Conditional Deletions of Hdc Confirm Roles of Histamine in Anaphylaxis and Circadian Activity but Not in Autoimmune Encephalomyelitis

Histamine is a chemical mediator derived from the decarboxylation of l-histidine by histidine decarboxylase (HDC) (1). It regulates a variety of physiological processes such as immune response, vascular permeability and dilation, bronchoconstriction, gastric acid secretion, and even neurotransmission (2). It acts via four transmembrane receptors called H1R to H4R (3). A broad range of antihistamines, targeting these receptors individually, has been developed and has become widespread medication. H1R antagonists are the most common and versatile, used to relieve allergies and nausea as well as to induce sleep. Since the early 1980s, there has been increasing evidence that histamine plays a role and could be targeted in autoimmune diseases, including those of the CNS such as multiple sclerosis (MS) and neuromyelitis optica spectrum disorder (4–12). However, because of contradicting results, the source and significance of histamine in these demyelinating diseases remain unclear.

Clinical studies have shown that MS patients have increased histamine levels in cerebrospinal fluid and increased H1R expression in demyelinated plaques (13). Moreover, the use of H1R antihistamines has been associated with a decrease in MS risk (9). Consistently, making use of the MS model experimental autoimmune encephalomyelitis (EAE), functional studies have shown that mice lacking H1R (14–16), H2R (17), or all four histamine receptors (18), as well as mice treated with an H1R antagonist (19, 20), develop milder EAE. In contrast, mice lacking HDC (18, 21), H3R (22), or H4R (23) develop more severe EAE. These opposite observations could be explained by the multiple sources and functions of histamine; for example, leukocyte-derived histamine could promote autoimmunity by acting on different targets (e.g., myeloid cells, T cells, and endothelium), whereas the histaminergic neuronal pathway could provide a negative regulatory feedback by stimulating the hypothalamic–pituitary–adrenal (HPA) axis, resulting in the secretion of immunosuppressive glucocorticoids. In support of this argument, multiple studies have reported histamine as a modulator of the HPA axis (24–27), and HDC-deficient mice have been shown to suffer from an endocrine imbalance, including reduced levels of adrenocorticotropic hormone (28).

The most well-known sources of histamine in the immune system are mast cells and basophils (29). Surprisingly, mast cell–derived histamine is not required for EAE (30), suggesting the existence of another source, yet to be discovered. Interestingly, in our microarray datasets, we found that Hdc mRNA is abundant in neutrophils infiltrating the spinal cord of EAE mice (31). In addition, others have reported that neutrophils can store histamine and upregulate its synthesis via HDC when stimulated (32–35). The mechanisms of action of neutrophils in EAE are still unclear, but evidence suggests that they contribute to the breakdown of the blood–brain barrier (36). Whether histamine mediates this effect is an open question.

The main objectives of this study were as follows: 1) to validate neutrophils as a source of histamine in EAE and 2) to dissect the roles of histamine from different sources by generating Hdc conditional knockout (cKO) mice. Surprisingly, we found no effect of Hdc deletion on EAE, regardless of whether it was ubiquitous or limited to neutrophils or other cell populations. However, we did observe classic effects on IgE-mediated anaphylaxis as well as on circadian activity, establishing the legitimacy of our cKOs for re-exploring the multiple facets of histamine.

As previously described to null the Hdc allele (37), we chose to delete exon 8, which codes for an essential part of the catalytic domain (Fig. 2A). As no target guide sequences were found suitable in intron 8 for genome editing, we inserted the second loxP sequence in intron 9, thus deleting exon 9 as well as exon 8.

Single-guide RNA (sgRNA) sequences that passed the off-target threshold were chosen with the help of Massachusetts Institute of Technology Zhang laboratory Optimized CRISPR Design Tools (crispr.mit.edu). LoxP sequences were inserted in a nonconserved section of the genome at least 200 bp from exon–intron boundaries to minimize interference with splicing. Selection of targeted nonconserved sequences was accomplished by analysis of the mouse genome version GRCm38/mm10 using University of California Santa Cruz Genome Browser tools: 1) Placental Mammal Basewise Conservation by PhyloP and 2) Multiz Alignment Genome.

