我想要这篇章Biomaterial-mediated modulation of oral microbio

　　你好根据你的需要，学术堂为你查找到《Biomaterial-mediated modulation of oral microbiota synergizes with PD-1 blockade in mice with oral squamous cell carcinoma》这篇全文PDF，下面展示部分内容，查看全文请在文末自行下载，希望对你有所帮助！

Biomaterial-mediated modulation of oral microbiota synergizes with PD-1 blockade in mice with oral squamous cell carcinoma

Di-Wei Zheng1,3, Wei-Wei Deng2,3, Wen-Fang Song1, Cong-Cong Wu2, Jie Liu2, Sheng Hong1,Ze-Nan Zhuang1, Han Cheng1, Zhi-Jun Sun?2?and Xian-Zheng Zhang-1

　　Because a host’s immune system is affected by host–microbiota interactions, means of modulating the microbiota could beleveraged to augment the effectiveness of cancer therapies. Here we report that patients with oral squamous cell carcinoma(OSCC) whose tumours contained higher levels of bacteria of the genus Peptostreptococcus had higher probability of long-termsurvival. We then show that in mice with murine OSCC tumours injected with oral microbiota from patients with OSCCs, antitu-mour responses were enhanced by the subcutaneous delivery of an adhesive hydrogel incorporating silver nanoparticles (whichinhibited the growth of bacteria competing with Peptostreptococcus) alongside the intratumoural delivery of the bacteriumP. anaerobius (which upregulated the levels of Peptostreptococcus). We also show that in mice with subcutaneous or orthotopicmurine OSCC tumours, combination therapy with the two components (nanoparticle-incorporating hydrogel and exogenousP. anaerobius) synergized with checkpoint inhibition with programmed death-1. Our findings suggest that biomaterials can bedesigned to modulate human microbiota to augment antitumour immune responses.

　　mmunotherapy can trigger substantial immune responses against various cancers 1–4. Yet for oral squamous cell carcinoma (OSCC),the 5 yr survival rate remains around 50% (refs.5,6). Owing to high antigen loads in OSCC, immunotherapy is expected to have considerable therapeutic effects 7,8. However, the responsiveness of OSCC to pembrolizumab immunotherapy is only 14.6% (ref.9).

　　The contradiction highlights an urgent need for strategies to elicit more efficient immune responses in OSCC 10. Recently, several studies have reported an association between the regulation of symbiotic bacteria and the responsiveness of the host immune sys-tem 11–15. This suggests that the modulation of the oral microbiota could improve immune responses to OSCC. However, there is still a paucity of research exploring the relationship of OSCC, oral microbiota and the immune system.

　　In this Article, we report that a silver nanoparticle (AgNP)-based hydrogel enhances immunotherapy via the modulationof oral microbiota. First, we searched for correlations in thesurvivorship of patients with OSCC and their expression ofimmune-related proteins and microbial abundance in their oralcavity. We found a genus of bacteria (Peptostreptococcus) that couldactivate the immune system and improve the prognosis of patients.

　　Then we sought to develop a strategy to regulate the oral micro-biota to activate anticancer immune responses16. Via screeningwe selected AgNPs, which inhibit the proliferation of otherbacteria17,18but allow the proliferation of Peptostreptococcus.

　　Subsequently, we formulated AgNPs into a mucous adhesivehydrogel (Agel) for the modulation of the oral microbiota. Wealso studied the therapeutic effects of the combination of Ageland an immune checkpoint inhibitor in mice intratumourallyinjected with different oral microbiota from patients, as well as in amurine model of chemically induced OSCC (via 4-nitroquinoline-oxide (4-NQO)).

　　Results

　　Screening and identification of OSCC-associated bacteria. First,we collected cancerous and paracancerous samples (mucosa) frompatients with OSCC. The microbial content and abundance of thesesamples were analysed through 16 S rDNA sequencing (Fig. 1a). Usingpermutational multivariate analysis of variance (PERMANOVA) ofprincipal coordinates analysis (PcoA), we failed to observe a signifi-cant difference in microbiota in the tumours and paracancer tissuesof patients with OSCC (Fig. 1a). However, we observed the enrich-ment of Peptostreptococcus in the tumours (Fig. 1b,c). Subsequently,we used a fluorescence in situ hybridization (FISH) probe toexamine the presence of these bacteria in paraffin-embedded tissues.

　　As shown in Fig. 1d, for the contents of other bacteria, no obvi-ous differences were found between cancerous and paracanceroustissues. In sharp contrast (Fig. 1e), the level of Peptostreptococcus inOSCC tissues (n = 199) was higher than that in oral dysplasia (n = 63)and oral mucosa (n = 30). As shown in Fig. 1f, the content of intra-tumoural Peptostreptococcus was significantly associated with thesurvivorship of patients (P = 0.002). Previous studies have also foundthat Peptostreptococcus is enriched in OSCC; however, the role ofthese bacteria remains unknown19. Compared with patients who hada low intratumoural level of Peptostreptococcus (n = 95), the overallsurvival rate in the Peptostreptococcus-enriched group (n = 96) wasbetter than in the Peptostreptococcus-deficient group (hazard ratio,HR = 0.4229; P = 0.002). From these results, we concluded thatPeptostreptococcus is enriched in the tumour mass and that it mayplay an important role in OSCC progression and treatment.

　　Interaction between Peptostreptococcus and host immunity.

　　We hypothesized that tumour inhibition by Peptostreptococcusmight be related to direct cytotoxicity by the bacteria. However,when SCC7 cells (murine OSCC cell lines) were co-cultured withPeptostreptococcus anaerobius (P. anaerobius, ATCC 27337), noobvious inhibition was observed. A strain of P. anaerobius isolatedfrom the oral cavity of patients with OSCC also displayed limitedcytotoxicity (Supplementary Fig. 1). The regulation of the immunesystem has been recognized as the primary route by which micro-biota affects cancers 20. As shown in Fig. 2a, the maturation of den-dritic cells (DCs) was induced by P. anaerobius. The bacteria ledto reproducible immune activation of bone mesenchymal DCsobtained from different mice. However, only intact bacteria led tosuperior immune-activation effects. Neither the bacteria’s capsularpolysaccharide nor their secretions in the culture medium showedsimilar performance in in vitro experiments. This was consistentwith the phenomenon that some bacteria can promote anticancerimmunity 21. The immune effects of Peptostreptococcus might bemediated by a multitude of functions.

