The NORAD-Pumilio regulatory axis in the evolution of inclusion body myositis

Abstract

Sporadic inclusion body myopathy (sIBM) is the most common progressive degenerative skeletal muscle disease of older people with poorly understood pathogenesis. We assessed the involvement and impact of the noncoding RNA activated by DNA damage (NORAD) and its binding partner Pumilio (PUM) proteins on the evolution of the disease process in muscle biopsies from sIBM patients with mild or severe histopathologic changes. A disease severity–dependent increase in NORAD and PUM2 expression was present along with subsarcolemmal formations of NORAD-PUM (NP) bodies in the sIBM muscles. Similarly, Norad, Pum1, and Pum2 expression levels were upregulated as disease severity increased with age in VCP-A232E mice, a model for hereditary IBM. Crossbreeding VCP-A232E mice with Norad^−/− mice leading to generation of VCP mutants with either single copy or the absence of Norad showed that Norad suppresses disease progression in the VCP-A232E model in a dose-dependent manner. Norad expression in the muscle from Pum2 overexpressing mice revealed significantly increased Norad transcripts in response to high Pum2 expression and the presence of NP bodies coalescing around muscle nuclei. Collectively, these findings provide strong evidence that elevated NORAD expression is a protective, compensatory response to disease-related induction of PUM2 activity in muscle.

Introduction

Sporadic inclusion body myositis (sIBM) is the most common acquired myopathy after the age of 50 with a prevalence of 180 per million adults.^1,2 The mean age at onset of symptoms is approximately 60 years with a range from the fourth to the ninth decade. It is characterized by progressive asymmetric weakness predominantly affecting the quadriceps and/or finger flexors. Loss of ambulation, hand grip, and dysphagia are major complications of the disease.³ The characteristic pathologies of sIBM muscle fibers are autophagic rimmed vacuoles, accumulation of ubiquitin-positive multiprotein aggregates containing misfolded proteins in the β-pleated sheet conformation of amyloid,⁴ and mononuclear cell inflammation around intact-appearing fibers.^5,6 In addition, abnormalities in the content and distribution of mitochondria, abundant ragged blue fibers with succinic dehydrogenase enzyme (SDH) histochemistry and cytochrome C oxidase (COX)-negative fibers are commonly seen.^7–9 Moreover, the mtDNA copy number in IBM muscle is reduced significantly compared to age-matched controls and deep sequencing and quantitation of mtDNA variants revealed that IBM muscles had markedly increased levels of large deletions and duplications and increased somatic single nucleotide variants.^10–12

The pathogenesis of sIBM is complex and remains poorly understood. It consists of a relationship between degenerative and inflammatory pathways that can be seen in aging.^13–18 Antigen-driven, clonally restricted, cytotoxic T cells are the main feature of the inflammatory component surrounding intact appearing muscle fibers. Abnormal protein homeostasis with protein misfolding, aggregation, and dysfunctional autophagy represent the degenerative component.^13–20 The role of a coexisting inflammatory component in the setting of these cellular abnormalities has been under long debate. With the recognition of anti-cytosolic 5′ nucleotidase 1A autoantibodies, which can be detected in 30%-60% of patients,²¹ the autoimmune feature of sIBM pathogenesis has become more substantial. It has been hypothesized that the inflammation is secondary to myodegenerative changes with an immune response recognizing modified proteins as foreign.²² It is not yet known whether inflammation may induce or aggravate the observed degeneration in sIBM, however. Nevertheless, T-cell depletion in a xenograft model of sIBM was shown not to alter muscle degenerative pathology, with persistence of rimmed vacuoles, mitochondrial pathology, and nuclear clearance and cytoplasmic aggregation of TAR DNA-binding protein 43 (TDP-43).²³ Currently, sIBM remains refractory to immunosuppressive treatment^17,24; therefore, a better understanding of the pathobiology and evolution of the disease process should lead to the identification of novel therapeutic targets.