We used single-stranded donor oligonucleotides (ssODN) 180 base–long (Ultramers, Integrated DNA Technologies) as directed-homology recombination templates for insertion of loxP sequences flanked by 70-base homology arms. To facilitate allele identification, Nhe1 and EcoR1 restriction sites were inserted upstream and downstream of loxP sequences into intron 7 and 9, respectively. The loxP sites in the ssODN integrated at the Cas9 cleavage site to prevent recognition of the target sequence by sgRNA and recleavage by Cas9.

sgRNAs were generated by PCR using sgRNA oligonucleotide templates (Supplemental Table I) with px330-U6-Chimeric_BB-CBh-hSpCas9 plasmid (Addgene) as DNA template coding for transactivating CRISPR RNA. sgRNA oligonucleotide templates contained the following elements (5′ to 3′): T7 RNA polymerase promoter, target sequence, and sgRNA transactivating CRISPR RNA scaffold. The PCR amplicons were gel-purified with the MinElute Gel Extraction Kit (QIAGEN) and used as templates to further amplify sgRNAs using the MAXIscript T7 Transcription Kit (Thermo Fisher Scientific). The resulting RNAs were purified by phenol/CHCl3 extraction and isopropanol precipitation. The RNA pellets were dissolved in RNase-free water and quantified.

Microinjection of zygotes from superovulated females mated with males C57BL6/N were performed at the Goodman Cancer Research Centre Transgenic Facility, as previously described (38). Briefly, the microinjected mixture was comprised of sgRNAs (50 ng/μl), Cas9 mRNA (50 ng/μl; Sigma), and ssODN (1 μM). Zygotes were microinjected with CRISPR–Cas9 components with an average injection volume of 1–2 pl per embryo. The injected zygotes were cultured overnight in droplets of potassium simplex optimized medium under mineral oil in a 35-mm dish at 37°C in a 5% CO₂ incubator until the two-cell stage and then transferred into the oviducts of 0.5-d pseudopregnant CD-1 females.

C57BL/6, Tek-cre (39), Mrp8-cre (40), Nestin-cre (41), CMV-cre (42), and Ai14 (43) mice were obtained from The Jackson Laboratory and crossed with Hdc^fl/fl mice (see above; registered as Hdc^em1Luval in the Mouse Genome Informatics database). 2D2 mice (44) were also obtained from The Jackson Laboratory. Genotypes were confirmed by PCR using the primers listed in Supplemental Table I. All mice were kept on a C57BL/6 background, housed under specific pathogen-free conditions, and fed Envigo diet 2018. All protocols were approved by the Laval University Animal Protection Committee and followed the guidelines of the Canadian Council on Animal Care.

Mice (8–10 wk old, mixed sexes) were immunized via two 100-µl s.c. injections at the flank level of MOG emulsion. Emulsion was composed of either 300 µg of MOG_35–55 peptide or 500 µg of bMOG protein (31) dissolved in saline and mixed with an equal volume of complete Freund’s adjuvant (MilliporeSigma) supplemented with 500 µg of killed Mycobacterium tuberculosis H37 RA (Difco Laboratories). Mice were injected i.p. with 20 µg/kg of pertussis toxin (PTX) (List Biological Laboratories) immediately after and 2 d postimmunization.

Mice (8–10 wk old, mixed sexes) were i.p. injected with 20 × 10⁶ encephalitogenic cells. These were isolated from abdominal lymph nodes and spleens of mice killed 8 d after active EAE induction and then cultured for 2 d in DMEM with MOG_35–55 (15 μg/ml), murine IL-12 (5 ng/ml, R&D Systems), murine IL-23 (20 ng/ml, R&D Systems), heat-inactivated HyClone Bovine Growth Serum (10%, Thermo Fisher Scientific), modified Eagle medium nonessential amino acids (1%, Wisent Bio Products), penicillin (100 U/ml), streptomycin (100 μg/ml), and amphotericin B (250 ng/ml).

2D2 mice (8–10 wk old, mixed sexes) received two i.p. injections of PTX (20 μg/kg) at a 2-d interval.

