Characterization of the CA-CDs

The CA solution was transparent and showed little fluorescence under 365 nm UV lamp, while the prepared CA-CDs solution was pale yellow under light and showed yellow-green fluorescence under 365 nm UV irradiation (Fig. S1). Transmission electron microscopy (TEM) image showed that the CA-CDs had relative uniform size and good dispersion, with an average size of 9.70 ± 0.10 nm (Fig. 1A and B). The CA-CDs exhibited excitation-dependent characteristics (Fig. 1C). At the optimal excitation wavelength of ~ 360 nm, the emission wavelength of the CA-CDs was ~ 420 nm. The fluorescence decay curve of the CA-CDs was fitted according to the second-order decay index function (Fig. 1D). Two fluorescence lifetimes (τ) were obtained, indicating that the CA-CDs had two fluorescence centers (Table S1). The average lifetime of the CA-CDs was ~ 3.24 nm and the quantum yield was ~ 0.8%.

The CA solution had three absorption peaks at ~ 216 nm, ~ 291 nm, and ~ 314 nm, respectively (Fig. 1E), indicating the presence of benzene ring and conjugated carbonyl group in the compound. The wide absorption at ~ 258 nm and the absorption peak at ~ 280 nm of the CA-CDs solution revealed the presence of π–π*, n–π* electronic transitions, indicating the presence of C = C, C = O, and C-O on the CA-CDs surface [35]. Fourier transform infrared spectroscopy (FTIR) showed that CA had absorption peaks at 3417, 3228, 1639, 1524, 1293, 1210 and 1117 cm− 1, which were related to the expansion and contraction vibrations of the functional groups O-H, C-H, C = O, C = C and C-O (Fig. 1F). While for the CA-CDs, the absorption peak at ~ 3380 cm− 1 belonged to the expansion vibration of O-H bond, the absorption peak at ~ 2920 cm− 1 belonged to the presence of C-H bond, the absorption at ~ 1600 cm− 1 belonged to the expansion vibration of C = O, the absorption at ~ 1500 cm− 1 was the expansion vibration of benzene ring skeleton, and the absorption at ~ 1196 cm− 1 was the expansion vibration of C-O. XPS full-spectra scan showed that both CA and the CA-CDs were composed of C and O elements (Fig. S2 and Fig. 1G). The high-resolution C1s spectra of the CA-CDs had distinct peaks at 284.8 eV, 286.3 eV, and 288.5 eV, which related to the presence of C = C/C-C, C-O, and C = O, respectively (Fig. 1H) [36]. The characteristic peaks at ~ 531.9 eV and ~ 533.5 eV in the high-resolution O1s spectra correspond to the C-O and C = O groups, respectively (Fig. 1I). The atomic ratios of C and O in CA were ~ 71.65% and ~ 28.35%. The corresponding atomic ratios of the CA-CDs were ~ 78.14% and ~ 21.86%. It suggested that hydrothermal synthesis caused the partial chemical bonds of CA to break and reassemble during the fabrication of the CA-CDs. The change in the atomic ratio was the result of the formation of CDs through carbonization [37]. The above results indicated that the CA-CDs retained part of the structure of CA, and the surface was rich in -OH and -COOH groups.

Fig. 1

Structural characterization of the CA-CDs. (A) (B) TEM image and size distribution of the CA-CDs, the scale bar was 50 nm. (C) Fluorescence spectra of the CA-CDs at different excitation wavelengths of 320 ~ 520 nm. (D) Fluorescence attenuation curve of the CA-CDs. (E) UV-Vis absorption spectra of CA and the CA-CDs (0.2 mg/mL) in aqueous solution. (F) FT-IR spectra of CA and the CA-CDs. (G) High-resolution XPS spectra of the CA-CDs: C1s spectra (H) and O1s spectra (I)

The stability of nanomedicine is an important indicator to measure whether it can be applied. First, the stability of the CA-CDs in a wide concentration range of NaCl (0–1.0 mol/L) and KCl (0–1.0 mol/L) was tested. It can be seen that the increase in NaCl and KCl concentrations hardly affected the fluorescence intensity of the CA-CDs (Fig. S3A and S3B). The results showed that the CA-CDs had good salt resistance, and their structure and properties would be less affected in organisms (the concentration of normal salts in normal body fluids is ~ 0.15 mol/L). The CA-CDs were irradiated at 365 nm ultraviolet for 180 min. The fluorescence intensity of the CA-CDs also changed little, implying that the CA-CDs had good photobleaching resistance (Fig. S3C). In addition, considering the use of the CA-CDs is as a drug, the stability had been also tested in the phosphate-buffered saline, RPMI 1640 media, fetal bovine serum and penictomycin mixture (Fig. S4). There were some little changes of the fluorescence properties of the CA-CDs. It suggested that the CA-CDs were stable in complicated bioenvironments.

