The CD14 (−159 C/T) SNP is associated with sCD14 levels and allergic asthma, but not with CD14 expression on monocytes

LPS-ligation to CD14/TLR-4 on monocytes/macrophages triggers the production of IL-12-family cytokines. IL12/18 promote TH1-differentiation, counteracting the TH2-driven asthma. Therefore, CD14 modulation could alter the TH2-differentiation and should be taken into account when studying asthma. To analyse the alteration in CD14 levels and its association with CD14 (−159 C/T) SNP (rs2569190) in Caucasian adults with stable allergic asthma, we performed a cross-sectional study (277 healthy subjects vs. 277 patients) where clinical parameters, CD14 values and the CD14 (−159 C/T) SNP were studied. Apart from typical biomarkers, we found an increment of neuron-specific enolase (NSE) in allergic asthma, probably linked to monocyte activity. Indeed, we evidenced increased monocyte numbers, but lower CD14 expression and normalised sCD14 values in patients. Moreover, we noticed an association of the T allele (P = 0.0162) and TT genotype (P = 0.0196) of the CD14 SNP with a decreased risk of allergic asthma and augmented sCD14 levels. In conclusion, monocyte CD14 expression and normalized sCD14 values were reduced in stable state asthmatics, and this could be related to the presence of an expanded CD14low monocyte subset. This study also demonstrates that the CD14 (−159 C/T) polymorphism is a risk factor for moderate-severe allergic asthma in adult Caucasians.


Results
Demographic and clinical characteristics of the study population. In this work a case-control study was performed, where adult patients with both intermittent-mild and moderate-severe allergic asthma were recruited. These patients were in a clinically stable state, had a well controlled disease, and the majority were non-smokers under treatment with inhaled corticosteroids (Table 1). FEV1 (%) and FEV1/FVC ratio (%) values are described in Table 1, showing decreased levels in MSAA (moderate-severe allergic asthmatics) compared to IMAA (intermittent-mild allergic asthmatics) ( Table 1). Although 66.4% of AA came from rural areas, most of them had no animals or just dogs/cats; only 11% were taking care of farm animals.
According to their allergic disease state, patients had significantly augmented levels of eosinophils and total IgE compared to HC, but there were no significant changes regarding disease severity (Table 1). IgE showed a positive correlation with eosinophil, monocyte and, to a lesser extent, lymphocyte blood count, underlying the relevance of these subsets in allergic asthma pathogenesis ( Table 2). Since activation of eosinophils and macrophages has been associated to enhanced NSE levels under some pathological conditions 26,27 , we also undertook the measurement of this enzyme in serum samples. As Table 2 shows, it was found a positive correlation of IgE, eosinophils and monocytes with NSE (Table 2), which prompted us to examine the utility of this parameter as an additional marker of allergic asthma. As shown in Fig. 1a, NSE levels were augmented in AA compared to HC and tend to be around 14.6% higher in men than in women (Fig. 1a). Moreover, the area under the curve (AUC) of the ROC plot for NSE levels was close to AUC of total IgE and higher than AUC of blood eosinophils (absolute values) (Fig. 1b).
No alterations were appreciated for C-reactive protein (CRP), IgG, IgA, IgM, or tumour necrosis factor (TNF) in AA. The influence of age was also taken into consideration, with a negative correlation with FEV1%, FEV1/ FVC, IgE and leukocyte subtypes (mostly eosinophils), but a small positive association with TNF, CRP, mCD14 and sCD14 (Table 2).
Allergic asthma enhances the number of peripheral blood monocytes, but causes a reduction of mCD14 in these cells. Complete blood count revealed an increased number of leukocytes in AA (HC 6.44 (5.39-7.75) × 10 3 cells/μL vs. AA 6.93 (6.00-8.23) × 10 3 cells/μL; P < 0.001). These differences were partially dependent on some innate subsets like eosinophils and monocytes, but not neutrophils or basophils (Table 1). Moreover, men had higher numbers of monocytes in peripheral blood than women (P < 0.0001), and asthma caused a slight but significant elevation of monocytes in both IMAA and MSAA (Fig. 2a). mCD14 was mainly expressed by monocytes (>90% CD14 + ) (Supplementary Figure 1), and asthma produced a decrease in the percentage of CD14 + monocytes (and mean fluorescence intensity/MFI values), without changes between IMAA and MSAA ( Fig. 2b and c). Indeed, when a ROC curve was constructed for mCD14 values (%), the AUC gave a value of 0.7881 (whole asthmatics), which underlines the potential use of this parameter as an allergic asthma marker (Fig. 2d), even after segregating the patients in IMAA and MSAA.
