Innate and Adaptive Immune Defects in Chronic Pulmonary Aspergillosis

We evaluated the expression of biomarkers of innate and adaptive immune response in correlation with underlying conditions in 144 patients with chronic pulmonary aspergillosis (CPA). Patients with complete medical and radiological records, white cell counts, and a complete panel of CD3, CD4, CD8, CD19, and CD56 lymphocyte subsets were included. Eighty-four (58%) patients had lymphopenia. Six (4%) patients had lymphopenia in all five CD variables. There were 62 (43%) patients with low CD56 and 62 (43%) patients with low CD19. Ten (7%) patients had isolated CD19 lymphopenia, 18 (13%) had isolated CD56 lymphopenia, and 15 (10%) had combined CD19 and CD56 lymphopenia only. Forty-eight (33%) patients had low CD3 and 46 (32%) had low CD8 counts. Twenty-five (17%) patients had low CD4, 15 (10%) of whom had absolute CD4 counts <200/μL. Multivariable logistic regression showed associations between: low CD19 and pulmonary sarcoidosis (Odds Ratio (OR), 5.53; 95% Confidence Interval (CI), 1.43–21.33; p = 0.013), and emphysema (OR, 4.58; 95% CI; 1.36–15.38; p = 0.014), low CD56 and no bronchiectasis (OR, 0.27; 95% CI, 0.10–0.77; p = 0.014), low CD3 and both multicavitary CPA disease (OR, 2.95; 95% CI, 1.30–6.72; p = 0.010) and pulmonary sarcoidosis (OR, 4.94; 95% CI, 1.39–17.57; p = 0.014). Several subtle immune defects are found in CPA.


Introduction
The saprophytic, opportunistic, and ubiquitous airborne moulds of the genus Aspergillus are capable of causing a wide spectrum of bronchopulmonary diseases in a range of susceptible individuals [1]. Aspergillus fumigatus is the usual culprit; A. flavus, A. niger, and A. terreus are the other three clinically significant species encountered in mycology settings of the over 300 authenticated species of Aspergillus [2,3]. Invasive pulmonary aspergillosis (IPA), characterised by hyphal angioinvasion, is a life-threatening infection that occurs in severely immunocompromised patients with prolonged and/or profound neutropenia or T-cell dysfunction [4]. In the hypersensitive host, conidia inhalation initiates an allergic response culminating in allergic bronchopulmonary aspergillosis (ABPA) in individuals dedicated to the clinical care of CPA patients and others with aspergillosis referred from all over the UK. Patients referred from 1 April 2009 to July 2016 were included in this audit.

Patient Selection
We collected all the available results of immunological biomarkers of T, B, and natural killer cells, a panel consisting of cluster of differentiation CD3, CD4, CD8, CD19, and CD56. This is a standard panel of CDs routinely requested at the NAC for patients suspected of having immunodeficiency. Patients were then enrolled in this audit if they met the diagnostic criteria for CPA previously described: (1) chronic (duration >3 months) pulmonary or systemic symptoms (e.g., productive cough, haemoptysis, dyspnoea, fatigue, weight loss), (2) radiological evidence of a progressive (over months or years) pulmonary lesion with surrounding inflammation (e.g., cavitation, infiltration, and pleural thickening), (3) no major apparent immunocompromising factors (e.g., AIDS, leukaemia, or transplantation), (4) serological or microbiological evidence of Aspergillus, and (5) absence of alternative diagnosis explaining the findings [26,27].
Only CPA patients with a complete panel of CD markers are included in the study. The other mandatory data required for inclusion were: (1) blood indices (total white cell, lymphocytes, neutrophils, and monocyte counts), (2) radiological information (computed tomography (CT) images and thoracic X-ray), (3) pertinent medical records indicating underlying pulmonary and systemic conditions, and (4) use of immunosuppressant medications. Only data available in written and electronic medical records was collected. Figure 1 shows the study schematic demonstrating the inclusion and exclusion criteria.

