nature nanotechnology Letter https://doi.org/10.1038/s41565-022-01260-8 Graphene oxide elicits microbiome- dependent type 2 immune responses via the aryl hydrocarbon receptor Received: 17 December 2020 Guotao Peng 1, Hanna M. Sinkko1,2, Harri Alenius 1,2, Neus Lozano 3 , Kostas Kostarelos 3,4, Lars Bräutigam5 & Bengt Fadeel 1 Accepted: 12 October 2022 Published online: 12 December 2022 The gut microbiome produces metabolites that interact with the aryl Check for updates hydrocarbon receptor (AhR), a key regulator of immune homoeostasis in the gut1,2. Here we show that oral exposure to graphene oxide (GO) modulates the composition of the gut microbiome in adult zebrafish, with significant differences in wild-type versus ahr2-deficient animals. Furthermore, GO was found to elicit AhR-dependent induction of cyp1a and homing of lck+ cells to the gut in germ-free zebrafish larvae when combined with the short-chain fatty acid butyrate. To obtain further insights into the immune responses to GO, we used single-cell RNA sequencing to profile cells from whole germ-free embryos as well as cells enriched for lck. These studies provided evidence for the existence of innate lymphoid cell (ILC)-like cells3 in germ-free zebrafish. Moreover, GO endowed with a ‘corona’ of microbial butyrate triggered the induction of ILC2-like cells with attributes of regulatory cells. Taken together, this study shows that a nanomaterial can influence the crosstalk between the microbiome and immune system in an AhR-dependent manner. The increasing exploitation of nanomaterials including graphene-based beyond through several distinct mechanisms2. Recent work has shown materials necessitates a comprehensive evaluation of the potential that BA regulates the aryl hydrocarbon receptor (AhR) and its target effects of these materials on human health4. However, although the genes (including genes encoding members of the cytochrome P450 interactions of nanomaterials with the immune system have been family, such as CYP1A1) in the liver and intestine6, whereas other stud- addressed, their impact on the microbiome of the host remains to ies have implicated BA in the modulation of AhR expression or the be understood. Furthermore, studies are needed to address whether expression of other AhR ligands7,8. AhR controls intestinal epithelial graphene-based materials or other nanomaterials modulate immune cell regeneration, mediates anti-inflammatory responses and modu- responses via effects on the microbiota and/or its metabolites. The lates type 3 innate lymphoid cell (ILC3) polarization2. The activation microbiome, our ‘forgotten organ’, is involved in the regulation of mul- of AhR also facilitates the induction of tolerogenic regulatory T cells tiple signalling pathways in the host, and there exists bidirectional (Tregs), and a recent study has shown that the AhR pathway regulates communication between the gut microbiome and the immune sys- the ILC2–ILC3 balance in the gastrointestinal (GI) tract to ensure an tem5. The gut microbiome produces numerous metabolites including appropriate immune response against pathogens9. Thus, alterations short-chain fatty acids (SCFAs) such as acetate (AA), butyrate (BA) of the gut microbiota could lead to changes in the pool of AhR ligands, and propionate (PA)5. SCFAs, in turn, signal to host cells in the gut and subsequently affecting gut immunity in the host. Institute of Environmental Medicine, Karolinska Institutet, Stockholm, Sweden. 2Human Microbiome Research Program (HUMI), University of Helsinki, 1 Helsinki, Finland. 3Catalan Institute of Nanoscience and Nanotechnology (ICN2), Bellaterra, Spain. 4National Graphene Institute, and Faculty of Biology, Medicine & Health, University of Manchester, Manchester, UK. 5Comparative Medicine, Karolinska Institutet, Stockholm, Sweden. e-mail: bengt.fadeel@ki.se Nature Nanotechnology | Volume 18 | January 2023 | 42–48 42 Letter https://doi.org/10.1038/s41565-022-01260-8 a GO: 50 µg l–1 (low dose) c Fusobacteriota 500 µg l–1 (high dose) 100 # Control Relative abundance (%) 80 Low GO ** High GO 60 0 1 2 3 4 5 6 7 Days 40 WT versus AhR deficient 16S rRNA-seq 20 0 WT ahr2+/– b d Proteobacteria Patescibacteria WT (control) # Planctomycetota 50 ** Control Relative abundance (%) WT (low GO) Actinobacteriota 40 Low GO WT (high GO) Bdellovibrionota High GO 30 Verrucomicrobiota +/– ahr2 (control) Bacteroidota 20 ahr2+/– (low GO) Firmicutes 10 Fusobacteriota ahr2+/– (high GO) 0 Proteobacteria WT ahr2+/– 0 20 40 60 80 100 Phylum relative abundance (%) e f g WT Control Control Low dose Low dose ahr2+/– High dose High dose dbRDA2 (10.8%) dbRDA2 (2.4%) dbRDA2 (8.5%) dbRDA1 (36.0%) dbRDA1 (43.8%) dbRDA1 (23.1%) Fig. 1 | AhR-dependent changes in the gut microbiome of adult zebrafish. using Tukey’s multiple comparisons test was used to analyse the statistical a, Experimental design for the seven-day exposure regimen in adult zebrafish differences (Fusobacteriota, **p = 0.0065, #p = 0.0265; Proteobacteria, (WT and ahr2+/−). b, The most abundant bacteria phyla of the gut microbiota **p = 0.0055, #p = 0.0186). e, Supervised analyses of the microbiota composition among genotypes and treatments. Each bar represents the average of six between the two genotypes. f,g, Impact of GO on gut microbiota composition individuals in each condition. c,d, Relative phylum abundance of Fusobacteriota among WT (f) and ahr2+/− (g) zebrafish. dbRDA, distance-based redundancy (c) and Proteobacteria (d) in WT versus ahr2+/− fish exposed to GO. The error analysis. Differential abundances of ASVs are shown in Supplementary Fig. 5. bars represent the mean values ± s.d. of six individuals. Significant differences Credit: fish in a, Adobe Stock. between the treatments and genotypes are shown. Two-way analysis of variance In this study, we first determined whether graphene oxide (GO) goblet cells in animals exposed to GO (50 µg l–1) (Supplementary Fig. 4). can modulate the gut microbiota composition, and whether the AhR Furthermore, 16S rRNA gene sequencing revealed that the phyla Pro- plays a role in shaping the microbiota in unexposed or exposed animals. teobacteria and Fusobacteria dominated the intestinal microflora in To this end, we used the zebrafish (Danio rerio) as a model10. We previ- both WT and AhR-deficient (ahr2+/−) zebrafish (Fig. 1b). Exposure to ously reported on the structural properties of GO in E3 medium11 (Sup- GO shifted the relative abundance from Fusobacteria to Proteobac- plementary Fig. 1 shows the corresponding results in other relevant teria, and this effect was more pronounced in AhR-deficient animals media). GO was determined to be endotoxin free before exposing (Fig. 1c,d). Interestingly, 16S rRNA gene sequencing revealed a sig- the animals (Supplementary Fig. 2). We exposed wild-type (WT) and nificant shift in the ratio of Firmicutes to Bacteroidetes following AhR-deficient zebrafish to GO (50 or 500 µg l–1) continuously for seven oral exposure of mice to GO (2.5 mg kg−1 per day for seven days)14. days at which time intestines were dissected and samples were har- The present data showed that the gut microbiota composition was vested for 16S rRNA gene sequencing (Fig. 1a). A transmission electron significantly different between the two genotypes (R2, 21%; p = 0.0004) microscopy (TEM) analysis of the dissected intestines showed that GO and between the exposure groups (R2, 26%, p = 0.0001) (Fig. 1e and was present in the gut lumen admixed with bacteria (Supplementary Supplementary Tables 1 and 2). GO exposure explained 46% of the Fig. 3). High-magnification images revealed GO sheets in close apposi- variation in microbiota composition in WT animals (Fig. 1f), whereas tion with microvilli (low dose) and signs of cellular uptake of GO (high in AhR-deficient animals, GO exposure explained 34% of the variation dose). Epithelial cells at mucosal barriers represent the very first line (Fig. 1g). Differential abundance testing of the amplicon sequence of immune defense12, and the hyperplasia of goblet cells—specialized variants (ASVs) of the bacterial 16S rRNA gene sequences in WT and mucin-secreting epithelial cells—is a hallmark of type 2 immunity13. AhR-deficient animals exposed to GO are summarized in Supplemen- We determined the expression of goblet cells by using the Alcian blue tary Fig. 5. Notably, in WT animals, high-dose exposure resulted in an and periodic acid–Schiff reagent and observed a significant increase in enrichment of Vibrio, Pseudomonas and Aeromonas, consistent with Nature Nanotechnology | Volume 18 | January 2023 | 42–48 43 Letter https://doi.org/10.1038/s41565-022-01260-8 previous studies of known AhR agonists15,16, whereas in AhR-deficient cells were upregulated in the GF embryos (Supplementary Fig. 10), animals, similar effects were noted in both low-dose and high-dose whereas genes related to ILC1 (tbet, ifn-γ) and ILC3 (rorca, il22) cells groups. However, the relative abundance of individual members of the were not upregulated. To confirm the induction of lck in situ, transgenic gut microbiota does not necessarily correlate with their immunomodu- Tg(lck:GFP) zebrafish larvae raised under CV and GF conditions were latory effects17. No differences in ASV abundance were noted between exposed to GO+BA. We found that GO+BA prompted the homing of males and females (Supplementary Table 1). In contrast, another work18 lck+ cells to the gut (Fig. 3f). This was only observed in GF zebrafish found that chronic exposure (25 days) to a high dose (5 mg l–1) of GO (Fig. 3g shows the representative images and Supplementary Fig. 11 elicited differences between female and male zebrafish at the phylum shows the visualization of red fluorescent BA in the gut and lck (green) and genus levels. In sum, week-long exposure to GO