IL-7/IL7R axis dysfunction in adults with severe community-acquired pneumonia (CAP): a cross-sectional study

Community-acquired pneumonia (CAP) is a worldwide leading cause of death. Recognized risk factors in some severe cases have not been identified. Lymphocytopenia has been frequently described in CAP. Since IL-7, membrane-bound receptor (IL7Rα;CD127) and soluble IL7Rα (sIL7R) are critical in lymphocytes homeostasis, in this work we aimed to evaluate the involvement of the IL-7/IL7Rα axis in the severity of adult CAP, since it has not been explored. The IL7Rα SNPs rs6897932, rs987106, and rs3194051 SNPs in IL7α were genotyped, the systemic expression of the IL7R gene, sIL7R, IL-7, and levels of peripheral IL7Rα+ T lymphocytes were quantified in 202 hospitalized CAP cases. rs3194051GG was more frequent in non-survivors than in survivors; rs987106TT was more frequent and rs3194051AA less frequent in patients at intensive care unit (ICU) than in those not admitted to ICU. IL7Rα gene expression was lower in non-survivors than in survivors, and in severe than in mild cases. CD3+CD127+ lymphocytes were lower in severe than in mild cases; in non-survivors than in survivors and in ICU than in non- ICU admitted cases. sIL7Rα plasmatic levels were higher in non-survivors than in survivors, and in severe than in mild cases. rs6897932CC, rs987106AA and rs3194051GG carriers showed the highest while rs6897932TT showed the lowest sIL7Rα levels. The AUC of sIL7Rα levels predicting 30-day mortality was 0.71. Plasma IL-7 levels were lower in ICU-admitted than in not ICU-admitted and in non-survivors than in survivors. No additional association was detected. In conclusion, rs3194051GG and rs987106TT IL7R genotypes were associated with a poorer prognosis. A significant association between sIL7R levels and SNPs of the IL7R gene is described for the first time in adult CAP. Increased plasmatic sIL7R could contribute to identifying adult CAP cases at risk of death.

www.nature.com/scientificreports/ abdominal ultrasound imaging and/or computed tomography scan); chronic kidney damage (serum creatinine level ≥ 2 mg/dL); cardiac insufficiency (heart failure with reduced ejection fraction (≤ 45%)); neurologic disease (nervous system disorders including epilepsy, Parkinson's Disease, Alzheimer's Disease, and stroke); smoking; alcohol consumption (> 80 g per day yes/no); clinical presentation (fever, respiratory symptoms); chest radiographic patterns (consolidation and pleural effusion), and laboratory parameters (i.e., hemogram, plasmatic concentration of glucose, sodium and blood urea nitrogen). The severity of patients was evaluated according to PSI 3 and CURB-65 1 clinical scores applied during the first 48 h after enrollment: IV and V groups of PSI and ≥ 3 of CURB-65 score were considered as severe, while I and II groups of PSI and ≤ 2 of CURB-65 score were considered as mild. Several outcome variables were analyzed: death up to 30 days post-discharge, admission to ICU, use of mechanical ventilation (MV), and supplemental oxygen therapy. Basal values of different parameters were determined in 22 asymptomatic adults without respiratory infection at least 45 days before enrollment and with no pathogen detected in microbiological screening. The study was approved by the Committee on Research in Human Beings of Faculty of Medicine, University of Chile; the Ethics Committee of Clínico U. Chile Hospital, and the Ethics Committees of the South and North Metropolitan Health Service. Written informed consent was provided by all patients at enrolment. The study was performed in accordance with the principles of the Declaration of Helsinki.
