ECM stabilization of MXene platforms is concentration dependent and impact mechanical behaviour

With the fabrication of the biohybrid substrates optimized, we next sought to assess the impact of MXene on mechanical and physical properties. Young’s modulus and hardness increased proportionally with the % w/w content of MXene, with an enhanced elasticity and stiffness of the biohybrid films (Fig. 2a, b) which overcomes the poor mechanical performance that is typical of biological-derived polymers in their naïve state⁴². Wettability of the films is an important criterion of biomaterials as it impacts surface proteins adsorption⁴³, and consequently, cell adhesion^44,45. Cells adhere preferentially to moderately hydrophilic surfaces, as hydrophilicity enhances protein absorption⁴⁶. With an increasing content of MXene, substrate hydrophilicity increases (Fig. ​(Fig.2c),2c), in line with previous works that attributed the hydrophilicity of MXene to surface terminal groups as hydroxyl, oxygen or fluorine^35,47, We demonstrate that these groups are present in these biohybrid films by Raman spectra (Fig. ​(Fig.1b).1b). Among other parameters that influence material interaction with cells, swellability is another key feature to be considered⁴⁸. With increasing amounts of MXene, a statistically significant decrease in water uptake ratio was observed. (Fig. ​(Fig.2e).2e). Additionally, the stability of the biohybrid films were assessed by subjecting them to a bulk degradation test where the films were incubated at 37°C in PBS. It was found that all the groups retained more than 95% of their mass after 13days (Fig. ​(Fig.2d).2d). However, 75% w/w MXene biohybrid films exhibited a continual drop in mass which became notable from day 4, indicating that the bulk collagen network was insufficient to contain the MXene flakes within the film.

Open in a separate window

Fig. 2

Mechanical and physical analysis.

a Young’s modulus (GPa) of 0%, 30%, 45%, 60% and 75% w/w MXene films; b Hardness (GPa) of 0%, 30%, 45%, 60% and 75% w/w MXene films; c Contact angle (°) for 0%, 30%, 45%, 60% and 75% w/w MXene; d Weight percentage (%) after dissolution test, evaluated in 5 time points (1day, 4days, 7days, 10days, and 13days) of 0%, 30%, 45%, 60% and 75% w/w MXene films; e Swelling ratio (%) evaluated in 4 time points (1h, 4h, 8h, 24h) of 0%, 30%, 45%, 60% and 75% w/w MXene films. Statistical analysis was performed using a one-way ANOVA with Tukey’s multiple comparison test (except for (d), where two-way ANOVA was performed), where a resultant p-value of less than or equal to 0.05 marked with * was considered significant. # and & represent statistical significance (p<0.05) between the indicated group and all other groups using one-way ANOVA with Tukey’s posthoc test. All graphical data represents the mean with error bars defining ±standard deviation.

MXene concentration is proportional to the conductivity of the biohybrid platforms in both hydrated and crosslinked states

With the impact of MXene on the mechanical and physical performances of the biohybrid platforms established, conductivity was assessed within four different conditions. While investigating the dependence of the electrical conductivity on the MXene content, we also assessed whether the cross-linking process could interfere with the electrical conduction, and if the hydration state (i.e. dry or wet) of the biohybrid films played a major role. The impact of MXene on conductivity is clear, and was also proportional to the MXene concentration employed, ranging from 0.1Sm⁻¹ for 30% w/w MXene to over 1.5Sm⁻¹ for 75% w/w MXene (Fig. ​(Fig.3a).3a). No significant differences were detected between the four different environmental condition, suggesting that EDAC-NHS cross-linking of the collagen or the hydration state had a minor role in the conduction of the electrical signal. Incidentally, the conductivity of the biohybrid MXene-collagen platforms matches closely and mildly surpasses the conductivity values of native tissues such as myocardium (0.2~0.8Sm⁻¹)^49–51, skeletal muscle (0.2~0.4Sm⁻¹)^52,53 and grey matter tissue (0.1~0.3Sm⁻¹)^54,55. Upon examination, the specific capacitance at four different scan rates, noted that with increasing MXene concentrations, the capacitance also increased (Fig. ​(Fig.3d),3d), highlighting the influence of MXene concentration on energy storage in this platform. Previous works have document that redox capacitance is a major contributor to MXene capacitance, causing the electrochemical adsorption of ions in the surface and consequently charge-transfer⁵⁶. Here, it is observed that scan rate is inversely proportional to specific capacitance: as the scan rate increases, specific capacitance decreases (Fig. ​(Fig.3c).3c). To understand this, slowing down the scan rate, causes electrolytes to penetrate more thoroughly into the surface area (which is large and rough/textured), making greater contact with the internal surface of the material. As a result, more charge is stored on the surface and the measured capacitance increases, becoming closer to the intrinsic capacitance of the material⁵⁷. Similarly, the impedance of the biohybrid platforms verifies the capacitor-like behaviour of the MXene, with impedance decreasing as frequency increases. Specifically, MXene lowers the impedance of the biohybrid system, resulting in a lower opposition to current flow, and ultimately a higher conductivity (Fig. ​(Fig.3b).3b). Taken together, the conductivity, capacitance and impedance characterisation clearly demonstrate the electroconductive nature of the material, confirming the strong electroconductive ability of the biohybrid platforms.

