Background
Pancreatic ductal adenocarcinoma (PDAC) is among the deadliest cancers with a five-year survival of only 11% (SEER) [
1]. The high mortality of pancreatic cancer patients is explained by the lack of early diagnostic markers leading to late-stage diagnosis of the disease and the high rate of treatment resistance [
2]. Standard of care includes resection for about 20% of patients with localized disease who are eligible and treatment with cytotoxic chemotherapies such as gemcitabine or combination agents like FOLFIRINOX (5-FU, oxaliplatin, folinic acid, and irinotecan). This combination, when tolerated, improves progression-free survival by 11 months [
3]. Despite the aggressive treatment of this disease with combinations of cytotoxic drugs, the vast majority of patients do not have an enduring response to the treatment. Through improvements in precision cancer treatments based on genetic mutations, implementation of combination therapies like FOLFIRINOX, and advances in immunotherapy over the last decade the five-year survival has increased from 6 to 11%, nevertheless, pancreatic cancer remains extremely deadly. These factors combined with the increased incidence of risk factors like diabetes and obesity. It is predicted that pancreatic cancer will be the second leading cause of cancer deaths by 2030 [
4].
Doubling the five-year survival rate is remarkable progress, but novel treatment strategies are still needed. One effective strategy is to identify cellular markers or pathways that are enriched in pancreatic cancer cells. For example, over 90% of PDAC tumors have a mutation in the KRAS gene, however, no KRAS targeting treatments are available and existing targeted therapies like PARP inhibitors [
5] and EGFR inhibitors [
6] are limited to a subset of patients. It is clear that targeted therapies are impactful for the eligible patients with significant improvement in progression-free survival among patients receiving a targeted treatment [
7]. Unfortunately, patients who do not respond to standard chemotherapeutics and are not eligible for existing targeted therapies are left with few options. Thus, there is significant interest in understanding the development of resistance, and in pursuing targets that might sensitize resistant cells to existing treatments.
Our previous transcriptomic study identified genes whose expression was positively or negatively correlated with patient survival and also linked to in vitro response to gemcitabine, a common PDAC treatment. We identified ANGPTL4 among those genes whose expression in patient tumors is associated with poor survival and whose knockdown in PDAC cell lines could increase sensitivity to gemcitabine [
8]. ANGPTL4 is a member of the family of angiopoietin-like proteins that were first described for their role in angiogenesis [
9,
10]. ANGPTL4 can be proteolytically cleaved [
11] and the two products each have their own functions. The N-terminal domain plays a role in lipid metabolism and genetic variation in this domain is linked to cardiovascular disease risk. The C-terminal domain has been implicated in tumorigenesis, the promotion of proliferation, and wound healing [
11‐
13].
Overall, the role of ANGPTL4 in cardiovascular disease and lipid metabolism has been much better described than its roles in cancer. ANGPTL4 has a described role in known cancer pathways including its ability to regulate CREB, FOS, and STAT3 via ERK signaling [
14,
15]. ANGPTL4’s ability to alter metabolism and ATP abundance can also impact drug transport [
15]. The picture, however, is not perfectly clear since ANGPTL4 expression seems to have different impacts in cancers of different primary sites. For example, methylation and downregulation of ANGPTL4 are associated with progression and metastasis in colon cancer [
16] but overexpression is associated with progression and poor prognosis in breast cancer [
17]. Breast cancers of the triple negative subtype may be different since ANGPTL4 overexpression in that context has been associated with inhibition of invasion [
18]. In pancreatic cancer,
ANGPTL4 overexpression has been associated with tumorigenesis [
19], cellular resistance to chemotherapy [
8], hypoxia response, and poor patient outcomes [
20]. The complicated role for ANGPTL4 motivates our further exploration of its function in pancreatic cancer.
Here, we describe the role of ANGPTL4’s in PDAC by exploring the molecular and cellular changes associated with altered activity of ANGPTL4, the impact of ANGPTL4 on chemoresistance, and the potential for inhibiting downstream pathways driven by ANGPTL4 activation to sensitize tumor cells to treatment. We show that overexpression of ANGPTL4 leads to chemoresistance, increased migratory potential, and proliferation. Our transcriptomic analysis revealed altered gene expression signatures downstream of ANGPTL4 overexpression that are linked to epithelial to mesenchymal transition (EMT) and predict patient outcomes. We showed that knockdown of downstream effectors including APOL1 and ITGB4 reversed resistance to treatment and reduced migratory potential. The expression of these genes is also associated with patient survival. These data support the hypothesis that ANGPTL4 and its downstream pathways are potential therapeutic targets for the reversal of treatment resistance in pancreatic cancer.
