Skip to main content
Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Exposure to a firefighting overhaul environment without respiratory protection increases immune dysregulation and lung disease risk

  • Stephen J. Gainey,

    Roles Conceptualization, Data curation, Formal analysis, Investigation, Writing – original draft, Writing – review & editing

    Affiliation Department of Animal Sciences, University of Illinois, Urbana, Illinois, United States of America

  • Gavin P. Horn,

    Roles Conceptualization, Funding acquisition, Methodology, Project administration, Resources, Writing – review & editing

    Affiliation Illinois Fire Service Institute, Champaign, Illinois, United States of America

  • Albert E. Towers,

    Roles Data curation, Investigation

    Affiliation Division of Nutritional Sciences, University of Illinois, Urbana, Illinois, United States of America

  • Maci L. Oelschlager,

    Roles Data curation, Investigation

    Affiliation Department of Pathology, Program in Integrative Immunology and Behavior, University of Illinois College of Medicine, Urbana, Illinois, United States of America

  • Vincent L. Tir,

    Roles Investigation

    Affiliation Department of Pathology, Program in Integrative Immunology and Behavior, University of Illinois College of Medicine, Urbana, Illinois, United States of America

  • Jenny Drnevich,

    Roles Data curation, Formal analysis, Visualization

    Affiliation Roy J. Carver Biotechnology Center, University of Illinois, Urbana, Illinois, United States of America

  • Kenneth W. Fent,

    Roles Conceptualization, Resources

    Affiliation Division of Surveillance, Hazard Evaluations, and Field Studies, National Institute for Occupational Safety and Health, Cincinnati, Ohio, United States of America

  • Stephen Kerber,

    Roles Conceptualization, Resources

    Affiliation Director, UL Firefighter Safety Research Institute, Columbia, Maryland, United States of America

  • Denise L. Smith,

    Roles Conceptualization, Resources

    Affiliations Illinois Fire Service Institute, Champaign, Illinois, United States of America, Department of Health and Human Physiological Sciences, Skidmore College, Saratoga Spring, New York, United States of America

  • Gregory G. Freund

    Roles Conceptualization, Investigation, Project administration, Resources, Supervision, Writing – original draft, Writing – review & editing

    Affiliations Department of Animal Sciences, University of Illinois, Urbana, Illinois, United States of America, Division of Nutritional Sciences, University of Illinois, Urbana, Illinois, United States of America, Department of Pathology, Program in Integrative Immunology and Behavior, University of Illinois College of Medicine, Urbana, Illinois, United States of America


Firefighting activities appear to increase the risk of acute and chronic lung disease, including malignancy. While self-contained breathing apparatuses (SCBA) mitigate exposures to inhalable asphyxiates and carcinogens, firefighters frequently remove SCBA during overhaul when the firegrounds appear clear of visible smoke. Using a mouse model of overhaul without airway protection, the impact of fireground environment exposure on lung gene expression was assessed to identify transcripts potentially critical to firefighter-related chronic pulmonary illnesses. Lung tissue was collected 2 hrs post-overhaul and evaluated via whole genome transcriptomics by RNA-seq. Although gas metering showed that the fireground overhaul levels of carbon monoxide (CO), carbon dioxide (CO2), hydrogen cyanine (HCN), hydrogen sulfide (H2S) and oxygen (O2) were within NIOSH ceiling recommendations, 3852 lung genes were differentially expressed when mice exposed to overhaul were compared to mice on the fireground but outside the overhaul environment. Importantly, overhaul exposure was associated with an up/down-regulation of 86 genes with a fold change of 1.5 or greater (p<0.5) including the immunomodulatory-linked genes S100a8 and Tnfsf9 (downregulation) and the cancer-linked genes, Capn11 and Rorc (upregulation). Taken together these findings indicate that, without respiratory protection, exposure to the fireground overhaul environment is associated with transcriptional changes impacting proteins potentially related to inflammation-associated lung disease and cancer.


