Control of mRNA stability contributes to low levels of nuclear poly(A) binding protein 1 (PABPN1) in skeletal muscle

Abstract

Background

The nuclear poly(A) binding protein 1 (PABPN1) is a ubiquitously expressed proteinthat plays critical roles at multiple steps in post-transcriptional regulation ofgene expression. Short expansions of the polyalanine tract in the N-terminus ofPABPN1 lead to oculopharyngeal muscular dystrophy (OPMD), which is an adult onsetdisease characterized by eyelid drooping, difficulty in swallowing, and weaknessin the proximal limb muscles. Why alanine-expanded PABPN1 leads to muscle-specificpathology is unknown. Given the general function of PABPN1 in RNA metabolism,intrinsic characteristics of skeletal muscle may make this tissue susceptible tothe effects of mutant PABPN1.

Methods

To begin to understand the muscle specificity of OPMD, we investigated thesteady-state levels of PABPN1 in different tissues of humans and mice.Additionally, we analyzed the levels of PABPN1 during muscle regeneration afterinjury in mice. Furthermore, we assessed the dynamics of PABPN1 mRNA decay inskeletal muscle compared to kidney.

Results

Here, we show that the steady-state levels of both PABPN1 mRNA and protein aredrastically lower in mouse and human skeletal muscle, particularly those impactedin OPMD, compared to other tissues. In contrast, PABPN1 levels are increasedduring muscle regeneration, suggesting a greater requirement for PABPN1 functionduring tissue repair. Further analysis indicates that modulation of PABPN1expression is likely due to post-transcriptional mechanisms acting at the level ofmRNA stability.

Conclusions

Our results demonstrate that PABPN1 steady-state levels and likely control ofexpression differ significantly in skeletal muscle as compared to other tissues,which could have important implications for understanding the muscle-specificnature of OPMD.

Background

RNA-binding proteins regulate all steps of RNA biogenesis and play a key role inpost-transcriptional regulation of gene expression [1]. Key players among these RNA-binding proteins are the poly(A)-bindingproteins, which modulate 3′-end formation of mRNA transcripts [2]. The nuclear poly (A)-binding protein 1 (PABPN1) is a ubiquitously expressedprotein in eukaryotes that binds with high affinity to polyadenosine RNA [2]. PABPN1 has critical roles at multiple steps in post-transcriptionalregulation of gene expression. The best characterized role of PABPN1 is in mRNApolyadenylation, where PABPN1 stimulates poly(A) synthesis by direct interaction withthe nascent mRNA poly(A) tail and the poly(A) polymerase [3, 4]. In a subsequent regulatory step, PABPN1 decreases poly(A) polymeraseelongation activity following addition of 250 adenine residues, thereby dictating thelength of the poly(A) tail added to mRNA transcripts [4]. Moreover, PABPN1 is also involved in regulation of alternative cleavage andpolyadenylation by suppressing weak proximal polyadenylation signals [5], which can influence both gene expression and the structure of the proteinproduced [6]. Finally, PABPN1 has been implicated in the polyadenylation-dependent pathwayof RNA decay, which targets non-protein coding genes such as the long non-coding RNAs(lncRNAs) [7]. Thus, PABPN1 modulates a number of processes that are critical forcontrolling gene expression.