Mice were weighed and scored daily for typical symptoms as follows: 0, no visual sign of disease; 0.5, partial tail paralysis; 1, complete tail paralysis; 1.5, weakness in one hind limb; 2, weakness in both hind limbs; 2.5, partial hind limb paralysis; 3, complete hind limb paralysis; 3.5, partial forelimb paralysis; 4, complete forelimb paralysis; and 5, dead or sacrificed for humane reasons. Hdc^Cmv cKOs and their controls were also scored for atypical symptoms as follows: 0, no visual sign of disease; 1, hunched or staggered walking; 2, ataxia or slight head tilt; 3, severe head tilt, slight axial rotation, or cannot stay upright; 4, severe axial rotation or spinning; and 5, death or sacrificed for humane reasons.

To assess visual acuity of mice before and after EAE induction, the optokinetic response was measured with the OptoDrum system (Striatech, Tübingen, Germany). As previously described (45), mice were placed individually in the middle of a platform surrounded by four computer screens on which moving gratings were displayed at increasing spatial frequencies. The reflex movement of the head in the direction of the grating motion was monitored to determine the spatial frequency threshold (in cycles/degree) of the optokinetic reflex. The function of each eye was separately evaluated by changing the direction of the visual stimulus.

Mice (16–20 wk old, male) were injected i.p. with 10 µg of monoclonal anti-DNP IgE (Sigma-Aldrich) dissolved in sterile saline. Twenty-four hours later, immune complex (IC) challenge was initiated by i.v. injection of 500 µg of human serum albumin–conjugated DNP (Sigma-Aldrich) under light anesthesia. Body temperature was monitored every 10 min during IC challenge for 1 h. Temperature was obtained by rectal probe using a Physitemp TH-5 Thermalert clinical monitoring thermometer equipped with a RET-3 ISO rectal probe.

Mice (18–19 wk of age in average, male) were individually housed in Digital Ventilated Cages (Tecniplast) for 21 d (7 d of acclimation and 14 d of monitoring). Cages were distributed as to avoid location biases between groups. Locomotion data were collected and analyzed using DVC Analytics (Tecniplast).

Mice were s.c. injected with 1 ml of 3% thioglycolate medium (BD Difco). Four hours later, peritoneal cells were harvested with two washes of 2.5 ml of HBSS (Wisent Bio Products). Cells were spun, resuspended in HBSS containing 0.1% BSA, and then purified by negative selection with the EasySep Mouse Neutrophil Enrichment Kit (STEMCELL Technologies). Neutrophils (5 × 10⁵) were cultured in a 96-well microplate for 24 h in RPMI 1640 (Wisent Bio Products) containing 10% HyClone Bovine Growth Serum, 1× Antibiotic–Antimycotic Solution (Wisent Bio Products), and 55 μM β-mercaptoethanol. Supernatants were collected, centrifuged to remove cells, and kept at −80°C until histamine quantification.

To isolate tissue containing histaminergic neurons, brains were dissected by removing the cerebellum, olfactory bulb, and cerebral cortex. The remaining tissue (mainly the hypothalamus and thalamus) were powdered on dry ice with a BioPulverizer (BioSpec Products) and used for quantitative RT-PCR (RT-qPCR) and ELISA.

Total RNA was extracted either from brains or neutrophils with EZ-10 Spin Column Animal Total RNA Miniprep Kit (Bio Basic) or from spinal cords with TRI Reagent (Sigma-Aldrich). First-strand cDNA was generated from 1–5 µg of RNA using Superscript III (Invitrogen) with random hexamers and 20-mer oligo-dT primers and then purified using the GenElute PCR Clean-Up Kit (Sigma-Aldrich). The product (20 ng) was analyzed using the LightCycler 480 system with the SYBR Green I Master mix and primers (listed in Supplemental Table II) and according to the manufacturer’s instructions (Roche Diagnostics). PCR conditions consisted of 45 cycles of denaturation (10 s at 95°C), annealing (10 s at 60°C), elongation (14 s at 72°C), and reading (5 s at 74°C). The number of mRNA copies was determined using the second derivative method and a standard curve of Cp versus logarithm of the quantity (46). The standard curve was established using known amounts of purified PCR products (10, 10², 10³, 10⁴, 10⁵, and 10⁶ copies) and a LightCycler 480 v1.5 program provided by the manufacturer (Roche Diagnostics). In Fig. 1, quantitative PCR results were presented in absolute quantities after confirmation that there was no difference between Hprt1-normalized and nonnormalized data.

