Expanding the Molecular and Pathologic Spectrum of HSPB8 Myopathy and Distal Motor Neuropathy

Abstract

Objectives

HSPB8 variants cause myopathy, distal motor neuropathy, and Charcot-Marie-Tooth disease. We describe 2 patients who expand the molecular and pathologic spectrum of HSPB8 disorder.

Methods

We reviewed clinical and laboratory data and performed molecular dynamics simulations to explore variant effect.

Results

Patient 1 is an adult man presenting with childhood-onset, distal lower limb weakness, followed by proximal weakness. EMG detected predominant myopathic and neurogenic changes in upper and lower limbs, respectively. Biopsy revealed myopathy with rimmed vacuoles in the supraspinatus and neurogenic changes in the tibialis anterior. He carries a novel, predicted deleterious HSPB8 heterozygous variant, c.185G>A (p. Gly62Asp). Patient 2 is an adult man presenting with distal, asymmetric, progressive lower limb weakness that extended to proximal and neck muscles. Quadriceps biopsy showed myopathy with rimmed vacuoles and protein aggregates, especially TIA1, p62, and TDP-43. TIA1 aggregates were more prominent than Z-disk protein accumulation. He carries a known HSPB8 pathogenic variant, c.421 A>G (p.Lys141Glu). Molecular dynamics simulations suggested that p.Gly62Asp may exert its effects through post-translation modifications while p.Lys141Glu may disrupt dimerization.

Discussion

HSPB8 p.Gly62Asp is the first N-terminal variant associated with myopathy. TIA1 aggregates, more prominent than Z-disk myofibril aggregates, suggest that p.Lys141Glu may affect stress granule dynamics more than Z-disk integrity.

Introduction

Heat shock protein family B member 8 (HSPB8) is a ubiquitous chaperone preferentially expressed in muscle and neurons, crucial for preventing misfolded protein aggregation.^1,2 It is part of the chaperone-assisted selective autophagy complex, which is relevant for maintenance of Z-disk integrity. HSPB8 has 3 domains (D): N-terminal (NTD), alpha-crystallin (ACD), and C-terminal (CTD). HSPB8 variants can cause myopathy, distal hereditary motor neuropathy (dHMN), or Charcot-Marie-Tooth disease type 2L.³ Currently, known variants causing HSPB8 myopathy are frameshift in CTD resulting in elongated protein, except p.Lys141Glu.^3-7 Recently, HSPB8 myopathy was referenced in the Online Mendelian Inheritance in Man as myofibrillar myopathy type 13 with rimmed vacuoles.

We report 2 patients with HSPB8-associated neuromuscular disorder who expand the disease spectrum, including the first NTD variant causing myopathy. We describe findings of molecular dynamics (MD) simulations to explore variants impact on protein structure and dynamics.

Methods

Patients' data were reviewed (eMethods). To investigate effects of patients' HSPB8 variants on structure and dynamics, structural models were generated using AlphaFold 3, the gold standard for ab initio molecular modeling,⁸ and refined using MD simulations (eMethods). For validation, we used RoseTTAFold for structure prediction.⁹

Standard Protocol Approvals, Registrations, and Patient Consents

The study was approved by the Mayo Clinic Institutional Review Board (IRB). Written informed consent was obtained from patient 1 and his brother (IRB#13-007054). Patient 2's consent was waived for the retrospective data review (IRB13-004243/IRB#25-005895). Authorization was obtained from patients for disclosure of any recognizable person in photographs, videos, or other information that may be published in the journal.

Data Availability

Anonymized data not published within this article will be made available by request from any qualified investigator.

Results

Patient 1

A man in his early 60s, born to nonconsanguineous asymptomatic parents (deceased), presented with progressive distal leg weakness since age 10. Around age 40, he noted progressive proximal arm and axial weakness. He had diabetes for a decade. Examination revealed severe (Medical Research Council, MRC grade 1–2) asymmetric weakness of the iliopsoas, calf, and toe flexors; moderate (MRC grade 3–4-) weakness of other lower limb muscles except the quadriceps; mild (MRC 4- to 4+) weakness of the supraspinatus and infraspinatus; and moderate abdominal wall muscle weakness. Lower limb distal sensation was reduced. Achilles tendon reflexes were absent.

Electromyography (EMG)/nerve conduction studies (NCSs) showed findings of myopathy in upper limbs, polyradiculopathy in lower limbs, and length-dependent sensory neuropathy (eTable1). Serum creatine kinase (CK) was normal. ECG and echocardiogram were normal. Pulmonary function tests and overnight oximetry revealed reduced maximal respiratory pressures and features of sleep-disordered breathing, respectively. Supraspinatus biopsy (Figure 1, A–G) revealed myopathy with rimmed vacuoles. Tibialis anterior biopsy (Figure 1H) showed neurogenic changes and a fiber harboring rimmed vacuoles. The patient carries a novel heterozygous HSPB8 (NM_014365.2) variant of uncertain significance, c.185G>A (p. Gly62Asp), absent in the 69-year-old asymptomatic brother. Other sequenced genes (eMethods) causing protein aggregate myopathies showed no pathogenic or suspected pathogenic variants.

