Regulation of microtubule growth rates and their impact on chromosomal instability

ABSTRACT

Microtubules are polymers of α/β tubulin dimers that build the mitotic spindle, which segregates duplicated chromosomes during cell division. Microtubule function is governed by dynamic instability, whereby cycles of growth and shrinkage contribute to the forces necessary for chromosome movement. Regulation of microtubule growth velocity requires cell cycle-dependent changes in expression, localization and activity of microtubule-associated proteins (MAPs) as well as tubulin post-translational modifications that modulate microtubule dynamics. It has become clear that optimal microtubule growth velocities are required for proper chromosome segregation and ploidy maintenance. Suboptimal microtubule growth rates can result from altered activity of MAPs and could lead to aneuploidy, possibly by disrupting the establishment of microtubule bundles at kinetochores and altering the mechanical forces required for sister chromatid segregation. Future work using high-resolution, low-phototoxicity microscopy and novel fluorescent markers will be invaluable in obtaining deeper mechanistic insights into how microtubule processes contribute to chromosome segregation.

Introduction

Microtubules are hollow polymers of α/β tubulin heterodimers. Tubulin dimers are arranged in a head-to-tail orientation, creating a polarized microtubule lattice with α-tubulin exposed at the minus end and β-tubulin exposed at the plus end [1]. Microtubules are essential components of the eukaryotic cytoskeleton and orchestrate a variety of cellular processes, including meiotic and mitotic cell divisions, intracellular trafficking and formation of cilia and axonemes [2].

Mitosis occurs when replicated chromosomes are segregated into new daughter cells, a process that relies on a large array of microtubules known as the mitotic spindle. Mitotic spindle assembly initiates with microtubule nucleation primarily from centrosomes (in most cells) and results in the formation of a bipolar structure that captures the chromosomes through the establishment of stable microtubule attachments to kinetochores [3,4]. Throughout mitosis, oscillatory chromosome movements progressively align sister chromatids at the metaphase plate [3,4]. Once all chromosomes are successfully aligned, sister chromatids are separated and pulled apart toward opposite cell poles during anaphase, and cell division is completed by cytokinesis, when the cytoplasm of new daughter cells is physically partitioned [3,5].

Although the term “cytoskeleton” alludes to a static structure, microtubules are instead highly dynamic, with turnover rates reported to average between 200 s to 345 s during interphase [6,7] and 11 s to 54 s during mitosis [6,7]. Microtubule turnover is caused by continuous cycles of growth and shrinkage – a property known as dynamic instability [8]. Microtubule growth occurs by the addition of GTP-bound tubulin dimers, followed by GTP hydrolysis at the β-tubulin site once dimers are incorporated into the lattice [9]. Although both ends of the microtubule can switch from cycles of growth and shrinkage, microtubule minus ends display slower growth dynamics than plus ends and are often stably anchored to sites of microtubule nucleation [10]. Actively growing microtubules form a stable “GTP-cap” that recruits several microtubule-associated proteins (MAPs) and motor proteins [9,11]. The concerted action of MAPs and motor proteins further determines whether microtubules continue to grow by the addition of new GTP-tubulin dimers or whether the GTP-cap is lost when GTP hydrolysis outpaces the addition of GTP-tubulin, causing the microtubule to release GDP-bound tubulin and shrink in a process known as “catastrophe” [9]. A rapidly shrinking microtubule can often resume a new cycle of growth, in a phenomenon known as “rescue”. Microtubule growth and shrinkage allow for the search and capture of kinetochores during spindle formation [4] and generate mechanical forces [12] that are essential for chromosome segregation during cell division [13].

Microtubules in the mitotic spindle are organized as an antiparallel array consisting of three observable microtubule subpopulations: kinetochore-microtubules, non-kinetochore (or pole-to-pole microtubules) and astral microtubules (or cytoplasmic microtubules in budding yeast), all of which have crucial roles on spindle morphology and function [14]. Kinetochore-microtubules bind to chromosomes by attaching to kinetochores, thereby forming a stable connection known as k-fiber that mediates chromosome movement and alignment as well as sister chromatid separation during anaphase [14]. Non-kinetochore microtubules include all microtubules that populate the region between both spindle poles other than kinetochore microtubules, and this subgroup undergoes extensive sliding that helps provide stability to the spindle by ensuring spindle pole separation. Astral microtubules radiate from the spindle poles toward the cell cortex, thereby anchoring the spindle to the cortex [14].

In addition to displaying fundamentally different functions, microtubule subpopulations differ greatly in their dynamic properties. An estimated 10–30 kinetochore-microtubules bind to individual kinetochores in mammalian cells [15,16] and can nucleate either from spindle poles [17] or directly at kinetochores [18] (reviewed in Prosser and Pelletier [19]). Throughout mitosis, kinetochore-microtubules maintain a steady balance between microtubule polymerization and depolymerization at the kinetochore, poleward flux [20,21] as well as depolymerization at the spindle poles, resulting in relatively slow turnover rates [7]. Non-kinetochore microtubules are thought to be nucleated throughout the spindle rather than exclusively at the poles [22]. They are characterized by very high turnover rates [6,7] as well as poleward flux properties [23]. Microtubule poleward flux is an evolutionarily conserved mechanism characterized by the continuous movement of microtubules toward the spindle poles. This process is generally coupled with adding new tubulin subunits at the microtubule plus ends and their removal at the minus ends at the spindle poles [20]. Astral microtubules similarly maintain high turnover rates [24], although this subset of microtubules is nucleated exclusively at the spindle poles and appears to display no poleward flux [25].