In the past few years, long noncoding RNAs (lncRNAs) have received considerable attention because of their emerging biological functions and complex molecular mechanisms. This has stimulated interest in uncovering their roles in disease pathogenesis. Long intergenic noncoding RNA LINC000657 caught the attention of Lee et al because of its unusual high conservation and abundance (>300 copies per cell).²⁵ They showed that this noncoding RNA activated by DNA damage (NORAD) is required for maintaining chromosomal stability and proper mitotic divisions in human cells.²⁵ Furthermore, in a collaborative study, NORAD deletion in mice was found to result in genomic instability and mitochondrial dysfunction, leading to a multisystem degenerative phenotype resembling premature aging.²⁶ In mouse tissues, Norad is the preferred target of Pumilio RNA-binding proteins. Upon loss of Norad, Pumilio hyperactively represses key genes required for mitosis and mitochondrial function.^25,26 Moreover, NORAD is able to out-compete thousands of other transcripts to bind and regulate the activity of PUM proteins by nucleating the formation of liquid-like cytoplasmic ribonuclear protein (RNP) condensates, termed NORAD-PUM (NP) bodies.²⁷ PUM proteins bind an extensive network of mRNAs and repress protein expression by inhibiting translation and promoting mRNA decay. Therefore, PUM1/2 regulatory networks appear to be involved in diverse cellular processes and molecular pathways.²⁸

In this study, we focused on investigating the involvement and impact of the NORAD-PUM1/2 regulatory axis on the evolution of the disease process in muscle biopsies from sIBM patients with mild or severe histopathologic changes. We observed a disease severity-dependent increase of NORAD transcripts along with increased PUM2 expression. In addition, we asked 2 questions: (1) is NORAD upregulation a compensatory response to impaired cellular function, or detrimental to the segmental viability of muscle? and (2) does NORAD regulate PUM1/2 activity through NP body formation in IBM muscle? Answers to these questions are the bases for this report.

Methods

Histopathological analysis

Muscle biopsy specimens

In total, 39 biopsy specimens from histopathological diagnosis of IBM, dermatomyositis (DM), vacuolar myopathy (VM), and specimens with no pathologic diagnosis as controls were available for molecular studies following deidentification protocol. These were archived biopsy specimens, obtained between 2015 and 2023. All were mounted in gum tragacanth and initially frozen in isopentane cooled in liquid nitrogen. A standard battery of stains included hematoxylin and eosin (H&E), modified Gomori trichrome, succinic dehydrogenase, NADH dehydrogenase, Cytochrome C oxidase, ATPase (pH 4.2, 4.6, and 9.4), and Congo red. For each case, clinical information was available in the clinical history section of biopsy reports to establish age at onset, initial distribution of symptoms, pattern of muscle involvement, ambulatory status, disease progression, and serum creatine kinase (CK) levels. All severe IBM biopsies met the histopathological diagnostic criteria,⁵ confirming the clinical diagnosis of suspected IBM. The components of diagnostic criteria were the presence of muscle fibers with rimmed vacuoles, congophilic inclusions, and multifocal endomysial inflammation surrounding or invading intact-appearing muscle fibers. The IBM biopsies were divided into 2 subgroups as “mild” and “severe” based on the occurrence of muscle fibers with rimmed vacuoles and those fibers illustrating abnormalities in the content and distribution of mitochondria with oxidative enzyme histochemistry as ragged blue fibers and COX negative/severely deficient fibers. Biopsies displaying “rare” or “a few” muscle fibers (<4 per section) with rimmed vacuoles along with rare/a few ragged blue fibers or COX negative fibers were grouped as mild IBM (with probable IBM clinical diagnosis), which represented the patient population with mean age of 61.2 ± 10.6 (n = 6, all females). Severe IBM biopsies representing more advanced cases showed numerous fibers with rimmed vacuoles and numerous ragged blue and COX-negative fibers. Connective tissue increase with fat replacement was more prominent in this group from a patient population with mean age of 68.2 ± 4.8 (n = 9; 7 males, 2 females) at the time of biopsy. Variable degrees of multifocal endomysial inflammation surrounding or invading intact fibers were present in all cases without distinguishing features between groups. Age and sex comparable control specimens (66.9 ± 4.2, n = 7; 5 males, 2 females) with no pathologic diagnosis were identified, based on the clinical history of myalgias, muscle cramps or suspected myositis with no CK elevation or myopathic changes by electrophysiology. As disease controls, the DM group included biopsies having small regenerating fibers with typical perifascicular localization and variable extent of perivascular inflammation; this patient population had a mean age of 58.0 ± 16.4 (n = 10; 6 males, 4 females). The common histopathologic findings in the VM group with mean age of 51.5 ± 9.4 (n = 5; 3 males, 2 females) included numerous fibers with prominent rimmed vacuoles, congophilic inclusions, increase of connective tissue with fat replacement in focal areas with marked fiber size variability in the absence of or with minimal inflammatory changes. One biopsy showed fibers containing amorphous material accumulations as well as internal cytoarchitecture disorganization, compatible with protein aggregate/myofibrillary myopathy; in another biopsy, clusters of large fibers with prominent structural changes displaying numerous cytoplasmic bodies were present as distinguishing features. In the remaining 3 cases, the genetic test results were available in 2: presence of a copy gain in the CHCHD10 gene and a point mutation in VCP gene (R155H).

