The nuclear envelope environment and its cancer connections

The nuclear envelope and cell motility

In cancer, dynamics in cell migration and adhesion are particularly important to the process of metastasis, in which cancer cells escape from the primary tumour and enter the vasculature (intravasation); travel to distant sites; and extravasate to invade a new niche. A change to any key structural element of the nucleus has the potential to affect cell migration simply by increasing the plasticity of nuclear shape and by allowing changes to cell form that are required for squeezing the cell through restricted spaces⁷. This is illustrated by the case of neutro phils, in which low levels of lamins and other nuclear envelope structural components result in a lobulated nucleus that is adapted to further deformation that can be encountered as the neutrophil carries out its functions, such as during egress from blood vessels⁸.

A growing appreciation for the existence of a protein complex that spans across the nuclear envelope, referred to as the linker of nucleoskeleton and cyto-skeleton (LINC) complex^9,10 (FIG. 2), has substantially added to the conceptual framework for understanding the role of nuclear envelope-associated proteins in cell migration⁷. This complex involves interactions between the nuclear lamina and integral membrane proteins of the INM, between luminal domains of INM and ONM proteins, and between regions of ONM proteins that face the cytoplasm and components of the cytoskeleton. Experimental manipulation of components of the LINC complex and associated proteins, such as emerin, affects perinuclear actin organization, nuclear shape, nuclear rigidity and positioning of the centrosome with respect to the nucleus^7,11,12. Cancer-associated alterations in the expression of emerin¹³, lamin A and lamin B (TABLE 1) probably perturb a similar range of features. In turn, the concomitant alterations in nuclear plasticity, cell polarity cues and cell movement may contribute to the propensity for invasion and metastasis. The importance of cellular tension and tissue stiffness in tumorigenesis¹⁴ further highlights the importance of studying the function of the LINC complex, and its role in sensing and in mediating responses to extracellular force, in the context of cancer. Interestingly, nuclear deformation through the disruption of the LINC complex can be uncoupled from changes in the expression of genes, such as early growth response 1 (Egr1) and immediate early response 3 (Ier3; also known as Iex1), that are induced by mechanosensitive signalling in mouse embryonic fibroblasts¹⁵, which underscores the complexity of the coordinated response to changes in extracellular forces.

Components of the nuclear envelope affect cell migration, and its balance with adhesion, in additional ways beyond the paradigms of nuclear structure and physical connectivity between the nucleus and the cytoplasm. For example, the overexpression of lamin A, which increased the motility of a colon cancer cell line, resulted in the upregulation of the actin-binding protein T-plastin (also known as PLS3) and the down-regulation of E-cadherin, and cytoskeletal and adhesion molecules that are involved in regulating cell migration¹⁶. The depletion of the nucleoporin NUP153 leads to alterations in the organization of A-type lamins and to alterations in the localization of the INM protein SUN domain-containing protein 1 (SUN1), and this is accompanied by impaired cell migration¹⁷. Determining whether a role for NUP153 in the integrity of the LINC complex underlies defective migration is worth further investigation.

However, NUP153 — as well as other nucleoporins — may contribute in additional ways to cell motility, as nucleoporins interact with important players in cell migration^17–23. For example, NUP358 (also known as RANBP2) is reported to form a cooperative interaction with kinesin 2 to regulate the localization of adenomatous polyposis coli (APC) to the cell cortex¹⁹. This pore protein has also been implicated in regulating RAP family GTPases by binding a guanine nucleotide exchange factor, exchange protein directly activated by cyclic AMP 1 (EPAC1; also known as RAPGEF3). The depletion of NUP358 was found to enhance cAMP-induced cell adhesion in OVCAR3 ovarian cancer cells, which is consistent with biochemical characterization showing that NUP358 inhibits EPAC1 from activating RAP1B²².

Collectively, these observations expand the functional horizon of nuclear pore proteins beyond serving as the gateway to the nucleus. Moreover, spatial organization is often a crucial aspect of the roles for nuclear envelope components in regulating cell migration and adhesion. For these reasons, it will be important to track not only changes in the expression of nuclear envelope components, but also changes in nuclear envelope organization, to better understand the relationship between the nuclear envelope and the capacity for tumour dissemination and invasiveness.

