Saturday, October 1, 2022
HomeNanotechnologyBoosting nutrient starvation-dominated cancer therapy through curcumin-augmented mitochondrial Ca2+ overload and obatoclax-mediated...

Boosting nutrient starvation-dominated cancer therapy through curcumin-augmented mitochondrial Ca2+ overload and obatoclax-mediated autophagy inhibition as supported by a novel nano-modulator GO-Alg@CaP/CO | Journal of Nanobiotechnology

[ad_1]

Characterizations of synthesized GO-Alg@CaP/CO

At the acidic conditions, GO was capable of immobilizing on the abundant carboxyl groups of Alg via a facile EDC/NHS-induced covalent reaction. First, BCA assays was established to access the successfully immobilization of GO. Figure 1G provided the standard linear calibration curve and GO concentration can be calculated from the following formula (Eq. 1) (R2 = 0.9956):

$${\text{C}} \left( {{\text{GO}}} \right)\left( {\frac{{{\text{mg}}}}{{{\text{mL}}}}} \right) = 0.8177 \times {\text{OD}} – 0.1454.$$

(1)

Fig. 1
figure 1

Particle sizes of A CaP, B Alg@CaP and C GO-Alg@CaP/CO. These insets showed the images of the corresponding aqueous solution. Transmission electron microscopy (TEM) of D CaP, E Alg@CaP and F GO-Alg@CaP/CO (scale bars were 200 nm). G The standard content curves of GO based on BCA assays. H Variation of glucose concentration and I pH value in glucose solution (1 mg/mL) after the addition of GO, GO-Alg and GO-Alg@CaP/CO with the identical GO concentration (1 mg/mL)

After diluted 10 times, the OD (Optical Density) of GO-Alg was 0.28 and GO concentration was then identified with the above formula as 0.084 mg/mL. Eventually, the immobilization efficiency (IE) of GO was calculated as 84% with the following formula (Eq. 2):

$${\text{IE}}\left( {{\text{GO}}} \right)\left( \% \right) = \frac{{{\text{GO}}\;{\text{quantified}}\;{\text{in}}\;{\text{the}}\;{\text{GO-Alg}}\;{\text{solution}} \left( {{\text{mg}}} \right)}}{{{\text{Total}}\;{\text{GO}}\;{\text{added}} \left( {{\text{mg}}} \right)}}.$$

(2)

The results of the correlational analysis were echoed that GO could be easily and effectively immobilized on Alg. To carry this idea one step further, amide N–H stretch at 3700–3500 cm−1, amide C=O stretch at 1690–1630 cm−1, and amide I band and II band at 1500–1560 cm−1 were shown in GO-Alg at FT-IR spectrum (Additional file 1: Fig. S1).

The next section of the results was concerned with GO-Alg@CaP. It can be seen from Additional file 1: Table S1, there were clear trends of increasing in the average particle size and decreasing in the zeta potential of GO-Alg@CaP with different ratios (GO-Alg/CaP) ranged from 0/10 to 50/10. Overall, an implication of this was the possibility that GO-Alg could be successfully involved with CaP due to the affinity between Ca2+ and Alg. This view was further confirmed in the GO-Alg@CaP/CO spectra (Additional file 1: Fig. S1), the bands ranged from 1400–1700 cm−1 and 3000–3500 cm−1 were attributed to GO-Alg. The bands appearing in the range of 800–1200 cm−1 could be ascribed to –PO43− bonds of CaP. Similarly, the characteristic peaks of Alg were also shown in the Alg-CaP spectra (Additional file 1: Fig. S1). Additional file 1: Table S1 also presented a random polydispersity index (PDI) which should be considered. The sizes, zeta potentials and PDI were all profoundly large when the ratio was over 5/10 due to the high viscosity of the Alg, which was not neglected for Alg to produce larger emulsion droplets leading to larger particles. Eventually, 5/10 (GO-Alg/CaP) was chosen for he further experiments with an ideal property. Similarly, the final GO concentration of GO-Alg@CaP was also confirmed with BCA assay as 80% IE.

