We used an improved facile one-pot synthesis method to prepare DHA@Zif-8 NPs and pure Zif-8. We carefully screened and optimized the reaction conditions of reaction time, temperature, solvents, and the weight ratio of target drug molecules (DHA) to metal ions (Zn²⁺) and organic linkers (MIM). As the NPs formed, we loaded DHA into the framework of Zif-8 to circumvent the drawback of DHA’s poor water solubility. To investigate the sizes and morphologies of these NPs, we performed scanning electron microscopy (SEM) and transmission electron microscopy (TEM). The pure Zif-8 NPs and DHA@Zif-8 NPs both had uniform monodispersed particles and rhombic dodecahedral shapes. The average size of the particles was about 100 nm, as shown in Figure 1. We used the successfully prepared DHA@Zif-8 NPs as the basis to apply DHA clinically to treat Giardia.

To evaluate the in vitro antigiardial effects of DHA@Zif-8 NPs, we monitored the proliferation inhibition effect through cell counting. Because nanomaterials must be biocompatible for biomedical applications, we evaluated whether pure Zif-8 NPs demonstrated cellular toxicity on G. lamblia. As shown in Figure 2A, Zif-8 NPs exhibited negligible cytotoxicity toward G. lamblia at concentrations ranging from 0 to 500 μM, which indicated that Zif-8 NPs possess excellent biocompatibility. Compared with equivalent free DHA, the DHA@Zif-8 NPs had a much better proliferation inhibition effect on G. lamblia. The concentration range was between 0 and 200 μM for 24 or 48 h (Figures 2B,C). The IC₅₀ value of free DHA on G. lamblia at 24 h was 233.3 μM and at 48 h was 221.7 μM. The DHA@Zif-8 NPs possessed IC₅₀ values of 94.7 μM at 24 h and 31.1 μM at 48 h, which were significantly different from the free DHA.

During the encystation process of Giardia trophozoite, the changes in the expression of several genes are signs that cysts are forming. For example, the expression of α-tubulin, β-tubulin, and actin was decreased, and the expression of Cwp2 was increased. Metronidazole has been reported to promote the differentiation of Giardia trophozoites into cysts, which leads to resistance to metronidazole. Our results showed that the DHA@Zif-8 NP-treated group reduced cyst formation (Figure 2D). In the metronidazole-treatment group, the gene expression level of Cwp2 increased and the expression levels of actin, α-tubulin, and β-tubulin decreased. In the DHA@Zif-8 NP-treated group, the expression level of Cwp2 was increased as well as the expression levels of actin, α-tubulin, β-tubulin, and actin (Figure 2E). According to these results, G. lamblia cyst formation can be inhibited by DHA@ Zif-8 NPs.

The effective endocytosis of DHA@Zif-8 NPs may explain this enhanced toxicity. To improve drug effects, the endocytosis of NPs is essential. We used confocal laser scanning microscopy (CLSM) to detect the cellular uptake behavior of DHA@Zif-8 NPs by G. lamblia. We used a blue channel by Hoechst 33258 (a nucleus staining dye) to visualize the cell nuclei. To facilitate fluorescence observation, we embedded the DHA with Nile red (NR) into the Zif-8 NPs. After treatment of 50 μM of DHA@Zif-8 NPs, we observed a strong red fluorescence signal that was distributed mainly in the cytoplasm (Figure 2F). This result indicated that DHA@Zif-8 NPs passed into the cytoplasm from across the cell membrane. After a prolonged incubation time from 1 to 5 h, fluorescence intensity was significantly enhanced. This result showed that DHA@Zif-8 NPs possessed a time-dependent internalization, which validated the sustained uptake of the prepared DHA@Zif-8 NPs in successfully treating G. lamblia.

