An in vivo brain-bacteria interface: the developing brain as a key regulator of innate immunity

Infections have numerous effects on the brain. However, possible roles of the brain in protecting against infection, and the developmental origin and role of brain signaling in immune response, are largely unknown. We exploited a unique Xenopus embryonic model to reveal control of innate immune response to pathogenic E. coli by the developing brain. Using survival assays, morphological analysis of innate immune cells and apoptosis, and RNA-seq, we analyzed combinations of infection, brain removal, and tail-regenerative response. Without a brain, survival of embryos injected with bacteria decreased significantly. The protective effect of the developing brain was mediated by decrease of the infection-induced damage and of apoptosis, and increase of macrophage migration, as well as suppression of the transcriptional consequences of the infection, all of which decrease susceptibility to pathogen. Functional and pharmacological assays implicated dopamine signaling in the bacteria–brain–immune crosstalk. Our data establish a model that reveals the very early brain to be a central player in innate immunity, identify the developmental origins of brain–immune interactions, and suggest several targets for immune therapies.

Innate immunity provides a first line of defense against pathogens1 and plays a crucial role in initiating adaptive immune responses.2 Most of the cell types of hematopoietic origin and genes known to be involved in mammalian innate and adaptive immunity have been identified in X. laevis,3,4 making the frog model an important emerging tool for understanding the cell and molecular biology of immunity.1,5,6 The immune response in the Xenopus embryo is provided uniquely by innate immunity during the first 12 days (d) post-fertilization,7 allowing studies of response to infection by wide range of pathogens without the confounding influences of adaptive immune component.5,8–11

An important emerging field concerns bidirectional intercommunication between two ‘super-systems’, the immune response and the brain, with implications for both basic evolutionary/cell biology and for biomedicine.12–14 Cytokines and other inflammation-related molecules affect vagal afferents or directly on brain, controlling aspects of behavior.15 Exciting recent studies integrating cognitive endpoints and cell biology of lymphatic vessels in the mature brain reveal the link between immune cells and brain functions.2,16–19 These studies reveal the adult brain as a regulator of adaptive immune response.20,21 The immune system has been proposed as a ‘seventh sense’,22 receiving information from pathogen agents to inform the central nervous system. Studies using invertebrate models suggest that neural circuits receive and integrate stimuli coming from pathogens, via G protein-coupled receptors (GPCRs), to guide the immune response.23–25

Several key open questions remain in this fascinating field. First, while much work has been done on the role of the CNS in adaptive immunity, the interplay between brain and innate immunity is still poorly understood. Second, most of the data come from adult organisms, and the developmental origins of brain–immune interactions are largely mysterious. Last, an obvious limitation of mammalian models is the difficulty of performing loss-of-function studies that unequivocally show that the brain controls the immune system.26 To address these knowledge gaps, and to identify intervention strategies for innate immune function, we exploited a unique system for probing the interactions of brain and immunity in embryogenesis under normal conditions and when challenged with human pathogenic bacteria.

We developed a Xenopus model in which we could study brain-dependent events in embryogenesis: the brain is removed during early embryonic stages, but the animal can be kept alive and development continues. The ability of this vertebrate, a popular model for numerous biomedical contexts,27–34 to survive and develop without a brain provides a unique opportunity to understand the role of the brain in diverse systems-level outcomes. Our prior research into brain-dependent developmental signaling revealed that the nascent brain, even before being fully formed, plays an instructive role in patterning somitic muscle and peripheral neural networks.35,36 Here, we use this brainless vertebrate model, with intact spinal cord and peripheral innervation, to demonstrate an unknown role of the brain: regulating the early innate immunity in the presence or absence of infection. Our data, synthesizing a molecular comparison of infected and uninfected animals under normal, brainless, and tail-regenerative conditions, reveal the profound influence of the brain, in part mediated by dopamine signaling, upon susceptibility and response to pathogenic challenge at the cellular, molecular, and organism-wide levels.

We previously showed that the uropathogenic E. coli UTI89 readily colonizes X. laevis embryos when infected at blastula or gastrula stages,9 and that survival rates at 4–5 days after infection could be used as a readout of the degree of activation of the innate immune system. Here, we investigated the role of brain-derived signals in innate immunity and susceptibility by asking how animals respond to systemic infection in the absence of a brain (Fig. 1a).

