Brain ventricular enlargement in human and murine acute intermittent porphyria

Abstract

The morphological changes that occur in the central nervous system of patients with severe acute intermittent porphyria (AIP) have not yet been clearly established. The aim of this work was to analyze brain involvement in patients with severe AIP without epileptic seizures or clinical posterior reversible encephalopathy syndrome, as well as in a mouse model receiving or not liver-directed gene therapy aimed at correcting the metabolic disorder. We conducted neuroradiologic studies in 8 severely affected patients (6 women) and 16 gender- and age-matched controls. Seven patients showed significant enlargement of the cerebral ventricles and decreased brain perfusion was observed during the acute attack in two patients in whom perfusion imaging data were acquired. AIP mice exhibited reduced cerebral blood flow and developed chronic dilatation of the cerebral ventricles even in the presence of slightly increased porphyrin precursors. While repeated phenobarbital-induced attacks exacerbated ventricular dilation in AIP mice, correction of the metabolic defect using liver-directed gene therapy restored brain perfusion and afforded protection against ventricular enlargement. Histological studies revealed no signs of neuronal loss but a denser neurofilament pattern in the periventricular areas, suggesting compression probably caused by imbalance in cerebrospinal fluid dynamics. In conclusion, severely affected AIP patients exhibit cerebral ventricular enlargement. Liver-directed gene therapy protected against the morphological consequences of the disease seen in the brain of AIP mice.

The observational study was registered at Clinicaltrial.gov as NCT02076763.

Introduction

Acute intermittent porphyria (AIP, MIM: #176000) is an autosomal dominant metabolic disease caused by the deficient activity of the porphobilinogen deaminase enzyme (PBGD; HGNC: 4982) in the liver, the third enzyme of the heme biosynthesis pathway (1–3). Patients exhibit neurovisceral attacks associated with marked overproduction and accumulation of porphyrin precursors, δ-aminolevulinic acid (ALA) and porphobilinogen (PBG). The liver has been confirmed as the major source of porphyrin precursors in AIP, and orthotopic liver transplantation effectively resolves the disease (4).

Acute attacks are characterized by severe abdominal pain, nausea, vomiting, hyponatremia, hypertension, tachycardia and bladder dysfunction, probably due to autonomic neuropathy. More severe symptoms include motor weakness and sensory disorders (5). Seizures, confusion, insomnia, anxiety and depression are central nervous system (CNS) manifestations present in some cases. And alterations similar to those occurring in posterior reversible encephalopathy syndrome (PRES) in association with cerebral vasoconstriction and edema develop in only a minority of cases (6–12). PRES typically improves with time or when AIP treatment is initiated and decreases the urinary excretion of ALA and PBG (10). Given the reportedly very low blood–brain barrier (BBB) permeability of the ALA compound (13–17), these data suggest the manifestations of PRES are presumably due to cerebral vasoconstriction (10) and the transient breakdown of the BBB and concomitant cytotoxic effect of porphyrin precursors of hepatic origin (7).

The aim of this work was to analyze brain structure in 8 patients with AIP suffering recurrent acute attacks (≥2 hospitalizations during the previous year) without clinical PRES. In addition, we studied the brain structural changes that occurred in AIP mice subjected to recurrent porphyric attacks for 8 months. Half of the mice were treated with a therapeutic recombinant adeno-associated virus vector (rAAV-PBGD), which overexpresses PBGD specifically in the liver (18).

Results

Brain and cerebral ventricular volumes were measured in 8 patients with severe AIP and 16 age- and sex-matched controls. Total brain volume (tBV) showed no differences between cases and controls but we observed in the former a significant increase in ventricular brain volume (VBV) and in the percentage of VBV per tBV (VBV/tBV) (Table 1 and Supplementary Material, Fig. S1A and B). This percentage did not correlate with age, disease duration or urine porphyrin precursor levels but showed a strong inverse correlation with the number of courses of heme injection (Spearman r = − 0.87; one-tailed P-value = 0.004). The genotype of the human peptide transporter 2 (PEPT2, UniProtKB-Q16348), which is highly expressed in choroid plexus epithelia and affects the tightness of binding of ALA (19) was also studied in these patients (Supplementary Table). The 4 patients with the highest ALA affinity PEPT2^*1/^*1 genotype showed no significant difference in VBV or the percentage of VBV/tBV when compared with the 4 individuals carrying PEPT2^*1/^*2.