Even as personal protective equipment (PPE) improves [1], the incidence and mortality from cancer in firefighters increases and is a leading cause of death [2]. Epidemiological evidence shows that firefighters have a greater risk of cancer when compared to the general population [2,3]. Firefighters in the United States respond to 1.2–1.4 million fires each year including approximately 475,000–500,000 structure fires [4]. Exposure to toxicants is possible during live fire responses, which can result in biological absorption of polycyclic aromatic hydrocarbons (PAHs) and benzene [57] and inhalation of carbon monoxide [8] and hydrogen cyanide [9]. Interestingly, in 2010, the International Agency for Research on Cancer (IARC) classified occupational exposure during firefighting as possibly carcinogenic to humans [10]. Part of the rational for this classification results from the lack of genotoxicity studies in animals that involves exposure to smoke from the combustion of structural materials. Even with substantial upgrades to PPE, such as SCBAs and turnout gear technology, firefighters are imperiled if SCBAs are compromised, not worn or removed [5,1113]. While exposure risk is minimized with PPE [14,15], PPE usage is not universal for all phases of a response.

Currently, the highest risk of toxicant exposure appears to be during overhaul, since initial fire suppression is usually associated with heavy smoke and the obvious need for SCBA [16]. During overhaul, time spent searching for unextinguished fire inside structures can exceed 30 minutes and is most often coupled to improper or little use of respiratory protection [11,16]. Unfortunately, failure to use PPE during overhaul can result in contact with concealed carcinogens (like asbestos) due to fire- or firefighting-dependent structural damage [16]. In addition, smoke and/or fume inhalation is most prevalent during this period due to frequent abandonment of SCBA [17,18]. While the U.S. Fire Service has gained traction in limiting removal of SCBA, firefighters still make their own determination on when to utilize it based on heat stress, comfort or visual indications of clear air. Therefore, the purpose of this study was to examine the impact of unprotected respiratory exposure to the fireground during overhaul on mouse lung gene expression. It should provide insight to potential pathways linked to lung cancer development.

Materials and methods


The use of animals (S1 Checklist) was in accordance with the recommendation in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health and an Institutional Animal Care and Use Committee (IACUC) approved protocol (Protocol #15099) at the University of Illinois. C57BL/6J male mice (10 weeks old) were purchased from Jackson Laboratories (Bar Harbor, ME). Mice were group-housed (4 per cage) in shoebox cages (length 29.9 cm; width 18.4 cm; height 12.5 cm) and allowed free access to food and water, unless otherwise noted. Housing temperature (22 °C) and humidity (45–55%) were controlled as was a 12/12 h reversed dark-light cycle (light = 1000–2200 h). Animals were euthanized for tissue collection using CO2. Total number of mice used was 54.

Live-fire scenario setup

Firefighting activities were conducted in a purpose-built live-fire research test structure. The structure, based on a design by a residential architectural company, was representative of a home constructed in the mid-twentieth century with walls and doorways separating all rooms and 2.4m ceilings. The structure had an approximate floor area of 111 m2 with 8 total rooms. Interior finish in the burn rooms was protected by gypsum board on the ceiling and walls. Furnishings were acquired from a lone source to afford inter-scenario standardization. The bedrooms, where the fires were ignited, were appointed with a double bed (covered with a foam mattress topper, comforter and pillow), stuffed chair, side table, lamp, dresser and flat screen television. Floors were covered with polyurethane foam padding and polyester carpet. Fires were ignited using the stuffed bedroom chair via remote ignition comprised of a match book electrically energized by fine wire heating. Each resultant flaming fire could grow until it approached early ventilation-limitation. Based on national averages, fire department dispatch was between 4–5 min after ignition for all scenarios. The structure was repaired/rebuilt after each scenario.

Firefighting and overhaul

A team of 12 firefighters battled the fires involving two fully involved bedrooms. As soon as the fire was suppressed, and interior operations were completed (two simulated trapped occupants removed), mice were transported into the burned structure as overhaul operations by firefighters were initiated.

Mouse groups/transport/housing

For each cohort (n = 18 mice/cohort), mice were placed into three groups (n = 6 each): 1) control (C) group, which never left the animal housing facility; 2) fireground (FG) group, which was taken to the fireground but placed in a portion of the structure that was uninvolved with the fire and overhaul activities: and 3) overhaul (OH) group, which was taken to the fireground and placed in the interior of the structure during overhaul (as described above). Three cohorts of mice were used, one for each of the three experiments performed on three separate days at approximately the same time of day (0800–0900). The FG and OH mouse groups were transported to the fireground, arriving 30 min prior to firefighting and were placed on a table approximately 25 m from the structure while active fire was being fought by the firefighters. Mice were housed in shoebox cages wrapped in heat-resistant AB Technology Group Knitted Fiberglass Plain Tape (Ogdensburg, NY) and 3M Silver Foil Tape 3340 (Maplewood, Minnesota), on 3.5 sides. Interior cage temperature was recorded using a Fisher Scientific (Hampton, NH) digital probe thermometer, and animals were visually monitored every 5 min throughout the exposure period for signs of pain or distress. Mouse groups were returned to the animal care facility 15 minutes after the conclusion of overhaul.

