Human bronchial epithelial cells (16HBE14°−) were obtained from Professor Dieter Gruenert (UCSF). Cell culture vessels were coated with 100 μg/ml bovine serum albumin fraction V (Sigma Aldrich, Dorset, UK), 0.5 μg/ml Cultrex mouse collagen IV (R&D Systems, Abingdon, UK), and 1 μg/ml human fibronectin (EMD Millipore, Hertfordshire, UK) before seeding. Cells were grown in Iscove's Modified Dulbecco's Medium (IMDM), GlutaMAX™ supplemented with 10% FBS (fetal bovine serum) and 1% Penicillin-Streptomycin (Thermo-Fisher, Loughborough, UK). WI-38 fetal lung fibroblasts and Hela cells were obtained from ATCC. These cells were grown in Dulbecco's modified Eagle's medium (DMEM) with 10% FBS and 1% Penicillin-Streptomycin (Thermo-Fisher, Loughborough, UK). All cells were grown in a heated humidified incubator at 37°C, 5% CO_2. Cells were detached with 0.05% Trypsin-EDTA solution (Thermo-Fisher, UK) and seeded at 5 × 10⁴ cells/ml for 12-well plates and allowed to grow for 24 h before infection or stimulation experiments. RNAzol-RT and Saponin were all obtained from Sigma-Aldrich (Dorset, UK).

Antibodies against Citrullinated Histone H3 (ab5103), PAD2 (ab16478), PAD4 (ab128086), PAD3 (ab183209), Tubulin and β-actin (ACTN05) were purchased from Abcam (Cambridge, UK). The anti-peptidyl-citrulline antibody, clone F95 (MABN328) was from EMD Millipore (Hertfordshire, UK). Antibodies against total histone H3 (1B1B2) and (96C10) were from Cell Signaling Technologies (Hertfordshire, UK). PAD1 (HPA062294) was from Sigma Aldrich (Dorset, UK), and PAD2 (12110-1-AP) was from Proteintech (Manchester, UK). All secondary antibodies used were raised in goat. F(ab')2 anti-Rabbit IgG A647, F(ab')2 anti-Mouse IgG PE (A10543), anti-Mouse IgG A594 (A11032) and anti-Mouse IgG A488 (A11029) were from Thermo Fisher Scientific (UK). Anti-Mouse IgM Texas Red was from VectorLabs (Peterborough, UK) and Anti-Rabbit IgG FITC was from Sigma Aldrich (Dorset, UK).

For flow cytometry, the following mouse anti human antibodies CD3 PEcy7, CD16 PerCP-Cy 5.5, CD19 BV786, CD56 BV650, and CD14 PE were obtained from BD Biosciences (Oxford, UK).

Human venous blood was collected in accordance with local institutional ethical and health & safety regulations. Peripheral blood mononuclear cell (PBMC) and granulocyte cell fractions were separated from the blood of healthy donors using dextran sedimentation (Pharmacosmos, Reading, UK) followed by discontinuous isotonic Percoll gradient centrifugation (GE Healthcare, Amersham, UK). Granulocytes and PBMC were washed twice in 1x DPBS without calcium or magnesium before resuspension in RPMI 1640 (Thermo-Fisher, Loughborough, UK). Purity was established by morphological assessment and FACS analysis. PBMC were seeded in triplicate at 2 × 10⁵ cells/ well in round bottom 96-well plates (Corning, Flintshire, UK) and rested for 2 h in RPMI 1640 5% FBS before RV1B exposure for 1 h in serum free RPMI media.

Peptides were assembled using the Fmoc/tBu solid-phase peptide synthesis approach (39) using either model 433A (Applied Biosystems, CA, USA) or model Liberty (CEM Corporation, NC, USA) automated peptide synthesizers followed by cleavage in the trifluoroacetic acid (TFA)/phenol/thioanisole/ethanedithiol/water (10:0.75: 0.5:0.25:0.5, w/w) mixture at 25°C for 90 min followed by precipitation with cold diethyl ether. The crude peptides were purified by preparative reversed-phase high-pressure liquid chromatography (RP-HPLC). The peptide purity (>98%) was confirmed by analytical RP-HPLC, and the masses were confirmed by mass spectrometry. Following lyophilization, the purified peptides were obtained in the form of their TFA salts. Stock solutions (5 mg/mL) were prepared in ultrapure DNAse/RNAse free water (Thermo-Fisher, Loughborough, UK) and stored in aliquots at −80°C. The sequences of the peptides used in this study are shown in Table 1.

