Skip to main content
PMC home page
PLOS ONE logoLink to PLOS ONE
. 2013 Nov 13;8(11):e80795. doi: 10.1371/journal.pone.0080795

Ambient Air Pollution Associated with Suppressed Serologic Responses to Pneumocystis jirovecii in a Prospective Cohort of HIV-Infected Patients with Pneumocystis Pneumonia

Robert J Blount 1,*, Kpandja Djawe 2,3, Kieran R Daly 2,4, Leah G Jarlsberg 1, Serena Fong 8, John Balmes 1,5, Robert F Miller 6,7, Peter D Walzer 2,3,4, Laurence Huang 1,8; on behalf of the International HIV-associated Opportunistic Pneumonias (IHOP) Study.
Editor: Sten H Vermund9
PMCID: PMC3827464  PMID: 24236202

Abstract

Background

Ambient air pollution (AAP) may be associated with increased risk for Pneumocystis pneumonia (PCP). The mechanisms underlying this association remain uncertain.

Objectives

To determine if real-life exposures to AAP are associated with suppressed IgM antibody responses to P. jirovecii in HIV-infected (HIV+) patients with active PCP, and to determine if AAP, mediated by suppressed serologic responses to Pneumocystis, is associated with adverse clinical outcomes.

Methods

We conducted a prospective cohort study in HIV+ patients residing in San Francisco and admitted to San Francisco General Hospital with microscopically confirmed PCP. Our AAP predictors were ambient air concentrations of particulate matter of < 10 µm in diameter (PM10) and < 2.5 µm in diameter (PM2.5), nitrogen dioxide (NO2), ozone (O3), and sulfur dioxide (SO2) measured immediately prior to hospital admission and 2 weeks prior to admission. Our primary outcomes were the IgM serologic responses to four recombinant P. jirovecii major surface glycoprotein (Msg) constructs: MsgC1, MsgC3, MsgC8, and MsgC9.

Results

Elevated PM10 and NO2 exposures immediately prior to and two weeks prior to hospital admission were associated with decreased IgM antibody responses to P. jirovecii Msg. For exposures immediately prior to admission, every 10 µg/m3 increase in PM10 was associated with a 25 to 35% decrease in IgM responses to Msg (statistically significant for all the Msg constructs), and every 10 ppb increase in NO2 was associated with a 19-45% decrease in IgM responses to Msg (statistically significant for MsgC8 and MsgC9). Similar findings were seen with exposures two weeks prior to admission, but for fewer of the Msg constructs.

Conclusions

Real life exposures to PM10 and NO2 were associated with suppressed IgM responses to P. jirovecii Msg in HIV+ patients admitted with PCP, suggesting a mechanism of immunotoxicity by which AAP increases host susceptibility to pulmonary infection.

Introduction

Greater than three million deaths per year are attributable to ambient air pollution (AAP), and AAP is the 9th leading factor contributing to worldwide burden of disease according to the 2010 Global Burden of Disease Study [1]. AAP is well established as a risk factor for a wide array of cardiopulmonary diseases, and there is emerging data that it is associated with pulmonary infections. For instance, a number of observational studies have found AAP to be associated with increased emergency department visits and hospital admissions for community acquired pneumonia [2-12]. AAP may also contribute to the morbidity of other pulmonary infections, although data in this area is limited. It is also unclear if the effects of AAP are present and perhaps magnified in immunosuppressed populations, such as those with HIV, who are already at increased risk for pulmonary infections [13,14].

Pneumocystis pneumonia (PCP) continues to contribute to substantial morbidity and mortality in immunosuppressed children and adults worldwide [15], and a better understanding of environmental risk factors contributing to PCP incidence and disease severity is needed. In one recent case-crossover study Djawe et al. observed an association between high AAP levels and increased hospital admissions for PCP among HIV-infected (HIV+) patients [16]. The underlying mechanisms for this increased susceptibility for PCP hospitalizations with AAP exposure remain unknown, but mechanisms of AAP induced immunotoxicity seem most plausible. As Pneumocystis is extremely difficult to grow in culture, we have developed an enzyme-linked immunosorbent assay (ELISA) using recombinant Pneumocystis jirovecii major surface glycoprotein (Msg) to assess serologic responses to P. jirovecii [17-21]. We selected this assay for the current study because it has been validated in diverse cohorts worldwide [22,23] and has been informative in assessing immunologic responses to Pneumocystis in those with HIV infection [24].

The mechanisms by which ambient air pollution increases susceptibility to pulmonary infection are not well defined. Single pollutant controlled exposure studies using cells, animals and human subjects have found that ambient air pollutants alter innate lung immunity at multiple levels [25] including altered mucociliary function [26], respiratory epithelial cell dysfunction [27,28], impaired alveolar macrophage phagocytosis [29-32], and surfactant protein A and D dysfunction [33]. However, the effects of AAP on humoral immunity and serologic responses to pulmonary infection, and the immunotoxic effects of real-life exposures to AAP remain unclear.

To address these areas of uncertainty we proposed two hypotheses for this study. First, real-life exposures to AAP suppress IgM antibody responses to P. jirovecii Msg, and second, AAP exposures, mediated by suppressed serologic responses to P. jirovecii Msg, are associated with adverse clinical outcomes. To test these hypotheses we conducted a prospective cohort study of HIV+ patients residing in San Francisco and hospitalized at San Francisco General Hospital with confirmed PCP.

Methods

Study population

From June 2000 to September 2007 we enrolled consecutive HIV+ adults (age > 18 years) hospitalized at San Francisco General Hospital with suspected PCP into a longitudinal cohort [34]. From this cohort, those with a definitive microscopic diagnosis of PCP and residence within the San Francisco city limits were eligible for inclusion in our prospective study. We excluded those with prior enrollment into the cohort and those without an available serologic specimen within 2 days of hospital admission. Hospital admission criteria for PCP included a supplemental oxygen requirement and/or poor oral intake, while discharge criteria included oxygen saturations > 90% while breathing room air and adequate oral intake.

