Carcinoembryonic antigen (CEA)-related cell adhesion molecules are co-expressed in the human lung and their expression can be modulated in bronchial epithelial cells by non-typable Haemophilus influenzae, Moraxella catarrhalis, TLR3, and type I and II interferons

Background The carcinoembryonic antigen (CEA)-related cell adhesion molecules CEACAM1 (BGP, CD66a), CEACAM5 (CEA, CD66e) and CEACAM6 (NCA, CD66c) are expressed in human lung. They play a role in innate and adaptive immunity and are targets for various bacterial and viral adhesins. Two pathogens that colonize the normally sterile lower respiratory tract in patients with chronic obstructive pulmonary disease (COPD) are non-typable Haemophilus influenzae (NTHI) and Moraxella catarrhalis. Both pathogens bind to CEACAMs and elicit a variety of cellular reactions, including bacterial internalization, cell adhesion and apoptosis. Methods To analyze the (co-) expression of CEACAM1, CEACAM5 and CEACAM6 in different lung tissues with respect to COPD, smoking status and granulocyte infiltration, immunohistochemically stained paraffin sections of 19 donors were studied. To address short-term effects of cigarette smoke and acute inflammation, transcriptional regulation of CEACAM5, CEACAM6 and different CEACAM1 isoforms by cigarette smoke extract, interferons, Toll-like receptor agonists, and bacteria was tested in normal human bronchial epithelial (NHBE) cells by quantitative PCR. Corresponding CEACAM protein levels were determined by flow cytometry. Results Immunohistochemical analysis of lung sections showed the most frequent and intense staining for CEACAM1, CEACAM5 and CEACAM6 in bronchial and alveolar epithelium, but revealed no significant differences in connection with COPD, smoking status and granulocyte infiltration. In NHBE cells, mRNA expression of CEACAM1 isoforms CEACAM1-4L, CEACAM1-4S, CEACAM1-3L and CEACAM1-3S were up-regulated by interferons alpha, beta and gamma, as well as the TLR3 agonist polyinosinic:polycytidylic acid (poly I:C). Interferon-gamma also increased CEACAM5 expression. These results were confirmed on protein level by FACS analysis. Importantly, also NTHI and M. catarrhalis increased CEACAM1 mRNA levels. This effect was independent of the ability to bind to CEACAM1. The expression of CEACAM6 was not affected by any treatment or bacterial infection. Conclusions While we did not find a direct correlation between CEACAM1 expression and COPD, the COPD-associated bacteria NTHi and M. catarrhalis were able to increase the expression of their own receptor on host cells. Further, the data suggest a role for CEACAM1 and CEACAM5 in the phenomenon of increased host susceptibility to bacterial infection upon viral challenge in the human respiratory tract.

able to mediate homophilic and heterophilic interactions and can trigger actin cytoskeleton re-organization [25,26]. CEACAM1-S can thus not only mediate adhesion as well as internalization of pathogens [27], but can also control the signaling activities of the long CEACAM1 isoforms [28].
Chronic obstructive pulmonary disease (COPD) is a major cause of morbidity and mortality and is expected to be the third leading cause of death, and the fifth leading cause of disability by 2020 [29]. The Global Initiative for Chronic Obstructive Lung Disease defines COPD as characterized by persistent airflow limitation that is usually progressive and associated with an enhanced chronic inflammatory response in the airways and the lung to noxious particles or gases. Exacerbations are associated with increased airway and systemic inflammation, and evidence suggests that 70% may be caused by microorganisms [30]. Two pathogens that are associated with acute exacerbations and the progression of COPD are the CEACAM-interacting pathogens M. catarrhalis and H. influenzae, which often colonize the mucosa of the lower human respiratory tract in patients with COPD [31][32][33]. M. catarrhalis and H. influenzae express structurally unrelated outer membrane proteins, ubiquitous surface protein A1 (UspA1), and P5homologous adhesin (P5), respectively, that share the ability to bind to the extracellular immunoglobulin V (IgV)like domain of human CEACAM1 [5,8]. The interaction of CEACAM1 with M. catarrhalis results in reduced TLR2-initiated inflammatory responses of primary pulmonary epithelial cells [16]. CEACAM5 and CEACAM6 can mediate bacterial adhesion as well [5,7,8,34]. All three CEACAMs in human airway epithelia can therefore be of importance for the colonization of the lower airways and have a role in acute exacerbations. Since the lower respiratory airways are normally sterile and protected by mucociliary clearance, CEACAMs expressed here are most likely to encounter bacteria in medical conditions leading to dysfunction of the mucociliary clearance, such as COPD [35]. To date, a comprehensive analysis of (co-) expression patterns of CEACAM1 isoforms, CEACAM5 and CEACAM6 in the different lung tissues is lacking.
