Introduction

The ever-growing number of diseases, in which PLS3 is involved, highlights the importance of this F-actin-binding and -bundling protein, with a broad spectrum of cellular pathways (Fig. 1). It seems that the expression of PLS3 is tightly regulated since knockout or mutations cause osteoporosis, while overexpression seems to trigger osteoarthritis and various types of cancer. Instead in several neuromuscular disorders, such as spinal muscular atrophy (SMA), amyotrophic lateral sclerosis (ALS) and CHP1-associated ataxia, PLS3 overexpression acts as a protective modifier. In this review, we present the current knowledge on the gene expression and protein function, the various cellular functions, in which PLS3 is involved and the disorders associated with PLS3 levels (decreased or increased) (Fig. 2).

Fig. 1
figure 1

PLS3-associated disorders and main cell types involved. a PLS3 is involved in a variety of diseases, which associate with the PLS3 levels in a cell. Low protein abundance leads to osteoporosis, while increased levels are associated with cancer and osteoarthritis. In some neurodegenerative diseases, where F-actin levels are reduced, PLS3 overexpression acts protective. Green letters imply a protective role of PLS3 while red letters highlight PLS3 as disease driving protein in the depicted disorders. PLS3 plastin 3, SMA spinal muscular atrophy, ALS amyotrophic lateral sclerosis. b PLS3 fulfills distinct functions within different cell types. Osteoclasts, osteoblasts, osteocytes and chondrocytes are the target cells within the bone disease spectrum, which are influenced when PLS3 is dysregulated. In some neurodegenerative disorders, where motor neurons (e.g., SMA) or Purkinje cells (e.g., ataxia) are affected, overexpression of PLS3 showed a protective effect. Involvement of PLS3 in cancer is highly divers and includes different kinds of solid tissues as well as hematopoietic and lymphatic cancers. The figure was created with BioRender.com

Full size image
Fig. 2
figure 2

Cellular function of PLS3. The main function of PLS3 is its role in F-actin-binding and -bundling activity and, thus, in F-actin dynamics. Consequently, PLS3 is involved in endocytosis, cell motility, cell adhesion, mechanotransduction, pathogen infection, Ca2+ homeostasis, exocytosis, vesicle trafficking, axonal local translation and others. PLS3 protein is depicted as red circle. ER endoplasmic reticulum, ECM extra cellular matrix. The figure was created with BioRender.com

Full size image

Gene and expression

In 1980, Bretscher & Weber identified fimbrin (later renamed to plastin), a 68 kD protein isolated from small intestine of the chicken [1]. Plastins are a family of actin-binding and -bundling proteins consisting of three tissue-specific paralogs PLS1, PLS2, and PLS3 localized in humans on chromosomes 3, 13 and X, respectively [2]. The three isoforms share approximately 70% nucleotide sequence identity [3]. PLS1 (plastin 1, I-plastin) is expressed in the intestine, colon, kidney and hair cell stereocilia of the inner ear, PLS2 (plastin 2, L-plastin) in hematopoietic cell lineages and many types of cancer cells, and PLS3 (plastin 3, T-plastin), the most abundant isoform, in cells from solid tissues [2, 3]. In lower organisms, such as yeast, only one isoform is present, known as fimbrin or Sac6. Before 2008, papers referred to fimbrin or T-plastin, and only after that, the correct name of plastin 3 has been established in publications.

The expression of PLS3 is of utmost interest in the context of several medical conditions, such as spinal muscular atrophy (SMA), osteoporosis, osteoarthritis as well as a large number of cancers. While the regulation of the PLS3 expression in different tissues is still elusive, several mechanisms have been proposed, which do not necessarily exclude each other (Table 1). Early reports stated that PLS3 expression is limited to solid tissues [4]. However, elevated expression levels were found in blood samples in about 5% of healthy individuals [5]. PLS3 shows high expression in the fetal and adult spinal cord and PC12 cells during neuronal differentiation, where it accumulates in growth cones [5]. PLS3 mRNA is enriched and locally transcribed in the axonal compartment of motor neurons [6].

Table 1 Mechanisms that were proposed to regulate the expression of PLS3 in cancer
Full size table

PLS3 is located in both, humans and mice, on the X-chromosome next to the macrosatellite DXZ4 [7]. In women, approximately 15% of genes escape the X-chromosomal inactivation (XCI) and are, therefore, biallelically expressed [8]. Another 15% of genes variably escape XCI in a tissue-specific manner [8,9,10]. PLS3 is a facultative escape gene in humans but not in mice [11]. The macrosatellite DXZ4 is essential for XCI and has a highly variable repeat number of 50–100 copies of a 3 kb repeat monomer [12]. Interestingly, on the inactive X-chromosome, DXZ4 is hypomethylated and binds to the transcriptional regulator protein CCTF-binding factor (CTCF), both features of chromatin regulation [13]. It has been hypothesized that the copy number of DXZ4 may modulate the escape of genes in its molecular neighborhood and thereby the expression level of PLS3 in females [14].

Accordingly, in SMN1-deleted asymptomatic women of spinal muscular atrophy (SMA) discordant families, the PLS3 expression is elevated in blood and Epstein–Barr Virus transformed lymphoblastoid cells but not in their SMA-affected siblings, indicating a sex-specific protective effect (see more details in Sect. 3.6.1) [5]. Moreover, mutations or knockout of PLS3 mainly affect men or male mice, while women or female mice are either very mildly or not affected, further supporting the sex specificity of PLS3 expression [15, 16].

Function

The main function of PLS3 is in F-actin-binding and -bundling. Consequently, PLS3 is involved in all processes dependent on F-actin dynamics such as cell motility, focal adhesion, cell division, endocytosis, neurotransmission, vesicle trafficking, axonal local translation, intracellular calcium PLS3-dependent processes, pathogen infection and others (Fig. 2).

