Abstract

   Rapid drug clearance and off-target effects of therapeutic drugs can
   induce low bioavailability and systemic side effects and gravely
   restrict the therapeutic effects of inflammatory bowel diseases (IBDs).
   Here, we propose an amplifying targeting strategy based on orally
   administered gallium (Ga)-based liquid metal (LM) nano-agents to
   efficiently eliminate reactive oxygen and nitrogen species (RONS) and
   modulate the dysregulated microbiome for remission of IBDs. Taking
   advantage of the favorable adhesive activity and coordination ability
   of polyphenol structure, epigallocatechin gallate (EGCG) is applied to
   encapsulate LM to construct the formulations (LM-EGCG). After adhering
   to the inflamed tissue, EGCG not only eliminates RONS but also captures
   the dissociated Ga to form EGCG-Ga complexes for enhancive
   accumulation. The detained composites protect the intestinal barrier
   and modulate gut microbiota for restoring the disordered enteral
   microenvironment, thereby relieving IBDs. Unexpectedly, LM-EGCG
   markedly decreases the Escherichia_Shigella populations while
   augmenting the abundance of Akkermansia and Bifidobacterium, resulting
   in favorable therapeutic effects against the dextran sulfate
   sodium-induced colitis.
     __________________________________________________________________

   Liquid metal improves gut dysbacteriosis and obliterates RONS to
   alleviate IBDs via a reassembly enhanced accumulation strategy.

INTRODUCTION

   Gut microbiota dysbiosis, featuring an immoderate proliferation of
   profitless microbes and a decrease in beneficial bacteria, plays an
   important role in the progress of multiple diseases, such as
   inflammatory bowel diseases (IBDs) and tumors ([41]1, [42]2). Although
   conventional therapeutic approaches for IBDs have mainly focused on
   suppressing hyperactive immune responses in inflamed tissues, a growing
   body of evidence has indicated that reprograming gut microbiota is
   highly beneficial to IBD alleviation ([43]3–[44]5). Accordingly,
   numerous efforts have been made to develop effective strategies to
   modulate gut microbiota for IBD alleviation ([45]6–[46]9). However, IBD
   treatment remains a challenge since the first-line drugs show side
   effects and limited therapeutic efficiency due to the complicated
   etiology and pathogenesis of IBDs ([47]4, [48]10–[49]12). Besides, the
   neglected overproduced reactive oxygen and nitrogen species
   (RONS)–induced serious inflammatory response and the poor efficacy of
   intestinal flora regulation have markedly offset the therapeutic
   effects of IBDs ([50]13, [51]14). Therefore, exploiting orally
   administered agents that can precisely reprogram gut microbiota and
   synchronously eliminate RONS is of great significance for IBD
   management.

   As an indispensable nutrient for bacterial survival, proliferation, and
   bacterial infection, iron is involved in many vital metabolic
   processes, including energy generation and DNA synthesis
   ([52]15–[53]17). Biological systems of bacteria are unable to
   distinguish gallium from iron ions due to their analogous structural
   features and physicochemical properties ([54]18, [55]19). As a result,
   bacteria would specifically take in Ga^3+ through iron ion uptake
   pathways in the absence of iron. Of note, Ga^3+ cannot be reduced under
   physiological conditions, making Ga^3+ a potential “tool” for bacterial
   regulation ([56]18, [57]20–[58]22). Correspondingly, gallium-based
   compounds have been applied to inhibit pathogenic bacteria without
   causing bacterial resistance ([59]23–[60]25). For instance, ultrasmall
   nonantibiotic gallium-indocyanine green nanoparticles were applied to
   eradicate biofilms and against drug-resistant bacterial liver abscesses
   ([61]26). Therefore, designing nano-agents for delivering Ga^3+ to
   infection sites would be a promising strategy for bacterial modulation.

   Gallium (Ga)-based liquid metal (LM), characterized by favorable
   plasticity and excellent biocompatibility ([62]27, [63]28), has
   received increasing attention in stretchable electronic devices,
   microfluidic systems, and biomedical applications ([64]29–[65]31). The
   extraordinary photothermal conversion efficiency, large specific
   surface area, and special metallic surface properties of LM have
   rendered LM-based formulations as excellent candidates for drug
   delivery platforms and photothermal agents in cancer therapy
   ([66]32–[67]36). Recently, gallium-based LMs have attracted tremendous
   interest as potential anti-infection agents for their satisfactory
   bacterial-modulation performance ([68]37–[69]39). For example, an
   LM-based soil system can enrich the gut bacterial diversity and rectify
   bacterial dysbiosis under a pathophysiological condition ([70]40), and
   gallium-based LMs have been applied to construct acid-responsive
   micromotors against the Helicobacter pylori–caused infections ([71]41).
   Nevertheless, despite a few inspiring applications, gallium-based LM in
   cancer therapy or bacterial regulation is still at its preliminary
   stage.

   Epigallocatechin gallate (EGCG), a major bioactive polyphenol in green
   tea ([72]42), can not only coordinate with metal ions but also adhere
   to the biomolecules owing to its polyphenolic hydroxyl structure that
   has anti-oxidation and strong adhesivity ([73]43–[74]45). Here, we
   proposed an EGCG-coated LM nano-agent (LM-EGCG) to efficiently
   eliminate RONS and modulate the dysregulated microbiome for remission
   of IBDs by amplification targeting strategy based on reassembly of
   Ga^3+ and EGCG. After LM-EGCG degradation ([75]Fig. 1), the
   dissociative EGCG could not only adhere to electropositive inflamed
   tissue to eliminate RONS but also capture dissociated Ga^3+ to form
   metal polyphenol complex, thereby improving Ga^3+ accumulation to
   modulate gut microbiota. Besides, the impaired intestinal barrier was
   protected and the overall abundance and diversity of gut microbiota
   were increased after oral admonition of LM-EGCG, resulting in favorable
   therapeutic outcomes against the dextran sulfate sodium (DSS)–induced
   colitis. LM-EGCG markedly decreased pathogenic Escherichia_Shigella
   populations while augmenting the abundance of Akkermansia and
   Bifidobacterium, which are crucial microbes in gut homeostasis. This
   work proposes an amplifying targeting strategy based on the reassembly
   of EGCG and Ga^3+ for efficient remission of IBDs, which broadens the
   horizons of Ga-based LM in biomedical applications.

Fig. 1. Schematic illustration of the construction of LM-EGCG and the
mechanism for colitis treatment.

   [76]Fig. 1.
   [77]Open in a new tab

   (A) LM-EGCG was successfully constructed by one-pot ultrasonic
   treatment and orally administered for IBD therapy. (B) The process of
   pH-mediated degradation of LM-EGCG in the stomach and reassembly of
   EGCG with Ga^3+, thereby modulating the dysregulated microbiome,
   intestinal barrier, and immune responses for alleviating inflammation
   diseases.

RESULTS

Construction and characterization of LM-EGCG

   Before LM-EGCG synthesis, the performance of Ga^3+ and organic ligands
   to form complexes in the intestinal fluid simulation was investigated.
   Predictably, Ga^3+ promptly formed complexes at the existence of those
   of ligands containing polyphenolic hydroxyl structure, especially EGCG,
   in intestinal fluid simulation (fig. S1). Therefore, we further
   inquired into the variation of the mixture of EGCG and Ga^3+ in
   different conditions. Individual EGCG or Ga^3+ would not form
   macroscopic complex precipitation in simulated gastric fluid, simulated
   intestinal fluid, and simulated colonic fluid (SCF) ([78]Fig. 2A).
   Because of the complexation of EGCG with Ga^3+ in alkalescent SCF, the
   hydrodynamic size of the mixture containing EGCG and Ga^3+ increased to
   approximately 2 μm ([79]Fig. 2C) and a great deal of buff floccule
   appeared in the mixture, which increased with longer placement time
   ([80]Fig. 2B). Encouraged by the phenomenon that EGCG could capture
   Ga^3+ to form complex in SCF, we synthesized the complex of EGCG with
   Ga (Ga-EGCG) and the EGCG-coated LM nano-agent according to our
   previous report ([81]21). Transmission electron microscopy (TEM) images
   revealed that the size of LM-EGCG was about 250 nm, and the thickness
   of EGCG shell was approximately 60 nm ([82]Fig. 2D). Moreover, the
   relative content of EGCG in LM-EGCG was analyzed using
   thermogravimetric analyzer (TGA), which revealed that almost a 21%
   weight decrease occurred in LM-EGCG ([83]Fig. 2E), indicating about 21%
   loading efficiency of EGCG.

