Histamine Promotes Pseudomonas aeruginosa Biofilm Formation and Renders P. aeruginosa Biofilms More Resistant to Gentamicin and Azithromycin

The Gram-negative bacterium Pseudomonas aeruginosa is a major pathogen in chronic lung diseases including cystic fibrosis (CF) [1] and non-CF bronchiectasis [2]. Individuals with CF become transiently infected with P. aeruginosa before being colonized with this bacterium. Once colonized by P. aeruginosa, lung function declines rapidly as illustrated by respiratory failure, the main cause of death [3].

To establish a chronic infection in the airways, P. aeruginosa must adapt to the hostile lung environment such as oxidative stress, low iron concentrations, the innate immune response of the host, and aerosolized antibiotic treatments. P. aeruginosa attachment to a surface and subsequent biofilm formation are hallmarks of the capacity of this bacterium to cause persistent infections [4]. A biofilm is a structured consortium of bacteria, embedded in a self-produced polymer matrix forming a protective layer that keeps antibiotics and immune components from clearing the bacteria. Biofilms can vary in their makeup from a population of one bacterial species to a complex community of many microorganisms [5].

Biofilm formation can occur on both natural and artificial surfaces, most commonly on the skin, respiratory, and urinary tracts in humans, and on the surfaces of catheters [6] and various other medical apparatus or contact lenses. Formation of P. aeruginosa biofilms is a complex process which can be divided into five different phases: initial adhesion of planktonic bacterium, early attachment, formation of young biofilms, mature biofilms, and finally, P. aeruginosa dispersal [5]. These phases in P. aeruginosa biofilm formation are dependent on two main bacterial appendages: type IV pilus and the flagella [7]. Bacteria move in a viscous milieu by a flagella-dependent process called swarming [8] or exhibit surface-associated movement through twitching motility mediated by functional type IV pili [9]. Swarming-negative PAO1 mutants are impaired in their capacity to form biofilms, indicating a close link between swarming motility and PAO1 biofilms [8]. However, the relationship between bacterial motility and biofilms is complex. This is well illustrated by the fact that airway environments select for populations of P. aeruginosa which have lost flagella [10, 11] and twitching motility due to mutations in type IV pili genes [12]. Thus, expression of the flgE gene encoding for the immunogenic flagellar hook protein FlgE, which is key for the synthesis of flagella and swimming motility, is reduced during biofilm maturation. Such decrease in the expression of FlgE in biofilms represents adaptations that modify the structure of P. aeruginosa biofilms rendering the biofilms more resistant to antibiotics [11]. It is plausible that the requirement or not of bacterial motility varies during the different stages of biofilm formation.

Formation of biofilms has been characterized in mouse models of P. aeruginosa airways infection. For example, lung homogenates from mice chronically infected with the NH57388A strain have two distinct P. aeruginosa colony morphologies, mucoid NH57388A (forming large colonies) and small colony variants (forming smaller colonies) resulting from adaptations to the CF lung environment [13].

The microbiome of CF and non-CF bronchiectasis airways is complex. It is conceivable that bacterial metabolism results in products that promote bacterial biofilms. One of these molecules is histamine which is produced via decarboxylation of l-histidine by histidine decarboxylases (HDCs) [14]. Indeed, respiratory tract bacteria found in acute exacerbations of chronic bronchitis, CF, or pneumonia including P. aeruginosa, Klebsiella pneumonia, B. catarrhalis, and H. parainfluenzae generate large amounts of histamine in vitro [15]. In support for a role of histamine-producing bacteria in the pathogenesis of chronic airways diseases, it has been shown that sputa from individuals with chronic obstructive bronchitis or CF contain bacteria synthesizing histamine [16]. P. aeruginosa express histamine receptors (TlpQ, PctC, and PctA) [17] which are chemoreceptors enabling communities of bacteria to communicate with one another for biofilm synthesis. Therefore, histamine produced by the bacteria could represent a previously unappreciated evolutionarily conserved molecular dialogue between bacteria and the host, whereby production of histamine would help bacteria to form biofilms and resist destruction by the immune system of the host or antibiotics. We then hypothesized that histamine promotes P. aeruginosa adhesion, biofilm formation, and resistance to antibiotics.

