ToF-SIMS and TIRF microscopy investigation on the effects of HEMA copolymer surface chemistry on spatial localization, surface intensity, and release of fluorescently labeled keratinocyte growth factor

Abstract

The need for direct biomaterial-based delivery of growth factors to wound surfaces to aid in wound healing emphasizes the importance of interfacial interactions between the biomaterial and the wound surface. These interactions include the spatial localization of growth factor, the surface intensity of growth factor in contact with the wound, and the release profile of growth factor to the wound surface. The authors report the use of time-of-flight secondary ion mass spectrometry to determine the relationship between biomaterial surface chemistry and the spatial localization of growth factor. They have implemented a novel application of total internal reflectance fluorescence (TIRF) microscopy to measure the surface intensity and release of growth factor in contact with a glass substrate that has been used to model a wound surface. Detailed information regarding TIRF experiments has been included to aid in future studies regarding the biomaterial delivery to interfaces. The authors have evaluated the effects of (hydroxyethyl)methacrylate (HEMA) homopolymer, 5.89% methyl methacrylate/HEMA, and 5.89% methacrylic acid/HEMA surface chemistry on the spatial localization of AlexaFluor 488-labeled keratinocyte growth factor (AF488-KGF), AF488-KGF surface intensity at the copolymer surface, and release to a glass substrate. KGF is known to promote re-epithelialization in wound healing. The results show that the two copolymers allow for increased surface coverage, surface intensity, and release of AF488-KGF in comparison to the homopolymer. It is likely that differences in these three aspects could have a profound effect on the wound healing response.

I. INTRODUCTION

Biomaterial development for the delivery of growth factors to wounds to aid in wound healing involves direct and long-term contact between the biomaterial and the wound surface.¹ Interfacial interactions of importance for wound healing efficacy can include the spatial distribution of growth factors at the biomaterial surface, their surface intensity, and the release profile to the wound over time. The evaluation and improvement of biomaterials for wound healing must, therefore, focus on the relationship between biomaterial surface chemistry and these three aspects.

The relationship between surface chemistry and the spatial distribution of growth factors can be determined using time-of-flight secondary ion mass spectrometry (ToF-SIMS) under ultrahigh vacuum (UHV) conditions. In contrast, the surface intensity and release profiles of these biomaterials must be evaluated under hydrated conditions at the body temperature (37 °C). However, it is important to determine the surface intensity and release profile from a biomaterial when it is in contact with the wound. Previous studies have determined surface intensity and concentration with methods such as x-ray photoelectron spectroscopy under UHV conditions.^2–4 Release profiles have been determined with bulk fluorescence release assays measuring released growth factors from a biomaterial into solution.^5,6 In contrast, we have specifically modeled the interaction between the biomaterial and the wound by measuring the surface intensity of growth factor at the biomaterial surface in contact with a glass substrate and release of growth factor onto a glass substrate using total internal reflectance fluorescence (TIRF) microscopy. The wound surface has been modeled by the glass substrate. While a glass substrate differs highly in viscoelastic properties from healing wound tissue, it does provide a direct method of measuring the delivery of growth factors to an interface. This is a novel use of TIRF microscopy for the study of biomaterial-wound interactions. The use of a fluorescently labeled growth factor for both ToF-SIMS and TIRF studies has allowed for the use of identical samples for each study. All samples for analysis were prepared by allowing for copolymers to take up a fixed concentration of growth factor from a solution. The release of growth factor from the hydrogel copolymers onto the glass substrate under hydrated conditions at 37 °C was used to model the release of growth factor to a wound interface. Given the novelty of this ToF-SIMS/TIRF microscopy combination study, detailed information regarding each experiment type has been included below.

(Hydroxyethyl)methacrylate (HEMA) based hydrogel copolymers are of interest in the delivery of growth factors given their ability to allow diffusion of biomolecules into their porous network at room temperature and release of these biomolecules at body temperature.⁷ HEMA hydrogels also have viscoelastic properties representative of tissues.⁷ HEMA hydrogel copolymers (herein referred to as copolymers) are of interest due to improvements in physical and mechanical properties such as hydrophobicity, Young's modulus, and glass transition temperatures in comparison with HEMA homopolymers.⁸ However, the surface chemistry of these copolymers is often drastically different from the homopolymer due to the enrichment of lower surface free energy polymer components.⁹ In the case of copolymers developed for the delivery of growth factors, we hypothesize that changes in surface chemistry due to copolymer formulation can influence the spatial localization, surface intensity, and release profiles of growth factors.

