Lipopolysaccharide-induced expansion of histidine decarboxylase-expressing Ly6G+ myeloid cells identified by exploiting histidine decarboxylase BAC-GFP transgenic mice

Histamine is a biogenic amine that is chiefly produced in mast cells and basophils and elicits an allergic response upon stimulation. Histidine decarboxylase (HDC) is a unique enzyme that catalyzes the synthesis of histamine. Therefore, the spatiotemporally specific Hdc gene expression profile could represent the localization of histamine-producing cells under various pathophysiological conditions. Although the bioactivity of histamine is well defined, the regulatory mechanism of Hdc gene expression and the distribution of histamine-producing cell populations in various disease contexts remains unexplored. To address these issues, we generated a histidine decarboxylase BAC (bacterial artificial chromosome) DNA-directed GFP reporter transgenic mouse employing a 293-kb BAC clone containing the entire Hdc gene locus and extended flanking sequences (Hdc-GFP). We found that the GFP expression pattern in the Hdc-GFP mice faithfully recapitulated that of conventional histamine-producing cells and that the GFP expression level mirrored the increased Hdc expression in lipopolysaccharide (LPS)-induced septic lungs. Notably, a CD11b+Ly6G+Ly6Clow myeloid cell population accumulated in the lung during sepsis, and most of these cells expressed high levels of GFP and indeed contain histamine. This study reveals the accumulation of a histamine-producing myeloid cell population during sepsis, which likely participates in the immune process of sepsis.

Histamine is a biogenic amine that elicits allergic and anaphylactic responses in a number of pathophysiological conditions 1,2 . In immunological cells, histamine is primarily synthesized in mast cells and basophils, in which histamine is stored in intracellular granules 3 . Once these cells receive external activating stimuli, the stored histamine is released extracellularly and provokes an allergic response by enhancing vasodilation and increasing vascular permeability 4 . The antigen-IgE immune complex bound to a high-affinity IgE receptor (FcεRI) on the surface of mast cells and basophils is the most potent stimulus that activates the release of histamine from these cells. Subsequently, histamine promotes the progression of allergic and inflammatory diseases such as anaphylaxis and bronchial asthma 3,5,6 .
Histidine decarboxylase (HDC) is a primary and unique enzyme that catalyzes the synthesis of histamine through the decarboxylation of the amino acid L-histidine in mouse and human [7][8][9][10] . Indeed, Hdc gene-deficient mice show a significant decrease in plasma histamine levels, underscoring the essential requirement of HDC for the biosynthesis of histamine 11,12 . It has been reported that LPS (lipopolysaccharide) induces Hdc mRNA