Clinical and demographic data collection

After obtaining written, informed consent, we administered a standardized questionnaire to all participants, followed them throughout their hospitalization, and performed chart abstraction upon discharge from the hospital to obtain the following clinical data: age, sex, race/ethnicity, residential address, cigarette smoking status, prior history of PCP, antiretroviral medication and PCP prophylaxis, CD4 cell count, HIV viral load, length of hospital stay, admission to intensive care unit, intubation, and in-hospital mortality.

Ambient air pollutant measurements

We evaluated criteria pollutants most implicated in cardiopulmonary disease: particulate matter of < 10μm (PM10) and < 2.5μm in diameter (PM2.5), nitrogen dioxide (NO2), ozone (O3), and sulfur dioxide (SO2). The Bay Area Air Quality Management District monitors air pollutant concentrations in San Francisco at a centrally located monitoring station on 10 Arkansas Street for U.S. Environmental Protection Agency (EPA) regulatory compliance purposes. We used these publically available data to assign AAP exposures for each participant immediately prior to hospital admission and 2 weeks prior to admission. We chose these two exposure periods a priori based on biological plausibility of the timing of these exposures being associated with serologic responses to P. jirovecii at admission. For each period, we calculated the average exposure over 3 days. For the period immediately prior to admission, we averaged exposures from the day of admission, 1 day prior to admission, and 2 days prior to admission (days 0, -1, and -2). For the period 2 weeks prior to admission, we averaged exposures from days 14, 15, and 16 prior to admission (days -14, -15, and -16). The air pollutant metrics used in the analysis were as follows: daily mean PM10 and PM2.5 concentrations (μg/m3), daily maximum 1-hour NO2 concentrations in parts per billion (ppb), daily maximum 8-hour O3 concentrations (ppb), and daily maximum 1-hour SO2 concentrations (ppb).

Recombinant antigens

Using a λgt11 clone of Msg gene obtained from human infected lung as our template, we amplified the gene segment Msg2015-3332 (subscript denotes number of nucleotides) using PCR with AmpliTaq enzyme (Applied Biosystems, Carlsbad, CA, USA) [18]. We then inserted the amplified gene segment into the pET30 E. coli expression system (Novagen, Madison, WI, USA) to generate MsgC1 recombinant protein. We purified this protein by affinity chromatography. Using a similar process and P. jirovecii-infected human lung as our template, we generated MsgC3, MsgC8, and MsgC9 recombinant proteins from the same gene region as for MsgC1 [20]. These MsgC constructs are representative of the most antigenically conserved carboxyl-terminus region of the major surface glycoprotein [17].

IgM ELISA

We collected serum from each subject within two days of hospital admission and stored the sera at -80°C until shipment to University of Cincinnati for IgM ELISA analysis. Using a previously described protocol developed in our lab [21], we measured serum IgM responses to the four recombinant MsgC constructs; MsgC1, MsgC3, MsgC8, and MsgC9. We tested subject serum specimens and standard reference sera against the recombinant Msg constructs, using phosphate-buffered saline (PBS) without Msg as the negative control. We corrected the reactivity of each serum specimen to Msg by subtracting the reactivity of each serum specimen to PBS (mean optical density (OD) with Msg – mean OD without Msg) and quantified the results using methods described by Bishop and Kovacs [35]. We prepared a standard serum with specificity for each Msg construct by mixing the sera from 4 to 6 specimens with high reactivity for the specific construct. We selected these specimens from banks of sera from blood donors and HIV+ patients. The standard pool for each Msg construct was defined as having a value of 100 U in 100 μl of a 1:100 dilution. We used the same standard pools throughout the study. From the standard pool, we generated a standard curve for each Msg construct on each day the assay was used. We used this curve to calculate the units of reactivity to the Msg construct. We diluted test serum samples at 1:100 to fit the linear portion of the standard curves. Taking into account the dilution, we then calculated units of reactivity.

Ethics approval

The study was approved by the institutional review boards at the University of California San Francisco and the University of Cincinnati.

Statistical analysis

We evaluated associations between ambient air pollutant concentrations and IgM antibody responses to Msg constructs using multivariable tobit regression analysis for left censored data. Our predictors were exposures to PM10, PM2.5, NO2, O3, and SO2. Our primary outcome variables were log-transformed serum IgM antibody levels (in units of reactivity) to each Msg construct (MsgC1, MsgC3, MsgC8, and MsgC9). We log-transformed outcome variables to best conform to regression model assumptions. We tested for the effect modification of CD4 cell count and active cigarette smoking using interaction terms in the tobit regression analysis. We selected these variables as potential effect modifiers based on prior literature [24,36] and the biologic plausibility of their interactions with air pollutants in their associations with IgM responses to Msg constructs. Our pre-specified criteria for reporting effect modification were a p < 0.20 and a difference between stratum specific regression coefficients of at least 20%. We scaled and exponentiated regression coefficients to express logged outcomes as percent change in IgM response to each Msg construct per 10 unit increase in air pollutant concentration. We also used multivariable tobit regression analysis to evaluate the associations between clinical factors (CD4 cell count, viral load, and active cigarette smoking) and IgM responses, adjusting for a priori selected potential confounders (PM10, NO2, smoking, and CD4 cell count).

We evaluated associations between air pollutant (PM10, PM2.5, NO2, O3, and SO2) concentrations and clinical outcomes using multivariable linear regression for the continuous outcome (length of hospital stay) and logistic regression for the binary outcomes (ICU admission, intubation, and in-hospital mortality), evaluating for effect modification of log-transformed CD4 cell count and active cigarette smoking as detailed above.

