Early elevations of RAS protein level and activity are critical for the development of PDAC in the context of inflammation

Abstract

The KRAS^G12D mutation was believed to be locked in a GTP-bound form, rendering it fully active. However, recent studies have indicated that the presence of mutant KRAS alone is insufficient; it requires additional activation through inflammatory stimuli to effectively drive the development of pancreatic ductal adenocarcinoma (PDAC). It remains unclear to what extent RAS activation occurs during the development of PDAC in the context of inflammation. Here, in a mouse model with the concurrent expression of Kras^G12D/+ and inflammation mediator IKK2 in pancreatic acinar cells, we showed that, compared to KRAS^G12D alone, the cooperative interaction between KRAS^G12D and IKK2 rapidly elevated both the protein level and activity of KRAS^G12D and NRAS in a short term. This high level was sustained throughout the rest phase of PDAC development. These results suggest that inflammation not only rapidly augments the activity but also the protein abundance, leading to an enhanced total amount of GTP-bound RAS (KRAS^G12D and NRAS) in the early stage. Notably, while KRAS^G12D could be further activated by IKK2, not all KRAS^G12D proteins were in the GTP-bound state. Overall, our findings suggest that although KRAS^G12D is not fully active in the context of inflammation, concurrent increases in both the protein level and activity of KRAS^G12D as well as NRAS at the early stage by inflammation contribute to the rise in total GTP-bound RAS.

1. Introduction

PDAC is the major form of pancreatic cancer, a highly lethal disease that is projected to rise as the second leading cause of cancer-related mortality in the next decade [1]. The development of PDAC is a multifactorial process in which both genetic and environmental factors are at play [2–5]. The oncogenic KRAS mutations occur in as high as 90% of PDAC cases [6], in concordance with the essential role of mutant KRAS as a driver of pancreatic tumorigenesis [7–10]. However, KRAS mutations are not exclusive to cancer patients, suggesting that the mutations in KRAS per se are not sufficient for PDAC development [11, 12]. In support, studies using genetically engineered mouse models have shown that mutant KRAS alone in adults is insufficient to drive PDAC, underscoring the notion that a second hit is required for full-blown PDAC development [9, 13]. Furthermore, studies have shown that inflammatory insults, such as camostat, caerulein, lipopolysaccharides (LPS), and overexpression of IκB kinase 2/β (IKK2) or cyclooxygenase-2 (COX-2), rapidly accelerated the development of oncogenic KRAS-mediated pancreatic pathologies [4, 8, 9, 14–18]. It was thus reported that mutant KRAS and inflammation promote self-perpetuating positive feedforward loops that amplifies RAS activity to pathological levels sufficient for the development of PDAC [19–21].

The RAS gene family encodes the small guanine nucleotide-binding protein (small G protein, 21 kDa) with KRAS, NRAS, and HRAS as the most commonly mutated genes in human cancers. RAS proteins have GTPase activity and function as binary molecular switches that cycle between an active guanosine triphosphate (GTP)-bound state and an inactive guanosine diphosphate (GDP)-bound state. The GTP binding and activation of RAS signal proteins promote various downstream signaling pathways, such as the RAF-ERK1/2 pathway, PI3K-AKT pathway, and Ral-GEFs pathway, leading to an array of cellular responses including cell proliferation, survival, migration, and metabolic reprogramming [2]. It has long been thought that oncogenic RAS mutations, such as G12D and G12V in KRAS, serve to maintain RAS in a GTP-bound and constitutively active state, leading to uncontrolled cell growth and malignant tissue transformation [7, 10]. On the other hand, it has been shown that the activity of mutant KRAS is only a fraction of that predicted in a full GTP-occupied state and can be further enhanced by inflammatory stimuli in the pancreas [19]. Although inflammation and oncogenic KRAS cooperate to rapidly promote PDAC development, it remains unclear to what extent KRAS activity is necessary for the initiation and development of PDAC in this context.

Here we report that the expression of an endogenous level of mutant Kras^G12D/+ in pancreatic acinar cells for one month had negligible effects on both the abundance of KRAS protein and the pathogenesis of the pancreas, in comparison to the IKK2 mice and the control fElas^CreERT mice, although KRAS^G12D protein expression and activity were observed. Over an extended period, the expression of Kras^G12D/+ led to a 2.5-fold increase in KRAS^G12D protein levels with a slight rise in activity, resulting in an overall elevation of GTP-bound KRAS^G12D. This suggests that the increased levels of KRAS^G12D protein are the primary contributing factor. By contrast, when coupled with IKK2 overexpression, both the protein level and activity of oncogenic KRAS^G12D rapidly increased during the early induction stage, which together augmented the overall pool of GTP-bound KRAS^G12D that was maintained throughout the extended induction stage. This implies that, under inflammatory conditions, the protein abundance and activity of KRAS^G12D reach their maximal levels rapidly at the early stage. Notably, although inflammation is highly effective in promoting KRAS^G12D-mediated PDAC development with high penetrance, only a fraction of KRAS^G12D is in the GTP-bound active form. Furthermore, even in pancreatic cancer cells with homologous KRAS^G12D/G12D, less than 50% of KRAS^G12D in its total protein pool is in the GTP-bound active state. Additionally, the changes in the protein level and activity of NRAS, but not HRAS, followed a similar pattern to that of KRAS^G12D. Overall, our findings suggest that IKK2 and KRAS^G12D cooperate to rapidly elevate the protein level and activity of both KRAS^G12D and NRAS, contributing to an overall augmentation in the abundance of the GTP-bound KRAS^G12D and NRAS, leading to the extensive formation of PanIN lesions at the early stage. Nevertheless, a fully GTP-bound active state of KRAS^G12D is not observed during PDAC development in the context of inflammation.

