tRNA Queuosine Modification Enzyme Modulates the Growth and Microbiome Recruitment to Breast Tumors

Background: Transfer RNA (tRNA) queuosine (Q)-modifications occur specifically in 4 cellular tRNAs at the wobble anticodon position. tRNA Q-modification in human cells depends on the gut microbiome because the microbiome product queuine is required for its installation by the enzyme Q tRNA ribosyltransferase catalytic subunit 1 (QTRT1) encoded in the human genome. Queuine is a micronutrient from diet and microbiome. Although tRNA Q-modification has been studied for a long time regarding its properties in decoding and tRNA fragment generation, how QTRT1 affects tumorigenesis and the microbiome is still poorly understood. Results: We generated single clones of QTRT1-knockout breast cancer MCF7 cells using Double Nickase Plasmid. We also established a QTRT1-knockdown breast MDA-MB-231 cell line. The impacts of QTRT1 deletion or reduction on cell proliferation and migration in vitro were evaluated using cell culture, while the regulations on tumor growth in vivo were evaluated using a xenograft BALB/c nude mouse model. We found that QTRT1 deficiency in human breast cancer cells could change the functions of regulation genes, which are critical in cell proliferation, tight junction formation, and migration in human breast cancer cells in vitro and a breast tumor mouse model in vivo. We identified that several core bacteria, such as Lachnospiraceae, Lactobacillus, and Alistipes, were markedly changed in mice post injection with breast cancer cells. The relative abundance of bacteria in tumors induced from wildtype cells was significantly higher than those of QTRT1 deficiency cells. Conclusions: Our results demonstrate that the QTRT1 gene and tRNA Q-modification altered cell proliferation, junctions, and microbiome in tumors and the intestine, thus playing a critical role in breast cancer development.


Introduction
Transfer RNAs (tRNAs) read the genetic code and are essential for the proliferation, fitness, and adaptation of the cells in the body. Emerging roles of tRNA in various human diseases have been implicated through genetic mutations, post-transcriptional modifications, overexpression, and RNA-protein interactions. tRNA-derived small RNA fragments (tRFs) have been shown to be associated with breast cancer in the extracellular vehicles and blood, indicating that tRFs may be useful of QTRT1-KO cells was absent (Figure 1b), which was further confirmed by genomic sequencing of the PCR products ( Figure S1a). The reduced QTRT1 expression in the knockout clone was persistent in different passages of cells, indicating that we established stable QTRT1-KO in human MCF7 cells. To confirm the role of QTRT1 in breast cancer cells, we further knockdown QTRT1 in the MDA-MB-231 breast cancer cells ( Figure S2a,b). This model will help us to understand whether a knockdown of QTRT1 is sufficient to induce all of these phenotypes.
Cancers 2020, 12, x FOR PEER REVIEW 3 of 22 amplification of QTRT1-KO cells was absent (Figure 1b), which was further confirmed by genomic sequencing of the PCR products ( Figure S1a). The reduced QTRT1 expression in the knockout clone was persistent in different passages of cells, indicating that we established stable QTRT1-KO in human MCF7 cells. To confirm the role of QTRT1 in breast cancer cells, we further knockdown QTRT1 in the MDA-MB-231 breast cancer cells ( Figure S2a,b). This model will help us to understand whether a knockdown of QTRT1 is sufficient to induce all of these phenotypes. (e,f) Cell proliferation markers of PCNA (proliferating cell nuclear antigen) and Ki67 in WT and QTRT1-KO MCF7 cells were detected using Western blot (E) and immunofluorescence staining (F), respectively. Immunofluorescence staining of Ki67 and DAPI were performed in the cells, and Ki67 index (Ki67 stained cells/total cells) was calculated. Mean ± SD, n = 3; ** p-value < 0.01, *** p-value < 0.001, two-tailed Welch's t-test. We measured the Q-modification levels of the MCF7 wildtype (WT) and QTRT1-KO cells using the standard APB gel method ( Figure S1b). We found that the WT cells had nearly stoichiometric levels of Q-modification, whereas our QTRT1-KO cells showed significantly reduced Q-modification levels (~25%). This incomplete elimination of Q-modification may be derived from either incomplete removal of potential QTRT1 splice isoforms in our cells, or the presence of an unknown gene or pseudogene that can also Q-modify tRNA. In any case, our KO cells did show a significant reduction of Q-levels, consistent with our experimental design.

Knockout of QTRT1 Suppressed Cell Proliferation and Migration
After establishing the QTRT1-KO single clone in MCF7 cells, we examined cell proliferation by the MTT assay, which assesses viable cell metabolic activity. We found that cell proliferation was significantly reduced in QTRT1-KO MCF7 cells (p < 0.01) ( Figure 1d) compared with its parental cells. The suppressed proliferation ability in the cells was further verified by significantly decreasing the expression of PCNA and Ki67, the markers for cell proliferation, in QTRT1-KO MCF7 cells (Figure 1e,f).
The adherent cells have the ability to migrate and heal wounds, like the re-epithelialization of a skin scratch. To investigate the role of QTRT1 in the migration of breast cancer cells, we investigated the wound healing ability of these cells. We found that cells with suppressed QTRT1 expression healed the wound area significantly slower than that of the WT cells (Figure 2a,b). QTRT1 knockout in MCF7 cells significantly reduced cell migration at all the timepoints post wound healing (Figure 2a,b). Three-day post wounding the cells, the WT MCF7 cells healed around 20% of the wound area, whereas the QTRT1-KO cells healed less than 10% of the wound area (p < 0.05). After 9 days, WT cells healed 100% of the wound area compared with around 40% in QTRT1-KO MCF7 cells (p < 0.01) (Figure 2a,b).
The suppression of cell proliferation and migration of QTRT1 deletion were confirmed with the knockdown of QTRT1 in MDA-MB-231 breast cancer cells ( Figure S2c-g).