substantially in the thymus and GI tract). affected the gut microbiota composition in adult zebrafish, which, in Importantly, a recent single-cell transcriptional analysis revealed turn, was modulated by the AhR. the presence of ILC-like cells in the gut of adult zebrafish25. However, To further investigate the importance of the microbiome, we gen- ascribing specific cellular phenotypes on the basis of bulk analysis of erated germ-free (GF) zebrafish embryos on a WT and AhR-deficient gene expression by qPCR is challenging. Therefore, to identify ILC-like background (Supplementary Fig. 6). Previous work has demonstrated cells in zebrafish embryos and to determine the impact, if any, of GO+BA that bacterial communities from adult zebrafish intestines synthesize on these cells, we performed single-cell RNA sequencing (scRNA-seq) all the three main SCFAs19, and BA has also been detected in situ (in adult on cells from whole embryos as well as on cells enriched for lck (Sup- zebrafish)20. Importantly, using the human intestinal epithelial cell line plementary Fig. 12a,b). To this end, cells from WT embryos and lck+ cells HT-29, we found that only BA (not AA or PA) was able to trigger AhR enriched from Tg(lck:GFP) embryos raised under GF conditions and activation (Supplementary Fig. 7). Similarly, others have shown that exposed to GO+BA or not were collected (Supplementary Fig. 12c,d) and BA induced CYP1A1 expression in HT-29 cells6. Using TEM, we could subjected to scRNA-seq using the 10x Genomics technology, as detailed also show that GO (30 μg ml–1) was internalized by differentiated HT-29 in Methods (Supplementary Figs. 13 and 14 show the quality control cells following a 24 h exposure, in contrast to a previous study in which (QC) of these transcriptomics data). The transcriptomics analysis of differentiated Caco-2 cells were used to model the GI barrier21. Thus, whole zebrafish embryos allowed us to identify an lck+ cell population GO is taken up by intestinal cells in vitro and in vivo (as shown above) in the control samples with markers of both T cells and ILCs, whereas and may therefore serve to ‘deliver’ BA to these cells. The exposure of in GO+BA-exposed zebrafish, two separate lck+ cell populations were GF zebrafish larvae (5 days post fertilization (dpf)) to GO (5 µg ml–1) identified, out of which one corresponded to T cells and the other to for 24 h resulted in the presence of GO in the GI tract, as evidenced by ILC-like cells (Supplementary Fig. 15). Furthermore, the integrated TEM (Fig. 2a) and Raman spectroscopy/microscopy (Fig. 2b). The TEM analysis of both samples revealed the expansion of a cell population analysis showed that GO was located on the surface of the epithelial cell expressing markers of pancreas and liver in the GO+BA-exposed lar- microvilli. GO exposure did not adversely affect the overall architecture vae (Extended Data Fig. 1). This may imply that the presence of GO of the GI tract, although a certain degree of microvilli and epithelial cell with a ‘corona’ of SCFAs in the gut is sensed as ‘food’, leading to the membrane damage was noted. We then monitored the induction of induction of genes encoding digestive enzymes (for example, serine cyp1a as a marker of AhR activation using reverse transcription quanti- proteases) and genes involved in lipid metabolism (Extended Data tative polymerase chain reaction (RT-qPCR). To this end, embryos were Fig. 1). We also noted that cyp1a was induced in the cluster identified exposed to GO (10, 30 and 50 µg ml–1), BA (0.5, 1.5 and 2.5 mM) or to a as intestinal cells in GF embryos exposed to GO+BA, in line with the combination of GO and BA (GO+BA) at the indicated concentrations for results obtained in Tg(cyp1a:GFP) zebrafish. However, ILCs comprise 24 h. We used the high-affinity AhR ligand 6-formylindolo[3,2-b]carba- only a small fraction of lymphocytes present at mucosal barriers26. To zole (FICZ)22 as a positive control. FICZ triggered a marked induction refine our approach, we, therefore, performed scRNA-seq on lck+ cells of cyp1a in conventional (CV) and GF embryos, and this was nullified sorted from GF Tg(lck:GFP) larvae. Our analysis showed that a distinct in ahr2−/− embryos (Fig. 2c–f). GO and BA, alone or in combination, cell population (cluster) displaying markers of ILCs was present in the had no or modest effects in CV embryos. In contrast, in GF embryos, control (Fig. 4a), in line with a previous study in which ILC-like cells were the combined exposure to GO+BA triggered a significant (~20-fold) identified in dissected intestines of adult zebrafish25. The relevant genes induction of cyp1a, and this was not seen in ahr2−/− embryos (Fig. 2d,f), are shown in Fig. 4b, and the red box delineates cluster 4 (corresponding confirming the role of AhR. We also utilized transgenic Tg(cyp1a:GFP) to ILC-like cells). Furthermore, on exposure to GO+BA, a cluster corre- zebrafish derived under GF conditions and could show that GO+BA sponding to ILC-like cells could be identified (Fig. 4c) which, in turn, was elicited cyp1a induction in the GI epithelium (Fig. 2g), whereas red shown to comprise ILC2-like cells (nitr+gata3+il4+il13+) and ILC3-like cells fluorescent BA was detected in the gut lumen. FICZ also prompted (nitr+rorc+il17a/f1+il22+) as well as ILC2 cells with attributes of regulatory cyp1a induction in the liver and gut, whereas the response to BA alone ILC-like cells known as ILC210 cells27 (nitr+gata3+foxp3a+il10+) (Fig. 4d). or GO alone was less pronounced (Supplementary Figs. 8 and 9). Feature plots of the ILC-like cluster in GO+BA-exposed larvae (Fig. 4c,d, Next, we asked whether GO alone or in combination with the micro- cluster 8) are displayed in Fig. 4e. The corresponding feature plots for bial metabolite BA would elicit immune responses in zebrafish embryos the control sample are shown in Supplementary Fig. 16 (note that the (Fig. 3a–d). Indeed, we observed a profound induction (~60-fold) of lck gene encoding IL-10 is not present in this cluster). Furthermore, the (lymphocyte-specific protein tyrosine kinase) in GF embryos exposed specific markers in the cell population corresponding to ILC210-like to GO+BA, whereas neither GO or BA alone had any effect. This was cells in exposed larvae are shown in Supplementary Fig. 17. It is noted not the case in CV embryos nor did we detect the induction of lck in that although it has been hypothesized that a regulatory ILC population AhR-deficient zebrafish. Furthermore, neither AA nor PA (alone or in exists28, such cells (that is, an ILC subset expressing FOXP3) have thus combination with GO) upregulated lck in GF zebrafish (Fig. 3e). Since far not been identified in mice or humans29. However, IL-10-producing we applied embryos at 5 dpf, corresponding to the early larval stage, ILC2 cells have been associated with regulatory activities29. It is, thus, the immune responses may be mainly attributed to the innate immune relevant to note that we identified a subset of cells in GF zebrafish with system23. Hence, GO+BA triggered the induction of lck, a molecular markers of ILC2 cells along with il10 and foxp3. These cells were also marker that is shared by all the three innate lymphoid cell (ILC) sub- found to express il1rl1 (also known as st2), encoding a receptor for types24, in a strictly AhR-dependent manner, and this was observed IL-33 (Supplementary Fig. 17). Previous studies in mice have shown that only in GF zebrafish. Furthermore, gene expression profiling showed IL-10-producing ILC2 cells can be generated following the activation by that the transcription factor genes, namely, gata3 and stat6, and the alarmins such as IL-33 and retinoic acid27. However, the gene encoding cytokine-encoding genes, namely, il4 and il13, corresponding to ILC2 IL-33 is absent from the zebrafish genome, and one may speculate that Nature Nanotechnology | Volume 18 | January 2023 | 42–48 44 Letter https://doi.org/10.1038/s41565-022-01260-8 a Control (GF) GO+BA (GF) c d WT-CV-cyp1a WT-GF-cyp1a * 140 * 60 * * Relative mRNA expression level Relative mRNA expression level * # 120 * 40 # # # # # # # # # # # 100 # # # # 20 # # 80 3 3 1 µm 1 µm 1 µm * 2 2 *** * 1 * * b 1 # # ** *** 0 0 i ii BA 50 BA 50 G 10 G 30 G 10 G 30 O +B .5 O B .5 BA .5 G 30 A0 5 BA .5 O B .5 +B 2.5 +B 2.5 O + .5 O B .5 FI 5 FI 5 G rol O B .5 G ol O B 5 30 BA .5 O B 1.5 O +B .5 O +B .5 Z Z O B . . . G 30+ A1. C C G 10 A2 G 10 + A2 0 G + 2 0 G 0 + A2 G 30+ A0 A2 A2 G 0 + A1 1 G 0 A2 tr G 30 A2 O O O O O O t 50 A 50 A on on 3 B C C 3 3 O G G e f 200 µm 200 µm 3 ahr2–/–-CV-cyp1a 3 ahr2–/–-GF-cyp1a Relative mRNA expression level Relative mRNA expression level iii 2 2 # # D Intensity (a.u.) G 1 1 ### # # # * * # # # # # # ** ****** # # # *** # ** *** # 90 µm 0 0 2D G ol G 10 G 30 BA 50 5 O B 1.5 O B .5 O +B .5 O B 5 O B .5 O +B .5 +B 2.5 FI 5 Z G ol G 10 G 30 BA 50 5 O B 1.5 O B .5 O +B .5 O B 5 O B .5 O +B .5 +B 2.5 FI 5 Z 0. G 0+ A1. . 