In this study, we are specifically interested in the evaluation of HEMA copolymers for delivery of keratinocyte growth factor (KGF or Fibroblast Growth Factor 7, FGF7). Prior work from our group and other groups has shown that biomaterial-based delivery of KGF can assist in re-epithelialization in in vitro and in vivo wound closure assays.^10–12 These copolymers are being developed for eventual use in direct delivery of KGF to wounds to assist with re-epithelialization and expedite wound closure.

We have previously reported that HEMA based copolymers composed of 5.89% methyl methacrylate (MMA) in HEMA (MMA/HEMA) and 5.89% methacrylic acid (MAA) in HEMA (MAA/HEMA), and the pure HEMA homopolymer have distinguishable surface chemistry and present surface-bound KGF in different orientations.¹³ These differences in orientation may potentially influence the bioactivity of KGF in in vivo assays. MMA/HEMA, MAA/HEMA, and HEMA were investigated in order to distinguish the effects of adding a hydrophobic polymer component (MMA) from the effects of adding a hydrophilic polymer component (MAA) to HEMA.¹³ Our work in this study has continued to use these three copolymer types.

A. ToF-SIMS analysis to evaluate the effects of copolymer surface chemistry on AF488-KGF spatial localization

In this study, we have utilized ToF-SIMS imaging to evaluate the effects of surface chemistry of these copolymers on the spatial localization of AlexaFluor 488 labeled KGF (AF488-KGF). ToF-SIMS uses a primary ion source to bombard a surface under UHV conditions, which produces secondary ions representative of the first few nanometers of the surface.¹⁴ The source can be rastered over a region to create ion images. In this study, high mass resolution ToF-SIMS imaging using a Bi₃⁺⁺ liquid metal ion gun as a primary source has been performed in order to study the copolymers and their interactions with KGF under the same conditions used in our previously published work.¹³

Imaging of proteins at surfaces was accomplished by mapping the CN⁻ ion in a negative mode which is representative of the amide linkage found in proteins. Further confirmation of protein signatures was also accomplished by mapping the CNO⁻ ion. Unique negative ions representative of MMA and HEMA were used for mapping the distribution of MMA across the HEMA surface in negative ion ToF-SIMS images. Alternatively, unique ions of MAA were better represented in a positive ion mode which was used for mapping the distribution of MAA across the HEMA surface. The positive ion mode is ideal for the identification of specific amino acids as opposed to negative ion mode which allows for the identification of general signatures of proteins.¹³ However, the use of specific amino acids may give skewed results regarding protein localization given that our prior work has shown that certain amino acids are found in higher relative intensities at the different copolymer surfaces due to differential interactions between KGF and the surface.¹³ For the purposes of this study, we aimed to obtain ToF-SIMS images that reflected the general signature of AF488-KGF distribution which could be directly compared to TIRF images. In order to obtain a general signature of AF488-KGF in positive ion mode, mapping of AF488-KGF at the MAA/HEMA surface was carried out by summing ion images of detected amino acids known to be present in the structure of KGF.¹⁵ The same general signatures of AF488-KGF at the MAA/HEMA surface were observed by mapping the CN- ion in negative ion mode. The relationship between the distribution of MMA and MAA at the surface was then related to the distribution of AF488-KGF. Using this information, the influence of copolymer chemistry on spatial localization of AF488-KGF taken up from solution was determined.

B. TIRF measurements of surface intensity and release of AF488-KGF from the copolymers

While we chose to evaluate copolymer surface chemistry and AF488-KGF distribution under UHV conditions using ToF-SIMS, these copolymers are intended to deliver KGF, while hydrated, to wound interfaces at body temperature (37 °C). To study this system, measurements of surface intensity and release of AF488-KGF were done using TIRF microscopy of hydrated copolymers at 37 °C. TIRF spectroscopy has previously been extensively used to measure the kinetics of adsorption and desorption of proteins from solution onto the surfaces of glass substrates.^16–19 However, to the best of our knowledge of current literature, TIRF microscopy has had limited use in the study of the spatial distribution of proteins and subsequent release from biomaterial surfaces until now. TIRF microscopy produces images representative of the first ∼100 nm of a surface through utilization of a laser objective that images the sample of interest in contact with a glass substrate.²⁰ We believe that this TIRF microscopy setup is a valuable tool with broad applicability in biomaterial systems intended for delivery of growth factors to wounds or other interfaces.