Results
Generation of Hdc-GFP transgenic mice. The entire Hdc locus encompasses a more than 25-kb genomic region, and the distribution of regulatory elements has rarely been identified. Therefore, we presumed that a broader range of the Hdc gene locus would be required to monitor endogenous Hdc gene expression. Hence, we used a BAC clone, RP23-40N15, which contains the entire set of mouse Hdc gene locus along with approximately 120-kb of 5′ and 148-kb of 3′ extended flanking sequence (Fig. 1A). We introduced a GFP reporter cassette in frame with the translation initiation codon of the first exon by means of homologous recombination in E. coli strain EL250 19 . After successful BAC recombination and deletion of the neomycin resistance cassette, the modified BAC DNA construct was injected into fertilized BDF1 ova. Subsequently, we generated two lines of Hdc-GFP transgenic mice (line#1 and line#2) and confirmed that both lines were GFP-and CAT (chloramphenicol acetyltransferase cassette)-positive by PCR-genotyping ( Fig. 1B and Supplementary Fig. S1). Both lines of Hdc-GFP mice stably transmitted the transgene for more than five generations, and thereafter, we subjected the mice to analysis. Genomic quantitative (q) PCR analysis at 5 sites in the Hdc locus demonstrated that approximately 4-5 copies of the Hdc locus containing all the exons were integrated (line#1, Fig. 1C). In the 5′ flanking region, 8 copies of the 10-kb upstream region were integrated (line#1, Fig. 1C). While line#2 carries the transgenic GFP DNA, integration of either 5′-or 3′-distal flanking sequences was not detected by genomic qPCR. We surmise that line#2 harbors the proximal Hdc sequences and the GFP reporter DNA. GFP expression pattern in Hdc-GFP mice. To address whether the Hdc BAC-directed transgenic GFP reporter recapitulates the endogenous Hdc expression profile, we first examined the GFP fluorescence in the brain, stomach and peritoneal cavity cells (PECs) since these tissues and cellular fractions contain the majority of the canonical histamine-producing cells 3 . We harvested PECs by peritoneal lavage with phosphate buffered saline (PBS) and dissected the brain, stomach and other tissues from the Hdc-GFP mice. We first analyzed whole tissue GFP fluorescence using an in vivo imaging system (IVIS) and found that line#1 exhibited robust GFP expression in the brain, stomach and PEC suspension of the peritoneal lavage ( Fig. 2A). Line#2 also showed GFP fluorescence in the brain and stomach at a lower level than line#1 ( Fig. 2A). Quantitative analysis demonstrated that the GFP fluorescence in the brain and stomach of line#1 was approximately 1.68-fold and 10.11-fold higher than that of line#2, respectively (Fig. 2B). Line#1 showed strong GFP fluorescence in PECs, while line#2 rarely showed GFP fluorescence in the PECs by IVIS analysis (Fig. 2B). Other tissues, including heart, lung, thymus, liver, intestine, kidney and spleen, rarely showed GFP fluorescence either in line#1 or line#2 ( Fig. 2A,B).
Histamine is also synthesized in a certain type of neuroendocrine cells (enterochromaffin-like cells, ECL cells) in the gastric mucosa and stimulates parietal cells to secrete gastric acid 22 . In further IVIS observations, GFP fluorescence was predominantly detected in the corpus of the stomach, which contains the vast majority of the gastric ECL cells 23 , whereas GFP was rarely detected in the pyloric regions ( Fig. 2A). Therefore, we quantitatively examined the mRNA expression levels of endogenous Hdc and transgenic GFP separately in the GFP-high corpus region and in the GFP-low pyloric region ( Fig. 2A). As anticipated, significant mRNA levels of the endogenous Hdc and transgenic GFP were both exclusively detected in the corpus region, while the pyloric region expressed neither endogenous Hdc nor transgenic GFP (Fig. 2C,D). Consistently, GFP immunoreactivity was detected in the small polygonal cells that were localized in the gastric mucosa beneath the epithelium, which represents the typical histological characteristics of ECL cells (Fig. 2E). These results indicate that the GFP fluorescence level faithfully recapitulates the region-specific endogenous Hdc expression level in the stomach.
In the brain, histamine is produced in the tuberomammillary nucleus (TMN) of the hypothalamus and participates in a number of neurobehavioral processes 24,25 . Indeed, immunohistochemical analysis detected intense www.nature.com/scientificreports www.nature.com/scientificreports/ To further clarify the GFP expression pattern in the histamine-producing hematopoietic cells, we separated bone marrow cells and PECs into multiple distinct fractions ( Supplementary Fig. S2). Representative histogram data showed that line#1 expressed robust GFP fluorescence in approximately 64.6% of the bone marrow basophils (c-kit − FcεR + ) and 91.3% of the peritoneal mast cells (c-kit + FcεR + ) (Fig. 4A,B). Line#2 also showed GFP fluorescence in approximately 63.5% of the bone marrow basophils and in 69.1% of the peritoneal mast cells, albeit at a low mean fluorescence intensity (MFI) relative to that of line#1 (Fig. 4A,B). On average, line#1 and line#2 Notably, we found that approximately 50% of the CD11b + Gr-1 + immature myeloid cell (IMC) fraction of the line#1 bone marrow expressed GFP fluorescence ( Fig. 4A,C), indicating that the GFP reporter directed by the Hdc-GFP transgene monitors endogenous Hdc expression in IMCs. Line#2 showed lower levels of GFP in the IMCs, macrophages and granulocytes than line#1, indicating that the reporter in line#2 is preferentially expressed in mast cells and basophils (Fig. 4C,D). Because line#1 showed a higher GFP fluorescence intensity in a wider spectrum of the Hdc-expressing cell lineages than line#2, we mainly utilized line#1 for the subsequent analysis.
Hdc BAC-directed transgenic GFP expression identifies the LPS-induced histamine-producing cells. While the pathophysiological activity of histamine in sepsis has been described 17 , the secretory origin of histamine in the damaged tissues during sepsis remains unknown. To address this issue, we examined histamine-producing cells in a murine model of LPS-induced sepsis using Hdc-GFP mice. We previously reported that systemic inflammation peaks at 4 hours after LPS administration 21 . Thus, we monitored whole body GFP fluorescence at 4 hours after LPS administration using IVIS. Notably, we found that the most robust GFP signal was detected in the lungs of the Hdc-GFP mice (Fig. 5A). Ex vivo analysis of the tissues confirmed that GFP fluorescence was most strongly induced in the lung, while other tissues, including liver, intestine and kidney, showed a marginal increase in GFP fluorescence upon sepsis induction (Fig. 5A). Quantitative RT-PCR analysis demonstrated that the endogenous Hdc mRNA level was increased 6.1-fold in the lung at 4 hours after LPS administration (Fig. 5B). Furthermore, the stomach showed constitutively high-level fluorescence irrespective of the LPS administration (Fig. 5A).
To address what cellular population expresses GFP in the septic lung tissue, we prepared a cellular suspension from the septic lung tissues and analyzed the cell surface marker phenotype by flow cytometry. We found that GFP was predominantly expressed in the CD45-positive hematopoietic cell population but rarely expressed in the CD45-negative cells in the lung at steady state (Fig. 5C,D). After LPS administration, the GFP-high population was significantly increased (approximately 1.8-fold) in the CD45-positive cells, while the CD45-negative cells showed no GFP induction (Fig. 5C,D). These results indicate that the CD45-positive hematopoietic cells are responsible for HDC production in the lung in response to LPS treatment.
Ly6G + histamine-producing myeloid cells in the LpS-treated lung. Next, we further examined cellular lineage of the GFP-expressing hematopoietic cells in the lung upon LPS administration. Flow cytometry analysis showed that the GFP-high population was significantly increased (approximately 2.7-fold) in the IMCs (Gr1 + CD11b + ) after LPS administration (Fig. 6A). In contrast, the GFP-high proportion of other hematopoietic cells, including mast cells and basophils, rarely changed upon LPS administration (Fig. 6A). There have been a series of reports showing that myeloid lineage immune cells, other than mast cells and basophils, produce histamine after LPS administration 13,28 . To further clarify the identity of the Hdc-GFP-expressing cells, we precisely examined the GFP expression profile in each myeloid cell lineage using multiple surface markers, i.e., Ly6G (neutrophils), Ly6C (monocytes), F4/80 (macrophages) and CD11b (a myeloid lineage marker). We eventually found that Ly6G + Ly6C low CD11b + F4/80 − cells, a subset of neutrophils (referred to as Ly6G + cells hereafter), predominantly expressed GFP in the lung upon LPS administration (Fig. 6B). Neutrophils comprise the largest pool of circulating white blood cells and are recruited rapidly to sites of inflammation 29 . Therefore, we next addressed whether the lung-accumulated GFP-positive cells are tissue-localized (noncirculating) cells or circulating bloodborne leukocytes. To this end, we conducted intravascular (i.v.) staining of hematopoietic cells by i.v.-injecting an anti-CD45 antibody right before the flow cytometry analysis and thereby discriminated between circulating  30 . We found that almost all of the Ly6C + and Ly6G + fractions in the lung were derived from bloodborne CD45 i.v. + circulating leukocytes (Fig. 6C). We next examined the GFP fluorescence levels separately in the Ly6C + and Ly6G + cells in the peripheral blood. The GFP histogram showed that the nearly half of the Ly6G + fraction expressed a high level of GFP fluorescence (49.8% of Ly6G + cells), while the Ly6C + fraction expressed a low level of GFP fluorescence (53.1% of Ly6C + cells) (Fig. 6D). These data indicate that the majority of the GFP-high cells are bloodborne Ly6G + cells and that the Ly6G + cells most likely serve as a major source of histamine in the LPS-induced lung. Importantly, the Ly6G + GFP-high population was significantly expanded (approximately 4.5-fold) upon LPS administration in peripheral blood (Fig. 6E). The GFP median fluorescent intensity (MFI) was further increased (1.1-fold) in the Ly6G + cells upon LPS administration (Fig. 6F). To further clarify whether the GFP-positive Ly6G + cells produce histamine, we Hdc-GFP mice. Hematopoietic cells were separated into the following seven distinct fractions: FcεR − c-kit + ; progenitor cells, FcεR + c-kit + ; mast cells, FcεR + c-kit − ; basophils, Gr1 + CD11b − ; granulocytes, Gr1 + CD11b + ; immature myeloid cells (IMCs), Gr1 − CD11b + ; macrophages, c-kit − Gr1 − CD11b − ; others (hematopoietic cells other than those above). See Supplementary Fig. S2 for the separation strategy. Data are presented as the means ± SEM in the bar graphs.
sorted out the Hdc-GFP-positive or negative cells within the Ly6G + fraction and quantified histamine production by mass spectrometer. We found that the Hdc-GFP + Ly6G + fraction contained higher level of histamine than the Hdc-GFP − Ly6G + fraction did (0.13 vs 0.05 pmol/10 5 cell, respectively) (Fig. 7). Ly6G + neutrophils from the wild type mice (WT Ly6G + ) also showed histamine production (0.096 pmol/10 5 cell), which amounted below the levels of the Hdc-GFP + Ly6G + cells. Whole peritoneal cells from the vehicle-treated wild type mice (WT PEC) contained substantial level of histamine (1.70 pmol/10 5 cell) and served as a positive control for this analysis (Fig. 7). These results indicate that the transgenic Hdc-GFP successfully enriches the high histamine-producing cells in the Ly6G + neutrophils.