In the present study, we found CEACAM1, CEACAM5, and CEACAM6 expression on all pulmonary epithelia of the majority of the tested 19 individuals. Expression patterns were not dependent on COPD, smoking status and granulocyte infiltration. In NHBE cells, CEACAM1 expression was enhanced upon exposure to interferons, the TLR3 agonist polyinosinic:polycytidylic acid (poly I:C), M. catarrhalis, and NTHi. However, there were no differences in the induction of the different CEACAM1 isoforms tested (CEACAM1-4L, CEACAM1-3L, CEACAM1-4S and CEACAM1-3S). CEACAM5 expression was increased by interferon-gamma and CEACAM6 was unaffected by all treatments tested.

Acquisition and processing of human lung specimens
Tissue was obtained from surgical specimens, patients underwent surgery for lung resection to treat lung cancer. A positive vote of the ethics committee of the University of Heidelberg and informed consents were obtained. The resected tissue was fixed in formalin and embedded in paraffin using standard procedures [36].

Immunohistochemical analysis
Immunohistochemical staining for CEACAM1, CEACAM5, CEACAM6 and CEACAM8 using 20 μg/ml of the monoclonal antibodies B3-17, 5C8C4, 1H7-4B and 6/40c, respectively, was performed on paraffin wax sections obtained from 19 different human lung sections. Sections were deparaffinized (Histoclear 3 × 5 min, ethanol 100% 5 min, ethanol 96% 5 min, ethanol 80% 5 min, ethanol 70% 5 min) and rehydrated (H 2 O 2 × 5 min). Rabbit serumblocking (2% in PBS) was performed before the sections were incubated with primary antibody over night at 4°C, and with HRP-coupled secondary rabbit antibody for 1 h at RT. Then the nickel-glucose oxidase development technique was performed to enhance the DAB chromogen of the peroxidase immunohistochemistry. The specimens were counterstained with Calcium Red for 1-2 minutes to visualize the tissue structure. The result was a distinct purple/black staining product. As negative control served an IgG1 istotype matched control antibody. Non-cancer tissues from the sections were used for analysis.

Acquisition of bronchoalveolar lavage fluid (BALF)
BALF samples were obtained from 181 individuals in whom a bronchoscopy was performed for different diagnostic or therapeutic purposes. All individuals underwent bronchoscopy following standard diagnostic procedures at the department of infectious diseases and pulmonary medicine at the Charité -Universitätsmedizin Berlin with informed consents from the patients. The study was approved by the ethical committee of the Charité and all samples were available as residual material without any personal information or clinical data. BALF was centrifuged at 1,200 rpm and supernatants were then collected and stored in aliquots at −80°C until processing for ELISA.
CEACAM1-, CEACAM5-, and CEACAM6-specific Sandwich-ELISA Frozen BALF samples were thawed and centrifuged at 2000 × g and 4°C for 10 min. Then CEACAM1, CEACAM5, and CEACAM6 were assessed in the supernatants. 96 well micro titer plates (MaxiSorb TM plates, Nunc) were coated for 2 h at room temperature with 3 μg/ml rabbit anti-CEA-antibody (Dako) diluted in PBS. After washing the plate twice with 0.05% Tween (Carl Roth GmbH) in PBS all unbound sites were blocked for 2 h at room temperature with PBS containing 3% bovine serum albumin (Carl Roth GmbH). To quantify the different CEACAMs in the samples appropriate standard curves were prepared by making serial dilutions of recombinant human CEACAM1-Fc, recombinant human CEACAM6-Fc and purified CEA (from 0.25 ng/ml to 100 ng/ml). The standard and the undiluted samples were incubated over night at 4°C and washed three times. Then 20 μg/ml C5-1X (anti-CEACAM1), 5C8C4 (CEACAM5), or 1H7-4B (CEACAM6) were added. Plates were washed three times with PBS and supplemented with HRPcoupled goat anti mouse antibody (Dianova) for 2 h followed by three washing steps. Then 100 μl TMB-X-tra substrate (Biotrend Chemikalien GmbH) were added and incubated for up to 30 min. The reaction was stopped by 20 μl of 2 N H 2 SO 4 (Carl Roth) and optical densities were read at 450 nm in a microplate reader (Tecan). All antibodies, samples and standard curves were diluted in PBS containing 1.5 % BSA. The linear ranges for CEACAMs 1, 5, and 6 were 0.25 ng/ml -100 ng/ml.
Preparation of aqueous phase cigarette smoke extract (CSE) CSE was prepared as described [37]. Briefly, smoke from one cigarette (10 mg Tar, 0.8 mg nicotine; Camel Filters, Japan Tobaco International) with the mouthpiece filter removed was bubbled through 10 ml complete medium using a peristaltic pump (P-1, GE Healthcare) and filtered through a 0.22 μm filter. Suction was regulated, so that sidestream smoke developed during the entire combustion, lasting 5 min for each cigarette. The filtered CSE was regarded as 100%. CSE was used within 30 min of preparation at a final concentration of 4%. Lau et al. showed that NHBE cells were induced to secrete IL-8 and remained viable after tratment with 4% CSE for 24 h [37].
