Changes in the prevalence and biofilm formation of Haemophilus influenzae and Haemophilus parainfluenzae from the respiratory microbiota of patients with sarcoidosis

Background Healthy condition and chronic diseases may be associated with microbiota composition and its properties. The prevalence of respiratory haemophili with respect to their phenotypes including the ability to biofilm formation in patients with sarcoidosis was assayed. Methods Nasopharynx and sputum specimens were taken in 31 patients with sarcoidosis (average age 42.6 ± 13), and nasopharynx specimens were taken in 37 healthy people (average age 44.6 ± 11.6). Haemophili were identified by API-NH microtest and by the matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) system. Biofilm was visualised by crystal violet staining and confocal scanning laser microscopy (CSLM). The statistical analysis was performed with Statgraphics Plus for Windows. Results In total, 30/31 patients with sarcoidosis and 31/37 healthy people were colonized by Haemophilus influenzae (6/30 vs. 1/31) and Haemophilus parainfluenzae (28/30 vs. 31/31) in the nasopharynx. The overall number of nasopharyngeal haemophili isolates was 59 in patients with sarcoidosis and 67 in healthy volunteers (H. influenzae 6/59 vs. 1/67, P = 0.05; H. parainfluenzae 47/59 vs. 65/67, P = 0.0032). Moreover, the decreased number of H. parainfluenzae biofilm-producing isolates was shown in nasopharyngeal samples in patients with sarcoidosis as compared to healthy people (19/31 vs. 57/65, P = 0.006), especially with respect to isolates classified as strong and very strong biofilm-producers (8/31 vs. 39/65, P = 0.002). Conclusions The obtained data suggest that the qualitative and quantitative changes within the respiratory microbiota concerning the overall prevalence of H. influenzae together with the decreased number of H. parainfluenzae strains and the decreased rate of H. parainfluenzae biofilm-producing isolates as compared to healthy people may be associated with sarcoidosis.

Understanding the role of microbiota composition is a new frontier of human biology and the contemporary direction in the investigation of physiological or pathological phenomena of health or diseases [5][6][7]. Qualitative and quantitative shifts or perturbation in the microbiota can lead to the development of diseases. Microbiota monitoring and modification may be useful for determination of health, thus providing new means of protection and/or of intervention, and data interpretation [8][9][10]. The microbiota components predominantly colonizing the respiratory mucosa without causing any disease symptoms can occasionally cause respiratory infections. Besides possible positive or negative interactions between commensals and pathobionts as well as other potential pathogens, microbiota can play an important role for the human host organism in preventing of respiratory and invasive infections [11]. Additionally, microbiota disturbance can contribute to acquisition and carriage of pathogens, can predispose to viral co-infection, especially in people with an immature or damaged immune system.
The human-restricted respiratory tract microbiota representatives are Haemophilus influenzae with significant pathogenicity and opportunistic commensal H. parainfluenzae [12,13]. They may be etiologic agents of invasive or opportunistic diseases [14][15][16]. H. influenzae, both the encapsulated (mainly serotype b -Hib) and non-encapsulated (nontypeable H. influenzae -NTHi) strains have also been associated as potential pathogens with chronic or recurrent and invasive diseases (e.g. bacteremia or sepsis, otitis media, chronic bronchitis, and community-acquired pneumonia) often reported in children and rarely in adults. H. parainfluenzae, as an opportunistic bacteria, less often reported as an etiologic agent of infectious diseases, may cause systemic or other respiratory infections (e.g. epiglottitis, meningitis, bacteremia or sepsis, bronchitis, chronic obstructive pulmonary disease, and infective endocarditis).
The human microbiota is a reservoir of opportunistic and potential pathogens (pathobionts), including haemophili, living mainly in a diverse community of biofilm [17][18][19]. Biofilm as a structure of microbial community enveloped in a polymeric matrix and adhered to both natural and synthetic surfaces may be regarded as a phenotypic adaptation and protective or pathogenic factor in many infections, depending on the condition [11,[20][21][22]. It was found to be a form of microbial life important both in colonization and in chronic and recurrent or acute diseases such as otitis media and pneumonia caused by NTHi species [17]. Biofilm is estimated to be involved in about 65 % of human infections with bacterial etiology [23]. Adhesive properties, as well as biofilm formation by microoorganisms together with its intrinsic antimicrobial resistance, exopolysaccharide production and quorum sensing are factors allowing for adaptation to host organism [20]. Both H. influenzae and H. parainfluenzae have been found to be a biofilm-forming bacteria.