Total genomic DNA from purified neutrophils was quantified by CyQuant (Invitrogen). PCR was performed using forward Hdc intron 7 and reverse Hdc intron 9 primers (Supplemental Table I) with Phusion polymerase (Thermo Fisher Scientific) in a reaction containing 1× GC buffer, 3% DMSO, 0.5 μM of each primer, and 0.2 μM dNTPs. A touchdown PCR was performed with the following conditions: 1) 98°C for 30 s, 2) 98°C for 10 s, 3) 70°C for 15 s, 4) 72°C for 45 s, 5) step 2–4 for 10 cycles with a decrease of 0.5°C for step 3 at every cycle, 6) 98°C for 10 s; 7) 64°C for 15 s, 8) 72°C for 45 s, 9) step 6–8 for 30 cycles, and 10) 72°C for 10 min. The PCR product was then loaded onto a 1% agarose gel dissolved in Tris–borate–EDTA buffer and supplemented with SYBR Safe DNA Gel Stain (Invitrogen). Expected band sizes were as follows: wild-type allele, 2023 bp; floxed allele, 2103 bp; and excised allele, 821 bp.

Histamine was quantified in neutrophil supernatants or brain samples using an ELISA kit (Beckman Coulter), as per the manufacturer’s instructions. Neutrophil supernatants were collected from purified neutrophil cultures obtained as described above. Histamine was extracted from brain powder (20–40 mg) with 10 µl of 0.2N HClO4 per mg of tissue ground with a microsample homogenizer (PRO Scientific). The homogenates were centrifuged for 5 min at 10,000 g and ∼4°C. The supernatants were neutralized with 1 μl of 1.5 M NaOH per mg of tissue and centrifuged for 1 min at 10,000 g and ∼4°C. Histamine quantity was normalized to either the mass of brain tissue or the quantity of DNA from neutrophils at the time of supernatant collection, as measured by CyQuant assay.

Corticosterone was quantified in urine samples using an ELISA kit (Arbor Assays). Samples were collected on days 0, 4, 8, and 12 postimmunization with MOG_35–55 and diluted 50-fold in diluted assay buffer before carrying out the ELISA as per the manufacturer’s instructions.

Mice were anesthetized and exsanguinated by transcardial perfusion with saline. Spinal cords were harvested and minced with razor blades in Dulbecco’s PBS containing calcium and magnesium. The cell suspensions were then digested for 45 min at 37°C with 0.13 U/ml Liberase (Roche Diagnostics) and 50 U/ml DNase (MilliporeSigma) in Dulbecco’s PBS, filtered through 70-µm cell strainers, and separated from myelin debris by centrifugating at 1,000 × g in 35% Percoll (GE Healthcare). For immunostaining, cells were incubated on ice for 5 min with rat anti-CD16/CD32 Ab (BD Biosciences, clone 2.4G2, 5 µg/ml) and Fixable Viability Dye eFluor 455UV (eBioscience, 1:1000) followed by a 30-min incubation with primary Ab mixture (Supplemental Table II). Cells were washed and resuspended in PBS before being analyzed/sorted with a FACSAria II flow cytometer (BD Biosciences). Before analysis, the following quality control checks were performed using FlowJo (Tree Star): 1) debris were removed using forward scatter area and side scatter area; 2) doublets were removed using forward scatter area and forward scatter height; and 3) dead cells positive for the viability dye were removed. Gates were based on fluorescence minus one controls. To control for variation in cell yield, CNS cell counts were normalized to CD45⁻ cells, whose number is proportional to the sample size.

Single-cell suspensions from pooled spinal cords and brains were stained with cell surface marker Abs as described above and then stained for Hdc mRNA with a type-4 Hdc probe using the PrimeFlow Kit (Invitrogen), according to the manufacturer’s 96-well plate protocol.

