Dual Acidic pH‐Responsive Post‐Crosslinked E‐Spun Nanofibrous Scaffolds Exhibiting Enhanced Disassembly and Release for Localized Cancer Treatment

ABSTRACT

The development of multifunctional electro‐spun nanofibrous mats has been explored as an effective platform for implantable localized cancer treatment because they provide a supportive matrix to regenerate tissues and deliver encapsulated anticancer drugs to suppress the proliferation of cancer cells. However, conventionally designed nanofibrous scaffolds present several critical challenges, particularly delayed and uncontrolled release of drug molecules due to inefficient crosslinks. Herein, we report a robust approach with the synthesis of a new phenyl diboronic acid crosslinker bearing an acid‐cleavable Schiff base linkage to develop dual acidic pH‐responsive degradable e‐spun nanofibrous mats. The dual smart mats crosslinked with both acid‐labile imine and boronic ester linkages, being structurally stable in a physiological pH, rapidly degrade through their acid‐catalyzed hydrolysis in an acidic environment, leading to the enhanced release of encapsulated anticancer drugs. Furthermore, doxorubicin‐loaded mats have desired antitumoral activity and hemocompatibility, while empty ones are biocompatible. These results demonstrate the great potential of dual acidic pH‐responsive degradable e‐spun nanofibrous mats as implantable localized drug delivery scaffolds for cancer chemotherapy.

Dual acidic pH‐responsive degradable e‐spun nanofibrous mats crosslinked with acid‐labile imine and boronic ester linkages that exhibit rapid degradation and enhanced release of encapsulated anticancer drugs through their acid‐catalyzed hydrolysis and their doxorubicin‐loaded mats having desired antitumoral activity and hemocompatibility, as promising implantable localized delivery scaffolds for cancer chemotherapy.

1. Introduction

Because of unique structural and functional properties such as high aspect ratio, large pore size, and high molecular orientation [1, 2, 3], electro‐spun (e‐spun) nanofibers have been extensively explored in various applications [4, 5, 6, 7, 8], particularly including localized cancer treatment [9, 10, 11]. Nanofibrous mats encapsulated with anticancer therapeutics serve as not only implantable scaffolds, providing supportive matrix for tissue regeneration, but also drug depot implants, suppressing the proliferation of tumor tissues following the surgical excision of solid tumors [12, 13]. These implantable scaffolds enable the localized and controlled release of anticancer therapeutics, maintain therapeutic concentrations at the tumor site, and eliminate cell remnants undetected during surgery [14]. Owing to these features, well‐controlled nanofibrous mats are attracted as less toxic alternatives to treat cancers accessible in surface‐level tissues such as cervical and skin cancers, thus circumventing the adverse effects of chemotherapeutic agents that stemms from their non‐selective nature and systemic exposure following intravenous administration.

A variety of synthetic and hydrophobic polymers have been explored to fabricate effective e‐spun fibrous scaffolds [15, 16]. Despite these advances, hydrophilic and naturally occurring polymers, particularly poly(vinyl alcohol) (PVA), have been used for implantable and localized cancer therapy due to their high water‐absorbing capability and good biocompatibility [17, 18, 19]. Numerous strategies have been proposed to develop effective nanofibrous scaffolds encapsulated with anticancer drugs such as doxorubicin (DOX) [20, 21, 22]. Strategies include the incorporation of superparamagnetic nanoparticles for hyperthermia [23] and porphyrin photosensitizer for photodynamic therapy [24] as well as polymeric nanoassemblies to facilitate the encapsulation of DOX [25]. Additionally, the embedding of lignin nanoparticles [26] as well as the fabrication of core/shell [27, 28] and multilayer [29] nanofibers have been proposed to develop e‐spun mats encapsulated with hydrophobic drugs such as paclitaxel and 5‐fluorouracil. Due to the use of permanent crosslinks [30] such as ester [31, 32, 33] and carbamate [34] bonds, however, these nanofibrous scaffolds present critical challenges to be addressed, such as instability in biological environments and premature drug loss or burst release caused by insufficient or absent crosslinking as well as low loading efficiency and uncontrolled release of drugs.