Methods
Cell culture
HEK293FT cells (ThermoFisher #70,007), MIA PaCa-2 cells (ATCC #CRM-CRL-1420), and Panc-1 (ATCC #CRL-1469) were cultured in D10 media: DMEM (Lonza #12-614Q) supplemented with 10% FBS, and 0.5% Penicillin-Streptomycin. All cell lines were maintained at 37 °C and 5% CO2. Cells were cryopreserved with the addition of 10% DMSO (EMD #MX1458-6) to D10 media.
Plasmids
LentiSAMv2 (Addgene #92,062) and lenti-MS2-p65-HSF1-Hygro (Addgene #89,308) were used to generate stable cell lines for gene activation. pMD2.G (Addgene #12,259) and psPAX2 (Addgene #12,260) were used to facilitate viral packaging of sgRNA vector plasmids.
sgRNA cloning
sgRNA oligos were designed and cloned into their respective plasmids as described previously (ANGPTL4: 5’-CACGGCCCTGGGGATGCCAAACTGTGG-3’ and NTC: 5’-ACGGAGGCTAAGCGTCGCAA-3’) [
21].
DsiRNA
IDT TriFECTa RNAi kit was used per manufacturer’s protocol. The DsiRNA sequences used were as follows: ANGPTL4 (IDT hs.Ri.ANGPTL4.13.1), APOL1 (IDT hs.Ri.APOL1.13.3), and ITGB4 (IDT hs.Ri.ITGB4.13.3). 100,000 cells were seeded per well of a 12 well tissue culture treated plate 24 h prior to transfection. Cells were transfected using RNAiMax (ThermoFisher #13778-030) following manufacturer’s recommended protocol. As directed in the TriFecta kit (IDT #hs.Ri.HDAC1.13), TYE 563 transfection efficiency control, positive HPRT-S1 control, and negative (DS NC1) scrambled sequence control were utilized. Functional assays were performed 48 h after transfection. Expression was validated with each transfection with IDT PrimeTime qPCR Assay system according to manufacturer’s recommendations on an Agilent QuantStudio 6 Flex Real-Time PCR system (ANGPTL4: Hs.PT.58.40146104, ACTB: Hs.PT.39a.22,214,847, GAPDH: Hs.PT.39a.22,214,836, and HPRT: Hs.PT.58v.45,621,572).
Protein quantification using AlphaLISA
Briefly, cells were seeded at 25,000 cells in 50uL of media per well of a 96-well tissue-culture treated plate in triplicate per cell type. The following day, the media from each well was transferred to a non-treated 96-well plate and 25ul of AlphaLISA Lysis Buffer (PerkinElmer, #AL001C) was added per well. Plates were then shaken at 250 RPM at room temperature for 10 min. Lysates were immediately used. Samples or standards were transferred to a 384 well AlphaPlate (PerkinElmer, #600,535, lot:8220–21,331) and the AlphaLISA ANGPTL4 manufacturer’s kit protocol (PerkinElmer, #AL3017HV, lot:2,905,318) was followed for standards, lysate, and supernatant (media) except for the following: lids were covered in foil to prevent light exposure when possible and during incubation periods plates were shaken at 150 RPM for the recommended time. The AlphaLISA plate was read on a BioTek Synergy H5 following previously published protocols [
22]. Data were analyzed in R (version 4.1.2) and GraphPad Prism 9.
RNA-sequencing
Cells were seeded at a density of 0.6 × 10^6 cells per well of a 6-well plate in triplicate. The following day media was removed and 3ml of media per well with or without 1.5nM of gemcitabine (Sigma #G6423) was added and incubated for 24 hours. Media was removed, cells were washed with PBS (Gibco #10010072), and released from the plate with 2 ml of TrypLE (Gibco #12604-021) per well. After centrifugation, pellets were frozen at -80 C until RNA extraction. For RNA extraction, of cell pellets 350 ul of RL Buffer plus 1% BME from the Norgen Total RNA extraction kit and extraction proceeded per manufacturer’s instructions including use of the DNase kit (Norgen # 37500, 25720). RNA quality was verified with the Agilent BioAnalyzer RNA Nano 600 kit (cat# 5067 − 1512) with the RIN range between 9.2–10. RNA-sequencing libraries were made using Lexogen QuantSeq 3’ mRNA-Seq Library Prep Kit FWD for Illumina kit (cat# 015.24) with 250 ng of RNA input. They were pooled and sequenced on an Illumina NextSeq 500 instrument with 75 bp single-end reads. Read counts averaged 4 million reads and 93.65% of bases exceeding Q30. Lexogen’s BlueBee integrated QuantSeq data analyses pipeline was used for trimming, mapping, and alignment and DESeq2 was used for differential expression [
23]. Heatmaps were generated using iDEP.95 [
24].
