Double‐Layer Microneedle Patch Loaded with HA‐PBA‐QCT for Management of Paclitaxel‐Induced Peripheral Neuropathic Pain

Abstract

Chemotherapy‐induced neuropathic pain (CINP) is a common adverse effect of antineoplastic drugs, often leading to dose reduction, treatment delays, or cessation of chemotherapy. Chemotherapy agents, like paclitaxel (PTX), damage the somatosensory nervous system by inducing neuroinflammation and oxidative stress, resulting in the sensitization of sensory neurons. Quercetin (QCT), known for its anti‐inflammatory, antioxidant, and neuroprotective properties, is investigated for various neurological disorders. This work creates phenylboronic acid‐modified hyaluronic acid (HA‐PBA) gels with incorporated QCT and fabricates a double‐layer microneedle (MN) patch using an HA‐PBA‐QCT complex in the needles and HA/polyvinyl alcohol (PVA) as the substrate. The crosslinking between PVA and HA‐PBA‐QCT enables a controlled, sustained release of QCT upon application. This work applies these QCT‐loaded microneedle (QMN) patches to the instep skin of PTX‐treated mice, which exhibits mechanical allodynia and cold hyperalgesia. Biweekly applications of the QMN patches significantly reduce pain responses. This analgesic effect is associated with the modulation of satellite glial cell activity, decreased macrophage infiltration, and reduced TNF‐α and IL‐6 levels in dorsal root ganglia (DRGs). Additionally, the treatment improves cellular antioxidant capacity, indicated by upregulated Nrf2 and catalase in DRGs. Overall, these findings suggest that double‐layer QMN patches offer long‐term anti‐inflammatory and antioxidant benefits, potentially alleviating CINP in patients.

The phenylboronic acid‐modified hyaluronic acid (HA‐PBA) gels incorporating quercetin (QCT) are used in a double‐layer microneedle (MN) patch for sustained transdermal drug release. The QCT‐loaded microneedle (QMN) patch delivers QCT to mice with paclitaxel‐induced neuropathic pain, alleviating pain responses by reducing neuroinflammation and oxidative stress. The patch modulates satellite glial cell (SGC) activity and macrophage infiltration and enhances antioxidant capacity in dorsal root ganglia (DRGs) and sciatic nerves.

1. Introduction

Chemotherapy‐induced neuropathic pain (CINP) predominately manifests as ongoing or spontaneous pain and hypersensitivity to mechanical and/or cold stimuli in hands and feet, affecting 19% to over 85% of patients receiving chemotherapy.^{[ 1 ]} CINP is a major dose‐limiting side‐effect that results in reduced dosage and dosing frequency or even impediment of the therapy regimen, thus compromising the efficacy of chemotherapy. Paclitaxel (PTX) is one of the first‐line chemotherapeutic agents and is primarily used to treat breast, lung, and ovarian cancers. ≈68% of patients reported neuropathic pain in the first month after PTX treatment, and 30% still experienced the pain after 6 months.^{[ 2 ]} The current medical treatments, including nonsteroidal anti‐inflammatory drugs, opioids, corticosteroids, and antidepressants, for PTX‐induced CINP are marginally effective and have several adverse effects that raise concerns.^{[ 3 ]}

PTX causes neurotoxicity by stabilizing microtubules in sensory neurons, therefore interfering with axon transport and thus resulting in neuropathy.^{[ 4 ]} Meanwhile, PTX can cause damage to mitochondria, which leads to the over‐production of reactive oxygen species (ROS) in sensory neurons.^{[ 4} , ^{5 ]} It has been widely reported that the administration of ROS scavengers and the activation of antioxidant pathways exhibit analgesic effects, further highlighting the pivotal role of oxidative stress in the development and maintenance of CINP.^{[ 6} , ⁷ , ^{8 ]} Neuroinflammation in peripheral sensory neurons is also closely related to CINP. Multiple types of pro‐inflammatory chemokines, for example, CB2 alongside chemokine (C‐C motif) ligand 2 (CCL2), are released from primary sensory neurons upon PTX exposure, which further triggers the activation of surrounding satellite glial cells (SGCs) as well as infiltration of macrophages.^{[ 9} , ¹⁰ , ^{11 ]} During the period of inflammation, the expressions of numerous cytokines, for example, tumor necrosis factor α (TNF‐α) and interleukin‐6 (IL‐6), are elevated, which can sensitize sensory neurons and thus contribute to CINP.^{[ 12 ]} Therefore, neuroprotection from PTX‐induced oxidative stress and regulation of cytokines and inflammation have become some of the most putative therapeutic strategies in current preclinical studies.^{[ 13 ]}

Quercetin (QCT) is a naturally occurring flavonoid glycoside and is well‐known for its potent antioxidant and anti‐inflammatory functions. As an antioxidant, QCT can directly scavenge superoxide anion radicals and thereby inhibit lipid peroxidation, which commonly occurs in pathological nervous systems.^{[ 14 ]} In addition, QCT can modulate antioxidant enzyme activities, such as glutathione, catalase, and glutathione peroxidase.^{[ 15 ]} Regarding its anti‐inflammatory functions, QCT is capable of decreasing TNF‐α and IL‐6 by inhibiting the activation of NF‐κB and MAPK pathways.^{[ 16 ]} In the treatment of neurological diseases, QCT effectively inhibited microglial recruitment and activation.^{[ 17} , ^{18 ]} Meanwhile, QCT exerted neuroprotective effects and prevented neuronal death by decreasing oxidative stress.^{[ 19} , ^{20 ]} Due to its function in neuroprotection and reducing neuroinflammation, QCT has exhibited therapeutic potential in a plethora of neurological and neurodegenerative diseases, including Parkinson's disease, Alzheimer's disease, multiple sclerosis, seizures, etc.^{[ 21} , ^{22 ]} However, QCT is extremely hydrophobic, which results in its poor water solubility and low absorption. Additionally, in a physiological environment, QCT is unstable and can be degraded or oxidized, leading to a short biological half‐life.^{[ 23 ]} Therefore, in treating CINP, a more practical method would deliver QCT conveniently and effectively to alleviate both regional pain and systemic neuropathy.

