3D‐Printed Microfluidic Platform for Creating Porous Nanofibrous Microspheres to Regulate Cell Response and Enhance Tissue Regeneration

Abstract

Porous nanofibrous microspheres (PNMs) present a versatile and minimally invasive strategy for tissue regeneration, combining biomimetic morphology, tunable structure, and injectability. While self‐assembly and co‐axial electrospray are explored for PNM fabrication, these methods are limited in compositional versatility and production scalability. Here, a 3D‐printed microfluidic platform is presented that enables large‐scale fabrication of PNMs with precise control over size, pore architecture, and morphology. PNMs can be functionalized with bioactive molecules through UV crosslinking, enhancing their regenerative potential by promoting osteogenesis in human bone marrow stromal cells (hBMSCs), angiogenesis in human umbilical vein endothelial cells (HUVECs), and exerting anti‐inflammatory effects on macrophages. Subcutaneous implantation in rats demonstrates that PNMs support cell infiltration, minimize fibrosis, and facilitate tissue integration, achieving complete cell penetration and tissue incorporation within 14 days. These findings establish PNMs as versatile, scalable, and customizable platforms, ideal for applications as injectable drugs or cell carriers, as well as powders, offering promising solutions for wound healing and tissue regeneration.

A 3D‐printed microfluidic platform enables the scalable production of porous nanofibrous microspheres (PNMs) with controlled size and porosity. These PNMs, functionalized with bioactive molecules, enhance osteogenesis, angiogenesis, and anti‐inflammatory responses, demonstrating significant potential for tissue regeneration. This streamlined approach offers a versatile solution for advancing regenerative medicine applications.

1. Introduction

Microspheres fabricated from synthetic or natural polymers represent a transformative approach for addressing irregular‐shaped tissue damage, owing to their customizable size, morphology, and material composition.^{[ 1 ]} Their injectability facilitates minimally invasive delivery and precise defect filling, and by incorporating bioactive molecules with tailored release profiles, microspheres can actively promote tissue regeneration and healing.^{[ 2} , ^{3 ]} Furthermore, their ability to self‐assemble into complex structures enhances their adaptability to varied anatomical geometries, providing substantial promise for clinical applications.^{[ 4 ]}

Hydrogel microspheres have been extensively utilized in tissue engineering due to their capacity to encapsulate cells and release bioactive agents in a controlled manner.^{[ 5 ]} However, their limited surface‐to‐volume ratio and restricted diffusion properties can impede the functionality of encapsulated cells. Additionally, hydrogel microspheres often face challenges with host tissue integration^{[ 6} , ^{7 ]} and control over degradation rates,^{[ 8 ]} which constrain their efficacy and long‐term application in tissue repair. Nanofibrous microspheres (NMs), on the other hand, have emerged as promising biomaterials due to their high surface‐to‐volume ratio, which mimics the nanoscale architecture of the extracellular matrix and promotes cell adhesion, migration, and differentiation.^{[ 3} , ⁶ , ⁹ , ¹⁰ , ¹¹ , ¹² , ¹³ , ¹⁴ , ^{15 ]} These properties have positioned NMs as exemplary platforms for cell and drug delivery in tissue engineering. However, the limited pore size of most NMs restricts the distribution of seeded cells to the microsphere surface, hindering cell infiltration post‐implantation and limiting their effectiveness in promoting tissue regeneration. To address this, porous nanofibrous microspheres (PNMs) have been proposed as ideal injectable scaffolds, offering enhanced cell loading capacity and regenerative potential.^{[ 2} , ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ , ²¹ , ^{22 ]}

Several fabrication techniques for NMs and PNMs have been developed, including emulsification coupled with self‐assembly,^{[ 2} , ⁶ , ¹³ , ¹⁴ , ¹⁵ , ¹⁹ , ²¹ , ²² , ²³ , ²⁴ , ^{25 ]} electrospraying,^{[ 3} , ⁹ , ¹⁰ , ¹¹ , ¹⁶ , ^{26 ]} and microfluidic approaches.^{[ 12 ]} Self‐assembly methods often require precise control over polymer synthesis and can be influenced by the molecular structure, which limits their compositional versatility.^{[ 22 ]} Additionally, the poor solubility of certain natural polymers in organic solvents and the need for high temperatures for processing add complexity and restrict material choices.^{[ 15} , ^{18 ]} While self‐assembly can produce PNMs, controlling pore size, pore number, and uniformity remains challenging, making large‐scale, reproducible production difficult and labor‐intensive.^{[ 2} , ²² , ^{23 ]} Electrospraying, particularly coaxial electrospraying, has demonstrated potential for producing PNMs with more controlled sizes and morphologies.^{[ 16} , ^{18 ]} This method provides flexibility in material selection and pore customization, addressing limitations of self‐assembly. However, challenges persist in scaling up production and ensuring uniform size distributions, primarily due to low solution flow rates and the formation of satellite droplets. Moreover, constructing coaxial spinnerets for generating PNMs requires manual assembly, making it labor‐intensive and complex. Microfluidic technology has emerged as a versatile and efficient tool for fabricating microspheres from various materials, addressing many limitations of traditional methods. The laminar flow regime characteristic of microfluidic devices enables precise control over microsphere size and morphology.^{[ 12} , ¹⁷ , ²⁰ , ²⁷ , ²⁸ , ^{29 ]} Moreover, microfluidic platforms facilitate high‐throughput production^{[ 30 ]} and the generation of diverse microsphere architectures, such as core‐shell,^{[ 31} , ³² , ^{33 ]} porous,^{[ 32} , ³⁴ , ^{35 ]} and multicompartmental structures.^{[ 36} , ^{37 ]} These features make microfluidic technology an attractive approach for scalable and consistent fabrication of advanced biomaterials for regenerative medicine. While several groups have utilized microfluidic devices to fabricate NMs^{[ 12 ]} and porous poly(lactic‐co‐glycolic acid) (PLGA) microspheres,^{[ 17} , ^{20 ]} the production of PNMs using this technology has, to our knowledge, not yet been demonstrated.

In this study, we present a 3D‐printed microfluidic platform tailored for the fabrication of PNMs with precisely controlled size, pore dimensions, and pore distribution. This platform employs a streamlined design that integrates inner and outer air flows with a short nanofiber suspension, enabling high‐throughput production while maintaining tunability and reusability. Furthermore, we conjugate bioactive molecules to the PNMs to support specific cellular functions, including osteogenesis, angiogenesis, and anti‐inflammatory responses. The effectiveness of these PNMs in tissue regeneration is evaluated through subcutaneous implantation in rats. Our innovative approach represents a significant leap forward, offering a robust and scalable solution for creating advanced biomaterials with broad potential applications in tissue engineering and regenerative medicine.