Open in a separate window

Fig. 3

Electroconductive property.

a Conductivity (S m⁻¹) of 30%, 45%, 60% and 75% w/w MXene films, performed before and after the cross-linking step, in both cases in dry and hydrated conditions; b impedance (Omh) of 30%, 45%, 60% and 75% w/w MXene films; c specific capacitance (F g⁻¹) of 30%, 45%, 60% and 75% w/w MXene films; d Specific capacitance curves (F g⁻¹) of 30%, 45%, 60% and 75% w/w MXene films evaluated at scan rate equal to 10mVs⁻¹. Statistical analysis was performed using a one-way ANOVA with Tukey’s multiple comparison test, where a resultant p-value of less than or equal to 0.05 marked with * was considered significant. # Represents statistical significance (p<0.05) between the indicated group and all other groups using one-way ANOVA with Tukey’s posthoc test. All graphical data represents the mean with error bars defining ±standard deviation.

Biohybrid platforms show high biocompatibility, allowing cell spreading and proliferation

To assess the initial biocompatibility of the substrates, mouse embryonic fibroblasts C3H10 were employed. Biohybrid films with 75% w/w MXene were excluded for the experiments from this point due to their instability in aqueous media at 37°C (Fig. ​(Fig.2d).2d). C3H10 cells were highly viable with the biohybrid platforms with a negligible number of dead cells at day 7 (Fig. ​(Fig.4a).4a). DAPI/F-actin (for nuclei/cytoskeletal skeleton respectively) staining determined that C3H10 cell morphology was not impacted by the concentration of MXene. Nuclei retained a circular appearance, while cells spread flat on the substrates, yielding confluent monolayers with excellent cell-cell interactions (Fig. ​(Fig.4b).4b). Additionally, C3H10 cells cultured in all Mxene biohybrid platforms showed a significant increase in metabolic activity (assayed using AlamarBlue™ reduction) from day 3 to day 7 (Fig. ​(Fig.4c),4c), another parameter that indicates cell viability. Meanwhile, besides there being no statistically significant increase in the metabolic activity of C3H10 cells cultured in collagen from day 3 to day 7, C3H10 cells grown in MXene biohybrid platforms at day 7 show greater viability and higher metabolic activity than ones cultured in the collagen. This outcome has been confirmed by the C3H10 cell count, where the addition of MXene led to increased number of C3H10 cells on the substrates when compared with collagen (0% MXene) (Fig. ​(Fig.4d4d).

Open in a separate window

Fig. 4

Cytocompatibility of C3H10 cells on biohybrid platform.

a Live/Dead (scale bar: 200µm) at day 7; b nuclei/F-actin staining (respectively, DAPI and phalloidin) (scale bar: 100µm) at day 7; c metabolic activity which was assessed via AlamarBlue™ reduction assay at day 3 and day 7; d Quantification of viability (cells mm⁻²) at day 7. Statistical analysis was performed using a one-way ANOVA with Tukey’s multiple comparison test, where a resultant p-value of less than or equal to 0.05 was considered significant.* (<0.05), ** (<0.01), *** (0.005) and ****(<0.001) signify a statistically significant difference between the groups indicated. All graphical data represents the mean with error bars defining ±standard deviation.