Microbiological study. An immunochromatographic test (Binax NOW, Portland, Oregon, USA) was used to detect Streptococcus pneumoniae and L. pneumophila serogroup 1 antigens in urine. For the detection of other respiratory pathogens, total genetic material was extracted from nasopharyngeal aspirate/swabs samples (300 µL) by Favorprep™ Viral Nucleic Acid Extraction kit (Favorgen®), according to manufacturer's instructions, and after was quantified in an EPOCH® spectrophotometer. Human respiratory viruses such as respiratory syncytial virus, influenza A and B viruses, adenovirus, metapneumovirus, bocavirus, coronavirus (HCoV 229E; HCoV OC43, HCoV-NL63, and HCoV-HKU1), parainfluenza virus, rhinovirus/enterovirus, and bacterias (Chlamydophila pneumoniae, Mycoplasma pneumoniae and Legionella pneumophila) were detected by Respiratory Multi well system r-gene Real Time RT-PCR kit Argene®, according to the manufacturer's instructions in a MIC® thermocycler (BMS®).
Only Chilean people were studied. This population is over 90% admixture Caucasian and South amerindian people 16 .
Allele-specific PCR was applied for genotyping SNP rs987106, using a specific forward primer to each allele (A or T) and a common reverse primer as described above in "Primer design". Two reactions were performed for each sample. Final volume (10 μl) was prepared with DNA template ( www.nature.com/scientificreports/ expression was determined by real time RT-PCR assay based on SYBR-Green detection. TUBB was used as gene for normalization 17 . cDNAs were obtained from 2 µg RNA using M-MLV Reverse Transcriptase (Promega ® ) and random primers (Promega ® ), according to the manufacturer's guidelines. Primers to IL7R mRNA were designed using the Genbank sequence NM_002185.4 and the software Beacon Designer 8.02 (PREMIER Biosoft ® ) (F:  5'-CAG CAA TGT ATG AGA TTA -3'; R: 5'-ATG GTT AGT AAG ATA GGA T-3'). The PCR reaction was performed using the KAPA SYBR © Fast qPCR (Kapa Biosystems ® ) with cDNA (2 µl) in a final volume (10 µl), according to the Eco™ Real-Time PCR system (Illumina ® ) protocol. PCR amplification was performed at 95 °C for 3 min, and  45 cycles: 95 °C for 3 s, 53 °C for 15 s and 72 °C for 3 s followed by the melting curve (55-95 °C). Gene expression fold changes (FC) were determined using the 2 −ΔΔCT method 18 . The group used as control is indicated in each analysis. Statistical analysis. Continuous variables were described as median and interquartile range (IQR) and categorical variables as frequency and percentage. Differences between groups were performed using the Chi square test for categorical data, and Student´s t and Mann-Whitney test for continuous variables, as appropriate.

Quantification of membrane bound IL7Rα (CD127) expression on T cells and of peripheral
Parameters were compared according to predictor variables age, sex, and comorbidities. Days of illness were also analyzed in some parameters-as indicated in "Results"-because the immune response is dynamic over time.
The area under the receiver operating characteristic curve (AUC) for predicting 30-day mortality and admission to ICU was calculated for plasmatic sIL7Rα. The association for each SNP of the IL7R with severity score and clinical outcomes (supplemental oxygen therapy, days of illness, ICU admission and death) in adults with CAP was evaluated through logistic regression model: crude (model 1), adjusted by age and sex (model 2) and adjusted by age, sex, and severity score (model 3). Odds ratios (ORs) and respective 95% confidence interval (CI) were reported as measure of associations. Hardy-Weinberg equilibrium of each SNP and association studies were tested using Armitage's trend test (http:// ihg. gsf. de/ cgi-bin/ hw/ hwa1. pl). Alleles, genotypes and haplotype analysis were carried out with the Unphased 3.1.5 software 19 , as previously described 19

Results