Animals

VCP-A232E mice, referred to as Vv mice in this article, were purchased from The Jackson Laboratory (Strain: 034127). Since the transgene is X-linked, only male mice were used throughout the study to avoid the inconsistencies that may be observed in the hemizygous female mice due to X-inactivation. This model carries human VCP cDNA with A232E mutation. The Norad^−/− mouse model previously generated.²⁶ VCP-A232E and Norad^−/− mice were used to generate the mouse models that express human VCP cDNA with A232E mutation along with either heterozygous or homozygous deletion of Norad gene leading to 2 different models that we refer as VvNn (VCP-A232E; Norad^+/−) and Vvnn (VCP-A232E; Norad^−/−). Genotyping protocol by The Jackson Laboratory or by JT Mendell Lab were used to confirm the genotypes of VCP-A232E and Norad^−/−, respectively. C57BL/6 mice were used as WT controls. Overdosage of xylazine/ketamine was used to euthanize the mice; quadriceps, tibialis anterior, and gastrocnemius muscles were collected from the 2- to 3-month-old mice (young cohort) and 10- to 12-month-old mice (old cohort). All animal experiments were performed according to the guidelines approved by The Abigail Wexner Research Institute at Nationwide Children’s Hospital Animal Care and Use Committee, which operates in full accordance with the Animal Welfare Act and the Health Research Extension Act (IACUC approval number: AR18-00123). The Pum2; rtTA3 mouse model, which has enforced Pum2 expression was generated in the JT Mendell Lab.²⁶ Archived gastrocnemius muscle tissue blocks from this model were used for expression analysis and immunohistological detection of NP bodies.

Histological analysis of skeletal muscle

Routine histopathological stain H&E, and stain for mitochondrial enzyme histochemistry for SDH were performed in fresh-frozen sections from mouse skeletal muscle using established protocols. Three images representing deep, intermediate, and superficial zones of the muscles were taken at 20× magnification using an Olympus BX41 microscope and SPOT camera. The muscle fiber size distribution histograms (number/cross-sectional area) were derived from analysis of an average number of 393.3 ± 32.3 fibers per mouse (n = 3 per cohort). For the number of fibers with internal nuclei, an average of 400 ± 78.4 fiber per mouse (n = 4-7 per cohort) were quantified and for the number of fibers with abnormal mitochondrial distribution and content, an average of 479.5 ± 89.1 fibers per mouse (n = 3 per cohort) were analyzed.

RNA isolation and quantitative reverse transcription PCR

Total RNA was extracted from frozen muscle blocks (20-µm-thick sections, 20 sections per block) using Mini RNeasy Plus Universal Kit (Qiagen, 74134). cDNAs were synthesized using ProtoScript II First Strand cDNA Synthesis Kit (New England BioLabs, E6560L). Primer sets (synthesized by IDT) for NORAD (human, F: AAGCTGCTCTCAACTCCACC, R: GGACGTATCGCTTCCAGAGG; mouse, F: AGACCTCATCCTGTGAACCT, R: CCTCTTGAA CATGGAACTGCTA), PUM1 (human, F: CTTCAAGACACATGCAACGC, R: TTGTCGGAGTGAAAGGTAAGG; mouse, F: CTTCAAGACACATGCAACGC, R: TTGTCGGAGTGAAAGGTAAGG); and PUM2 (human, F: TCGACCACTACCATTTGTCAG, R: TTCCTCCTCTCCAAGTCCAA; mouse, F: AGAGATATACCGACCACTACCG, R: CAAGCCAAACTCCAGGACAT). All quantitative PCR (qPCR) was performed using PowerUp SYBR Green Master Mix (ThermoFisher, A25777) according to the manufacturer’s instructions. qPCR assays were performed using QuantStudio 6 Flex (Applied Biosystem). Expression data were normalized to mouse Gapdh mRNA level and data were analyzed by the ΔΔCt method.