Protein structure

The three plastin isoforms seem to have different effects on the actin cytoskeleton organization depending on the cell type [17]. Plastins are monomeric proteins with a single polypeptide chain composed of two tandem repeats of actin-binding domains (ABD1 and ABD2) [18]. Each of these ABDs encompasses a pair of ~ 125 residue calponin homology (CH) domains (ABD1: CH1 and CH2, ABD2: CH3 and CH4) [19]. Each of the CH domains consists of four major α-helices. While three α-helices form a loose bundle of helices, the fourth is orientated perpendicular to the major bundles [20]. At their N-terminal end, plastins possess a Ca2+-binding regulatory domain (RD) of approximately 100 amino acids. The RD contains two EF-hands and a calmodulin-binding motif (CBM) [3, 21, 22]. The RD is connected with the ABD core via a linker [23]. PLS3 consists of 630 amino acids and the molecular mass of the protein is 70.8 kDa. In humans, the homology between PLS1 and PLS2 is 73%, 75% between PLS1 and PLS3, and 79% between PLS2 and PLS3 [3]. Comparisons of the plastin isoforms between humans and mice show 94% identity for PLS1, 96% for PLS2, and 99% for PLS3 [3].

Plastins cross-link actin filaments into bundles, which are higher-order assemblies, through the tandem pair in their ABDs [24, 25]. So far, only the crystal structures of the ABD1 and ABD2 actin-binding cores are resolved [24, 26]. A homology-based model of the PLS3 actin-binding core has been generated using the Phyre2 web portal for protein modeling, prediction, and analysis [23]. The full-length crystal structures of all plastins remain to be determined.

F-actin dynamics/ cell motility

The main function of plastins is the binding and bundling of F-actin [3, 23, 27, 28]. F-actin is a polymer composed of globular (G-actin) subunits. These microfilaments are the main components of the cytoskeleton [29]. The actin cytoskeleton is a complex network whose dynamic formation influences numerous fundamental physiological cellular processes, such as focal adhesion, cell motility, endo- as well as exocytosis, mechanotransduction and cell division [30]. The F-actin network underlies a tight regulation and must be assembled, maintained, and disassembled in 3D at the correct time and place, and with proper filament organization and dynamics [31]. Thereby, the dynamic formation of the actin cytoskeleton is dependent on various F-actin-binding and -bundling proteins [32].

Plastins are involved in the assembly and organization of the actin cytoskeleton through their two ABDs. Binding of each ABD to two separate actin filaments promotes the formation of a bundle with a center-to-center distance of ~ 120 Å between the filaments [28]. However, the exact mechanism by which the ABD1 and ABD2 bind and bundle actin is poorly understood. It is assumed that both ABDs differently interact with actin [26]. Actin-binding is mediated through ABD1, while bundling is facilitated through both ABDs [33]. Furthermore, the plastin isoforms differentially interact with F-actin resulting in distinct F-actin organization [17, 20]. It has been suggested that PLS3 binding via both ABD1 and ABD2 is essential at the leading edge and focal adhesions [23]. The F-actin-binding and -bundling activity of the ABD domains is Ca2+ dependent, as the binding of Ca2+ by EF-hands inhibits the interaction with F-actin [33, 34].

The ABD cores form a compact and rather globular structure in an antiparallel arrangement induced by the contact between the CH1 and CH4 domains at the N- and C-terminal ends [3, 26]. This conformational structure seems to be highly dynamic and dependent on the presence of Ca2+ [35]. The conformational plasticity of CH2, located within the structurally polymorphic ABD1, influences the diverse functions of different actin assemblies [22]. This indicates that the structural plasticity is related to the function of the plastin isoform. Furthermore, the crystal structure indicates that the two ABDs harbor different binding characteristics with F-actin [20, 33].

As an F-actin-binding and -bundling protein, PLS3 plays an essential role in cell motility [36]. The movement of a cell is activated by adhesion to the extracellular matrix, executed through membrane remodeling at the leading edge and controlled by the actin cytoskeleton [37, 38]. PLS3 promotes membrane protrusion and cell migration to overcome gaps in the extracellular matrix and when adhesion is spatially gapped [39]. In the skin, PLS3 influences the basement membrane assembly [40]. Furthermore, PLS3 is involved in membrane trafficking under hypoxia [41]. All these functions arise from the interaction with the actin cytoskeleton. It has been speculated that PLS3 serves as a mechanical link between the actin polymerization network at the front of the cell and the myosin motor activity in the cell body [36]. PLS3 modulates the actin dynamics and generates force independent of cross-bridge formation mediated by actin related protein 2/3 complex (Arp2/3) [42]. The Arp2/3 complex facilitates the actin polymerization which is essential to generate pushing forces capable of deforming the cell membrane [31].

Influence of intracellular calcium on PLS3-dependent processes

The crucial role of Ca2+ for F-actin-bundling by plastins has been a subject of research for a long time [43]. Especially, in terms of PLS2, studies focused on its Ca2+-dependent functions regarding actin-bundling and specifically T-cell activation and motility [34, 44], reviewed in Babich, Burkhardt [45], Morley [46].

Interestingly, comparisons of the three plastin isoforms show a much lower amino acid identity at the N-terminal EF-hand domains than the higher conserved actin-binding domains. Thus, several studies compared the plastin isoforms regarding their Ca2+ sensitivities and Ca2+-dependent functions [33, 47,48,49].

EF-hands of both PLS2 and PLS3 were shown to consist of alpha-helix-rich sequences and underlie conformational changes upon Ca2+-binding [47]. PLS3 has a lower sensitivity for Ca2+ as indicated by only a slight change of the EF-hands’ secondary structure in contrast to PLS2. Also, higher concentrations of Ca2+ are required for structural changes of the EF-hand motifs of PLS3 [47]. Congruently, monitoring the disassembly of plastin/F-actin bundles induced by Ca2+, PLS3 has been proven to be less sensitive to Ca2+ than both PLS1 and PLS2 [33]. Further analysis has shown a difference in affinity to Ca2+ within the two EF-hands of PLS3 (one high and one low affinity site) in contrast to PLS2 comprising two high-affinity sites [47].

Generally, the presence of Ca2+ reduces the ability to bundle F-actin in all three isoforms [33]. Further studies on the functional linkage between the EF-hands and the linker region CBM revealed inhibition of rapid proteolytic cleavage at the CBM upon Ca2+ stimulation, emphasizing a Ca2+-dependent linkage of these two regions in all three human isoforms. Focusing on the impact of Ca2+ on the ABDs, inequality is found: one ABD, most likely ABD1, binds F-actin independent of the presence of Ca2+; whereas at the other ABD, most likely ABD2, F-actin-binding is inhibited by Ca2+ bound by the EF-hands [33].