MXene lends anti-bacterial and anti-fouling property to biohybrid platforms

Antimicrobial capacity of a biomaterial is a feature of importance, since bacterial infection can lead to implant failure and the development of an anti-fouling biomaterial is significant considering the advent of multidrug bacterial resistance^58,59. The primary causative pathogen of implant related infections is Staphylococcus aureus and here, we found that the biohybrid films displayed inhibited gram positive S.aureus Newman attachment and proliferation after 24h (Fig. ​(Fig.5a).5a). With increasing MXene concentrations, a significantly lower number of bacteria attached to the biohybrid films when compared to collagen films (Fig. ​(Fig.5b).5b). Similarly, the absolute area occupied by bacteria reduced with increasing MXene concentration (Fig. ​(Fig.5c).5c). This outcome is in line with other previous studies^60,61 that attribute the anti-bacterial effect to the anionic surface^[46] or to hydrogen interactions between the oxygenate groups of Ti₃C₂T_x MXene and the bacteria membrane’s lipopolysaccharides^47,62.

Open in a separate window

Fig. 5

Antimicrobial evaluation of MXene/collagen type I biohybrid films.

a Live/Dead™ BacLight™ (scale bar: 200µm) after 24h with live bacteria in green and red bacteria in red; b count of S.aureus Newman strain (both live and dead) attached on the biohybrid platforms; c field of view area of % with S.aureus present after 24h. Statistical analysis was performed using a one-way ANOVA with Tukey’s multiple comparison test, where a resultant p-value of less than or equal to 0.05 was considered significant. * Signifies a statistically significant difference between the groups indicated. All graphical data represents the mean with error bars defining ±standard deviation.

Biohybrid films facilitate cardiomyocyte growth

After having demonstrated the impact of MXene concentration on the mechanical, electrical and cell biocompatibility properties, as well as their antimicrobial capability, we next sought to assess the suitability of these biohybrid films for cardiomyocytes. Cardiomyocytes are responsive to electrical stimuli, and have been reported to proliferate and have improved differentiation when in contact with electroconductive biomaterials^7,10. Therefore we extracted neonatal rat cardiomyocytes and seeded onto biohybrid substrates containing 0%, 30%, 45% and 60% MXene and cultured for 7days (Fig. ​(Fig.6a).6a). In this study, primary neonatal rat cardiomyocytes retained an expression of connexin-43 (Cx43), cardiac troponin T and sarcomeric α-actinin throughout all the groups (Fig. 6b, c). Additionally, on day 7, a significantly higher number of cells were found to be attached to the biohybrid films when compared with collagen films (Fig. ​(Fig.6d),6d), with increased cell spreading correlating with an increased MXene concentrations (Fig. ​(Fig.6e6e).

Open in a separate window

Fig. 6

Neonatal rat cardiomyocytes study.

a Schematic illustrating the setup of the experiment with extracted neonatal rat cardiomyocytes seeded onto biohybrid substrates containing 0%, 30%, 45% and 60% MXene and cultured for 7days; b row of immunofluorescent staining micrographs with nuclei in blue (via DAPI), cardiac troponin T (cTnT) in green and connexin43 (cx43) staining in red (scale bar: 20µm) at day 7; c row of immunofluorescent staining micrographs with nuclei in blue (via DAPI /Sarcomeric α-actinin green and F-actin staining by staining with Rhodamine Phalloidin (scale bar: 10µm)) at day 7; d quantification of viability (cells mm⁻²) at day 7; e cardiomyocytes spreading quantification (µm² cell⁻¹). Statistical analysis was performed using a one-way ANOVA with Tukey’s multiple comparison test, where a resultant p-value of less than or equal to 0.05 was considered significant. *Signifies a statistically significant different between the groups indicated. All graphical data represents the mean with error bars defining ±standard deviation.

Scanning Electron Microscopy (SEM)

Biohybrid platforms were mounted on aluminium stubs with a conductive carbon tape (Ted Pella, USA), and a gold–palladium layer of approximately 5nm was sputter coated on the sample surface. Characterisation with SEM was performed using a Zeiss Ultra Plus (Carl Zeiss) microscope fitted with a Gemini column, SE detector and ESB detector. The typical acceleration voltage used was 1.5kV. Energy dispersive x-ray spectroscopy (EDX) spectra were acquired using the EDAX detector at an acceleration voltage of 15kV, and acquired in the range of 0–10kV.