Detection of NP bodies

Human biopsy samples (IBM, n = 4 and Control, n = 4) and gastrocnemius muscle samples harvested from mice (Pum2; rtTA3, n = 3 and WT, n = 3) were sectioned at 8 µm thickness using a cryostat. Cy3-conjugated probes spanning the bases 4743—5159 and 4363—4765 for human NORAD (NCBI Reference Sequence: NC_000020.11) and mouse Norad (NCBI Reference Sequence: NC_000068.8), respectively, were ordered from Eurofins Genomics. Simultaneous multicolor detection of NORAD and PUM2 protein (NBP2-97161AF488, Novus Biologicals for human; ab92390, Abcam followed by Cy2-conjugated secondary antibody [ab6940, Abcam] for mouse) was performed following a previously published protocol.²⁹

Eight to 10 randomly selected representative Z-stacked images were taken at 60×-magnification (6× zoom) for each muscle using a Nikon AX R confocal microscope; NIS-Elements (v5.3) software was used to assess NP bodies. The number and volume of objects that are positive for Norad probe (Cy3), PUM2 antibody (Alexa Fluor 488), and both probe and PUM2 antibody (for colocalization) were determined for each image and averages of these images were used for final analysis.

Results

Involvement of the NORAD-Pumilio axis during the evolution of myopathic process in sIBM

Our prior studies in Norad^−/− mice revealed that Norad is essential for maintenance of tissue homeostasis in vivo. In post-mitotic Norad-deficient tissues, PUM hyperactivity leads to mitochondrial dysfunction, increased reactive oxygen species, and premature aging.²⁶ These observations directed us to examine if the NORAD-Pumilio axis plays a role in the pathobiology of sIBM. We first investigated the expression levels of NORAD and Pumilio proteins in muscle biopsies from sIBM patients with typical clinical presentations of finger flexor and/or knee extensor involvement with or without history of dysphagia. Based on the severity of histopathological features, we divided these biopsies into mild or severe subgroups and found a disease severity–dependent increase in NORAD expression compared to age comparable normal biopsies with no pathological features (Figure 1A). While there was no significant change in PUM1 expression levels, we found a disease severity-dependent increase in PUM2 transcripts accompanying the changes in NORAD expression (Figure 1B and C). The NORAD increase was not exclusive to sIBM muscle, as a similar change was observed in biopsies from patients with histopathological diagnosis of vacuolar myopathy (VM) with no inflammation and DM with focal-perifascicular areas of muscle fiber regeneration, necrosis, and perivascular inflammation (Figure 1D). Moreover, NORAD increase was also associated with increases in PUM1/2 expression in VM and DM which were statistically significant for PUM2 as observed in sIBM (Figure 1E and F).

Figure 1.

Disease activity related increase in (A) NORAD expression in muscle biopsies from sporadic IBM patients showing mild (mIBM) and severe (IBM) pathology compared to control muscle with no pathologic diagnosis. Increased (B) PUM 1 and (C) PUM 2 levels from the same samples compared to control muscle. Increased (D) NORAD expression in biopsies from vacuolar myopathy (VM) and dermatomyositis (DM) cases and increased (E) PUM 1 and (F) PUM 2 levels from the same diseased samples compared to control muscle are shown. Representative confocal images showing NORAD probe-positive staining in (G) IBM and (H) control muscle samples. (I) shows NORAD probe-positive staining around nucleus at high magnification and (J) colocalization of NORAD probe with PUM2 as NP bodies in muscle from an IBM patient. Scale bars: G, H = 50 µm; I, J = 5 µm. Data are represented as mean ± SEM, 1-way ANOVA, *P < .05, **P < .01, ***P < .001. n numbers: Control: 6 to 7, mIBM: 6, IBM: 8-9, DM: 9-10, VM: 5.

Previously, it has been shown that NORAD is able to outcompete thousands of other transcripts to bind and regulate the activity of PUM proteins by nucleating the formation RNP condensates, termed NP bodies, in an in vitro setting.²⁷ As the next step, we developed a protocol and investigated if NORAD regulates PUM2 activity by forming intracytoplasmic NP bodies in fresh frozen muscle tissue from severe sIBM biopsies. We found that NORAD probe-positive staining was co-localized with PUM2 antibody staining predominantly as perinuclear and punctate aggregates, representing NP bodies (Figure 1G-J). Quantitative analysis of confocal Z-stacked images from IBM and control biopsies confirmed colocalization of NORAD and PUM2 protein in NP bodies. Furthermore, we found that the mean individual NP body size in the IBM biopsies was significantly higher than in controls (IBM: 2.3 ± 0.30 µm³ vs control: 0.59 ± 0.27 µm³, P = .00575), constituting a significantly greater total NP body volume per unit area in sIBM compared to control muscle (50.8 ± 9.7 vs 22.4 ± 3.8 µm³, P = .035, n = 4 biopsies analyzed for each group). Mander’s Overlap Coefficient measuring degree of overlap between channels (values close to +1 indicate strong colocalization, values near 0 no correlation) was also higher in the IBM cases (0.980 ± 0.006 vs 0.934 ± 0.008, P = .0048).

