Vimentin’s side gig: regulating cellular proteostasis in mammalian systems

Abstract

Intermediate filaments (IFs) perform a diverse set of well-known functions including providing structural support for the cell and resistance to mechanical stress, yet recent evidence has revealed unexpected roles for IFs as stress response proteins. Previously it was shown that the type III IF protein vimentin forms cage-like structures around centrosome-associated proteins destined for degradation, structures referred to as aggresomes, suggesting a role for vimentin in protein turnover. However, vimentin’s function at the aggresome has remained largely understudied. In a recent report, vimentin was shown to be dispensable for aggresome formation, but played a critical role in protein turnover at the aggresome through localizing proteostasis-related machineries, such as proteasomes, to the aggresome. Here we review evidence for vimentin’s function in proteostasis and highlight the organismal implications of these findings.

Introduction

Maintaining protein homeostasis (proteostasis) is critical for an organism’s ability to properly function and avoid disease (Douglas and Dillin 2010). To ensure proteostasis is maintained, cells have evolved three primary systems which are thought to respond to elevated levels of misfolded, damaged or mutant proteins: chaperones, the ubiquitin-proteasome system, and autophagy. Chaperones are a class of proteins that, amongst many functions, are able to bind to proteins and assist in their refolding or target them for degradation by protein degradation systems (Barral, Broadley et al. 2004). The ubiquitin-proteasome system involves a network of proteins that first identify proteins destined for degradation, label them with polyubiquitin, and subsequently target them to proteasomes for proteolysis (Collins and Goldberg 2017). Finally, autophagy consists of the loading of larger protein inclusions into autophagosomes which then can fuse with acidified lysosomes to denature proteins via an acidic environment and lysosomal proteases (Wong and Cuervo 2010). Collectively, in healthy organisms and in many pathological contexts, these mechanisms are sufficient to maintain proteostasis. However, there are also many situations in which these systems fail to sufficiently turn over protein which ultimately result in disrupted proteostasis and impaired cellular viability (Douglas and Dillin 2010).

When proteostasis is disrupted, cells may employ additional mechanisms to ensure proper cellular function is maintained. One of these mechanisms involves the trafficking of proteins destined for degradation by dynein motor proteins along microtubules to the centrosome, which is surrounded by a cage composed of the IF vimentin. This structure is referred to as the aggresome (Johnston, Ward et al. 1998, Johnston, Illing et al. 2002, Kawaguchi, Kovacs et al. 2003, Iwata, Riley et al. 2005) (Fig. 1). The terms “inclusion bodies” and “aggresomes” have been used somewhat interchangeably to describe protein aggregates that are dispersed throughout the cytoplasm. However, here we define the “aggresome” as a distinct cytoplasmic structure rich in proteins destined for degradation present at the centrosome/nuclear bay that is dependent on microtubules to form and which resides within a cage comprised of IFs (Johnston, Ward et al. 1998). Since the aggresome was first identified in 1998, much has become clear about mechanisms driving aggresome formation and clearance, and the numerous healthy and pathological contexts in which the aggresome is utilized by cells (French, van Leeuwen et al. 2001, Johnston, Illing et al. 2002, Kawaguchi, Kovacs et al. 2003, Olzmann, Li et al. 2008, Xu, Graham et al. 2013, Morrow, Porter et al. 2020). However, the role of vimentin at the aggresome has remained largely unexplored.

Figure 1 –

Schematic of aggresome formation. During conditions where proteostasis becomes disrupted, proteins destined for degradation (red) are trafficked by dynein (purple) and adapter proteins (yellow) along microtubules (white) to the centrosome (orange), accompanied by a collapse of the IF vimentin (green) and a redistribution of proteasomes (dark blue) to the nuclear bay.