Fig. 2. LM-EGCG exerts a satisfactory RONS elimination rate and Ga^3+ remains
at the inflamed colon in DSS-treated mice.

   [84]Fig. 2.
   [85]Open in a new tab

   (A and B) Photographs of (A) the mixture in simulated gastric fluid
   (SGF), simulated intestinal fluid (SIF), and SCF and (B) the mixture in
   SCF at various time points. (I) Ga^3+, (II) EGCG, and (III) Ga^3+ +
   EGCG. (C) Hydrodynamic size of Ga^3+, EGCG, and Ga^3+ + EGCG in SCF.
   (D) Transmission electron microscopy (TEM) images of LM, Ga-EGCG, and
   LM-EGCG. Scale bars, 200 nm. (E) TG analysis of LM and LM-EGCG. (F and
   G) Evaluation of the (F) •O[2]^− and (G) reactive nitrogen species
   (RNS) elimination rate of LM, EGCG, and LM-EGCG. (H) The RNS
   elimination capacities of LM, EGCG, and LM-EGCG at various time points.
   (I) Evaluation of the intracellular reactive oxygen species (ROS)
   clearance performance of LM, EGCG, and LM-EGCG. Scale bar, 100 μm. (J
   and K) In vivo fluorescence imaging after oral administration of
   Cy5.5-labeled LM-EGCG, and (K) the corresponding semiquantitative mean
   fluorescence intensity (MFI). (L) Ga content in the stomach, small
   intestine, and colon from healthy or DSS-treated mice after oral
   administration of LM-EGCG. (M) WT of samples Ga foil and Ga[2]O[3], the
   mixture of EGCG and Ga in SCF, and the colonic contents from IBD mouse
   after oral gavage of LM-EGCG. (N) Ga K-edge XANES spectra of reference
   samples (Ga foil and Ga[2]O[3]), LM-EGCG, the complex collected in SCF,
   and the colonic contents from IBD mouse after oral gavage of LM-EGCG.
   (O) The corresponding Fourier transformation of Ga K-edge x-ray
   absorption fine structure (EXAFS) spectra in k-space. (P) Fourier
   transforms (FT) of k3-weighted Ga K-edge EXAFS spectra. (Q)
   Energy-dispersive x-ray spectroscopy element mapping images of the
   colonic contents from IBD mouse after oral gavage of LM-EGCG. Scale
   bar, 200 nm. Data are presented as means ± SD (n = 3). ****P < 0.0001,
   analyzed by one-way ANOVA. a.u., arbitrary units.

Elimination of RONS

   The •O2^−-scavenging capacity of LM-EGCG was first evaluated using
   nitro blue tetrazolium (NBT). The absorbance of NBT solution treated
   with EGCG was markedly decreased with the increasing of LM-EGCG
   concentration (fig. S2), indicating that •O2^− in NBT solutions was
   eliminated ([86]Fig. 2F). As expected, the absorbance signal of
   LM-treated NBT solution was also decreased, indicating that LM consumes
   •O2^− owing to its surface reducibility. Compared to EGCG or LM, more
   than 90% of •O2^− was obliterated in LM-EGCG–treated NBT solution,
   suggesting the favorable antioxidative properties of LM-EGCG.
   Subsequently, we further investigated the reactive nitrogen species
   (RNS)–scavenging performance of LM-EGCG using
   2,1,1-diphenyl-2-picrylhydrazyl radical (DPPH). The elimination rate of
   RNS was positively associated with the concentrations of LM-EGCG
   ([87]Fig. 2G). Benefitting from the remarkably reduced capacity and
   unbound state, EGCG could efficiently eliminate RNS. Besides, because
   of the nanostructure of LM-EGCG, the RNS clearance efficiency of
   LM-EGCG was lower than dissociative EGCG at low concentrations, whereas
   more than 95% of RNS was effectively eliminated by LM-EGCG at the
   concentration of 100 μg ml^−1. Correspondingly, more than 90% RNS in
   the LM-EGCG solution was depleted after incubation for 27 min ([88]Fig.
   2H). These results demonstrate that LM-EGCG can effectively eliminate
   RONS.

   Encouraged by the favorable RONS clearance efficiency of LM-EGCG, we
   further assessed its ability to scavenge intracellular RONS. Reactive
   oxygen species (ROS)–sensitive molecule (2′,
   7′-dichlorodihydrofluorescein diacetate, DCFH-DA) probe evaluation
   experiments showed that compared to phosphate-buffered saline
   (PBS)–treated NCM460 cells, attenuated green fluorescence signals from
   DCFH-DA were detected in LM-treated cells, while negligible
   fluorescence was observed in EGCG- or LM-EGCG–treated NCM460 cells
   ([89]Fig. 2I). Similarly, the RNS fluorescence probe [nitric oxide and
   3-amino, 4-aminomethyl-2′, 7′-difluorescein diacetate (DAF-FM DA)]
   evaluation experiments confirmed that, in contrast to brilliant
   fluorescence in NCM460 cells treated with PBS or LM, inconspicuous
   green fluorescence emitted in NCM460 cells after the treatments of EGCG
   or LM-EGCG (fig. S3). These results indicate that EGCG or LM-EGCG could
   effectively clear intracellular RONS.

Accumulation of the complex at the colitis site

   Before colitis treatment, the accumulation capacity of the complex of
   EGCG and Ga^3+ in DSS-induced IBD mice was assessed by tracking the
   fluorescence from cyanine 5.5 (Cy5.5)–labeled EGCG. Compared to
   EGCG-treated mice ([90]Fig. 2J), the strongest fluorescence was
   detected in colonic tissues in those of IBD mice after oral gavage of
   LM-EGCG ([91]Fig. 2K). Besides, negligible fluorescence was observed in
   other organs including the heart, liver, spleen, lung, and kidney,
   while the strongest fluorescence was detected in colonic tissues,
   demonstrating an enhanced aggregation of EGCG and Ga^3+ at inflammatory
   colonic tissues (fig. S4). Subsequently, we quantitatively estimated
   the Ga^3+ accumulation at the colitis site by inductively coupled
   plasma optical emission spectrometry (ICP-OES) analysis. As assumed,
   plenty of Ga^3+ was detected in the colon of the colitis mice compared
   to healthy mice after being orally administrated with LM-EGCG for 12
   hours ([92]Fig. 2L). Only a small amount of Ga^3+ was detected in the
   stomach and small intestine of both healthy or IBD mice. Besides,
   negligible Ga^3+ remained in the stomach and small intestine of IBD
   mice after being treated with Ga^3+, LM, or LM-EGCG (fig. S5A). In
   contrast, massive Ga^3+ was detected in colon tissues of IBD mice after
   being treated with LM-EGCG, indicating that EGCG could capture Ga^3+ to
   form complexes and augment the retention of Ga^3+ at the colitis site.
   To verify that Ga^3+ principally accumulated at the inflammatory tissue
   of the colon, the main organs (heart, liver, spleen, lung, and kidney)
   were collected and analyzed by ICP-OES. Except for gastrointestinal
   tissue, scarce Ga^3+ was detected in the main organs through blood
   circulation after oral gavage of Ga^3+, LM, or LM-EGCG (fig. S5, B to
   D). These results demonstrate that EGCG could pre-target colonitis and
   capture Ga^3+ to form a complex and augment the retention of Ga^3+ at
   the colitis sites.