Medical Records
We performed thorough hand searching of clinical case notes and correspondences. A standardised data collection sheet was used. Details recorded included confirmation of CPA diagnosis, age, and gender; date of enrolment into clinical care at NAC, underlying pulmonary disorder(s), if any, known systemic co-morbidities, chronic medication, especially immunosuppressive therapy, and band categorization.

Radiology
Electronic Images and reports were obtained from picture archiving and communication system (PACS) software accessed through the Centricity Enterprise Web version 3.0 (General Electric Healthcare, Barrington, IL, USA) at UHSM. Lung windows of both conventional and high resolution multi-slice thoracic computed tomography (CT) images with or without intravenous contrast administered, and plain chest X-ray images were electronically obtained. Baseline images at the time of diagnosis were examined. We also reviewed all the electronic reports of the images submitted by consultant radiologists. Data obtained included: (1) pulmonary involvement (unilateral or bilateral disease), (2) lobe(s) involved, (3) cavitation, (4) evidence of pleural thickening, and (5) presence of aspergilloma.

Data Management
Data from the patients' files and immunology reports were summarised and entered in a printed data collection sheet. The electronically retrieved data from SunquestICE ® desktop and PACS web version 3.0 software together with data on the data collection sheet were thereafter entered into Microsoft Excel 2010 spread sheet (Microsoft Corp., Redmond, WA, USA). The Excel workbook was password protected to restrict access to patients' identifiable information or unlikely alteration of the collected data. Patients' identifiable information entered into the Excel work book included the NHS and RM2 numbers, date of birth, and initials of their names. All this identifiable information was deleted after assigning study numbers to the subjects and prior to exportation to SPSS for statistical analysis. Encrypted flash drives were used for data transfer between the NHS trust computers, e-mailing of patients' data was considered unacceptable.

Statistical Analysis
All statistical analyses were performed using statistical package for the social science (SPSS) software version 22.0 (IBM Corp., Armonk, NY, USA). Unless stated otherwise, statistical significance was at the 5% level for all analyses. Summary statistics were presented in terms of means, ranges, and standard deviations for continuous data. Non-normally distributed variables were presented in terms of medians and ranges. Pearson's correlation was used to test the strength of association for normally distributed variables (CD3, CD4, and lymphocytes) and Spearman's rank correlation was used if one or both variables were not normally distributed (CD8, CD19, and CD56). CD and CBC variables were dichotomised into low (CD19, CD56, CD4, CD8, CD4:CD8 ratio, lymphocyte counts) or high (total white cell count, neutrophils, CD4:CD8 ratio, monocytes) and normal. Pearson's chi-squared tests and Fisher's exact probability tests were performed as appropriate to assess for associations between the CDs, CBC parameters, and radiological and clinical (co-morbidities) characteristics. Logistic regression analyses were performed with multiple covariates to determine independent predictors. Odds ratios and 95% confidence intervals (CIs) were recorded.

Ethical Consideration
This audit was a service evaluation and, as such, informed consent was not required. However, all the Caldicott principles of transfer and handling of patients' identifiable information were observed. The study was registered with the audit department at UHSM.

Demographics, Baseline Radiological and Clinical Characteristics
One hundred and forty-four patients with complete data were analysed. Eighty-five patients were male (59%), and the median age at time of testing was 60 years (range: 22-84). Eighty-seven (60%) had unilateral disease with 64% (n = 56) of the unilateral disease affecting the right lungs of the patients. Ninety-one (63%) patients had multiple cavities in their lungs with fungal ball (aspergilloma) occupying cavities of up to 71 (49%) patients. Upper lobes were the most affected lobes in this cohort of patients (81%, n = 117).