Bacteria were injected into the blastocoel of embryos at 4.5 × 103 ± 7.8 × 103 cfu/embryo. First, we evaluated the survival rates at stage 48 (4–5 d after infection) for embryos subjected to the different infection conditions: infected with E. coli UTI89 (UTI condition) vs. not-infected (NI condition), belonging to each intervention or experimental group at st. 25 (Fig. 1b–e, Supplementary Fig. S1): intact brain or control embryos (Control or Ctrl; Fig. 1b, c), brain removal (brainless or BR−; Fig. 1d, e; Supplementary Fig. S1a, b), resection of a piece of cervical spinal cord (SC–; Supplementary Fig. S1c, d), tail-bud amputation (Tailless or Tail–), and muscle lysis via Simvastatin treatment (Simv; Supplementary Fig. S1h, j). At least, five biological replicates or dish of embryos (r = 5, n = 40 embryos per replicate, N total = 200 embryos) were used per each group. Results were consistent among the replicates for each condition and group. For NI embryos, survival rates were similar among groups, ranging from 92 to 100% (99 ± 1% for Ctrl, 96 ± 3% for BR−, 99 ± 1% for SC−, 99 ± 1% for Tail−, and 97 ± 5% for Simv; one-way ANOVA P > 0.05; r = 5 per group; Fig. 1f), suggesting any further differences among groups cannot be due to the surgical intervention itself. E. coli infection caused significant differences in the survival rates among groups (one-way ANOVA P < 0.01; Fig. 1g), indicating its utility as infection-survival assay. For intact or Ctrl embryos the mean infection-survival rate was 50 ± 20% (r = 8).

Tail– animals expressed significantly higher survival rates with respect to Ctrl (71 ± 14%, Bonferroni’s posttest P = 0.018, r = 8), apparently resulting from the induced regenerative response.9,37 Conversely, infected animals developing without a brain (BR–) exhibited a low survival rate, with a significant drop to 16 ± 8% (Bonferroni’s posttest P < 0.0001, r = 8). The survival rate for SC– embryos (37 ± 15%) was not significantly different than Ctrl group (P > 0.9999, r = 8), indicating that the effect was not due simply to surgical damage or the requirement for intact CNS overall. As a control for possible effects of general tissue distress on outcomes, and to ascertain potential impact of muscle structure on survival phenotypes,36 we tested Simvastatin treatment (Simv; Supplementary Fig. S1h–j). Simvastatin is a drug well-known to give severe collateral myotoxic effects in human patients38 and zebrafish embryos.39,40 It induces muscle disorganization similar to that observed after brain removal in Xenopus embryos, allowing independent characterization of the effects of brain removal vs. of muscle tissue lysis and disorganization. Simv group’s survival was similar to that of Ctrl (52 ± 9%; P > 0.9999, r = 5), demonstrating that the survival capacity and susceptibility to bacterial infection is not non-specifically modulated by a stressful combination of infection and strong tissue stressor.

We next wanted to ensure that the effects of the brain ablation were not due to the removal of migrating neural crest cells (NC; Supplementary Fig. S1e–g). Survival after NC ablation in absence of infection was similar to the other groups, 98 ± 3% (r = 3, n = 40, N = 120), without any significant death related to the intervention itself. UTI infection in NC– embryos lead to a mean survival rate of 60 ± 13% (r = 3, n = 40, N = 120), similar to that reached in Ctrl animals (61 ± 30%; r = 3, n = 40, N = 120) and significantly different from BR– group (11 ± 12%; r = 3, n = 40, N = 120; one-way ANOVA P < 0.05; Bonferroni’s test for NC– vs. Ctrl P = 0.9365; NC− vs. BR– P = 0.458). Of the several tissues and organs we targeted, surgically and via lytic toxins, the brain was unique in its effects in promoting survival.

To further characterize the role of brain in modulating overall susceptibility to bacterial infection, we assessed the bacterial load in living embryos within each experimental group at 2 d after infection. Microbial analysis showed that the pathogen load was maintained among the different experimental groups, with an average of 3.24 × 106 cfu/ml (one-way ANOVA P > 0.05; Fig. 1h: the average initial dose of UTI E. coli injected was 4.5 × 103 cfu/ml). Since the presence of brain improves survival but does not significantly reduce pathogen load, we analyzed our groups with an HPRI metric—an index that takes into account pathogen load as well as survival (Fig. 1i). Differences in HPRI index demonstrated that developing without an intact Central Nervous System (CNS; brain or spinal cord) significantly decreases tolerance41,42 and increases the susceptibility to infection (one-way ANOVA P < 0.01).

