Visual Overview
Abstract
AraC-FdUMP[10] (CF10) is a second-generation polymeric fluoropyrimidine that targets both thymidylate synthase (TS), the target of 5-fluorouracil (5-FU), and DNA topoisomerase 1 (Top1), the target of irinotecan, two drugs that are key components of FOLFIRNOX, a standard-of-care regimen for pancreatic ductal adenocarcinoma (PDAC). We demonstrated that F10 and CF10 are potent inhibitors of PDAC cell survival (in multiple cell lines including patient-derived lines) with IC50s in the nanomolar range and are nearly 1,000-fold more potent than 5-FU. The increased potency of CF10 relative to 5-FU correlated with enhanced TS inhibition and strong Top1 cleavage complex formation. Furthermore, CF10 displayed single-agent activity in PDAC murine xenografts without inducing weight loss. Through a focused drug synergy screen, we identified that combining CF10 with targeting the DNA repair enzyme, poly (ADP-ribose) glycohydrolase, induces substantial DNA damage and apoptosis. This work moves CF10 closer to a clinical trial for the treatment of PDAC.
Implications: CF10 is a promising polymeric fluoropyrimidine with dual mechanisms of action (i.e., TS and Top1 inhibition) for the treatment of PDAC and synergizes with targeting of DNA repair.
Visual Overview: http://mcr.aacrjournals.org/content/molcanres/19/4/565/F1.large.jpg.
This article is featured in Highlights of This Issue, p. 541
Introduction
Pancreatic ductal adenocarcinoma (PDAC) is the third leading cause of cancer-related death in the United States, with a 5-year survival rate of 10% (1). While thousands of PDAC genomes have been whole-exome sequenced and core pathways have been identified as targets (2, 3), our best current options to treat PDAC, besides surgical resection, remain with cytotoxic chemotherapies. Clinical phase III trials in the metastatic setting show improved survival with a combination of cytotoxic agents called FOLFIRINOX (leucovorin, 5-fluorouracil/5-FU, irinotecan, oxaliplatin) which has become a standard-of-care option, particularly in healthy, high performance status patients with metastatic PDAC (4–6). Still, improvement of overall survival is limited to just a few months, all patients responding to treatment eventually succumb to the disease, and many experience detrimental side effects such as neutropenia and extreme fatigue, indicating an unmet need for novel therapeutic strategies.
While 5-FU is an integral component of FOLFIRINOX, the anticancer activity of this regimen is likely suboptimal because 5-FU is inefficiently metabolized to FdUMP (7, 8), the thymidylate synthase (TS) inhibitory metabolite that is responsible for 5-FU's anticancer activity. To overcome the limitations of 5-FU, we are developing polymeric fluoropyrimidines (F10, CF10) that we anticipate being more efficiently converted to the TS-inhibitory metabolite FdUMP by virtue of their higher fluoropyrimidine content per molecule of 5-FU. Another expected benefit of these molecules is that they may generate relatively lower levels of ribonucleotide metabolites that are responsible for 5-FU's systemic toxicities that can become serious and life threatening in some patients. FdUMP[10] (F10) dually targets TS and DNA topoisomerase 1 (Top1; refs. 9–12), which are well-established targets in PDAC by virtue of their targeting by 5-FU and irinotecan in FOLFIRINOX. In addition, F10 achieves strong antitumor activity while causing fewer systemic toxicities than 5-FU in multiple preclinical cancer models (11, 13–15).
In this study, we investigate a promising second-generation polymeric fluoropyrimidine AraC-FdUMP[10] (CF10), as a potential new treatment for PDAC. CF10 includes a non-native nucleotide (AraC) at the 3′-terminus which we believe will enhance core F10 stability by differential recognition by exonucleases (16) and anticancer activity through independent mechanisms of action. CF10 also includes PEG5 at the 5′-terminus, a modification known to promote binding to plasma proteins and increase circulation of molecules in vivo (17). We demonstrate therapeutic advantages for CF10 in PDAC models relative to both 5-FU and the prototype fluoropyrimidine polymer F10. In addition, we found that CF10 could be synergistically combined with a small molecule inhibitor of the DNA repair factor poly (ADP) glycohydrolase, poly(ADP) ribose glycohydrolase (PARG), suggesting the potential of novel combination therapy regimens that synergistically target DNA repair.