　　Next, we studied the expression of a series of immuneresponse-related proteins, including CD8, IDO, CD47, Foxp3,B7-H3, TIM-3, ICOS, LAG-3 and programmed death-1 (PD-1) in tumours of patients. We used multivariate linear regression toestimate the independent association between the intratumourallevels of Peptostreptococcus and of immune-response-relatedproteins (Supplementary Fig. 2). Interestingly, the expressionof intratumoural CD8 was found to be highly correlated withPeptostreptococcus density (Fig. 2b). As the bacterial density in thetumour increased, the corresponding level of CD8 cells becamehigher. Using another FISH probe targeting Peptostreptococcus alsoshowed that intratumoural Peptostreptococcus levels were posi-tively correlated with long-term survival and intratumoural CD8levels (Extended Data Fig. 1). As shown in Supplementary Fig. 3,results from the Human Protein Atlas database showed that patientswith a high expression of CD8A and CD8B had a better progno-sis than CD8-low patients (P = 0.0028 and P = 0.007, respectively).

　　Considering the dominant role played by CD8+T cells in anticancer immunity 22, we deduced that Peptostreptococcus inhibited tumourprogression by inducing an effective immune response.

　　In view of the effects of Peptostreptococcus in clinical samples,we then used SCC7-tumour-bearing mice. Tumour-infiltratingimmune cells were collected for flow cytometry. As shown in Fig.

　　2c, we observed significantly upregulated T cells (CD3+) and DCs(CD11c+CD80+CD86+). This demonstrated that P. anaerobius acti-vated antigen-presenting cells, recruited tumour-killing T cells andreduced immunosuppression after colonization of the tumours. Wethen co-incubated bone mesenchymal DCs (BMDCs) with P. anaero-bius and used transcriptomic analysis to study the effects of bacteriaon DCs. By comparing gene-expression levels before and after bac-terial co-culture, we selected differential genes with a threshold of P < 0.05 (Supplementary Fig. 4). To find significantly enriched path-ways, we performed annotation and functional enrichment on the basis of the Kyoto Encyclopedia of Genes and Genomes (KEGG). Asshown in Fig. 2d, we observed that differential genes were signifi-cantly enriched in NOD-like receptor (NLR) and Toll-like receptorsignalling pathways (TLR). Commonly, TLRs/NLRs always mediatetranscriptional responses of antigen-presenting cells towards intra-cellular pathogens, and the engagement of TLR/NLR-associated proteins can induce anticancer immune responses 23,24. As shown inFig. 2e,f, the genes associated with TLR, such as Irf7, Il1b and Tnf,were significantly upregulated (P < 0.0001, two-way analysis of vari-ance (ANOVA)). Almost all NLR-pathway-associated genes werealso upregulated.

　　To verify the effects of TLR/NLR pathways in Peptostreptococcus-mediated immune responses, HEK-293 cells were trans-fected with NOD2, TLR4, TLR7 or TLR9 genes together with apELAM-luciferase reporter expression vector 25. After the activa-tion of the receptor, the bioluminescent signal was detected. As shown in Fig. 2g, the expression of TLRs (TLR4, TLR7 and TLR9)or NOD2 resulted in the activation of the firefly luciferase gene inresponse to P. anaerobius co-incubation. Additionally, the additionof NOD2 inhibitor blocked the activation of antigen-presentingcells. In contrast, the use of TLR inhibitors alone or in combinationdid not effectively inhibit DC maturation (Fig. 2h). We concludedthat Peptostreptococcus elicited an immune response by activatingthe TLR/NLR pathways. However, even when the TLR pathway isinhibited, Peptostreptococcus-induced immune responses can stilloccur through other compensatory routes 26.

　　We also studied the anticancer effects of P. anaerobius. As shownin Extended Data Fig. 2a, we found that the intratumoural injectionof P. anaerobius significantly inhibited the growth of subcutaneousSCC7 tumours (Extended Data Fig. 2a). Experiments were furtherperformed in T-cell-deficient BALB/c nude mice (Extended DataFig. 2b). In this model, the transplantation of P. anaerobius onlyshowed weak inhibition, which indicated that the bacteria-mediatedimmune effect was indeed the main reason for the anticancer effect of P. anaerobius.

　　However, when the salivary microbiota from patients withOSCC was intratumourally injected, the injection of P. anaerobiusshowed no significant therapeutic effects (Extended Data Fig. 2c,d).

　　PCR analysis showed that the intratumoural content of P. anaero-bius in the saliva + P. anaerobius group was similar to that in thesaliva group (Extended Data Fig. 2e). This phenomenon was con-trary to what was previously observed in tumours without bacteriainjection. Recent studies have indicated that commensal microbiotacan prevent the colonization of exogenous microbes 27. We suspectedthat the pre-existence of oral microbes prevented the colonization ofPeptostreptococcus.

　　Screening and designing materials to promote the growthof Peptostreptococcus. First, the study of the effects of differ-ent antibiotics on the proliferation of P. anaerobius showed thatP. anaerobius was sensitive to the antibiotics (Supplementary Fig.5). Nanomaterials such as polysaccharide nanoparticles 14, C3N4 nanosheets 16 and metal colloidal nanocrystals 28, are known to regu-late the behaviour, metabolism and growth of bacterial species. Wetherefore selected a range of nanomaterials, including AgNP, goldnanoparticles (AuNP), Fe3O4 nanoparticles, C3N4 nanosheets, cel-lulose nanoparticles, graphene nanosheets, ZnO nanoparticlesand chitosan nanoparticles, for co-incubation. We observed thatAgNP significantly inhibited the proliferation of various bacte-ria species except P. anaerobius (Fig. 3a). AgNP displayed no sig-nificant effects on the growth of P. anaerobius. Hence, we deducedthat AgNP might inhibit the growth of endogenous microbes andallow Peptostreptococcus to gain growth advantages in the micro-biota. We then collected saliva samples from patients. As shown inFig. 3b, the treatment with AgNP caused an increase in the level ofPeptostreptococcus.

　　In general, the antibacterial effects of nanoparticles areoften related to the generation of free metal ions 29. Here, AgNO3 was co-incubated with P. anaerobius, Escherichia coli (E. coli),Bacillus thuringiensis (B. thuringiensis) and Staphylococcus aureus (S. aureus). Even at a high AgNO3 concentration of 1 μg ml-1 , thegrowth of P. anaerobius remained unaffected, but the proliferationof other bacteria, including E. coli, S. aureus and B. subtilis wassignificantly inhibited. As shown in Fig. 3c, P. anaerobius showedthe lowest intracellular Ag concentration among the four bacteria.