Boronic acid can react with two adjacent hydroxyl groups on the B‐ring of the QCT molecule to form cyclic boronate esters.^{[ 24 ]} Based on the reaction between QCT and boronic acid‐functionalized materials, nanodrugs have been developed to control the release of QCT.^{[ 25} , ^{26 ]} In the present study, we found that QCT can interact with phenylboronic acid (PBA)‐modified hyaluronic acid (HA‐PBA) to form HA‐PBA‐QCT complexes. Moreover, we previously developed HA‐PBA that crosslinked with polyvinyl alcohol (PVA) to form an HA‐PBA‐PVA hydrogel.^{[ 27 ]} The HA‐PBA‐PVA hydrogel also has inherent antioxidative and neuroprotective functions that protect nerves and prevent chronic pain after peripheral nerve injury.^{[ 28} , ^{29 ]} Additionally, the HA‐PBA‐PVA hydrogel can control a sustained release of flavonoid drugs, such as curcumin.^{[ 29 ]} Microneedle (MN) administration is a minimally invasive method for transdermal drug delivery. However, to date, the application of MN in neuropathic pain has not been well exploited.^{[ 30} , ^{31 ]} Thus, we further incorporated HA‐PBA‐QCT into an MN patch, which has a double‐layer structure with needle tips containing the HA‐PBA‐QCT component and a second layer composed of HA/ polyvinyl alcohol (PVA). The acute release of QCT from HA‐PBA‐QCT needle tips occurred immediately upon insertion. Meanwhile, gelation at the interface of PVA and HA‐PBA‐QCT encapsulated QCT within the crosslinking network, which extended the release profile of QCT. We further applied the QCT‐loaded microneedle (QMN) patches onto the foot instep skin of mice with PTX‐induced CINP. This treatment exerted persistent analgesic effects, reducing hypersensitivity to both mechanical and cold stimuli, primarily due to the anti‐inflammatory and neuroprotective functions of QCT. The overall experimental design and detailed schematics are outlined in Figure 1 .

Figure 1.

Schematic of QMN fabrication and application for treating CINP. A) Preparation of the HA‐PBA‐QCT microneedle; B) The QMN ameliorated PINP by regulating SGC activation, neuron sensitization, and macrophage infiltration in the peripheral nervous system.

2. Results

2.1. QCT Reduced Inflammation and Oxidative Stress in PTX‐Treated SGCs In Vitro

We first investigated the effects of PTX and QCT on SGC behaviors. SGCs are specialized glial cells found in sensory ganglia such as DRGs.^{[ 32 ]} SGCs surround the neuronal cell bodies and play crucial roles in both supporting neurons and modulating neuroimmune responses.^{[ 33 ]} Since alterations in SGCs and their interactions with DRG neurons have been shown to be significantly affected by PTX and closely related to elevated pain behavior in CINP evaluation,^{[ 34 ]} primary mouse SGCs isolated from DRGs were used as model cells to investigate the effects of QCT in vitro. The cytotoxicity of QCT and PTX on SGCs was first measured by CCK‐8. QCT and PTX treatments did not cause significant SGC death at concentrations lower than 30 and 0.5 µm, respectively (Figure 2A,B). We then chose 30 µm QCT and 0.5 µm PTX as safe concentrations for further in vitro experiments. The imbalance between increased ROS production and cellular antioxidants caused by chemotherapy drugs plays a significant role in inflammation and neuronal hyperexcitability, contributing to neuropathic pain. To evaluate the antioxidant effect of QCT under PTX challenge, 3 groups (control, PTX, and PTX+QCT) were generated, for which mouse SGCs were treated with no drug, 30 µm QCT, or 30 µm QCT together with 0.5 µm PTX for 1 day. The total ROS in the cells was measured using a Dichlorofluorescein diacetate (H2DCFDA) Cellular ROS Assay Kit. The ROS level was strongly elevated in the PTX‐treated SGCs compared to the control group. The increased ROS in SGCs was mitigated by the addition of QCT (Figure 2C). This prevention effect of QCT against a PTX‐induced ROS increase was further confirmed by confocal imaging of SGCs stained with DCFDA (Figure 2D). The number of SGCs with elevated ROS levels, indicated by green fluorescence, increased in the PTX‐treated group but was significantly reduced by QCT (Figure 2D). This suggested that QCT served as an ROS scavenger in SGCs under PTX treatment. Inhibition of neuronal inflammation is crucial for the exacerbation of CINP.^{[ 35 ]} To further investigate the antioxidant and anti‐inflammatory effects of QCT, SGCs treated with PTX and QCT were collected and were quantitatively measured by qPCR for their gene expressions of some inflammatory and anti‐oxidative factors. The relative gene expression of TNF‐α, a well‐studied pro‐inflammatory cytokine, after PTX treatment was significantly increased compared to that observed in the control group. This PTX‐induced increase in TNF‐α expression was effectively reduced by co‐treatment with QCT (Figure 2E). Nuclear factor erythroid 2‐related factor 2 (Nrf2), a key transcription factor regulating multiple antioxidative and anti‐inflammatory genes, and catalase, a crucial cell‐protective antioxidant enzyme that scavenges hydrogen peroxide into water and oxygen, were also evaluated in SGCs. After the PTX challenge, cellular catalase expression was significantly reduced, but this effect was mitigated by co‐treatment with QCT. Although PTX treatment did not significantly alter Nrf2 expression, the addition of QCT significantly enhanced NRF2 levels compared to the control group (Figure 2E). We also assessed the expressions of the immune regulating cytokine transforming growth factor‐beta (TGF‐β) and immune cell recruiting CCL2, but no significant differences were observed (Figure 2E). Overall, QCT showed antioxidant and anti‐inflammatory effects against PTX treatment.

Figure 2.