In addition to playing a key role in RNA metabolism, PABPN1 is of significant clinicalinterest as mutations in the PABPN1 gene lead to oculopharyngeal musculardystrophy (OPMD) [8]. This disease is caused by a small GCN trinucleotide expansion in the codingregion of PABPN1, resulting in the expansion of a stretch of 10 alanines to 12to 17 alanines in the PABPN1 N-terminus. OPMD is a late onset, autosomal dominantdisease characterized primarily by progressive eyelid drooping (ptosis) and difficultiesin swallowing [9]. Additional weakness is noted in proximal limb, facial and other extraocularmuscles [10–12]. Disease progression is variable between patients and complications includechoking, regurgitation, aspiration and pneumonia. The pathologic hallmark of the diseaseis the presence of nuclear aggregates of PABPN1 in muscle [13, 14]. Given the ubiquitous expression and general function of PABPN1 in RNAmetabolism [15], how mutations of this post-transcriptional regulatory factor cause amuscle-specific disease is unclear. PABPN1 is essential for both mRNA biogenesis as wellas proliferation and differentiation of myogenic precursor cells, suggesting a criticalrole in muscle regeneration and maintenance [16]. Skeletal muscle is highly specialized for contraction and has uniquecharacteristics compared to other tissues, such as being highly regenerative andcomprised of multinucleated post-mitotic cells, which suggests that intrinsiccharacteristics of this tissue may make it more vulnerable to the effects of mutantPABPN1 than other tissues that are not affected in OPMD.

To begin to understand the muscle specificity of OPMD, we investigated the steady-statelevels of the PABPN1 protein in different tissues. We find that the steady-state levelsof PABPN1 are drastically lower in skeletal muscle compared to other tissues.Strikingly, craniofacial muscles, which are affected in OPMD, show the lowest levels ofPABPN1. We also found that PABPN1 is upregulated during muscle repair after injury.Studies of mRNA stability indicate that regulation of PABPN1 expression is likely due todistinct post-transcriptional mechanisms in different tissues. Taken together, ourresults demonstrate that PABPN1 steady-state levels and likely control of expressiondiffer significantly in skeletal muscle as compared to other tissues, which could haveimportant implications for understanding the muscle-specific nature of OPMD.

Methods

Animals and primary muscle cell culture

Experiments involving animals were performed in accordance with approved guidelinesand ethical approval from Emory University’s Institutional Animal Care and UseCommittee. Adult male C57BL/6 mice between 2 to 6 months of age were used inexperiments. To induce regeneration, gastrocnemius muscles of male mice were injectedwith 40 μl of 1.2% BaCl₂[17] and collected 2, 5 and 14 days after injury. Primary myoblasts werederived from the hindlimb muscles of mice and cultured to >99% purity as previouslydescribed [18]. Cells were maintained in growth media (GM: Ham’s F10, 20% fetalbovine serum, 5 ng/ml basic fibroblast growth factor (bFGF), 100 U/ml penicillin G,100 mg/ml streptomycin) in a humidified 5% CO₂ incubator at 37°C oncollagen-coated dishes. For histologic analyses, serial 10 μm sections werecollected along the length of the muscle and stained with hematoxylin and eosin.Images were obtained using Axiovert 200 M microscope with a 0.30 NA 10× or20× Plan-Neofluar objective (Carl Zeiss MicroImaging, Inc., Oberkochen, Germany)and camera (QImaging, Surrey, Canada) with OpenLab 5.5.2 (PerkinElmer, Waltham,MA).

Immunoblot analysis

Tissues were homogenized in radioimmunoprecipitation assay (RIPA)-2 buffer (50 mMTris-HCl pH 8.0, 150 mM NaCl, 1% NP-40, 0.5% deoxycholic acid, 0.1% SDS) withprotease inhibitors (Complete Mini Tablets, Roche, Pleasanton, CA). Equal amounts oftotal protein were resolved by SDS-PAGE, transferred to nitrocellulose and thedesired protein was detected by immunoblotting with appropriate antibodies andenhanced chemiluminescence. PABPN1 levels in human tissues were analyzed usingINSTA-blot (IMGENEX, San Diego, CA) membranes. The primary antibodies andconcentrations used were as follows: anti-PABPN1 antibody 1:5,000 [16], anti-Histone H3 1:1,000 (Abcam, Cambridge, MA), anti-heat shock protein90 (HSP90) 1:1,000 (Santa Cruz Biotechnology, Dallas, TX) and anti-tubulin 1:5,000(Sigma-Aldrich, St. Louis, MO). Anti-mouse or anti-rabbit IgG 1:5,000 (JacksonImmunoResearch, West Grove, PA) were used as secondary antibodies.