The objectives of the present study were: the analysis of the correlations of diagnostic results in patients with sarcoidosis based on simple regression, haemophili isolation in nasopharyngeal and sputum specimens, antimicrobial resistance determination in H. influenzae and H. parainfluenzae clinical isolates, biofilm production by clinical isolates of these species together with the analysis of its structure.

Patients
A group of 31 adult patients (average age 42.6 ± 13) with a suspicion of sarcoidosis who were diagnosed in 2011 at the Chair and Department of Thoracic Surgery (Medical University of Lublin, Poland), participated in the study. The selection criterion was sarcoidosis, which was diagnosed with clinical findings suggesting an incidence of this disease. Patients were directed for diagnosis because of radiological findings such as: lymphadenectomy or tumour of mediastinum, or the presence of small nodules and infiltrations in the lung parenchyma, sclerosis, thickening or fibrosis discovered in CT scans. Multivariable demographic, clinical, radiographic and histological data were collected on the basis of the patients' questionnaires and information protocol.
All patients were diagnosed by means of bronchoscopy, mediastinoscopy or/and lung biopsy. Before the procedure blood samples were collected for standard blood tests (basic metabolic panel and complete blood count). The obtained tissue samples were evaluated by the same pathomorphologist. The histopathological findings were usually described as tuberculosis like granulation which could be considered as sarcoidosis in accordance with clinical changes.
A control group of 37 healthy volunteers (average age 44.6 ± 11.6) who agreed to participate in the survey was also included. They did not suffer from respiratory infections and had not received an antimicrobial therapy for at least three months prior to the examination or had not been admitted to hospital for at least two years.
Written informed consent for participation was obtained from people who agreed to take part in the study and filled out the survey. The Ethics Committee of the Medical University of Lublin approved study protocol (KE-0254/75/2011).

Microbiological processing of haemophili isolates
A total of 31 nasopharyngeal swabs and 31 sputum specimens were taken from patients with sarcoidosis on the day of hospitalization or a day after. Additionally, 37 nasopharyngeal specimens were collected from healthy people.
After incubation (48 h, 35 ± 2°C, 5 % CO 2 ) the colonies with morphological differences were identified independently on selective HAEM-medium (Haemophilus-chocolate-agar, bioMérieux, France). The growth of bacteria in the form of individual colonies or from abundant to very abundant number of morphologically different colonies on Chocolate agar was observed. Initially biochemical identification and biotyping of 192 Gram-negative isolates (125from patients with sarcoidosis and 67from healthy people) was carried out using the API-NH microtest (bioMérieux). The phenotypes of haemophili isolates were differentiated based on various observable properties in the growth morphology (e.g. the shape and size of the colony, smooth or rough surface, texture, colony elevation), on a set of biochemical reactions (according to API NH results) and antimicrobial susceptibility results. API-NH is a standardized system for the identification of Neisseria, Haemophilus (and related genera) and Moraxella (B.) catarrhalis, which uses microtests and a specially adapted database. Next, for the species differentiation the matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS, Bruker Daltonik, Germany) system was used according to the procedure described earlier [24]. Software used for data acquisition was MALDI-Biotyper 3.0 (Bruker Daltonik) software. The species which were not identified as H. influenzae or H. parainfluenzae were classified as other Haemophilus spp.

