Lipopolysaccharide (LPS) is a component of Gram-negative bacterial cell wall. Once recognized by the immune system, LPS elicits a proinflammatory response (Zhang and Ghosh 2000) and has thus become extensively used in research for this purpose (e.g., van Dam et al. 1992; Gatti and Bartfai 1993; Laye et al. 1994; Breder et al. 1994; Quan et al. 1999). In particular, inflammation within the brain (neuroinflammation) can be obtained by central or peripheral administration of LPS (Rivest 2003). However, Tiwari et al. (2016) recently reported not observing neuroinflammatory effects of peripheral LPS administration (via repeated intraperitoneal injections) to rats. Why this discrepancy with the previous literature?

While it is common knowledge that negative results are less frequently published (Fanelli 2012), a close look at the literature (Table 1) demonstrates that the results of this study have an alternative explanation. In Table 1, I have summarized only the studies cited within Tiwari et al. (2016), where peripheral injection of LPS was used. Based on this table, it is possible to make two important observations: (1) within the range of doses of LPS administered in those studies (from 100 to 10,000 μg/kg), Tiwari et al. (2016) used one of the lowest doses (125 μg/kg, i.e., about 100 times less then the high end of the spectrum) and (2) studies using low doses of LPS have found neuroinflammatory effects within a short period of time post-injection (few hours), while studies using high doses have found neuroinflammatory effects in both the short- (few hours) and long-term (months). Tiwari et al. (2016) collected brain samples at a relatively long time period post-injection.

Table 1 Articles cited by Tiwari et al. (2016), where peripheral LPS administration was employed and respective examples of the alteration in the neuroinflammatory markers quantified
Full size table

Most studies in Table 1 do not reveal whether low doses of LPS elicit long-term responses, but there are two studies that do so. In the study by Biesmans et al. (2013), where several LPS doses were tested, the authors quantified neuroinflammatory effects of one intraperitoneal injection of LPS over time. The findings reveal that, at the dose of 630 μg/kg, a dose five times higher than that of Tiwari et al. (2016), several effects within the brain have already subsided at 24 h post-LPS injection. Similar results were obtained by Spulber et al. (2012), at a lower dose (330 μg/kg). Tiwari et al. (2016) collected their brain samples at 48 h after their last administration of LPS. Given the low dose and the time of collection, they should no longer observe neuroinflammatory changes, which they did not. Notably, a study in rats (Quan et al. 1999) found inflammation in certain brain regions 2 h post intravenous administration, using a dose of LPS more than ten times lower than the one used by Tiwari et al. (2016). Therefore, while the authors write that their finding of a lack of neuroinflammatory effect cannot be attributed to dosage or short- versus long-term effects of LPS, it is likely that it can in fact be attributed to the combination of these two variables. Thus, their conclusion that “LPS (i.p.) administration is devoid of any neuroinflammatory effects” should be placed in the context of the dose used and the timing they chose to collect their samples. Finally, the title of their article “Redefining the role of peripheral LPS as a neuroinflammatory agent…” must be considered carefully—the role of LPS cannot be redefined based solely on a single study where a low dosage has been administered and where samples are collected after a long time period has elapsed. Combined, the study by Tiwari et al. (2016) and the literature summarized here should serve as a warning for future studies about the importance of considering both dose and timing when neuroinflammatory effects are expected.