To gain further insight into the process of Ca2+-binding to the EF-hands as well as into kinetic and structural details on Ca2+-regulated domains of plastins, three biophysical methods have been used: surface plasmon resonance studies, isothermal titration calorimetry studies, and nuclear magnetic resonance spectroscopy [48]. In terms of Ca2+-regulated function, a crucial role of the regulatory helix 5 (H5) linker region localized between the second EF-hand motif and the first CH1 domain of ABD1 has been suggested, which may even differ between the plastin isoforms. Precisely, the H5 region of PLS3 (TPL-EF-H5 construct) is not displaced by the bee venom peptide melittin, unlike a possible displacement in PLS2 (LPL-EF-H5 construct) [48].

Another approach assessed the Ca2+-coordination structures of PLS2 and PLS3 and their synthetic peptide analogs by Fourier transform infrared spectroscopy [49]. Based on the results, an association between the aggregation tendency of the two Ca2+ binding sites of PLS3 and its lower sensitivity to Ca2+ was suggested [49].

The importance of well-regulated Ca2+ homeostasis for the proper function of PLS3 is underlined by several findings of disturbed Ca2+ regulation associated with loss of its functionality, shown in PLS3 rescue in SMA models [50] or osteogenesis imperfecta (OI)-associated PLS3 mutations [23].

Endocytosis

The role of PLS3 in endocytosis has been first described in budding yeast [51]. Knockout of the PLS3 ortholog, Sac6, causes massive impairment of endocytosis; especially of the receptor-mediated internalization of the pheromone α-factor [51]. Actin-bundling is a crucial process required for proper endocytosis in which Sac6-mediated bundling of actin filaments forms a framework in the early stages of this process [52]. Actin organization in cortical actin patches, which are dynamic actin structures within the inner faced layer of the cell membrane in yeast, dependent on actin-bundling proteins Sac6 and Scp1 (ortholog of human SM22). Loss of both proteins results in tremendous defects in patch biogenesis (increased patch lifetime), while Sac6 loss alone causes random movement or patch disassembly, with the absence of fast movements [52]. In filamentous fungi, Ashbya gossypii and Aspergillus nidulans, the polarized hyphal tip growth and endocytosis are Sac6/fimbrin A dependent, and their lack causes reduced rates of endocytic uptake [53, 54]. This highlights the importance of PLS3 orthologs as actin regulators and the requirement of functional actin dynamics to fulfill proper extension growth across species. Additionally, Sac6 along with other actin-binding and -bundling proteins (such as Abp85) are required for the uptake of the maltose transporter in yeast [55]. Moreover, fimbrin cross-linkers, twisting the actin filaments, provide approximately one-sixth of the energy required for endocytosis [56, 57]. Importantly, Arp2/3-actin networks are crucial to generate the force during membrane invaginations. Those networks are disrupted in Sac6-deficient cells and consequently result in disturbed endocytosis [56].

In fission yeast, it has been proposed that fimbrin selectively regulates the access to actin filaments for other actin-binding proteins, such as the tropomyosin Cdc8p [58]. Accordingly, in fimbrin-deficient cells, mislocalization of tropomyosin to actin patches results in increased patch lifespan and decreased motility [58]. Fimbrin is also able to displace tropomyosin from actin filaments lowering the inhibition of myosin I by tropomyosin and thereby ensuring motor activity [59].

The role of PLS3 in endocytosis caught first attention in human studies when PLS3 has been identified as a protective modifier of SMA [5]. In the context of SMA, both F-actin levels and Ca2+- homoeostasis are reduced, two processes crucial for endocytosis, but restored by PLS3 overexpression [5, 60,61,62,63,64]. Importantly, endocytosis is a key process in neurons as they rely on constant refilling of the recycling vesicle pool in the presynapse. Apart from other reasons, endocytosis is a main cellular process disturbed in SMA, contributing to disturbed function of neuromuscular junctions (NMJs) [61, 65,66,67,68].

Another approach assessed defects in endocytic pathways in a SMA model and hypothesize that PLS3 increases the expression of endocytic proteins by supporting the availability of ribonuclear protein granules containing the required components for their translation [68]. Local axonal translation is dependent on transport of mRNAs and microRNAs along the axons to the growth cone of polarized neurons. Thereby, mRNAs and microRNAs not only hitchhike on late endosomes and lysosomes but they also act as hubs for local translation, finally contributing to the function and maintenance of neurons and neuronal circuits [69, 70]. Congruently, PLS3 directly interacts with activated RAB5, an early endosomal protein, and thereby regulates endocytic activity in mammalian cells [71]. It is, therefore, tempting to postulate that PLS3 overexpression may restore impaired vesicle trafficking in SMA and consequently local translation.

Signaling

The dynamic regulation of F-actin assembly and the interaction with the actin cytoskeleton is regulated by the coordinated activation of actin assembly factors through different signaling cascades [31]. It is accepted that PLS3 is involved in these signal transduction pathways [72]. Early studies show that the up- or down-regulation of PLS3 influences different signaling pathways. Downregulation of PLS3 has been shown to inhibit the p38 mitogen-activated protein kinase (MAPK) signaling pathway in MDA-MB-231 cells, which mediates apoptosis [73]. In keratinocytes, PLS3 is involved in the calcineurin/nuclear factor of activated T-cell (NFAT) pathway, which is a major regulator of cell migration [74]. In osteocytes, PLS3 levels regulate the expression of RELA proto-oncogene, NFΚκB subunit. PLS3 interacts with NFκB repressing factor (NKRF) facilitating its translocation into the nucleus, and thus the transcription of nuclear factor-activated T cells c1 (NFATC1) and an important factor in osteoclastogenesis [16]. Through the PI3K/AKT signaling pathway, PLS3 regulates tumor progression by promoting the proliferation and migration of cancer cells [75].