Characteristics of studied population. Carriers of the rs987106TT and the rs3194051AG genotypes were more frequent in patients hospitalized in ICU ( Table 2). The G allele in this last SNP is a risk allele according to Armitage's test for association studies, with an odds ratio of 2.97 (p = 0.0003) between ICU-no ICU groups and of 2.52 with respect to the A allele between severe/mild cases according to PSI (p = 0.004). In addition, the rs3194051GG genotype was also more frequent in non-survivors than in survivor patients ( Table 2). No differences were identified in the SNP rs6897932, except a higher number of carriers of the C allele and CC genotype among severe than in mild patients according to CURB-65. On the other hand, carriers of the rs3194051 AA genotype were significantly more frequent in patients who did not were admitted to the ICU (Table 2). Multivariable logistic regression models were performed to assess if the genotype GG of rs3194051 was associated with the severity of illness compared to the other genotypes (GA, AA, or any of them (AA/GA)). rs3194051GG had a significant positive relationship with admission to ICU even in models adjusted by age, sex, and severity (Table 3). rs3194051GG was also positively associated with dead in relation to any of other genotypes in adjusted models by age and sex, but it did not reach statistical significance when severity was included in the model. These analyzes were not possible with the PSI score because of the small sample size. www.nature.com/scientificreports/ According to age and sex, carriers of the C allele and CC genotype in the rs6897932 and the A allele in the rs987106 were significantly more frequent in older than in younger patients. This difference remained significant among men, but not among women (see Supplementary Table 2 online). In addition, the rs6897932CC and rs987106AA genotypes were more frequent in females than in males, but only in < 65 years patients. No differences were detected in the SNP rs3194051.
IL7R haplotypes. The association between the haplotypes in the three SNPs of the IL7Rα gene and CAP severity was studied. The CAA allelic combination (rs6897932-rs987106-rs3194051) was less frequent in patients admitted to the ICU; the CTA haplotype was associated with dead and the CTG haplotype with admission at ICU (Table 4). The TTA haplotype was associated with mild illness, but only when CURB-65 score is applied.  www.nature.com/scientificreports/ Since age and comorbidities are factors associated to severity also were analyzed. Certain haplotypes were associated with any comorbidities and CAA allelic combination was more frequent in adults over 65 years ( Table 4).
Analysis of the genotypes of all the SNPs studied showed that the CC-TT-AG combination was significantly more frequent in severe cases than in mild cases, and the CC-TT-GG genotype in the severe by CURB-65 and non-survivors patients (Table 5). So, G allele in the SNP rs3194051 is associated with severity of the CAP in adults.
IL7Rα gene expression. IL7Rα gene expression was lower in 13 non-survivors adults than in 112 survivors with CAP (FC: 0.26 vs 1.0; p = 0.01), but similar between 22 patients in ICU and 107 cases without admission to ICU (p = 0.6). According to severity of CAP defined by clinical scores, gene expression was lower in 57 severe cases than in 39 mild cases by PSI (FC: 0.27; p = 0.009), but similar between cases classified by CURB-65.
In relation to genotypes of the SNPs studied, expression of IL7Rα gene was significantly lower in 58 carriers of the CT genotype than in 54 carriers of the CC genotype of the rs6897932 (FC: 0.28; p = 0.01). Although differences were not statistically significant, expression was lower in 66 carriers of the AT and 28 of the TT genotypes than in 35 carriers of the AA genotype in the rs987106 (FC: 0.33 and 0.25, respectively) (p > 0.1) and higher in 30 carriers of the AG genotype than in 97 carriers of the AA genotype in the rs3194051 (FC: 1.79; p = 0.5).
IL7Rα gene expression was lower in 81 adults over 65 years old with CAP than in 48 adults < 65 years (FC: 0.55 vs 1.0, respectively), and was higher in 62 men than in 67 women (FC: 2.31 vs 1.0, respectively), although a statistical significance was not gotten. No differences were found comparing days of illness or pathogen detected.  (Fig. 1B). Also CD3 + TL was lower in non-survivors patients than in survivors (median: 35.3% vs 46.6% [IQR: 29.9-48.2 and 35.2-60.8; p = 0.1), but statistical significance was not reached. CD3 + lymphocytes proportion was lower in severe patients than in mild cases according to PSI

Expression of membrane-bound IL7Rα (CD127) on T lymphocytes. T lymphocytes expressing
α-receptor membrane-bound IL7Rα (CD127) were studied by flow cytometry. Significantly lower proportion of the CD3 + CD127 + lymphocytes was recorded in patients as compared to asymptomatic adults (median: Although expression of membrane bound IL7Rα (CD127) in CD4 + and CD8 + T cells was quantified in few patients, a significant lower proportion of CD4 + CD127 + cells, but no of CD8 + CD127 + cells, were detected in severe cases as compared to mild cases classified by PSI (Fig. 4A,B).
