A recent report by Damjanov et al (1) described nonsignificant therapeutic effects of a selective p38 MAPK inhibitor in patients with rheumatoid arthritis (RA), and a recent editorial by Genovese (2) discussed the disappointing outcomes of several clinical trials of p38 inhibitors in RA and in other chronic inflammatory diseases. Blockade of this signaling pathway has consistently had little or no clinical benefit and has resulted in only transient suppression of markers of inflammation. It has been argued that the lack of efficacy of p38 inhibitors “must lie in one or more biologic adaptations that allow escape from this pathway” (2). Therefore, “to best understand the failure of this class of therapies, we need to better understand this biologic adaptation” (2).
MAPK p38 inhibitors were first discovered using simple in vitro screens for compounds that blocked production of tumor necrosis factor (TNF) and interleukin-1 (IL-1) in myeloid cells stimulated with lipopolysaccharide (LPS). Since that initial identification of p38 as a promising therapeutic target, many increasingly potent and specific inhibitors have been generated. Together with genetic and molecular approaches, these compounds have been used in investigations to determine exactly what p38 does, and how. It was soon recognized that p38 not only promotes expression of proinflammatory cytokines but is also activated by them. This potential feed-forward loop creates the danger of self-amplifying, runaway inflammatory responses. Such an outcome is prevented, in part, by p38-dependent antiinflammatory mechanisms. For example, the p38 pathway controls the expression of important antiinflammatory factors that limit inflammation and contribute to its resolution; consequently, long-term blockade of the pathway may have little net benefit. It also participates in crosstalk that constrains the activity of other proinflammatory pathways, with the result that p38 inhibitors may enhance other proinflammatory processes. To use a crude metaphor, p38 has one foot on the gas and one foot on the brake. An awareness of this characteristic may help us to understand the disappointing outcomes of clinical trials of p38 inhibitors.
Inflammatory stimuli (e.g., IL-1, TNF, or ligands of the Toll-like receptors) engage their diverse receptors, and all promote activation of the transforming growth factor β–activated kinase 1 (TAK1) signaling complex (Figure 1). TAK1 then activates MKKs and IKK, which is responsible for the activation of the transcription factor NF-κB. The MKKs phosphorylate and activate both p38 and JNK, another member of the MAPK family. (There are 4 isoforms of p38: α, β, γ, and δ. Herein we generally refer to p38α, the major ubiquitous isoform and the most important regulator of inflammatory responses. Most, although not all, inhibitors are active against p38α and p38β.) Once phosphorylated and activated, p38 phosphorylates its own substrates, which include downstream kinases such as MAPKAPK2 (MK2).
Together, the NF-κB, JNK, and p38 pathways regulate the expression of inflammation mediators, including cytokines and chemokines, cyclooxygenase 2, and many other proteins. Both transcriptional and posttranscriptional effects of p38 are critical to this coordinated response. Many inflammatory effectors are encoded by messenger RNA (mRNA) that contain destabilizing AUUUA motifs. The RNA binding protein tristetraprolin (TTP) recognizes these sequence motifs and directs mRNA for rapid degradation (3). The antiinflammatory function of TTP is well illustrated by the erosive arthritis found in TTP−/− mice, which is largely caused by overexpression of TNF. When activated by p38, MK2 phosphorylates and inactivates TTP, allowing accumulation and translation of mRNA for TNF and other inflammatory cytokines.
Figure 1 illustrates antiinflammatory mechanisms that are controlled by the p38 pathway. First, the expression of IL-10, a powerful antiinflammatory cytokine with an important role in the resolution of inflammation, is strongly dependent on p38 (4, 5). Ironically, p38 inhibitors could have also been identified by screening for compounds that blocked IL-10 expression. Second, blockade of p38 signaling can reduce expression of TTP and interrupt the normal “off” phase of inflammatory gene expression that involves the degradation of mRNA encoding mediators of inflammation (6, 7). Third, inhibiting p38 may augment JNK activity by reducing the expression of MAPK phosphatase 1. This enzyme dephosphorylates and inactivates both p38 and JNK, and hence restrains inflammatory responses (8–10). Finally, in a similar way, p38 inhibitors can prevent the negative feedback control of TAK1 activity and thus enhance activity of JNK and NF-κB pathways (11).
In clinical trials, “escape” from, or lack of, antiinflammatory effects may involve disruption of one or more of these antiinflammatory mechanisms. It is likely that therapeutic targeting of kinase pathways will often run the risk of perturbing negative feedback and crosstalk mechanisms, whose function is to constrain cellular responses. It might be worth considering whether the endogenous feedback mechanisms are disturbed in chronic diseases such as RA, in which mediators of inflammation continue to be expressed in the absence of a recognizable stimulus. If the breakdown of negative feedback loops contributes to chronicity, it may be possible to exert antiinflammatory effects by repairing or reinforcing these natural “off” mechanisms, rather than by simply blocking the p38 pathway.