2. Results

2.1. Development of a 3D‐Printed PNM Generator

We successfully designed and fabricated a 3D‐printed PNM generator using a digital light processing (DLP) 3D printer, capitalizing on its ability to create complex fluid channel geometries. Despite the advantages of 3D printing, achieving high‐resolution micron‐scale fluid channels posed significant challenges due to limitations in the printing process. Specifically, micron‐size hollow structures were prone to blockage during fabrication.^{[ 38 ]} To address this issue, we employed a hybrid approach by integrating a 34‐Gauge syringe needle into the 3D‐printed design. The needle, with an inner diameter of 60 µm and an outer diameter of 250 µm, was strategically embedded into the 3D‐printed part to function as a micron‐sized inner air channel (Figure  1 ). This innovative integration circumvented the resolution constraints of the 3D printing technique, enabling the generation of fine‐scale fluid channels essential for PNM fabrication.

Figure 1.

Schematic illustrating the fabrication procedures of peptide‐conjugated PNMs and their applications in regulating cell response. 1) Preparation of short nanofiber suspension: Cryocutting of aligned electrospun PLGA:GelMA fiber mats and homogenization of short nanofibers and gelatin. 2) Preparation of PNMs: Manipulation of flow rates of nanofiber suspension, inner air, and outer air through a 3D printed porous microsphere generator, freeze‐drying and GA crosslinking. 3) Preparation of peptide‐conjugated PNMs: BMP‐2, QK, or AF‐1 peptides were tethered onto PNMs using UV crosslinking. 4) Regulation of cell response: PNMs conjugated with BMP‐2, QK, and AF‐1 peptides enhance osteogenesis in hBMSCs, stimulate angiogenesis in HUVECs, and exhibit anti‐inflammatory properties in macrophages. Created with BioRender.com.

2.2. Influence of Inner Air Flow Rate (Q_i ) on PNM Formation

The morphology and size of PNMs are influenced by multiple parameters, including the flow rates of inner air (Q_i ), outer air (Q_o ), and nanofiber suspension (Q_s ). These parameters collectively impact pore characteristics, such as the number and size of pores, and the overall diameter of PNMs. We first examined the effect of Q_i on PNM formation. We maintained Q_o and Q_s constant at 12 and 1 mL min⁻¹, respectively, while varying Q_i from 2 to 10 mL min⁻¹. Scanning electron microscopy (SEM) images revealed that Q_i primarily governs pore formation and the internal structure of PNMs (Figure  2a). At lower Q_i , PNMs exhibited a multicompartmental and highly porous internal structure. As Q_i increased, the internal structure transitioned to a hollower configuration with fewer compartments. Interestingly, the diameter of PNMs (D) remained consistent at ≈800 µm across all tested Q_i values (Figure 2b). However, the diameter of the pores (D_pore ) on the surface expanded with increasing Q_i (Figure 2b). Furthermore, the number of pores on PNMs was observed to be ≈10 when Q_i ranged from 2 to 8 mL min⁻¹ (Figure 2c). At Q_i = 10 mL min⁻¹, the number of pores decreased significantly to ≈4, demonstrating a dependence of pore characteristics on the inner air flow rate. These findings underscore the critical role of Q_i in tuning the structural properties of PNMs.

Figure 2.

Effects of Q_i on PNM morphology and size. a) SEM images of PNMs, including cross‐sectional views shown in the bottom row. b) D and D_pore as a function of Q_i . c) Number of pores per microsphere. The Q_o and Q_s were maintained at 12 and 1 mL min⁻¹, respectively. (n = 15).

2.3. Influence of Outer Air Flow Rate (Q_o ) on PNM Formation

We then investigated the effect of varying Q_o on PNM formation. To do this, Q_o was adjusted from 8 to 14 L min⁻¹, while Q_i and Q_s were kept constant at 6 and 1 mL min⁻¹, respectively. SEM analysis demonstrated that changes in Q_o significantly impacted the internal structure, D, and pore characteristics of PNMs (Figure  3a). As Q_o increased from 8 to 14 L min⁻¹, the internal structure of the PNMs became progressively hollower (Figure 3a), and D decreased substantially from ≈1200 to 700 µm (Figure 3b). The number of pores per microsphere decreased from ≈15 to 8 (Figure 3c) as Q_o increased, demonstrating a reverse relationship between pore density and Q_o . Interestingly, while the number and overall diameter of the PNMs were affected by Q_o , D_pore on the surface remained relatively consistent across different Q_o values. These results highlight the importance of outer air flow rate in modulating the size and pore density of PNMs.

Figure 3.

Effects of Q_o on PNM morphology and size. a) SEM images of PNMs, including cross‐sectional views shown in the bottom row. b) D and D_pore as a function of Q_o . c) Number of pores per microsphere. The Q_i and Q_s were held constant at 6 and 1 mL min⁻¹, respectively. (n = 15).

2.4. Influence of Nanofiber Suspension Flow Rate (Q_s ) on PNM Formation

We further investigated the impact of Q_s on the formation of PNMs. Q_s was varied from 0.5 to 2 mL min⁻¹, while Q_i and Q_o were maintained at 6 and 12 L min⁻¹, respectively. SEM analysis indicated that Q_s significantly influenced the internal structure, D_pore , and the number of pores per microsphere (Figure  4a). As Q_s increased, the internal space of the PNMs became more filled. Specifically, when Q_s was increased from 0.5 to 2 mL min⁻¹, D_pore decreased substantially from ≈180 to 80 µm (Figure 4B). Concurrently, the number of pores per microsphere increased from ≈6 to 10 (Figure 4C), demonstrating that higher Q_s values led to a denser pore structure. The D did not show significant changes with varying Q_s . These findings underscore the importance of nanofiber suspension flow rates in controlling pore characteristics, suggesting potential strategies for optimizing PNM properties for specific biomedical applications, such as tissue scaffolding and drug delivery systems.

Figure 4.

Effects of Q_s on PNM morphology and size. a) SEM images of PNMs, including cross‐sectional views shown in the bottom row. b) D and D_pore as a function of Q_s . c) Number of pores per microsphere. The Q_o and Q_i were maintained at 12 and 6 mL min⁻¹, respectively. (n = 15).