Comparisons between TIRF and ToF-SIMS images were done to confirm that the representations produced by both imaging modalities were similar. Fluorescence intensity of AF488-KGF at the copolymer surface in contact with glass slides was used to determine the surface fluorescence intensity of AF488-KGF. The surface intensity of AF488-KGF for each copolymer and homopolymer was then compared.

Release experiments were designed to measure the intensity of AF488-KGF released over 10 min time intervals onto the glass substrate, which models the wound surface. This experimental setup aimed to accurately model the amount of KGF directly released to wound tissue in contact with the copolymer, and not take into account the total amount of KGF contained within the copolymer, or account for KGF leaked to the surroundings or loosely bound to the copolymer surface or the glass substrate. Released KGF was considered to be KGF adsorbed onto the glass substrate that remained after rinsing to remove loosely bound protein. Our measurements did not attempt to quantify or make conclusions regarding the total amount of AF488-KGF that was being released from the copolymers.

Fluorescence intensity of released AF488-KGF onto fresh glass substrates every 10 min for a 1 h time course was used to create release profiles for each copolymer. Copolymers were kept hydrated throughout the measurements by keeping phosphate buffered saline (PBS) on the glass substrate. The process of transferring copolymers to new fresh glass substrates for each interval ensured that only AF488-KGF released onto the glass substrates during the 10 min interval was measured. For example, AF488-KGF may have been released into PBS on the glass substrates or may have been loosely bound to the copolymer at the time of transfer to the next glass substrate. Furthermore, the inclusion of protein released to the PBS in the TIRF image resulted in saturation of the TIRF signal and prevented accurate intensity measurements. Cumulative released AF488-KGF over an hour time course was also compared for each copolymer and homopolymer.

We report the effects of HEMA copolymer formulation on the distribution and spatial localization of AF488-KGF at each surface using ToF-SIMS imaging, and on surface intensity and amount of released AF488-KGF to a glass substrate using TIRF microscopy. The methods used in this study are applicable to any system involving biomaterial-based protein delivery to a surface such as a wound.

II. EXPERIMENT

2-(Hydroxyethyl)methacrylate (HEMA, contains ≤50 ppm monomethyl ether hydroquinone as an inhibitor, SKU 477028), MAA (contains 250 ppm monomethyl ether hydroquinone as an inhibitor, SKU 155721), MMA (≤30 ppm monomethyl ether hydroquinone as an inhibitor, SKU 55909), trimethylolpropane trimethacrylate (TMPTMA, contains 600 ppm monomethyl ether hydroquinone as an inhibitor, SKU 246808), 2,2′-azobis(2-methylpropionitrile) (AIBN, SKU 441090), benzoin methyl ether (BME, 96%, SKU B8703), PBS (pH 7.4, liquid, sterile filtered and suitable for cell culture, SKU 806552), and glycerol (SKU G9012) were purchased from Sigma-Aldrich. Keratinocyte growth factor was purchased from Prospec (CYT-219), and the Protein Microscale Labeling Kit was purchased from Invitrogen (#A30006). #1.5 glass coverslips were purchased from Fischerbrand (12-544-D).

A. Hydrogel copolymer preparation

0.5% crosslinked HEMA hydrogels were prepared by using HEMA, 0.5 vol. % TMPTMA, and 0.2% BME dissolved in glycerol as previously described.¹³ The components were mixed, degassed, and injected in between silanized glass slides that were separated by a 1.1 mm thick Teflon spacer and polymerized under ultraviolet light for 30 min. MMA/HEMA and MAA/HEMA hydrogels were prepared using 5.89 mol. % of MMA and 5.89 mol. % MAA, respectively, and 0.5 vol. % TMPTMA, 94.1 mol. % HEMA, and 0.128 mol. % AIBN. The components were mixed and injected in between silanized glass slides that were separated by a 1.1 mm thick Teflon spacer and polymerized for 12 h at 50 °C and then for 24 h at 70 °C. All hydrogels were removed from the glass slides and washed three times at 70 °C in triply distilled water.

B. Preparation of AlexaFluor 488-KGF and of samples for ToF-SIMS studies and TIRF studies

KGF was labeled with the AlexaFluor 488 tetrafluorophenyl ester and purified using the Protein Microscale Labeling Kit. The concentration of purified AlexaFluor 488-KGF (AF488-KGF) was determined using a NanoDrop by measuring absorbance at A280 and A494 according to calculations described in the amine-reactive probes catalog provided by Invitrogen. The concentration of AF488-KGF was determined to be 2 μM. Copolymer samples were prepared by incubating dry 1 × 1 cm² copolymers into 1 ml solutions of 1.5 nM AF488-KGF prepared from 0.1× PBS for 24 h. After 24 h, all copolymer samples were dried in an oven at 50 °C for two hours and then dried in a vacuum box overnight prior to ToF-SIMS and TIRF analysis. Samples were kept in the dark at all times.

