P4HA1 mutations cause a unique congenital disorder of connective tissue involving tendon, bone, muscle and the eye

Clinical findings

The proband (P1) is an 8-year-old male born to non-consanguineous parents of Chinese descent. First concerns arose prenatally at a routine ultrasound at 20 weeks of gestation, which showed contractures of the joints, with the knees and elbows in a fixed position in all fetal scans (images are not available). Prenatal movements were present on all subsequent ultrasound examinations. He was born at 40 weeks and 3 days by Cesarean section due to failure to progress. Apgars were 7 and 9. Birth weight was 3320 grams (50^th percentile for ethnic Chinese newborns), and birth length was 48 centimeters (8.5^th percentile for ethnic Chinese newborns). Both axial and appendicular hypotonia were noted with particularly pronounced low tone distally in the wrists without any movement except for thumb flexion and extension. There were no swallowing or breathing concerns. He had delayed motor development and started sitting when placed at age 9 months, standing at age 1.5 years, taking steps with support at age 2 years, walking in braces at age 3 years and walking independently at age 4 years.

On examination at age 7 ½ years, P1 had high myopia [−14.50 (OD) and −19.50 (OS)], macrocephaly (head circumference > 98% for age), low weight (weight < 3% for age) and small stature (height <3% for age). He had mild dysmorphic features with a broad nasal bridge, flat midface, mild frontal bossing, and retroverted ears, as demonstrated in clinical images taken at 6 ½ years of age (Fig. 1A and B) . He had camptodactyly, with notable webbing of fingers and absent flexor creases (Fig. 1C and D). He had flat arches and prominent calcanei. There was facial weakness with bilateral ptosis, endgrade limitation of upgaze, ocular proptosis, a high-arched palate and a transverse smile (Fig. 1E). Skin was soft and velvety. He had hypotonia which was prominent distally. He had joint contractures at the hips, knees, elbows, wrists, and the metacarpophalangeal and proximal interphalangeal joints of the hands, with hypermobility of the distal interphalangeal joints of the hands as well as hypermobility of the ankles and toes. Thumbs were maintained in a position of adduction (Fig. 1C and D). Despite his hand positioning, he was able to draw a very detailed picture with a regular pen, holding the pen between the thumb and palm of the hand. Deep tendon reflexes were absent, except at the biceps (1+). He stood with a lordotic and wide-based stance and had thoracic spinal rigidity with mild, asymmetric (left-greater-than-right-sided) scapular winging.

P1 demonstrated a notable improvement in strength over time. At the time of first clinical assessment at 3 ½ years of age, he was unable to maintain strength against gravity for neck and trunk flexion which subsequently improved to antigravity strength by 6 ½ years of age. Manual muscle testing revealed that muscle strength was full proximally and had improved distally, between 6 ½ years and 7 ½ years [Medical Research Council (MRC) grades] as follows: elbow extension 4/5 to 4+/5; wrist flexion 3/5 to 4/5; wrist extension 3/5 to 4/5; and hand grip 3+/5 to 4+/5 (within range of motion). Lower extremity distal strength remained weak with dorsiflexion 2/5, ankle inversion 3/5 and ankle eversion 3/5. At 7 ½ years, P1 was able to arise from the floor by rolling from supine to prone and arising without using his arms to a broad stance. He ambulated with a waddling-type gait with his left foot internally rotated and bilateral foot drop. He could accelerate his walking gait but could not achieve a run. Cognition was normal, and speech was hypophonic but fluent.

Serum creatine kinase (CK) level was performed at 1 year of age and was found to be mildly elevated (358 IU/L). Skeletal radiographic survey at age 3 years 9 months showed diffuse osteopenia, Wormian bones (intra-sutural bones) of the skull, and thinning with slight bowing and waviness of the bones with subtle metaphyseal changes (Fig. 2A–C), which seems to have improved on a repeat survey at age 6½ years (Fig. 2D). Muscle magnetic resonance imaging (MRI) performed of the lower extremities at 6 ½ years revealed relative preservation of the muscles above the knee except for some evidence of abnormally increased T1 signal within the vastus lateralis muscle bilaterally (Fig. 2E) compared to healthy control (Fig. 2G). In contrast, there was evidence of severely increased signal on T1-weighted sequences in all muscles below the knees bilaterally, suggestive of replacement of skeletal muscle with connective tissue and/or adipose tissue (Fig. 2F) compared to healthy control (Fig. 2H). Electromyography performed of the biceps brachii, vastus lateralis and tibialis anterior muscles was reported as normal and noted the very small size of the tibialis anterior muscle.