We used SEM to observe the ultrastructural changes to explore the proliferation inhibition effects of Zif-8 NPs on G. lamblia. As shown in Figure 3AI, the cross-section of the Giardia trophozoite appeared pear-shaped, with two oval vesicular nuclei. The staining plasmids around the nucleus are evenly distributed at the inner edge of the nuclear membrane. Single-layer arranged small vesicles could be seen under the plasma membrane. Flagella are emitted from the matrix, and the sugar bodies in the cytoplasmic core are dense and evenly distributed. In the coronal section of the trophozoites, the cross-sectional view of the round nucleolus and flagella in the nucleus, the microtubules of the sucker around the axoneme of the tail flagella can be observed in the control group. After DHA@Zif-8 NPs treatment (Figures 3AII,III), the G. lamblia showed swelling expansion of the endoplasmic reticulum. Dissolution and vacuoles appeared in the cytoplasm. The microtubules of the sucker disintegrated. The cytoplasm is almost completely depleted, the nuclear chromatin edge gathers, perinuclear space widens, the plasma membrane breaks, and the ultrastructure of the nuclear membrane disintegrates.

Reportedly, ART and its derivatives have demonstrated antiparasitic and antitumor activity by creating reactive oxygen species (ROS) in parasites and cancer cells. In this study, we performed flow cytometry to detect ROS in DHA@Zif-8 NP-treated G. lamblia to understand the potency of these treatments in generating ROS. Compared with the control group, the DHA@Zif-8 NP-treatment group generated more ROS. These results verified that DHA delivery by Zif-8 NPs to G. lamblia was improved and triggered significant inhibition in vivo following more ROS production (Figure 3B,C).

To verify the gene expression in G. lamblia before and after treatment with DHA@Zif-8 NP and to further explore the antigiardial mechanisms of DHA@Zif-8 NPs, we used RNA sequencing (RNA-seq). We used formaldehyde agarose gel electrophoresis and ultraviolet spectrophotometry to analyze the total RNA extracted from the DHA@Zif-8 NP-treated group (D) and untreated group (C) for integrity and quality. According to these results, the RNA obtained was nondegraded, complete, and suitable for RNA-seq analysis. We used principal component analysis (PCA) to assess the complete dataset of the samples. The PCA results showed that the samples were separated into two clusters. Correlations were calculated by the Pearson correlation coefficient and visualized in unclustered heatmaps. This demonstrated that the various groups had a distinct directionality based on similarities in gene expression (Figures 4A,B,C). We also identified the differentially expressed genes (DEGs). The volcano plots showed the DEG results from the treatment with DHA@Zif-8 NP. We used the Benjamini–Hochberg method to perform multiple testing corrections with a 1.5-fold change cutoff and a corrected p-value of <0.05. We compared the RNA-seq data from the C and D, and we verified 219 DEGs, among which 126 were downregulated and 123 were upregulated.

Using the RNA-seq data, we conducted pathway enrichment analysis and performed GO for 219 DEGs following treatment with DHA@Zif-8 NP. The GO classification covers three domains: biological processes (BPs), molecular functions (MF), and cellular components (CCs). A high proportion of DEGs in the BP domain was related to biological regulation, immune system processes, cellular processes, metabolic processes, and responses to stimuli. The enriched parts in the CC domain cover the cell, cell part, membrane, and organelle. The catalytic activity, binding, molecular function regulator, and transcriptional activator activity were affected primarily in the MF domain (Figure 5A). On the basis of the DEG results, we performed enrichment analysis and KEGG functional classification. According to the KEGG functional classification analysis, the treatment with DHA@Zif-8 NPs affected signal transduction as well as amino acid, carbohydrate, and lipid metabolism translation (Figure 5B). The KEGG pathway enrichment revealed that ribosome biogenesis in eukaryotes, phosphatidylinositol signaling system, and spliceosomes were affected (Table 1).

Next, we selected some of the essential genes related to the metabolism to verify the expression levels in DHA@Zif-8 NP-treated or -untreated G. lamblia by quantitative reverse transcription polymerase chain reaction (qRT-PCR) (Figure 5C). According to the results of qRT-PCR, hgsA, Gnpnat, Nthl1, sac1, agpat9l, hspB, and DDB_G0267588 were downregulated. The RNA-seq results were the same as the gene expression.