0. G 0+ A1. . C C G + 2 G 30 A0 G 10+ A2 G 30 A2 A2 G + 2 G 30 A0 G 10+ A2 G 30 A2 A2 tr tr O O O A O A O O 30 A 50 A 30 A 50 A on on B B C C 3 3 500 1,000 1,500 2,000 2,500 3,000 3,500 G G Wavenumber (cm–1) g BF CYP1A BA Merged CYP1A BA BF CYP1A BA Merged CYP1A BA Fig. 2 | GO plus BA trigger AhR-dependent CYP induction in GF zebrafish. of comparisons between control and the indicated treatments (*p < 0.05, a, GO visualized by TEM analysis. The arrows indicate GO sheets interacting **p < 0.01, ***p < 0.001), and for comparisons between BA versus GO+BA with microvilli in the gut of GF zebrafish larvae exposed to 5 µg ml–1 of GO for (#p < 0.05, ##p < 0.01, ###p < 0.001). g, Visualization of cyp1a induction using 24 h. Scale bars, 1 μm. b, Light microscopy ((i) and (ii)) and Raman confocal Tg(cyp1a:GFP) zebrafish larvae under GF conditions following exposure to the mapping (iii) to verify the presence of GO in the gut. The analysis was done on combination of GO (30 μg ml–1) and resorufin butyrate (5 μM). BA (red) was found 5 dpf zebrafish exposed to 5 µg ml–1 of GO for 24 h. The spectra shown represent in the gut lumen, and cyp1a induction (green) was noted in the GI epithelial cells the average of 10,000 spectra across the whole area scan. c,d, Relative mRNA (Supplementary Figs. 8 and 9 show additional positive and negative controls). expression of cyp1a in WT-CV (c) and WT-GF (d) larvae. e,f, Relative mRNA The upper and lower rows are from two different individuals. The pseudo-three- expression of cyp1a in ahr2−/− CV (e) and ahr2−/− GF (f) larvae. FICZ was used as a dimensional images were generated with the 2.5D tool in ZEN 3.0, and the highest- positive control. Data are presented as mean values ± s.d. of three independent intensity values are represented by the greatest extension in the z axis. BF, bright experiments (n = 3). Student’s t-test (two sided) was used for the analysis field. Scale bars, 50 μm. another ‘IL-33-like’ factor may be involved. Thus, our studies provided not affect the development of the latter cells (reviewed in another work2). evidence for the existence of ILC-like cells in zebrafish embryos and sug- Here we uncovered a novel aspect of gut immunity where a nanomate- gested plasticity within ILC lineages in GF zebrafish exposed to GO+BA. rial (GO) in combination with a microbial metabolite (BA) was shown to Further studies are required to functionally characterize these cells. elicit AhR-dependent type 2 immune responses in GF zebrafish with the Previous studies in mice have shown that the AhR plays a role in the induction of ILC2-like cells displaying attributes of regulatory cells. In maintenance and function of ILC3s in the gut2. However, GF conditions did conclusion, this study demonstrated that GO influences the crosstalk Nature Nanotechnology | Volume 18 | January 2023 | 42–48 45 Letter https://doi.org/10.1038/s41565-022-01260-8 a WT-CV-lck b WT-GF-lck e WT-GF-lck 80 ** Relative mRNA expression level Relative mRNA expression 8 60 Relative mRNA expression 100 *** 80 ****** ** 40 6 60 * ** level 40 20 8 level 4 20 * 8 6 * * 6 4 2 4 2 * 2 0 0 0 G trol G 10 G 30 BA 50 O BA1.5 G 30 BA .5 G 30 A .5 G 30 +BA .5 G O10+BA 1.5 G 30+BA2.5 50 BA .5 A2 5 .5 G rol G 10 G 30 BA 50 A1.5 G 30 BA2.5 G 30 BA .5 G 30 BA .5 G 10 BA 1.5 G 30 BA .5 50 BA .5 A2 5 .5 G rol BA 0 AA .5 O P .5 O B 5 O A 5 +P 2.5 .5 +B 2. +B 2. G 30+ A2. G 0+ A2. 3 0 O +B 2 0 0 O + 0 O + 2 O + 2 O + 2 2 2 A2 O O O O t O O t O 30 A on on on B O + C C O + C 3 O O O G G G c d f 25 ** ahr2–/–-CV-lck ahr2–/–-GF-lck Relative mRNA expression level 8 8 Relative mRNA expression 20 lck+ cells per gut 6 6 15 level 4 4 10 ns 2 2 5 ** * ** * * * ** ** ** ** *** *** ** 0 0 0 G rol G 10 G 30 BA 50 A1.5 G 30 BA2.5 G 30 BA .5 G 30 BA .5 G 10 BA 1.5 G 30 BA .5 50 BA .5 A2 5 .5 l A l A G rol G 10 G 30 BA 50 BA0.5 G 30 BA2.5 G 30 BA .5 G 30 BA .5 G 10 BA 1.5 G 30 BA .5 50 BA .5 A2 5 .5 tr tr +B 2. +B +B +B 2. 0 O + 0 O + 2 O + 2 1 O + 0 O + 2 O + 2 C C O O O t O O t O on on V_ F_ O O B G G G O + C O + C O + C O + V_ F_ G C O O G G g CV GF BF GFP Merged BF GFP Merged (i) (iii) lck+ lck+ Thymus Thymus GI tract GI tract (ii) (iv) + lck+ lck Thymus Thymus GI tract GI tract lck+ lck+ Fig. 3 | GO plus BA trigger AhR-dependent homing of lck+ cells in GF fish. between control and exposed larvae (**p = 0.0014). f, Quantification of lck+ cells a–d, Relative mRNA expression of lck in WT-CV (a), WT-GF (b), ahr2−/−-CV (c) and homing to the gut. A significant increase in lck+ cells in the gut was observed on ahr2−/−-GF (d) zebrafish larvae on exposure to GO alone, BA alone or GO+BA at GO+BA exposure under GF conditions, but not in CV zebrafish. Student’s t-test the indicated concentrations. Supplementary Fig. 10 shows the additional gene (two sided) was used for the analysis of comparisons between control and the profiling results. Data are presented as mean values ± s.d. of three independent treatments (ns = no significant difference; **p = 0.0055). The numbers of lck+ experiments (n = 3). Student’s t-test (two sided) was used for the analysis cells were quantified based on seven individuals per group. g, Visualization of lck+ of comparisons between control and the indicated treatments (*p < 0.05, cells using Tg(lck:GFP) zebrafish larvae under CV and GF conditions exposed as **p < 0.01, ***p < 0.001). e, PCR analysis of lck following the exposure to GO and/ follows: (i) CV fish (control), (ii) CV fish (GO+BA), (iii) GF fish (control), (iv) GF fish or various SCFAs in WT-GF zebrafish larvae. AA, acetic acid; BA, butyric acid; (GO+BA) (Supplementary Fig. 11 shows the experiments with resorufin butyrate). PA, propionic acid. Student’s t-test (two sided) was used for the comparison Scale bars, 100 μm. Nature Nanotechnology | Volume 18 | January 2023 | 42–48 46 Letter https://doi.org/10.1038/s41565-022-01260-8 a Control (GF) b 10 CLP 9 ILC-like cells % cells 20 1 8 0 0 25 1 7 4 2 50 5 10 2 6 75 Cluster 3 0 ILC precursor like 4 5 Avg. exp. tSNE_2 3 8 5 4 2.0 9 6 1.5 0 3 7 1.0 −20 6 8 2 CHILP 9 0.5 CD4+/CD8+T cells 1 0 10 0 HSPC 7 −40 n r tb 1 a 2a cc rc cd a cd −1 ni a ni 1k 4a ni 5 ga 9 st 3 6 3 il1 il4 1 ifn 2 g1 fa gz fb n k fo kl.1 3a 0 il7 zb ag /f m 16 r9 8 il2 tr yc ta il1 tr at il1 tn 4 tr ro tn id 7a xp tr ni r m −25 0 25 50 tSNE_1 c GO+BA (GF) d 12 12 11 40 % cells 10 CD4+T cells 0 0 HSPC 1 9 9 25 2 8 50 20 3 0 CLP 7 75 Cluster 4 5 6 Avg. exp. tSNE_2 3 6 5 2.0 0 6 4 1 7 4 1.5 8 3 1.0 7 2 9 CHILP 10 2 0.5 11 −20 11 1 0 CD8+T cells 10 8 ILC210 12 0 ILC progenitor 5 ILC2 and ILC3 NK-like: nitr+nkl.1+gzmk+ ILC-like cells n r tb 1 a 2a cc rc cd a cd −1 ni a ni 1k 4a ni 5 ga 9 st 3 6 3 il1 il4 1 ifn 2 g1 fa gz fb n k fo kl.1 3a 0 il7 zb rag /f m 16 r9 8 il2 tr yc ta il1 tr at il1 tn 4 tr ro tn id xp 7a tr ni m −25 0 25 tSNE_1 e (key) nitr4a nitr9 il1rl1 40 40 40 2.5 40 20 20 3 20 2.0 20 0.6 tSNE_2 tSNE_2 tSNE_2 tSNE_2 ILC2 and ILC210 like 2 1.5 0.4 0 0 0 0 ILC3 like 1.0 1 0.2 −20 −20 −20 0.5 −20 0 0 0 −25 0 25 −25 0 25 −25 0 25 −25 0 25 tSNE_1 tSNE_1 tSNE_1 tSNE_1 gata3 stat6 il4 foxp3a 40 4 40 40 40 2.5 3 20 3 20 20 2.0 20 tSNE_2 tSNE_2 tSNE_2 tSNE_2 2 2 2 1.5 0 0 0 0 1 1.0 1 −20 1 −20 −20 0.5 −20 0 0 0 0 −25 0 25 −25 0 25 −25 0 25 −25 0 25 tSNE_1 tSNE_1 tSNE_1 tSNE_1 il10 rorc il17a/f1 il22 40 40 40 40 1.6 4 4 20 1.2 20 20 3 20 2 tSNE_2 tSNE_2 tSNE_2 tSNE_2 3 0 0.8 0 2 0 2 0 1 −20 0.4 −20 1 −20 1 −20 0 0 0 0 −25 0 25 −25 0 25 −25 0 25 −25 0 25 tSNE_1 tSNE_1 tSNE_1 tSNE_1 Fig. 4 | scRNA-seq analysis of lck+-enriched cells collected from GF zebrafish. (d). e, Feature plots of the ILC-like cluster in GO+BA-exposed larvae (cluster 8 in a,c, Two-dimensional projection of tSNE analysis of 10x RNA-seq data showing c and d), which, in turn, is shown to comprise ILC2-like cells (nitr+gata3+il4+il13+) the heterogeneity of lck+ cells in controls (a) and GO+BA-exposed embryos and ILC3-like cells (nitr+rorc+il17a/f1+il22+), as well as ILC2 cells with attributes of (c). b,d, Dot plots show the average expression level of target genes in each of regulatory ILC-like cells (ILC210 cells) (nitr+gata3+foxp3a+il10+). Supplementary the clusters in control (b) and GO+BA-exposed larvae (d). The size of the dots Fig. 17 provides additional information and Supplementary Fig. 16 shows the indicates the percentage of cells within the cluster that express the gene in feature plots of ILC-like cell markers in the control sample (corresponding to question. The red boxes delineate cluster 4 (corresponding to ILC-like cells) in cluster 4 in a and b). control (b) and cluster 8 (corresponding to ILC-like cells) in GO+BA-exposed fish between the gut microbiome and immune system with the induction of system ‘senses’ GO+BA as a pathogen. This has important implications for a type 2 immune response. Type 2 immunity is best known for its protec- our understanding of the hazard potential of graphene-based materials tive role against helminth infections, as well as for its pathogenic role in and other nanomaterials and places AhR at the nexus of the bidirectional allergic diseases such as asthma30. Our findings imply that the immune communication between the gut microbiome and innate immune system. Nature Nanotechnology | Volume 18 | January 2023 | 42–48 47 Letter https://doi.org/10.1038/s41565-022-01260-8 Online content 17. Rolig, A. S., Parthasarathy, R., Burns, A. R., Bohannan, B. J. Any methods, additional references, Nature Portfolio reporting sum- M. & Guillemin, K. Individual members of the microbiota maries, source data, extended data, supplementary information, disproportionately modulate host innate immune responses. Cell acknowledgements, peer review information; details of author contri- Host Microbe 18, 613–620 (2015). butions and competing interests; and statements of data and code avail- 18. Jia, P. P. et al. Chronic exposure to graphene oxide (GO) induced ability are available at https://doi.org/10.1038/s41565-022-01260-8. inflammation and differentially disturbed the intestinal microbiota in zebrafish. Environ. Sci. Nano 6, 2452–2469 (2019). References 19. Cholan, P. M. et al. Conserved anti-inflammatory effects and 1. Stockinger, B., Di Meglio, P., Gialitakis, M. & Duarte, J. H. The aryl sensing of butyrate in zebrafish. Gut Microbes 12, 1824563 (2020). hydrocarbon receptor: multitasking in the immune system. Annu. 20. Okazaki et al. Microbiome alteration in type 2 diabetes mellitus Rev. Immunol. 32, 403–432 (2014). model of zebrafish. Sci. Rep. 9, 867 (2019). 2. Rothhammer, V. & Quintana, F. J. The aryl hydrocarbon 21. Kucki, M. et al. Uptake of label-free graphene oxide by Caco- receptor: an environmental sensor integrating immune 2 cells is dependent on the cell differentiation status. J. responses in health and disease. Nat. Rev. Immunol. 