The nuclear envelope and fidelity of cell division

NPCs have important roles in cell division in the canonical context of nucleocytoplasmic trafficking by restricting the nuclear entry and exit of key mitotic regulators, such as p27 (also known as KIP1) and cyclin B^24,25, as well as by directing selective export of mRNAs that encode cell cycle regulators²⁶. The NPC-associated sentrin-specific protease 1 (SENP1)²⁷ is also reported to influence cell cycle progression by regulating the expression of cyclin-dependent kinase (CDK) inhibitors²⁸; and lamin A and lamina-associated poly-peptide 2α (LAP2α) actively participate in cell cycle control by regulating RB-mediated transcriptional control of target genes^3,29 (discussed below). Moreover, the nuclear envelope, including the underlying lamina and the NPCs, undergoes dramatic remodelling during each cell division: signalling in mitosis leads to nuclear envelope disassembly, and, following chromosome segregation, to coordinate reformation³⁰. Errors in this process can contribute to various alterations ranging from aneuploidy to defective nuclear structure that have the potential to alter cellular function. There is now evidence for surveillance mechanisms of correct post-mitotic reformation of NPCs. Specifically, disrupting the assembly of the basket-like feature on the nuclear face of the NPC resulted in persistent aurora kinase B activation and delayed the completion of cytokinesis³¹. This role for aurora kinase B is probably deregulated in tumours in which aurora kinase B is over-expressed or, conversely, in which this kinase is inhibited by chemotherapy (for example, AZD1152, GSK1070916A and other compounds in clinical trials)³². Mechanisms that are involved in cell cycle-associated remodelling of the nuclear envelope are also important to consider for their potential to reveal new chemotherapeutic targets, as blocking a step in this process could provide a novel strategy to stop cell proliferation.

During interphase, emerin, nesprins and A-type lamins have roles in coordinating the anchorage of centrosomes to the ONM⁹. This contributes to microtubule organization and cellular orientation, which is important for migration, but probably also affects the central role of centrosomes in organizing and positioning the mitotic spindle. Notably, treatment with farnesyltransferase inhibitors (FTIs) results in defective centrosome separation, which causes profound aberrations in mitosis. Preventing lamin B1 farnesylation (BOX 1) was shown to specifically contribute to this phenotype³³. It is plausible that defects in targeting lamin B1 to nuclear membranes and consequently deficient integration of lamin B1 into the LINC complex affects centrosome separation, but this has not yet been investigated. Although FTIs are in clinical development for the treatment of cancer, the germane downstream targets have been elusive since it was discovered that oncogenic RAS mutants are not the targets, as originally intended³⁴. A connection between FTIs, lamin B1 and impaired cell division may provide a crucial clue in the effort to understand the relevant targets and to further develop this therapeutic strategy.

At the G2/M cell cycle transition, two nucleo porins have roles in tethering centrosomes to the nuclear envelope^35,36. During prophase, these interactions are thought to contribute to the role of microtubules in nuclear envelope breakdown³⁰ and probably also to re-position sister centrosomes to opposite sides of the nucleus^35,36. The importance of coordinating these events has been illustrated by a recent observation that cells in which centrosomes do not separate fully before nuclear envelope breakdown have a greater frequency of mis-segregated chromosomes³⁷. The NPC additionally contains factors with important regulatory roles during mitosis that are released when the NPC disassembles at the end of prophase. For example, NPC-associated SUMO enzymes, such as the SUMO E3 ligase NUP358 and the SUMO proteases SENP1 and SENP2, are implicated in the sumoylation and the desumoylation of several mitotic factors^38–41. The sumoylation of proteins such as topoisomerase IIα and CENPE is thought to regulate their localization at mitosis^38,39. Targeting SENP2 to the NPC has been shown to restrain its catalytic activity⁴², which suggests a paradigm for the importance of this localization for both SENP1 and SENP2. This further implies that over-expression of these SENPs may override such spatial control of their activity. Interestingly, premature desumoylation of topoisomerase IIα by SENP1 or SENP2, which could either impair targeting of topoisomerase IIα to inner centromeres³⁸ or upregulate its decatenation activity⁴³, is proposed to lead to aneuploidy⁴⁴, which hints at a possible pathogenic mechanism when SENP1 is overexpressed in cancers^45,46.