To generate viral stocks, HRV-16 and HRV-1B (Public Health England Virus Collection, Salisbury, UK) were propagated in Hela cells grown at 33°C/5% CO₂. Hela cells were tested and shown to express high levels of ICAM-1 and LDLR by flow cytometry. After 5 days, HRV infected cell or non-infected cells (“Hela Lysate” control) were subjected to three freeze-thaw cycles, harvested and spun to remove cellular debris (3,000 × g for 30 min) and frozen at −80°C.

Virus stocks were concentrated and partially purified (40, 41). Briefly, infected cell lysates were concentrated by centrifugation with a 100,000-molecular weight cut-off Vivaspin-20 filter (2,000 rpm for 2 h at 4°C; GE Healthcare, Amersham, UK). Soluble factors from cell origin pass through the filter while the virus is concentrated. Concentrated virus was then layered on a 30% sucrose layer (30%, w/v, pH 7.2 in DPBS) and ultracentrifugated in Quick-Seal® Polypropylene 8 mL tubes using a SW32.1Ti swinging bucket rotor (Beckman Coulter) for 3 h at 25,000 rpm at 16°C. Cell pellets were washed, resuspended in DPBS and aliquoted at −80°C.

Virus titrations were performed in sub-confluent WI-38 cultures in 96-well plates using serial dilutions from the viral stocks. Cells were infected for 2 h in serum free DMEM, washed to remove inoculum and left for 6 days in DMEM 5% FBS at 33°C/5% CO₂ until CPE assessment was performed and TCID₅₀ calculated using the Spearmen & Kärber algorithm and Reed Muench method (47).

Total RNA was extracted from 12-well plates using RNAzol RT. RNA concentrations were determined using a NanoDrop 1000 spectrophotometer (Thermo Fisher Scientific, Loughborough, UK). RNA integrity was checked using a 2100 Bioanalyzer (Agilent, UK) with RIN values routinely >8. Total RNA (1 μg) was transcribed to cDNA with a Precision RT all-in-one mix kit (PrimerDesign Ltd, Southampton, UK). cDNA (25 ng), 250 nM specific primers (Table 2) and SYBR green mastermix (PrimerDesign Ltd, Southampton, UK) were used to perform qPCR reactions in total volume of 20 μL in a StepOnePlus instrument (Applied Biosystems). A panel of six human reference genes were evaluated using a geNorm kit (PrimerDesign Ltd, Southampton, UK) and qbase⁺ (Biogazelle, Belgium) software. The geometric means of ACTB and GAPDH genes were selected as the most stable combination for normalization in 16HBE14°− cells. The 2^−ΔΔCt method was used and data is represented as fold change over untreated cells.

To determine viral RNA copies in cells, total RNA was extracted as indicated above, and a one-step RT-qPCR was performed using 5 ng of total RNA as template together with a Precision OneStepPLUS Master Mix (PrimerDesign Ltd, Southampton, UK), and specific primers and FAM-taqman probes included in the Human Rhinovirus kit (PrimerDesign Ltd, Southampton, UK). RNA viral copies were extrapolated from the Ct values generated using the standard curve.

ELISA measurements of IL-6, IL-8, and CCL-5 were performed according to manufacturer's instructions (R&D Systems, Abingdon, UK).