Failure to properly build the spindle causes cells to develop persistently high rates of chromosome segregation errors through several cell divisions – a phenomenon known as chromosomal instability (CIN) that causes aneuploidy [26,27] and can be a driver of tumorigenesis [28] and developmental disorders [29,30]. Importantly, seminal work has shown that disrupted microtubule dynamics, including altered microtubule turnover [26] and growth rates [27,31], contribute to the development of CIN. In this review, we revisit the mechanisms governing the regulation of microtubule growth velocities in different cell types and throughout the cell cycle. We also discuss the role of microtubule growth rates on the maintenance of chromosomal stability and highlight recent technical advances that may overcome methodological bottlenecks that currently limit our ability to measure all features of microtubule dynamics with spatiotemporal precision.

Microtubule growth rates throughout the cell cycle and across cell types

Microtubule dynamics change substantially throughout the cell cycle [32], and early work in cultured mammalian cell lines Ptk1 and BSC demonstrated that microtubule turnover rates increase dramatically during mitosis compared to interphase [6]. Later, at the metaphase-anaphase transition, microtubule dynamics once again change, and turnover rates have been shown to decline up to 5-fold during anaphase compared to metaphase [7,33]. These precise changes in turnover rates indicate that fine-tuned regulation of microtubule dynamics during cell cycle transitions is important for spindle assembly and function. Subsequent work in Xenopus egg extracts [34] and mammalian LLCPK-1 cells [24] suggested that elevated turnover rates in mitosis may be attributable to increased catastrophe rates [24,34] and decreased rescue rates [24]. More specifically, it was found that catastrophe rates increased 2-fold, and rescue rates decreased nearly 4-fold during mitosis [24].

The available data on cell cycle-dependent regulation of microtubule growth rates are much less consistent. Interphase microtubule growth rates were found to be as low as 4.5 μm/min in Ptk2 cells [35,36] and as high as 26.5 μm/min in the human fibrosarcoma cell line HT1080 [31] (Figure 1). During mitosis, growth rates as low as 9.6 μm/min in mouse embryos [52] and as high as 30 μm/min in human adenocarcinoma HeLa cells [47] have been reported (Figure 1). In non-mammalian systems, growth rates were shown to vary between 1.7 to 2.2 μm/min in budding yeast [41] and between 10.2 to 44 μm/min in C. elegans embryos [42,43,59] (Figure 1). Moreover, although some studies report an increase in microtubule growth rates during mitosis [34,51], other reports have instead shown a decline in growth velocities during mitosis [31,41,48] (Figure 1).

Figure 1.

Microtubule growth rates across cell types and organisms.Heatmap of microtubule growth rates in different cell lines/organisms and cell cycle stages. Letters in boxes correspond to growth measurements performed in distinct microtubule subpopulations within the mitotic spindle, where (A) represents astral microtubules and (S) represents central spindle microtubules. Values in boxes represent reported microtubule growth rates in correspondent studies for the following cell lines/organisms: A498 [37]; A549 [38,39]; BS-C-1 [40]; S. cerevisiae [41]; C. elegans embryo [42,43]; Caco-2 [27]; Caov-3 [37]; CHO [44]; ESC and Dif (differentiated)-ESC (Kletter et al., BioRxiv 2024 [45]); H460 [46]; HCT116 [27]; HeLa [47,48]; HMEC-1 [39]; HPAEC [49]; HT1080 [31,38,50]; HT29 [27]; HUVEC [39]; LLCPK [24,44,51]; LS1034 [46]; MCF7 [27]; mouse embryo [52]; Newt lung epithelial cells [17,39,53]; NIH-3T3 [38]; Ptk [35,54,55]; RKO [27]; RPE-1 [27,31,56]; SW48 [27]; SW480 [27]; SW620 [27]; SW837 [27]; U118 [38]; U20S [57]; S. pombe [58].

A possible reason for these variations could be the specific characteristics of different cell lines, which result in distinct microtubule growth dynamics. Consistent with this idea, previous work has shown that chromosomally unstable cancer cell lines display faster microtubule growth rates than chromosomally stable cell lines [27], indicating that microtubule dynamics may be deregulated in cancer cells. This supports the idea that cell line-specific factors may influence the regulation of microtubule growth dynamics. We will further discuss in detail the link between microtubule growth rates and CIN in later sessions of this review.

Differences in cell geometries and local microtubule densities may also account for variations in growth dynamics. Elegant work using cell-free extracts and hydrogel micro enclosures to simulate defined cytoplasmic volumes has shown that smaller cytoplasmic volumes result in higher microtubule densities [60]. Moreover, increased microtubule densities cause a decline in microtubule growth rates, perhaps suggesting that cell size and shape may directly regulate microtubule growth dynamics [60]. Interestingly, local variations in microtubule growth rates have been observed in Ptk1 migrating cells, where microtubule growth rates were found to be between 1.4 and 2.2 μm/min faster at the migrating leading edge as compared to the center of the cell [54,55]. Similarly, microtubule growth rates were found to be much faster in astral microtubules (44 μm/min) [43] as compared to non-astral microtubules (between 12 to 16 μm/min) [42,43,59] of C. elegans mitotic spindles, suggesting that diverse microtubule subpopulations may undergo differential regulation of growth dynamics (Figure 1). Whether local variations in microtubule densities may account for these differences in growth rates throughout the cytoplasm and among different microtubule subpopulations is currently unknown and warrants future research.

