Clinical findingsDevelopment

The boy was born via cesarian section in 36th week of gestation due to HELLP-syndrome of the mother [weight 3060 g (25 P., -0.66 z), length 54 cm (85 P., -1.03 z), head circumference 35 cm (52 P., -0.04 z)] he suffered from episodes of apnea and bradycardia until 4 months of age and needed feeding via nasogastric tube for 6–8 weeks postnatally. Motor development was delayed, as he was able to sit with 22 months, to stand with 24 months and to walk with 30 months. Achievement of speech development milestones was also delayed: two words at the age of 15 months and no further development afterwards. In line with this, cognitive development is severely impaired (IQ 50/SON-R). Communication could be established using sign language. Neurologically, also at his last appointment at the age of 12 years in 2023, he presented with muscular hypotonia, hyperextensible joints and soft muscles. Subsequently, he had developed a broad-based gait pattern and difficulties in climbing stairs. The mother reported episodes suspicious of absence seizures with upward gaze deviation and tremor. EEG was also normal.

Diagnostic workup

At his first presentation in our department with the age of 15 months (Fig. 1A and B), the boy showed a severe motor development disorder with pronounced weakness in the upper extremities and generalized muscular hypotonia. A positive decrement (median nerve) was identified upon repetitive nerve stimulation (Fig. 1C). Based on this pathological decrement and his clinical features fitting to a congenital myasthenic syndrome, at the age of 20 months treatment with pyridostigmine (3.8 mg/ Kg bodyweight) was initiated and the boy initially (four weeks after first treatment) showed a benefit in muscle strength of lower extremities and improvement of motor development: he was able to drink on his own and roll from back to stomach as well as to sit unsupported and put himself into upward position. Repeatedly blood tests with creatine kinase (last CK-value: 161 U/l, ref < 165 U/l), basic and extended metabolic screening produced normal results. An MRI of the brain showed no pathologies (data not shown). EEG was normal. Electroneurography measuring nerve conduction velocity showed normal values.

Fig. 1

Patient’s photograph at the age of 15 months. A Facial dysmorphia with Cupid’s bow lip. B Doubled crease of M. pectoralis is shown. C Positive decrement of the median nerve upon repetitive stimulation. D Schematic representation of the diagnostic work-up along with clinical findings in our PTHS patient

However, as no persisting improvement was achieved upon the treatment with pyridinstigmine, a muscle biopsy was performed at the age of 2.5 years to exclude an underlying myopathic disease.

With disease progression, our follow-up clinical examinations revealed dysmorphia with persisting Cupid’s bow lip (Fig. 1), talipes varus of both feet and doubled crease of the pectoralis muscle. The patient was diagnosed with PTHS at the age of 8 years.

A schematic representation of the timeline of the diagnostic workup and clinical findings is presented in Fig. 1D.

Microscopic findings

Due to suspected muscular involvement, a muscle biopsy was performed at two years of age. Histological investigation of the muscle collected from right vastus lateralis muscle did not show any striking myopathologies. However, results of ATPase 4.3 reaction showed a moderate type I fiber predominance, whereby these fibers were occasionally organized in groups (Fig. 2). An enzyme histochemistry revealed no COX-negative fibers but indicated increased subsarcolemmal mitochondrial content in some fibers by nicotine adenine dinucleotide hydride (NADH) and combined Cytochrome C oxidase and Succinate dehydrogenase reaction (COX-SDH). Immunohistological-based examination of protein abundances (including Caveolin 3, Dystrophin 1–3, α-/ β-Dystroglycan, α-/ β-/ γ-/ δ- Sarcoglycan, Laminin α2, Laminin α5, Emerin, Dysferlin, Collagen IV and VI, Utrophin, neonatal myosin) of the muscle showed normal results (data not shown). Given that in the central nervous system an impact of TCF4 abundance on fibroblast growth factor production (necessary for neuronal function) was shown [16], by performing immunofluorescence studies on the patient-derived muscle biopsy, we additionally investigated the characteristics of muscle fibroblasts: Co-staining of Vimentin (fibroblast marker) and CD90 (marker of activated fibroblasts) revealed an approximate complete co-expression of both proteins in muscle biopsy specimen derived from two controls whereas in the biopsy derived from our patient a co-expression could only be detected in roughly 50% of fibroblasts with a higher proportion solely showing immunoreactivity for CD90 but not Vimentin (Fig. 3).