19, Nanobiotechnol. 15, 46 (2017). 184–197 (2019). 22. Rannug, A. & Rannug, U. The tryptophan derivative 3. Ebbo, M., Crinier, A., Vély, F. & Vivier, E. Innate lymphoid cells: 6-formylindolo[3,2-b]carbazole, FICZ, a dynamic mediator of major players in inflammatory diseases. Nat. Rev. Immunol. 17, endogenous aryl hydrocarbon receptor signaling, balances cell 665–678 (2017). growth and differentiation. Crit. Rev. Toxicol. 48, 555–574 (2018). 4. Fadeel, B. et al. Safety assessment of graphene-based materials: 23. Trede, N. S., Langenau, D. M., Traver, D., Look, A. T. & Zon, L. I. The focus on human health and the environment. ACS Nano 12, use of zebrafish to understand immunity. Immunity 20, 367–379 10582–10620 (2018). (2004). 5. Nicholson, J. K. et al. Host-gut microbiota metabolic interactions. 24. Björklund, A. K. et al. The heterogeneity of human CD127+ innate Science 336, 1262–1267 (2012). lymphoid cells revealed by single-cell RNA sequencing. Nat. 6. Korecka, A. et al. Bidirectional communication between the aryl Immunol. 17, 451–460 (2016). hydrocarbon receptor (AhR) and the microbiome tunes host 25. Hernández, P. P. et al. Single-cell transcriptional analysis reveals metabolism. npj Biofilms Microbiomes 2, 16014 (2016). ILC-like cells in zebrafish. Sci. Immunol. 3, eaau5265 (2018). 7. Rosser, E. C. et al. Microbiota-derived metabolites 26. Simoni, Y. et al. Human innate lymphoid cell subsets possess suppress arthritis by amplifying aryl hydrocarbon receptor tissue-type based heterogeneity in phenotype and frequency. activation in regulatory B cells. Cell Metab. 31, 837–851.e10 Immunity 46, 148–161 (2017). (2020). 27. Seehus, C. R. et al. Alternative activation generates IL-10 producing 8. Yang, W. et al. Intestinal microbiota-derived short-chain fatty type 2 innate lymphoid cells. Nat. Commun. 8, 1900 (2017). acids regulation of immune cell IL-22 production and gut 28. Jegatheeswaran, S., Mathews, J. A. & Crome, S. Q. Searching immunity. Nat. Commun. 11, 4457 (2020). for the elusive regulatory innate lymphoid cell. J. Immunol. 207, 9. Li, S. et al. Aryl hydrocarbon receptor signaling cell intrinsically 1949–1957 (2021). inhibits intestinal group 2 innate lymphoid cell function. Immunity 29. Spits, H. & Mjösberg, J. Heterogeneity of type 2 innate lymphoid 49, 915–928.e5 (2018). cells. Nat. Rev. Immunol. 22, 701–712 (2022). 10. Douglas, A. E. Simple animal models for microbiome research. 30. Gause, W. C., Rothlin, C. & Loke, P. Heterogeneity in the initiation, Nat. Rev. Microbiol. 17, 764–775 (2019). development and function of type 2 immunity. Nat. Rev. Immunol. 11. Peng, G. et al. Nitric oxide-dependent biodegradation of 20, 603–614 (2020). graphene oxide reduces inflammation in the gastrointestinal tract. Nanoscale 12, 16730–16737 (2020). Publisher’s note Springer Nature remains neutral with regard to 12. Hammad, H. & Lambrecht, B. N. Barrier epithelial cells jurisdictional claims in published maps and institutional affiliations. and the control of type 2 immunity. Immunity 43, 29–40 (2015). Open Access This article is licensed under a Creative Commons 13. McKenzie, G. J., Bancroft, A., Grencis, R. K. & McKenzie, A. N. Attribution 4.0 International License, which permits use, sharing, A distinct role for interleukin-13 in Th2-cell mediated immune adaptation, distribution and reproduction in any medium or format, responses. Curr. Biol. 8, 339–342 (1998). as long as you give appropriate credit to the original author(s) and the 14. Chen, H. et al. Acute oral administration of single-walled carbon source, provide a link to the Creative Commons license, and indicate nanotubes increases intestinal permeability and inflammatory if changes were made. The images or other third party material in this responses: association with the changes in gut microbiota in article are included in the article’s Creative Commons license, unless mice. Adv. Healthcare Mater. 7, e1701313 (2018). indicated otherwise in a credit line to the material. If material is not 15. Chen, L. et al. Dysregulation of intestinal health by included in the article’s Creative Commons license and your intended environmental pollutants: involvement of the estrogen receptor use is not permitted by statutory regulation or exceeds the permitted and aryl hydrocarbon receptor. Environ. Sci. Technol. 52, use, you will need to obtain permission directly from the copyright 2323–2330 (2018). holder. To view a copy of this license, visit http://creativecommons.org/ 16. Sun, Y. et al. Activation of aryl hydrocarbon receptor by dioxin licenses/by/4.0/. directly shifts gut microbiota in zebrafish. Environ. Pollut. 255, 113357 (2019). © The Author(s) 2023 Nature Nanotechnology | Volume 18 | January 2023 | 42–48 48 Letter https://doi.org/10.1038/s41565-022-01260-8 Methods the regional committee for animal experiments in Stockholm (ethical Characterization of GO permit no. 14049-2019). Zebrafish embryos carrying a point mutation GO prepared by Hummers’ method was obtained from Graphenea in ahr2 (ahr2hu3335) were generated at the Wellcome Sanger Institute and (Spain). The full physicochemical characterization of the GO suspen- provided by the European Zebrafish Resource Center at Karlsruhe Insti- sions in H2O and E3 medium has been previously reported11. Here we tute of Technology. Offspring were raised to adulthood and genotyped provide information on GO suspensions in H2O and different cell cul- for the ahr2hu3335 point mutation with DNA isolated from fin clips34. DNA ture media, that is, Dulbecco’s modified Eagle medium (DMEM) (used was extracted using the QIAamp DNA Mini Kit (QIAGEN) and amplified for the HT-29 cell line) and Roswell Park Memorial Institute (RPMI) (used with the point mutation detection primers (ahr2-mut-F, TATTGCTAG- for primary macrophages). TEM and atomic force microscopy was GCAGAGAGCAC; ahr2-mut-R, GATGTCTTCTGTGATGATTTCG) using performed as previously described11. Briefly, carbon-film-coated grids the DreamTaq Green PCR Master Mix (ThermoFisher). The PCR product were pretreated with a glow discharge using a current of −25 mA and was purified with ExoSAP-IT Express reagent (Applied Biosystems), and for 30 min duration. Ten microlitres of the solutions at 50 µg ml–1 were loaded on the ABI 3730 PRISM DNA analyser (Applied Biosystems) for deposited on the grid, removing excess sample after 2 min of deposi- DNA sequencing. Zebrafish determined to be WT (ahr2+/+), heterozy- tion time. TEM images were acquired at 80 kV, and the size distribution gous (ahr2+/−) or homozygous (ahr2−/−) for the point mutation in ahr2 analysis was determined using ImageJ software version 1.5. Atomic were used for further experiments (adult zebrafish, 4.5 months old; force microscopy images were acquired using a Bruker Multimode 8 larvae, 5 dpf). atomic force microscope in the tapping mode with an OTESPA probe. Adult zebrafish exposure Endotoxin detection Adult zebrafish of different genotypes (WT, ahr2+/− and ahr2−/−) were GO suspensions were evaluated for endotoxin content by using the continuously exposed to GO (50 µg l–1 or 500 µg l–1) for seven days. TNF-α expression test based on primary human monocyte-derived For WT, six female and six male fish were included, whereas for the macrophages, as previously described31. The cells were isolated from other genotypes, three female and three male fish were included. The buffy coats obtained from the Karolinska University Hospital. The zebrafish were housed together before the genotyping (see above) samples are completely anonymized, and the data cannot be traced and were then housed in separate fish tanks for one week before the back to the individual donors. Human monocyte-derived macrophages exposures to GO. During the seven-day exposure, the fish were fed were exposed to GO (5–75 μg ml–1) for 24 h and cell viability was evalu- once per day in the morning with an approximately equal amount for ated using the alamarBlue assay (ThermoFisher Scientific)32. Human each tank. The fish water was refreshed 1 h after the feeding and GO was monocyte-derived macrophages were then incubated for 24 h with added to the exposure groups. At day 7, the fish were sacrificed with GO at 25 μg ml–1 in the presence or absence of the lipopolysaccharide tricaine, and the intestines were dissected under a stereomicroscope inhibitor polymyxin B (10 μM). Lipopolysaccharide (0.01 µg ml–1) was and fixed in 2.5% glutaraldehyde in 0.1 M phosphate buffer at pH 7.4 used as a positive control. Cell culture supernatants were collected and stored at 4 °C for TEM analysis (WT), or in 4% formaldehyde for and TNF-α levels were determined by enzyme-linked immunosorb- histopathological examination (WT) or stored at −80 °C for subse- ent assay (MabTech). A standard curve was established based on quent 16S rRNA gene sequencing (WT, ahr2+/−). The ahr2+/− zebrafish lipopolysaccharide-induced TNF-α. were used for the analysis of the gut microbiome composition as the survival of some ahr2−/− fish in the high-dose GO exposure group AhR reporter cells was compromised. HT-29-Lucia AhR reporter cells established from the human HT-29 colon adenocarcinoma cell line were obtained from InVivoGen. The cells Histopathology were initially cultured in DMEM supplemented with 4.5 g l–1 glucose, After fixation in 4% formaldehyde for at least 24 h, the intestines were 2 mM l-glutamine, 10% foetal bovine serum (FBS), 100 U ml–1 penicil- dehydrated in ethanol and embedded in paraffin using an embed- lin, 100 μg ml–1 streptomycin and 100 μg ml–1 Normocin. Following ding station (Tissue-Tek, Sakura Finetek). The paraffin-embedded at least two passages, the growth medium was supplemented with tissues were then cut using a microtome (Microm