2. Materials and methods

2.1. Genetically engineered mouse models

Kras^LSL-G12D/+ mice bearing the conditional knock-in of mutant Kras^G12D were obtained as described [21]. fElas^CreERT mice (called Cre after TM), which express tamoxifen-regulated Cre recombinase under a full-length Elastase promoter specifically in pancreatic acinar cells, were developed as previously described [22]. Kras^LSL-G12D/+ mice and fElas^CreERT mice were cross-bred to generate fElas^CreERT;Kras^LSL-G12D/+ double-transgenic mice (called Kras^G12D/+ after TM induction) for targeted expression of Kras^G12D/+ in pancreatic acinar cells at nearly 100% efficiency. Mice with the conditional expression of the constitutive active IKK2 mouse model (IKK2^LSL-f/f or IKK2), which bears a construct containing loxP-GFP-stop-loxP, were crossed with fElas^CreERT mice to generate fElas^CreERT;IKK2^LSL-f/f double-transgenic mice (called IKK2 after TM) for targeted expression of IKK2, has been described [8]. Kras^LSL-G12D/+;fElas^CreERT and IKK2^LSL-f/f;fElas^CreERT mice were cross-bred to generate fElas^CreERT;IKK2^LSL-f/f;Kras^LSL-G12D/+ triple-transgenic mice (called KIC after TM) for targeted expression of both IKK2 and Kras^G12D/+ in pancreatic acinar cells. All animal experiments were reviewed and approved by the Stony Brook University Institutional Animal Care and Use Committee (IACUC) and the University of Texas at El Paso IACUC.

2.2. Animal treatment

fElas^CreERT, IKK2, Kras^LSL-G12D/+, and IKK2;Kras^LSL-G12D/+ mice were given TM by intraperitoneal injection for five days to fully activate Cre recombinase in pancreatic acinar cells. All mice were kept in a standard animal facility condition with a temperature of 21 ± 2 °C, relative humidity of 50 ± 15%, and a 12 h light-darkness cycle with water and standard chow diet ad libitum. After one month or 10 months of TM induction, mice were euthanized, and pancreatic tissues were harvested for further experiments. To exam the impact of high-fat diet (HFD) on IKK2 expression, fElas^CreERT or Kras^LSL-G12D/+;fElas^CreERT mice were treated with TM at about 60 days of age to induce the expression of Cre or KRAS^G12D/+ and then randomly divided into two groups fed either a normal diet (ND) or an HFD for a duration of 10 weeks before being euthanized for pancreatic tissues for further analysis.

2.3. Pancreatic cancer cell lines

The human pancreatic cancer cell lines BxPC-3, PANC-1, SU86.86, and AsPC-1 were obtained from the American Type Culture Collection (ATCC). PANC-1 cells were derived from human patient-derived primary PDAC. AsPC-1 cells were from nude mouse xenografts initiated with cells from the ascites metastasis of a 62-year-old female Caucasian patient with pancreas adenocarcinoma. SU86.86 cells were derived from a liver metastasis of human PDAC. BxPC3 cells were from a 61-year-old female with moderate differentiation PDAC. All these cell lines were cultured in DMEM supplemented with 10% fetal bovine serum, 100 u/ml penicillin, and 100 ug/ml streptomycin (Invitrogen).

2.4. Pancreatic histopathological analysis

Mouse pancreatic tissues were quickly dissected and fixed in a 10% formalin solution overnight. Formalin-fixed tissues in plastic cassettes were dehydrated in alcohol gradients and embedded in paraffin for tissue sectioning. All 5-μm-thick tissue sections were dewaxed in xylene, rehydrated through reversed ethanol gradients, and processed to staining with H&E.

2.5. Immunohistochemistry

The immunohistochemical (IHC) staining experiment was performed on pancreatic sections as described [4, 18]. Briefly, the 5 μm-thick pancreas sections were deparaffinized, rehydrated, and then subjected to antigen retrieval in a citrate buffer (pH 6.0) in a pressure cooker for 8 minutes followed by blocking the endogenous peroxidases with 0.3% H₂O₂ for 15 minutes. Slides were then blocked using the normal serum for 30 minutes to prevent nonspecific binding and then incubated with primary antibodies for COX-2 (1:600; Cell Signaling #12282S), p65 (1:500; Cell Signaling #8242), CD3 (1:150; Santa Cruz Biotechnology sc-20047), Pax-5 (1:150; Santa Cruz Biotechnology sc-13146), Ly-6G (1:100; Santa Cruz Biotechnology sc-53515), IL-1a (1:150; Santa Cruz Biotechnology sc-9983), F4/80 (1:100; eBioscience^™ #14–4801-82) and IKK2 (1:200; Novus Biologicals NB100–56509SS) at 4°C overnight. After washing, the sections were incubated with the appropriate biotinylated secondary antibodies (Vector Laboratories, CA, USA) for 1 hr, washed again in PBS (Phosphate buffered saline) containing 0.05% Tween 20 (PBS-T), incubated with ABC reagent (Vector Laboratories, CA, USA) for 30 min, and then reacted with diaminobenzidine (DAB, Vector Laboratories, CA, USA). Sections were viewed on an Olympus IX70 microscope. The resulting sections were then counterstained with hematoxylin. Fiji ImageJ software and GraphPad Prism were used for quantification and statistical analyses.

2.6. Alcian Blue Staining

Alcian blue staining was performed for PanIN lesions as described [17, 23]. Briefly, pancreatic tissue sections were hydrated in distilled water, processed with 3% acetic acid, and incubated in Alcian blue solution (Sigma-Aldrich, Louis, MO, USA) for 45 minutes at room temperature, followed by a wash in running water and quick nuclear-fast red staining. Five random, non-overlapping images were obtained under microscopy at a magnification of x100. For each image, the Alcian blue positive area and total pancreatic area were scanned using Fiji ImageJ, and the percentage of the Alcian blue positive area was calculated.

2.7. Western blot analysis

To evaluate target protein levels, cell and tissue lysates were separated by SDS-PAGE and analyzed by immunodetection. Briefly, cell pellets or snap-frozen tissues were homogenized in ice-cold MLB (Mg²⁺ Lysis Buffer) (Millipore, MA, USA) containing complete Protease Inhibitor Cocktail (Roche, Germany). Tissue homogenates were centrifuged at 12,000 g for 10 minutes at 4°C, to collect the supernatant. Protein lysates from tissues were aliquoted to determine protein concentration using a protein assay dye reagent concentrate (Bio-Rad, CA, USA). The lysates were separated by SDS-PAGE and then transferred to Polyvinylidene difluoride (PVDF) membranes. The membranes were rinsed with TBS containing 0.05% Tween 20 (TBS-T) and probed with the following antibodies against IKK2 (1:1000; NB100–56509SS, Novus Biologicals), KRAS (1:200; sc-30; Santa Cruz), HRAS (1:200; sc-35, Santa Cruz), NRAS (1:200; sc-31, Santa Cruz), β-Actin (1:1000; sc-47778, Santa Cruz), RAS^G12D (mutant specific, 1:1000; #14429S, Cell Signaling Technology), and AKT (1:2000; #9272S, Cell Signaling Technology). The membranes were then washed with TBS-T and probed with the respective secondary antibodies conjugated to horseradish peroxidase for one hour at room temperature. Autoradiography or the Odyssey Imaging System (LiCor Biosciences, Lincoln, NE) was used to visualize protein bands. ImageJ densitometry software was used to quantify individual bands against an internal control protein.