2.5. Conjugation of BMP‐2, QK and AF‐1 Peptides to PNMs

To enhance the functions of PNMs for regulating cell response, we conjugated bioactive molecules to PNMs through photocrosslinking of the methacrylic group in gelatin methacryloyl (GelMA) and octenyl alanine (OCTAL) in the modified peptides following a similar protocol in our previous studies.^{[ 3} , ^{18 ]} To confirm the successful conjugation of bone morphogenetic protein‐2 (BMP‐2)‐mimicking peptides, vascular endothelial growth factor (VEGF)‐mimicking QK peptides, and anti‐inflammatory peptide 1 (AF‐1) onto PNMs, we used fluorescently labeled peptides (BMP‐2‐OCTAL‐fluorescin isothiocyanate (FITC), QK‐OCTAL‐tetramethylrhodamine isothiocyanate (TRITC), and AF‐1‐OCTAL‐cyanine5 (Cy5)). Confocal laser scanning microscopy (CLSM) was employed to visualize the UV‐crosslinked peptides (UV exposure time = 20 min) on the PNMs, with BMP‐2 (green), QK (red), and AF‐1 (cyan) peptides uniformly distributed across the microsphere surfaces (Figure  5a–c; Videos S1–S3, Supporting Information). Next, we investigated the release kinetics of BMP‐2, QK, and AF‐1 peptides after conjugation to PNMs, varying the UV crosslinking time (3, 10, and 20 min) (Figure 5d–f). The release rates of all peptides were tunable by adjusting the UV crosslinking time. For example, when the UV crosslinking time was 3 min, over 80% of BMP‐2 peptides were released within 10 days. In contrast, extending the UV crosslinking time to 20 min resulted in less than 10% release over the same period. Furthermore, the release kinetics varied slightly depending on the peptide type. Cumulative release analysis revealed a nonlinear release pattern for BMP‐2 and a linear release profile for QK and AF‐1 peptides. These findings suggest that peptide‐conjugated PNMs allow for efficient and controllable release of BMP‐2, QK, and AF‐1 peptides, demonstrating their potential for sustained and controlled delivery in biomedical applications.

Figure 5.

BMP‐2, QK, and AF‐1 peptides conjugated PNMs. a–c) Z‐stack confocal microscopy images of BMP‐2 a), QK b) and AF‐1 c) peptides conjugated PNMs. The fluorescence images illustrate the distribution of BMP‐2, QK, and AF‐1 peptides on PNMs. Green: BMP‐2 peptides. Red: QK peptides. Cyan: AF‐1 peptides. d–f) Cumulative release curves of BMP‐2 d), QK e) and AF‐1 f) peptides from BMP‐2, QK, and AF‐1 peptides conjugated PNMs as a function of UV crosslinking time. (n = 2).

2.6. BMP‐2 Peptide‐Conjugated PNMs Regulate Osteogenic Differentiation

To assess the impact of BMP‐2 peptide conjugation on PNMs in facilitating osteogenesis in human bone marrow stromal cells (hBMSCs), we cultured hBMSCs on both BMP‐2‐conjugated and non‐conjugated PNMs. After a 4‐day proliferation phase, the proliferation of hBMSCs was measured using a CCK‐8 assay. Then, osteogenic differentiation was initiated and continued for 21 days as illustrated in Figure  6a. Notably, hBMSCs cultured on BMP‐2‐conjugated PNMs exhibited increased proliferation compared to those on unconjugated ones (Figure 6b). CLSM images show immunofluorescence staining of osteocalcin in differentiated hBMSCs on Days 14 and 21 (Figure 6c; Videos S4–S7, Supporting Information). Minimal osteocalcin expression was observed in the absence of BMP‐2 conjugation, while BMP‐2‐conjugated PNMs induced robust osteocalcin expression. Additionally, the mRNA expression of osteogenic and bone turnover markers, including RUNX2, OCN, BMP‐2, COL1A1, ALP, and RANKL, was significantly higher in hBMSCs cultured on BMP‐2‐conjugated PNMs (Figure 6d–i). These findings conclusively demonstrate that BMP‐2 peptide conjugation on PNMs substantially enhances the osteogenic differentiation of hBMSCs.

Figure 6.

Osteogenic differentiation of hBMSCs seeded on PNMs. a) Schematic illustrating the experimental design. hBMSCs were cultured on PNMs for 4 days to allow proliferation before initiating osteogenic differentiation. b) Proliferation of hBMSCs cultured on control and BMP‐2‐conjugated PNMs. (n = 3) c) Confocal microscopy images showing osteocalcin expression in hBMSCs seeded on PNMs after 14 and 21 days of culture, with or without BMP‐2 peptide conjugation. Red: Osteocalcin. Blue: DAPI. d–i) mRNA expression levels of osteogenic markers, including RUNX2 d), OCN e), BMP‐2 f), COL1A1 g), ALP h), and RANKL i). (n = 4, ^****: p ≤ 0.0001, ^***: p ≤ 0.001, ^**: p ≤ 0.01, ^*: p ≤ 0.05).

2.7. QK Peptide‐Conjugated PNMs Regulate Angiogenic Differentiation

To evaluate the potential enhancement of angiogenesis in human umbilical vein endothelial cells (HUVECs), we cultured cells on PNMs with and without QK peptide conjugation (Figure  7 ). CLSM images show immunofluorescence staining of CD31, an endothelial cell marker, in HUVECs on Days 3, 7, and 14 (Figure 7a–f; Videos S8–S13, Supporting Information). On Day 3, HUVECs displayed a tightly packed, cobblestone‐like morphology, regardless of QK peptide presence on PNMs (Figure 7a,d). By Day 7, HUVECs showed a more elongated and outstretched morphology, irrespective of QK peptide conjugation (Figure 7b,e). Notably, by Day 14, HUVECs cultured on QK peptide‐conjugated PNMs demonstrated enhanced angiogenic responses, including the formation of a microvascular network (Figure 7c,f). Except on Day 3, the length of the long axis of HUVECs on QK‐conjugated PNMs was significantly greater than that on control PNMs (Figure 7g). Proliferation of HUVECs was measured using a CCK‐8 assay, revealing improved proliferation on QK‐conjugated PNMs compared to control PNMs (Figure 7h). To further support these findings, mRNA expression levels of key genes regulating angiogenesis, including VEGF, VEGF‐R1, VEGF‐R2, VEGF‐C, TGF‐β, and ANG‐1, were assessed. The results showed elevated expression of all these genes in HUVECs cultured on QK peptide‐conjugated PNMs compared to control PNMs (Figure 7i–n). These findings indicate that QK peptide conjugation on PNMs enhanced the angiogenic differentiation of HUVECs, providing valuable insights for potential applications in tissue engineering and regenerative medicine.

Figure 7.

Angiogenic differentiation of HUVECs seeded on PNMs without and with conjugation of QK peptide. a–f) Confocal microscopy images showing CD31 expression in HUVECs which were cultured on PNMs without a–c) and with d–f) the conjugation of QK peptides for 3, 7, and 14 days. Green: CD31. Blue: DAPI. g) Length of long axis of HUVECs as a function of culture time. (n = 35–248 cells) h) Proliferation of HUVECs cultured on control and QK conjugated PNMs. (n = 3) i–n) mRNA expression of angiogenic markers including VEGF i), VEGF‐R1 j), VEGF‐R2 k), VEGF‐C l), TGF‐β m), and ANG‐1 n). (n = 4, ^****: p ≤ 0.0001, ^***: p ≤ 0.001, ^**: p ≤ 0.01, ^*: p ≤ 0.05).