Expression of CDs, Total and Differential White Cell Counts
The distribution of the CDs (CD3, CD4, CD8, CD19, and CD56), total white cell, neutrophil, lymphocyte, and monocyte counts in this group of patients are summarised in Table 3. The laboratory reference range for CD4:CD8 ratio is 0.9-1.9 (Beckman Coulter (UK) Ltd.). The calculated mean CD4:CD8 ratio was 1.55 from our reference laboratory values of CD4 and CD8. A median CD4:CD8 ratio of 1.9 (range: 0.4-6.9) was obtained from the 144 patients. This was just at the upper limit of the reference range and well above the calculated mean CD4:CD8 ratio. Sixty-one (42%) patients had CD4:CD8 ratio within the reference range, 68 (47%) patients had ratios above 1.9, and only 15 (10%) patients had ratios of less than 0.9 (Table 3).
A low CD4 count was more often seen in patients with rheumatoid arthritis: 3 (75%) compared to 22 (16%) of patients without rheumatoid arthritis (Fisher's exact test p = 0.017). Likewise, IL-17 deficiency (2 (100%) compared to 23 (16%) of patients with normal IL-17 cytokine levels) was associated with a low CD4 count (Fisher's exact test p = 0.029). Pulmonary sarcoidosis was also associated with a low CD4 count: 8 (62%) patients versus 17 (13.0%) of the patients without pulmonary sarcoidosis (Fisher's exact test p < 0.001). Multivariable logistic analysis did not reveal any statistically significant associations between low CD4 and the above factors.
Pulmonary sarcoidosis was associated with lymphopenia, where 11 (85%) of the 13 patients had low lymphocyte counts compared to 73 (56%) without pulmonary sarcoidosis (p = 0.044). Age was negatively associated with CD3, CD8 and lymphocyte counts. However, these associations were not statistically significant (p = 0.244, p = 0.703, p = 0.401, respectively) There were no significant associations elicited between the CDs and gender, presence or absence of aspergilloma, pleural thickening, or lung involvement.

