Using a flow- and volume-controlled inhalation, we were able to improve the deposition of LPS in the lung and to elicit a pronounced and significant inflammatory response in healthy volunteers using a low dose of 20,000 E.U. (2 μg) LPS. When repeating the procedure after a 4 week washout period, the overall inflammatory response was shown to be reproducible; however, we did observe a small decline in the neutrophil/monocyte ratio after the second LPS challenge. Treatment with the PDE4-inhibitor Roflumilast for 5 days changed the expression of HLA-DR and CD86 on sputum macrophages, but did not result in a significant attenuation of the inflammatory cell influx. Our data suggests that the low-dose challenge required for the use of GMP-grade endotoxin is suitable for “proof-of-concept” studies of novel compounds targeting neutrophilic and monocytic airway inflammation.
Regulatory authorities increasingly control the origin and production of substances used for airway provocation. In Germany only GMP-grade LPS, such as CCRE produced by the NIH Clinical Center, is allowed to be used for these purposes. This material is also increasingly being used for studies in the USA [9, 12]. As this material is of limited availability, an improved deposition and a nebulizer with a small residual volume were essential for our study. Flow controlled inhalation of nebulized aerosols can greatly improve deposition . Although the AKITA® inhalation system is a commercially available device with integrated flow control, we could not use this system in this study, as the death volume of its jet nebulizer is too large. Therefore we used the Aeroneb solo nebulizer (Inspiration Medical), which creates the aerosol using a high-frequency vibrating membrane with 1000 precision-cut openings, working basically like a micro-pump system. Here, the residual volume of LPS was generally below 100 μL. This nebulizer was combined with a mass-flow control unit that limited the inhalation flow and applied air only during the end of each inspiration. The controlled inhalation most likely increased the lung deposition of LPS by avoiding the unwanted deposition in the mouth and pharynx.
Inhalations of low LPS doses were safe and well tolerated. Only a small decrease in lung function was detected, but the affected subjects did not report any symptoms. The extent of systemic effects was in the expected range, with an increase in body temperature of less than 1°C. The blood total leukocyte and neutrophil count increased, but this is known to occur even after exposure to a low dose of LPS, that does not elicit a detectable change in the composition of airway leukocytes .
The main focus was the analysis of induced sputum. Compared with a previous study that used the same dose of LPS , the neutrophilic response was more pronounced. The 6 h time point after LPS inhalation has been used frequently to assess the inflammatory effect; however, there are conflicting results with respect to the maximal effect. In a study by Doyen et al., the peak neutrophil cell count was detected 24 h after the challenge , which is in line with data from endobronchial LPS challenges. A more pronounced effect at 6 h was recently shown by Aul and coworkers , but this was in healthy smokers.
Using induced sputum to assess the inflammatory response to LPS, we assessed cytospin slides after the first LPS challenge. While the neutrophil influx was clearly detectable, we also saw increased numbers of smaller macrophages and monocytes. Therefore, we included flow cytometry into our analysis of sputum composition after the second LPS challenge and measured the proportion of monocytes using CD14 staining. Comparison of the flow cytometry data with the mean cytospot cell count of two independent observers showed a good correlation for macrophages and neutrophils. We also found a fairly good relationship between CD14-positive cells and the sum of monocytes and small macrophages, supporting our approach to count these cells together (Additional file 1: Figure S4). For a more detailed flow cytometric analysis of induced sputum please refer to Lay et al. .
Looking at the data derived from the cytospot analysis, we observed only a small increase in the proportions of monocytes and small macrophages after the first, but a significant increase after the second LPS challenge. This effect could partly be responsible for the lower neutrophil proportion detected after the repeated LPS challenge. However, changes in cell proportions are difficult to interpret. We would, therefore, recommend using the cumulative response to LPS consisting of neutrophils, monocytes, and small macrophages as an additional outcome in future LPS trials. The reproducibility of this cumulative response compared with baseline was also better than for sputum neutrophils alone.