In vitro inhibition effect of the CA-CDs on BCPAP cells

Based on the above good physical and chemical properties of the CA-CDs, the biological activity of the CA-CDs in vitro was studied by Cell-Counting-Kit-8 method (CCK-8). After 48 h of co-incubation, low-dose CA-CDs showed a significant inhibitory effect on the proliferation of the papillary thyroid cancer cells BCPAP (Fig. 2A). As the dosing concentration increased, the cell viability decreased obviously. At concentrations of 12 µg/mL and 16 µg/mL of the CA-CDs, inhibition rate of the BCPAP cells reached ~ 73% and ~ 79%, respectively. In contrast, CA showed little effect on the cell viability of the BCPAP cells (Fig. 2B). As a natural herbal medicine, CA should have many biological activities. However, due to the poor water solubility and low cell membrane permeability of CA molecules, the bioavailability and medicinal value of CA molecules have been seriously hindered. The above experiments have demonstrated that the CA-CDs had certain hydrophilic groups on the surface, could be uniformly dispersed in water, and retained part of the structural of CA. Thus, the CA-CDs may possess certain biological activity. When the CA-CDs were co-incubated with the BCPAP cells, they can be engulfed through the cells in large quantities owing to their nanometer size and surface functional groups, and thereby exerted efficient drug effects (Fig. 2C) [7, 8]. In addition, CCK-8 analysis demonstrated low biotoxicity of the CA-CDs on normal thyroid cells Nthy-ori-3, hepatic normal cells LO2, and renal normal cells HEK-293T under the same conditions (Fig. 2D, E and F). These results suggested that the CA-CDs could be considered as a potential nanodrug against papillary thyroid carcinoma.

Fig. 2

Effects of the CA-CDs on viability, cycle, and apoptosis of the BCPAP cells. (A) Effect of the CA-CDs on the viability of the BCPAP cells. (B) Effect of CA on the BCPAP cell viability. (C) Statistic diagram comparing the effects of CA and the CA-CDs on the BCPAP cell activity. Effect of the CA-CDs on the cell activity of Nthy-ori-3 (D), LO2 (E) and HEK-293T (F). (G) .Flow cytometry cycle diagram of the BCPAP cells after 48 h treatment with the CA-CDs (0 µg/mL, 4 µg/mL, 8 µg/mL, 12 µg/mL and 16 µg/mL). (H) Apoptosis of the BCPAP cells after 48 h treatment with the CA-CDs (0 µg/mL, 4 µg/mL, 8 µg/mL, 12 µg/mL and 16 µg/mL). *p < 0.05; **p < 0.01; ***p < 0.001 compared to Control

The cell cycle is closely related to cell proliferation. Flow cytometry showed that most of the BCPAP cells of Contral group (CA-CDs 0 µg/mL) were in the G0/G1 phase, with fewer cells in the S phase and the smallest proportion of cells in the G2/M phase. The proportion of S phase of the BCPAP cells increased after 48 h treated with different concentrations of the CA-CDs (4 µg/mL, 8 µg/mL, 12 µg/mL and 16 µg/mL) (Table S2). In particular, after treatment with the CA-CDs at concentrations of 12 µg/mL and 16 µg/mL, the proportion of the BCPAP cells in S phase was as high as ~ 46% and ~ 43% (p < 0.001, p < 0.01), indicating that the CA-CDs could arrest the BCPAP cells in S phase (Fig. 2G). After 48 h of incubation with the CA-CDs (4 µg/mL, 8 µg/mL, 12 µg/mL and 16 µg/mL), the apoptosis of the BCPAP cells was detected by a flow cytometry. The proportion of early apoptosis and late apoptosis of the BCPAP cells in the Contral group (CA-CDs 0 µg/mL) was ~ 1.07% and ~ 1.58%, respectively. The proportion of early apoptosis and late apoptosis of the BCPAP cells in the maximum administration treatment group (CA-CDs 16 µg/mL) was ~ 5.31% and ~ 30.16%, respectively (Table S3), indicating the significantly increase of apoptosis rate (p < 0.05, p < 0.001). The results showed that the low-dose CA-CDs could significantly induce the BCPAP apoptosis, and the apoptosis rate was positively correlated with the dose of the CA-CDs (Fig. 2H).