Apart from monocytes, neutrophils and lymphocytes also contain CD14 + cells, but the expression of this marker is much lower than in monocytes (Supplementary Figure 1). Because of this, we assessed the differences between neutrophils and lymphocytes-associated mCD14 levels in AA and HC, but they were not influenced by the presence of allergic asthma (data not shown).
Reduction of relative but not absolute sCD14 levels in allergic asthma. Despite the decreased levels of mCD14 in monocytes from AA, we did not reach significant differences in the serum concentration of sCD14 in IMAA, MSAA and HC (data not shown). Moreover, when a cut-off value of total IgE (80 IU/mL) was selected to maximize the true positive rate (70.76%) and minimize the false positive rate (15.88%) (Fig. 1b), sCD14 levels remained unaltered in donors with IgE < 80 IU/mL or ≥80 IU/mL.
As our results were supporting the presence of altered numbers of monocytes in asthmatic patients, and monocytes were the most likely source of sCD14, we evaluated the correlation between the absolute count of peripheral blood monocytes and the serum concentration of sCD14, with negative results (Table 2). In spite of this, sCD14 levels could still be influenced by the combined effects of increased monocyte count and decreased SCIENtIFIC RepoRts | (2018) 8:4147 | DOI:10.1038/s41598-018-20483-1 expression of mCD14 on this leukocyte subset (phenotype). Accordingly, sCD14 values were normalised with respect to the absolute number of circulating monocytes to maximize the effect of cell phenotype, finding in this case a significant reduction of sCD14 in AA compared to HC even after segregating by disease severity (Fig. 2e). We also detected higher levels of relative sCD14 in women compared to men (data not shown).
We next examined the mCD14-sCD14 correlation. In order to analyse the strength of this relationship, we have taken into account the absolute and normalized values of sCD14, as well as we have analysed the data as a whole or after segregating our results according to sex or donor group (healthy or diseased). However, we have found no connection between both parameters ( Table 2, whole set of mCD14-sCD14 pairs), which underlines an independent regulation of both molecules. Apart from the monocyte count, alternative reasons might explain this low correlation, like other sCD14 sources (e.g. hepatocytes) or the existence of different genetic backgrounds. Regarding the first possibility, there were no differences in CRP or TNF between HC and AA (P > 0.05), or evidence of respiratory infection, even though there is very small correlation between CRP or TNF and sCD14 (Table 2). Therefore, we next analysed the influence of the CD14 (−159 C/T) SNP genotype (rs2569190) on mCD14/sCD14 levels.
Association between CD14 (−159 C/T) SNP and allergic asthma risk. A summary of allergic asthma risk according to the different genetic models is presented in Table 4. We found an association between T allele and a decreased allergic asthma risk in the overall allergic asthma population (T vs. C: OR = 0.74, 95% CI = 0.59-0.94, P = 0.0162). Moreover, when AA were subset by disease severity (IMAA and MSAA vs. HC), this association was maintained only for MSAA (OR = 0.70, 95% CI = 0.54-0.92, P = 0.0136), while IMAA lost the allelic association (Table 4).
Regarding to the other genetic models, a significant association of CD14 (−159 C/T) and allergic asthma risk was found between the homozygotes TT vs. CC (OR = 0.55, 95% CI = 0.34-0.89, P = 0.0196), and almost reached significance in a dominant model (TT + TC vs. CC: OR = 0.67, 95% CI = 0.45-0.98, P = 0.0511) ( Table 4). Furthermore, after segregating by disease severity, it was found an association of allergic asthma risk with this SNP only in MSAA according to a recessive model (TT vs. TC + CC: OR = 0.61, 95% CI = 0.38-0.96, P = 0.0442) or TT vs. CC genotype comparisons (OR = 0.49, 95% CI = 0.28-0.85, P = 0.0163). Therefore, the results suggested that T allele and TT homozygote individuals have decreased risk of allergic asthma compared with C allele and CC homozygote carriers, respectively.