Materials and Methods
Cell line culture conditions
All cells lines were purchased from ATCC. They were authenticated by short tandem repeat profiling at least twice per year and confirmed negative for Mycoplasma contamination at least once per month. Cells were cultured according to ATCC specifications with the addition of prophylactic plasmocin (InvivoGen). Patient-derived cell lines (PDX) were obtained from Talia Golan and were cultured as described previously (18, 19). For all experiments, cell line passage number was kept below 20.
Chemical compounds
F10 was synthesized and characterized as described previously (20, 21). CF10 was synthesized and provided by William Gmeiner (22) and both F10 and CF10 were prepared in saline. PDDX-04 was provided by Joseph Salvino (The Wistar Institute, Philadelphia, PA), and was prepared in DMSO. A list of other commercially available compounds used in the study is supplied in Supplementary Table S1.
5-day PicoGreen survival assay and Bliss additivity analyses assay
A total of 1,000 to 3,000 cells per well were plated in 96-well plates. 5-day drug treatment of cells was performed in sextuplet the following day. Pico green assay was carried out as described previously (19, 23). For Bliss additivity analyses, survival percentages were calculated and then input into the Combenefit program as described previously (19, 23, 24).
10-day colony growth inhibition assay
A total of 1,000 to 5,000 cells per well were plated in 6-well plates and drug treatments were carried out the following day. After 5 days, cell culture media were changed, and new drug was added for another 5 days. Visualization of colonies was performed as described previously (19, 23) and colony area coverage was quantified using the Colony Area plugin for ImageJ (RRID:SCR_003070; ref. 25).
In vivo xenograft experiment and dosing
Six-week-old female aythmic nude mice (Envigo) were injected subcutaneously with 5 × 106 MIA PaCa-2 cells suspended in a 1:1 mixture of Matrigrel (Corning) and 1× PBS on both left and right flanks. Tumors progressed to 100 mm3 before mice were randomized into treatment groups. Mice were then treated either with F10 (200 mg/kg), CF10 (200 mg/kg), or vehicle control (saline) three times per week, alongside tumor volume measurement [L × (W × W)/2] and mice weight recording. After completion of dosing (3 weeks for F10 and 5 weeks for CF10), mice were euthanized and tumors were extracted for further analyses. All animal studies and handling were performed and approved according to Institutional Animal Care and Use Committee guidelines.
IHC for Ki67 and cleaved caspase-3 from formalin-fixed paraffin-embedded tumor tissue
For Ki67 IHC staining formalin-fixed paraffin-embedded tumor tissues, antigen retrieval was performed on the Roche Ventana Discovery ULTRA staining platform using Discovery CCI for 36 minutes. Primary immunostaining was performed using a predilute Ki-67(Roche 790-4286) and incubated at 41°C for 32 minutes. Secondary immunostaining used a horseradish peroxidase multimer cocktail (Roche) and immune complexes were visualized using the ultraView Universal DAB (diaminobenzidine tetrahydrochloride) Detection Kit (Roche). Slides were then washed with a Tris-based reaction buffer (Roche) and counterstained with Hematoxylin II (Roche) for 4 minutes. For cleaved-caspase-3 IHC staining, a similar protocol was followed using a caspase-3–specific antibody (Biocare, PP229AA). For both IHC stains, analyses of positive tumor tissue from five randomly selected fields per tumor were done via ImageJ thresholding analysis.
Western blotting for thymidylate synthase classic complex, total cell PARylation, γH2AX, and cleaved caspase-3
A total of 7.5 × 105 to 1 × 106 cells were plated in 10 cm dishes and cells were treated with drugs the following day for indicated timepoints. Cell lysis and Western blotting were performed as described previously (19, 23). Antibodies used are supplied in Supplementary Table S2.
Immunofluorescence imaging of stalled Top1 complexes
Imagining of stalled Top1 complexes was performed using a specific antibody (MABE1084, Sigma-Millipore) as described previously (26, 27). Slides were imaged with a Nikon A1R confocal microscope with a 60× oil objective. Quantification of Top1 foci was done via ImageJ by counting at least 100 cells and cells were declared positive if they had >50 foci per nuclei.