2.8. RAS activity assay

Levels of active GTP-occupied RAS in cells and mouse pancreata were measured by using an RAF pull-down assay kit (Millipore, MA, USA) as previously described [19]. Briefly, cell pellets or snap-frozen pancreatic tissues were homogenized on ice in MLB buffer with a complete Protease Inhibitor Cocktail. Cellular debris was removed by centrifuging at 12,000g for 10 minutes at 4°C. Protein concentrations were determined. About 500 μg of lysates were incubated for 1 h at 4°C with agarose beads coated with RAF-RBD binding domain, and the beads were washed 4 times with wash buffer. Active RAS was analyzed by immunoblotting with K-, N-, H-RAS, and KRAS^G12D-specific antibodies. For comparison with the corresponding total RAS proteins, 20–40% of total lysates used for pulldown were loaded to adjacent wells. AKT and β-Actin were used as internal controls. The intensity of the bands generated from the Western blot assay was quantified using ImageJ software, and the fold changes relative to RAS activity from fElas^CreERT mice or the designated control sample were determined.

2.9. PCR and DNA sequencing

To verify the genetic profile of heterozygous or homozygous mutations at KRAS codon 12, genomic DNAs of human pancreatic cancer cell lines BxPC-3, PANC-1, SU86.86, and AsPC-1 were isolated using AllPrep DNA/RNA/miRNA Universal kit (REF 80224, Qiagen, Hilden, Germany) according to the manufacturer’s protocol. PCR was performed with the isolated genomic DNAs as templates and primers of forward (5’-GAT GTC ACA ATA CCA AGA AAC CC-3’) and reverse (5’- TCT GCA GTC AAC TGG AAT TTT CA3’) at 94 °C for 3 min followed by 38 cycles of 94 °C for 20s, 55 °C for 20s and 72 °C for 1.5 min and one extension at 72 °C for 5 min. The PCR product was individually isolated from the band in 1.6% agarose gel with GeneJET Gel extraction kit (REF K0691, Thermo Scientific), and sequenced with primer 5’-AAA AAG ATT GTC TTT TAG GTC CA-3’ at the DNA sequencing core facility of Stony Brook University. Results were presented with Chromas sequencing analysis software.

2.10. Gene Expression Analysis

The gene expression levels of KRAS, NRAS, HRAS in mouse pancreatic tissues were analyzed by RT-qPCR. Total RNA was isolated from mouse pancreas using TRIzol reagent (Ambion, Life Technologies, Camarillo, CA). RNA was further purified by DNase digestion and recovered with RNeasy kit (Qiagen, 74104). Reverse transcription was performed by using first-strand complementary DNA synthesis. Quantitative PCR was conducted using QuantiFast SYBR Green PCR Kit (Qiagen, 204056). The following primers were used: Kras (NM_021284.6, forward: 5’-TGC CTA TGG TCC TGG TAG GGA-3’; revers: 5’-AGG CAT CGT CAA CAC CCT GT-3’); Nras (BC058755.1, forward: 5’-GAT GGT GAG ACC TGC CTG CT-3’; revers: 5’-CAG AGG AAC CCT TCG CCT GT-3’); Hras (BC011083.1, forward: 5’-TGG AGG CGT GGG AAA GAG TG-3’; revers: 5’-TGA CCA CCT GTT TCC GGT AGG-3’). The expression level of 18S serves as the internal control.

2.11. GEPIA Database Analysis

The gene expression levels of IKK2 in patients with pancreatic adenocarcinoma were evaluated based on the GEPIA database (http://gepia.cancer-pku.cn/detail.php?gene=ikk2), which is a new interactive website based on genotype-tissue expression (GTEx) and TCGA data.

2.12. Statistical analysis

Comparison between two groups was analyzed by Student’s t-test unless otherwise indicated. Results were expressed as group means ± SD. *, p<0.05; **, p<0.01; ***, p<0.001; ****, p<0.0001. ns, not significant. GraphPad Prism 6.0 were used for statistical analysis.

3. Results

3.1. Oncogenic KRAS^G12D and IKK2 synergistically promote the rapid formation of PanIN lesions

While oncogenic KRAS mutations, such as KRAS^G12D/V, stipulate early events essential for the development of pancreatic cancerous lesions in most pancreatic cancer patients, evidence suggests that in adults, the presence of an endogenous level of oncogenic KRAS alone is insufficient. A second hit, such as inflammatory stimuli, is required to drive pancreatic tumorigenesis with high penetrance. To gain deeper insights into the degree of oncogenic KRAS activation required for accelerated pancreatic cancer development in the context of inflammation, we employed a mouse model denoted as KIC, comprising Kras^LSL-G12D/+;IKK2^LSL-f/f;fElas^CreERT. In this model, IKK2 overexpression and an endogenous level of Kras^G12D/+ expression are concurrently induced specifically in pancreatic acinar cells by tamoxifen (TM). Control groups consisted of mice carrying Kras^LSL-G12D/+;fElas^CreERT (referred to as Kras^G12D/+ after TM), IKK2^LSL-f/f;fElas^CreERT (referred to as IKK2 after TM), or fElas^CreERT (referred to as Cre after TM) (Figure 1A). Of note, only one allele of the Kras^G12D/+ mice carries the Kras^G12D mutation, while the other allele expresses wild-type Kras. Overexpression of IKK2, the inhibitor of nuclear factor kappa B kinase subunit beta known to activate the nuclear NFκB transcriptional network, leads to robust inflammatory and immune responses capable of inducing pancreatitis, contributing to the development of invasive PDAC [14].