Fig. 2

Light microscopic findings on TCF4-mutant muscle. H&E as well as Gomori stains show normal morphology. Increase of mitochondria in the subsarcolemmal region of some muscle fibres is indicated by NADH and COX-SDH stain. COX-negative fibers as not present. ATPase stain (pH 4.3) shows moderate increase of type I fibers occasionally organized in groups. This finding was confirmed by immunofluorescence studies of MYH7 labelling type I fibers. COX: cytochrome c oxidase; SDH: succinate dehydrogenase; HE: Hematoxylin and Eosin; Gomori: Gomori-Trichorme stain; NADH: Nicotinamid Adenin Dinucleotid Hydrid. Scale bars: 50 µm and in immunofluorescence stain of MYH7 100 µm

Fig. 3

Immunofluorescence-based studies of myofibroblasts on TCF4-mutant muscle. Reduced co-localization of CD90 and Vimentin in the muscle biopsy specimen of the PTHS patient compared to sex- and age-matched normal disease controls (NDC). Nuclei are visualized by DAPI staining. Co-localization is indicated by yellow arrows whereas areas solely immunoreactive for Vimentin are highlighted with green arrows. Scale bars: 50 µm

Molecular genetic findings

Conventional karyotyping and array CGH were unremarkable. The boy was tested negative for Angelman syndrome and panel diagnostic for encephalopathic epilepsies including Rett syndrome and Coffin-Lowry syndrome. The analysis of point variants in genes relevant for PTHS did not reveal any likely pathogenic or pathogenic variants. However, the analysis of copy number variants revealed a possible deletion of exons 15 and 16 of TCF4 [(NM_003199.2): c.(1146 + 1_1147-1)_(1468 + 1_1487-1)del]. This deletion was confirmed by MLPA analysis and shown to be absent in both parents.

Proteomic signature of TCF4-mutant muscle

A proteomic profile was performed on whole protein extract of the muscle biopsy of the PTHS patient and three controls and enabled the robust quantification of 2243 proteins (Fig. 4A and B). Among these, we identified 99 statistically significant dysregulated proteins: 87 Proteins were upregulated, 12 proteins downregulated (Fig. 4C & Supplementary Table 1). Of these dysregulated proteins, 17 are involved in B-cell-depending immune response, including immunoglobulins, but also unspecific humoral immune system, such as complement factors (Table 1 & Fig. 4D). Notably, 7 dysregulated proteins modulate oxidative stress burden (Table 2) and further 7 affected proteins are localized to mitochondria (Table 3). After filtering the proteomic data for significantly dysregulated proteins with known involvement in specific diseases, we identified different proteins associated with phenotypical features of PTHS: These proteins include NAXE (NADPH-hydrate epimerase) and HMGCS2 (Hydroxymethylglutaryl-CoA synthase). Pathogenic NAXE variants are associated with PEBEL (encephalopathy, progressive, early-onset, with brain edema and/or leukoencephalopathy) (OMIM #617,186) [17]. Mitochondrial HMG-CoA synthase deficiency caused by bi-allelic pathogenic variants may be associated with psychomotoric retardation (OMIM # 605,911). GSS (Glutathione synthetase) is related to glutathione synthetase deficiency, which go along with central nervous damage [18]. Psychomotoric retardation and muscle weakness are associated with CAH2 (carbonic anhydrase II)-deficiency [19]. Remarkably some of the upregulated proteins are in line with inflammatory processes. In particular, 15 immunoglobulin subtypes were found to be increased in the context of possible B-cell activation (Table 1). Two complement factors (C4A and C4B) were found to be upregulated according to humoral immune response. None of those dysregulated proteins is notoriously associated with neuromuscular diseases on a genetic base, but B cell and complement activation are well known hallmarks in a variety of muscular diseases.