Discussion
In this study, we established that significant CD3, CD4, CD8, CD19, and CD56 lymphopenia occurs in patients with CPA. Fifteen (10%) patients without documented HIV infection had very severe CD4 suppression, that is, CD4 <200 cells/µL. Furthermore, we found poor immune homeostasis with immune response inclined towards inflammatory signals, as shown by high CD4:CD8 ratio among these patients. Our findings are in keeping with the currently budding concept that mild immunological suppression and genetic polymorphisms are important factors leading to the development of CPA [12,23]. Severe immunosuppression in the form of chronic granulomatous disease (CGD) and profound neutropenia is well known to predispose to invasive pulmonary aspergillosis [28] and so is T H 2 biased immune mediated pathological inflammation in ABPA [29].
To our knowledge, we present for the first time in literature a study that describes the expression of biomarkers of T-cells (CD3, CD4, and CD8), B-cells (CD19), and natural killer cells (CD56) in a large population of patients with CPA.
It is striking to note that the majority of patients had just one or two identifiable underlying diseases and two (1.4%) patients had no identifiable risk factors. In these patients, one might expect non-pulmonary systemic factors, possibly genetic or immune defects, to play a role in the development of CPA. Interestingly, patients previously treated for cancer (cancer survivors) had lower CD3 counts than other patients, suggesting a prolonged immune suppressing effect following cancer treatment (or a pre-disposition to lung cancer) that could contribute to CPA development. In addition, a low CD3 count was associated with multicavitary as opposed to single cavitary disease. Therefore, more immunosuppression may lead to more extensive disease.
A substantial number of patients had subnormal counts of individual CD subsets. Table 4 shows the numbers of such patients, which was most frequently seen for CD56 (13%), followed by CD19 (7%). In both cases, the total lymphocyte count was usually in the normal range, although not always. Combined low CD19 and CD56 counts were seen in an additional 15 patients (10%), also with generally normal lymphocyte counts, and therefore subnormal CD19 and/or CD56 counts were found in 43 patients (30%). Isolated subnormal CD8 counts were seen in only four patients (3%), and none had isolated subnormal CD4 or CD3. While "compensatory" monocytosis was seen in some patients, this was by no means universal, and no patients had monocytopenia. No published literature currently exists on these cellular biomarkers in CPA patients to compare with our results. Our findings suggest multiple heterogeneity and likely several pathways to CPA development.
CD3, CD4, CD8, and CD19 lymphocyte subset counts were significantly correlated ( addition, a low CD3 count was associated with multicavitary as opposed to single cavitary disease. Therefore, more immunosuppression may lead to more extensive disease. A substantial number of patients had subnormal counts of individual CD subsets. Table 4 shows the numbers of such patients, which was most frequently seen for CD56 (13%), followed by CD19 (7%). In both cases, the total lymphocyte count was usually in the normal range, although not always. Combined low CD19 and CD56 counts were seen in an additional 15 patients (10%), also with generally normal lymphocyte counts, and therefore subnormal CD19 and/or CD56 counts were found in 43 patients (30%). Isolated subnormal CD8 counts were seen in only four patients (3%), and none had isolated subnormal CD4 or CD3. While "compensatory" monocytosis was seen in some patients, this was by no means universal, and no patients had monocytopenia. No published literature currently exists on these cellular biomarkers in CPA patients to compare with our results. Our findings suggest multiple heterogeneity and likely several pathways to CPA development.
CD3, CD4, CD8, and CD19 lymphocyte subset counts were significantly correlated ( ƿ ranged from 0.42 to 0.92, all p < 0.001), indicating severe overall T-cell and B-cell lymphopenia in the peripheral blood of these patients and inadequate cellular and humoral responses in the adaptive arm of the immune system. However, there was no statistically significant correlation between CD3 and CD56 counts (ƿ = 0.10, p = 0.21) and CD19 and CD56 counts (ƿ = 0.15, p = 0.077). This is theoretically expected, as CD56 and the other CDs in this study are from different lineages. However, more significantly, it provides an insight that deficits in both innate and adaptive arms of the immune system are involved in the pathogenesis of CPA.
In our study, 13 (9%) patients had a histologically proven diagnosis of pulmonary sarcoidosis, comparable to 7.1% in a previous study at the same centre [ [11]]. Ten (77%) of those with sarcoidosis in our study had low CD19 counts, nine (69%) had low CD3 counts, and eight (62%) had low CD4 counts. Sweiss et al. (2010) reported significant lymphopenia involving CD4, CD8, and ranged from 0.42 to 0.92, all p < 0.001), indicating severe overall T-cell and B-cell lymphopenia in the peripheral blood of these patients and inadequate cellular and humoral responses in the adaptive arm of the immune system. However, there was no statistically significant correlation between CD3 and CD56 counts ( enetic or immune defects, to play a role in the previously treated for cancer (cancer survivors) had ing a prolonged immune suppressing effect following cancer) that could contribute to CPA development. In h multicavitary as opposed to single cavitary disease.