Cells were lysed in ice-cold RIPA buffer (150 mM NaCl, 1.0% IGEPAL® CA-630, 0.5% sodium deoxycholate, 0.1% SDS, 50 mM Tris, pH 8.0) from Sigma Aldrich (Dorset, UK), supplemented with protease inhibitors (cOmplete™ Mini, EDTA-free; Roche, West Sussex, UK). Lysates were subjected to sonication (3 bursts of 10 s) and centrifuged at 12,000 x g for 15 min at 4°C. Proteins in the soluble fraction were quantified using a Bradford Assay (Sigma Aldrich, Dorset, UK) and heated (95°C for 5 min) in SDS-PAGE loading buffer. Equal amounts of protein were separated by electrophoresis in 4–20% TGX mini-protean gels (Biorad, Hertfordshire, UK). Samples were transferred to PVDF membranes (Millipore) using a wet blotting system and visualized with Ponceau S staining (Thermo Fisher, Loughborough, UK). Membranes were blocked in 10% skimmed milk powder (Marvell) in Tris Buffered Saline (TBS) 0.1% Tween-20 (TBS-T) for 1 h at room temperature. The following primary antibodies were diluted in 3% milk TBS-T and incubated overnight at 4°C; α-citrullinated histone H3 (1/2,500), α-PAD2 (1/1,000), α-PAD4 (1/2,000), α-total histone H3 (1/2,000), α-β-actin (1/4,000). To detect protein citrullination, the F95 antibody was used at 1/1,000 with 5% BSA in TBS-T as blocking solution. Membranes were incubated with α-Mouse IgG 800CW; α-Mouse IgM 800CW or α-rabbit IgG 680IRdye secondary antibodies (LI-COR Biosciences Odyssey® Infrared Imaging System, Cambridge, UK). Alternatively, horseradish peroxidase (HRP)–conjugated antibodies (Agilent, Cheshire, UK) and ECL Plus reagents (GE Healthcare, Amersham, UK) were used. Image Studio Lite software (Licor) was used to quantify protein bands.

Cells were seeded on glass chambered cell culture slides (Corning, Flintshire, UK) at 2 × 10⁴ cells/well. After the indicated exposures, cells were washed with DPBS and fixed with 4% PFA for 20 min at room temperature (RT). Cells were permeabilized with 0.1% Triton X-100 and blocked with PBS 5% normal goat serum (blocking buffer). Primary antibodies were diluted in blocking buffer and incubated overnight at 4°C. F95 and PAD2 antibodies were used at 1/100 dilution. After washing, goat anti-rabbit or mouse IgM secondary antibodies diluted in blocking buffer (1/1,000) were incubated with samples for 1 h at RT. Final mounting was performed in Vectashield aqueous mounting medium containing DAPI (Vector Laboratories, Peterborough, UK), prior to visualization on a Zeiss LSM 880 AxioObserver Z1 confocal fluorescent microscope (wavelengths: 405, 488, 594 nm, laser power 2%).

16HBE14°− cells were seeded at a density of 1 × 10⁵ cells/well in 12-well plates in duplicate wells per experimental condition. After treatment, cells were washed, detached with 0.05% Trypsin-EDTA, placed in FACS tubes and fixed with 4% PFA at room temperature for 15 min. Cells were washed with PBS and permeabilised by washing twice in PBS with 0.05% saponin. After a blocking step in 5% goat serum, primary antibodies were left overnight at 4°C in PBS with 0.05% Saponin and 5% goat serum. Fluorochrome conjugated secondary antibodies were added for 1 h at RT, and after final washing, cells were acquired in a FACS Celesta instrument (BD Biosciences).

Previous studies have shown that LL-37 is susceptible to citrullination by PAD enzymes, with arginines 7, 29, and 34 in the sequence being preferentially targeted (30, 35). To assess the impact of citrullination on the antiviral activity of LL-37, we generated variants of LL-37 peptides with varying numbers of arginine residues replaced with citrulline (Table 1). We then tested the ability of the modified peptides to inhibit HRV infection using the same prophylactic regime established previously (36). Briefly, RV1B virus was incubated with native or modified LL-37 peptides and after 2 h incubation the peptide/virus mixture was placed in contact with 16HBE14°− bronchial epithelial cells for 1 h to allow infection. After washing, cells were left for 24 h at 33°C to allow viral replication. Cells were harvested and HRV copy number quantified by qPCR. The addition of native LL-37 reduced HRV copy number by ~80% at both 10 and 30 μg/ml (Figures 1A,B).

Substitution of one arginine residue to citrulline (LL-37_1cit,) did not alter the antiviral effect at either concentration tested. The substitution of two arginine residues (LL-37_2cit) reduced LL-37 activity against HRV at 10 μg/ml (Figure 1A). However, this effect was not observed at the higher (30 μg/ml) concentration of LL-37_2cit (Figure 1B). Notably, LL-37_3cit lacked all antiviral activity at 10 μg/ml displaying a statistically significant difference to native LL-37. However, at 30 μg/ml, LL-37_3cit, retained some activity against HRV. When all arginines in LL-37 were substituted with citrulline (LL-37_5cit,) the antiviral effect was significantly reduced when compared to native LL-37 at 30 μg/ml, and with an apparent, but non-statistically significant loss of antiviral activity at 10 μg/ml. We also assessed a peptide with identical amino acid composition to LL-37 but with a scrambled sequence (SCR) as a control for exogenous peptide supplementation and to confirm that effects of LL-37 were sequence specific (19). The addition of LL-37-SCR at 30 μg/ml did not cause a reduction in viral copy number (Figure 1B).