Quantification of soluble IL7R. Soluble IL7R (sIL7R) is associated with the homeostasis of the IL-7/ IL7R axis. In this study, plasmatic levels of sIL7Rα were higher in asymptomatic adults than in adults with CAP, although statistical significance was not developed (median: 26 (Fig. 5A). Also levels of sIL7R were significantly higher in adults ≥ 65 years than in patients < 65 years (median: 24.6 vs 17.5 [IQR: 16.9-35.0 and 11.8-25.4]; p = 0.001) (Fig. 5B). The non-survivors patients showed Table 5. Genotypes of SNP rs6897932, rs987106 and rs3194051 of the IL7Rα gene in 202 adults with CAP. a (n haplotype/n total haplotypes). b P values were calculated using the likelihood-ratio Chi-squared test for allelic or genotypic frequencies of each SNP in each group.  (Fig. 5C). sIL7R also was higher in the 29 patients with mechanical ventilation than in 124 without MV, but the difference was not significant (median: 29.4 ng/ml vs 21.39 ng/ml; p = 0.05). Significant differences in levels of sIL7R were detected in the patients according to the genotypes of the three SNP studied. Thus, the highest levels were detected in carriers of rs6897932CC, rs987106AA, and rs3194051AA genotypes; while the lowest concentrations were observed in adults with rs987106TT, rs3194051AA genotypes, and the lowest of all, the rs6897932TT carriers (Fig. 6).
Moreover, to determine whether plasma concentrations of the sIL7R correlate with the level of membranebound receptor expression on T cells (CD3 + CD127 + lymphocytes), ratio of membrane to soluble receptor was evaluated, being significantly lower in severe than in mild cases by both PSI and CURB-65, and in adults ≥ 65 years than in adults < 65 years. Although not statistically significant, ratio was also lower in non-survivors than in survivor patients and in cases in ICU than in without admission to ICU (Table 6). www.nature.com/scientificreports/ Although the linear regression model showed no relationship between sIL7R concentration and days of illness (coefficient −0.02; 95% CI: −0.06 to 0.014; p = 0.29), given sIL7R levels were significantly higher in 95 patients with less 7 days of illness than in 57 cases with > 7 days (median: 23.8 vs 19.4 ng/ml; p = 0.03), the analyses were repeated excluding the latter cases. Significant differences remained in all comparisons previously shown, except in higher levels of sIL7R between carriers of AG vs AA and GG vs AA genotypes in the rs3194051. On the contrary, the greatest levels of sIL7R in carriers of CC vs CT genotypes in the rs6897932 and AA vs AT genotypes in the rs987106 achieved statistical significance (see Supplementary Table 3

Concentration of IL-7.
Since IL-7 activates its receptor and plays a transcendental role in T cell homeostasis, IL-7 levels were evaluated. Significantly higher levels of plasmatic cytokine IL-7 were detected in adults with CAP than in asymptomatic adults. Levels were lower in patients in ICU, with oxygen therapy and in the non-survivors patients, although the last difference was not statistically significant (Table 7). No significant differences were detected in IL-7 levels by gender, age, or severity according to both clinical scores (Table 7), or by genotypes either (see Supplementary Table 4 online).
No correlation between plasmatic IL-7 and sIL7R concentrations was found in 123 CAP cases studied (Spearman's rank correlation coefficients, r = −0.