　　Additionally, X-ray photoelectron spectroscopy (XPS) also revealedthe existence of Ag (0) on the surface of P. anaerobius (SupplementaryFig. 6a). For the other three bacteria, no obvious peak correspond-ing to Ag (0) was observed (Supplementary Fig. 6b–d). Therefore,we hypothesized that the tolerance mechanism of P. anaerobius might be the release of a reducing substance, such as electron accep-tors, to reduce Ag+into Ag (0)30. Transmission electron microscopy(TEM) revealed obvious morphological changes in P. anaerobiusafter the treatment. AgNPs with a diameter of ~150 nm were foundon the surface of P. anaerobius. Generally, the purported mecha-nism is illustrated in Fig. 3d. Briefly, Peptostreptococcus mightreduce the concentration of Ag+ through reduction reactions, mak-ing Peptostreptococcus more resistant to AgNPs than other bacteria.

　　We developed AgNPs into a hydrogel formulation thatadhered to the oral mucosa. We prepared a Schiff-base-mediatedtissue-adhesion hydrogel, mixed from chitosan and polyaldehyd-edextran (Supplementary Fig. 7)31,32. As shown in Fig. 3e, the mix-ture of these two polymer solutions formed a viscous hydrogel witha light-brown colour. The brown of the hydrogel also indicatedthe formation of a Schiff base. Then, the cyanine5 (Cy5)-labelledAgel was applied to the oral cavity of mice, and in vivo fluorescenceimaging was used to measure the changes in fluorescence intensityover time. The animals’ ability to eat was not affected during thetreatment. After 24 h, strong Cy5 fluorescence was still observed intheir oral cavities (Fig. 3e). In particular, a layer of hydrogel with~50 ?m thickness was observed on the surface of mucosa tissues.

　　In the 4-nitroquinoline N-oxide (4-NQO)-induced murine OSCCmodel, some green-fluorescent AgNPs dispersed in the tissue.

　　We imaged the dry Agel with scanning electron microscopy(SEM). As shown Fig. 3f, there are several nanoscale humps on thehydrogel surface.

　　We also examined the effects of Agel in a complex microbialenvironment. Agel was implanted in murine SCC tumours withoral microbiota from patients. Experiments with flow cytometryalso showed significant upregulation of DC maturation in drain-ing lymph nodes (Fig. 3g). Furthermore, we also tested the treat-ment effects of Agel + P. anaerobius in an orthotopic 4MOSC1tumour model. As shown in Fig. 3h, treatment with Agel +P. anaerobius inhibited the growth of tumours to a certain extent. Incontrast, intratumoural injection of P. anaerobius showed obviouseffects (Extended Data Fig. 2). However, in the orthotopic tumourwith symbiotic microbiota, P. anaerobius alone had no therapeu-tic effect. Only with modulation of the oral microbiota with Ageldid P. anaerobius induce anticancer immune responses and inhibittumour growth (Fig. 3h and Extended Data Fig. 2).

　　In vivo therapeutic effects of Agel. We measured the therapeuticeffects of Agel in murine SCC7 tumours intratumourally injectedwith oral microbiota from patients. First, oral microbiota samplesfrom patients were collected and tested. Subsequently, accordingto the content of Peptostreptococcus, the samples were divided intothree groups of high, medium and low bacterial levels (Fig. 4a).

　　After transplanting these three groups of microbes into murineSCC7 tumours, anti-PD-1 antibody (aPD-1) and Agel were admin-istered for treatment (Fig. 4b).

　　As shown in Fig. 4c, although the growth rates of tumours withoral microbiota from different patients were slightly different, theuse of Agel or Agel + aPD-1 still inhibited tumour growth. In threeexperiments, the Agel + aPD-1 treatment displayed tumour inhi-bition rates of 77.2%, 73.4% and 85.5%, respectively. A markedincrease in CD8+T cells (CD3+CD8+) was observed. In two of theseexperiments, the content of CD8+T cells was statistically improved.

　　We speculate that this might be because the high Peptostreptococcuscontent in the tumour itself triggered an effective immune response.

　　Similarly, in mice with oral microbiota with middle-level con-tent of Peptostreptococcus, Agel + aPD-1 also showed satisfactorytumour-suppressive effects, together with significantly improvedintratumoural CD8+T cells. Treatment with Agel + aPD-1 also sup-pressed tumour growth (Fig. 4d). Notably, significant increases inCD8+T-cell infiltration were observed in the three experiments.

　　Subsequently, we investigated the efficacy of Agel + aPD-1 onmurine tumours intratumourally injected with saliva from patientswith low levels of Peptostreptococcus. Unfortunately, the combi-nation therapy only displayed limited therapeutic effects in thesegroups. There was no significant improvement from this combi-nation treatment, whether in tumour suppression or in the levelsof infiltrating CD8+T cells (Fig. 4e). Exogenous P. anaerobius wasgiven to upregulate the intratumoural levels of Peptostreptococcus,followed by the injection of Agel and aPD-1 (ref.33). In sharp con-trast, the infiltration of CD8+T cells was significantly improvedafter the administration of exogenous P. anaerobius. As shown inFig. 4e, in three experiments, the combination treatment of Agel,exogenous P. anaerobius and aPD-1 led to high tumour-inhibitionrates: 85.7, 88.1 and 91.4%, respectively. This suggests that the com-bination treatment of Agel + exogenous P. anaerobius + aPD-1may have clinical relevance for patients with OSCC with low oralcontent of Peptostreptococcus. We verified the tumour-suppressingeffects of the combination treatment via in vivo bioluminescenceimaging (Fig. 4f). SCC7 cells with the firefly luciferase reportergene (SCC7luc) were used to construct a murine OSCC model. Here,the growth of tumours was monitored with the bioluminescence of SCC7luc cells. The mice treated with Agel + P. anaerobius + aPD-1 showed the smallest tumour volume and the weakest biolumines-cence intensity. However, both mono-treatments (aPD-1 or Agel)and the combination of Agel + aPD-1 showed limited therapeuticeffects, although strong bioluminescence was seen in some cases.