Several studies have demonstrated that PLS2 can be phosphorylated at positions Ser 5 and Ser7 during leukocyte activation by various stimuli [3]. Phosphorylation of Ser5 is mediated by cAMP-dependent protein kinase A, while the relevant kinase for Ser7 is still controversially discussed. In osteoclasts, phosphorylation of PLS2 on Ser5 and Ser7 increases the F-actin-bundling capacity [76] and the avidity for cellular F-actin and F-actin-binding activity [77]. In macrophages, PLS2 is phosphorylated exclusively on Ser5 by stimulation with bacterial lipopolysaccharide [78]. Ser5 is only found in PLS2 and not conserved in other plastin isoforms. To date, no phosphorylation of PLS3 is known and further studies are required [72].

Disease involvement

PLS3 levels require a tight balancing in the cell. Aberrantly increased levels are associated with cancer and osteoarthritis. Instead in several neurodegenerative conditions associated with decreased F-actin levels, PLS3 overexpression acts protective. Instead, deletions or loss-of-function mutations in PLS3 cause osteoporosis in humans and mice. Finally, S-nitrosylation of PLS3 associates with thoracic aortic dissection (Fig. 1). Moreover, PLS3 is involved in infection and pathogen entry.

Cancer

Malignancies in solid tissues

A rising number of biomedical studies investigated PLS3 in the context of the most common malignancies (colorectal, prostate, breast, gastric, and lung cancer) and explored its potential use as a biomarker. Metastases in distant organs are the most common cause of death in these solid tissue tumors. The cell migration and invasion of tumor cells are driven by the modulation of the actin cytoskeleton. PLS3 is highly expressed during epithelial–mesenchymal transition (EMT) in circulating tumor cells (CTCs) in colorectal cancer. Elevated levels of PLS3 during EMT in CTCs negatively regulate expression levels of epithelial genes, and positively regulate the levels of mesenchymal genes and thereby increase the invasiveness of colorectal cancer [79]. Several mechanisms that could explain the regulation of PLS3 in CTCs had been suggested. Chromosomal instability, a hallmark of cancer formation, can trigger a copy number gain of Xq23 leading to PLS3 overexpression [79, 80]. A study in colorectal cancer cells states that PLS3 is a downstream target of Lamin A, which is a risk factor for this type of cancer. PLS3 in turn downregulates E-cadherin leading to increased invasiveness during EMT [81]. Furthermore, genetic polymorphisms modifying PLS3 levels have been discussed. The intronic polymorphism rs871773 C > T is associated with a low recurrence time of colorectal cancer, while the rs6643869 polymorphism correlates with the time of tumor recurrence in women and has been identified in CTCs [82,83,84].

PLS3 overexpression is found in two-third of CTCs from breast cancer patients in different stages of EMT and is associated with poor overall survival, especially in patients with luminal type A or triple-negative breast cancer [85]. Silencing of PLS3 in MDA-MB-231 triple-negative breast cancer cells increases the sensitivity towards the anti-cancer agent paclitaxel [73]. A recent study has published a gene panel consisting of seven genes (PLS3, MGB1, HER2, CK19, CDH1, CDH2, and VIM) that can be used to discriminate EMT stages in CTCs tested by qPCR [86]. For prostate cancer, a gene panel with 14 genes has been developed including PLS3, VIM, and CDH2, and in 9.5% of patients an increased expression of PLS3 has been found [87]. Interestingly, LOXL1 downregulates CDH1 and upregulates VIM, CDH2, PLS3, and SNAI2 in gastric cancer cells, where expression of LOXL1 correlates with EMT [88]. In gastric cancer, increased expression of PLS3 indicates a poor prognosis and is associated with cancer differentiation, the depth of tumor invasion and EMT [89]. Chromatin immunoprecipitation (ChIP) assays in patients with gastric cancer have shown that the transcription factor ZNF471 acts as a tumor suppressor gene. ZNF471 binds the PLS3 promoter and suppresses its expression. Methylation of the CpG-methylation site of the ZNF471 promoter is a useful prognostic marker for overall survival in gastric cancer patients [90]. High PLS3 expression is also a feature of pancreatic cancer cells. Here, PLS3 acts as an oncogene and its overexpression triggers the activation of the PI3K/AKT signaling pathway leading to cancer progression [75]. In non-small-cell lung cancer (NSCLC), increased PLS3 plasma levels are a predictor of poor survival. PLS3 expression is of therapeutic value and can predict the responsiveness to treatment with Nivolumab—a PD-1 monoclonal antibody—in NSCLC patients [91]. Moreover, UV-light- and cisplatin-resistant tumors show elevated PLS3 expression [92, 93]. Downregulation of PLS3 in human liver cancer cells increased the sensitivity of these cells to cisplatin [94]. Cells lacking PLS3 expression are sensitive to DNA damage and silencing of PLS3 led to increased damage by UV light. By this, expression of PLS3 could be used as a therapeutic marker in irradiation therapy [94]. Together, the mentioned studies suggest an association of increased PLS3 expression during EMT and a high prognostic potential of PLS3 as a biomarker in malignancies of solid tissues. The expression levels of PLS3 represent a useful marker for cancer prognosis especially as part of a gene panel.

Malignancies of the hematopoietic and lymphatic system

Elevated PLS3 expression is a negative prognostic marker for acute myeloid leukemia (AML), the most common acute leukemia in adults, while knockdown of PLS3 increases survival in vivo [95]. In Sézary Syndrome (SS), an aggressive, rare form of cutaneous T-cell lymphomas (CTCL), circulating CD4+ T-cells show increased expression of PLS3, TWIST1 and GATA6 compared to normal CD4+ T-cells [96, 97]. The promoter regions of all three genes have been found to be hypomethylated in SS CD4+ T cells indicating an epigenetic regulation of the expression levels. This is of large interest, since all three genes are located on different chromosomes [98,99,100,101]. Therefore, PLS3 has been suggested as a biomarker for SS CD4+ T cells and it is associated with an unfavorable disease outcome [102,103,104].