The histamine ELISA kit was from IBL International GmbH/Tecan (Hamburg, Germany). Histamine (Ref: H7125) and diamine oxidase (DAO) (D7876) were purchased from Merck Life Science (UK). Powder growth medium for BHI broth (CM1135) was purchased from Oxoid (Basingstoke, UK). NaCl (S/3120/65) was purchased from Fisher Scientific (UK), and tryptone (Ref: T7293) and yeast extract (Ref: 92144) were from Merck Life Science (UK). The Nunclon Delta surface 96-polystyrene well plates used for biofilm assays were from Thermo Fisher Scientific (Roskilde, Denmark).

The standard laboratory-adapted P. aeruginosa PAO1 strain ATCC 15692 (sourced from the UK Health Security Agency) or the mucoid strain DK1-NH57388A [18] were inoculated either in BHI broth or LB medium (2.5 g yeast extract, 5 g tryptone, 5 g NaCl for 500 mL H₂O) and grown at 37°C in a glass flask under orbital shacking. After 18 h, the bacterial solution was adjusted to an O.D. value at 600 nm of 0.03. 100 μL of these diluted suspensions were added to 96-well polystyrene microplates in the absence or presence of DAO, histamine, or a combination of DAO and histamine. The polystyrene 96-well plates were incubated in a humidified atmosphere at 37°C. After 18 h, the supernatants were removed by aspiration, and the plates washed twice with PBS (400 μL), after which, the adhered bacteria were fixed with methanol for 10 min, followed by staining with crystal violet (0.1% crystal violet in 10% methanol) for 15 min. The plates were then washed twice with PBS (400 μL). The stain was eluted with a solution containing 50% ethanol and 0.1 m sodium citrate (pH 4.2). Using a 96-well plate reader, the O.D. values of the elution were measured at 570 nm. The mean O.D. values were calculated from four wells (quadruplicates) for each experimental condition.

A stock solution of DAO (10 mg/mL) was prepared by dissolving the DAO powder in BHI broth or LB medium, and the solution was then sterile filtered. A stock solution of histamine (10⁻¹m) was prepared in water and sterile filtered. The DAO solution was used immediately after preparation.

The haemolymph of 18 h surviving larvae, inoculated with either PBS or the PAO1 strain (2-5 CFU), was collected in 1.5 mL tubes on ice and protected from light to minimize melanization. 0.01 m sodium citrate (anti-coagulant) was added. The samples were pooled and subjected to centrifugation (2,000 g, 10 min). The supernatants, free of cells, were collected, and the concentration of histamine in the cell-free supernatants was measured using an ELISA kit (see below).

BHI broth, LB medium, or larvae haemolymph were used for quantification of histamine using an ELISA kit by following the protocol provided by the manufacturer. Concentration of histamine was calculated by referring to a standard curve.

PAO1 or the mucoid NH57388A strains were inoculated in LB medium or BHI broth and grown overnight in a glass flask under orbital rotation at 37°C. After 18 h, bacterial suspensions were adjusted to an O.D. value at 600 nm of 0.06–0.07. Different effectors diluted in growth media were added to the flasks including histamine (10⁻⁷m–10⁻⁴m) or DAO (10 mU/mL). The flasks were shaken at 37°C and, after each hour, an aliquot (1 mL) was collected, transferred to a plastic cuvette and the optical density at 600 nm was measured. LB medium or BHI broth served as a reference. For later time points, aliquots were diluted in medium prior to measuring the O.D. values.