　　We also found that treatment with Agel + P. anaerobius + aPD-1significantly prolonged the survival of SCC7-tumour-bearing mice(Fig. 4g).

　　On the basis of these results, we speculate that treatment withAgel could be applied as a personalized therapy, on the basis of thecomposition of the oral microbiota. In patients with higher oralPeptostreptococcus content, aPD-1 + Agel could be used directlyfor treatment. However, for patients with lower Peptostreptococcuscontent, oral microbiota transplantation might be used beforeaPD-1 + Agel administration.

　　In a 4-NQO-induced spontaneous OSCC model (Fig. 5a)34,after various treatments the mice were anaesthetized and thetumours in their oral cavities were counted. As shown in Fig. 5b,macroscopic lesions were more prominent in PBS or in the aPD-1-treated group compared with the Agel + aPD-1-treated group 35,36.

　　In Agel + aPD-1-treated mice, we observed fewer tumour-likenodules, as well as smaller nodule volumes on their tongues. Onthe basis of histopathology examination, we classified these nodulesamples as dysplasia, invasive carcinoma and in situ carcinoma (Fig. 5c and Supplementary Fig. 8). The combined treatment ofAgel + aPD-1 significantly (P = 0.023) inhibited the developmentof murine SCC (invasive and in situ carcinoma). However, thesetreatments did not significantly affect the occurrence of dysplasia.

　　Then, CD8+T cells and mature DCs in draining lymph nodeswere analysed via flow cytometry. As shown in Fig. 5d, there wassignificant upregulation of mature DCs (P = 0.016) and cytotoxicT cells (P = 0.03). By using a FISH probe to visualize the existenceof intratumoural Peptostreptococcus, we saw a marked increase inPeptostreptococcus density. Semi-quantitative studies also indicate asignificant increase in the amount of Peptostreptococcus after Agel +aPD-1 treatment (Fig. 5e).

　　We also studied whether the direct administration ofP. anaerobius without Agel would produce a therapeutic effect inthe 4-NQO model (Extended Data Fig. 3a). In the original experi-mental design, we gave mice oral bacteria every 3 d for 1 month.

　　After seven oral administrations of P. anaerobius, reduced move-ment and eating, or even dyspnea, were observed. Therefore, wesuspended the administration of P. anaerobius. It should be pointedout that these adverse effects occurred only after a high dose ofP. anaerobius was given. With the Agel, only a low dose of P. anaero-bius was needed. We observed that both P. anaerobius + aPD-1 andP. anaerobius treatments displayed significant therapeutic effects(Extended Data Fig. 3b,c). In these two groups, fewer carcino-mas and increased CD8+T cell levels in lymph nodes (ExtendedData Fig. 3d) were observed. This phenomenon suggested thatP. anaerobius inhibited the progression of OSCC in the spontaneoustumour models. We also estimated the activation of CD8+T cells bymeasuring intratumoural levels of CD8+IFN-γ+cells. As shown inSupplementary Fig. 9, the synergy of Agel and aPD-1 significantlyimproved the secretion of cytokines from T cells.

　　Previous studies have shown that AgNPs can disturb the gutmicrobiota, or even induce dysbiosis 37. Here we profiled the overallstructure of the gut microbiota before and after Agel treatmentvia 16 S ribosomal DNA sequencing 38. A circos plot was used torepresent the interaction between Agel and the gut microbial com-munity. The Agel treatment displayed no obvious influence on thedominant bacterial genera in the gut (Extended Data Fig. 4). In viewof this, we speculate that this might result from the low Ag dose andthe adhesiveness of the hydrogel.

　　We also evaluated changes in the oral microbiota after Agel treat-ment. As shown in Fig. 5f,g, the treatment with Agel changed theoral microbiota of mice (analysis of similarities, ANOSIM, R = 1,P = 0.004). In addition to the effect on Peptostreptococcus, Ageltreatment also significantly downregulated Streptococcus (P = 0.012)and significantly increased the levels of Pseudomonas (P = 0.012),Achromobacter (P = 0.0009) and Cutibacterium (P = 0.012). Theoral pathogens in periodontitis and caries, such as Porphyromonasand Fusobacterium, were unaffected. Although dysbiosis of the oralmicrobiota may cause periodontitis and caries, OSCC is the mostserious disease of the oral cavity. It may therefore be acceptable todisrupt the oral microbiota to treat OSCC. However, owing to alack of sufficient understanding of the properties of the oral micro-biota, it is still unclear whether Agel affecting the oral microbiotamay cause other side effects. Subsequent studies on the possible sideeffects will be needed.

　　The excretion profiles for the Agel were studied by measuringtotal silver levels on urine and faeces. The faecal clearance wasfound to be the main excretion pathway for AgNP rather thanurinary excretion. Of the total amount of AgNP administered, morethan 95% was cleared via faeces (Supplementary Fig. 10a). Bloodbiochemistry and routine examination also suggest that Agel haslimited side effects (Supplementary Fig. 10b,c). Collectively, theseresults suggest that Agel does not lead to overt acute toxicity.

　　Colitis, dermatitis and hepatitis are the most problematic andcommon immune-mediated side effects associated with cancerimmunotherapy 39. Pathological analysis from hematoxylin andeosin staining did not find obvious inflammatory cell infiltra-tion (Supplementary Fig. 11). However, because the immune sys-tems of humans and mice are very different, further evaluation ofimmune-related side effects, also in non-human primate models,will be needed.

　　The low clinical response rate of aPD-1 is an important reasonfor restricting its use for the treatment of OSCC. Therefore, wefurther investigated whether the synergistic treatment with Agel andP. anaerobius could overcome this shortcoming. We verified theefficacy of Agel + P. anaerobius (Pa) on 4MOSC1 and 4MOSC2orthotopic tumours 40. As shown in Fig. 5h,i and SupplementaryFig. 12, the Agel + Pa + aPD-1 treatment inhibited the growth of4MOSC1 tumours, but aPD-1 alone had little effect on 4MOSC2tumours. However, the use of Agel + Pa significantly improved theanticancer effect of aPD-1 in 4MOSC2 tumours. This supports theclinical potential of Pa-based strategies for the treatment of patientswith different PD-1 treatment states.