C. ToF-SIMS analysis

Negative and positive secondary ion spectra and images were acquired on an ION-TOF V instrument (IONTOF, GmbH, Munster, Germany) using a 25 keV Bi₃⁺⁺ primary ion source kept under static conditions (primary ion dose <10¹² ions/cm²). The current measured to be 0.3 pA and 45 scans were taken over 5 min for each image acquired. A ∼20 keV pulsed flood gun was used for charge compensation. All analysis was done using the IONTOF setting “high current bunched mode” which is a high mass resolution mode. Copolymer samples doped with AF488-KGF images were collected over a 250 × 250 μm² region (256 × 256 pixels). The ion beam was moved to a new spot on the sample after acquiring each set of spectra and images. Spectra were acquired over a range of 0–847 m/z, and mass resolution (m/Δm) at an m/z of 27 was between 5500 and 7000 for all samples collected. Negative ion spectra were mass calibrated using CH⁻, OH⁻, and C₂H⁻. Positive ion spectra were mass calibrated using CH₃⁺, C₂H₃⁺, and C₃H₅⁺. Mass calibration errors were kept under 20 ppm.

All ion images shown in this study were first reconstructed using the Ion-ToF surfacelab software. Images were then exported as a BIF6 file and imported into the NESAC/BIO NBToolbox imagegui program written for MATLAB 2011b (MathWorks, Inc., Natick, MA). All images were normalized by total ion intensity per pixel.

D. TIRF imaging and sample preparation

Imaging was performed on an Eclipse Ti inverted microscope (Nikon) equipped with a motorized laser TIRF illumination unit, a 60× Apochromat TIRF 1.49 NA objective (Nikon), an iXon Ultra EMCCD camera, and a laser merge module (LMM5; 15 Spectral Applied Research). Samples containing AF488-KGF were excited with a 488-nm diode laser and imaging of the sample's emission was done using a Chroma ET525/50 m band-pass emission filter. The AlexaFluor-488 label has a fluorescence excitation maximum at 490 nm and an emission maximum at 525 nm.²¹ The microscope and associated hardware were controlled with MicroManager which can be downloaded at https://micro-manager.org/. Imaging for each sample was centered on the middle of the region covered by the copolymer piece. MicroManager was then used to collect nine images in a 3 × 3 grid surrounding this central point for each sample type. Microscope settings for each channel were kept consistent across all samples to allow comparison of released protein intensities and surface intensities. #1.5 glass coverslips were used as glass substrates. All #1.5 glass coverslips, herein referred to as glass slides, used in microscopy experiments were sonicated for 10 min in 95% ethanol, washed three times with triply distilled water, and sonicated an additional 10 min in triply distilled water before being dried under a stream of nitrogen. Cleaned glass was stored in a covered polystyrene Petri dish for further use. Copolymers were hydrated in 1 ml of 0.22 μm-filtered PBS for 45 min at room temperature and cut into 3 mm² pieces for analysis. Copolymer pieces were then placed on the surfaces of recently cleaned glass slides. The region of the glass covered by the copolymer piece was indicated with a permanent marker to ensure that the correct region was imaged later.

1. Measurement of surface intensity of AF488-KGF at each copolymer surface

TIRF images of the hydrated copolymer surfaces in contact with glass slides were captured using an Eclipse Ti inverted microscope (Nikon) equipped with an Okolab cage incubator set to 37 °C. Nine images were taken for each sample.

2. Measurement of AF488-KGF release from each copolymer to a glass slide

For release experiments, hydrated copolymers that were placed on glass slides were placed in a large Petri dish in a 37 °C oven for 10 min intervals. 100 μl of PBS was also placed on the glass slide containing the copolymers in order to keep them hydrated. To limit desiccation of the copolymers, a damp paper towel was placed next to the glass slides inside the Petri dish. However, noticeable desiccation of the copolymers was observed beyond an hour, limiting release experiments to one hour.