A newly discovered complex between the mitotic regulator transforming acidic coiled-coil-containing protein 3 (TACC3) and the tumour suppressor tuberin (which is encoded by TSC2) was found to reside at the nuclear envelope partly through an interaction of TACC3–tuberin with NUP62 (REF. 47). Deficiency of either TSC2 or TACC3 in mouse embryonic fibroblasts results in strikingly aberrant nuclear morphology; however, further investigation is needed to determine whether this is due to the disruption of a direct role at the nuclear envelope or whether it is a consequence of aberrancies that arise in mitosis⁴⁷. If it is a result of aberrations that arise in mitosis, then perhaps NUP62 serves to sequester, rather than direct, the activity of TACC3–tuberin. Interaction with the nuclear pore can both restrict the spatial and temporal activity of mitotic regulators, and poise them for action. This is illustrated by knockdown of the nucleoporin translocated promoter region (TPR), which is a binding partner of the mitotic spindle assembly checkpoint proteins MAD1 and MAD2, which resulted in diminished recruitment of MAD1 to kinetochores and a compromised spindle assembly checkpoint in HeLa cells⁴⁸.

The roles of nuclear envelope proteins also extend to the establishment of the mitotic spindle apparatus and chromosome segregation. Although these are not functions of the nuclear envelope per se, they are important to consider here, as a change in the nuclear periphery will often lead to a change in the fidelity of chromo-some inheritance, which is a common feature of cancer cells that contributes to tumour phenotype and, at least in some cases, to tumorigenesis⁴⁹. There are now many well-characterized examples of NPC components that have roles during mitosis. This includes roles for NUP358 at kinetochores and for NUP98 in regulating the APC/C (anaphase promoting complex; also known as the cyclo-some)⁵ and in regulating mitotic microtubule dynamics to promote spindle assembly⁵⁰. NUP88, which is frequently overexpressed in cancers, as well as TPR, which is found in oncogenic translocations⁵¹, have also been proposed to participate in proper mitotic spindle function⁵. Mice with reduced expression of NUP358 or of RNA export 1 homologue (RAE1), which is a binding partner of NUP98, have chromosome segregation defects, severe aneuploidy and are susceptible to carcinogen-induced tumour formation^38,52,53. Although aneuploidy is likely to be a contributing factor, its relationship to tumorigenesis is complex, with reduced expression of NUP358, but not of RAE1, leading to spontaneous tumours in mice and the simultaneous reduction of NUP98 expression with RAE1 increasing aneuploidy but not tumour susceptibility⁵². The multifunctional role of these nucleoporins clearly creates individual profiles of deregulation when their expression levels are altered, culminating in different outcomes.

Interestingly, remnants of the nuclear membrane itself, along with associated proteins, are also emerging as active participants in mitosis^54,55. A role in regulating spindle morphogenesis for the nuclear membrane and associated proteins may explain why the downregulation of emerin leads to cytokinesis failure and polyploidy¹³ or this may perhaps relate to the connection between the effects of FTIs and lamin B1 (REF. 33). In either of these cases, decreased levels or mistargeting of nuclear membrane-associated proteins may impair a direct role in chromosome segregation, resulting in lagging chromosomes that in turn prevent proper progression through cytokinesis. Although attributing phenotype to a specific role is difficult — as is the case with many multi functional proteins — it is evident that proteins of the nuclear periphery contribute at multiple levels to coordinating cell division. Indeed, a screen of transmembrane nuclear envelope proteins for roles in regulating the cell cycle has recently revealed eight novel candidates, reinforcing the importance of delving further into this link⁵⁶. It is clear that cancer-associated changes in the composition and/or morphology of the nuclear periphery have the potential to create mitotic defects and that such errors can contribute to chromosome instability, which is often seen in tumours. Mis-coordinated chromosome segregation, however, can also lead to the formation of lobulated, malformed nuclei⁵⁷. Therefore, even low-frequency errors in chromosome segregation may be further reinforced and augmented over time by altering nuclear envelope-dependent cell cycle events.

The nuclear envelope and gene expression

The nuclear envelope environment serves as a scaffold for numerous transcription factors (including, receptor-regulated SMADs (R-SMADS), sterol regulatory element-binding protein 1 (SREBP1), FOS and OCT1 (also known as POU2F1)), chromatin remodelers (for example, CREB-binding protein (CBP; also known as CREBBP) and histone deacetylase 3 (HDAC3)) and regulators of transcription (such as RB, inhibitor of growth protein 1 (ING1), germ cell-less (GCL; also known as GMCL1), protein phosphatase 1 (PP1), PP2, SENP1, SENP2 and ubiquitin-conjugating enzyme 9 (UBC9))^3,27,58,59. Coordinated nuclear import, sequestration, post-translational modification and release of these proteins into the nucleoplasm are integrated at the nuclear envelope to regulate mRNA expression⁶⁰. Similarly, the nuclear periphery — lamin B1 in particular — also modulates the expression of microRNAs (miRNAs)⁶¹. This additional layer of control is especially noteworthy as one of the most affected miRNAs in lamin B1 (Lmnb1)^Δ/Δ cells, miR-31, is thought to have important roles in cancer. For example, miR-31 expression is elevated in colorectal cancer and its expression is correlated with more advanced stage of disease⁶². In the context of breast cancer, however, miR-31 expression was found to inversely correlate with metastasis, and experimentally inducing its expression in xenograft models of metastasis caused the regression of metastatic lesions^63,64. The pleiotropic targets of this miRNA probably make the context of its expression crucial, but overall these observations highlight the importance of the connection between miRNA expression and lamin B1.