Fig. 1. Oncogenic KRAS^G12D and IKK2 synergize to promote rapid PanIN lesions.

(A) Experimental scheme: fElas^CreERT(Cre), IKK2, Kras^G12D/+, and IKK2;Kras^G12D/+(KIC) mice at 25–30 days of age were treated with TM, and the pancreata were harvested after 1 month of TM treatment. (B) H&E staining (upper panel) and Alcian Blue staining (lower panel) of pancreatic tissue sections of the Cre, IKK2, Kras^G12D/+, and KIC mice after one month of TM induction. Scale bar, 200 μm. n=3 each for Cre, IKK2, and Kras^G12D/+ groups, n=7 for KIC group. (C). Quantitative analysis of Alcian Blue positive stains in Figure 1B. n=3 for Cre, IKK2, and Kras^G12D/+ groups, n=7 for KIC group. (D-E) Western blot analysis and quantification of IKK2 in the indicated pancreata of Cre, IKK2, Kras^G12D/+, and KIC mice one-month post-TM. Scale bar, 200 μm. n=3 for each group. (F-I) IHC analysis and quantification of p65 and COX-2 levels in the pancreatic tissue sections of Cre, IKK2, Kras^G12D/+, and KIC mice. Representative images with magnification (insets) were shown. Scale bar, 200 μm. n=3 for each group. Quantitative results were expressed as group means ± SD. *, p<0.05; ***, p<0.001; ****, p<0.0001.

One month after TM induction, histopathological analysis with H&E staining and PanIN lesion assessment with Alcian blue staining for acidic mucin on pancreatic tissue sections revealed that IKK2 overexpression alone or an endogenous level of Kras^G12D/+ expression alone did not lead to notable pancreatic histological alterations when compared to the normal Cre control mice. However, the concurrent expression of IKK2 and oncogenic Kras^G12D/+ for one month led to rapid and extensive PanIN lesions, characterized by the prominent neo-ductal structures and extensive Alcian blue positive staining, in comparison to the Cre, IKK2, or Kras^G12D/+ mice (Figures 1B–C). Western blot data showed that the protein levels of IKK2 were marked upregulated in the pancreata of both IKK2 and KIC mice compared to that in Cre and Kras^G12D/+ mice (Figures 1D–E). Of note, the IKK2 level in KIC mice was significantly lower than that in IKK2 mice, suggesting that the concurrent expression of oncogenic KRAS and IKK2 reduced the IKK2 protein level. IKK2 is a main activator of the inflammatory transcription factor NF-κB. Immunohistochemistry (IHC) staining showed that overexpression of IKK2 led to the accumulation of nuclear p65, a component of the active NF-κB complex. Interestingly, the nuclear p65 level was significantly higher in the pancreata of KIC mice compared to that in IKK2 mice, despite the latter having a notably higher level of IKK2 protein. This suggests that IKK2 and oncogenic KRAS synergistically enhanced the amount of nuclear p65. In contrast, Cre or Kras^G12D/+ mice did not exhibit detectable p65-positive staining (Figures 1F–G). Furthermore, levels of COX-2, a downstream effector of the activated NFκB inflammatory pathway, were elevated at the neo-ductal sites of the KIC pancreata compared to that of the Cre or Kras^G12D/+ mice (Figures 1H–I). Interestingly, the level of COX-2 protein in IKK2 mice was comparable to that of the Cre or Kras^G12D/+ mice, suggesting that the expression of IKK2 alone did not lead to a rapid increase in the level of COX-2 protein.

3.2. IKK2 and oncogenic KRAS^G12D synergize to promote pancreatic immune cell infiltration

To assess whether the simultaneous expression of IKK2 and oncogenic KRAS^G12D could promptly induce immune cell infiltration, we conducted IHC staining to examine protein markers associated with various types of immune cells. Our findings revealed that CD3, indicative of peripheral T-cell infiltration, was notably elevated in the pancreata of KIC mice compared to that in IKK2 mice. In contrast, the expression of KRAS^G12D alone in Kras^G12D/+ mice did not lead to a significant increase in CD3⁺ T-cell infiltration compared to Cre mice (Figures 2A–B). Similarly, B cells, detected by Pax-5⁺ IHC staining, exhibited an elevated presence in KIC mice, but not in Kras^G12D/+ mice, when compared to Cre mice (Figures 2A & C). Additionally, neutrophils marked by Ly-6G, proinflammation assessed by IL-1a, and macrophages detected by F4/80, were all significantly increased in the pancreata of KIC mice. In contrast, these markers did not exhibit notable elevations in the pancreata of Kras^G12D/+ mice when compared to IKK2 and Cre mice (Figures 2A, 2D–F). These observations strongly indicate that IKK2 and oncogenic KRAS collaborate synergistically in promoting immune cell infiltration into the pancreata, whereas oncogenic KRAS alone does not have this effect.

Fig.2. KRAS^G12D and IKK2 promote the infiltration of immune/inflammatory cells into the pancreata.

(A) IHC analysis of different immune cell specific markers CD3, Pax-5, Ly-6G, IL-1α, and F4/80 in the pancreata of Cre, IKK2, Kras^G12D/+, and KIC mice. Scale bar, 200 μm. n=3–4 for each group. (B-F) Quantitative analysis of IHC stains for the immune cell markers as indicated in (A). Quantitative results were expressed as group means ± SD. *, p<0.05; ****, p<0.0001.