   Next, to investigate the chemical state and coordination environment of
   the complex, we collected colonic contents from the IBD mice after oral
   gavage of LM-EGCG to further analyze the status of gallium by using
   edge-extended x-ray absorption fine structure (EXAFS), high-angle
   annular dark-field scanning (HAADF) TEM, and energy-dispersive x-ray
   spectroscopy (EDS) element mapping. According to the Fourier
   transformation into the R-space, the extended Ga K-edge EXAFS curves
   revealed that the main peak of the colonic contents emerged at ∼1.4 Å,
   which was similar to Ga-O in the metal polyphenol complex in SCF rather
   than Ga-Ga in Ga foil or LM-EGCG (∼2.5 Å) ([93]Fig. 2M). As shown in Ga
   K-edge x-ray absorption near-edge structure (XANES) spectra, the
   absorption edge of colonic contents is nearest to the complex of EGCG
   and Ga in SCF, indicating that Ga species integrated with lighter atoms
   and positively charged with a chemical state. Consistent with the
   result in Ga K-edge XANES spectra, EXAFS wavelet transform (WT)
   analysis and Fourier transforms of k3-weighted Ga K-edge EXAFS spectra
   revealed the direct coordination of Ga-to-O atom without the formation
   of Ga-Ga metal–derived crystalline structures in colonic contents
   ([94]Fig. 2, O and P). In addition, as the TEM and HAADF revealed, the
   morphology of the colonic contents was much different from LM-EGCG, and
   the signal of Ga element was obviously detected in the colonic contents
   from IBD mice after being treated with LM-EGCG by EDS element mapping
   ([95]Fig. 2Q). Collectively, these results convincingly demonstrated
   the in situ coordination of Ga and EGCG in IBD mice after oral gavage
   of LM-EGCG.

Effective alleviation of DSS-induced colitis

   Next, we evaluated the therapeutic efficacy of LM-EGCG against
   DSS-induced acute colitis ([96]Fig. 3A). In comparison with other
   treatments, LM-EGCG evidently restrained the body weight loss of
   DSS-induced colitis mice after the secondary treatment ([97]Fig. 3D)
   and reduced the hematochezia occurrence during treatment (fig. S6). In
   addition, the colon length of mice treated with LM-EGCG was much longer
   than those treated with 5-aminosalicylic acid (5-ASA), EGCG, Ga^3+ LM,
   or Ga-EGCG ([98]Fig. 3, B and C). Moreover, compared to other
   treatments, scarcely any red fluorescence from dihydroethidium (DHE)
   staining was observed in colonic tissues in IBD mice after being
   treated with LM-EGCG (fig. S7A). As shown in the results of terminal
   deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick
   end labeling (TUNEL), consistent with mice in the healthy group,
   ignorable green fluorescence was detected in LM-EGCG–treated mice (fig.
   S7B), confirming that LM-EGCG could protect the intestinal villus from
   further injury. Furthermore, LM-EGCG treatment furthest augmented the
   expression level of proliferating cell nuclear antigen in DSS-induced
   colitis mice (fig. S7C), suggesting that the impaired colonic tissues
   were renovated after oral gavage of LM-EGCG.

Fig. 3. LM-EGCG efficiently mitigates DSS-induced colitis in a murine model.

   [99]Fig. 3.
   [100]Open in a new tab

   (A) The C57BL/6 mice were orally administered with water, 5-ASA, EGCG,
   Ga^3+, LM, Ga-EGCG, or LM-EGCG on days 5, 7, and 9 after being treated
   with water or 2.5% DSS-containing water for 7 days. (B and C)
   Photographs of colon tissues obtained after various treatments on day
   10 and (C) the corresponding colon length (n = 6). (D) Daily body
   weight changes in each group for 10 days. (E) Bio-TEM images of colon
   tissues obtained from healthy, DSS-treated, or DSS + LM-EGCG–treated
   mice. Scale bar, 1 μm. (F and G) The colon tissues were collected for
   immunofluorescence staining of ZO-1 (pink), claudin-1 (green), and
   occludin (red) at the end of treatment and (G) the corresponding
   semiquantitative MFI analysis. Scale bar, 100 μm. (H to M) Relative (H)
   myeloperoxidase (MPO), (I) interleukin-6 (IL-6), (J) tumor necrosis
   factor–α (TNF-α), (K) IL-1β, (L) IL-10, and (M) transforming growth
   factor–β (TGF-β) levels in the collected colon tissues by enzyme-linked
   immunosorbent assay (ELISA) analysis (n = 5). (N) Schematic diagram of
   the mechanism of LM-EGCG for colitis remission. Data are presented as
   means ± SD. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001,
   analyzed by [(C), (G), and (H) to (M)] one-way or (D) two-way ANOVA.
   ns, not significant.

   Furthermore, the colon tissues were analyzed by Bio-TEM after treatment
   to assess intestinal villus recovery of DSS-induced colitis mice
   ([101]Fig. 3E). Compared to healthy mice that fed with water, extensive
   damage to intestinal villi was revealed in the mice fed with 2.5% DSS
   in drinking water, indicating apparent damages to the intestinal tissue
   of DSS-induced colitis mice. Notably, the length or morphology of
   damaged intestinal villi in DSS-induced colitis mice was restored after
   being treated with LM-EGCG, revealing LM-EGCG’s therapeutic effects
   against DSS-induced colitis. Meanwhile, immunofluorescent staining
   analysis on expression of tight junction-correlative proteins,
   including zonula occludens-1 (ZO-1) (pink; fig. S8), claudin-1 (green;
   fig. S9), and occludin (ted; fig. S10), was used to assess the repair
   effects of various treatments on DSS-inflamed colonic epithelium
   ([102]Fig. 3F). Notably, brightest pink, green, and red fluorescence
   were concurrently observed on the healthy mice and DSS-induced colitis
   mice that orally administered with LM-EGCG, while relatively feeble
   fluorescence was detected in other control groups, including 5-ASA,
   EGCG, LM, and Ga-EGCG treatment ([103]Fig. 3G), indicating that LM-EGCG
   could normalize the expression of the tight junction-associated
   proteins and efficiently repair the DSS-inflamed intestinal epithelial
   barrier in DSS-induced colitis mice.

Remission of colonic inflammation

   The overexpression of pro-inflammatory cytokines, such as interleukin-6
   (IL-6), IL-1β, and tumor necrosis factor–α (TNF-α), could activate an
   inflammatory cascade response in IBD. We thus analyzed the trends of
   pro-inflammatory or anti-inflammatory cytokines expression levels in
   colonic tissues after various treatments using enzyme-linked
   immunosorbent assay (ELISA). As expected, compared to other control
   groups, DSS-induced colitis mice orally administered with LM-EGCG not
   only furthest decreased the local level of IBD-associated
   myeloperoxidase (MPO) in the colon ([104]Fig. 3H) but also reduced the
   expression of pro-inflammatory cytokines, including IL-6, TNF-α, and
   IL-1β, while increasing the expression of anti-inflammatory cytokines
   IL-10 and transforming growth factor–β (TGF-β) ([105]Fig. 3, I to M).
   To further evaluate the effect of LM-EGCG on inflammatory regulation in
   DSS-induced colitis mice, we detected the local levels of a
   pro-inflammatory M1 macrophages’ marker, inducible nitric oxide
   synthase (iNOS; red), and an M2 macrophages’ marker, arginase-1 (Arg-1;
   green), in colon tissue by immunofluorescence analysis (fig. S11).
   Consistent with the result of ELISA, low levels of Arg-1 and iNOS were
   detected in DSS-induced colitis mice with oral administration of
   LM-EGCG compared to other control groups. Besides, according to the
   results in immunohistochemistry analysis, LM-EGCG evidently reduced the
   CD86 level, which was similar to those of colonic tissues in healthy
   mice (fig. S12), suggesting that LM-EGCG could evidently extinguish the
   hyperactive inflammatory response in DSS-induced colitis mice. These
   results strongly indicate that LM-EGCG could efficiently lower the
   pro-inflammatory cytokines and inactivate inflammatory responses,
   thereby relieving DSS-induced IBDs.

