Physicochemical characterization of TLNTs
TLNTs were extracted from the juice of fresh tea leaves by differential centrifugation and further purified by sucrose density gradient ultracentrifugation. Driven by the sucrose gradient, TLNTs were mostly present at the interface of 30/45% (Fig. 1A). The transmission electron microscopy (TEM) and the atomic force microscopy (AFM) were employed to detect the morphology of the purified TLNTs. As shown in Fig. 1B, C, these NTs presented exosome-like spherical particles with an average particle size of 70 nm. The further dynamic light scattering (DLS) results revealed that TLNTs had a hydrodynamic particle size of 166.9 nm, a uniform size distribution (PDI = 0.100) and a negatively charged surface (− 28.8 mV), as shown in Fig. 1D. The difference among particle sizes determined by the TEM, AFM and DLS might be ascribed to the phenomenon that TLNTs are fully dehydrated prior to applying them for imaging, while they stay in highly swollen and wet states during the DLS test.
The lipidomic analysis implied that TLNTs were primarily composed of phosphatidylcholine (PC, 40%), phosphatidylmethanol (Pme, 15%), phosphatidylethanol (Pet, 11%), triglyceride (TG, 11%), phosphatidic acid (PA, 7%), diacylglycerol (DG, 6%) and monogalactosyldiacyglycerol (MGMG, 4%) (Fig. 1E). All these molecules were amphiphilic, and they established the structural foundation for stability of TLNTs. Furthermore, protein compositions in TLNTs were investigated by using liquid chromatography coupled with tandem mass spectrometry assays. It was found that 446 kinds of proteins were present in TLNTs (Fig. 1F). The biological functions of these proteins were analyzed by using the Gene Ontology (GO) database. We found that over 40 kinds of proteins related to cancers, metabolic processes, immune diseases and cell components (Fig. 1G). According to the GO database and Genes and Genomes (KEGG) Annotation Path Analysis, TLNTs contained 11 kinds of proteins associated with cellular processes, metabolic processes, cell parts and catalytic activity (Fig. 1H).
Tea leaves are enriched with active small molecular constituents, such as polyphenols and flavonoids, which possess the capacity to impact on cell apoptosis, cell migration and immune responses [20,21,22]. Therefore, we quantified the contents of the active small molecules in TLNTs by high performance liquid chromatography-tandem mass spectrometry (HPLC-MS/MS). As presented in Fig. 1I, J, TLNTs contained large amounts of the well-documented anti-tumor polyphenols and flavonoids, including EGCG [23], vitexin-2-O-rhamnoside [24], vitexin [25], myricetin-3-O-rhamnoside [26], kaempferol-3-O-galactoside [27] and myricetin [28]. The results relevant to the TLNT compositions collectively provide a primary foundation for their application in cancer treatment.
In vitro cellular uptake and anti-tumor activities of TLNTs
Efficient cell internalization of NTs is a prerequisite for exerting anti-tumor effects. The CLSM images revealed that the control cells (without NT treatment) exhibited no green fluorescence signals. On the contrary, after co-incubation for 5 h, almost all the cells showed obvious green signals (TLNTs), which were predominantly distributed in the cytoplasm (Additional file 1: Fig. S1). Furthermore, the cellular uptake efficiencies were quantified by FCM. It was observed that both cellular uptake percentages and mean fluorescence intensities (MFIs) of DiO-labeled TLNTs increased over time (Additional file 1: Fig. S2), and more than 80% of cells internalized TLNTs after 5-h co-incubation. These observations demonstrate that TLNTs are preferentially internalized by 4T1 cells and mainly present in the cytoplasm.