To evaluate the association of IgM responses with clinical outcomes, we used linear regression for the continuous outcome (length of hospital stay) and logistic regression for the binary outcomes (ICU admission, intubation, and in-hospital mortality), adjusting for log-transformed CD4 cell count, active cigarette smoking, and PCP prophylaxis as a priori selected potential confounders. Finally, we performed a mediation analysis to determine the indirect effects of IgM responses on clinical outcomes compared with the direct effects of air pollutants on clinical outcomes.

Results

Subjects and ambient air pollution

We enrolled 283 consecutive HIV+ patients with suspected PCP into the larger cohort study. A definitive microscopic diagnosis of PCP was made in 180 of these patients, 162 of whom were being enrolled into the cohort with PCP for the first time (Figure 1). Of these, 88 subjects with PCP had serum available within 2 days of admission on which IgM serologic analysis was performed. The majority were male (n=77, 88%), white (n=48, 55%), and active cigarette smokers (n=59, 67%) (Table 1). Only 10 (11%) were taking antiretroviral therapy (ARV) and eight (9%) taking PCP prophylaxis at the time of admission. The median CD4 cell count was 33 cells/μl (interquartile range 10-77 cells/μl) and the mean log HIV RNA level was 11.3 copies/ml (standard deviation ± 2.0 copies/ml), reflecting a severely immunosuppressed cohort with uncontrolled viremia. Overall, nine patients required ICU admission (10%) and four (5%) died during the hospitalization. Among first-time enrollees with confirmed PCP (n=162), we compared those patients excluded (n=74) to those included in the study (n=88) and found no statistically significant differences in baseline clinical characteristics or clinical outcomes (data not shown).

Figure 1. Participant enrollment.

Figure 1

Open in a new tab

Table 1. Cohort demographics and clinical characteristics (n=88).

Characteristic
n (%)*
Median age (IQR) 40 (36–46)
Male sex 77 (88)
Race/ethnicity: White 48 (55)
Black 29 (33)
Other 11 (13)
Active tobacco smoker 59 (67)
Homeless 16 (18)
ARV use 10 (11)
PCP prophylaxis use 8 (9)
Median CD4 cell count, cells/μl (IQR) 33 (10–77)
Log viral load ± SD 11.3 ± 2.0
Median length of stay, days (IQR)† 7 (5-10)
Admission to ICU 9 (10)
Intubation 7 (8)
In-hospital mortality 4 (5)
Open in a new tab

SD indicates one standard deviation; IQR, interquartile range; ICU, intensive care unit; ARV, antiretroviral medication.

* Results in parentheses represent percentages unless otherwise indicated.

† Among those who survived hospitalization (n=84)

All study participants resided within the San Francisco city limits. Specific addresses were available for 65 subjects (74%) (Figure 2). All of these resided within 9 km of the monitoring station, with the majority (n=36, 55%) residing within a 3km radius of the station. Three-day average exposures to ambient air pollutants immediately prior to and 2 weeks prior to hospital admission varied considerably over the study period (Figures 3, 4), and were much less than the maximum allowable concentrations as defined by National Ambient Air Quality Standards (NAAQS) (Table 2). SO2 concentrations were particularly low throughout the study period.

Figure 2. Map of San Francisco showing geographic distribution of participants (n=65).*.

Figure 2

Open in a new tab

Legend: Blue circle— EPA compliance air quality monitoring station; Black circle—study site (San Francisco General Hospital); Red circles—residential area for each participant.

*Residential addresses not available for 23 subjects.

Map created at www.openstreetmap.org. Cartography licensed as CC-BY-SA.

Figure 3. Three-day average particulate concentrations immediately prior to each participant’s hospital admission.

Figure 3

Open in a new tab

Figure 4. Three-day average NO2, O3, and SO2 concentrations immediately prior to each participant’s hospital admission.

Figure 4

Open in a new tab

Table 2. Mean air pollution exposure concentrations for study participants compared with NAAQS (n=88).

Pollutant concentration ± SD
Pollutant Averaging time Immediately prior to admission*       2 weeks prior to admission† NAAQS
PM10 (μg/m3) 24-hour 22.6 ± 12.9 21.6 ± 11.1 150
PM2.5 (μg/m3) 24-hour 11.3 ± 8.3 9.7 ± 7.5 35
NO2 (ppb) 1-hour 18.0 ± 10.5 15.9 ± 8.3 100
O3 (ppb) 8-hour 19.4 ± 9.9 21.0 ± 7.9 75
SO2 (ppb) 1-hour 2.1 ± 1.4 1.9 ± 1.2 75
Open in a new tab

SD indicates one standard deviation; NAAQS, National Ambient Air Quality Standards; PM10, particulate matter of < 10 µm in diameter; PM2.5, particulate matter of < 2.5 µm in diameter; NO2, nitrogen dioxide; O3, ozone; SO2, sulfur dioxide.

* 3-day average: the day of admission, 1 day prior to admission, and 2 days prior to admission

† 3-day average: days 14, 15, and 16 prior to admission

Association between AAP and IgM

Elevated 3-day average ambient air concentrations of PM10 and NO2 immediately prior to hospital admission were associated with decreased IgM responses to all of the analyzed Msg constructs (Table 3). For every 10 μg/m3 increase in PM10 there was a 25% decrease in IgM response to MsgC1 (p=0.002), a 35% decrease in IgM to MsgC3 (p=0.009), a 33% decrease in IgM to MsgC8 (p=0.009), and a 25% decrease in IgM response to MsgC9 (p=0.02). We found a similar trend for NO2: elevated ambient air concentrations of NO2 were associated with decreased IgM responses to all the Msg constructs. These results reached statistical significance for MsgC8 and MsgC9. For every 10 ppb increase in NO2 there was a 45% decrease in IgM response to MsgC8 (p<0.001) and a 44% decrease IgM response to MsgC9 (p<0.001). On the other hand, we found no statistically significant associations or consistent trends between PM2.5, O3, or SO2 immediately prior to hospital admission and IgM responses to Msg (data not shown).