Larvae (twenty per group with a weight between 250 and 350 mg) were injected with the P. aeruginosa PAO1 strain (2 CFU or 5 CFU/larvae) in the absence or presence of histamine (10⁻⁸m). These concentrations of PAO1 were chosen because they did not cause 100% mortality over the 50-h time period, a prerequisite to observe potentiation of the death of the organism by histamine. Inoculation of larvae was performed by injection into the haemocoel, through the last left proleg, using an insulin syringe of the following: 10 μL of the bacterial solution (2 or 5 CFU) in sterile PBS and a second injection of 10 μL of PBS, 10 μL of the bacterial solution (2 or 5 CFU) in PBS and a second injection of 10 μL of histamine (10⁻⁸m) in PBS, 10 μL of PBS followed by a second injection of 10 μL of PBS (controls). Thus, all larvae received two 10 μL injections. Larvae were placed at 30°C for up to 50 h and were assessed at regular intervals for viability and melanization. Larvae were considered dead when they displayed no movement in response to puncture with a microneedle. We could not test the effect of histamine on the virulence of the mucoid NH57388A strain in G. mellonella because its injection into the larvae led to the rapid death of the organism (within 1–2 h) even at a very low CFU (2 CFU).

To compare the means of multiple groups (control versus treatments), we used the analysis of variance (ANOVA) statistical test. We assessed whether the p value corresponding to the F-statistic test of one-way ANOVA was lower than 0.05 to evaluate if one or more pairs of groups were significantly different. When it was the case, we used a Tukey HSD test and a Scheffé multiple comparison test to identify which of the pairs of treatments are significantly different from each other. To further confirm statistical significance, we also compared treatment groups relative to controls simultaneously by using Bonferroni correction to adjust probability values.

To compare the means and standard deviations of two separate samples and to assess the significance of differences between the groups, we used a paired Student’s t test as the data were obtained from the same experiments. The non-parametric statistical Mantel-Cox test is a test that compares the survival rates of two or more groups and was used to assess the effect of histamine on larvae survival.

P. aeruginosa strains adhere and form biofilms on solid surfaces. Hydrophobic substrata such as polystyrene support P. aeruginosa adhesion [19]. The extent to which P. aeruginosa strains adhere to polystyrene surfaces and grow as biofilms is dependent on the nutrient composition of the medium. The Brain Heart Infusion (BHI) broth, rich in nutrients, is optimum for P. aeruginosa adhesion on polystyrene surfaces, biofilm growth, and maturation [20]. Luria-Bertani (LB) and nutrient broths are also rich in nutrients but are less effective than BHI broth to promote P. aeruginosa adhesion and biofilm growth. The reason for this is not clear. We hypothesized that BHI broth contains a substance which facilitates P. aeruginosa biofilms formation. BHI broth is made up of extracts from beef/pig heart and calf brain. We noted that these organs are rich in histamine. Indeed, histamine is present in caudate-putamen and cortical regions of the brain [21]. High concentrations of histamine are found in both myocardial tissue and the stellate ganglia which innervate the heart [22].

Next, we investigated whether histamine played a role in biofilm formation. To this end, we tested the ability of the P. aeruginosa strain PAO1 in BHI broth to form biofilms on polystyrene in the absence or presence of DAO. This enzyme catalyses oxidative deamination of histamine, leading to the formation of the biological inactive product imidazole acetaldehyde [23]. We found that in the presence of DAO (2.5 mU/mL and 10 mU/mL), the amount of PAO1 biofilms were reduced by ∼2-fold (p < 0.01) compared to the amount of biofilms formed in the absence of DAO (shown in Fig. 1a). It is important to note that there is no endogenous DAO activity in the BHI broth as the medium has been autoclaved. The fact that only partial, but significant, inhibition was obtained could be explained by the fact that imidazole-4-acetaldehyde is a potent inhibitor of DAO [23], and biofilm genes are also regulated independently from histamine. To verify that the effect of DAO is due to the catabolism of histamine, we measured in parallel the concentration of histamine in the BHI broth in the absence or presence of DAO. The concentration of the diamine was ∼140 ng/mL which corresponds to ∼1.3 μm. Addition of DAO (10 mU/mL) to the BHI broth reduced by ∼3-fold the concentration of histamine (shown in Fig. 1b).