Knockout of QTRT1 Altered Cell Adhesion and Tight Junctions
Cell adhesion plays an essential role in regulation of fundamental cellular process, such as cell proliferation and migration [11]. To elucidate the mechanism of how suppressed QTRT1 expression leads to altered proliferation and migration of breast cancer cells, we investigated the expression of markers of cell adhesion proteins, including β-catenin and E-cadherin. E-cadherin is also known as an anticancer protein. We found that knockout of QTRT1 significantly increased expression of E-cadherin and β-catenin and in the QTRT1-KO MCF7 cells, compared with the WT cells ( Figure 2c).

Knockout of QTRT1 Decreased Tumor Growth and Altered Tight Junctions Regulators In Vivo
The changes of breast cancer cells in vitro have been replicated using direct orthotopic or heterotopic injection of cells into the mouse [15][16][17][18][19]. To investigate the impact of QTRT1 knockout on tumor outgrowth, the xenograft nude mouse model (n = 10 mice per group) was established by subcutaneous bilateral injection of QTRT1-KO or WT breast cancer cells. Individual mice may develop two, one, or zero tumors. We found that the total tumor number, tumor volume, and weight were significantly reduced in mice injected with QTRT1-KO MCF7 cells compared to the WT cells (p < 0.01) ( Figure 3a). The BrdU index (number of cells stained with BrdU/number of total cells) was significantly lower in mice injected with QTRT1-KO cells than that of WT cells (p < 0.01) (Figure 3b), suggesting less cell proliferation in the QTRT1-KO cells in vivo. The expression of PCNA, another important proliferation regulator of cells, was also significantly decreased in tumors derived from QTRT1-KO MCF7 cells, compared with WT cells (Figure 3c,d).

Knockout of QTRT1 Decreased Tumor Growth and Altered Tight Junctions Regulators In Vivo
The changes of breast cancer cells in vitro have been replicated using direct orthotopic or heterotopic injection of cells into the mouse [15][16][17][18][19]. To investigate the impact of QTRT1 knockout on tumor outgrowth, the xenograft nude mouse model (n = 10 mice per group) was established by subcutaneous bilateral injection of QTRT1-KO or WT breast cancer cells. Individual mice may develop two, one, or zero tumors. We found that the total tumor number, tumor volume, and weight were significantly reduced in mice injected with QTRT1-KO MCF7 cells compared to the WT cells (p < 0.01) (Figure 3a). The BrdU index (number of cells stained with BrdU/number of total cells) was significantly lower in mice injected with QTRT1-KO cells than that of WT cells (p < 0.01) (Figure 3b), suggesting less cell proliferation in the QTRT1-KO cells in vivo. The expression of PCNA, another important proliferation regulator of cells, was also significantly decreased in tumors derived from QTRT1-KO MCF7 cells, compared with WT cells (Figure 3c    As TJ proteins regulate several key signaling pathways in cancer process and development [21], we tested the markers of junction proteins in the tumors. Western blot analysis showed that the QTRT1-KOinduced tumors had significantly increased E-cadherin and β-catenin, but ablated expression levels of claudin-5 ( Figure 3e). Immunofluorescence staining further showed increased β-catenin in tumors from QTRT1-KO injected mice, compared to that in WT MCF7 breast cancer cells (Figure 3f).
The suppression of tumor growth and alteration of TJ proteins in the tumors were also found in the QTRT1 knockdown MDA-MB-231 cells, using the xenograft nude mouse model ( Figure S3). These data suggest that a knockdown of QTRT1 should be sufficient to induce all of these phenotypes.

Bacteria Were Detected in Tumors from the Nude Mice
Two probes, EUB338 and Bfi826, were used for detecting total bacteria and Butyrivibrio fibrisolvens-related clones, which are in the gastrointestinal microbiome and ubiquitously present in the gastrointestinal tract of humans and animals, respectively. Fluorescence in situ hybridization with (f) Immunofluorescence staining of β-catenin was performed on tumors from both QTRT1-KO-injected mice and WT-injected mice. Proteins and DNA were stained with mouse monoclonal anti-β-catenin and anti-mouse Alexa Fluor 488 antibody and DAPI, respectively. Scale bar is 20 µm.
As TJ proteins regulate several key signaling pathways in cancer process and development [21], we tested the markers of junction proteins in the tumors. Western blot analysis showed that the QTRT1-KO-induced tumors had significantly increased E-cadherin and β-catenin, but ablated expression levels of claudin-5 ( Figure 3e). Immunofluorescence staining further showed increased β-catenin in tumors from QTRT1-KO injected mice, compared to that in WT MCF7 breast cancer cells (Figure 3f).
The suppression of tumor growth and alteration of TJ proteins in the tumors were also found in the QTRT1 knockdown MDA-MB-231 cells, using the xenograft nude mouse model ( Figure S3). These data suggest that a knockdown of QTRT1 should be sufficient to induce all of these phenotypes.