Vimentin is a type III IF comprised of a central rod domain with a head and tail domain on either side, and is expressed in numerous cell types throughout the body in many organisms (Perreau, Lilienbaum et al. 1988, Schaffeld, Herrmann et al. 2001, Herrmann and Aebi 2004, Danielsson, Peterson et al. 2018). Vimentin is able to polymerize either by itself or heteropolymerize with other IFs into non-polar unit length filaments which can assemble into full length filaments (Steven, Hainfeld et al. 1983, Herrmann, Haner et al. 1996). Two decades ago, vimentin knockout (KO) mice were created and observed to develop and reproduce with no obvious phenotypes, suggesting that vimentin may be dispensable for the organism’s general viability (Colucci-Guyon, Portier et al. 1994). However since then, numerous reports have emerged suggesting that although vimentin is not important for early-development, vimentin becomes critical during a response to a wide variety of challenges (Table 1) (Terzi, Henrion et al. 1997, Rogel, Soni et al. 2011, Cheng, Shen et al. 2016, Danielsson, Peterson et al. 2018). For example, vimentin is now recognized as a critical player in directional migration during wound healing (Rogel, Soni et al. 2011). Vimentin also has been observed to dynamically upregulate in response to challenges such as heat shock and cadmium chloride treatment (Vilaboa, Garcia-Bermejo et al. 1997). Thus, the precedent of vimentin being a stress response protein supports the notion that vimentin could be important for recovering proteostasis in the cell after a cellular stress which impairs proteostasis. Here we review investigations of vimentin’s function in maintaining proteostasis at the aggresome, and discuss the implications of vimentin’s function at the aggresome in both healthy organisms and in numerous pathologies and diseases.

Table 1 –

Diverse phenotypes in vimentin KO organisms/cells

Vimentin KO Phenotype	Source
Immune response
Reduced replication of murine cytomegalovirus	(Roy, Kapoor et al. 2020)
Reduced activation of the inflammasome	(Xiao, Xie et al. 2018)
Reduced fibroblast proliferation, keratinocyte differentiation, and wound healing	(Cheng, Shen et al. 2016)
Reduced lung inflammation following bleomycin exposure	(dos Santos, Rogel et al. 2015)
Longer tail bleeding time	(Da, Behymer et al. 2014)
Increased acute colitis	(Mor-Vaknin, Legendre et al. 2013)
Impaired microglia activation	(Jiang, Slinn et al. 2012)
Impaired glial scar formation (in combination with GFAP KO)	(Pekny, Johansson et al. 1999)
Structural support
Increased mechanical stress generation	(van Loosdregt, Weissenberger et al. 2018)
Increased actin stress fiber assembly and contractility	(Jiu, Peranen et al. 2017)
Death in response to renal mass reduction	(Terzi, Henrion et al. 1997)
Impaired flow-induced dilation in mesenteric resistance arteries	(Henrion, Terzi et al. 1997)
Reduced cytoplasmic stiffness	(Guo, Ehrlicher et al. 2013)
Reduced lymphocyte rigidity	(Brown, Hallam et al. 2001)
Higher expression of sub-endothelial basement membranes	(Langlois, Belozertseva et al. 2017)
Reduced collagen production	(Challa and Stefanovic 2011)
Compromised endothelial integrity	(Nieminen, Henttinen et al. 2006)
Development
Cerebellar development defects and impaired motor coordination	(Colucci-Guyon, Gimenez et al. 1999)
Disrupted vascular smooth muscle cell differentiation	(van Engeland, Suarez Rodriguez et al. 2019)
Impaired mammary gland development	(Peuhu, Virtakoivu et al. 2017)
Disrupted endothelial differentiation of embryonic stem cells	(Boraas and Ahsan 2016)
Decreased axonal outgrowth and uptake of C3bot	(Adolf, Leondaritis et al. 2016)
Altered arterial remodeling	(Schiffers, Henrion et al. 2000)
Disrupted Notch signaling and impaired angiogenesis	(Antfolk, Sjoqvist et al. 2017)
Intermediate filament network formation
Disrupted GFAP network	(Galou, Colucci-Guyon et al. 1996)
Disrupted nestin polymerization	(Park, Xiang et al. 2010)
Impaired desmin filament formation	(Geerts, Eliasson et al. 2001)
Motility/Migration
Increased motility of caveolin-1 vesicles	(Shi, Fan et al. 2020)
Impaired directional cell migration	(Vakhrusheva, Endzhievskaya et al. 2019)
Miscellaneous
Disrupted subcellular localization of Na-glucose transporters	(Runembert, Couette et al. 2004)
No obvious phenotypes in vimentin KO mice	(Colucci-Guyon, Portier et al. 1994)
Decreased astrocyte activation (in combination with GFAP KO)	(Wilhelmsson, Faiz et al. 2012)
Reduced ability to respond to proteotoxic stress	(Morrow, Porter et al. 2020)
Increased rearrangement of the mitochondrial genome	(Bannikova, Zorov et al. 2005)
Reduced fibroblast proliferation rate and inability to immortalize	(Tolstonog, Shoeman et al. 2001)