Prior studies had noted the high drug loading efficiency of CaP. GO-Alg@CaP also endowed a large capacity for the drug loading which was not restricted in the GO-Alg coating. Moreover, the percentages of Cur and Obatoclax entrapped into GO-Alg@CaP were both almost 90%. A probable explanation was that GO-Alg interaction with CaP provided an excellent natural polysaccharide coating that also elicited a special adhesive affinity to drugs. And Additional file 1: Table S2 had set out the basic characterizations of the final nano-formulations, including that GO-Alg@CaP/CO had a size distribution of 204.1 nm with a PDI of 0.227. Even though larger particle size of GO-Alg@CaP could be also obtained from Fig. 1A–F in accordance with previous results. On the contrary, for the morphology of GO-Alg@CaP/CO, it was almost certain from Fig. 1D, E that GO-Alg could partially control the irregular aggregation of CaP, suggestive of the regular spherical shape. We determined the stability of GO-Alg@CaP/CO under biological conditions during 5 days as following. As exhibited in Additional file 1: Fig. S3, the size change of GO-Alg@Cap/CO was almost negligible at physiological conditions of pH 7.4 for 5 days, suggestive of good stability for the intravenous injection. Moreover, Additional file 1: Fig S4A, B showed that only 35% of Obatoclax and 28% Curcumin were released from GO-Alg@CaP/CO at the pH 7.4 over 240 h, while nearly 50% of Obatoclax and Curcumin could be released at pH 5.2, indicating that the pH-responsive dissociation of CaP which was in accordance with our previous studies. Taken together, these results held that GO-Alg@CaP/CO would be promising nanocomplexes for the following application.

Catalytic ability measurement of synthesized GO-Alg@CaP/CO

The glucose content was established by a DNS method through a linear fit curve between glucose concentration and UV–Vis absorption (540 nm) (Additional file 1: Fig. S2). By incubating glucose solution with different preparations containing the identical GO concentration, as shown in Fig. 1H, I, the glucose content and pH value demonstrated a clear trend of decreasing with time extended, revealing the successful conversion of glucose into gluconic acid together with pH value dropped from ~ 7.4 to ~ 3.0 and glucose decrease from 1 to ~ 0.4 mg/mL, respectively. Another important finding was that GO-Alg and GO-Alg@CaP/CO exhibited similar catalytic efficacy but suggested weaker activity in comparison with GO, indicating that GO interaction with Alg would alter the catalytic activity of GO. Therefore, incomplete reaction might be a major factor because the immobilized GO could not completely be involved with glucose. In this view, these changes were reasonable without significant difference and the activities of GO in GO-Alg and GO-Alg@CaP/CO were also well preserved with quickly decreased pH ranged from 7.4 to 3.2 and gradually dropped glucose from 1 to 0.48 mg/mL.

The cellular uptake of GO-Alg@CaP/C

As elicited in Additional file 1: Figs. S5, S6, the cells treated with Cur still showed weaker green fluorescence compared with GO-Alg@CaP/C for different time points. Moreover, the co-localized green (GO-Alg@CaP/C) with red (lysosomes) fluorescent after 6 h noted that the cellular uptake of GO-Alg@CaP/C was attributed to endocytosis. It seemed possible that nano-complexes could improve the solubility of Cur which might facilitated the final cellular uptake, indicating that nano-complexes GO-Alg@CaP/C endowed increased cellular internalization efficiency.