Bone disorders

Osteoporosis

Dynamic regulation of the actin cytoskeleton by actin-binding and -bundling proteins like PLS3 [3, 20] is particularly crucial in the musculoskeletal system to instantly adapt to environmental changes through mechanotransduction [105, 106]. Mechanotransduction is the conversion of mechanical signals into cellular response, which is assumed to play an essential role in several pathologies of the musculoskeletal system, such as osteoporosis [107, 108] and osteoarthritis [109,110,111]. In chicken, fimbrin is detected in the dendrites of osteocytes [105, 112], which play a key role in mechanotransduction [113]. Therefore, PLS3 could influence the mechanical signal transformation [108].

In 2013, pathogenic variants in PLS3 have been reported to be associated with osteoporosis including fractures in men and mild osteoporosis in women [108]. Moreover, PLS3 showed a significant association with osteoporosis in women upon menopause, suggesting that mutations in PLS3 not only associate with the monogenic but also the complex genetic trait of osteoporosis [108]. Osteoporosis is a multifactorial disease that is dependent on hormonal, environmental as well as genetic factors [114, 115]. Low bone mineral density (BMD) seems to be genetically determined in 50–85% of cases [114, 116,117,118]. This emphasizes the genetic component as a determining factor for BMD and its relating fracture risk. Primary (hereditary) osteoporosis usually becomes symptomatic in childhood and is, therefore, referred to as early-onset osteoporosis. The clinical features are characterized by low BMD (age normalized average score (Z-score < 2.0)), occurring vertebral compression fractures (VCFs), or low-trauma fracture history [119, 120]. The most common form of monogenic osteoporosis is OI [120], which additionally affects extraskeletal features like blue sclera, joint hypermobility, and deafness [15, 114, 121]. Around 85–90% [15, 114] of OI cases are linked to a dysregulation in type I collagen, although the list of OI-associated genes is increasing [15, 114, 122,123,124,125].

Up to date, 27 mutations in PLS3 are associated with early-onset osteoporosis (Table 2; Fig. 3). Nevertheless, rare variants also indicate OI traits [108, 126, 127], which is why PLS3 was also included as a genetic cause for OI according to Van Dijk, Sillence [121]. However, several studies have reported variants that lead to severe skeletal abnormalities among women, resembling the osteoporotic phenotype of men [108, 115, 128,129,130]. This huge variation in heterozygous women is suggested to be caused by X-inactivation of the mutant allele or PLS3 escaping X-inactivation which is why women are less severely affected [15, 126, 129, 131, 132]. Even though bone morphometry was very heterogeneous [16, 126, 133,134,135,136], most male patients showed peripheral fractures, low BMD, VCFs, especially in the thoracic spine, and low bone turnover rate, while only a few developed extraskeletal OI traits [108, 126,127,128, 131, 137,138,139,140], developmental delay [15, 127] or neuromuscular abnormalities, like waddling gait [108, 126, 131, 141]. So far, no specific biomarkers have been identified, which can distinguish genetic factors of osteoporosis, although microRNAs are getting more and more popular as functional markers [120, 142]. This makes it difficult to diagnose PLS3 mutations, although genetic analysis would be important to assess bone fragility risk within families. Thus, to find genetic causes of bone disorders, it is recommended to use gene panel screening containing known monogenic osteoporosis genes in fracture-prone children to identify genetic factors influencing bone health and to reveal family risk [114, 143,144,145].

Table 2 Clinical findings of patients with X-linked early-onset osteoporosis caused by PLS3 mutations
Full size table
Fig. 3
figure 3

Schematic representation of the PLS3 gene and corresponding protein domains. Illustrated are the causative mutations for osteoporosis and the resulting amino acid changes that were previously published. All depicted mutations were modified according to HGMD Professional (NM_005032.7 (GRCh38)). References and clinical characterization are shown in Table 2. Note, RNA containing frameshift and nonsense variants (except for the last exon 16), most likely undergo nonsense-mediated mRNA decay and thus no protein is produced. E EF-hand motifs, CBM calmodulin-binding motif, RD regulatory domain, CH calponin-homology domain, ABD actin-binding domain, Core actin-binding core domain, SNP single-nucleotide polymorphism, CNV copy number variation, NA not available; ★ = PLS3 isoform including 18 exons, accession No.: NG_012518. The figure was created with BioRender.com

Full size image

In summary, most of the osteoporosis-related PLS3 variants are frameshift [108, 122, 135, 137, 138, 144] or nonsense mutations resulting in premature termination codons [108, 128, 130, 140], which are followed by mRNA decay [108, 128, 131, 135]. Next to that, partial deletions within PLS3 were identified [126, 140, 146]. Surprisingly, no phenotypic differences were seen between mutations causing deletion or truncation of PLS3 protein or missense or in frame insertion mutations resulting in mutated PLS3 protein, most likely because all types of mutations cause a loss of function of PLS3 [108, 128,129,130,131, 135, 140, 147].

Studies of missense mutations in full-length PLS3 report impairments in F-actin-bundling ability or defects in Ca2+ sensitivity, which lead to localization shifts within the cell and can also disturb Ca2+-dependent structural rearrangements necessary for actin dynamics [23]. Impaired actin dynamics caused by mutations within the ABD2 of PLS3 result in inefficient F-actin-bundling properties, without influencing actin-binding by ABD1 [23]. In turn, this specific mutation interferes with PLS3′s actin-cytoskeletal association and thus, its localization at the lamellipodia and leading edge [23]. Considering these findings, it could be hypothesized that the proper intracellular localization of PLS3 is dependent on its actin-bundling function [23]. Similarly, in another study, a PLS3 mutation, disrupting the ABD interface, results in conformational changes and defective actin-binding and -bundling function [140].

A cellular mislocalization and impaired Ca2+-dependent actin-associated distribution of PLS3 were also related to mutations influencing the Ca2+ sensitivity, causing either hypo- or hypersensitivity to Ca2+. Misregulated PLS3 cycling could thereby result in impaired bone homeostasis and might explain the osteoporotic phenotype in these patients [23].

To analyze the proposed functions of PLS3 regarding mechanotransduction [5, 15, 108, 120, 143, 144, 148], Ca2+ regulation [23, 33, 149], osteoblastic bone mineralization [15, 120, 126, 131, 134, 143, 144, 150], osteoclastogenesis [16] and vesicular trafficking [15, 16, 61, 151], various PLS3 mutations have been investigated and mouse model studies established to specifically investigate the effect of PLS3 loss or overexpression [16].

