Biological systems respond to environmental cues with two distinct output modes. In an analog output mode the response is linear and proportionate to the strength of the agonist (, left bottom panel). In a digital output mode the response is all or none at an individual cell level (, right bottom panel). With the increasing strength of the agonist, the number of cells but not signaling intensity per cell increases (, top right panel). One of the interesting properties of the ERK1/2 signaling module is its ability to switch back and forth between analog and digital output mode (100
). Digital systems manifest hysteresis. In physics hysteresis is history dependence of physical systems (103
). This frequently applies to magnetic materials -- as the external field with the signal from the microphone is turned off, the little magnetic domains in the tape do not return to their original configuration, thus storing memory (music). Systems that display hysteresis can toggle between two alternative stable steady states. This is known as system bistability. A bistable system exists in two states—an “on” state, and an “off” state. These two states are transitioned through an unstable intermediate state (101
). Early examples of biological bistable systems include the lambda phage lysis-lysogeny switch and the hysteretic lac repressor system (104
). Lisman in 1985 first suggested that a bistable system could serve as a self-sustaining biochemical memory (105
Figure 2 Top panel: A diagrammatic presentation of analog and digital output of pERK1/2 with the increasing concentration of an agonist. Bottom panel, left: A monostable system showing a linear dose-response relationship. Right: A bistable system showing a non-linear (more ...)
Bistability arises from a positive feedback loop or a mutually inhibitory, double-negative feedback loop (102
). Ferrell and associates have mathematically shown that when the strength of positive feedback exceeds a certain limit (f=0.08), the system shows hysteresis and becomes bistable (106
). In confirmation they have shown that progesterone-induced activation of MAPK and Cdc2 induces a self-sustained activation mechanism in frog oocytes, which is dependent upon a positive feedback loop (107
). Disruption of the positive feedback loop at the level of c-Mos (Raf-1) abrogates the signaling memory. ERK1/2 bistability is not unique to oocyte as it has also been studied in mammalian cells. Using a combination of computational and biological experiments Bhalla and Iyengar showed that a single stimulation of fibroblasts with PDGF led to ERK1/2 MAPK activation, whose duration was dependent upon a positive feedback loop through cytoplasmic phospholipase A2 (cPLA2) and protein kinase C (PKC) (108
). The duration was shortened in the presence of cPLA2 and PKC inhibitors. The duration of ERK1/2 signal is also determined by the level of phosphatases that are present in the cells. ERK1/2 is regulated by MAPK specific phosphatases as well as non-specific phosphatases. The authors showed that at a low level expression of MKP1 ERK1/2 showed a switch-like “off’ and “on” response, which is typical of digital output from a bistable system. In contrast, at a high level expression of MKP1 the system showed a gradual time and dose-dependent response, typical of analog output from a monostable system. In T cells digital vs analog output is determined by the nature of the agonist used. Superantigen-loaded dendritic cell stimulation results in digital output whereas stimulation with SDF-1 elicits an analog output (109
). KSR1, a scaffold protein, plays a crucial role in switching the output from analog to digital. Bistability is not unique for ERK1/2. Its presence has also been demonstrated in other MAPK family members (110
) as well as upstream activators—Ras (111
) and Sos (112
). Since many biological processes have built-in negative and positive feedback loops, they all could potentially function as bistable systems.
As mentioned, bistability can function as a biochemical memory. For this reason, the ERK1/2 signaling pathway has been studied in the CNS for its potential role in memory. Indeed, ERK1/2 and one of its downstream effector CREB have been shown to play a critical role in neuronal plasticity, long-term potentiation (LTP) and memory formation (reviewed in ref. 113
). Mice with functional ERK1/2 deficiency have impaired protein synthesis as well as selective defects in LTP and memory consolidation (115
). The definition of memory is the retention of an acquired (learned) signal. In neuronal networks, the result of this signal is expression of an activation path whose physiological manifestation is a particular pattern of neural firing. System memory in organs outside the central nervous system (CNS) is rarely invoked and studied. There are memory T cells and B cells but their mechanism is different from that in the CNS. T and B cell memory is maintained by the persistence of antigen-specific cell clones and can be evoked by a specific antigen during a later encounter. Whether or not these cells utilize a signaling memory is unknown.