Discussion
To the best of our knowledge, this is the first study of the genic/protein expression IL-7/IL7R axis in adults with CAP, looking for a new potential severity factor. High plasmatic levels of soluble receptor IL-7 (sIL7R) and two IL7R gene polymorphisms studied were associated to CAP severity. Thus, the non-survivors patients had higher www.nature.com/scientificreports/ plasmatic levels of sIL7R than the recovered patients, as well as non-significant lower levels of both the genic expression of IL7R and IL-7, and of circulating CD3 + TL and CD3 + CD127 + TL. These results agree with previous reports on the regulatory role of sIL7R in the capture of IL-7 6 . Its increase produces a diminishing of the circulating IL-7 level, which interacts with the lymphocyte's membrane receptor, determining lymphopenia. The increase in membrane receptor levels could be due to attempts to compensate for the decrease in IL7. CD4 + 127 +  www.nature.com/scientificreports/ more than CD8 + 127 T lymphocytes were affected, which is very important given the essential role of these cells in the immune response. The significant relationship between high sIL7R levels and lethality risk in CAP has also been described in patients suffering sepsis, where cases with lower sIL7R have better outcomes, proposing its use as a risk biomarker 20 . Actually, in our study those cases with fatal sepsis had higher levels of sIL7R than those recovered (medians: 30.8 vs 19.7 ng/ml; p = 0.01) and the increased sIL7R levels persisted when the case severity was defined by other parameters, such as ICU admission, PSI and CURB-65 severity scores. AUC of sIL7R for fatality was similar to those Score index (AUC = 0.71), being similar to the obtained in septic cases (AUC: 0.774-0.846). Furthermore, it is interesting to outstand that cases classified as moderated by PSI and CURB-65 had sIL7R levels similar to severe cases and also over the mild cases levels, (median: 22.41, 25.76 and 17.18 ng/ml, respectively), suggesting that those cases should have been managed as severe ones; some of them were after admitted to ICU or passed on.
In addition, two IL7R gene polymorphisms studied were associated to CAP severity. rs3194051GG genotype carriers displayed five times more lethal outcomes than cases with no GG. These patients also predominated among the cases admitted to UCI, as well as rs987106TT genotype carriers. GG genotype has been associated with autoimmune pathologies like type I diabetes 21 and multiple sclerosis 11 , diseases that were not reported in the cases studied. Both SNPs are located in a potential regulatory zone whose effects have not been defined 14 and which it is different to those related to mutations observed in the combined severe immunodeficiency, disease that is not present in the adults studied.
The different production of sIL7R is the more noticeable effect of polymorphisms 12,22 . In this study, changes in the sIL7R levels were detected in the three SNP according to the genotype. Thus, the plasmatic level was higher in rs3194051GG, rs987106AA and rs6897932CC genotypes, while the lowest was in the carriers of rs6897932TT. www.nature.com/scientificreports/ It is striking that the highest levels are detected in the genotype associated with severe and fatal cases, while the lowest levels are identified in the genotype TT, described previously as a protector for multiple sclerosis 11 , through of reduced exon splicing and production of soluble IL7R 23,24 .
Although the plasmatic levels of sIL7R were associated to genotypes, no association was observed between the genotypes and the expression of IL7R gene in blood. This last result should be interpreted with caution by two limitations of this study: first, the analyzed mRNA corresponded to the receptor attached to the membrane and not to the soluble one, and second because RNA is extracted from all blood cells, which does not detect changes in expression in a certain cell type. Association between genotypes and receptor expression in the LT membrane was also not detected. It would be interesting to explore whether this condition also occurs in other types of leukocytes such as monocytes. Recently, a key role of the latter cells has been raised in the biological pathway of diseases associated with cytokine IL-7 and its receptors 24 .