Norad and Pumilio expression increases with disease severity in the hIBM mouse model

As the next step, we examined Norad and Pum1/2 expression levels in a mouse model of hereditary IBM (hIBM), the VCP-A232E mouse (Vv), which is modeled from a human mutation in valosin containing protein/p97 gene (VCP). The Vv mice are known to present the earliest histopathological changes in muscle around 3 months of age.³⁰ Older mice around 9-12 months of age were reported to show muscle fiber size variability with increased internal nuclei, occasional muscle fibers with rimmed vacuoles, and intracytoplasmic accumulation of TDP-43, which is ubiquitin-positive.³⁰ Our studies confirmed that compared to 3 months of age, 1-year-old Vv mice showed a significant increase in the number of muscle fibers with internal nuclei compatible with increasing cycles of necrosis and regeneration over time (Figure 2A-C). Additionally, an increase of fibers with abnormal subsarcolemmal accumulation of mitochondria was detected using mitochondrial enzyme histochemistry (SDH) (Figure 2D-F). As in the sIBM muscle biopsies, we examined Norad and Pum1/2 expression levels in this model at 2 different stages of muscle histopathology: early stage (3 months of age) versus late stage (1 year of age). For comparison, we also examined the Norad and Pum1/2 expression profiles temporally in WT C57Bl/6 mice. Norad and Pum1/2 expressions in WT mice at 3, 10, and 24 months of age showed an age-dependent decrease in Norad expression with a significant increase in Pum1 and Pum2 transcripts in aged muscle (Figure 2H and I). Previously, we showed that Norad deficiency is associated with Pumilio hyperactivity and premature aging.²⁶ Contrasting with this normal aging process, but similar to the pattern that we observed in the sIBM patient biopsies, we found a disease severity-dependent increase in Norad and Pum2 expression, comparing 3-month-old Vv mice to 1-year-old animals (Figure 2H and I). These findings, combined with the presence of age-related progressive histopathology in this model, support our observations in sIBM patient biopsies and confirm that there is a disease activity-related increase in NORAD expression during the progression of a myopathic disease, which may represent a compensatory response to elevated PUM2 levels.

Figure 2.

Representative H&E-stained images from gastrocnemius muscle of young (A) and old (B) VCP-A232E (Vv) mice showing increased internal nuclei and fiber size variability with aging. (C) shows percent distribution of fibers with internal nuclei (n = 5 for 2-3 months old, n = 7 for 1-year-old Vv mice). SDH enzyme histochemical stain on muscle sections from young (D) and old (E) Vv mice. (F) shows marked increase in number of fibers with abnormal mitochondria content and distribution (arrows in E) with aging (n = 3 for each cohort). (H-I) show relative mRNA expression levels of Norad, Pum1, and Pum2 in young and old Vv mice and age-matched controls. n numbers: C57BL/6 (2-3 mos): 6, C57BL/6 (1 yr): C57BL/6 (2 yrs): 8, Vv (2-3 mos): 9, Vv (1 yr): 11-12 Data are represented as mean ± SEM, unpaired t-test, 1-way ANOVA, 2-way ANOVA, *P < .05, **P < .01, ***P < .001, ****P < .0001. Scale bars: A, B = 50 µm; D, E = 30 µm.

Evidence that Norad compensates for disease-associated Pum2 upregulation in the hIBM model

