Cigarette smoking differentially affects immunoglobulin class levels in serum and saliva: An investigation and review

The aim of the present study was to compare concentrations of IgG, IgA, IgM and IgD in both serum and saliva samples from smoking and non‐smoking individuals using a protein microarray assay. The findings were also compared to previous studies. Serum and saliva were collected from 48 smoking male individuals and 48 age‐matched never‐smoker male individuals. The protein microarray assays for detection of human IgG, IgM, IgA and IgD were established and optimized using Ig class‐specific affinity‐purified goat anti‐human Ig‐Fc capture antibodies and horseradish peroxidase (HRP)‐conjugated goat anti‐human Ig‐Fc detection antibodies. The Ig class specificity of the microarray assays was verified, and the optimal dilutions of serum and saliva samples were determined for quantification of Ig levels against standard curves. We found that smoking is associated with reduced IgG concentrations and enhanced IgA concentrations in both serum and saliva. By contrast, smoking differentially affected IgM concentrations—causing increased concentrations in serum, but decreased concentrations in saliva. Smoking was associated with decreased IgD concentrations in serum and did not have a significant effect on the very low IgD concentrations in saliva. Thus, cigarette smoking differentially affects the levels of Ig classes systemically and in the oral mucosa. Although there is variation between the results of different published studies, there is a consensus that smokers have significantly reduced levels of IgG in both serum and saliva. A functional antibody deficiency associated with smoking may compromise the body's response to infection and result in a predisposition to the development of autoimmunity.


| 475
TARBIAH eT Al. non-smokers, Igs have been quantified by ELISA, 15 immunodiffusion assay, 16 nephelometry, [17][18][19] radioimmunoassay 20 and turbidimetry. 21,22 Investigations of Ig levels in the saliva of smokers and non-smokers have been carried out using mainly ELISA, [23][24][25][26][27][28][29] although immunodiffusion assays [30][31][32] or turbidimetry 33 have also been used. These studies have generated varying results concerning the effects of smoking on the concentrations of different Ig classes (IgG, IgA, IgM) in serum or saliva, reporting that smoking is associated with increased, decreased or unchanged Ig class concentrations compared to non-smoking controls. However, these studies have not investigated simultaneously the effects of smoking on Ig class concentrations in both serum and saliva in the same individuals. In addition, few of these studies have investigated the effects of smoking on levels of IgD 20,22 : this is expressed particularly in the upper aerodigestive tract in humans, 34 which is heavily exposed to CS in smokers.
The aim of the present study was therefore to compare concentrations of IgD, as well as IgG, IgA and IgM, in both serum and saliva samples from smoking and non-smoking individuals. Furthermore, this was undertaken using a protein microarray assay to quantify the Ig class levels-this relatively recent technique was not used in any of the previous studies referred to above. [15][16][17][18][19][20][21][22][23][24][25][26][27][28][29][30][31][32][33] The protein microarray used was essentially a miniaturized version of a sandwich/capture ELISA. However, in comparison with ELISAs, microarrays require smaller sample volumes, are more sensitive and have a greater dynamic range. These factors make microarrays a cheaper and potentially more sensitive alternative to ELISAs for the large-scale detection of known proteins, including Igs. 35

| MATERIALS AND METHODS
The study was conducted in accordance with the Basic & Clinical Pharmacology & Toxicology policy for experimental and clinical studies. 36

| Subjects
Serum and saliva were collected from 48 smoking male individuals (median age = 30 years; age range 19-51) and 48 age-matched never-smoker male individuals (median age = 32.5 years; age range 20-56), all of whom were blood bank volunteers at King Abdulaziz Hospital, Jeddah, Saudi Arabia. As healthy blood bank volunteer donors at the time of sample collection, all the individuals fulfilled the following criteria: no history or evidence of significant illnesses/ diseases, including lung/respiratory tract, heart/chest, liver, kidney, nervous system, diabetes and cancer; no infectious diseases or recent vaccinations, fever, flu-like illness, sore throat or antibiotic treatment; no recent dental extractions; no recent surgery, blood transfusion or organ transplantation; and no recent anaemia, high blood pressure or abnormal pulse.
The serum and saliva samples were aliquoted and stored at −20°C. They were transported on ice to Queen's Medical Centre, Nottingham, UK, and then stored frozen until assayed. Ethical approval was granted by The Biomedical Research Ethics Committee, Ministry of Higher Education, King Abdulaziz University, Faculty of Medicine. The smoking individuals smoked a median of 20 cigarettes/d (range 1-40) and had smoked for a median of 10 years (range 2-33).
Donors ceased from eating and drinking (and smoking) about 1 hour prior to donation. Smoker and non-smoker donors were requested to provide 10 mL of blood and 5 mL of unstimulated saliva after signing an informed consent form. Blood samples were collected in plain serum separation tubes (BD Vacutainer™ Venous Blood Collection Tubes: SST™ Serum Separation Tubes: Hemogard). The tubes were centrifuged at 2000 × g for 10 minutes. Saliva was collected over a period of 5-10 minutes. After cleaning the mouth with mouthwash and rinsing with water over 5 minutes, donors provided saliva into plain separation tubes (Thermo Fisher Scientific) and the tubes kept on ice. Then, the saliva samples were centrifuged at 2600 × g for 15 minutes at 4°C. Samples were aliquoted and immediately frozen at −20°C.

