The PERKs of ER homeostasis for hematopoietic stem cells Free

In this issue of Blood, Zheng et al¹ show that the unfolded protein response (UPR) sensor PERK (protein kinase R-like endoplasmic reticulum kinase) is dispensable for steady-state hematopoiesis, but promotes mammalian target of rapamycin (mTOR) signaling and hematopoietic stem cell (HSC) proliferation when protein quality controls fail, compromising their self-renewal.

Studies in the 1970s and 1980s showed that glucose starvation and protein misfolding triggered a shared set of stress-induced genes.² This led to the identification of the UPR pathway, initially dissected in yeast, that detects misfolded proteins in the endoplasmic reticulum (ER) and induces a transcriptional program to restore homeostasis. The UPR is activated by diverse stressors, including nutrient deprivation, hypoxia, and oxidative or proteotoxic stress. This activation produces different cell fate outcomes depending on the type, severity, and duration of stress: generally, an initial adaptive response turns proapoptotic under strong or sustained stimulation.³ The mammalian UPR is mediated by 3 ER sensors: IRE1α (inositol-requiring enzyme 1α), ATF6 (transcription factor 6), and PERK. PERK is a kinase that canonically phosphorylates eukaryotic translation initiation factor 2 subunit α (eIF2α), which reduces global protein synthesis while promoting translation of stress-response transcripts.

HSCs must maintain quiescence to preserve regenerative capacity to support life-long blood production. Intact protein quality control systems are critical to sustain integrity of the HSC pool. A propensity of HSCs to undergo apoptosis following ER stress was first demonstrated with chemically induced ER stress using tunicamycin (an inhibitor of N-linked glycosylation) and thapsigargin (which depletes ER calcium), which activate PERK and induce HSC apoptosis through ATF4 and CHOP (C/EBP homologous protein).⁴ In this case, UPR attenuation improved HSC function in xenotransplantation assays. In contrast, low-level ER stress can activate protective PERK signaling, as recently observed in vacuoles E1 enzyme X-linked autoinflammatory somatic syndrome, a systemic inflammatory disorder caused by a somatic mutation in the ubiquitin-activating enzyme UBA1 that creates proteotoxic stress in HSCs.⁵ In this context, PERK signaling supports the survival of mutant HSCs. These findings bracket the range of PERK responses from apoptosis under high stress to survival under mild stress.

Zheng et al now define a third, noncanonical mode of PERK activity.¹ First, the authors establish that PERK is largely dispensable during homeostatic hematopoiesis. Mice lacking PERK show normal HSC numbers, multilineage differentiation, and long-term reconstitution capacity. These findings, replicated across both Mx1-Cre and Vav1-Cre systems, establish that PERK does not play a major role in steady-state hematopoiesis.

Then, Zheng et al test the role of PERK in HSCs during ER stress. To induce ER stress, they use models of Sel1L or Hrd1 deletion in HSCs, which blocks ER-associated degradation (ERAD) machinery, a process that identifies and targets misfolded ER proteins for proteasomal degradation.⁶ The authors show that blocking ERAD leads to ER stress, loss of HSCs, reduced chimerism, and impaired hematopoietic reconstitution. However, these deleterious consequences are rescued by PERK deletion, indicating that HSC loss under stress is mediated by PERK. Although all 3 arms of the UPR are activated on deletion of ERAD components, only deleting PERK, or disabling its kinase function, restores HSC quiescence and function. This establishes PERK as the dominant UPR effector in this setting. Mechanistically, PERK activation correlates with increased mTOR signaling as observed by elevated phosphorylation of 4EBP1 and S6, whereas Hrd1/PERK double knockout attenuates mTOR complex 1 activity. These data indicate that in ERAD-deficient conditions, PERK promotes HSC depletion, in part by amplifying mTOR and proliferation.

Taken together, the literature discussed here supports a model in which PERK calibrates HSC fate based on the level and source of stress. Under high (chemically induced) and moderate (ERAD deficiency) PERK activation, HSCs can be cleared through apoptosis or proliferative depletion, whereas low ER stress (mutated UBA1) can induce protective signaling in HSCs. Further research is needed to illuminate the full range of outcomes that PERK can effectuate as a stress sensor that matches the cellular response to the degree of proteotoxic challenge.

The findings presented by Zheng et al highlight the importance of proteostasis for HSCs and add to our understanding of the factors involved in cell fate decisions under proteotoxic stress. They also raise new questions. What physiological conditions activate the UPR in HSCs in vivo? For example, do inflammation,⁵ aging-associated changes in autophagy,⁷ or oncogenic transformation⁸ activate UPR-dependent survival or attrition? Can PERK inhibition be used to preserve HSC function during myelosuppressive therapy, improve expansion in vitro, or selectively target UBA1-mutated HSCs? As our understanding of the proteostasis network in HSCs expands and we uncover its dysregulation in blood disorders, these and other therapeutic opportunities are on the horizon.

Conflict-of-interest disclosure: The authors declare no competing financial interests.