Nucleoporins take part in the regulated transport of transcription factors, sometimes in a direct manner, as has been described for the transforming growth factor-β (TGFβ)-responsive transcription factor SMAD2 (REF. 65). Transcription factor export from the nucleus can also be a key step in the regulation of gene expression; thus, nucleoporins that interface with export machinery make important contributions in this regard. The cytoplasmic aggregates of NUP88 that are often seen when this nucleo porin is overexpressed in tumours could divert CRM1 (also known as EXP1) from its normal interactions at the pore, thereby impairing export-mediated downregulation of transiently activated transcription factors such as nuclear factor-κB (NF-κB)^66,67.

Nucleoporins are also intimately connected to chromatin, as they take part in regulating transcription^68–70. There is evidence for two classes of nucleoporin–chromatin interactions: those that take place at the NPC and those that occur in the nucleoplasm with a dynamic population of nucleoporins⁷⁰. Although details of the determinants of nucleoporin–chromatin association and the exact roles of nucleoporins at this site are still unknown, one pathway that is emerging as a modulator of the connection between nucleoporins and chromatin in mammalian cells is protein acetylation. Histone deacetylase (HDAC) inhibitors were shown to alter NUP93-associated chromosome regions⁷¹, and, in cardiomyocytes, HDAC4 was reported to counteract the association of nucleoporins that are reactive with the monoclonal antibody mAb414 (which primarily recognizes NUP358, NUP214, NUP153 and NUP62) and certain target genes⁷². Moreover, NUP155 was discovered to be a partner of HDAC4, and this interaction was found to be important for the inhibitory effect of HDAC4 on the expression of these target genes. A deeper understanding of the interplay between nucleoporins and HDACs is relevant to the application of HDAC inhibitors as a cancer therapy⁷³.

Although there are roles for mobile, nucleoplasmic nucleoporins in regulating transcriptional activity at interphase, the nuclear periphery remains an important environment for the regulation of gene expression^74–77. In Drosophila melanogaster and human cells, A-type lamins and NPCs have been reported to possess boundary activity, which probably allows the coexistence of both transcriptionally permissive and repressive chromatin configurations at the nuclear periphery^78,79. Experimental depletion of lamin B1 in mammalian cells results in blebs of the nuclear envelope that are enriched with A-type lamin and are, moreover, where gene-rich DNA tends to be concentrated. Notably, these lamin-defined microdomains are not, however, associated with active transcription, but rather they appear to contain stalled RNA polymer-ase II (REF. 76). Nuclear blebs with similar features have recently been reported in prostate cancer (see below)⁸⁰. These and other observations underscore that changes at the nuclear envelope and in its morphology are likely to be accompanied by alterations in gene expression. Such transcriptional deregulation may in some cases be a protective adaptation, but it also holds the potential of contributing to the progression from normal to tumorigenic cell growth.

The nuclear envelope and signalling pathway

Many signalling molecules associate with the nuclear envelope, which plays a part in coordinating multiple signalling pathways, including those downstream of TGFβ, insulin-like growth factor (IGF), epidermal growth factor (EGF), WNT and NOTCH^3,81,82. These signalling pathways are often deregulated in cancer, leading to aberrant cell proliferation that is uncoupled from normal cues and control; changes at the nuclear envelope have the capacity to play a part in this deregulation.

Components of the nuclear periphery can differentially modulate signalling pathways. This point is exemplified by the canonical WNT signalling pathway, which leads to the stabilization and translocation of β-catenin to the nucleus, where it partners with members of the T cell factor (TCF)/ lymphoid enhancer factor (LEF) family of transcription factors to activate specific genes. Whereas nesprin 2 promotes the nuclear localization of β-catenin, emerin seems to have the opposite effect^83,84. The specific context may be important for these differing roles, but it is evident that perturbing the expression of these nuclear envelope proteins leads to defective WNT signalling and severely disrupts cellular proliferation and differentiation programmes^60,83,85. The regulation of WNT signalling by the nuclear envelope is further highlighted by evidence that additionally implicates the sumoylation of transcription factor 4 (TCF4) — which is stimulated by NUP358 and counteracted by SENP2 — in promoting TCF4–β-catenin interaction and transcriptional activity⁸⁶.