Bioactive function of QCT in reducing ROS and inflammation in vitro in SGCs. A,B) Mouse SGCs were cultured with QCT and PTX in different titrations for 1 day, and the SGCs were evaluated by a cell viability test (n = 8–9). C) Total ROS levels of mouse SGCs in the Control group, treated with 0.5 uM PTX, or further treated with 30 uM QCT for 1 day (n = 6). D) H2DCFDA fluorescence staining in mouse SGCs after being treated with saline, 0.5 uM PTX with dissolved empty microneedles, or further treated with the PTX and with dissolved microneedles loaded with 30uM QCT for 1 day (Scale bar = 100 µm). E) Gene expressions of TNF‐alpha, NRF2, Catalase, TGF‐beta, and CCL2 in SGCS measured by qPCR analysis (n = 4). *p < 0.05, **p < 0.01, ***p < 0.001, ns: no significant difference.

2.2. Formation of HA‐PBA‐QCT Complexes

We mixed either HA or HA‐PBA in DI water with a QCT ethanol solution. QCT mixed into the HA solution precipitated immediately, while the mixture of QCT and HA‐PBA did not lead to significant precipitation. Since PBA can form dynamic covalent bonds with cis‐diol groups of QCT (Figure 3A),^{[ 36} , ^{37 ]} we determined the binding rate by dissolving the lyophilization‐created HA‐PBA‐QCT composite in DMSO solution, in which free QCT would be dissolved in DMSO and HA‐PBA‐QCT would be kept in precipitation. We found that increasing the feeding ratio of QCT decreased the binding rate (HA‐PBA: QCT for 10:1, 10:2, and 10:5 yielded binding rates of 37.1 ± 4.7%, 18.2% ± 2.4%, and 6.4 ± 0.6%, respectively). However, the loading efficiency remained unchanged (HA‐PBA: QCT for 10:1, 10:2, and 10:5 had loading efficiencies of 3.6 ± 4.3%, 3.6% ± 0.4%, and 3.2 ± 0.3%, respectively). QCT is poorly soluble in aqueous solution and forms needle‐shaped crystals in water (Figure 3C), which were absent in the HA‐PBA solution. Instead, mixing QCT and HA‐PBA resulted in the formation of particle‐like structures, which increased in size with a higher feeding ratio of QCT (Figure 3C). The formed particles exhibited fluorescent signals due to the labeling of HA‐PBA (Figure 3B), indicating that particle formation is dependent on the binding reaction between HA‐PBA and QCT. SEM imaging also showed the formation of particles at nano‐ and micro‐scales when HA‐PBA and QCT were mixed at a 10:1 or 10:2 ratio. However, the particle shapes became less distinct at a higher QCT ratio (Figure 3C) and could last for at least 7 days (Figure S1, Supporting Information). Based on DLS analysis, the sizes of the HA‐PBA‐QCT particles were broadly distributed. The mixture of HA‐PBA and QCT at 10 to 1 resulted in polydispersity, with 81% of particle sizes at 818.0 ± 117.5 nm, while feeding HA‐PBA and QCT at 10 to 2 resulted in monodispersity in size at 1030.7 ± 194.0 nm (Figure 3E). Consistent with the imaging analysis, further increasing the QCT feeding ratio would lead to even broader dispersity and aggregation (Figure 3D).

Figure 3.

Formation of HA‐PBA‐QCT particles. A) Schematic of QCT interacting with HA‐PBA. B) The binding rate and loading rate of QCT in multiple HA‐PBA: QCT ratios (n = 4); ***p < 0.001, ns: no significant difference. C) Optical microscope imaging of HA‐QCT, HA‐PBA, and mixed particles in multiple HA‐PBA: QCT ratios (Scale bar = 10 µm). D) IF (Scale bar = 50 µm) and SEM (Scale bar = 5 µm) images of HA‐PBA or mixed particles in multiple HA‐PBA: QCT ratios. E) Quantitative analysis of the size of HA‐PBA (grey) and HA‐PBA‐QCT particles in ratios of QCT at 10:1 (blue) and 10:2 (orange) by DLS analysis.

2.3. Preparation and Characterization of Double‐Layer QMN

To fabricate a double‐layer‐MN patch that controls QCT release, we first filled the needle layer compartment with HA‐PBA or HA‐PBA‐QCT via the centrifuge method. We then added a mixed solution of HA and PVA to form the substrate as the second layer of the patch. The MN patches were formed and then peeled off the mold after dehydration (Figure 4A(a–c)). Representative SEM images of HA‐PBA‐QCT MNs are shown in Figure 4A(d–i). The surface of the HA‐PBA‐QCT MNs did not show the needle‐shaped crystals of QCT, which were evident in the MNs composed of HA and QCT (Figure 4A(g–i)). This indicates that the HA‐PBA and QCT are bound together in the HA‐PBA‐QCT MNs. To characterize the double‐layer structure of the MNs, HA‐PBA was labeled with rhodamine (Rho‐HA‐PBA), and the substrate solution was mixed with FITC. Signal distribution analysis showed that Rho‐HA‐PBA was present throughout the needles, from base to tip, while the substrate solution, comprised of HA and PVA, was primarily localized at the needle base (Figure 4B). The top and side views of the fabricated HA‐PBA‐QCT MN patch are shown in Figure 4Aa,C. To determine the penetration capacity of the MNs, the MN patches were pressed into an 8‐layer paraffin film so that the impressions left on each layer were able to be used to calculate the insertion depth. More than half of the needles left impressions in the third layer (≈375 µm), and the maximum insertion depth reached the fourth layer (≈500 µm) (Figure 4D). Meanwhile, the addition of PVA into the substrate solution showed an increased penetration capacity. To further test the insertion ability in mouse skin, the MN patches were applied to dorsal skin in vivo. The MNs created visible indentations, which were observed by wiping off the patch 1 h after application (Figure 4E). In addition, the MN patches were also applied to isolated dorsal skin and instep skin for histological staining. The staining image revealed that the MNs caused deformation and penetrated the stratum corneum (Figure 4F).