Stimuli‐responsive degradation (SRD) through chemical transition involves the introduction of labile covalent bonds into the design of nanomaterials [35, 36, 37, 38, 39, 40, 41]. When needed, the labile linkages are cleaved in response to triggers such as pH change and light as well as reductive, oxidative, and enzymatic reactions, causing the degradation of nanomaterials, thus leading to the enhanced release of encapsulated drug molecules [42, 43, 44]. Although SRD has been significantly explored for the construction of controlled drug delivery nanoparticles such as block copolymer‐based nanoassemblies and nanogels [45, 46, 47], it has been less studied for the fabrication of hydrophilic and water‐soluble e‐spun nanofibers crosslinked with degradable covalent bonds through a post‐crosslinking method [48, 49, 50]. We recently developed an elegant approach exploring boronic ester chemistry to fabricate e‐spun nanofibrous mats crosslinked with degradable boronic ester bonds [51, 52, 53]. Our approach utilizes the synthesis of novel phenyl diboronic acid crosslinkers for post‐crosslinking e‐spun nanofibrous mats. The mats crosslinked with boronic ester crosslinks degraded in response to acidic and alkali pHs, reactive oxygen species and glucose, stimuli commonly found in chronic wounds, thus leading to the enhanced release of encapsulated antibiotics.

However, the cleavage of boronic ester bond in acidic pH turned to be slow, resulting in delayed degradation of the nanofibrous mats [52]. Toward implantable localized cancer chemotherapy, the mats could require rapid and tunable degradation in tumor environments. Extracellular compartment in tumor is slightly acidic, typically ranging at pH 6.5–6.9 due to high metabolic activity and poor perfusion, while the intracellular compartment of cancer cells, particularly endosomes and lysosomes, is more acidic at pH 4.5–5.5. However, cytoplasm remains physiological pH 7.0–7.2 to optimize survival and proliferation due to the overexpression of membrane proton pumps, including Na⁺/H⁺ exchangers (NHE1), monocarboxylate transporters (MCT1/4) exporting lactate and carbonic anhydrases (CA9) [54, 55, 56]. We envisioned that the development of a novel diboronic acid crosslinker incorporating an acid‐labile linkage enables the fabrication of dual acidic pH‐degradable nanofibrous mats exhibiting rapid degradation, thereby enhancing the release of anticancer drug molecules through acid‐catalyzed hydrolysis of both the acid‐labile linkage and the boronic ester bonds. Schiff base linkage is known to be acid‐labile, cleaved to the corresponding aldehyde and primary amine bonds [57, 58, 59]. Its acid‐catalyzed hydrolysis is rapid and effective in tumoral and endo/lysosomal acidic pH environment.

Herein, we report a robust approach with the synthesis of a new phenyl diboronic acid crosslinker bearing acid‐cleavable Schiff base linkage (DBA‐I) and develop dual acidic pH‐responsive degradable e‐spun nanofibrous mats (BE/I‐PVA). As illustrated in Scheme 1, the dual smart mats crosslinked with boronic ester linkages and labeled with benzoic imine linkages were dimensionally stable in a physiological pH, but rapidly degraded for the enhanced release of encapsulated anticancer drugs through acid‐catalyzed hydrolysis of both mainly Schiff base linkages and boronic ester bonds in an acidic environment. DBA‐I crosslinker was synthesized by a facile coupling reaction of 4‐formyl phenyl boronic acid with ethylene diamine through the formation of in situ Schiff base linkage and turned to be effective to crosslink PVA nanofibers with degradable BE bonds through post‐crosslinking approach. The crosslinked mats were characterized for their structural and mechanical properties. Their acidic pH‐responsive degradation and enhanced release of encapsulated DOX were comprehensively studied. DOX‐loaded mats (BE/I‐PVA/DOX) were further evaluated for their antitumoral activity and hemocompatibility to demonstrate their potential as implantable localized drug delivery scaffolds for cancer chemotherapy.

SCHEME 1.

Our approach with the synthesis of a new phenyl diboronic acid crosslinker bearing acid‐cleavable Schiff base linkage (DBA‐I) to develop dual acidic pH‐degradable e‐spun nanofibrous mats loaded with doxorubicin (DOX) anticancer drug (BE/I‐PVA/DOX), exhibiting the enhanced release profiles of encapsulated DOX molecules, thus demonstrating their potential as implantable localized drug delivery scaffolds for cancer chemotherapy.