In vitro Ca2+ production by GO-Alg@CaP/CO

Prior to detecting Ca2+ in mitochondria, free Ca2+ in 4T1 cells was stained with Fluo-3 AM to be visualized by CLSM. It was because there had been an apparent increase of green fluorescence intensity in all nano-complexes with CaP. Figure 2 revealed that CaP could give rise to high free Ca2+ concentration in 4T1 cells compared to free drugs and ctr groups. As expected, CaP could be collapsed in cells and eventually lead to producing intracellular overmuch free ionic calcium. Based on the above analysis, the intramitochondrial Ca2+ concentration alteration was then visualized by CLSM with the help of a Rhod-2 AM staining assay. In particular, it was somewhat surprising that the observed difference between GO-Alg@CaP/CO group (without Cur) and GO-Alg@CaP group (with Cur) was significant. Although GO-Alg@CaP/CO group (without Cur) and GO-Alg@CaP group (with Cur) endowed with similar intracellular ionic calcium in the Fig. 2B, C, there was a significant difference in above two groups by flow cytometry in Fig. 2D, E, indicating that Cur might attribute to inhibit expelling of Ca2+ and induce the mitochondrial Ca2+ accumulation. It was also verified in the relevant studies that Cur might attribute to inhibit expelling of Ca2+ and induce the mitochondrial Ca2+ accumulation24. To further illustrate these findings, the co-localization of Ca2+ and mitochondria was clearly demonstrated in Fig. 3. Notwithstanding the high red fluorescence intensity, no significant co-localization of Ca2+ and mitochondria in GO-Alg@CaP group was observed which was consistent with above findings. Furthermore, there were distinctly overlaps between green fluorescent (mitochondria) and red fluorescent (Ca2+) in both GO-Alg@CaP/C and GO-Alg@CaP+Cur, highlighting a noticeable increase of free Ca2+ in mitochondria induced by Cur.

Fig. 2
figure 2

Disruption of intramitochondrial Ca2+ homeostasis A intracellular Ca2+ production and mitochondrial Ca2+ concentrations of 4T1 cells treated with different preparations for 6 h. Scale bars were 50 μm. BE Flow cytometry results of 4T1 cells incubated with GO-Alg@CaP and GO-Alg@CaP/C. Data are mean ± SD, n = 3; *P < 0.05, **P < 0.01 and ***P < 0.001

Fig. 3
figure 3

Biodistribution of Ca2+ in mitochondria in 4T1 cells after incubation with different preparations for 6 h. The nucleus and mitochondria were stained with Hoechst 33342 and Mito-Green, respectively. The scale bar was 10 μm

Mitochondrial disfunction induced by GO-Alg@CaP/CO

To carry these results one step further, JC-1 staining was used to visualize mitochondrial membrane potentials (MMP). High MMP could be associated with J-aggregates shown as red fluorescence. On the contrary, low MMP would represent a monomer with green fluorescence. To assess MMP, the relative levels of green and red fluorescence intensities were compared in different preparations. As depicted in Additional file 1: Fig. S7, different nano-complexes exhibited lower red fluorescence than free drugs and ctr groups. Moreover, GO-Alg@CaP/CO group pervaded the remarkable increased green fluorescence suggesting the lowest MMP, which was similar with the analysis of free Ca2+ in mitochondria.

