Tasmanian devil cathelicidins exhibit anticancer activity against Devil Facial Tumour Disease (DFTD) cells

The Tasmanian devil (Sarcophilus harrisii) is endangered due to the spread of Devil Facial Tumour Disease (DFTD), a contagious cancer with no current treatment options. Here we test whether seven recently characterized Tasmanian devil cathelicidins are involved in cancer regulation. We measured DFTD cell viability in vitro following incubation with each of the seven peptides and describe the effect of each on gene expression in treated cells. Four cathelicidins (Saha-CATH3, 4, 5 and 6) were toxic to DFTD cells and caused general signs of cellular stress. The most toxic peptide (Saha-CATH5) also suppressed the ERBB and YAP1/TAZ signaling pathways, both of which have been identified as important drivers of cancer proliferation. Three cathelicidins induced inflammatory pathways in DFTD cells that may potentially recruit immune cells in vivo. This study suggests that devil cathelicidins have some anti-cancer and inflammatory functions and should be explored further to determine whether they have potential as treatment leads.

At 0 h, 100 μL of each peptide dilution was then added to each plate in quadruplicate for all timepoints (12, 18, 24 and 36 h), resulting in a final peptide concentrations of 500 μg/mL, 250 μg/mL, 125 μg/mL, 62.50 μg/mL, 31.25 μg/mL and 15.62 μg/mL.To ensure that the solvent did not affect absorbance values, a vehicle control was also included for the maximum incubation period (36 h).An untreated growth control of RPMI 1640, a positive control of dimethylsulfoxide (DMSO) (15%) and sterility control of RPMI 1640 only (no cells) were also included for each time point.
For the 12 h timepoint, 10% alamarBlue was immediately added to each well of the plate after addition of the peptide.For the remaining timepoints, plates were incubated for an additional 6, 12 or 24 h after addition of the peptides, then 10% alamarBlue added to each well.For all timepoints, after the addition of alamarBlue cells were incubated for a further 12 h, resulting in a total peptide incubation period of 12, 18, 24 and 36 h.Absorbance of alamarBlue for all timepoints was determined at 570 nm and 630 nm on a Biotek 800 TS microplate reader.Cell viability was calculated according to manufacturer's instructions and expressed as a percentage of cell survival compared to the untreated growth control.
A one-sample, one-tailed t-test was used to test for significant difference in viability between treated cells and the negative growth control.A one sample, two-tailed t-test was also conducted to ensure there was no significant difference in viability between the negative growth control and the cells treated with the solvent only.Statistical analysis was conducted in R (R Development Core Team 2021).

Mechanism of action.
RNAseq was used to characterize the mechanisms underlying the anticancer activity of the peptides.Confluent DFT1 1426 cells (passage number 25) were re-suspended at a concentration of 5 × 10 5 cells/ml and 1 mL seeded into each well of a sterile flat-bottom polystyrene 12-well plate (Corning).
Following incubation for 24 h at 35 °C 5% CO 2 , media was aspirated from the cells and replaced with 500 μL of RPMI 1640 with 2 mM l-glutamine, 10% FBS and 10% AmnioMax II.500 μL of each 1 mg/mL peptide solution (Saha-CATH1 to 7) was then added to the plate in triplicate, to give a final peptide concentration of 500 μg/ mL.This was the highest concentration tested in the previous cytotoxicity assay and hence would likely lead to maximal peptide activity, enabling precise analysis of changes in gene expression.A vehicle control was also added to the plate in triplicate.
The plate was then incubated for a further 10 h at 35 °C in 5% CO 2 .This incubation period was selected as the cytotoxicity assay indicated it was sublethal for most of the peptides and would maintain sufficient cell viability to extract intact RNA.Media was then aspirated from the cells, which were washed twice with Dulbecco's phosphate buffered saline (DPBS) (Sigma-Aldrich).
Total RNA was extracted from each replicate of the peptide-treated and vehicle control cells (n = 3) using the RNeasy mini kit (Qiagen) with cell lysis directly in each well of the 12 well plate.Total RNA was quality assessed using the RNA nano 6000 kit on the Agilent Bioanalyzer with all samples displaying a RIN score between 5 and 9.In total, 24 RNA samples corresponding to three treatments per peptide (Saha-CATH1 to 7) and a vehicle control were submitted to Ramaciotti Centre for Genomics (The University of New South Wales) for sequencing.Illumina TruSeq mRNA libraries were prepared for all samples, which were sequenced as 2 × 150 bp paired-end reads across an SP flowcell on the NovaSeq6000.This resulted in 27 to 50 million raw reads per sample.
Raw reads were quality assessed using FastQC v0.11.8 47 , then quality and length trimmed using Trimmomatic v0.39 48 using default parameters.Trimmed reads for each treatment (n = 3) and the control (n = 3) were aligned to the Tasmanian devil reference genome v7.0 (NCBI: GCA_000189315.1 49 using STAR v2.7.8a 50.Alignments for each treatment were summarized into gene counts using featureCounts in the subread package v1.5.1 51 . Gene counts were used as input for differential expression analysis in R 52 .Genes with less than 50 counts across all samples were removed from the analysis, as these were unlikely to be biologically relevant and lacked sufficient statistical power.Firstly, the data was normalized by trimmed mean of M values (TMM) using edgeR v3.32.1 53 to account for any composition bias between libraries and to give an effective library size for downstream analysis.Multidimensional scaling (MDS) was used to check for variation between treatments using limma v3.36.0 54 .Expression levels were then normalized using upper-quartile normalization in EDAseq v2.24.0 55 to account for differences in distribution between lanes, such as sequencing depth.Differential expression analysis was performed using voom in the limma v3.36.0 package 54 .
For each treatment, a false discovery rate (FDR) cutoff of 0.02 was applied and genes that were up or downregulated greater than 1.5× fold were selected for Gene Ontology (GO) and Ingenuity Pathway Analysis (IPA) 56 .Over-representation analysis of Biological Processes was conducted in clusterProfiler v3.18.1 57 .Statistical significance was adjusted for multiple comparisons using the Benjamini-Hochberg method, and terms were considered significant when p-adj < 0.05.To remove general terms, gene sets larger than 200 were removed, and the simplify function was used to remove redundant GO terms.Further pathway analysis was conducted in IPA (Qiagen) 58 , using mammal as the species type and nervous system for the tissue/cell type.