HM 360, Marshall Zeocin (100 µg ml–1). For the AhR activity assay, cells were detached Scientific). The thickness of each slice was 5 µm. The slices were depar- with trypsin, centrifuged and resuspended in the test medium without affinized following the steps of xylene for 5 min, xylene for 5 min, Normocin and Zeocin at a density of 2.8 × 105 cells ml–1. The cells were 100% ethanol for 3 min, 95% ethanol for 3 min and distilled water for exposed to AA, BA or PA (Sigma-Aldrich); FICZ (Sigma-Aldrich) (200 nM) 3 min. Sections were stained with haematoxylin and eosin for general was used as a positive control. Then, cell supernatants were transferred morphological examination, and Alcian blue and periodic acid–Schiff into a 96-well plate, and QUANTI-Luc assay solution was added. The reagent (ThermoScientific) for goblet cell identification using a Zeiss measurements were recorded using a Tecan Infinite F200 plate reader. Axioplan microscope equipped with an Olympus SC30 digital camera. The numbers of goblet cells per villus are presented as average results HT-29 cell differentiation of six slices per condition. HT-29 cells were cultured in RPMI 1640 supplemented with 10% FBS, 2 mM l-glutamine, 100 U ml–1 penicillin, 100 μg ml–1 streptomycin and Ultrastructural analysis 2 g l–1 glucose for 21 days, as described33. The cell medium was changed TEM analysis35 of the GI tract of control and exposed animals was per- every other day until the cells reached confluence. Thereafter, the formed on 5 dpf larvae and dissected intestines of adult zebrafish. Fol- medium was changed every day for 21 days. The differentiated cells lowing the primary fixation, samples were rinsed with 0.1 M phosphate were then exposed to GO (30 μg ml–1) for 24 h. Then, the cells were buffer followed by post-fixation in 2% osmium tetroxide in 0.1 M phos- washed, harvested using trypsin–EDTA (0.25%) and fixed in 2.50% phate buffer at pH 7.4 at 4 °C for 2 h. The samples were then ethanol glutaraldehyde in 0.1 M phosphate buffer (pH 7.4). The samples were dehydrated stepwise followed by stepwise acetone/LX-112 infiltration then processed for TEM analysis, as described in the ‘Ultrastructural and finally embedded in LX-112. Semi- and ultrathin sections were analysis’ section. prepared on a Leica EM UC7 ultramicrotome. The ultrathin sections were then contrasted with uranyl acetate followed by Reynolds lead Zebrafish genotyping citrate and examined using a Hitachi HT7700 transmission electron Zebrafish were housed and experiments were conducted in compliance microscope operating at 100 kV. Digital images were acquired using a with national ethical guidelines, and the present study was approved by 2k × 2k Veleta charge-coupled device camera. Nature Nanotechnology Letter https://doi.org/10.1038/s41565-022-01260-8 16S rRNA gene sequencing medium was refreshed daily. At day 4, the hatched embryos were used Total DNA of the zebrafish gut samples was extracted using the QIAamp for sterility validation. Gnotobiotic zebrafish were validated through DNA Mini Kit (QIAGEN). The amplicon PCR was then performed with two different approaches: bacterial growth on Luria broth plates and bacterial universal primers to target the V3 and V4 regions of the 16S DNA amplification by bacterial universal primers. To this end, ten rRNA gene using the forward primer TCGTCGGCAGCGTCAGATGT- embryos were randomly selected from the culture flasks. For the first GTATAAGAGACAGCCTACGGGNGGCWGCA and the reverse primer approach, the embryos were washed and homogenized with 200 μl GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAGGACTA CHVGGG- sterile medium. Then, 100 μl homogenate was spread on the Luria broth TATCTAATCC. The KAPA HiFi HotStart ReadyMix (KAPA Biosystems) plate and incubated at 37 °C overnight. The bacterial colony formation was used for the PCR reactions. Index PCR was performed with the was checked on the next day. For the second approach, DNA of the Nextera XT Index Kit (Illumina), followed by cleanup with AMPure collected embryos was extracted using the QIAamp DNA Mini Kit (QIA- XP beads. The concentrations of the DNA libraries were quantified GEN). DNA was amplified with bacterial universal primers (515F/806R) by an Agilent 2100 Bioanalyser (Agilent Technologies). DNA libraries using the DreamTaq Green PCR Master Mix (ThermoFisher). The PCR were then sequenced on the Illumina MiSeq system to generate raw product was resolved by electrophoresis, and Midori Green Direct paired-end reads (2 × 150 bp). (NIPPON Genetics Europe) was used for the visualization of the DNA. The gel image was captured using a Gel Doc EZ System (Bio-Rad). Gene sequencing analysis Preprocessing. Primers were removed from sequences using cutadapt Zebrafish larvae exposure (version 2.9) and sequences were quality checked by FastQC (version WT and ahr2−/− adult zebrafish were maintained at 28.0 ± 0.5 °C on a 0.11.9) and combined using multiqc (version 1.9.dev0). ASVs were 14 h:10 h light/dark cycle in the fish breeding circulatory system at inferred from the bacterial 16S rRNA gene sequences were using the the zebrafish core facility at Karolinska Institutet. Two pairs of male/ DADA2 package36 (version 1.14.1). Forward and reverse reads were trun- female fish were placed in a single mating tank with a divider one day cated after 282 and 222 bp, respectively, and further filtered using the before spawning. Spawning was triggered by removing the divider in function filterAndTrim with default options, with the exception that the morning. Embryos were collected after 2 h, washed and then trans- the maximum expected error rate was set at 2. The rest of the functions ferred to the E3 medium in a Petri dish. Healthy and fertilized embryos in the DADA2 pipeline were performed with default options, with the at the same developmental stages were selected and raised up to 5 dpf. exception that before sample inference, that is, removing sequencing Zebrafish larvae at 5 dpf (CV WT, GF WT, CV ahr2−/− and GF ahr2−/−) were errors, all the samples were pooled rather than handled sample wise exposed to GO or BA or GO+BA for 24 h. Each treatment was performed (default). The ASVs of 16S rRNA gene sequences were assigned to taxa in three replicates, and each replicate sample contained ten larvae. using the SILVA taxonomic training data formatted for DADA2 (ver- FICZ (200 nM) was used as a positive control for cyp1a induction. After sion 138)37. Before the normalization step, non-bacterial sequences the exposure, the samples were fixed in 2.5% glutaraldehyde for TEM were removed. ASVs were then normalized using cumulative sum analysis or stored at −80 °C for RT-qPCR, as detailed above. scaling38. The steps in the analysis of the data from preprocessing to further downstream analyses were done in the R environment (3.6.2, Raman confocal analysis R Core Team, 2019). The presence of GO in the GI tract of zebrafish larvae was monitored by Raman confocal analysis11. Briefly, larvae exposed to GO for 24 h Statistical analysis. Before unsupervised and supervised analyses were washed, anaesthetized in 0.01% tricaine solution and positioned of the preprocessed and normalized data, ASVs occurring only in one in 1% low-melt agarose on glass slides. Samples were then dried on a sample were removed. To investigate the overall variation in the gut plate heater at 50 °C. Raman analysis was performed using a confocal microbiota of WT and AhR-deficient zebrafish, principal coordinate Raman microscope (WITec alpha300 system) with a laser of 532 nm analysis was performed using the function cmdscale in the R package wavelength set at an integration time of 0.5 s and ×60 magnification. vegan39. Supervised analyses were, in turn, used to study the effect of The scan area for each sample was adjusted to 50 × 50 µm2. The spectra GO exposure on the gut microbiota of WT and ahr2+/− animals, taking shown represent the average of 10,000 spectra recorded across the into account the gender of the animals. Permutational multivariate whole area scan. GO could be detected on the basis of its characteristic analysis of variance (PERMANOVA) and distance-based redundancy Raman signature, that is, the D band (1,354 cm−1), G band (1,582 cm−1) analysis (dbRDA) were performed using the functions adonis and and 2D band (2,690 cm−1). dbrda, respectively, in the R package vegan. In the PERMANOVA and dbRDA analyses, exposure, gender and genotype were applied RT-qPCR analysis as categorical variables to supervise the microbiota composition. RNA extraction was performed on 5 dpf larvae (WT versus ahr−/−) Between-sample Bray–Curtis distances were used in principal coor- exposed to GO+BA or not using the RNeasy Mini Kit (QIAGEN). RNA dinate analysis, PERMANOVA and dbRDA. Statistical significances concentration was quantified by a NanoDrop spectrophotometer were based on 9,999 random permutations. To determine which ASVs (ThermoFisher). Total RNA (500 ng) was reverse transcribed using were differently abundant in exposures and genotypes, the function the iScript Advanced cDNA Synthesis Kit (Bio-Rad). The transcription fitFeatureModel was used in R package metagenomeSeq (retrieved of target genes was quantified using a QuantStudio 5 Real-Time PCR from https://cbcb.umd.edu/software/metagenomeSeq). System (Applied Biosystems). The reaction mixtures were formulated using Maxima SYBR Green/ROX qPCR Master Mix (ThermoScientific). GF zebrafish derivation Thermal cycling conditions were as follows: 95 °C for 10 min, 40 cycles The generation of GF zebrafish followed previously established pro- of three-step amplification of 15 s at 95 °C, 30 s at 60 °C and 30 s at 72 °C. tocols40. In brief, 2 hpf embryos were transferred to Petri dishes with Primers (Supplementary Table 3) were purchased from