After each 10 min interval, the copolymer pieces were removed from the glass slides and transferred to new clean glass slides and returned to the oven. After transferring the copolymers, the old glass slides were immediately rinsed with 1 ml of PBS and dried under a stream of nitrogen to remove any loosely bound protein from the sample. This process was carried out to prevent any protein that had released into the 100 μl of PBS contained on the glass slide from adsorbing onto the glass slide after the 10 min interval. If this step was skipped, the additional amount of protein adsorbed onto the glass due to protein in the PBS was much larger than the amount of protein released directly onto the glass slide from the copolymer. Therefore, this step was crucial to ensure that only protein released from the copolymer directly onto the glass slide was measured. Samples were stored in the dark at 4 °C until they were all imaged with TIRF microscopy at the conclusion of the experiment and images were captured using an Eclipse Ti inverted microscope (Nikon) equipped with an Okolab cage incubator set to 37 °C. Nine images were taken for each sample.

3. Preparation of surface intensity and release plots

a. Calculation of fluorescence intensity for TIRF images

For each plot, image analysis and plotting was carried out using an in-house python script. Briefly, the mean pixel intensity value for each image was taken. Each mean value was background corrected by (1) subtracting off the mean pixel intensity of nine images taken of only freshly cleaned glass (no copolymer) and (2) subtracting off the mean pixel intensity of nine images taken of the corresponding copolymer containing no protein. Next, the mean and standard deviation of all images for each copolymer type were computed and plotted. All one-way ANOVA were performed using the f_oneway function of version 0.18.1 Scipy.Stats python package.

b. Preparation of surface intensity and release plots

Protein surface intensity plots were computed using TIRF images taken of the copolymer surface mentioned above. ANOVA was performed pairwise between each copolymer type. Protein release plots were computed using TIRF images taken of the glass slides containing delivered protein mentioned above. To produce release plots for the one hour time course, fluorescence intensity was plotted as a function of time. To produce the cumulative release plots, the mean fluorescence intensity of all previous time points was summed and added to the current time point's mean intensity. The cumulative standard deviation for each time point was computed by taking the square root of the sum of squared standard deviations of the current and all previous time points. One-way ANOVA was performed pairwise between each copolymer type for every time point.

III. RESULTS

A. Effect of HEMA copolymer surface chemistry on AF488-KGF spatial localization

Figure 1 maps the distribution using high mass resolution images of each added copolymer component at each copolymer surface (column 1), the distribution of HEMA at each copolymer surface (column 2), and the distribution of AF488-KGF (column 3) at each surface. The literature derived ions CH₃O⁻ and C₂H₃O₂⁻ were used to map out MMA and HEMA, respectively.²² The literature derived ions C₆H₉O₂⁻, and C₇H₉O₂⁻ were used to map out MAA.²³

Fig. 1.

High mass resolution negative ion images of each added copolymer component (column 1), HEMA at each copolymer surface (column 2), and AF488-KGF distribution (column 3) for (a) MMA/HEMA, (b) HEMA, and (c) MAA/HEMA.

For the MMA/HEMA copolymer, Fig. 1(a) shows that MMA is distributed uniformly throughout the surface with weak streaking (column 1) and HEMA has higher intensity regions streaked at the surface (column 2). The pattern of AF488-KGF avoids these high intensity HEMA and MMA regions and preferentially adsorbs to lower intensity HEMA and MMA regions (column 3). The reason for this pattern is unclear. For the HEMA homopolymer, Fig. 1(b) shows that AF488-KGF is found in aggregates distributed across the HEMA surface (column 3). For the MAA/HEMA copolymer, Fig. 1(c) shows that AF488-KGF is found in large domains or aggregates about 40–50 μm in size (column 3). However, in contrast to the MMA/HEMA copolymer, the distributions of literature derived ions representative of MAA do little to explain the pattern of domains or aggregates of AF488-KGF at the MAA/HEMA surface (column 1). Furthermore, while manual analysis of all ion images in the 0–200 m/z range for all three copolymers confirmed that the CH₃O⁻ distribution was characteristic for MMA/HEMA, manual analysis was unable to confirm that the C₆H₉O₂⁻, and C₇H₉O₂⁻ distributions were characteristic for MAA/HEMA (full reconstructions not shown). This is likely because both HEMA and MAA have been reported to produce carboxylate-containing ions in negative mode SIMS spectra, and C₆H₉O₂⁻ and C₇H₉O₂⁻ contain carboxylate moieties.²³