The influence of the CD14 (−159 C/T) SNP on CD14 levels. Although we have shown augmented peripheral blood monocyte count and decreased levels of mCD14 (Fig. 2a-c), the number of monocytes was not influenced by the CD14 (−159 C/T) SNP genotype. We also failed to detect any change in mCD14 related to the SNP genotype (data not shown). In contrast (and regardless of whether they belong to the control group or to the asthmatic population), CC genotypes and to a lesser extent TC genotypes had lower concentrations of sCD14 (absolute values) than TT carriers, while no significant differences were observed between TC and CC subjects (Fig. 3a). This association was maintained between TT and CC carriers when sCD14 levels where normalised by the absolute count of monocytes (Fig. 3b). Therefore, this SNP could be partially responsible for the reduction of normalised sCD14 levels in AA and influence the severity of this disease (Fig. 2e).  Table 2. Spearman correlation matrix of the study population. CRP, C-reactive protein; IgE, immunoglobulin E; NSE, neuron specific enolase; P, patient population; TNF, tumour necrosis factor. *P < 0.05, **P < 0.01, ***P < 0.001. # % of CD14 + monocytes.

Discussion
In the current study, we report data supporting an increase of NSE and monocytes and a down-modulation of their mCD14 expression in allergic asthma regardless of disease severity. In addition, we detect a decrease of  normalised sCD14 values in serum samples from asthmatics, suggesting the expansion of a CD14 low monocyte subset and the influence of the CD14 (−159 C/T) SNP genotype. Indeed, we evidence an association of the T allele and TT genotype of CD14 (−159 C/T) polymorphism with reduced risk of moderate-severe allergic asthma. AA in our study have an atopic disease with eosinophilia, monocytosis and elevated levels of IgE and NSE. NSE is the neuronal isomer of the glycolytic enzyme 2-phospho-D-glycerate hydrolase, and a typical biomarker of small cell lung cancer 28 . Nevertheless, changes in non-malignant inflammatory lung diseases have also been found 27,[29][30][31][32] , since this enzyme can translocate towards the cell surface upon proper stimulatory signals to enhance a proinflammatory response 33 . Our results support the presence of higher NSE levels in men, as previously reported by Collazos et al. 29 , but contrary to this work our asthma patient cohort displays above-normal NSE levels in serum. Monocytes/macrophages appear to be a possible source of NSE 27,33 , and increased numbers of monocytes as well as a correlation of them with NSE levels were demonstrated in our study. However, other plausible sources of NSE are eosinophils or injured epithelial cells during pulmonary infiltration 26 , while the neuronal distress or hypoxia occurring in the lung also could play a role during the disease 29 .
Apart from other cells as T and B lymphocytes, eosinophils, basophils or neutrophils, monocytes are gaining importance as regulators of inflammation in asthma and as key players in the pathogenesis [34][35][36] . Our results show the expansion of this subset in AA regardless the severity of the symptoms (i.e., IMAA and MSAA), as well as a reduction of mCD14, a monocyte marker associated to asthma 3 . In contrast, some authors have described no differences in the staining for mCD14 37,38 , even though these could be the result of a low statistical sample size. Reduced levels of mCD14 or TLR4 in AA makes biological sense 39 , as signal transduction through CD14/TLR4 leads to IL-12 secretion, a powerful inducer of TH 1 polarization 13 . Therefore, attenuated mCD14 levels on antigen presenting cells (APCs) could favour TH 2 -driven allergic asthma 4,5,13,16 . This diminished number of mCD14 molecules on monocytes could arise as a result of several, and not mutually exclusive, mechanisms: (a) altered transcription/translation rates affecting protein abundance; (b) expansion of CD14 low monocyte subsets; (c) a vesicle-or enzymatic-mediated mechanism that release mCD14 from monocytes and should also affect sCD14 concentration.