| Human Ig class microarrays
Affinity-purified goat anti-human Ig-Fc capture antibodies and horseradish peroxidase (HRP)-conjugated goat antihuman Ig-Fc detection antibodies provided in the Bethyl Laboratories Human IgG/IgA/IgM/IgD ELISA Quantitation Sets (obtained from Cambridge Bioscience) were used to establish the microarray assay for quantifying human IgG, IgA, IgM and IgD, respectively. As accurate quantification of each Ig class requires the capture and detection antibodies to be highly specific for the relevant class, this was initially confirmed by running the Ig class-specific ELISAs according to the manufacturer's instructions and demonstrating that each capture/detection antibody pair bound only the relevant class of purified Ig (obtained from Athens Research Technology); for example, the anti-IgG-Fc capture and detection antibodies detected human IgG, but not IgA, IgM or IgD (Figure 1).
To transpose the assay to the microarray platform, optimization studies were undertaken to determine the optimal printing buffer, blocking buffer, and concentrations of each capture antibody and detection antibody. The optimized microarray assays for detection of human IgG, IgA, IgM and IgD were then performed as follows: Capture antibody goat anti-human Ig-Fc (100 μg/mL in phosphate-buffered saline [PBS]-Trehalose [50 mmol/L]) was loaded onto a 384-well plate (Genetix) and printed in double quadruplicate in a 16 × 16 array format onto a poly-l-lysine-coated glass slide (Thermo Fisher) using a Biorobotics Microgrid II arrayer (Microgrid) and a silicon contact pin (Parallel Synthesis Technologies). During printing, the array chamber was set to 20°C and 60% humidity, and the spot diameter was 315 μm. Slides were held under a vacuum overnight. On the next day, the slides were blocked with bovine serum albumin (BSA) blocking buffer with shaking for 1 hour at room temperature. After incubation, the slides were washed 3 times with PBS containing 0.05% Tween-20 on the shaker for 3 minutes of each wash. Ig standards, serum and saliva samples were added and incubated for 1 hour.
Following incubation, the slides were washed as described above. Then, 100 μL of the HRP-conjugated detection antibody was added and the slides were incubated for 1 hour at room temperature on a shaker. The slides were washed, as described above, and 100 μL biotinylated-tyramide amplification reagent was applied to the slides and incubated for 10 minutes at room temperature. After the incubation, the slides were washed as described above, and 100 μL Cy5conjugated streptavidin (1:1000 in 3% BSA) was applied to each block. The slides were incubated for 15 minutes on the shaker in the dark. Finally, the slides were released from the holder, washed in ultra-pure water for 5 minutes and centrifuged for 3 minutes. After processing, slides were immediately scanned in a GenePix 4200AL microarray scanner, to measure the fluorescence (Cy5 Fluor 635).
The microarray assays had intra-assay coefficient of variation (CV) <10% and inter-assay CV < 15%. The Ig class specificity demonstrated in the ELISA format was maintained in the microarray assays, as exemplified for detection of IgG ( Figure 2).

| Statistical analysis
The data were not normally distributed, and therefore, comparisons between groups were made using the Kruskal-Wallis test for non-paired data. Statistical analyses were performed using GraphPad Prism 7.

| Determining optimal sample dilutions for Ig class quantification
The microarray assays for detection of human IgG, IgM, IgA and IgD were established and optimized as described in the Material and Methods section. Experiments were undertaken to determine the optimal dilutions of human serum for accurate quantification of each Ig class. As shown in Figure  3, the serum dilutions that generated signals corresponding to the sensitive regions of the standard curves were 1:40 000 for IgG and IgA, and 1:200 for IgM and IgD. A saliva dilution of 1:2 was suitable for detection of salivary Igs (data not shown).