Biofilm detection
Biofilm formation was examined during the stationary culture in vitro in 24-well polystyrene microplates (24 F-Well Microplates, Thermo Scientific™ Nunc™, Denmark) using a 0.1 % crystal violet (CV) stain as previously described [26]. The Tripticasein Soy Broth (TSB, Biocorp, Poland) supplemented with Haemophilus Test-Medium Supplement (HTMS, Oxoid) designated as TSB + HTMS was used. Overnight cultures were diluted in TSB + HTMS-medium and standardized at 570 nm with an initial optical density of OD 570~0 .08 ± 0.02 (~0.5 McFarland-standard) using a microplate reader ELx800 (BioTek Inc., USA). Next, 500 μl of the microbial suspension was inoculated for each well and incubated (35°C, 24 h, 5 % CO 2 ). The growth of haemophili was assessed by measuring the OD 570 . Nonadherent cells were removed by rinsing the wells with sterile water. The biofilm was detected with OD 570 according to the method based on staining with 500 μl 0.1 % CV. Each isolate was tested in triplicate in three series. TSB + HTMS without bacteria was incubated under the same conditions and served as blank control.
Haemophili were classified as biofilm-producers as described elsewhere [26]. The experiments were performed in triplicate and the results were averaged. For the purposes of a detailed analysis of the obtained results the classification of biofilm producers was introduced on the basis of criteria proposed by Stepanović et al. [27] and modified by Kosikowska et al. [26]. They defined the cut-off an optic density OD (OD c ) for the microtiterplate test as three standard deviations above the mean OD of the negative control. The bacteria were classified as follows: OD ≤ OD c -non-producers (category 0); OD c < OD ≤ 2 x OD c -weak producers (category 1); 2 x OD c < OD ≤ 4 x OD cmoderate producers (category 2); 4 x OD c < OD ≤ 8 x OD cstrong producers (category 3), 8 x OD c < ODvery strong producers (category 4). OD c was 0.1 ± 0.02 in our experiments. Additionally, H. influenzae ATCC10211, H. parainfluenzae ATCC7901, H. parainfluenzae ATCC33392 and H. parainfluenzae ATCC51505 reference strains were used as biofilm producers. All tests were carried out three times and the results were averaged.
To visualize H. influenzae and H. parainfluenzae biofilm formation in a 24 h culture in 24-well polystyrene microplates, the inverted microscope Axiovert 200 M equipped with LSM5 Pascal Head (Carl Zeiss, Germany, with magnification 200×) was used. To obtain images of the biofilm, cultures were stained with Bacterial Live/Dead ® BacLight™-L7012-kit (Invitrogen, USA) accordingly to the manufacturer's procedure. Biofilm was stained with Live/Dead ® L7012 kit with two components, Syto9 and propidium iodide (PI), which are nucleic acid stains. The Syto9 stain penetrates through the intact and damaged cell wall, intercalates into DNA, and emits green fluorescence (detection of both live and dead bacteria). The PI stain diffuses only across the dead and damaged cell wall, intercalates into DNA and emits red fluorescence (detection of dead bacteria). The use of a combination of these two dyes is that PI displaces Syto9 (it had a higher affinity) reducing its fluorescence. Thus, the bacteria with intact cell membranes stain fluorescent green, whereas with damaged cell membranes stain fluorescent red. Biofilm formation by 7 H. influenzae and 18 H. parainfluenzae randomly selected clinical isolates taken in patients with sarcoidosis was screened using confocal microscopy. The thickness of biofilm (μm) and measurement area covered by biofilm (%), content of live cells (%) and biofilm area covered by live cells (%) were detected.
Overnight cultures were standardized in TSB + HTMSmedium. Then 500 μl/well of cultures was added (for each strain in quadruplicate). After incubation (35°C, 24 h, 5 % CO 2 ) the content of the wells was removed and each well was washed three-four times with 0.85 % NaCl. Then, 500 μl/well of 0.85 % NaCl was added and then stained with 1.5 μl of Live/Dead-kit solution. Microphotographs were taken in the green and/or red channel in a Confocal Laser Scanning Microscope LSM5-PASCAL (CLSM; Carl Zeiss, Germany) after incubation for 15 min in the dark at room temperature. This experiment was repeated twice.
The planimetric measurement of the biofilm was performed based on the microphotographs taken at 50× or 200× magnification in a two-dimensional scan (2D). Biofilm parameters were calculated using ImageJ-1.43e software (Wayne Rasband, National Institutes of Health, USA). A three-dimensional (3D) image of biofilm was reconstructed using the CLSM (200× magnification).

