Cytokine RT-qPCR and ddPCR for immunological investigations of the endangered Australian sea lion (Neophoca cinerea) and other mammals

Measurement of cytokine gene expression by reverse transcription quantitative polymerase chain reaction (RT-qPCR) is used widely to assess the immune system of animals and to identify biomarkers of disease, but its application is limited in wildlife species due to a lack of species-specific reagents. The free-ranging endangered Australian sea lion (Neophoca cinerea) experiences significant clinical disease and high pup mortality due to intestinal hookworm infection. Developing immunological tools specific to the species will aid in the assessment of drivers of disease and its impact in population demographics. This study describes the development and validation of cross-reactive RT-qPCR assays to measure five important cytokines involved in innate and Th1/Th2 responses (IL-6, TNFα, IFNγ, IL-4 and IL-10) in unstimulated blood samples from a range of different mammalian species including the Australian sea lion. All RT-qPCR assays efficiencies ranged between 87% (Ovis aries TNFα) and 111% (Bos taurus IL-10) and had strong linearity (R2). IL-4 and IFNγ gene expression for N. cinerea fell below the dynamic range (and therefore quantifiable limits) of RT-qPCR assays but were able to be quantified using the novel droplet digital PCR (ddPCR). This study delivers new immunological tools for eco-immunologists studying cytokine gene expression in wildlife species and is to our knowledge, the first cytokine ddPCR approach to be reported in a pinniped species.


INTRODUCTION
Wildlife species are exposed to a wide range of stressors, often increasing their susceptibility to disease. The endangered Australian sea lion (Neophoca cinerea) experiences high pup mortality rates at some colonies in southern Australia, limiting population growth and likely contributing to population decline (Goldsworthy, 2015;Goldsworthy et al., 2009;Marcus, Higgins & Gray, 2014;Shaughnessy et al., 2011). Disease caused by hookworms (Uncinaria sanguinis) has been identified as a significant contributor to this trend (Marcus, Higgins & Gray, 2014). Given this endemic pathogen is prevalent at 100% in pups, an understanding of the host response is likely to be informative when evaluating the potential for factors such as anthropogenic pollution, resource or genetic limitations to impact susceptibility to disease. In every species, disease outcomes result, in part, from the interaction between multiple immune cell types and specialised secretory molecules such as acute phase proteins, hormones and cytokines that work together as a network in up-regulated and down-regulated pathways. Cytokines play an essential role in both the initiation and maintenance of the immune response against pathogens, and their variations serve as indicators to assess the immune system of animals and as biomarkers of disease (Fonfara et al., 2008;Fonfara, Siebert & Prange, 2007;Murtaugh et al., 1996;Sitt et al., 2016). Wildlife researchers within the growing discipline of eco-immunology, are using approaches such as the reverse transcription quantitative polymerase chain reaction (RT-qPCR) to measure the expression of cytokine messenger RNA (mRNA) in order to understand the immune system in a wild context (Bowen et al., 2006;Bowen et al., 2012;Brock, Murdock & Martin, 2014;Shoda, Brown & Rice-Ficht, 1998). Although some progress has been made, further research is needed to evaluate the impacts of threats on individual fitness and species resilience (Bowen et al., 2012;Puech et al., 2015;Spitz et al., 2015), so as to guide management decisions aimed toward protecting threatened populations.
The majority of cytokine gene expression studies in pinnipeds have been in phocids or involved stimulated peripheral blood mononuclear cells (PMBC) (Das et al., 2008;Fonfara et al., 2008;Levin, 2018). The remote locations and challenging logistics associated with sampling the Australian sea lion and many other wildlife species generally precludes cell separation and culture methods and rather requires the use of whole blood stored in RNA preservative. RT-qPCR is the gold standard for relative gene expression and has become widely used in wildlife (Bowen et al., 2012;Bustin et al., 2005;Funke et al., 2002;Lau et al., 2015;Lehnert et al., 2019) but is limited when applied to samples with low gene transcript concentrations and/or the presence of PCR inhibitors. The more recently developed droplet digital PCR (ddPCR) (Bio-Rad, Hercules, CA, USA) has potential to overcome these limitations by partitioning a normal PCR reaction into thousands of droplets, in which fluorescent dye-based end-point PCRs occur independently, thereby increasing the likelihood of detecting low abundance targets by decreasing the effect of interfering substances and PCR biases (Rački et al., 2014). These features can allow for more precise, reproducible and statistically significant results when working with low levels of nucleic acid and variable amounts of contaminants (Baker et al., 2018;Hindson et al., 2013;Rački et al., 2014;Taylor, Laperriere & Germain, 2017) but the technique has not yet been widely applied in immunology studies of free-ranging animals.
The purpose of this study is to develop and validate RT-qPCR assays to measure, by relative quantification via the delta Ct method, five important cytokines involved in innate and Th1/Th2 responses (IL-6, TNFa, IFNγ, IL-4 and IL-10) in the Australian sea lion. A consensus sequence approach was taken, and the primers' performance in diverse domestic species was confirmed to illustrate their suitability as candidates for evaluation in future immunological studies of other mammalian wildlife species. Additionally, a subset of these primers was adapted for use in species-specific ddPCR to permit quantification of IL-4 and IFNγ mRNA, which were found to be in low-abundance in blood samples collected from Australian sea lion pups.