3.3. The levels of KRAS^G12D and NRAS proteins are significantly increased in the pancreata of KIC mice

Since oncogenic RAS activation serves as the fundamental molecular event in the initiation and progression of pancreatic cancer, we first evaluated the protein levels of total RAS, which may contribute to the overall quantity of GTP-bound active RAS in the pancreata of these mice. Our findings indicated a noteworthy elevation in total RAS protein levels in the pancreata of KIC mice. This increase was not observed in Kras^G12D/+ and IKK2 mice compared to the Cre control mice (Figures 3A–B). As members of the RAS family, KRAS, NRAS, and HRAS are all implicated in human cancers, we assessed their possible contributions to the heightened total RAS protein levels by using specific antibodies targeting different RAS isoforms. A significant five-fold increase in total KRAS protein was observed in the pancreata of KIC mice, but not in Kras^G12D/+ or IKK2 mice, compared to the Cre control (Figures 3C–D). Since we did not detect a noticeable rise in total RAS or KRAS protein levels in the Kras^G12D/+ mice following one month of induction, we evaluated the levels of mutant KRAS^G12D protein using KRAS^G12D specific antibody. Our data revealed that mutant KRAS^G12D was readily detectable in the pancreata of Kras^G12D/+ mice compared to that of IKK2 or Cre mice, suggesting that the knock-in allele in Kras^G12D/+ mice functioned as intended. Of note, we observed a significant three-fold increase in KRAS^G12D protein levels in the pancreata of KIC mice compared to those of Kras^G12D/+ mice (Figures 3E–F), indicating that the concurrent expression of Kras^G12D/+ and IKK2 augmented the abundance of KRAS^G12D protein in the pancreata of KIC mice. Furthermore, the expression level of NRAS, but not HRAS, exhibited a significant fourfold increase in the pancreata of KIC mice, but not in Kras^G12D/+ or IKK2 mice, compared to Cre mice (Figures 3G–I). This suggests that the augmented NRAS protein level could also contribute to the overall elevation of total RAS protein levels observed in the pancreata of KIC mice, as shown in Figure 3A. The absence of HRAS response aligns with prior studies indicating a limited contribution of HRAS to pancreatic cancer [24].

Fig. 3. Levels of KRAS^G12D and NRAS proteins are significantly increased in the pancreata of KIC mice.

(A-B) Western blot analysis and quantification of the protein levels of total RAS in the indicated pancreata of Cre, IKK2, Kras^G12D/+, and KIC mice one month post TM. AKT served as a loading control. n=5 for each group. (C-D) Western blot analysis and quantification of the protein levels of KRAS in the indicated pancreata one month post TM. AKT served as a loading control. n=8 for Cre and IKK2 groups, n=11 for Kras^G12D/+ and KIC groups. (E-F) Western blot analysis and quantification of the protein levels of KRAS^G12D in the indicated pancreata one month post-TM. n=12 for each group. (G-I) Western blot analysis and quantification of the protein levels of NRAS (G, H, n=9) and HRAS (G, I, n=5) in the indicated pancreata one month post TM. Note that the AKT bands shown in Figure 3G are the same as the ones in Figure 3A as the same samples were used. (J) Quantitative RT-qPCR analysis of Kras, Nras and Hras mRNA levels in the pancreatic tissues of Cre, IKK2, Kras^G12D/+, and KIC mice. 18S mRNA level serves as internal control. Quantitative results were expressed as group means ± SD. **, p<0.01; ***, p<0.001; ****, p<0.0001.

To ascertain whether the heightened RAS protein levels are a result of transcriptional activation of various RAS isoforms, we conducted quantitative RT-PCR analyses. Our results showed that the mRNA levels of both KRAS and NRAS were markedly upregulated by approximately 20 and 17 fold, respectively, in the pancreata of KIC mice compared to that of the IKK2, Kras^G12D/+, or Cre mice. In contrast, no statistically significant differences were observed in HRAS mRNA levels among the groups (Figure 3J). This indicates the presence of a transcriptional regulatory mechanism contributing to the observed elevation in both KRAS^G12D and NRAS when IKK2 and KRAS^G12D are concurrently expressed. Collectively, these findings imply that the protein levels of KRAS and NRAS are significantly increased in the KIC pancreata, which contributes significantly to the heightened total RAS protein levels in response to the combined effects of KRAS^G12D and inflammation. Of note, the expression of Kras^G12D/+ alone had no discernible effect on NRAS or HRAS protein levels even one month after the endogenous level of Kras^G12D/+ expression.

3.4. The concurrent short-term expression of IKK2 and Kras^G12D/+ leads to a rapid increase in the overall abundance of GTP-bound KRAS^G12D

To investigate whether inflammation promotes rapid PanIN lesions by facilitating the transition of KRAS^G12D to its fully active GTP-bound state in the pancreata of KIC mice during the early induction stage, we employed a RAF pull-down assay, followed by Western blotting, to evaluate the overall abundance of the GTP-bound RAS, a critical component in intracellular signaling for cell survival, growth, and transformation. Our data showed a marked increase in the total amount of RAS-GTP, with an approximately eightfold increase in the KIC mice, a fourfold increase in the Kras^G12D/+ mice, and no discernible change in the IKK2 mice, in comparison to the Cre mice one month after TM induction (Figures 4A–B). These results, combined with the data from Figures 3A–B, suggest that the simultaneous expression of IKK2 and oncogenic Kras^G12D/+ for one month led to a significant increase in both the protein abundance and the overall amount of GTP-bound RAS compared to the expression of KRAS^G12D/+ alone. By employing an anti-KRAS^G12D-specific antibody, we noted a significant rise in the amount of GTP-bound KRAS^G12D, approximately 5.2 times higher in the pancreata of KIC mice compared to Kras^G12D/+ mice (Figures 4C–D). When considered in conjunction with the data presented in Figures 3E–F, the elevated absolute quantity of the KRAS^G12D-GTP form observed in the KIC mice (Figure 4C) can, in part, be attributed to the heightened KRAS^G12D protein level. However, in the Kras^G12D/+ pancreata, the levels of GTP-bound KRAS^G12D were increased without a corresponding increase in KRAS protein level, as shown in Figures 3C–D, suggesting that the overall rise in the amount of the GTP-bound KRAS^G12D might be attributable to the increased KRAS^G12D activity rather than an increase in KRAS protein abundance.

Fig. 4. The concurrent short-term expression of IKK2 and Kras^G12D/+ rapidly enhances the overall abundance of GTP-bound KRAS^G12D.

