Going through changes – the role of autophagy during reprogramming and differentiation Free

Organisms are always engaged in a relentless quest to adapt to ever-shifting habitats, as a failure to do so may result in their demise (Dinneny et al., 2008). Adaptation requires the activation and termination of a diverse array of cellular processes, which are dependent on perception of a myriad of signals. The orchestration and rearrangement of these different cellular processes or programs can be referred to as temporary reprogramming, which is characterized by being transient and thus reversible. In this context, it can be argued that whenever cells need to accommodate change, they undergo reprogramming (Davière and Achard, 2016; Hafner et al., 2019; Koo and Guan, 2018; Rodriguez et al., 2020; Xu et al., 2017; Zhang et al., 2018).

In addition to transient reprogramming, cells can also undergo a more definitive type of reprogramming. Under normal developmental conditions, young cells integrate endogenous and exogenous cues that help them to shape their developmental pathway into maturity, reaching what are often assumed to be terminal identities (Aydin and Mazzoni, 2019). This journey of identity acquisition is achieved through a suppression and/or dampening of the innate plasticity capacity (stemness) of the cell and the progressive differentiation into a final, mature identity (Di Stefano, 2022). Although cellular differentiation stifles this plasticity, it is still possible for differentiated cells to undergo fate changes and acquire new identities, for instance, during regeneration of damaged tissue (Kanne et al., 2022; Kopp et al., 2016; Li et al., 2017; Miao et al., 2020). Changes in cellular fate can occur in different ways. Mature cells can transdifferentiate into another mature cell type without reverting into stem cells (Merrell and Stanger, 2016). Likewise, mature cells can also undergo dedifferentiation and acquire stemness (to varying degrees), as seen during formation of induced pluripotent stem cells (iPSCs) (Sang et al., 2018).

Regardless of the classification of cellular reprogramming, research both in animals and plants has mostly been focused at the epigenetic and transcriptional level (Churko et al., 2017; Ikeuchi et al., 2015; Ishikawa et al., 2019; Iwafuchi-Doi and Zaret, 2014; Iwase et al., 2021; Kim et al., 2018; Lazure et al., 2023; Miao et al., 2020; Miyamoto et al., 2015; Papp and Plath, 2013; Roche et al., 2017; Vainorius et al., 2023). In contrast, the role of proteostatic mechanisms during reprogramming events remains comparatively less understood. However, it should come as no surprise that to enact reprogramming, cells need to terminate certain cellular programs and degrade signatures of previous states and fates (Mizushima and Komatsu, 2011). The main candidates for proteome remodeling during reprogramming are the ubiquitin–proteasome system (UPS) and autophagy. The proteasome, with its high specificity mediated by protein ubiquitylation, has been linked to stem cell formation (Buckley et al., 2012; Floyd et al., 2015; Vilchez et al., 2012). In contrast, autophagy was initially only considered to be a starvation-induced mechanism for the degradation of random cytoplasmic portions to supply metabolic building blocks to sustain cellular functions under stress (Yang and Klionsky, 2013), and, therefore, its role during reprogramming was mostly overlooked. Subsequent studies have shown that autophagy is engaged upon the perception of a wide range of stimuli to facilitate termination of unnecessary programs, to allow novel cellular programs to be executed correctly and to deal with stress (Rodriguez et al., 2020). Moreover, the identification of diverse cargo adaptors has revealed that autophagy can selectively degrade targets (Acheampong et al., 2020; Nicot et al., 2014; Nolan et al., 2017; Stephani et al., 2020; Svenning et al., 2011; Wu et al., 2021; Zhou et al., 2013), thus allowing autophagy to participate in reprogramming events as part of a more targeted approach to facilitate the transition to different cellular states.

In this Review, we focus on the role of autophagy during cellular reprogramming in plants and draw parallels to what is known in other organisms.