Although inter-cell line specificities may account for some of the variation observed in reported microtubule growth rates, it is worth noting that variations among different studies have also been observed. For instance, microtubule growth rates were shown to diverge among different studies from 13 to 30 μm/min in mitotic HeLa cells [47,48], 14 to 26.9 μm/min in interphase HT1080 cells [31,38,50] and 8 to 19 μm/min in interphase LLCPK cells [44,51] (Figure 1). It is possible that these discrepancies may be a result of different assays or analysis methods applied for the assessment of microtubule growth rates. Both the choice of the fluorescently labeled construct used for microtubule growth measurements (e.g. EB proteins vs tubulin) and the type of analysis method used (manual measurements with kymographs vs automated tip tracking) could, in principle, influence measured growth rates. Alternatively, external factors such as specific cell culture media additives/supplements or temperature fluctuations could potentially influence microtubule growth rates and might explain the observed variations among different reports.

MAPs and microtubule growth rates

How microtubule dynamics are modulated throughout the cell cycle is not entirely understood. However, growing evidence suggests that the cell cycle-dependent activity of MAPs contributes to this process. The first MAPs were identified by purifying tubulin from brain extracts using cycles of assembly and disassembly, which resulted in the isolation of higher molecular weight components of the microtubule cytoskeleton named tau [61], MAP1 and MAP2 [62]. Further methodological developments allowed for the purification of tubulin from HeLa cells [63] and led to the discovery of two non-neuronal MAPs, later described as MAP4 [64] and MAP7 [65]. Since then, a variety of other MAPs has been detected in fractions of polymerized tubulin (reviewed in Bodakuntla et al. [66]), and advances in molecular cloning have allowed for detailed investigation of their identities and functions on microtubule dynamics both in vitro and in vivo.

Since their discovery in the 1970s, multiple MAPs have been studied in different cellular contexts (Supplementary Table S1) and are involved in regulating microtubule growth rates. However, the extent of their impact on growth velocity varies greatly depending on the MAP and the cell cycle stage (Supplementary Table S1). Notably, misregulation of ACF7, CLASP2, Katanin, MARK1/3, Stathmin [57] and the checkpoint kinase CHK2 [27] have all been shown to increase microtubule growth rates >3 μm/min above wild-type levels. Reversely, reductions of >3 μm/min have been reported when APC2 [57], AURKB [56], chTOG/XMAP215 [27,48,57], CKAP2 [31], CLIP170 [57], Doublecortin [57] and heat-stable brain MAPs [40] were misregulated in cells.

The regulatory role of certain MAPs on microtubule growth rates is cell cycle-dependent. For instance, knockdown of Adenomatous polyposis coli (APC) was reported to decrease microtubule growth rates during mitosis with minimal effects during interphase [57] (Supplementary Table S1). Conversely, overexpression of the chromosomal passenger complex (CPC) component Survivin increased microtubule growth rates during interphase with no observable changes in mitosis [56] (Supplementary Table S1). These differences in MAP activity throughout the cell cycle could be associated with cell cycle-dependent changes in subcellular localization. For instance, the CPC component Survivin is centrosomal during interphase and re-located to kinetochores during mitosis [67]. These changes in subcellular localization may potentially explain why Survivin regulates microtubule growth dynamics strictly during interphase [56].

Alternatively, changes in MAP activity could also be caused by cell cycle-dependent regulation of gene expression or post-translational modifications, particularly phosphorylation driven by mitotic kinases [68]. Consistent with this idea, a large-scale proteomic and phosphoproteomic analysis of human RPE-1 cells revealed cell-cycle-dependent variations in abundance and phosphorylation of multiple MAPs (Rega et al., BioRxiv 2024 [69]). Cell-cycle-dependent phosphorylation of MAPs such as EB2 [70], XMAP215 [71,72], TPX2 [73] and MCAK [74,75] was shown to be required for the regulation of microtubule dynamics and spindle assembly during mitosis. The spindle protein CKAP2 is expressed only at the late stages of the cell cycle – starting at the S- to G2 transition and throughout mitosis [31,69]. Moreover, a cascade of phosphorylation events promotes a shift in CKAP2 localization from spindle microtubules to chromatin during late anaphase before its degradation in telophase/G1 [31,76–78]. Taken together, these effects could explain the primarily mitotic role of CKAP2 in microtubule growth regulation, with less pronounced effects detected in interphase [31] (Supplementary Table S1).

Mechanisms of microtubule growth and nucleation regulation by MAPs

With the advent of single-molecule in vitro reconstitution [79], cryo-electron tomography [80] and computational modeling approaches [81], important insights have been obtained into the biochemical mechanisms by which MAPs contribute to microtubule growth regulation. One possible mode by which MAPs could promote microtubule polymerization is by increasing the effective concentration of tubulin at the microtubule end. This has been proposed to be the case for XMAP215/chTOG [82,83], in which a two-step process of electrostatically steered diffusion-to-capture followed by electrostatic stabilization of GTP-tubulin dimers could facilitate tubulin incorporation into the microtubule end (Venables et al., BioRxiv 2018 [83] (Figure 2(a)). Tethered TOG domains were shown to bind to a curved conformation of tubulin dimer at the microtubule plus end and to another curved tubulin dimer located proximally to the microtubule end, thereby facilitating tubulin binding to the plus end [84] (Figure 2(a)).