Northern blot analysis

Total RNA from tissues, primary cell cultures and fluorescence-activated cell sorting(FACS)-sorted cells was isolated using TRIzol (Invitrogen, Carlsbad, CA) according tothe manufacturer’s protocol. Northern blotting was performed as describedpreviously by Ausubel et al. [19]. DNA probes were generated by polymerase chain reaction (PCR) using customprimers (PABPN1-F 5′-CCCAGGCAATGCTGGCCCAGTGATCATGTCTC-3′ and PABPN1-R5′-CTAGCCCGGCCCCTGTAGATTCGACCCCGGGGC-3′, c-Myc-F5′-GAACTTCACCAACAGGAACTATGACCTCG-3′ and c-Myc-R5′-GGTGTCTCCTCATGCAGCACTAGG-3′), primers from SA Biosciences, Valencia,CA (peroxisome proliferator-activated receptor gamma coactivator 1α(PGC1α); PPM03360E, glyceraldehyde 3-phosphate dehydrogenase (GAPDH); PPM02946E)or by Ambion, Austin, TX (QuantumRNA Classic II 18S). PCR products were labeled with[α-³²P]dCTP using a random primer DNA labeling system (Invitrogen,Carlsbad, CA).

FACS

Mononucleated cells were enzymatically isolated from gastrocnemius muscles 3 daysafter BaCl₂ injury and fluorescently labeled with antibodies to CD31 andCD45 (PE), Sca-1 (PE-Cy7), and alpha-7-integrin (AlexaFluor 649). Propidium iodidestaining was used to gate out dead cells from the sort. Myoblasts(CD31^-/CD45^-/Sca-1^-/alpha-7-integrin⁺/PI^-)and non-myogenic cells(CD31⁺/CD45⁺/Sca-1⁺/alpha-7-integrin^-/PI^-)were collected using a FACSAria II (Becton-Dickinson, Franklin Lakes, NJ). Isolatedcells were then processed for RNA extraction.

Quantitative reverse transcription (RT)-PCR

cDNA synthesis from 100 ng RNA was performed using M-MLV reverse transcriptase(Invitrogen, Carlsbad, CA). mRNA levels were determined by real-time PCR using the iQSYBR Green (Bio-Rad, Hercules, CA) and iCycler iQ Real-Time Detection System andsoftware (Bio-Rad, Hercules, CA). The relative levels of PABPN1 were determined bythe ΔΔCt method and normalized to the housekeeping gene HPRT1. Primers werefrom SA Bioscences, Valencia, CA (PABPN1: PPM25445A, HPRT1: PPM03559E).

mRNA decay

To analyze mRNA stability in vivo, mice were injected intraperitoneally withactinomycin D (Sigma-Aldrich, St. Louis, MO) at 2.5 μg/g and quadriceps musclesand kidney were collected 1, 2, 4 and 6 h later. To measure mRNA stability in primarymyoblasts in vitro, 5 μg/ml actinomycin D was added to the growthmedium and cells were harvested 0.5, 1, 2 and 4 h later. Total RNA was extracted fromtissues or cells and analyzed by northern blot and half-lives were determined bydensitometry.

5′ and 3′ RACE

In order to determine the 5′ and 3′UTRs of PABPN1 transcripts, we usedthe 5′ and 3′ rapid amplification of cDNA ends (RACE) system (Invitrogen,Carlsbad, CA), respectively. Total RNA from either muscle or testis was used as atemplate according to the manufacturer’s instructions. PCR products were clonedinto the pCR2.1 vector (TOPO TA cloning, Invitrogen, Carlsbad, CA) and sequenced byBeckman Coulter Genomics, Danvers, MA.

Statistical analysis

Statistical analysis to determine significance between two groups was performed usinga Student’s t test. One-way analysis of variance (ANOVA) was used forcomparisons between multiple groups as appropriate. All statistical analyses wereperformed using GraphPad Prism 5.0 for Macintosh (GraphPad Software). Differenceswere considered to be statistically significant at P <0.05.