Scratch-wound assay
Cells were seeded at 100,000 cells per well of a 96-well plate. After 24 h uniform scratches were made across the diameter of the wells using a multichannel pipette with 200ul pipette tips and even pressure applied across the wells to cause a wound. The media was then vacuumed off and 200ul of new media was added. Cells were imaged on a Lionheart FX Automated Microscope every 8 h for 48 h. Cell culture growth conditions of 37 °C and 5% CO2 were maintained throughout. Forty images were taken of each scratch (4 wide by 10 long) with no overlap, autofocus, and auto brightness.
Images were integrated using R and the “tiff” package. Images were further processed using GIMP 2.10. ImageJ and the MRI_wound_Healing_Tool.ijm macro plugin [
25] were used to process the images and calculate the area of the wound the cells have not covered. Relative wound closure over time using time 0 for each condition as the control was plotted and a curve line equation was formed by fitting the curve to a non-linear fit one phase decay least squares fit with Yo = 0 as a constraint. This base equation (Y=(Y0 - Plateau)*exp(-K*X) + Plateau) where Y0 is the Y value when X (time) is zero, Plateau is the Y value at infinite times, and K is the rate constant, was used to determine the value of X or half the time it takes to close the wound. Data were analyzed using GraphPad Prism 9.
Drug resistance screening
Cells were seeded in 96-well plates at 2000 cells/well. Seeded cells were treated with a range of gemcitabine concentrations. Cells were treated again 48 h later. The number of viable cells surviving drug treatment were measured with CellTiter-Glo (Promega #G7571) 24 h after the last drug treatment per the manufacturer’s protocol using a BioTek Synergy H5 plate reader. Sample size ranged from 4 to 11.
Pathway analysis
1198 DEG from the ANGPTL4 OE vs. KD analysis were imported into the KEGG Mapper-Color [
26] and Panther GO Enrichment Analysis [
27]. The following parameters were used for KEGG: Search mode: hsa, used uncolored diagrams, included aliases. The following parameters were used for PANTHER: overrepresentation test, homo sapiens reference list, GO biological process complete, Fisher’s exact test, and calculate FDR. Top pathways were merged manually and drawn using BioRender.
Overall survival analysis (OS)
To conduct survival analysis, clinical and RNA-seq expression data was retrieved from The Cancer Genome Atlas for 178 PDAC (TCGA-PAAD) patients (
https://portal.gdc.cancer.gov/). Data were normalized using the R package DESeq2 and differentially expressed genes with an FDR < 0.1 were used to generate Kaplan-Meier survival curves. We classified tissues based on their average expression of a given gene set (bottom 25%, middle 50%, and top 25% of gene expression). We compared the patients with the lowest and highest quartile of average gene expression and performed survival analysis. Survival curves and analyses were generated using the “ggplot2”, “survminer”, and “survival” R packages. P values were generated using a log rank test.
Recurrence free survival (RFS)
A Kaplan Meier curve of recurrence-free survival survival plots for ANGPTL4 was created using GEPIA (
http://gepia.cancer-pku.cn/) single gene analysis. The relevant parameters were as follows: Disease-free Survival (RFS), Group Cutoff: Quartile (75% high, 25% low), Hazards Ratio: Yes, 95% Confidence Interval: Yes, Axis Units: Months, and datasets Selection: PAAD.
Correlation analysis
Rank-based correlation coefficients were computed for each of the DEG with log2 fold change greater than 0.7 and baseMean > 10 (MP2_ANGPTL4_KD vs. MP2_ANGPTL4_OE with or without treatment) for (1) all TCGA-PAAD data DEG per gene vs. ANGPTL4 expression and (2) all TCGA-PAAD data DEG gene versus OS time. These correlation values were used to generate a list of genes that are co-expressed with ANGPTL4 is overexpressed.
Correlation coefficients and p-values were computed using the “Hmisc” (version 4.6-0) R package. We classified tissues based on their ANGPTL4 expression of a given gene set (bottom 25%, middle 50%, and top 25% of gene expression). We computed correlation matrices using the expression data for patients with the lowest and highest quartile of ANGPTL4 expression which was used to divide samples into quantiles with highest and lowest average gene expression; based on expression of these 42 genes. (40 genes were measured in TCGA: KDM7a and STN1 were omitted).