Gene
GenBank ensemble sequences Primers (5′->3′) (based on ** marked sequences) Annealing/ extension ( C, s) (Macrogen, Seoul, South Korea) were selected from a previous publication that used Genbank sequences for Canis familiaris (Peters et al., 2004) and its performance was optimised to the study conditions.

Blood collection, RNA extraction and reverse transcription
Blood samples (0.5 mL) from domestic dog (C. familiaris, n = 4), cattle (Bos Taurus, n = 3) and sheep (Ovis aries, n = 3) were collected from brachial, tail and jugular veins, respectively into EDTA tubes (Sarstedt, Nümbrecht, Germany) and then centrifuged at 5,000×g for up to 3 min. Plasma was removed using a sterile disposable pipette, and the remaining red blood cells and buffy coat were resuspended in 1,300 µl RNAlater TM (Applied Biosystems, Carlsbad, CA, USA) and then transferred into cryovials to approximate field storage conditions. Samples were initially stored at 4 C for 2-4 days and then at −20 C until RNA extraction was performed within 12 months of blood collection. RNA extractions from the same species were combined and used as pooled samples for further applications.
In 2010, four blood samples (0.5 mL) collected from the brachial vein from N. cinerea pups for a previous study (Marcus, Higgins & Gray, 2014) (Government of South Australia Department of Environment, Water and Natural Resources; Wildlife Ethics Committee approvals 3-2008 and 3-2011 and Scientific Research Permits A25088/4-5) were immediately transferred into EDTA tubes (Sarstedt, Nümbrecht, Germany) and processed as described above. Samples were initially stored at 4 C for 2-4 days and then at −20 C until RNA extraction was performed in 2018. RNA extractions were combined and used as pooled samples for further applications.
For RNA extractions, samples in RNAlater TM were thawed at room temperature, centrifuged at 16,000×g for 1 min, and the supernatant of RNAlater TM discarded from the cell pellet. Total RNA extraction was performed on the cell pellet using the RiboPure TM -Blood Kit (Ambion, Carlsbad, CA, USA) according to the manufacturer's instructions, such that the equivalent of 200 µl of whole blood was used per extraction. The RNA concentration and purity were assessed (A 260 /A 280 ) using a NanoDrop 1000, Thermo Scientific TM (Waltham, MA, USA) and RNA stored as multiple aliquots at −80 C for subsequent use.
For analysis, aliquots of RNA were thawed on ice, and two sequential DNase treatments were performed using the RNase-free DNase I (provided in the RNA extraction kit) to eliminate genomic DNA (gDNA). Reverse transcription was performed on 50-100 ng of RNA template in a 20 µl reaction with the RevertAid First Strand cDNA Synthesis Kit (Thermo Fisher TM , Carlsbad, CA, USA), with a combination (50:50) of random hexamer and oligo (dT) 18 primers to improve the sensitivity of cDNA synthesis (Ferrante, Hunter & Wellehan, 2018;Gallup, 2011). cDNA was stored as multiple aliquots at −20 C for subsequent use.