Results
Four Tasmanian devil cathelicidins (Saha-CATH3, 4, 5 and 6) significantly decreased cell viability by more than 50% over 36 h compared to the growth control at 500 μg/mL.Saha-CATH5 displayed the most rapid cytotoxic activity against DFT1 cells, and reduced cell viability to less than 0% at all time points (p-value < 3.63E−09) (Fig. 1).This negative viability is likely caused by the change in media pH over the incubation period.Cell viability was calculated by measuring the amount of reduced alamarBlue (AR).This calculation requires a correction factor to allow for the oxidized substrate present in the media.In cases of very low survival, the change in pH between the media of the treated cells and the media used to calculate the correction factor can result in a negative AR value.Saha-CATH3 (p-value < 2.44E−06) and Saha-CATH6 (p-value < 7.12E−05) required an 18-h incubation period before exhibiting a similar level of toxicity to Saha-CATH5.Saha-CATH4 also significantly reduced cell viability (p-value = 0.04) but was slower acting, requiring 36 h to induce significant cytotoxicity by more than 50%.
Differential expression (DE) analysis identified 12,402 DE genes out of a total of 15,548 across all seven treatments when compared to the control.Most of these were in the Saha-CATH5 treatment (11, 514 or 74.05%).The other toxic treatments had between 10 and 20% DE genes-Saha-CATH3 had 1, 963 (12.63%),Saha-CATH4 had 2, 915 (18.75%) and Saha-CATH6 had 2, 419 (15.56%).All the DFT1 cells treated with the non-toxic peptides Saha-CATH1, 2 and 7 had less than 1% DE genes, with Saha-CATH7 having 0 under the quality filters chosen.Due to the low number of differentially expressed genes, these treatments were excluded from further analysis.