Table 3. Percent change in IgM responses to Msg constructs per 10 unit increase in pollutant exposure immediately prior to hospital admission and 2 weeks prior to admission (n=88).

MsgC1
MsgC3
MsgC8
MsgC9
Pollutant (units)     Exposure period (days prior to admission)* % (95%CI) p value % (95%CI) p value % (95%CI) p value % (95%CI) p value
PM10 (μg/m3) 0 to -2 -25 (-41 to -9.4) 0.002 -35 (-61 to -9.1) 0.009 -33 (-57 to -8.5) 0.009 -25 (-46 to -4.1) 0.02
-14 to -16 0.64 (-25 to 26) 0.96 -45 (-73 to -18) 0.002 -34 (-62 to -6.9) 0.02 -22 (-49 to 4.0) 0.10
NO2 (ppb) 0 to -2 -19 (-41 to 3.2) 0.09 -30 (-62 to 3.4) 0.08 -45 (-68 to -23) <0.001 -44 (-63 to -24) <0.001
-14 to -16 -15 (-44 to 14) 0.30 -41 (-77 to -5.3) 0.03 -63 (-84 to -42) <0.001 -37 (-65 to -10) 0.008
Open in a new tab

Msg indicates major surface glycoprotein; CI, confidence interval; PM10, particulate matter < 10 μm in diameter; NO2, nitrogen dioxide.

* Where the day of hospital admission is indicated by “day 0”

Similarly, we found that elevated 3-day average concentrations of PM10 and NO2 in San Francisco 2 weeks prior to admission were associated with decreased IgM responses to MsgC, but were statistically significant for fewer MsgC constructs (Table 3). Every 10 μg/m3 increase in PM10 predicted a 45% decrease in IgM response to MsgC3 (p=0.002) and a 34% decrease in IgM to MsgC8 (p=0.02). Likewise, every 10 ppb increase in NO2 predicted a 41% decrease in IgM response to MsgC3 (p=0.03), a 63% decrease in IgM response to MsgC8 (p<0.001) and a 37% decrease in IgM response to MsgC9 (p=0.008). As with exposures immediately prior to admission, we did not find statistically significant associations between PM2.5, O3, or SO2 two weeks prior to admission and IgM responses to Msg.

Effect of cigarette smoking on the association between AAP and IgM

Active cigarette smoking appeared to augment the immunosuppressive effects of PM10 and NO2 exposure 2 weeks prior to admission on IgM responses to Msg (data stratified by smoking status not shown). For every 10 μg/m3 increase in PM10 smokers experienced a larger decline in IgM responses to all the constructs compared with nonsmokers. This interaction was statistically significant for MsgC1 (p=0.04), MsgC3 (p=0.03), MsgC8 (p=0.01), and of borderline statistical significance for MsgC9 (p=0.07). A similar interaction was seen between cigarette smoking and NO2 exposure 2 weeks prior to admission. For every 10 ppb increase in NO2, smokers experienced a larger decline in IgM responses to all of the Msg constructs compared with nonsmokers. This interaction was statistically significant for MsgC8 (p=0.03) and MsgC9 (p=0.01). Smoking, on the other hand, did not appear to be directly associated with IgM responses in bivariate analyses (data not shown) or after adjusting for CD4 cell count, PM10, and NO2 (see Table 4).

Table 4. Percent change in IgM responses to Msg constructs by clinical factors (n=88).*.

MsgC1
MsgC3
MsgC8
MsgC9
Clinical factors % (95%CI) p value % (95%CI) p value % (95%CI) p value % (95%CI) p value
CD4 cell count† 12 (2.5 to 22) 0.02 9.4 (-7.3 to 26) 0.27 16 (1.5 to 31) 0.03 9.3 (-3.1 to 22) 0.14
Viral load† -2.3 (-8.0 to 3.4) 0.42 -5.9 (-15 to 3.6) 0.22 -6.8 (-15 to 1.1) 0.09 -0.5 (-8.0 to 7.0) 0.90
Active smoking# 2.3 (-58 to 62) 0.94 34 (-105 to 172) 0.63 -18 (-89 to 52) 0.61 -9.8 (-79 to 59) 0.78
Open in a new tab

* Percent change in IgM responses to Msg constructs per 50% increase in CD4, 50% increase in viral load, or for active cigarette smoking status

† Adjusted for active smoking, and PM10 and NO2 immediately and 2 weeks prior to admission

# Adjusted for CD4 cell count, and PM10 and NO2 immediately and 2 weeks prior to admission

Indicators of more intact immunity were associated with increased IgM responses to Msg (Table 4). For every 50% increase in CD4 cell count there was a 9.3% to 16% increase in IgM responses to Msg, statistically significant for MsgC1 (p=0.02) and MsgC8 (p=0.03). On the other hand, for every 50% increase in viral load there was a non-statistically significant decrease in IgM responses to Msg of 0.5% to 6.8%.

Association between AAP and clinical outcomes

Higher IgM responses to all the Msg constructs were associated with shorter length of hospital stay, with each 50% increase in IgM associated with a 0.36 to 0.90 day decrease in length of stay (Table 5). However, this was statistically significant only for MsgC8 (p=0.03). IgM responses to Msg were not associated with other clinical outcomes (ICU admission, intubation, or in-hospital mortality), nor were 3 day average air pollution exposures immediately and 2 weeks prior to hospital admission associated with length of stay, ICU admission, intubation, or in-hospital mortality. Similarly, the ambient air pollutants were not associated with other surrogates for PCP severity, such as serum lactate dehydrogenase (LDH). We also evaluated the association of each ambient air pollutant with CD4 cell count as the clinical outcome, and found no statistically significant associations between AAP and CD4 cell count (data not shown).