The development of tolerance could also be a reason, why the response to the second LPS challenge was lower. This phenomenon is well known [24–26], however, it has not been seen this clearly in a LPS inhalation trial before. Loh et al. suggested that tolerance would be visible only at low doses of LPS , but there are no studies available looking at repeated low-dose LPS challenges. In a recently published paper by Aul et al. , a dose comparable to the one used by Loh et al. was inhaled and no tolerance was detected, But Aul and colleagues challenged healthy active smokers, who are constantly exposed to LPS from cigarette smoke at relevant levels , and might therefore show a more homogeneous response. While this would be in favour for including only active smokers into LPS challenge proof-of-concept studies, the level of acute smoking is generally not easy to control and could bias drug effects in numerous ways. Based on our results, it appears to be advisable to include a screening LPS challenge, when planning proof-of-concept studies with healthy subjects. This was not actually tested in our trial; however, we did not see a further decline in neutrophils in the third LPS challenge, therefore, a bias due to a tolerance effect appears to be limited to the second LPS challenge in healthy subjects. In addition, the reproducibility between the second LPS challenge (LPS 2) and the LPS challenge after treatment (LPS Tx) was clearly better than between LPS 1 and LPS 2.
Interestingly, we did not see an attenuated response in the second challenge with respect to IL-8 and MPO levels in the sputum supernatant, indicating that either the sputum supernatant analysis is less sensitive or that other mechanisms than simply chemo-attraction are involved in determining the cellular response to LPS.
Comparison of sputum of 11 subjects collected on average more than 3 months (median 111, minimum 56, maximum, 157 days) after the last LPS inhalation revealed a higher mean neutrophil count as compared to the baseline visit of this study (median between baseline sputum inductions: 203 days (minimum: 191, maximum: 245). Despite this, neutrophil percentages were highly correlated (r = 0.86). It could be speculated that the reason for the increase is season-related, as baseline sputum of this study was obtained during late summer 2011 (August/September) and the repeated measurements were taken in early and colder springtime 2012 (March/April).
In this study, we investigated the effect of a 5 day treatment with Roflumilast, a duration of treatment during which a steady-state level in serum can be achieved [28, 29]. In primates treated with a comparable dose (7 μg/kg body weight per day for 5 days), a small decline in BAL neutrophil numbers and percentage was observed . In our study, the effect of Roflumilast treatment on sputum neutrophil percentage was small and only significant when the results after treatment were compared with the first LPS challenge. This is in line with data of COPD patients and of asthma patients after allergen challenge, obtained in two studies in which the treatment period exceeded 14 days, but did not find an effect on the percentage of sputum neutrophils [31, 32]. Furthermore, Roflumilast was not able to change the relative cellular composition of BAL in healthy subjects after 4 weeks of treatment and segmental LPS challenge, while it reduced absolute neutrophil numbers .
With respect to the total sputum cell count and the neutrophil cell count, we observed the lowest values after Roflumilast treatment, but this decline did not reach statistical significance compared with the second LPS challenge. Based on the data in primates, we hypothesized that a 5 day treatment duration, would prove efficacious on neutrophil cell numbers. Nevertheless, the small effect on cell numbers seen in our study is compatible with the effects seen in the above mentioned studies [4, 31, 32]. The strongest effect on sputum neutrophil cell numbers was observed after 4 weeks of treatment. Notably, even a 4 or 2 week treatment with Roflumilast did not have profound effects on sputum inflammatory mediators, which is in line with our results. The decrease of IL-8 in COPD was borderline significant, and Roflumilast did not change sputum IL-8 and MPO after allergen challenge in asthmatic patients [31, 32]. Finally, it could be speculated that in sputum, which has a higher baseline neutrophil count than BAL (approximately 20–30% compared to < 3%), there is less room for improvement of a given treatment upon endotoxin challenge and therefore effects on neutrophils are more difficult to demonstrate.
Interestingly, we found an increase in the expression of HLA-DR and CD86 on sputum macrophages. This increase was unexpected. While we can only speculate on the Roflumilast driven mechanism, we interpret this consistent finding in all subjects as an indicator for treatment compliance. Our study design was not randomized, as a clear sequence of experiments was required, in order to answer our questions. A randomized sequence would have been likely to show a larger treatment effect, but this would have been biased by an unnoticed tolerance effect.