(A-B) Changes in the overall levels of GTP-loaded active RAS (RAS-GTP) in the pancreata of the indicated mice one-month post-TM as revealed by Western blots. AKT was used as a loading control. n=9 for Cre, n=10 for IKK2, Kras^G12D/+, and KIC. (C-D) Changes in the overall levels of pancreatic GTP-loaded active KRAS^G12D one-month post-TM. AKT was used as a loading control. n=7 for Kras^G12D/+ and KIC groups. (E) Changes in the overall levels of pancreatic GTP-loaded active HRAS (HRAS-GTP) one-month post TM. n=3 for each group as indicated. The pulmonary tissue protein sample serves as a positive control. (F-G) Changes in the overall levels of pancreatic GTP-loaded active NRAS (NRAS-GTP) one-month post-TM. n=6 for each group. (H-I) The ratio of active KRAS^G12D-GTP to the total KRAS^G12D protein (KRAS^G12D activity) in the pancreata of Kras^G12D/+ and KIC mice one-month post-TM. Active GTP-bound RAS fraction (denoted as ‘A’) was precipitated by RAF affinity from 200 μg of total protein lysates and loaded on the same SDS-PAGE gel with 40 μg of total protein (indicated as ‘T’). A and T were probed with an anti-KRAS^G12D-specific antibody to detect GTP-bound KRAS^G12D fraction from A and total KRAS^G12D protein from T, respectively. The data were quantified by densitometry of the Western blot bands to obtain KRAS^G12D activity, which is the ratio of GTP-bound KRAS^G12D to total KRAS^G12D protein. n=7 for each group. Note that the AKT bands shown in Figures 4A and 4E are the same as the ones shown in Figures 3A and 3G. The AKT bands in 4C and 4F are also the same as the same samples were used. Quantitative results were presented as group means ± SD. *, p<0.05; **, p<0.01; ***, p<0.001; ****, p<0.0001.

Additionally, we conducted a comparative assessment of the amount of GTP-bound NRAS and HRAS using their respective specific antibodies. HRAS-GTP was not detected in the pancreatic tissues of any mouse groups, consistent with the negligible levels of HRAS protein in contrast to the lung tissues, which exhibit high levels of HRAS protein (Figure 4E), indicating a minimal role played by HRAS in the development of pancreatic cancer in the KIC mice [24]. In contrast, we observed a 3.5-fold increase in the NRAS-GTP level in the pancreata of the KIC mice compared to that of the Kras^G12D/+ mice (Figures 4F–G). Of note, the levels of both NRAS-GTP and HRAS-GTP in the pancreata of the Kras^G12D/+ mice showed no significant difference compared to those of the IKK2 or Cre mice (Figures 4E–G).

It should be noted that the overall quantity of GTP-bound KRAS^G12D is determined by both the total KRAS^G12D protein level and its activity, which is determined by the ratio of GTP-bound KRAS^G12D to the total KRAS^G12D protein. In order to ascertain whether oncogenic KRAS^G12D is fully active in the context of inflammation, we quantitatively measured KRAS^G12D activity by separating the total protein sample into two parts: one part of 40 μg proteins was directly loaded onto an SDS-PAGE gel to detect the total KRAS^G12D protein, while the other part, comprising 200 μg of proteins, underwent immunoprecipitation using RAF affinity beads before being loaded onto the same SDS-PAGE gel for the detection of the GTP-bound RAS. We subsequently conducted Western blots using a KRAS^G12D-specific antibody along with protein normalization. Our data revealed that approximately 32% of the total KRAS^G12D protein is in the GTP-bound form in the Kras^G12D pancreata, whereas in the KIC pancreata, this proportion increases to 48% (Figures 4H–I), suggesting that inflammation elevated KRAS^G12D activity; however, the GTP-bound KRAS^G12D was far less than the expected 100%, indicating that KRAS^G12D itself remains partially active even in the context of inflammation at the early stage (e.g., one-month post TM). These data, in conjunction with Figures 4C–D and Figures 3E–F, suggest that in the pancreata of the Kras^G12D mice, the heightened amount of GTP-bound KRAS^G12D primarily arises from increased KRAS^G12D activity. In contrast, in the KIC mice, this increase is attributable to both elevated activity and higher protein abundance of KRAS^G12D in the early stage.

3.5. KRAS^G12D remains partially active even after a prolonged concurrent induction of Kras^G12D/+ and IKK2

In order to assess whether inflammation could lead to a further increase in the protein level and activity of oncogenic KRAS^G12D over a relatively extended period, we evaluated changes in protein levels of KRAS^G12D in the pancreata of Kras^G12D/+ and KIC mice ten months post TM vs. one month post TM (Figure 5A). Results showed that expression of oncogenic Kras^G12D/+ alone for 10 months resulted in the formation of PanIN lesions; however, concurrent expression of Kras^G12D/+ and IKK2 led to PDAC (Figure 5B). There was an approximately 3.5-fold increase in the levels of KRAS^G12D protein over the course of 10 months compared to one month post-TM induction in Kras^G12D/+ mice. In contrast, the levels of KRAS^G12D protein in KIC mice remained as elevated at ten months as they were at one-month post-TM (Figures 5C–D). Similarly, the protein levels of KRAS and NRAS showed notable increases of 7.6 and 5-fold, respectively, after 10 months compared to one-month induction in the Kras^G12D/+ mice. In sharp contrast, in the KIC mice, the abundance of both KRAS and NRAS proteins appeared to have already reached the peak at the early stage, and these levels were not significantly altered over the course of 10 months post-TM (Figures 5C, E–F). These findings indicate that IKK2 and oncogenic KRAS^G12D cooperate to rapidly elevate KRAS^G12D protein levels to promote extensive PanIN lesions, while KRAS^G12D alone is considerably less efficient, requiring a longer latency to reach a comparable high level.

Fig. 5. KRAS^G12D remains partially active after a long-term expression of IKK2 and Kras^G12D/+.

(A). Experimental scheme: One-month-old male and female Kras^G12D/+ and KIC mice were treated with TM, and the pancreata were collected after 1 month and 10 months of TM treatment. (B) H&E staining of pancreatic tissue sections of the Kras^G12D/+ and KIC mice 10 months after TM induction. Scale bar, 200 μm. n=6 for Kras^G12D/+, n=13 for KIC. (C-F) Western blot and quantitative densitometry analyses of protein levels of different RAS isotypes including KRAS^G12D, KRAS, and NRAS in the Kras^G12D/+ and KIC pancreata 10 months post TM compared to that one-month post TM. AKT was used as a loading control. Cre served as a negative control. n=11–12 for Kras^G12D/+ and KIC mice one-month post TM. n=3 for Kras^G12D/+ and KIC 10 months post TM. (G-H) The ratio of active KRAS^G12D-GTP to the total KRAS^G12D protein (KRAS^G12D activity) in the pancreata of Kras^G12D/+ and KIC mice 10 months after TM. n=3 for each group. Results were presented as group means ± SD. ns, not significant.