The degree of mCD14 down-modulation on monocytes suggests the preferential expansion of a small CD14 low subset and not a globally altered transcription/translation rate. This, for example, is in line with the increased percentage of CD14 −/low monocytes upon in vitro culture in the presence of TSLP, an important cytokine in allergic asthma 40 . Monocytes are heterogeneous, with major (CD14 high ) and minor (CD14 low ) subsets 32 . CD14 high ("classical") monocytes display a CD16/FcγRIIIphenotype, while the less frequent CD16 + subset consists of both intermediate (CD14 high CD16 + ) and non-classical (CD14 low CD16 + ) subpopulations 33 . CD16 + monocytes, particularly the intermediate subset, are expanded in inflammation, severe asthma or upon allergen challenge 33,35 , in line with their pro-inflammatory nature 41 . A major constraint of our study is that we have not analysed CD16, but our results show a significant down-modulation of mCD14 in allergic asthma, which appears to rather support the expansion of CD14 low CD16 + monocytes. These cells (non-classical subset) express high levels of CD80, CD86, and CD163, suggesting a high antigen presenting capability 42 . Furthermore, non-classical monocytes are in an advanced differentiation stage and they evidence high invading ability to infiltrate and differentiate into M2-type macrophages 43 , a subset related to allergic inflammation 44 .
CD14 can be released to medium from hepatocytes as an acute phase protein 45 . Although we saw a small correlation between sCD14 and CRP or TNF, the levels of these two last molecules had no changes between AA and HC, and our patients were in a steady-state of the disease. Excluding the hepatocyte origin, monocytes are the most likely cell source of sCD14. Down-modulation of mCD14 in monocytes from AA does not fit with either where there is no relationship between monocyte count and sCD14 or mCD14-sCD14 correlation 16 . Therefore, our results only make sense considering a puzzling scenery with an elevation of monocyte numbers and enhanced frequencies of both CD14 high 34 , but also CD14 low (our results) subsets of monocytes in asthma. Indeed, in our study only normalised serum levels of sCD14 were significantly reduced in patients. In agreement, sCD14 levels have been inversely correlated with IL-4-production 11 , total IgE 11 , or asthma severity 46 . However, some authors have detected higher levels 12 or no differences 7,15 of baseline sCD14 in peripheral blood from asthmatics. Hence, we cannot rule out the contribution of many potential confounding factors that explain these different results, like gene-gene or gene-environment interactions 11,15 . One of the most studied CD14 polymorphisms in asthma is the CD14 (−159 C/T) SNP (rs2569190) 11 . Previous studies investigating the association of this SNP with allergic asthma yielded variable results regarding to the strength and direction of association 3,17,24 . These contradictory results can be explained by differences in ethnicity, low sample size, age of patients or gene-environment interaction 21,25 . We performed our study in a well-defined population (Caucasian, adults, allergic asthmatics and mostly non-farmers), with a high sample size (277 AA vs. 277 HC), and two different disease severity grades (IMAA and MSAA). In agreement with others 11,17 , we show an association of the frequency of the C allele and the CC genotype with allergic asthma (whole asthmatics). More interesting, this association is also related to the disease severity, as it is only maintained in MSAA and, within this group, in severe asthmatics. Moreover, the risk of having moderate-severe allergic asthma (but not intermittent-mild asthma) is lower in carriers of the T allele (T vs. C) and TT genotype, following either a recessive model (TT vs. TC + CC) or after comparing TT vs. CC homozygotes. As other works have shown 5,11,12,18,19 , we evidence augmented sCD14 levels in subjects carrying the TT genotype but no association of this polymorphism with mCD14 levels on monocytes. This suggests an adverse role for the C allele, the CC genotype and the presence of low levels of sCD14/mCD14 in allergic asthma or atopy 11,46 , especially among adult and atopic subjects exposed to low levels of endotoxin, like our cohort 13 .