Table 1

Total brain and ventricle volumes in 8 patients with severe AIP and 16 healthy controls

Patients with severe AIP									Healthy controls
ID	Age	Gender (M/F)	Total brain volume (tBV)[cc]		Ventricular brain volume (VBV)[cc]		VBV/tBV [%]		Gender	Age	ID
P1	47	F	1137		51.4		4.52
				1188		33.3		2.81	W	47	C1
				1081		24.3		2.25	W	48	C2
P2	63	F	1107		37.8		3.41
				1299		38.0		2.93	W	61	C3
				1277		32.6		2.55	W	63	C4
P3	53	M	1299		59.1		4.56
				1303		28.1		2.16	M	55	C5
				1198		31.9		2.67	M	55	C6
P4	42	M	1244		53.2		4.28
				1472		38.9		2.64	M	42	C7
				1424		38.3		2.69	M	44	C8
P5	40	F	1289		38.8		3.01
				1111		31.5		2.83	W	40	C9
				1150		30.2		2.63	W	41	C10
P6	38	F	980		38.4		3.91
				1168		22.6		1.93	W	37	C11
				1074		28.7		2.67	W	38	C12
P7	51	F	1040		25.0		2.40
				1212		28.8		2.37	W	52	C13
				1148		26.9		2.35	W	55	C14
P8	33	F	1153		35.7		3.10
				1295		27.4		2.12	W	35	C15
				1324		26.0		1.96	W	33	C16
Mean	46.6		1156	1233	42.41	30.46	3.65	2.47		46.62
Standard deviation	9.6		115	116	11.2	4.9	0.79	0.32		9.33
The Mann–Whitney test			0.14		0.01		0.001			0.8

Patients with severe AIP (P1-8), Healthy controls (C1-16), W: women and M:men, cc: Cubic centimeters.

No differences were observed in the percentage of VBV/tBV parameter between the 4 hypertensive and the 4 normotensive patients (Supplementary Table). Notably, all the patients showed high plasma Endothelin-1 (EDN1) values when compared to healthy volunteers and to asymptomatic AIP carriers exhibiting normal urinary excretion of porphyrin precursors (Fig. 1). Among the 8 patients with AIP, plasma EDN1 levels showed a weak correlation with urinary ALA (Spearman r = 0.68; one-tailed P-value = 0.04) but not with PBG excretion (Spearman r = 0.14; one-tailed P-value = 0.38).

Figure 1

Plasma endothelin-1 levels in healthy volunteers and patients with frequent attacks and asymptomatic AIP. Plasma Endothelin-1 levels were determined by ELISA immunoassay in 39 healthy volunteers (23 women and 16 men, age range: 29–65 years and mean 45 years), 8 patients (6 women and 2 men) with severe AIP and 8 women asymptomatic PBGD gene mutation carriers into long-term biochemical remission. The value in each patient with severe AIP corresponds to the average obtained (4 to 12 data/patient measured every 2 months) measured during the observational study NCT02076763. Results were plotted as mean ± SD. AIP, acute intermittent porphyria.

Cerebral blood flow (CBF) was measured twice in two of the 8 patients with severe AIP. In Patient 1 (a 47-year-old female), CBF experienced a reduction of 6.2% in the second study compared to the first (7 months apart) (Fig. 2A and B). At the time of the second measurement, this patient was emerging from an acute attack with a high increase of urinary values of porphyrin precursors (ALA: from 11.9 to 27.4 mmol ALA/mol creat; PBG: from 14.6 to 67 mmol PBG/mol creat.). On the other hand, Patient 2 (a 63-year-old female) was hospitalized because of an acute porphyria attack and elevated levels of urinary ALA (14.3 mmol/mol creat.) and PBG (52.8 mmol/mol creat.) just before the first CBF determination (Fig. 2C) and the second study was performed 8 years later when her clinical condition was significantly improved (Fig. 2D) (Urinary excretion: 8.8 mmol ALA/mol creat.; 14.4 mmol PBG/mol creat.). In this case, we observed an increase of 22.1% in CBF at the time of the second study. These results suggest that CBF decreases during acute attacks of porphyria perhaps due to increased levels of porphyrin precursors.