Figure 2.

Mechanisms of microtubule growth regulation by MAPs. (a) Microtubule polymerization can be promoted by increasing the effective concentration of tubulin at microtubule ends. Tethered TOG domains of the MAP XMAP215 bind to a curved conformation of tubulin at the microtubule end and to another dimer near the end, thereby facilitating tubulin incorporation at the microtubule end [83,84]. MAPs such as EB1 [85–87] were shown to form biomolecular condensates with tubulin that increase the effective concentration of tubulin at the microtubule end and facilitate dimer incorporation and growth. (b) Tubulin concentrations can be determined by the regulation of tubulin mRNA degradation. Soluble tubulin reversibly binds to and sequesters the ribosome-binding factor TTC5, thereby preventing tubulin mRNA degradation [88]. Dissociation of tubulin from TTC5 induces tubulin mRNA deadenylation and decay [88]. (c) MAPs can regulate microtubule growth dynamics by modifying conformational configurations of tubulin dimers or the microtubule lattice. Assembly-competent tubulin dimers display a 12° inter-dimer curvature that must be straightened for lattice incorporation. The microtubule depolymerase MCAK stabilizes the curved conformation of tubulin, which induces its detachment from the lattice and results in microtubule depolymerization [89,90]. Actively growing microtubule ends are characterized by long tapers, and microtubules grown in the presence of the intraluminal protein CSPP1 were found to have shorter tapered ends, suggesting that CSPP1 possibly regulates microtubule growth by modulating tapered end conformations [91]. (d) Microtubules grown in the presence of high levels of GDP-tubulin display large growth fluctuations [92]. This has been attributed to GDP-tubulin acting as a “microtubule poison”, thereby hampering growth. MAPs could, in principle, regulate microtubule growth dynamics by modulating GTP-hydrolysis rates, although a direct effect of MAPs on GTP hydrolysis has yet to be identified.

An alternative mechanism by which local tubulin concentrations may be increased is through the formation of biomolecular condensates (Figure 2(a)). The centrosome has been suggested to form through biomolecular condensation [93] that displays a 10-fold enrichment in tubulin concentrations compared to the cytoplasm, with peak values of 470 μM at centrosomes as opposed to 47 μM throughout the cytoplasm [94]. In addition, several recent reports have shown that MAPs such as TPX2 [95], EB1 [85–87], EB3-CLIP-170 [96], MCAK and TIP150 [97] may form condensates with tubulin and, by doing so, may regulate microtubule nucleation, growth and cell division. Rather than being stably incorporated within the biomolecular condensate, the current model proposes that tubulin would act as a “client” protein, with high mobility and higher energetic stability at the droplet interface instead [93,96]. This would thus increase the availability of tubulin for biochemical reactions at high concentrations, for instance, at +TIP condensates [96] and at microtubule organizing centers [93].

Paradoxically, the enrichment in tubulin concentrations at the centrosome does not necessarily result in higher microtubule nucleation capacity, as evidenced by relatively low microtubule nucleation rates at centrosomes (averaged at 90–113 microtubules per minute during mitosis [31,98]) as compared to those detected in vitro with tubulin concentrations equivalent to those found in the cytoplasm (~250 microtubules/min at 50 μM tubulin) [99]. It is, therefore, possible that the centrosome does not simply promote microtubule nucleation and may instead act as a “microtubule nucleation regulating center.” In support of this notion, landmark RNAi screens in C. elegans embryos showed that centrosomes do not nucleate microtubules to saturation, and instead, microtubule outgrowth is counteracted by the depolymerizing kinesin-13 member KLP-7 [100].

Alternatively, tubulin levels could be modulated globally by factors that induce or prevent tubulin mRNA degradation. Recent work has shown that sequestration of ribosome-binding factor TTC5 by soluble tubulin prevents it from inducing tubulin mRNA degradation. Conversely, loss of TTC5 sequestration by tubulin causes constitutive activation of TTC5, resulting in lower tubulin levels and, subsequently, chromosome segregation defects [88] (Figure 2(b)). It is possible that factors that facilitate or prevent sequestration of TTC5 by tubulin could ultimately determine tubulin concentrations, although such factors are yet to be identified.

Another mechanism by which MAPs can regulate microtubule growth dynamics is by promoting or stabilizing conformational configurations of tubulin dimers or microtubule ends that either favor or disfavor polymerization. Assembly-competent tubulin dimers possess an intra-dimer kink that must be straightened upon incorporation into the lattice (reviewed in Brouhard and Rice, 2014 [101]). Actively growing microtubule ends are characterized by tapered ends (Figure 2). The taper results from both the decreasing curvature of tubulin dimers as they become part of the lattice as well as staggered protofilament ends. The tapered end structure may help stabilize the growing end by allowing the binding of additional tubulin dimers and other associated proteins, contributing to microtubule elongation and stabilization. Microtubule depolymerases such as the kinesin-13 MCAK can stabilize the curved conformation of tubulin, thereby inducing its detachment from the lattice and promoting depolymerization [89,90,102] (Figure 2(c)). Moreover, microtubules grown in the presence of the intraluminal protein CSPP1 were found to have shorter tapered ends [91] (Figure 2(c)). The absence of tapered ends is consistent with CSPP1’s inhibitory effect on microtubule growth [91]. CSPP1 binding to either the minus or the plus end stabilizes the microtubule end and even protects from depolymerization by MCAK [103]. This may support a general mechanism whereby MAPs modulate microtubule tapered end conformations either directly or indirectly to regulate growth dynamics.

