Isolation and characterisation of hGFs-MVs

Enriched MVs were characterised by cup-shaped MVs morphology by TEM (Fig. 2a). MVs and cell lysate were positive for CD9 by ELISA (Fig. 2b). The histogram of MVs NTA results showed MVs size ranging from 50 to 700 nm, with a few MVs peaks at 131 nm, 185 nm, 271, and 429 nm (Fig. 2c). MVs mean and mode were smaller than 200 nm (Fig. 2d), which has 77 µg MV protein (Fig. 2e), with MV particles at 4.6 × 109 and purity of 2 × 107 MV particles per µg protein (Fig. 2f).

Fig. 2

MVs characterisation. a Cup-shaped EV morphology by TEM. b MVs-enriched protein CD9 quantification by ELISA. c–f MVs size histogram and d mean/mode, e MVs protein and f MVs purity of MVs particles per µg protein. g EV surface marker characterisation using a multiplex exosome kit

To provide a more comprehensive understanding of hGFs-MVs, we conducted a multiplex analysis to investigate the surface markers of EVs, as illustrated in Fig. 2g. The results showed that hGFs-MVs exhibited elevated levels of CD9, CD63, and CD81, commonly recognised as EV markers. Notably, CD63 displayed a higher expression compared to CD9 and CD81. Furthermore, our analysis revealed the presence of CD44 and CD29 in hGFs-MVs (Fig. 2g). These results collectively indicate that hGFs-MVs express characteristic EV markers.

Few studies have investigated fibroblast cell-derived MVs. In the present study, hGFs-MVs size peaked at ~ 130 nm, which differed from published MVs sizes derived from a gingival epithelial cell line (~600 nm [24, 25]), hCMEC/D3 endothelial cells, and RAW 264.7 macrophages (~200 nm [38]). The MVs size difference between our study and other cell source-derived MVs may be attributed to different MVs isolation methods and varied cell sources. The current study used 16,000 g for 20 min, while other studies enriched epithelial cells MVs at 25,000 g for 30 min [24, 25] and endothelial/macrophage cell lines at 20,000 g for 45 min. The MVs enrichment method is a critical factor for obtaining a specific MVs population and is considered a key parameter the for MVs function [17].

Surface topography and distribution of MVs loaded on Ti and TNTs

Uni-directional microgrooves were observed on as-received Ti wires (Fig. 3a) that mimic the micro-machining lines on the surface of clinical dental implants [39]. Anodisation of these Ti wires in adequately aged ethylene glycol electrolyte at 80 V for 60 min yielded TNTs of diameter 110 ± 5 nm and 5.99 ± 0.428 µm in length (Fig. S1, Supplementary information). These nanotubes were uniformly self-ordered on the entire electrolyte-exposed region of Ti wires during anodisation. Furthermore, the characteristic nanotubular arrangement, open-tops, and inter-nanotube gaps are seen. It is noteworthy that despite the curvature of the Ti wires, the anodic film or TNTs formed on the surface appeared to be stable. This could be attributed to the use of aged electrolytes and reduced water-containing electrolytes [36, 37]. Dimensions of TNTs play a major role in drug loading and releasing abilities. In the current study, we fabricated TNT with higher diameters to facilitate the substantial loading of MVs (diameter 50–200 nm).

Fig. 3

Surface characterisation of various implant surfaces. a Top view SEM images of Ti wire, TNT, MVs-Ti, and MVs-TNT; b confocal images of DiO-labelled MVs loaded on Ti and TNT. Top view and side view confocal images are 3D stacked images of the MVs loaded on the substrates to demonstrate uniform loading on the curved surface

MVs were loaded on the surface of Ti and TNTs by immersing them in PBS for 1 h at room temperature. The MVs appeared to be distributed over the entire surface of the Ti substrates, as shown in Fig. 3a. DiO-tagged MVs were loaded and observed under confocal microscopy to further observe the distribution of MVs on the substrate surface (Fig. 3b). Interestingly, MVs showed a higher affinity towards Ti surfaces than the nanotubular surface. Earlier reports demonstrate that Ti surfaces have lower hydrophilicity compared to nanostructures [40]. Furthermore, exosomes and EVs have a high affinity towards hydrophobic surfaces, which makes techniques like hydrophobic interaction chromatography (HIC) suitable for EV isolation. This affinity of vesicles towards hydrophobic surfaces might be attributed to their enhanced activity towards the lesser hydrophilic Ti surface [41].

In vitro release of MVs from Ti and TNTs

MVs released from Ti and TNTs was visualised on days 0, 1, 3, and 7 using confocal microscopy (Fig. 4). The presence of MVs on the Ti and TNTs could be observed continually until day 7. A slight increase in MVs on the TNT surface was observed on day 3, which might correspond to the release of MVs from within the nanotubes and the inter-tubular spaces. The amount of protein measured in the release solution on day 1 demonstrated no significant difference in the amount of MVs released between Ti and TNTs. In contrast, MV-TNT showed significantly higher release than Ti on day 3 (Fig. 4c), although the amount of protein dropped substantially in both groups by day 7. Interestingly, while a trend towards higher levels of CD9+MVs was observed for TNT, there was no significant difference between the Ti and TNT groups throughout the 7-day release period (Fig. 4d).