We next explored the question of whether increased NORAD expression in myopathic muscle protects from disease pathology and if so whether NORAD upregulation is a compensatory response to increased PUM2 levels. We first tested if Norad insufficiency accelerates disease progression in the Vv mouse model. When we cross-bred Vv with Norad^−/− (nn) mice, we observed that disease was accelerated in both VCP-A232E; Norad^+/− (VvNn) mice and VCP-A232E; Norad^−/− (Vvnn) mice, with more pronounced changes in the latter (Figure 3). Our histopathological characterization studies of the Vv mutant confirmed the onset of pathological changes in muscle to be around 3 months of age (Figure 2), and as expected, muscles from 6 weeks old Vv mice were free of ongoing pathology (Figure 3A). At the same age of 6 weeks, however, Vvnn muscle was predominantly composed of small size fibers with an increase in fiber density (the number of fibers per unit area) compared to age-matched Vv and VvNn mice (Figure 3A-D). In addition, SDH enzyme histochemistry revealed an increase in the number of muscle fibers showing abnormalities in the content and distribution of mitochondria, which was most pronounced in Vvnn compared to Vv and VvNn mice, inversely correlating with Norad copy number in each group (Figure 3E-H). Moreover, in the Vvnn mice, ongoing muscle fiber necrosis/phagocytosis and regeneration were commonly seen and muscle fiber loss with fat replacement was prominent in some samples (Figure 3I-L). In contrast, myopathic changes were more subtle with scattered, rare muscle fiber necrosis, regeneration, and fat replacement in the VvNn mice at 6 weeks of age (Figure 3M-O). At 1 year of age, however, we found significant increases in the number of fibers with internal nuclei and abnormal mitochondria content and distribution in VvNn mice compared to age matched Vv counterparts (Figure S1A-F). These data demonstrate that Norad suppresses disease progression in the Vv model in a dose-dependent manner.

Figure 3.

Representative H&E-stained images from gastrocnemius muscle of (A) VCP-A232E (Vv), (B) VCP-A232E; Norad^+/− (VvNn), and (C) VCP-A232E; Norad^−/− (Vvnn) mice at 6 weeks of age showing (D) significant increase in muscle fibers with diameter < 25 µm and increased fiber density per unit area in the Vvnn mice. Images with SDH enzyme histochemical stain on muscle from (E) Vv, (F) VvNn, and Vvnn (G) mice showing muscle fibers with increased subsarcolemmal mitochondria (arrows) in F and G. (H) Marked increase in number of fibers with abnormal mitochondria content and distribution are seen in VvNn and Vvnn compared to Vv. Representative gastrocnemius muscle sections with H&E stain from Vvnn mice at 6 weeks of age show (I) muscle fiber necrosis, phagocytosis, (J) regenerating basophilic fibers (arrow), and (K) muscle fiber loss with fat replacement, which was prominent in (L) some samples. Images of gastrocnemius muscle from VvNn mice at 6 weeks of age show scattered, rare (M) necrotic or (N) basophilic regenerating fibers and (O) subtle fat replacement. Data are represented as mean ± SEM, n = 3 for all experimental cohorts, 1-way ANOVA, 2-way ANOVA, *P < .05, **P < .01, ***P < .001, ****P < .0001. Scale bars: A-C = 50 µm; 30 µm for other images.

Correlating with these histopathological findings, Pum1/2 expression was similar in muscles from 2- to 3-month-old Vv and VvNn mice, which exhibit similar histopathology at this time point (Figure 4 and Figure S1). In contrast, we found that both Pum1 and Pum2 levels were significantly elevated in young Vvnn muscle compared to age-matched Vv and VvNn counterparts, correlating with the severity of histopathologic changes. Furthermore, we observed that Pum1 and Pum2 expressions in VvNn muscle at 1 year of age were elevated compared to Vv muscle at this time point, also correlating with disease severity. Thus, Pum1/2 expression tracks closely with histopathology in the Vv model.

Figure 4.

Relative mRNA expression levels of (A) Norad, (B) Pum1, and (C) Pum2 in young (solid bars) and old (crossed bars) Vv, VvNn, and Vvnn mice. Data are represented as mean ± SEM, n numbers: Vv (2-3 mos): 9, Vv (1 yr): 11 to 12, VvNn (2-3 mos): 5-6, VvNn (1 yr): 3, Vvnn (2-3 mos): 6, unpaired t-test, 1-way ANOVA, *P < .05, **P < .01, ***P < .001, ****P < .0001.

To determine whether increased Norad expression in myopathic muscle is a direct result of increased in Pum1/2 expression, we employed a previously reported transgenic mouse model with inducible Pum2 expression. We previously showed that the administration of doxycycline (dox) to Pum2; rtTA3 double transgenic mice starting around 2 months of age resulted in highly accelerated aging within 2 months, resembling to Norad^−/− mice at 1 year of age.²⁶ Skeletal muscle from Pum2; rtTA3 mice exhibited mitochondrial abnormalities with a global reduction in COX activity together with scattered necrotic and regenerating fibers.²⁶ Examination of Norad expression in the gastrocnemius muscle from Pum2; rtTA3 mice at 15 weeks of age revealed significantly increased Norad transcripts along with high Pum2 expression compared to age matched WT mice (Figure 5A and B). Furthermore, Norad and PUM2 staining of fresh frozen muscle sections demonstrated the presence of NP bodies coalescing around nuclei in muscle from Pum2; rtTA3 mice (Figure 5C-G), similar to our observations in sIBM muscle biopsies from patients. Collectively, these findings provide strong evidence that elevated NORAD expression is a protective, compensatory response to disease-related induction of PUM2 activity.