Finally, MAPs could regulate microtubule growth dynamics by directly modulating GTP hydrolysis (Figure 2(d)). Previous work using interference reflection microscopy and computational modeling has shown that, as opposed to steady microtubule growth observed in the absence of GTP hydrolysis, microtubules grown in the presence of GTP hydrolysis displayed larger growth fluctuations [92]. These fluctuations were attributed to the presence of GDP-tubulin that could act as a “microtubule end poison,” thereby hindering growth [92] (Figure 2(d)). Consistent with this notion, microtubules grown in the presence of tubulin mutants with reduced GTP hydrolysis rates display an extended GTP cap, higher nucleation capacity, and reduced catastrophe rates, suggesting that the GTP state favors microtubule growth [104,105]. Although a direct effect of MAPs on GTP hydrolysis has yet to be identified, these results support a conceptual mechanism whereby MAPs could suppress or induce GTP hydrolysis to regulate microtubule growth rates.

Tubulin post-translational modifications and the regulation of microtubule dynamics

In addition to direct impacts of MAPs on the regulation of microtubule dynamics, post-translational modifications (PTMs) of tubulin also modulate microtubule dynamics either directly or indirectly. Most tubulin PTMs include the covalent addition of amino acids such as glutamate, glycine and tyrosine [106]. Glutamates and glycines are added to the intrinsically disordered C-terminal tail of either α- or β-tubulin, whereas tyrosines are added to the carboxy-end of α-tubulin after initially being removed during the detyrosination-tyrosination cycle [106]. Additional modifications can be included in residues in the tubulin body, such as acetylation, methylation, phosphorylation, palmitoylation, SUMOylation and polyamination [106]. The following paragraphs will detail the contributions of the most commonly studied tubulin PTMs – namely, polyglutamylation, tyrosination/detyrosination and acetylation – to the regulation of microtubule dynamics. Comprehensive reviews on the tubulin code can be found in Roll-Mecak, 2020 [106] and Janke & Magiera 2020 [107].

Polyglutamylation is abundant on stable microtubule networks, including in axonemes and axons, and is thought to be a negative regulator of microtubule growth [108]. In vitro-grown microtubules assembled from tubulin with long glutamate chains were found to have reduced growth rates and increased catastrophe rates, as compared to those assembled with monoglutamylated or deglutamylated tubulin [108]. Polyglutamylation has also been shown to impact the activity of MAPs. The velocity and processivity of kinesins 1 and 2 were found to be increased as a function of polyglutamylation [109]. Moreover, glutamylation was found to tune the activity of microtubule severing protein spastin in a nonlinear biphasic manner [110]. Glutamate chains of up to ~ 8 glutamates enhance spastin’s severing activity, whereas longer glutamate chains display an inhibitory effect [110]. Reversely, MAPs can also modulate tubulin polyglutamylation levels. Recent work has shown that the neuronal MAP DCX acts synergistically with the tubulin polyglutamylase TTLL11 to regulate α-tubulin polyglutamylation levels in neurons [111].

The α-tubulin detyrosination-tyrosination cycle is also an important regulator of microtubule dynamics, and is essential for spindle assembly and chromosome movement during cell division [112,113]. Detyrosinated microtubules have lower turnover rates and are, therefore, more stable and long-lived than tyrosinated microtubules, which are more dynamic [114]. Interestingly, polymer dynamics of in vitro-assembled microtubules harboring α-tubulin tail modifications is unchanged as compared to unmodified microtubules [115], suggesting that the regulatory role of detyrosination-tyrosination on microtubule dynamics is mostly indirect, possibly by modulating the recruitment and activity of MAPs. Consistent with this, multiple proteins were shown to preferentially bind to either tyrosinated [112,116–118] or detyrosinated [112] microtubules. This appears to be essential for the regulation of protein activity. For instance, tubulin detyrosination was shown to decrease kinesin-1 processivity and increase kinesin-2 velocity and processivity [109]. Moreover, the activity of the microtubule depolymerizing protein MCAK – which preferentially binds to tyrosinated microtubules – is suppressed by microtubule detyrosination [117]. This may contribute to the higher turnover rates observed in tyrosinated vs detyrosinated microtubules [117]. Disruption of tubulin detyrosination-tyrosination has been shown to cause spindle positioning defects [118], disrupt CENP-E-driven chromosome congression [113] and impair NDC80 and CLASP2-dependent establishment of the kinetochore-microtubule interface [112]. Overall, the current data suggests that the adverse effects of disrupted tubulin detyrosination-tyrosination are likely due to altered localization and activity of multiple MAPs, rather than a direct impact of α-tubulin tail modifications on microtubule dynamics.