It is well established that PLS3 is highly abundant in the dendrites of osteocytes, which not only control mechanosensing but also regulate osteoblast and osteoclast activity, and an impairment of this cell system might explain the resulting osteoporotic phenotype [136, 152]. Indeed, in osteocytes, mutations in PLS3 interfere with cellular signaling, resulting in imbalanced bone homeostasis [121, 129]. In line with this, data acquired from PLS3 mutation carriers showed increased levels of apoptotic osteocytes together with abnormal gene expression of osteocyte-related genes [153].

Because PLS3 mutations cause childhood-onset osteoporosis associated with low BMD, it has been further hypothesized that the bone-forming osteoblasts might be disturbed by PLS3 loss [129, 143]. In healthy bone conditions, PLS3 abundance increases upon osteoblast differentiation [131, 154]. However, in some patients with PLS3 mutations, matrix mineralization and osteoblast number are decreased [126, 129, 131, 155]. The resulting imbalanced levels of osteoblasts and osteoclasts, favoring bone resorption, might result in osteoporosis.

For differentiation and bone formation, osteoblasts rely on high Ca2+ concentrations, which is in part regulated by the interaction of the Ca2+-binding proteins PLS3 and PLS2. Through their interaction, bound Ca2+ is released to increase the intracellular Ca2+ level when the extracellular Ca2+ level is low [149]. However, mutations in Ca2+ binding domains (EF-hands) of PLS3 weakened the interaction of PLS3 with PLS2, as revealed by PLS2-TRAP co-immunoprecipitation and western blot analysis, and thus might contribute to the osteoporotic phenotype [149].

To enable further analyses, a ubiquitous Pls3 KO mouse model has been generated that is comparable to loss-of-function mutations observed in OI patients [16]. In 3-month-old Pls3 KO animals reduced cortical and trabecular thickness and decreased bone parameters is seen. In line with humans, the osteoporotic phenotype in male mice is more pronounced than in females [16]. Importantly, the NFκB pathway, an essential pathway for osteoclastogenesis, is strongly influenced upon PLS3 loss [16]. NFκB signaling is initiated through binding of receptor activator of NFκB ligand (RANK-L) to its receptor RANK on the surface of osteoclasts, activating the transcription of nuclear factor-activated T cells c1 (Nfatc1) [156]. Besides, Nfatc1 expression is negatively regulated by the NFκB-repressing factor, NKRF. As PLS3 binds to NKRF, Pls3 KO results in insufficient repression of Nfatc1 nuclear translocation, and therefore in increased osteoclast differentiation [16]. Moreover, podosomes, specific F-actin ring structures in osteoclasts, important for resorptive activity, migration, and adhesion, are structurally disturbed in Pls3 KO mice, which might be due to increased depolymerization of F-actin and, thus, instability of the cytoskeleton [16].

When investigating the effects of elevated PLS3 levels in mice carrying homozygously a human PLS3 transgene under an ubiquitously expressing promotor and integrated into the Rosa26 locus on chromosome 6, a thickening of the cortical bone is seen in 3-month-old female PLS3 OE mice compared to the controls [16]. This led to increased stiffness and breaking force in the females, which might be due to impaired bone renewal [16, 116]. In contrast to Pls3 KO, PLS3 OE might cause increased translocation of NKRF to the nucleus, which in turn results in inhibition of Nfatc1 transcription and consequently inhibits osteoclastogenesis [16]. These PLS3 effects might be additionally influenced by impaired intracellular vesicle dynamics [61]. Similar to the Pls3 KO, also in PLS3 OE mice defective podosome formation is detected, probably caused by reduced F-actin disassembly [16]. In summary, it is assumed that an optimal amount of PLS3 is essential to sustain bone health. Neither removal of PLS3, resulting in bone abnormalities, nor an increased level of PLS3, leading to a hyperosteotic phenotype, seems to be beneficial for bone homeostasis but rather detrimental.

Osteoarthritis

In 2015, increased PLS3 levels have been reported in chondrocytes from patients with osteoarthritis compared to healthy individuals [157]. Osteoarthritis is a multifactorial disease that affects the whole joint and is defined by degeneration of the articular cartilage, synovial inflammation and increased bone mass in the subchondral bone [158]. This is in line with the observed elevated cortical thickness in PLS3 OE mice, which might be due to inhibited osteoclastogenesis [16]. Furthermore, the RANKL–RANK pathway plays a crucial role not only in osteoarthritis in general [159,160,161], but seems to be stressed in osteoarthritis patients showing increased PLS3 levels [157]. Further, differentially upregulated pathways in osteoarthritis patients are related to the actin cytoskeleton, endocytosis, TGF-β, MAPK, TNF-α, WNT and general metabolism processes [157]. As mentioned before, an interference of the mechanotransduction process is a crucial factor associated with the pathogenesis of osteoarthritis, where also the impact of PLS3 on the actin cytoskeleton is suggested to play an essential role [109,110,111]. Recent findings have shown that patients with varus gonarthrosis-induced osteoarthritis show higher levels of PLS3 in the medial compartment compared to the less affected lateral knee compartment. Besides, PLS3 was also predominantly found at the direct force-facing contact side of the respective joints [162]. These data might confirm previous assumptions that PLS3 seems to be involved in the mechanotransduction system and its enhanced level is associated with osteoarthritis. So far, less is known about the effect of PLS3 deletions on cartilage. Another study has shown no effect on the intervertebral disc of osteoporosis patients linked to loss of function and PLS3 deletions [148]. The influence of PLS3 overexpression, PLS3 loss, and PLS3 mutations on articular cartilage itself remains elusive and the underlying mechanism affecting cartilage health still needs to be explored.

Since about 5% of the general population show increased PLS3 expression in blood [5], it would be interesting to study if the increased expression of PLS3 in blood is associated with an increase of PLS3 in cartilage, that seems to be associated with osteoarthritis, and thus can act as a biomarker for osteoarthritis as well [162].