Real-time PCR optimisation and validation
Optimisation and validation parameters were achieved following recommendations for qPCR assays from Bustin et al. (2009) (MIQE guidelines). All qPCR assays were performed using SYBR Green (SsoAdvanced TM Universal SYBR Ò Green Supermix, BioRad) on a CFX96 Real-Time cycler (Biorad, Hercules, CA, USA). Cycling conditions were optimised by annealing temperature (T a ) gradient and evaluation of three primer concentrations (100, 200 and 300 nM). Optimal parameters were determined based on those that yielded a single, sharp peak in the melt curve analysis with the lowest quantification cycle (Ct) for each primer pair (Table 1).
Amplifications were performed in white 8-strip PCR tubes (BioRad, California, USA), following manufacturer's instructions for the SYBR Green Supermix in a 20 µl reaction with each primer pair at optimised concentration and 4 µl of cDNA template. In addition, "no-reverse transcription" controls (NRT), and "no template" controls (NTC) of RNase-DNase free water, were included in each run to check for gDNA contamination and the formation of primer-dimers, respectively. Under identical qPCR cycling conditions, reactions were validated across cDNA templates from N. cinerea, C. familiaris, B. taurus and O. aries. Amplification conditions were 95 C for 1 min (1 cycle); 95 C for 10 s and 60 C for 20 s (40 cycles). After each cycling protocol, a melt curve analysis was generated by heating from 65 C to 95 C with 0.5 C increments for 5 s to confirm the absence of non-specific products or primer dimers and define melting temperatures (T m ) for each amplicon. The size of each qPCR product was confirmed by 2% agarose gel electrophoresis and identity of the amplicon was further confirmed by DNA sequence analysis (Macrogen, Seoul, South Korea) and comparison with nucleotide sequences of terrestrial and marine mammals using the NCBI BLAST programme (Altschul et al., 1990).
Confirmed qPCR products were removed from the plates and diluted for further use as a template in standard curves. Amplification efficiencies for each gene of interest (GOI) and the reference gene (GAPDH) were determined for each species through a standard curve using serial dilutions of qPCR product with a minimum of five standards with the dilution extending at least to that producing a Ct of 34. The efficiency (E) and linearity (R 2 ) of each primer pair were calculated on these curves using the Bio-Rad CFX Maestro software 1.1 (BioRad, Hercules, CA, USA) with automatic threshold settings. Linear regression of the qPCR standard curves was recalculated with Microsoft Excel (Microsoft, Redmond, WA, USA; 2016). Primer pairs with efficiencies of 100% ± 10% and R 2 value > 0.96 were considered optimised for qPCR (IL-4, IL-6, IL-10, IFNγ and TNFa) following MIQE recommended ranges. The limit of quantification (LoQ) was based on the linear operating range of each assay (Berdal & Holst-Jensen, 2001). Assays that showed linear and efficient amplification but that produced Ct values above the dynamic range of qPCR (Ct alues >35) when applied to N. cinerea pup blood samples, were selected for subsequent development of novel ddPCR. As the assays were developed for use in relative expression studies using the delta Ct method, absolute quantification and limits of detection (LOD) were not derived using quantified standards. Quantified standards were, however used for direct comparison of sensitivity of ddPCR vs qPCR in IL-4 and IFNγ assays.