To assess viral particle release from host cells, supernatants collected from the same experiments were serially diluted and used to infect WI-38 cells to measure viral titer by TCID₅₀ assay. In accord with the viral copy number assessment, virus titer in cell supernatants decreased with LL-37 addition (Figure 1C) and a gradual loss of this effect was observed when two or more arginine residues were substituted with citrulline.

To address whether the impact of citrullination on LL-37 function also extended to the antibacterial activity of the peptide (30), 30 μg of native and modified peptides were incubated in direct contact with Staphylococcus aureus (S. aureus) for 2 h, before enumeration on an agar plate after 24 h growth. A 97% decrease in S. aureus CFU/ml was observed in the presence of LL-37 (Figure 1D), an antibacterial effect that was gradually reduced with each arginine to citrulline substitution, with LL-37_2cit, LL-37_3cit, and LL-37_5cit all exhibiting statistically significant differences to native LL-37 (p ≤ 0.05, p ≤ 0.01, and p ≤ 0.001; respectively). Taken together, these data indicate that citrullination of LL-37 profoundly affects its direct antiviral and antibacterial activity.

HRV infections are known to induce pro-inflammatory cytokine and chemokine release from epithelial cells (2) and elevated cytokines, in particular IL-8, correlate well with disease severity (9). Thus, we examined the potential for native and citrullinated forms of LL-37 to modulate HRV-induced pro-inflammatory cytokine release. We assessed mRNA expression of IL-8 (Figure 2A), IL-6 (Figure 2B), and CCL5 (Figure 2C) by qPCR. Infection with RV1B induced a marked increase in IL-8 (17-fold over untreated), IL-6 (7-fold) and CCL5 (33-fold). Addition of LL-37 reduced virus induced IL-8 mRNA expression by 81% (^*p ≤ 0.05) and CCL5 by 84% (^***p ≤ 0.01) and also resulted in a marked reduction of 71% in IL-6 mRNA levels (^*p ≤ 0.05). In agreement with the effects of citrullinated LL-37 on viral copy number, LL-37_1cit and LL-37_2cit caused a reduction in the levels of IL-8, CCL5, and IL-6 mRNA induced by RV1B. Citrullination of three arginine residues in LL-37_3cit further reduced the immunomodulatory activity of the peptide, whereas LL-37_5cit completely lacked the ability to reduce IL-8, IL6, or CCL5 mRNA, showing statistically significant differences when compared to native LL-37. Addition of the LL37SCR control peptide did not result in any measurable changes in cytokine expression in response to HRV.

To assess changes at the protein level, we examined cell supernatants by ELISA (Figures 2D–F). Notably, HRV infection induced an increase in CCL5 secretion compared to untreated cells, while the addition of LL-37 reduced this effect. The effects of LL-37_1cit were indistinguishable from those of native LL-37, whereas LL-37_2cit, and LL-37_3cit showed a moderate loss in their ability to reduce HRV-induced CCL5 secretion. LL-37_5cit exhibited no effect at any of the concentrations tested. However, in contrast to gene expression, the data quantifying IL-6 and IL-8 secretion was not statistically significant at this timepoint. Taken together, this data suggests that LL-37 can modulate some virus-induced proinflammatory cytokine gene expression and protein release while total citrullination of the peptide abrogates this function.