Vimentin’s function at the aggresome

The identification of vimentin’s presence at the aggresome more than 2 decades ago suggested a new role for this IF. Despite this observation, while many studies have utilized vimentin as a marker of aggresome formation, few have addressed vimentin’s function in proteostasis at the aggresome. To address this question, we recently investigated vimentin’s function at the aggresome in mouse primary hippocampal neural stem cells (NSCs) (Morrow, Porter et al. 2020). Interestingly, vimentin KO NSCs were still able to form an aggresome, demonstrating that vimentin is not essential for aggresome formation. However, vimentin KO NSCs displayed a decreased capacity to recover from impaired proteostasis both after a transient challenge with the proteasome inhibitor MG132 in vitro and during NSC quiescence exit in vitro and in vivo, a time when NSCs must clear a wave of proteins to activate and enter the cell cycle. Vimentin KO NSCs not only displayed reduced viability after a transient pulse with the proteasome inhibitor MG132, but also an increased accumulation of aggregated proteins. Vimentin KO NSCs compensated for defects in protein clearance by increasing autophagy, however, were still unable to recover to the same extent as WT NSCs. Further, we identified through co-immunoprecipitation that vimentin bound several types of proteostasis-related machineries at the aggresome, suggesting that vimentin is critical for efficient protein turnover by acting as a scaffold for these machineries at the aggresome. Immunostaining and proximity ligation assays confirmed that proteasomes, identified in the co-immunoprecipitation, were enriched at the aggresome in WT NSCs, but not in vimentin KO NSCs. Thus, this suggested that vimentin’s function at the aggresome was, at least in part, to localize proteasomes to the aggresome for efficient protein turnover. However, the extent to which this mechanism drives the phenotypes reported in vimentin KO NSCs is not fully clear as numerous other proteostasis machineries, such as ribosomal proteins and chaperone proteins, were also pulled down by vimentin in NSCs. Vimentin KO in NSCs also resulted in a loss of nestin and Glial fibrillary acidic protein (GFAP) IFs, making it difficult to understand if the phenotypes observed in NSCs were vimentin-specific or due to the loss of these other IF proteins. Interestingly, there was no compensation by any other IF in the absence of vimentin. Together, these findings suggest that vimentin’s repositioning of proteasomes to the aggresome in NSCs following impaired proteostasis is critical to essential cellular functions.

Similar to our hypothesis that vimentin is critical for positioning cellular components in the cell, it has also been proposed that vimentin is a regulator of autophagosome and lysosome distribution in HEK293 cells, and that vimentin potentiates autophagy (Biskou, Casanova et al. 2019). This study utilized treatment with the compound Withaferin A (WFA), which binds vimentin and inhibits filament formation, to perturb the filamentous vimentin network and probe for downstream consequences. WFA treatment induced a redistribution of vimentin protein to the aggresome along with autophagosomes and lysosomes. Further, WFA perturbed autophagy through disruption of autophagosome-lysosome fusion (Biskou, Casanova et al. 2019). However, WFA has been found to be cytotoxic even in vimentin KO cells, suggesting that WFA is not specific to targeting vimentin (Bargagna-Mohan, Hamza et al. 2007). For example, WFA also has been observed to perturb the cell’s microtubule network (Grin, Mahammad et al. 2012). WFA’s prevention of vimentin polymerization also could result in increases in vimentin degradation which could overwhelm protein turnover systems and lead to measurable perturbations in autophagy such as what was reported. This scenario is further supported by the recent observation that vimentin can be degraded by autophagy (Park, Yoon et al. 2020). Thus, it would be interesting to reexamine these findings in a vimentin KO or KD cell line which may be less limited by the non-specific activities of WFA.