DNA fiber assay and high-throughput alkaline comet assay
Cells were treated with either DMSO or CF10 for 16-hours and analyzed for changes in replication dynamics via DNA fiber assay or DNA strand breaks via comet assay as described previously (28).
siRNA-mediated knockdown of PARG
A total of 500,000 to 600,000 cells were plated in 10 cm dishes and were transfected the following day. A total of 10 nmol/L of either siControl or siPARG siRNA were prepared with OPTI-MEM (Corning) and Lipofectamine RNAiMAX (Thermo Fisher Scientific) as per manufacturer's instructions. After 48 hours of transfection, cells were split for downstream experiments and a separate pellet was taken for protein validation via Western blot analysis with a PARG antibody supplied in Supplementary Table S2.
Statistical analyses
All statistical analyses were performed with the GraphPad Prism software (version 8.3.0, RRID:SCR_003070). For all in vitro experiments, three biological replicates were used for analysis and appropriate two-way ANOVA tests with Bonferroni correction were conducted with P < 0.05 being considered as significant except where stated otherwise. For in vivo xenograft experiments, a standard t test with Bonferroni correction on the final day of measurement was conducted via GraphPad Prism.
Additional experimental reagents are supplied in Supplementary Materials and Methods.
Results
CF10 has potent antitumor activity in PDAC cells
We first wanted to assess the in vitro and in vivo efficacy of CF10 and results from 5-day cell survival assays showed that CF10 had significantly lower IC50 values compared with F10 and 5-FU for inhibiting PDAC cell growth in vitro (Fig. 1A; Table 1; Supplementary Fig. S1A). We then conducted a 10-day colony growth inhibition assay to determine CF10′s impact on longer-term growth potential (Supplementary Fig. S1B). Quantification of colony area revealed that CF10 was significantly more potent than F10 and 5-FU at inhibiting colony growth in PANC1 (P < 0.01) and ASPC1 (P < 0.0001) cells, with a trend toward significance relative to F10 in MIA PaCa-2 cells (Fig. 1B).
CF10 has potent antitumor activity in PDAC cells. A, Representative graphs of 5-day survival assay with quantification of relative cell number via PicoGreen in the indicated PDAC cell lines. B, Colony area percentage quantification from 10-day colony growth inhibition assays in the indicated cell lines. C, Graph of individual MIA-PaCa 2 flank-tumor volumes (n = 8 mice/arm) after 35 days of CF10 treatment (200 mg/kg 3×/week) as compared with vehicle (saline) control with graph of changes in relative weight after initiation of CF10 treatment versus vehicle. D, Representative images of Ki67 and cleaved caspase-3 IHC staining of tumor tissues from CF10- and vehicle-treated mice with quantification of IHC staining. For each in vitro experiment, three independent experiments were conducted, and errors bars represent SEMs. Significance was tested via two-way ANOVA with Bonferroni correction for in vitro assays and Student t test with Bonferroni correction for in vivo assays denoted by *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.
IC50 values with SD for F10, CF10, and 5-FU across all PDAC cell lines examined. P values indicate differences between CF10 IC50s compared with F10 and 5-FU.
We next investigated the in vivo efficacy of CF10 in a murine flank xenograft study. Because neither CF10 nor F10 were previously tested in PDAC models, we initiated a pilot study with F10 to determine whether it was efficacious at a dose similar to previous studies in other tumor types (200 mg/kg 3×/week for 3 weeks; refs. 14, 15, 29). We found that F10 significantly reduced tumor volume (P < 0.05) compared with vehicle and was well tolerated, as F10-treated mice did not experience a decrease in body weight over the duration of the study (Supplementary Fig. S1D). Quantification of Ki67 and cleaved caspase-3 (cCas3) IHC of F10-treated tumor tissue showed a significant reduction in Ki67 staining (P < 0.001) and increased cCas3 staining (P < 0.0001; Supplementary Fig. S1E) compared with vehicle, indicating that F10 treatment reduced proliferation and induced apoptosis in PDAC cells in vivo. A similar in vivo study with CF10 revealed significant reduction in tumor volume (P < 0.01) with no decrease in mouse body weights, indicating that CF10 was well tolerated and effective (Fig. 1C; Supplementary Fig. S1C). Ki67 and cCas3 IHC staining of CF10-treated tissues showed similar results as for F10 (P < 0.0001 for both Ki67 and cCas3; Fig. 1D).