Thoracic aortic dissection

In terms of its role in diseases, PLS3 has also been shown to be involved in the causative mechanisms of thoracic aortic dissection in mice [163]. In more detail, S-nitrosylation of PLS3, induced by angiotensin II and mediated by inducible nitric oxide synthase, promoted the formation of a complex of PLS3 with cofilin, which depolymerizes actin and facilitates actin dynamics, and plectin, which links various cytoskeletal proteins and is important in formation of cell junctions in endothelial cells. This process was detected to aggravate the development of thoracic aortic dissection via pathological angiogenesis as well as disruption of the adherent junctions at the endothelial barrier [163].

Neurodegeneration

PLS3 is highly abundant in certain areas of the brain especially in the hippocampus but also in the spinal cord [164]. It is strongly upregulated during the development and maturation of motor neurons pointing towards an important role in these processes [66]. PLS3 is an important interaction partner of several proteins or protein complexes involved in neurodegeneration disorders such as SMA [5, 61], ataxia [66], amyotrophic lateral sclerosis (ALS) [68], and Charcot–Marie–Tooth (CMT) [165] (Fig. 1).

Spinal muscular atrophy

One of the most unexpected findings was the discovery that PLS3 overexpression acts as a protective modifier of SMA [5]. SMA is primarily a neurodegenerative motor neuron disorder and the most common cause of genetic death in infancy. It is caused by deletions and functional loss of SMN1, while the copy gene SMN2 strongly modulates the disease severity [166].

Rarely, individuals carrying a homozygous deletion of SMN1 and three or four SMN2 copies are fully asymptomatic, in contrast to approximately 99% of individuals with such a genotype who develop SMA. This suggests that protective modifier(s) in the genome of these subjects counteract the detrimental effect of reduced SMN levels [5, 65, 167]. PLS3 is the first human SMA protective modifier identified by transcriptome analysis of differentially expressed genes. All asymptomatic, but none of the symptomatic siblings of six discordant families showed an up to ~ 40-fold upregulation of PLS3 in lymphoblastoid cell lines (where PLS3 usually is not expressed). Instead, the same individuals show no difference in PLS3 expression in fibroblasts, suggesting a tissue-specific regulation [5]. Generation of iPSCs from fibroblasts of two discordant families showed that the expression of PLS3 is highly elevated in asymptomatic individuals but not in symptomatic siblings [167]. This is also found in motor neurons (MNs) differentiated from iPSCs [167]. The mechanism behind is still not fully understood.

Moreover, three additional SMA protective modifiers, coronin-1C (CORO1C), a Ca2+-dependent protein involved in actin dynamics, and two Ca2+ sensor proteins, neurocalcin delta (NCALD) and calcineurin-like EF-hand protein 1 (CHP1) have been identified [61, 65, 66]. Similar to PLS3, CORO1C acts protective when upregulated, whereas NCALD and CHP1 act protective when being downregulated as shown in various animal models [61, 65, 66]. Importantly CORO1C and CHP1 interact with PLS3. All modifiers are involved in endocytosis and restore reduced endocytosis in SMA (see section of 2.4).

PLS3 overexpression acts protective not only in humans with SMN1 mutations but also in zebrafish, worm, fly and mouse SMA models [5, 61, 64, 168,169,170]. In all systems, PLS3 overexpression ameliorates or counteracts the major hallmarks of SMA pathology. Reduced SMN level impairs F-actin dynamics, which might be due to a disturbed transport of β-actin mRNA along the axons [63]. F-actin dynamics is pivotal for cellular integrity and is involved in cellular shape, migration, vesicular trafficking, RNA translation, and endocytosis, among others [171]. In highly polarized MNs, all these processes are particularly crucial. In severely affected SMA models, numerous F-actin-dependent processes including axonal growth, axonal connectivity at the NMJ, neurotransmission, F-actin caging, synaptic vesicle recycling as well as proprioceptive input at MN somata are reduced [60, 64, 172,173,174]. All these processes were either restored or ameliorated by PLS3 overexpression as shown across various SMA animal models [61, 64].

Unexpectedly, while impaired NMJ function and motoric abilities are ameliorated in severely affected mice, survival is prolonged by only a few days [64]. To generate a more comparable situation as found in discordant SMA families (where 3–4 SMN2 copies are present, but never only 2 copies as in the severe SMA mouse model), an intermediate-like SMA mouse phenotype has been generated by subcutaneous injection of a low dose of SMN-ASOs (Nusinersen, 30 µg at P2 and P3) [61]. This slightly increased SMN level rescues inner organ function and doubles the survival rate. These mice still die due to MN loss with 1 month of age. In contrast, additional overexpression of human PLS3, from a homozygous PLS3 transgene [64], rescues MN and NMJ function and significantly prolongs survival (60% survive > 250 days; 30% > 400 days), clearly proving the protective effect of PLS3 overexpression in an intermediate SMA mouse model [61]. Besides, a gene therapy approach using adenovirus-associated virus (AAV9)-PLS3, in combination with low dose SMN-ASO, ameliorates the SMA phenotype in mice and prolongs survival [175, 176]. These experiments provide strong evidence that PLS3 is a genuine protective modifier and highlight the power of combinatorial therapies in SMA.

Exo- and endocytosis are crucial for neurotransmission in neurons. To properly maintain neurotransmission, endocytic uptake is essential to replenish the recycling pool, which supplies vesicles to the readily releasable pool (RRP) [177]. At NMJs of severely affected SMA mice, the organization and number of docked vesicles on the presynaptic site are significantly reduced causing a decrease of neurotransmitter release. Moreover, the RRP size is significantly reduced in SMA, and the depletion and refilling time constants of this pool tend to be slower [64, 172, 178]. F-actin is essential in all types of endocytosis [179] and its inhibition reduces endocytosis in neurons under high frequency stimulation [180]. Indeed, reduced SMN levels dramatically decreased endocytosis in vitro, as well as the FM1-43 uptake at NMJ level in SMA mice. Instead, PLS3 overexpression fully rescued endocytosis to similar levels as observed in controls [61]. Moreover, endocytic uptake of FM1-43 in the presynaptic terminal of NMJs upon electrical stimulation was significantly reduced in SMA and fully restored to control levels in SMA-PLS3 OE mice [61].