We assessed the experimental infection by monitoring the weight of mice and fecal Giardia cysts. We found that mice had a slight decrease in weight, and Giardia cysts in the fecal matter were successfully confirmed as infection. Following DHA@Zif-8 NP treatment, we observed a significant decrease in the G. lamblia cyst count in the fecal samples from the infected treated mice relative to the infected nontreated mice. In addition, we did not observe any significant difference between high doses of DHA@Zif-8 NPs and MTZ. Following a high dose of DHA@Zif-8, the most obvious effect was the disappearance of the parasite from the stool and a more than 95% reduction after treatment for 12 days (Figure 6A). The DHA@Zif-8 NP-treated groups’ survival rate was significantly better than the infected groups (Figure 6B). After the experiment, the treatment group’s body weight was the same as the control group, whereas the model group’s body weight was lower (Figure 6C).

We conducted an H&E examination of sections of the small intestines in the mice in the control group. We observed normal brush border and villous architecture with an average length and width of the villi. In the infected group, we observed that the structure of the intestinal mucosa had changed as evident in atrophy and villous shortening. Along with fusion, this contributed to increased goblet cells, a blunting of most villi with desquamation, a decrease in the villous height–to–crypt length ratio, and infiltration of the lamina propria with inflammatory cells (e.g., lymphocytes and eosinophils) with diffuse loss of the brush border of the microvillous surface area and necrosis. In the DHA@Zif-8 NP-treated mice, we observed a marked improvement in histopathological lesions in the form of complete healing of the intestinal mucosa. In addition, mucosal ulceration was absent, villous architecture was normal, and the brush border was preserved. As shown in Figure 6D, inflammatory cell infiltration of the G. lamina disappeared, thus proving that DHA@Zif-8 NPs play a therapeutic role against giardiasis.

We purchased the chemicals and solvents from commercial sources. Unless indicated otherwise, all were used without further treatment. Molecular Devices supplied the SpectraMax iD5 Multi-Mode Microplate Reader. Becton, Dickinson, and Company (BD) (Franklin Lakes, NJ, USA) supplied the Accuri C6 flow cytometer. We used a JEOL JEM-1011 electron microscope (acceleration voltage of 100 kV) for the TEM images.

We followed the method given by Li (39) to perform the synthesis of DHA@Zif-8. We dissolved 75 mg of zinc nitrate hexahydrate in 2.5 mL of deionized water in a typical experiment. In other cases, we dissolved 165 mg of 2-methyl imidazole in 4.5 mL of methanol and 10 mg of DHA in 0.5 mL of DMF. We added the aqueous zinc nitrate solution to the 2-methyl imidazole and DHA solution under continuous stirring at room temperature. As a result, the synthesis solution turned turbid quickly. The DHA@Zif-8 NPs formed after 3 min. To obtain DHA@Zif-8 NPs, we centrifuged the solution at 12,000 rpm for 10 min. To remove the unreacted reagents, we washed the DHA@Zif-8 NPs with methanol three times. We collected the supernatants to measure encapsulation efficiency and drug loading content of DHA. Following treatment with 0.2% NaOH aqueous solution at 50°C for 30 min converted DHA in the supernatant into a UV-absorbing compound. We used the UV–vis spectrometer to measure the detection wavelength at 290 nm. After freeze-drying, we stored the products at −20°C until further use. We used the method to synthesize the Zif-8 NPs as a control.

We collected C2, trophozoite of G. lamblia isolates, from a patient in southwest China. We used modified trypticase yeast extract iron-serum-33 medium (TYI-S-33), which was composed of 2% casein digest, 1% yeast extract, 1% glucose, 0.2% NaCl, 0.2% l-cysteine, 0.02% ascorbic acid, 0.2% K₂HPO₄, and 0.06% KH₂PO₄. We used borosilicate glass screw-cap culture tubes to supplement this with 10% heat-inactivated bovine serum (Hangzhou Sijiqing Biological Engineering Materials Company) and 0.05% bovine bile (Sigma) at pH 6.8. Without shaking, we inoculated the culture with 4 × 10³ trophozoites per 4 mL tube at 37°C three times a week.