Figure 5.

Elevated Norad (A) and (B) Pum2 mRNA levels in the gastrocnemius muscle of Pum2; rtTA3 double transgenic mouse model at 15 weeks of age 6 weeks following enforced Pum2 expression (Pum2-OE). Representative confocal images showing Norad probe-positive staining in (C) Pum2-OE and (D) control gastrocnemius muscle samples. High magnification mages show (E) Norad probe-positive staining, (F) PUM2 antibody staining, and (G) their colocalization as NP bodies coalescing around nucleus in the Pum2-OE muscle. Scale bars: C, D = 20 µm; I, J. Data are represented as mean ± SEM, 2-way ANOVA, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Discussion

sIBM is the most common late onset myopathy with complex pathogenesis that includes degenerative, inflammatory, and mitochondrial components. Since the discovery of Congo red-positive amyloidogenic inclusions in sIBM biopsies,⁴ an accumulating large body of evidence has now established that the histopathological phenotype of sIBM muscle has similarities to the brains of Alzheimer and Parkinson disease patients, the 2 most common neurodegenerative diseases associated with aging as risk factor.^{6,15,22,31–33} The hallmarks of the degenerative process, common in these myodegenerative and neurodegenerative disorders include pathological protein aggregation, aberrant proteostasis, and altered energy homeostasis.

In the sIBM muscle, the presence of mtDNA sequence changes, clonal expansion of mtDNA deletions, decreased mtDNA copy number, decreased mitophagy, mitochondria proliferation, and oxidative phosphorylation defects have long been recognized.⁹ In addition, there is extensive evidence linking decreased mitochondria function to aging-associated phenotypes.³⁴ This includes a previous demonstration that mice expressing defective mitochondrial DNA polymerase, which consequently accumulate mtDNA mutations and deletions, exhibit a premature aging phenotype³⁵ with many similarities to that seen in Norad-deficient mice.²⁶ These observations prompted us to investigate whether the NORAD-Pumilio axis plays a role in the evolution of the disease process in sIBM. PUM proteins bind an extensive network of mRNAs and repress protein expression by inhibiting translation and promoting mRNA decay. Therefore, PUM1/2 regulatory networks are involved in diverse cellular processes and molecular pathways.²⁸ PUM proteins promote proliferation and differentiation³⁶ and are therefore essential for growth and development while dysregulation of PUM activity contributes to neurodegeneration, movement disorders, epilepsy, infertility, and cancer. As examples, haploinsufficiency of PUM1 was shown to result in an increase in Ataxin1 levels, leading to progressive motor dysfunction and degeneration of Purkinje cells, features typical of SCA1.³⁷ A missense PUM1 mutation was shown to be associated with adult-onset ataxia whereas PUM1 mutations leading to haploinsufficiency cause developmental delay and seizures.³⁸

NORAD is the preferred binding partner of PUM2 in human cells and mouse tissues.²⁵ We previously showed that Norad-deficiency resulting in PUMILIO hyperactivity triggers the coordinated repression of a large set of PUM2 target transcripts with key roles in mitochondrial function and homeostasis, associated with widespread structural and functional mitochondrial defects.²⁶ These findings were corroborated by others reporting that elevated PUM2 levels in aged mice impaired mitochondrial homeostasis.³⁹ In this study, we found a disease-severity-dependent increase of NORAD and PUM2 transcripts in muscle biopsies from IBM patients; this pattern was also present in the VCP-A232E mutant, a mouse model for hIBM. The presence of this disease severity-dependent increase in Norad and Pum2 expression in this model raised the question of whether the Norad increase might be a compensatory transcriptional adaptation to counteract the increase in Pum2 expression. As we demonstrated using Pum2; rtTA3 transgenic mice, Norad expression indeed increases in response to enforced Pum2 expression, providing compelling evidence for a compensatory, protective role of Norad to control a disease-induced increase in Pum2 activity that may be harmful for cellular energy homeostasis. In fact, prior studies have shown that disease-induced or age-related increased oxidative damage leads to an increase in PUM2 expression. Under conditions of oxidative stress, such as hydrogen peroxide treatment or interleukin-1β stimulation, PUM2 expression is induced, which subsequently impairs mitochondrial function and promotes further cell damage.^39,40 Multi-omics and cross-species analyses showed that PUM2 inhibits the translation of the mRNA encoding for the mitochondrial fission factor, thereby impairing mitochondrial fission and mitophagy, which are critical for mitochondrial quality control and tissue homeostasis. This mechanism appears highly conserved, maintained in Caenorhabditis elegans by the PUM2 ortholog, PUF-8.³⁹