Oxidation of histamine by DAO generates H₂O₂ [23]. We excluded that the reduction of PAO1 biofilms was indirectly mediated by H₂O₂, which has antimicrobial properties, for at least two reasons: Firstly, the decrease in biofilm formation brought about by DAO occurred regardless of whether or not sodium pyruvate (a scavenger of H₂O₂) was present in the BHI broth (shown in Fig. 1c). Secondly, P. aeruginosa possesses two catalases encoded by katA and katB genes which protect planktonic and biofilm cells against H₂O₂ [24]. We also showed that heat inactivation of exogenously added DAO totally abolished the ability of the enzyme to block biofilm formation (shown in Fig. 1d).

We carried out similar experiments with the mucoid NH57388A strain, which originates from an individual with CF and can be prone to form biofilms in lungs [18]. In contrast to what we observed with the PAO1 strain, addition of DAO (1–10 mU/mL) to the BHI broth did not significantly reduce mucoid NH57388A strain biofilm formation (shown in Fig. 1e). Addition of a high concentration of histamine (10⁻⁴m) to the LB medium led to a ∼2-fold (p < 0.01) increase in mucoid NH57388A strain biofilms, an effect abrogated by DAO (shown in Fig. 1f). These results show that the concentration of histamine in the BHI broth (∼1.3 μm) is sufficient to promote PAO1 strain biofilms but not mucoid NH57388A strain biofilms. As a control, we showed that addition of DAO to the BHI broth did not influence the planktonic growth of P. aeruginosa PAO1 strain (shown in Fig. 1g, left panel) indicating that reduction of biofilms by DAO is not due to modification in the bacterial growth. Likewise, DAO had no effect on the growth of the mucoid strain in BHI broth (shown in Fig. 1g, right panel).

To confirm the role of histamine in promoting biofilms, we investigated whether exogenously added histamine would augment biofilms. To test this, we used LB medium instead of BHI broth because the concentration of histamine in this medium (0.080 μm) is 16 times lower than in the BHI broth.

We found that histamine induced a dose-dependent (10⁻⁸–10⁻⁴m) increase in mucoid NH57388A strain biofilms (shown in Fig. 2a). A significant effect was observed with concentrations of histamine ranging from 10⁻⁵m to 10⁻⁴m (p < 0.01). Histamine also induced PAO1 biofilms (shown in Fig. 2b). A significant effect (p < 0.01) was obtained with a concentration of histamine as low as 10⁻⁸m. As a control, we showed that augmented P. aeruginosa biofilms in response to histamine was not due to augmented bacterial growth. Indeed, addition of histamine (10⁻⁷m–10⁻⁴m) to the LB broth had no effect on the planktonic growth of the mucoid NH57388A strain (shown in Fig. 2c, left panel) or the PAO1 strain (shown in Fig. 2c, right panel).

Next, we investigated whether histamine played a role in P. aeruginosa biofilm’s resistance to antibiotics. To this end, the P. aeruginosa strains were grown on polystyrene wells for 48 h to form biofilms. Thereafter, histamine (10⁻⁴m) was added or not added to the wells in the presence of antibiotics, either gentamicin or azithromycin. After 24 h, the amount of biofilm was measured (shown in Fig. 3a). Gentamicin or azithromycin are two antibiotics used to treat P. aeruginosa infections. For instance, azithromycin has been shown to improve the clearance of P. aeruginosa biofilms in a chronic rat lung infection model [25]. We found that gentamicin (5–100 μg/mL) (shown in Fig. 3b, left panel) or azithromycin (50–300 μg/mL) (shown in Fig. 3b, right panel) reduced the mucoid NH57388A biofilms in a dose-dependent manner. Interestingly, in the presence of histamine (10⁻⁴m), gentamicin (10–100 μg/mL) or azithromycin (50–300 μg/mL) were less effective in reducing biofilms (p< 0.01 or p < 0.05). Similarly, gentamicin (10–100 μg/mL) or azithromycin (150–300 μg/mL) were less effective in reducing PAO1 biofilms (p< 0.01 or p < 0.05) in the presence of histamine (10⁻⁴m) (shown in Fig. 3c). We concluded that histamine renders mature P. aeruginosa biofilms more tolerant to antibiotics.