2. Results and Discussion

2.1. Synthesis of DBA‐I and DBA‐I‐PE

Figure 1a depicts our approach to synthesize a new DBA‐I crosslinker bearing two acid‐cleavable imine linkages through the in situ formation of imine bonds. The approach involves the reaction of the aldehyde group of 4‐formyl phenyl boronic acid (FPBA) with the primary amine group of ethylenediamine (EDA) in dried MeOH at room temperature. An excess FPBA (2 equivalent to the amine group of EDA) was designed to avoid the formation of undesired mono‐substituted by‐product. The target product, DBA‐I, was characterized by various spectroscopic techniques. ¹H NMR spectrum shows a peak at 8.34 ppm (b) corresponding to the newly formed imine proton, a peak at 3.88 ppm (a) to ethylene spacer protons, and the peaks at 7.64–7.83 ppm (c and d) to aromatic protons (Figure 1b). Their integral ratio was quantitative to the number of their corresponding protons. Solid‐state ¹¹B NMR displays a broad dipolar‐coupled symmetric resonance signal in −10–40 ppm with two distinct peaks at 11.7 and 17.4 ppm, which correspond to trigonal sp²‐hybridized boron atoms (Figure 1c). Furthermore, FT‐IR spectrum shows two important absorption bands at 1642 cm⁻¹ corresponding to imine (C═N) stretching mode and 1330 cm⁻¹ to B─O─C stretching mode in boronic acid group (Figure 1d). These spectroscopic analyses, along with ¹³C NMR (Figure S1) and HRMS (Experimental) results, confirm the synthesis of DBA‐I in 95% yield. For model studies, the pinacol ester derivative of DBA‐I (DBA‐I‐PE) was synthesized by a facile coupling reaction of DBA‐I with pinacol (2.5 mole equivalent excess to the boronic acid group) (Figure 1a). ¹H NMR analysis displays a peak at 1.31 ppm corresponding to methyl protons of pinacol moieties (d′) (Figure 1b). ¹¹B NMR shows a broad quadrupolar‐coupled resonance signal in −10–40 ppm with distinct peaks at 28.3, 20.6, 15.1, 7.1 7.1, and 2.1 ppm, which appear to be different from those for DBA‐I (e.g., a diboronic acid) (Figure 1c). FT‐IR spectrum shows the characteristic absorption band at 1354 cm⁻¹ corresponding to B─O─C stretching mode in boronic ester moiety, along with the band at 1642 cm⁻¹ to imine (C═N) group (Figure 1d). Combined with ¹³C NMR (Figure S2) and HRMS (Experimental), these results confirm the synthesis of DBA‐I‐EP in 92% yield.

FIGURE 1.

Synthesis (a), ¹H NMR spectra (b), solid‐state ¹¹B NMR spectra (c), and FT‐IR spectra (d) of DBA‐I crosslinker bearing two acid‐labile imine bonds and DBA‐I‐PE, a pinacol ester of DBA‐I. Note that residual solvents are denoted by x, y, and z: water (x), DMSO (y), and THF (z).

2.2. Fabrication of Acid‐Degradable BE/DBA‐I E‐Spun PVA Nanofibrous Mats

The standard conditions for electrospinning established in our laboratory [51, 52], were applied to fabricate well‐defined, droplet‐free, spaghetti‐like PVA e‐spun nanofibers. They include aqueous 10 wt.% PVA solution, BD needle 25G 1 ½ (0.15 mm inner diameter), voltage of 25 kV, flow rate of 0.85 mL h⁻¹, and distance nozzle‐collector of 9 cm.