To quantitatively evaluate the activity of KRAS^G12D ten months after the initial induction of Kras^G12D/+ or IKK2;Kras^G12D/+, we conducted experiments similar to those described in Figure 4H. Our data showed that about 40% of KRAS^G12D protein was in the active KRAS^G12D-GTP form in the pancreata of KIC mice, compared to 34% in Kras^G12D/+ mice (Figures 5G–H).

3.6. KRAS^G12D is partially active in patient-derived PDAC cells

To corroborate our results that KRAS^G12D is partially active in pancreatic cancer, we evaluated KRAS^G12D activity in patient-derived pancreatic cancer cells with varying mutant KRAS contents, including BxPC-3, PANC-1, AsPC-1, and SU86.86 cells. Our DNA sequencing analysis showed the presence of only wild-type KRAS (KRAS^wt) in BxPC-3, both KRAS^wt and KRAS^G12D in PANC-1 and SU86.86 cells, and KRAS homozygous KRAS^G12D/G12D in AsPC-1 cells (Figure 6A). Correspondingly, we conducted a RAS activity assay as described in Figure 4H. As predicted, we only detected total KRAS but not KRAS^G12D, KRAS^G12D-GTP, or KRAS-GTP in BxPC-3 cells, indicating that without KRAS^G12D, the wildtype KRAS remained inactive (Figure 6B). In PANC-1 cells, the active GTP-loaded forms of KRAS^G12D and KRAS constituted approximately 28% of the total KRAS^G12D protein and about 12.5% of the total KRAS protein, respectively (Figures 6B–D), while in SU86.86 cells, the activities of both KRAS and KRAS^G12D were about 40%. These data indicate that although G12D mutation does augment KRAS^G12D activity, the proportion of GTP-bound KRAS^G12D relative to total KRAS^G12D is less than the anticipated 100%. Furthermore, in AsPC-1 cells carrying only mutant KRAS^G12D, we observed that only about 47% of KRAS^G12D was in the GTP-bound form. This proportion also falls significantly short of the predicted 100%, signifying that KRAS^G12D maintains a state of partial activity even if both alleles are mutated (Figures 6B–D).

Fig. 6. KRAS^G12D is partially active in human PDAC cells.

(A) KRAS mutation status at G12 codon (red rectangle) in BxPC-3, PANC-1, AsPC-1, and SU86.86 cells by DNA sequencing. (B) Western blot analysis for the ratio of active KRAS-GTP to the total KRAS (KRAS activity), IKK2, and the ratio of active KRAS^G12D-GTP to the total KRAS^G12D proteins (KRAS^G12D activity) in BxPC-3, PANC-1, AsPC-1, and SU86.86 cells. β-actin serves as a loading control. (C) Quantitation of the ratio of active KRAS-GTP to total KRAS protein. (D) Quantification of the ratio of active KRAS^G12D-GTP to total KRAS^G12D proteins. The experiment in Figure 5B was repeated three times. Results were presented as group means ± SD. (E) Changes of IKK2 levels in tumors from patients diagnosed with PAAD (name in the TCGA database for PDAC) compared to normal tissues in the GEPIA (gene expression profiling interactive analysis) database (http://gepia.cancer-pku.cn/detail.php?gene=ikk2). The Y-axis represents the relative gene expression level. *, p < 0.05. PAAD, pancreatic adenocarcinoma. (F) IHC staining of IKK2 was conducted on pancreatic tissues obtained from the Cre or Kras^G12D/+ mice fed either a ND or an HFD. Scale bar, 100 μm. (G) Summary of protein levels and activities of KRAS^G12D in the pancreata of Kras^G12D/+ and KIC mice in the short-term vs. long-term post-TM induction. After one month of induction, the expression of oncogenic KRAS^G12D/+ alone increased its activity without affecting the pancreatic pathology; however, when oncogenic KRAS^G12D/+ was co-expressed with IKK2, both the protein level and activity of KRAS^G12D increased rapidly, leading to the extensive formation of PanIN lesions. Over the 10-month induction period, the protein levels of oncogenic KRAS^G12D increased, resulting in a higher overall abundance of GTP-bound KRAS^G12D and the development of PanIN lesions compared to the one-month induction in the pancreata of KRAS^G12D/+ mice. In contrast, the combined expression of oncogenic KRAS^G12D/+ and IKK2 did not lead to further increases in protein level and activity compared to the one-month induction. This suggests that a rapid augmentation of both protein abundance and activity of KRAS^G12D at the early stage play crucial roles in determining the overall quantity of GTP-bound KRAS^G12D, leading to the development of PDAC under inflammation. ↔, remaining unchanged from the previous stage.

Given the significant effects of IKK2 on the augmentation of the protein level and activity of KRAS^G12D, we analyzed IKK2 expression levels in human PDAC cell lines and pancreatic cancer patient samples. Our data showed that IKK2 protein was readily detectable in various human PDAC cell lines, including BxPC-3, Panc-1, Aspc-1, and Su86.86 cells (Figure 6B). IKK2 mRNA levels in tumor tissues of PDAC patients (annotated as PAAD within TCGA) were significantly higher than those in normal tissues. (http://gepia.cancer-pku.cn/detail.php?gene=ikk2) (Figure 6E). Studies have shown that exposing Kras^G12D/+ mice to chronic HFD led to an elevated level of GTP-bound RAS, which accelerates oncogenic KRAS-mediated pathologies [4]. To assess the potential role of IKK2, we further analyzed its expression level in this mouse model under chronic HFD treatment. IHC analysis revealed a substantial increase in IKK2 protein levels in the pancreata of Kras^G12D mice fed an HFD compared to those fed a ND or Cre mice fed either a ND or an HFD (Figure 6F). These findings imply that the elevation in IKK2 levels may play a significant role in the rapid development of PDAC promoted by obesogenic HFD.