| Serum Ig class concentrations in smokers and non-smokers
The serum Ig concentrations for the smokers and nonsmokers are shown in Figure 4A-D. The IgG levels in smokers were very significantly lower than those in nonsmokers (P < 0.0001; Figure 4A); conversely, the serum IgM levels were very significantly higher in smokers than non-smokers (P < 0.0001; Figure 4B). The serum IgA levels were also significantly higher in smokers (P < 0.05; Figure 4C), but the smokers' IgD concentrations were significantly lower than in the non-smokers (P < 0.05; Figure  4D). Figure 4E-H shows that, as in serum, IgG concentrations in the saliva of smokers were significantly lower than in nonsmokers (P < 0.05; Figure 4E), and salivary IgA concentrations were significantly higher in smokers (P < 0.05; Figure  4G). By contrast, IgM levels were very significantly lower in the saliva of the smokers compared to the non-smokers (P < 0.0001, Figure 4F), which is the opposite of the findings for the serum levels of IgM described above ( Figure 4B). There was no significant difference in the very low concentrations of salivary IgD between smokers and non-smokers ( Figure 4H).

| Salivary Ig class concentrations in smokers and non-smokers
There were no statistically significant correlations between serum and salivary Ig levels in the same individuals (either smokers or non-smokers) for any of the four Ig classes measured (data not shown). Also, there were no significant correlations between smoking pack-years and Ig class levels in either serum or saliva samples.

| DISCUSSION
Our findings indicate that smoking is associated with reduced IgG concentrations in both serum and saliva, and enhanced IgA concentrations in both serum and saliva. By contrast, smoking differentially affected IgM concentrations-causing increased concentrations in serum, but decreased concentrations in saliva. We found smoking to be associated with decreased IgD concentrations in serum, and not to have a significant effect on the very low IgD concentrations in saliva.

F I G U R E 2 Investigation by microarray of cross-reactivity
between IgM, IgA and IgD standards with anti-IgG. Anti-human IgG was used as the capture antibody, and an HRP-conjugated goat antihuman IgG-Fc antibody served as the detection antibody. No crossreactivity was observed between the IgM, IgA or IgD standards with an anti-human IgG capture antibody and an HRP-conjugated goat antihuman IgG antibody as the detection antibody. IgG standard curves were generated on three different occasions. The values for IgD were completely undetectable and therefore do not appear on this figure Tables 1 and 2 summarize the findings in the present study, together with the results reported in previous studies (quoted in the Introduction) that investigated the effects of smoking on levels of IgG, IgM, IgA and IgD in serum (Table  1) or in saliva ( Table 2). As mentioned in the Introduction, unlike the present study, these previous studies did not investigate Ig concentrations in both serum and saliva; few determined IgD concentrations (two for serum and none for saliva); and none used microarray technology. The advantages of the microarray platform include its sensitivity and broad dynamic range (0-65 000 fluorescence units, compared to the 0-4 optical density units in ELISA), and the increased number of replicates that can readily be performed within an assay together with detailed feature analysis enabling poorly performing replicates to be legitimately excluded, hence reducing replicate variation.
Considering all the studies summarized in Tables 1 and 2 that investigated the effects of smoking on IgG levels, 6/7 (including the present study) reported a significant decrease in IgG concentrations in the serum of smokers; also 2/3 studies (including the present study) reported a significant reduction in IgG levels in the saliva of smokers. Thus, there is a strong consensus of evidence that smoking is associated with reduced IgG levels both systemically and in the oral mucosa. It would be of interest to examine the effects of smoking on IgG subclasses, especially as these are known to be affected in COPD. 37 By contrast, there is no clear consensus for the effects of smoking on IgA levels, with different studies reporting an increase, a decrease or no effect of smoking on IgA concentrations in either serum (Table 1) or saliva ( Table 2). The reasons for this variation are unclear: it does not appear to be related to different types of assays being used (eg nephelometry vs turbidimetry vs ELISA) since, for example, amongst the seven studies that used ELISA to determine salivary IgA levels, one reported an increase, three reported no change, and three reported a decrease in smokers compared to nonsmokers. Other possible factors leading to the variations could include ethnicity, amount of smoking and relative detection of IgA1 vs IgA2.
Five studies listed in Table 1 reported no significant effects of smoking on serum levels of IgM, whereas a very significant