In the TYI-S-33 medium, pH 6.8, we grew Trophozoites of G. lamblia at 37°C for 48 h. This medium was supplemented with 0.05% bovine bile and 10% heat-inactivated bovine serum. We detached the trophozoites in the logarithmic phase by chilling. To detect growth inhibition of DHA@Zif-8 NPs on G. lamblia, we subcultured the G. lamblia (5 × 10⁴) trophozoites with TYI-S-33 medium at 37°C for 24, 48, and 72 h in a 4 mL tube in the presence of various concentrations of DHA@Zif-8, DHA, or pure Zif-8 NPs. We counted the parasites every 24-h period for consecutive days. To make each reading, we placed a tube of cultured G. lamblia on ice for 20 min. Then, we inverted the tube gently to detach the trophozoites. We centrifuged the tubes for 10 min at 2,000 g. After we discarded the supernatants, we resuspended the pellets in 1 mL of PBS. Following homogenization, we took out the aliquots to count the parasites in the Neubauer chamber. We performed three experiments independently.

We grew G. lamblia for 48 h at 37°C and used chilling PBS to detach the trophozoites. We subcultured the trophozoites (5 × 10⁴) in a 4 mL tube with encystation medium TYI-S-33 (pH 7.1), which was supplemented with 0.5 mg/mL bile. After the G. lamblia reached the logarithmic growth period, we changed the medium to TYI-S-33 (pH 7.8) and added 10 mg/mL bile for 24 h. We next changed the medium to TYI- and added S-33 (pH 7.1) with 0.5 mg/mL bile for 24 h. We separated the G. lamblia into DHA@Zif-8 NP-treated, metronidazole-treated, and encystation groups (100 μg/mL). We used deionized water to split trophozoites and incomplete cysts, added 1 mL of distilled water, and used the blood cell counting plate to calculate the rate of cyst formation. After we extracted the total RNA, we reverse-transcribed it into single-stranded DNA. We used qPCR to detect the gene expression level of actin, Cwp2, α-tubulin, β-tubulin, HSP70, and GAPDH (see Supplementary Table S1 for the primer sequences).

We treated trophozoites for 48 h with various concentrations of DHA@Zif-8 NPs. Following DHA@Zif-8 NP treatment, we collected the trophozoites and fixed the sample in PBS (pH 7.3) with ice-cold 2.5% glutaraldehyde. We removed the fixative and washed each trophozoite two times with the cold PBS. Then, we used graded ethanol for dehydration, propylene oxide for immersion, and epoxy resin to embed the sample. We used lead citrate/uranyl acetate to double-stain ultra-thin sections, which we then examined by TEM.

We used a ROS assay kit (Beyotime, Shanghai, China) to assess the accumulation of intracellular ROS after DHA@Zif-8 NP treatment. We treated the trophozoites with various concentrations of DHA@Zif-8 NPs before incubation with 10 μM DCFH-DA for 30 min at 37°C. Then, we collected the samples and suspended them in ice-cold PBS. After deacetylation, DCFH-DA was oxidized by ROS and formed 2,7-dichlorofluorescein (DCF). We used flow cytometry at an excitation wavelength of 488 nm and an emission wavelength of 525 nm to measure the DCF fluorescence intensities.

We examined the DHA@Zif-8 NPs internalized by G. lamblia using a fluorescence microscope. We seeded the trophozoites in six-well culture plates at a density of 5 × 10⁴ cells per well. We put a sterile coverslip in each well and allowed them to adhere for 24 h. We treated the trophozoites with DHA@Zif-8 NPs at 37°C for 0.5, 2, and 4 h. We carefully removed the supernatant and washed the trophozoites three times with PBS. We fixed the cells with 500 μL of 4% formaldehyde at room temperature for 20 min and used PBS to wash them two times. We used a blue channel for Hoechst 33258 and a red channel for NR to visualize the trophozoites with a fluorescence microscope.