P. aeruginosa produces histamine when cultured in vitro [15]. We investigated whether histamine is produced in a living organism in response to P. aeruginosa infection. To this end, we inoculated the larvae of G. mellonella with the PAO1 strain (2 CFU). After 18 h, the haemolymph of six surviving larvae were collected, pooled, and histamine concentration measured. Histamine concentration in the haemolymph of larvae inoculated with PBS was 5.4 ng/mL (∼5 × 10⁻⁸m) and reached 15.6 ng/mL (∼1.4 × 10⁻⁷m) (∼3-fold increase) in larvae inoculated with the PAO1 strain (2 CFU) (shown in Fig. 4a). To investigate whether histamine has an effect on the pathogenicity of PAO1, 20 larvae were inoculated through the proleg with 2 or 5 CFU of PAO1 and subsequently histamine (10⁻⁸m) or PBS were injected via the same route. We then measured larval survival over time. We found that histamine significantly reduced the survival rate of the larvae. In larvae inoculated with 2 CFU of PAO1, around 80% of the larvae survived after 24 h. In contrast, the survival rate was only 40% for the larvae inoculated with 2 CFU of PAO1 and histamine 10⁻⁸m (shown in Fig. 4b, left panel). In larvae inoculated with 5 CFU of PAO1, around 40% of the organisms survived after 24 h. When the larvae were inoculated with a combination of histamine (10⁻⁸m) and 5 CFU of PAO1, the survival rate decreased to 20% (as shown in Fig. 4b, right panel).

As a control, we showed that inoculation of histamine (10⁻⁸m) alone or PBS did not lead to the death of G. mellonella (tested in 10 larvae over a 48-h incubation period). These data demonstrate that as a result of PAO1 infection, histamine is produced in G. mellonella and augments P. aeruginosa virulence.

Gram-negative bacteria producing histamine, including P. aeruginosa, are present in the airways of individuals with CF or other chronic lung diseases [15, 16]. Bacteria colonizing the airways are not the only source of histamine. Indeed, it was shown that neutrophils are major producers of airways histamine in mice mycoplasma pneumonia [26]. These findings indicate that histamine may contribute to the pathogenicity of respiratory pathogens in the lower respiratory tract.

We provided evidence for a role of histamine in the regulation of P. aeruginosa biofilms. Firstly, we showed that converting biologically active histamine into a biologically inactive form by means of oxidation of the diamine by DAO, reduced biofilm formation by the PAO1 and the mucoid NH57388A strains. Secondly, exogenously added histamine increased biofilm formation. Notably, we observed a difference between the two P. aeruginosa strains in terms of histamine dependency to form biofilms. For the PAO1 strain, nanomolar concentrations of histamine were sufficient to promote biofilms, whereas for the mucoid NH57388A strain, micromolar concentrations were required. The PAO1 strain expresses three different histamine receptors, namely, TlpQ, PctA, and PctC [17]. TlpQ is a receptor activated by nanomolar concentrations of histamine. Therefore, TlpQ may be involved in the regulation of PAO1 biofilms. The TlpQ gene (European Nucleotide Archive, CAI9909172), with no loss-of-function mutations at the locus, is also present in the genome of the P. aeruginosa NH57388A strain (GenBank accession number cp003149). Based on these observations, the mechanism underlying the selective augmentation of NH57388A strain biofilms in response to micromolar concentrations of histamine remains unclear. One possibility may be that the TlpQ expression levels are relatively low in the mucoid strain, leading to histamine receptors with lower affinities for histamine (PctA and PctC) playing a dominant role in biofilm formation.

Interestingly, when histamine was added 48 h after the biofilms were made, the diamine had no effect on the biofilm mass. This result may suggest that histamine plays an important role during the early 24 h formation of biofilms by favouring the initial adhesion of bacteria to the hydrophobic surface and/or by production of the extracellular matrix.