Immunoprecipitations from cell culture supernatants NHBE cell culture supernatants from confluent cells were harvested after 48 h. CEACAM1, CEACAM5 and CEACAM6 were precipitated from 10 ml supernatant using 20 μl protein G-Sepharose 4 Fast Flow (GE Healthcare) and 3 μg of monoclonal antibodies B3-17, 2H8-5, and 13H10, respectively. An IgG1 isotype control was included. Immunoprecipitations were analyzed by Western blot using anti CEA polyclonal antibody that cross-reacts with all three CEACAMs and HRP-coupled goat anti-rabbit IgG. Analysis was done using the Fusion FX System (Peqlab) and SuperSignal West Pico Chemiluminescent Substrate (Thermo Scientific).

RT-PCR and qPCR analysis
To analyze the gene expression of CEACAM1, CEACAM5, and CEACAM6, total RNA was extracted from 2×10 6 NHBE cells using the Qiagen RNeasy mini kit. Residual genomic DNA was removed by on-column incubation with DNaseI (Qiagen). A NanoDrop D-1000 Spectrophotometer (Thermo-Fisher Scientific) was then used to assess the amount and quality of the isolated RNA samples. First-strand complementary DNA (cDNA) was synthesized from 2 μg of RNA using the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems). To detect the expression of the CEACAM genes by PCR, specific primers for each target were designed using the on-line primer-BLAST tool of the National Center for Biotechnology Information (NCBI, http://www.ncbi.nlm.nih.gov/ tools/primer-blast/). Possible secondary structures at the primer binding sites were taken into account by characterizing the nucleotide sequence of the regions of interest using the Mfold algorithm [44]. Primer design across different exon boundaries allowed for specific primer pairs targeting 7 isoforms of CEACAM1. Primers were also designed for CEACAM5, CEACAM6, two housekeeping genes, and IFNβ-1a.
PCRs of the cDNAs were carried out on a S1000TM Thermal Cycler (BioRad) in a 25 μl reaction volume containing 0.2 μM primers, 1 U Taq DNA polymerase (5-Prime) and 200 μM dNTPs. Thermal conditions included an initial 95°C denaturation step for 3 min, and then 35 cycles of 10 s at 94°C, 30 s at 60°C and 30 s at 72°C. The resulting PCR products were separated on an ethidium bromide containing agarose gel and visualized under a UV-transiluminator to confirm the expected amplicon size. To verify the identity of amplified PCR fragments of the CEACAM genes and the different splicing variants of CEACAM1, PCR-products were purified and sequenced (Eurofins MWG Operon, Ebersberg, Germany). Sequences were aligned with the corresponding NCBI-reference transcripts using ClustalX [45].
To quantify the relative expression of each gene, we used a CAS-1200 pipetting robot (Qiagen) to set up the qPCRreactions and a Corbett Rotor-Gene 6000 (Qiagen) as Real-Time qPCR apparatus. Each sample was analyzed in duplicate in a total reaction volume of 20 μl containing 10 μl of 2× SensiMix SYBR Master Mix (Bioline) and 0.2 μM of each primer. The cycling conditions included an initial step of 95°C for 10 min followed by 40 cycles of 95°C for 15 s, 60°C for 20 s and 72°C for 20 s. For each experiment, an RT-negative sample was included as a control. Melting curve analysis and size verification by electrophoresis was used to confirm the specificity of the qPCR reactions.
The relative expression of the target genes was analyzed using the Pfaffl method [46]. The expression levels were normalized to the geometric mean of 2 housekeeping genes: hypoxanthine phosphoribosyltransferase1 (HPRT1) and peptidylpropyl isomerase B (PPIB). The stability of the housekeeping genes was assessed using the BestKeeper algorithm [47]. Relative differences in mRNA expression between different experimental conditions were analyzed by pair-wise fixed randomization tests using the REST 2009 software [48].

CEACAM expression in human lung
Expression and co-expression patterns of CEACAM1, CEACAM5 and CEACAM6 in the human lung were analyzed. Paraffin sections of lung tissues from 19 donors who underwent surgery for lung resection to treat lung cancer (Table 1) were stained with an IgG control antibody and antibodies specific for CEACAM1, 5 and 6, as well as with an antibody specific for CEACAM8, which is solely expressed in human granulocytes. Noncancer tissues from these sections were used for analysis. Figure 1 shows representative CEACAM stainings of lung sections for the respective antibodies. Table 2 gives an overview of the number of positive tissues in the specimens. CEACAM1 was expressed in 12 out of 19 lungs, CEACAM5 in 18 out of 19 and CEACAM6 in all 19 lungs examined ( Figure 1 and Table 2). As expected, granulocyte-specific CEACAM8 ( Figure 1D, Table 2) showed strong staining if infiltrated granulocytes appeared (data not shown) and did not give any staining above IgG background in granulocyte-free lung tissues (Table 2).