4. Discussion

Mutant KRAS is prevalent in patients with pancreatic cancer and is a critical driver in the development of PDAC [2, 6]. However, studies have demonstrated that KRAS^G12D or KRAS^G12V alone is insufficient to drive PDAC with high penetrance [4, 8, 9, 13, 17, 19]. Furthermore, it has been found that various stimuli or insults, such as inflammation or a high-fat diet challenge, expedite mutant KRAS-mediated pancreatic neoplasia, highlighting the necessity of a second event for the development of PDAC [4, 8–10, 14, 16, 17, 19, 22, 25, 26]. Although it has been understood that mutated KRAS exhibits heightened activity, a pivotal and unanswered question pertains to the degree to which inflammation enhances KRAS^G12D activity, thereby expediting the swift development of PDAC.

Through our investigation utilizing a mouse model co-expressing KRAS^G12D and IKK2 in pancreatic acinar cells, we have unveiled several new insights into how inflammation augments the overall amount of GTP-bound oncogenic KRAS^G12D, contributing to the development of PDAC. Firstly, the short-term expression of endogenous levels of Kras^G12D/+ alone heightened KRAS^G12D activity with minimal changes in KRAS protein levels and pancreatic pathogenesis. Secondly, prolonged expression of Kras^G12D/+ alone leads to an elevation in KRAS^G12D protein levels, with only a slight elevation in KRAS^G12D activity compared to short-term induction. This suggests that the substantial rise in KRAS^G12D protein quantity, rather than its activity, is the primary factor contributing to the overall augmentation of total GTP-bound KRAS^G12D over time in the absence of an inflammatory stimulus. Thirdly, the co-expression of KRAS^G12D and IKK2 induces a rapid and significant surge in both the protein level and activity of KRAS^G12D during the initial phase. This combined effect leads to a saturation point for the total quantity of GTP-bound KRAS^G12D, which does not further increase over an extended period (Figure 6G). Fourthly, despite the potent role of inflammation in promoting oncogenic KRAS-mediated invasive PDAC, KRAS^G12D remains partially in the GTP-bound active state under all tested conditions. This suggests that complete activation of KRAS^G12D is not a prerequisite for PDAC development even under inflammation. Finally, in patient-derived PDAC cells, even with a homozygous mutation in KRAS^G12D alleles, the GTP-bound form constitutes less than 50% of the total protein. This confirms that full activation of the KRAS^G12D protein is not imperative for PDAC cells.

Several previous studies have demonstrated that inflammation acts cooperatively with mutant KRAS to drive PDAC with high penetrance. Studies from the Jacks group showed that inflammation could promote neoplasia by altering the fate of the differentiated cells, which are typically refractory to oncogenic stimulation [27]. Similarly, the Barbacid group observed that mice expressing Kras^G12V specifically in adult pancreatic acinar cells failed to develop any PanIN lesions or PDAC unless subjected to inflammatory challenges [13]. Mechanistically, studies from the Logsdon group suggested that inflammatory stimuli trigger an NF-κB-mediated positive feedback mechanism that amplifies RAS activity to pathological levels, leading to accelerated development of PDAC [8]. Our data support the presence of a feedforward loop resulting from the synergistic interaction between KRAS^G12D and inflammation to enhance overall GTP-bound KRAS^G12D. We have shown that pancreatic NF-κB is significantly activated in the KIC mice. In addition, a significant upregulation of mRNA levels for both KRAS and NRAS in the pancreata of KIC mice was observed, suggesting that IKK2 overexpression contributes to the increase in RAS transcription. Thus, our findings reveal that inflammation rapidly enhances the protein level and activity of KRAS^G12D. This dual augmentation results in an adequate quantity of GTP-bounded KRAS^G12D, leading to the development of extensive PanIN lesions during the early stage of induction (Figure 6G). In addition, there was an increase in KRAS activity but without detectable pathological changes one month post TM induction in the pancreata of Kras^G12D/+ mice, indicating that the elevation in RAS activity precedes histological changes.

Currently, it remains unclear whether inflammation fully activates mutant KRAS^G12D. Our study provides several lines of evidence to address this important question: (1) In the pancreata of Kras^G12D/+ mice alone, only approximately one-third of mutant KRAS^G12D was active; (2) Inflammation not only induced a rapid increase in KRAS^G12D activity but also elevated its protein levels, resulting in an overall rise in the quantity of GTP-loaded Kras^G12D; and (3) Even in the presence of inflammation, only a fraction of KRAS^G12D was in the GTP-bound form.

The specific roles of NRAS and HRAS in PDAC development have yet to be fully elucidated. In our study, we discovered that mutant KRAS^G12D collaborates with IKK2 to augment both the total protein levels and activity of RAS at the early stage of induction, which together attribute to the increased pool of both GTP-bound KRAS^G12D and NRAS. Notably, we observed no alterations in the protein level and activity of HRAS, indicating that HRAS may play a negligible role in PDAC development, even in the presence of inflammation. Nevertheless, further research is warranted to elucidate the precise role of NRAS in KRAS^G12D-mediated PDAC development in an inflammatory milieu.

In summary, our findings underscore that both the augmentation of KRAS^G12D activity and the increase in KRAS^G12D protein levels play pivotal roles in determining the overall quantity of the GTP-bound active form of KRAS^G12D required for the rapid development of PDAC (Figure 6G). Even though KRAS^G12D pool does not reach full activation under inflammatory conditions, a significant increase in KRAS^G12D protein levels appears to compensate for the partial activation of KRAS^G12D. Consequently, this compensatory effect significantly contributes to the overall increase in the quantity of GTP-bound KRAS^G12D necessary for PDAC development.

Highlights.

Short-term Kras^G12D/+ expression alone causes limited increases in KRAS^G12D activity not abundance.

Long-term Kras^G12D/+ expression elevates total pool of GTP-bound KRAS^G12D via increasing KRAS^G12D protein abundance.

Kras^G12D/+ and IKK2 synergistically cause early maximization of both protein abundance and activity of KRAS^G12D with extensive PanIN lesions.

Kras^G12D/+ and IKK2 together also enhance protein abundance and activity of NRAS at early stage.

Even under inflammation only a fraction of KRAS^G12D protein exists in an active, GTP-bound state.