Atmospheric data collection

Air concentrations of carbon monoxide (CO), carbon dioxide (CO2), hydrogen cyanide (HCN), hydrogen sulfide (H2S), and oxygen (O2) gases were quantified with a MX6 iBrid (Industrial Scientific; Pittsburg, PA) portable personal gas monitor. The meter was placed on top of the mouse cages in one of the fire rooms being overhauled by firefighters.

RNA extraction and fragment analysis

Mouse lungs were harvested 2 hrs after the OH group was removed from the overhaul environment and immediately placed in Qiagen RNAlater (Valencia, CA). RNA was extracted using the Qiagen miRNeasy Mini Kit including DNAase. RNA integrity was determined using an Applied Biosystems Fragment analyzer (Foster City, CA); all 54 samples had RQN score >7 and were defined as acceptable.

Illumina RNA sequencing

RNAseq libraries were prepared with Illumina TruSeq Stranded RNA Sample Prep Kit (San Diego, CA) resulting in 5’ to 3’ strand-specific libraries. A single library was prepared from each sample. All libraries were then quantitated by qPCR and sequenced on seven lanes for 101 cycles using an Illumina HiSeq2500 100nt single-end read with the TruSeq SBS sequencing v3 kit. Fastq files were processed and demultiplexed with bcltofastq 1.8.4.

RNAseq data and statistical analysis

Raw reads were checked for quality using FASTQC (v 0.11.2) then trimmed and filtered using Trimmomatic (v 0.33) to remove residual adapter content and low-quality bases (Phred quality score < 28). Trimmed/filtered reads were aligned to NCBI’s Mus musculus GRCm38.p3 genome and gene model annotation release 105 using STAR (v 2.4.2a). Post-alignment gene counts were then determined using featureCounts (v 1.4.3-pl) with multi-mapping reads excluded.

The gene-level read counts were then imported into R (v. 3.4.3) for statistical analyses. TMM normalization (Robinson and Oshlack 2010) in the edgeR package (Robinson et al. 2010; v 3.20.6) was used to normalize the counts to log2-transformed counts per million (logCPM), using the cpm function with prior count = 3. 25,525 genes without logCPM > log2 (1) in at least 5 samples were filtered out, leaving 16,261 genes to be analyzed for differential expression. TMM-values were re-calculated as well as logCPM normalized values with prior.count = 3 to use in down-stream analyses and visualizations.

Clustering of samples to check for outliers and batch effects was done using Principle Components Analysis [19]. We then performed surrogate variables analysis (sva) [20,21] using the sva package (v 3.26.0) [22],) to detect and remove artifacts like batch effects by creating eight surrogate variables (sv). The sv were added to the statistical model for the 3 treatment groups and differential expression testing [23] using the limma package’s (v 3.34.5) [24] “trend” approach because the variation in library sizes was less than the recommended 3-fold maximum [25]. A one-way ANOVA across the 3 groups was calculated, along with all three pairwise comparisons. Multiple hypothesis testing adjustment was done separately for each test using the False Discovery Rate (FDR) method [26]. While the sva method was judged to be the best way to correct the overall FG vs OH comparison for individual fire and other partially confounded batch effects, it does not allow us to pull individual FG vs OH comparisons for each fire. Therefore, we also made a separate statistical analysis for the 9-different treatments X fire groups + seven estimated surrogate variables and pulled out pairwise FG vs OH comparisons within each fire. Because we were mainly interested in comparing the numbers of genes differentially expressed between fires, we performed a global FDR correction across the three comparisons to ensure that a gene with the same raw p-value in different fires ended up with the same FDR p-value.

Functional annotation was taken from Bioconductor’s [27] package (v 3.5.0) using the respective Entrez Gene ID from NCBI. KEGG pathways were downloaded directly from using the KEGGREST package (v 1.18.0). Over-representation testing was done on KEGG pathways for specified gene sets using the GOstats (v 2.44.0) [28] and Category (v 2.44.0) packages. Statistical significance was assumed at FDR p < 0.05 unless otherwise noted.