X-ray diffraction

Films were characterized with XRD, using a Bruker D8 Discovery X-ray Diffractometer in θ/2θ configuration, in the range of 3−75°, at 2° min⁻¹. D-spacing was calculated from the Ti3C2Tx (002) reflection according to the formula for Bragg’s law:

Fabrication of MXene-collagen biohybrid platforms

Since Ti₃C₂T_x MXene presents forces binding that keep flakes together, to reduce the surface tension a gelatin solution (Sigma-Aldrich) was used as surfactant. Gelatin facilitates the formation of a colloidal solution when collagen and MXene are mixed. Gelatin was dissolved at 40°C to obtain a solution 4mgml⁻¹, and sterilized by autoclaving at 120°C. The MXene stock solution etched using HF, at concentration of 40mgml⁻¹ was added to the gelatin solution to obtain a final concentration of 20mgml⁻¹. The solution was homogenised by vortex and sonification, then 3 repeated washings were carried out by centrifuging the mixture at 5000rpm for 45mins at room temperature to centrifuge down MXene flakes, decant the excess of gelation in the supernatant and resuspended to wash again.

Porcine collagen type I was dissolved in 0.05M acetic acid (HAc) (pH≈3) overnight at a T<6°C under rotation at 20rpm. The collagen solution, with a concentration 15mgml⁻¹, was neutralized using approximately 1μl 10M sodium hydroxide (NaOH). Then, the MXene/Gelatin solution at 20mgml⁻¹ concentration was mixed with the collagen solution. MXene-Collagen solutions were made at 30%, 45%, 60% and 75% w/w MXene, maintaining for each group a final solution concentration of 15mgml⁻¹. Finally, the solution was vortexed before being drop casted to ensure equal dispersion of the MXene. 200μl of solutions were drop casted in the glass coverslips with a diameter of 12mm, then they were annealed in the oven at 40°C for 2h.

Collagen crosslinking

To stabilize films EDAC (1-Ethyl-3-(3-dimethylaminopropyl)carbodiimide) and NHS (N-Hydroxysuccinimide), was used to covalently crosslink the collagen structure as described in previously works with minor changes⁶⁷.

Bulk degradation

Dry samples were submerged in 1x Phosphate Buffered Saline (1xPBS) at 37°C with the media refreshed each day. Following incubation, samples were washed twice with distilled water and dried in the oven. The percentage change in weight at each time point was evaluated according to the following equation (Eq. 1):

where W₀ is the dry starting mass and W_t is the dry mass at time t. The percentage of weight loss was evaluated at 5 different time points (1day, 4days, 7days, 10days, and 13days).

Swelling test

Dry samples were submerged in X1PBS and incubated at 37°C and refreshed every day. Following incubation, samples were washed twice with distilled water. The swelling ratio of each sample was calculated using the equation (Eq. 2):

where W_s and W_d are the swollen weight and the dry weight respectively. The swelling ratios were evaluated at 3 different time points (1h, 4h, 8h, 24h).

Contact angle measurement

Water contact angle was measured through a custom-made 3D printed instrument that allows to place a droplet of 10μl of DI-water on a glass slide coated with biomaterial films. The droplet was allowed to settle for 10seconds before taking a photograph using a Dino-Lite digital microscope (Dino-lite AM4113ZT USB Microscope). Images were analysed in ImageJ to determine contact angle.

Conductivity measurement

The electroconductivity measurement was carried out through four point method as previously discussed⁶⁸ using a custom built setup available open source⁶⁹.

Electrochemistry characterization

Samples were prepared by drop casting the MXene/Collagen dispersion onto indium tin oxide (ITO) coated glass. Electrochemical measurements consisting of cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS) were carried out using a VMP-300 potentiostat (Bio-Logic, France). CVs were obtained at scan rates from 10 to 100mVs⁻¹, with potentiostatic EIC measurements carried out at 0V and −0.5V, using a frequency range of 1MHz to 100mHz and sinus amplitude of 10mV.

Raman spectroscopy

Raman spectra were acquired in backscattering configuration using an inVia confocal Renishaw Raman microscope equipped with diode (785nm) lasers excitation. The incident beam was focused by a Leica microscope with a 50x magnification objective and short focus working distance; incident power was kept <2mW to avoid sample damage. Spectra were baseline corrected using commercial software prior to analysis. The mapping was performed by raster scanning in the streamline mode at various x-axis steps.

Nanoindentation

Nanoindentation was performed using Nano Indenter XP (Now, Keysight Technologies), a mechanical microprobe system that automatically evaluates the hardness and the elastic modulus. an increasing load from 0 to 10 mN, with an increasing step of 0.5 µN was applied.