The fabricated e‐spun nanofibrous mats were then subjected to post‐crosslinking with the newly synthesized DBA‐I crosslinker, forming dimensionally stable BE/I‐PVA mats crosslinked with BE bonds. Dried e‐spun nanofibrous mats were immersed in DMF, a poor solvent to PVA, containing excess DBA‐I as a mole equivalent ratio of BA/2OH = 2/1 at room temperature for 48 hrs. The ratio of BA/2OH = 2/1 was chosen based on our previous investigation on how the amount of a diboronic acid crosslinker can influence crosslinking and its efficiency [51]. One BA group of DBA‐I is expected to react with two OH pendant groups in PVA chains, forming a single BE crosslink (Figure 2a). A gravimetric analysis following the change in the mass of the fibers before and after the treatment with DBA‐I was conducted. %DBA‐I incorporation into nanofibers (e.g., the amount of DBA‐I incorporated into the fibers relative to the amount initially included in the recipe) was determined to be 2.1 ± 0.6%, and %OH mol reacted (e.g., the mole equivalent ratio of OH groups that reacted with boronic acid groups) was 16.8 ± 8.5%, meaning that 83% OH groups in PVA chains remained unreacted in the nanofibrous mats. The plausible reason for relatively low extent of conjugation of pendant OH groups could be attributed to their occurrence of reaction with BA groups mostly on the surfaces of nanofibers due to limited access of DBA‐I to inside nanofibers in DMF. Promisingly, the gel content of the mats turned out to be as high as 85%, measured in PBS solution (pH 7.4) for 24 h.

FIGURE 2.

Schematic presentation of BE/I‐PVA mats crosslinked with DBA‐I through the formation of BE bonds (a), SEM image and size distribution histogram of uncrosslinked PVA mats (b) and crosslinked BE/I‐PVA mats (c). For comparison, FT‐IR spectra (d), solid‐state ¹¹B NMR spectrum (e), and stress‐strain curves by universal tensile measurement (f).

Further, SEM analysis in Figure 2b,c indicates that both BE/I‐PVA mats and their uncrosslinked counterparts are well‐defined and intact as spaghetti‐like nanofibrous forms. Their average diameters appeared to be similar as 132 ± 30 nm for BE/I‐PVA mats and 134 ± 28 nm for uncrosslinked ones. FT‐IR spectra in Figure 2d and Figure S3 (full scan) shows the characteristic absorption band at 1304 cm⁻¹ corresponding to B─O─C stretching mode in the BE moiety. Promisingly, the spectrum also shows the band at 1642 cm⁻¹ corresponding to imine stretching vibration mode, indicating that imine bonds remained intact during crosslinking. Solid‐state ¹¹B NMR spectrum in Figure 2e shows the distinctive quadrupolar asymmetric resonance signals at 3.2, 10.4, 17.5, and 24.3 ppm. Further to our spectroscopic analysis, the mechanical properties of BE/I‐PVA mats were measured using a universal tensile tester. Figure 2f compares the tensile strength of 9.1 kPa, Young's modulus of 5.4 kPa, and elongation at break of 4.3% for BE/I‐PVA mats, while the tensile strength of 8.5 kPa, Young's modulus of 5.1 kPa, and elongation at break of 9% for uncrosslinked mats. This result suggests that the mechanical properties of mats are enhanced upon crosslinking through the formation of BE bonds.

Overall, these results suggest that our approach exploring BE chemistry is highly efficient to crosslink PVA nanofibrous mats with a small molecule diboronic acid crosslinker through the formation of BE crosslinks, achieving dimensionally stable e‐spun nanofibrous mats.

In a separate experiment, the needle sizes as its gauge numbers were varied in e‐spinning process. The fabricated e‐spun nanofibers were subjected to crosslinking with DBA‐I under similar conditions. With an increasing needle gauge number from 18 to 25G, the average diameter of the fibers decreased from 280.0 to 131.7 nm (Figure S4) and interestingly, %OH mol reacted increased from 5.5% to 19.7% (Figure S5). This could be attributed to the decrease in diameters, e.g., increase in surface areas of nanofibers. This result confirms the relationship of diameters with crosslinking density of BE/‐I PVA mats, suggesting that crosslinking reaction mainly occurs on the surfaces of nanofibers and their mats.

2.3. Model Studies for Acidic pH‐Responsive Degradation

The fabricated BE/I‐PVA mats were post‐crosslinked with DBA‐I bearing acid‐degradable imine bonds through the formation of BE bonds. Prior to their acidic pH‐responsive degradation studies, small molecule DBA‐I (a boronic acid crosslinker) was first examined for acid‐catalyzed hydrolysis of imine bond using ¹H NMR spectroscopy in DMSO‐d₆. As schematically illustrated in Figure 3a, the imine bond is cleaved to corresponding aldehyde and primary amine groups in acidic pH. ¹H NMR spectra show the gradual increase of the aldehyde peak (b′) at 10.03 ppm, while the gradual decrease of the imine peak (b) at 8.36 ppm, as a consequence of the cleavage of imine bonds (Figure 3b; Figure S6). Using their integral ratios, %hydrolysis of imine bond in DBA‐I was determined at pD = 0.5. As seen in Figure 3c, it rapidly increased to >65% in 15 min and gradually increased to completion (>98%) in 120 min.