Data are expressed as mean ± SEM. In general, means were compared with nonparametric (Wilcoxon or Kruskal–Wallis) or parametric tests (one-way ANOVA with two-tailed Student t test or Dunnett multiple comparison test; two-way ANOVA with two-tailed Student t test) when data were continuous, normally distributed (Anderson–Darling test), and of equal variance (Levene test). EAE prevalence curves were constructed using the Kaplan–Meier method and compared by Wilcoxon test. EAE severity curves were compared by two-way ANOVA with repeated measures using rank-transformed scores. Correlation between variables was determined using the Pearson test. The sexes were grouped in the figures as the conclusions were the same regardless of whether the statistical analyses were done on mixed or single-sex groups. All these analyses were performed with either JMP 14 (SAS Institute) or GraphPad Prism 8.3 (GraphPad Software). A p value < 0.05 was considered significant.

We have previously observed that neutrophils express high levels of Hdc mRNA (31). To characterize the expression of Hdc in EAE, we performed RT-qPCR on spinal cords from mice with different forms of EAE (i.e., induced either by active immunization, passive transfer of encephalitogenic T cells, or injection of PTX to 2D2 mice expressing a MOG-specific TCR). Hdc was increased in all of these forms (Fig. 1A) and correlated with a time-dependent increase of Ly6g, an indicator of neutrophil infiltration (Fig. 1B). Further, RT-qPCR performed on FACS-purified cells revealed that Hdc was present in neutrophils but not in other infiltrating leukocytes (Fig. 1C). We did not observe a difference in Hdc expression between neutrophil subpopulations expressing or not ICAM1 (Fig. 1C), a marker that distinguishes extravasated from intravascular neutrophils (31). This suggests that Hdc is constitutively expressed in neutrophils. To further analyze the cell specificity of Hdc expression, Hdc mRNA was detected by flow cytometry using PrimeFlow. Validating RT-qPCR results, this assay showed that up to 70% of Hdc⁺ cells from EAE spinal cords were neutrophils (Fig. 1D). Thus, we conclude that neutrophils are the major source of HDC (and therefore of histamine) in the spinal cord of EAE mice.

In search of a mouse model to assess the role of histamine specifically in neutrophils, we generated a cKO of the Hdc gene using a CRISPR–Cas9 strategy with the aim to delete exons 8 and 9 (Fig. 2A). Exon 8 was the main target because it codes for the binding site of pyridoxal 5′-phosphate, an essential coenzyme for the decarboxylation of histidine to histamine (47, 48). After obtaining the desired mouse, we proceeded to validate the functionality of the floxed allele by crossing with three well-established Cre-expressing mouse strains: the Mrp8-cre, Nestin-cre, and Tek-cre strains, which allow gene deletion in neutrophils, neurons, or hematopoietic plus endothelial cells, respectively. Specific cre-mediated excision of the Hdc allele was confirmed by PCR on purified neutrophils (Fig. 2B) and brain samples (data not shown) followed by sequencing of the Hdc amplicons (data not shown). RT-qPCR analysis confirmed that conditional Hdc knockdown with Mrp8-cre (Hdc^Mrp8 cKOs) abolished Hdc expression in neutrophils without affecting expression in brain tissue (Fig. 2C). Tek-driven knockdown (Hdc^Tek cKOs) abolished Hdc expression in neutrophils as expected but also partially decreased expression in brain samples (Fig. 2C). This can be explained by the concomitant expression of Hdc and Tek in cerebral endothelial cells and pericytes, as observed in a public single-cell RNA sequencing dataset (Supplemental Fig. 1). Surprisingly, nestin-driven knockdown (Hdc^Nes cKOs) abolished Hdc expression not only in the brain but also partially in neutrophils (Fig. 2C). Although nestin expression has not been reported in neutrophils, it is known to occur in endothelial precursors (49, 50). Given that hematopoietic stem cells hail from the endothelial lineage (51), at least some hematopoietic cells could have inherited the Hdc deletion. Supporting this hypothesis, cell lineage tracing analysis of the Nestin-cre line bred with a tdTomato Ai14 reporter mouse showed that half of endothelial cells from spinal cords and brains were tdTomato⁺ (Supplemental Fig. 2B). Hematopoietic cells were found to be 5–40% tdTomato⁺, depending on the subsets (Supplemental Fig. 2B). Nevertheless, the majority of tdTomato⁺ cells in naive mice were found in the neural cell fraction (i.e., CD31⁻CD45⁻) (Supplemental Fig. 2C), thus confirming that Hdc^Nes cKO is a suitable model to assess the function of HDC in neurons. Finally, we validated our RT-qPCR results with the help of histamine ELISA assays (Fig. 2D). We therefore conclude that the generated conditional Hdc floxed allele is functional and can be used for cell-specific excision.