Discussion
In this study, we demonstrated that Hdc-BAC-directed reporter GFP expression recapitulated endogenous Hdc expression in mast cells, basophils, gastric enterochromaffin-like (ECL) cells and hypothalamic tuberomammillary nucleus neurons (TMNs), all of which are known histamine-producing cells 3 . We also demonstrated that the GFP expression level faithfully reproduced the induced Hdc expression in the lung and peripheral blood during lipopolysaccharide (LPS)-induced sepsis. Furthermore, we confirmed that the Hdc-GFP + cells indeed contained a high level of histamine, while the Hdc-GFP − cells did not. Thus, these series of results clearly indicate that the Hdc BAC-directed GFP fluorescence faithfully labels the endogenous histamine-producing cells.
Despite the well-recognized pathophysiological activity of histamine, the primary secretary origin of histamine during sepsis remains unknown. In the current study, we demonstrated that Hdc-GFP + Ly6G + , a www.nature.com/scientificreports www.nature.com/scientificreports/ www.nature.com/scientificreports www.nature.com/scientificreports/ subpopulation of neutrophils, predominantly expanded in the lung and peripheral blood upon sepsis induction. Interestingly, it has been reported that LPS induces histamine production in human neutrophils rather than basophils, monocytes and lymphocytes 31 . Collectively, these findings support our notion that the Hdc-GFP + Ly6G + neutrophils constitute a major histamine-producing population during sepsis.
The LPS-induced HDC catalyzes de novo synthesis of histamine, which has been shown to aggravate inflammation 15,16,28 . However, a number of other reports suggest that the induced histamine plays various immunomodulatory roles 32 . For instance, histamine polarizes the Th2-immune response, which enhances IL-10 production and inhibits inflammation via H 2 receptors 33,34 . Another report also showed immune-suppressive roles of histamine, in which the induced histamine negatively regulates acute inflammation in a bacterial peritonitis model and delays the clearance of bacteria 35 . It would be of particular value to assess how the de novo synthesized histamine in the Hdc-GFP + Ly6G + neutrophils modulates the local inflammation in the septic lung, which would significantly extend our understanding of the immunomodulatory roles played by histamine.
There have been only a limited number of reports that describe the regulatory mechanism of the Hdc gene in immunocompetent cells. Previously, we generated plasmid-based GFP reporter transgenic mice harboring the approximately 1-kb promoter region of the Hdc gene. Thereafter, we observed that the GFP reporter expression was weak and not specific to the histamine-producing cells, indicating that the 1-kb proximal promoter region is not sufficient to recapitulate the endogenous Hdc expression in vivo 36 . Line #2 of the current Hdc-GFP mice, which lack the distal flanking sequences of the Hdc locus, still showed recapitulation of the Hdc expression in the mast cells and basophils. This result may indicate that line#2 of the Hdc-GFP mice carries sufficient regulatory sequences for endogenous Hdc expression in mast cells and basophils. However, line#2 does not express the reporter in other myeloid lineage cells. These data imply that putative mast cell/basophil-specific regulatory elements are located beyond the 1-kb promoter but relatively close to the Hdc gene. Other myeloid lineage-specific enhancers could be localized in the far distal flanking sequences contained in the Hdc BAC clone. A recent report showed that the basic helix-loop-helix leucine zipper transcription factor MITF (microphthalmia-associated transcription factor) binds to an enhancer in the −8.8 kb upstream region of the mouse Hdc gene and upregulates Hdc gene expression in mast cells 37 , which supports our hypothesis. A more detailed study into the regulatory mechanism of Hdc gene expression will provide additional insights into allergic and inflammatory reactions and lead to possible therapeutic avenues for histamine-related diseases.