Taken together, the results demonstrate the unique effects of brain removal on susceptibility to infection: survival depends on the presence of the brain regardless of any tissue injury, and the absence of brain makes embryos likely to succumb to bacterial infection.

To understand the increased susceptibility to infection in BR– animals, we conducted a longitudinal assay, scoring survival rates and characterizing apoptosis at several time points after infection and surgery (r = 3, N = 120 per group; Fig. 2a, b). Embryos were analyzed at 2–4 h post-surgery (hps; corresponding to st. 26–27), 8 hps (st. 30), 24 hps (st. 35), 36 hps (st. 40), 48 hps (st. 42) and the end stage-48 time point (4 days ps). For all groups, the first wave of death was detected at 24 hps (or 36 h post infection, hpi; Supplementary Table S1), specially for BR– animals, which went from 87 ± 6% at st. 30 to 54 ± 12% at st. 35 (two-way ANOVA P < 0.01; Bonferroni’s posttest P = 0.0005 for BR– st. 25 vs. st. 35). For Ctrl animals, this peak (or first time point with significant differences respect to st. 25) was reached at st. 40 (48 hpi), with a drop in the survival rate to 68 ± 23%. At st. 40, BR– animals showed a dramatic decreased survival rate (19 ± 7%), leading to significant differences with the rest of the groups, which kept constant until the end of the experiment (Bonferroni’s posttest P < 0.01 for BR– vs. C; see Supplementary Table S1 for over-time values at each stage and per each group). To characterize the cellular mechanisms behind the infection-induced death, we studied the patterning and number of apoptotic cells in Ctrl and BR– animals at the same longitudinal period (Fig. 2b). We used the antibody against Cleaved Caspase-3 (Asp175; CC3, which detects levels of activated caspase-3) as it has been extensively shown to be a reliable marker for apoptosis in Xenopus embryos.43–45 For both Ctrl and BR– groups, apoptosis followed a successive positive slope from st. 26–27, reaching a peak at st. 35, when significant differences are found in the number of cells between them (41 ± 11 vs. 55 ± 11 CC3-positive cells per area in Ctrl vs.

BR–, respectively; two-way ANOVA P < 0.01; Bonferroni’s posttest P = 0.0018; Fig. 2c, d). Morphological analysis of CC3 expression revealed that UTI Infection seems to affect mostly the gut region, characterized by a focus of highly reactivity to CC3 marker, which is focused on a posterior domain in Ctrl animals and occupying a more extensive area in BR– embryos (Fig. 2c, d). Animals at later stages in both groups expressed abnormalities in this region, but with decreased apoptosis, pointing this damage on the digestive domain as indicative of the UTI infection and defining the lethal consequences of infection to occur during the first 48 hpi. From st. 35, apoptosis starts to decrease, being constant and without differences between groups until the end of the experiment (Fig. 2e, f). Taking together, these results show that death after infection-surgery occurs sequentially over the time, as a direct consequence of increased weakness, and it does not happen at the first immediate hours after surgery (st. 26–30). Interestingly, this analysis also shows that the peak of death in absence of brain occurs 24 h earlier than when the brain is present (at st. 35 vs. st. 40), with a significant increased apoptosis, supporting the protective effects of brain against the lethal consequences of the infection.

Next, we asked whether the protective effects of the brain were due to action on the development of the primitive immune system (regardless of infection, as it does with muscle and nerve development) or whether, conversely, the brain acted on the active response to infection (implying differences between Control and BR– animals only after infection). In Xenopus, the earliest known marker of the primitive myeloid is spib-a (spiba). Spib- is a highly conserved ETS transcription factor that marks the primitive myeloid-cell lineage and is required for its development.3,10 spib expression in myeloid precursors has been characterized in Xenopus,10 mouse,46 and humans.47 To specifically probe the role of the brain in establishing key cell types involved in innate immunity, we studied the spiba-positive10 cells in UTI and NI early-staged embryos (st. 28), subjected to the different surgical interventions (Ctrl, BR–, SC–, Tail–; Supplementary Fig. S2).

