**3. Results**

#### *3.1. Preparation and Characterization of Cu-MOF/GOD@HA*

The preparation process of Cu-MOF/GOD@HA was shown in Figure 1A as described in the experimental section. GOD was firstly loaded into Cu-MOF by amide reaction between –COOH on the Cu-MOF and –NH2 on GOD. Then, Cu-MOF/GOD was coated by HA to avoid the leakage of GOD and to improve its biocompatibility as well as the targeting ability. The CDT mechanism of Cu-MOF/GOD@HA was shown in Figure 1B. After endocytosis into MCF-7 cells, i.e., the tumor cells, CDT process based on Cu-MOF/GOD@HA was triggered sequentially by GSH "AND" H2O2 in the cancer cell interior. In addition, GOD in Cu-MOF catalysis Glu to supply H2O2 to improve CDT efficacy.

Transmission electron microscope (TEM) images in Figure 2A,B illustrate the Cu-MOF with diameter of 63.32 ± 9.12 nm. SEM image and Elemental mapping of C, N, O and Cu of Cu-MOF confirmed the successful preparation (Figure S1A,B). From the EDS spectra from SEM, Cu-MOF were consisted of 1% Cu, 2% N, 5% O and 92% C. As shown in Figure 2C, the shells of loaded GOD and coated HA were 5 nm. Compared with Cu-MOF, the FT-IR spectra of Cu-MOF/GOD, Cu-MOF/GOD@HA and GOD showed bands at 1550 cm–1, attributed to vibrational stretches characteristic of GOD, indicating the successful loading of GOD (Figure S2A). Furthermore, the TGA analysis of Cu-MOF and Cu-MOF/GOD@HA further confirmed the GOD loading and HA coating (Figure S2B). From the TGA analysis, Cu-MOF/GOD@HA were consisted with 15.32% HA, 21.56% GOD and 63.12% Cu-MOF.

**Figure 2.** (**A**) TEM images of Cu-MOF. Scale bar: 50 nm. (**B**) The size distribution of Cu-MOF as measured by TEM image. ( **C**) TEM images of Cu-MOF/GOD@HA (Scale bar: 25 nm).

In order to confirm the successful preparation of Cu-MOF/GOD@HA, the UV-vis absorption spectra of the reaction mixture were recorded. The absorption-change of Cu-MOF, Cu-MOF/GOD and Cu-MOF/GOD@HA in the 600–900 nm region, indicating the successful assembly of Cu-MOF/GOD@HA is clearly seen in Figure 3A. The color of Cu-MOF and Cu-MOF/GOD@HA changed from sky blue to green, also indicating successful loading of GOD and coating of HA. Zeta potential values of the corresponding nanocomposites obtained in each step are shown in Figure 3B, where the Cu-MOF aqueous solution exhibits a potential of −10.03 ± 0.16 mV. The loading of GOD reduces the zeta potential to −4.84 ± 0.19 mV, implying the deprotonation of the –NH2 of the GOD. The increase of zeta potential to −7.84 ± 0.28 mV due to the presence of the –OH of HA [32,33]. Furthermore, the X-ray diffraction (XRD) spectra of Cu-MOF, Cu-MOF/GOD and Cu-MOF/GOD@HA indicates that the loading of GOD and coating of HA did not influence the crystal structure of Cu-MOF (Figure S3).

**Figure 3.** (**A**) UV-vis absorption spectra of Cu-MOF, Cu-MOF/GOD and Cu-MOF/GOD@HA (The concentrations of pending test samples were 100 μg mL–1). Inset in (**A**) shows the photographs of the solutions containing Cu-MOF, Cu-MOF/GOD and Cu-MOF/GOD@HA. (**B**) Zeta potential values of Cu-MOF, Cu-MOF/GOD and Cu-MOF/GOD@HA. (The concentrations of pending test samples are 200 μg mL–1).

#### *3.2. Depletion of GSH and Generation of H2O2*

In order to confirm the oxidation-reduction reactions between Cu-MOF and GSH, molar equivalents of Cu-MOF and GSH were mixed in an aqueous solution. It is clearly illustrated in Figure S4 that the absorption of Cu-MOF completely disappeared after reaction of Cu-MOF and GSH for ca. 2 min along with the formation of Cu(I)-MOF, and the color of the mixture of Cu-MOF and GSH turned from sky blue to white. This observation obviously demonstrated that the incubation of Cu-MOF with GSH facilitates the reduction of divalent Cu(II) in the Cu-MOF by GSH. In addition, compared with TEM image of Cu-MOF, Cu-MOF was collapsed after treatment with 10 mmol L–<sup>1</sup> GSH (Figure S5), proving the reaction between GSH and Cu-MOF.

