(A) Schematic representation of human 4-hydroxyphenylpyruvate dioxygenase-like protein (HPDL) with the mitochondrial targeting sequence (MTS), the two vicinal oxygen chelate (VOC) domains, and the three Fe binding sites. All the HPDL variants described in the literature are reported. The variants identified in our study are shown in bold black type, and the novel ones in bold red type. (B) Immunoblot analysis of HPDL in skin fibroblasts from patients and control (CTRL) individuals. Determination of β-tub content was used as a normalization control. The variants identified in each patient were the following: Gly50Asp/Gly50Asp in F-A-ID1, F-A-ID2, F-E-ID6, F-M-ID12; Phe31Leu/Gly278Ser in F-B-ID3; Ser49Arg/Ser49Arg in F-C-ID4; Gly50Asp/Ser173Cys in F-D-ID5; Trp358Gly/Trp358Gly in F-G-ID7, F-G-ID8, F-G-ID9; Ala86ArgfsTer45/Ile266Thr in F-H-ID10; Leu234GlyfsTer94/Ser172Asn in F-L-ID11. (C) Oxygen consumption rate (OCR) measured using a Seahorse XF analyzer. HPDL patients presented altered OCR with reduced maximal respiration and reduced spare respiratory capacity (SRC). (D) Increased reactive oxygen species (ROS) production in HPDL patients' skin fibroblasts versus controls after TBHP (tert-butyl hydroperoxide, an inducer of ROS production in cells) administration (RM = regular medium). (E) (left) Plasma glial fibrillary acidic protein (GFAP) levels in HPDL patients and controls. Boxplots within violin plots show median (thick horizontal line), interquartile range (thin horizontal lines), and whiskers extending to 1.5 × IQR. Statistical analyses demonstrated significantly higher GFAP levels in HPDL patients compared with age-matched controls. Correlation plots between GFAP levels and (center) the localization variants within the putative domains of HPDL protein, (right) patients' phenotypes (mild/intermediate/severe). (F) Dysregulated metabolites identified in patients' serum through variable importance in projection (VIP) score analysis. Pink bars represent metabolites found to be decreased in HPDL patients, while green bars represent metabolites found to be increased. In all graphs, asterisks denote statistically significant differences: *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ***P < 0.001, whereas “ns” indicates non-significant results.

Representative neuroimaging in HPDL. Brain MRI findings in five HPDL patients (rows A–E). As indicated using roman numerals: T1-weighted mid-sagittal images (I), coronal T2-weighted images at the level of the medulla passing through the olivary nuclei (II) and cerebellar hemispheres (III), axial T2-FLAIR (IV) and T2-weighted (V) images at the level of the bulb. Patient F-H-ID10 (row D) is already reported elsewhere.43 Patients F-A-ID1 (row E) and F-A-ID2 (row B) did not show significant MRI alterations (only a slightly dysmorphic posterior portion of the corpus callosum in F-A-ID1, row E, see yellow arrowhead). Patient F-D-ID5 (row C) had very mild cerebellar atrophy with slight ventriculomegaly and moderate global cerebral atrophy. Patient F-C-ID4 (row A) showed a hypoplastic corpus callosum (yellow arrowhead: column I) and global cerebral atrophy. Patients F-C-ID4 and F-H-ID10 had cerebellar atrophy, T2 hyperintensities in the medulla oblongata (blue arrow: columns II, IV and V, rows A and D), middle cerebellar peduncle atrophy with T2 hyperintensity in the sub-dentate white matter (orange arrow: column III, rows A and D), and reduction of the cerebellar white matter volume and ventriculomegaly

In vivo studies in zebrafish. (A) Burst activity in zebrafish embryos at 30 hpf (hours post-fertilization) is decreased in hpdl-F0 versus controls (hpdl-F0, n = 70; ctrls, n = 83). (B) Automated analysis of spontaneous motor activity revealed an increase in swim distance and velocity in hpdl-F0 at 120 hpf compared with control siblings (hpdl-F0, n = 106; ctrls, n = 106). (C) Detection of dying cells by acridine orange staining in head and tail of control and hpdl-F0 embryos at 24 hpf. Arrows indicate the higher number of dying cells in hpdl-F0 embryos compared with controls. (D) (left) Representative fluorescence images of reactive oxygen species (ROS) generation in zebrafish larvae at 48 hpf (left). Quantitative analysis of ROS production (right). In all graphs, asterisks denote statistically significant differences: *P ≤ 0.05, **P ≤ 0.01.

Drug tests. (A) Analysis of reactive oxygen species (ROS) production in two 4-hydroxyphenylpyruvate dioxygenase-like protein (HPDL) patients' fibroblasts after 1 week of treatment with 4-hydroxybenzoic acid (4HB) 1 mM and 4-hydroxymandelic acid (4HMA) 5 mM. Data showed a significant effect of 4HMA on ROS production, while no effects of 4HB treatment were observed. Data are the averages of two experiments. (B) 4HB 1 μM treatment improved burst activity of hpdl-F0 embryos versus controls; 4HMA 5 μM treatment normalized the phenotype of hpdl-F0 embryos (controls, n = 114; controls+4HB, n = 115; controls+4HMA, n = 51 hpdl-F0, n = 85; hpdl-F0+4HB, n = 74; hpdl-F0+4HMA, n = 42). (C) Locomotor behavior (distance moved and velocity) of hpdl-F0 versus controls at 5 dpf (days post-fertilization) was rescued with 4HMA 5 μM treatment, but not with 4HB 1 μM treatment (controls, n = 107; controls+4HB, n = 114; controls+4HMA, n = 111 hpdl-F0, n = 69; hpdl-F0+4HB, n = 57; hpdl-F0+4HMA, n = 44). In all graphs, asterisks denote statistically significant differences: *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001, whereas “ns” indicates non-significant results.