Cytocompatibility

To assess the biocompatibility of the material, direct cytotoxic tests were performed. C3H10 mouse embryonic fibroblasts were cultured in growth media prepared using Dulbecco’s Modified Eagle’s Medium (DMEM) low glucose (Sigma-Aldrich) containing 10% v/v fetal bovine serum (FBS) (Gibco® by Life Technologies) and 2% v/v penicillin streptomycin (Pen-Strep) (Sigma-Aldrich) at 37°C with 5% CO2. Films were sterilized with multiple washings in 70% ethanol and exposed to UV light. Following rinsing with PBS, they were incubated in the growth media for 24h. 20,000 cells were seeded on to the samples and media was replaced once every two days, up to day 7. Cell viability, through live-dead assay, was evaluated at day 3 and 7 with the live and dead assay adopting a solution of 2μlml⁻¹ ethidium monodimer-1 and 0.5μlml⁻¹ calcein and 2μlml⁻¹ in PBS. The films were observed using a scanning confocal microscope. Metabolic activity, through AlamarBlue™ assay, was carried out according to the manufacture’s protocol at day 3 and 7 by adding the AlamarBlue cell viability reagent (Invitrogen by Thermo Fisher Scientific) directly to the samples to obtain a 9:1 ratio of cell culture media to AlamarBlue™. In addition, a negative control was included that consisted solely of DMEM low glucose with 10% FBS and 1% Pen-Strep, and 10x of the reagent. The control will eliminate the effect of background absorbance from the cell culture media. The samples were incubated for 4h at 37°C with 5% CO₂. Approximately 100μl aliquots from each well were transferred into a 96-well plate and placed into a plate reader (BioTek, Synergy HT) to read the corresponding absorbance at 570 and 600nm.

To visualize the morphology of cells in the films, staining was carried out to detect nuclei and F-actin filaments. C3H10 cells were cultured on samples with MXene concentrations of 0%, 30%, 45% and 60%. At day 7 the cells were fixed in 4% paraformaldehyde (PFA) for 10min. Films were permeabilized with 0.5% Triton X-100, which was dissolved in PBS, by incubation at room temperature for 5min. Following incubation, films were washed three times in PBS, with a 5-min incubation between each wash. The working solution was prepared by combining 2μl of phalloidin (1μlml⁻¹), 2μl of 4’,6-Diamidine-2’-phenylindole dihydrochloride (DAPI) (1mgml⁻¹) (Sigma-Aldrich) and 2ml of PBS together. Both the nucleus and F-actin filaments can be shown under fluorescent imaging. The working solution was added on top of the films to completely submerge the scaffold in the solution. Samples were protected from light and incubated for 30min. They were then washed three times in PBS before being mounted on glass slides. Glass slides were covered with microscope cover glasses and edges were sealed with clear nail polish. Fluorescent imaging was carried out with a fluorescent microscope.

Anti-bacterial study

The anti-bacterial properties of the films were assessed using S. aureus Newman (gram positive), a clinical isolate of osteomyelitis, at a concentration of 5×10⁵ CFU ml⁻¹ as recommended by the Clinical and Laboratory Standards Institute. The films were seeded with 5×10⁵ CFU ml⁻¹ of S. aureus Newman in Brain Heart Infusion (BHI) broth and incubated at 37°C for 24h prior to beginning LIVE/DEAD and alamarBlue™ assays. The antibiofilm and contact-kill potential of the MXene films against S. aureus Newman bacteria was evaluated using LIVE/DEAD™ BacLight™ bacterial viability kit for microscopy & quantitative assays (Invitrogen, USA) was used as per the manufacturer’s protocols following 24h of S. aureus culture on the films in BHI broth. After culture, the films were rinsed three times with PBS, fixed using 10% formalin solution for 30min at 4°C, followed by three more rinses with PBS. The films were incubated with the stain for 3min at room temperature and then rinsed again with PBS. The stained scaffolds were imaged using confocal microscopy and analysed using ImageJ to determine the percentage of bacterial cell area on the scaffolds.

Ethics approval

Ethics approval for extraction of neonatal rat cardiomyocytes was granted by the Trinity College Dublin Animal Research Ethics Committee as part of a governing approval for extraction of primary cells from neonates. Ethical approval for the use of commercial iPSCs is not necessary in Trinity College Dublin.