Overhaul environmental conditions

Table 1 shows that mouse cage temperature averaged 31.6 °C with differences between test fires ranging from 28.3–33.9 °C. Peak temperatures ranged from approximately 30.6–40.6 °C and occurred as the mice cages were introduced into the structure. Peak concentrations for CO2, HCN, H2S, and minimum level of O2 did not exceed the 10-hour NIOSH TWA levels. Peak CO did exceed NIOSH STEL and OSHA PEL TWA levels and remained under the NIOSH ceiling recommendation. Qualitatively, the overhaul environment appeared visually clear during each overhaul in contrast to dense smoke during the fire itself (data not shown).

Table 1. Environmental measurements for OH group mice during overhaul respective of mouse cohort/fire.

Principal component analysis (PCA) demonstrate distinct separations between C, FG and OH groups

To determine the transcriptomic relationship between exposure/control groups, high-throughput sequencing was used to delineate global gene expression. Table 2 indicates the number and types of 41,786 gene entities in the genome and the 16,261 genes remaining after filtering. Principle Components Analysis clustering after removing the effects of the eight surrogate variables (Fig 1) shows significant separation between all three groups (C, FG, and OH) based on the distance between clusters plotted on PC1 and PC2.

Table 2. Number of different gene entities in NCBI Mus musculus GRCm38.p3 gene annotations.

Fig 1. Principal component analysis of control, fireground, and overhaul gene expression data following surrogate variables removal.

Principal components 1 and 2 are shown with control samples are represented by circles (red color). The fireground samples are represented by squares (lime color) and the overhaul samples represented as diamonds (blue color). The numeric labels 1, 2, and 3 indicate the cohort.

Gene expression in lung after overhaul exposures is markedly different from fireground exposures

Table 3 shows the number of significant differentially expressed genes overall and broken down by each of the three fires. Overall, mice exposed to the overhaul environment resulted in a dramatically differential gene expression than mice kept at the fireground, modulating 3,852 genes. However, it is also apparent there was significant fire-to-fire variation in the gene expression, ranging from 3,460 on Fire 1 to 698 on Fire 3, although the majority of significantly changed genes on Fire 1 were trending the same direction on Fires 2 and 3, leading to overall FG vs. OH significance.

Table 3. Number of genes significantly up- or down- regulated (FDR p-value < 0.05) by overhaul (OH) exposure compared with fireground (FG).

Table 4 highlights these differentially expressed genes that display a greater than ± 50%-fold change (FC). This list consists of 43 up-regulated and 43 down-regulated genes. Importantly, the top 5 up-regulated genes link to cancer or immunomodulation, including calpain 11 (Capn11), immunoglobulin kappa chain variable 5–43 (Igkv5-43), immunoglobulin heavy constant alpha (Igha), immunoglobulin heavy variable 1–26 (Ighv1-26), and immunoglobulin heavy constant gamma 2B (Ighg2b) [2933]. In correlate, several down-regulated genes are important to immune and cancer defense, specifically tumor necrosis factor (ligand) superfamily member 9 (Tnfsf9), tumor necrosis factor receptor superfamily member 13c (Tnfrsf13c), and S100 calcium binding protein A8 (S100a8) [3436].

Table 4. List of significant differentially expressed genes for FG vs OH.

Heatmap display shows global gene expression differences in lung tissue exposed to the overhaul environment

The 3,852 significantly expressed genes in the group exposed to the overhaul environment vs fireground (Table 3) were visualized in a heatmap to see the expression patterns across all three groups. Two main heatmap patterns were apparent based on differences in exposure (Fig 2). Pattern 1 shows marked similarities in the C and FG groups when compared to the OH group for the genes located in the purple and black bars (n genes = 1,122 and 1,017, respectively). Pattern 2 shows marked similarities in the C and OH groups compared to the FG group for genes located in the yellow and green bars (n genes = 839 and 874, respectively). Finally, KEGG (Kyoto Encyclopedia of Genes and Genomes) analysis (Fig 3) of the black, purple, and both (black and purple combined) gene clusters from Fig 2 shows 22 significantly over-represented cellular pathways [37]. The black cluster contained 68% (15/22) of the over-represented pathways while the purple cluster contained 23% (5/22) with 9% (2/22) overlapping between both clusters.

Fig 2. Heatmap showing changes in global gene expression.

The overall expression patterns across all three treatments groups were visualized for the 3,852 genes with OH vs. FG FDR p-value < 0.05 using a heatmap. Each row represents one gene and each column is one individual mouse, grouped by treatment. The color scale represents standard deviations from the mean expression level across all samples with greater expression represented in red and lesser expression by blue relative to the mean.

Fig 3. Significant over-represented KEGG pathways amongst overhaul treated samples.