to more extensive disease. ubnormal counts of individual CD subsets. Table 4 as most frequently seen for CD56 (13%), followed by count was usually in the normal range, although not ts were seen in an additional 15 patients (10%), also nd therefore subnormal CD19 and/or CD56 counts normal CD8 counts were seen in only four patients r CD3. While "compensatory" monocytosis was seen niversal, and no patients had monocytopenia. No cellular biomarkers in CPA patients to compare with heterogeneity and likely several pathways to CPA ubset counts were significantly correlated ( ƿ ranged severe overall T-cell and B-cell lymphopenia in the uate cellular and humoral responses in the adaptive as no statistically significant correlation between CD3 D19 and CD56 counts (ƿ = 0.15, p = 0.077). This is her CDs in this study are from different lineages. sight that deficits in both innate and adaptive arms of enesis of CPA.
logically proven diagnosis of pulmonary sarcoidosis, t the same centre [ [11]]. Ten (77%) of those with , nine (69%) had low CD3 counts, and eight (62%) had d significant lymphopenia involving CD4, CD8, and = 0.10, p = 0.21) and CD19 and CD56 counts ( non-pulmonary systemic factors, possibly genetic or immune defects, to play a role in the development of CPA. Interestingly, patients previously treated for cancer (cancer survivors) had lower CD3 counts than other patients, suggesting a prolonged immune suppressing effect following cancer treatment (or a pre-disposition to lung cancer) that could contribute to CPA development. In addition, a low CD3 count was associated with multicavitary as opposed to single cavitary disease.
Therefore, more immunosuppression may lead to more extensive disease.
A substantial number of patients had subnormal counts of individual CD subsets. Table 4 shows the numbers of such patients, which was most frequently seen for CD56 (13%), followed by CD19 (7%). In both cases, the total lymphocyte count was usually in the normal range, although not always. Combined low CD19 and CD56 counts were seen in an additional 15 patients (10%), also with generally normal lymphocyte counts, and therefore subnormal CD19 and/or CD56 counts were found in 43 patients (30%). Isolated subnormal CD8 counts were seen in only four patients (3%), and none had isolated subnormal CD4 or CD3. While "compensatory" monocytosis was seen in some patients, this was by no means universal, and no patients had monocytopenia. No published literature currently exists on these cellular biomarkers in CPA patients to compare with our results. Our findings suggest multiple heterogeneity and likely several pathways to CPA development.
CD3, CD4, CD8, and CD19 lymphocyte subset counts were significantly correlated ( ƿ ranged from 0.42 to 0.92, all p < 0.001), indicating severe overall T-cell and B-cell lymphopenia in the peripheral blood of these patients and inadequate cellular and humoral responses in the adaptive arm of the immune system. However, there was no statistically significant correlation between CD3 and CD56 counts (ƿ = 0.10, p = 0.21) and CD19 and CD56 counts (ƿ = 0.15, p = 0.077). This is theoretically expected, as CD56 and the other CDs in this study are from different lineages.
However, more significantly, it provides an insight that deficits in both innate and adaptive arms of the immune system are involved in the pathogenesis of CPA.
In our study, 13 (9%) patients had a histologically proven diagnosis of pulmonary sarcoidosis, comparable to 7.1% in a previous study at the same centre [ [11]]. Ten (77%) of those with sarcoidosis in our study had low CD19 counts, nine (69%) had low CD3 counts, and eight (62%) had low CD4 counts. Sweiss et al. (2010) reported significant lymphopenia involving CD4, CD8, and = 0.15, p = 0.077). This is theoretically expected, as CD56 and the other CDs in this study are from different lineages. However, more significantly, it provides an insight that deficits in both innate and adaptive arms of the immune system are involved in the pathogenesis of CPA.
In our study, 13 (9%) patients had a histologically proven diagnosis of pulmonary sarcoidosis, comparable to 7.1% in a previous study at the same centre [11]. Ten (77%) of those with sarcoidosis in our study had low CD19 counts, nine (69%) had low CD3 counts, and eight (62%) had low CD4 counts. Sweiss et al. (2010) reported significant lymphopenia involving CD4, CD8, and CD19 positive cells among sarcoidosis patients, and this was unrelated to their medical treatment but rather more to disease pathology [30]. CPA complicates sarcoidosis, with estimate global burden of over 70,000 individuals [31]. However, no statistically significant CD56, CD8 lymphopenia was seen among CPA patients with sarcoidosis in our study, in contrast to the work of Sweiss and colleagues.
In our study, impaired IL-17 production was statistically associated with low CD4 (Fisher's exact test p = 0.029). This is consistent with the experimental work of Doffinger et al. (2014) that showed a significant impairment of IFN-γ, IL-12, and IL-17 production in a majority of patients with CPA, suggesting a major involvement of T H 1/T H 17 and potentially NK-cell subsets [32]. IL-12, also known as T-cell stimulating factor, is produced by antigen presenting cells, in particular dendritic cells and macrophages, and neutrophils to a lesser extent. It stimulates the differentiation and production of IFN-γ and TNF-α by naive T-cells and natural killer cells [33]. Both CD4 + T H 1 cells and CD8 + T-cells produce IFN-γ, which is critical in host defence against pulmonary infections [17].