Statistical analysis
The statistical analysis was performed with Statgraphics Plus for Windows, Version 4.1 (Statistical Graphics Corp. 1999, Statpoint Technologies, Inc. Warrenton, Virginia, USA). To assess the relationship between the variables a simple regression analysis was used. The aim of the statistical analysis was preliminary determination of the relationship between individual factors and the dependent variables. The relationship between age/gender and different clinical manifestations and laboratory findings were studied. The hypotheses were raised: H 0 -there is no association between age/gender and different clinical manifestations and laboratory findings, and H 1 -there is an association between age/gender and different clinical manifestations and laboratory findings. Quantitative variables were presented as mean (± standard deviation, SD) and median values. In some studies Fisher's test was evaluated. The level of P < 0.05 was usually considered as statistically significant.

Characteristics of patients
The main characteristics of patients with sarcoidosis are summarized in Table 1. Demographic and clinical data were obtained from the patients' files and based on a questionnaire conducted for the presented variables. "Fatigue" was self reported as being unable to go any further. Susceptibility to "fatigue" was associated with a short walk on a flat surface, or with stair walk to the first or second floors. Symptoms such as shortness of breath, wheezing, cough, expectoration, hypertension, coronary artery disease, allergy, asthma, and recurrent infections (≥3/year) were also considered. Almost 26 % of the patients with sarcoidosis showed recurrent infections in childhood and had other symptoms such as anaemia, celiac disease, or endocrinological disorders. Some of them had two parallel disorders, for example hypertension and coronary artery disease (about 13 %), and/or allergy (about 10 %).
There was a positive correlation (P < 0.1) between gender and a tumor of the lung/mediastinum (c = 0.4) or fluid in the pleural cavity (c = 0.32) in patients with sarcoidosis ( Table 2). These changes mostly occurred in males rather than in females. A positive correlation was demonstrated between gender and CRP (C-reactive protein; P = 0.01, c = 0.44), and in elderly patients between the disease and visible changes (e.g. sclerosis, thickening or fibrosis) in the imaging diagnostics of the chest (P

Number of haemophili isolates
From one to six different phenotypes of haemophili were isolated in the nasopharynx and/or in the sputum   15.4 %, P = 1.000) were also found, but they were not statistically significant.
Morphometric parameters of biofilm formed by 7 H. influenzae and 18 H. parainfluenzae isolates taken from the patients with sarcoidosis were assessed by means of CLSM technique ( Table 7). The biofilm formed by H. influenzae isolates had the thickness of 14.2 ± 3.5 μm and covered 71.6 ± 5.6 % of the measured area. The content of living cells was from 8.6 to 95.2 % (average 71.6 ± 4.4 %) and it was 75.2 ± 4.8 % in the biofilm area. The biofilm formed by H. parainfluenzae isolates had the thickness of 20.02 ± 4.3 μm and covered about 50.6 ± 4.8 % of the measured area. The content of living cells was from 42.8 to 99.9 % (average 80.9 ± 4.8 %) and it was about 77.9 ± 8.98 % in the biofilm area.
The biofilm formed by one of H. parainfluenzae nasopharyngeal isolate taken from a patient with sarcoidosis   Abbreviations: Am ampicillin, AmC amoxycillin/clavulanate, Caz ceftazidime, Ctx cefotaxime, Sam ampicillin-sulbactam, Sxt trimethoprim/sulfametoxazole, Te tetracycline was revealed by CLSM image (Fig. 2). The biofilm area formed by living (Fig. 2a) and dead (Fig. 2c) cells is presented. The textural parameters were detected as the grey scale intensity of the biofilm formed by live (Fig. 2b) and dead (Fig. 2d) cells.

Discussion
Several environmental agents interacting with social and genetic factors and including immune responses to microbial components rather than an infection per se have been considered to play a role in the pathogenesis of sarcoidosis [1,2,4]. In our study changes in the healthy conditions, in the imaging diagnostics of the chest and in levels of three major immunoglobulins (IgG, IgM and IgA) as well as CRP value with respect to sarcoidosis and the age and gender were shown (Tables 1 and 2).