Therapeutic mechanisms of LM-EGCG on IBD

   To further illustrate the therapeutic mechanism of the LM-EGCG to
   DSS-induced colitis, transcriptome analysis of colonic tissues was
   conducted at the end of treatment. As shown in [106]Fig. 4 (A and B),
   compared to the healthy mice, 966 up-regulated genes and 1138
   down-regulated genes in DSS-induced colitis mice were detected.
   Meanwhile, 744 genes were up-regulated, and 960 genes were
   down-regulated after the regulation of LM-EGCG. The results in the
   Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment
   analysis showed that the IL-17, inflammatory response, and TNF
   signaling pathways are overexpressed in DSS-induced colitis mice
   compared to healthy mice or LM-EGCG–treated IBD mice ([107]Fig. 4, D
   and E), suggesting that those of signaling pathways are highly
   associated with the therapeutic mechanisms of LM-EGCG. Furthermore,
   compared to healthy mice, the normalized heatmap revealed that the
   genes related to inflammatory response were highly expressed in
   DSS-induced colitis, which was evidently decreased after LM-EGCG
   treatment ([108]Fig. 4C). Similarly, LM-EGCG treatment decreased the
   expression level of the genes related to ROS in DSS-induced colitis
   mice ([109]Fig. 4F). Besides, according to KEGG pathway function
   enrichment, the LM-EGCG treatment increased the expression level in Wnt
   signaling pathway, which was beneficial to repair of intestinal
   epithelial damage ([110]46), in comparison with DSS-induced colitis
   mice ([111]Fig. 4G). Synchronously, the tight junction-associated genes
   were markedly up-regulated in DSS-induced mice under LM-EGCG treatment
   ([112]Fig. 4H). Consistently, the above evidence convincingly indicated
   that LM-EGCG could effectively eliminate oxidative stress, relieve
   inflammation, and recover the intestinal barrier.

Fig. 4. Transcriptomic analysis of DSS-induced colitis in mice after the
regulation of LM-EGCG.

   [113]Fig. 4.
   [114]Open in a new tab

   (A and B) Volcano plot exhibiting the differentially expressed genes in
   (A) the comparison of DSS with the healthy group and (B) the DSS versus
   the LM-EGCG–treated group from RNA sequencing data (red: up-regulated
   genes; green: down-regulated genes). (C) Heatmap analysis of the
   expression of the inflammatory response genes in mice after being
   treated with LM-EGCG. (D and E) Kyoto Encyclopedia of Genes and Genomes
   pathway enrichment analysis of the up-regulated genes in (D) the
   comparison of DSS with the healthy group and (E) in the DSS versus the
   LM-EGCG–treated mice. The 20 most significantly enriched pathways were
   shown. (F to H) Heatmap analysis of the expression of the (F) response
   to ROS, (G) Wnt signaling pathways, and (H) cell tight junction
   signaling pathways in the healthy, IBD, LM-EGCG–treated mice. Data
   represent at least five independent replicates.

Modulation of LM-EGCG on gut microbiome

   Reinstating disordered gut microbiota via imposing restrictions on the
   abnormal expansion of cantankerous microbes is beneficial to relieving
   IBD. Encouraged by the favorable alleviation of colonic inflammation,
   we investigated the gut microbiome variation by 16S ribosomal RNA
   (rRNA) gene sequencing in the V4 region of fecal samples from the
   experimental mice after various treatments on day 10. Compared to
   DSS-induced colitis mice treated with EGCG or 5-ASA, bacterial richness
   (observed operational taxonomic unit richness) and diversity (Shannon
   and Simpson) in DSS-induced colitis mice treated with Ga-based
   nano-agents were notably increased and similar to healthy mice
   ([115]Fig. 5, A to D). The principal coordinates analysis revealed that
   the composition of gut microbiota in DSS-induced colitis mice treated
   with LM-EGCG was closest to the healthy mice ([116]Fig. 5F).
   Furthermore, the analysis of microbiota composition revealed that
   LM-EGCG treatment decreased the relative abundance of
   Enterobacteriaceae and Bacteroidaceae but increased the proportion of
   Lactobacillaceae and Verrucomicrobiaceae ([117]Fig. 5E).

Fig. 5. LM-EGCG regulates the gut microbiome dysbiosis and improves the
intestinal microenvironment.

   [118]Fig. 5.
   [119]Open in a new tab

   Feces from the treated mice were collected for gut microbiome analysis
   by 16S rRNA sequencing. (A) Estimation of microbial community observed
   operational taxonomic unit (OTU) richness. (B to D) Alpha-diversity was
   displayed by (B) chao1, (C) Shannon, and (D) Simpson. (E) Relative
   abundance of the gut microbiome is presented at family-level taxonomy.
   (F) Principal coordinates analysis (PCoA) illustrating the β-diversity
   of the gut microbiome. Each point represents each mouse. (G to J)
   Relative abundance of selected taxa (G) Escherichia_Shigella, (H)
   Bacteroides, (I) Akkermansia, and (J) Bifidobacterium. (K and L)
   Photographs of bacterial colonies of E. coli after being treated with
   different concentrations of Ga^3+, and (L) the corresponding colony
   count. (M) Short-chain fatty acid levels in feces from experimental
   mice after various treatments by using gas chromatography–mass
   spectrometry. Data are presented as means ± SD (n = 5). *P < 0.05, **P
   < 0.01, ***P < 0.001, and ****P < 0.0001, analyzed by one-way ANOVA.

   Unexpectedly, among the variation of microbiota composition, the
   overexpansion of Escherichia_Shigella (known to induce diarrhea and
   IBDs) and Bacteroides (known to be associated with the progression of
   DSS-induced IBDs) was terminated in DSS-induced colitis mice with oral
   administration of LM-EGCG ([120]Fig. 5, G and H), while the populations
   of Akkermansia muciniphila (AKK; known to protect intestinal barrier
   functions) and Bifidobacterium (known for increasing expression levels
   of junction proteins) were increased ([121]Fig. 5, I and J). Moreover,
   for further investigation of the impact of Ga^3+ on gut microbiota, we
   counted the colony-forming units (CFUs) of Escherichia coli after Ga^3+
   treatment ([122]Fig. 5K). Approximately 60% of E. coli was inhibited as
   the concentration of Ga^3+ reached 64 mg liter^−1 ([123]Fig. 5L and
   fig. S13). On the contrary, compared with the aseptic culture medium
   containing Ga^3+ (fig. S14A), the AKK-inoculated culture medium was
   evidently turbid (fig. S14B). Meanwhile, the corresponding OD[600]
   (optical density at 600 nm) of AKK was almost unchanged with the
   increase of Ga^3+ concentration, indicating that Ga^3+ showed an
   insignificant impact on the growth and proliferation of AKK (fig.
   S14C). Collectively, these results indicate that LM-EGCG is capable of
   ameliorating disordered gut microbiota, and is beneficial for improving
   colitis.

   Short-chain fatty acids (SCFAs), produced by probiotic gut microbiota,
   have commendable anti-inflammatory capacities and can enhance the
   function of the regulatory T cell to prevent an excessive immune
   response for effective amelioration of colitis ([124]47, [125]48).
   Encouraged by the gut microbiome modulation, the feces SCFA levels in
   experimental mice were investigated after the secondary gavage of
   LM-EGCG. As shown in [126]Fig. 5M, LM-EGCG treatment significantly
   increased the total SCFA levels in comparison with water-treated IBD
   mice. Particularly, LM-EGCG treatment markedly improved the abundance
   of acetic, n-butyric, and propionic acid, which approached those of
   mice in the healthy group ([127]Fig. 5M and fig. S15). Moreover, the
   production of branched SCFAs, including isobutyric, isovaleric, and
   2-methylbutyric acid, were also elevated after LM-EGCG treatment (fig.
   S15). In short, LM-EGCG treatment ameliorated the gut microbiota to
   promote the generation of SCFA levels, which could contribute to the
   amelioration of colitis in mice.