Autophagy inhibition evaluation of GO-Alg@CaP/CO

The complex correlation of autophagy and energy deficiency encouraged us to further investigate autophagic responses after different treatments. As a typical autophagosomal biomarker, both LC3 and P62 were widely considered to reveal the autophagy level in cells. The relationship of LC3-II and autophagy was complex and intertwined, it can be exclusively determined that the autophagy level was inversely related to P62 on the basis of the improved level of LC3-II. The observed increase of LC3 II/LC3 I and reduction of P62 in GO and GO-Alg@CaP groups were visualized in Fig. 4 and the repeated experiments were shown in Additional file 2. It can therefore be assumed that the high level of autophagy was stimulated in starvation therapy. Moreover, Alg@CaP and Alg@CaP/C groups also pointed out slight accumulation of autophagosomes together with the upregulation of LC3 II. It was possible to hypothesize that harsh conditions including mitochondrial Ca2+ overload would improve autophagy level as well. More importantly, very little autophagosomes were found in these groups (Fig. 4D), which was in agreement with above discussion. Comparatively, significant increase of LC3 II/LC3 I and P62 was also observed in Obatoclax treated groups (Fig. 4A–C), characteristic of the lower autophagy level. As shown in Fig. 4D, the increase in the quantity of autophagosomes were both observed in GO-Alg@CaP group and GO-Alg@CaP/CO group and the latter exhibited the most autophagosomes absolutely. Taken together, above results further revealed that GO-Alg@CaP enhanced the accumulation of autophagosomes by the activation of autophagic flux instead of blocking it. On the contrary, GO-Alg@CaP/CO could cause quantity of autophagosomes retention by blocking lysosomal degradation. These results were consistent with our previous studies and suggested that GO-Alg@CaP/CO could suppress the existing protective mechanism such as autophagy which would compensate for the energy deficiency in the harsh conditions.

Fig. 4
figure 4

A P62, LC3-I, and LC3-II expression in 4T1 cells with a western-blotting analysis under different groups. Matching gray-scale analysis of B LC3-II/LC3-I and C P62. D TEM images of autophagosome. Red arrows and circles implied the autophagosome. Data are mean ± SD, n = 3; *P < 0.05, **P < 0.01, and ***P < 0.001 vs control. The scale bar was 5 μm

In vitro cytotoxicity and apoptosis of GO-Alg@CaP/CO

Figure 5A first pointed out that blank carriers (Alg@CaP) were almost non-toxic with around 90–100% cell viability. Thereafter, the concentration gradient of each and every drug was set up to compare the efficacy differentiation between free drugs and nano-complexes. In general, the negative correlation between the cell viability and the drug concentration were apparently presented in Fig. 5B–D. A comparison of the free drugs and nano-complexes revealed striking discussions as following. Cell viability of GO-Alg@CaP reported significantly more than GO at GO concentration of 0.02 μg/mL (Fig. 5B) which was consistent with previous catalytic ability analysis. It was because the GO-Alg@CaP endowed lower catalytic ability to result in much less cell death than GO. From the graph in Fig. 5B, both GO and GO-Alg@CaP would elicit a sharp lethality more than 85% at the desired concentration of GO (0.04 μg/mL), revealing that the glucose of cells was almost depleted by GO inducing cytotoxicity. In order to comprehensively assess the boosting nutrient starvation efficacy, lower GO concentration (0.01 μg/mL, 20% inhibitory concentration) was considered to eliminate the sole contribution to cytotoxicity by GO. Compared with Cur and Alg@CaP/C, the well efficacy of Alg@CaP/C was demonstrated (Fig. 5C) which was also in accordance with above better cellular uptake. Similarly, 5 μg/mL of Cur (20% inhibitory concentration) was selected for the following experiments. As for Obatoclax and Alg@CaP/O, negligible cytotoxicity with 0.01 μg/mL was found (Fig. 5C) which could be identified for further evaluation. For further emphasizing the augmented efficiency of GO-Alg@CaP/CO, the evaluation of cytotoxicity was finally established for different treatments. The observed highest lethality in GO-Alg@CaP/CO treated cells with 18% cell viability which could be attributed to the combined effects of several treatments (Fig. 5E). The differences between the sum of each treatment and GO-Alg@CaP/CO also suggested this view.

