Isolation procedure
Human IM are less accessible than AM, which is why IM have in the past mostly been characterized using animal models [
4,
24]. Our approach for MΦ isolation from human lung interstitial was based on a previously described method for isolation of epithelial cells [
13] and allows parallel isolation of AM, IM and epithelial cells. The digestion procedure that we used slightly differed from those previously described for isolation of human IM [
6,
7].
We are aware that donor specifics such as medication or smoking behaviour might alter MΦ functions. In fact, smoking has been shown previously to increase basal levels of TNF-α, IL1 or IL8 in human AM [
25] and to cause changes in morphology and surface molecule expression in rat AM [
4]. Nevertheless, our results reveal many similarities of AM from lung tissue compared to AM from bronchoalveolar lavage described in the literature, as detailed below.
CD14 and HLA-DR expression
Studies using primary rat AM and IM suggest that AM and IM do not differ in CD14 expression [
4]. We were able to show that CD14 is marginally expressed in human AM as well as IM. Low CD14 expression was reported previously for human AM obtained from bronchoalveolar lavage [
28,
29]. CD14 expression by human IM is not described in the literature, but our results resemble findings reported for other tissue macrophage populations [
28].
Several studies using rat, mouse as well as human MΦ reported a higher MHC-II expression in IM compared to AM [
4,
5,
7,
30]. Our own results resemble those findings, indicating that IM are more involved in acquired immune response than AM.
Toll-like receptor expression
In the present study, expression of TLR1-10 mRNA levels were examined in both AM and IM for the first time. TLR1-10 mRNA expression by AM obtained from bronchoalveolar lavage was previously described by Maris et al. [
35]. According to this study, TLR1, TLR4, TLR7, TLR8 are strongly and TLR2, TLR6 weakly expressed, whereas TLR3, TLR5, TLR9 and TLR10 were not detectable. Our results for TLR1, TLR4, TLR6, TLR8 and TLR10 expression by AM resemble those observations. In contrast to Maris et al., we observed a strong expression of TLR2 mRNA which is in line with results by Suzuki et al. [
36] as well as ourselves [
16]. Contrary to Maris et al. we were able to detect TLR3, TLR5 and TLR9 mRNA, suggesting a higher sensitivity of our assay. In fact, our real time PCR analysis is linear over 8 orders of magnitude, down to a concentration of 10
-6 attomole/μl.
It was long believed that both human monocytes as well as human macrophages do not express TLR9 [
37]. This view was supported by examinations like that by Miettinen et al. [
38], failing to detect TLR9 mRNA by Northern Blot in human macrophages. By now, the presence of TLR9 in MΦ was confirmed by Fenhalls et al. [
39], Juarez et al. [
40], and ourselves [
16] by immunhistochemichal detection, Western blot analysis, flow cytometry, real time RT-PCR, and evidence of TLR9 functionality.
No significant differences between AM and IM concerning the TLR mRNA expression profile were found. Still, we can not exclude that TLR protein expression or localization differ in the different macrophage populations, which might both cause a differential cell reaction upon ligand binding.
Cell reaction upon TLR activation
Previous studies using rat MΦ revealed that upon TLR4 stimulation AM express higher amounts of the proinflammatory cytokine TNF-α compared to IM [
4,
9]. Our data suggest that this is also true for human AM and IM. Moreover, we were able to detect significant differences between AM and IM concerning IL10 as well as IL6 expression.
Although IL10 is one of the most important anti-inflammatory mediators in human immune response [
41,
42], its expression by human AM and IM has not been investigated before. AM display low basal IL10 levels, which might allow efficient defence against inhaled particles and pathogens, whereas the fast induction of IL10 upon LPS treatment suggests an autoregulatory mechanism. As for IM, basal IL10 mRNA and protein levels were found to be significantly higher than in AM and to increase after LPS treatment. A recently published study comparing murine AM and IM showed that IL10 levels were markedly higher in IM [
5]. Our data demonstrate that this is also true for human IM. In the animal model, IM were shown to inhibit lung DC maturation and migration in an IL10-dependent manner, thereby preventing Th2 sensitization to harmless inhaled antigens [
5]. Our findings suggest that this might also be exhibited by human IM, indicating that IM play a crucial role in immune homeostasis.
IL6 has proinflammatory as well as anti-inflammatory properties. Studies using knockout mice demonstrated that in innate immunity IL6 acts predominantly as an antiinflammatory cytokine, mainly by attenuating the synthesis of proinflammatory cytokines [
43,
44]. Moreover, IL6 is involved in the specific immune response by upregulating B-cell differentiation, T-cell proliferation, and antibody secretion [
44]. The high constitutive expression of IL6 that we found in IM both on mRNA and protein level indicates that IM display a pronounced immunoregulatory capacity and suggests that they are more involved in specific immune responses.
IL1ra is a major antiinflammatory cytokine that functions as a specific inhibitor of the two other functional members of the IL-1 family, IL-1a and IL-1β [
41,
45]. Our data demonstrate that IM secrete higher amounts of IL1ra when compared to AM, both at baseline and upon TLR4 activation. In contrast, LPS induced secretion of proinflammatory cytokines was low in IM when compared to AM (IL1β, IL12p40) or even completely absent (IL12p70 and IFNγ). These findings clearly underline the anti-inflammatory phenotype of IM previously described in the literature based on data obtained from murine or rat MΦ [
4,
5,
9,
24].
Despite the weak expression of TLR9 in AM and IM, cells reacted strongly upon stimulation with mycobacterial DNA. Methylation or digestion of mycobacterial DNA as well as chloroquine pretreatment lead to an abrogation of the macrophage response [
16], which indicates that gene expression upon treatment with isolated DNA is not due to contaminants in DNA preparations, but due to TLR9 activation. It has been shown previously for human
in vitro differentiated MΦ as well as a mouse MΦ cell line that TLR9 activation is higher upon treatment with bacterial DNA than after oligonucleotide treatment, which might be due to oligonucleotide structure [
16,
46,
47]. Moreover, it has been reported for human
in vitro differentiated macrophages as well as for AM from BAL that DNA from virulent strains has a lower potential to activate TLR9 in MΦ than DNA from attenuated strains [
16]. Several studies indicate that the virulent H37Rv strain is able to methylate cytosines whereas the H37Ra strain is not [
48,
49], which might explain why H37Rv DNA fails to activate TLR9.
We were able to show that IM are less responsive to bacterial DNA than AM concerning TNF-α induction, which resembles our findings for TLR4 activation. Similar to the results of the TLR4 activation experiments, IL6 as well as IL10 expression were much higher in IM compared to AM, which clearly underlines the immunoregulatory function of IM.
Moreover, IL10 was only induced in IM, but not in AM, upon mycobacterial DNA treatment. Absence of IL10 induction after TLR9 activation by mycobacterial DNA has been reported before [
40] and might be part of a mechanism of AM to overcome the immunosuppressive environment of the alveoli. Lung epithelial cells have been shown to constitutively express IL10, which is accompanied by an impaired responsiveness of AM towards IL10 [
50]. In the same study, it was also observed that activation of human AM through TLR2, TLR4 or TLR9 leads to inhibition of IL10 receptor function associated with a reduced ability to activate STAT3. As IL10 is known to induce its own transcription
via several positive feedback loops involving STAT3 [
51,
52], a low capacity of AM to activate STAT3 might contribute to the lack of IL10 induction in AM upon TLR9 activation as well as to the low basal levels of IL10 in comparison to IM.