In order to further explore the inhibitory effect of the CA-CDs on the BCPAP cells, the entry of the CA-CDs into the BCPAP cells at different incubation times was monitored by fluorescence microscopy (Fig. 3). In brightfield, it could be seen that with the extension of incubation time, the BCPAP cells gradually shrank and became rounded, and the cell morphology was disrupted. It can be seen that the CA-CDs can effectively enter the cells and emitted green and red fluorescence under different light irradiation. After co-incubation for 0 ~ 24 h, the fluorescence intensity of the system gradually increased, indicating that the number of intracellular CA-CDs increased with time. The fluorescence intensity of the system was the highest at 24 h, indicating that the BCPAP cells had the highest uptake of the CA-CDs. From 24 to 48 h, the fluorescence intensity of the system decreased, and the fluorescence was almost undetectable at 48 h. This may be due to the structural destruction of the intracellular CA-CDs by reacting with the cell matrix, accompanied with the gradual shrinkage and death of the BCPAP cells. The change of fluorescence intensity in the incubation system suggested that the inhibitory effect of the CA-CDs on the BCPAP cells was time-dependent. In addition, the uptake of the CA-CDs by the LO2 cells and the Nthy-ori-3 cells at different times was also observed (Fig. S5 and S6). The LO2 cells and the Nthy-ori-3 cells exhibited similar uptake properties of the CA-CDs, compared with the BCPAP cells. It implies that the CA-CDs can also be taken up by the normal cells, but performed little inhibition on these cells.

Fig. 3

Distribution of the CA-CDs in BCPAP cells with time. Fluorescence microscopy images of the CA-CDs (30 µg/mL) and the BCPCP cells under brightfield and fluorescence fields (fluorescence excitation wavelengths 488 nm, 405 nm and 516 nm) after co-incubation for different times (0 h, 12 h, 24 h and 48 h). The scale bars were 100 μm

Reactive oxygen species (ROS) are also an important factor in cancer treatment research. Elevated ROS levels can induce apoptosis through both endogenous and exogenous pathways. Here, the changes in ROS fluorescence intensity in the BCPAP cells after treated with different concentrations of the CA-CDs (4, 8, 12, and 16 µg/mL) were observed by fluorescence microscopy (Fig. 4A). Compared with the control group (0 µg/mL CA-CDs), the average intracellular ROS red fluorescence intensity was much stronger when the concentration of the CA-CDs reached 12 µg/mL (p < 0.01). When the concentration of the CA-CDs reached 16 µg/mL, the intracellular ROS red fluorescence was further significantly enhanced. The images in brightfield also showed that with the increase of the CA-CDs administered concentration, the morphology of the BCPAP cells shrunk and rounded, and the number of the cells decreased obviously (Fig. S7). These results showed that the CA-CDs could inhibit the BCPAP cells by triggering a large amount of ROS production in cells. After 48 h of treatment with the CA-CDs (0, 4, 8, 12, and 16 µg/mL), the mitochondrial membrane potential (MMP) changes of the BCPAP cells were observed by fluorescence microscopy (Fig. S8). It was found that the MMP orange-red fluorescence intensity in the BCPAP cells decreased significantly after 16 µg/mL CA-CDs treatment.

The transwell assay was used to evaluate the effects of the CA-CDs on the migration and invasion of the BCPAP cells (Fig. 4B). Compared with the control group (CA-CDs, 0 µg/mL), the migration and invasion ability of the BCPAP cells was significantly reduced after treatment with different concentrations of the CA-CDs (4 µg/mL, 8 µg/mL, 12 µg/mL and 16 µg/mL). When the concentration of the CA-CDs reached 12 and 16 µg/mL, the migration and invasion ability of the BCPAP cells was inhibited by ~ 65% and ~ 94%, respectively (p < 0.001). It suggested the CA-CDs could significantly and effectively weaken the migration and invasion ability of the BCPAP cells, which might reduce the possibility of cancer cell metastasis and spread.