2.8. AF‐1 Conjugated PNMs Regulate Anti‐Inflammatory Effects of Macrophages

For effective tissue regeneration, promoting a proper inflammatory response is crucial, as chronic inflammation can hinder the regeneration process. We further tested the effects of AF‐1 conjugated PNMs on macrophage response. Human monocytes were cultured for 5 days to differentiate into human macrophages. From Day 5, exogenous factors such as lipopolysaccharide (LPS) (to trigger a pro‐inflammatory response), dexamethasone (DEX) (to promote an anti‐inflammatory response), or a combination of both LPS and DEX were added to the culture medium, and treatment with AF‐1 conjugated or control PNMs was initiated (Figure  8a). Since morphological differences between M1 and M2 macrophages are difficult to distinguish microscopically (Figure 8b), we assessed macrophage response by measuring nitrite release using the Griess assay (Figure 8c) and quantifying tumor necrosis factor‐ α (TNF‐α) and interleukin‐10 (IL‐10) levels via ELISA assay kits (Figure 8d,e). AF‐1 conjugated PNMs significantly reduced nitrite production (a stable metabolite of nitric oxide) and TNF‐α (a pro‐inflammatory cytokine) in LPS‐stimulated macrophages. The level of IL‐10, an anti‐inflammatory cytokine, increased in response to LPS as part of a negative feedback loop to prevent excessive inflammation. However, after treatment with AF‐1‐conjugated PNMs, IL‐10 levels significantly decreased (Figure 8e).

Figure 8.

Anti‐inflammatory effects of AF‐1 peptide‐conjugated PNMs (AF‐1 PNMs). a) Schematic illustrating the experimental design. Human monocytes were cultured for 5 days to differentiate into macrophages. From Day 5, exogenous factors such as LPS, DEX, and a combination of LPS and DEX were added to the culture medium, and treatment with PNMs was initiated. b) Representative microscope images of macrophages treated with LPS and PNMs. c) Nitrite release measurement using the Greiss assay. (d‐e) Quantification of TNF‐α d) and IL‐10 e) levels via ELISA. f–j) mRNA expression levels of inflammatory markers, including TNF‐α f), IL‐1β g), IL‐6 h), IL‐10 i), and ARG1 j). (n = 4, ^****: p ≤ 0.0001, ^***: p ≤ 0.001, ^**: p ≤ 0.01, ^*: p ≤ 0.05).

Additionally, mRNA expression of TNF‐α, interleukin‐1β (IL‐1β), interleukin‐6 (IL‐6), IL‐10, and arginase 1 (ARG1) decreased significantly following AF‐1 conjugated PNM treatment of LPS‐stimulated macrophages (Figure 8f–j), indicating potent anti‐inflammatory effects. The reduction in pro‐inflammatory cytokines (e.g., TNF‐α, IL‐1β, IL‐6) suggests that AF‐1 conjugated PNMs effectively suppress macrophage activation and their production of pro‐inflammatory cytokines. While IL‐10 is an anti‐inflammatory cytokine, its downregulation in this context may be due to the strong suppression of the overall inflammatory response by AF‐1‐conjugated PNMs. Additionally, the decrease in ARG1 expression, which is associated with M2 macrophage polarization and tissue repair, suggests that AF‐1 conjugated PNMs may shift macrophages toward a less inflammatory phenotype, promoting resolution of inflammation.

2.9. Subcutaneous Implantation of PNMs

To explore the structural effect of NMs in tissue repair, we conducted subcutaneous implantation studies in rats using solid NMs (Figure S1) and PNMs with hollow and porous internal structures. We utilized Hematoxylin and Eosin (H&E) (the first two rows) and Masson's Trichrome (MT) (the last two rows) stains to assess cellular activity and collagen deposition at Days 7 and 14 postimplantation (Figure  9 ). H&E highlights cell nuclei in dark purple and cytoplasmic or extracellular matrix in pink. Yellow circles mark areas of interest. At Day 7, solid NMs exhibited limited cellular infiltration, with small clusters of inflammatory or immune cells indicating an early‐stage tissue response. Hollow NMs displayed moderate cellular infiltration around the edges of their structures, with cellular activity localized near the hollow regions. In contrast, PNMs demonstrated significant cellular infiltration, as their porous architecture facilitated better cell penetration, likely supporting early remodeling or inflammatory responses. By Day 14, solid NMs showed increased cellular activity, with regions of cell infiltration and tissue organization becoming more evident. Hollow NMs exhibited enhanced cellular migration and tissue remodeling within and around the hollow spaces. PNMs continued to show significant tissue ingrowth and remodeling, with their porous structure promoting a high degree of cellular activity and integration. Collagen deposition was further assessed using Masson's trichrome (MT) staining. At day 7, solid NMs showed minimal collagen deposition, suggesting early remodeling was limited to the microsphere periphery. Hollow NMs showed moderate collagen deposition localized around the hollow spaces. PNMs exhibited high collagen deposition in areas of active cellular infiltration, indicating robust extracellular matrix remodeling driven by their porous structure. On Day 14, solid NMs displayed a moderate increase in collagen deposition compared to Day 7, localized to the peripheral sites. Hollow NMs demonstrated enhanced collagen deposition along the hollow edges, suggesting progressive tissue remodeling. PNMs showed extensive collagen deposition, with widespread remodeling throughout the microsphere structure, indicating that their porous architecture enables extracellular matrix production within the microspheres. Overall, PNMs demonstrated superior outcomes, characterized by high cellular infiltration, robust extracellular matrix remodeling, and effective integration with host tissues. Notably, increased blood vessel formation within the PNMs suggests improved vascularization. These findings highlight PNMs as the most favorable environment for efficient tissue repair, likely due to their high cell permeability and porous architecture. This reinforces the notion that the porous structure of PNMs facilitates cell infiltration and vascularization, ultimately promoting more effective tissue regeneration.

Figure 9.

Subcutaneous implantation of developed microspheres in rats. Representative H&E (first two rows) and MT (last two rows) staining images showing the presence of solid, hollow, and porous nanofibrous microspheres (yellow circles) at Day 7 and their complete integration at Day 14. Yellow arrowheads: the site of the microspheres. Scale bars = 1 mm (low magnification, top row), 100 µm (high magnification, bottom row).