BAC modification and generation of transgenic mice.
A 293-kb BAC clone, RP23-40N15, which harbors all mouse Hdc exons with the 120-kb 5′ and 148-kb 3′ flanking sequences, was subjected to homologous recombination in E. coli as previously described 38 . The targeting vector was constructed with the 5′ homologous region of the 1.1-kb 5′-promoter region of Hdc, while the 3′ homologous region included 1.1-kb of sequence in the 1 st intron (Fig. 1A). After homologous recombination and deletion of the neomycin cassette, the resulting clone was referred to as Hdc-GFP and subjected to transgenic microinjection in fertilized eggs. All mice were handled according to the intravascular staining of hematopoietic cells. Intravascular staining to discriminate between bloodborne and tissue-localized hematopoietic cells was performed as previously described 30 . Briefly, 0.5 μg of anti-CD45 antibody was administered by intravascular injection, and then 3 min later, flow cytometry analysis was performed.
Quantitative real-time pcR. Genomic DNA was purified by phenol: chloroform: isoamyl alcohol (Nacalai Tesque, Kyoto, Japan). Total RNA was purified by Micro Smash MS-100 (TOMY SEIKO, Tokyo, Japan) and Sepasol ® -RNA I Super G (Nacalai Tesque, Kyoto, Japan). cDNA was synthesized by ReverTra Ace ® (TOYOBO, Osaka, Japan). Quantitative real-time PCR was performed with THUNDERBIRD ® SYBR ® qPCR Mix (TOYOBO, Osaka, Japan) on a CFX96 Touch TM Detection System (Bio-Rad Laboratories, Hercules, CA). The genomic DNA was normalized to the value of the Gata2 −2.8 kb locus. The mRNA expression level was normalized to the Gapdh expression level. The primers used in the quantitative real-time PCR are listed in Supplementary Table 1.