Figure 4.

Preparation and characterization of the double‐layer QMN patch. A) Representative images of the MN patch. a: Photo of a fabricated QMN patch; b,c: Morphology of the tips of the HA‐PBA patch observed by an optical microscope; d–f: SEM images of the needles on the HA‐PBA‐QCT patch captured at different magnifications; g–i: Zoomed‐in SEM images of the surfaces of the HA‐PBA, HA‐QCT, and HA‐PBA‐QCT patches. B) Double‐layer structure of the MN imaged by a confocal microscope. C) Size image of the fabricated MN patch. D) Penetration capacity of the MN detected by paraffin film (n = 3). ***p < 0.001. E) Process of the microneedle penetration test on mouse dorsal and instep skin in vivo. (F) Histological analysis of mouse dorsal and instep skin after the penetration test (Scale bar = 100 µm).

2.4. QCT Release Kinetics from the MN Patch in Acute and Chronic Phases

To determine the dissipation of MNs after insertion, Rho‐HA‐PBA MN patches, with a substrate containing either 0, 3, or 5% PVA, were thumb‐pressed into a skin‐mimicking agarose gel (Figure 5A). Generally, the fluorescent signals decreased over time. The rate of dissipation depended on the PVA concentration in the substrate; a higher PVA level resulted in more concentrated signals and maintained the signal intensity for longer (Figure 5B). At 24 h after insertion, a bulk gel persisted at the base of the MN (Figure S2, Supporting Information), indicating that PVA stabilized the Rho‐HA‐PBA by forming a gel. This contributed to the extended and slowed dissipation of the MNs. MN patches with 3% PVA in the substrate were used in subsequent tests due to their acceptable flexibility. QMN patches were pressed into agarose gel to determine the QCT release profile (Figure 5C). After being transferred into 0.5% Tween 20 and PBS solution, the HA/PVA substrate dissipated into the surrounding aqueous environment within 1 h. MNs composed of HA and QCT exhibited rapid dissipation, with most QCT (90.3 ± 2.6%) released within 7 h. In contrast, MNs containing HA‐PBA‐QCT demonstrated a two‐phase release profile of QCT, with a relatively fast release (25.1 ± 2.6%) on the first day, followed by a sustained release (59.7 ± 4.9%) over at least 7 days. The two‐phase profile may be attributed to the presence of both HA‐PBA‐QCT complexes, which can dissociate quickly, and the HA‐PBA‐QCT‐PVA gel, which provides a sustained release over the long term (Figure 5D).

Figure 5.

QCT release kinetics from the QMN patch in acute and chronic phases. A) Dissipation of MNs containing 0, 3, or 5% of PVA, determined by skin‐mimicking agarose gel (Scale bar = 500 µm). B) Quantification of the dissipation rate (n = 6). *p < 0.05, **p < 0.01, ***p < 0.001, ns: no significant difference. (C) QCT release rate of HA‐QCT and HA‐PBA‐QCT MN patches in agarose gel for 7 days (n = 6) D) Image of MN patch dissipation for 7 days.

2.5. Application of QMN Patches Ameliorated Mechanical and Cold Allodynia in Mice with CINP

Chemotherapeutics are known to cause damage to peripheral nerves, particularly in the extremities, leading to pain in areas like the fingers, hands, and feet. Thus, to mimic clinical medication, we applied the MN patch to the instep skin of the paw in mice with PTX‐induced neuropathic pain (Figure S3, Supporting Information and Figure 6A). By labeling HA‐PBA with Cy7.5, In‐Vivo Imaging System (IVIS) results showed that the MN patches were attached to the skin and could remain for at least 3 days (Figure 6B). To evaluate the therapeutic function of QMN patches in relieving pain, QMN patches were applied twice a week for 2 weeks after the full development of CINP. Mechanical allodynia and cold hypersensitivity induced by CINP were determined by von Frey and acetone tests, respectively. The timeline and protocol for inducing the CINP in‐vivo model are detailed in Figure S3, Supporting Information. Following continuous PTX injections, mice developed allodynia in response to non‐nociceptive mechanical stimuli on day 8, and this condition persisted for the subsequent 2 weeks (Figure 6C). The application of QMN patches on days 9 and 11 effectively reduced the mechanical and cold hypersensitivity. Subsequent applications maintained and further enhanced the analgesic effects. PTX treatment also triggered hyperalgesia to cold stimuli, evidenced by an increased frequency and duration of the typical pain response to acetone spraying. QMN patch treatment showed a pain‐relieving effect after just two applications, with subsequent treatments restoring the pain response to healthy levels (Figure 6C). Interestingly, the analgesic effects of the QMN patch application were also observed in the contralateral paws (Figure S4, Supporting Information), suggesting a systemic distribution of QCT after transdermal delivery by the QMN patch.

Figure 6.

Application of the QMN patch ameliorated mechanical and cold allodynia in mice with PINP. A) Process of MN application on the mouse instep skin in vivo. B) Retention test of the Cy7.5‐labeled QMN by IVIS imaging for 3 days after attaching it to mouse paw skin. C) Withdrawal threshold of the CINP mouse hind paw, measured by a von Frey test on ipsilateral paws (n = 5). D) Response frequency of mice to cold stimuli, measured by an acetone test on ipsilateral paws (n = 5). E) Response duration of mice to cold stimuli, measured by an acetone test on ipsilateral paws (n = 5). Statistical significance for Sham versus PTX: *p < 0.05, **p < 0.01, ***p < 0.001; Sham versus QMN: #p < 0.05, ##p < 0.01, ###p < 0.001; PTX versus QMN: &p < 0.05, &&p < 0.01, &&p < 0.001.