Having demonstrated that Hdc is strongly expressed in neutrophils (Fig. 1), key players in the development of active EAE (36), we hypothesized that the loss of Hdc in neutrophils of Hdc^Mrp8 cKOs and in all leukocytes of Hdc^Tek cKOs would attenuate the clinical course of active EAE. In contrast, we considered that the loss of Hdc in neurons of Hdc^Nes cKOs could potentially worsen EAE, given that histaminergic neurons have been shown to regulate the immunosuppressive HPA axis (24–27). As shown in Fig. 3A and 3B, no significant difference in the prevalence and severity of EAE was noted between all three cKOs and their wild-type controls (data of both sexes were combined as no sex-dependent difference was observed; see Supplemental Dataset 1). There was also no significant difference in the cumulative clinical score, the day of onset, and the day of peak (Kruskal–Wallis test, p = 0.076, 0.093, or 0.860, respectively). To corroborate these findings, we analyzed populations of immune cells infiltrating the spinal cord 16 d post–EAEonset by flow cytometry. Consistently, no intergenotype difference was observed (Fig. 3C). Furthermore, corticosterone levels varied regardless of genotype (Fig. 4), suggesting that histaminergic neurons are not potent regulators of the HPA axis and supporting the lack of effect on EAE development. We excluded the possibility that histamine could modulate EAE when derived from an alternative cellular source by repeating this experiment with mice (Hdc^Cmv cKOs) having a ubiquitous (heritable) deletion of Hdc (Fig. 2E, 2F); these mice developed EAE normally as evidenced by their typical and atypical symptoms (Fig. 3D–G), degree of CNS leukocyte infiltration (Fig. 3H), and loss of visual acuity (Fig. 3I). We also excluded the possibility that histamine plays a role in a mechanistically different model of EAE by immunizing Hdc cKO mice with the B cell–dependent Ag bMOG (instead of the classic B cell–independent MOG peptide); no difference was observed in cell-specific knockouts (Fig. 3J, 3K) as well as total HDC^Cmv knockouts (prevalence, log-rank test, p = 0.17; clinical score, two-way ANOVA with repeated measures, p = 0.47). Altogether, these results indicate that histamine does not significantly impact the pathophysiological mechanism of EAE.

To explore the importance of histamine expression in neutrophils while confirming anew the functionality of our models, we submitted our cKO mice to IgE-mediated anaphylactic shock, a histamine-dependent type-I hypersensitivity reaction (52). As expected, the total knockout Hdc^Cmv cKOs showed complete resistance to systemic anaphylaxis, with body temperature variations statistically equal to those of saline controls (Fig. 5). Hdc^Tek and Hdc^Mrp8 cKOs recorded a similar decrease in the magnitude of their systemic shock when compared with controls (Fig. 5). Not surprisingly, there was no observed resistance in Hdc^Nes cKOs (Fig. 5). These results indicate that IgE-mediated anaphylaxis involves histamine from different sources: 1) neutrophils as Hdc^Mrp8 cKOs were partially resistant, and 2) a bone marrow–independent cell type, most likely mast cells as Hdc^Cmv cKOs were fully resistant, whereas Hdc^Tek cKOs were partially resistant. It was beyond our objective to further study mast cells, which play a well-accepted role in IgE-mediated anaphylaxis (53, 54) and which may not be targeted effectively by the Tek-cre strategy as they have a dual embryonic origin and their replenishment in adulthood is largely independent from the bone marrow (55).

The influence of histaminergic neurotransmission on wakefulness and activity is well established (29, 56–58). To further validate the usefulness of our cKO mice, we evaluated their spontaneous locomotor activity for two weeks using digital cages. As anticipated, total and neuron-specific deletion of Hdc reduced nighttime (active phase) locomotion levels (Fig. 6). Hdc deletion in neutrophils only or in all leukocytes and endothelial cells had no consequence on activity levels (Fig. 6). As it is fully expected that histamine’s effect on wakefulness would be achieved by its expression in the brain, these observations confirm the suitability of our cKO model for the study of histaminergic neurons.