Plants are recognized as having a remarkable capacity to regenerate whole tissues and even complete organisms from a few cells. In vascular plants, this regenerative capacity is derived from stem cell niches located in the apical meristems, which can generate all cell types, thus supporting post-embryonic organ development (Ikeuchi et al., 2019). In root tissue, cells rapidly divide in the meristem until they reach the transition zone, where cell division slows down, and cells instead increase in volume (Motte et al., 2019). Cells thus engage in active growth in the elongation zone until they reach a terminal size and enter the differentiation zone, where they acquire maturity (Burkart et al., 2022; Dolan et al., 1993; Sabatini et al., 2003; van den Berg et al., 1997). Interestingly, although post-embryonic root formation is imprinted before cells leave the elongation zone, in a region referred to as the oscillation zone, novel roots mostly form in the differentiation zone, and thus these novel lateral roots (LRs) are formed through naturally induced cellular reprogramming (Banda et al., 2019). In the plant model Arabidopsis thaliana (and dicots in general), LRs are derived from the xylem-pole pericycle (XPP) cells, which retain a certain level of stemness. This stemness enables XPP cells to rapidly undergo cell fate changes to become lateral root founder cells (LRFCs) and continue differentiating to form a completely new meristem (Casero et al., 1993; Laskowski et al., 1995; Möller et al., 2017; Parizot et al., 2008; von Wangenheim et al., 2016).

At a molecular level, LR formation is controlled by the phytohormone auxin and requires a signaling node composed of SOLITARY ROOT (SLR, also known as IAA14), the AUXIN REPONSE FACTOR proteins ARF7 and ARF19, and LATERAL ORGAN BOUNDARIES DOMAIN proteins (LBD proteins, also known as ASL proteins) (De Smet et al., 2010). In the absence of auxin, SLR represses ARF7 and ARF19 to prevent LR initiation (Fukaki et al., 2002). High auxin levels promote SLR degradation via the UPS (Gray et al., 2001), consequently releasing ARF7 and ARF19 to enact the transcriptional reprogramming that is necessary for XPP cells to initiate the cell fate changes that lead to post-embryonic organ development (Okushima et al., 2007). Specifically, ARF7 and ARF19 coordinate many of the steps during LR formation, including the initial selection of which XPP cells are primed to become LRFCs, re-entry into cell division, cell proliferation and LR emergence (Goh et al., 2019; Ito et al., 2016a; Okushima et al., 2007; Perianez-Rodriguez et al., 2021).

The cellular reprogramming events that induce LR formation can be exploited in vitro to form plant iPSC masses (calli), and these calli can subsequently be used to regenerate organs (roots and shoots) and even whole plants through the culturing of tissue explants in media with different ratios of auxin and the phytohormone cytokinin, which promotes cell division and cytokinesis (Valvekens et al., 1988). As the programs needed for formation of LRs and plant iPSCs are similar (Atta et al., 2009; Sugimoto et al., 2010), many of the key factors have already been identified, and their regulation at the epigenetic and transcriptional level has been characterized. Precisely as is seen in mammals with the overexpression of the Yamanaka transcription factors (Takahashi and Yamanaka, 2006), plant-specific transcription factors can be overexpressed to yield ectopic stem cell formation (Fan et al., 2012; Li et al., 2017; Xu et al., 2018). Likewise, epigenetic reprogramming is also essential to achieve pluripotency in plants (Kim et al., 2018), indicating overall that many of the steps on the path to achieving stemness are similar across kingdoms.

Unlike flowering plants, which evolved pericycle cells retaining a certain level of stemness, bryophytes have apices housing a solitary stem cell without an accompanying cell niche. The process of stem cell formation in bryophytes has been extensively investigated in mosses and liverworts (Ishikawa and Hasebe, 2022). If a leaf of the moss Physcomitrium patens is excised, differentiated cells at the damaged margin reprogram into apical stem cells (Ishikawa et al., 2011). Thus, in contrast to what is seen in mammals and flowering plants, where iPSC generation requires either overexpression of master transcription factors (Takahashi and Yamanaka, 2006; Xu et al., 2018) or the application of exogenous phytohormones (Atta et al., 2009; Sugimoto et al., 2010; Valvekens et al., 1988), in P. patens, wounding alone is sufficient to trigger iPSC formation (Ishikawa et al., 2011; Kanne et al., 2022; Li et al., 2017; Rodriguez et al., 2020). Interestingly, the LIN28 protein, in conjunction with other factors, can induce formation of pluripotent stem cells from fibroblast cells (Yu et al., 2007), and similarly in moss, the forced expression of LIN28 homologs and the transcription factor STEMIN1 enhances somatic reprogramming (Li et al., 2017). Therefore, this model plant might be a highly suitable tool to determine how to generate iPSCs under more physiologically relevant conditions applicable for both animals and flowering plants (Ishikawa et al., 2019).