Sigma-Aldrich. sterile E3 medium, supplemented with ampicillin (100 μg ml–1), kana- The transcription level of each target gene was normalized to rpl13. mycin (5 μg ml–1) and amphotericin B (250 ng ml–1), and incubated at The relative mRNA expression level was calculated relative to control 28 °C. At 50% epiboly up to the shield stage (6 hpf), the embryos were using the 2−ΔΔCt method. surface disinfected with 0.1% polyvinylpyrrolidone-iodine for exactly 2 min, followed by 0.003% sterile bleach immersion for 18 min. The Tg(cyp1a:GFP) reporter strain embryos were then rinsed with sterile E3 medium, transferred to flasks Transgenic Tg(cyp1a:GFP) zebrafish41 were provided by the China and incubated at 28 °C. The viability was monitored, and the sterile Zebrafish Resource Center. GF zebrafish larvae were generated as Nature Nanotechnology Letter https://doi.org/10.1038/s41565-022-01260-8 described above and exposed at 5 dpf to GO (30 µg ml–1), red fluores- Single-cell RNA sequencing cent resorufin butyrate (Sigma-Aldrich) (5 µM) and a combination of The samples were loaded on a 10x GemCode Single-Cell Instrument GO and resorufin butyrate for 24 h. FICZ (Sigma-Aldrich) (200 nM) (10x Genomics) to generate single-cell gel beads in emulsion (GEMs), was used as a positive control. After exposure, the larvae were washed, and libraries were constructed using the Chromium Next GEM Single anaesthetized in 0.01% tricaine and positioned in 1% low-melt aga- Cell 3’ GEM, Library & Gel Bead Kit v3.1 (10x Genomics). Briefly, GEMs rose for analysis by confocal microscopy (Zeiss LSM880, ZEISS). The were generated by combining barcoded Single Cell 3’ v3.1 Gel Beads, a bright-field images were acquired with the transmitted light detector Master Mix containing cells and partitioning oil onto Chromium Next (T-PMT). The green and red fluorescence images were captured under GEM Chip G. Following GEM generation, the gel bead was dissolved, 488 and 561 nm to visualize cyp1a and butyrate, respectively. The primers containing an Illumina TruSeq Read 1, 10x Barcode, unique images were analysed in ZEN 3.0 software, blue edition (ZEISS). The molecular identifier and poly(dT) sequence were released, and the cells 2.5D view tool (ZEN) was used to generate pseudo-three-dimensional were lysed. The barcoded, full-length cDNA was synthesized, purified images and the highest-intensity values are represented by the greatest and amplified by PCR for library construction. Dual-indexed libraries extension in the z axis. containing the P5 and P7 primers used in Illumina amplification were prepared for an estimated 5,000 nuclei per sample. Paired-end, dual Tg(lck:GFP) reporter strain indexing sequencing of libraries was conducted on a NovaSeq 6000 Transgenic Tg(lck:GFP) zebrafish42 were obtained through the Euro- sequencing system (Illumina). Cell Ranger 6.0.1 (10x Genomics) pipe- pean Zebrafish Resource Center. CV and GF larvae were exposed at 5 dpf lines (cellranger mkfastq and cellranger count) were used to convert to a combination of GO (30 µg ml–1) and butyrate (2.5 mM) for 24 h. GO Illumina Base call files to FASQT format, align sequencing reads to the and BA were pre-incubated for 1 h before the exposure. After the expo- zebrafish reference genome GRCz11 and generate feature-barcode sure, the larvae were washed, anaesthetized and positioned for analysis matrices. The generated feature-barcode matrices were used for the by confocal microscopy (Zeiss LSM880, ZEISS), as described above. subsequent analysis. The numbers of lck-positive cells homing to the gut were manually quantified based on seven fish per group. Resorufin butyrate (5 µM) was Data analysis of 10x Genomics data also applied to better visualize the interactions between GO, butyrate The analysis of the 10x Genomics data was performed using the Seu- and lck-positive cells under GF conditions. The z-stack analysis was rat toolkit (version 4.0.6) (available at https://satijalab.org/seurat/ performed with the interval of 1 µm of each slice, and z projections index.html) in the R environment (RStudio, version 4.2.0). A standard were made with Fiji (ImageJ)43. preprocessing workflow was applied, including QC metrics, data nor- malization and scaling, and the detection of highly variable features. Zebrafish dissociation and cell sorting Specifically, Chromium Single Cell 3’ samples with unique feature GF WT (AB) zebrafish larvae were exposed at 5 dpf to the combination of counts over 6,000 or less than 200 were filtered out. In addition, GO (30 µg ml–1) and BA (2.5 mM) for 24 h. GO and BA were pre-incubated the percentage of mitochondrial content was set to be less than 10%. for 1 h before the exposure. Twenty larvae were used as one replicate, After QC, 3,115 single cells from the WT control sample and 3,012 and four replicates, that is, eighty larvae in total, were used for each single cells from the WT GO+BA sample (experiment 1) and 2,312 condition. After the exposure, zebrafish larvae were dissociated for single cells from the Tg(lck:GFP) control sample and 2,669 single cells single-cell suspensions following the published protocol44. Briefly, from the Tg(lck:GFP) GO+BA sample (experiment 2) were used for the zebrafish larvae were euthanized with 0.01% tricaine for 5 min, col- downstream analysis. The raw counts that passed the QC were centred lected in a 1.5 ml tube and washed three times with phosphate-buffered by a scale factor of 10,000 and log transformed. The highly variable saline. The dissociation was initiated by adding 500 μl of pre-warmed features were detected using the ‘FindVariableFeatures’ command in enzyme mix containing 460 μl of 0.25% trypsin–EDTA (Gibco) and 40 μl Seurat by directly modelling the mean–variance relationship inherent of collagenase (Sigma-Aldrich) (100 mg ml–1), followed by mechanical in the single-cell data and 2,000 features per dataset were returned45. dissociation using P1000 and then P200 pipette tips on a heat block A linear transformation (scaling) was applied to the data before the at 30 °C until tissues were no longer visible (about 10 min). The dis- principal component analysis. Only the previously determined highly sociation was then stopped by adding 800 μl DMEM supplemented variable features were used as the input for the calculation of principal with 10% FBS. The cell pellets were collected by centrifugation at components using the ‘RunPCA’ command in Seurat. The ‘JackStraw- 1,000 rpm for 5 min at room temperature, followed by washing with Plot’ and ‘ElbowPlot’ commands were applied to visualize the rank- phosphate-buffered saline. The cells were then resuspended in 0.5 ml ing of the principal components. Significant principal components DMEM + 10% FBS. Four replicates from each condition were pooled showing strong enrichment of features with low p values were selected together at this step and filtered through a 40 μm nylon mesh, with an for the subsequent clustering analysis46. Specifically, 20 and 15 prin- additional washing step with DMEM + 10% FBS. The cell suspension was cipal components were identified for the first and second RNA-seq then stained for 10 min at room temperature with the fluorescent DNA experiment, respectively. The t-stochastic neighbour embedding dye DRAQ7 (Invitrogen) (3 μM) to allow for the exclusion of non-viable transformation47 was achieved by the ‘RunTSNE’ command in Seu- cells by using fluorescence-activated cell sorting (BD FACSAria III, BD rat. The positive marker genes in each cluster compared with all the Biosciences) operating with FCS Express software version 7.0 (DeNovo remaining cells were identified using the ‘FindAllMarkers’ command. Software). DRAQ7– cells of each sample were sorted into tubes contain- An identified feature requires to be detected at a minimum percentage ing DMEM + 10% FBS and were immediately placed on ice and pro- of 0.25 in each of the cluster of cells and differentially expressed (on ceeded further for scRNA-seq analysis, as described below. In addition average) with a Log2FC threshold of 0.25 between the clusters. The to the sorting of cells from WT embryos, cells were also sorted from cell clusters were annotated according to the Zebrafish Information GF transgenic Tg(lck:GFP) zebrafish to enrich lck+ cells for scRNA-seq. Network (https://zfin.org/) and on the basis of current literature25,48,49. Fifty larvae were used as one replicate, and four replicates, that is, two The visualization of the marker expression was demonstrated by heat hundred larvae in total, were used for each condition. The exposure, maps, violin plots, feature plots and dot plots using the ‘DoHeatmap’, and the procedures for single-cell dissociation and DRAQ7 staining, ‘VlnPlot’, ‘FeaturePlot’ and ‘DotPlot’ commands, respectively. The were the same as for the WT zebrafish. However, the gating strategy scRNA-seq integration analysis was further performed to identify the was based on forward scatter, DRAQ7 and GFP. DRAQ7−GFP+ cells were cell types that are present in both datasets and to find cell-type-specific sorted into tubes containing DMEM + 10% FBS and were immediately responses to the stimulation50. The normalization and identification placed on ice, and processed for scRNA-seq analysis. of variable features for each dataset were independently performed Nature Nanotechnology Letter https://doi.org/10.1038/s41565-022-01260-8 as described above. The features that were repeated variables across 42. Langenau, D. M. et al. In vivo tracking of T cell development, the datasets were selected for integration using the ‘SelectIntegration- ablation, and engraftment in transgenic zebrafish. Proc. Natl Features’ command in Seurat, followed by ‘FindIntegrationAnchors’ Acad. Sci. USA 101, 7369–7374 (2004). and ‘IntegrateData’ commands to create an integrated