Fig. 5
figure 5

A Relative viability of 4T1 cancer cells cultured with blank carriers (Alg@CaP) at different concentrations. Relative viability of 4T1 cancer cells cultured with different treatments. B GO and GO-Alg@CaP, C Cur and Alg@CaP/C, and D obatoclax and Alg@CaP/O. E Relative viability of 4T1 cancer cells cultured with different preparations at desired concentrations for 24 h. F Intracellular ATP content analysis of 4T1 cancer cells cultured with different preparations at desired concentrations for 48 h. G Apoptosis assays of 4T1 cells in different groups with a flow cytometric analysis. Data are mean ± SD, n = 3; *P < 0.05, **P < 0.01 and ***P < 0.001 vs. control

The results obtained from Fig. 5G pointed out that each single treated cells’ late apoptosis was lower than ~ 5%. However, we found that about 31.06% of GO-Alg@CaP/CO treated cells died of late apoptosis which was highest upon all groups, indicating that this high efficacy was stemmed from complex synergistic effects upon substances such as GO, Cur, CaP and Obatoclax. In accordance with above cytotoxicity evaluation, the apoptosis of GO-Alg@CaP/CO and GO-Alg@CaP/C were also not only the summarization of each single treatment. Together these results provided important insights into the GO-Alg@CaP/CO for enhanced efficacy.

To clarify energy metabolism perturbations, the results obtained from the preliminary quantified intracellular ATP levels were summarized in Fig. 5F. Distinct decreases of ATP production were observed in GO, GO-Alg@CaP, while no significant change was observed in Alg@CaP group, demonstrating that blockade of glucose supply by GO was capable of exacerbating of energy deprivation. Also, Cur and Alg@CaP/C groups could lead to mitochondrial dysfunction through Ca2+ efflux inhibition via Cur which would cut off the energy sources from mitochondria consistent with the previous discussion, revealing a considerable decrease in the ATP level. The graph presented a slight fall in the ATP level at Obatoclax and Alg@CaP/O groups, which was stemmed from another energy deprivation source (autophagy inhibition) caused by Obatoclax. As a consequence of the trebling additive efficacy of energy depletes, it appeared that a sharp drop in the ATP level of GO-Alg@CP/CO group.

In vivo antitumor efficacy of GO-Alg@CaP/CO

The hemocompatibility was first investigated to confirm in vivo application of GO-Alg@CaP/CO. As shown in Additional file 1: Fig. S8, even high concentration of nano-complexes around 1000 μg/mL was endowed with a relative low hemolysis rate of 0.87% after 24 h incubated, indicating extraordinary hemocompatibility.

Substantial cellular uptake results about GO-Alg@CaP/CO led us to further explore in vivo distribution of nano-complexes. IR780 was loaded as a fluorescent marker in nano-complexes and then injected into tumor-bearing mice via tail vein. The repeated experiments were shown in Additional file 1: Fig. S17, and it was apparent that the accumulation of nano-complexes at tumor sites was observed. There was a clear trend of increasing of fluorescence intensity at tumor sites within 12 h but a slight decreasing after 24 h. In response to these results, the fluorescence intensities of organs and tumor were detected ex vivo (Fig. 6B, C). A clear high retention of GO-Alg@CaP/CO at tumor sites was noted, revealing their efficient passive EPR effect (enhanced permeability and retention effect) at tumor sites. Overall, this was an important criterion for the next therapeutic evaluation of GO-Alg@CaP/CO.

Fig. 6
figure 6

A In vivo fluorescence images of 4T1 tumor-bearing mice that were intravenously injected with Alg@CaP/I at different time points and B ex vivo fluorescence images of major organs and tumors at 24 h post injection. C Histogram analysis of relative average radiant efficiency of liver, lungs, and tumor. D TUNEL-stained tumor slices collected from 4T1 tumor-bearing mice after treatments for 15 days. The scale bar was 50 μm. E Blood analysis of mice at 15 days after various treatments. F Relative volume changes and G mice body weight changes in different groups. Data are mean ± SD, n = 5; *P < 0.05, **P < 0.01, and ***P < 0.001 vs. control