CEACAM5 displayed frequent and intense staining in bronchial and alveolar epithelium ( Figure 1B, Table 2). The former tissue was always positive when found in the respective specimen (17/17), the latter in 16 out of 19 samples. Also, pleura and blood vessel endothelium were often CEACAM5 positive (12/17 and 15/19, respectively). A weaker and less frequent CEACAM5 expression was found in the adventitia submucosa (11/17).
CEACAM6 showed an abundant and strong staining in the alveolar epithelium of all lungs examined ( Figure 1C, Table 2). The bronchial epithelium of most specimens also displayed a strong CEACAM6 expression (15/18). In some cases, the pleura exhibited a weak CEACAM6 staining (6/ 18). Adventitia submucosa and blood vessel endothelium did not show any CEACAM6 expression.

CEACAM co-expression patterns in human lung tissues
The co-expression patterns of CEACAM1, CEACAM5 and CEACAM6 in the different lung tissues of the 19 Chronic obstructive pulmonary disease (COPD) 8 Hypertension 10 Granulocyte infiltration 9 specimens were also analyzed ( Table 3). In alveolar epithelium, CEACAM6 was present in all specimens; therefore, only CEACAM1 and CEACAM5 expressions varied. CEACAM1 was mostly co-expressed with both, CEACAM5 and CEACAM6. In bronchial epithelium, CEACAM5 was expressed in all specimens; only CEACAM1 and CEACAM6 expression varied. Both proteins were co-expressed with CEACAM5 alone and with each other. The adventitia submucosa was always negative for CEACAM6. CEACAM1 and CEACAM5  Paraffin sections of human lung tissue from surgical specimens of lung cancer patients (N=19) were stained for CEACAM1, CEACAM5, CEACAM6, CEACAM8 and an unspecific control IgG. Figure 1 shows representative pictures of non-cancer tissues for these stainings. CEACAM expression in non-cancer tissues was analyzed for alveolar epithelium, bronchial epithelium, adventitia submucosa, pleura and blood vessel endothelium. Since not all tissue types were found for all specimens, the number of positive specimen and the number of total specimen including the respective tissue are given. *Assessment of total staining includes alveolar epithelium, bronchial epithelium, adventitia submucosa, pleura and blood vessel endothelium and excludes present blood cells due to intrinsic peroxidase activity. § Four lung sections showed enhanced background staining for the control IgG or the anti-CEACAM8 antibody.
were expressed alone or co-expressed. When the pleura was positive for CEACAM6, at least one of the two other CEACAMs were co-expressed. Most of the CEACAM6-negative specimens showed pleura staining for CEACAM5, and some for CEACAM5 and CEACAM1. Blood vessel endothelium was negative for CEACAM6 and showed staining for CEACAM1 and CEACAM5 either alone or together (Table 3). Coexpression patterns showed no correlation to COPD, smoking status and granulocyte infiltration in total staining analysis or in the individual lung tissues (alveolar epithelium, bronchial epithelium, adventitia submucosa, pleura and blood vessel endothelium; data not shown).
CEACAM expression in human lung does not correlate with COPD, smoking status or granulocyte infiltration It was then examined whether the presence or expression level of CEACAM1, CEACAM5 and CEACAM6 were linked to COPD or smoking status. Table 4 shows the relative staining intensity of the three CEACAMs. Total staining of the paraffin sections, including alveolar epithelium, bronchial epithelium, adventitia submucosa, pleura and blood vessel endothelium, was assessed in the range of 0 (no staining) to 3 (strong staining). However, no correlation was found for either the presence (data not shown) or the expression levels (Table 4) of the tested CEACAMs according to two-tailed Student's t-test. Next, the interdependency of CEACAM1, CEACAM5 and CEACAM6 expression with granulocyte infiltration was analyzed, but again no correlation was detected (Table 4). Also, the correlation of CEACAM expression with COPD, smoking status and granulocyte infiltration in the individual lung tissues (alveolar epithelium, bronchial epithelium, adventitia submucosa, pleura and blood vessel endothelium) was examined, but no significant differences were found (data not shown).

Soluble CEACAMs are present in human airways
Next, the presence of soluble CEACAM1, CEACAM5 and CEACAM6 in bronchoalveolar lavage fluid (BALF) was evaluated. Therefore BALF samples were examined by ELISA (Figure 2A). CEACAM1 was found with an average concentration of 10 ng/ml (± 9 ng/ml SD) in 4 out of 181 BALF samples (2.2%). CEACAM5 was present in 78.5% of BALF samples with an average concentration of 11 ng/ml (± 13.6 ng/ml SD) and CEACAM6 in all Paraffin sections of human lung tissue from surgical specimens of lung cancer patients (N=19) were stained for CEACAM1, CEACAM5, and CEACAM6. Figure 1 shows representative pictures for these stainings. Non-cancer tissues were analyzed for the co-expression of CEACAM1, CEACAM5 and CEACAM6. Numbers of specimens with the respective pattern denominated in column 1 are given. *Assessment of total staining includes alveolar epithelium, bronchial epithelium, adventitia submucosa, pleura and blood vessel endothelium, and excludes present blood cells due to intrinsic peroxidase activity. §One or more stainings for CEACAM1, CEACAM5 or CEACAM6 are missing for this tissue in a specimen preventing an evaluation. Paraffin sections of human lung tissue from surgical specimens of lung cancer patients (N=19) were stained for CEACAM1, CEACAM5, CEACAM6, CEACAM8 and an unspecific control IgG. Figure 1 shows representative pictures for these stainings. Relative total stainings (including alveolar epithelium, bronchial epithelium, adventitia submucosa, pleura and blood vessel endothelium) were assessed in the range of 0 (no staining) to 3 (strong staining). Numbers given are mean intensity and standart deviation. No significant differences were found in comparison to the respective control groups according to two-tailed Student's t-test.
