Measures of LPS-induced inflammation were broadly consistent with our previous study, where subjects inhaled LPS but did not undergo active/sham leukapheresis , and therefore it is unlikely that cells present in BAL fluid of the subjects involved in this study were altered or activated, or that measures of inflammation were significantly changed by sham leukapheresis. Sham leukapheresis has been shown to have no effect on monocyte chemotaxis or migration .
Our original classification of inducible and resident PMLC subpopulations was based upon the variable expression of the monocyte markers CD14 and CD16, and the clear shift in the ratio of these cells in LPS-mediated experimental acute lung inflammation. We previously postulated that iPMLC may be sequestered from the circulation alongside neutrophils during self-limiting lung inflammation and, in contrast, that rPMLC were endogenous to the lung and contributed to macrophage homeostasis i.e. by undergoing proliferation and macrophage differentiation . Our observation in the current study that rPMLC have a significantly increased expression of the mature macrophage markers CD206 (mannose receptor), CD71 (transferrin receptor) and 25 F9 (mature macrophage marker) and a significantly increased expression of the proliferation antigen Ki67, compared to iPMLC, is broadly supportive of our previous hypothesis, although this cannot be proven definitively, given the limitations and the nature of the current study. Conversely, iPMLC display a less mature, monocyte-like phenotype and have a significantly lower proliferative capacity compared to rPMLC. Therefore, it seems unlikely that the primary function of iPMLC is to contribute to regenerative demand and cellular turnover, and our data suggest that a pro-inflammatory function is more likely attributable to these cells, at least at the time point studied. This is supported by our cytokine data.
There are a number of limitations to this study that merit discussion. These principally stem from two problems. Firstly, flow sorting of cells from complex inflammatory populations in the human lung inevitably yields a relatively small and finite number of cells for study. Secondly, repeat bronchoscopy on the same day raises considerable ethical issues, and bronchoalveolar lavage itself leads to an inflammatory response in the lung after the procedure – these considerations inevitably meant that we could only study one time point in the process, and no inferences can be made regarding the time course and dynamics of any of the processes studied.
As such, due to the low numbers of rPMLC present in BAL fluid in this model it was not possible to isolate these cells by FACS in order to determine their basal cytokine production or response to steroid treatment, although this would be of interest for future studies. Due to the technical complexities associated with using FACS to isolate highly pure cell subpopulations from BAL fluid for in vitro growth and treatment with dexamethasone (i.e. samples of less than 95% purity were discarded), it was only possible to grow sorted cell populations from 8 of out 14 subjects. Notably, we were still able to demonstrate a significant response by AM to dexamethasone treatment, whereas iPMLC did not show any response to dexamethasone. The phagocytic capacity of individual PMLC subpopulations is of interest, although it was not possible to determine this due to limitations in the assay conditions. It is important to highlight that, despite significant differences in cell surface antigen expression by iPMLC and rPMLC, a degree of heterogeneity exists within both PMLC subpopulations, and it is plausible that functional crossover may occur. We have previously reported differential cell counts from cytocentrifuge preparations of BAL fluid from subjects who have inhaled LPS and then undergone active leukapheresis compared to those receiving sham leukapheresis , which was in accordance with the secondary endpoint of our RCT of pulmonary neutrophilia in BAL fluid. A discrepancy exists in the reported cellular composition of BAL fluid when comparing data from cytocentrifuge preparations compared to lineage analysis using polychromatic flow cytometry reported herein. However, we have shown that PMLC subpopulations are similar to lymphocytes and neutrophils in size, and it may be difficult to consistently distinguish PMLC from other lineages using cytocentrifuge preparations alone. Our data show that PMLC represent a significant proportion of the cells present in BAL fluid following LPS inhalation, similar in number to neutrophils at 9 hours. This observation was unexpected and implies that PMLC may play a significant role in the acute inflammatory response. It also implies that future studies of the cell content of BAL fluid should consider using flow cytometry to quantify the cellular composition of BAL fluid, taking into account PMLC subpopulations, according to the protocols described herein.
Experimental inhalation of LPS models some of the early features of acute lung injury (ALI) such as neutrophilia in peripheral blood and in the alveolar space, and protein leak across the alveolar-capillary membrane. The use of corticosteroids to treat ALI, and the optimal timing of administration, remains extremely controversial [8–11], and we felt it would be relevant to assess corticosteroid-responsiveness in distinct cell populations isolated from BAL fluid in volunteers who had inhaled LPS. iPMLC produced comparable concentrations of the inflammatory cytokines IL-6, IL-8 and TNFα compared to AMs, yet displayed a significantly reduced responsiveness to dexamethasone. Our data suggest that corticosteroids may potentially down-regulate secretion of inflammatory cytokines from AMs, but are unlikely to influence cytokine release from iPMLC. However, as discussed previously, the time point selected for study could have influenced these results. The BAL cells were collected at 9 hours after inhalation of LPS and the addition of dexamethasone was thus commenced at a relatively late stage; it remains possible that differences in dexamethasone-induced suppression of cytokine release observed between AM and iPMLC reflects a difference in the timing of peak cytokine release by both cell types. Corticosteroids inhibit apoptosis of neutrophils , the predominant alveolar cell in ALI, and this has been proposed as a potential reason for lack of therapeutic efficacy in this setting. In the current study we did not have sufficient cells to determine whether corticosteroids directly affect apoptosis of AMs and iPMLCs, and we recognise that any such effect could influence our findings.
In summary, our data show that after LPS challenge iPMLC constitute a significant proportion of the cellular component of BAL fluid. iPMLC have a low capacity for proliferation or phagocytosis and are capable of production of similar, albeit low, levels of pro-inflammatory cytokines compared to AM. Interestingly, unlike AM, iPMLC appear to be unaffected by corticosteroid treatment in vitro, at least in our experimental system. Further data are clearly required to understand the physiological or pathological relevance of these cells, and to determine whether their function and phenotype change as the inflammatory process proceeds. Nevertheless, we believe our findings represent a useful first step, suggesting the potential for iPMLC to play a role in the early regulation of inflammatory responses in the human lung.