　　Discussion

　　There are opportunities at the intersection of research in biomateri-als, the microbiota and host diseases. In this work, we have shownthat AgNPs allowed Peptostreptococcus to proliferate while inhibit-ing the growth of competing bacteria. In mouse models of OSCC,an AgNP-bearing hydrogel, Agel, sustainably adhered to the oralcavity and regulated the oral microbiota. After supplementationwith P. anaerobius, the hydrogel led to substantial therapeutic effectsin OSCC-bearing mice with low levels of Peptostreptococcus.

　　This work has some important limitations. First, although AgNPshave been approved by the US Food and Drug Administration(FDA) for use as antimicrobial dressings, using AgNPs as an oralspray could cause liver toxicity and inflammation. Any long-termtoxicity of Agel should be assessed. Second, it has been shown thatPeptostreptococcus may cause colon dysplasia by increasing thelevel of intestinal inflammation 41. However, inflammation causedby bacteria could be beneficial to immunotherapy for established tumours 42,43. Considering the complexity of bacteria-mediated anti-cancer immune responses, in-depth mechanistic research will beneeded. Third, we have assessed the therapeutic effects of Agel inmurine tumour models. Studies in non-human primates will helpdetermine the clinical potential of this therapeutic strategy. Fourth,because the inhibition of bacteria by AgNPs is dose-dependent,high concentrations of AgNP may inhibit the proliferation ofbacteria; hence, the dosage for Agel, especially in large-animalmodels, will need to be optimized.

　　methods

　　Materials. LPS (from E. coli O111:B4) and 4-NQO were purchased fromSigma-Aldrich. NOD-IN-1, resatorvid and hydroxychloroquine sulfatewere purchased from MedChemExpress. AgNO3 , chloroauric acid, sodium borohydride, FeCl3·6H2O, FeCl2·4H2O, chitosan and citric acid were purchasedfrom Aladdin. Dextran T40 and dextran T100 were purchased from ShanghaiYuanye Biotechnology. NH3·H2O, glutaraldehyde, NaIO4 and urea were purchased from Sinopharm Chemical Reagent. Cellulose nano-crystals (~15–25 nm diameter), ZnO nanoparticles (10 nm), TiO2 nanoparticles (10–20 nm), CeO2 nanoparticles (2–6 nm), CuO nanoparticles (10–50 nm), SnO2 nanoparticles(10–30 nm) and graphene nanosheets (~1.0 nm thickness) were purchased fromXFnano. MTT (3-[4,5-Dimethylthiazol-2-yl]-2,5-diphenyltetrazolium-bromide),d-luciferin potassium salt (luciferin) and cell counting kit-8 (CCK-8) weresupplied by Beyotime Biotechnology. InVivoMAb anti-mouse PD-1 (clone RMP1–14) was purchased from Bioxcell. pGL3-ELAM-luc, pcDNA3-TLR9-YFP,pcDNA3-TLR4-YFP and pcDNA3-TLR7-YFP were gifs from Doug Golenbock(Addgene_13029, Addgene_13642, Addgene_13018 and Addgene_13022).

　　pECMV-Nod2-m-FLAG was purchased from Miaolingbio.

　　Antibodies. Antibodies used for flow cytometry were as follows: anti-mouseCD45 (BD Pharmingen, APC-Cy7, Cat: 557659, Clone: 30-F11); anti-mouseCD3e (Invitrogen, FITC, Cat: 11-0031-82, Clone: 145-2C11); anti-mouse CD4 (Invitrogen, eFluor 450, Cat: 48-0042-80, Clone: RM4-5); anti-mouse CD8a(Invitrogen, APC, Cat: 17-0081-81, Clone: 53-6.7); anti-mouse CD11c (Invitrogen,FITC, Cat: 11-0114-81, Clone: N418); anti-mouse CD80 (Invitrogen, PE, Cat:12-0801-81, Clone: 16-10A1); anti-mouse CD86 (Invitrogen, APC, Cat: 17-0862-81,Clone: GL1); and Fixable Viability Dye (Invitrogen, eFluor 506, Cat: 65-0866-14).

　　Sample collection. Saliva and tumour samples were obtained from patients whounderwent surgeries in the School and Hospital of Stomatology, Wuhan University.

　　The Medical Ethics Committee of the School and Hospital of Stomatologyapproved this study, and informed consent was obtained from the patients beforethey underwent surgery. All the samples used in this study were diagnosed asOSCC by pathologists. The fresh OSCC tumour samples were collected andimmediately stored in liquid nitrogen. The follow-up records of eight patientswere absent and these data were excluded in Fig. 1f. The saliva samples wereobtained from patients suffering from OSCC and were collected before themorning meal. Saliva (500 μl) was added into the Salivette (Sarstedt), followedby centrifugation at 1,000 g for 3 min. The filtered saliva was collected for furtherexperiments.

　　Animal experiments. The study was performed with the approval of theInstitutional Medical Ethics Committee of the School and Hospital of Stomatology.

　　All mouse experimental procedures were performed in accordance with theRegulations for the Administration of Affairs Concerning ExperimentalAnimals approved by the State Council of the People’s Republic of China. Forall animal experiments, mice were housed under controlled conditions of ~50%humidity, ~25 °C temperature and a 12 h light/dark cycle. Isoflurane was used foranaesthesia. All experimental animals were 6 weeks old at the beginning of theexperiments. At the end of all experiments, animals were euthanized under CO2anaesthesia.

　　DCs activation. BMDCs were differentiated from bone mesenchymal stem cells(BMSCs)4. BMSCs were washed out of the bone marrow cavity of female C3H-HeNmice and cultured in RPMI-1640 medium containing GM-CSF (20 ng ml-1) andIL-4 (10 ng ml-1). After 6 d differentiation, the as-prepared BMDCs were collectedfor further experiments.

　　DCs were then co-cultured with P. anaerobius (10^7 CFU ml-1), LPS (1 μg ml-1),polysaccharide from P. anaerobius (10^7 CFU ml-1) and the medium of P. anaerobius (diluted to 1%) for 24 h, after which the cells were collected for flow cytometryand transcriptomic analysis. In the in vitro mechanism experiment, DCswere pre-treated with inhibitors including NOD-IN-1 (25 μg ml-1), resatorvid(100 μg ml-1) and hydroxychloroquine sulfate (40 μg ml-1) for 24 h. Then,P. anaerobius was added to the medium for an additional incubation of 24 h. Flowcytometry was used to study the maturation of DCs. A bacterial polysaccharideextraction kit (Bestbio) was used to separate the polysaccharide of P. anaerobius.