The highest proportion of severe cases and deaths detected among those over 65 is consistent with the risk of severity described in older adults. They also had higher levels of sIL7R and lower percentage of CD127 + T lymphocytes in relation with the younger cases, which agree with the role of IL-7/IL7R axis in the immunosenescense 6 . However, the IL7R expression was not significantly diminished in older adults as it has been published 25 , and further studies are necessary. On the other side, since sIL7R levels are increased over 65 years old and all the fatal  www.nature.com/scientificreports/ cases were over that age, it could be argued that the high level was due to the age factor. However, among those over 65 years of age, sIL7R levels were significantly lower in recovered than in non-survivors cases (median: 23.3 vs 33.2 ng/ml; p = 0.004). Therefore, the old age would not be the only cause of the high level of the sIL7R. There were no gender differences, but some parameters showed differences when they were analyzed according to age. Thus, while the proportion of CD3 + and CD127 + T lymphocytes in females was lower over 65 years old than in younger ages, the genotype rs987106CC in men was significantly more frequent in cases over 65 years old. There are no explanations for this fact or its influence in the CAP outcome.
Although the IL7/IL7R axis has been associated with certain autoimmune conditions like type 1 diabetes mellitus, rheumatoid arthritis and multiple sclerosis 26 , the more frequent comorbidities in patients over 65 years old does not explain the higher sIL7R concentration detected. They represent other type of pathologies, mostly cardiovascular and neurologic diseases, and no differences were observed in the levels of sIL7R between cases with and without these comorbidities. Therefore, the presence of comorbidities would not interfere with the level of sIL7R as severity biomarker in adult CAP. Furthermore, its determination can be quantitatively performed with a routine technique such as ELISA, which it can be easily implemented in clinical hospitals.
Plasmatic concentrations of IL-7 had no variation in relation to severity conditions described or with genotypes or sIL7R plasmatic levels, unlike that described in sepsis 20 . Probably, the SNP-associated mechanisms affecting IL-7 levels are more complex, and independent mechanisms of sIL7R regulate IL-7 levels 23 . This needs to be resolved before implementing recombinant human IL7 therapy as proposed given severe CAP-associated lymphopenia 27 .
There were no differences in sIL7R and IL-7 in relation to pathogens detected. Just a higher IL-7 level in viral versus bacterial CAP was detected, probably explained by the role of IL-7 in promoting antiviral activity of T lymphocytes 14 .
A limiting condition of the study was the patient evaluation just one time, which impedes to have a dynamic vision of quantified parameters. However, only differences in sIL7R were detected according to the time of evolution, being significantly lower in cases with over one week of evolution. No differences were seen in cases with more or less than 3 days of evolution, neither between 1 and 3 days as in sepsis 20 . Another limitation of the study could be the mixed analysis of patients with and without detected pathogen; however, the absence of differences in the demographic and clinical characteristics, in the evolution and in the parameters studied between the two groups (data not shown) suggest that the results would not be affected by the detection of pathogens.
The results in our study explore alterations in the IL-7/sIL7R/mIL7R axis in adult community-acquired pneumonia. The aforementioned axis formulated in relation to autoimmune diseases like multiple sclerosis, includes lower levels of sIL7R and augmented relationship of mIL7R/sIL7R 26 . For adult CAP with risk of severity we propose a pathogenic model characterized by a pattern of a high level of sIL7R and a diminished IL-7 concentration, which tend to decrease T lymphocyte activity and therefore, the control of the infection. www.nature.com/scientificreports/ The development of new biomarkers of severity could improve therapy of CAP in adults. In this study, parameters associated with severity were identified which, although currently have limited clinical application, it is of interest to explore their usefulness in conjunction with other biomarkers, since the complexity of the pathogenesis of the neumonia, a panel of different biomarkers is probably required, including genetic and immunological parameters.
In conclusion, SNPs of IL7R would be related to the severity of adults with CAP, since unfavorable IL7R genotypes (rs987106 TT, and rs3194051GG) had a worse prognosis. Our results constitute the first description of a significant association between sIL7R levels and SNPs in rs3194051, rs6897932 and rs987106 in adults with CAP. The increase in plasma level of sIL7R can contribute to identify adults with CAP in risk of death. Their routine implementation and even automated quantification, by the ELISA technique can be easily executed in hospitals.