Similar observations were also done by other authors [28][29][30][31]. Hiperglobulinemia is frequently observed in patients with sarcoidosis [30]. Changes observed in this group of patients [29,30] and in elderly people [31] suggested an association between age, race, sex and the immune system's defense as well as microbials presence. According to Buckley et al. [29], in patients with sarcoidosis the increase in IgG was significant in white patients, and the increase in IgM concentration was significant only in black patients, especially in black woman. Cagatay et al. [32] found higher than normal laboratory values for immunoglobulins IgG, IgA and IgM. They have noted that IgG and IgA levels were significantly higher in the group of routinely checked patients without any clinical   Distribution of biofilm-producers was detected using the crystal-violet (CV) method. The cut-off OD (OD c ; here: 0.1 ± 0.02) was defined as three standard deviations above the mean OD 570 of the negative control. The categories of biofilm-producers: 0 -non-producers (OD ≤ OD c ); 1, weak (OD c < OD ≤ 2xOD c ); 2, moderate (2xOD c < OD ≤ 4xOD c ); 3, strong (4xOD c < OD ≤ 8xOD c ); 4, very strong (8xOD c < OD) producers, according to [26]   symptoms and during established the activity of the disease. Drent et al. [33] have identified a high CRP concentration associated with severe fatigue in sarcoidosis.
The human microbiota has multidirectional effects on the host's health and changes in its composition may have important consequences for human pathophysiology and disease development [5,7,8,34]. According to literature, bacterial 16S rRNA from H. influenzae as well as Moraxella catarrhalis was detected in sarcoid fluid samples [35]. Cantwell's review [3] pointed to the presence of other bacterial species (e.g. Mycobacterium spp., staphylococci, Propionibacterium acnes) in the biopsy of sarcoid tissue, blood, and skin samples taken from patients with sarcoidosis.
According to our results, higher rate of nasopharyngeal colonization by H. influenzae, but not by H. parainfluenzae (P = 0.05 vs. P = 0.49) was found in patients with sarcoidosis as compared to colonization in healthy people. It suggests that sarcoidosis can be regarded as a factor predisposing for colonization by H. influenzae. Besides, H. influenzae was frequently found in the sputum samples taken in patients with sarcoidosis even if this bacterial species was absent in the nasopharynx. In our opinion, it suggests the role of H. influenzae in the lower respiratory tract colonization and inflammation particularly in chronic diseases, e.g. in patients with sarcoidosis. In the literature [36], it has been documented that a combination of pathogenic mechanisms of bacteria and defects in host defense may allow this species to their migration into the lower respiratory tract. This may result in chronic colonization and/or in acute exacerbations of airway disease.
Despite the high and a similar number of people colonized by H. parainfluenzae in the nasopharynx of healthy people and patients with sarcoidosis (83.8 % vs. 96.8 %), we observed a decreased number of this species isolates with different phenotypes, differentiated on the basis of growth morphology, biochemical characteristics and biotypes. In 30 patients with sarcoidosis 47 isolates of H. parainfluenzae were identified, while in 31 healthy people -65 isolates of this species. H. parainfluenzae isolates with biotypes I and II were found to occur most frequently, similarly in patients with sarcoidosis and in healthy people (Table 3). As found by other authors [37][38][39], these biotypes constituted most of H. parainfluenzae isolates in patients with other respiratory diseases such as chronic bronchitis or cystic fibrosis.
During our studies we compared the antimicrobial susceptibility of nasopharyngeal H. parainfluenzae isolates obtained in patients with sarcoidosis and in healthy people ( Table 4). Resistance of these isolates to tetracycline (P = 0.041) and trimethoprim/sulfametoxazole (P = 0.297) in patients with sarcoidosis was higher compared to that in healthy people. In contrast, resistance to beta-lactams (P = 0.214), including ampicillin (P = 1.000) was lower in patients with sarcoidosis compared to that in healthy people. These differences may be due to higher consumption of a given group of antimicrobials in a defined population. The growing antibiotic resistance and reduction or elimination their effectiveness is one of the world's most pressing public health problems [40].