Acetylation of α-tubulin at lysine 40 (K40) occurs in the microtubule lumen through the catalytic activity of tubulin acetyltransferases [119]. K40 acetylation has been shown to increase microtubules’ resistance against mechanical breakage [120] and it decreases the rates of spontaneous nucleation in vitro [121], with no detectable effects on microtubule polymerization in cells [122]. K40 acetylation is most frequently associated with long-lived, stable microtubules, although whether K40 acetylation promotes microtubule stabilization or is simply a marker of stable microtubules is still under debate. Initial evidence suggested that tubulin acetylation could indeed promote microtubule stabilization [123,124]. Reduction of tubulin acetylation through overexpression of deacetylase HDAC6 increased the susceptibility of microtubules to drug-induced depolymerization [123]. Reversely, increasing tubulin acetylation levels through depletion of HDAC6 hampered microtubule dynamics [124]. However, the possible stabilizing activity of tubulin acetylation has been put in question by findings that HDAC6 itself can inhibit microtubule growth [125]. In addition to acetylation of α-tubulin K40, β-tubulin K252 can also be acetylated, and this was found to impair tubulin polymerization, likely by weakening the α–β intradimer interaction and causing conformational changes that disfavor tubulin incorporation [126].

Microtubule growth rates and chromosomal instability

Accurate spindle assembly is key for proper chromosome segregation during cell division, and alterations in spindle microtubules are known to negatively impact mitosis. For instance, decreased microtubule turnover rates disrupt the establishment of kinetochore-microtubule attachments [26], which in turn leads to chromosome alignment defects [127–129] and lagging chromosomes [130,131]. Altered sliding of non-kinetochore microtubules leads to abnormally shaped spindles that cause chromosome mis-segregation [132–134], whereas defects in astral microtubules cause spindle mis-positioning [135,136]. Importantly, errors in chromosome segregation can often result in aneuploidy [26,27], which is a stepping stone for the development of a variety of human diseases, including developmental disorders [29,30] and tumorigenesis [28].

Significant effort has been made to uncover the molecular players driving spindle assembly and chromosome segregation. However, much less is known about how fine-tuned regulation of microtubule dynamics contributes to chromosomal stability. Chromosomally unstable cancer cell lines were shown to display increased microtubule growth rates, ranging from ~ 18 to ~22 μm/min during mitosis, as compared to an average of between 15.6 to 16.4 μm/min in chromosomally stable cells [27]. This increase in growth rates correlated with high aneuploidy rates [27]. Restoring microtubule growth velocities to normal levels through partial depletion of chTOG [27] or treatment with sub-nanomolar concentrations of Taxol [137] suppressed chromosome mis-segregation [27,137] and decreased aneuploidy rates [27]. Thus, it has been proposed that elevated microtubule growth rates contribute to the development of CIN and aneuploidy.

Previous work has shown that replication stress and DNA damage signaling pathways may contribute to the increase in microtubule growth rates observed in chromosomally-unstable cancer cells. Chromosomally-unstable cancer cells were found to have mild levels of replication stress [137], possibly as a result of upregulation of replication origin firing genes in those cells [138]. Increasing either replication stress [137] or origin firing [138] in chromosomally-stable cells increased microtubule growth rates during mitosis and caused chromosome mis-segregation and CIN [137,138]. Moreover, microtubule growth velocities, chromosome mis-segregation and CIN could all be restored by decreasing replication stress in chromosomally-unstable cells [137]. The impacts of replication stress on microtubule dynamics and CIN could be related to mitotic roles of DNA damage repair components, such as the tumor-suppressor protein BRCA1, which is frequently misregulated in cancer cells [27,139]. It has been shown that the mitotic kinase Chk2 phosphorylates BRCA1 at S988, and this CHK2-BRCA1 phosphorylation axis is required for inactivation of Aurora kinase A through PP6C-SAPS3 phosphatase [27,139]. Loss of CHK2 or PP6C-SAPS3 promotes Aurora A-mediated phosphorylation of BRCA1 at S308, which inhibits its mitotic function and promotes increased microtubule growth rates, chromosome mis-segregation and CIN [27,139]. Therefore, misregulation of any component of the CHK2-BRCA1-PP6C-Aurora A axis in cancer cells could contribute to altered microtubule growth rates and CIN in those cells.

More recently, it has been demonstrated that knock-out of the spindle protein CKAP2 sharply reduces microtubule growth rates during mitosis and increases aneuploidy rates [31], suggesting that excessively low microtubule growth rates can also contribute to CIN, possibly explaining why mis-regulation of CKAP2 is associated with tumor cell proliferation and invasiveness [140]. Interestingly, plotting microtubule growth velocities as a function of aneuploidy rates results in a near-parabolic distribution, where either high or low growth rates correlate with high levels of aneuploidy (Figure 3). In fact, a narrow window of growth velocities that ranges between ~ 13 to ~17 μm/min seems to be optimal for low aneuploidy rates (Figure 3). This opens an interesting possibility that fine-tuned regulation of microtubule growth rates is required for genome stability and is consistent with evidence that altered expression of MAPs that regulate microtubule dynamics is often associated with tumorigenesis (reviewed in Wattanathamsan & Pongrakhananon [141]). Future work exploring a wider range of microtubule growth velocities by, for instance, depleting or overexpressing different MAPs that exert varying impacts on microtubule growth rates is warranted in order to determine optimal microtubule growth velocities required for the maintenance of chromosomal stability.

Figure 3.