Fig. 4

(A) After 48 h of treatment with different concentrations of the CA-CDs, the fluorescence photos of ROS in the BCPAP cells and the corresponding mean fluorescence intensity change quantitative column chart. The scale bars were 200 μm. (B) Migration and invasion dyed images of the BCPAP cells, and the corresponding statistic diagram of the number of migrating cells after 48 h treatment with different concentrations of the CA-CDs. The scale bars were 500 μm. *p < 0.05; **p < 0.01; ***p < 0.001 compared to Control

Effect of the CA-CDs on the expression of MAPK pathway-related proteins in the BCPAP cells

After 48 h treatment of the BCPAP cells with the CA-CDs (16 µg/mL), transcriptomics analysis was performed between the CA-CDs group (CA-CDs 16 µg/mL) and the Control group (CA-CDs, 0 µg/mL). A volcano plot of transcriptome differential gene screening results between the CA-CDs group (n = 5) and the Control group (n = 5) were plotted (Fig. 5A). Compared with the Control group, 7596 differential genes were screened out in the CA-CDs group. Among them, 2783 genes were up-regulated and 4813 genes were down-regulated. The two groups of differential genes (SOS1, KRAS, BRAF, MAPK1, ELK1, GCNT3 and HSPA6) were randomly selected for correlation analysis. It was found that the data of the two groups (n = 5) were reproducible and there were differences between the groups, which could be used for subsequent analysis (Fig. 5B). For the aim to test the data reliability of the transcriptome, the randomly selected differential genes SOS1, KRAS, BRAF, MAPK1 and ELK1 in Fig. 5B were verified by qRT-PCR (Fig. 5D). It can be seen that the down-regulation trend of these five genes was obvious, which was consistent with the trend of transcriptome sequencing analysis in Fig. 5B. KEGG enrichment analysis was performed on all differential genes, and the top 20 KEGG pathways with the highest enrichment were selected for plotting (Fig. 5C). The signaling pathways with a high degree of differential gene enrichment were Human papillomavirus infection, mitogen-activated protein kinase (MAPK) signaling pathway, proteoglycans in cancer and human immunodeficiency virus 1 infection.

The WB was used to analyze the regulatory effect of the CA-CDs on the MAPK classical signaling pathway in the BCPAP cells (Fig. 5E and F). Compared with the Control group, there was little expression content change in the BRAF, MEK1 and ERK1/2 proteins in the BCPAP cells, while the expression contents of KRAS, p-BRAF, p-MEK1 and p-ERK1/2 proteins were significantly reduced. It is speculated that the CA-CDs could simultaneously inhibit the proliferation of the BCPAP cells and promote the apoptosis of the BCPAP cells by inhibiting the expression of the continuous RAS–RAF–MEK–ERK proteins in the MAPK classical signaling pathway.

Fig. 5

Anti-tumor mechanism experiments of the CA-CDs treated the BCPAP cells after 48 h. (A) A volcano plot of transcriptome differential gene screening results of CA-CDs group and Control group (n = 5). Green dots represent down-regulated genes, and red dots represent up-regulated genes. (B) Data correlation analysis heat map of CA-CDs group and Control group (n = 5). (C) Bubble map of the top 20 KEGG pathways with the highest differential gene enrichment. (D) The relative mRNA levels of SOS1, KRAS, BRAF, MAPK1 and ELK1 genes in BCPAP cells in CA-CDs group and Control group, detected by qRT-PCR. (E) WB images of MAPK classical pathway-related proteins of BCPAP cells in CA-CDs group and Control group. The Control group was on the left and the CA-CDs group was on the right. (F) The statistic histogram of relative protein expression

Inhibitory effect of the CA-CDs on poorly differentiated human papillary thyroid tumors in vivo