3. Discussion

PNMs can serve as injectable drug/cell carriers or fillers, demonstrating great potential in wound healing and tissue engineering.^{[ 2} , ¹⁶ , ¹⁷ , ¹⁸ , ²⁰ , ²¹ , ²² , ³⁹ , ^{40 ]} In this study, we demonstrated a 3D‐printed microfluidic platform for successful fabrication of PNMs with tunable size, pore structure, and morphology. The microfluidic‐based fabrication method offers significant advantages for generating PNMs in a large scale without restricting to certain compositions and with precise control over particle size, pore size, and pore density. These characteristics remain challenging to achieve using previous techniques like phase separation‐induced self‐assembly and electrospraying. For example, NMs produced by self‐assembly are limited to star shaped polymer (e.g., poly(_L‐lactic acid) (PLLA)) and few natural polymers (e.g., gelatin, chitosan, chitin, cellulose, collagen).^{[ 12} , ¹³ , ¹⁵ , ¹⁹ , ²² , ²⁴ , ⁴¹ , ^{42 ]} Among them, only star‐shape PLLA was demonstrated to form NMs with smooth surface, hollow, and sponge structures by varying arm numbers and arm lengths, which was mainly controlled by molecular architecture and the hydroxyl density.^{[ 24 ]} Although co‐axial electrospraying has been used to produce PNMs with various pore structures and compositions, it is difficult to scale up the production using this technique.^{[ 16} , ^{18 ]} In contrast, in this study, by simply varying Q_i , Q_o , and Q_s , we can obtain NMs with smooth surfaces, and hollow, and sponge structures.

The 3D‐printed PNM generator leverages a controlled interplay of inner air, nanofiber suspension, and outer air streams to generate droplets containing stabilized air bubbles. According to previous studies, the typical size of air bubbles generated by a submerged nozzle in liquid (oriented upward) is estimated based on a force balance equation, showing that bubble diameter increases with the inner airflow rate.^{[ 43 ]} Although our nozzle is oriented downward, the trend in bubble size remains consistent (Figure 2). At lower Q_i , smaller bubbles form due to the dominance of surface tension and moderate air pressure, leading to a multi‐pore internal structure within the PNMs. In contrast, at higher Q_i , shear‐driven coalescence leads to the formation of a dominant central bubble, resulting in a hollow core structure. The outer surface of the PNMs consistently exhibits multiple pores across tested flow conditions, attributed to bubble interactions at the droplet interface.

The stabilization of individual bubbles within the nanofiber suspension is primarily driven by the presence of gelatin, which acts as a surfactant, reducing surface tension and forming a protective layer around the bubbles. This surfactant effect prevents the coalescence of adjacent bubbles, ensuring a uniform distribution within the droplet, which maintains the integrity of the internal structure. The coflowing geometry of the device ensures precise control over the droplet and bubble formation processes, promoting the formation of consistent pore structures in the resulting PNMs.

The diameter of PNMs is determined by a delicate interplay of forces governing the droplet generation process. The outer air flow generates a drag force (F_D ) on the nanofiber suspension at the nozzle, which, together with the droplet's weigh (F_W ), must overcome the capillary force (surface tension) (F_C ) resisting detachment. The force generated by the inner airflow can be considered negligible because the velocity of the outer air (V_{outer air} ) is significantly greater than that of the inner air (V_{inner air} ). The force balance ensures that droplets are pinched off only when the combined forces exceed the capillary resistance, as described by F_D + F_W > ΨF_c where Ψ is the Harkins–Brown correction factor that accounts for the sustained liquid on the nozzle after droplet release.^{[ 30} , ^{44 ]} This mechanism enables precise control of droplet size by adjusting Q_o which directly influences F_D . As observed, increasing Q_o results in smaller PNM diameters (D) due to the higher F_D . Additionally, as demonstrated in our previous research, an alternative approach to controlling the diameter of PNMs involves modifying the nozzle diameter (d), which affects the capillary force.^{[ 30 ]} Smaller nozzles produce smaller droplets, consistent with the relationship D ∝ d ^1/3. For example, decreasing the nozzle size by half results in approximately a 21% reduction in the droplet diameter, highlighting the sensitivity of this parameter to nozzle dimensions.

Upon contact with liquid nitrogen, the nanofiber suspension droplets are rapidly solidified, trapping air bubbles in their spatial configuration and preserving the porous morphology. The rapid freezing also fixes smaller bubbles at the droplet's surface, leading to the formation of surface pores as some bubbles escape or partially evaporate. The resultant microspheres thus exhibit dual porosity: an internal architecture influenced by the inner air flow rate and an external porous surface governed by the interaction of bubbles with the droplet interface. These findings highlight the critical role of fluid dynamics, interfacial phenomena, and material properties in tailoring the structural features of PNMs for potential applications in tissue regeneration and drug delivery.

The cross‐linking of nanofibers plays a critical role in stabilizing their structure while maintaining biocompatibility for biomedical applications. In this study, we employed GA vapor crosslinking due to its effectiveness in forming stable covalent bonds without compromising the fibrous architecture. While GA is widely used, concerns regarding its potential cytotoxicity necessitate careful optimization. To mitigate toxicity, we implemented post‐treatment steps, including thorough drying and washing, to reduce residual GA. Previous studies have demonstrated that such post‐processing significantly minimizes cytotoxic effects, making GA‐crosslinked nanofibers viable for biological applications.^{[ 45} , ^{46 ]} However, alternative crosslinking strategies, such as genipin—a naturally derived crosslinker known for its lower toxicity—could be explored in future studies to further enhance biocompatibility. Thermal crosslinking was not considered due to its potential to alter the nanofiber morphology and mechanical properties, which are critical for maintaining the integrity of the microspheres. By carefully selecting and optimizing crosslinking methods, we aim to balance structural stability with biocompatibility for potential biomedical applications.

The 3D‐printed PNM generator ensures uniformity and scalability, making it highly suitable for large‐scale production of the PNMs. Compared to self‐assembly‐based phase inversion,^{[ 2} , ²¹ , ^{22 ]} which is labor‐intensive and constrained by material compatibility, and electrospraying,^{[ 16} , ^{18 ]} which poses challenges in scaling and process optimization, our microfluidic approach enables precise control over PNM size and porosity while maintaining high reproducibility. Additionally, porogen leaching and emulsion templating, though commonly used for generating porous structures,^{[ 19} , ²⁰ , ^{32 ]} often suffer from incomplete porogen or residual surfactants removal and difficulty in achieving well‐interconnected pores, which can lead to structural inconsistencies and reduced mechanical stability. In contrast, our method produces highly uniform and reproducible porous structures without requiring post‐processing for porogen removal. The 3D‐printed PNM generator achieves high‐throughput fabrication, as demonstrated in Video S14 (Supporting Information). Under the tested condition (Q_i = 2 mL min⁻¹, Q_o = 12 L min⁻¹, Q_s = 1 mL min⁻¹), the device produced 29 ± 0.9 PNMs/s (n = 3), equivalent to ≈1700 PNMs min⁻¹. Furthermore, the integration of a multi‐nozzle design into the PNM generator holds the potential to significantly increase the production rate.

