in culture (C). Data in (C) are expressed as the mean ± SEM.

Figure 5.. Association between NF1 and p53 in PDAC

(A) Oncoplot showing coding genetic alterations in NF1 and p53 in human PDAC samples found in public datasets. (B) Numbers of overlapping or independent PDAC mutations in NF1 and TP53 in samples shown in (A). TP53 mutational status is not known in the 14 samples with NF1 mutations. (C–E) NF1 or p53 expression was analyzed by immunohistochemistry (C and D) or co-immunofluorescence (E) using three different human PDAC TMAs. Representative pictures of normal areas, PanINs, and PDAC lesions are shown. Scale bars: 25 μm. The data shown in (D) represent the mean ± SEM of the percentage of overlapping samples with high or low levels of NF1 and p53 expression in the three TMAs (n = 3). The chi-square test was used to test the independence of the samples. (F) Expression of NF1 mRNA was analyzed by qRT-PCR using 3-month-old mice with the genotypes highlighted by the scheme (top). Mice were analyzed 4 weeks following treatment with vehicle or tamoxifen (Tam) (bottom). Data are expressed as the mean ± SEM (n = 6; 3 females and 3 males).

Figure 6.. p53 regulates NF1 gene expression

(A) MIA PaCa-2-Dox-HA-p53 cells were treated with Dox for the indicated times and analyzed for NF1 and HA-p53 expression by immunoblotting. (B) MEFs from PML^+/+ or PML^−/− mice were analyzed for the binding of PML to the NF1 promoter by ChIP and agarose gel. (C) MEFs from PML^+/+ or PML^−/− mice were transfected with the NF1^Luc reporter in the absence or presence of p53 mutants and analyzed for luciferase activity (n = 6). (D–G) PML was deleted from MIA PaCa-2-Dox-HA-p53 cells via CRISPR-Cas9, reconstituted with a CRISPR-Cas9-resistant PML mutant, and then treated with Dox for 48 h (n = 6). Cell extracts were analyzed for NF1 and HA-p53 expression by immunoblotting (D). Chromatin was analyzed for the binding of PML and p53 to the NF1 promoter by ChIP and agarose gel (E) or qPCR (F and G). (H–J) Pancreatic chromatin from control or KC mice with early PanIN (2 months) or advanced PDAC lesions (12 months) was analyzed for the binding of PML and p53 to the NF1 promoter by ChIP and agarose gel (H) or qPCR (I and J) (n = 6; 3 females and 3 males). Data in (C), (F), (G), (I), and (J) are expressed as the mean ± SEM.

Figure 7.. Concurrent ablation of NF1 and p53 culminates in PDAC formation and progression

(A) Schematic model of MADM^NF1/p53 conditional activation in pancreas. (B) Representative pictures of 10-month-old MADM and MADM^NF1/p53 mice (female and male) and their corresponding pancreas when MADM^NF1/p53 mice developed invasive PDAC. (C) FFPE pancreatic tissues from 10-month-old mice with the indicated genotypes were analyzed by H&E or immunohistochemistry using anti-CK19 antibody. Representative pictures are shown (n = 7–16; 3–9 females and 4–7 males). Scale bars: 25 μm. (D) Kaplan-Meier analysis of survival of mice with the indicated genotypes (n = 16; 7–9 females and 7–9 males). (E and F) FFPE pancreatic tissues from MADM^NF1/p53 mice at different ages (1–12 months) were subjected to H&E or DAPI staining and analyzed for GFP (green), TdTomato (red), and nuclei (blue). Representative pictures are shown. Scale bars: 50 μm (E). The green/red (G/R) ratio was determined by quantifying all cells in the section (F). Data are expressed as the mean ± SEM (n = 12; 6 females and 6 males). (G) Model depicting how NF1 interacts with Kras or p53 to influence PDAC pathogenesis and progression.