Over-representation testing completed based on purple and black gene set clusters, as well as both clusters combined, identified from heatmap (Fig 2) using Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways. Twenty-two pathways had more significant genes than expected by chance in at least one of the three comparisons (raw p-value < 0.005). The color of the box represents the –log10(p-value) to give more significant values darker color while the actual p-values are printed inside each box; the grey box indicates no genes in the black cluster mapped to that pathway.


Fire suppression is associated with high rates of duty-related sudden cardiac death [38,39]. In addition, firefighters are at increased risk for developing lung disease [40,41]and cancer [2,3,42]. While the etiologies of lung disease and cancer are thought to be linked to toxicant exposure during fire suppression and overhaul activities [16,43], mechanistically little is known about why firefighters show these increased incidences or what aspects of firefighting exacerbates disease risks. Using a mouse model of exposure sans airway protection, the impact of environmental exposure during overhaul on lung gene expression was assessed to better define pathways that are potentially critical to firefighter-related chronic illnesses. Our major finding is that working in an overhaul environment without breathing protection is associated with changes in transcripts with links to respiratory diseases including asthma, COPD and cancer. Importantly, these changes occurred in the absence of obvious increases in the poisonous gases HCN and H2S.

As expected, mice absent from the fireground (C group) or on the fireground but well distanced from the overhaul activities (FG group) showed greater similarity in gene expression than mice exposed to the overhaul environment (OH group). Interestingly, the C vs. FG group comparison showed more gene expression dissimilarity than anticipated (Fig 2). While transportation stress in mice is a well-described phenomenon [44,45], the magnitude of this effect from a gene transcription perspective is not currently known, but appears to need further study. In addition to transportation stress, the FG group was also exposed to fire apparatus lights, fireground sounds, and, potentially, light smoke. All or anyone of these could be a potential confound.

Gene listed in Table 4 are associated with immune and inflammatory pathways differentially expressed in the OH vs FG group. Interestingly, 50% of these genes were downregulated. An expected upregulation of pro-inflammatory genes, [46] downstream of NF-κB, was not observed, which was unexpected. In fact, the two principal cytokine mediators of innate immunity [47], namely IL-1βnot shown) and TNF were down-regulated. In contrast, a small group of cancer-associated genes were up-regulated during overhaul including: Calpain 11 (Capn11), RAR-Related Orphan Receptor Gamma (Rorc), and Deoxyribonuclease II Beta (Dnase2b) [29,48,49]. This pattern of gene expression accounts for the overrepresentation of pathways linking immune dysregulation to cancer (Fig 3), and suggests that working in the overhaul environment without airway protection, even when visibly “clear”, poses a danger to lung health. These findings may also add insight into the increased incidence of respiratory diseases and cancer that is reported in the fire service [50,51].

Other than CO, gases were measured at levels below the recommended limits for an 8–10 hr occupational exposure even when peak values were factored. CO, however, exceeded the NIOSH REL TWA (10 hour), and CO peak levels were above OSHA PEL TWA (8 hour). Since CO never approached the NIOSH ceiling value of 200 ppm and peak values were well below the IDLH of 1200 ppm [52,53], many fire services would clear firefighters for SCBA face piece removal in the overhaul conditions experienced by mice in this study. Additionally, for fire departments without quantitative requirements, the qualitative observation that the area was visibly clear of smoke would allow for unmasking.

Given that CO levels did not appear to correlate with the differential gene expression observed, several other well-described fireground contaminants could be responsible for the results including benzene and polycyclic aromatic hydrocarbons [54]. Unfortunately, portable monitoring for said toxicants was not available to this study. In sum, changes in lung gene expression appear relatively substantial in the unprotected mouse during overhaul. Further investigation is warranted to better understand how activation of potentially deleterious pathways in the mouse lung translate to pulmonary diseases in individuals exposed to the fireground.


We thank Dr. Alvaro Hernandez and his team at the Carver Biotechnology Center, High-Throughput Sequencing and Genotyping Unit for assistance in preparing libraries with RNA templates, fragment analysis and processing RNA for high-throughput sequencing.


The findings and conclusions in this paper are those of the author(s) and do not necessarily represent the official position of the National Institute for Occupational Safety and Health, Centers for Disease Control and Prevention.