Subsequently, the in vitro inhibitory effects of TLNTs on the proliferations of various tumor cell lines were investigated by the MTT assay. After co-incubation with TLNTs, the viabilities of CT-26 cells, MCF-7 cells and 4T1 cells gradually decreased with the increased TLNT concentrations and prolonged incubation (Fig. 2A). Meanwhile, we calculated the half maximal inhibitory concentrations (IC50) of TLNTs to compare the anti-proliferatory capacities of TLNTs (Additional file 1: Table S1). It was detected that after co-incubation for 24 h, the IC50 value of TLNTs against CT-26 cells was 1.2 and 1.8-fold higher than those against MCF-7 cells and 4T1 cells, respectively, and strikingly, became 2.5- and 220.8-fold higher, just forty-eight hours after incubation. These findings clearly demonstrate that TLNTs have a stronger capacity to suppress the proliferation of breast tumors (especially for 4T1 cells) than that of colon tumors. Therefore, TLNTs were used to inhibit the growth of breast tumor cells (4T1 cells). Subsequently, the pro-apoptotic properties of TLNTs against 4T1 cells were assessed by FCM. As shown in Fig. 2B, the apoptotic percentage of the control cells (without TLNT treatment) was 2.3%, whereas the apoptotic percentages of cells receiving the treatment of TLNTs for 4 and 8 h increased to 25.4% and 67.0%, respectively. It is widely believed that tumor metastasis is the main cause of death, and its fundamental step is tumor cell migration [29]. As a cheap and highly reproducible method, the cell scratch assay was performed to determine the suppressive effect of TLNTs on tumor cell migration. As seen in additional file 1: Fig. S3, the treatment of TLNTs significantly inhibited the migration of 4T1 cells, in comparison with the control cells.
In vitro anti-tumor mechanism of TLNTs
Previous studies indicated that polyphenols and flavonoids could increase oxidative stress in tumor cells [30, 31]. Accordingly, we determined the produced amounts of intracellular ROS in the TLNT-treated cells. It was observed that after the TLNT treatment, green fluorescence signals gradually increased, and they were mainly present in the interior of the entire cells, including nucleus (Fig. 2C, D). These observations imply that large amounts of intracellular ROS are produced in the TLNT-treated cells, which might damage the fundamental substances involved in life activities of tumor cells [32, 33]. Mitochondria play important roles in various cell activities such as energy supply, metabolism and apoptosis [34]. To determine the damage effect of TLNTs on mitochondria, the mitochondrial membrane potential of the TLNT-treated 4T1 cells was evaluated using a mitochondrial probe (JC-1). Figure 2E showed that the negative control cells exhibited dark yellow fluorescence signals, indicating that these cells had largely intact mitochondrial membranes, while the positive control cells and the TLNT-treated cells appeared green, which suggested their damaged mitochondrial membranes, resulting in the decreased mitochondrial membrane potential. In all, these results demonstrate that TLNTs can efficiently induce mitochondrial damages.
Furthermore, we performed PI staining to investigate whether TLNTs could induce cell cycle arrest. It was determined in Fig. 2F that the percentages of the TLNT-treated 4T1 cells in G0/G1 phase and S phase remarkably increased, and those in G2/M phase greatly decreased. These results indicated that the TLNT treatment could block DNA replication in tumor cells and eventually cause cell death. The progression of cell cycle is regulated by various cycle proteins, including Cyclin A, Cyclin B and Cyclin D. These three kinds of proteins are involved in S/G2, G2/M and G1/S transitions of cell cycles, respectively. It was discovered that the intracellular amounts of Cyclin A, Cyclin B and Cyclin D were greatly decreased with the treatment of TLNTs, and the decreased trend was positively correlated with the treatment time (Fig. 2G), suggesting that TLNTs could inhibit cell cycle progression through down-regulation of the critical cycle proteins.