Additionally in line with the view that vimentin can function through binding and localizing of cellular components to different regions within the cytoplasm, vimentin has been reported to increase activity of the intracellular calcium channel inositol 1,4,5-trisphosphate (IP3) receptor type 1 (IP3R1) through sequestering the negative IP3R1 regulator, IP3R1-interacting protein released with IP3 (IRBIT), to the aggresome (Bauer, Hudec et al. 2012). However, this role for vimentin was suggested to be detrimental to the cell’s capacity to maintain proteostasis, as increased IP3R1 activity was previously connected with increased aggregation of overexpressed mutant Huntington protein in Neuro-2a cells (Bauer, Hudec et al. 2011). In line with this model, vimentin overexpression in Neuro-2a cells increased levels of overexpressed aggregated Huntingtin protein and mild vimentin knock-down (KD) decreased levels of overexpressed aggregated Huntingtin protein (Bauer, Hudec et al. 2012). Whereas overexpressing proteins such as vimentin could be detrimental to a cell by introducing an overabundance of proteins that interfere with cellular processes or compete with endogenous proteins that require degradation, how vimentin KD results in an increased capacity to maintain proteostasis in Neuro-2a cells overexpressing Huntingtin protein is less clear compared with our model in which vimentin is beneficial for the cell’s capacity to maintain proteostasis (Morrow, Porter et al. 2020). As mutant Huntingtin has an increased propensity to aggregate, rendering it resistant to degradation by the proteasome, vimentin’s function in localizing proteasomes to the aggresome may be less consequential for degrading these proteins (Thibaudeau, Anderson et al. 2018). Further, due to the complexity and variety of dynamically regulated pathways that are able to compensate for impaired nodes of the cell’s proteostasis network, it is also possible that weak KD of vimentin could be acting as a type of hormesis for the cell which then becomes exacerbated in a stronger KD or full vimentin KO cells. Finally, as the authors utilized both Neuro-2a cells and HeLa cells, both of which are cancer cell lines, it may be interesting to see if similar effects are seen in primary cells, as cancer cells may have modified their methods of responding to disruptions in proteostasis (see below a further discussion on cancer in part II).

Vimentin and the aggresome also can be asymmetrically inherited during mitosis (Rujano, Bosveld et al. 2006, Ogrodnik, Salmonowicz et al. 2014, Moore, Pilz et al. 2015, Morrow, Porter et al. 2020). Thus, any functions that vimentin may serve in interphase, such as what is described above, can be asymmetrically distributed between two daughter cells after mitosis and could lead to distinct cell behavioral outcomes. Indeed, the daughter cell inheriting more vimentin and the aggresome is associated with a relatively longer cell-cycle length and an increased chance of undergoing apoptosis (Ogrodnik, Salmonowicz et al. 2014, Moore, Pilz et al. 2015). However, it remains unclear if inheriting more vimentin and the aggresome would always be bad for the cell. For example, daughter cells inheriting more vimentin-associated proteins, such as proteasomes, could be better equipped to handle specific circumstances compared to daughter cells with less inherited proteasomes (Morrow, Porter et al. 2020). It has not yet been determined whether daughter cells which inherit more or less vimentin have different capacities to maintain proteostasis after mitosis.

Together, while the field has now answered several key questions about vimentin’s function in proteostasis, this topic remains largely understudied. Additional research is needed to reconcile unresolved discrepancies, as numerous elements could factor into conflicting findings, such as: cell type, KO of vimentin as opposed to KD of vimentin, and different techniques to measure or enhance protein accumulation (endogenous labeling of aggregated proteins as opposed to mutant Huntingtin overexpression). What these studies do agree upon is that vimentin interacts with a diverse set of proteins and that, whether for the better or the worse, perturbing vimentin potentiates the cell’s capacity to maintain proteostasis.

Organismal implications for vimentin’s function at the aggresome

Although vimentin KO mice retain the ability to develop and reproduce normally, vimentin plays numerous critical roles throughout the body, and can be upregulated in pathological contexts (Danielsson, Peterson et al. 2018). As mounting evidence supports a role for vimentin in the cell’s proteostasis network, vimentin’s action at the aggresome could be a putative mechanism underlying previously established vimentin KO phenotypes.