Amelioration of colitis after antibiotic pretreatment

   Then, we assessed the efficacy of LM-EGCG against colitis in mice after
   pretreatment with a cocktail of antibiotics. The mice were orally
   administered with a cocktail of antibiotics for 5 days to disrupt gut
   commensal microbes before being fed with DSS ([128]Fig. 6A). Although
   LM-EGCG could suspend the further shortening of colon length, the colon
   was still significantly shorter than the colon of healthy mice
   ([129]Fig. 6, B and C). LM-EGCG treatment increased the expression of
   tight junction-associated proteins ([130]Fig. 6, D and E), including
   ZO-1 (pink; fig. S16), claudin-1 (green; fig. S17), and occludin (red;
   fig. S18), and restored the damaged intestinal villi ([131]Fig. 6F).
   Compared to control groups, DSS-induced colitis mice treated with
   LM-EGCG not only furthest decreased the local level of IBD-associated
   MPO in the colon ([132]Fig. 6G) but also reduced the expression of
   pro-inflammatory cytokines, including IL-6, TNF-α, and IL-1β, while
   increasing the expression of anti-inflammatory cytokines IL-10 and
   TGF-β ([133]Fig. 6, H to L). Furthermore, LM-EGCG treatment markedly
   weakened the fluorescence from the pro-inflammatory M1 macrophage
   marker, iNOS (red), and the M2 macrophage marker, Arg-1 (green), in
   comparison with other control groups (fig. S19). Together, LM-EGCG
   treatment could weaken the inflammatory response and restore the
   intestinal barrier in DSS-induced colitis mice, whereas the efficacy of
   LM-EGCG against DSS-induced colitis was discounted without the
   assistance of modulated gut microbiome ([134]4, [135]49).

Fig. 6. LM-EGCG mitigates colitis in a murine model that was pretreated with
a cocktail of antibiotics.

   [136]Fig. 6.
   [137]Open in a new tab

   (A) C57BL/6 mice were pretreated with a cocktail of antibiotics for 5
   days and then orally administered with water or EGCG, Ga^3+, or LM-EGCG
   on days 5, 7, and 9 after being treated with water or 2.5%
   DSS-containing water for 7 days. (B and C) Photographs of colon tissues
   obtained after various treatments on day 10 and (C) the corresponding
   colon length (n = 6). (D and E) The colon tissues were collected for
   immunofluorescence staining (D) of ZO-1 (pink), claudin-1 (green), and
   occludin (red) at the end of treatment, and (E) the corresponding
   semiquantitative MFI analysis. Scale bar, 100 μm. (F) Bio-TEM images of
   colon tissues obtained from healthy, DSS-treated, or DSS +
   LM-EGCG–treated mice. Scale bar, 1 μm. (G to L) Relative (G) MPO, (H)
   IL-6, (I) IL-1β, (J) TNF-α, (K) IL-10, and (L) TGF-β levels in the
   collected colon tissues by ELISA analysis (n = 5). Data are presented
   as means ± SD. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001,
   analyzed by one-way ANOVA.

Amelioration of severe colitis with delayed treatment

   Encouraged by the excellent gut microbiota regulation and colitis
   amelioration capacity of LM-EGCG, we further investigated LM-EGCG
   against the severe colitis induced by DSS. Before LM-EGCG treatment,
   the mice were fed with 2.5% DSS containing water for 10 days ([138]Fig.
   7A). At the end of DSS feeding, various agents were orally
   administrated on days 10, 12, and 14. LM-EGCG treatment halted the
   sharp downward tendency of body weight of colitis mice, which then
   restored to the beginning of the DSS feeding (fig. S20).
   Understandably, LM-EGCG not only renovated the impaired colon tissues
   ([139]Fig. 7B), which approached the healthy mice ([140]Fig. 7C), but
   also commendably increased the expression of the tight
   junction-associated proteins ([141]Fig. 7, D and E), including ZO-1
   (pink; fig. S21), claudin-1 (green; fig. S22), and occludin (red; fig.
   S23). Furthermore, compared to control groups, DSS-induced colitis mice
   treated with LM-EGCG not only recovered the damaged intestinal villi
   ([142]Fig. 7F) but also maximally decreased the local level of
   IBD-associated MPO in the colon ([143]Fig. 7G) and reduced the
   expression of pro-inflammatory cytokines, including IL-6, TNF-α, and
   IL-1β. Simultaneously, the expression of anti-inflammatory cytokines
   IL-10 and TGF-β was increased ([144]Fig. 7, H to L). Similarly, LM-EGCG
   treatment sharply weakened the fluorescence from the pro-inflammatory
   M1 macrophage marker, iNOS (red), and the M2 macrophage marker, Arg-1
   (green), in comparison with other control groups (fig. S24).
   Satisfactorily, LM-EGCG treatment could weaken the inflammatory
   response and restore the intestinal barrier in DSS-induced colitis
   mice, thus relieving DSS-induced IBD.

Fig. 7. LM-EGCG exerts satisfactory therapeutic efficacy in a murine model of
DSS-induced colitis with a delayed therapeutic setting.

   [145]Fig. 7.
   [146]Open in a new tab

   (A) C57BL/6 mice were orally administered with water or EGCG, Ga^3+, or
   LM-EGCG on days 10, 12, and 14 after being treated with water or 2.5%
   DSS-containing water for 10 days. (B and C) Photographs of colon
   tissues obtained after various treatments on day 10 and (C) the
   corresponding colon length (n = 6). (D and E) The colon tissues were
   collected for immunofluorescence staining of ZO-1 (pink), occludin-1
   (green), and occludin (red) at the end of treatment and (E) the
   corresponding semiquantitative MFI analysis. Scale bar, 100 μm. (F)
   Bio-TEM images of colon tissues obtained from healthy, DSS-treated, or
   DSS + LM-EGCG–treated mice. Scale bar, 1 μm. (G to L) Relative (G) MPO,
   (H) IL-6, (I) IL-1β, (J) TNF-α, (K) IL-10, and (L) TGF-β levels in the
   collected colon tissues by ELISA analysis (n = 5). Data are presented
   as means ± SD. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001,
   analyzed by one-way ANOVA.

Biocompatibility of LM-EGCG

   First, the cell viability after incubation with various agents was
   investigated. Markedly, more than 80% of NCM460 cells were alive when
   the concentration of Ga^3+ reached 400 μg ml^−1 ([147]Fig. 8A).
   Correspondingly, almost all NCM460 cells were alive after being treated
   with 100 μg ml^−1 EGCG (the loading rate of EGCG from LM-EGCG was about
   20%) ([148]Fig. 8B). As expected, Ga-EGCG or LM-EGCG exhibited no
   impact on the growth and proliferation of NCM460 cells ([149]Fig. 8C).
   Consistent with the results detected by cytotoxicity assay, massive
   green fluorescence was observed from living cells while scarcely any of
   red fluorescence was detected in those of NCM460 cells even the
   concentration of LM-EGCG reached up to 800 μg ml^−1 ([150]Fig. 8E and
   fig. S25), convincingly demonstrated that Ga^3+ or LM-EGCG was
   nontoxic. In addition, compared with the control group, no apparent
   hemolysis was observed in those blood samples incubated with EGCG,
   Ga^3+, Ga-EGCG, or LM-EGCG, respectively, even though the concentration
   of the agents arrived at 400 μg ml^−1 ([151]Fig. 8F and fig. S26).

Fig. 8. LM-EGCG shows favorable biosecurity.

   [152]Fig. 8.
   [153]Open in a new tab

   (A to C) In vitro cell viabilities of NCM460 cells (normal colonic
   epithelial cells) after incubation with different concentrations of (A)
   Ga^3+, (B) EGCG, and (C) Ga-EGCG or LM-EGCG for 24 hours. (D) Ga
   content in feces after oral administration of Ga^3+, LM, or LM-EGCG.
   (E) Fluorescence images of Ga^3+, Ga-EGCG, or LM-EGCG–treated NCM460
   cells stained with calcein AM (green fluorescence, living cells) and PI
   (red fluorescence, dead cells). Scale bar, 100 μm. (F) Photographs of
   the microcentrifuge tubes containing blood samples with various
   treatments. (I) H[2]O, (II) EGCG, (III) Ga^3+, (IV) Ga-EGCG, and (V)
   LM-EGCG. (G to I) Renal function was evaluated by monitoring the (G)
   creatinine (Crea), (H) blood urea (Urea), and (I) glucose levels in the
   blood after various treatments. (J to N) Liver function was evaluated
   by monitoring the important markers of (J) alanine aminotransferase
   (ALT), (K) total protein (TP), (L) alkaline phosphatase (ALP), (M)
   albumin (ALB), and (M) total bilirubin (TBIL) levels in the blood after
   various treatments.