In order to further investigate whether the distribution of MAA at the HEMA surface could explain the distribution of AF488-KGF, mapping of MAA, HEMA, and AF488-KGF was carried out in the positive ion mode and is shown in Fig. 2. The literature derived positive ions C₄H₉⁺, C₄H₁₀⁺, and C₄H₁₁⁺ were used for mapping of MAA as shown in Figs. 2(a)–2(c), and the literature derived positive ion C₂H₅O⁺ was used for mapping of HEMA as shown in Fig. 2(d).¹³ Our previous work on principal component analysis of the three copolymer surfaces in the positive ion mode showed that the C₄H₉⁺, C₄H₁₀⁺, and C₄H₁₁⁺ ions were responsible for separating MAA/HEMA from MMA/HEMA at the 95% confidence level.¹³ Manual analysis of all ion images in the 0–200 m/z range for all three copolymers confirmed that the distinct C₄H₉⁺, C₄H₁₀⁺, and C₄H₁₁⁺ distributions were characteristic for MAA/HEMA. The C₄H_n⁺ ions likely arise from fragmentation of the polymer backbone after loss of the carboxylic acid. The distribution of AF488-KGF at the MAA/HEMA surface is mapped in Fig. 2(e) by summing the images of nine amino acids listed in Table 1 in the supplementary material.²⁷ These amino acids have been previously used in our studies of KGF at these copolymer surfaces, as well as other ToF-SIMS studies of proteins, and have m/z ratios known to not overlap with ions arising from the copolymer surfaces.^13,24 Figure 2(e) indicates a large domain of AF488-KGF similar to the AF488-KGF domain seen in Fig. 1(c) (column 3). The location where AF488-KGF is found in the highest intensity overlaps with the high intensity region of MAA shown in Fig. 2(a)–2(c). Additionally, the relative intensity of C₂H₅O⁺ in Fig. 2(d) gradually decreases in the region where the C₄H_n⁺ ions have high intensity, but does not disappear completely. These data suggest that incomplete phase segregation of MAA and HEMA may cause the formation of a domain about 40–50 μm in size where AF488-KGF preferentially adsorbs. Preferential protein adsorption to partially phase segregated areas over pure domains has been previously observed in x-ray photoemission electron microscopy studies of human serum albumin adsorption at polyethylene oxide-polystyrene blend surfaces.²⁵ This effect is because partially phase segregated areas have the lowest free energy in comparison to pure domains.²⁵ Figure 2(e) also suggests that the intensity of AF488-KGF decreases in regions where the sample is mostly composed of HEMA. The negative ion images of MAA/HEMA in Fig. 1(c) support the possibility of partial phase segregation because the C₂H₃O₂⁻ ion (column 2) shows regions of attenuated ion intensity as well.

Fig. 2.

High mass resolution positive ion images (normalized by total ion intensity) of the MAA/HEMA surface. (a) C₄H₉⁺ ion image, (b) C₄H₁₀⁺ ion image, (c) C₄H₁₁⁺ ion image, (d) C₂H₅O⁺ ion image, and (e) overlay of nine amino acids representing AF488-KGF distribution.

By comparing the images shown in Fig. 1 that have been normalized by total ion intensity, we observe that AF488-KGF has more uniform surface coverage at the MMA/HEMA surface than at the HEMA surface. There are also regions of higher relative intensity of AF488-KGF at the MMA/HEMA surface in comparison to the HEMA surface. AF488-KGF is localized to larger domains at the MAA/HEMA surface. At the MAA/HEMA surface, the relative intensity of AF488-KGF is similar to that of HEMA with the exception of the region to the upper right of the image which has a high relative intensity of AF488-KGF.

To confirm that the distributions of CN⁻ are in fact due to AF488-KGF, ion images of CNO⁻ at each copolymer surface are presented in Fig. 1 in the supplementary material (column 1).²⁷ To additionally ensure that these signatures are not arising from common hydrocarbon-containing contaminants such as dust, ion images of C₂H⁻ and C₂H₂⁻ are presented in Fig. 1 in the supplementary material (columns 2 and 3) for all copolymers.²⁷ C₂H₂⁻ had a negligible contribution at the MMA/HEMA surface and was not plotted. In all cases, the images of the hydrocarbon ions looked significantly different from the images of CN⁻ and CNO⁻ and confirm that CN⁻ and CNO⁻ represent the distribution of AF488-KGF.