Nebulized tobramycin, azithromycin, or gentamicin dramatically improve chronic respiratory diseases; however, lung infections cannot be eradicated once P. aeruginosa has colonized the airways. One of the reasons for the limited effect of antibiotics could be that P. aeruginosa strains form biofilms, which are largely resistant to antibiotics and to the immune system of the host. We found that gentamicin and azithromycin have anti-biofilm activity against P. aeruginosa PAO1 and mucoid NH57388A strains. Both antibiotics reduced in a dose-dependent manner 48 h established biofilms. Interestingly, in the presence of histamine, the anti-biofilm activities of gentamicin or azithromycin were significantly reduced. One possible mechanism whereby histamine could make the biofilms more tolerant to antibiotics may be through modification of the structure of P. aeruginosa mature biofilms, as exemplified by loss of flagella [11]. This hypothesis needs to be explored in the future.

It was shown that promotion of P. aeruginosa virulence by histamine requires the LysR-type transcriptional regulator HinK. HinK activates the promoters of genes involved in histamine uptake and metabolism, iron acquisition, and Pseudomonas quinolone signal (PQS) biosynthesis [27]. Interestingly, besides its role as a quorum-sensing signalling molecule, PQS generation ameliorates the adaptation of bacteria to environments and causes resistance to environmental stress [28]. Therefore, improved resistance of P. aeruginosa strains to antibiotics in the presence of histamine may be mediated via the alkyl-quinolone PQS. Finally, we provided evidence that histamine is produced in a living organism in response to P. aeruginosa infection. We cannot determine the source of histamine in the haemolymph of G. mellonella, in response to P. aeruginosa inoculation. However, we believe that the bacterium contributes significantly to the production of histamine in the living organism for at least two reasons. Firstly, P. aeruginosa produces histamine when grown in vitro [15]. Secondly, the Gram-negative bacterium Acinetobacter baumannii wild type, but not the A. baumannii strains in which the hdc gene has been inactivated, synthesizes histamine upon inoculation into G. mellonella [29]. It is not known how P. aeruginosa strains synthesize histamine because these Gram-negative bacteria lack a recognizable hdc gene. P. aeruginosa strains may have an hdc gene unrelated to the hdc genes of other Gram-negative bacteria or have an alternative way of producing histamine. Interestingly, bacterial arginine decarboxylases (involved in the production of putrescine) slowly decarboxylate l-histidine to produce histamine [30]. Furthermore, P. aeruginosa strains express a biosynthetic arginine decarboxylase (speA) gene. Thus, P. aeruginosa may produce histamine via an hdc-independent and speA-dependent pathway. We have demonstrated a role of histamine in bacterial virulence by showing that injection of a dose of histamine, which promotes P. aeruginosa biofilms in vitro, accelerated the death of G. mellonella.

Histamine potentiates P. aeruginosa biofilm formation and thus renders mature biofilms more resistant to gentamicin and azithromycin, two antibiotics used to treat P. aeruginosa airways infections in CF individuals. The ability of histamine to promote biofilms in vitro may explain at least in part, the augmented virulence of the PAO1 strain in vivo upon administration of histamine into G. mellonella. Our findings imply that reducing airways histamine in individuals with chronic airways diseases such as CF may be a promising new therapeutic strategy to improve the efficacy of antibiotics towards P. aeruginosa and thereby reduce bacterial burden.

We thank Prof Niels Høiby (University of Copenhagen, Denmark) for providing the mucoid NH57388A strain; Amy Parke and Conall Hughes (Queen’s University Belfast) for measuring biofilms; Amal El Banna for performing the work in G. mellonella; and Dr. Clara Radulescu for measuring histamine concentrations.

Ethics approval was not required because it did not involve any animals or human subjects. The use of G. mellonella does not require ethical approval.

The author declares no conflict of interest.

This work was financially supported by the Centre for Biomedical Science Education, Queen’s University Belfast through funding of honour projects (intra-mural funding).

Karim Dib performed experiments, supervised the work of honour student projects, collected and analysed the data, and wrote and submitted the manuscript.

The data will be made publicly available without any restrictions.