BALF samples with at an average concentration of 93 ng/ ml (± 48 ng/ml SD). In order to approach soluble CEACAM1 expression in human lung with an alternative method, normal human bronchial epithelial (NHBE) cells from healthy donors were used for transcription analysis. RT-PCR and subsequent sequencing of the obtained PCRproducts confirmed the expression of three soluble CEACAM1 isoforms, CEACAM1-4C1, CEACAM1-3 and CEACAM1-3C2 ( Figure 2B). However, mRNAs of all three isoforms were not expressed at sufficient levels to allow for reliable quantification analysis by qPCR. To test for expressed and shedded soluble CEACAMs in NHBE cells, CEACAM1-, 5-, and 6-immunoprecipitations from cell culture supernatants were analyzed. Western blot analysis revealed soluble CEACAM6 and low amounts of soluble CEACAM1, but no soluble CEACAM5 ( Figure 2C) in NHBE cell culture supernatants.

CEACAM1 isoforms are up-regulated by interferons and TLR3 agonist poly I:C
In order to investigate the regulation of CEACAMs by cigarette smoke or in the presence of acute inflammation in human lung, NHBE cells were stimulated as described below and the expression of CEACAMs was analyzed using RT-PCR and qPCR. Untreated confluent NHBE cells expressed CEACAM5 and CEACAM6, as well as the CEACAM1 isoforms CEACAM1-4L, CEACAM1-4S, CEACAM1-3L and CEACAM1-3S. This was shown through RT-PCR analysis with pairs of primers specific for each CEACAM and isoforms ( Table 5). Identity of the obtained PCR-products for each CEACAM and isoforms was confirmed by sequencing. The intensity of the resulting bands after electrophoretic separation suggests that both CEACAM1 isoforms bearing the long cytoplasmic domain were expressed to a much lesser extent than the two short CEACAM1 isoforms. To address short-term effects of smoking on epithelial CEACAM expression, NHBE cells were exposed to CSE. To assess the tie of CEACAM expression in human lung cells to innate immune responses, NHBE cells were treated with IFNα, IFNβ, IFNγ, TNFα, and the TLR agonists MALP-2 (TLR2/6), poly I:C (TLR3) and flagellin (TLR5). Then, RNA from these cells was isolated and qPCR analysis was used to check for differences in CEACAM1-4L, CEACAM1-3L, CEACAM1-4S, CEACAM1-3S, CEACAM5 and CEACAM6 mRNA levels. To ensure an accurate normalization of their expression levels, we tested two candidate housekeeping genes for their stability across the unstimulated and stimulated samples. PPIB (SD (±Ct)=0.24; CV(%Ct)=1.24) as well as HPRT1 (SD(±Ct)= 0.34; CV(%Ct)=1.42) showed a highly stable expression, allowing for normalization with the geometric means of both genes.  The four transmembrane CEACAM1-isoforms CEACAM1-4L, CEACAM1-3L, CEACAM1-4S and CEACAM1-3S were basically co-regulated by all reagents that elicited a difference in transcription levels (Table 5, Figure 3B). IFNα significantly upregulated the expression of CEACAM1-4L (3.7-fold, p < 0.05), CEACAM1-4S (4.8-fold, p < 0.05) and CEACAM1-3S (4.3-fold, p < 0.001), whereas the 3.5-fold upregulation of CEACAM1-3L failed to reach significance (p = 0.124). IFNβ and IFNγ significantly increased the expression of all four splicing variants of CEACAM1 (p < 0.05 in all cases). In the case of IFNβ we could observe a slightly higher up-regulation of the two short isoforms CEACAM1-4S (5.2-fold) and CEACAM1-3S (4.7-fold) than the long isoforms CEACAM1-4L (4.4-fold) and CEACAM1-3L (3.0-fold). On the other side, IFNγ seemed to favor the two long isoforms CEACAM1-4L (6.3-fold) and CEACAM1-3L (5.2-fold) when compared to the short variants CEACAM1-4S (4.2-fold) and CEACAM1-3S (3.5-fold). However, no significant alteration in the ratios between long and short CEACAM1 isoforms was observed. Poly I:C treatment also upregulated all four isoforms of CEACAM1 in a significant manner (p < 0.001 in all cases) with expression ratios ranging between 4.9 and 7.0-fold ( Figure 3B). CSE, TNFα, MALP-2 and flagellin had no effect on any of the four CEACAM1 transcripts despite of a significant increase of the IL-8 mRNA levels, which served as positive control of their efficacy (data not shown).