B. Effect of copolymer choice on AF488-KGF surface intensity

Figure 3 shows high mass resolution ToF-SIMS images of CN⁻ representing AF488-KGF at each copolymer surface (column 1) next to TIRF images of AF488-KGF at each copolymer surface in hydrated samples in contact with a glass slide at 37 °C (column 2). The ToF-SIMS and TIRF images show similar representations of AF488-KGF distribution at the surfaces. For example, the high surface coverage “streaking” patterns of AF488-KGF at the MMA/HEMA surface observed in Fig. 1(a) are also observed in the TIRF image of the surface [Fig. 3(a)]. Similarly, small localized aggregates of AF488-KGF at the HEMA surface [Fig. 3(b)] and large domains of AF488-KGF at the MAA/HEMA surface [Fig. 3(c)] are observed in the TIRF images of the surface. To ensure that the patterns of protein observed in the TIRF images did not arise from autofluorescence, copolymers without protein were imaged under the same conditions as the copolymers with protein. Images of the copolymers with protein next to images of the blank copolymers scaled identically are provided in Supplementary material Fig. 2 (columns 1 and 2)²⁷ and indicate that the contribution from autofluorescence is negligible. Column 3 of Fig. 2 (Ref. 27) in the supplementary material has been rescaled to show the very weak autofluorescence signal that is detected from copolymer surface.

Fig. 3.

High mass resolution ion images of CN⁻ representing the distribution of AF488-KGF (column 1) and TIRF images (column 2) of AF488-KGF at the surface of MMA/HEMA (a), HEMA (b), and MAA/HEMA (c). Scale bar in TIRF images is 20 μm.

In addition to showing the patterns of AF488-KGF distribution observed in the ToF-SIMS images, the TIRF images also show that all of the copolymers contain aggregates of protein at the surface. However, these aggregates are only seen at the HEMA surface in the ToF-SIMS images. This is potentially because AF488-KGF aggregates at the HEMA surface are found within the first few nanometers of the surface that are penetrated by the primary ion beam in ToF-SIMS, while MMA/HEMA and MAA/HEMA contain aggregates farther from the surface and can only be visualized by TIRF microscopy which penetrates ∼100 nanometers into the surface.

A plausible explanation for the difference in the location of the aggregates is differences in the 3D structure of each copolymer. Given that all copolymers are synthesized using a TMPTMA crosslinking process known to cause pore formation,²⁶ the observed aggregates may be protein contained in pores. Consequently, these pores may be present at different depths in the copolymers due to the 3D structure and different rates of crosslinking. In-depth 3D structural characterization of the copolymers and protein distribution to investigate this possibility is the subject of future efforts.

Figure 4 shows the effect of copolymer composition on the surface fluorescence intensity of AF488-KGF measured from the TIRF images. Figure 4 shows that the HEMA/MMA and HEMA/MAA copolymers contain a higher surface intensity of AF488-KGF in comparison to HEMA as indicated by one-way ANOVA at the 0.05 level (f = 42.38, p < 0.05, f = 20.89, p < 0.05, respectively). Values of surface fluorescence intensity in Fig. 4 have been corrected for autofluorescence by subtracting fluorescence intensities of each blank copolymer from the copolymer containing protein. Combined interpretation of these AF488-KGF surface intensity results with Fig. 4 suggests that the distribution of MMA and MAA at the MMA/HEMA and MAA/HEMA surface recruits more AF488-KGF to the surface in comparison to AF488-KGF at the HEMA surface.

Fig. 4.

Comparison of fluorescence intensity of AF488-KGF for each copolymer surface. HEMA is indicated by HEMA, MAA/HEMA is indicated by MAA, and MMA/HEMA is indicated by MMA.

C. Effect of copolymer choice on AF488-KGF release

Profiles of fluorescence intensity of AF488-KGF released from each copolymer onto fresh glass slides every 10 min imaged by TIRF microscopy are shown in Fig. 5(a). The release profiles indicate that the AF488-KGF released from the MMA/HEMA and MAA/HEMA copolymer onto the glass slides is higher in comparison to HEMA released onto the glass slides for the first 10 min of release as indicated by one-way ANOVA (f = 9.34, p < 0.05 and f = 0.07, p < 0.05 respectively). The release profiles also indicate that the release of AF488-KGF to the glass slide from MMA/HEMA levels off by 40 min to a constant released amount every 10 min while the release of AF488-KGF to the glass slide from MAA/HEMA and HEMA increases again at 50 and 60 min, respectively.

Fig. 5.

(a) Fluorescence intensity of AF488-KGF released from each copolymer onto fresh glass slides every 10 min and (b) fluorescence intensity of cumulative release of AF488-KGF onto glass slides over an hour. HEMA is indicated by HEMA, MAA/HEMA is indicated by MAA, and MMA/HEMA is indicated by MMA.