To compare the prevalence and spatial distribution of the spiba-positive cell population, we characterized four independent regions: face, tail, dorsal flank, and ventral flank. At least, ten different embryos, from different replicates, per each group and condition were used for quantification (r = 3, n = 40, N = 120; Supplementary Fig. S2a, b, t–w and Methods for details of quantification). NI embryos belonging to each experimental group possessed a similar number and distribution of positive cells, mainly located in the ventral region or hematopoietic organ at these stages (Supplementary Fig. S2c, d). In NI embryos, the only intervention that resulted in significant differences at the site injury from intact embryos was tail-bud amputation (one-way ANOVA P < 0.05; Bonferroni’s posttest P = 0.032; Supplementary Fig. S2f). The number of spiba+ cells in Tail– animals was especially higher in the posterior regions, occupying the bud and amputation plane. No differences from Ctrl embryos were detected for any area in BR– animals, suggesting the brain, in the absence of infection, is not necessary to maintain this population of primitive myeloid precursors.

Next, we analyzed the same cell population in embryos that were previously infected with E. coli UTI89 (UTI condition; Supplementary Fig. S2j–n). Neither intervention (brain, SC or tail removal) nor body region showed significant differences in number of spiba+ cells within infected embryos. Comparing effects of the infection on number of cells within each experimental group (Supplementary Fig. S2o–r), we detected a significant decrease in the total number of myeloid precursors as consequence of bacteria injection for all experimental conditions (two-way ANOVA P < 0.05). This generic drop detected in st. 28 embryos indicates that infection with the human uropathogenic E. coli affects the primitive myeloid cells at early stages of immune system development. Intriguingly, tail amputation performed after infection, which showed the highest survival rates, did not induce a higher proliferation of spiba+ cells, suggesting that the higher survival percentage observed in this group might not due to the presence of more myeloid progenitor cells.

Taken together, our results reveal the negative impact of infection at very early embryonic stages and demonstrate that the positive contribution of the brain to surviving infection is not mediated by brain-induced changes in the number of early myeloid-cell precursors.

Considering that the early brain does not control the primitive formation of myeloid cells at early stages, we decided to evaluate the dynamic behavior or migration of immune cells at later stages. We labeled a gene coding for a protein expressed by Xenopus embryonic macrophages, mmp7, that mediates extracellular matrix remodeling, a crucial aspect of their migration.11,48 Homologs of this gene-protein have been extensively used to study macrophage behavior.10,11,48 mmp7-ISH was performed for the four experimental groups (Ctrl, BR–, SC–, Tail–) per each infection condition (UTI vs. NI) in st. 36 embryos. At least, ten different embryos, from different replicates, per each group and condition were used for quantification (r = 3, n = 40, N = 120; Fig. 3). In absence of infection (NI, Fig. 3c–g), the mmp7+ population was only significantly different at the injury site for Tail– animals, as seen with the earlier myeloid spiba+ cells, with an increased number of mmp7+ cells accumulated in the bud tail or amputation plane respect to Ctrl embryos (one-way ANOVA P < 0.01; Bonferroni’s posttest P < 0.0001; Fig. 3d). No differences were detected for any area between Ctrl and BR– animals, suggesting that the brain, in absence of infection, does not affect macrophage migration in the developing embryo. However, in embryos infected early in development with E. coli (UTI, Fig. 3h–l), absence of brain led to a significant increase in the number of mmp7+ cells in ventral areas respect to Ctrl animals (one-way ANOVA P < 0.05; Bonferroni’s posttest P = 0.0464 for BR– vs. Ctrl at ventral area; Fig. 3k), represented by the accumulation of this population along the ventral hematopoietic niche (Fig. 3m, n).

The ventral clumping of mmp7-positive cells in BR– after infection was not detected in SC– and Tail– embryos, resulting in significant differences between BR– and Tail– groups (Bonferroni’s posttest P = 0.0225 for BR– vs. Tail– at ventral area; Fig. 3k), and demonstrating that the effect was not due simply to surgical damage or the need for an intact CNS.