FIGURE 3.

Degradation through acid‐catalyzed hydrolysis of acid‐labile linkages (imine and boronic ester bonds) (a,d), overlaid ¹H NMR spectra before and after being exposed to acidic pH (b,e), and %hydrolysis of labile linkages over incubation time (c,f) for DBA‐I (a–c) and DBA‐I‐PE (d–f).

To further compare the acid‐catalyzed hydrolysis of imine with boronic ester bonds, DBA‐I‐PE was examined. As illustrated in Figure 3d, both imine and BE bonds are cleaved to the corresponding aldehyde and primary amine as well as boronic acid and diol groups in acidic pH, respectively. ¹H NMR spectra for the mixture in DMSO‐d₆/THF at pD = 0.3 show the change in characteristic peaks, particularly the peak (b) at 8.28 ppm equivalent imine proton and the peak (b′) at 1.14 ppm corresponding to methyl protons in pinacol ester moieties (Figure 3e; Figure S7). %Hydrolysis calculated from their integral ratios reached >99% completion for imine bond, which can be compared with 53% for BE bond after 120 min (Figure 3f). This result suggests more rapid hydrolysis of the imine bond compared with BE group in acidic pH.

2.4. Acidic pH‐Responsive Degradation of BE/I‐PVA Mats

Given our model studies with DBA‐I and DBA‐I‐PE, we investigated acidic pH‐responsive degradation of BE/I‐PVA mats. Figure 4a illustrates their degradation through acidic pH‐responsive cleavage of both imine linkage and BE crosslinks. Pieces of BA/I‐PVA mats were incubated in McIlvaine buffer solutions at acidic pH 5.4 and 6.4, compared with physiological pH 7.4. Gravimetric analysis was first employed to follow the %degradation of the mats. As compared in Figure 4b, %degradation increased rapidly at both acidic pHs over incubation time, e.g., 80% at pH 5.4 and 62% even at pH 6.5 after 24 h. For BE‐PVA mats crosslinked with a phenyl diboronic acid crosslinker (named as DBA‐E, a structural analogue of DBA‐I without imine bonds), their %degradation was reported to be <6% at pH 5.4 [52]. Hence, the greater %degradation of BE/I‐PVA mats is attributed to the cleavage of mainly imine bonds, along with BE bonds in acidic pHs. At pH 7.4, the %degradation of BE/I‐PVA mats was as low as 28% in 24 h. Considering that %degradation was 15% for BE‐PVA mats at pH 7.4, the %degradation of BE/I‐PVA mats is presumably due to mainly the cleavage of BE bonds.

FIGURE 4.

For BE/I‐PVA mats, degradation through acidic pH‐responsive cleavage of both imine linkage and BE crosslinks, %degradation over incubation time at acidic pHs 5.4, 6.5, and 7.4 (b), digital images (scale bar = 2 cm) and SEM images (scale bar = 10 µm) at pH 5.0 (c) and 7.4 (d), and FT‐IR spectra after the incubation of 120 min at pH 5.0 and 7.4 (e).

Further to our gravimetric analysis, acidic pH‐responsive degradation of mats was investigated with analytical techniques. As illustrated in Figure 4c, when exposed at pH 5.0, the BE/I‐PVA mats appeared lost their opacity and became transparent. They also shrunk and their dimensions decreased by 77.2% at pH 5.0 in 120 min, estimated using Image J software. SEM analysis reveals that the mats lost their spaghetti‐like nanofiber morphologies, forming aggregates of PVA chains. In contrast, the mats incubated at pH 7.4 remained opaque and retained their spaghetti‐like fiber morphologies (Figure 4d). FT‐IR spectrum in Figure 4e shows the complete disappearance of the band at 1642 cm⁻¹ (imine vibrational mode) and the decrease in the band at 1304 cm⁻¹ (B‐O‐C vibration mode in boronic ester bond). Solid‐state ¹¹B NMR in Figure S8 turned out to be similar to that of DBA‐I crosslinker (see Figure 1c), but much lower intensity, which could be attributed to the presence of residual moieties of FPBA as a degraded product generated upon the cleavage of BE bonds.