Table 5. Change in length of hospital stay (days) per 50% increase in IgM response to Msg among those who survived to hospital discharge (n=84).*.

IgM response to Msg
Length of stay
Construct Geometric mean, units (95% CI) # days (95%CI) p value
MsgC1 14.2 (11.0–18.4) -0.36 (-1.2 to 0.53) 0.42
MsgC3 3.8 (2.8–5.1) -0.53 (-1.3 to 0.21) 0.16
MsgC8 3.5 (2.6–4.6) -0.90 (-1.7 to -0.10) 0.03
MsgC9 5.7 (4.3–7.6) -0.57 (-1.3 to 0.19) 0.14
Open in a new tab

* Adjusted for log CD4 cell count, active cigarette smoking, and PCP prophylaxis.

Discussion

To our knowledge, this is the first prospective cohort study to evaluate the association of ambient air pollution exposures on serologic responses to Pneumocystis jirovecii and clinical outcomes in HIV+ patients hospitalized with confirmed PCP. We found that elevated ambient concentrations of PM10 and NO2 were associated with suppressed IgM responses to recombinant Pneumocystis major surface glycoprotein, and that this suppression of responses may be associated with increased length of hospital stay. Smoking appeared to augment the immunosuppressive effects of PM10 and NO2.

PM10 and NO2 were specifically associated with suppressed IgM responses, while PM2.5, O3, and SO2 were not. When designing this study, we hypothesized that particulate matter would play a significant role in altering serologic responses. Particle deposition in the lungs is size-dependent; so-called “fine” particles < 2.5 μm in aerodynamic diameter are inhaled to the alveolar level, whereas somewhat larger particles from 2.5 to 10 μm in diameter (the so-called “coarse” fraction) tend to deposit on lower airway mucociliary surfaces. As PM10 includes both fine and coarse particles it is possible that PM10 exerts its immunotoxicity in both airway and alveolar compartments. Coarse particles encountering bronchus associated lymphoid tissue, for instance, could impair antigen presenting cells, resulting in decreased activation of the humoral immune system and suppressed serologic responses, while fine particles could exert immunotoxic effects on alveolar macrophages, also resulting in impaired humoral activation. Periodic northeasterly (“Diablo”) winds bring coarse particulate matter of crustal origin into San Francisco. Such particulate matter can be highly bioactive. It is possible that the greater magnitude and variability in PM10 exposure afforded by these local weather patterns contributed to more significant findings for PM10 than PM2.5. Similarly, NO2 could be immunotoxic at the alveolar level or, freely diffusible, directly toxic to B lymphocytes. SO2 levels were clearly too low in this study to be implicated in immunologic derangement or clinical outcomes. San Francisco has exceedingly low SO2 pollution due to the absence of coal-powered power plants and industry in the city as well as California regulations requiring the use of low sulfur fuel. Ozone concentrations are also very low in San Francisco due to prevailing westerly winds that carry ozone precursors from motor vehicle emissions eastward across the Bay. PM10 and NO2 levels were statistically significantly associated with most, but not all, Msg constructs. It is unclear, for instance, why associations between NO2 and the Msg constructs MsgC8 and MsgC9 were more statistically significant than for MsgC1 and MsgC3.

We selected the timing of the periods during which air pollutant exposures were assigned to subjects based on a priori hypothesized mechanisms of immunotoxicity. We posited that pollutants with capacity to cross the alveolar capillary interface into blood circulation (e.g., NO2) could have early impact on B cell production of Pneumocystis jirovecii specific IgM. On the other hand, we hypothesized that coarse particulate matter, as mentioned above, would exert its effects predominantly via impaired antigen presentation and delayed humoral activation, an effect that might take 1 to 2 weeks to fully realize. Supportive of these mechanistic hypotheses, we found PM10 and NO2 exposures both immediately prior to hospital admission and 2 weeks prior to admission to be associated with suppressed IgM responses. These were 3-day averaged exposures during the two periods, and it remains unclear whether peak exposures might have independent associations with IgM suppression. However, it would be technically challenging to test this hypothesis in San Francisco given the relative absence of high peak pollutant levels.

As expected, smoking appeared to interact with PM10 and NO2 to augment suppression of serologic responses to Msg, a finding supported by prior studies. For instance, Crothers et al., using similar recombinant Msg and ELISA protocols, recently found that, among a cohort of active heavy smokers, smoking was associated with decreased serologic responses to Pneumocystis Msg compared with non-smoking controls [36]. Many of the pollutants in cigarette smoke are similar to those in ambient air pollution, but at much greater concentrations. Given these properties, we expected smoking to also be independently associated with suppressed IgM, which it was not in this study. It was the smoking—AAP interaction, and not smoking alone, that was associated with suppressed IgM responses. Why we failed to find an independent effect of smoking in this study is unclear.

Although cellular immune responses are predominantly involved in effectively countering Pneumocystis jirovecii infections, it is well known that humoral immunity plays an essential role [37-46]. For instance, B cell deficient mice succumb to Pneumocystis infection much more rapidly than their immunocompetent controls [41]. In humans, cellular immunosuppression, caused by immunosuppressive medications and conditions such as HIV infection, play a prominent role in the increased risk for PCP, but deficiencies of humoral responses, such as congenital X-linked agammaglobulinemia, may also impart an increased risk [47,48]. Given the known importance of humoral immunity of Pneumocystis infection in animals and humans, we hypothesized that ambient air pollution exposures, mediated by suppressed IgM responses to Pneumocystis Msg, would be associated with adverse clinical outcomes. Although we found a trend for decreased IgM responses associated with increased length of hospital stay, the precise reasons for this finding are unclear as length of stay is influenced by numerous factors. IgM responses were not associated with the other measured clinical outcomes, nor were ambient air pollutants associated with clinical outcomes. It is important to note that the number of clinical endpoints in our study were too few to adequately test this clinical hypothesis.