ddPCR assays for Neophoca cinerea qPCR products obtained from the amplification of N. cinerea samples with the IL-4 and IFNγ qPCR primers formerly described in this study were sequenced and aligned using the CLC Main Workbench 6.9.1 (Qiagen, Redwood City, CA, USA). NCBI Primer-BLAST programme (Ye et al., 2012) was used to design N. cinerea primer-probe pairs for ddPCR following recommended parameters from the Droplet Digital TM PCR Applications Guide (www.bio-rad.com). The primers and probes sequences (Macrogen, Seoul, South Korea) are listed in Table 2. The in silico tool 'PCR Primer Stats' (http://www.bioinformatics.org/sms2/pcr_primer_stats.html) was used to evaluate primer-probe pairs melting temperature, GC content and secondary structures (Stothard, 2000). Probes were labelled with FAM (F) fluorophore and quenched with non-fluorescent black hole quenchers number 1 (BHQ-1).
Instruments, reagents and consumables for the ddPCR workflow were supplied by Bio-Rad (Bio-Rad, Hercules, CA, USA). Optimal ddPCR annealing temperatures for IL-4 and IFNγ assays were defined by performing a temperature gradient in the annealing/ extension step of the thermal cycling protocol as suggested by ddPCR guidelines (Huggett et al., 2013). All ddPCR optimisation assays were performed using the ddPCR TM Supermix for Probes (no dUTP) master mix in a C1000 Touch TM Thermal Cycler. The fluorescence difference between NTC and positive samples within a single run was used to set a threshold between negative and positive droplets. Positive droplets show increased fluorescent amplitude when compared to the negative droplets and contain at least one copy of the target per sample. The optimal annealing temperature for these assays was defined as the one giving the largest difference in fluorescence between negative and positive droplets (Table 2; Fig. 1) (Huggett et al., 2013).
ddPCR master mix reactions included 10 µl of the Supermix, one µl of each primer and probe, six µl of cDNA from pooled N. cinerea samples and RNase-DNase free water to complete a 20 µl total volume reaction ( Table 2). PCR mix and Droplet Generation oil for Probes were added to corresponding wells in the Droplet Generator DG8 TM Cartridge and covered with DG8 TM Gaskets following the manufacturer's instructions. Droplets were generated by using the QX100 TM Droplet Generator, gently transferred to a clean 96-well PCR plate and sealed using a PCR Plate Sealer. Amplification was performed using the following cycling conditions: an initial enzyme activation period of 10 min at 95 C, followed by 40 cycles consisting of denaturation at 95 C for 30 s and annealing/extension step of 59 C for 1 min, and followed by an enzyme deactivation period of 10 min at 95 C and a 4 C indefinite hold. The overall ramp rate was set at 2 C s −1 . Droplets were read as positive or negative with a QX200 Droplet Reader, and further analysis was performed using the Quanta-Soft Analysis Pro TM software (Bio-Rad). The software, based on fluorescence amplitude, establishes a threshold between positive and negative droplets. Positive droplets are converted to copy numbers in the PCR mix based on Poisson algorithms (Droplet Digital TM PCR Applications Guide).
To assess linearity (R 2 ), efficiency (E) and to compare the limits of detection (LoD) and quantification (LoQ) of qPCR and ddPCR, IL-4 and IFNγ consensus sequences were selected to synthesise double-stranded DNA standards (gBlocks Ò gene fragments; Integrated DNA Technologies, Singapore). Each DNA standard was resuspended in Tris EDTA buffer (The Bosch Institute, Faculty of Medicine, The University of Sydney) to reach a final concentration of 10 ng µl −1 according to the manufacturer's instructions with subsequent storage at −20 C. For qPCR, a calibration curve (regression line) was performed with 10-fold serial dilutions of the standard ranging from 10 6 to one target copies µl −1 and six technical replicates for each dilution. For ddPCR, 5-fold serial dilutions ranged from 200 to 0 target copies µl −1 with three replicates for each dilution. LoD and LoQ were calculated with Microsoft Excel (2016) as 3.3 and 10 times, respectively, the standard deviation of the y-intercept of the regression line, divided by the slope of the corresponding calibration curve (FDA, 1996;Mohamad, 2018).