Isolation and handling of neonatal rat cardiomyocytes

Primary CMs were isolated from neonatal rat hearts (nrCM) using a protocol adapted from the Pierce Primary Cardiomyocyte Isolation Kit from Thermo Scientific (88281, Thermo Fisher Scientific). After isolation, cells were seeded immediately on films. Cells were seeded at a density of 1×10⁶ cells per film and were cultured in the Pierce Primary Cardiomyocyte culture media.

Immunofluorescent staining for cardiac markers was performed with an internal protocol at day 7. Briefly, the cell membrane was lysed in 1v/v% Triton X-100 (T8787, Sigma-Aldrich), then samples were blocked with a PBS-based solution consisting of 4w/v% BSA (A2153–50G) for stabilisation, 0.05v/v% Triton X-100 to enhance permeabilisation, 0.05v/v% Tween20 (P1379, Sigma-Aldrich) to reduce surface tension. Cardiac troponin T (cTnT, ab8295 Abcam), connexin 43 (Cx43, ab11370 Abcam) and sarcomeric α-actinin (α-act, EA53 Thermo Fisher) were used as primary antibodies and diluted 1:300 (cTnT and Cx43) or 1:100 (α-act) in antibody dilution buffer containing PBS with 1w/v% BSA, 0.05v/v% Triton X-100, 0.05v/v% Tween20. Secondary antibodies were diluted 1 in 250 in antibody dilution buffer. The secondary antibodies used for this study were Alexa Fluor® 594 for Cx-43 (ab150076, Abcam) and Alexa Fluor® 488 for cTnT and α-act (ab150113, Abcam). Rhodamine Phalloidin (00027, VWR) at concentration of 5µml⁻¹ was incubated together with the secondary antibody for α-act. DAPI solution was made from a 1000X stock concentrate (32670, Sigma-Aldrich). All washings were performed with PBS and 0.05v/v% Tween20. Micrographs were obtained using an OlympusIX83 epifluorescence microscope (Olympus, Germany). Cell spreading was calculated analysing Nuclei—Sarcomeric—α-actinin staining images through ImageJ.

iPSC culture and cardiomyocyte derivatizatio

iPSCs (SFCi55-ZsG—tagged with ZS Green⁷⁰) were expanded in MTESR media (Stem Cell®) grown on Matrigel™ coated plates and cultured to 70–80% confluency before splitting at a ratio of 1:5 using Versene (Gibco®). Cardiomyocyte differentiation was induced using the extensively described GiWi protocol⁷¹, Briefly, 90–95% confluent iPSCs (day 0) were treated with 10µM CHIR99021 in RPMI media supplemented with B27 minus insulin (differentiation media) for 24h. On day three, the cells were treated with 5µM IWP2 for 48h, in differentiation media. On day five, the media containing IWP2 was replaced with un-supplemented differentiation media and finally from day seven, the cells were cultured in RPMI media supplemented with B27 with insulin (maintenance media). Spontaneous beating was noted from day eight of the differentiation and the iPSC-derived cardiomyocytes were lifted and seeded on the substrates on day twenty-one of the differentiation.

Electric field stimulation

Starting from day 3, iPSC-CMs seeded on substrates were stimulated with a custom-made bioreactor⁶³ with a biphasic 2ms long pulse of ±2.5V and a frequency of 2Hz. Cells were stimulated 1h per day until day 7. Immunofluorescent staining for cardiac markers (Cx43 and Sarcomeric α-actinin) was performed at day 7 (mentioned above). Cell spreading, cell count and cx43 intensity were calculated analysing, respectively, Nuclei—Sarcomeric—α-actinin and cx43 staining images through ImageJ.

This work was supported through the Advanced Materials and Bioengineering Research (AMBER) Centre (SFI/12/RC/2278_P2), partly supported by the European Regional Development Fund. The SEM imaging for this project was carried out at the Advanced Microscopy Laboratory (AML), Trinity College Dublin, Ireland. The AML (http://www.tcd.ie/crann/aml) is an SFI supported imaging and analysis centre, part of the CRANN Institute and affiliated to the AMBER centre. The authors also thank Prof Tatiana Perova and Dr Khairul Hoque for assistance with Raman spectroscopy. This work was also partially supported by an SFI-HRB Wellcome Trust-ISSF Award Institutional Strategic Support Fund (Ref No. 204814/Z/16/Z), and the Irish Research Council (Project No. GOIPG/2019/818).