Results

PABPN1 levels are lower in skeletal muscle compared to other tissues

A better understanding of the mechanisms that underlie OPMD pathology can be obtainedby analyzing the function of PABPN1 in skeletal muscle. To begin to identifymuscle-specific properties of PABPN1, we first examined the expression of PABPN1across different tissues. Immunoblot analysis revealed that PABPN1 steady-statelevels vary significantly among mouse tissues, with skeletal muscle displaying thelowest levels of PABPN1 (Figure  1A). The low abundance ofPABPN1 in skeletal muscle could result from skewed misrepresentation of this proteinwithin the protein pool by the uniquely high levels of cytoplasmic proteinscomprising the contractile machinery in this tissue. However, relatively similarlevels of both the nuclear protein histone H3 [20] and the cytoplasmic protein HSP90 [21] were observed between muscle and other tissues, suggesting that thenuclear protein fraction is not under-represented in muscle. Furthermore, analysis ofPABPN1 levels among different mouse muscles revealed even lower levels of thisprotein in the craniofacial muscles (masseter, tongue and pharynx), some of which aremuscles primarily affected in OPMD patients [12], compared to other muscles of the body (Figure  1B). Significantly lower levels of PABPN1 in muscle as compared to othertissues were also observed in human samples (Figure  1C),suggesting that the low levels of this protein in muscle are not species-specificfindings, and this may have physiologic implications for humans.

Figure 1

Nuclear poly(A) binding protein 1 (PABPN1) levels are low in all skeletalmuscles. Lysates prepared from different (A) mouse tissues (50μg of total protein per lane), (B) mouse muscles (150 μg oftotal protein per lane) or (C) human tissues (20 μg of totalprotein per lane) were immunoblotted with anti-PABPN1 antibody. Histone H3 andheat shock protein 90 (HSP90) were used as loading controls for mouse samples.Amido black staining was used as the loading control for human samples.Immunoblots are representative of at least three independent sets oftissues.

To examine whether the expression of PABPN1 is regulated at the protein or RNA levelwe performed northern blot analysis (Figure  2). Thisanalysis revealed a strong correlation between the low levels of PABPN1 protein andthe low abundance of PABPN1 transcript in mouse skeletal muscle (Figure  2B), suggesting that control of PABPN1 expression occurs at theRNA level, either by transcriptional or post-transcriptional means. As previouslyreported, PABPN1 has two major mRNA variants, a 2.1 kb and a 1.4 kb transcript(Figure  2) [22, 23]. The 2.1 kb transcript, which was detected in all tissues but was presentat low levels in muscle (Figure  2B), utilizes a distalpolyadenylation site 851 bp downstream of the stop codon (Figure  2A) [23]. The 1.4 kb represents the transcript that uses a proximal polyadenylationsite 66 bp downstream of the stop codon (Figure  2A) [23]. This 1.4 kb mRNA variant was the predominant transcript only in testis,but was also found in other tissues at much smaller amounts (Figure  2B). Interestingly, the levels of the 1.4 kb PABPN1 transcriptwere very high in testis, which correlates with the very high levels of PABPN1protein observed in this tissue (Additional file 1: FigureS1). Furthermore, with the exception of testis, no significant variation in the ratiobetween the two mRNA isoforms was observed in the analyzed tissues. Northern blottingalso revealed a band of approximately 3.6 kb that was observed in all tissues (datanot shown). This transcript was previously reported and suggested to be either atranscript of a related gene [22] or generated by a distinct PABPN1 promoter [23]. We performed both 5′RACE and 3′RACE from kidney, muscle andtestis but failed to identify any novel PABPN1 transcript other than the 1.4 kb and2.1 kb mRNAs variants, suggesting the 3.6 kb band might indeed represent a transcriptfrom a related gene. Together, our results from immunoblotting and northern blottingreveal low steady-state levels of PABPN1 mRNA and protein in skeletal muscle, whichis indicative of either a decrease in PABPN1 transcription or altered mRNA stabilityin this tissue.