Bacteria Were Detected in Tumors from the Nude Mice
Two probes, EUB338 and Bfi826, were used for detecting total bacteria and Butyrivibrio fibrisolvens-related clones, which are in the gastrointestinal microbiome and ubiquitously present in the gastrointestinal tract of humans and animals, respectively. Fluorescence in situ hybridization with probes EUB338 and Bfi826 highlighted the bacteria replication in the tumors from the nude mice. The relative abundance of bacteria in tumor induced from WT cells was significantly higher than that of probes EUB338 and Bfi826 highlighted the bacteria replication in the tumors from the nude mice. The relative abundance of bacteria in tumor induced from WT cells was significantly higher than that of QTRT1-KO cells using both EUB338 and Bfi826 probes (p < 0.01 and p < 0.05, respectively) ( Figure  4a,b). The bacteria's existence in the tumors was confirmed with the tumor samples from the mice challenged with QTRT1-KD and WT MDA-MB-231 cells ( Figure S4a).

Altered Gut Microbiome in the Nude Mice Injected with WT MCF7 Cells
We investigated whether gut bacteria have been changed in the nude mice injected with WT MCF7 cells or QTRT1-KO cells, using the data of 16S rRNA gene sequencing. We focused on hypothesis testing of two diversity measures: alpha diversities (Chao 1 and Shannon diversity) and beta diversity (Bray-Curtis dissimilarity). The analysis of the bacterial 16s rRNA gene sequencing showed that the intestinal microbiome community was different between the mice injected with WT and QTRT1-KO cells (Figure 4c-e; Figure S2).
There were significant changes between MCF7-WT and MCF7-KO groups at the bacterial species level (p = 0.0002). The Shannon values of bacteria species was statistically decreased (FDR = 0.0028) post MCF7-KO cancer cell injection (mean ± SD: 2.1 ± 0.3; median: 2.1), compared to before injection (mean ± SD: 2.5 ± 0.1; median: 2.5). The same trend was found in the animals post-injection of MCF7-WT cells (mean ± SD: 2.1 ± 0.2; median: 2.1) and pre-injection (mean ± SD: 2.4 ± 0.2; median: Beta diversity is to show differences or similarities among samples. We performed a permutational multivariate ANOVA (PERMANOVA) to detect the differences of Bray-Curtis dissimilarity among groups. The breast cancer cell injected groups had significantly different variations between WT MCF7 and QTRT1-KO groups (p = 0.0028) (Figure 4c). Our data showed that the groups accounted for 76% variations of the Bray-Curtis dissimilarities (p = 0.01) (Figure 4c). We then used a nonparametric procedure analysis of similarity (ANOSIM) based on a permutation test for analyzing amongand within-group similarity. ANOSIM further confirmed that the compositional dissimilarity of post-injection of WT MCF7 cells was significantly decreased compared with that before injection, but not of QTRT1-KO cells (p = 0.001). Principal coordinate analysis (PCoA) was used for the visualization of Bray-Curtis dissimilarity. In Figure 4d, PCoA showed that before injection, the ranges of the dissimilarities between WT and KO mice overlapped, whereas the ranges of bacterial dissimilarities increased in mice post injection.
Several core bacteria were markedly changed in mice post cell injection, such as Lachnospiraceae, Lactobacillus, and Alistipes (Figure 4e; Figures S5 and S6). The butyrate-producing genera Lachnospiraceae and Ruminococcaceae, where the genus could enhance producing butyric acid to protect the body from colon cancer [22] and triple-negative breast cancer (TNBC) [23], were found suppressed in the mouse model post breast cancer cell injection. The proportions of Lactobacillus, one of the common colonic microbiota, are frequently correlated with intestinal diseases like Crohn's disease [24], obesity, type 2 diabetes, and cancer [25], and breast cancer tumor development [26] was found decreased as well. We also found a change in abundance of Alistipes, which may have disease-promoting activities with gut diseases and relatively high incidence in the nipple aspirate fluid from people with breast cancer [27][28][29].
The 16S rRNA gene sequencing was also performed using the fecal samples from mice injected with the MDA-MB-231 cells ( Figure S4b-d). We observed altered microbiome in tumors and intestines of nude mice injected with MDA-MB-231 breast cancer cells. Microbial community was different between the mice injected with WT and QTRT1-deficient MDA-MB-231 cells. Our data demonstrated that the bacterial species level between WT and QTRT1-deficient groups had differences in alpha and beta diversities, which are potentially associated with breast cancer status and progression.

Inflammation Enhanced the Bacteria in Tumors of Nude Mice
Based on the results above, we further hypothesized that inflammation may lead to increased bacteria population in tumors. Cancer-related inflammation facilitates unlimited replicative potential, independence of growth factors, resistance to growth inhibition, escape of cell death, enhanced angiogenesis, and tumor extravasation [33,34]. Dysfunction of microbiome (dysbiosis) and bacterial metabolites could impact the tumor microenvironment and enhance chronic inflammation, thus affecting tumor growth and further stimulating bacterial multiplication [35]. The CD68-positive and CD-11b-positive inflammation-related cells were immunostained in the tumors from nude mice injected with WT or QTRT1-KO breast cancer cells, indicating an immune response inside the tumors (Figure 5a). The number of CD68-positive and CD11b-positive cells in the tumors from the mice injected with WT MCF7 cells was significantly higher than that of QTRT1-KO cells (p < 0.05) (Figure 5a). We further investigated the cytokines and chemokines levels in the plasma of the studied animals. MCP-3, the chemotactic gradients implicated in the process of inflammation, and interleukin 6 (IL-6), a pro-inflammatory cytokine, were found markedly increased in the mice injected with WT cells, compared with that of QTRT1-KO cells (Figure 5b). We did not find significant changes of IL-10 or TNF-α (Figure 5b).