Extracellular signalling can also lead to functional alterations at the nuclear envelope. NPC density can change in response to hormonal stimulation^87,88, according to the phase of the cell cycle^89,90 and in response to cell signalling⁹¹. One pathway that illustrates the complex crosstalk between extracellular signals and NPCs is the MAPK–ERK pathway. Following mechanical stimulation that is induced by cell stretching, MAPK signalling is activated and leads to increased numbers of NPCs⁹¹. Activation of this pathway can also lead to functional modulation of nucleoporins. NUP50, NUP153 and NUP214 were found to be phosphorylated by ERK, and their interaction with the soluble transport receptor importin-β is decreased in the presence of active ERK, which is concomitant with a decreased ability of the receptor to accumulate in the nucleus⁹². This indicates that ERK negatively regulates importin-β-dependent import into the nucleus; exactly how this shift in homeostasis would affect tumour biology in the case in which ERK signalling is activated is not yet clear. In the case of TPR, phosphorylation by ERK2 stabilizes interaction with this kinase, which perhaps allows TPR to serve as a scaffold for further phosphorylation events⁹³. In addition, direct interaction of activated ERK1–ERK2 with TPR, as well as NUP153, is thought to promote nuclear translocation of activated ERK1–ERK2, independently of soluble trafficking receptors^93–95. Enhanced nuclear entry of ERK1–ERK2 in cancer cells versus cells cultured from normal tissue has been proposed to result from differences in nucleoporin levels. Experimentally decreasing levels of NUP153 in the ovarian cancer cell lines OVCAR-10 and SKOV3 attenuated the expression of FOS, which is a downstream target of the ERK pathway, without altering levels of activated ERK⁹⁶.

Components of the INM are also implicated in ERK1– ERK2 regulation; for example, depletion of LEM domain-containing 2 (LEMD2; also known as NET25) was shown to result in elevated levels of phosphorylated ERK1 following EGF exposure⁹⁷ — raising the issue of whether deregulation of LEMD2-dependent events at the nuclear envelope could also contribute to the hyperactivity of the EGF pathway in the context of cancer. Changes in the number of NPCs, in nucleoporin partnerships and in the global architecture of the nuclear envelope⁹⁸, suggest that there is a shift in the landscape of the nuclear envelope in conjunction with MAPK–ERK activation. There remains much to understand in terms of variables, such as acute versus chronic signalling and cellular context. In addition to the phosphorylation of nuclear periphery proteins, ERK1 and ERK2 can also physically compete with RB for interaction with lamin A, which thereby displaces RB into the nucleoplasm for phosphorylation and thus promotes entry into the cell cycle⁹⁹. Emerin and LAP2β, the phosphorylation of which may lead to functional modulation³, are additional components of the nuclear envelope that are targeted by multiple signalling cascades (such as EGF and fibroblast growth factor (FGF) pathways) and kinases (for example, SRC and ABL)³.

Whatever the mechanism by which a tumour cell acquires the hallmark folding and invagination of the nuclear envelope (FIG. 1), one consequence with respect to signalling may be increased sensitization to the key second-messenger molecule calcium. Involutions of the nuclear envelope that occur under non-pathogenic conditions have been suggested to enhance nuclear transduction of calcium signals^98,100. Calcium-mediated signalling is altered in multiple cancers¹⁰¹ and is important in regulating key cell cycle factors, as well as survival and apoptosis pathways. Thus, this is a prime opportunity to apply approaches that are used in other fields to elucidate how cancer-associated changes in nuclear architecture affect calcium signalling. Our knowledge to date, however, already indicates that morphological and molecular changes at the nuclear envelope that are associated with cancer are likely to be intertwined with alterations in intracellular signalling.