　　This procedure was performed according to the manufacturer’s instructions.

　　Anticancer effect of P. anaerobius in SCC7 tumour-bearing mice. FemaleC3H-HeN mice were anaesthetized and 3 × 10^5 SCC7 cells in 100 μl RPMI wereinjected subcutaneously into the dorsal sides of mice. When tumour volumesof mice reached about 100 mm 3, the mice were randomly divided into differentgroups. Subsequently, 10^6 CFU P. anaerobius (ATCC 27337) was injected intratumourally into mice. One week after the dosing, the mice were euthanized.

　　Flow cytometry was used to analyse the (CD3+CD4+and CD3+CD8+) and M2 macrophages (CD11b+F4/80+CD206+) in tumours. The tumour-draining lymphnodes were collected for analysis of the maturation of DCs (CD11c+CD80+CD86+).

　　The type strain P. anaerobius (ATCC 27337) was obtained from American TypeCulture Collection.

　　In vivo mechanism studies. Female C3H-HeN mice were anaesthetized and 3 × 10^5 SCC7 cells in 100 μl RPMI were injected subcutaneously into the dorsalsides of mice. When the tumour volume of mice reached about 100 mm^3, themice were randomly divided into different groups. Then, 50 μl pre-treated salivawas injected intratumourally into mice. The day of microbiota transplantationwas denoted as day 0. On the 1st day, 100 μl Agel with a silver content of 68 μgwas injected subcutaneously into the dorsal sides of mice. Additionally, 10^6 CFU P. anaerobius was injected intratumourally into mice on the first day of thetreatment. One week after the dosing, the mice were euthanized. Flow cytometrywas used to analyse the CD3+CD4+and CD3+CD8+ in tumours. The tumour-draining lymph nodes were collected for analysis of the maturation of DCs(CD11c+CD80+CD86+).

　　Co-incubation of bacteria with nanomaterials. Eight kinds of nanoparticles (AgNP, AuNP, Fe3O4, cellulose, C3N4, ZnO, graphene and chitosan) with the same concentration of 20 μg ml-1 were individually co-cultured with differentbacteria, including B. thuringiensis, B. subtilis, E. coli, S. aureus and P. anaerobius.

　　P. anaerobius needed to be cultured for up to 24 h before testing. Afterco-incubating bacteria with nanomaterials for 6 h, absorbance was measuredto monitor the proliferation of bacteria. For the co-incubation of oral microbiotawith AgNP, 100 μl purified saliva sample was added to 1 ml BHI medium.

　　Meanwhile, AgNP was added to the above medium, and the final concentrationof AgNP in the medium was 20 μg ml-1.

　　Synthesis of Agel. Dextran T100 (1 g) was dissolved in 40 ml deionized water. Thesolution was bubbled with nitrogen gas for 30 min. Then 0.142 g KIO4 was added to the solution and the mixture was stirred under N2 at room temperature for 12 h.

　　The solution was added dropwise into methanol. The forming white precipitate(polyaldehyde dextran, PAD) was collected. The prepared PAD was then driedunder vacuum for 24 h. AgNO3(0.57 g) and 0.226 g dextran T100 were added to133 ml deionized water. The solution was irradiated with UV-light (12 W) at roomtemperature for 2 h. Then, centrifugation (3,920 g, 30 min) was performed to obtainAgNP. The hydrogel was prepared by mixing PAD (20%) and chitosan (1%). Tween80 (1%) and carboxycellulose were added to promote transmucosal absorption inorthotopic tumour models. Briefly, AgNP were dispersed in the chitosan solution.

　　Immediately, 500 μl PAD solution and 500 μl AgNP-containing chitosan solutionwere mixed. In all in vivo experiments, Agel was injected as soon as possible.

　　UV–visible spectroscopy (Shimadzu UV-3600 UV–Vis spectrophotometer) wasused to measure changes in the absorption of the mixture before and after thegelation. The resulting hydrogel was lyophilized for TEM (JEM-2100, JEOL) andSEM (SIGMA, Zeiss) observations.

　　In vivo personalized treatments. Female C3H-HeN mice were anaesthetizedand 3 × 10^5SCC7 cells in 100 μl of RPMI were injected subcutaneously into thedorsal sides of mice. When tumour volumes of mice reached about 100 mm^3,the mice were randomly divided into 4 or 5 groups, with 3 mice per group. Then,50 μl pre-treated salivary bacteria was injected intratumourally into mice. The dayof microbiota transplantation was denoted as day 0. On day 1, 100 μl Agel with asilver content of 68 μg was injected subcutaneously into the dorsal sides of mice.

　　On days 3, 5 and 7 of the treatment, aPD-1 was administered intraperitoneally ata dosage of 10 mg kg-1. At the end of the experiment, the mice were euthanized. Atthe same time, the tumour tissues of the mice were collected for FISH staining andflow cytometry analysis. Tumour size and mice weight were measured immediatelybefore various treatments. Tumour volume was defined as V = L × W^2× 0.5, whereL and W are the longest and shortest diameters of tumours, respectively. In murinetumour models intratumourally injected with microbiota containing low levels ofPeptostreptococcus, 10^6 CFU P. anaerobius was injected intratumourally into miceon the first day of the treatment. There were no changes in the dose and durationof administration of Agel and aPD-1. In long-term animal experiments, thetreatment was repeated every 2 weeks.

　　To visualize the growth of the tumour, SCC7 cells with firefly luciferasereporter gene (SCC7 luc) were used to construct a murine OSCC model.

　　Female C3H-HeN mice were anaesthetized and 3 × 10^5 SCC7 luc cells in 100 μl of RPMI were injected subcutaneously into the dorsal sides of mice. Whenthe tumour volumes of mice reached about 100 mm^3, the mice were randomlydivided into 4 or 5 groups, with 3 mice per group. Then, 50 μl pre-treated saliva(low Peptostreptococcus level) was injected intratumourally into mice. Therewere no changes in the dose and duration of administration of Agel, aPD-1and P. anaerobius. Twenty minutes before bioluminescent imaging, intraperitoneal(i.p.) injection of d-luciferin (10 mg ml-1 in 100 μl PBS) was performed. Forhumane reasons, animals were euthanized when the solid tumour volumeexceeded 2,000 mm^3.