Disrupted Tight Junctions in the Colon of Mice Injected with WT MCF7 Cells
Intestinal tight junction [36] proteins, including claudins and zona occludens (e.g., ZO-1), have an important role in maintaining the integrity of the gut epithelium. Claudin-5 and claudin-2, both belonging to the claudin family, have totally different functions and tissue specificity in breasts and intestine. Claudin-5 is highly expressed in breast cancer patients with high-risk metastasis and reoccurrence. It was examined to verity the impact of QTRT1 knockout on the breast cell proliferation and migration (Figure 3). Claudin-2 is a leaky protein regulating intestinal permeability [37,38]. Dysfunction of intestinal TJs, including changes in protein distribution and expression level, could result in increased gut permeability and chronic inflammation [21]. Leaky gut is also known to further enhance dysbiosis and inflammation [37,39]. Thus, we investigated the intestinal TJs of the nude mice that were injected with WT or QTRT1-KO MCF7 cells. We found that the expression of the TJ protein ZO-1 was significantly reduced in the colon of nude mice injected with WT cells (Figure 6a). The structure disruption of ZO-1 was visualized in the colon of WT cell injected mice, whereas the ZO-1 distribution was intact in the colon of mice injected with QTRT1-KO cells (Figure 6a). Claudin-2 is considered a leaky protein that increases the intestinal permeability [37,38]. Interestingly, we found a significant increase of claudin-2 in the colon of WT cell injected mice, suggesting enhanced leaky gut (Figure 6b). In contrast, the colons of mice injected with QTRT1-KO cell showed significantly reduced claudin-2 ( Figure 6b).

Disrupted Tight Junctions in the Colon of Mice Injected with WT MCF7 Cells
Intestinal tight junction [36] proteins, including claudins and zona occludens (e.g., ZO-1), have an important role in maintaining the integrity of the gut epithelium. Claudin-5 and claudin-2, both belonging to the claudin family, have totally different functions and tissue specificity in breasts and intestine. Claudin-5 is highly expressed in breast cancer patients with high-risk metastasis and reoccurrence. It was examined to verity the impact of QTRT1 knockout on the breast cell proliferation and migration (Figure 3). Claudin-2 is a leaky protein regulating intestinal permeability [37,38]. Dysfunction of intestinal TJs, including changes in protein distribution and expression level, could result in increased gut permeability and chronic inflammation [21]. Leaky gut is also known to further enhance dysbiosis and inflammation [37,39]. Thus, we investigated the intestinal TJs of the nude mice that were injected with WT or QTRT1-KO MCF7 cells. We found that the expression of the TJ protein ZO-1 was significantly reduced in the colon of nude mice injected with WT cells (Figure 6a). The structure disruption of ZO-1 was visualized in the colon of WT cell injected mice, whereas the ZO-1 distribution was intact in the colon of mice injected with QTRT1-KO cells (Figure 6a). Claudin-2 is considered a leaky protein that increases the intestinal permeability [37,38]. Interestingly, we found a significant increase of claudin-2 in the colon of WT cell injected mice, suggesting enhanced leaky gut (Figure 6b). In contrast, the colons of mice injected with QTRT1-KO cell showed significantly reduced claudin-2 (Figure 6b).