The nuclear envelope and genome stability

Studies of laminopathies have highlighted the importance of lamins to DNA repair. Mutation or aberrant expression of lamin A compromises DNA repair pathways, with defective recruitment of DNA double-strand break (DSB) repair proteins, including RAD50 and RAD51, and aberrant recruitment of the nucleotide excision repair factor xeroderma pigmentosum complementation group A (XPA) to constitutive DSBs detected as γH2AX foci¹⁰². The expression of lamin A mutants that are associated with muscular dystrophy or progeria in cultured cells also results in the mislocalization of the DNA damage response kinase ataxiatelangiectasia and Rad3-related (ATR)¹⁰³. Furthermore, response to DNA damage caused by cisplatin or ultraviolet (UV) irradiation is impaired in the presence of lamin A mutants¹⁰³. Several layers of regulation in the DNA damage response intersect with lamin: loss of lamin A function reduces the transcription of RAD51 and BRCA1 (REF. 104), and also results in the downregulation and mislocalization of the tumour suppressors RB^105,106 and ING1 (REF. 107), as well as the DNA repair factor p53-binding protein 1 (53BP1)¹⁰⁸. Thus, changes in lamin levels (TABLE 1) or even lamin organization (as may be the case when lobulated polymorphic nuclei are observed in tumours) could reflect defects in the ability of a cell to respond to DNA damage and the accumulation of DNA lesions. This deregulation increases the potential of acquiring mutations that might contribute to tumorigenesis or further augment the aggressiveness of the tumour cells, but may also promote sensitivity to chemotherapeutic strategies that capitalize on an impaired ability for DNA repair.

Proteins of the NPC also aid in coordinating DNA repair. In yeast, many observations point to the NPC as a hub for SUMO- and ubiquitin-mediated regulation of DNA repair^27,109. In addition, Nup1p, Nup2p and Nup60p are phosphorylated in response to DNA damage that is induced by treatment with the DNA-alkylating agent methyl methanesulphonate, and these proteins have been identified as direct targets of Rad53p (the yeast homologue of the human checkpoint kinase CHK2), which functions in triggering cell cycle checkpoint arrest in response to DNA damage¹¹⁰. Replication stress and inhibition of Chk2 were found to be lethal in zebrafish larvae bearing a mutation in the nucleoporin Elys¹¹¹. In mice, tissue-specific disruption of ELYS in intestinal epithelium leads to elevated activation of the DNA damage response, along with crypt cell apoptosis and disruption of intestinal epithelium morphology¹¹². Overexpression of the UBC9-binding motif of the mammalian nucleo porin NUP358 compromises RAD51 recruitment to DNA damage-induced foci¹¹³; whereas, NUP153 was recently identified in a screen for factors important for 53BP1 recruitment to DNA damage-induced foci following ionizing radiation¹¹⁴. Follow-up experiments showed that NUP153 is crucial to the nuclear import of 53BP1, although whether this is the full scope of its role remains to be seen. In a different study, lowering the level of NUP62 expression in cultured ovarian carcinoma cells was found to result in cisplatin resistance¹¹⁵. Although the molecular pathway underlying the role of NUP62 in this context has not yet been elucidated, these results point towards the importance of nucleoporins in cellular responses to DNA damage.

The nuclear envelope environment is also of particular importance to telomere maintenance. The fibro-blasts from patients with Hutchinson–Gilford progeria syndrome (HGPS) (BOX 3) have aberrantly short telomeres¹¹⁶. Interestingly, overexpression of either the mutant form of lamin A found in progeria, which accumulates at the INM with a persistently farnesylated tail, or wild-type lamin A, leads to telomere shortening¹¹⁷, indicating that the balanced expression of lamin A, as well as the integrity of the nuclear lamina, are important for telomere maintenance. Lamin A, together with its binding partners LAP2α and barrier to autointegration factor (BAF; also known as BANF1), are known to participate in tethering telomeres to the nuclear periphery^118,119. Molecular details of the roles for components of the nuclear periphery in telomere maintenance are better characterized in yeast, in which the recruitment of telomeres by INM proteins is proposed to promote telomere addition by telomerase; whereas, the recruitment of telomeres to NPCs is implicated in telomerase- independent recombination mechanisms that are the last resort when telomeres are severely eroded¹²⁰. A recent study has highlighted a role of sumoylation in the dynamics of telomere tethering to the nuclear periphery¹²¹. With many players now identified, this area of research is poised to yield further mechanistic information that is important for the spatiotemporal control of DNA repair. Given that genomic instability and telomere attrition are typical features of tumour cells from many cancer types, connections between the nuclear envelope and DNA repair and telomere maintenance are a highly relevant consideration in the assessment of phenotypes that accompany changes in nuclear morphology.