PNMs can be further functionalized with biological molecules to regulate cell response using various surface conjugation approaches. In this study, we selected PLGA:GelMA as the base material for PNMs due to its combination of mechanical stability and bioactivity. PLGA provides structural integrity, while GelMA offers functional sites for cell adhesion and enzymatic degradation, making it well‐suited for tissue engineering applications. The presence of GelMA also enables facile peptide conjugation via photocrosslinking, allowing for precise control over bioactive molecule immobilization. The covalent surface functionalization with peptides has been demonstrated in our previous studies.^{[ 3} , ^{18 ]}

In this work, we demonstrated the conjugation of BMP‐2, QK, and AF‐1 peptides to PNMs via UV crosslinking of the OCTAL group and methacrylic group. The biological functionality of PNMs was extensively evaluated in vitro and in vivo. AF‐1‐conjugated PNMs exhibited potent anti‐inflammatory effects on modulating macrophage response, reducing the secretion of pro‐inflammatory cytokines like TNF‐α and IL‐6, and minimizing NO production, which can create a favorable immune environment conducive to tissue healing. Additionally, BMP‐2 and QK peptides conjugated PNMs enhanced osteogenic differentiation of hBMSCs and angiogenic differentiation of HUVECs, respectively, further demonstrating their multifunctionality. These findings suggest that PNMs can provide not only a structural framework for repair but also a biologically active environment that regulates positive cell response and promotes tissue regeneration. This study demonstrated the individual conjugation of each type of biological molecule to PNMs; however, multiple types of therapeutics could also be immobilized on PNMs with tailored release kinetics for each molecule by sequentially adjusting the UV crosslinking times. For example, to better accommodate the stages of alveolar bone regeneration, we could produce PNMs with simultaneous conjugation of BMP‐2, QK, and AF‐1 peptides with QK peptide release first for vascularization, AF‐1 peptide release after a week or so to reduce inflammation, and constant release of BMP‐2 peptide to promote bone growth. Alternatively, we could administrate a mixture of PNMs with surfaces functionalized by different bioactive molecules, each exhibiting distinct release kinetics.

The in vivo implantation studies demonstrated the superior performance of PNMs in supporting cell infiltration and reducing fibrotic responses compared to solid NMs. PNMs exhibited complete integration with newly formed tissue and surrounding tissues within 2 weeks postimplantation, leaving minimal fibrotic residues and fostering a conducive environment for tissue repair. Looking ahead, the ability of signaling molecules‐conjugated PNMs to promote vascularization, modulate immune response, and facilitate tissue formation, likely underpins their enhanced capacity for tissue regeneration.

The findings from this study lay the groundwork for further exploration of PNMs in regenerative medicine. Future research could focus on optimizing PNM design for specific tissue types, incorporating additional biofunctional molecules to prevent infection and target tissue healing pathways, and scaling up the microfluidic fabrication process for large‐scale production. Investigating the long‐term effects of PNMs on tissue remodeling and regeneration in preclinical models will also be critical for their translational success. Additionally, while the current study demonstrated the ability of peptide‐conjugated PNMs to promote osteogenesis and angiogenesis, and positively regulate inflammation in vitro, further studies could explore their potential in other therapeutic contexts, such as bone, skin, cartilage, or cardiac tissue regeneration. Leveraging advanced in vivo imaging techniques such as micro‐computed tomography (micro‐CT), magnetic resonance imaging (MRI), and ultrasound to better understand the in vivo degradation dynamics and host‐tissue interactions of PNMs would provide valuable insights for refining their design.

4. Conclusion

In summary, this study demonstrated a versatile and scalable 3D‐printed microfluidic‐based approach for the fabrication of PNMs with precise control over the size, pore structure, and morphology. PNMs can be easily functionalized with biological molecules through surface conjugation techniques, enabling enhanced osteogenic, angiogenic, and anti‐inflammatory properties in vitro. PNMs also promoted cell infiltration and tissue integration after subcutaneous implantation in rats when compared to solid NMs. Additionally, the minimal fibrosis and seamless tissue integration observed in vivo underscore their biocompatibility and potential for further exploration in wound healing and tissue regeneration applications. These findings establish PNMs as a minimally invasive therapy platform for tackling challenges in tissue repair and wound healing, while also paving the way for their applications in diverse regenerative therapies.

5. Experimental Section

PNM Fabrication and Functionalization

The fabrication of peptide‐conjugated PNMs involved a multi‐step process (Figure 1). Aligned PLGA:GelMA nanofibers were first generated on a rotating mandrel and crosslinked using glutaraldehyde vapors to enhance their stability. To prepare the short nanofiber suspension, cryocut nanofibers were homogenized with gelatin. Using a custom‐designed 3D‐printed porous microsphere generator, PNMs were fabricated by manipulating the flow rates of the nanofiber suspension, inner air, and outer air. To enhance their bioactivity, BMP‐2, QK, or AF‐1 peptides were tethered onto the PNMs via UV crosslinking. These peptide‐conjugated PNMs were subsequently evaluated for their ability to promote osteogenesis in hBMSCs, angiogenesis in HUVECs, and anti‐inflammatory effects in macrophages.

Nanofiber Suspension Preparation

An aligned nanofiber mat composed of PLGA (50:50 monomer ratio, 26269‐10, Polysciences, Warrington, PA) and GelMA was fabricated using electrospinning (Figure 1). For electrospinning, a suspension of 12% PLGA:GelMA (1:1 w/w) was prepared by dissolving the polymers in hexafluoroisopropanol (HFIP, Sigma‐Aldrich, St. Louis, MO) and loaded into a syringe fitted with a 21‐G needle. The electrospinning parameters included an applied voltage of 10 kV, a flow rate of 1 mL h⁻¹, and a needle‐to‐collector distance of 10–15 cm. Following fabrication, the nanofiber mat was crosslinked by exposure to glutaraldehyde (Sigma‐Aldrich, St. Louis, MO) vapors for 24 h. The crosslinked mat was then frozen in deionized water, cryo‐cut at −20 °C, freeze‐dried, and stored at 4 °C for subsequent use. The mean length of short nanofibers is 20 ± 12 µm (Figure S2, Supporting Information). To prepare the short nanofiber suspension, 200 mg of the cryo‐cut nanofibers underwent plasma treatment for 5 min. The treated nanofibers were then dispersed in 9 mL of deionized water and homogenized using a 20 kHz probe sonicator (Q500, Qsonica, Newtown, CT) under ice‐cold conditions. The sonication process involved on/off cycles of 10/20 s at 20% amplitude for 1 h, ensuring uniform dispersion of the nanofibers. The homogenized suspension was then allowed to equilibrate to room temperature. Separately, 10 mg of gelatin type A (Sigma‐Aldrich, St. Louis, MO) was dissolved in 1 mL of deionized water and subsequently added to the homogenized nanofiber suspension. The resulting mixture was vortexed for 3 min to ensure uniform incorporation of gelatin, producing a short nanofiber‐gelatin suspension with a nanofiber concentration of 20 mg mL⁻¹ and a gelatin concentration of 1 mg mL⁻¹. This suspension was used for the fabrication of PNMs.