When the proband (P1) was 4 years of age, a subsequent male pregnancy (P2) was found to have bilateral clubfeet on fetal ultrasound at 19 weeks of gestation. A follow-up ultrasound performed at 21 weeks of gestation revealed abnormal posturing and a flaccid appearance of the hands and feet as well as a lack of fetal tone (Fig. 2I and J). Amniotic fluid volume appeared normal, and estimated fetal weight was 405 grams (20% percentile). The genetic etiology as well as the clinical prognosis of P1’s condition had remained undetermined at the time of this subsequent pregnancy, which the parents elected to terminate.

Family history was significant for two prior miscarriages and high myopia in the mother [−8 (OD) and −10 (OS)] and low myopia in the father [−2 (OD) and −2.5 (OS)]. There is one unaffected sibling age 12 months, who does not appear to be nearsighted; however, she has not had formal opthalmological testing. The maternal aunt and grandfather are reported to have moderate myopia, each between −4 and −6 diopters. There is a paternal uncle with low myopia [−2 (OD) and (−2.5 (OS)]; however, both paternal grandparents are not nearsighted.

Mutation analysis

Whole exome sequencing (WES) of P1 and P2 followed by variant filtering for call quality, rarity and predicted pathogenicity yielded two heterozygous mutations in P4HA1: a heterozygous splice donor site mutation (c.1543 + 2 T > G; NM_001017962.2) predicted to cause skipping of exon 12, and a heterozygous two-base pair insertion in exon 9 (c.1323_1324insAG; NM_001017962.2) predicted to result in a frameshift and premature stop codon (Fig. 3). Neither mutation had been previously reported, and neither was present in dbSNP, NHLBI EVS or ExAC Browser (16,17). Targeted parental segregation testing confirmed that the splice donor site mutation was inherited from the carrier mother, and the insertion mutation was inherited from the carrier father.

The effect of these mutations at the transcript level was further analysed using cDNA synthesized from RNA of dermal fibroblast of patients (P1, P2) and carrier parents. A pair of primers spanning the exon 12 transcript was used to confirm the effect of the maternal mutation on splicing (Supplementary Material, Table S1). Gel electrophoresis of the PCR fragment revealed an additional smaller band of equal intensity as the normal band in samples of P1, P2 and their mother, compared to the father and disease control (data not shown). Sequencing confirmed that the smaller band was derived from a mutant transcript in which exon 12 was skipped.

As P4HA1 exon 9 and exon 10 undergo mutually exclusive alternative splicing (5,13), cDNA synthesized from RNA of dermal fibroblasts were used to analyse the effect of the paternal two base pair insertion with specific primers for the exon 9 splice form (Supplementary Material, Table S1 and Fig. S3). In the exon 9 retaining transcript the two nucleic acid insertion in exon 9 is expected to result in a frameshift generating a premature stop codon within exon 9, so that the transcript is predicted to potentially be degraded by nonsense-mediated decay (NMD). Using cDNA prepared from untreated fibroblasts of P1, P2 and carrier father, the mutation was not detectable, consistent with the presence of NMD. Mutation analysis with the cDNA prepared from cycloheximide (CHX) treated fibroblasts to suppress NMD, then confirmed the presence of the c.1323_1324insAG mutation in both P1, P2 and carrier father.

Muscle biopsy analysis

Muscle histology from P1’s biopsy at age 2 years from the left quadriceps showed myofiber atrophy, with mild type 1 hypotrophy, an increase in interstitial endomysial connective tissue and mild fibrosis (Supplementary Material, Fig. S1). Immunohistochemical analysis of muscle tissue from P1 and P2 showed reduced immunoreactive staining for collagen IV as well as laminin γ1 in the muscle basement membrane (Fig. 5A).

C-P4H activity assay and western blot

The amount of total C-P4H activity in patient-derived fibroblasts was found to be reduced to approximately 50% of age-matched controls. There was only a slight reduction (approximately 10%) of activity in the mother’s fibroblasts and a normal activity in the father’s fibroblasts (Fig. 4A). C-P4H α(I) Western blot revealed a significant reduction of C-P4H α(I) protein in both affected siblings but no significant change in either parent compared to controls (Fig. 4B). In contrast, there was no significant difference in protein amount for C-P4H α(II) by Western blot analysis among all tested samples, indicating C-P4H α(II) expression was not up-regulated to compensate for the reduction of C-P4H α(I) expression (Fig. 4B).