Collectively, all these similarities between plant and animal reprogramming should not be surprising given that core eukaryotic cellular mechanisms are conserved between plants and animals (Sang et al., 2018).

Most of the examples in the literature regarding a role for autophagy in reprogramming come from animal research, where autophagy has been shown to impact different steps of the cellular decision making required to achieve and maintain stemness (Guan et al., 2013) (Fig. 1). As stated above, for somatic cells to reprogram, they need to undergo a series of steps aimed at both unlocking stem cell programs and terminating programs that define mature identities. Defining a specific sequence of steps necessary for somatic cells to become stem cells is complicated; reprogramming is a dynamic process that depends on the identity of the somatic cell undergoing reprogramming (David and Polo, 2014). Thus, the order of steps followed to obtain stemness and the speed at which these steps are achieved might differ for each cell type, or even within the same cell type due to stochastic fluctuations in cellular identity (Buganim et al., 2013; Hanna et al., 2009). Still, stemness acquisition in mammalian cells has been roughly classified into two phases: an initial, stochastic phase and a late, deterministic phase (Buganim et al., 2013). The initial phase comprises a burst of transcriptional reprogramming, removal of molecular determinants defining cellular identity of the somatic cell, remodeling of organelles and cytoplasm, metabolic reprogramming, and re-entry into active cell division. After this, a second wave of transcriptional reprogramming leads to expression of pioneer stemness factors and initiation of pluripotency acquisition. This step, which is sometimes considered an intermediate or transition step between early and late reprogramming, is still stochastic in nature and is often a bottleneck for reprogramming efficiency (Buganim et al., 2013).

The late stages of reprogramming, in which cells undergo maturation and stabilization, are characterized by deterministic expression of key pluripotency transcription factors, which allows cells to achieve pluripotency and have their epigenetic memory reset, and confers the ability to sustain self-renewal (Buganim et al., 2013; David and Polo, 2014).

In the sections below, we specifically look at how autophagy impacts on the capacity of somatic cells to achieve and maintain stemness, as well as how autophagy affects differentiation of stem cells into more mature cell types.

Evidence indicates that autophagy-mediated proteome remodeling is activated during the early steps of iPSC reprogramming, as seen for mouse fibroblasts (Wang et al., 2013; Wu et al., 2015) and during salivary gland renewal upon tissue damage (Orhon et al., 2022). However, it has also been recognized that autophagic activity must be tightly regulated during reprogramming: early autophagic activation during pluripotency acquisition appears to be transient in nature, and incorrect activation at critical steps could even be detrimental to reprogramming efficiency (Wang et al., 2013). In accordance with this, it has been shown that the Yamanaka transcription factors induce an early burst of autophagic activity, but in later stages, autophagy reduces reprogramming efficiency (Wu et al., 2015). This temporal restriction of autophagy during reprogramming events could be partially explained by the fact that autophagy is inhibited during mitosis (Eskelinen et al., 2002; Furuya et al., 2010), and thus erroneous activation of autophagy might interfere with stem cell proliferation. However, an overview of the literature reveals conflicting findings, which complicates our understanding of the precise role of autophagy during reprogramming. For instance, activation of autophagy during influenza infection has been found to decrease pluripotency of human stem cells (Zahedi-Amiri et al., 2019).

In plants, we have previously shown that autophagy inhibition substantially impairs stem cell proliferation in Arabidopsis, with atg2 and atg5 loss-of-function mutants showing a reduction of ∼80% in the production of stem cell masses compared to the production by control plants (Rodriguez et al., 2020). We rationalized that this impaired iPSC capacity of atg mutants was because of a reduced capacity to modulate their proteome and thus allow activation of the pluripotency program (Fig. 2). In accordance with this, mass spectrometry analyses revealed that upon treatment with auxin and subsequently with the phytohormone cytokinin, which antagonizes auxin, ATG2-deficient plants displayed remnants of the ‘auxin-induced proteome’, which had been removed in wild-type plants (Rodriguez et al., 2020). It should be noted that the defects in plant iPSC generation that are observed for atg mutants can be partially circumvented by prolonging cultivation of the obtained stem cell masses in iPSC medium (Rodriguez et al., 2020). This suggests that although stemness acquisition can occur in the absence of autophagy, this degradation mechanism is necessary to optimize cellular reprogramming and efficient stem cell formation (Rodriguez et al., 2020).