2.5. Loading and Enhanced Acidic pH‐Responsive Release of DOX

The encapsulation of DOX, a clinically used anticancer drug, in mats were examined. The standard protocol for e‐spinning was employed for aqueous 10% PVA solution containing 0.48 wt.% DOX. The fabricated DOX‐loaded mats, named BE‐PVA/DOX, were then subjected to crosslinking with DBA‐I crosslinker, yielding BE/I‐PVA/DOX mats. The loading efficiency of DOX in mats was determined to be >98% for both uncrosslinked PVA/DOX and crosslinked BE/I‐PVA/DOX mats, suggesting that both our e‐spinning and crosslinking processes are reliable to encapsulate DOX (see Figures S9 and S10 and our detailed analysis in Supporting Information). SEM analysis reveals that the fabricated BE/I‐PVA/DOX (0.48%) had an average diameter of 156 ± 37 nm, which is larger than bare BE/I‐PVA counterparts (Figure 5a). Their tensile strength of 10.9 kPa, Young's modulus of 66.3 kPa, and elongation at break of 3.3%, which can be compared with the tensile strength of 9.1 kPa, Young's modulus of 5.4 kPa, and elongation at break of 4.3% for BE/I‐PVA (counterparts without Dox) (Figure 5b). This result suggests that the mechanical properties of the mats are enhanced with encapsulated Dox.

FIGURE 5.

SEM image of BE/I‐PVA/DOX mats (a), stress‐strain curves of BE/I‐PVA/DOX and BE/I‐PVA by universal tensile measurement (b), %release of DOX from BE/I‐PVA/DOX mats over incubation time in acidic pHs 5.0 and 6,4 and neutral 7.4 (c). Viability of HEK293, HFF‐1, and HeLa cells incubated with BE/I‐PVA nanofibrous mats (d) and HeLa cells with BE/I‐PVA/DOX mats, uncrosslinked PVA/DOX mats, and free DOX, determined by resazurin reduction assay (e). Hemocompatibility of BE/I‐PVA/DOX mats evaluated through quantification of hemoglobin release via UV/Vis spectroscopy at λ = 577 nm (f). Inset in (f): the digital images of supernatants denoted as 1, 2, 3, and 4 corresponding to the negative control, BE/I‐PVA, BE/I‐PVA/DOX (0.48 wt.%), and positive control, respectively.

Give our comprehensive characterization, BE/I‐PVA/DOX mats were evaluated for the release of DOX at pH 5.4 and 6.4, compared with 7.4, using fluorescence spectroscopy. For quantitative analysis, correlation curves with emission intensity of DOX at λ_em = 590 nm over the concentration of DOX in the same media at pH 5.4, 6.4, and 7.4 were constructed (Figure S11). As seen in Figure 5c, %DOX release rapidly increased in both acidic pHs, reaching 45% at pH 5.4 and 43% at pH 6.4 in 6 h, while it gradually increased at pH 7.4, reaching 30% in 6 h. Such enhanced release of DOX in acidic pHs could be attributed mainly to acid‐catalyzed hydrolysis of imine bonds. Furthermore, the release mechanism of DOX from BE/I‐PVA/DOX mats was investigated with Peppas‐Sahlin model which is known to be a kinetic equation used to analyze drug release from polymer delivery systems by separating it into Fickian diffusion and Case II relaxation transfer [60, 61, 62]. As seen in Figure S12, all release profiles at three pHs are well‐fitted to the model (adjusted R² values >0.95). K₂ kinetic constant for Case II relaxational contribution is negative at all three pHs, suggesting that the release of DOX is more likely to undergo through Fickian diffusion. Diffusion coefficient (m) values are <0.45–0.5 at acidic pHs, suggesting that DOX release could obey Fickian diffusion, while m value is 0.69 at pH 7.4, suggesting that DOX release could follow non‐Fickian diffusion.