Figure 2

Nuclear poly(A) binding protein 1 (PABPN1) mRNA levels are low in skeletalmuscle. (A) Structure of the 2.1 kb and 1.4 kb PABPN1 transcripts (solidboxes represent coding regions and open boxes non-coding regions). (B)Total RNA from different mouse tissues was analyzed by northern blot using aPABPN1 probe. Two different exposures, short and long, are shown. 18S rRNA wasprobed as a loading control. Figure is representative of at least threeindependent sets of tissues.

PABPN1 levels are increased during muscle regeneration

Adult skeletal muscle is comprised primarily of post-mitotic myofibers, however, itis a highly regenerative tissue that undergoes extensive repair after injury(Figure  3A) [24]. In the earliest phases of muscle regeneration, inflammatory cells invadethe tissue to remove dead tissue. Subsequently, large numbers of proliferativemyoblasts derived from resident stem cells undergo differentiation and fusion to formnew myofibers. Although PABPN1 levels are very low in adult muscle tissue, levelswere significantly increased during the period of extensive cellular proliferation,differentiation and fusion that occurs 2 to 5 days after muscle injury (Figure 3B). However, 14 days after injury, when muscle architecturewas restored (Figure  3A), PABPN1 levels were again low(Figure  3B). A similar pattern of upregulation wasobserved for the cytoplasmic poly(A) binding protein, PABPC1 [2], during muscle regeneration, however the levels of the heat shock proteinHSP90 remained constant over the time course. This result indicates that increasedlevels of poly(A)-binding proteins during muscle regeneration are not unique toPABPN1. To determine whether the increased levels of PABPN1 observed at 2 to 5 daysafter injury were due in part to myoblasts, we used flow cytometry and specificantibodies to isolate myoblasts and non-myogenic cells (including inflammatory cells)from mouse muscles 3 days after injury. As we were unable to perform immunoblots forPABPN1 on the small amount of cells isolated by flow cytometry, we used quantitativeRT-PCR to examine PABPN1 transcript levels in sorted cells compared to muscle tissue.Similar to what we observed for PABPN1 protein, PABPN1 transcript levels wereincreased approximately fivefold in injured compared to uninjured muscle(Figure  3C). We also found that PABPN1 mRNA levels wereexceptionally high in both sorted myoblasts(CD31^-/CD45^-/Sca-1^-/alpha-7-integrin⁺)and non-myogenic cells(CD31⁺/CD45⁺/Sca-1⁺/alpha-7-integrin^-)compared to uninjured muscle tissue (Figure  3C). Thesedata suggest that myoblasts significantly contribute to increased PABPN1 levels inregenerating muscle. We conclude that PABPN1 levels are not static in muscle butrather modulated by the physiologic state of the tissue, suggesting a greaterrequirement for PABPN1 function during tissue repair.

Figure 3

Nuclear poly(A) binding protein 1 (PABPN1) levels are increased duringmuscle regeneration in part due to increased levels in myoblasts. (A)Representative hematoxylin and eosin stained sections of gastrocnemius musclesat different times after BaCl₂ injury are shown. (B) Lysateswere prepared from gastrocnemius muscles at different times after injury andwere immunoblotted with anti-PABPN1, PABPC1 or heat shock protein 90 (HSP90)antibodies (n = 3 per timepoint). (C) Total RNA was obtained fromuninjured and injured muscle tissue (three independent samples) as well asfluorescence-activated cell sorting (FACS)-sorted myoblasts and non-myogeniccells obtained 3 days after muscle injury (pooled from five mice). PABPN1 mRNAlevels were determined using real-time polymerase chain reaction (PCR) andhypoxanthine-guanine phosphoribosyltransferase (HPRT) mRNA was used as aninternal control. Amount of PABPN1 mRNA relative to uninjured muscle is shown;n = 3. Data are mean ± SD; *P <0.05 vs uninjured muscle.