   Thereafter, in consideration of the favorable degradation property in
   the acid environment of LM ([154]50), we detected the Ga content in
   blood, urine, and excretion from normal mice after being orally treated
   with LM-EGCG. Apparently, insignificant Ga^3+ was detected in blood and
   urine from DSS-induced colitis mice after oral gavage of LM-EGCG (fig.
   S27). The majority of Ga^3+ in LM or LM-EGCG was excreted via excrement
   in healthy mice after oral gavage of the above agents for 24 hours
   ([155]Fig. 8D). Moreover, after the delayed treatment, the hematoxylin
   and eosin (H&E) staining analysis of major organs (heart, liver,
   spleen, lung, and kidney) revealed that no pathological abnormalities
   in the organs were detected, suggesting that LM-EGCG had no obvious
   tissular toxicity on those of experimental mice during the treatment
   (fig. S28). Furthermore, the liver and kidney functions were applied to
   further assess the biosafety of LM-EGCG by detecting the crucial
   detection indicators containing creatinine (Crea), blood urea (Urea),
   glucose (GLU), alanine aminotransferase (ALT), total protein (TP),
   alkaline phosphatase (ALP), albumin (ALB), and total bilirubin (TBIL)
   ([156]51). As important biomarkers of kidney function, Crea, Urea, and
   GLU were rectified to normal range after the remission of enteritis in
   comparison with other control groups ([157]Fig. 8, G to I). Similarly,
   the important markers of liver function including ALT, TP, ALP, ALB,
   and TBIL in LM-EGCG–treated mice were much closer to the healthy mice,
   demonstrating that LM-EGCG had no visible hepatic toxicity in mice
   ([158]Fig. 8, J to N). Collectively, these results confirmed that the
   constructed LM-EGCG had negligible side effects and good
   biocompatibility for biomedical applications.

DISCUSSION

   IBDs, including ulcerative colitis and Crohn’s disease, are chronic
   inflammatory diseases of the gastrointestinal tract, which will
   markedly lower the quality of life of patients and have been an
   emerging challenge for public health worldwide ([159]52, [160]53). The
   overproduced ROS and overexpansion of profitless bacteria are typical
   pathological features of colitis ([161]54). The hyperactive ROS not
   only provokes the inflammatory cells to produce a great deal of ROS
   during the progress of colitis to hurt the intestinal epithelial cells
   but also restrains the colonization of beneficial bacteria at the
   inflammatory site ([162]55). The abnormal expansion of harmful
   bacteria, such as E. coli, will occupy the ecological niches to prevent
   the colonization of beneficial bacteria and the reprogramming of the
   microbiome during inflammation, thereby impeding the recovery of the
   colitis microenvironment ([163]56, [164]57). In our work, the reductive
   EGCG, a polyphenol in green tea ([165]42), was applied to eliminate the
   overproduced ROS. Meanwhile, Ga-based LM was used as an antibiotic-free
   intestinal flora regulator to modulate the dysregulated microbiome for
   effective remission of DSS-induced colitis. Our results indicate that
   the strategy of working along both lines to eliminate the overproduced
   ROS and simultaneously modulate intestinal flora could efficiently
   ameliorate the colitis microenvironment.

   Targeting drug delivery in response to the pathological
   microenvironment provides opportunities for diminishing the off-target
   systemic side effects and improving the therapeutic effect of IBD
   treatments. The inflamed colon is characterized by the imbalance of the
   gut microbiome, hyperactive immune responses, elevated levels of ROS,
   and enrichment of positively charged proteins ([166]58, [167]59).
   Exploiting ROS-induced targeting retention of therapeutic agents at
   inflammatory sites is an optional approach to meet the medical need of
   improving drug bioavailability ([168]60, [169]61). Green tea polyphenol
   (EGCG) has a polyhydroxy structure that can adhere to the innate amine,
   thiol, carboxyl groups, and electropositive biomolecule making EGCG a
   potential therapeutic drug for gastrointestinal disease
   ([170]43–[171]45). In addition to the adhesion behavior, the EGCG would
   be polymerized during the process of scavenging ROS. In this study, we
   used EGCG as a pre-targeting module to accumulate at the inflammation
   site for scavenging the overproduced ROS and synchronously capture the
   metal ions (Ga) to form metal-phenolic complexes in situ at the colonic
   microenvironment. The result revealed that the detained metal-phenolic
   complexes elevated the gathering of EGCG and Ga at the colitic area to
   balance the oxidative stress and protect the damaged intestinal barrier
   from graver injury.

   Gut microbiota dysbiosis is closely related to the progress of colitis,
   and increasing evidence has denoted that regulating the gut microbiota
   is highly beneficial to the alleviation of IBDs ([172]62, [173]63). The
   inconducive microbes became dominant bacteria and the population of
   beneficial bacteria decreased during the occurrence of colitis
   ([174]64, [175]65). Toward the goal of frenzied expansion, the
   pathogenic bacteria require a lot of nutriment to meet their high level
   of metabolism. As an indispensable nutrient for bacterial survival,
   proliferation, and bacterial infection, iron is involved in many vital
   metabolic processes, including energy generation and DNA synthesis
   ([176]66). The biological systems of bacteria are unable to distinguish
   gallium from iron ions and will specifically take in gallium through
   iron ion uptake pathways, making gallium a potential “tool” for
   bacterial regulation. We used Ga-based LM, which could be degraded in
   an acidic environment and released Ga^3+, to modulate the gut
   microbiome for the alleviation of colitis. Our results showed that,
   with the assistance of gut microbiome modulation, the designed LM
   nano-agents could efficiently relieve DSS-induced colitis.

   In summary, in consideration of the poor targeting accumulation
   capacity of conventional drugs that are difficult to concurrently
   modulate the dysregulated microbiome and simultaneously obliterate
   overproduced RONS in IBDs, we propose an amplifying targeting strategy
   to enhance the accumulation of therapeutic agents at the colitis site
   for IBDs remission. First, we validated the treatment efficacy of
   LM-EGCG against colitis in the process of forming enteritis. Then, to
   examine the contribution of intestinal flora to colitis, we examined
   the efficacy of LM-EGCG against colitis in mice after pretreatment with
   a cocktail of antibiotics. After that, the efficacy of LM-EGCG in
   delayed therapy was assessed on mice after being fed with 2.5% DSS
   containing water for 10 days. The consequences of toxicological
   evaluation demonstrate that LM-EGCG had negligible side effects on the
   experimental mice, indicating that LM is a promising candidate for a
   broad variety of biomedical applications including IBD alleviation.
   However, the lack of large animal models in the evaluation of
   therapeutic efficacy is one limitation of the current study due to the
   limitations of the mouse model in elucidating complex human pathology.
   We will further investigate the therapeutic efficacy of an LM-based
   system against colitis in large animals. In addition, in-depth studies
   of the molecular mechanisms in gut microbiome modulation in situ and
   systematic biocompatibility evaluation are necessary before the further
   application of LM-based nano-agents.