2.6. Application of QMN Patches Reduced Inflammation and Oxidative Stress in Drgs of Mice with CINP

To further investigate the alleviation of allodynia by QMN patches, we conducted IF staining of DRGs isolated from mice from the three groups. TRPV1, a marker of pain sensitization in sensory neurons, was used to assess the effects. In the PTX group, the TRPV1 expression in the DRGs was significantly elevated compared to the sham group. However, this increase was mitigated by QMN treatment (Figure 7A). Oxidative stress in the DRGs of the PTX group mice was also evident, as shown by DHE staining, a marker for intracellular ROS, particularly superoxide (Figure S5, Supporting Information). The PTX‐induced enhancement of the DHE signal was reversed by the QMN treatment, suggesting that QMN patches effectively reduce PTX‐induced hypersensitivity and ROS. This is consistent with the pain‐suppressing effects observed in the behavioral tests. Chemotherapy‐induced inflammation and elevated ROS levels often result in the release of pro‐inflammatory factors, leading to the recruitment of immune cells, particularly macrophages, to the damaged tissues and nerves.^{[ 11} , ^{38 ]} To assess this inflammatory response during CINP, we stained the DRGs for ionized calcium‐binding adaptor molecule 1(Iba‐1), a protein marker commonly used to identify microglia and macrophages. PTX treatment significantly increased Iba‐1 expression in DRGs, indicating macrophage accumulation and infiltration, which was prevented by QMN treatment (Figure 7B). We further validated these findings in sciatic nerve staining for DHE and CD68, a glycoprotein highly expressed in macrophages and used as a pan marker for these cells. The QMN treatment significantly prevented the PTX‐induced ROS pressure and macrophage infiltration (Figure 7C). To confirm the anti‐inflammatory and antioxidant properties of the QMN patches, we used qPCR to measure the relative expressions of related gene expressions in DRGs isolated from the three groups. Consistent with earlier findings, the PTX treatment significantly increased the gene expressions of pro‐inflammatory cytokines, that is, TNF‐alpha and IL‐6, while the QMN treatment mitigated these increases. In line with the in vitro qPCR results, the NRF2 expression remained unchanged by PTX but was upregulated by QMN. Catalase expression was reduced by both PTX and PTX combined with QMN. TRPV1 expression, elevated after PTX treatment, was significantly inhibited by QMN (Figure 7E), further confirming the protective effects of QMN against PTX‐induced hypersensitivity, inflammation, and oxidative stress.

Figure 7.

Application of the QMN patch reduced neuroinflammation and oxidative stress in DRGs in mice with PINP. A) IF staining of TRPV1 (red) and Tuj1 (green) in isolated mouse DRGs. B) IF staining of Iba‐1 in isolated mouse DRGs. C) IF staining of DHE (red) and CD68 (green) in isolated mouse sciatic nerves. D) Quantification of the fluorescent signal intensity of the TRPV1 in (A), the positive area of Iba‐1 in (B), and the fluorescent signal intensity of CD68 and DHE in (C) (normalized to Sham) (n = 5). E) In vivo qPCR results of isolated mouse DRGs (n = 4). *p < 0.05, **p < 0.01, ***p < 0.001, ns: no significant difference.

3. Discussion

Peripheral neuropathic pain induced by PTX, or other antineoplastic agents, is closely linked to inflammation, extending from the peripheral skin, nerve, and DRGs to the central nervous system.^{[ 39 ]} In the hind paw skin, the PTX treatment increased ROS levels in the epidermis. Elevated ROS levels potentially contribute to the loss of intraepidermal nerve fibers in mice with CINP and may also affect nerve terminals by inducing ectopic discharges and sensitizing nociceptors.^{[ 40 ]} In the DRGs, PTX treatment induced neuroinflammation, characterized by upregulated expressions of proinflammatory cytokines, including TNF‐α and IL‐6. DRG SGCs play critical roles in the inflammatory response because they can be directly activated by chemotherapeutic agents.^{[ 41} , ^{42 ]} SGCs, which support and maintain peripheral sensory neurons, were found to be disrupted by PTX treatment, leading to increased oxidative stress in vitro and in DRG tissue by downregulating catalase. Additionally, macrophages can infiltrate peripheral nerves, producing inflammatory mediators and ROS.^{[ 43 ]} Consequently, sensory neurons and nociceptors become sensitized, as indicated by the overexpression of nociceptive receptors like TRPA1 and TRPV1.

The anti‐inflammatory and neuroprotective properties of QCT have been documented in neurodegenerative and neuroinflammatory diseases, including Alzheimer's disease, Parkinson's disease, and strokes.^{[ 18} , ⁴⁴ , ⁴⁵ , ⁴⁶ , ^{47 ]} In CINP, QCT has been shown to stabilize mast cells, preventing histamine release and TRPV1 sensitization.^{[ 48 ]} Our study demonstrated that QCT also stabilized SGCs during PTX treatment, reducing inflammatory responses and oxidative stress in the DRGs and peripheral nerves. In previous studies, we identified HA‐PBA as a biocompatible material for drug delivery with intrinsic neuroprotective and antioxidant functions.^{[ 28 ]} QCT, which contains abundant hydroxyl groups, particularly the 3′,4′‐dihydroxy group on the catechol structure, can react with boronic acid to form irreversible covalent bonds.^{[ 49 ]} In this study, the reaction between QCT and HA‐PBA led to the formation of nano‐ to microparticles. We further incorporated these HA‐PBA‐QCT complexes into the MN tips followed by casting with HA/PVA to form the core of the needle and the substrate layer of the patch. Based on calculations, about 50 µg of QCT could be loaded into the MNs. Compared to QCT in HA‐based MNs, the QMN exhibited a two‐phase release profile upon insertion into a moist agarose gel. In the first phase, QCT was primarily released from free HA‐PBA‐QCT complexes in the needle tips. In the second phase, the HA‐PBA‐QCT that was close to the HA/PVA core crosslinked to form a gel network, entrapping QCT and resulting in a prolonged release phase. The crosslinking process between HA‐PBA‐QCT and PVA may occur during fabrication and application. When the PVA substrate layer is added, the PVA solution may dissolve some of the HA‐PBA‐QCT at the interface, forming crosslinkages. Upon insertion of the MNP into an aqueous environment, such as tissue, HA‐PBA‐QCT near the PVA core may also form boronate ester linkages due to the reaction between PBA groups and hydroxyl groups in PVA, resulting in the crosslinking.