Statistical analysis
Statistical analysis was conducted in R (version 3.6.1 and R version 4.1.2 for RNAseq analysis). The following R packages and software were used for analysis:
survival (version 1.2.1335) [
28]
survminer (version 0.4.9) [
29]
ggplot2 (version 3.3.6) [
30]
DESeq2 (version 1.24.0) [
31]
pheatmap (version 1.0.12) [
32]
Hmisc” (version 4.6-0) [
33]
tiff (version 0.1–11) [
34]
Discussion
Here we have described how overexpression of
ANGPTL4 in pancreatic cancer contributes to disease progression and resistance. We have shown that overexpression of this gene and protein leads to cellular resistance to gemcitabine, one of the most commonly used chemotherapeutics in PDAC. It follows that if overexpression leads to chemoresistance, increased expression of this gene would also be associated with poor patient outcomes, and we confirmed that in independent patient cohorts. The goal of our study is to further understand the role of this gene in chemoresistance in the hopes that a mechanistic understanding will facilitate the development of new treatment strategies. Our transcriptomic analysis revealed that overexpression of
ANGPTL4 broadly impacts transcriptomic profiles in pancreatic cancer cells. This is perhaps not surprising given a rather large body of literature describing not only ANGPTL4’s role in cancer but also cardiovascular disease risk and metabolism [
9]. In fact, ANGPTL4’s role in cancer can be difficult to summarize because it has been shown to have both protective and promoting effects in different cancer types [
13]. Our initial findings highlighted the potential importance of this gene and existing literature was not sufficient to understand how this gene functions in pancreatic cancer. Using our transcriptomic data, we narrowed our analysis to a list of
ANGPTL4-impacted genes that are also responsive to gemcitabine treatment in vitro. We identified 42 genes associated with
ANGPTL4 overexpression and gemcitabine response and determined that the expression of those genes effectively predicted patient outcomes.
Further narrowing our focus, we identified APOL1 and ITGB4 among the genes that are associated with ANGPTL4 overexpression, gemcitabine resistance, and patient survival and explored them further. We found that knockdown of either APOL1 or ITGB4 increased sensitivity to gemcitabine in cells overexpressing ANGPTL4. This finding confirms that inhibition of APOL1 or ITGB4 can reverse resistance associated with ANGPTL4 expression in pancreatic cancer cells. One unexpected finding was that knockdown of these genes also reduced expression of ANGPTL4 itself. Rather than our initial hypothesis that expression of ANGPTL4 was correlated with APOL1 or ITGB4 because they were regulated by ANGPTL4, this data supports the hypothesis that ANGPTL4 is regulated, directly or indirectly, by APOL1 and ITGB4 and these genes could be involved in a feedback loop that explains our results.
Considering
APOL1 first, APOL1 (apolipoprotein L1) is part of the HDL lipid complex which plays a key role in lipid metabolism [
44]. APOL1 can activate NOTCH1 signaling leading to reduced proliferation and increased apoptosis [
42]. NOTCH1 also upregulates ABCC1, a transporter that can also lead to multidrug resistance [
45]. In other contexts, NOTCH1 also has been shown to stabilize and activate PPARγ, a transcription factor that promotes transcription of ANGPTL4 [
46]. These results suggest a possible mechanism by which knockdown of
APOL1 may reduce
ANGPTL4 expression and further suggests that inhibition of
APOL1, especially in patients with high expression of
ANGPTL4 might reduce cellular capacity for migration and proliferation. Additional studies would need to be performed to demonstrate whether the link between ANGPTL4 and APOL1 requires NOTCH signaling.
ITGB4 knockdown also reduces
ANGPTL4 overexpression. ITGB4 (integrin β4, ɑ6β4) is an integrin protein that plays a role in cell-extracellular matrix adhesion. It is upregulated in several cancers including pancreatic cancer and has been linked to poor prognosis, high tumor grade, lymph node metastasis, and drug response [
47‐
49]. Knockdown of
ITGB4 increases cisplatin sensitivity in lung cancer models [
47] suggesting that
ITGB4 is relevant for resistance to multiple chemotherapeutics. ITGB4 may promote receptor tyrosine kinase (RTK) activation through ERBB2 [
50]. RTKs generally promote proliferation, invasion, and sustained cell migration, and ITGB4 has a synergistic effect in combination with increased RTKs expression [
50,
51]. In our study, we observed increased expression of
CSF3, VEGFA, HBEGF, and
DDR2, which all play a role in the epithelial-to-mesenchymal transition in cells overexpressing
ANGPTL4. An alternate hypothesis is that
ITGB4 promotes cell migration and invasion through regulation of the MEK1-ERK1/2 signaling cascade [
49]. Reduction of
ITGB4 in the context of overexpressed
ANGPTL4 could also affect this pathway, explaining the reduced migratory potential of
ITGB4 KD cells in the cells overexpressing
ANGPTL4. Additional transcriptomic and biochemical studies are needed to determine which of these pathways are most relevant for the chemoresistance phenotypes.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit
http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (
http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.