Discussion
Here we aimed to identify cathelicidin candidates for future development as anti-DFTD therapeutics.Four Tasmanian devil cathelicidins (Saha-CATH3, 4, 5 and 6) were toxic to DFT1 1426 cells at high concentrations: Saha-CATH3 at concentrations ≥ 62.5 µg/mL; Saha-CATH4 at 500 µg/mL; Saha-CATH5 at concentrations ≥ 125 µg/ mL and Saha-CATH6 at concentrations ≥ 250 µg/mL.Previous studies have shown that two of these peptides (Saha-CATH3 and 4) are non-toxic to human cell line A549 44 .The other two (Saha-CATH5 and 6) reduced cell viability at high concentrations (500 µg/mL), resulting in 42% and 59% cell survival respectively 44 .At the same concentrations and incubation period in this experiment, Saha-CATH5 and 6 resulted in 0% cell survival (Fig. 1).Interestingly, the fastest acting peptide has also shown broad-spectrum antimicrobial activity 44 .One of the toxic peptides (Saha-CATH4) has not previously shown antimicrobial activity in vitro.This may indicate that Saha-CATH4 also has antimicrobial activity against pathogens that have not been tested or is slower acting against the strains it has been tested against.
Saha-CATH3, 4 and 5 induced cell cycle arrest at the G1/S phase.The cell cycle is composed of four phases (M, G1, S, G2) during which DNA is replicated and the cell is divided in two 59 .Cell progression from one phase to the next is primarily triggered by cyclin dependent kinases (CDKs).Although CDKs are constantly expressed, they only become activated upon binding with a specific cyclin subunit 60 .Cyclin expression oscillates throughout the four phases to coordinate the cell cycle 61 .The transition between the G1 and S phase is triggered by cyclins D and E and signifies the point at which cell growth ceases and DNA replication begins 62,63 .It is accompanied by an increase in proteins that form the pre-replication complex (pre-RC) required for chromosomal replication such as ORC1 64 .Both the cyclins and elements of the pre-RC are overexpressed in tumours and have been identified as potential therapeutic targets [65][66][67][68][69][70][71] .
Many of these elements were downregulated in DFT1 1426 cells treated with Saha-CATH3, 4 and 5 (Table 1).This aligns with the activity of cathelicidins from other species that also suppress genes involved in DNA replication and cell cycle progression 72,73 .The data suggests that Saha-CATH3, 4 and 5 were inducing cell cycle arrest at the G1/S phase and should be further explored as potential therapeutics.
Alongside cell cycle arrest, Saha-CATH5 also regulated the ERBB and Hippo signaling pathways.It is particularly interesting that the ERBB3 gene was significantly downregulated by Saha-CATH5.This receptor has undergone copy gains in the DFTD genome 15 and the ERBB-STAT3 axis has been identified as an important driver of DFTD 17 .Although DFTD cells are sensitive to multiple RTK inhibitors, they appear to respond most to those acting via this pathway 15 .This suggests Saha-CATH5 may be acting via a similar mechanism.
The oncogenic activity of RTK signaling pathways is often amplified by the formation of a positive feedback loop with the Hippo pathway 74,75 .When the Hippo pathway is dysregulated, the transcription factor YAP1 accumulates in the nucleus 76,77 .This can cause overexpression of many genes involved in proliferation and survival 78 .WWC3 attenuates this process by phosphorylating YAP1 to prevent its nuclear translocation 79 .WWC3 has undergone hemizygous deletion in DFTD, potentially resulting in overexpression of YAP1 15 .Therefore, the high toxicity of Saha-CATH5 against DFT1 1426 cells may be due to the peptide targeting the synergistic activity of both RTK signalling and YAP1 expression.RTK inhibition has not yet been documented for any cathelicidins.The human cathelicidin LL-37 has even shown the opposite by activating two RTKs (EGFR and ERBB2), promoting tumour progression in certain cell lines 23,26 .However, RTK inhibitors are a major class of therapeutics against a wide variety of cancers 80,81 .This reveals an important property of Saha-CATH5 that has strong potential for future drug development as a therapeutic for DFTD.
GO analysis of Saha-CATH6 treatment suggested ER stress.These include glycosylation inhibition, calcium depletion and elevated reactive oxygen species (ROS) [82][83][84][85] .ER stress is induced when excessive misfolded proteins start to accumulate and disrupt homeostasis 84 .The unfolded protein response (UPR) aims to alleviate this stress by reducing protein translation, increasing chaperone expression, and degrading malformed proteins.If prolonged, the UPR leads to apoptosis 86 .
Other enriched GO terms in Saha-CATH6-treated DFT1 cells also indicated signs of ER stress.For example, protein hydroxylation was downregulated, suggesting attenuation of ER function that is characteristic of a stress response.Increased mRNA catabolism may be indicative of Regulated Ire1-Dependent Decay (RIDD).This process is upregulated during the UPR and involves degrading mRNAs to reduce the burden on the ER 87 .GO terms also suggested that translation was increased.Although this may appear contradictory as global translation is reduced during ER stress, the expression of ER chaperones and ERAD components is upregulated 85 .This is one of three main pathways activated by the UPR.It attenuates global translation, while increasing expression of protective proteins [88][89][90] .The other two main pathways involved in the UPR are activated by inositolrequiring enzyme 1-a (IRE1a) and activating transcription factor 6 (ATF6) 86 .The increased mRNA catabolism

Figure 1 .
Figure 1.Changes in DFT1 1426 cell viability over time at peptide concentrations of 500 µg/mL.Cell viability is expressed as a percentage of cell survival compared to the untreated growth control.The mean values ± SD (error bars) of the assay performed in quadruplicate are reported.3 of the toxic peptides (Saha-CATH3, Saha-CATH5 and Saha-CATH6) reduced cell viability to 0% after 18 h, while Saha-CATH4 reduced cell viability by 50% after 36 h.The other 3 the peptides (Saha-CATH1, Saha-CATH2 and Saha-CATH7) did not reduce cell viability.

Figure 2 .
Figure 2. GO terms downregulated in (a) SahaCATH3 treatment, (b) SahaCATH4 treatment and (c) SahaCATH5 treatment.Terms associated with cell cycle and DNA repair/checkpoints were downregulated in all three.In Saha-CATH5 treatment, terms associated with ERBB and YAP1 signalling were also downregulated.

Figure 3 .
Figure 3. GO terms (a) upregulated and (b) downregulated in Saha-CATH6 treatment.Terms associated with an immune response were upregulated.Treatment also indicated signs of ER stress.

Table 1 .
DE genes in SahaCATH3, 4 and 5 treatments involved in DNA replication and cell cycle progression.