We would like to emphasize that increased NORAD and PUM2 expression is not exclusive to sIBM muscle as we observed a similar increase in biopsies from patients with histopathological diagnosis of vacuolar myopathy and DM as well as in the VCP-A232E mutant mice, a model for hIBM. As anticipated, mitochondrial dysfunction is not unique to sIBM and has been shown to associate with other idiopathic inflammatory, acquired, or hereditary myopathies to varying degrees.⁴¹ A major cluster of downregulated genes related to mitochondria activities, translating into mitochondrial abnormalities, increased ROS and decreased OXPHOS were found in DM, as revealed by transcriptome analysis.⁴² Not surprisingly, the predominantly affected perifascicular areas of DM muscle, subjected to ischemic damage were also the regions where most mitochondrial alterations can be observed.⁴³ In addition, pathogenic VCP mutations can induce mitochondrial abnormalities including uncoupled respiration, reduced mitochondrial membrane potential, decreased ATP production, and impaired mitochondrial quality control.^44,45 Collectively, these observations suggest that the NORAD-Pumilio axis is an important upstream regulator of the bioenergetics of muscle and that NORAD is playing a protective role during the evolution of myopathic process.

NORAD is known as the preferred target for PUM proteins, controlling and regulating the activity of PUM proteins through the formation of cytoplasmic NP bodies.²⁷ As anticipated, we were able to detect NORAD probe-positive staining, co-localized with PUM2 antibody as perinuclear and punctate aggregates in sIBM patient biopsies representing NP bodies. We were intrigued by the close association/co-localization of NP bodies with subsarcolemmal “residual bodies,” also known as lipofuscin bodies, which are commonly observed in sIBM biopsies as well as in other aging post-mitotic tissues including aged neurons. It is likely that NORAD levels could correlate with the amount of residual/NP body formations quantitatively. Although theoretically lipofuscin could contain RNA-binding proteins among other components, to our knowledge this is the first demonstration of PUM2, complexed with NORAD, and structurally associated with residual bodies. Further studies may explore the mechanistic pathophysiological evolution of NP and residual body relationship, and whether the onset of formation of these subcellular structures may coincide.

It is now well established that PUM2 is induced upon aging and acts as a negative regulator of lifespan and mitochondrial homeostasis.^26,39 As we illustrated here in the WT mice, the Pum2 increase is accompanied with an age-related gradual decline in Norad expression levels. The contrasting pattern of the NORAD-PUM2 relationship between normal physiological aging process ultimately leading to death and what we observed as a compensatory/protective increase of NORAD to control PUM2 activity during the evolution of myopathic process is intriguing and will require further studies for understanding the underlying mechanisms of these 2 different conditions.

In summary, our findings indicate that NORAD upregulation has a compensatory/delaying role in the pathobiology of hIBM disease progression and in other myopathic diseases. Using the VCP-A232E mutant, Pum2-overexpressing, and Norad-deficient mouse models, we uncovered evidence that Norad increases in response to Pum2 induction during disease pathogenesis, suggesting that the NORAD-Pumilio axis may play an important regulatory role in the optimal maintenance of bioenergetics and stress responses of muscle in both normal physiological and disease states. Our study strongly suggests that NORAD plays a protective role in the pathobiology of IBM disease progression and likely has a similar role in other myopathic conditions as well. We also predict that overexpression of NORAD, or PUM2 RNA silencing, using viral or nonviral vectors during strategically determined stages of the disease process could be beneficial for controlling disease activity and progression. In this respect, we anticipate that NORAD blood levels could serve as a predictive biomarker for disease activity and response to therapy.

Supplementary material

Supplementary material is available at Journal of Neuropathology & Experimental Neurology online.

Conflicts of interest

The authors report no competing interests.

Funding

This study was funded by the Technology Development Fund by Office of Technology Commercialization at Nationwide Children’s Hospital (Award No.: AWD50001263 to Z.S.), the Welch Foundation (I-1961 to J.T.M.), and the National Institutes of Health (R01CA282036 and U01CA305105 to J.T.M.). J.T.M. is an Investigator of the Howard Hughes Medical Institute.

Data availability

The data that support the findings of the study are available as supplementary material.