In vivo bio-distribution of TLNTs
I.v. injection is a widely used drug administration route in the preclinical treatment of various cancers, while alternatively, oral administration is the most preferable approach for patients in terms of safety, noninvasiveness, satisfactory compliance and cost-effectiveness [35]. As these two approaches possessed their own merits, we comparatively investigated the in vivo bio-distribution profiles of TLNTs via i.v. injection and oral administration. Initially, TLNTs were labeled with a DiR, administrated to mice bearing 4T1 tumors and processed by the IVIS spectrum imaging system. To determine whether fluorescence dye could remain in the TLNTs during the passage through the acidic stomach, we investigated its release profile from TLNTs. It was found that less than 20% of fluorescence dye was released from TLNTs after 6-h incubation in the stomach stimulating solution (pH 2.5) (Additional file 1: Fig. S4). As shown in Fig. 3A, TLNTs gradually accumulated in the tumors irrespective of drug administration approaches. We further found that the strongest fluorescence intensity of tumors after i.v. injection and oral route was readily found at the time point of 24 and 48 h, respectively (Additional file 1: Fig. S5). Nevertheless, the maximal fluorescence signals were present in the small intestine at the time point of 6 h for both administration routes, and then these signals tended to fade away.
To clarify the adsorption sites of TLNTs in the GIT after oral administration, mice were gavaged with DiO-labeled TLNTs (3 mg protein/kg mice), and their GITs were excised, sectioned and stained with DAPI. As visualized in Fig. 3B, there were a few green fluorescence signals in the mucosa of stomachs, ceca and colons, which appeared in the surface layer of the duodenum mucosa. It was worth noting that strong green signals suffused the mucosa of the jejunum and ileum, demonstrating that TLNTs might be adsorbed into the circulatory system through these two sections of the GIT (jejunum and ileum).
In vivo anti-breast tumor effect of TLNTs
There is growing evidence that the intestinal microbiota plays a critical role in the initiation and development of various diseases [9, 10, 36]. We employed ATBs to investigate whether the intestinal microbiota exerted impacts on the therapeutic outcomes of oral nanomedicines against breast cancer. Mice bearing subcutaneous breast tumors were established and divided into 7 groups, namely the control group, the TLNT (i.v., low)-treated group, the TLNT (i.v., high)-treated group, the TLNT (i.v., high, ATB)-treated group, the TLNT (oral, low)-treated group, the TLNT (oral, high)-treated group and the TLNT (oral, high, ATB)-treated group.
Body weights and tumor volumes of various mouse groups were recorded during the treatment with TLNTs. Figure 4A showed that there was no apparent difference detected in body weights among the control group and the treatment groups, and all the mice did not exhibit any abnormalities during the entire treatment period. Compared with the control group, the mean tumor volumes were 1.1- and 2.6-fold smaller for the TLNT (i.v.)-treated group and 1.8- and 2.2-fold smaller for the TLNT (oral)-treated group, at a 1.5 and 3 mg/kg protein dose, respectively, on day 15 (Fig. 4B), which are in line with morphological changes in tumor weights and sizes (Fig. 4C, D). Meanwhile, we found that the anti-breast tumor effect of TLNTs (i.v. or oral) greatly decreased after the treatment of broad-spectrum ATBs. It was worth noting that the control group showed negligible variations in spleen weights compared with the TLNT-treated groups, except the TLNT (oral, high)-treated group (Fig. 4E). The spleen weight variation is a result of systemic inflammatory responses [37], and the inflammation is associated with the development of tumorgenesis [38]. Therefore, the strong anti-tumor outcomes of the TLNT (oral, high)-treated group might contribute to the capacity of TLNTs to alleviate the inflammatory responses. These results collectively demonstrate that the TLNT (oral, high)-treated group exhibits comparable anti-tumor effects as the TLNT (i.v., high)-treated group.
Next, we performed a histological assay to evaluate the proliferation profiles of tumor cells. As shown in Fig. 4F, H&E staining images implied that the TLNT treatment caused obvious decreases in the tumor cell amounts in the tumor tissues sections. Moreover, TUNEL staining results indicated that obvious red fluorescence signals were shown in all groups receiving the treatment of TLNTs, particularly in the TLNT (i.v., high)-treated group and the TLNT (oral, high)-treated group, while there was the other way around in the control group (Additional file 1: Fig. S6). These findings reveal that the treatment of TLNTs with high dosages apparently has a stronger capability to cause the apoptosis and retard the growth of tumor cells regardless of the administration approaches (i.v. injection and oral administration).