Applying knowledge of the nuclear envelope

As our knowledge of the molecular nature of the nuclear envelope environment grows, can this be applied to create more precise, objective measurements to classify tumours or to make prognostic predictions? Given the myriad determinants of nuclear morphology, a simple molecular signature that corresponds to altered shape is unlikely. Compensatory changes and cell type-specific context will influence how the alteration of a nuclear envelope component manifests. Nonetheless, the intimate connection between nuclear envelope morphology and cancer suggests that the expression levels of key proteins at the nuclear envelope may provide an informative parameter in tumour detection and characterization. Several components of the nuclear periphery — lamins, LAP2 and emerin — have already been assessed as bio-markers in a wide range of cancer types (for example, see REFS 13,122–125) (TABLE 1). NUP88 was serendipi tously discovered as a cancer biomarker^126,127, and has since been found in many studies to be overexpressed in malignant tissues^127–133. More generally, lamins, certain nucleoporins and SENP1 have emerged in unbiased screens as biomarkers of interest for their change in levels found in cancer (for example, see REFS 45,134–136). Intriguingly, not only was lamin B1 found in a proteomic approach to be upregulated in hepatocellular tumours¹³⁷, but mRNA encoding lamin B1 was also detected in the blood circulation, and its detection in plasma indicated early stage hepatocellular carcinoma with 76% sensitivity and 82% accuracy¹³⁸.

Absolute levels, of course, do not always reflect the functional status of a particular protein. In this regard, it is interesting that in a comparative phosphoproteomic analysis using subclones of a breast tumour cell line with differing metastatic properties, phosphorylation of lamin A was found to be associated with metastatic potential¹³⁹. Although the mechanistic consequence of this modification in tumour cells has not yet been pursued, one function ascribed to lamin A phosphorylation is in regulating the process of lamina assembly and disassembly during the cell cycle¹⁴⁰. It is also interesting to note that certain viral infections cause the phosphorylation of lamins, which triggers local lamina reorganization that affects nuclear envelope integrity (for example, see REF. 141).

The assessment of the components of the nuclear envelope as prognostic indicators has been limited (TABLE 1), and again does not reflect a simple pattern^81,142. Elevated expression of various lamins was found to correlate with indicators of poor prognosis in liver (lamin B1)¹³⁸, prostate (lamin A)¹³⁶ and colorectal cancer (lamin A)¹⁶. By contrast, downregulation of A-type lamins correlates with poor prognosis in gastric carcinoma¹⁴³ and large B cell lymphoma¹⁴⁴. Changes in the levels of lamins at the nuclear envelope, either an increase or a decrease, may cause an imbalance that undermines their function. Consistent with this, comparative proteomic analysis in progression models of both fibrosarcoma and colorectal carcinoma point to an inverse correlation between lamin A expression and malignant progression^145,146; whereas, lamin A levels increased in more advanced stages of ovarian cancer¹⁴⁷. Collectively, these studies indicate that, when considering lamin expression, deviation from the norm, rather than an increase or a decrease, could be used as a marker of nuclear envelope deregulation. However, overexpression of a distinct marker of the nuclear envelope, NUP88, consistently correlates with an aggressive tumour phenotype and poor prognosis in numerous cancers^{128,129,131,132,148}.

The long list of translocations that involve nucleoporins also underscores the link between NPC-associated alterations and tumorigenesis. Although many of the resulting fusion proteins have roles that are independent of the NPC (see REF. 51 for a detailed review), altered function of the NPC may contribute to pathogenicity. In the case of NUP214–ABL, targeting of this fusion protein to the NPC proved important for its transforming activity¹⁴⁹, and NUP98 fusion proteins have been found to divert CRM1 from crucial roles in nuclear export¹⁵⁰. In addition to their ancillary use for diagnosis, detecting the presence of these nucleoporin fusions can have implications for therapeutic stratification, as has been shown for the t(6;9) DEK–NUP214 fusion, which is associated with aggressive acute myelogenous leukaemia¹⁵¹.

As with other biomarkers, a particular combination — which could include, but need not exclusively include, components of the nuclear envelope — might be most informative. This combination might be derived from unbiased analysis or could be directed towards markers that reflect a particular feature of the tumour or its environment that is relevant to consider in conjunction with alterations at the nuclear envelope. As an example, tracking both EGF receptor and lamin levels might be more predictive than tracking either factor alone. Another strategy is to combine the assessment of a nuclear envelope component with a functional attribute of the tumour.