Prior TB and NTM infection were not statistically associated with CD4 or CD8 lymphopenia. Active TB infection is known to cause CD4 and CD8 lymphopenia, and T-cell exhaustion and depletion [18,34]. Loss of CD4 + T-cells results in progressive primary TB infection, reactivation of latent TB infection (LTBI), and enhanced susceptibility to re-infection but not susceptibly to TB [34]. Both CD4 + and CD8 + T-cells have been reported to recover towards normality after successful anti-TB treatment [35]. T-cell exhaustion is a known syndrome implicated in a number of chronic infections [36]. T-cell exhaustion has not been previously reported in CPA. Perhaps the profound lymphopenia indicates a possibility of such a phenomenon in CPA. Recently, clinical studies have shown that T-cell immunity impairment is associated with an increased susceptibility to A. fumigatus infection [16].
Sixty-eight (47%) patients had high CD4:CD8 ratio, an indication of a pro-inflammatory biased immune signal among these patients. CD4 + T-lymphocytes are important mediators of host inflammatory/anti-inflammatory responses, with the balance between T H 1/T H 17 and T H 2 phenotypes dictating disease pathology [37]. Aspergillus-host interactions trigger the cells of the innate immune system that subsequently elicit the adaptive arm to regulate the balance between pro-inflammatory and anti-inflammatory signals [14]. Thirty-eight (26.4%) of our patients had neutrophilia. Neutrophils and alveolar macrophages are important human innate cells against Aspergillus infection through phagocytosis and secretion of the neutrophil extracellular traps (NETs) [38]. An experimental study by Smith et al. (2015) showed increased levels of macrophage-derived neutrophil chemoattractant pro-platelet basic proteins, which is thought to be pathological in ABPA and CCPA patients [39].
Sixty-two (43%) patients had low CD19 counts, demonstrating profound humoral immune deficiency. Contrarily, over 90% of CPA patients have positive Aspergillus-specific IgG, a key criterion for diagnosis [26,40,41]. The possible explanation here could be that the few patients with negative anti-Aspergillus-specific IgG antibody tests had very low CD19 counts. We used absolute CD19 values below reference range as "low", but this might not be critically low enough to impair antibody production. Page et al. (2016) reported two cases without any Aspergillus IgG in multiple assays and referred to them as "seronegative CPA", since they had overt CPA but negative titres for Aspergillus antibody; ineffective antibody response to Aspergillus due to underlying immune deficit was a suggested explanation [42]. In a recent study, only 47.6% of our CPA patients were shown to have adequate response 3 months after administration of pneumococcal 23-valent polysaccharide vaccine [43].
Severe asthma and allergic bronchopulmonary aspergillosis (ABPA) are risk factors for development of CPA in a minority of our patients [11]. Consistent with our findings, a recent study has shown no significant difference in CD3, CD8, and CD56 expression in patients with asthma and healthy controls [44].
This study, however, has a number of limitations. First, absolute T-and B-lymphocyte counts can be influenced by many factors such as infections and medications compared to the relatively stable percentage total lymphocyte counts [45]. CPA is a chronic infection; therefore, it cannot be established if CPA is the cause or the consequence of the immunological alterations. Second, with the exception of CD4 counts, it is difficult to establish precise levels of clinically significant CD3, CD8, CD19, and CD56 lymphopenia. This makes it difficult to translate these results in clinical practice, as clinically relevant depression in CD counts might be quite different from the lower limit of the laboratory reference range. Third, the retrospective nature of the study made it difficult to access other potentially vital information such as concomitant corticosteroid or immunosuppressive medications (although infrequently used in our patients). However, significant immunosuppression is an exclusion from the diagnosis of CPA, so even if present, such medications are unlikely to have a radical impact on the results. Fourth, some immunological abnormalities are associated with underlying pulmonary disorders, including sarcoidosis, but these parameters have not been well studied in other underlying disorders for CPA. Therefore, the specificity of the findings requires substantial additional work. Fifth, only a single biomarker of natural killer cells (CD56) and B-cells (CD19) were used in this study and yet CD16 and CD57 are also expressed by NK-cells and CD20 and CD22 for B-cells. Lastly, we cannot exclude all systemic co-morbidities if not already known by the referring physician.
The present study is the first of its kind to provide direct evidence of considerable defects in both the innate and adaptive arm of the immune system in the majority of patients with CPA. Routine analyses of these biomarkers, preferably both absolute and percentage T-and B-lymphocyte counts are advised to detect subtle immunological deficits in this group of patients. This could help tailor individualised management of CPA patients.
This study could benefit in the future design of a matched prospective study to evaluate the trend, prognostic and therapeutic significance of these biomarkers in patients with CPA compared to patients with underlying pulmonary conditions that do not have CPA. Particular patient groups within CPA need further study, notably those who are seronegative for Aspergillus IgG.