Fig. 4

Evaluation of MVs release from Ti and TNTs. Confocal microscopy images of DiO-labelled MVs (green), loaded on a Ti and b TNTs at day 0, 1, 3, and 7 of immersion in PBS. Top view and side view confocal images are 3D stacked images of the MVs loaded on the substrates to demonstrate uniform loading on the curved surface. c Quantification of protein using BCA assay to quantify MVs release; d quantification of CD9 via ELISA to observe MVs release from Ti substrates. **p < 0.002

Cellular uptake of MVs on Ti substrates by keratinocytes

Several studies have demonstrated that nanotubular and nanoporous structures can enhance the attachment and proliferation of osteoblasts [40] and fibroblasts [5]; however, limited studies have investigated the influence of anodised nanotopography on keratinocytes. In dentistry, keratinocytes play a significant role in implant stability by maintaining an epithelial seal at the transmucosal interface with the oral cavity [42].

Smith et al. observed that human epithelial keratinocytes showed poor interaction and attachment with TNTs of 70–90 nm diameter, compared to Ti surfaces, attributed to the cuboidal architecture of keratinocytes [43]. However, in the current study, we did not observe differences in the attachment of keratinocytes on Ti and surfaces (Fig. 5). Furthermore, we quantified the percentage of green fluorescence using ImageJ from the images in an attempt to quantify the MVs uptake; however, no significant difference was observed between the MV-loaded groups (Fig. S2, Supplementary information).

Fig. 5

Keratinocyte attachment and cellular uptake of MVs loaded on various Ti substrates observed by confocal microscopy. Red indicates actin filaments, blue indicates cellular nuclei, and green indicates the DiO-labelled MVs. Scale bar indicates 100 µm

Earlier reports indicate that the nanotopography can facilitate the alignment of osteoblasts and fibroblasts along the underlying nano-topography [44]. Han et al. demonstrated that nanopores of diameter 66 nm can modulate molecular signalling cascades, enhancing the formation of mature focal adhesion points and aiding in elongating gingival fibroblasts along the underlying topography [4]. Similarly, titania nanopores of diameter 40–70 nm were previously observed to modulate the attachment and alignment of human osteoblasts along the underlying nanotopography [40]. In the current study, an effect of the underlying topography on the attachment or alignment of gingival keratinocytes was not observed.

A recent study by Poyraz et al. revealed that while keratinocytes cultured on random/aligned electrospun polycaprolactone (PCL) sheets demonstrated increased cellular proliferation, they did not exhibit alignment along the underlying topography [45]. Cells like fibroblasts and osteoblasts usually produce F-actin stress fibers in response to surface topography; however, this effect was not observed on the keratinocytes on PCL sheets, and they exhibited a clump-like morphology [45]. We have observed similar results where keratinocytes retained their cobblestone morphology while forming dense cell-cell connections. The inherent non-directional cuboidal cellular structure and short lamellipodia of keratinocytes [46] and the need to form strong cell-cell interactions for cell proliferation and signalling [45] might be responsible for its poor response to topographical cues [43]. Detailed analysis of the effect of topography on keratinocyte cellular functions and cell signalling cascades is lacking and would be an interesting avenue for further exploration.

We determined keratinocyte uptake of MVs using DiO-labelled MVs. The encapsulation of MVs by keratinocytes was observed on day 1 and day 3 for both Ti and TNT substrates. Interestingly, while the cells were attached uniformly on Ti and TNT surfaces, they appeared to attach in denser clusters surrounding the MVs on MV-Ti and MV-TNT. MSC cultures on exosome-immobilised Ti have previously demonstrated enhanced cell attachment, increased cellular spreading, and large cellular sizes compared to bare Ti [8]. Similarly, by day 3 in the current study, keratinocytes on MV-loaded substrates exhibited slightly larger size with stretched cuboidal morphology. MSCs demonstrated well-extended filopodial extensions on exosome-loaded Ti substrates. However, we did not observe such morphological variations for keratinocytes in response to MVs, perhaps due to the cuboidal structure and short lamellipodia of these cells [46].

Quantification of cytokine and chemokine released by keratinocyte in response to MVs

Cytokines and chemokines, such as IL-6, IL-7, IL-6, IL-1α, and MCP-1, play a significant role in immunomodulation by stimulating B cells and T cells and modulating the migration of neutrophils, monocytes, and macrophages [47]. Upon implant fixation, the host body responds to the implant material. It begins the secretion of various inflammatory cytokines, and if left unchecked, it can lead to inflammation of the surrounding tissues and even implant failure. One can regulate immuno-inflammatory responses at the implant site by modulating the levels of these chemokines. Recently, Bi et al. hypothesised that pocket epithelium in a periodontal disease condition could release MVs that can modulate the inflammatory response of fibroblasts [24]. In the current study, we evaluated the influence of MV-loaded Ti substrates on the inflammatory cytokine response of gingival keratinocytes.