　　4-NQO-induced oral carcinogenesis. Female C57BL/6 mice (6 weeks old)were housed and drinking water containing 100 μg ml-1 4-NQO was given for16 weeks. All diet and water were made available ad libitum. Subsequently, the oralcavities of the mice were examined. Eighteen mice with lesions were randomlydivided into 3 groups. Normal drinking water was given. Agel (100 μl) with asilver content of 68 μg was evenly applied to the oral cavity of the female C57BL/6mice. Two days later, aPD-1 was administered intraperitoneally at a dosage of10 mg kg-1. Both Agel and aPD-1 were given once a week for 4 weeks. Then, micewere euthanized. Tongues were collected for pathological examination and FISH staining. The tumour-draining lymph nodes were collected for flow cytometrystudies. When the mice had difficulty eating due to tumour growth, the experimentwas ended.

　　4MOSC1 and 4MOSC2 tumour models. 4MOSC1 and 4MOSC2 cells (10^6)were injected into the tongue of female C57BL/6 mice (6 weeks old). On the fifthday after injection, mice with tumours were randomly divided into differentgroups (n = 5). Various treatments were started when the tumours grew to 10 mm^3. P. anaerobius (10^7 CFU) and 100 μl Agel were smeared on the tongues of mice with a cotton swab. The mice were treated by i.p. injection with aPD-1(10 mg kg-1). Hydrogel and antibody were administered every 3 d. The experimentwas ended when the mice showed obvious eating difficulties due to tumourgrowth. Subsequently, the mice were euthanized and their lymphoid and tumourswere removed for flow cytometry. Skin, intestine and liver were collected forpathological analysis.

　　Reporting Summary. Further information on research design is available in theNature Research Reporting Summary linked to this article.

　　Data availability

　　The main data supporting the findings of this study are available within the paperand its Supplementary Information. The raw and analysed datasets generatedduring the study are too large to be publicly shared, but they are available forresearch purposes from the corresponding authors on reasonable request. Sourcedata for Figs. 4 and 5, and Extended Data Fig. 2 are provided with this paper.

　　Eukaryotic transcriptome and bacterial 16 S rRNA sequencing data are availablefrom the NCBI Sequence Read Archive (accession numbers: PRJNA759007 andPRJNA758237).

　　Received: 10 November 2020; Accepted: 10 September 2021;

　　Published: xx xx xxxx

　　References

　　1. Chen, Q. et al. In situ sprayed bioresponsive immunotherapeutic gel forpost-surgical cancer treatment. Nat. Nanotechnol. 14, 89–97 (2019).

　　2. Hoos, A. Development of immuno-oncology drugs — from CTLA4 to PD1to the next generations. Nat. Rev. Drug Discov. 15, 235–247 (2016).

　　3. Min, Y. Z. et al. Antigen-capturing nanoparticles improve the abscopal efectand cancer immunotherapy. Nat. Nanotechnol. 12, 877–882 (2017).

　　4. Nam, J. et al. Cancer nanomedicine for combination cancer immunotherapy.Nat. Rev. Mater. 4, 398–414 (2019).

　　5. Bray, F. et al. Global cancer statistics 2018: GLOBOCAN estimates ofincidence and mortality worldwide for 36 cancers in 185 countries.CA Cancer J. Clin. 68, 394–424 (2018).

　　6. Leemans, C. R., Braakhuis, B. J. M. & Brakenhof, R. H. Te molecularbiology of head and neck cancer. Nat. Rev. Cancer 11, 9–22 (2011).

　　7. Krishna, S. et al. Human papilloma virus specifc immunogenicity anddysfunction of CD8(+) T cells in head and neck cancer. Cancer Res. 78,6159–6170 (2018).

　　8. Badoual, C. et al. PD-1-expressing tumor-infltrating T cells are a favorableprognostic biomarker in HPV-associated head and neck cancer. Cancer Res.73, 128–138 (2013).

　　9. Cohen, E. E. W. et al. Pembrolizumab versus methotrexate, docetaxel, orcetuximab for recurrent or metastatic head-and-neck squamous cellcarcinoma (KEYNOTE-040): a randomised, open-label, phase 3 study. Lancet393, 156–167 (2019).

　　10. Mandal, R. et al. Te head and neck cancer immune landscape and itsimmunotherapeutic implications. JCI Insight 1, e89829 (2016).

　　11. Garrett, W. S. Cancer and the microbiota. Science 348, 80–86 (2015).

　　12. Mager, L. F. et al. Microbiome-derived inosine modulates response tocheckpoint inhibitor immunotherapy. Science 369, 1481–1489 (2020).

　　13. Riquelme, E. et al. Tumor microbiome diversity and composition infuencepancreatic cancer outcomes. Cell 178, 795–806 (2019).

　　14. Zheng, D. W. et al. Phage-guided modulation of the gut microbiota of mousemodels of colorectal cancer augments their responses to chemotherapy.Nat. Biomed. Eng. 3, 717–728 (2019).

　　15. Gopalakrishnan, V. et al. Gut microbiome modulates response to anti-PD-1immunotherapy in melanoma patients. Science 359, 97–103 (2018).

　　16. Zheng, D. W. et al. Optically-controlled bacterial metabolite for cancertherapy. Nat. Commun. 9, 1680 (2018).

　　17. Mao, C. Y. et al. Photo-inspired antibacterial activity and wound healingacceleration by hydrogel embedded with Ag/Ag@AgCl/ZnO nanostructures.ACS Nano 11, 9010–9021 (2017).

　　18. Dong, X. et al. Bioinorganic hybrid bacteriophage for modulation of intestinalmicrobiota to remodel tumor-immune microenvironment against colorectal cancer. Sci. Adv. 6, eaba1590 (2020).

　　19. Zhao, H. et al. Variations in oral microbiota associated with oral cancer.Sci. Rep. 7, 11773 (2017).

　　20. Zheng, B. B. et al. Bacterium-mimicking vector with enhanced adjuvanticity for cancer immunotherapy and minimized toxicity. Adv. Funct. Mater. 29,1901437 (2019).

　　21. Patel, R. B. et al. Development of an in situ cancer vaccine via combinationalradiation and bacterial-membrane-coated nanoparticles. Adv. Mater. 31,1902626 (2019).

　　22. Gibney, G. T., Weiner, L. M. & Atkins, M. B. Predictive biomarkers forcheckpoint inhibitor-based immunotherapy. Lancet Oncol. 17, E542–E551(2016).