Microtubule growth rates and aneuploidy. Scatterplot demonstrating the correlation between microtubule growth rates and percentage of cells with abnormal chromosome numbers (aneuploidy). Data compiled from studies measuring microtubule growth rates during mitosis in either wild-type cells or cells harboring depletions of proteins that impair microtubule growth velocity (see data analysis section for details).

How do altered microtubule growth dynamics cause CIN?

How altered microtubule growth rates lead to CIN is currently unknown. However, optimal microtubule growth velocities may be required for proper kinetochore-microtubule establishment. Merotelic attachments – where a single kinetochore is attached to microtubules originating from both spindle poles – are the major drivers of anaphase lagging chromosomes and, thus, aneuploidy [130,131]. Many MAPs known to regulate microtubule polymerization or depolymerization in vitro also modulate kinetochore-microtubule attachments and/or movements, such as chTOG/XMAP215 [142,143], CKAP2 [31,99], CLASP [144,145], Augmin [146] and mitotic kinesins such as MCAK [147,148] and Kif18A [149,150]. It has been shown that cells with low microtubule growth rates resulting from CKAP2 knock-out display elevated proportions of kinetochore-microtubule mis-attachments [31]. Similarly, high microtubule growth velocities caused by either CHK2 knock-out or AURKA overexpression lead to decreased microtubule turnover rates [27], which is indicative of defective kinetochore-microtubule error correction [26].

Optimal microtubule polymerization may be required for the establishment of microtubule bundles at the kinetochore interface. Consistent with this idea, treatment of cells with low doses of microtubule growth inhibitor BAL27862 was shown to decrease kinetochore-microtubule occupancy from ~ 20 to ~ 13 microtubules per kinetochore, and to impair the segregation of merotelically-attached sister chromatids [151] (Figure 4(a)). In untreated cells, anaphase lagging chromosomes resulting from merotelically-attached kinetochores were more likely to be correctly segregated, whereas, in BAL27862-treated cells, those lagging chromosomes were often mis-segregated [151] (Figure 4(a)).

Figure 4.

Regulation of microtubule growth rates at the kinetochore. (a) Optimal microtubule growth rates may be required for the establishment of kinetochore-microtubule bundles. Kinetochores attached to ~ 20 microtubules successfully segregate merotelically-attached sister chromatids due to higher force differential within the kinetochore that allows it to be pulled towards the correct cell pole [151]. Cells treated with the microtubule destabilizing agent BAL27862 display reduced microtubule occupancy (~13 microtubules/kinetochore), which likely reduces the force differential within the kinetochore and drives the mis-segregation of merotelically-attached kinetochores [151]. (b) Microtubules grown in vitro and bound to kinetochore beads are kept in close proximity by coordination of growth and pausing events in response to tension [152]. In this example, two microtubules (MT 1 and MT 2) initially subjected to equal tension at the kinetochore interface undergo intrinsic variabilities in growth rates, where MT 2 stochastically grows faster than MT 1. The increase in tension exerted on MT 1 favors its growth by delaying pausing events and accelerating exit from pause, thereby restoring equal tension and allowing microtubules to remain in close proximity.

In addition to the regulation of microtubule growth rates, tight control of microtubule depolymerization is also required for kinetochore-microtubule dynamics. Disrupting microtubule depolymerization through depletion of the microtubule depolymerizing protein MCAK increases the rates of merotelic and syntelic attachments – in which both sister kinetochores are attached to microtubules emerging from the same spindle pole [148]. MCAK is thought to stimulate microtubule dynamics by promoting microtubule catastrophe and suppressing rescue [35]. Therefore, microtubule dynamics at the plus end would thus increase the likelihood that microtubules correctly attach to kinetochores [148]. The coordination of microtubule growth and catastrophe rates is thus a major determinant of kinetochore-microtubule establishment and error correction.

Consistent with this idea, an elegant study using dual laser traps demonstrated that intrinsic variabilities in microtubule growth rates and pausing events are coordinated by mechanical forces exerted by microtubule-kinetochore tension [152]. By using a dual laser trap system, the authors found that pairs of microtubule tips growing in vitro maintain close proximity when sharing a single force load on kinetochore beads (Figure 4(b)) [152]. This is achieved by tight coordination of microtubule growth rates and pausing events in response to tension. When one microtubule from the pair grows faster than the other, tension on the faster-growing microtubule decreases, whereas tension on the slower-growing microtubule increases (Figure 4(b)) [152]. This increase in tension at the slower-growing microtubule delays pausing events and accelerates exit from pause, thereby favoring microtubule growth and allowing for microtubules to remain in proximity (Figure 4(b)) [152]. It is, therefore, a very likely possibility that mis-regulation of MAPs that control microtubule growth rates and pausing events may disrupt the coordination of growth velocity and tension within the k-fiber, thereby leading to kinetochore-microtubule mis-attachments that cause CIN.

Concluding remarks and future directions

Our knowledge of microtubule dynamics has grown substantially since dynamic instability was first proposed 40 years ago [8] (Figure 5). A plethora of factors influence microtubule dynamics, including morphological properties (cell size and shape), microtubule subpopulations, cell-cycle dependent changes in MAP expression and activity, tubulin post-translational modifications and even disease-specific phenotypes such as replication stress and protein misregulation found in many cancers (Figure 5). Multiple MAPs have been discovered and their functions characterized, providing important insights into how regulation of microtubule dynamics contributes to cell division and the maintenance of chromosomal stability. The development of fluorescently-labeled microtubule-tracking constructs and advances in imaging technologies have enabled the visualization and quantification of highly dynamic microtubule processes, such as growth and turnover, in live cells. These approaches have led to key discoveries that expand our understanding of how CIN can emerge, namely that decreased microtubule turnover [26] and altered microtubule growth velocities [27,31] are associated with CIN.