X-ray photoelectron spectroscopy (XPS) was also used to investigate the oxidationreduction reactions between Cu-MOF and GSH. XPS is a surface sensitive technique. Compared with Cu-MOF (Figure S6A), an obvious N1s peak was observed in the XPS spectra of Cu-MOF/GOD and Cu-MOF/GOD@HA (Figure S6B,C), further demonstrating the occurrence of GOD loading. After treatment by GSH, an obvious N1s peak appears in Cu-MOF due to the residues of GSH (Figure S6D). Furthermore, after treatment of GSH, the XPS analysis of Cu(I)-MOF only shows Cu(I) (932.7, 952.4 eV) peaks (Figure S6E), also

confirming the reduction of Cu-MOF by GSH. Moreover, Figure 3A also shows that after loading with GOD and coating with HA, the UV-vis absorption of between 600–900 nm is significantly reduced. Interestingly, the proportion of Cu(I) on the surface of Cu-MOF have increased after loading with GOD and coating with HA. As shown in Figure S6F–H, the valence state of Cu is "+1" or "+2", since the Cu 2p XPS spectra displayed peaks on Cu(II) (934.4 eV, 954.2 eV)/Cu(I) (932.7 eV, 952.4 eV). The existence of Cu(II)/Cu(I) redox pair provide a grea<sup>t</sup> potential for Fenton-like reactions. In the paramagnetic chemical state (Figure S6F–H), the Cu 2p3/2 XPS spectrum is quantitatively analyzed, and the Cu(II)/Cu(I) ratio of Cu-MOF is 1.2872. The Cu(II)/Cu(I) ratios on the surface of Cu-MOF/GOD and Cu-MOF/GOD@HA in Cu 2p3/2 XPS spectra are 1.0376 and 0.8744, respectively. Obviously, after loading with GOD and coating with HA, the proportion of Cu(I) on the surface of Cu-MOF increased, indicating that Cu(II) was reduced to Cu(I) during the assembly process on the surface of Cu-MOF. This phenomenon may shorten the reaction time from Cu-MOF to Cu(I)-MOF.

After reduction of Cu-MOF by GSH, Cu(I) moiety catalyzes H2O2 to generate ·OH via a Fenton-like reaction. The Cu(I)-MOF-involved CDT processes are investigated and demonstrated by observing the degradation methylene blue (MB) [34]. ·OH induced degradation of MB which may result in a significant decrement on the absorption of MB at the maximum wavelength of 663 nm. Figure 4A shows that the degradation results of MB under different conditions by incubation for 40 min. In the presence of only Cu(I)-MOF or H2O2, virtually no degradation of MB was observed. On the contrary, a remarkable degradation of MB was recorded by introducing H2O2 into the mixture containing MB and Cu(I)-MOF. In addition, in the presence of free Cu(I) and H2O2, no remarkable degradation of MB was observed within 40 min. Thus, compared with free Cu(I), Cu(I) activated from Cu-MOF has a higher catalytic performance. Moreover, Figure S7 shows that Cu-MOF could not react with H2O2 to generate ·OH to reduce the degradation of MB. This provides a clear evidence for the occurrence of the Cu-MOF involving in the cascade reactions in CDT process. First, divalent Cu(II) in Cu-MOF was activated by GSH to convert to monovalent Cu(I) in Cu(I)-MOF. Subsequently, Cu(I)-MOF catalyzes the generation of ·OH from H2O2. The absorption spectra of MB illustrating its time-dependent degradation is shown in Figure 4B. It is obvious that MB was completely degraded after incubation for 40 min.

The depletion of GSH by Cu-MOF and the generation of H2O2 by GOD were investigated. First of all, Cu-MOF were treated by excess GSH. The residual GSH was indicated by 5,5-dithiobis (2-nitrobenzoic acid) (DTNB), a kind of sulphydryl (–SH) indicator at different time points [35]. With the extended time, the characteristic absorbance at ~412 nm decreased, indicating the depletion of GSH. As shown in Figure S8A, the depletion capacity of GSH exhibits a concentration-dependent model. The fluorescence signals of Ampliflu Red were influenced by the change of H2O2 concentration [36]. Therefore, Ampliflu Red can be used to investigate the generation of H2O2. As shown in Figure S8B,C, the generation of H2O2 exhibits a time-dependent model and a Glu-concentration-dependent model. The catalytic reaction of GOD with glucose to produce gluconic acid and H2O2 will cause the pH of the aqueous solution to change. Therefore, after mixing Cu-MOF or Cu-MOF/GOD and Glu for different time, the pH change can reflect the catalytic performance of GOD in Cu-MOF/GOD. As shown in Figure S9, the pH values show a time-dependent mode and a concentration-dependent mode (Glu). These results indicate that GOD loaded in Cu-MOF/GOD still can catalyze Glu to produce H2O2.

**Figure 4.** Experimental demonstration of the mechanism for the CDT process with Cu-MOF as the nano-therapeutic agent. OH generated from the Fenton-like reaction between Cu(I)-MOF and H2O2 leads to the degradation of MB. (**A**) MB degradation after incubation for 40 min in the presence of Cu(I)-MOF, H2O2, H2O2+Cu(I)-MOF and H2O2+free Cu(I), respectively. (**B**) The time-dependent degradation process of MB in the presence of Cu(I)-MOF. MB: 5 μg mL–1, H2O2: 10 mmol L–1, Cu(I)-MOF: 200 μg mL–<sup>1</sup> (Cu(I): 0.494 μg mL–1), free Cu(I): 0.494 μg mL–1, time interval: 5 min.