To unravel the tumor inhibition mechanism of TLNTs, the transcriptome analysis of tumor tissues was conducted with assistance from the Majorbio Company. As presented in Fig. 5A, B, in comparison with the control group, 329 up-regulated genes and 203 down-regulated genes were identified in the TLNT (i.v., high)-treated group, and 367 up-regulated genes and 691 down-regulated genes were identified (fold change ≥ 2 and p < 0.05) in the TLNT (oral, high)-treated group, respectively. The Venn diagram (Fig. 5C) indicated that the control group shared the similar expression of 12,998 genes with the TLNT (i.v., high)-treated group and 12,606 genes with the TLNT (oral, high)-treated group. We also found that 538, 251 and 98 genes were exclusively expressed in the control group, the TLNT (i.v., high)-treated group and the TLNT (oral, high)-treated group, respectively. Interestingly, principal component analysis (PCA) revealed that genes were differentially expressed in the control group, the TLNT (i.v., high)-treated group and the TLNT (oral, high)-treated group, respectively (Additional file 1: Fig. S7). Our results also revealed that tumoricidal action genes related to anti-tumor immune responses were up-regulated in tumor tissues in the TLNT-treated groups (Fig. 5D). Other mark genes included NOS2, CCL4, CXCL9, and IL-10. The NOS2-encoded inducible nitric oxide synthase (iNOS) that mediated the tumoricidal activity and produced high output nitric oxide were clearly upregulated in the TLNT-treated groups [39]. Likewise, the upregulated CCL4 and CXCL9 genes could suppress tumors by actively recruiting CD8+ T cells [40] and regulating immune cell migration, differentiation and activation, correspondingly [41], while IL-10 could exhibit anti-tumor activity by enhancing NK cell activity [42].
To further investigate the biological functions of differentially expressed genes (DEGs), we performed GO analysis by querying each DEG in tumors from different mouse groups against the GO database, leading to the top 20 GO enrichment terms of DEGs. It was found that the responses to regulate cell proliferation occupied the strongest enrichment degree, since it possessed the largest number and was also involved in the regulation of a cell cycle (Fig. 5E), indicating that TLNTs mainly exerted their anti-tumor function through the interferences of cell cycle and cell proliferation. As reported, polyphenols and flavonoids have growth-inhibiting effects on a variety of tumor cells [43], which is consistent with our results (Fig. 4B, C). The Kyoto encyclopedia of genes and genomes (KEGG) pathway analysis reflected that TLNTs exerted anti-proliferation and pro-apoptosis against breast tumor cells mainly through the cellular signal pathways related to the cytokine-cytokine receptor interaction, JAK-STAT and cell cycle (Fig. 5F).
Accumulating evidence demonstrates that the intestinal microbiota plays a critical role in the development, metastasis and treatment responses of various tumors [44]. Thus, we investigated whether TLNTs could affect the homeostasis of the intestinal microbiota. As shown in Fig. 6A, the α-diversity Simpson index reflected the improved diversity of the intestinal microbiota with the treatments of TLNTs (i.v., high and oral, high). Furthermore, principal coordinates analysis (PCoA) suggested that there were dramatic alterations in the microbiome of the tumor control group, in comparison with that of the healthy control group (Fig. 6B). Moreover, Venn diagrams showed that 55, 29, 19 and 17 unique operational taxonomic units (OTUs) were found in the healthy control group, the tumor control group, the TLNT (i.v., high)-treated group and the TLNT (oral, high)-treated group, respectively (Fig. 6C). It was also found that mice bearing breast tumors had increased OUT numbers in their intestinal microbiota compared with the healthy control group (Fig. 6D). To further determine how the TLNT treatments to influence gut microbiota, we analyzed their compositions at the genus level. Figure 6E showed increases in the abundance of the total fecal bacteria for all the TLNT-treated mice. At the phylum level, the statistically significant higher Bacteroidetes/Firmicutes ratios were detected in the TLNT-treated group, compared with the tumor control group (Fig. 6F).