CF10 directly inhibits TS/Top1, disrupts replication fork dynamics, and is more DNA-directed than 5-FU
With the in vitro and in vivo efficacy of CF10 assessed, we next wanted to validate whether previously known mechanisms of action seen with F10 were active with CF10 in our PDAC models. Western blot analysis for TS inhibitory covalent complex (TS CC; refs. 30–32) revealed that CF10 or F10 at 10 nmol/L for 24 hours in MIA PaCa-2 cells strongly induced formation of the TS CC (Fig. 2A). 5-FU at a similar dose of 10 nmol/L did not induce TS CC formation, although at 10 μmol/L, a concentration that greatly exceeded the 10-fold difference in fluoropyrimidine content of CF10, 5-FU–induced TS CC. These findings were replicated in PANC1 cells. Inhibition of Top1 in PDAC cells was assessed by Top1 cleavage complex (Top1cc) immunofluorescence as described previously (26). Quantification of Top1cc-positive foci in MIA PaCa-2 and PANC1 cells revealed that 24 hours after treatment both CF10 and F10 significantly increased the number of Top1cc-positive nuclear foci compared with the no treatment control (P < 0.001; Fig. 2B) and to a level similar to the topotecan positive control (26, 33). 5-FU at a similar dose to F10- and CF10-induced lower levels of Top1cc, suggesting less incorporation of 5-FU generated FdUMP into DNA. This finding prompted us to investigate whether 5-FU generated Top1ccs could be caused through an RNA-mediated mechanism through 5-FU conversion to FUMP as opposed to FdUMP. To assess this, we conducted a similar Top1cc immunofluorescence assay but supplemented MIA PaCa-2 or PANC1 cells with uridine to slow incorporation of FUMP into RNA and found that only 5-FU had a statistically significant decrease in Top1cc generation compared with all other conditions (Supplementary Figs. S2–S4). Taken together, these data suggest independent mechanisms of Top1cc inhibition between F10, CF10, and 5-FU that involve DNA-mediated effects with F10 and CF10 compared with RNA-mediated effects with 5-FU.
CF10 directly inhibits TS/Top1, disrupts replication fork dynamics. and is more DNA directed than 5-FU. A, Representative Western blots detecting the TS CC after treatment in indicated PDAC cell lines with F10, CF10, and 5-FU at indicated concentrations for 24 hours. B, Representative immunofluorescence images with an antibody specific for Top1cc in indicated PDAC cell lines treated at indicated concentrations of F10, CF10, 5-FU for 24 hours and Topotecan for 1 hour with quantification graphs to the right. C, Schematic depicting the dose and timing scheme for DNA fiber experiment, representative images from PANC1 cells used in DNA fiber analysis (yellow arrows indicate terminal replication forks), and graphs for quantification of fork velocity and percent terminal forks in PANC1 cells. D, Quantifications of median bliss synergy scores in the indicated cell lines after cotreatment of F10, CF10, and 5-FU with thymidine and uridine. For each experiment, three independent experiments were conducted, and errors bars represent SEMs. Significance was tested via two-way ANOVA with Bonferroni correction for Top1cc immunofluorescence and Student t test with Bonferroni correction for DNA fiber and thymidine/uridine supplementation assays denoted by *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.
Because both TS inhibition and Top1cc formation could affect replication fork progression (34), we next investigated CF10′s effects on DNA replication fork dynamics via DNA fiber assay. Analysis of DNA fiber length and structure via immunofluorescence in PANC1 cells revealed that CF10 significantly reduced fork velocity (P < 0.0001) and increased the number of terminal replication forks (P < 0.001; Fig. 2C), indicating that CF10 induces replication stress in PDAC cells. These findings demonstrate CF10 is cytotoxic to PDAC cells via dual targeting of TS and Top1, which disrupts replication fork progression and causes replication stress. To further investigate the origin of CF10′s DNA-directed cytotoxic mechanism and its improved potency relative to 5-FU, we performed thymidine and uridine rescue experiments. Previous studies demonstrated that F10-induced apoptosis was reversed by exogenous thymidine, but not uridine (35). In contrast, 5-FU–induced apoptosis was reversed with uridine consistent with an RNA-directed mechanism (36, 37). We employed Bliss additivity analysis (19, 24) to examine whether the combination of CF10 with thymidine affected cell viability. We observed that thymidine antagonized both CF10 and F10 cytotoxicity in MIA PaCa-2 and ASPC1 cells (Fig. 2D; Supplementary Figs. S5 and S6). However, thymidine was a less potent antagonist of CF10, consistent with a non-TS–dependent cytotoxic mechanism that results from AraC release, a non-TS–dependent DNA damaging nucleoside analog. Consistent with a non-TS–directed cytotoxic mechanism, the 5-FU plus thymidine combination displayed relatively less antagonism relative to F10 or CF10. In contrast to our thymidine antagonism studies, Bliss additivity analysis revealed only mild antagonism when either CF10 and F10 were combined with uridine, but much stronger antagonism between 5-FU and uridine (P < 0.05; Fig. 2D). We believe that these supplementation studies highlight the increased “DNA-directed” versus “RNA-directed” effects of CF10 relative to 5-FU in PDAC cells, consistent with an improved therapeutic index.