The intragroup comparisons (Fig. 3o–r), analyzing embryos from the same surgery intervention but infected vs. not-infected conditions, and similarly to spiba+ population, demonstrated that the overall number of mmp7+ cells was lower in infected embryos, regardless of experimental intervention. This result confirms that the human uropathogenic E. coli targets the early innate immune system. Interestingly, the differences previously detected between Ctrl and Tail– embryos in the tail area injury site, due to the tail injury itself, with a significant macrophage mobilization to the amputation plane (tail-bud area) observable only in absence of infection (Fig. 3d vs. Fig. 3i), were also displayed when comparing the effects of infection within the Tail– group (Fig. 3r). A significant difference in the number of macrophages at the tail region or amputation site is detected between not-infected and UTI-infected Tail− embryos (Bonferroni’s posttest P = 0.0487 for Tail− NI vs. Tail− UTI at tail region), indicating that the macrophage accumulation to the amputation plane in this group is entirely impaired in presence of bacterial infection. Our results show that the absence of the brain during development leads to defects in the response against infection, specifically the accumulation of mmp7+ cells in the ventral niche reveals defects in macrophage migration.

Given the observed effects of the brain on the response of the developing innate immune (∼st. 30), we next analyzed effects of the brain’s presence on the fully-developed innate system5,7,49–51 and the surrounding innervation, in embryos at later stages of development (∼st. 44–48; see Supplementary Note 1 for more details about embryogenesis of the immune system in Xenopus).

Immunofluorescence using the XL-2 antibody8 revealed the distribution of all Xenopus leukocytes (monoclonal antibodies specific for particular leukocyte subpopulations are not currently available for these species).1 We quantified XL2+ cells along the middle and posterior parts of the animal body in not-infected and infected st. 44–45 Ctrl, BR–, SC–, Tail–, and Simv embryos (Fig. 4a–h). In addition, we included the NC– group in the XL2 analysis to study whether this intervention led to similar phenotypes of leukocyte migration than those ones expressed by BR– animals (Fig. 4e, f). At least, ten different embryos, from different replicates, per each group and condition were used for quantification (r = 3, n = 40, N = 120). Our results for NI animals showed similar leukocyte patterns and number among all the groups, being tail amputation the only intervention that provoked a differential increase of XL2+ cells in the tail region (two-way ANOVA P < 0.01; Bonferroni’s posttest P < 0.0001 for Tail– vs. the rest of the groups; Fig. 4b). Then, we analyzed the XL2 expression in UTI-infected animals subjected to the different interventions. Remarkably, the population of XL2-positive cells after infection for all groups was 2–4-fold greater than in absence of infection, with leukocytes leaving the central blood vessels and spreading along the entire fin. UTI-infected BR– animals expressed the lowest XL2-positive population (19 ± 7 vs. 49 ± 10 XL2-positive cells per area in BR– vs. Ctrl, respectively), with significant differences respect the rest of interventions (Bonferroni’s posttest P < 0.0001 for BR– vs. the rest of the groups).

Thus, these results confirmed our observations at early stages of development. While in absence of bacterial threat, brain is not required for specifying leukocyte number, in presence of bacterial threat brain is required to induce an appropriate response-migrating behavior of the immune cells.

Next, we analyzed the transgenic xlurp::GFP animals, which express GFP in myeloid cells (mainly monocyte/macrophage and granulocyte/neutrophils at these stages)5,52 at st. 46–48 of Ctrl, BR–, SC–, Tail– and Simv embryos. At least, ten different embryos, from different replicates, per each group and condition were used for quantification (r = 3, n = 40, N = 120; Fig. 4i–o, Supplementary Fig. S3, and Supplementary Video S1). One conspicuous feature observed in BR– animals was the presence of longitudinal-like patches of high-density of GFP-positive cells along the dorsal and ventral fin, not detected in the somite or central regions of the body. Consequently, we decided to evaluate the number of immune cells for each experimental group on two independent tail regions: center and periphery, with somite region and fin region, respectively (Fig. 4i–l). The central region after tail removal (Tail–: 97 ± 27 GFP+ cells) exhibited a marked increase in the number of myeloid cells (normalized to area) compared with Ctrl embryos (59 ± 28 GFP+ cells; Dunn’s posttest P = 0.0373; Supplementary Fig. S3a, n), mainly focused at the region close to the amputation plane or injury site, indicating that even weeks after injury, myeloid cells are still invading the injured tail area. This effect was, conversely, not detected after brain removal, as more immune cells were not present either in the face region surrounding the injury site nor the central area. In the peripheral region, absence of the brain provoked a drastic increase in the number of myeloid cells at these late stages, forming defined patches that branched off the fin of the tail, at long distance from the injury site (from 30 ± 21 GFP+ cells in Ctrl group to 86 ± 15 GFP+ cells in BR– group; two-way ANOVA P < 0.05; Bonferroni’s posttest P < 0.0001).

