Molecular Evaluation of Endoplasmic Reticulum Homeostasis Meets Humoral Immunity

Abstract

The biosynthesis of about one third of the human proteome, including membrane receptors and secreted proteins, occurs in the endoplasmic reticulum (ER). Conditions that perturb ER homeostasis activate the unfolded protein response (UPR). An 'optimistic' UPR output aims at restoring homeostasis by reinforcement of machineries that guarantee efficiency and fidelity of protein biogenesis in the ER. Yet, once the UPR 'deems' that ER homeostatic readjustment fails, it transitions to a 'pessimistic' output, which, depending on the cell type, will result in apoptosis. In this article, we discuss emerging concepts on how the UPR 'evaluates' ER stress, how the UPR is repurposed, in particular in B cells, and how UPR-driven counter-selection of cells undergoing homeostatic failure serves organismal homeostasis and humoral immunity.

Keywords: B cell development; RIDD; antibody production; endoplasmic reticulum; proteostasis; unfolded protein response.

Figure 1. The UPR circuitry

The UPR in humans is signaled at least via three branches: PERK, IRE1α, and ATF6α. Upon activation by ER stress, PERK mediates the phosphorylation of eIF2α, which leads to translational attenuation. A few transcripts are exempted from translational attenuation, in particular, the expression of the PERK downstream effectors, ATF4 and ATF5 is enhanced when eIF2α is phosphorylated. Targets of ATF4 and ATF5, such as the ATF4 target CHOP, are induced as a consequence. Since next to PERK a few other stress-responsive kinases (PKR, HRI, and GCN2) can drive phosphorylation of eIF2α as well, the effector program downstream of the transcription factors ATF4, ATF5, and CHOP is referred to as the Integrated Stress Response. Upon its activation IRE1α removes an intron from unspliced XBP1 mRNA (XBP1^U), yielding spliced XBP1 mRNA (XBP1^S), which encodes the transcription factor XBP1. Next to XBP1 mRNA, IRE1α can target other RNA molecules in a process referred to as RIDD. Upon activation ATF6α is severed in its transmembrane region, releasing its cytosolic domain, ATF6α-p50, which serves as a transcription factor that, like XBP1, drives expression of ER chaperones and ERAD components. As such, the ER protein folding machinery as well as its machinery to dispose of misfolded ER client proteins expands. Thus, the combined efforts of the three UPR branches aim at ER homeostatic readjustment [6].

Figure 2. A unified model for ratiometric ER stress sensing and UPR activation

The way the UPR sensors are activated has been highly debated. Since UPR activation coincides with the ER chaperone BiP dissociating from the UPR sensors, it was proposed that titrating BiP away from those sensors by unfolded ER client proteins would be sufficient for UPR activation [18,19]. A contrasting model was put forward when it was shown that unfolded proteins can serve as activating ligands of UPR sensors, namely that the extent of UPR activation would tie in with the extent of unfolded protein associating with the sensors [20]. However, ER stress sensing by the UPR is commensurate with the ratio of unfolded ER client protein levels over BiP [8]. Therefore, the two models are not contradictory, but complementary, since the ratio of UPR sensors bound to unfolded protein over UPR sensors bound to BiP, by essence, would serve as the best readout for the ratio of unfolded protein over BiP. Thus, we posit that the two models can be unified in a single model of ratiometric UPR sensing [8,9].

Figure 3. RIDD and the transitioning from an ‘optimistic’ to a ‘pessimistic’ UPR

When ER homeostatic readjustment is successful, it entails that the UPR has ensured that BiP levels eclipse those of its unfolded ER client proteins again [8]. Yet, when unfolded proteins persistently eclipse BiP, PERK and IRE1α are chronically maximally activated [9]. Since persistent IRE1α activity entails that the pre-existing pool of unspliced XBP1^U mRNA is depleted (as all will be converted into spliced XBP1^S mRNA), IRE1α automatically commits to RIDD as its activity will be directed against other substrates [36]. As such, splicing-to-RIDD transitioning is a readout for persistent maximal IRE1α activation and, hence, for persistent ER stress. Splicing-to-RIDD transitioning can serve as a molecular device for the cell to switch from being ‘optimistic’ to ‘pessimistic’ when ER homeostatic readjustment cannot be achieved. The onset of splicing-to-RIDD transitioning depends also on the ratio of the preexisting levels of XBP1 mRNA over those of IRE1α, which are developmentally determined. An established example of how splicing-to-RIDD transitioning can initiate pro-apoptotic pathways is that TXNIP, which is a downstream target of PERK (via ATF5), will activate the inflammasome. Normally, the TXNIP transcript is targeted for degradation by miR-17, but once IRE1α commits to RIDD of miR-17, TXNIP is de-repressed. Thus, PERK and IRE1α activity converge in jointly driving a pro-apoptotic UPR [13,34,35]. Yet, the outcome of splicing-to-RIDD transitioning, and, hence, of the UPR turning ‘pessimistic’ may differ depending on the cell type, as other RIDD targets come into play.

Figure 4. How (re-)purposing the IRE1α/XBP1 UPR branch serves humoral immunity

The IRE1α/XBP1 branch of the UPR is key for the humoral immune response, since antibody production is compromised when either IRE1α, or XBP1, or both, are ablated [,,–46]. We argue that the purposing and repurposing of the IRE1α/XBP1 branch shapes the B cell lineage at least at three stages. Firstly, when pro-B cells undergo VDJ recombination and somatic hypermutation (SHM): the resulting HCs in many cases are unfit to fold and/or to assemble with SLC [53]. Likewise, recombination and SHM of the Ig-κ/Ig-λ loci in pre-B cells often results in LCs that cannot assemble with the HC into a functional BCR [61]. In either case of unsuccessful HC/(S)LC folding and/or assembly, ER homeostatic failure ensues, and IRE1α may serve as the executor of ‘HC toxicity’ by initiating pro-apoptotic pathways upon splicing-to-RIDD transitioning. Secondly, upon antigen stimulation, B lymphocytes massively induce XBP1 expression [14,42] to an extent that basal IRE1α activity likely yields sufficient XBP1 protein to drive ER expansion in absence of ER stress. As such, an antibody production plant is ready before the onset antibody production [38]. Finally, once bulk antibody production is underway, the HC often accumulates in the ER to (re-)activate the IRE1α/XBP1 branch, such that ensuing XBP1 production reinforces ER function. In the plasma cell, splicing-to-RIDD transitioning does no longer herald apoptosis, since μ_s mRNA is a key RIDD target. Thus, when all XBP1^U mRNA is consumed, RIDD serves to curtail excess secretory HC expression [14,46].