The CNS memory is classified into episodic memory and semantic memory (118
). Episodic memory is the explicit memory of events and context, which includes time, space, and associated emotions. Semantic memory is knowledge independent of context. It is a form of declarative memory that requires repetition for consolidation (119
). The repetition of the memory-forming event for consolidation is an important concept. Indeed, when one carefully examines the development of a chronic disease such as asthma, repetition of the initial inciting trigger seems to be a common finding. Many children develop wheezing during a viral infection. This usually resolves without sequelae. However, if the wheezing recurs on subsequent infections, this recurrence usually heralds the development of chronic asthma (120
). Once developed the affected children begin to experience wheezing and other asthma symptoms when exposed to diverse environmental stimuli without regard to the viral infection. This suggests that repetition of the inciting trigger leads to the formation of a system memory, which then drives the process in a self-perpetuated manner. We applied this logic of repeated stimulation inducing system memory in signaling studies. We asked if repeated stimulation would lead to the development of a self-perpetuated mechanism of activation of ERK1/2 (formation of bistability). We tested this hypothesis using airway epithelial cells.
A single stimulation of epithelial cells with cytokines and growth factors (IL-4, IL-6, IL-13, eotaxin, and EGF) causes rapid ERK1/2 activation, which returns to baseline in 24 hr. Interestingly, repeated stimulation on 3 consecutive days leads to sustained activation of ERK1/2 but not JNK, p38 or STAT6 (122
). The ERK1/2 activation lasts for 3-7 days and depends upon a positive feedback mechanism involving spry 2 and Fyn. Repeated stimulation leads to increasing expression of the adapter protein Spry 2, which directly activates Fyn kinase. The activation of Fyn creates a positive feedback loop by further activating ERK1/2 () (122
). It also creates a double negative feedback loop by phosphorylating and inhibiting the E3 ubiquitin ligase Itch (123
). Itch physiologically ubiquitinates JunB, a downstream ERK-inducible transcription factor (124
). By inhibiting Itch Fyn stabilizes JunB and further augments spry 2 production. Spry 2 is known to inhibit another E3 ubiquitin ligase Cbl, which is responsible for receptor degradation (125
). By inhibiting Cbl spry 2 prolongs receptor signaling. This is another example of a double negative feedback loop. Thus, there are multiple positive and double negative feedback loops that can perpetuate ERK1/2 signaling in the cell. Overexpression of spry 2 induces and its genetic deletion abrogates ERK1/2 bistability. The sustained pERK1/2 is excluded from the nucleus and partially compartmentalized to the Rab5+ endosomal compartment. Consequently, many genes which are induced during a single stimulation are silent upon repeated stimulation. The epigenetic regulation may also be involved as demonstrated in the repeated stimulation model of macrophages using LPS (128
). We have shown that there is only a small group of genes which remains active in repeatedly stimulated epithelial cells and this group includes regulators of MAPK signaling (122
). We hypothesize that these MAPK regulators contribute to sustained ERK1/2 signal.
A schematic presentation of positive and double negative feedback loops leading to ERK1/2 bistability
What are the benefits of sustained but compartmentalized ERK1/2 activation in epithelial cells? We have shown that cells with this sustained endosomal pERK1/2 manifest resistance against growth factor withdrawal-induced cell death (122
). The survival pathway likely involves endosomal proteins as cells with reduced amounts of Rab5 demonstrate increased apoptosis. Cells with sustained endosomal pERK1/2 are also primed for heightened cytokine production. Prolonged cell survival and cellular priming are two important features of asthma. We examined the biological relevance for ERK1/2 bistability in asthma (122
). Epithelial cells from human asthma and from the mouse model of chronic asthma manifest increased pERK1/2. A significant part of this pERK1/2 is associated with Rab5+ endosomes. The increase in pERK1/2 is associated with a simultaneous increase in spry 2 expression in these tissues. The results from this study suggest that the ERK1/2 pathway develops bistability upon repeated stimulation. We speculate that ERK1/2 bistability serves as a signaling memory for epithelial priming of the immune system in chronic asthma.