Despite several inspiring studies that have labeled histamine as a potential therapeutic target for autoimmune diseases, histamine’s role in MS is still unclear. This is due, in part, to conflicting results in EAE and to the lack of models accounting for the array of histamine sources (until now, unknown in EAE) and how each of these can produce different, even confounding effects. After re-exploring these issues using a new transgenic model, we come to the following conclusion: although neutrophils become the main producers of histamine in the nervous system during EAE, histamine does not affect the course of the disease, regardless of its source. In trying to validate our model, we discovered, as a bonus, that neutrophil-derived histamine plays a significant role in IgE-mediated anaphylaxis.

Contrary to previous studies (14–17, 19–23), we demonstrate that histamine, regardless of its source, does not play a role in EAE. Variables such as environment and genetic background affect immune reactions involved in EAE (59). Despite precautions that may be taken to ensure homogeneity between experimental groups, these variables are often overlooked or compromised in exchange for easy access to control animals. Mindful of this, we generated our controls within our transgenic colonies to minimize genetic divergence and environmental influences among the groups. In addition, we supported our conclusions by repeating EAE experiments many times, by testing two different Ags (MOG peptide and bMOG protein) that induced moderate symptoms (comparable to those observed in other studies), by combining multiple behavioral and cytometric assays, and by using a rigorous statistical approach (in which the scores were normalized according to the day post–EAE onset), allowing us to analyze disease severity separately from the prevalence. Thus, we can be confident of our data’s credibility.

We are the first, to our knowledge, to report that neutrophils contribute to IgE-mediated anaphylaxis by producing histamine. However, our study likely underestimates this contribution because neutrophil-derived histamine was not totally abolished in Hdc^Mrp8 mice, despite full abolishment of Hdc mRNA expression, a limitation also reported for another neutrophil-specific cre model (31, 60), which is ostensibly because of an inherent feature of the neutrophil not the animal model (e.g., HDC may be synthesized and stored early in neutrophil development before Mrp8-driven excision). Furthermore, although neutrophils are known to possess all the attributes to respond to Ab–Ag complexes, the mechanism by which they release histamine remains largely unexplored. This response may be involved in cerebral malaria (61) and other IgE-mediated diseases (62). Because neutrophils can also secrete histamine when exposed to other stimuli (e.g., microbial agents) (32–35), we cannot exclude a role for neutrophil-derived histamine in numerous conditions involving neutrophils [e.g., neuromyelitis optica spectrum disorder, an IgG-mediated demyelinating disease that could be alleviated with antihistamines (63)]. In light of all this, our Hdc^Mrp8 cKO mouse will be a valuable tool to assess the importance of neutrophil-derived histamine in a variety of neurologic and nonneurologic disorders.

Consistent with the literature (29, 56–58), our observation that neuron-specific deletion of Hdc decreases locomotor activity validates our mouse model for studies on the histaminergic neuronal system. This system is implicated in diverse neurologic functions (e.g., sleep, cognition, and feeding) and pathologies (e.g., narcolepsy, Alzheimer and Parkinson diseases, schizophrenia, and addiction) (29). One function that deserves further investigation is the HPA axis, which was suggested to be regulated by the histaminergic system (24–27). However, the current study does not support this hypothesis as mice with a neuron-specific deletion of Hdc had normal levels of corticosterone over the course of EAE. Of course, we cannot exclude that the histaminergic system regulates the HPA axis in circumstances other than EAE.

This study offers a new transgenic mouse line allowing the conditional deletion of the Hdc gene. This model leads us not only to refute the concept that histamine plays a role in EAE, but also to unveil a role of neutrophil-derived histamine in IgE-mediated anaphylaxis. This mouse model will prove useful in future studies to re-explore the various inflammatory and neurologic roles of histamine.

The authors have no financial conflicts of interest.

We thank the McGill Integrated Core for Animal Modeling at the Goodman Cancer Research Centre, Montreal, for providing services for the creation of the conditional Hdc knockout mouse line.