Real-time qPCR design and validation
The primer pairs designed in this study measured cytokine gene expression in mammalian species in separate evolutionary clades, representing terrestrial (domestic dog, sheep and cattle) and the Australian sea lion. mRNA sequences from qPCR products showed high homology to available mammalian sequence data, with values ranging from 86 to 100% (Table 3). The integrity of isolated RNA was demonstrated in all blood samples from every species by the amplification of GAPDH mRNA (Ct 25 ± 3.1, mean ± SD). Optimal qPCR parameters allowed amplification of IL-4 (Australian sea lion and dog), IL-6 (Australian sea lion and dog), IL-10 (all four targeted species), TNFa (all four targeted species), IFNγ (Australian sea lion and dog) and GAPDH (all four targeted species), using the same cycling protocol, with a combined annealing and extension step at 60 C. The defined optimal parameters for the assays are shown in Table 1. In addition, the presence of a single specific product was confirmed by melt curve analysis (Figs. S1-S4), agarose gel electrophoresis and sequencing. The absence of non-specific products and primer-dimers was confirmed (Figs. S1-S4). No amplification was detected in NRT controls and NTC. Two consecutive DNAse treatments were required to eliminate evidence of gDNA in NRT controls. All assay efficiencies ranged between 87% (Ovis aries TNFa) and 111% (Bos taurus IL-10) and had strong linearity (R 2 ) ( Table 4). Although sample sizes were too small for comparison, the low expression of IL-4 and IFNγ in N. cinerea was consistent with levels in canine samples that we assessed (Ct 32 ± 2.5, n = 4).