data assay. The 43. Schindelin, J. et al. Fiji: an open-source platform standard workflow described above, including scaling data, principal for biological-image analysis. Nat. Methods 9, component analysis, clustering and t-stochastic neighbour embed- 676–682 (2012). ding transformation was also performed on the integrated data. The 44. Bresciani, E. et al. An efficient dissociation protocol for generation ‘split.by’ argument was used to visualize the two conditions side by of single cell suspension from zebrafish embryos and larvae. side, and the ‘subset’ argument was used to plot the data in a specific MethodsX 5, 1287–1290 (2018). cluster. The transcriptomics data from both analyses are deposited 45. Brennecke, P. et al. Accounting for technical noise in in ArrayExpress at EMBL-EBI. single-cell RNA-seq experiments. Nat. Methods 10, 1093–1095 (2013). Statistical analysis 46. Macosko, E. Z. et al. Highly parallel genome-wide expression Experiments were performed at least three times, each in triplicate profiling of individual cells using nanoliter droplets. Cell 161, for each condition. Statistical differences were analysed using the 1202–1214 (2015). Student’s t-test (GraphPad Prism version 8.2.0). The data presented 47. Van der Maaten, L. & Hinton, G. Visualizing are mean values ± standard deviation (s.d.). The differences between data using t-SNE. J. Mach. Learn. Res. 9, groups or treatments were considered significant when p < 0.05. The 2579–2605 (2008). analysis of the 16S rRNA data and scRNA-seq data is described above. 48. Wagner, D. E. et al. Single-cell mapping of gene expression landscapes and lineage in the zebrafish embryo. Science 360, Reporting summary 981–987 (2018). Further information on research design is available in the Nature Port- 49. Xia, J. et al. A single-cell resolution developmental folio Reporting Summary linked to this article. atlas of hematopoietic stem and progenitor cell expansion in zebrafish. Proc. Natl Acad. Sci. USA 118, Data availability e2015748118 (2021). The 16S rRNA gene sequencing data are deposited at NCBI (accession 50. Stuart, T. et al. Comprehensive integration of single-cell data. Cell no. PRJNA682318) and the two scRNA-seq datasets are deposited at 177, 1888–1902 (2019). ArrayExpress (accession nos. E-MTAB-11984 and E-MTAB-11991). Source data are provided with this paper. Acknowledgements This work was supported by the European Commission through References the Graphene Flagship (grant nos. 785219 and 881603) (B.F.). 31. Mukherjee, S. P. et al. Detection of endotoxin contamination of We thank L. Haag (Electron Microscopy Core Facility, Karolinska graphene based materials using the TNF-α expression test and Institutet) for assistance with the TEM analysis, and J. Sommertune guidelines for endotoxin-free graphene oxide production. PLoS (RISE Research Institute of Sweden, Stockholm) for assistance with ONE 11, e0166816 (2016). the Raman analysis. We also thank the Zebrafish Core Facility at 32. Mukherjee, S. P., Kostarelos, K. & Fadeel, B. Cytokine profiling Karolinska Institutet for the zebrafish husbandry. We acknowledge of primary human macrophages exposed to endotoxin-free the Bioinformatics and Expression Analysis (BEA) Core Facility graphene oxide: size-independent NLRP3 inflammasome at Karolinska Institutet for 16S rRNA gene sequencing, and the activation. Adv. Healthcare Mater. 7, 1700815 (2018). Biomedicum Flow Cytometry Core Facility for assistance with 33. Hekmati, M. et al. A morphological study of a human the fluorescence-activated cell sorting. We thank the Eukaryotic adenocarcinoma cell line (HT29) differentiating in culture. Single Cell Genomics Facility at SciLifeLab, Stockholm, for expert Similarities to intestinal embryonic development. Cell Differ. Dev. assistance with the scRNA sequencing, and the Swedish National 31, 207–218 (1990). Infrastructure for Computing (SNIC) at Uppsala Multidisciplinary 34. Goodale, B. C. et al. AHR2 mutant reveals functional diversity of Center for Advanced Computational Science (UPPMAX) for providing aryl hydrocarbon receptors in zebrafish. PLoS ONE 7, e29346 the computational resources. R. Francis at the National Bioinformatics (2012). Infrastructure Sweden (NBIS) provided advice on the scRNA-seq data 35. Mukherjee, S. P. et al. Graphene oxide elicits membrane lipid analysis. Graphenea (Spain) is acknowledged for the provision of the changes and neutrophil extracellular trap formation. Chem 4, GO samples. We thank the Wellcome Sanger Institute (UK) for the ahr2 334–358 (2018). mutant zebrafish strain, and China Zebrafish Resource Center (China) 36. Callahan, B. J. et al. DADA2: high-resolution sample inference and European Zebrafish Resource Center (Germany) for the provision from Illumina amplicon data. Nat. Methods 13, 581–583 (2016). of the Tg(cyp1a:GFP) and Tg(lck:GFP) zebrafish strains used in the 37. Quast, C. et al. The SILVA ribosomal RNA gene database project: present study. improved data processing and web-based tools. Nucleic Acids Res. 41, D590–D596 (2013). Author contributions 38. Paulson, J. N., Stine, O. C., Bravo, H. C. & Pop, M. Differential G.P. performed all the experiments, analysed the data including abundance analysis for microbial marker-gene surveys. Nat. single-cell transcriptomics data and drafted the paper. H.M.S. Methods 10, 1200–1202 (2013). performed the analysis of the 16S rRNA gene sequencing data, 39. Oksanen, J. et al. vegan: community ecology package. R supervised by H.A., and both of them drafted the paper. N.L. package version 2.5-6. https://CRAN.R-project.org/package= characterized the GO samples, under the supervision of K.K. L.B. vegan (2019). provided advice on the in vivo experiments. B.F. conceived and 40. Pham, L. N., Kanther, M., Semova, I. & Rawls, J. F. Methods for coordinated the study, analysed the data and wrote the paper. generating and colonizing gnotobiotic zebrafish. Nat. Protoc. 3, All the co-authors approved the final version of the paper. 1862–1875 (2008). 41. Kim, K. H. et al. Cyp1a reporter zebrafish reveals target tissues for Funding dioxin. Aquat. Toxicol. 134–135, 57–65 (2013). Open access funding provided by Karolinska Institutet. Nature Nanotechnology Letter https://doi.org/10.1038/s41565-022-01260-8 Competing interests Correspondence and requests for materials should be addressed to The authors declare no competing interests. Bengt Fadeel. Additional information Peer review information Nature Nanotechnology thanks Desheng Pei Extended data is available for this paper at and Stefan Oehlers for their contribution to the peer review of https://doi.org/10.1038/s41565-022-01260-8. this work. Supplementary information The online version contains supplementary Reprints and permissions information is available at material available at https://doi.org/10.1038/s41565-022-01260-8. www.nature.com/reprints. Nature Nanotechnology Letter https://doi.org/10.1038/s41565-022-01260-8 Extended Data Fig. 1 | Integrated analysis of wild-type (WT) germ-free (GF) the feature plots showing the expression of genes involved in lipid metabolism control versus GO+BA samples. (a) 2D projection of the tSNE analysis showing (apoda.2) and proteolysis (ela3l, ela2l, ela2, ctrb1, prss1, prss59.1, prss59.2), lck+ lymphocytes (cluster 6) and the cluster corresponding to the liver and respectively, in the GF control samples (b,d) versus GO+BA samples (c,e). pancreas (cluster 14) which is noticeably expanded in the GO+ BA fish. Below are Nature Nanotechnology nature research | reporting summary Corresponding author(s): Prof. Bengt Fadeel, Karolinska Institutet Last updated by author(s): Sep 30, 2022 Reporting Summary Nature Research wishes to improve the reproducibility of the work that we publish. This form provides structure for consistency and transparency in reporting. For further information on Nature Research policies, see our Editorial Policies and the Editorial Policy Checklist. Statistics For all statistical analyses, confirm that the following items are present in the figure legend, table legend, main text, or Methods section. n/a Confirmed The exact sample size (n) for each experimental group/condition, given as a discrete number and unit of measurement A statement on whether measurements were taken from distinct samples or whether the same sample was measured repeatedly The statistical test(s) used AND whether they are one- or two-sided Only common tests should be described solely by name; describe more complex techniques in the Methods section. A description of all covariates tested A description of any assumptions or corrections, such as tests of normality and adjustment for multiple comparisons A full description of the statistical parameters including central tendency (e.g. means) or other basic estimates (e.g. regression coefficient) AND variation (e.g. standard deviation) or associated estimates of uncertainty (e.g. confidence intervals) For null hypothesis testing, the test statistic (e.g. F, t, r) with confidence intervals, effect sizes, degrees of freedom and P value noted Give P values as exact values whenever suitable. For Bayesian analysis, information on the choice of priors and Markov chain Monte Carlo settings For hierarchical and complex designs, identification of the appropriate level for tests and full reporting of outcomes Estimates of effect sizes (e.g. Cohen's d, Pearson's r), indicating how they were calculated Our web collection on statistics for biologists contains articles on many of the points above. Software and code Policy information about availability of computer code Data collection MiSeq Control Software v. 4.0 was used to collect 16S rRNA gene sequencing data from intestines of adult zebrafish. QuantStudio5 Real-Time PCR System Software was used to collect gene expression