Figure 2

Reduced cerebral blood flow (CBF) during an acute attack in two women with severe AIP. A) Sagittal view of ASL perfusion MRI scan showing CBF map of AIP Patient 1 at the age of 45 years old and B) 7 months later at the end of an acute attack needing hospital admission and heme arginate administration (one month post-admission). Although with high levels of porphyrin precursor excretion, the patient did not show porphyria symptoms at the time of the first scan. AIP patient 2 had chronic hypertension. C) The first scan was performed during an acute attack that required hospital admission; and the second scan D) in an intercrisis period but with levels of excretion of urinary heme precursors over normal range (normal urinary levels are <4 mmol ALA/mol creat. and <1.5 mmol PBG/mol creat.). CBF is expressed as mean in ml min⁻¹ 100 g⁻¹.

To determine whether the CNS changes present in our patients were also reproduced in AIP mice we used magnetic resonance imaging (MRI) to analyze the size of cerebral ventricles in these animals. Unlike patients, AIP mice exhibit 30% of normal PBGD activity (Supplementary Material, Fig. S2A) and the urine ALA and PBG levels are only slightly elevated (Supplementary Material, Fig. S2B and C), unless a barbiturate is given (Supplementary Material, Fig. S2D and E). Indeed, we found that cerebral ventricular enlargement was also observed in AIP mice of 1 year of age that never received phenobarbital (Fig. 3). These data suggest that high ALA and PBG accumulation may not be essential for ventricular enlargement in AIP mouse strains with a strong PBGD deficiency in all tissues. Of note, we observed that recurrent phenobarbital-induced acute attacks on alternate weeks for 8 months further exacerbated ventricular enlargement (Fig. 3B, AIP + Phen). This could not be attributed to any direct toxic effect of repeated barbiturate administration since this drug did not significantly modify VBV in wild-type animals (Fig. 3B, WT + Phen). Importantly, in animals treated with rAAV-PBGD recurrent phenobarbital administration failed to increase the size of cerebral ventricles (Fig. 3). Indeed, the administration of a single dose of rAAV-PBGD protected AIP mice against hepatic aminolevulinate synthase (ALAS1) over-expression (Supplementary Material, Fig. S2F) and the consequent accumulation of heme precursors (Supplementary Material, Fig. S2D and E). These data suggest that the regulation of the hepatic heme synthesis pathway had been restored.

Figure 3

Ventricular enlargement in a 1-year-old wild-type and AIP mice with or without rAAV2/5-PBGD liver gene therapy. To document the role of the liver in CNS involvement, a cohort of AIP mice were injected with a single dose of recombinant adeno-associated virus vector (5 × 10¹² gc/kg, i.v. of rAAV2/5-PBGD). A) Representative T2-weight MRI in one animal from each of the wild-type (WT), AIP and AIP treated with rAAV-PBGD gene therapy (AIP-GT) groups. B) Ventricular volume measurements were performed as indicated in material and methods. Ventricular enlargement was exacerbated in AIP mice receiving biweekly phenobarbital challenges for 8 months. LV, Lateral Ventricles. WT, wild-type; AIP, acute intermittent porphyria; Phen, phenobarbital; GT, gene therapy. ^*P < 0.05; ^***P < 0.001 vs wild-type group. ^###P < 0.001 vs wild-type group receiving recurrent phenobarbital challenge.

A more in-depth morphological analysis was performed in order to determine whether a loss of neuronal tissue was the origin of the ventricular enlargement. Histological examination of brain slices from 6- to 7-month-old mice, which received three consecutive phenobarbital challenges did not show any evidence of ischemic areas or neuronal damage or death (Fig. 4A–D). In the gray matter of AIP mice, we observed a more dense neurofilament pattern in layers located in the periventricular areas, but no changes in neuronal cellularity or abnormalities in neural fiber structures (Fig. 4E–H). These data suggest that the enlarged ventricular size found in AIP mice cannot be related to brain atrophy or neuronal cell loss.

Figure 4

Brain histological analysis in consecutive coronal sections of wild-type and AIP mice to determine neuronal damage. Classical histological analysis to label neuronal populations and identify the basic neuronal structure (Nissl staining) was performed in wild-type (A,A´ and B,B´) and AIP (C,C´ and D,D´) mice (6– to 7– months old at sacrifice) after three consecutive phenobarbital challenges. High and medium molecular weight neurofilaments were stained by immunohistochemistry in wild-type (E, E´ and F, F´) and AIP (G, G´ and H, H´) mice in consecutive coronal sections. Ventricular atrophy and a narrowing of neurofilament layers located in the nearest region of the ventricle were observed in AIP mice. Scale bar: 1000 μm in A-H panels and 200 μm in A´-H´ panels.

