Background
Bacillus anthracis is a gram-positive, aerobic, spore-forming, rod-shaped bacterium which causes a virulent disease, anthrax. The three primary forms of the disease are due to three different mechanisms of exposure: ingestion (gastrointestinal), contact (cutaneous) or inhalation (inhalational) [
1]. Inhalational anthrax is the most life-threatening form of the disease [
2,
3], and was the type seen during the recent bioterrorism attacks [
4].
Inhalational anthrax is characterized by a rather unique finding in that the inhaled spores do not vegetate and cause disease at the site of entry [
5,
6]. Instead, spores are rapidly and efficiently phagocytosed by alveolar macrophages and dendritic cells, and carried through lung tissue to the regional lymph nodes [
5,
7‐
9]. It is only after the pathogen escapes the lung that dissemination occurs, following transit to the thoracic lymph nodes. There is a significant delay, as long as 43 days, between exposure and clinical disease, implying that there is temporary containment of the pathogen, likely by the innate immune system [
10].
Alveolar macrophages play a central role in the innate immune system and are the first line of defense against inhaled pathogens. They are the most prominent resident cells that not only engulf and kill infectious agents, but also produce numerous modulators of the inflammatory response to recruit and activate additional cells of the immune system. Alveolar macrophages also provide a link to the adaptive immune system since they function as antigen presenting cells. Previous studies of the transcriptional response of the murine macrophage-like RAW 264.7 cell line to
B. anthracis (Sterne) spore infection provided initial insight into macrophage responses [
11] and
B. anthracis adaptation to the host microenvironment [
12], but correlation of these results with human alveolar macrophage responses has not been performed.
Our earlier studies examined the interaction of
B. anthracis spores with human alveolar macrophages (HAM) obtained by bronchoalveolar lavage (BAL). We studied the initial events after exposure to spores beginning with the rapid internalization of spores by the macrophages. Spore exposure rapidly activated the mitogen-activated protein kinase (MAPK) signaling pathways ERK, JNK, and p38. This was followed by transcriptional activation of cytokine and primarily monocyte chemokine genes and the data was confirmed at the level of translation [
13].
In the current study, we infected the HAM obtained by BAL with
B. anthracis (Sterne) spores and performed Affymetrix Human Genome U133 Plus 2.0 Array (Santa Clara, CA) to provide a comprehensive view of the innate immune response of the alveolar macrophage to the pathogen. In comparing the expression pattern of spore-exposed with mock-infected cells, our analysis identified TNF-α and NF-κB as key components of the innate immune response to
B. anthracis. Many (48) of the genes affected by spores shared c-Rel transcription regulatory element, a member of NF-κB family of transcription factors. Our findings are consistent with our previous, more limited evaluation of the immune response by HAM to spores using ribonuclease protection assay (RPA) and ELISA [
13]. In addition, our results show the actions of a significant number of poorly annotated genes, indicating that much of the response occurs due to currently unknown mechanisms.
This study is the first detailed microarray analysis to describe the HAM response to B. anthracis spores. It provides a tool for investigators to use for diagnostic and therapeutic purposes and points toward a number of unknown processes as being important in the macrophage response.
Discussion
The alveolar macrophage represents a major defensive mechanism against infection of the lung. Although some details are clear, much is unknown about how these cells respond to pathogens such as B. anthracis. In this study a systematic investigation of gene expression of HAM infected by B. anthracis spores was undertaken to map out the full response and to identify the full range of genes and pathways involved. The majority of genes that were differentially expressed in response to spore infection were upregulated. Among the upregulated genes, we identified chemokine ligand, apoptosis, and interestingly, keratin filament genes. Central hubs regulating those upregulated genes were TNF-α and NF-κB and their ligands/receptors. Other well known players were IL-1α, IL-18 and others. Many (48) of the spore-induced genes shared c-Rel TREs. C-Rel is a member of the NF-κB family of transcription factors. Other TREs shared in common among spore induced genes were c-Myb, CP2, Barbie Box, E2F and CRE-BP1. In contrast, genes downregulated in cells infected by B. anthracis spores did not form well defined ontological groups and networks, neither did they share common transcription regulatory elements, possibly reflecting that down-regulation occurs by different mechanisms than at the transcriptional level.