Meijer and colleagues¹⁵² have recently carried out such an analysis in colon cancer, looking at lamin A expression in conjunction with microsatellite stability, which indicates the functional status of the mismatch repair pathway¹⁵². Interestingly, they found that, in microsatellite-stable stage III tumours, if LMNA mRNA was low, 100% of patients who had not received adjuvant chemotherapy had disease recurrence versus 37.8% of patients with the same profile except for high LMNA expression (P < 0.01). This finding needs to be confirmed in an independent data set, but it points to additional variables that may help to resolve discrepant conclusions of biomarker studies that looked exclusively at LMNA in cohorts that were mixed with respect to microsatellite stability, tumour grade and status of adjuvant therapy.

The morphology of the nucleus is particularly informative in pathological assessment because it gives an integrated view of many layers of cellular regulation that affect cell growth and behaviour. Thus, to derive more benefit in the clinical arena from our growing knowledge of the nuclear envelope environment, another tactic is to image key features of the nuclear periphery with more precision. For example, the detection of nuclear envelope components can give a much more detailed view of the contours of the nucleus than can be obtained from standard histological stains, as has been shown by tracking the INM proteins emerin and lamin¹⁵³. And, although immunohistochemistry reveals this pattern, the detection of specific nuclear envelope components by immunofluorescence gives a more detailed view¹⁵³ (FIG. 1) and is becoming more readily applied in the clinical setting¹⁵⁴. There are indications in the literature that higher NPC numbers are associated with more aggressive tumours^155–158. The quantitative assessment of NPCs is now possible with high-resolution light microscopy (for example, see REF. 90) and could be used to evaluate this more thoroughly and to determine its potential use as a prognostic tool. Beyond absolute levels of proteins in the nuclear envelope environment, their precise localization more fully reflects whether they are functioning normally. Consistent with this, the prevalence of lamin B-deficient microdomains of the nuclear envelope was found to increase with Gleason grade in prostate cancer⁸⁰. Overall, the information we are gaining in terms of the composition and function of the nuclear envelope opens up several new strategies to make greater use of the longstanding connection between nuclear morphology and cancer phenotype.

Conclusions and future perspectives

Currently, our knowledge suggests that changes in the nuclear envelope environment can propagate to alterations in signalling, as well as the control of transcription, cell migration and polarity, genome stability and cell division (FIG. 3). The nuclear envelope environment is clearly tightly interconnected, both physically and functionally, with many aspects of cell biology. For example, nuclear shape is highly sensitive to the function of the endoplasmic reticulum (ER), including lipid production¹⁵⁹; thus, cancer-associated changes in nuclear shape may additionally implicate an alteration in ER function, lipid homeostasis and/or lipid signalling. As the site of secretory and transmembrane protein synthesis, the ER is also crucial for post-translational protein glycosylation and for the folding of many proteins, including those located in the nuclear envelope. Hypoxia and nutrient depletion, which are characteristics of many tumour microenvironments, can compromise these functions of the ER and can lead to the accumulation of misfolded proteins, creating a condition generally termed ER stress¹⁶⁰. Not only would this probably influence the range of nuclear envelope proteins (and, in turn, the structure and function of the nucleus), but physical continuity between the nuclear envelope and the ER also means that the lumen of the nuclear envelope would be directly subjected to stress-associated changes that should be evaluated for their effect on nuclear envelope function.

Recent advances support the view that the nuclear envelope is a site where many aspects of cell function are coordinated. Nuclear envelope components have vital roles both in this environment and sometimes — through dynamic spatial and temporal regulation — at other sites in the cell. Changes that are associated with alterations of the nuclear envelope can undermine normal cell function and have the potential to augment the ability of tumour cells to adapt and evolve. With the availability of molecular tools for detection and analysis, systematic approaches can now be carried out to more directly address the ramifications of discrete changes of the nuclear envelope on tumour growth. For example, mice expressing different levels or mutant alleles of lamins could be crossed with genetically engineered mouse models of cancer¹⁶¹. The inducible expression of protein (or, similarly, the expression of mutants or loss of expression through genetic deletion) could be used to disrupt the normal organization of the nuclear envelope in order to probe the contribution of such changes to different stages in carcinogenesis (such as, initiation versus metastasis). Such models would also be of use in understanding how the response to DNA-damaging agents and other chemotherapeutics is altered by changes at the nuclear envelope. As pathologists have known for years, even subtle alterations in the morphology of the nuclear envelope can have profound implications on cellular function. The knowledge of underlying pathways and molecules has given us new insight into how such perturbations may influence tumour cell function and has laid the groundwork for future research to further decipher and capitalize on the connections between the nuclear envelope and cancer.