Prior cellular studies had shown synergistic therapeutic effects as a consequence of the trebling additive efficacy of energy depletes of GO-Alg@CaP/CO. To further evaluate their therapeutic effect in vivo, 4T1 tumor-bearing mice were used. Fast increase of the tumor volumes in saline and single treatment groups was visualized in Fig. 6F, Additional file 1: Figs. S9 and S11, while the growth of tumors in GO-Alg@CaP/C and GO-Alg@CaP/CO groups was slightly restrained as considerable energy depletes routes boosting starvation therapy. We tried to detect the ATP in tumor issues after GO-Alg@CaP/CO treatment in Additional file 1: Fig. S12. In accordance with precious in vitro ATP assay, distinct decreases of ATP production were observed after 15 days GO-Alg@CaP/CO treatment, demonstrating that GO-Alg@CaP/CO was capable of exacerbating of energy deprivation in tumor issues. Solid evidence about efficient synergistic starvation therapy (tumor inhibition rate of 85% for GO-Alg@CaP/CO group) was found from the above graph which was in agreement with the previous cellular results, demonstrating that the blockage of autophagy by Obatoclax could further potentiate energy deprivation in the tumor harsh conditions. We further provided the protein level of LC3 and P62 in tumor after nanocomposites treatments (GO-Alg@CaP and GO-Alg@CaP/CO) to show solid evidence for autophagy evaluation in vivo. As shown in Additional file 1: Fig. S13, slight decrease of P62 and significant increase of LC3 II revealed the high level of autophagy in starvation therapy. Moreover, both significant increase of LC3 II and P62 were also observed in GO-Alg@CaP/CO group, indicating the autophagy was effectively inhibited in accordance with cellular results.

TUNEL immunofluorescence staining and H&E staining were applied in the follow-up experiment of the therapeutic evaluation. As illustrated in Fig. 6D, GO-Alg@CaP/CO group exhibited highest green fluorescence intensity (revealing dead cells) than other groups, proving the remarkable tumor cell apoptosis. Furthermore, very little cell nucleus was found in GO-Alg@CaP/CO group form the Additional file 1: Fig. S10, demonstrating the same conclusions with the above results. We provided H&E staining images of whole organs (Heart, Liver, Spleen, Lung and Kidney) after treatment in Additional file 1: Fig. S14 and short-term short-term blood biochemical assay data such as 24 h and 7 days after nanoparticle injection in Additional file 1: Fig. S15. The H&E staining of the major organs showed no obvious physiological morphology changes in mice. Delightedly, no evident changes were demonstrated in various serum biochemistry indexes including Liver Function and Renal Function. All evidence suggested that the nanocomposites GO-Alg@CaP/CO would be safely administered for tumor therapy. In the end, no significant fluctuations among mice weight and blood biochemical data were observed in all groups (Fig. 6E, G), confirming negligible toxic and side effects. We added flow analysis in spleen to confirm the biosafety of GO-Alg@CaP/CO. Remarkably, GO-Alg@CaP/CO treatment caused an increase in tumor-infiltrating lymphocytes (TILs, CD3+), especially regulatory T cells (Tregs) including CD4+, CD8+ and CD4+CD8+ within the tumor (Fig. 7). CD4+/CD8+. The increased Tregs in the distant tumor could trigger antitumor immunity which would facilitate tumor treatment of GO-Alg@CaP/CO. And we further provide the ratio of CD4+/CD8+ in Additional file 1: Fig. S16, there was no significant change between Ctr group and GO-Alg@CaP/CO, indicating no obvious damage in immune system after treatment.

Fig. 7
figure 7

A Representative flow cytometric analysis (left) and quantification (right) of CD3+CD8+ and CD3+CD4+ T cells in the tumor. B Representative flow cytometric analysis gating on CD3+ cells (left) and relative quantification (right) of CD4+CD8+ T cells and CD3+cells. Data are mean ± SD (n = 3). *P < 0.05; **P < 0.01; ***P < 0.001

[ad_2]

Source link

RELATED ARTICLES

LEAVE A REPLY

Please enter your comment!
Please enter your name here

Most Popular

Recent Comments