F I G U R E 4 Comparing the Ig concentrations in serum samples and saliva samples from smokers and non-smoker individuals using a Mann-
Whitney test revealed significant differences. (A) Serum IgG concentrations were lower in smokers than in non-smoker individuals (P < 0.0001). (B) Serum IgM concentrations were higher in smokers than in non-smoker individuals (P < 0.0001). (C) Serum IgA concentrations were higher in smokers than in non-smoker individuals (P < 0.05). (D) Serum IgD concentrations were lower in smokers than in non-smoker individuals (P < 0.05). (E) Salivary IgG concentrations were lower in smokers than in non-smokers (P < 0.05). (F) Salivary IgM concentrations were lower in smokers than in non-smokers (P < 0.0001). (G) Salivary IgA concentrations were higher in smokers than in non-smokers (P < 0.05). (H) Salivary IgD concentrations were not significantly different between smokers and non-smokers increase in smokers was observed in the present study. This may be because, in the present study, several smokers had particularly high serum IgM concentrations, plus very little variation was found in the serum IgM levels of non-smokers ( Figure  4B). On the other hand, the present study agrees with 2/3 other studies in finding reduced levels of IgM in the saliva of smokers compared to non-smokers. Overall, the present study agrees with the consensus view of other studies that smoking has different effects on IgM levels in serum and saliva.
The concentrations of IgD are very low in serum and extremely low in saliva, with an insufficient number of studies for a consensus view on the effects of smoking to be drawn. There is, therefore, clear scope for further studies of the effects of smoking on IgD levels in both serum and saliva. The importance of IgD in the upper aerodigestive tract in humans is indicated by the high numbers of IgD-producing B cells in tonsils, adenoids, salivary and lachrymal glands, and nasal mucosa, whereas IgD-producing B cells are much less frequent in peripheral lymph nodes, spleen, bone marrow and intestinal mucosa. 34 Furthermore, IgD recognizes a range of pathogenic microbes often found in the upper aerodigestive tract of humans, consistent with a protective role for IgD in this location. 34 A better understanding of the effects of smoking on IgD is therefore important. Although the effects of tobacco smoking on IgE levels were not the subject of the present study, this is clearly of importance, not least since smoking exacerbates allergic conditions. Numerous previous studies have shown that tobacco smoking is associated with elevated levels of serum IgE [38][39][40][41][42][43] ; it would also be of interest to determine the effects of smoking on salivary IgE.
The lack of correlation between smoking pack-years and Ig class levels in serum and saliva that we observed is broadly consistent with the findings of others. One study reported a significant inverse correlation between salivary IgA concentrations and current level of smoking, but no correlation between smoking and salivary IgG or IgM concentrations. 23 By contrast, another study reported there to be no correlation between salivary IgA levels and frequency or duration of smoking. 33 With regard to serum Igs, no significant correlation was found between IgD levels and duration of smoking. 20 Another study reported a significant correlation between serum IgG concentrations and the number of cigarettes smoked per day; however, these authors also pointed out that there was a clearly significant difference in serum IgG levels between non-smokers and those smoking a relatively low number of cigarettes (≤10 per day). 17 They also reported a low significance for the correlation between IgM concentrations and level of smoking, but no significance for the correlation of IgA levels and smoking. 17 Overall, our findings, and these previous studies, are consistent with the proposal that any level of cigarette smoking affects Ig levels in serum and saliva, with smoking per se being the main factor, regardless of whether the level of smoking is low, moderate or high.
A variety of mechanisms could be involved in the differential effects of smoking on the concentrations of Ig classes in serum and saliva: these might include direct effects on B cells and indirect effects on T cells and antigen-presenting cells, which could affect Ig class switching and/or differential survival of naïve B cells or memory B cells. 3 In this regard, we have preliminary evidence (see Appendix S1) that the in vitro production of IgG, IgM and IgA by human peripheral blood mononuclear cells (PBMCs) or purified B cells stimulated with anti-CD40 plus IL-4 and pokeweed mitogen is inhibited by cigarette smoke extract or nicotine. It is interesting that pure nicotine appears to suppress Ig production to the same extent as CSE, suggesting that nicotine is an important contributor to the Ig-suppressive effect of whole CSE. Furthermore, these observations are consistent with the investigations of Skok et al in murine models into the effects of nicotinic acetylcholine receptors (nAChRs) expressed by immature and mature B cells. [12][13][14] They deduced that the activity of nAChR promotes the survival and development of immature B cells in the bone marrow, but suppresses B-cell activation in response to antigenic challenge in secondary lymphoid tissues like the spleen. 