Discussion
In the current study, we demonstrated that QTRT1 plays a crucial role in breast cancer cell proliferation and growth both in vitro and in vivo. The proliferation and collective migration of QTRT1-KO MCF7 breast cancer cells were markedly suppressed, compared with its parent cells. Meanwhile, QTRT1 knockout leaded to changes in the expression of E-cadherin, β-catenin, and claudin-5, which are important cell junction proteins and facilitate cells adhesin formation. These differences were further confirmed in the xenograft nude mouse model in vivo. Furthermore, we found several core bacteria, such as Lachnospiraceae, Lactobacillus, and Alistipes, markedly altered in mice post injection with breast cancer cells. QTRT1 deficiency cells had reduced relative abundance of bacteria in tumors, compared to wildtype cell-induced tumors. All these changes in both cell culture and animals model levels were verified with another cell line of MDA-MB-231, which further strengthened our conclusions.
In our study, QTRT1 knockout or reduction could significantly increase the expression of β-catenin and change its location from inside the cells to the membrane, which participates in the formation, maintenance, and function of adherence junctions by linking cadherin to the actin cytoskeleton [40]. E-cadherin, one of the classical cadherins, which plays a crucial role in cell integrity and polarity, was also found upregulated in the QTRT1-KO cells. Thus, both the changes of E-cadherin and β-catenin have an impact on cell adhesion, combination, and polarity, and they further reduce the proliferation and migration of MCF7 cells after QTRT1 knockout. Claudins are major adhesion molecules in tight junctions and are strongly expressed in various cancers [41], and claudin-5 is highly expressed in breast cancer patients [15]. Interestingly, we found downregulation of claudin-5 in QTRT1-KO MCF7 cells. Suppressed cell proliferation in QTRT1 knockout contributed to marked reduction of the tumor growth, number, volume, and weight. The queuine-insertase complex, including QTRT1 (queuine tRNA-ribosyltransferase 1) and QTRTD1 (queuine-tRNA ribosyltransferase domain containing 1) subunits, has cytosolic proteins that could incorporate queuine base into tyrosyl, asparaginyl, aspartyl, and histidyl tRNAs. It is possible that modification of QTRT1-knockout on the Q-modified tRNAs could further impact cell gene transcription. As beta-catenin and tight junction proteins are important in cell gene transcription, it is reasonable that these proteins were changed after QTRT1 deficiency. However, more experiments should be performed to verify these possibilities. In the QTRT1-KO cells, we showed significantly reduced Q-modification levels (~25%). This incomplete elimination of Q-modification may be derived from either incomplete removal of potential QTRT1 splice isoforms in our cells, or the presence of an unknown gene or pseudogene that can also Q-modify tRNA. In any case, our KO cells did show a significant reduction of Q-levels, consistent with our experimental design. Q-modification most likely affects translation of a subset of genes. Since our primary goal was to investigate depletion of Q-modification impact on breast tumor growth, a detailed investigation on how translation of specific transcripts is affected will be a major goal in our follow-up study.
Interestingly, we found elevated levels of the microbiome in the tumors from WT-MCF7 cell injected mice. Increased inflammation and disrupted tight junctions in the colon of mice injected with WT MCF7 cells may contribute to accelerated breast cancer development (Figure 6c). As previously reported, the human and animal breasts are not sterile, rather containing a diverse and unique bacterial community. The breast microbiome is distinct from that of other body sites [42]. In vivo, we investigated the bacterial profile in the tumor in situ and in fecal samples. We detected the existence of bacteria in the tumors grown in the xenograft nude mice injected with QTRT1-KO and WT MCF7 cells. Meanwhile, we found that the enriched bacteria in tumors from WT MCF7 cells correlated with increased inflammation inside the tumors. Our data suggest bacterial replication in the tumor is associated with the changed microenvironment. Although it is still unclear whether the microbial differences are the consequences or causes of breast cancer or other breast diseases, the changes in the composition of breast microbiota could contribute to disease development and progression [42]. Similarly, our findings indicate that the QTRT1-associated altered microbiome and microenvironment may contribute to tumor growth.
The source of the bacteria in tumors may shed light on breast tumor development. B. fibrisolvens belongs to the genus Butyrivibrio and is part of the gastrointestinal microbiome of mammals and ubiquitously present in the gastrointestinal tract of many animals, including humans and mice [43]. These bacteria inside the tumors could be translocated from the intestine directly. Indeed, the intestinal lumen and outer mucus layer are the location of an overwhelming majority of gut bacteria, but it seems reasonable to envisage that a very small but significant minority of bacteria could occasionally breach the intestinal epithelium and quickly arrive at other organs or tissues, especially in the nude mice, which are immune deficient, used in our study. Recently, more and more data support the transfer of intestinal bacterial products and bacteria translocation directly or indirectly in some diseases, such as liver disease [44,45]. This hypothesis may be supported by the redistribution of ZO-1 and increased leaky protein claudin-2 in the colon of nude mice, which is related with intestinal permeability [46]. Tight junction proteins are key players in epithelial barrier function in inflammatory bowel disease [46][47][48][49][50]. Furthermore, the gut microbiome community of the mouse model markedly changed after breast cancer cell injection and tumor growth, compared with pre-injection, based on 16S ribosomal RNA gene amplification data. Also, taxonomic richness and diversity differences were also found between the mice injected with WT and KO cells. Intestinal dysbiosis, especially in the mice injected with WT cells, should further enhance dysfunctions in intestine. Therefore, the alteration of TJ proteins and microbiome components in the intestine should enhance bacterial translocation by regulating intestinal bacterial interactions with the mucosal surface, which may support bacterial translocation from the intestine to other locations or tumors. Increased cytokines and chemokines in the plasma of mice injected with WT MCF7 cells also support the idea that inflammation and leaky gut promote bacterial dysbiosis [51] and bacterial translocation.
Our data on the reduced butyrate-producing genera Lachnospiraceae and Ruminococcaceae in the mice with tumor growth support the role of intestinal dysbiosis in promoting breast cancer. Bacteria that enhance producing butyric acid are known to protect the body from colon cancer [22] and triple-negative breast cancer (TNBC) [23]. Intestinal Lactobacillus are frequently correlated with diseases, like Crohn's disease [24], obesity, type 2 diabetes, cancer [25], and breast cancer tumor development [26]. We also found Lactobacillus decreased in mice with injection of breast cancer cells. Interestingly, we found a change in the abundance of Alistipes, which may promote digestive diseases, and relatively high incidence in the nipple aspirate fluid from people with breast cancer [27][28][29]. The potential origin of part of the breast tissue microbiome may be translocated from the gastrointestinal tract in addition to the skin via the nipple-areolar orifices [52]. Our data suggest the potential translocation of the gut microbiome to breast tumors.
The knockout of QTRT1 in MCF7 cells could impact the cell proliferation ability, which influences the microenvironment of the tumors and could feedback to affect the bacterial multiplication in the tumor. Moreover, the inflammation investigated in the tumors with related biomarkers (CD68 and CD11b) may be related with the bacterial growth in the tumors and microenvironment modification. More studies need to be done to investigate the relationship between the microbiome in tumors and tumor growth, and the source of these microorganisms. Another hypothesis was that the bacteria found in the tumors might be transferred from the skin, which touches the surrounding feces in the mouse cage. In the future, studies are also needed to understand the mechanisms of how QTRT1 interacts within TJ proteins to influence cell proliferation and tumor growth.