  1. 1. Boorady LM, Barker J, Lee Y-A, Lin S-H, Cho E, Ashdown SP. Exploration of Firefighter Bunker Gear Part 2: Assessing the Needs of the Female Firefighter [Internet].
  2. 2. Daniels RD, Kubale TL, Yiin JH, Dahm MM, Hales TR, Baris D, et al. Mortality and cancer incidence in a pooled cohort of US firefighters from San Francisco, Chicago and Philadelphia (1950–2009). Occup Environ Med. 2014;71: 388–397. pmid:24142974
  3. 3. LeMasters GK, Genaidy AM, Succop P, Deddens J, Sobeih T, Barriera-Viruet H, et al. Cancer Risk Among Firefighters: A Review and Meta-analysis of 32 Studies. J Occup Environ Med. 2006;48: 1189–1202. pmid:17099456
  4. 4. Haynes HJG. Fire Loss in the United States During 2016 [Internet]. 2017. www.nfpa.research.
  5. 5. Keir JLA, Akhtar US, Matschke DMJ, Kirkham TL, Chan HM, Ayotte P, et al. Elevated Exposures to Polycyclic Aromatic Hydrocarbons and Other Organic Mutagens in Ottawa Firefighters Participating in Emergency, On-Shift Fire Suppression. Environ Sci Technol. 2017;51: 12745–12755. pmid:29043785
  6. 6. Wingfors H, Nyholm JR, Magnusson R, Wijkmark CH. Impact of Fire Suit Ensembles on Firefighter PAH Exposures as Assessed by Skin Deposition and Urinary Biomarkers. Ann Work Expo Heal. Oxford University Press; 2018;62: 221–231. pmid:29236997
  7. 7. Fent KW, Eisenberg J, Snawder J, Sammons D, Pleil JD, Stiegel MA, et al. Systemic Exposure to PAHs and Benzene in Firefighters Suppressing Controlled Structure Fires. Ann Occup Hyg. 2014;58: 830–45. pmid:24906357
  8. 8. Satran D, Henry CR, Adkinson C, Nicholson CI, Bracha Y, Henry TD. Cardiovascular Manifestations of Moderate to Severe Carbon Monoxide Poisoning. J Am Coll Cardiol. 2005;45: 1513–1516. pmid:15862427
  9. 9. Purser DA, Grimshaw P, Berrill KR. Intoxication by cyanide in fires: a study in monkeys using polyacrylonitrile. Arch Environ Health. 1984;39: 394–400. pmid:6098227
  10. 10. IARC Working Group on the Evaluation of Carcinogenic Risks to Humans. Meeting (2007 : Lyon F, International Agency for Research on Cancer. Painting, firefighting, and shiftwork. International Agency for Research on Cancer; 2010.
  11. 11. Burgess JL, Nanson CJ, Bolstad-Johnson DM, Gerkin R, Hysong TA, Lantz RC, et al. Adverse respiratory effects following overhaul in firefighters. J Occup Environ Med. 2001;43: 467–73. Available: pmid:11382182
  12. 12. Fent KW, Alexander B, Roberts J, Robertson S, Toennis C, Sammons D, et al. Contamination of firefighter personal protective equipment and skin and the effectiveness of decontamination procedures. J Occup Environ Hyg. 2017;14: 801–814. pmid:28636458
  13. 13. Laitinen J, Mäkelä M, Mikkola J, Huttu I. Fire fighting trainers’ exposure to carcinogenic agents in smoke diving simulators. Toxicol Lett. 2010;192: 61–65. pmid:19576276
  14. 14. Raimundo AM, Figueiredo AR. Personal protective clothing and safety of firefighters near a high intensity fire front. Fire Saf J. Elsevier; 2009;44: 514–521.
  15. 15. Baxter CS, Hoffman JD, Knipp MJ, Reponen T, Haynes EN. Exposure of Firefighters to Particulates and Polycyclic Aromatic Hydrocarbons. J Occup Environ Hyg. 2014;11: D85–D91. pmid:24512044
  16. 16. Bolstad-Johnson DM, Burgess JL, Crutchfield CD, Storment S, Gerkin R, Wilson JR. Characterization of firefighter exposures during fire overhaul. AIHAJ. 61: 636–41. Available: pmid:11071414
  17. 17. JANKOVIC J, JONES W, BURKHART J, NOONAN G. ENVIRONMENTAL STUDY OF FIREFIGHTERS. Ann Occup Hyg. Oxford University Press; 1991;35: 581–602.
  18. 18. Austin CC, Dussault G, Ecobichon DJ. Municipal firefighter exposure groups, time spent at fires and use of self-contained-breathing-apparatus. Am J Ind Med. Wiley-Blackwell; 2001;40: 683–692. pmid:11757045