When C-P4H α (I) Western blot analysis was performed with longer gel electrophoresis, two bands were revealed in the two affected siblings and the carrier mother (Supplementary Material, Fig. S2A). One band is of normal molecular weight and the other is smaller. As the c.1323_1324insAG mutation in paternal allele is located in alternatively spliced exon 9, the exon 10 containing transcript is expected to be intact. The normal sized band in two affected siblings is predicted to represent the protein product of the paternal exon 10 splice form, but with reduced expression level compared with age-matched disease controls which have both exon 9 and exon 10 splice forms. Based on the mutation analysis, the smaller band is predicted to correspond to the internally deleted protein translated from the maternal allele with exon 12 skipping. In both affected siblings, the smaller band had similar reduced intensity as the normal sized exon 10 splice form from the paternal allele; however, in the heterozygous mother, the smaller mutant band is much weaker than the normal band. The relatively lower expression level of the smaller mutant band in both affected siblings and their carrier mother indicates that the internally deleted C-P4H α(I) may be prone to degradation. We also confirmed by EndoH digestion that the molecular weight difference between the two bands is not due to abnormal N-glycosylation in the ER. (Supplementary Material, Fig. S2B)

Posttranslational modification and thermal stability of type I collagen

Type I collagen, the major collagen in skin and bone, would be predicted to also be affected by the reduction in C-P4H α(I) function. Migration of type I collagen synthesized from P1 and P2 fibroblasts, examined by SDS-urea PAGE, showed that the alpha 1 chain of type I collagen from both P1 and P2 fibroblasts migrated slightly faster than the control (Fig. 5B). As variations in collagen migration are frequently caused by changes in hydroxylation and glycosylation, hydroxylysine and hydroxyproline levels in the cultures were measured by amino acid analysis. Collagen proline 4-hydroxylation levels were decreased from 43.8% in control to 34% in P1, consistent with a decrease in C-P4H α(I). Interestingly, however, lysine hydroxylation was increased from 20.8% in control to 32.6% in P1 (Fig. 5C). This increase in lysine hydroxylation is not caused by slower collagen folding, since an intracellular assay of collagen folding did not reveal a difference, nor was collagen secretion found to be reduced in the context of the altered modification (Supplementary Material, Fig. S4).

Proline 4-hydroxylation is known to contribute significantly to collagen thermal stability, as collagens lacking 4-Hyp are not stable at 37ºC (18). Differential scanning calorimetry (DSC) revealed ∼1.3 °C decrease in the thermal stability of type I collagen in the patient samples. A broadened and skewed collagen denaturation peak indicated the presence of a heterogeneous mix with respect to extent and/or sites of the posttranslational modification (Fig. 5D).

Differential expression of the alternatively spliced exon 9 and 10 of P4HA1 in muscle tissue and dermal fibroblasts from individuals of different ages

Tissue and age-dependent differential expression of the two P4HA1 alternatively spliced forms containing either exon 9 or 10 was analysed with a semi-quantitative assay based on unsaturated PCR followed by Cac8I (specific for exon 9) or BlpI (specific for exon 10) digestion. This analysis revealed that in cultured dermal fibroblasts derived from individuals of different ages, the exon 9 and 10 splice forms were expressed at similar level at all tested ages (ranges from 8 months to adult, average ratio of exon 9/exon 10 = 0.92, SEM = 0.03) (Fig. 4C). In contrast, the expression level of the exon 9 splice form was higher compared to the exon 10 splice form in muscle biopsy tissue across all tested ages (ranges from 4 months to adult). However, the difference was more pronounced in muscle samples from individuals of younger age (<4.5 year, average ratio of exon 9/exon 10 = 4.22, SEM = 0.59) than those in individuals older than age 9 years (average ratio of exon 9/exon 10 = 2.05, SEM = 0.09) (Fig. 4C). This result suggests that the exon 9 splice form is the major form expressed in muscle tissue, in particular in younger ages (below approximately age 4.5 years), while the difference between the two alternatively splice forms becomes less significant with increasing age, even though the exon 9 splice form remains the major form.

Patient recruitment and sample collection

The study was approved by the Institutional Review Board of the National Institute of Neurological Disorders and Stroke, National Institutes of Health (Protocol 12-N-0095). Written informed consents were obtained by a qualified investigator. Medical history was obtained, and clinical evaluations were performed as part of the standard neurologic evaluation. DNA and tissue were obtained based on standard procedures.

Exome sequencing and mutation analysis

Whole exome sequencing on DNA obtained from blood (P1) and cultured skin fibroblast (P2) was performed at the NIH Intramural Sequencing Center using the Illumina (San Diego, CA) TruSeq Exome Enrichment Kit and Illumina HiSeq 2000 sequencing instruments. Variants were analysed using Varsifter and Seqr and searched for in dbSNP, NHLBI EVS, Exome Aggregation Consortium (ExAC Browser) (17,38).

Targeted P4HA1 mutation analysis was further performed with cDNA prepared from dermal fibroblast, which were treated with 0.2 mg/ml cycloheximide (CHX) (Sigma, Saint Louis, Missouri) for 19 h to suppress nonsense-mediated mRNA decay. Primers were designed to amplify the specific transcript spanning the alternatively splice exon 9 or exon 10 and the transcript spanning exon 12 (Supplementary Material, Table S1).