In addition to the augmented size of cerebral ventricles, AIP mice showed a 39% reduction in CBF as compared to wild-type animals as revealed by arterial spin labeling (ASL) (Fig. 5A). Importantly, CBF returned to normal in AIP mice treated with rAAV-PBGD recalling the improvement of CBF occurring in patients upon remission of symptoms. In mice undergoing recurrent phenobarbital-induced attacks, CBF measurement performed 24 h after the last challenge showed a reduction in both wild-type and AIP animals (Fig. 5A).

Figure 5

Cerebral blood flow (CBF), plasma endothelin-1 levels and the expression of endothelin receptor Type B in wild-type and in AIP mice with or without rAAV2/5-PBGD liver gene therapy. A) CBF values were measured by ASL-MRI in total brain in the same groups of animals presented in Fig 3. Phenobarbital administration modified cerebral perfusion in all groups of animals. B) Plasma Endothelin-1 levels determined by ELISA immunoassay in wild-type and AIP mice treated or not with PBGD-liver gene therapy. Phenobarbital administration increased Endothelin-1 levels in all groups of animals. C) Hepatic expression of endothelin receptor type B (endrb) gene before and after a 4-d period of phenobarbital challenge in the indicated groups of animals. Endrb expression was measured in D) hippocampal, E) hypothalamus, and F) cortex brain regions in the indicated groups of mice that never received phenobarbital. AIP, acute intermittent porphyria; PBGD, porphobilinogen deaminase; GT, gene therapy; rAAV5, recombinant adeno-associated viral vector serotype 5.

Regarding edn1, plasma levels were similar in wild-type and AIP mice with or without rAAV-PBGD and in all groups the values increased similarly upon phenobarbital challenge (Fig. 5B). Given that endothelin receptor type B (EDNRB) plays a major role in the clearance of circulating EDN1 (20), we measured the expression of this receptor in the liver and different brain regions (Fig. 5C–F). The expression levels of the ednrb gene in the liver of AIP mice was lower than in wild-type and AIP mice treated with rAAV-PBGD, although the differences only reached statistical significance in the last group (Fig. 5C). Of note, while the administration of phenobarbital did not modify its liver expression in AIP mice, it caused an increase in those mice that received liver-directed gene therapy (Fig. 5C).

An interesting finding was that AIP mice exhibited an increase of ednrb in the CA1 hypothalamus region which is located adjacent to the ventricle space in the mouse brain (Fig. 6A-A´ and B-B´). The transcription levels of the ednrb gene were higher in the hippocampus (Fig. 5D) and hypothalamus (Fig. 5E) regions of AIP mice but levels in the brain cortex (Fig. 5F) were similar in both groups indicating that changes in ednrb expression were limited to specific regions of the brain of AIP mice. Of interest, ednrb over-expression in the CA1 hypothalamus was reversed in animals treated with rAAV-PBGD (Fig. 5E and Fig. 6C-C`).

Figure 6

Representative hippocampus sections of a wild-type mouse (A), an AIP mouse (B) and AIP mice treated with PBGD-liver gene therapy. Paraformaldehyde-fixed mouse brains from animals presented in figure 4 were immunostained for Endothelin B receptor (EDNRB) and glial fibrillary acidic protein (GFAP). EDNRB immunoreactivity was strong in the hippocampus in AIP mice but greatly reduced in AIP mice injected with a liver-specific rAAV-PBGD vector. EDNRB colocalized extensively with GFAP (the hallmark of reactive astrocytes). Alexa Fluor-labeled secondary antibody TOPRO-3 was used for nuclear staining. Scale bar: 650 μm in A, B and C panels and 20 μm in A’, B′ and C′ panels.

Discussion

Brain involvement in patients with severe AIP has been poorly characterized. Different reports have shown that CNS disturbances in AIP are non-specific and occur without a clear morphological counterpart (21). Current neuroanatomical understanding is based on post mortem histology studies (22) and brain MRI analysis in a small number of AIP patients that showed manifestations compatible with PRES (6–12,23–25). In the present report we explore CNS involvement in 8 patients with severe AIP, none of whom exhibited seizures or PRES, a condition which has been previously reported to be associated with ventricular enlargement (6–8,11,12,26). In 7 of our patients we found a significant increase of both ventricular volume and percentage of VBV/tBV. None of our patients showed any sort of cognitive impairment or the urinary incontinence (Supplementary Table) frequently found in patients with ventricular enlargement because of other causes such as normal pressure hydrocephalus (27).