  19. 19. Ringnér M. What is principal component analysis? Nat Biotechnol. Nature Publishing Group; 2008;26: 303–304. pmid:18327243
  20. 20. Leek JT, Storey JD. A general framework for multiple testing dependence. Proc Natl Acad Sci U S A. National Academy of Sciences; 2008;105: 18718–23. pmid:19033188
  21. 21. Leek JT, Storey JD. Capturing Heterogeneity in Gene Expression Studies by Surrogate Variable Analysis. PLoS Genet. Public Library of Science; 2007;3: e161. pmid:17907809
  22. 22. Leek JT, Johnson WE, Parker HS, Jaffe AE, Storey JD. The sva package for removing batch effects and other unwanted variation in high-throughput experiments. Bioinformatics. 2012;28: 882–883. pmid:22257669
  23. 23. Smyth GK, Hall Institute E. Linear Models and Empirical Bayes Methods for Assessing Differential Expression in Microarray Experiments. Stat Appl Genet Mol Biol. 2004;3. pmid:16646809
  24. 24. Ritchie ME, Phipson B, Wu D, Hu Y, Law CW, Shi W, et al. limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res. Oxford University Press; 2015;43: e47–e47. pmid:25605792
  25. 25. Law CW, Chen Y, Shi W, Smyth GK. voom: precision weights unlock linear model analysis tools for RNA-seq read counts. Genome Biol. BioMed Central; 2014;15: R29. pmid:24485249
  26. 26. Benjamini Y, Hochberg Y. Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing [Internet]. Journal of the Royal Statistical Society. Series B (Methodological). WileyRoyal Statistical Society; 1995. pp. 289–300.
  27. 27. Huber W, Carey VJ, Gentleman R, Anders S, Carlson M, Carvalho BS, et al. Orchestrating high-throughput genomic analysis with Bioconductor. Nat Methods. 2015;12: 115–121. pmid:25633503
  28. 28. Falcon S, Gentleman R. Using GOstats to test gene lists for GO term association. Bioinformatics. 2007;23: 257–258. pmid:17098774
  29. 29. Huang Y, Wang KKW. The calpain family and human disease [Internet]. TRENDS in Molecular Medicine. 2001. Available: http://tmm.trends.com1471
  30. 30. Iacovelli S, Hug E, Bennardo S, Duehren-von Minden M, Gobessi S, Rinaldi A, et al. Two types of BCR interactions are positively selected during leukemia development in the E -TCL1 transgenic mouse model of CLL. Blood. 2015;125: 1578–1588. pmid:25564405
  31. 31. Takahara T, Matsuo K, Seto M, Nakamura S, Tsuzuki S. Synergistic activity of Card11 mutant and Bcl6 in the development of diffuse large B-cell lymphoma in a mouse model. Cancer Sci. Wiley-Blackwell; 2016;107: 1572–1580. pmid:27560392
  32. 32. Sage PT, Ron-Harel N, Juneja VR, Sen DR, Maleri S, Sungnak W, et al. Suppression by TFR cells leads to durable and selective inhibition of B cell effector function. Nat Immunol. Nature Publishing Group; 2016;17: 1436–1446. pmid:27695002
  33. 33. Janz S, Jones GM, Müller JR, Potter M. Genomic Instability in B-Cells and Diversity of Recombinations That Activate c-myc. Springer, Berlin, Heidelberg; 1995. pp. 373–380.
  34. 34. Shao Z, Schwarz H. CD137 ligand, a member of the tumor necrosis factor family, regulates immune responses via reverse signal transduction. J Leukoc Biol. Wiley-Blackwell; 2011;89: 21–29. pmid:20643812
  35. 35. Losi CG, Silini A, Fiorini C, Soresina A, Meini A, Ferrari S, et al. Mutational Analysis of Human BAFF Receptor TNFRSF13C (BAFF-R) in Patients with Common Variable Immunodeficiency. J Clin Immunol. 2005;25: 496–502. pmid:16160919
  36. 36. Su Y, Xu F, Yu J, Yue D, Ren X, Wang C. Up-regulation of the expression of S100A8 and S100A9 in lung adenocarcinoma and its correlation with inflammation and other clinical features. Chin Med J (Engl). 2010;123: 2215–20. Available:
  37. 37. Jin L, Zuo X-Y, Su W-Y, Zhao X-L, Yuan M-Q, Han L-Z, et al. Pathway-based Analysis Tools for Complex Diseases: A Review. Genomics Proteomics Bioinformatics. 2014;12: 210–220. pmid:25462153
  38. 38. Soteriades ES, Smith DL, Tsismenakis AJ, Baur DM, Kales SN. Cardiovascular Disease in US Firefighters. Cardiol Rev. 2011;19: 202–215. pmid:21646874
  39. 39. Kales SN, Soteriades ES, Christophi CA, Christiani DC. Emergency Duties and Deaths from Heart Disease among Firefighters in the United States. N Engl J Med. 2007;356: 1207–1215. pmid:17377158
  40. 40. Sheppard D, Distefano S, Morse L, Becker C. Acute effects of routine firefighting on lung function. Am J Ind Med. Wiley-Blackwell; 1986;9: 333–340.
  41. 41. Scannell CH, Balmes JR. Pulmonary effects of firefighting. Occup Med. 1995;10: 789–801. Available: pmid:8903749
  42. 42. Sama SR, Martin TR, Davis LK, Kriebel D. Cancer incidence among massachusetts firefighters, 1982–1986. Am J Ind Med. Wiley-Blackwell; 1990;18: 47–54.
  43. 43. Fritschi L, Glass DC. Firefighters and cancer: Where are we and where to now? Occup Environ Med. BMJ Publishing Group; 2014;71: 525–526. pmid:24996680
  44. 44. Towers AE, York JM, Baynard T, Gainey SJ, Freund GG. Mouse Testing Methods in Psychoneuroimmunology 2.0: Measuring Behavioral Responses. Methods in molecular biology (Clifton, NJ). 2018. pp. 221–258. pmid:29705851
  45. 45. Olfe J, Domanska G, Schuett C, Kiank C. Different stress-related phenotypes of BALB/c mice from in-house or vendor: alterations of the sympathetic and HPA axis responsiveness. BMC Physiol. BioMed Central; 2010;10: 2. pmid:20214799
  46. 46. Walters K-A, Olsufka R, Kuestner RE, Cho JH, Li H, Zornetzer GA, et al. Francisella tularensis subsp. tularensis Induces a Unique Pulmonary Inflammatory Response: Role of Bacterial Gene Expression in Temporal Regulation of Host Defense Responses. Rottenberg ME, editor. PLoS One. Public Library of Science; 2013;8: e62412. pmid:23690939
  47. 47. Dantzer R, O’Connor JC, Freund GG, Johnson RW, Kelley KW. From inflammation to sickness and depression: When the immune system subjugates the brain. Nat Rev Neurosci. 2008;9. pmid:18073775
  48. 48. Tosolini M, Kirilovsky A, Mlecnik B, Fredriksen T, Mauger S, Bindea G, et al. Clinical impact of different classes of infiltrating T cytotoxic and helper cells (Th1, th2, treg, th17) in patients with colorectal cancer. Cancer Res. American Association for Cancer Research; 2011;71: 1263–71. pmid:21303976
  49. 49. Ribeiro FR, Paulo P, Costa VL, Barros-Silva JD, Ramalho-Carvalho J, Jerónimo C, et al. Cysteine-Rich Secretory Protein-3 (CRISP3) Is Strongly Up-Regulated in Prostate Carcinomas with the TMPRSS2-ERG Fusion Gene. Batra SK, editor. PLoS One. Public Library of Science; 2011;6: e22317. pmid:21814574
  50. 50. Guidotti TL. Mortality of urban firefighters in alberta, 1927–1987. Am J Ind Med. Wiley-Blackwell; 1993;23: 921–940.
  51. 51. Ribeiro M, de Paula Santos U, Bussacos MA, Terra-Filho M. Prevalence and risk of asthma symptoms among firefighters in São Paulo, Brazil: A population-based study. Am J Ind Med. Wiley-Blackwell; 2009;52: 261–269. pmid:19117017
  52. 52. Criteria for a recommended standard… occupational exposure to carbon monoxide. [Internet]. 1972 Jan.
  53. 53. Clayton GD, Clayton FE, Allan RE (Ralph E., Patty FA (Frank A. Patty’s industrial hygiene and toxicology. Wiley; 1991.
  54. 54. Fent KW, Evans DE, Booher D, Pleil JD, Stiegel MA, Horn GP, et al. Volatile Organic Compounds Off-gassing from Firefighters’ Personal Protective Equipment Ensembles after Use. J Occup Environ Hyg. 2015;12: 404–414. pmid:25751596