Muscle biopsy analysis

Cold methanol fixed 9 μm muscle cross-sections were incubated with primary antibodies at 4 °C overnight (Collagen IV (MAB1430, Millipore, Billerica, MA), Laminin γ1 (L9393, Sigma, Saint Louis, Missouri), the antibody labeling was detected with secondary antibodies for 1 h at room temperature [Alexa 568-conjugated goat anti-mouse IgG and Alexa 488-conjugated goat anti-rabbit IgG (Thermofisher, Rockford, IL)]. Prepared muscle sections were imaged with a Leica SP5 confocal microscope (Leica, Wetzlar, Germany).

C-P4H activity assay and western blot

Cultured fibroblasts isolated from the two affected siblings, the parents and age-matched controls were washed twice with PBS and lysed with 0.1M NaCl, 0.1M glycine, 0.1% Triton X-100, 10 mM Tris pH 7.8 supplemented with EDTA-free complete protease inhibitors (Roche, Indianapolis, IN). Soluble fraction was obtained by centrifugation and the protein concentration was measured using the Bradford assay (Bio-Rad, Hercules, CA). Total C-P4H activity was determined by an assay that measures formation of 4-hydroxy [¹⁴C] proline in a [¹⁴C]proline-labeled protocollagen substrate consisting of nonhydroxylated pro-α chains of chick type I procollagen (39).

EndoH (New England Biolabs, Ipswich, MA) digestion was performed according to manufacturer’s instructions. Aliquots of the lysates were analysed by Western Blot using 10% SDS-PAGE under reducing conditions followed by a transfer to PVDF membrane (Millipore, Billerica, MA) and incubation with antibodies against P4HA1 (12658-1-AP, Proteintech Group, Rosemont, IL) and P4HA2 (15). In order to resolve P4HA1 protein products from maternal and paternal allele, a 16cm long gel was used instead of a 5.5cm long one used in other assays.

Posttranslational modification and thermal stability of type I collagen

Cultured fibroblasts from the two affected siblings and a age-matched control were grown to confluence in 6-well culture dishes. Cells were incubated in serum-free media for 2 h, then labeled with 437.5 μCi/ml L-[2,3,4,5-³H] proline in serum-free media for 18 h. Media and cell layer collagens were collected, precipitated by ammonium sulfate, and analysed by 6% SDS-urea-PAGE. The media and cell layer samples were balanced for equal signal to examine difference in type I collagen migration. Amino acid analysis was performed on secreted type I collagen by high pressure liquid chromatography (AIBiotech, Richmond, VA) to quantitate levels of 4-hydroxyproline, proline, hydroxylysine and lysine.

Collagen secreted by cultured fibroblasts was purified by pepsin treatment (0.1 mg/ml in 0.5 M acetic acid) followed by two rounds of selective precipitation with 0.7 M NaCl. Thermal stability (denaturation thermograms) of collagen suspended in 0.2 M sodium phosphate, 0.5 M glycerol, pH 7.4 (to prevent aggregation) was measured by differential scanning calorimetry in a Nano III DSC instrument (Calorimetry Sciences Corporation, Lindon, Utah) at 0.125 °C/min heating rate from 10 to 50 °C (40).

Inclusion of the alternatively spliced exon 9 and exon 10 of P4HA1 in muscle tissue and in dermal fibroblasts from individuals of different ages

To study the expression of exon 9 and exon 10 P4HA1 splice forms, we designed a semi-quantificative assay based on unsaturated PCR and splice form specific restrictive enzyme (RE) digestion (Supplementary Material, Table S1 and Fig. S3). cDNA was prepared from total RNA isolated from muscle tissue or from dermal fibroblast cultures of individuals of different ages. Both exon 9 and exon 10 alternatively spliced transcripts were unsaturated-PCR amplified simultaneously with a pair of common primers that span both exon 9 and exon 10. Equal amounts of PCR product were digested with either Cac8I or BlpI (New England Biolabs, Ipswich, MA). Cac8I has a cutting site in exon 9, but not in exon 10, while BlpI can cut exon 10 sequence only. Therefore, the uncut band from Cac8I digestion is a transcript from the exon 10 containing splice form (exon 10 splice form for short), the uncut band from BlpI digestion is a transcript from the exon 9 containing splice form (exon 9 splice form for short). Exon 9 and exon 10 splice form specific PCR products were used as RE digestion control to monitor the digestion efficiency. After the RE digestion was complete according to digestion controls, cut and uncut bands were separated in 2% agarose gel. Images of the separated bands were taken with Geldoc (Biorad, Hercules, CA), and intensity of the following bands was measured using image studio software (Li-Cor Inc, Lincoln, NE): exon 9 splice form (BlpI uncut band) and exon 10 splice form (Cac8I uncut band). The relative ratio of two splice forms is calculated as: exon 9 splice form/exon 10 splice form.