12,13 In this regard, wild-type mice (expressing nAChR) were found to produce higher levels of natural antibodies than nAChR-knockout mice, but wild-type mice produced lower levels of antigen-specific IgG antibodies than the nAChR-knockout mice when challenged. 12 This may help to explain our observation that smokers, compared to non-smokers, have very significantly lower levels of serum IgG (most of which is derived from antigen-stimulated B cells), but very significantly higher levels of serum IgM (that may include a higher proportion of natural antibodies). By contrast, different processes are likely to determine salivary Ig, where we found both IgG and IgM to be decreased in smokers.
It should also be noted, however, that the suppression by CSE and nicotine of IgA and IgM production by polyclonally stimulated PBMCs and B cells (see Appendix S1) is in contrast to our findings that IgA levels were increased in the sera and saliva of smokers, as were the levels of IgM in smokers' sera. This suggests that the in vitro findings cannot fully explain the effects of CSE and nicotine in smokers in vivo, which are likely to be much more complex. For example, the increased occurrence of infections in the upper respiratory tract of smokers 44 is likely to stimulate the production of IgA as a major mucosal Ig. This complexity is also reflected in the effects of cigarette smoking on antibodies specific to particular antigens, as distinct from total Ig class levels. For example, smokers were reported to have raised levels of anti-pneumococcal IgG antibodies, presumably due to an increased occurrence of pneumococcal infections of the respiratory tract. 45 By contrast, smokers and non-smokers were found to have the same prevalence and titres of antibodies to Escherichia coli and Candida albicans, possibly because these infections are more associated with the gut than the respiratory tract. 16 At another level, smoking was reported to reduce the avidity but not the quantity of IgG anti-human papillomavirus (HPV) antibodies generated in HPV-vaccinated female individuals. 46 Thus, numerous direct and indirect effects of cigarette smoking are likely to contribute to the changes in total Ig and specific antibody levels in smokers.
In addition to the qualitative effects of smoking on Ig class levels, the quantitative changes in Ig concentrations associated with smoking are an important consideration in relation to the possible functional consequences. IgG constitutes most of the Ig in blood, with a concentration about four times higher than IgA and ten times higher than IgM. Thus, the 25%-30% reduction in serum IgG concentrations associated with smoking observed in the present study could be particularly biologically significant. Similarly, the reduced levels of IgG and IgM in the saliva of smokers could be significant despite the slight rise in salivary IgA, even though IgA has the highest concentrations in saliva. The concentrations of IgD in both serum and saliva are too low to affect the overall Ig concentrations significantly.
Overall, the reduction in Ig levels associated with smoking might reasonably be considered as a form of moderate secondary immunodeficiency. 47,48 Similarly, the associations between smoking and autoimmune diseases 5 might result from smoking inducing a degree of immunodeficiency (including reduced Ig levels) that causes inadequate control of infective agents which, in turn, promote the development of autoimmunity. 3,49 This model provides a plausible link between smoking and infections as two environmental factors both known to be associated with autoimmune diseases. 5,49 In addition to its association with autoimmune diseases, cigarette smoke is the main risk factor for COPD. 4 In this context, others have shown that about 25% of COPD patients have total serum IgG levels below the lower limit of normal, 50 consistent with our findings in smokers without COPD. This is associated with an increased incidence of exacerbations and hospitalizations (usually because of respiratory tract infections), and this is mainly in patients with deficiencies in IgG1 and/or IgG2 subclasses. 37 This is consistent with reports that intra-venous Ig (IVIG) replacement therapy reduces respiratory infections and exacerbations in COPD. 51,52 Administration of IVIG is also beneficial in a range of autoimmune diseases, 53 and substantial evidence exists for the occurrence of autoimmunity in COPD-particularly the production of autoantibodies 54,55 ; thus, IVIG may help to reduce inflammation as well as infection in COPD.
To conclude, our data show reduced levels of IgG and raised levels of IgA, in both the serum and saliva of smokers compared to never-smokers. By contrast, IgM and IgD levels were differentially affected in serum and saliva by smoking. The clear evidence for reduced IgG levels associated with smoking in this and other studies may relate to the link between smoking and both autoimmune diseases and COPD.