　　23. Rodell, C. B. et al. TLR7/8-agonist-loaded nanoparticles promote thepolarization of tumor-associated macrophages to enhance cancerimmunotherapy. Nat. Biomed. Eng. 2, 578–588 (2018).

　　24. Chao, Y. et al. Combined local immunostimulatory radioisotope therapy andsystemic immune checkpoint blockade imparts potent antitumor responses.Nat. Biomed. Eng. 2, 611–621 (2018).

　　25. Hemmi, H. et al. Small anti-viral compounds activate immune cells viathe TLR7 MyD88-dependent signaling pathway. Nat. Immunol. 3, 196–200(2002).

　　26. Long, X. et al. Peptostreptococcus anaerobius promotes colorectal carcinogenesisand modulates tumor immunity. Nat. Microbiol. 4, 2319–2330 (2019).

　　27. Martens, E. C., Neumann, M. & Desai, M. S. Interactions of commensal andpathogenic microorganisms with the intestinal mucosal barrier. Nat. Rev.Microbiol. 16, 457–470 (2018).

　　28. Peng, B., Zhang, X. L., Aarts, D. G. A. L. & Dullens, R. P. A.Superparamagnetic nickel colloidal nanocrystal clusters with antibacterialactivity and bacteria binding ability. Nat. Nanotechnol. 13, 478–482 (2018).

　　29. Le Ouay, B. & Stellacci, F. Antibacterial activity of silver nanoparticles: asurface science insight. Nano Today 10, 339–354 (2015).

　　30. Silver, S. Bacterial silver resistance: molecular biology and uses and misusesof silver compounds. FEMS Microbiol. Rev. 27, 341–353 (2003).

　　31. Conde, J., Oliva, N., Zhang, Y. & Artzi, N. Local triple-combination therapyresults in tumor regression and prevents recurrence in a colon cancer model.Nat. Mater. 15, 1128–1138 (2016).

　　32. Conde, J., Oliva, N. & Artzi, N. Implantable hydrogel embedded dark-goldnanoswitch as a theranostic probe to sense and overcome cancer multidrugresistance. Proc. Natl Acad. Sci. USA 112, E1278–E1287 (2015).

　　33. Zhou, S. B., Gravekamp, C., Bermudes, D. & Liu, K. Tumor-targeting bacteriaengineered to fght cancer. Nat. Rev. Cancer 18, 727–743 (2018).

　　34. Strati, K., Pitot, H. C. & Lambert, P. F. Identifcation of biomarkers thatdistinguish human papillomavirus (HPV)-positive versus HPV-negative headand neck cancers in a mouse model. Proc. Natl Acad. Sci. USA 103,14152–14157 (2006).

　　35. Deng, W. W. et al. LAG-3 confers poor prognosis and its blockade reshapesantitumor response in head and neck squamous cell carcinoma.Oncoimmunology 5, e1239005 (2016).

　　36. Rao, L. et al. Platelet-facilitated photothermal therapy of head and necksquamous cell carcinoma. Angew. Chem. Int. Ed. 57, 986–991 (2018).

　　37. Ma, Y. B. et al. Sex dependent efects of silver nanoparticles on the zebrafshgut microbiota. Environ. Sci. Nano 5, 740–751 (2018).

　　38. Zheng, D. W. et al. An orally delivered microbial cocktail for the removal ofnitrogenous metabolic waste in animal models of kidney failure. Nat. Biomed.Eng. 4, 853–862 (2020).

　　39. Perez-Ruiz, E. et al. Prophylactic TNF blockade uncouples efcacy andtoxicity in dual CTLA-4 and PD-1 immunotherapy. Nature 596, 428–432(2019).

　　40. Wang, Z. et al. Syngeneic animal models of tobacco-associated oral cancerreveal the activity of in situ anti-CTLA-4. Nat. Commun. 10, 5546 (2019).

　　41. Tsoi, H. et al. Peptostreptococcus anaerobius induces intracellular cholesterolbiosynthesis in colon cells to induce proliferation and causes dysplasia inmice. Gastroenterology 162, 1419–1433 (2017).

　　42. Coutzac, C. et al. Systemic short chain fatty acids limit antitumor efect ofCTLA-4 blockade in hosts with cancer. Nat. Commun. 11, 2168 (2020).

　　43. Zheng, J. H. et al. Two-step enhanced cancer immunotherapy with engineeredSalmonella typhimurium secreting heterologous fagellin. Sci. Transl. Med. 9,eaak9537 (2017).

　　acknowledgements

　　This work was supported by the National Key Research and Development Programof China (2019YFA0905603) and the National Natural Science Foundation of China(51833007, 51988102, 51690152 and 81874131). We thank Q.-M. Chen (SichuanUniversity, P. R. China) for the kind gift of the SCC7 cell lines, and the Wuhan Instituteof Biotechnology (Wuhan, P. R. China) for the technical support. 4MOSC1 and 4MOSC2cells were kindly gifted by S. Gutkind (University of California, San Diego, USA) throughthe material transfer agreement (SD2017-202).

　　author contributions

　　D.-W.Z., W.-W.D., Z.J.S. and X.-Z.Z. conceived the project and designed theexperiments. D.-W.Z., W.-F.S. and S.H. synthesized materials. W.-F.S. performed in vitro microbiological experiments. D.-W.Z., W.-W.D. and H.C. collected and analysed thedata. C.-C.W. and W.-W.D. performed in vitro cell experiments. W.-W.D., J.L. and W.-F.S.performed in vivo experiments. D.-W.Z., W.-W.D., W.-F.S., Z.-N.Z., Z.-J.S. and X.-Z.Z.co-wrote the manuscript. All authors discussed the results and reviewed the manuscript.

　　Competing interests

　　The authors declare no competing interests.

　　additional information

　　Extended data is available for this paper at https://doi.org/10.1038/s41551-021-00807-9.

　　Supplementary information The online version contains supplementary materialavailable at https://doi.org/10.1038/s41551-021-00807-9.

　　Correspondence and requests for materials should be addressed toZhi-Jun Sun or Xian-Zheng Zhang.

　　Peer review information Nature Biomedical Engineering thanks Christian Jobin, ZhuangLiu and Bo Xiao for their contribution to the peer review of this work.

　　Reprints and permissions information is available at www.nature.com/reprints.

　　Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims inpublished maps and institutional affiliations.

　　? The Author(s), under exclusive licence to Springer Nature Limited 2021