CF10 and PDDX-04 synergy produces unresolved DNA damage which triggers apoptosis
Because combinations of drugs are historically more effective than single agents for PDAC treatment (5, 38), we next conducted a focused drug synergy screen with CF10 and various potential combinatorial partners using Bliss additivity analysis. Combinations of CF10 with standard-of-care PDAC agents such as 5-FU, irinotecan, oxaliplatin, or gemcitabine in either MIA PaCa-2 or PANC1 cells did not yield robust synergy (data not shown). We then screened inhibitors of PARP1 and PARG, olaparib and PDDX-04, respectively, two key enzymes involved in DNA damage repair (39, 40). PARP1 is important for Top1cc repair and PARP inhibitors produce increased sensitivity to Top1 inhibitors (41, 42). Furthermore, PARP1 is important for the resolution of replication fork stress (40), which our DNA fiber data indicate CF10 produces. PARG opposes PARP1 function by removing PARylation, and we recently reported studies demonstrating that PARG inhibition sensitized PDAC cell lines to DNA damaging agents (19, 43, 44) We observed antagonism for the CF10 plus olapraib combination but intriguingly, the CF10 plus PDDX-04 combination displayed strong synergy in both MIA PaCa-2 and PANC1 cells (Fig. 3A; Supplementary Fig. S7A). To gain further insight into the mechanism for this synergy, we determined separately the relative contributions of the F10 and AraC components. Bliss additivity analysis of all possible combinations of PDDX-04, F10, and AraC revealed extensive synergy for the PDDX-04 plus F10 combination relative to PDDX-04 plus AraC and mild antagonism between F10 plus AraC in both MIA PaCa-2 and PANC1 cells (Fig. 3A; Supplementary Fig. S7A and S7B), suggesting that the bulk of synergy between CF10 and PDDX-04 is indeed due to the fluoropyrimidine component. We validated the observed synergy between CF10 and PDDX-04 via siRNA oligos targeted against PARG (siPARG). siPARG treatment sensitized MIA PaCa-2 cells to both CF10 and F10, confirming the specificity of effects seen with the CF10 plus PDDX-4 combination (Table 2; Supplementary Fig. S8A).
CF10 and PDDX-04 synergy produces unresolved DNA damage which triggers apoptosis. A, Representative Bliss synergy plots of indicated PDAC cell lines concurrently treated with serial dilutions of CF10 plus olaparib, CF10 plus PDDX-04 and F10 plus PDDX-04. B, Representative immunofluorescence images of MIA PaCa-2 cells from an alkaline comet assay after treatment with CF10, PDDX-04 or their combination for 16 hours at the indicated concentrations and quantification of DNA tail moments in indicated PDAC cell lines. C, Representative Western blot analysis for PAR, γH2AX, and cleaved caspase-3 with appropriate loading controls in indicated PDAC cell lines treated with CF10, PDDX-04 or their combination for 48 hours with quantifications below. For each experiment, three independent experiments were conducted, and errors bars represent SEM. Significance was tested via two-way ANOVA with Bonferroni correction and denoted by *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.
IC50 values with SD from 5-day survival assay of MIA PaCa-2 cells treated with CF10 or F10 after 48 hours of siRNA knockdown of PARG compared with control scrambled siRNA. P values indicate differences between siCon and siPARG cells.