Many reports have demonstrated general reduced cellular function in vimentin KO cells (Galou, Gao et al. 1997, Terzi, Henrion et al. 1997, Vilaboa, Garcia-Bermejo et al. 1997, Lundkvist, Reichenbach et al. 2004, Perez-Sala, Oeste et al. 2015, Boraas and Ahsan 2016, Cheng, Shen et al. 2016). For example, vimentin KO fibroblasts display a slower proliferation rate (Cheng, Shen et al. 2016), and vimentin KO embryonic stem cells have a reduced capacity to differentiate down a endothelial lineage (Boraas and Ahsan 2016). These findings suggest that vimentin’s functions within the cell are important for maintaining basic cellular functions which could translate into the numerous phenotypes reported in vimentin KO organisms (Table 1) (Danielsson, Peterson et al. 2018). Although vimentin’s previously established mechanisms of action could contribute to these phenotypes, such as vimentin’s function in cell motility and adhesion, it is also possible that these phenotypes are the result of an impaired capacity to maintain proteostasis through mechanisms such as what is described above (Danielsson, Peterson et al. 2018). Closer examination for the presence of aggresome formation in these cells would shed light on whether vimentin’s function at the aggresome is playing a role in these phenotypes.

Vimentin’s function in proteostasis may also be important for an organism’s defense against pathologies and diseases throughout the body. For example, vimentin and the aggresome are studied in the lung where they respond to insults such as exposure to cigarette smoke or lung pathologies such as Chronic Obstructive Pulmonary Disease (COPD), which can display elevated ubiquitinated protein levels and impaired proteostasis (Min, Bodas et al. 2011, Tran, Ji et al. 2015, Shivalingappa, Hole et al. 2016). Further, many diseases and stimuli can induce disrupted proteostasis in the liver, some of which culminate in aggresome formation (French, van Leeuwen et al. 2001, Bardag-Gorce, Riley et al. 2004, French, Mendoza et al. 2016, French, Masouminia et al. 2017). Thus, there are plenty of pathological contexts which may warrant vimentin’s action at the aggresome. However, not all cell types throughout the body express vimentin. Therefore, it would be important to verify that the cell type being studied either already expressed vimentin or upregulated vimentin as a complement to aggresome formation to understand if vimentin is playing a role in these circumstances. In the event that vimentin is not expressed, other IFs could play an analogous role to vimentin. Indeed, in mouse motor neuron-neuroblastoma cells the neurofilament network was observed to collapse around aggresomes formed by overexpression of a truncated androgen receptor with a 112-glutamine repeat in vitro (Taylor, Tanaka et al. 2003). It remains unclear if this observation would translate to bonafide neurons in vivo. Further, as many intermediate filaments heteropolymerize, it is likely that many intermediate filament proteins (such as nestin and GFAP) could be similarly enriched in the cage surrounding the aggresome and that these proteins could be performing functions analogous to vimentin (Steven, Hainfeld et al. 1983, Morrow, Porter et al. 2020).

Further supporting the notion that vimentin plays a role in pathologies across the body, vimentin is also upregulated in tissues during several pathologies where proteostasis is disrupted. For example, neurons in a mouse model of Alzheimer’s Disease (Tg2576), which normally don’t express vimentin once mature, upregulate vimentin during disease progression (Levin, Acharya et al. 2009). Further, vimentin also has been found in human Alzheimer’s Disease amyloid plaques (Liao, Cheng et al. 2004, Rudrabhatla, Jaffe et al. 2011). Finally, vimentin expression is increased generally during aging in tissues, such as the brain (Xu, Gao et al. 2016, Benayoun, Pollina et al. 2019). A recent quantitative proteomic analysis during aging in the hippocampus, an area critical for learning and memory, found vimentin as one of 35 upregulated genes out of 4582 that were analyzed (Xu, Gao et al. 2016). One explanation may be that vimentin is upregulated in reactive astrocytes; however, it could also be that vimentin becomes activated to respond to the disruptions in proteostasis that are sustained in the brain during aging (Wang, Bekar et al. 2004).