In order to examine whether the positive responses of the CEACAM1 isoforms to poly I:C were caused by the up-regulation of interferons, poly I:C treated NHBE cells were analyzed for IFNα, IFNβ and IFNγ mRNA expression. IFNα and IFNγ mRNAs were not expressed at sufficient levels to allow qPCR analysis (data not shown). However, qPCR showed that poly I:C strongly induced an immediate IFNβ expression (780-fold increase after 4 h) that dissolved after 24 h ( Figure 3E). qPCR showed that IFNγ treatment also significantly increased CEACAM5 mRNA levels in NHBE cells by 4.1-fold ( Figure 3C). No difference in CEACAM5 mRNA levels were induced by CSE, IFNα, IFNβ, TNFα, or any of the TLR agonists. CEACAM6 expression levels in NHBE cells were not altered under any of the tested conditions ( Figure 3D).

Increased CEACAM1 and CEACAM5 mRNA levels lead to increased protein levels on NHBE cell surfaces
Indirect immunofluorescence showed that CEACAM1, CEACAM5 and CEACAM6 were expressed on the cell surface of untreated confluent NHBE cells ( Figure 4A). While CEACAM1 showed a low expression on most cells, CEACAM5 and CEACAM6, respectively, were only expressed by a subpopulation.
It was then tested whether the up-regulation of CEACAM1 transcripts upon treatment with IFNα, IFNβ, IFNγ, and poly I:C were also reflected by higher cell surface protein levels. FACS analysis confirmed an increased expression of CEACAM1 on the cell surface of NHBE cells under all four conditions ( Figure 4B). The CEACAM1-specific antibody used for flow cytometry did not allow discrimination between the different isoforms.
Also, an increased CEACAM5 cell surface expression upon IFNγ treatment could be confirmed by FACS analysis ( Figure 4C). Flow cytometry also verified that there was only a small subpopulation of NHBE cells that were CEACAM5 positive in untreated control cells. Upon IFNγ treatment, both, the number of CEACAM5positive cells and the expression level of CEACAM5 on the cell surface, were increased ( Figure 4C). However, it remained only a defined portion of NHBE cells that expressed CEACAM5. Structure of forward and reverse primers (5′-3′) of all target genes and isoforms and the expected size of their respective amplified PCR-products.

Non-typable Haemophilus influenzae (NTHi) and Moraxella catarrhalis up-regulate CEACAM1 expression
Next, the effects of acute NTHi and Moraxella catarrhalis infection on CEACAM1, CEACAM5 and CEACAM6 mRNA expression levels in NHBE cells were investigated ( Figure 5). qPCR analysis revealed no differences in CEACAM5 and CEACAM6 expression upon bacterial infection. The M. catarrhalis wild type strains 25238 and BBH18 as well as the NTHi wild type strains 2019 and 1128 enhanced the mRNA expression of all four transmembrane CEACAM1-isoforms to a similar degree in a co-regulatory manner ( Figure 5A,B,D,E). The mean induction of CEACAM1 transcription by M. catarrhalis strains was twice as high as by NTHi strains (3.5-5.5 fold vs. 1.9-2.8 fold). Since all four pathogens can bind to CEACAM1, we next tested whether this interaction was essential to the up-regulation of CEACAM1. To that end we used the M. catarrhalis UspA deletion mutant BBH18.1 and the NTHi P5 deletion mutant 1128f-, which both lack the respective CEACAM1-binding adhesin ( Figure 5C,F). Again, the infection with these strains induced an elevated CEACAM1 expression (4.0-4.9 fold and 1.9-2.4 fold, respectively) comparable to their parental strains, indicating a CEACAM1-independent, more general mechanism for this effect. We then tested whether the CEACAM1 upregulation might be due to an increase in interferons. Both

Discussion
Here we present the first comprehensive study based on immunohistochemistry demonstrating that CEACAM1, CEACAM5, and CEACAM6 are frequently co-expressed in several tissues of the human lung, including epithelia of the airways and alveoli. CEACAM expression was not connected to COPD, smoking status and granulocyte infiltration ( Figure 1, Tables 3 and 4). Despite the analysis of non-cancer tissues from the specimen, the fact that the lung sections used for immunohistochemical analysis were from patients that underwent lung resection to treat lung cancer might conceal a regulatory effect of COPD or smoking status on CEACAM expression, since CEACAM1, CEACAM5, and CEACAM6 have all been shown to be up-regulated in lung cancer [49][50][51][52][53]. Also, the inflammatory processes associated with cancers of the lung might have had an effect on the expression levels of the CEACAMs. For example, as discussed below, IFNγ can up-regulate CEACAMs 1, 5, and 6. Importantly, we show that the COPD-associated pathogens M. catarrhalis and NTHi can also upregulate CEACAM1 expression independent of their ability to bind to CEACAM1. The up-regulation by M. catarrhalis might be at least in part due to the induction of IFNβ production ( Figure 5G) and is in accordance with the observation of CEACAM1 up-regulation by pathogenic Neisseria in endothelial and epithelial cells [54,55]. CEACAM5 and CEACAM6 expression levels were not affected by M. catarrhalis and NTHi. However, both receptors were already expressed at high levels and in all specimens (Figure 1, Tables 2 and 3). Regarding colonization, the increase in CEACAM1 on epithelial cells upon bacterial challenge is likely to increase bacterial adherence and infection. This interaction would be facilitated by an impaired mucocilliary clearance, which is associated with later phases of COPD. Thus, CEACAM receptors might be at least in