MATERIALS AND METHODS

Reagents

   Gallium-indium eutectic (99.99% LM) was purchased from Alfa Esha
   (China) Chemical Co., LTD. Methoxy poly(ethylene glycol) thiol
   (mPEG-SH; M[w] = 5 kDa) was purchased from Shanghai Ponsure Biotech,
   Inc. EGCG, 5-ASA, 1, DPPH, NBT, and
   3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-2H-tetrazolium bromide (MTT)
   were purchased from Aladdin Biochemical Technology Co., Ltd. (Shanghai,
   China). DSS salt colitis grade (36,000 to 50,000 Da) was purchased from
   MP Biomedicals (Santa Ana, CA, USA). Dulbecco’s modified Eagle’s medium
   (DMEM) and fetal bovine serum (FBS) were purchased from Gibco
   Laboratories (Grand Island, NY, USA). DCFH-DA, DAF-FM DA, calcein-AM
   (acetoxymethyl), bicinchoninic acid (BCA) kit, and propidium iodide
   (PI) were purchased from Beyotime Biotech Inc. (Shanghai, China). Radio
   immunoprecipitation assay (RIPA) was purchased from Wuhan Boerfu
   Biotechnology Co., Ltd. (Hubei, China). ELISA kits were purchased from
   4A Biotech Co., Ltd. (Jiangsu, China). Metronidazole, vancomycin,
   neomycin, and ampicillin were obtained from Shanghai Macklin
   Biochemical Technology Co., Ltd. (China). Cyanine 5.5 NHS ester (Cy5.5)
   was obtained from DuoFluor Inc. (Hubei, China). Tryptone and yeast
   extract were purchased from Oxoid Ltd. (China). E. coli, AKK, and
   AKK-corresponding bacterial culture medium were obtained from Beijing
   Beina Chuanglian Institute of Biotechnology (China).

Experimental apparatus

   TEM images were carried out by JEOL JEM 2100F. Particle size was
   measured by a dynamic light scatterer (Malvern Zetasizer Nano ZS90).
   TGA was performed by Mettler TGA. Ultraviolet absorbance was measured
   by PerkinElmer VICTOR Nivo. In vivo imaging experiments were performed
   on IVIS imaging systems (PerkinElmer). Cell fluorescence was observed
   by inverted fluorescence microscopy (Olympus). Inductively coupled
   plasma mass spectrometry (ICP-MS) was measured by Agilent 7800.

Preparation of EGCG-coated liquid metal nano-agents

   First, the LM nanoparticles were prepared according to our previous
   work ([177]36). Briefly, LM (200 mg) was added into 15 ml of deoxidized
   ultrapure water (UPW) containing mPEG-SH (70 mg), and the mixture was
   ultrasonically treated with a 6-mmφ probe at 80% intensity for 40 min
   in an ice bath. Then, LM (200 mg) was added into 15 ml of deoxidized
   UPW containing EGCG (70 mg) or Cy5.5-labeled EGCG, and the mixture was
   ultrasonically treated with a 6-mmφ probe at 80% intensity for 40 min
   in an ice bath. The mixture was centrifuged at 1000 rpm for 10 min, and
   the supernatant was collected and followed with 8000 rpm for 10 min to
   obtain LM-EGCG. The nano-agents were washed with UPW three times and
   stored at 4°C for further use. For the thermal gravimetric analysis of
   LM-EGCG, the obtained LM and LM-EGCG (100 mg) nanoparticles were
   analyzed by a thermal gravimetric analyzer under nitrogen conditions
   after lyophilization, and the weight loss was recorded and analyzed.

Self-assembly of Ga^3+ and EGCG in simulated colonic fluid

   To explore the self-assembly behavior of Ga^3+ and organic ligands,
   Ga(NO[3])[3] (2 mg ml^−1) was dropwise added into SCF containing EGCG,
   dopamine, sodium alginate, hyaluronic acid, trehalose anhydrous, or
   synanthrin (2 mg ml^−1), respectively. Next, the mixture was
   photographed at different time points (0.5, 1, 2, and 4 hours) and
   determined by a dynamic light scatterer to evaluate the generation of
   the Ga and EGCG complex. Afterward, we further assessed the
   self-assembly of Ga^3+ and EGCG in SCF. First, Ga(NO[3])[3] (2 mg
   ml^−1) or EGCG (2 mg ml^−1) was separately added into SCF, and then
   another Ga(NO[3])[3] (2 mg ml^−1) solution was added into SCF
   containing EGCG (2 mg ml^−1). Then, the mixtures were photographed at
   different time points (0.5, 1, and 2 hours) and determined by a dynamic
   light scatterer.

RONS-scavenging ability of LM-EGCG

   The ROS-scavenging capacity of LM-EGCG was tested by NBT. First, NBT,
   riboflavin, and methionine were severally dissolved in 0.05 M PBS to
   form 0.6 mg ml^−1, 49 mg ml^−1, and 3 μg ml^−1 stock solutions. After
   that, 0.3 ml of reaction solution (PBS, EGCG, LM, or LM-EGCG,
   respectively) was added to 1.8 ml of PBS containing 0.3 ml of
   riboflavin, methionine, and NBT stock solutions. Then, the absorbance
   of the mixtures was measured at 560 nm after light irradiation for 20
   min. Posteriorly, DPPH was applied to investigate the RNS-scavenging
   performance. Twenty microliters of reaction fluid (PBS, EGCG, LM, or
   LM-EGCG, respectively) was mixed with 180 μl of DPPH (0.12 mg ml^−1,
   soluble in methanol) and incubated for 3 to 30 min in the dark. Then,
   the absorbance of the mixture at 517 nm was measured by PerkinElmer
   VICTOR Nivo.

Intracellular RONS-scavenging capacity of LM-EGCG

   The NCM460 and HUVEC cells were cultured in DMEM containing 10% FBS,
   penicillin (100 U ml^−1), and streptomycin (100 mg ml^−1) in a cell
   incubator at 37°C with 5% carbon dioxide. Intracellular ROS level was
   detected using DCFH-DA. First, 1 × 10^5 NCM460 cells were seeded in
   six-well plates and cultured for 24 hours, and then the culture medium
   was replaced by fresh DMEM containing LM, EGCG, or LM-EGCG (Ga
   concentration, 100 μg ml^−1). After incubation for 1 hour, the cells
   were treated with H[2]O[2] (600 μM) for another 1 hour. The cells were
   observed in a fluorescence microscope after being stained with DCFH-DA
   for 30 min. Synchronously, DAF-FM DA was used to detect intracellular
   NO levels. HUVEC cells were incubated with LM, EGCG, or LM-EGCG (Ga
   concentration, 100 μg ml^−1) for 1 hour, respectively. Then, the cells
   were stained with DAF-FM DA after H[2]O[2] incubation for 1 hour,
   followed by fluorescence microscope observation.

Animals

   C57BL/6 mice (aged 6 to 8 weeks) with an average body weight of 20 g
   were purchased from Hubei Beiente Biotechnology Co., Ltd. All of the
   animal experiments were performed following the guidelines approved by
   the Institution Animal Care and Use Committee at Tongji Medical College
   (IACUC), Huazhong University of Science and Technology (Wuhan, China),
   and the corresponding IACUC number is 3501.

In vivo fluorescence imaging

   For in vivo fluorescence imaging, first, the mice were fed with DSS in
   drinking water after being pre-treated with fluorescent-free feed for 1
   week to build the IBD model. To investigate the augmented accumulation
   of LM-EGCG at the colonic site, IBD mice or healthy mice were treated
   with Cy5.5-labeled EGCG or LM-EGCG (Ga concentration, 5 mg kg^−1) by
   gavage. After 12 hours, the main organs including the heart, liver,
   spleen, lung, kidney, and gastrointestinal tissues for those mice, were
   collected for fluorescence imaging.

ICP-MS analysis of Ga^3+ in mice

   To investigate the enhancive accumulation of Ga in the inflamed site,
   DSS-treated mice or healthy C57BL/6 mice (n = 3) were orally
   administrated with Ga(NO[3])[3], LM, or LM-EGCG solution (Ga
   concentration, 5 mg kg^−1). After 12 hours, the mice were euthanized
   and the main organs (heart, liver, spleen, lung, and kidney), stomach,
   small intestine, and colon tissues were collected for the ICP-MS test
   to analyze the Ga^3+ content. The colonic contents were collected for
   EDS elements imaging and EXAFS analysis. For the purpose of testing the
   state of LM-EGCG in the stomach and the potential excretion way from
   the urine, the stomach contents, blood, and urine in DSS-induced
   colitis mice were collected for ICP analysis after oral gavage of
   LM-EGCG at different time points.