Figure 5.

Summary of factors that influence microtubule growth rates. (a) Microtubule growth dynamics may be regulated by cell size and shape. Simulations of cytoplasmic volumes with hydrogel micro-enclosures demonstrated that smaller volumes result in higher microtubule densities [60] (i), which in turn, negatively regulate microtubule growth velocities [60] (ii). Microtubule growth rates are faster at the migrating leading edge of Ptk1 cells as compared to the cell center [54,55] (iii). (b) Different microtubule subpopulations may undergo differential regulation of microtubule growth dynamics. Astral microtubules grow much faster [43] than non-astral microtubules [42,43] in C. elegans mitotic spindles. (c) Microtubule growth rates can be influenced by cell cycle-dependent expression, subcellular localization and post-translational modifications of microtubule-associated proteins (MAPs). Expression of the MAP CKAP2 peaks from late S/early G2 through mitosis, coinciding with its primarily mitotic role [31,78,153]. (i). Subcellular localization of the CPC component survivin shifts from centrosomes during interphase to kinetochores during mitosis, possibly explaining why survivin regulates microtubule growth dynamics exclusively during interphase [56] (ii). XMAP215 is phosphorylated during mitosis, and in vitro phosphorylation of XMAP215 reduces its microtubule polymerase activity [72] (iii). (d) Tubulin post-translational modifications regulate microtubule dynamics by either directly modulating tubulin or polymer biochemical properties [108,120,121,126] or indirectly by modulating the recruitment or activity of MAPs [109,110,112,113,117]. (e) Cellular abnormalities frequently found in cancer cells can influence microtubule growth dynamics. Replication stress [137,138], misregulation of DNA damage repair components such as BRCA1 [27,139] and misregulation of MAPs [27,31] (i.e. overexpression or downregulation) were all found to alter microtubule growth rates during mitosis.

One aspect currently underexplored is how microtubule growth velocities change spatially within the cell and among different spindle microtubule subpopulations. Advances in lattice-light sheet microscopy that allow for fast-acquisition three-dimensional image generation with low phototoxicity [47], coupled with new fluorescent markers such as StayGold that enable imaging over several hours with minimal photobleaching [31,154] will undoubtedly expand our ability to quantify microtubule growth dynamics globally in cells. Another interesting recent development is a new marker of stable microtubules named StableMARK [155] that could be used to label kinetochore microtubules at the spindle. By co-labeling kinetochore-microtubules with StableMARK and microtubule plus-ends with EB comets, one could, in principle, quantify microtubule growth rates exclusively in this microtubule subpopulation, thus providing invaluable insights into the mechanics of microtubule growth dynamics at the kinetochore interface.

Another question that remains to be addressed is the impact of microtubule catastrophe rates on mitotic fidelity and CIN. This aspect is currently limited by the lack of reliable tools for measurements of catastrophe events in cells. Microtubule plus-end tracking with EB proteins can only be used for measurements of microtubule growth and nucleation rates, given that EB proteins dissociate from plus-ends once the microtubule disassembles. Therefore, new tools that allow for the visualization of microtubule shrinkage are certainly warranted. A new class of MAPs named EML2 has been recently found to track shrinking microtubules [116], highlighting an interesting possibility that EML2 could be used as a live cell fluorescent marker of microtubule catastrophe.

Our understanding of the role of microtubule dynamics on mitotic fidelity and CIN has vastly expanded since the development of tools for measurements of microtubule processes in cells, although outstanding questions remain. Over ~ 800 proteins have been found to either localize at or contribute to spindle assembly [156], although their relative roles in the regulation of microtubule dynamics remain to be discovered. Recent advances in microscopy techniques and future work with novel fluorescent markers will undoubtedly provide exciting new insights into nuances of microtubule dynamics during cell division and help uncover the mechanisms driving CIN.

Data analysis

The chart presented in Figure 3 was plotted using GraphPad Prism 9 (www.graphpad.com). Microtubule growth rates were obtained from corresponding studies for the following cell lines: HT1080 wild-type and CKAP2 knock-out (KO) [31], RPE-1 wild-type [27,31] and CKAP2 KO [31], HeLa [47,48], HCT116 wild-type, AURKA OE and CHK2 KO [27], RKO, SW48, SW837, SW620, SW480 wild-type and ch-TOG KO [27] and HT29 [27]. Aneuploidy rates were obtained from the corresponding studies [27,31], except for HeLa cells, in which information was obtained from the American Type Culture Collection (www.atcc.org).

Funding Statement

Work in S.B.’s laboratory is supported by grants from the Canadian Institutes of Health Research [PJT-156193 and PJT-189995], Natural Sciences and Engineering Research Council of Canada [NSERC; RGPIN-2017-04649] and CRBS BlueSky. We thank the reviewers for their contributions to the peer review of this work.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The authors confirm that the data supporting the findings of this study are available within the article.

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/15384101.2025.2485842