Importantly, although autophagy-mediated proteome remodeling is a core feature of unlocking stemness, there are only a few accounts of autophagy directly regulating pluripotency regulators like SOX2 and NANOG (Sharma et al., 2022; Zhou et al., 2022). In this context, we have recently found that auxin promotes selective autophagy of ARF7, and this is mediated by the ubiquitin cargo adaptor NBR1 (Ebstrup et al., 2023 preprint). Accumulation of ARF7 due to defective autophagy reduces the priming of LR founder cells and LR initiation (Ebstrup et al., 2023 preprint). These findings demonstrate that autophagy can specifically degrade important pluripotency regulators and, thus, can have a direct impact in modulating reprogramming and cell fate commitment.

The selectivity of autophagy is not restricted to targeted macromolecule degradation. Autophagy also executes directed organelle remodeling, which is essential for maintenance of homeostasis (Dunn et al., 2005; Esteban-Martínez et al., 2017; Hickey et al., 2023; Izumi et al., 2017; Li et al., 2014; Stephani et al., 2020; Sun et al., 2022). Accumulation of damaged or superfluous organelles causes a plethora of negative effects that impact cell health and plasticity, including generation of oxidative damage, decreased metabolic function and inflammation (López-Otín et al., 2013). Consequently, during the reprogramming of mature cells into stem cells, organelle and cell remodeling is necessary, and autophagy plays a central role in those process (Liu et al., 2020). One of the earliest examples of autophagy-mediated organelle degradation during dedifferentiation of somatic cells dates back to 1978, when evidence of autophagic degradation of melanosomes, ribosomes and multivesicular bodies during lens cell formation from epithelial cells in the eye of newts was reported (Yamada et al., 1978). Perhaps the most studied organelle remodeling event during reprogramming initiation is the remodeling of mitochondria, and there are several conflicting studies addressing the potential contribution of mitophagy to reprogramming success. Several reports indicate that mitophagy is essential for the acquisition and maintenance of pluripotency in hematopoietic cells and mouse embryonic fibroblasts (Fan et al., 2019; Ito et al., 2016b; Liu et al., 2016; Ma et al., 2015; Vazquez-Martin et al., 2016; Wang et al., 2013; Xiang et al., 2017). Somatic cells preferentially produce energy through mitochondrial oxidative phosphorylation, which in comparison to glycolysis is more efficient in generating ATP but produces more reactive oxygen species (ROS) (Riley et al., 2020). Somatic cell reprogramming into a pluripotent state is often associated with a switch from oxidative phosphorylation to non-oxidative glycolysis, which is most likely necessary to reduce levels of oxidative stress that are detrimental to self-renewal and pluripotency (Ito et al., 2004). Therefore, it is probable that the role of mitophagy during early reprogramming events is connected to the necessity to remodel the mitochondrial pool to a less mature state and decrease ROS stress (Vazquez-Martin et al., 2012). However, another study using mouse embryonic fibroblasts deficient in Atg5 has provided evidence that autophagy is not necessary for remodeling of mitochondria during reprogramming (Wu et al., 2015), and thus care should be taken when attributing a role for mitophagy in reprogramming.

There are also some reports of autophagy participating in remodeling of the cytoplasm; for instance, work on zebrafish muscle regeneration has shown that autophagy is activated within the first 16 h after wounding, and this activation is necessary for cytoplasmic remodeling leading to efficient cellular reprogramming and tissue recovery (Saera-Vila et al., 2016). Induction of autophagy has also been found to be important for remodeling of the cytoskeleton, namely microtubules, during axon regeneration in injured spinal cord (He et al., 2016). In connection to this, mechano-perception and autophagy have been found to share a signaling feedback loop to modulate cytoplasmic remodeling and achieve cellular plasticity during reprogramming (Totaro et al., 2019). These findings, together with those from Yamada and colleagues regarding activation of autophagy during lens regeneration in newts (Yamada et al., 1978), point to a general role of autophagy in cytoplasmic remodeling to ensure plasticity and tissue repair.