In summary, our findings show an increment in the serum levels of NSE, which could be used as a novel biomarker of allergic asthma. On the other hand, we also found a decrease in the expression of CD14 on monocytes from allergic asthmatic patients, probably related to an increase of CD14 low monocyte subset. Moreover, we evidence an association of the (−159 C/T) SNP in the CD14 promoter with allergic asthma, and a decreased risk of having moderate-severe allergic asthma in carriers of T allele and TT genotype. Furthermore, TT genotype is associated to higher levels of sCD14, pointing out a protective role for the T allele in this disease.

Methods
Subjects. The study population was recruited from January 2009 to December 2012, at the Unit of Pneumology and Allergy of the USC University Hospital Complex of Santiago de Compostela (CHUS) and the Pontevedra Hospital Complex (CHOP). This study population included 277 healthy controls (HC) and 277 allergic asthmatics (AA), consisting of 108 intermittent-mild allergic asthmatics (IMAA) and 169 moderate-severe allergic asthmatics (MSAA). Asthma and allergy diagnosis was confirmed according to Global Strategy for Asthma Management and Prevention criteria (GINA 2006, http://www.seicap.es/documentos/archivos/ GINA2006general.pdf). All patients were in a stable phase for at least 4 weeks before the study initiation. Forced vital capacity (FVC), forced expiratory volume in 1 second (FEV1), and the FEV1/FVC ratio, were measured. Asthma diagnosis was confirmed by a positive broncodilatator test (>12% of FEV1 change after salbutamol) or metacholine challenge. Allergic sensitization was confirmed through a skin prick test or serum IgE specific to frequent allergens. Other variables were measured: smoking, pets at home, residence (rural/urban), profession, comorbidities, age of symptoms onset, asthma control, or number of visits to emergency units, family doctors or  The iPLEX GOLD reactions were set up in a final 9 µL volume and contained 0.222 × iPLEX buffer Plus, 0.222 × iPLEX termination mix and iPLEX enzyme (1.35 U/reaction). A SBE primer mix was made to give a final concentration of each primer between 0.52 µM and 1.57 µM (Metabion International AG, Germany). The thermal cycling conditions for the reaction included an initial denaturation step at 95 °C for 30 seconds, followed by 40 cycles of 95 °C for 5 seconds, with an internal 5 cycles loop at 52 °C for 5 seconds and 80 °C for 5 seconds, followed by a final extension step of 72 °C for 3 minutes. The next step was to desalt the iPLEX Gold reaction products with Clean Resin following the manufacturer's protocol. The desalted products were dispensed onto a 384 Spectrochip II using an RS1000 Nanodispenser and spectra were acquired using the MA4 mass spectrometer, followed by manual inspection of spectra by trained personnel using MassARRAY Typer software, version 4.0. All assays were performed in 384-well plates, including negative controls and a trio of Coriell samples (Na10830, Na10831 and Na12147) for quality control. 10% random samples were tested in duplicate and the reproducibility was 100%.
Statistics. Descriptive data are presented as either median (interquartile range; IQR1-3) or percentages. To assess the significance of changes between AA and HC, Mann-Whitney U two-tailed test or Kruskal-Wallis One Way Analysis of Variance on Ranks followed by Dunn's multiple comparison test were used. Receiver Operating Characteristic (ROC) curves and Spearman's association test were also employed, while differences in proportions were assayed by the χ 2 test. To evaluate the association between CD14 (−159 C/T) SNP and the risk of allergic asthma, Odd's ratios (ORs) and 95% confidence intervals (CIs) were calculated according to different models: TT + TC vs. CC (dominant model), TT vs. TC + CC (recessive model), TT vs. CC, TC vs. CC, and T vs. C (allelic model). Hardy-Weinberg equilibrium (HWE) was calculated by using Pearson χ 2 test. All analyses and graphs were conducted using GraphPad Prism version 6.0 (GraphPad Software, Inc., San Jose, California, USA). Data are presented in box and whisker plots, where median, 25 and 75 quartiles, 5-95 percentiles (error bars) and anomalous values are shown. The statistical signification was defined as P < 0.05.