Our results demonstrating that citrullination has a negative impact on LL-37 function together with the increased presence of PAD enzymes observed in lungs under inflammatory conditions (29–31) led to the hypothesis that HRV infection may increase levels of PAD enzymes and thus citrullination in epithelial cells. We obtained RNA and protein from 16HBE14°− cells post HRV infection and assessed PAD enzyme expression. In our model, PADI2 mRNA was the most abundant isoform in uninfected 16HBE14°− cells, being expressed at higher levels than both PADI1 and PADI4. PADI3 mRNA levels were on the limit of detection (data not shown). Given that LL-37 is susceptible to PADI2 and PADI4 enzymatic activity (30), we focused upon these two isoforms. Infection of 16HBE14°− cells with HRV did not significantly alter mRNA levels of either PADI2 (Figure 3A) or PADI4 (Figure 3D). In contrast, protein levels of PAD2 (Figures 3B,C) and PAD4 (Figures 3E,F) appeared to moderately increase after 24 and 48 h of HRV infection, with comparable increases observed with the synthetic dsRNA analog Poly I:C as measured by FACS. Importantly, this effect was not observed when viral particles were UV-irradiated, suggesting that in the context of HRV infection, active replication of the virus is required for the increase in PAD protein observed. We used Western immunoblotting to analyse PAD2 protein levels in 16HBE14°− cells 24 h after infection and a moderate increase in PAD2 was observed after 24 h of HRV MOI 5 infection or treatment with Poly I:C (Figures 3G,H). This was confirmed through the use of confocal microscopy to analyse PAD2 protein levels in 16HBE14°− cells where a clear increase in PAD2 staining was also observed after 48 h of HRV MOI 5 infection, further confirming our observations by FACS and Western immunoblotting (Figure 3I). PAD4 protein levels, assessed by Western immunoblotting, remained unchanged (data not shown).

Interestingly, when Poly I:C was used as a synthetic analog of a dsRNA viral infection, there was a statistically significant and marked drop in PADI2 mRNA expression at 24 h and 48 h (p ≤ 0.05 and p ≤ 0.01, respectively; Figure 3A). A similar, statistically significant drop was also observed for PADI4 mRNA at 24 h (p ≤ 0.05, Figure 3D). However, protein levels of these two enzymes were moderately increased both at 24 and 48 h time points in response to Poly I:C administration at levels comparable to those elicited by HRV infection (Figures 3C,F,G,H).

Although epithelial cells are the main site of HRV replication, immune cells are key contributors to the inflammatory response to HRV. We aimed to assess whether HRV also increased PAD2 levels in PBMCs, primary cells with well-characterized levels of PAD2 enzyme (48, 49). We obtained PBMCs from healthy donors and exposed them to RV1B or poly I:C for 24 h or 48 h. PAD2 expression was subsequently assessed in different immune subsets by FACS analysis. We found that CD14⁺ CD16^Low monocytes (Figure 4C) expressed higher levels of PAD2 than CD19⁺ B-cells (Figure 4B) or than CD3⁺ T-cells (Figure 4A). Interestingly, HRV (MOI 5) infection or Poly I:C stimulation resulted in a general increase in the levels of PAD2 across the different PBMC subsets analyzed. T-cells, NK-T cells and B-cells (Figures 4D-G) showed an increase in PAD2 levels at 24 h post-stimulation (data not shown), while in NK cells and monocyte subsets the increase PAD2 was more pronounced at 48 h (Figures 4F,H,I). In addition, increases in PAD2 were abolished when UV-irradiated viral particles were used as control (data not shown). A statistically significant increase in PAD2 expression was observed after 48 h RV1B MOI 5 infection, in the CD14⁻ CD16⁺⁺ subset, confirming that HRV viral infection is also able to increase PAD2 levels in non-epithelial cells. These observations suggest that viral infection or binding of ligand to Toll-Like Receptor-3 (TLR-3) can result in altered PAD expression, which has the potential to impact on global protein citrullination.

To address whether the increased levels of PAD enzymes observed after HRV infection resulted in increased protein citrullination, we used a well-characterized monoclonal antibody (F95) for the specific detection of citrullinated peptides (29, 50). Using lysates from infected, poly I:C treated and untreated cells for immunoblotting with the F95 antibody, we determined that both HRV infection and poly I:C stimulation increased the intensity of citrullinated proteins at 35–40 kDa. In addition, a clear increase in a band ~15 kDa was observed in the presence of Poly I:C and HRV treatment (Figure 5A). Because histones display a molecular weight ~18 kDa and they are well-characterized substrates for PAD enzymatic activity (51, 52), we used an antibody to detect citrullinated histone H3 (Cit H3) on the same lysates used in Figure 5A to confirm that Histone H3 was being citrullinated. A clear increase in the citH3 band intensity was observed in the presence of Poly I:C or HRV stimulation (Figure 5B), confirming that citrullination of H3 is increased under these stimuli.

We used confocal microscopy to assess total citrullinated protein levels at different time points post HRV infection (Figure 5C). After 24 and 48 h of infection, a clear increase in F95 staining and protein citrullination was detectable in the presence of HRV (Figures 5C,D). Taken collectively these data suggest that HRV infection, or activation of TLR3, results in increased protein citrullination, including histone H3.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.