Given that wounding is an important early signal during tissue regeneration in both plants and animals. (Birnbaum and Sánchez Alvarado, 2008; Hoermayer et al., 2020; Ikeuchi et al., 2017; Owlarn et al., 2017; Zhang et al., 2019), it would be interesting to further explore exactly how autophagy is activated upon the perception of wounding and how this contributes to reprogramming and tissue regeneration. In this context, the abovementioned work in zebrafish provides evidence of early activation of autophagy to regenerate damaged muscle tissue (Saera-Vila et al., 2016). In that study, the authors show that autophagy is activated during the first 16 h after tissue injury, well within the critical 18 h time window in which wound-induced reprogramming can occur in this model organism (Saera-Vila et al., 2015). In plants, there are also links between autophagy and wound perception. For example, we have shown that autophagy is activated within 30 min of treating seedlings with wound-related signals such as exogenous ATP or the damage signal peptide PEP1 (encoded by PROPEP1) (Rodriguez et al., 2020). While this rapid autophagic activation has been shown to be linked to temporary reprogramming of somatic cells (Rodriguez et al., 2020), other studies have shown that PEP1 also regulates reprogramming of the vasculature (Dhar et al., 2021), a potential link that merits further investigation. In addition, our previous work has shown that wound-induced regeneration in the moss model P. patens is reduced when autophagy is impaired (Rodriguez et al., 2020). This is consistent with findings that autophagy is necessary for alveolar progenitor cells to replenish lung tissue in response to injury (Li et al., 2020). Again in P. patens, evaluation of autophagic activity upon wound-induced reprogramming indicates a robust activation of autophagy early in the reprogramming process (i.e., 24–48 h after damage; Kanne et al., 2022). This aligns with the abovementioned report showing that wounding is an important trigger for the induction of autophagy during salivary gland stem cell renewal (Orhon et al., 2022). Moreover, these studies reveal a potential evolutionarily conserved role for autophagy during the early stages of reprogramming and stem cell activation, as a significant increase in reprogramming efficiency can be achieved by either transcriptional or pharmacological stimulation of autophagic activity in different organisms (Kanne et al., 2022; Orhon et al., 2022). Interestingly, although autophagy is indeed activated during the early stages of reprogramming, it has been found that this activation is mainly restricted to cells adjacent to wounded areas, and that autophagy ceases in cells that have completed reprogramming (i.e. displaying tip growth; Kanne et al., 2022). Consequently, these data suggest that while autophagy is important for wound-induced iPSC formation in plants, it is both spatially and temporally regulated, and predominantly acts during the early stages of reprogramming, which is consistent with similar observations in mammalian iPSCs (Wang et al., 2013).