ddPCR assays for Neophoca cinerea
The ddPCR primers and probe assays designed in this study effectively amplified N. cinerea blood mRNA. The forward-reverse and probe sequences of each assay are listed in Table 2. Gene sequence data obtained for both primer-probe pairs was BLASTn compared against similar mammalian species genes, and the results show 98% homology with marine mammals (Table 5). No amplification was detected in NTC.
The optimal ddPCR annealing/extension temperature for IL-4 and IFNγ defined by a temperature gradient was 60 C and 58 C, respectively (Fig. 1). The defined optimal Table 3 Nucleotide identity of qPCR products for IL-4, IL-6, IFNγ, IL-10 and TNFa represented by the best match in BLASTn search (http://www.ncbi.nlm.nih.gov/BLAST/) in the four target species. parameters for the assays are shown in Table 2. LoD and LoQ were confirmed to be lower in ddPCR than in qPCR (Table 6; Fig. 2). Quantification of IFNγ and IL-4 in the pooled N. cinerea samples indicated 69 and 13 copies per reaction, respectively (Table 6).

DISCUSSION
Cytokine mRNA qPCR has become a broadly used and affordable tool to establish cytokine expression profiles in animals but has been limited by the lack of species-specific reagents (Ferrante, Hunter & Wellehan, 2018;Funke et al., 2002;Maher et al., 2014;Maissen-Villiger et al., 2016;Puech et al., 2015). In this study, optimisation and validation of SYBR green RT-qPCR assays, cross-reactive to diverse mammalian species, was performed to allow relative quantification of mRNA of five important cytokines (i.e. IL-4, IL-6, IL-10, IFNγ and TNFa). Primer sets for IL-4, IL-6 and IFNγ cross-reacted in dog and Australian sea lion samples, whereas those for IL-10 and TNFa also amplified across sheep and cattle. The broad cross-reactivity suggests that these primer sets are likely to cross react across several other mammalian species. Studies of basal cytokine levels in marine mammals are forthcoming but still limited (Beineke et al., 2007;Beineke et al., 2004;Ferrante, Hunter & Wellehan, 2018;Hofstetter et al., 2017) and, consistent with our study, others have encountered challenges with sensitivity. Sitt et al. (2016) determined that IL-4 transcripts were typically absent in killer whales (Orcinus orca) and levels of IL-4 in domestic pig (Sus scrofa) remained low even after lymphocyte stimulation (Murtaugh et al., 1996). In the present study, qPCR of mRNA from whole blood was quantifiable for IL-6, IL-10, TNFa and GAPDH but despite optimal specificity and efficiency, IL-4 and IFNγ Ct values for samples from N. cinerea pups were outside the limits of quantification (>35). Separation and stimulation of peripheral blood mononuclear cells (PMBCs) is often performed for the quantification of cytokine gene expression (Beineke et al., 2004;King et al., 1993;Levin et al., 2014) and yields much greater concentrations of target mRNA; but these methods are often not feasible under field conditions and may not represent natural expression as closely as cytokine levels in non-stimulated whole blood samples. The novel ddPCR assays overcame this limitation for samples collected from Australian sea lion pups. This method represents a very sensitive technique capable of assessing low abundant targets, confirmed by the lower LoD and LoQ than qPCR. Thus, the technique shows a lot of promise for cytokine gene expression studies involving low abundant targets or samples with carryover of inhibitors from RNA extraction or cDNA synthesis methods, as can occur when field samples associated with wildlife studies cannot be collected and stored under ideal conditions (Rački et al., 2014;Taylor, Laperriere & Germain, 2017).
It was not possible to ensure that the primer pairs developed in this study did not span exon-exon junctions. It is therefore recommended that a thorough removal of gDNA is performed when following the methods presented here. Similar to a previous study Schwochow et al. (2012), two DNase treatments were needed to ensure the removal of contaminating gDNA.
The set of primers developed in this study have potential applications to immunological studies across multiple species, including wildlife, expanding the toolkit for researchers in the future to identify immunological markers of innate, T-helper-1 and T-helper-2 pathways in mammals. To our knowledge, the ddPCR assay developed for Australian sea lions is the first one to be reported in a pinniped species and is presented as an alternative for samples that contain a low concentration of target or those that could be affected by inhibitors. However, further investigation is necessary to explore the full potential of this approach.

CONCLUSION
In summary, SYBR Green RT-qPCR assays were developed to quantify cytokine gene expression across diverse mammalian species. The diversity of species strongly suggests that the assays have potential for application beyond the Australian sea lion to many other threatened wildlife species. The sensitivity of methods described here indicates that most are of use in mRNA extracts from whole blood, increasing their utility for analysis of field samples, where immediate sample processing is limited. Conveniently, they can also be applied under the same optimised cycling conditions in their respective targeted species. The novel ddPCR methods described here enabled detection of low expressed genes, like IL-4 and IFNγ in N. cinerea pups, providing comparative advantages when working with unstimulated tissues and limiting sample volumes as in the case of fieldwork-based wildlife studies.

Author Contributions
María-Ignacia Meza Cerda conceived and designed the experiments, performed the experiments, analysed the data, prepared figures and/or tables, authored or reviewed drafts of the paper, and approved the final draft. Rachael Gray conceived and designed the experiments, authored or reviewed drafts of the paper, and approved the final draft. Damien P Higgins conceived and designed the experiments, analysed the data, authored or reviewed drafts of the paper, and approved the final draft.

Animal Ethics
The following information was supplied relating to ethical approvals (i.e., approving body and any reference numbers): The Government of South Australia, Department of Environment, Water and Natural Resources, Wildlife Ethics Committee approvals provided full approval for this research .

Field Study Permissions
The following information was supplied relating to field study approvals (i.e., approving body and any reference numbers): Fieldwork was approved by the Government of South Australia Department of Environment, Water and Natural Resources [Scientific Research Permits (A25088/4-5)].

Data Availability
The following information was supplied regarding data availability: Raw data are available in the Supplemental Files.

Supplemental Information
Supplemental information for this article can be found online at http://dx.doi.org/10.7717/ peerj.10306#supplemental-information.