Figure 1. Patient 1's Muscle Biopsies.

Supraspinatus biopsy (A–G). (A and B) Hematoxylin-eosin–stained sections showing internalized nuclei (A, asterisk), rare necrotic (A, arrowhead) fibers, rimmed vacuoles (B, arrows) in atrophic and nonatrophic fibers, and increased endomysial connective tissue (B, double dagger). (C) Gomori trichrome–stained section showing a fiber (asterisk) with a rimmed vacuole (arrow), which overreacts for p62 (D, arrow) and TDP43 (not shown). (E) Small p62-positive aggregates also occurring in nonvacuolated fibers (arrow). (F) Small punctuate TIA positivity occurring in the sarcoplasm of the same vacuolated fiber shown in (C). (G) Congo red–stained section viewed under rhodamine optics, demonstrating a small congophilic inclusion (arrow). No Z-disk myofibril protein aggregates were observed; exceedingly rare vacuolated fibers were faintly and diffusely overreacting for myotilin and alpha-B crystallin (not shown). (H) Tibialis anterior biopsy. ATPase-stained section (pH 9.4) showing marked predominance and grouping of type 1 fibers with only a few type 2 fibers (arrows). Magnification: ×40 (A–K) and ×10 (L).

Patient 2

A man in his late 30s, born to asymptomatic parents, manifested asymmetric calf weakness 10 years prior. He then developed calf atrophy and fasciculations, followed by generalized lower limb and lumbar paraspinal weakness. Examination revealed proximal upper limb weakness (MRC grade 4+), weakness of the neck flexors, and distal-predominant lower limb weakness (MRC grade 4- to 3+). Tendon reflexes were normal or reduced. Sensation was normal.

EMG/NCSs showed myopathic changes with fibrillation potentials (eTable2). Serum CK was 420 U/L (normal 39–308). ECG was normal. Vastus lateralis biopsy (Figures 2) revealed myopathy with rimmed vacuoles and sarcoplasmic protein aggregates. The amorphous structures observed in trichrome-stained sections often stained for TPD-43 and TIA1, or much less frequently for myotilin and alpha-B crystallin. Only small intravacuolar congophilic inclusions were present. NADH dehydrogenase–reacted sections showed ring fibers. There were no features of reinnervation. The patient carries a known HSPB8 pathogenic heterozygous variant, c.421A>G (c.Lys141Glu), not present in either parent, suggesting a de novo variant (Figure 3A).

Figure 2. Patient 2's Vastus Lateralis Biopsy.

(A) Hematoxylin-eosin–stained section showing a fiber with rimmed vacuoles (asterisk) and atrophic fibers (small fibers). (B) Congo red–stained section demonstrating small congophilic inclusions (bright red) within a vacuolated fiber (asterisk). No large congophilic inclusions, as often seen in classic myofibrillar myopathy, were seen. (C) Gomori trichrome–stained section highlighting amorphous, dark, small (asterisks), and large (arrowhead) inclusions and a rimmed vacuole (arrow), which (D) overreacts for TIA1. (E) TDP-43 positivity occurring in the large inclusion (arrowhead) and vacuole (arrow). (F) A different fiber in the Gomori trichrome–stained section, with inclusions conferring a lace-like appearance (arrowhead) overreacting for (G) TIA1 and (H) p62, but not for TPD-43 (not shown). (I) In other fibers, the amorphous areas (arrow) seen in Gomori trichrome–stained sections, overreact for (J) NADH dehydrogenase, (K) myotilin, and (L) alpha-B crystallin. Magnification: ×40.

Structure Prediction and MD Simulations

Figure 3A illustrates patients' variants and protein domains. HSPB8 structural models were generated using AlphaFold 3. RoseTTAFold for structure prediction produced a comparable model (data not shown). Owing to extensive disorder in the predicted dimer structure (Figure 3B), the system required prolonged equilibration. The simulation stabilized after approximately 400 ns, based on the root mean square deviation of backbone atoms in the ACD (Figure 3C). Therefore, the analysis was conducted using data from 400–800 ns. Interaction energies—total Coulombic and total van der Waals—were calculated (Figures 3, D and E). While van der Waals energies were similar between wild type and mutant, the Coulombic interaction energy was significantly more favorable in wild-type HSPB8.

Figure 3. Schematic Representation of HSPB8 Gene, Protein, Previously Reported and Novel Variants (A), and Molecular Dynamics Simulations of the HSBP8 Homodimer Structure (B–E).

(A). The top panel represents the gene with its 3 exons [exon 1: amino acid (aa) 1–123, exon 2: aa 123–144, exon 3: aa 144–196] while the bottom panel is a schematic representation of the protein structure. The distinct colors indicate the associated phenotype. The novel variant in patient 1 is indicated in red. NTD = N-terminal domain; ACD = alpha-crystallin domain; CTD = C-terminal domain. (B). Structural model of the HSBP8 homodimer generated using AlphaFold 3. The 2 chains are shown in green and cyan. Residue Gly62 is highlighted as a red sphere. The close-up view on the right emphasizes the hydrogen bonds formed by Lys141, Leu31, and Asp33. (C) Backbone RMS deviation of HSBP8 wild type and p.Lys141Glu mutant during an 800-ns MD simulation. (D). Total Coulombic energy and (E) total van der Waals energy of the well-folded region of HSBP8 (residues 93–170), calculated over the 400–800 ns segment of the MD simulations. Wild type (WT) is shown in black; the p.Lys141Glu mutant is shown in red.