PABPN1 mRNA is unstable in skeletal muscle

The two PABPN1 transcripts schematized in Figure  2A arisefrom the usage of two different polyadenylation sites. The smaller 1.4 kb transcriptcontains virtually no 3′UTR, whereas the longer 2.1 kb transcript harbors a3′UTR of 851 bp containing a putative mRNA AU-rich destabilizing element (ARE) [23]. We hypothesized that the 2.1 kb transcript, which is the most abundanttranscript in all tissues but testis and contains a putative ARE, may be subject topost-transcriptional regulation in different tissues. To assess if this transcript isdifferently regulated in muscle, we analyzed PABPN1 mRNA stability in muscle andkidney after blocking transcription in mice with actinomycin D. We observed that thehalf-life of the 2.1 kb PABPN1 transcript was significantly shorter in muscle (2.3 h)compared to kidney (>6 h) (Figure  4A,E; Table  1). As a control to demonstrate similar transcriptional inhibitionbetween both tissues, we analyzed the stability of PGC1α and GAPDH mRNAs, knownunstable and stable transcripts, respectively [25, 26]. As expected, PGC1α mRNA displayed a short half-life in both muscleand kidney (1.7 h and 2.6 h) compared to GAPDH mRNA (>>6 h in both tissues)(Figure  4A,E; Table  1). Theseresults demonstrate a strong correlation between the high steady-state levels ofPABPN1 protein and the stable transcript in kidney, whereas in skeletal muscle, thelow steady-state levels of PABPN1 correlate with the unstable PABPN1 transcript.

Figure 4

Nuclear poly(A) binding protein 1 (PABPN1) mRNA is unstable in muscle tissuebut stable in cultured myoblasts. (A) Total RNA was collected atdifferent timepoints after injection of actinomycin D to inhibit transcriptionand PABPN1 mRNA decay was analyzed by northern blot. Time courses are shown forsamples from muscle and kidney. Peroxisome proliferator-activated receptorgamma coactivator 1α (PGC1α) and glyceraldehyde 3-phosphatedehydrogenase (GAPDH), a known unstable and stable transcript, respectively,were probed as controls (n = 3 per timepoint). To visualize PABPN1 signal inmuscle samples, the blot was exposed significantly longer than for kidneysamples. (B) Total RNA was obtained from skeletal muscle (SM) andcultured primary mouse myoblasts (Mb) and PABPN1 mRNA levels were determinedusing real-time polymerase chain reaction (PCR) and hypoxanthine-guaninephosphoribosyltransferase (HPRT) mRNA was used as an internal control. Theamount of PABPN1 mRNA relative to skeletal muscle (SM) is shown; n = 3independent samples. Data are mean ± SD; *P <0.05 vs skeletalmuscle. (C) Protein extracts were prepared from SM and Mb andimmunoblotted with anti-PABPN1 antibody. HSP90 was used as a loading control.The immunoblot is representative of at least three independent samples.(D) Total RNA was collected at different timepoints after treatmentof cultured primary mouse myoblasts with actinomycin D and PABPN1 mRNA decaywas analyzed by northern blot; c-Myc and GAPDH, known unstable and stabletranscripts in myoblasts, respectively, were probed as controls. Averages ofdensitometric measurements of northern blot bands were used to determine mRNAdecay. The image is representative of at least three independent samples.(E) The decay profile of PABPN1 mRNA in muscle, kidney and culturedmyoblasts plotted as mRNA amount relative to timepoint T = 0 h (n = 3 samplesper timepoint). Data are mean ± SD.

Table 1 Nuclear poly(A) binding protein 1 (PABPN1) mRNA is unstable in muscle tissuebut stable in cultured myoblasts