The method of microarray data analysis utilized here used not only statistical comparisons but also considered biological properties of the data [
26] in order to find unbiased results covering the entire genome. By filtering the results using stringent criteria that emphasizes a large change in expression, the specificity of analysis was increased, which identified the most robust "beacons" driving responses to the pathogen. The results provide us with information about the major processes affected. Genes inferred as being present were then checked against the whole dataset to determine that they were expressed. They were removed if expression was not seen in the microarray. This approach removes much of the noise from the system and provides a highly reliable means to identify processes and gene networks. However, as discussed above, all such analyses are limited because the real new knowledge lies in the unannotated or poorly annotated genes, and the pathway and ontologic analyses only confirm that mechanisms studied previously are active in the system under investigation. Some new knowledge also results from showing how these well annotated processes may fit together.
Previous results of gene expression profiling of macrophage responses to
B. anthracis spores [
11,
12] have been performed using the murine RAW 264.7 macrophage-like cell line. Although it is difficult to directly correlate those results to primary HAM, several important similarities were observed. In both cases there was induction of genes relating to the immune response, apoptosis, and cytoskeleton organization and biogenesis. Specifically TNFα, IL-1α, colony stimulating factor, IFNγ, and NF-κB were induced in the mouse model [
11] and in the current study. On the other hand, other cytokines, for example RANTES (CCL5), is induced by spores only in HAM, and not in RAW 264.7 cells. In these cases, considering that HAM are freshly isolated from normal lung, results with these cells are more likely to reflect those occur during natural infection, than those found with RAW 264.7 cells. Comparison of the current results from microarray analysis with our previously published findings. [
13,
27] and with additional experimental assays (Table
5) showed high correspondence between gene expression and protein levels. Our previous publication [
13] demonstrated
B. anthracis spore-induced activation of the MAPK signaling pathways, and induction of several cytokines and chemokines. Consistent with that work, the current study demonstrated that the p38 MAPK signaling pathway was overrepresented by genes upregulated in spore infected cells, along with IL-6 and IL-10 associated signaling pathways. Previous results analyzing HAM infected with
B. anthracis spores for 6 hours at MOI = 1 further confirm the current findings for TNF-α and IL-1β. TNF-α mRNA, induced 11 fold in the current study, was induced 20 fold as determined by RPA. TNF-α protein was induced 74 fold as measured by ELISA. IL-1β mRNA, induced 2.6 fold in the current study, was induced 14 fold in our previous study as measured by RPA. IL-β protein was upregulated by 13-fold as previously determined by ELISA.
IL-6, IL-10, GM-CSF and IFN-γ were also induced 4- to 43-fold in the previous study as measured by the RPA and/or ELISA, and this is consistent with the results for RNA induction as determined in the current study by microarray. The current microarray analysis also identified several additional 2- to 5-fold differentially expressed interleukins and their receptors/binding proteins (IL-1F6, IL-1F7, IL-1F9, IL-1RL2, IL-20, IL-23A, IL-24, IL-29, IL-2RA, IL-32, IL-6ST, IL-7R, IL-9R, IL-11RA, IL-12B, IL-12RB1, IL-15RA, IL-18BP, IL-18R1, IL-19). Two colony stimulating factors, CSF2 and CSF3, were induced 10 and 22 fold, respectively. However, this fold change should be interpreted with caution, due to the high variability of gene expression level for CSF2 and CAF3.
In this study, we performed a static comparison of gene expression in mock- and
B. anthracis infected HAMs to identify potential hallmarks of inhalational anthrax. However, our results are similar to dynamic changes that occurred in a time course of infection of murine alveolar macrophages by
Aspergillius fumigates [
28]. Nine genes (CCL3, CCL4, CXCL2, EGR1, ICAM1, IL1A, NFATC1, NFKBIZ, TNF) responded dynamically to
A. fumigates infection and these were also identified as key players in
B. anthracis induced response. Interestingly, TNF, IL1A, EGR1 and NFκB were the same central players in Ingenuity generated pathways, as in our case (Figure
2). Thus our findngs in the static condition used is simlar to that seen in time course responses of alveolar macrophages to other pathogens.