While there are numerous scenarios in which vimentin’s activity at the aggresome could be interpreted as beneficial, there are also scenarios in which the organism sustains greater consequences from an optimally functioning vimentin-caged aggresome such as in cancers and viral infection. Aggresomes and vimentin are widely studied chemotherapeutic targets in combination with proteasome inhibitors in cancers such as multiple myeloma, breast cancer and pancreatic cancer (Nawrocki, Carew et al. 2006, Komatsu, Moriya et al. 2013, Mishima, Santo et al. 2015, Park, Yoon et al. 2020). Interestingly, indirectly inhibiting aggresome function synergistically increases cytotoxicity associated with proteasome inhibition in a pancreatic cancer xenograft (Nawrocki, Carew et al. 2006). Other studies have reported that vimentin KO tumors have an impaired ability to metastasize, which may be mediated by a reduced capacity to migrate (Liu, Lin et al. 2015). However, an alternative explanation may be that vimentin KO tumors have a decreased capacity to maintain proteostasis. Together, these findings combined with vimentin’s role in proteostasis suggest that vimentin could be viewed as a target for not only reducing tumor metastasis, but also the tumor’s ability to maintain proteostasis.

Aggresomes have also been implicated in the cell’s response to viral infection. While the precise mechanism(s) by which viruses benefit from hijacking the aggresome are not fully clear, cells with impaired aggresome formation display reduced viral propagation, suggesting that an optimally functioning aggresome would be beneficial for viruses and detrimental for the organism (Nozawa, Yamauchi et al. 2004, Liu, Shevchenko et al. 2005). Thus, vimentin also could be viewed potentially as a target to reduce the virus’ capacity to propagate. These examples suggest that an optimally functioning aggresome may provide cellular benefits that may not always be beneficial to the organism and that it may not always be desirable to utilize vimentin to increase the cell’s capacity to recover proteostasis.

There are many examples of vimentin and the aggresome dynamically responding to numerous different healthy and pathological stimuli. Recent data investigating vimentin’s role in maintaining proteostasis at the aggresome provides a new consideration for vimentin in a myriad of diverse cellular functions and phenotypes. Future research will be needed to fully understand how this mechanism functions with other established roles of vimentin in these diverse conditions.

Future Directions

Evidence suggests that vimentin plays numerous unique roles in the cell, many of which fall under the umbrella of the cell’s response to stress. While many components of vimentin’s function in these processes are known, the field is still limited by the absence of critical and effective tools. For example, currently there is no mouse line that allows for conditional knock-out of vimentin, but only a full vimentin KO mouse that has had vimentin removed throughout development in all cell types. Therefore, all studies to date in vimentin KO mice are limited by any compensatory mechanisms that this line has generated to adapt to vimentin KO, and in the inability to determine cell-specific effects. Further, many studies investigating vimentin’s presence and function at the aggresome have relied on visualizing vimentin through overexpression of vimentin fused to a fluorophore, which could also be inducing artificial phenotypes that are less reminiscent of endogenous vimentin. Despite technical limitations, it is clear that vimentin expression and distribution in the cell are dynamic and functionally important in many cell types in numerous contexts. Improved tools can be used to resolve discrepancies in the understanding of vimentin’s role in proteostasis and further reveal how in vitro studies in cell lines translate into organismal changes in vivo.

As vimentin and IFs are expressed in a tissue-specific manner, it will be interesting to determine how cell types that don’t express vimentin maintain proteostasis through compensation by other nodes of the proteostasis network. For example, it is well-established that impairments in the ubiquitin-proteasome system can lead to increases in autophagy and vice versa (Dikic 2017). Further, revealing how species such as Drosophila melanogaster were able to evolve in the absence of IFs may reveal mechanisms of compensation by other mechanisms (Bohnekamp, Cryderman et al. 2016). Both of these phenomena may be explained by the fact that many roles of vimentin involve increased resilience to cellular challenges that would only become important in specific settings. These roles are also often redundant and provide increased efficiency in a cellular process, rather than an essential function that the cell could not complete otherwise. Additionally, as expression of vimentin has been shown to be harmful for organisms in specific scenarios such as cancer or viral infection, it may be that organisms have carefully evolved to only use vimentin in places where they are needed most to minimize their potential to cause harm to the organism. Vimentin’s cell-type specific expression raises the prospect that not all cells are using all of the tools nature has provided them with to maintain proteostasis. Identifying the molecular determinants driving vimentin’s function in maintaining proteostasis could lead to development of targeted therapies that can modulate cellular proteostasis and ultimately improve diseases resulting from impaired proteostasis.