As shown earlier (Figure  3C), PABPN1 levels are modulatedduring muscle regeneration and myoblasts contribute in part to the increased levelsof PABPN1 during this process. We next investigated if the increase in PABPN1 levelsin myoblasts is accompanied by a corresponding increase in PABPN1 mRNA stability.Similar to myoblasts directly isolated from injured muscle, we found that culturedprimary mouse myoblasts displayed higher levels of PABPN1 mRNA compared to uninjuredmuscle tissue (Figure  4B). Consistent with this finding,steady-state levels of PABPN1 protein were also higher in cultured myoblasts comparedto muscle tissue (Figure  4C). We analyzed the stabilityof PABPN1 transcripts in cultured myoblasts. As observed in skeletal muscle tissue,the 2.1 kb PABPN1 transcript was the predominant transcript in myoblasts (data notshown). However, in contrast to muscle tissue, the 2.1 kb PABPN1 transcript wasextremely stable in myoblasts (Figure  4D,E; Table 1). As expected, c-Myc mRNA, a known unstable transcript inmyoblasts [27], had a short half-life compared to the much longer half-life for GAPDHmRNA, a known stable transcript (Table  1). These resultsindicate that the high levels of PABPN1 in cultured myoblasts, and likely inmyoblasts present during muscle regeneration, is due at least in part to the increasein PABPN1 mRNA stability in those cells compared to muscle tissue. Taken together,our results suggest that PABPN1 expression in different tissues or during muscleregeneration is regulated by a post-transcriptional mechanism that modulatestranscript stability.

Discussion

Studying PABPN1 specifically in skeletal muscle is critical for defining the mechanismswhich make this tissue uniquely susceptible to the mutation causing OPMD. Here, wereport that steady-state levels of PABPN1 mRNA and protein are low in skeletal muscleand that expression of PABPN1 in this tissue is controlled, at least in part, bypost-transcriptional regulation of RNA levels. We also demonstrate that PABPN1 levelsare modulated during muscle repair providing further support for regulation of PABPN1expression in this tissue.

PABPN1 is not the only ubiquitous protein with a general function in basic cellularprocesses whose expression level is variable among tissues [28, 29]. For example, the expression of histone H3A, transcription elongation factorA1 (TCEA1) and heterogeneous nuclear ribonucleoprotein (hnRNP) C is relatively constantamong different tissues, whereas levels of GAPDH, β-actin and histone H2A are amongthe most variable within tissues [29]. These differences in expression levels are most likely related to intrinsicproperties of individual tissues and reflect differences in metabolic activity andcellular structure.

The extremely low levels of PABPN1 in skeletal muscle compared to other tissues mayindicate a low requirement for this factor in basal muscle metabolism and maintenance.Skeletal muscle is distinctly characterized by multinucleated, post-mitotic cells with avery specialized function and low complexity transcriptome [30, 31]. In skeletal muscle, a small number of genes contribute to a large fractionof the total mRNA pool, with the ten most expressed genes in muscle accounting for 20%to 40% of the total mRNA [31]. The most abundant transcripts in skeletal muscle encode proteins involved incontraction, glucose metabolism, ATP production and ribosomal proteins [30, 31], consistent with the role of this tissue in movement and metabolism. Suchtranscripts encoding proteins involved in general cellular functions are usually stablewith low turnover [27, 32]. Therefore, the low complexity of the skeletal muscle transcriptomeassociated with low turnover of a significant fraction of its transcripts may explainwhy skeletal muscle has low requirements for a protein involved in mRNA metabolism suchas PABPN1.

Our data indicate the low levels of PABPN1 in skeletal muscle are, at least in part,determined at the level of regulation of PABPN1 transcript stability. Regulation of mRNAdecay rate is a key factor in determining the expression pattern of many genes allowingrapid adaptation to changing cellular requirements [33, 34]. PABPN1 levels increase significantly during skeletal muscle regenerationsuggesting a greater requirement for PABPN1 in myoblasts and non-myogenic cells such asinflammatory cells, which may be due to their highly proliferative status and to a morecomplex transcriptome compared to uninjured muscle tissue. As the increased levels ofPABPN1 in regenerating muscle correlate with an increased transcript stability inmyoblasts and subsequent increase in the steady-state levels of PABPN1 transcript, wesuggest that skeletal muscle employs a post-transcriptional mechanism to control PABPN1levels according to the tissue requirements.

mRNA decay rates are modulated by an interplay of specific stabilizing or destabilizingfactors with the transcript, such as RNA-binding proteins and/or miRNAs and theirassociated enzymes [35]. One of the most studied post-transcriptional pathways is orchestrated by avariety of RNA-binding proteins that interact with AU-rich elements (ARE) within the3′UTR of mRNAs [33, 36] and many unstable mRNAs expressed in muscle contain AU-rich elements in their3′UTRs [27]. As the main 2.1 kb PABPN1 transcript expressed in skeletal muscle harbors anARE in the 3′UTR, we speculate this pathway is a strong candidate for the controlof PABPN1 levels in skeletal muscle.