Inflammation and immune response genes are important in cellular defense responses. This is consistent with our findings, as TREM1 signaling was the most significant pathway identified and was represented by NF-κB, TNF-α and interleukin members. TREM1 belongs to the Immunoglobulin (Ig) family of cell surface recettors and is selectively expressed on blood neutrophils, monocytes and macrophages. It is known that TREM is mediated by a transmembrane adaptor molecule DNAX-activating protein 12 (DAP12), leading to proinflammatory immune responses. The natural ligand for TREM1 is however, unknown. TREM1 signaling is associated with Toll like receptor (TLR) signaling and with a second major class of PRR - the NACHT-LRR receptors (NLR), which recognize intracellular microorganisms. Thus TREM1 acts as an indispensable link connecting (and, possibly, enhancing) signals from both major pathways of pattern recognition- extracellular TLR receptors and the intracellular NLR proteins.
The findings presented here provide additional evidence of the involvement of NLR and TLR signaling in the response to
B. anthracis. Luminex assays show that IL-1β protein is released from spore exposed HAM. Others have shown that cooperation between MyD88-dependent (TLR) and MyD88-independent (NLR) signaling pathways was required for
B. anthracis spore mediated IL-lβ induction. Also, both TLR and NLR signaling pathways are important in IL-1β induction and subsequent processing by inflammasome formation [
29,
30]. This, together with the implication of TREM1 signaling in the HAM response to spores confirms the importance of TLR and NLR in this process.
TNF and NF-κB have long been implicated in inflammation and immune response under various conditions [
31‐
33]. In our system TNF and other members of TNF superfamily were strongly upregulated indicating their involvement in the response to
B. anthracis. Other genes in the list add to the picture of cellular defense and death triggered by spores. Several interleukins (IL-1α, IL-17RB, IL-18) and TNF have been reported to mediate cell apoptosis and death [
34‐
36], although whether this transcriptional upregulation results in damage to the alveolar macrophage is yet to be determined.
NF-κB is a regulator of TNF and interleukin signaling [
37,
38]. The transcriptional regulatory element for NF-κB is present in many of the genes overrepresented in cells infected by spores. This may be a key to initiation of the response to
B. anthracis and a possible target for enhancement of cellular defenses against this pathogen.
This conclusion is also based on our transcriptional regulatory element analysis by PAINT, which assesses genes of interest regardless of their functional classification and considers only TREs shared among them. In this analysis many of the genes upregulated on spore infected cells shared the c-Rel regulatory sequence, a TRE activated by NF-κB. The possibility that this could be a chance finding is low, not only because the stringency was set at p < 0.05, but also because the false discovery rate was set to <0.1. Thus our findings indicated that NF-κB is likely a central regulator in the responses by HAM to B. anthracis spores.
In silico computer analysis is based on existing information about gene functions and interactions and provides little information about genes that are differentially regulated but poorly annotated. These genes represent undiscovered mechanisms and new knowledge as to processes operating during infection. One possibility for the asymmetry in publications about gene function is that the most likely genes to be discovered are those functioning in all or most cells or tissues. These represent core functions such as apoptosis and cell division. Genes that function in a cell-specific manner are less likely to be discovered, particularly if they occur in cells or tissues that are not heavily investigated. Cell or tissue specificity is likely to be conferred in genes that may be expressed in only a very few cell types, which suggests they should have a higher probability of being poorly annotated. Supporting this is the finding that LRRC50 and three other poorly annotated genes were differentially expressed mainly in lung and head and neck cancers. Our findings thus add to the knowledge about the function of these genes. It is apparent, for example, that these four genes not only function in cancer of the respiratory tract, but also function in the lung response to
B. anthracis. These findings suggest that to understand the response of human alveolar macrophage to
B. anthracis will require investigating the functions of these poorly annotated genes (Additional File
1) and how they might be connected to the known pathways.
Another limitation imposed on the current analysis is that Ingenuity limits the number of genes in an individual network to 35. However, individual networks and pathways do not function in isolation. Rather, they are part of a larger system of interacting pathways and genes that make up the response to the pathogen [
39]. We observed this in the networks presented in Figure
2, where members of NF-κB and TNF-α interaction appeared in different networks. Therefore, such networks should not be treated as individual entities but rather as a complex network with key genes occupying the "hub" positions.
In summary, the alveolar macrophage shows a complex response to infection by B. anthracis spores. Based on the number of poorly annotated genes, only a portion of the response is known. That the expected responders such as NF-κB and TNF were seen gives credence to the importance of these poorly understood genes.
Authors' contributions
JPM and KC designed the study and analyzed the data. WW and MD drafted the manuscript. JLB, REH and KMC collected the data and participated to their interpretation. MD performed the statistical analysis. MD, WW and JPM participated to the redaction of the manuscript. All authors read and approved the final manuscript.