3D Printed PNM Generator

The design of the PNM generator was conceptualized using Solidworks (Dassault Systèmes, Vélizy‐Villacoublay, France), and the design was subsequently converted into an STL file for 3D printing. The microsphere generator, printed with resin (RES‐01–3007, EnvisionTEC, Dearborn, MI), employed a digital light processing (DLP) 3D printer (Vida, EnvisionTEC, Dearborn, MI). Following the printing process, uncured resin residues on the device were thoroughly washed with isopropyl alcohol, and the device was cured in a UV chamber for 20 min. To introduce air bubbles into NMs, a 34‐G syringe needle was affixed to the 3D‐printed device (Figure 1).

The PNM generator is designed with distinct channels for the fluid flow of nanofiber suspension, outer air, and inner air. External airflow was employed to induce the pinch‐off of nanofiber suspension, resulting in the creation of droplets with the desired size. Detailed information regarding the droplet pinch‐off process can be found in our previous study.^{[ 30 ]} Pore formation on the PNMs was achieved by adjusting the flow rate of inner air. Subsequently, the microspheres were collected in a liquid nitrogen bath and subjected to lyophilization for 24 h in a freeze dryer. The freeze‐dried microspheres were then crosslinked with glutaraldehyde (GA) vapors for 24 h to enhance the structural stability of the microspheres. The impact of fluid flow rates of inner air, outer air, and nanofiber suspension on the dimensions and morphology of PNMs was assessed through SEM (FEI Quanta 200, FEI Company) images.

Peptide Conjugation on PNMs

To enhance the biological functionality of PNMs, we conjugated bioactive peptides (GenScript Biotech, Piscataway, NJ), including a BMP‐2‐mimicking peptide, a VEGF‐mimicking QK peptide, and an anti‐inflammatory peptide 1 (AF‐1), onto the PNMs (Figure 1). Each peptide was custom‐engineered with a photoreactive OCTAL group (BMP‐2‐OCTAL, QK‐OCTAL, and AF‐1‐OCTAL), facilitating photo‐crosslinking with GelMA within the PNMs. For conjugation, PNMs were immersed in a peptide solution (100 µg mL⁻¹) containing Irgacure 2959 (2 µg mL⁻¹) as a photoinitiator. The samples were exposed to 365 nm UV light for varying durations (3, 10, and 20 min). Following crosslinking, the peptide‐conjugated PNMs were thoroughly washed with deionized water. Based on optimization, a UV crosslinking time of 20 min was selected for cell culture studies. The success of peptide conjugation was validated through CLSM using fluorescently labeled peptides (BMP‐2‐OCTAL‐FITC, QK‐OCTAL‐TRITC, and AF‐1‐OCTAL‐Cy5). This approach allowed visualization of peptide distribution on the PNMs. Additionally, the release profiles of peptides from PNMs, influenced by the UV crosslinking time, were quantified by measuring the fluorescence intensity of the fluorophores in the release media.

hBMSCs Culture

hBMSCs were obtained from Lonza (PT‐2501, Lonza) and cultured following the manufacturer's recommendations. For experiments, control, and BMP‐2 conjugated PNMs were seeded with hBMSCs at a concentration of 4.0 × 10⁵ cells mL⁻¹. The cells were allowed to attach to the PNMs for 30 min in a microcentrifuge tube before transferring the hBMSC‐laden PNMs to a 48‐well low‐attachment plate for continuous culturing. The hBMSCs were initially cultured in proliferation medium (PT‐3001, Lonza) for 4 days, after which the medium was switched to osteogenic differentiation medium (PT‐3002, Lonza) and maintained for 21 days. The medium was replaced every 2 days during the culture period. PNMs were harvested on days 10, 14, and 21 to assess the osteogenic differentiation of hBMSCs.

HUVECs Culture

HUVECs were obtained from ATCC (PCS‐100‐013) and cultured in vascular cell basal medium (PCS‐100‐030, ATCC) supplemented with an endothelial cell growth kit (PCS‐100‐041, ATCC), following ATCC's recommendations. Before cell seeding, PNMs were pre‐coated with 20 µg mL⁻¹ fibronectin at 4 °C overnight. The pre‐coated PNMs were then washed twice with 1× Dulbecco's phosphate‐buffered saline (DPBS) to remove any unbound fibronectin. For experiments, HUVECs were seeded onto the control and QK conjugated PNMs at a concentration of 4.0 × 10⁵ cells per mL and allowed to attach for 30 min in a microcentrifuge tube. The HUVEC‐laden PNMs were then transferred to a 48‐well low‐attachment plate for continuous culture. The medium was replaced every 2 days, and samples were collected on days 3, 7, and 14 to evaluate angiogenic differentiation of HUVECs.

Human Monocyte Culture

Human peripheral blood mononuclear cells (PBMCs) were obtained from the Elutriation Core Facility at the University of Nebraska Medical Center (UNMC). A total of 1.0 × 10⁶ cells were seeded into each well of a cell culture‐treated 12‐well plate. PBMCs were differentiated into macrophages by culturing in RPMI 1640 medium supplemented with 10% fetal bovine serum, 2 mm L‐glutamine, 100 µg/mL streptomycin, 100 U mL⁻¹ penicillin, and 50 ng mL⁻¹ macrophage colony‐stimulating factor (M‐CSF). On day 5, macrophages were exposed to the following treatment conditions: 1) Control (M‐CSF only), 2) LPS (M‐CSF with 100 ng mL⁻¹ LPS), 3) DEX (M‐CSF with 1 nm DEX), 4) LPS + DEX (M‐CSF with both LPS and DEX), 5) Control PNMs (M‐CSF only), 6) LPS + Control PNMs (M‐CSF with LPS and control PNMs), 7) AF‐1 PNMs (M‐CSF only), and 8) LPS + AF‐1 PNMs (M‐CSF with LPS and AF‐1 PNMs). After three additional days of treatment, cells, and supernatants were collected for further analysis.