To conduct the RNA-seq analysis, we used trophozoites treated with DHA@Zif-8 NPs and performed the analysis for the DHA@Zif-8 NP-treated or untreated groups on three biological replicates. After this treatment, we used a mirVana miRNA Isolation Kit (Ambion, Austin, TX, USA) to extract RNA using the manufacturer’s recommendations. We treated the sample with RNase-free DNase I (ThermoFisher Scientific, Waltham, MA, USA) and removed the contaminated DNA. We used a NanoDrop 2000 spectrophotometer (Nanodrop Technologies, Wilmington, DE, USA) to assess the extracted RNA quality.

We used the RNA Nano6000 Assay Kit on a Bioanalyzer 2,100 system (Agilent Technologies, Santa Clara, CA, USA) to detect RNA integrity. If the samples had an RNA Integrity Number (RIN) of more than seven, we used them to construct the sequencing libraries. We constructed the libraries with a TruSeq Stranded mRNA LTSample Prep Kit (Illumina, San Diego, CA, USA) following the manufacturer’s instructions. We used the Illumina sequencing platform (HiSeq™ 2,500) to sequence the libraries and generated 150-bp paired-end reads. We filtered the raw tag data. We removed reads containing adapters, poly-N, and low-quality reads from the raw data to obtain clean data. On the basis of the high-quality clean data, we conducted the downstream analyses.

We performed DEG analyses to compare the DHA@Zif-8 NP-treated and untreated groups. First, we mapped clean reads to the reference genome using hisat2. Then, we calculated the FPKM and read count values for each transcript (protein coding) with bowtie2. Next, we identified the DEGs using the DESeq functions estimateSizeFactors and nbinomTest. For significant differential expression, we set a fold change of >1.5 or < 0.5 and a p-value of <0.05 as the threshold. Finally, to explore the transcript expression patterns, we performed a hierarchical cluster analysis of the DEGs. We used R based on the hypergeometric distribution to perform GO enrichment and KEGG pathway enrichment analysis.

We used total RNA that was extracted from DHA@Zif-8 NP-treated and untreated trophozoites to characterize the expression of selected genes. We employed oligonucleotide primers that were gene-specific to perform qRT-PCR on RNA samples (Supplementary Table S1). We also used an iScript One-Step RT-PCR Kit with SYBR Green and CFX96 Touch Real-Time PCR Detection System (Bio-Rad, Hercules, CA, USA). All values were standardized to the control condition and normalized to β-actin. The standard error of the mean (SEM) was represented by error bars. We used Student’s t-test to assess statistical significance. After RT-PCR was complete, we obtained the C_t values from the ABI 7500 fast v2.0.1 software. To represent mRNA-fold change, we used the ^ΔΔCt method.

The Institutional Animal Care and Use Committee at Ji Lin Medical University approved all animal procedures (2024-LW001). The Changchun Yisi Laboratory Animal Technology Co., Ltd., provided the BALB/c mice, which we housed at a controlled room temperature and humidity with a 12 h photoperiod. Mice were provided with sterilized food and water. After 48–72 h, we harvested the G. lamblia trophozoites in TYI-S-33 medium in the log phase. One time every day, we used oral gavage to infect 50 of the mice with 1 × 10⁶ G. lamblia trophoblast and 1 × 10⁵ G. lamblia cyst in 0.2 mL of PBS. Giardia infection was monitored through daily fecal collection and cyst detection. We confirmed that the model was established successfully when the mice had higher Giardia colonization with the fecal Giardia cyst. We randomly divided the mice into seven groups. Each group had six mice: Giardia infection, metronidazole (10 μM /kg/d), free DHA (20 μM/kg/d), low dose (5 μM/kg/d), middle dose (10 μM/kg/d), and high dose of DHA@Zif-8 (20 μM/kg/d), and control. Mice was administered by intragastric administration once everyday. We collected fecal samples every day. After 15 days of observation, we killed the mice by cervical dislocation.

We removed the small intestine from each mouse aseptically. After we fixed the sample in 4% paraformaldehyde solution, it was embedded in paraffin. To monitor the kinetics of infection, we sliced and stained the sample with hematoxylin and eosin (H&E).

We performed all experiments at least three times. We expressed all results as the mean ± standard deviation. To determine the statistical difference between experimental and control groups, we used Student’s t-test. Significance is denoted by *p < 0.05.