Beta-lactam antibiotics are the most widely used antimicrobial agents during treatment of both communityacquired and hospital infections. The resistance to this group of antimicrobials in Haemophilus spp. usually is mediated by the production of beta-lactamases and the presence of altered penicillin-binding protein (PBP) with lowered affinity for these antibiotics as a target site [17,41]. The ampicillin-resistant, beta-lactamase positive H. parainfluenzae isolates from patients with sarcoidosis were found in our studies (Tables 4 and 5). This may have some implications including the possibility to exchange resistance genes within microorganisms [40][41][42][43][44][45].
According to literature, over the past years many authors detected beta-lactamases mainly in H. influenzae and rarely in H. parainfluenzae isolates taken from patients with respiratory tract infections as well as from healthy people [25,26,46]. It was shown that DNA mutation and rapid multiplication as well as transformation can be important mechanisms in the spread of drug resistance in haemophili, including the ampicillin resistance due to beta-lactamase production [45,46]. It seems that especially efficient in transformation were H. parainfluenzae cells with a highest ability to develop competence and transfer of resistance genes occurs via free DNA (from dead or lysed cells) during natural transformation from the medium by competent cells. It may explain the acquisition of resistance or resistance gene exchange with other microorganisms. Besides, resistance and reduced susceptibility to beta-lactams mediated by altered PBPs is also important in many bacterial pathogens, including beta-lactamase negative H. influenzae [47]. For this reason, there are different events that may contribute to the emerge of resistance: the acquisition of resistance genes (e.g. beta-lactamases) by conjugation or transformation; and inter-species recombination of the ftsI gene [47,48]. According to Gromkova et al. [46], most efficient in transformation among H. parainfluenzae strains was biotype II, followed by biotype I.
It was shown in this paper that H. parainfluenzae isolates selected in patients with sarcoidosis compared to isolates selected in healthy people had a lower ability for biofilm production (Table 6). It is possible that a reduction in the number of H. parainfluenzae strains capable of biofilm formation may contribute to an increased colonization by certain opportunistic pathogens like H. influenzae (the present results) or by other bacteria [3,4]. The CLSM technique revealed that biofilms formed in vitro by H. parainfluenzae and H. influenzae isolates taken from patients with sarcoidosis (Table 7, Fig. 2) had high content of live cells (average 72 to 81 %), suggesting the possibility of bacterial persistence and dispersal in vivo.
Biofilm may be regarded as a pathogenic or protective factor depending on the conditions [18,21]. Nontypeable H. influenzae [NTHi] biofilms were observed for both bacteria colonizing the tissue of human, as an etiologic agent which causes the infection or exacerbation of chronic respiratory diseases [49][50][51]. On the other hand, Clancy and Dunkley [52] showed that oral NTHi could enhance the mucosal protection and prevent exacerbations of a chronic obstructive pulmonary disease.
On the basis of our results we propose, that haemophili, mainly H. influenzae and H. parainfluenzae, would be microorganisms indicative in respiratory microbiota changes as well as healthy condition in patients with chronic diseases. We observed higher frequency of H. influenzae colonization in patients with sarcoidosis compared to healthy people (P = 0.05). All H. influenzae isolates were weak-or non-biofilm producers independently to source of isolation and peoples' health condition. Moreover, despite the high prevalence of H. parainfluenzae in the nasopharynx of people from both studied groups (P = 0.49), less number of isolates of this species obtained in nasopharyngeal samples in patients with sarcoidosis as compared to healthy people were classified as biofilm-producers (61.3 % vs. 87.7 %, P = 0.006), especially as strong and very strong biofilm-producers (25.8 % vs. 60 %, P = 0.002).

Conclusions
The obtained results suggest that sarcoidosis, associated with many different factors, may be partially due to the respiratory microbiota condition. This is the first study describing qualitative and quantitative changes in the respiratory microbiota in patients with sarcoidosis with the respect to H. influenzae and H. parainfluenzae biotypes and their ability for biofilm formation. The question whether the biofilm formed by these bacterial species is a causative or just a protective factor in recurrent or chronic diseases, e.g. sarcoidosis, requires further studies.