Regents and Cell Lines
The monoclonal antibody to human QTRT1 was purchased from Santa Cruz Biotechnology (Dallas, TX, USA). Monoclonal antibodies of β-catenin and E-cadherin from BD Transduction (San Jose, CA, USA) and β-actin were purchased from Sigma-Aldrich (St. Louis, MO, USA). Monoclonal antibodies of PCNA and CD11b (Integrin αM) were purchased from Santa Cruz Biotechnology (Dallas, TX, USA).
Claudin-5 and claudin-2 mouse monoclonal antibodies and ZO-1 rabbit monoclonal antibody were purchased from Thermo Fisher Scientific (Rockford, IL, USA). CD68 rabbit polyclonal antibody was purchased from Abcam (Cambridge, MA, USA). All chemicals were purchased from Sigma-Aldrich unless otherwise stated. The breast cancer cell lines, including MCF7 and MDA-MB-231, were provided by Dr. Tao Pan's lab at the University of Chicago.

Transfection and Selection
The highly aggressive breast cancer cell line human MCF7 cells were cultured in a 6-well tissue culture plate in antibiotic-free MCF7 cell growth medium (Eagle's minimum essential medium with 10% FBS, 10 µg/mL bovine insulin, and 10 nM β-estradiol; Invitrogen, Carlsbad, CA, USA) to 70%-80% confluence. Cells were transfected with 2 µg of QTRT1 Double Nickase Plasmid (sc-413456-NIC), 10 µL LTX Lipofectamine, and 2.5 µL PLUS Reagent (Invitrogen) per well (manufacturer's protocol). The plasmid encodes a D10A mutated Cas9 nuclease and a unique, target-specific 20-nt guide RNA (gRNA), which has greater specificity of gene knockout than the CRISPR/Cas9 KO plasmid counterpart. QTRT1 Double Nickase Plasmid-derived puromycin resistance gene was used for positive selection of transfected cells. Cells were selected with 1 µg/mL puromycin for 5 days when further cell death was not observed. After selection, cells were collected and serial diluted onto four 96-well plates. Single cells were expanded to obtain individual clones for further study. The culture of MDA-MB-231 cells was as described before [53], and transfection and selection were as described above.

Clone Validation
Individual clones after expansion to 6-well plate format were collected for clone validation. Cells were lysed in CelLytic™ M Lysis Reagent (Sigma-Aldrich, St. Louis, MO, USA) with a protease inhibitor cocktail (ThermoFisher, Waltham, MA, USA) overnight at 4 • C. The lysate was centrifuged at 16,000 g for 30 min, and 10 µg of supernatant was fractioned on 4-20% SDS PAGE gels, transferred to nitrocellulose, and screened by Western blot with QTRT1 antibody (Santa Cruz, CA, USA; sc-398918). This antibody (Santa Cruz, CA, USA; sc-398918) is specific for amino acids 111-136, exon 3, and a 5 fragment of exon 4. QTRT1Genomic DNA was isolated from edited clones and nonedited MCF7 control cells with DNAzol (ThermoFisher, Waltham, MA, USA) and tested with PCR using QTRT1-specific primers in exon 1 (F1: 5 end of the exon1: GGCGGGAGCAGCTACCCA) and intron downstream of exon 3 (Rev1_3: intron downstream of the exon 3: CCCGGCCTCAAGTGATCTTC). Clone validation of MD-MB-231 cells was performed as described above.

Cell Proliferation Assay
The WT and QTRT1 KO MCF7 cells were plated at a density of 1 × 10 4 cells per well in triplicate. After two days of culture, proliferation was evaluated by the MTT Cell Proliferation Assay Kit (Thermo Fisher Scientific, Rockford, IL, USA) according to the product's instructions [53]. Triplicate wells were counted for each time point, and the whole experiment was repeated three times. The evaluation of proliferation of WT MDA-MB-231 and QTRT1-KD (knockdown) MDA-MB-231 cells was performed as described above using the MTT Cell Proliferation Assay Kit.

Wound Healing Assay
Both wildtype (WT) and QTRT1-KO of MCF7 were plated on dishes (MatTek, Ashland, MA, USA) and cultured in humidified chambers with 5% CO 2 at 37 • C. When cells were grown to confluency, the wounds were scratched and analyzed as described before [53]. Migration of cells into wounded areas was captured every day. The values were the means of three independent wound fields from three independent experiments (n = 3). As described above, the wound healing assay was also performed on the WT and QTRT1-KD MDA-MB-231 cells.