LM-EGCG relieves DSS-induced IBD

   C57BL/6 mice were acclimatized 1 week before random assignment to
   experimental groups (n = 6). The mice in the healthy group were treated
   with normal water only, while the other groups were fed with drinking
   water containing 2.5% (w/v) DSS for 8 days (days 0 to 7). Then, IBD
   mice were orally administered with PBS, 5-ASA (40 mg kg^−1), EGCG (5 mg
   kg^−1), Ga(NO[3])[3], LM, Ga-EGCG, or LM-EGCG (Ga concentration, 5 mg
   kg^−1) on days 5, 7, and 9, respectively. The body weight of all mice
   was recorded daily over the 10-day experiments. On day 10, fecal
   samples from those mice were collected for 16S rRNA analysis. Briefly,
   DNA was extracted and examined by 1% agarose gel electrophoresis.
   Extracted DNA was then rendered PCR (ABI GeneAmp 9700) which was
   specific to the V3 and V4 regions of the 16S rRNA gene and used to
   construct 16S rRNA libraries for community analysis by using TruSeqTM
   DNA Sample Prep Kit. The libraries were sequenced using the Mi-Seq
   Illumina sequencer. Meanwhile, all mice were euthanized on day 10, and
   the colonic tissues were collected for Bio-TEM imaging,
   immunofluorescence staining (ZO-1, occludin-1, occludin, iNOS, and
   Arg-1), H&E staining, and ELISA analysis. Colon tissues were
   homogenized in 1 ml of RIPA containing protease inhibitor, and then
   subjected to centrifugation at 5000g for 15 min at 4°C. Then, the
   supernatant was retained, and the concentration of protein was measured
   by BCA. Next, MPO, IL-6, TNF-α, IL-1β, IL-10, and TGF-β levels were
   determined by ELISA kit (n = 5).

Transcriptome sequencing analysis

   The feces were collected for SCFAs detection by using gas
   chromatography–mass spectrometry at the end of treatment. Thereafter,
   the experimental mice were euthanized to collect the frozen colon
   sections for transcriptomic analysis. After successful extraction by
   the TRIZol method, RNA was dissolved by adding 50 μl of DEPC-treated
   water. Subsequently, total RNA was identified and quantified using a
   Qubit fluorescence quantifier and a Qsep400 high-throughput biofragment
   analyzer. After passing the library check, the different libraries were
   sequenced in Illumina after pooling them according to the effective
   concentration and the target sequencing output data volume, yielding
   150–base pair paired-end reads. Four types of fluorescently labeled
   dNTPs, DNA polymerase, and junction primers were added to the
   sequencing flow cell for amplification. Data quality control was
   performed using fastp to remove reads with adapters. Paired-end reads
   were removed under the following conditions: when the number of N in
   any sequencing read exceeded 10% of the length of that read and when
   any sequencing read contained low-quality bases (Q ≤ 20) exceeding 50%
   of the length of that read. Subsequent analyses were based on clean
   reads. DESeq2 was used for differential gene expression analysis
   between two groups, and Benjamini and Hochberg correction was applied
   to P values. Corrected P values and log[2] fold change were used as
   thresholds for noteworthy differential expression. Enrichment analysis
   was performed on the basis of the hypergeometric test, with
   pathway-based hypergeometric distribution testing for KEGG and GO
   analysis.

Ga^3+ influences bacterial proliferation

   E. coli colonies were inoculated into centrifuge tubes (under a sterile
   environment) containing 10 ml of LB medium (10 g of tryptone, 5 g of
   yeast extract, 10 g of NaCl, and 1000 ml of water) and shaken at 37°C
   for 20 hours. AKK colonies were seeded in fluid thioglycollate medium
   under an oxygen-free environment and cultured at 37°C for 3 days. To
   investigate the influence of Ga^3+ on bacteria, E. coli and AKK
   cultures were diluted to 1 × 10^4 CFU ml^−1 and treated with various
   concentrations of Ga(NO[3])[3] (Ga concentration, 51, 64, 80, and 100
   mg liter^−1, respectively). Bacteria were calculated by measuring the
   absorbance of suspension and the CFUs were quantified by the
   spread-plate method.

LM-EGCG alleviates antibiotics-pretreated IBD

   Before feeding with DSS, C57BL/6 mice (aged 6 to 8 weeks) were
   pretreated with a cocktail of antibiotics (metronidazole, 1 g liter^−1;
   vancomycin, 0.5 g liter^−1; neomycin, 1 g liter^−1; and ampicillin, 1 g
   liter^−1) in drinking water for 5 days. Then, IBD mice were orally
   administered with PBS, EGCG (5 mg kg^−1), Ga(NO[3])[3], or LM-EGCG (Ga
   concentration, 5 mg kg^−1) on days 5, 7, and 9, respectively. On day
   10, all mice were euthanized, and the main organs (liver, spleen,
   kidney, heart, and lung) were collected for H&E staining. Colonic
   tissues were collected for Bio-TEM imaging, immunofluorescence staining
   (ZO-1, occludin-1, occludin, iNOS, and Arg-1), H&E staining, and ELISA
   analysis.

Delayed treatment of IBD

   After being fed with DSS for 11 days, the mice were orally administered
   with PBS, EGCG (5 mg kg^−1), Ga(NO[3])[3], or LM-EGCG (Ga
   concentration, 5 mg kg^−1) on days 10, 12, and 14, respectively. On day
   15, all mice were euthanized, and the colonic tissues were collected
   for immunofluorescence staining (ZO-1, occludin-1, occludin, iNOS, and
   Arg-1), Bio-TEM imaging, H&E staining, and ELISA analysis.

Cytotoxicity evaluation of LM-EGCG

   The cytotoxicity of LM-EGCG against NCM460 cells was evaluated by
   live/dead staining and MTT assay. First, 1 × 10^5 NCM460 cells were
   seeded in six-well plates for 24 hours and treated with Ga(NO[3])[3],
   Ga-EGCG, LM, or LM-EGCG (Ga concentration, 200, 400, and 800 μg ml^−1,
   respectively) for 24 hours. Then, the cells were observed by a
   fluorescence microscope after being stained with calcein-AM and PI.
   Meanwhile, an MTT assay was conducted to assess the cytotoxicity of
   LM-EGCG. A total of 5 × 10^3 NCM460 cells were seeded in 96-well plates
   and cultured for 24 hours. Then, cells were treated with Ga(NO[3])[3],
   EGCG, Ga-EGCG, or LM-EGCG and incubated for 24 hours. After MTT
   addition for 4 hours, the reagent was discarded, and 150 μl of DMSO was
   added to the 96-well plates. Last, the OD values of each well were
   measured at 570 nm on an enzyme standard. The cell viability
   calculation is as follows: (OD[samples] − OD[DMSO])/(OD[control] −
   OD[DMSO]) × 100%.

Safety assessment of LM-EGCG

   The blood of healthy mice or IBD mice (n = 4) was collected after
   inflammatory therapy to detect the liver function index including ALT,
   TBIL, ALB, ALP, and TP and kidney function index containing Urea, Crea,
   and GLU variation. Besides, the red blood cells derived from healthy
   mice were used for hemolysis experiments. The red blood cells were
   treated with H[2]O, PBS, Ga(NO[3])[3], LM, EGCG, Ga-EGCG, and LM-EGCG
   with different concentrations (Ga concentration, 100, 200, and 400 μg
   ml^−1, respectively). After incubation for 1 hour, the mixture was
   centrifuged (1000g, 3 min) to observe the color of the supernatant. To
   estimate the excretion of LM-EGCG in the digestive tract, healthy
   C57BL/6 mice (n = 3) were oral gavage of Ga(NO[3])[3], LM, or LM-EGCG
   (Ga concentration, 200 μg ml^−1), respectively, and the feces from
   those of mice was collected at different time points for ICP-MS
   analysis.

Statistical analysis

   Statistical analysis was performed by Origin 9.0 or Prism 7.0
   (GraphPad, USA). All data were presented as means ± SD. The results
   were analyzed with one-way or two-way analysis of variance (ANOVA) with
   the least significance difference test among the three groups.
   Differences were considered statistically significant if P < 0.05 (*P <
   0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001).

Acknowledgments