Because F10 is a known inducer of DNA damage (9, 28, 29), we hypothesized that combining CF10 with an inhibitor of DNA repair (i.e., PARG) would mechanistically result in unresolved DNA damage followed by induction of cell death. Alkaline comet assay was performed with CF10 and/or PDDX-04 and quantification of DNA tail length revealed that the combination of CF10 and PDDX-04 caused significant increase in tail length as compared with either treatment alone in MIA PaCa-2 (P < 0.01) and PANC1 cells (P < 0.001; Fig. 3B). Western blot analysis at 48 hours also revealed increased total cell PARylation, γH2AX serine-139 phosphorylation, and cCas3 with the CF10 and PDDX-04 combination in MIA PaCa-2 (P < 0.0001) and PANC1 cells (P < 0.05; Fig. 3C), compared with either drug alone suggesting that breaks in DNA observed previously are unable to be repaired and ultimately trigger apoptosis. These data were validated by siPARG treatment when we observed that 48 hours after CF10 treatment, siPARG-treated MIA PaCa-2 cells exhibit higher levels of total cell PARylation, γH2AX, and cCas3 compared with siControl cells (P < 0.01; Supplementary Fig. S8B). Ultimately, we believe that these data support our hypothesis that the combination of CF10 and PDDX-04 induces substantial DNA damage that is not resolved in the context of pharmacologic or genetic PARG inhibition and eventually triggers apoptosis.
Discussion
Ultimately, this study represents the first investigations of the novel polymeric fluoropyrimidine CF10 for the treatment of PDAC and we believe that our data moves CF10 closer to a clinical trial. The identification of the novel CF10 and PDDX-04 combination invites more study, beyond Chk1 and Wee1 inhibition (28, 44), involving the exploration of CF10 with other DNA repair inhibitors. One limitation of our study is the lack of bioavailability of PDDX-04 which prevents in vivo validation of CF10-PDDX synergy, but future investigations will involve validation using CF10 combined with genetic models of PARG inhibition (19, 44) and next-generation bioavailable PARG inhibitors. Future directions will include understanding why CF10 appears to preferentially synergize with inhibition of PARG but not PARP1 which could be due in part to the differential consequences of inhibiting these two DNA repair factors (40).
Authors' Disclosures
T. Golan reports grants and personal fees from AstraZeneca and Merck MSD; personal fees from Abbvie, Teva, Bayer, Bioline, and Roche outside the submitted work. C.J. Yeo reports grants from NIH and other grants during the conduct of the study. W.H. Gmeiner reports grants from NIH and Wake Forest School of Medicine during the conduct of the study; in addition, W.H. Gmeiner has a patent for CF10 in colon cancer pending. J.R. Brody reports personal fees from Perthera LLC, DOD council/peer review, editor in chief positions, DLA Piper, and Faster Better Media outside the submitted work. No disclosures were reported by the other authors.
Authors' Contributions
A.O. Haber: Conceptualization, formal analysis, validation, investigation, visualization, methodology, writing–original draft, writing–review and editing. A. Jain: Conceptualization, supervision, investigation, visualization, methodology, writing–original draft, project administration, writing–review and editing. C. Mani: Investigation, visualization, methodology. A. Nevler: Conceptualization, formal analysis, supervision, validation, investigation, visualization, methodology, writing– original draft, writing–review and editing. L.C. Agostini: Conceptualization, investigation, visualization, methodology. T. Golan: Resources, validation. K. Palle: Conceptualization, investigation, visualization, methodology. C.J. Yeo: Resources. W.H. Gmeiner: Conceptualization, resources, supervision, funding acquisition, methodology, writing–original draft, project administration, writing–review and editing. J.R. Brody: Conceptualization, resources, supervision, funding acquisition, methodology, writing–original draft, project administration, writing–review and editing.
Acknowledgments
This work was supported by NIH-NCI R21CA218933 (to W.H. Gmeiner), R01CA212600-01 (to J.R. Brody), and in part by NIH U01CA224012 (to R.C. Sears), P30CA069533 (to B.J. Druker), and P30CA056036 (to K.E. Knudsen). J. R. Brody is also supported by funding from the Pancreatic Cancer Cure Foundation, Pancreatic Cancer Action Network-AACR Research Acceleration Network Grant, grant ID 15-90-25-BROD (to J.R. Brody), and the Richard Alan Parry Pancreatic Cancer Research Fund.
Footnotes
Note: Supplementary data for this article are available at Molecular Cancer Research Online (http://mcr.aacrjournals.org/).
Mol Cancer Res 2021;19:565–72
- Received November 17, 2020.
- Revision received December 16, 2020.
- Accepted February 10, 2021.
- Published first February 16, 2021.
- ©2021 American Association for Cancer Research.
References
Volume 19, Issue 4