In Vivo Nude Mice Model
The 8-week old, specific-pathogen-free, female BALB/c Nude mice (n = 40) were purchased from Charles River Laboratories (Wilmington, MA). Animals were housed in the Biologic Resources Laboratory (BRL) at the University of Illinois at Chicago [54] and utilized in accordance with the UIC Animal Care Committee [54], the Office of Animal Care and Institutional Biosafety (OACIB) guidelines, and the animal protocol (number ACC . The same and consistent diet was provided to the animals throughout the experiment. The animals were separated as two groups including mice for MCF7-WT cell injection (n = 10) and mice for QTRT1-KO MCF7 cell injection (n = 10). The xenograft model was established by subcutaneous bilateral injection with 1.2 × 10 6 cells in 200 µL of 50:50 Matrigel/PBS (phosphate-buffered saline) into the hind flank [55]. At 60 days post-tumor challenge, tumors and samples were harvested from the animals that were euthanized by IP injection of sodium pentobarbital (100 mg per kg body weight) followed by cervical dislocation. The weight of the tumors was scaled, and the tumor volume (V) was calculated with caliper measurements using formulas V = (W 2 × L)/2 [56]. BrdU staining was performed as previously described [57]. Protein expression in tumor tissues was detected by immunofluorescence staining and Western blot as described below. Wildtype and QTRT1-KD MDA-MB-231 cells (7.5 × 10 6 ) were also analyzed using the BALB/c Nude mice (n = 40), as described above, except that the tumors and samples were harvested at 30 days post cell injection.

Western Blot Analysis
Breast tissue was lysed in stocked lysis buffer [53]. Cultured cells were rinsed twice in ice-cold Hanks' balanced salt solution (Sigma-Aldrich, Saint Louis, MO, USA) and lysed in protein loading buffer then followed by sonication (Branson Sonifier, Danbury, CT, USA) and centrifugation [53]. The target proteins were detected by special primary antibody (1:1000) followed by secondary antibody conjugated to horseradish peroxidase at 1:5000 dilution. The blots were visualized by ECL chemiluminescence (Thermo Scientific, Rockford, IL, USA). All experiments were performed 3-5 times. Western blot bands were quantified using image analyzer (ImageJ, NIH, Bethesda, MD, USA).

Immunofluorescence and Confocal Imaging
Wildtype and knockout/knockdown cells were plated on fibronectin-coated glass coverslips and cultured with medium described above to monolayer in humidified chambers with 5% CO 2 at 37 • C. The cells were immunostained and imaged as described before [53]. All experiments were performed multiple times using independent biological replicates.
Fresh tumors were fixed in 10% neutral buffered formalin followed by paraffin embedding. For immunofluorescence staining [58], slides were incubated in 5% bovine serum albumin (BSA) with 0.1% goat serum in PBS for 1 h at room temperature to reduce nonspecific background. The samples were incubated overnight at 4 • C with primary antibody at 1:100 dilution. The sections were then incubated with secondary antibodies and DAPI for 1 h at room temperature, and they examined with confocal microscope as described before [59].

Immunohistochemistry
Tissues were fixed in 10% neutral-buffered formaldehyde for 2 h, transferred into 70% ethanol, and processed the next day by standard techniques. Immunohistochemistry for inflammation cells was performed on paraffin-embedded sections (4 µm) of tumors. Briefly, the paraffin sections were baked in an oven at 56 • C for 30 min. The sides were deparaffinized and rehydrated in xylene, followed by graded ethanol washes at room temperature. Antigen retrieval was achieved by boiling the slides in a microwave oven with 0.01 M, pH 6.0 sodium citrate buffer. Sides were then incubated in hydrogen peroxide (3% H 2 O 2 in PBS) for 10 min at room temperature, followed by incubation in 5% fetal bovine serum/PBS for 1 h. The inflammation cells were stained with polyclonal CD 68 antibody (Abcam, Cambridge, MA, USA) and monoclonal CD 11b antibody (Santa Cruz, Dallas, TX, USA).

Luminex Immunoassays
The 26 cytokines and chemokines of plasma samples were analyzed using the mouse cytokine and chemokine magnetic 26-ple ProcartalPlex Panel 1 (Thermo Fisher Scientific, Waltham, MA, USA) with 25 µL of samples according to the manufacturer's instructions using a 2 h incubation at room temperature. Samples were read on a MAGPIX TM system platform (Millipore Sigma, Burlington, MA, USA).

Microbial Sampling and Sequencing
The day before cell injection, fresh feces were collected directly from the mouse into the sterile tubes. At the end of the experiment, fresh fecal samples were isolated from the colon and placed into the specially prepared sterile tubes. The samples were kept at low temperature with dry ice and were sent to the UIC RRC (University of Illinois at Chicago Research Resources Center, Chicago, IL, USA) for genomic sequencing. The DNA of samples was extracted using the DNeasy Power Fecal Kit (Qiagen, Hilden, Germany) based on the manufacturer's instructions with a slight modification. The samples were heated at 65 • C for 10 min before homogenizing with FastPre-24 5G bead-beating device (MP Biomedicals, Solon, OH, USA) at 6 m/s for 40 s. The workflow for preparing samples for next-generation amplicon sequencing contains two independent PCR steps. The first stage PCR amplification was performed using primers containing CS1 and CS2 linkers (CS1_341F: 5 -ACACTGACGACATGGTTCTACAGTGCCAGCMGCCGCGGTAA-3 ; CS2_806R: 5 -TACGGTAGCAGAGACTTGGTCTGGACTACHVGGGTWTCTAAT-3 ) to the V3-V4 variable region of the 16S rRNA gene, while the second stage of PCR amplification was performed on the first stage of PCR products using the Fluidigm Access Array barcoded primers. The 16S rRNA gene metagenomic sequencing was performed using MiSeq according to the Illumina protocol.