part responsible for the colonization of the lower airways by M. catarrhalis and NTHi in COPD patients. This setting would also explain the association of bacterial colonization with progressive/ advanced disease despite the fact that we did not find a regulation of CEACAM expression by the chronic pathologic conditions COPD, smoking status and granulocyte infiltration (Figure 1, Tables 3 and 4). Further studies including identification of pathogens in the lower airways will be necessary to shed light on this aspect of CEACAM-pathogen interactions. It will be also important to take into account that an up-regulation of CEACAMs might be transient either due to temporary qualities of chronic/persistent disease or due to pathogen characteristics at distinct time points during the infection process. It will also be interesting to test human lung tissues for the presence of the other CEACAMs expressed on epithelial cells (CEACAM7, 18, 20 and 21 [56,57]) and to analyze their ability to interact with pathogens.
The up-regulation of membrane-bound CEACAM receptors might be necessary to counteract the presence of soluble receptor forms that might act as decoys to prevent bacterial/viral infections or immune evasion. Even though bacteria possess redundant targeting mechanisms, and bacterial adhesins often work in a sequential manner, Hill et al. showed that perturbing CEACAM1-, CEACAM5-and CEACAM6-based adhesion by using a CEACAM-binding peptide prevents host cell binding by M. catarrhalis, H. influenzae, N. meningitidis and N. gonorrhoeae in a cell culture model [7]. The amounts of soluble CEACAMs, with high concentrations of soluble CEACAM6 in all and lower concentrations of soluble CEACAM5 in 78.5% of the bronchoalveolar lavage fluid samples is mirrored in the expression levels of their membrane-bound counterparts in lung tissues (Figures 1,  2, 3, 4 and [58][59][60]). The fact that soluble CEACAM1 is all but absent in BALF (98% negative BALF samples) increases the importance of this receptor for colonization by pathogens despite its comparatively low expression levels in human lung.
In NHBE cells, CEACAM1 mRNA and protein levels were also increased by IFNα, IFNβ, IFNγ and the TLR3 agonist poly I:C ( Figure 3B). All membrane-bound CEACAM1 isoforms examined here were basically coregulated under all conditions that affected their expression level (Figures 3 and 5). However, some differences were found for the differential up-regulation of the CEACAM1 isoforms. Type I interferons (IFNα and IFNβ) favored the two short isoforms while the type II interferon IFNγ favored the long isoforms, and poly I:C favored the long and the short isoforms bearing 4 extracellular domains. Especially the former differences, even though not significant, are interesting with regard to the fact that in order to interfere with TLR signaling or with T-cell activation, binding of the Shp-1 phosphatase to the immunoreceptor tyrosine-based inhibitory motif (ITIM) within the cytoplasmic domain of the long CEACAM1 isoform was necessary [15,16,19]. The two tyrosine residues that are part of the ITIM can bind to protein tyrosine kinases (e.g. c-Src) as well as protein tyrosine phosphatases (e.g. Shp-2) -leading to the stimulation or the inhibition of cell-signaling pathways, respectively. Long and short CEACAM1 isoforms can appear as  monomers, dimers and oligomeric microclusters in the membrane [26,28]. Trans-homophilic binding between different CEACAM1 molecules increases cis-dimerization in the plane of the membrane via an allostery-based mechanism. Binding of CEACAM1-L to Shp-1 or c-Src is dependent on the balance between the oligomeric states as well as the ratio of long and short isoforms [26,28]. Interestingly, no cytoplasmic domain is necessary for the CEACAM1-mediated internalization of H. influenzae, M. catarrhalis and N. gonorrhea [27]. But bacterial engagement of CEACAM1 might, in addition to procuring adherence, also influence the cis-dimerization and subsequent inhibitory or activatory signaling of CEACAM1, either supporting pathogen colonization or host response. For example, the interaction of CEACAM1 with M. catarrhalis and N. meningitidis proteins resulted in reduced Tolllike receptor (TLR) 2-initiated inflammatory responses (the major mediator of M. catarrhalis-induced immune responses) of NHBE cells [16,61,62]. Part of the host response to microbial infection is the production of cytokines [63]. Type I interferons IFNα and IFNβ can be produced by most cell types and are the major effector cytokines of the host immune response against viral infections. However, the production of type I interferons is also induced in response to bacterial infections [64]. The type II interferon IFNγ plays an important role during the immune response to bacterial pathogens, but is also induced upon infection with viruses. It is produced predominantly by natural killer (NK) cells and natural killer T cells as part of the innate immune response and by Th1 CD4+ and CD8+ cytotoxic T lymphocyte effector T cells once antigen-specific immunity develops [65]. In the present study CEACAM5 is up-regulated in NHBE cells by IFNγ, but not by type I interferons, and CEACAM1 is increased by type I and type II interferons (IFNβ and IFNγ, Figure 3B and D). While for CEACAM1, CEACAM5 and CEACAM6 an upregulation by IFNγ and viral infections has been described in epithelial cells [55,[66][67][68][69][70], to our knowledge this is the first report that also type I interferons enhance CEACAM1 expression.