Immunofluorescence Staining

Cell‐cultured PNMs were fixed with 4% paraformaldehyde for 1 h at room temperature, followed by three washes with phosphate‐buffered saline (PBS). The samples were then blocked with 3% bovine serum albumin (BSA) in PBS containing 0.1% Triton X‐100 (PBST) for 1 h. For hBMSCs undergoing osteogenic differentiation, the samples were incubated overnight at 4 °C with a rabbit antiosteocalcin antibody (1:200, PA5‐96529, Thermo Fisher Scientific) diluted in 1% BSA in 0.1% PBST. For HUVECs, the samples were incubated overnight at 4 °C with a mouse anti‐CD31 antibody (1:200, ab9498, Abcam) prepared in the same blocking solution. After primary antibody incubation, the samples were washed three times with 1% BSA in 0.1% PBST. They were then incubated at room temperature for 3 h with appropriate secondary antibodies: Alexa Fluor 633‐conjugated anti‐rabbit antibody (1:200, A‐21070, Thermo Fisher Scientific) for hBMSCs and Alexa Fluor 488‐conjugated antimouse antibody (1:200, ab150113, Abcam) for HUVECs, both prepared in 1% BSA in 0.1% PBST. Following secondary antibody incubation, the samples were washed three times with 1% BSA in 0.1% PBST. Nuclei were stained by incubating the samples with 300 nm DAPI (4′,6‐diamidino‐2‐phenylindole) in PBS for 1 h, followed by three PBS washes. Finally, the samples were mounted in PBS on a glass‐bottom petri dish (LabTek, Grand Rapids, MI), covered with a glass coverslip, and imaged using a CLSM (ZEISS LSM 880, Carl Zeiss AG) equipped with 10× and 40× objectives.

Gene Expression Assay

Total RNA was extracted from osteoblasts, HUVECs cultured on PNMs, as well as from macrophages cultured in the presence of PNMs, using 1 mL of TRIzol Reagent (Thermo Fisher Scientific, CA) according to the manufacturer's instructions. After incubation at room temperature for 5 min, 700 µL of chloroform was added, and the samples were centrifuged to separate the phases. The aqueous layer was collected, and RNA was precipitated using ethanol. The resulting RNA pellet was washed, dried, and resuspended in DEPC‐treated water. The concentration and purity of the RNA were assessed using a Nanodrop spectrophotometer.

Complementary DNA (cDNA) was synthesized from 1 µg of total RNA using the Transcriptor First Strand cDNA Synthesis Kit (Roche) following the manufacturer's protocol. Real‐time PCR was performed in duplicate using a 96‐well PCR plate and a reaction volume of 25 µL per well. The reaction mix included SYBR Green detection dye (Applied Biosystems Group), and amplification was carried out on an ABI 7500 Fast and Real‐Time PCR system under the following cycling conditions: initial hold at 50 °C for 2 min, denaturation at 95 °C for 10 min, and 40 cycles of 15 s at 94 °C and 1 min at 60 °C. Primer concentrations ranged from 200 to 300 nmol L⁻¹. The primer sequences used in this study are listed in Table S1 (Supporting Information). Gene expression was calculated relative to control cells and normalized to the expression of 36B4, a housekeeping gene, under identical conditions using the 2^−ΔΔCT method.

Nitrite Release Assay

Nitrite production was quantified using a Nitrite Assay Kit (Sigma–Aldrich, St. Louis, MO) following the manufacturer's protocol. Supernatants collected from the various treatment conditions described previously were analyzed. In brief, 10 µL of each supernatant sample was mixed with the nitrite assay buffer supplied in the kit and added to a 96‐well plate. The absorbance was measured at 540 nm using a microplate reader. Nitrite concentrations in the samples were calculated using a standard nitrite calibration curve.

ELISA Assay

The concentrations of IL‐10 and TNF‐α in the collected supernatants were determined using Quantikine ELISA kits (R&D Systems, Minneapolis, MN) according to the manufacturer's instructions. Absorbance was measured at 450 nm using a microplate reader. The concentrations of IL‐10 and TNF‐α in the samples were calculated based on standard curves generated for each assay.

Subcutaneous Implantation and Histology

Subcutaneous implantation of microspheres was conducted with modifications to previously reported protocols. All animal procedures complied with guidelines approved by the Institutional Animal Care and Use Committee (IACUC) at the UNMC under protocol No. 17‐103‐11‐FC. Rats were anesthetized with 4% isoflurane in oxygen and maintained on 2% isoflurane during the procedure. A 4 × 4 cm² area on the dorsal region of each rat was shaved and cleaned three times with povidone‐iodine solution to establish an aseptic field. To maintain body temperature, rats were positioned on a heating pad throughout the surgery. Four small incisions were made to create subcutaneous pouches at paraspinal sites. NMs from each group (solid, hollow, and porous) were injected into individual pouches using a micropipette. The incisions were closed with surgical staples, anesthesia was discontinued, and the animals were monitored until full recovery before being returned to their cages. Animals were euthanized at one week and two weeks post‐implantation to collect and photograph the skin tissues for histological analysis. Skin samples were fixed in neutral buffered formalin and dehydrated through a graded ethanol series. The Tissue Core Science Facility at UNMC processed the samples for paraffin embedding, sectioning, and histological staining, including hematoxylin and eosin (H&E) and Masson's Trichrome staining.

Statistical Analysis

Data are presented as mean ± standard deviation. A one‐way ANOVA test was used for comparisons among multiple groups. Student's t‐test was used for pairwise comparisons. Statistical significance was described as ^*: p ≤ 0.05, ^**: p ≤ 0.01, ^***: p ≤ 0.001, and ^****: p ≤ 0.0001. Statistical analysis was carried out using GraphPad Prism (GraphPad Software, Boston, MA).

Conflict of Interest

The authors declare no conflict of interest.

Author Contributions

D.L. wrote the original draft, methodology, investigation, data curation, and conceptualization. H.Q.T., N.S.S., S.M.A., and Z.Y. wrote, reviewed, and edited, methodology, and data curation. A.C.K., R.A.R., and W.Z. wrote, reviewed, and edited. J.X. performed resources, supervision, project administration, funding acquisition, and conceptualization.

Supporting information

Supporting Information

Supplemental Video 1

Supplemental Video 2

Supplemental Video 3

Supplemental Video 4

Supplemental Video 5

Supplemental Video 6

Supplemental Video 7

Supplemental Video 8

Supplemental Video 9

Supplemental Video 10

Supplemental Video 11

Supplemental Video 12

Supplemental Video 13

Supplemental Video 14

Lee D., Tran H. Q., Sharma N. S., et al. “3D‐Printed Microfluidic Platform for Creating Porous Nanofibrous Microspheres to Regulate Cell Response and Enhance Tissue Regeneration.” Small 21, no. 41 (2025): 2502033. 10.1002/smll.202502033

Data Availability Statement

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