4. Discussion
In this systematic review and meta-analysis, we demonstrated the efficacy of TNF-α, IL-1β, and IL-6 as in vitro inflammatory biomarkers in the LPS-induced RAW 264.7 cell model, and showed that NO levels still are the gold standard test to trial anti-inflammatory agents.
As expected, NO levels increased significantly when RAW 264.7 cells were induced by LPS compared to non-induced cells, independently of cell density, but cytokine levels changed according to cell density. To ensure better screening for novel drugs, the trial could be initiated by screening for drugs showing NO levels higher than 20 µM by LPS-induced RAW 264.7 cells with statistical significance compared to non-induced cells. After this initial screening, it was found that all the cytokines studied were excellent inflammatory biomarkers, since in all the studies where NO produced by cells was > 20 µM, the cytokines were significantly increased. Moreover, all cytokine levels presented a positive correlation with the cell density used in the experiments. However, significant heterogeneity was found among the studies, mainly due to the different methodologies presented by the authors. Therefore, it is essential to standardize the inflammation model in RAW 264.7 cells, following some specific guidelines for in vitro experimentation.
Considering our results, it is important to highlight that due to the model chosen for the review and the characteristics of the included studies, a meta-analysis was performed. The model applied to use LPS as a phlogistic agent is one of the main components of gram-negative bacteria (Romerio & Peri, 2020). This agent has the characteristic of activating Toll-like receptors present in cells of the immune system, including the macrophages. In Gram-negative bacterial infections, for example, in the early stages of the sepsis process, this interaction between LPS and cellular receptors leads to the activation of intracellular signaling cascades activating many nuclear transcriptional factors. This results in massive secretion of several pro-inflammatory mediators, such as NOx, TNF-α, IL-1β and IL-6, the focus of our research (Ciesielska et al., 2021; Poltorak et al., 1998). For this reason, the parallel increase of NO and cytokines when culture cells are stimulated with LPS was already expected (Dai et al., 2018). However, an important technical detail in the cell culture must be taken into account, which is the incubation time of the cells after LPS induction. Although the action of TLR4 by LPS, ends with the production of many inflammatory mediators, their release by the cells, presents a different time-course of secretion (Hobbs et al., 2018). This fact can produce different results in the levels of these inflammatory mediators, making comparison between them unfeasible (Liu et al., 2017; Rahman & McFadden, 2011).
As previously commented in the introduction section, this model is widespread among research laboratories worldwide, as it provides rich information on the inflammatory response and helps initiate the search for new anti-inflammatory and immunomodulatory drugs (Elisia et al., 2018; Taciak et al., 2018). However, there is still a lack of standardization regarding LPS doses, cell induction time, and cell density used, among others. To resolve these shortcomings in the standardization of this inflammatory model, we were very judicious in the application of eligibility criteria and in the data extraction, in order to ensure the reliability of the results obtained.
Our results show, as expected, that there is an increase in NO production when RAW 264.7 cells are induced by LPS (Moore et al., 2019; Ranaweera et al., 2020). Thus, considering the characteristics of the studies included in the meta-analysis, it is plausible to affirm that the NO measurement in the model in question is sufficient for the purposes of initial inflammatory screening, considering that the production of the pro-inflammatory cytokines in question accompanies the same behavior as that observed with the determination of NO levels, at least, when the time of cell incubation after LPS stimulation is between 12h and 24h. This data is important because, to our knowledge, there is no standardization among researchers in regard to the choice of initial inflammatory mediators to measure in this model, in screening for the possible anti-inflammatory or immunomodulatory activity of some compound or extract from a natural product. Moreover, we also found that there is no consensus among researchers regarding the biomarkers used for screening new anti-inflammatory drugs, cell density, LPS concentration used, time of induction, and use of culture medium supplementation.
Thus, taking into account the information collected, we strongly advise that the use of NO measurements be standardized when used to screen for possible compounds with anti-inflammatory activity. Obviously, taking into account the particularities of each laboratory is extremely important, however, the use of LPS at a dose of 1ug mL-1, cell density between 1 and 9.9x105, and incubation time of between 12 and 24 hours at 37°C, appear to be the minimum requirements for this standardization (Supplementary Material S10).
Finally, regarding the risk of bias and heterogeneity of included studies in this meta-analysis, we believe that due to this lack of standardization among published papers, researchers have selected a range of different mediators for their experiments. In our study, it was demonstrated that only NO levels are sufficient to draw the initial conclusions when seeking to determine whether some compound should be submitted to the next levels of anti-inflammatory experimentation, when the culture of Raw 264.7 cells induced by LPS is used as an initial experimental platform. The use of a single marker is also economically viable, as the technique is easy to perform, with good reproducibility and low cost, besides being quicker than using several markers.
Regarding the strengths, in this study, all steps were performed in pairs and validated by an expert. Also, a meta-analysis and risk of bias were performed. No time was limited to literature search. Some limitations were considered regarding heterogeneity between studies and the rigorous inclusion criteria.