Fig. 4

Proteomic studies on whole protein extract of TCF4-mutant muscle. A Schematic representation of the applied workflow. B Overall statistics of PTHS skeletal muscle proteomics. C Volcano plot showing proteins identified as being decreased (orange dots) and increased (purple dots) in PTHS patient-derived muscle. D Results of proteomaps-based data analysis indicating activation of complement cascade and altered oxidative phosphorylation based on identified protein dysregulations

Table 1 Dysregulated proteins within PTHS patient-derived muscle biopsy associated with inflammatory processesTable 2 Dysregulated proteins within PTHS patient-derived muscle biopsy associated with oxidative stressTable 3 Dysregulated proteins within PTHS patient-derived muscle biopsy showing mitochondrial localizationResults of confirmational immunofluorescence studies on TCF4-mutant muscle

Given that our proteomic findings were indicative for activation of the complement cascade and B-cell involvement, further immunostaining studies were carried out to validate these findings in the muscle tissue. C5b-9 immunostaining studies revealed an increased expression on some larger vessels, but not on muscle fibers (Fig. 5). Also, HLA1 showed more intense expression on larger vessels (Fig. 5, white arrows) as well as an increased intensity on smaller capillaries (Fig. 5, yellow arrows). Only singular CD4 + T-cells were detected; however, the few detected cells were localized around the vessels (Fig. 5). Immunostaining for the lymphocyte marker CD45 only reveals single immunoreactive cells in the muscle tissue (Fig. 5), while immunoreactivity for the B-cell marker CD20 was not detected (data not shown). Immunostaining of HLA2 was unremarkable and staining for CD8 + T cells, as well as CD68 + macrophages and CD79A + plasma cells was negative (data not shown).

Fig. 5

Immunofluorescence studies on TCF4-mutant muscle. C5b-9 immunostaining shows an increased immunoreactivity at some vessels (white arrow). Similarly, HLA1 and CD4 showed an increase at vessels (white arrows) accompanied by a general increase within the extracellular space (yellow arrows) which are indicative for the presence of CD4 + helper T cells. Immunostaining of CD45 revealed a dot-like increased immunoreactivity within the extracellular space (white arrows). Scale bars: 50 µm

To study the effect of the exon deletion in TCF4 on the level of corresponding transcripts, quantitative PCR was carried our revealing a profound reduction of the gene expression levels, compared to the level detected in muscle of juvenile control individuals (Fig. 6).

Fig. 6

Transcript studies on TCF4-mutant muscle. Quantitative PCR studies on whole RNA-extracts converted to cDNA revealed a 42% reduction of TCF4 transcripts along with a reduction of C2 (-61%), C5 (-43%), CXCR5 (-17%) and NNMT (-40%). In contrast, level of further studied transcripts were increased: CXCL12 (+ 87%), ETFA (+ 37%), LDHB (+ 28%) and NDUFA12 (+ 28%). BNIP3 showed only a 5% increase

To validate proteomic findings, we additionally studied the level of transcripts corresponding to components of the complement cascade (C2, C5), B cell homing factor CXCL12 and it´s receptor CXCR5, as well as genes responsible for mitochondrial function and turnover (ETFA, NNMT, NDUFA12, BNIP3 and LDHB). Whereas level of C2 and C5 were profoundly decreased (61% and 43%, respectively), the chemokine CXCL12 was strongly increased, hinting to a recruitment of B cells into the muscle tissue. However, the respective receptor CXCR5 was only slight changed (-17%) compared to the level identified in control muscle (Fig. 6). Regarding the level of transcripts encoding for mitochondrial relevant proteins, only NNMT displayed a decreased (by 40%) whereas ETFA, NDUFA12 and LDHB showed an increased (37%, 28% and 28%, respectively) (Fig. 6). Level of BNIP3 remained almost stable compared to controls (Fig. 6).