Discussion

These patients expand the spectrum of HSPB8-associated neuromuscular disorders.

Patient 1 has a novel HSPB8 missense variant (c.185G>A, p.Gly62Asp), manifesting as myopathy with rimmed vacuoles (affecting upper limbs) and dHMN. While diabetes contributed to neuropathy, the longstanding leg weakness favors dHMN as the main etiology.

The novel HSPB8 variant pathogenicity is supported by the following: (1) p.Gly62 is a highly conserved amino acid from C. elegans to humans, suggesting its critical role in protein function; (2) p.Gly62Asp is predicted to be deleterious (PolyPhen2); (3) the variant is absent in a population database and the asymptomatic brother; (4) the phenotype of combined myopathy and dHMN is in keeping with HSPB8 disorder.³ The p.Gly62Asp is the first NTD variant linked to myopathy. Our MD simulations predict that p.Gly62 resides in a disordered region (Figure 3B) and, therefore, is unlikely to affect protein folding/stability. HSPB8 p.Gly62 is a predicted myristoylation site, and, although it is unknown whether HSPB8 is myristoylated, the variant could compromise myristoylation, a process relevant for protein stability and membrane localization.¹⁰ In addition, p.Gly62 is adjacent to p.Thr63, a putative phosphorylation site, and could affect HSPB8 phosphorylation, chaperone activity, or degradation.² In vitro biochemical experiments showed that HSPB8 exists as a mixture of monomers, dimers, and higher order oligomers and NTD variants affect this distribution.¹¹

Patient 1's myopathy is characterized by rimmed vacuoles. In contrast to the myopathologic changes in HSPB8 myopathy associated with variants in CTD or ACD, patient 1's biopsy showed no myofibrillar pathology. The distinct myopathologic changes could stem from the different variant effects. The heterogeneity of pathologic changes in different muscles or within the same muscle cannot be excluded.

Patient 2's myopathy is due to HSPB8 p.Lys141Glu, a variant previously described as causative of combined motor axonopathy and myopathy.⁴ HSPB8 p.Lys141Glu, which occurs at a variant hot spot, is the only variant located in the ACD so far associated with myopathy. Variants in the HSPB8 ACD were shown to reduce in vitro chaperone activity.⁴ Our MD simulations of an HSPB8 homodimer model revealed that p.Lys141Glu is energetically unfavorable for dimerization. Lys141 forms 2 hydrogen bonds with Leu31 and Asp33 (Figure 3B), and the variant disrupts those interactions, likely destabilizing the dimer and potentially promoting HSPB8 oligomerization.

Our patients' muscle biopsies, especially that of patient 2, showed TIA1 aggregates, as we had previously noted in a patient with HSPB8 myopathy caused by a CTD variant.⁵ TIA1 is a mRNA-binding protein and marker of stress granules. These are membrane-less ribonucleoprotein complexes that assemble in response to stress, segregating mRNA, and disassemble on stress resolution with re-establishment of protein translation.^12,13 Formation of aberrant stress granules is implicated in neurodegenerative diseases. Additional studies demonstrated that only a small proportion of aberrant stress granules are targeted by autophagy, and most stress granules disassemble in a process dependent on the chaperone complex HSPB8-BAG3-HSPA.¹⁴ Impairment of this complex leads to altered stress granule disassembly and aberrant translation restoration. Considering these findings, the TIA1 aggregates in the absence of large Z-disk myofibrils aggregates, as seen in classic myofibrillar myopathy,¹⁵ would suggest that patient 2's mutant HSPB8 may exert its pathogenic role by affecting stress granule dynamics more than Z-disk maintenance.

In conclusion, HSPB8 should be included in the genetic evaluation of patients with myopathy characterized by rimmed vacuoles and/or myofibrillar pathology, of childhood and adulthood onset, with or without associated motor axonopathy. Absence of Z-disk myofibrillar aggregates should not dissuade consideration of an HSPB8-neuromuscular disorder.

Author Contributions

B.N. Putko: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; analysis or interpretation of data. E.J. Sorenson: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; analysis or interpretation of data. G. Cui: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; analysis or interpretation of data. T. Liewluck: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data. Z. Niu: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data. W.J. Litchy: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data. G. Mer: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; analysis or interpretation of data. M. Milone: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; study concept or design; analysis or interpretation of data.

Study Funding

This work was supported by a grant from Minnesota Regenerative Medicine (P008848103) and the Muscular Dystrophy Association (MDA) (award ID: 497263) to M.M.

Disclosure

M. Milone has received personal compensation in the range of $500-$4,999 for serving on a Scientific Advisory Board for Cartesian Therapeutics on an unrelated topic. Go to Neurology.org/NG for full disclosures.