Daily or single oral administration of 10–100 mg kg⁻¹ QCT has been reported to reduce pain.^{[ 48} , ⁵⁰ , ^{51 ]} Given QCT's rapid clearance rate, with an elimination half‐life of intravenous QCT between 3.8 and 86 min,^{[ 52 ]} the equivalent dosage of QCT resulted in ≈3 µg mL⁻¹ of plasma QCT 2 h after administration, which further decreased to less than 1 µg mL⁻¹ within 6 h.^{[ 53} , ⁵⁴ , ^{55 ]} Here, we further evaluated the practical application of QMN patches on mice with CINP. Since pain often occurs in the extremities of patients, we applied the QMN to the instep skin of mice with CINP. Four QMN patch applications significantly alleviated mechanical and cold hypersensitivity. Microneedle patches with similar QCT loading have been reported to reduce ROS.^{[ 56} , ^{57 ]} The high efficacy of QMN observed in this study may result from more efficient transdermal drug delivery via MNs, which avoids the first‐pass metabolism that can lead to the excretion of more than 90% of orally administered QCT.^{[ 58 ]} Transdermal delivery through MNs enables drug penetration into deeper dermal layer, thus providing both topical and systemic effects.^{[ 49} , ⁵⁹ , ^{60 ]} In addition, the sustained release of QCT from the HA‐PBA‐QCT complexes and the HA‐PBA‐QCT‐PVA gel likely contributes to maintaining therapeutic concentrations. The release kinetics of QCT from QMN insertion into the skin and its biodistribution in the blood and organs require further investigation.

The application of QMN patches reduced macrophage numbers and oxidative stress levels in the DRGs and sciatic nerves of PTX‐treated mice. Consistent with histological analyses, qPCR assays revealed that QMN downregulated the expression of pro‐inflammatory factors TNF‐α and IL‐6 while upregulating Nrf2 expression in the DRGs. Nrf2 is essential for neuronal defense against oxidative stress, and QCT has been reported to activate the Nrf2‐antioxidant response element, thereby mitigating damage caused by neuroinflammation and oxidative stress.^{[ 61 ]} Reduced neuroinflammation and oxidative stress consequently decreased the sensitivity of sensory neurons following PTX treatment, as evidenced by the downregulation of TRPV1. Although QMN patches were effective at reducing inflammation and oxidative stress, further studies are needed to fully elucidate the underlying molecular mechanisms, particularly the specific pathways through which QCT modulates these processes in SGCs and macrophages. Our future research will also explore the effects of QMN on a broader range of pain modalities and assess its efficacy in chronic pain models over extended periods. While the controlled release profile of QCT was determined, additional studies are necessary to evaluate the pharmacokinetics of QCT released from the QMN patch, including its metabolism and distribution in the blood, organs, and tissues, to better understand its therapeutic potential.

4. Conclusion

In summary, the reaction between HA‐PBA and QCT successfully formed HA‐PBA‐QCT complexes. We developed a double‐layer MN patch with HA‐PBA‐QCT needles and a HA/PVA substrate. The crosslinking of PVA with HA‐PBA‐QCT enhanced the controlled and sustained release of QCT after insertion. The application of the QMN patches in PTX‐treated mice significantly alleviated pain responses to mechanical and cold stimuli. Additionally, transdermal QCT delivery via the QMN patches reduced neuroinflammation and oxidative stress in the DRGs and peripheral nerves by modulating SGC activity, macrophage recruitment, and the cellular antioxidant system. Overall, the QMN patches demonstrated excellent sustained‐release capabilities and high therapeutic efficacy in treating CINP.

5. Experimental Section

Mouse SGCs Isolation and Culture

Mouse SGCs from dorsal root ganglia (DRGs) of C57BL/6 mice aged 2–4 weeks were isolated and cultured.^{[ 62 ]} The mice were euthanized and perfused with cold HBSS (Gibco). The L4‐5 DRGs were dissected and placed in a tube with HBSS on ice. The DRGs were then transferred to a tube containing HBSS and papain solution warmed to 37 °C and were incubated for 20 min. After centrifuging at 400 rpm for 3 min and gently removing the supernatant, the DRGs were washed with 3 mL of warmed HBSS. Subsequently, a solution containing collagenase (Gibco), dispase (Gibco), and TrypLE (Gibco) at a ratio of 1:1:1 (500 ul/500 ul/500 ul) was added and incubated for 20 min at 37 °C. The cells were then centrifuged at 500 rpm for 5 min. Finally, the cells were resuspended in standard culture medium and plated in a poly‐D‐lysine coated flask (Thermo Fisher Scientific). Isolated SGCs were cultured in DMEM (HyClone) supplemented with 10% FBS (Gibco) and 1% penicillin/streptomycin (Invitrogen) at 37 °C with 5% CO₂. Characterization of the SGCs was conducted based on staining for SGC markers: Glial fibrillary acidic protein (GFAP), glutamine synthetase, and S100B. All animal handling protocols were approved by the Institutional Animal Care and Use Committee (IACUC) at the University of Nebraska Medical Center (UNMC).

In Vitro Cytotoxic Assay of PTX and QCT

The in vitro PTX and QCT cytotoxic assays were tested using Cell Counting Kit 8 (CCK‐8, Abcam). SGCs (30 000 cells cm⁻²) were seeded in a 96‐well plate and cultured in DMEM with 10% FBS and 1% P/S overnight. To determine the cytotoxicity of PTX, SGCs were treated with 0, 0.5, 1, 5, 10, 50, or 100 µm PTX for 1 day. To determine the cytotoxicity of QCT, SGCs were treated with 0, 10, 20, 30, 60, or 100 µm QCT for 1 day. The SGCs were then incubated with the CCK‐8 assay solution for 1 h and were measured by a microplate reader for the absorbance at 460 nm (Synergy H1, BioTek).