Besides directly impacting pluripotency regulators, proteome remodeling and organelle remodeling, there is evidence suggesting that the early burst in autophagic activity during stem cell formation could also be related to the high energetic demands that are necessary to ‘jump-start’ quiescent cells into pluripotency and proliferation (Tang and Rando, 2014). It is known that reprogramming events are energetically demanding and are a sink for cellular resources (reviewed in Tsogtbaatar et al., 2020), and the connection between metabolism and stem cell formation is well established (Ho et al., 2017). Stem cells in quiescent states are characterized by a low metabolic rate, and they need to undergo metabolic reprogramming to resume proliferation (Ryall et al., 2015). Metabolic reprogramming is an essential step towards stem cell activation and was initially reported for cancer cells; it often encompasses a shift from oxidative phosphorylation to aerobic glycolysis (discussed in Deberardinis and Chandel, 2020), which is a hallmark of activation of rapidly dividing cells (Esteban–Martínez et al., 2017). Since resuming cellular proliferation requires energy as well as building blocks such as amino acids, phospholipids and nucleotides, it is not unexpected that autophagy, due to its roles in recycling cellular building blocks and energy homeostasis, also participates in metabolic reprograming to activate stem cells. Accordingly, the catabolic function of autophagy has been linked to the increase in ATP levels that is necessary for stem cell activation (Tang and Rando, 2014). Furthermore, chaperone-mediated autophagy, a type of autophagy in which proteins are delivered directly into lysosomes by chaperones (Kaushik and Cuervo, 2018), has been reported to promote hematopoietic stem cell activation, owing to its impact on lipid metabolism and glycolytic capacity (Dong et al., 2021). In plants, there is recent evidence that physiological reprogramming during LR formation is also highly energy demanding (Stitz et al., 2023). Specifically, the serine-threonine kinase TOR, which is a master regulator of cellular metabolism, has been shown to be essential for resource allocation during XPP cell reprogramming necessary for de novo organogenesis during LR formation (Stitz et al., 2023). This important role of TOR during XPP reprogramming might be in conflict with autophagic activity during the same process, given that TOR activation often leads to downregulation of autophagy – a relationship that is conserved across kingdoms (Díaz-Troya et al., 2008). In this context, TOR kinases are known to directly phosphorylate, and thus regulate, essential autophagic components such as ATG13 proteins and ATG1 proteins (including ULK1 in mammals) (Kamada et al., 2000; Kim et al., 2011), reducing autophagic activity. In plants, like other organisms, TOR activation leads to inhibition of autophagy, and conversely, TOR inhibition leads to activation of autophagy (Pu et al., 2017). These lines of evidence all point to a potential conflict between activation of TOR and autophagy during LR formation. However, studies in mammals have shown that mTOR activity during reprogramming is temporally restricted, as early inhibition of mTOR (and concomitant activation of autophagy) enhances reprogramming efficiency, whereas activation of mTOR and consequent downregulation of autophagy promotes reprogramming efficiency at later stages of that process (Wang et al., 2013). It should also be noted that TOR-mediated regulation of autophagy is dependent on the context; for instance, TOR directly controls autophagy with regard to nutrient availability and osmotic stress, but autophagy functions in a TOR-independent manner during endoplasmic reticulum stress and oxidative stress (Pu et al., 2017). Moreover, activation of TOR and autophagy during XPP reprogramming could be temporarily uncoupled, as recent data has shown that ARF7 autophagic turnover displays an oscillatory rhythm, which indicates periods of high and low autophagic activity during LR formation (Ebstrup et al., 2023 preprint).

As highlighted here, a wealth of evidence strongly indicates that autophagy plays an important role during the dedifferentiation of somatic cells into a pluripotent state. Most of the evidence points to a positive role for autophagy during early reprogramming, where it allows somatic cells to transition into stem cells by facilitating metabolic reprogramming, energy balance, and remodeling of the organelle and proteome pool. Some data suggest that autophagy must be tightly controlled in a temporal manner to prevent deleterious effects to reprogramming efficiency, and thus more studies are needed to gain a better insight into the spatio-temporal resolution of autophagy activity during early reprogramming.

Autophagy has also been shown to be involved at later stages of stem cell development, with most of the data indicating that it is required for the maintenance of stemness in diverse types of tissues and organisms. For instance, autophagy is necessary for maintenance of human hematopoietic stem cells (Mortensen et al., 2011), epithelial stem cells in Hydra (Tomczyk et al., 2019), mouse skeletal muscle stem cells (García-Prat et al., 2016) and embryonic stem cells (Liu et al., 2017), and Drosophila testis stem cells (Sênos Demarco et al., 2020). In this context, the role of autophagy during stem cell maintenance involves clearance of misfolded proteins and the prevention of aging, which otherwise would promote loss of stemness (Chua et al., 2023). Indeed, basal autophagy has been shown to dramatically increase in neural stem cells, as compared to that in fully differentiated cells (Tabor-Godwin et al., 2012), which correlates with the necessity to continuously remove proteins and organelles that might force differentiation of those stem cells.

Importantly, in addition to its role in stemness maintenance, autophagy is also essential for the cell fate changes leading to differentiation. For instance, besides showing that autophagy is necessary for maintenance of stemness, Tabor-Godwin and colleagues have also shown that autophagy is engaged during differentiation of neural stem cells (Tabor-Godwin et al., 2012). Recent findings show that the ubiquitin cargo adaptor SQSTM1 (also known as p62) binds to NANOG (Zhou et al., 2022). Furthermore, disruption of ATG7 leads to NANOG accumulation, which in turns causes enhanced NANOG binding to target genes that are involved in differentiation of neurons, thus repressing or causing abnormal neural differentiation of ATG7-deficient stem cells (Zhou et al., 2022). Autophagy has also been shown to positively regulate differentiation of cardiac stem cells (Zhang et al., 2012a). However, as seen before, autophagy needs to be tightly controlled to prevent premature differentiation: in another study, the same authors found that fibroblast growth factor-mediated signaling inhibits autophagy to prevent differentiation of heart progenitor cells into cardiomyocytes (Zhang et al., 2012b). In contrast, recent findings have shown that a short-term inhibition of autophagy drives differentiation of iPSCs into endoderm, and this is in part due to SOX2 degradation via autophagy (Sharma et al., 2022). All these data suggest that autophagy must indeed be very tightly regulated to allow cell fate transitions to occur correctly, and that mistiming of autophagic activity can lead to either restriction of exit from pluripotency or premature differentiation.