data, i.e., RT-PCR data from zebrafish larvae. NovaSeq Control Software was used to collect the single-cell RNA sequencing data (10x Genomics). BD FACSDiva 9.0.1 was used to collect the fluorescence-activated cell sorting data. Data analysis 16S rRNA gene sequence analysis was performed with the following software/code: cutadapt (version 2.9), FastQC (version v0.11.9), multiqc (version 1.9.dev0), and analyses were performed in the R environment (3.6.2, R Core Team, 2019) using DADA2 package (version 1.14.1) with rRNA gene database Silva (version 138). Further statistical analyses for 16S rRNA gene sequence data were performed using R packages vegan and metagenomeSeq. Single-cell RNA sequencing analysis was pre-processed using Cell Ranger pipelines (cellranger mkfastq and cellranger count, version 6.0.1, 10x Genomics), and further analyzed using the Seurat (version 4.0.6) in the R environment (RStudio, version 4.2.0). FACS data was analysed using FCS Express™ v. 7.0 software (DeNovo Software, Pasadena, CA). ZEN (version 3.0) software including the 2.5D view tool (for the analysis of the confocal microscopy data) along with Fiji (ImageJ) and GraphPad Prism (version 8.2.0) were also used in this study. For manuscripts utilizing custom algorithms or software that are central to the research but not yet described in published literature, software must be made available to editors and reviewers. We strongly encourage code deposition in a community repository (e.g. GitHub). See the Nature Research guidelines for submitting code & software for further information. April 2020 1 Data nature research | reporting summary Policy information about availability of data All manuscripts must include a data availability statement. This statement should provide the following information, where applicable: - Accession codes, unique identifiers, or web links for publicly available datasets - A list of figures that have associated raw data - A description of any restrictions on data availability The scRNA-seq data (using whole zebrafish embryos and lck-sorted cells from zebrafish embryos) are deposited at ArrayExpress (accession no. E-MTAB-11984 and E-MTAB-11991, respectively) and the 16S rRNA gene sequencing data obtained in adult fish are deposited at NCBI (accession no: PRJNA682318). The zebrafish reference genome GRCz11 can be assessed at NCBI (GCF_000002035.6). Source data for Figures 1-3 and for Supplementary Figures S2, S4, S7, and S10 are provided. Field-specific reporting Please select the one below that is the best fit for your research. If you are not sure, read the appropriate sections before making your selection. Life sciences Behavioural & social sciences Ecological, evolutionary & environmental sciences For a reference copy of the document with all sections, see nature.com/documents/nr-reporting-summary-flat.pdf Life sciences study design All studies must disclose on these points even when the disclosure is negative. Sample size No statistical method was used to predetermine the sample size. For single-cell RNA sequencing experiments, each condition was performed with four pooled replicates and each replicate contained twenty (for the study using whole wild-type embryos) and 50 larvae (for the study using sorted cells from Tg(lck:GFP) embryos), which generated enough viable single cells for the sequencing. For other experiments using zebrafish larvae, each condition was performed in three replicates and each replicate sample contained ten larvae. Ten larvae generated sufficient total RNA for the RT-qPCR analysis. For experiments using adult zebrafish, three female and three male individuals were used in each genotype for each experiment. For the human cell lines, each condition was performed in triplicate in three independent experiments. Data exclusions For gene expression analysis in the ahr2-/- zebrafish, the data were collected from four individuals, two female and two male. The exclusion criteria are specified in the text (in short, some individuals were excluded due to poor survival in the ahr2-/- group exposed to high dose GO). Replication All attempts at data replication were successful as reported in the main text and figure legends and figures (and refer to Methods for details). Randomization For the zebrafish experiments (larvae and adult fish), the animals in each genotype were randomly allocated into experimental groups. For experiments using the human cell line, different cell passages were used for each biological replicate (denoted as independent experiments). Blinding The zebrafish with different genotypes were collected separately. The investigators were not blinded to group allocation. The 16S rRNA gene sequencing data and single-cell transcriptomics data are not affected by knowledge of sample identities (for statistical analysis, see Methods). Reporting for specific materials, systems and methods We require information from authors about some types of materials, experimental systems and methods used in many studies. Here, indicate whether each material, system or method listed is relevant to your study. If you are not sure if a list item applies to your research, read the appropriate section before selecting a response. Materials & experimental systems Methods n/a Involved in the study n/a Involved in the study Antibodies ChIP-seq Eukaryotic cell lines Flow cytometry Palaeontology and archaeology MRI-based neuroimaging Animals and other organisms Human research participants Clinical data April 2020 Dual use research of concern Eukaryotic cell lines Policy information about cell lines Cell line source(s) Human colon adenocarcinoma cell line HT-29 from Sigma-Aldrich, and HT-29-Lucia™ AhR reporter cell line from InVivoGen. 2 Authentication The HT-29 cell line was authenticated by the supplier using STR-PCR profiling. Reporter activity of the reporter cell line has nature research | reporting summary been verified (validated) by functional assays by the supplier; refer to validation results reported by the supplier InVivoGen. Mycoplasma contamination MycoAlert™ Mycoplasma Detection Kit (Lonza) was used to screen for mycoplasma contamination. Cell lines thus tested negative for mycoplasma contamination. The cell lines were regularly screened and at no time was mycoplasma detected. Commonly misidentified lines No commonly misidentified cell lines (i.e., through cross-contamination or other mechanisms, e.g., mislabeling) were used. (See ICLAC register) Animals and other organisms Policy information about studies involving animals; ARRIVE guidelines recommended for reporting animal research Laboratory animals Zebrafish (Danio rerio) were used in the current study including four different genotypes, i.e., wild-type (WT) AB strain (KI Zebrafish Core Facility), AhR deficient strain (ahr2hu3335) (Wellcome Sanger Institute, UK), CYP1A reporter strain (Tg(cyp1a:GFP) (China Zebrafish Resource Center), and LCK reporter strain (Tg(lck:GFP) (European Zebrafish Resource Center). For adult zebrafish, both female and male individuals were used at 4.5 months. For studies on zebrafish larvae, 5 day-post-fertilization (dpf) larvae were used. Wild animals The study did not involve the use of wild animals. Experiments were conducted at the Zebrafish Core Facility at Karolinska Institutet. Field-collected samples The study did not involve samples collected in the field. Experiments performed at the Zebrafish Core Facility at Karolinska Institutet. Ethics oversight The zebrafish study was approved by the Regional Committee for Animal Experiments in Stockholm (ethical permit no. 14049-2019). Note that full information on the approval of the study protocol must also be provided in the manuscript. Flow Cytometry Plots Confirm that: The axis labels state the marker and fluorochrome used (e.g. CD4-FITC). The axis scales are clearly visible. Include numbers along axes only for bottom left plot of group (a 'group' is an analysis of identical markers). All plots are contour plots with outliers or pseudocolor plots. A numerical value for number of cells or percentage (with statistics) is provided. Methodology Sample preparation The single cell suspensions were prepared from whole zebrafish larvae by enzymatic dissociation and mechanical pipetting as specified in the methods. The cell suspension was then stained with a fluorescent DNA dye DRAQ7 to exclude non-viable cells (Invitrogen Cat# D15106) at the dose of 3 μM for 10 min at room temperature before the fluorescence-activated cell sorting. Instrument BD FACSAria III, BD Biosciences, NJ, USA. Experiments were performed at the Biomedicum Flow Cytometry Core Facilty at KI. Software BD FACSDiva 9.0.1 was used to collect data. Data were analyzed using FCS Express™ v. 7.0 (DeNovo Software, Pasadena, CA). Cell population abundance For wild-type zebrafish, the percentage of viable cells collected was around 24% while for Tg(lck:GFP) zebrafish samples, the percentage of viable, GFP (high) positive cells collected was around 0.5%. The cell viability of the obtained cells was checked with trypan blue staining (Bio-Rad Laboratories), and the single cell suspensions were inspected under the light microscope. Gating strategy For wild-type, germ-free zebrafish samples, the gating strategy was based on the forward scatter and DRAQ7 as shown in the Supporting Information. Thus, the DRAQ7 negative (viable) cells were sorted for subsequent single-cell RNA sequencing. For the germ-free Tg(lck:GFP) zebrafish samples, the gating strategy was based on forward scatter, DRAQ7, and GFP (see Supporting Information). DRAQ7 negative, GFP positive cells were sorted for downstream single-cell RNA sequencing analysis. Tick this box to confirm that a figure exemplifying the gating strategy is provided in the Supplementary Information. April 2020 3