In Vitro ROS Evaluation

The intracellular ROS were evaluated by a 2′,7′‐dichlorofluorescein diacetate (H2DCFDA) Cellular ROS Assay Kit (Abcam). SGCs (30 000 cells cm⁻²) were seeded in a 96‐well plate and cultured overnight in DMEM with 10% FBS and 1% P/S. The SGCs were then treated for 1 day with only medium, medium containing 0.5 µm PTX, or medium containing 0.5 µm PTX and 30 µm QCT. After incubating the SGCs with the H2DCFDA assay solution for 1 h, the fluorescence intensity was measured by a microplate reader at 460 nm. The treated SGCs were also imaged using a Zeiss LSM 710 confocal laser scanning microscope (CLSM) at Ex488/Em525.

Preparation and Characterization of HA‐PBA‐QCT

Biochemical synthesis of HA‐PBA through a 4‐(4,6‐dimethoxy‐1,3,5‐triazin‐2‐yl)‐4‐methyl‐morpholinium chloride (DMTMM)‐mediated coupling reaction was based on our established procedures.^{[ 27 ]} Briefly, 5 mg mL⁻¹ of HA, 2.35 mg mL⁻¹ of PBA, and 4 mg mL⁻¹ of DMTMM solution was prepared in DI water. The pH of the solution was continuously adjusted to 6.5 while the mixture solution was continuously stirred at room temperature for 3 days. Unreacted components were removed from the system through dialysis. The HA‐PBA conjugates from the dialyzed solution were then lyophilized. To prepare HA‐PBA‐QCT, 2.5 mg mL⁻¹ HA‐PBA in a 20% ethanol solution was mixed with 2.5 mg mL⁻¹ QCT in ethanol at volume ratios of 10:1, 10:2, and 10:5. After gently mixing them for 30 min, the solutions were lyophilized and stored at −80 °C before use. The binding rate of HA‐PBA and QCT was tested by dissolving HA‐PBA‐QCT complexes in DMSO, and then the dissolved QCT was determined at an absorbance wavelength of 375 nm using a microplate reader. The QCT binding rate and loading efficiency were calculated through the following equations:

$$\begin{array}{ccc}\hfill \mathrm{Binding}\mathrm{rate}& & =\left(\mathrm{Weight}\mathrm{of}\mathrm{feeding}\mathrm{QCT}-\mathrm{Weight}\mathrm{of}\mathrm{dissolved}\mathrm{QCT}\right)/\hfill \\ & & \mathrm{Weight}\mathrm{of}\mathrm{feeding}\mathrm{QCT}\hfill \end{array}$$ (1)

$$\begin{array}{ccc}\hfill \mathrm{Loading}\mathrm{efficiency}& & =\left(\mathrm{Weight}\mathrm{of}\mathrm{feeding}\mathrm{QCT}-\mathrm{Weight}\mathrm{of}\mathrm{dissolved}\mathrm{QCT}\right)/\hfill \\ & & \mathrm{Weight}\mathrm{of}\mathrm{HA}-\mathrm{PBA}\hfill \end{array}$$ (2)

To determine the morphological properties of the HA‐PBA‐QCT complexes, HA‐PBA was labeled with a fluorescent dye, Rhodamine (Rho). The HA‐PBA‐QCT or Rho‐HA‐PBA‐QCT derived from lyophilization was suspended in DI water, which was then observed and imaged using an optical microscope or a confocal microscope (Zeiss 880 LSM). In addition, the morphology of the HA‐PBA‐QCT complexes was also imaged by a scanning electron microscope (SEM, FEI Quanta 200). The size distributions of the HA‐PBA‐QCT complexes were further measured via dynamic light scattering (DLS, Malvern Instruments).

Preparation and Characterization of the HA‐PBA‐Based Double‐Layered MN Patch

After lyophilization, HA‐PBA or the equivalent HA‐PBA‐QCT was suspended in DI water to achieve a final concentration of 150 mg mL⁻¹, which was subsequently added into a micro mold (Micropoint Technologies Pte. Ltd). Through centrifuging at 3000 rpm for 30 mins, the HA‐PBA or HA‐PBA‐QCT filled the MN space, and then any extra material was removed. The MN layer was then dried in an oven at 37 °C for 20 mins, followed by the application of a substrate layer consisting of a 50 mg mL⁻¹ HA solution with 0%, 3%, or 5% PVA. After centrifuging it at 3000 rpm for 30 min, the MN patch was thoroughly dehydrated in an oven at 37 °C for 2 h. The final MN product was gently peeled off the mold and stored in a desiccator. An optical microscope and SEM were used to identify the morphology of the MNs. In order to determine the double‐layer structure, Rho‐HA‐PBA was used for the MNs, and the substrate layer was mixed with FITC, which allowed for fluorescent signals in each layer to be imaged by a confocal microscope.

Penetration Ability of the HA‐PBA‐Based Double‐Layered MN Patch

The penetration ability of MNs was tested on an 8‐layer parafilm sheet with a 1 mm thickness. Briefly, an MN patch was thumb‐pressed onto the parafilm sheet, and the number of indentions caused in each layer was counted to indicate the penetration rate of the MNs. In addition, the MN patches were pressed onto the dorsal or foot instep skin of mice, followed by moistening by using saline for better attachment to the skin. The patch was removed 1 h after application to examine the MN punctures. Meanwhile, skin samples that had the MNs attached were fixed in 4% paraformaldehyde (PFA), dehydrated using ethanol, and made into 10 µm paraffin sections for Hematoxylin and Eosin (H&E) staining.