Autophagy also participates in the metabolic reprogramming that stem cells undergo during differentiation. For instance, evidence of autophagy-mediated metabolic reprogramming during stem cell differentiation has been reported for neural stem cells (Calvo-Garrido et al., 2019), trophoblast stem cells (Chakraborty et al., 2020) and planarians (González-Estévez et al., 2007). Interestingly, a programmed autophagic clearance of mitochondria is necessary to induce glycolysis during neuronal stem cell differentiation, and this can be achieved by chemical activation of autophagy (Esteban–Martínez et al., 2017).

In plants, we have shown that iPSC masses of atg mutants are able to undergo differentiation to form shoots if their incubation in callus-inducing medium (CIM) is extended from six to 21 days (Rodriguez et al., 2020). Surprisingly, the shift from CIM to shoot-inducing medium (SIM) causes the atg mutant iPSCs to enter a state of rapid proliferation, which even surpasses that of the wild-type plants (Rodriguez et al., 2020). Proteomic analysis of atg mutants during the transition from being grown in auxin-rich medium to being grown in cytokinin-rich medium shows that protein signatures from the auxin program persist but that some proteins associated with the new cytokinin program over-accumulate when autophagy is defective (Rodriguez et al., 2020). Thus, autophagy appears to be necessary to fine-tune the proteome during these reprogramming events by removing components of the old program and regulating the levels of proteins from the new program (Fig. 2) (Rodriguez et al., 2020). It is therefore possible that when atg mutants are changed from CIM to SIM, the initially smaller population of iPSCs begins to proliferate at a dramatic pace, consistent with the known role of cytokinin in promoting cell division (Schaller et al., 2015). Furthermore, the high levels of cytokinin in SIM might rescue the known hyposensitivity of atg mutants to this hormone (Acheampong et al., 2020), thus allowing stem cells derived from autophagy-deficient explants to proliferate and differentiate into shoots in SIM (Rodriguez et al., 2020). It should be noted that although atg mutants are able to produce stem cells, prolonged culturing leads to tissue oxidation and premature senescence of those tissues (Rodriguez et al., 2020), which is similar to what has been reported in animals (García-Prat et al., 2016).

There is clear evidence that proteostatic mechanisms are essential for the acquisition of pluripotency, and although the role of autophagy in this process was initially neglected, recent efforts have begun to elucidate its functional relevance. However, a number of key issues remain unresolved; for instance, although autophagy has been shown to be necessary for the acquisition of pluripotency (i.e. the early stages of reprogramming), there are only a few reports presenting a direct link between autophagy and the turnover of pluripotency factors such as SOX2 and NANOG in animals (Sharma et al., 2022; Zhou et al., 2022), and ARF7 and ARF19 in plants (Ebstrup et al., 2023 preprint). Thus, it will be exciting in the future to further investigate to what extent autophagy can directly interfere with reprogramming by controlling the selective turnover of other key pluripotency regulators. Conversely, the contribution of autophagy to physiological reprogramming in plants is less well characterized, and there are many exciting open research questions related to the precise timings of the activation and shutdown of autophagy during cell fate changes, the processes that are modulated to enable stemness and differentiation, and how organelle remodeling or metabolic reprogramming affect stem cell formation in plants. Answering these questions will allow us to fill gaps in our understanding of stem cell formation and to master regeneration of tissues and organs in plants, which could be extremely impactful for agricultural and biotechnological applications. Furthermore, given the superior capacity of plant cells to undergo reprogramming and the degree of conservation in reprogramming events between kingdoms, it is likely that study of plant model systems could yield unsuspected knowledge that would allow us to improve the reprogramming of mammalian cells.

We sincerely apologize to those whose work has not been cited owing to space constraints. The figures were created with BioRender.com.