A temporal association between bacterial and viral infections is often observed in the human upper respiratory tract [71,72]. Infection by opportunistic colonizers, including Haemophilus influenzae and Moraxella catarrhalis, increases considerably following influenza and/ or respiratory syncytial virus (RSV) infections [71,72]. COPD is often associated with viral infections, mostly by rhinovirus, and it was recently shown that these infections indeed precipitate secondary bacterial infections, particularly in COPD patients [73].
Importantly, the present study shows that in addition to the primarily viral induced type I interferons, TLR3 agonist poly I:C also increased CEACAM1 expression ( Figure 3B). TLR3 plays a key role in anti-viral immune responses and recognizes synthetic dsRNA like poly I:C and virus derived dsRNA contained in cells infected by positive-stranded RNA viruses and DNA viruses [74,75]. It was shown recently that poly I:C enhances the susceptibility to secondary pulmonary infections by gram-positive bacteria in a mouse model [76]. The positive-stranded RNA virus rhinovirus enhances CEACAM5 expression in human nasal epithelial cells and two negative-stranded RNA viruses, respiratory syncytial virus (RSV) and human parainfluenza virus 3 (HPIV-3), enhance CEACAM1 expression in A549 and NHBE cells [67,77]. Since the latter viruses have to be recognized via TLR7 or TLR8, this is the first indication that the pathogen receptor CEACAM1 might also be up-regulated by positive-stranded RNA viruses via TLR3.
CEACAM6 expression was not altered by any agent used in this investigation ( Figure 3D). This was probably due to its initially high expression level in the NHBE cells. Fahlgren et al. showed that the LS174T cell line that expressed high levels of CEACAM5 and CEACAM6 before IFNγ treatment did not show any enhanced expression after IFNγ exposure while IFNγ up-regulated both mRNAs in two cell lines, HT-29 and T84, that initially expressed low levels of CEACAM5 and CEACAM6 [66]. However, the very high and constant CEACAM6 expression in NHBE cells as well as in the lung specimens is in accordance with its proposed role as a surfactant [58,78].
The up-regulation via viruses and the effects of the inflammatory cytokine IFNγ imply a more general role for CEACAM1 and CEACAM5 in the inflammatory response to infection, and also in the spatial and temporal association between bacterial and viral infections. One of the underlying mechanisms for the colonization of COPD patients by M. catarrhalis and NTHi may consist of the up-regulation of specific host receptors, i.e. CEACAM1, by viral infections, as described for several CEA family receptors [55,66,68,77]. For pathogenic Neisseria it was demonstrated that increased CEACAM expression levels correlated with an increase in bacterial invasion [55,79]. Once established, M. catarrhalis and NTHi themselves are also able to increase the expression of their receptor CEACAM1. Whether the starting point is a viral or bacterial pathogen, the result is a continual cycle of infection-induced increase in cytokine levels and, subsequently, receptor expression which then promotes bacterial invasion. Since viruses also recruit CEACAMs, elevated CEACAM expression might aid viral infection as well [80,81].

Conclusions
CEACAM1 is a well-established receptor for bacteria, including the human pathogens M. catarrhalis and H.
influenza that both colonize about one third of all COPD patients and cause acute exacerbations. While we did not find a direct correlation between CEACAM expression and COPD, these COPD-linked bacteria were able to increase the expression of their own receptor on host cells. We also propose a role for the constitutively expressed CEACAM6 as well as CEACAM1 and CEACAM5 in the phenomenon of increased host susceptibility to bacterial infection upon viral challenge in the human respiratory tract, as it occours for example in COPD patients, since either enhanced CEACAM1 expression induced by viruses or the COPD-associated changes in the function of the mucociliary clearance that will allow an easy access to CEACAMs on respiratory epithelia are likely to increase bacterial colonization.