2.6. Antitumoral Activity and Hemocompatibility

The fabricated BE/I‐PVA/DOX mats were evaluated for potential application in implantable cancer treatment. To get an insight into the biocompatibility, the BE/I‐PVA mats, their counterpart mats with no DOX, were first assessed for their cytotoxicity with three cell lines, including human foreskin fibroblast HFF‐1 and embryonic kidney normal HEK293 normal cells as well as HeLa cancer cells. Their disk‐shaped mats were incubated with the cells for 24, 48, and 72 h. Their viability was determined using resazurin reduction assay. As seen in Figure 5d, the viability of three cell lines remained >90% up to 72 h, suggesting that the mats are biocompatible.

Given the promising results, BE/I‐PVA/DOX mats were further evaluated for their antitumoral activity. A series of mats containing various amounts of DOX ranging at 0.001–0.5 wt.% were fabricated. They were incubated with HeLa cells, along with free DOX as a control, for 48 hrs. As seen in Figure 5e, the IC₅₀ value of free DOX was determined to be 0.11 µg/mL (equivalent to 0.19 µm), which is in the range of IC₅₀ = 0.1–3.5 µm for Dox reported [63, 64, 65]. The IC₅₀ value for BE/I‐PVA/DOX was 0.26 µg/mL (equivalent to 0.45 µm), slightly greater than free DOX, which is plausibly due to the gradual release of encapsulated DOX from mats over time. As a control, the IC₅₀ value for uncrosslinked PVA/DOX was 0.16 µg/mL (equivalent to 0.28 µM), which is slightly smaller than crosslinked BE/I‐PVA/DOX mat, but slightly greater than free DOX. Importantly, the controlled release provided by the crosslinked BE/I‐PVA/DOX mats is expected to reduce peak drug concentrations, minimize systemic exposure, and limit potential toxicity, while maintaining effective cytotoxicity against HeLa cells. Overall, these results suggest that BE/I‐PVA/DOX mats exhibit good antitumoral activity and effectively suppress the proliferation of cancer cells, and thus tumor tissues, while BE/I‐PVA mats (no DOX) are biocompatible. Further, BE/I‐PVA/DOX mats were assessed for hemocompatibility (e.g., compatibility with blood) using a hemocompatibility assay where the mats were incubated with RBC suspension. As seen in Figure 5f, the hemolysis was determined to be 0.81 ± 0.1% for BE/I‐PVA mats (no DOX) and 0.98 ± 0.1% for BE/I‐PVA/DOX mats, both remaining below 5%, which is regarded to be hemocompatible.

3. Conclusion

We developed an effective approach to fabricate dual acidic pH‐degradable e‐spun nanofibrous mats labeled with both acid‐labile imine bond and BE crosslink. A new phenyl diboronic acid crosslinker, DBA‐I, was synthesized by a facile conjugation of an aldehyde‐containing boronic acid with an aliphatic diamine through in situ imine formation. The crosslinker DBA‐I turned out to be highly effective to crosslink e‐spun nanofibrous mats through the formation of BE bonds, confirmed by gravimetric, spectroscopic, and mechanical analysis. The crosslinked mats, along with DBA‐I crosslinker, exhibited rapid degradation through acid‐catalyzed hydrolysis of mainly Schiff base linkage, along with BE bond. The mats had high loading efficiency of DOX anticancer drug (>99%) and further showed the enhanced release of encapsulated DOX in tumoral and endo/lysosomal pH 5.0–6.9, compared with physiological pH 7.4. DOX‐loaded mats had antitumoral activity with IC₅₀ = 0.26 µg/mL and excellent hemocompatibility, while empty mats (with no DOX) were not cytotoxic. These results demonstrate that our approach has great potential to develop implantable localized nanofibrous scaffolds with enhanced and tunable degradation and drug release kinetics, offering great promise for application in post‐surgical cancer therapy.

Author Contributions

Sofia Nieves Casillas‐Popova: conceptualization (lead), formal analysis (lead), investigation (lead), validation (lead), writing – original draft‐preparation (lead), writing – revision & editing (supporting). Arianna Cirillo: investigation (supporting). Cameron D. Skinner: resources (equal). Dajana Vuckovic: resources (supporting). Jung Kwon Oh: conceptualization (supporting), funding acquisition (lead), supervision, writing – original draft‐preparation (supporting), writing – revision & editing (lead).

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File: adhm71251‐sup‐0001‐SuppMat.docx.

Data Availability Statement

The data that supports the findings of this study are available in the supplementary material of this article.