Cumulative release profiles are shown in Fig. 5(b) and represent the total amount of AF488-KGF released from the copolymer onto the glass slide at each time point. MMA/HEMA and MAA/HEMA release significantly higher AF488-KGF onto the glass slides in comparison to HEMA over the 60 min time course, as indicated by one-way ANOVA (f = 4.3763, p < 0.05, f = 6.0727, p < 0.05 respectively).

IV. DISCUSSION

The results of this study show that the use of 5.89% MMA/HEMA or 5.89% MAA/HEMA copolymers significantly change the spatial localization, surface intensity, and quantity of AF488-KGF released to a glass substrate in comparison to the HEMA homopolymer. Our results suggest that the uniform distribution of MMA at the MMA/HEMA surface allows the surface to recruit more AF488-KGF in comparison to the HEMA homopolymer due to the increased interaction between the copolymer surface and AF488-KGF. In contrast, the interaction between AF488-KGF and HEMA is limited to AF488-KGF localization into what are likely pores of HEMA. Alternatively, partial phase segregation of MAA from HEMA at the MAA/HEMA surface forms a low free energy domain which also promotes the increased adsorption of AF488-KGF to the surface.

Mapping of the copolymer surfaces using the distinctive ions representative of MMA and MAA shows that the distribution of the added copolymer varies greatly. These differences in surface chemistry result in different patterns of AF488-KGF localization at the surface. AF488-KGF at the MMA/HEMA surface is more uniform in comparison to the MAA/HEMA surface where AF488-KGF localizes to large domains. The surface intensity of AF488-KGF calculated at each surface from fluorescence intensity of TIRF images shows that both the MMA/HEMA and MAA/HEMA copolymers contain significantly higher AF488-KGF at the surface in comparison to the HEMA homopolymer at the surface. A higher quantity of AF488-KGF is also released onto the glass substrate at 37 °C over a 1 h time course for MMA/HEMA and MAA/HEMA in comparison to the HEMA homopolymer.

To summarize, the use of 5.89% MMA/HEMA or MAA/HEMA copolymers increases the surface intensity and release of AF488-KGF, specifically in the first 10 min of release to a glass substrate that models the interfacial interaction between the copolymer and a wound surface. It is thought that immediate treatment of KGF to traumatic wounds can greatly increase re-epithelialization and wound healing.²⁶ Therefore, increased release of KGF from the copolymers to a wound interface during the first 20 mins of treatment has the potential to improve re-epithelialization results in comparison to the HEMA homopolymer. However, the difference in the bioactivity of KGF distributed in large pockets at the MAA/HEMA surface in comparison to the bioactivity of KGF that is more uniformly distributed at the MMA/HEMA surface must still be determined using cell-based assays.

V. CONCLUSIONS

The novel combination of ToF-SIMS imaging and TIRF microscopy has allowed for conclusions about the relationship between surface chemistry of HEMA copolymer-based KGF delivery systems and KGF surface distribution, intensity, and release properties that would have otherwise not been possible. This has been accomplished through the use of fluorescently labeled KGF in both experiment types. The broader implications of this study suggest that small variations in copolymer surface chemistry can greatly influence the surface distribution, intensity, and release of growth factors. Given that a biomaterial delivery system to a wound involves direct and extended contact between the biomaterial and a wound, it is likely that changes in these variables could have a profound effect on wound healing response.

Mapping of the copolymer components and KGF distribution using ToF-SIMS imaging allowed for the understanding of how distributions of MMA and MAA at the HEMA surface direct KGF distribution. TIRF imaging of KGF distribution at the copolymer surfaces allowed us to conclude that KGF distribution at these copolymer systems remains relatively constant under both UHV conditions and hydrated conditions at body temperature and allowed for the calculation of surface fluorescence intensity of KGF at each surface. TIRF microscopy also allowed for the measurement of KGF release to an interface modeling the interaction between the copolymer-based KGF delivery systems and a wound surface. The combination of these techniques has informed us on differences in surface fluorescence distribution, intensity, and release profiles due to surface chemistry that may otherwise have gone undetected in commonly used bulk fluorescence assays measuring the release of KGF into a solution from each copolymer. This work has also potentially informed on the interpretation of future cell-based wound healing assays focused on determining how cells modeling wounded tissue may respond differently to the three HEMA copolymers given that they contain different spatial distributions of KGF and release different amounts of KGF to the copolymer-cell interface. Finally, this work has provided guidelines on performing combination ToF-SIMS/TIRF microscopy studies applicable to other biomaterial-based protein delivery systems intending to release growth factors to wound interfaces through the use of fluorescently labeled proteins.