Chp1-associated ataxia

PLS3 is a genuine interaction partner of CHP1, a Ca2+ sensor protein [66]. CHP1 is ubiquitously expressed, but particularly abundant in neuronal tissues. CHP1 is a negative regulator of calcineurin, the most important phosphatase dephosphorylating the dephosphins involved in endocytosis [181]. Biallelic mutations in CHP1 cause autosomal recessive spastic ataxia (SPAX9; MIM 618438) in humans and mice [182,183,184]. Moreover, PLS3 overexpression in homozygous Chp1 mutant mice delays but does not ameliorate the ataxic phenotype at an early disease stage by preventing axon degeneration of Purkinje neurons [185]. PLS3 overexpression increases membrane targeting of NHE1 [185], an important binding partner of CHP1 [186] that is also associated with ataxia when mutated, at an early disease stage [187]. Thus, PLS3 overexpression has a moderate protective effect on ataxia caused by Chp1 depletion and demonstrates its potential as a cross-disease modifier [185].

SOD1-associated amyotrophic lateral sclerosis

PLS3 overexpression in a C. elegans model of ALS, carrying the dominantly inherited G85R SOD1 mutant, proved to be beneficial [68]. PLS3 increased locomotion rate, ameliorated pharyngeal pumping defects and counteracted sensitivity to paralysis by aldicarb. Contrary, on a wild-type C. elegans background, PLS3 overexpression had a detrimental effect on locomotion rate, strengthening the observation that the balance of F-actin is essential and its overload in a cell or organisms may lead to the opposite effect [68]. Since ALS and SMA share a lot of common pathological features, as well as molecular and cellular commonalities, further investigation of a beneficial impact of PLS3 on other ALS-involved genes, e.g., C9ORF72 would be very interesting.

Infection and pathogen entry

Plastins are conserved from yeast to mammals. Within all species, they fulfill functions in actin dynamics rendering rapid actin rearrangements upon pathogen challenges in the host cell while harboring distinct functions within the pathogen itself. Pathogens such as Salmonella typhimurium, Shigella flexneri or Trichomonas vaginalis enter the host cells by remodeling the actin cytoskeleton [188,189,190,191]. S. tysphimurium can even enter nonphagocytic cells like endothelial cells, upon the stimulation of CDC42 and RAC-1, which activate signaling of the pathogen protein SipA and thereby increase PLS3′s bundling activity [188]. At the contact point of host cell and pathogen, this process triggers actin rearrangements while leading to the formation of nucleation zones. PLS3 is recruited to these zones and stabilizes F-actin filaments. Arising from those zones, elongated actin filaments form protrusions surrounding the pathogen [188, 189, 191]. Nucleated actin filaments are the source of G-actin through active depolymerization to fuel the leading tips for the construction of the protrusion [191]. Finally, the pathogen is indulged in a vacuolic structure and released into the cytoplasm of the host cell. During infection with the Hepatitis C virus (HCV), PLS3 has been shown to regulate pathogen replication. Knockdown of PLS3 results in decreased viral replication and is, therefore, a potential target for HCV therapy [192].

Moreover, fimbrin facilitates active actin-cytoskeleton organizational changes to, e.g., enable amoebic movements and protrusion formation [190]. In Rickettsia, PLS3 together with profilin, capping protein, and cofilin determine the actin tail length required for motility during infection. In this species, actin bundles underlying tail formation resemble the organization of cell protrusions [193].

In the fungal pathogen, Candida albicans, the PLS3 ortholog Sac6 associates with oxidative stress responses. Sac6 negatively regulates cytosol–nucleus transport of the key transcription factor Cap1 and thereby the expression of oxidative stress response genes [194].

Biomarker

Differential expression of PLS3 is a frequent characteristic of CTCs as well as primary tumors, while usually being absent from the hematopoietic system. This renders the protein as well as the mRNA as immaculate biomarker candidates. In a recent study, a peptide library has been screened for affinity to PLS3-overexpressing cancer cells. They identify the peptide TP1 (KVKSDRVC) and develop a fluorescein isothiocyanate-labeled TP1 molecule that is able to identify PLS3-overexpressing CTCs in peripheral blood [195].

With regard to SMA, PLS3 is probably of limited use as a biomarker since it is only expressed in a minority of the human population in the hematopoietic system [5]. A study tested the expression of six putative protein biomarkers (COMP, DPP4, CLEC3B, SPP1, VTN, and AHSG) in mice with or without overexpression of a human PLS3 transgene. The expression level of PLS3 did neither affect the amount of SMN nor did the other putative biomarkers, supporting the hypothesis that PLS3 acts as an independent protective modifier in SMA [196].

One pitfall of PLS3 is that its differential expression is linked to multiple different diseases. Comorbidity of other PLS3-associated medical conditions should always be considered when an analysis of PLS3 as a biomarker in peripheral blood is performed. In addition to that, as an X-linked gene, a sex-specific expression bias towards women has been observed [5, 197].

Conclusion

Despite PLS3 not being essential for the survival of a cell or organism, both conditions overexpression or knockdown of PLS3 can influence many cellular processes either causing, facilitating or rescuing certain pathological phenotypes. While the full knockout in humans and mice causes a disruption of the bone remodeling cells and, thus, osteoporosis, its overexpression in some neurodegenerative disorders such as SMA or ALS, where F-actin levels are decreased, acts as a protective modifier. However, in wild-type condition, overexpression of PLS3 is detrimental and causes osteoarthritis or facilitates cancer. PLS3 overexpression is strongly correlated with several cancer types, and has been recommended for use as biomarker in cancer and poor prognosis for survival. PLS3 has an important function in F-actin-binding and -bundling and as such is involved in a plethora of cellular processes dependent on F-actin dynamics. These include cell migration and growth, axonal and neurite outgrowth of polarized cells, axonal local translation, endocytosis, influence of intracellular calcium on PLS3-dependent processes, mechanotransduction, signaling, infection with pathogens and others. This ever-growing knowledge on PLS3 is crucial to wisely use PLS3 overexpression or PLS3 reduction as a therapeutic target.