Dissipation and QCT Release of MNs after Insertion into Agarose Gel

The patches with MNs made of Rho‐HA‐PBA and substrate layers containing either 0%, 3%, or 5% PVA were pressed onto a 3% agarose gel. The fluorescent signals at 10 min, 30 min, 2 h, and 24 h were acquired by a confocal microscope, and the intensity was measured using ImageJ software. To determine the QCT release profile, the patches with 3% PVA in the substrate layer and MNs made of either HA‐PBA‐QCT or a simple mixture of HA and QCT were pressed onto the agarose gel then transferred into 3 mL of 0.5% Tween 20 in PBS solution and incubated at 37 °C. The solution was collected and replenished at 0 h, 1 h, 3 h, 7 h, 1 d, 3 d, and 7 d. The dissolved QCT in the collected solutions at each time point, as well as the residual QCT within the agarose gel, were measured at day 7 using a microplate reader at an absorbance wavelength of 375 nm.

Retention of QMN In Vivo Visualized by Using an In Vivo Imaging System

To determine the retention of MNs in mice, HA‐PBA was labeled with a fluorescent dye, Cy‐7.5. The Cy‐7.5 labeled QMN patches were pressed onto the foot instep skin, followed by moistening and dehydration to ensure a strong attachment. Collars were applied to mice to prevent chewing. The fluorescent intensity of the MNs was visualized through an IVIS at 2 h, 1 day, and 3 days after application.

CINP Mouse Model and Treatments

The animal experiments involved in this study were approved by the IACUC at UNMC. Female C57BL/6 mice aged 56–62 days were ordered from Charles River Laboratories. After labeling them with ear tags and markings on their tails and weighing them (18–21 g), the mice were randomly divided into 3 groups: (1) Sham group (Sham) in which the mice were intraperitoneally (i.p.) injected with a vehicle solution (ethanol, Cremophor EL vehicle, and Saline at a 1:1:8 ratio) and were given MNs without QCT; (2) PTX‐induced neuropathic pain group, in which the mice were i.p. injected with a 0.8 mg mL⁻¹ PTX solution (in ethanol, Cremophor EL vehicle, and Saline at a 1:1:8 ratio) and were treated with MNs without QCT, and (3) QCT treatment group, in which the mice were i.p. injected with a 0.8 mg mL⁻¹ PTX solution and treated with a QMN. On Day 0, the baseline for the behavior test was determined for all the groups. PTX, or its dissolving solution, was injected in 4 doses on days 1, 3, 5, and 7. The experimental design and the timeline are shown in Figure S3, Supporting Information. MN patches were applied onto the skin of the right instep of the mice biweekly (days 9, 11, 15, and 18). On Day 22, after euthanasia, the ipsilateral sciatic nerves, DRGs, and foot instep skin were isolated from the mice for further tests.

Behavior Tests

Mechanical allodynia and cold sensitivity of the mice were assessed through behavior tests, including von Frey and acetone tests. Briefly, the sensitivity to tactile stimuli was assessed using a von Frey test. The mice were placed in boxes on top of a raised metal grid for at least 30 min before testing. Their movements were limited, but they remained unrestrained. The mice were stimulated by a series of von Frey filaments with exponentially increasing rigidity (from 0.008 to 2.0 g, Aesthesio) applied to their hind paws. Paw shaking, licking, or quick movement away from the filaments were considered to be indicators of a withdrawal reaction. Each mouse underwent five separate evaluations that were spaced 15 min apart. The mechanical withdrawal threshold was determined by analyzing the data from all five trials. Cold sensitivity was assessed using an acetone stimulation test. The mice were allowed a 30‐min acclimation period in the boxes and then were exposed to 50 µL of acetone on their ipsilateral hind paws. Paw shaking and licking were considered to be indicators of pain. The frequency of pain responses and response durations within 60 s after acetone application were measured.

Immunofluorescence Imaging

Mouse sciatic nerves, DRGs, and hind paw skin from each group were collected and fixed in 4% PFA at 4 °C overnight and incubated in 30% sucrose solution for cryoprotection for 1 day at 4 °C before being embedded in OCT for cryosection. After cryosectioning, the tissue slices were incubated in methanol for 10 min and rinsed with PBS for 10 min, repeating this three times. After blocking in 3% goat serum with 0.4% Triton X‐100 in PBS for 2 h, the slices were incubated with TNF‐alpha (1:1000; Abcam), TRPV1 (1:200; Abcam), CD68 (1:100, AbboMax), and TUJ1 (1:400; Abcam) first antibodies overnight at 4 °C. After washing them with PBS three times, the slices were stained with a secondary antibody (1:500; Alexa 488, Alexa 568) for 2 h and then stained with DAPI (1:1000, Thermo Scientific) for 30 min. For in vitro Dihydroethidium (DHE) staining, SGCs were incubated with 10 µm DHE for 1 h at 37 °C in the dark, followed by washing them three times with PBS. Finally, the fluorescence of the slices or SGCs was detected using a Zeiss LSM 710 CLSM.

RNA Isolation and Quantitative Polymerase Chain Reaction Analysis

The total RNA was isolated from SGCs or mouse DRGs for qPCR analysis using QIAshredder and RNeasy Mini Kit (QIAGEN), following the manufacturer's protocol. The purified RNA was then reverse transcribed into cDNA by Bio‐Rad iScript cDNA Kits for further PCR amplification by the StepOnePlus PCR System (Thermo Fisher Scientific) and SYBR Green Supermix (Bio‐Rad). The relative expression levels of the target genes (as indicated in Figure 1) were evaluated using the comparative Ct (2^−ΔΔCt) method.

Statistics

Statistical analysis was performed using GraphPad Prism 10 software. Student t‐tests were used in experiments with two groups, while in experiments with more than two groups, one‐way analysis of variance ANOVA was used with Tukey post‐hoc tests. Differences were denoted as * p < 0.05, ** p < 0.01, and *** p < 0.001.

Conflict of Interest

The authors declare no conflict of interest.

Supporting information

Supporting Information

Kong Y., Pan T., Liu B., Kuss M., Krishnan M. A., Alimi O. A., Shi W., Duan B., Double‐Layer Microneedle Patch Loaded with HA‐PBA‐QCT for Management of Paclitaxel‐Induced Peripheral Neuropathic Pain. Small 2025, 21, 2409748. 10.1002/smll.202409748

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.