Our study has several limitations. Ambient air pollution exposures were estimated from daily levels measured by the single EPA compliance air quality monitoring station in San Francisco. Fortunately, the majority of our study participants resided within a 3 km radius of this station such that using the air quality data from this station to estimate subject exposures was reasonable. Our sample size was small, and we were unable to include a number of patients with confirmed PCP due to lack of serum specimens drawn within 2 days of admission. However, we are confident that the excluded cohort did not significantly bias our study findings or alter our conclusions given the similar clinical characteristics and comparable clinical outcomes among included and excluded cohorts. Finally, San Francisco experiences relatively low levels of all the criteria ambient air pollutants. Perhaps we would have found more pronounced immunotoxic effects and clinical outcomes in a more polluted city.

In conclusion, elevated 3-day average PM10 and NO2 ambient air concentrations immediately prior to hospital admission and 2 weeks prior to admission were associated with suppressed IgM antibody responses to recombinant Pneumocystis jirovecii major surface glycoprotein in HIV+ patients hospitalized with confirmed PCP, and active cigarette smoking appeared to augment the immunosuppressive effects of these pollutants. These provocative preliminary findings suggest a mechanism of immunotoxicity by which ambient air pollution increases host susceptibility to pulmonary infection. A larger prospective cohort study with more precise individual-level exposure assessments and evaluating both IgM and IgG responses is needed to confirm these findings and to better assess their clinical significance.

Acknowledgments

We would like to acknowledge the patients, staff, and administration at San Francisco General Hospital who participated in this study. We would also like to acknowledge Joan Turner, Shary Eiser, Matthew Bensley, Alexandra Swartzman, and Brenna Roth for their assistance in subject enrollment and data collection, Dr. Judy Koch for her Msg assay contributions, and Dr. Mark Segal for his biostatistical support.

Funding Statement

The work was supported by the following: National Institute of Environmental Health Science (NIEHS) www.niehs.nih.gov, F32ES022582(Blount); National Heart, Lung, and Blood Institute (NHLBI) www.nhlbi.nih.gov, T32HL007185 (Sheppard, Nadel, Huang), K24HL087713 (Huang), R01HL090335 (Huang). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