This sprouted-network like patterning of the myeloid population in the fin region of BR– animals displayed a similar pattern to the one detected previously for the peripheral neural network in absence of brain.36

Thus, we decided to study the co-location of peripheral nerves with respect to this myeloid population (using Tub immunofluorescence on xlurp::GFP embryos). Morphological analysis revealed that, in central regions, myeloid cells of BR– exhibited different morphology, with respect to other areas or to Ctrl embryos, with rows of flattened cells, compatible with a macrophage-like network,5 following the internal neuropil.36 This distribution pattern, with myeloid cells in close proximity to the internal neuropil, was entirely absent in Ctrl embryos, where the central areas, with the highest number of GFP cells (Fig. 4i, j; Supplementary Fig. S3e–g) were not intensely occupied by nerves. In the peripheral region, a similar distribution pattern for both myeloid cells and peripheral nerves was detected in BR–, with scattered groups of immune cells in proximity to the highly disorganized neural networks (see the area circled within the dashed-white line as a illustrative example in Fig. 4m–o; Supplementary Fig. S3k–m for Ctrl). SC– animals exhibited a similar number and general distribution of myeloid cells as the Ctrl group but the peripheral-nerve phenotype, as revealed by Tub immunofluorescence, showed the same ectopic growth and sprouting of neural network as in the BR– animals and significantly different from the Ctrl group (Supplementary Fig. S3o–q). Simv-treated embryos did not show differences for any marker, myeloid and nerve phenotypes, with respect to the Ctrl group.

Taken together, our results confirmed that brain removal during early development leads to the ectopic presence of mature myeloid cells at later stages of embryogenesis invading the fin, in close proximity to the aberrant peripheral neural network. This aberrant distribution of immune cells occurs far from the distal anterior injury site and is not mediated by spinal cord pathway. Lack of brain thus produces an immunologically different phenotype than other interventions, such as the removal of a different piece of the body (which induces the increased presence of immune cells but only at the local area of injury), or a severe toxic stress (Simv treatment, which caused no differences in myeloid cells relative to controls), indicating the unique influence of the embryonic brain on immune cell behavior.

Given the differences found for the innate immune response with or without a brain, we sought to identify the transcriptional mechanisms underlying the effects. To characterize the transcripts that could be differentially regulated under each condition, we conducted RNA-seq and compared the transcriptome of Ctrl and BR– embryos, with or without infection (Fig. 5, Supplementary Tables S1–S7, Supplementary Figs. S4 and S5, Supplementary Note 2, and Supplementary Data 1 and 2).

We first asked which transcripts are controlled by infection in an intact normal embryo by comparing the transcriptome of NI vs. UTI Control animals (Ctrl NI vs. Ctrl UTI, black labels in Fig. 5a). A total of 535 DEGs were regulated by infection in intact embryos, and 152 were unique to this group. Conversely, this number was significantly increased up to 719 (and 338 unique) transcripts controlled by infection in brainless embryos (or BR– NI vs. BR– infected; orange labels in Fig. 5a), revealing that the presence of the brain suppresses the transcriptional consequences of infection about twofold. In the absence of infection (comparing Ctrl NI vs. BR– NI, blue labels in Fig. 5a), we found that 115 genes (and 37 unique) were responsive to brain removal per se. To characterize the transcriptional response specifically due to brain removal, and not to the removal of any organ in general, we compared the dataset of DEGs after brain removal from our study with the DEGs obtained after tail removal that were identified in two prior studies, in Xenopus53 and lizard.54 The tail is large appendage including a massive CNS component (spinal cord) and triggers a robust regenerative response, allowing us to identify and exclude from our analysis genes that are not specific to the removal of brain. After subtracting the common transcripts between brain and tail removal, we found that, from the initial 115 genes, 91 were unique to our dataset for brain removal and 24 were common with tail amputation (for the complete list of genes, see Supplementary Data 1, Brain Removal vs. Tail Removal). These data indicate that about 75% of the transcripts detected in our study after brain removal are unique to this intervention. The last comparison, Ctrl UTI vs. BR– UTI (green labels in Fig.

5a) revealed that the response to infection in the absence of brain affects the transcription of 218 (117 unique) genes.