Imaging of GFP fluorescence in vivo and ex vivo. In vivo imaging was conducted utilizing an in vivo
imaging system (IVIS) (PerkinElmer, Waltham, MA) as previously described 21 . Briefly, Hdc-GFP transgenic mice were anesthetized with isoflurane. Subsequently, the mice or separately dissected tissues were placed in a light-sealed chamber, and the GFP fluorescence was imaged for 1 to 10 s. Fluorescence emitted from various regions of the mouse was quantified with Living Image software (PerkinElmer, Waltham, MA).

LpS-induced sepsis. Sepsis was induced by intraperitoneal (i.p.) injection of lipopolysaccharide (LPS;
Sigma-Aldrich, St. Louis, Missouri) at a dose of 1 mg/kg or 10 mg/kg body weight. Four hours after the LPS injection, all analyses were performed.
Quantification of the intracellular histamine by mass spectrometer. For quantification of the intracellular histamine, each cell fraction from peritoneal lavage and peripheral blood was fixed with Cellcover (AL Anacyte Laboratories UG, Hamburg, Germany) and sorted by FACS Aria Fusion. Histamine level was quantified with a modified protocol from the previously reported methods using TSQ Vantage AM mass spectrometer (Thermo Fisher Scientific, Waltham, MA) 41,42 . Reversed-phase Scherzo SM-C18 (3 mm × 100 mm, 3 μm, Imtakt Corp. Japan) was used for the analytical column.

Statistical analysis.
Comparisons between 2 groups were made using the Student's t-test. The data are presented as the means ± SEM. For all analyses, statistical significance was defined as a value of p < 0.05. Data management and statistical analysis was performed using Excel (Microsoft, Redmond, WA) and GraphPad Prism8 software (GraphPad Software, San Diego, CA).

Data availability
The data generated or analyzed during this study are included in this published article and its Supplementary Information File. Published: xx xx xxxx