References

  • 1. Lim SS, Vos T, Flaxman AD, Danaei G, Shibuya K et al. (2012) A comparative risk assessment of burden of disease and injury attributable to 67 risk factors and risk factor clusters in 21 regions, 1990-2010: a systematic analysis for the global burden of disease study 2010. Lancet 380: 2224-2260. doi: 10.1016/S0140-6736(12)61766-8. PubMed: 23245609. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2. Schwartz J (1994) Air pollution and hospital admissions for the elderly in Birmingham, Alabama. Am J Epidemiol 139: 589-598. PubMed: 8172170. [DOI] [PubMed] [Google Scholar]
  • 3. Schwartz J (1994) PM10, ozone, and hospital admissions for the elderly in Minneapolis-St. Paul, Minnesota. Arch Environ Health 49: 366-374. doi: 10.1080/00039896.1994.9954989. PubMed: 7944569. [DOI] [PubMed] [Google Scholar]
  • 4. Fusco D, Forastiere F, Michelozzi P, Spadea T, Ostro B et al. (2001) Air pollution and hospital admissions for respiratory conditions in Rome, Italy. Eur Respir J 17: 1143-1150. doi: 10.1183/09031936.01.00005501. PubMed: 11491157. [DOI] [PubMed] [Google Scholar]
  • 5. Peel JL, Tolbert PE, Klein M, Metzger KB, Flanders WD et al. (2005) Ambient air pollution and respiratory emergency department visits. Epidemiology 16: 164-174. doi: 10.1097/01.ede.0000152905.42113.db. PubMed: 15703530. [DOI] [PubMed] [Google Scholar]
  • 6. Zanobetti A, Schwartz J (2006) Air pollution and emergency admissions in Boston, MA. J Epidemiol Community Health 60: 890-895. doi: 10.1136/jech.2005.039834. PubMed: 16973538. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7. Cheng MF, Tsai SS, Wu TN, Chen PS, Yang CY (2007) Air pollution and hospital admissions for pneumonia in a tropical city: Kaohsiung, Taiwan. J Toxicol Environ Health A 70: 2021-2026. doi: 10.1080/15287390701601020. PubMed: 18049991. [DOI] [PubMed] [Google Scholar]
  • 8. Chiu HF, Cheng MH, Yang CY (2009) Air pollution and hospital admissions for pneumonia in a subtropical city: Taipei, Taiwan. Inhal Toxicol 21: 32-37. doi: 10.1080/08958370802441198. PubMed: 18923947. [DOI] [PubMed] [Google Scholar]
  • 9. Harris AM, Sempértegui F, Estrella B, Narváez X, Egas J et al. (2011) Air pollution and anemia as risk factors for pneumonia in Ecuadorian children: A retrospective cohort analysis. Environ Health 10: 93. doi: 10.1186/1476-069X-10-93. PubMed: 22050924. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10. Katanoda K, Sobue T, Satoh H, Tajima K, Suzuki T et al. (2011) An association between long-term exposure to ambient air pollution and mortality from lung cancer and respiratory diseases in Japan. J Epidemiol 21: 132-143. doi:10.2188/jea.JE 20100098. PubMed: 21325732 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11. Neupane B, Jerrett M, Burnett RT, Marrie T, Arain A et al. (2010) Long-term exposure to ambient air pollution and risk of hospitalization with community-acquired pneumonia in older adults. Am J Respir Crit Care Med 181: 47-53. doi: 10.1164/rccm.200901-0160OC. PubMed: 19797763. [DOI] [PubMed] [Google Scholar]
  • 12. Zanobetti A, Schwartz J, Dockery DW (2000) Airborne particles are a risk factor for hospital admissions for heart and lung disease. Environ Health Perspect 108: 1071-1077. doi: 10.1289/ehp.001081071. PubMed: 11102299. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13. Schwartz J, Bellinger D, Glass T (2011) Expanding the scope of environmental risk assessment to better include differential vulnerability and susceptibility. Am J Public Health 101 Suppl 1: S88-S93. doi: 10.2105/AJPH.2011.300268. PubMed: 22021314. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14. Cheng MF, Tsai SS, Chiu HF, Sung FC, Wu TN et al. (2009) Air pollution and hospital admissions for pneumonia: are there potentially sensitive groups? Inhal Toxicol 21: 1092-1098. doi: 10.3109/08958370902744855. PubMed: 19852550. [DOI] [PubMed] [Google Scholar]
  • 15. Huang L, Cattamanchi A, Davis JL, den Boon S, Kovacs J et al. (2011) HIV-associated Pneumocystis pneumonia. Proc Am Thorac Soc 8: 294-300. doi: 10.1513/pats.201009-062WR. PubMed: 21653531. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16. Djawe K, Levin L, Swartzman A, Fong S, Roth B et al. (2013) Environmental risk factors for Pneumocystis pneumonia hospitalizations in HIV patients. Clin Infect Dis 56: 74-81. doi: 10.1093/cid/cis841. PubMed: 23042978. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17. Mei Q, Turner RE, Sorial V, Klivington D, Angus CW et al. (1998) Characterization of major surface glycoprotein genes of human Pneumocystis carinii and high-level expression of a conserved region. Infect Immun 66: 4268-4273. PubMed: 9712777. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18. Daly KR, Fichtenbaum CJ, Tanaka R, Linke MJ, O'Bert R et al. (2002) Serologic responses to epitopes of the major surface glycoprotein of Pneumocystis jiroveci differ in human immunodeficiency virus-infected and uninfected persons. J Infect Dis 186: 644-651. doi: 10.1086/341565. PubMed: 12195351. [DOI] [PubMed] [Google Scholar]
  • 19. Daly KR, Koch J, Levin L, Walzer PD (2004) Enzyme-linked immunosorbent assay and serologic responses to Pneumocystis jiroveci . Emerg Infect Dis 10: 848-854. doi: 10.3201/eid1005.030497. PubMed: 15200818. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20. Daly KR, Koch JV, Shire NJ, Levin L, Walzer PD (2006) Human immunodeficiency virus-infected patients with prior Pneumocystis pneumonia exhibit increased serologic reactivity to several major surface glycoprotein clones. Clin Vaccine Immunol 13: 1071-1078. doi: 10.1128/CVI.00140-06. PubMed: 17028210. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21. Djawe K, Huang L, Daly KR, Levin L, Koch J et al. (2010) Serum antibody levels to the Pneumocystis jirovecii major surface glycoprotein in the diagnosis of P. jirovecii pneumonia in HIV+ patients. PLOS ONE 5: e14259. doi: 10.1371/journal.pone.0014259. PubMed: 21151564. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22. Daly K, Koch J, Respaldiza N, de la Horra C, Montes-Cano MA et al. (2009) Geographical variation in serological responses to recombinant Pneumocystis jirovecii major surface glycoprotein antigens. Clin Microbiol Infect 15: 937-942. doi: 10.1111/j.1469-0691.2009.02716.x. PubMed: 19416292. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23. Walzer PD, Djawe K, Levin L, Daly KR, Koch J et al. (2009) Long-term serologic responses to the Pneumocystis jirovecii major surface glycoprotein in HIV-positive individuals with and without P. jirovecii infection. J Infect Dis 199: 1335-1344. doi: 10.1086/597803. PubMed: 19301979. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24. Blount RJ, Jarlsberg LG, Daly KR, Worodria W, Davis JL et al. (2012) Serologic responses to recombinant Pneumocystis jirovecii major surface glycoprotein among Ugandan patients with respiratory symptoms. PLOS ONE 7: e51545. doi: 10.1371/journal.pone.0051545. PubMed: 23284710. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25. Al-Hegelan M, Tighe RM, Castillo C, Hollingsworth JW (2011) Ambient ozone and pulmonary innate immunity. Immunol Res 49: 173-191. doi: 10.1007/s12026-010-8180-z. PubMed: 21132467. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26. Wolff RK (1986) Effects of airborne pollutants on mucociliary clearance. Environ Health Perspect 66: 223-237. doi: 10.1289/ehp.8666223. PubMed: 3519203. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27. Deng X, Zhang F, Rui W, Long F, Wang L et al. (2013) PM-induced oxidative stress triggers autophagy in human lung epithelial A549 cells. Toxicol Vitro 27: 1762-1770. doi: 10.1016/j.tiv.2013.05.004. [DOI] [PubMed] [Google Scholar]