The specific PABPN1 expression pattern observed in skeletal muscle may be an importantfeature that makes this tissue more susceptible than others to the mutations in PABPN1that cause the muscle-specific disease OPMD. Whether the alanine expansion in PABPN1leads to a gain-of function or loss-of-function of this protein is unknown [12, 15]. The nuclear aggregates observed in muscle of OPMD patients may exert toxiceffects in the tissue as hypothesized for other polyglutamine and polyalanine expansiondisorders [37–40]. However, as wild-type PABPN1 can form reversible aggregates in neurons inresponse to changes in cell physiology without overt pathology [41], the toxicity of PABPN1 nuclear aggregates is unlikely to be the exclusivecause of OPMD etiology. In the loss-of-function model of OPMD etiology, one mechanismthat could lead to a loss or decrease of PABPN1 function is an intrinsic reduction inPABPN1 activity caused by the alanine expansion. Although wild-type and mutant PABPN1appear to have similar polyadenylation activity in vitro[2], the effects of the alanine expansion on this or other PABPN1 functions havenot yet been addressed in the context of skeletal muscle in vivo. Anothermechanism that could explain a loss of PABPN1 function is the depletion of the solubleand functional fraction of PABPN1 by sequestration in the nuclear aggregates present inmuscle of OPMD patients. In fact, a recent study that examined PABPN1 transcript levelsin human muscle samples reported a decrease in steady-state levels of PABPN1 mRNA afterthe fifth decade of life, the common age for onset of OPMD symptoms [42]. This study also presented evidence that this decrease in PABPN1 transcriptlevels is accelerated in OPMD patients [42]. These recent findings support the idea that a loss of PABPN1 function couldcontribute to the muscle-specific pathology in OPMD. Consistent with this idea,overexpression of wild-type PABPN1 reduces the pathology caused by the expression ofalanine-expanded PABPN1 in both cell and mouse models of OPMD [43]. In these loss-of-function scenarios, the low level of PABPN1 in the skeletalmuscle may make this tissue specifically more vulnerable to a further decrease in thetotal amount of functional PABPN1 available as illustrated in the schematic inFigure  5. The resulting lowered levels of PABPN1 may then bebelow the threshold amount of PABPN1 activity that is required for proper tissuemaintenance leading to pathology.

Figure 5

Low levels of nuclear poly(A) binding protein 1 (PABPN1) in skeletal muscle maypredispose this tissue to the deleterious effects of alanine-expandedPABPN1. We show muscle has lower levels of PABPN1 compared to other tissuesin normal individuals (N) but these levels are adequate for normal tissuefunction. In patients with oculopharyngeal muscular dystrophy (OPMD), functionallevels of PABPN1 could be decreased in all tissues due to expression of mutantPABPN1. However, muscle-specific pathology ensues in autosomal dominant OPMDbecause the levels of PABPN1 fall below the threshold required to maintain propertissue function.

Conclusions

Our results demonstrate that PABPN1 steady-state levels and likely control of expressiondiffer significantly in skeletal muscle as compared to other tissues, which could haveimportant implications for understanding the muscle-specific nature of OPMD. We suggestthe low levels of PABPN1 observed in skeletal muscle may be an important aspect of thistissue that underlies OPMD pathology. Further studies are necessary to better comprehendthe mechanisms and implications of the regulation of PABPN1 expression in skeletalmuscle, which could open avenues for potential therapeutic approaches for OPMD.