Meanwhile, we found that the abundance of the typical beneficial bacteria and harmful bacteria varied. In particular, Alistipes, a potential pathogen, contributed to tumor pathogenesis [45], and its relative abundance greatly decreased in the TLNT (i.v., high)-treated group and the TLNT (oral, high)-treated group, relative to the tumor control group (Fig. 6G). It is documented that Oscillibacter has the capacity to elevate the levels of pro-inflammatory cytokines (e.g., IL-1β and IL-6), which contribute to the progression of breast tumors [46]. We found that Oscillibacter was enriched in the tumor control group, whereas their abundance was significantly decreased in the groups receiving the treatment of TLNTs (i.v., high) and TLNTs (oral, high) (Fig. 6H). Next, compared with the healthy control group, the tumor control group had the increased abundance of Desulfovibrio, which was reported to be a cancer-risk genus and could magnify the inflammation and cardiometabolic risks for patients with breast cancer [47] (Fig. 6I). Additionally, it was discovered that the treatments of TLNTs (i.v., high and oral, high) obviously increased the abundance of the beneficial bacteria, Alloprevotella (Fig. 6J) [48]. The above findings clearly imply that TLNTs are liable to modulate the intestinal microbiota by increasing the abundance of beneficial bacteria and decreasing the abundance of harmful bacteria, which are consistent with our previous study about the application of tea flower-derived NTs in the treatment of breast cancer via oral route [9]. Interestingly, the treatment of ATBs was found to significantly attenuate the tumor retardation effects of orally administered TLNTs (Fig. 4). It was reported that intestinal microbiome was essential for the activation of anti-tumor immune responses [49]. The weak anti-tumor activity of TLNTs (oral, high, ATB) implies that the abundance and diversity of intestinal microbiota are crucial for potentiating the anti-tumor immunity and further facilitate TLNTs to exert their anti-tumor activity after oral administration.
In vivo biosafety evaluation of TLTNs
In vivo biosafety of nanomedicines is an essential prerequisite for their medical translation [50] and thus was evaluated. After 4 doses of TLNTs, mouse body weights and organ indexes were recorded. The results indicated that no significant difference was found between the healthy control group and the TLNT (oral, high)-treated groups during the entire experimental period. Strikingly, we visualized that the TLNT (i.v., high)-treated group showed a decrease in body weights (Fig. 7A) and increases in liver indexes and spleen indexes (Fig. 7B), compared with the healthy control group. Although the treatment of TLNTs (i.v., high) did not influence the secreted amounts of pro-inflammatory cytokines (IL-6 and IL-12), this treatment approach resulted in the increased concentrations of the typical inflammatory cytokines (TNF-α) and complement 3, which were obviously higher than those of the healthy control group (Fig. 7C, D).
Nevertheless, we investigated the blood compatibility of TLNTs by a hemolysis assay. It was observed that these TLNTs could not result in hemolysis (Additional file 1: Fig. S8). Moreover, we evaluated the potential toxicity of TLNTs against the liver and spleen by several key indicators, alanine aminotransferase (ALT) and aspartate aminotransferase (AST), two primary indicators of the liver function, as well as urea nitrogen (BUN) and creatinine (CRE), serological indices of the kidney function. We found that the liver and the kidney functioned normally after oral administration of TLNTs, while in the TLNT (i.v., high)-treated group, their hepatorenal function were seen abnormal (Fig. 7E–H). These results demonstrate that TLNTs are, although weak at inducing hemolysis, capable of stimulating the immune system, producing hepatorenal toxicity and altering the hemogram after i.v. injection, suggesting that i.v. route might not be an appropriate approach for administrating plant-derived NTs. However, the systemic cytotoxicities of intravenously injected plant-derived NTs have not been fully elucidated in previous reports [51,52,53,54]. In the context of the TLNT (oral, high)-treated group, they showed no obvious variations in terms of body weights, organ indexes, hepatorenal toxicity and complement system activation compared with the healthy control group. These results demonstrate that the plant-derived NTs can be developed as a safe nanoplatform for the treatment of breast cacer via oral administration, rather than i.v. injection.