  • 28. Manzo ND, LaGier AJ, Slade R, Ledbetter AD, Richards JH et al. (2012) Nitric oxide and superoxide mediate diesel particle effects in cytokine-treated mice and murine lung epithelial cells--implications for susceptibility to traffic-related air pollution. Part Fibre Toxicol 9: 43-8977-9-43 PubMed: 23151036. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29. Zhou H, Kobzik L (2007) Effect of concentrated ambient particles on macrophage phagocytosis and killing of Streptococcus pneumoniae . Am J Respir Cell Mol Biol 36: 460-465. doi: 10.1165/rcmb.2006-0293OC. PubMed: 17079778. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30. Yang HM, Antonini JM, Barger MW, Butterworth L, Roberts BR et al. (2001) Diesel exhaust particles suppress macrophage function and slow the pulmonary clearance of Listeria monocytogenes in rats. Environ Health Perspect 109: 515-521. doi: 10.1289/ehp.01109515. PubMed: 11401764. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31. Yin XJ, Dong CC, Ma JY, Roberts JR, Antonini JM et al. (2007) Suppression of phagocytic and bactericidal functions of rat alveolar macrophages by the organic component of diesel exhaust particles. J Toxicol Environ Health A 70: 820-828. doi: 10.1080/15287390701209766. PubMed: 17454558. [DOI] [PubMed] [Google Scholar]
  • 32. Lundborg M, Dahlén SE, Johard U, Gerde P, Jarstrand C et al. (2006) Aggregates of ultrafine particles impair phagocytosis of microorganisms by human alveolar macrophages. Environ Res 100: 197-204. doi: 10.1016/j.envres.2005.08.007. PubMed: 16171796. [DOI] [PubMed] [Google Scholar]
  • 33. Mikerov AN, Umstead TM, Gan X, Huang W, Guo X et al. (2008) Impact of ozone exposure on the phagocytic activity of human surfactant protein A (SP-A) and SP-A variants. Am J Physiol Lung Cell Mol Physiol 294: L121-L130. PubMed: 17981957. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34. Fei MW, Kim EJ, Sant CA, Jarlsberg LG, Davis JL et al. (2009) Predicting mortality from HIV-associated Pneumocystis pneumonia at illness presentation: an observational cohort study. Thorax 64: 1070-1076. doi: 10.1136/thx.2009.117846. PubMed: 19825785. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35. Bishop LR, Kovacs JA (2003) Quantitation of anti-Pneumocystis jiroveci antibodies in healthy persons and immunocompromised patients. J Infect Dis 187: 1844-1848. doi: 10.1086/375354. PubMed: 12792860. [DOI] [PubMed] [Google Scholar]
  • 36. Crothers K, Daly KR, Rimland D, Goetz MB, Gibert CL et al. (2011) Decreased serum antibody responses to recombinant Pneumocystis antigens in HIV-infected and uninfected current smokers. Clin Vaccine Immunol 18: 380-386. doi: 10.1128/CVI.00421-10. PubMed: 21191078. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37. Empey KM, Hollifield M, Schuer K, Gigliotti F, Garvy BA (2004) Passive immunization of neonatal mice against Pneumocystis carinii f. sp. muris enhances control of infection without stimulating inflammation. Infect Immun 72: 6211-6220. doi: 10.1128/IAI.72.11.6211-6220.2004. PubMed: 15501746. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38. Furuta T, Ueda K, Fujiwara K, Yamanouchi K (1985) Cellular and humoral immune responses of mice subclinically infected with Pneumocystis carinii . Infect Immun 47: 544-548. PubMed: 3871426. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39. Gigliotti F, Haidaris CG, Wright TW, Harmsen AG (2002) Passive intranasal monoclonal antibody prophylaxis against murine Pneumocystis carinii pneumonia. Infect Immun 70: 1069-1074. doi: 10.1128/IAI.70.3.1069-1074.2002. PubMed: 11854184. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40. Harmsen AG, Stankiewicz M (1991) T cells are not sufficient for resistance to Pneumocystis carinii pneumonia in mice. J Protozool 38: 44S-45S. doi: 10.1111/j.1550-7408.1991.tb04798.x. PubMed: 1687826. [DOI] [PubMed] [Google Scholar]
  • 41. Marcotte H, Levesque D, Delanay K, Bourgeault A, de la Durantaye R et al. (1996) Pneumocystis carinii infection in transgenic B cell-deficient mice. J Infect Dis 173: 1034-1037. doi: 10.1093/infdis/173.4.1034. PubMed: 8603947. [DOI] [PubMed] [Google Scholar]
  • 42. Walzer PD, Rutledge ME (1981) Humoral immunity in experimental Pneumocystis carinii infection. I. serum and bronchial lavage fluid antibody responses in rats. J Lab Clin Med 97: 820-833. PubMed: 7014742. [PubMed] [Google Scholar]
  • 43. Wells J, Haidaris CG, Wright TW, Gigliotti F (2006) Active immunization against Pneumocystis carinii with a recombinant P. carinii antigen. Infect Immun 74: 2446-2448. doi: 10.1128/IAI.74.4.2446-2448.2006. PubMed: 16552076. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44. Wells J, Haidaris CG, Wright TW, Gigliotti F (2006) Complement and Fc function are required for optimal antibody prophylaxis against Pneumocystis carinii pneumonia. Infect Immun 74: 390-393. doi: 10.1128/IAI.74.1.390-393.2006. PubMed: 16368994. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45. Gigliotti F, Harmsen AG (1997) Pneumocystis carinii host origin defines the antibody specificity and protective response induced by immunization. J Infect Dis 176: 1322-1326. doi: 10.1086/514128. PubMed: 9359734. [DOI] [PubMed] [Google Scholar]
  • 46. Zheng M, Shellito JE, Marrero L, Zhong Q, Julian S et al. (2001) CD4+ T cell-independent vaccination against Pneumocystis carinii in mice. J Clin Invest 108: 1469-1474. doi: 10.1172/JCI13826. PubMed: 11714738. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47. Dittrich AM, Schulze I, Magdorf K, Wahn V, Wahn U (2003) X-linked agammaglobulinaemia and Pneumocystis carinii pneumonia--an unusual coincidence? Eur J Pediatr 162: 432-433. PubMed: 12700897. [DOI] [PubMed] [Google Scholar]
  • 48. Alibrahim A, Lepore M, Lierl M, Filipovich A, Assaád A (1998) Pneumocystis carinii pneumonia in an infant with X-linked agammaglobulinemia. J Allergy Clin Immunol 101: 552-553. doi: 10.1016/S0091-6749(98)70363-X. PubMed: 9564810. [DOI] [PubMed] [Google Scholar]

Articles from PLoS ONE are provided here courtesy of PLOS

RESOURCES