Sequencing Bioinformatics
After DNA sequencing, all possible raw paired-end reads were evaluated and merged using the PEAR (Paired-End reAd mergeR) software (http://www.exelixis-lab.org/web/software/pear.) [61]. To filter out low-quality reads, quality trimming was processed based on quality threshold (p = 0.01) and length parameters (minimum length = 225). The adapter/primer sequences and ambiguous nucleotides were trimmed from reads. Sequencing chimera filtering was carried out using bot reference and algorithm methods, with sequences that passed both these being retained and operational taxonomic units (OTUs) selected based on a 97% similarity representing each OTU, and both methods utilized the USEARCH (http://www.drive5.com/usearch) algorithm and the reference database silva_132_16S.97 [62]. Then, OTUs were annotated using the USEARCH (http://www.drive5.com/usearch) algorithm as compared with a reference database (silva_132_16S.97) [63]. An OTU table was generated from these results, and alpha and beta diversity metrics were generated, calculated, and visualized as described below.
Initially, 1,314,538 reads were assembled from the source sequencing data. After trimming, the number of reads diminished to 1,290,241, and then after chimera checking it reduced to 1,211,340, which was used as the operational OTUs in the alignment. The OTUs ranged from 54,043 to 65,772 for the individual sample, with a mean of 60,567. Taxonomic assignments of the microbiomes of the studied samples were obtained with the OTU data. On the phylum level, Bacteroidetes and Firmicutes accounted for the major part of the microbial population in all samples (96.4%-99.4%).

Statistical Analysis
Data shown in the bar figures were the average values from at least three independent experiments with the mean ± SD. All statistical tests were two-sided. It was considered statistically significant with p-value < 0.05. One, two, and three asterisks on the bars in the figures represent p-values < 0.05, < 0.01, and < 0.001, respectively. To investigate the dynamic effects of QTRT1-KO/KD on cell migration in breast cancer cells, the group effects from 1 to 9 days (Figure 2a, Figure S2f) were tested using generalized linear mixed models (GLMMs). The GLMMs were performed using SAS version 9.4 (SAS Institute, Inc., Cary, NC, USA); other statistical analyses were conducted by GraphPad Prism 5 (GraphPad Software, Inc., La Jolla, CA, USA).
The alpha and beta diversity indices including Shannon diversity [64,65], Chao1 richness [66], and Bray-Curtis dissimilarity [67] were calculated and visualized to show the microbial abundance and diversity. Bar plot and heatmap were used to visualize the identified "core" bacteria, which were defined as the set of species shared by (almost) all individuals with 0.1% relative abundance in more than 50% of the samples. Principal coordinate analysis (PCoA) plots were used for the visualization of Bray-Curtis dissimilarity [68]. To detect the differences of Bray-Curtis dissimilarity among groups, PERMANOVA was performed, followed by a variance homogeneity assumption testing to ensure the reliability of the PERMANOVA results. Then, a nonparametric procedure analysis of similarity (ANOSIM) based on a permutation test was used for analyzing among-and within-group similarity. The latest version of R and R packages of ampvis2, microbiome, phyloseq, and vegan were used for microbiome data analyses as we did in our book using R software [69].

Availability of Data and Material
The raw sequence data from 16S rRNA gene amplicon sequencing were submitted to NCBI SRA under BioProject accession number PRJNA599383 (https://www.ncbi.nlm.nih.gov/sra/PRJNA599383).

Conclusions
Conclusively, we have demonstrated the important functions of QTRT1/Q-tRNA modification in regulating development of breast cancer.. Knockout or reduction of QTRT1 could significantly suppress the proliferation and migration of breast cancer cells. The knockout of QTRT1 resulted in the loss of aggressiveness of cancer cells both in vitro and in vivo. Microbiome-dependent tRNA Q-modifications regulate breast tumor growth and recruit microbiome to breast tumors. We have identified the close relation between the microbiome and microenvironment in tumors and the intestine. Our findings provide new insights into manipulating QTRT1 as a potential gene in controlling tumor growth and development through the gut-microbiome-tumor axis.
Supplementary Materials: The following are available online at http://www.mdpi.com/2072-6694/12/3/628/s1, Figure S1. Verification of QTRT1 knockout in MCF7 cells by sequencing and Northern blot. Figure S2. Knockdown of QTRT1 suppressed MDA-MB-231 breast cancer cell proliferation, migration, and altered cell adhesion and tight junctions. Figure S3. Knockdown of QTRT1 suppressed tumor growth and altered tight junctions in tumors in vivo. Figure S4. Altered microbiome in tumors and intestines of nude mouse model. Figure S5. Heatmap of the bacterial percentage abundance in the intestine of nude mice. Figure S6. Relative abundance of intestinal bacteria detected with real-time PCR.
Author Contributions: J.S. secured research funding, designed the study, analyzed data, and finalized the manuscript; J.Z., R.L., Y.Z. and W.Z. performed experiments and collected and analyzed data in vitro and in vivo; Z.M. and T.P. generated the KO cells; J.Z. and J.S. drafted the article; Y.X. provided support for the study design and statistical analysis. All authors were involved in writing the manuscript and reviewed and approved the final version. All authors have read and agreed to the published version of the manuscript.