On day 1, MV-Ti showed significantly lower expression of IL-7, IL-6, and IL-8 compared to Ti, while on day 3, the level of MCP-1 was significantly reduced (Fig. 6). MV-TNT also showed a significant reduction of IL-7 and IL-8 compared to the TNT group. This indicated that the incorporation of MVs significantly reduces inflammatory responses. Studies have shown that inflammatory cytokines like IL6 and IL1α play a critical role in the inflammatory response of macrophages towards Ti particles. Their increased expression can activate osteoclastogenesis and lead to bone resorption in in vivo conditions [48]. Notably, by locally inhibiting the activity of these cytokines, inhibition of triggered osteolysis can be achieved [49]. In our study, we successfully demonstrated that MV-loaded Ti substrates could downregulate the expression of IL-6 and thereby help modulate the inflammatory response.

Fig. 6

Chemokine and cytokine secretion in keratinocytes on various Ti substrates. *p < 0.05, **p < 0.002, ***p < 0.0002, between groups

Lamers et al. demonstrated that nanoscale surface modifications on Ti implants did not elicit any detrimental inflammatory response in in vitro and in vivo conditions; instead, it enhanced the expression of IL-1β, TNF-α, and TGF-β that aided in accelerated wound healing [50]. We also observed similar results where inflammatory cytokine expression from keratinocytes on TNTs was not higher than on Ti substrates, indicating that the nanotubular surfaces do not illicit any detrimental immune responses.

The effect of MV loading on cytokine and chemokine expression

The effect of MVs on cytokine production of keratinocytes was further confirmed by evaluating gene expression using qPCR. The mRNA expression of profiles showed that MV-loaded Ti or TNT wire-reduced gene expression of chemokines and cytokines (Fig. 7). The expression of TNFα (Fig. 7a), IL-6 (Fig. 7c), and IL1α (Fig. 7b) in the MV-TNT group was significantly reduced compared to the TNT group on day 1. At day 3, the MV-Ti group led to decreased TNFα, IL1α, MCP-1 (Fig. 7d), and MIP-1α (Fig. 7e) expression in contrast to the Ti group. It is noted that MIP-1α expression was reduced in both MV-Ti and MV-TNT groups compared to Ti and TNT groups on day 1, respectively. This indicates that MV loading can reduce the inflammation response in keratinocytes.

Fig. 7

Chemokine and cytokine gene expression in keratinocytes on Ti and TNT wires with and without MV loading. *p < 0.05, **p < 0.002, ***p < 0.0002, ****p < 0.0001 between groups

The anti-inflammatory effect of hGFs-EVs was in line with the recently published anti-inflammatory effect of hGFs [51] or hGFs in conditioned media [52], suggesting that hGFs-MVs may recapitulate the function of its parent cell—hGFs. Although they are a less studied EV population, MVs may possess the anti-inflammatory role of their parent cells [19, 20, 53,54,55], and hGFs-MVs share a similar function to reported gingival epithelial cell line-derived MVs [24, 25]. In addition, hGFs-MVs in TNTs also have a similar role as MSC exosome-loaded TNTs that reduce inflammatory response [11]. Interestingly, while chemokine and cytokine secretions from keratinocytes on TNTs did not exhibit significant reductions, mRNA expression of TNFα, IL-6, and MIP-1α was lower than that in Ti groups on day 3. This suggests that TNTs also play a role in modulating cytokine expression. This observation aligns with findings by Necsu et al., who noted a similar pattern in which macrophages cultured on TNTs, with and without LPS stimulation, displayed reduced expression and secretion of inflammatory cytokines [56]. Notably, changes in morphology, surface roughness, and wettability associated with nanotopography have previously demonstrated their capacity to alleviate inflammatory responses [57,58,59]. For instance, Chun et al. demonstrated that hydrophilic nanostructures induce minimal production of inflammatory cytokines such as TNFα and IL-6 [57]. Furthermore, our group has previously illustrated how anodisation augments surface hydrophilicity, which may contribute to their ability to reduce cytokine expression [60].

The current study demonstrates that MVs isolated from hGF can modulate cytokine release, thereby modulating inflammatory responses of keratinocytes. Results show that the MVs on both bare and anodised surfaces play a significant role in modulating inflammatory response, while the nanotopography also demonstrated a role in modulating cytokine expression. Furthermore, it is well-established that implant nanotopography augments osseointegration and soft-tissue integration [61]. This aspect positions MV-loaded TNTs as an exceptionally promising surface modification, offering the potential to effectively mitigate inflammation while concurrently expediting the processes of osseointegration and soft-tissue integration. Such characteristics hold paramount importance in the context of dental implants, making MV-loaded TNTs an attractive candidate for further exploration and development.

This ‘proof-of-concept’ study is one of the first to show that hGF-derived MVs can be successfully loaded onto Ti and TNT wires and demonstrate an anti-inflammatory effect of hGFs-MVs–loaded TNTs on keratinocytes. In this pioneering attempt, we have used a cylindrical 3D Ti substrate (with micro-roughness, mimicking clinical implant geometry/surface) to demonstrate that the loading and release of MVs from nano-engineered implants can facilitate local therapy. This paves the way for future application of MVs in implantology to promote soft-tissue integration via immunomodulation of keratinocytes.