BCPAP cells were injected into the subcutaneous skin of BALB/c nude mice, and a xenogeneic tumor-bearing model was established. The tumor-bearing nude mice were randomly selected to three groups (n = 5) and injected through the tail vein: the normal saline (90 mg/kg, 5 times per week) as the Model group; the CA-CDs (2 mg/kg, 5 times per week) as the CA-CDs group; the clinical anti-thyroid tumor drug cyclophosphamide (CTX, 60 mg/kg, 3 times per week) was used as the CTX group. Normal mice were injected with normal saline (90 mg/kg, 5 times per week) in the tail vein as the Control group (n = 5). Normal mice were injected with the CA-CDs (2 mg/kg, 5 times per week) in the tail vein as the Control + CA-CDs group (n = 5). Under the same feeding condition, the experimental cycle was set 14 days, within which observed the indicators of the mice. On the 14th day, the subcutaneous tumors of nude mice were removed for detection and analysis. Compared with the Model group, the tumor volume and weight of mice in the CTX group and the CA-CDs group were significantly reduced (Fig. 6A, B and C). The tumor volume of mice in the CA-CDs group decreased more, compared with that in the CTX group (Fig. 6C). The results indicated that low-dose CA-CDs had a good inhibitory effect on human papillary thyroid tumors, and the inhibitory effect was better than that of the positive control CTX.

The body weight changes of mice in the Model group, Control group, CTX group, CA-CDs group, and Control + CA-CDs group were recorded to evaluate the safety of CA-CDs. The mice in the Control group, Control + CA-CDs group, Model group and CA-CDs group were in good condition, and their body weight was steadily increasing (Fig. 6D). The mice in the CTX group became malaise after injection with the drug, and their body weight was clearly lower than that of the mice in other groups. The body weight of mice in the Control + CA-CDs group was higher than that in the Control group. This also indicated that the CA-CDs had low toxicity and side effects to organism, and had good biological safety. On the 14th day, the major organs (heart, liver, spleen, lung, and kidney) of mice in each group were collected for histological analysis to evaluate pathological differences (Fig. S9). Compared with the Control group, there was little organ damage and inflammatory damage in the main organs of the mice in the CA-CDs group and the Control + CA-CDs group. It further demonstrated that the CA-CDs may have good biocompatibility in vivo.

H&E staining was used to observe the damage of human papillary thyroid tumor tissue in mice in each group (Fig. 6E). H&E staining showed that the tumor cells of mice in the Model group were tightly arranged and there were fewer apoptotic cells. A large number of apoptotic cells appeared in the tumor tissues of mice in the CTX group and the CA-CDs group. The cells were loosely arranged, and the number of apoptotic cells was significantly higher than that of the Model group. These results indicated that high-dose CTX and low-dose CA-CDs could promote tumor tissue necrosis. Tunel method is widely used to detect apoptosis in tissue sections. Tunel staining showed that the tumor cells of mice in the Model group were tightly arranged and blue, and apoptotic cells with brownish-yellow nuclei were almost absent. The tumor cells of the mice in the CA-CDs group were loosely arranged, the cell volume was small, and most of the cells were brown. The tumor cells of the mice in the CTX group were not much different from those in the CA-CDs group (Fig. 6E). The expression of markers related to differentiation and proliferation (CDK4 protein and Ki-67 protein markers) in human papillary thyroid tumors in each group was observed by immunohistochemistry (IHC) (Fig. 6E). The results showed that the CDK4 and Ki-67 were highly expressed in the Model group, moderately expressed in the CTX group, and lightly expressed in the CA-CDs group. The above results suggested that low-dose CA-CDs could promote apoptosis and inhibit cell proliferation in human papillary thyroid tumor tissues in vivo.

Fig. 6

Inhibitory effect of the CA-CDs on human papillary thyroid tumors in vivo. (A) Tumor images of the mice in the Model group, CTX group and CA-CDs group. (B) Statistic column chart of tumor weight of the mice in Model group, CTX group and CA-CDs group mice on the 14th day. (C) Statistic chart of tumor volume of the mice in Model group, CTX group and CA-CDs group mice from 1 to 14 days. (D) The calculated statistic chart of body statistical weight of mice in the Model group, Control group, CTX group, CA-CDs group and Control + CA-CDs group from 1 to 14 days. (E) H&E staining, Tunel staining, and immunohistochemical staining of tumor marker proteins (Ki-67 and CDK4) of mouse tumors in Model group, CTX group and CA-CDs group. The scale bars were 500 μm. *p < 0.05; **p < 0.01; ***p < 0.001 compared to Model