In the eight patients in our study, the percentage of VBV/tBV was inversely correlated with the courses of heme infusions, but was independent of the PEPT2 genotype, urine ALA and PBG levels or number of attacks documented as hospital admissions per month.

Arterial hypertension has been reported to be associated with enlarged ventricular volume (28); and in patients with AIP, acute attacks coexist with increased blood pressure (29) which in some cases may become chronic with advanced age and progression of the disease (30,31). However, in our series only 3 of the 7 patients with high tBV showed chronic hypertension indicating that other factors come into play to cause the cerebral ventricular dilation. Notably, our patients with severe AIP exhibited high plasma EDN1 levels when compared to healthy controls and to patients with AIP who had been asymptomatic for a long period of time. Plasma EDN1 levels showed a weak but significant correlation with the accumulation of ALA precursor in the urine of 8 severely affected patients. The relationship between these two parameters is also underlined by the fact that EDN1 values returned to normal in two Swedish patients who entered biochemical and clinical remission following liver and kidney transplantation (Supplementary Material, Fig. S3, 32). Moreover, decreased CBF has been found in the two patients studied during acute attacks. Given the association between the ALA accumulation and circulating EDN1, the latter may act as a potent vasoconstrictor agonist regulating cerebral microcirculation (33).

Despite being normotensive (34), AIP mice showed a reduced development of ventricular dilatation, which was alleviated by the correction of the metabolic defect using liver-directed gene therapy. Repeated phenobarbital challenge exacerbated ventricular enlargement in AIP mice. However, a direct neurotoxic effect of ALA and PBG is an unlikely factor because porphyrin precursors are markedly accumulated in their liver and plasma, but not in the CNS as has been previously reported (35).

Upon barbiturate challenge, AIP mice developed arterial hypertension (34), increased edn1 plasma levels and reduced CBF. In a previous report, we showed that restoration of PBGD levels in the liver of AIP mice protects against hypertension (34). Here, we showed that treatment with rAAV-PBGD restored CBF. Our findings suggest microcirculation dysfunction during porphyria attack. Similarly, a previous report showed increased local vasoconstrictor responses in the mesenteric arteries in this mouse model with significant vasodilatation after hemin administration (36).

Given that the liver is an important site for circulating EDN1 clearance through the EDNRB (20); our data indicate that the correction of the metabolic defect in the liver, as measured by normal hepatic alas1 expression and normalization of porphyrin precursors after the treatment with rAAV-PBGD, counteracts the vascular effects of edn1 in AIP mice through induction of hepatic ednrb expression. Although downregulation of this receptor has been described in liver disorders (37) further studies are needed to understand the mechanism that regulates the expression of the ednrb receptor in a porphyric liver.

In the brain of our AIP mice, the CA1 region of the hippocampus located adjacent to the ventricle space exhibited high expression of ednrb. It seems possible that overexpression of ednrb might be a compensatory mechanism counteracting vasoconstrictive forces (38–40). Interestingly, normalization of CBF, and ednrb expression in the CA1 region, occurred in AIP mice receiving liver gene therapy.

Overall, our findings suggest that small vessel dysfunction in association with systemic arterial hypertension occurring during acute porphyria attacks might cause an imbalance in CSF dynamics resulting in ventricular enlargement. The imbalance could be due to increased pulse pressure in the ventricles via choroid plexus pulsation (41), abnormal fluid re-absorption by arachnoid villi (42) or both.

In conclusion, in the brain of patients with severe AIP as well as in AIP mice, we found cerebral ventricle enlargement seemingly as a result of vascular dysfunction. In the AIP animal model and in the few patients who were analyzed, this abnormality is associated with reduced CBF. In AIP mice correction of the metabolic disturbance in the liver corrected CBF and prevented the progressive cerebral ventricular enlargement occurring in animals subjected to repeated acute porphyric attacks.

Materials and Methods

Selection of subjects

Eight Caucasian patients (2 men and 6 women), diagnosed with AIP by clinical, biochemical data and genetic confirmation of the PBGD gene mutation, all suffering from recurrent acute attacks, were enrolled in a prospective observational study (NCT02076763) designed to collect clinical and laboratory data to later compare baseline and post-treatment variables in a gene therapy Phase 1 clinical trial (43). Two different phases were designed, a pretreatment ‘observational phase’ and an ‘interventional phase (NCT02082860)’, so that each patient served as her/his own control. Exclusion criteria for the NCT02076763 study were pregnancy, drug addiction or alcohol abuse, acute or chronic liver disease, kidney disorders, severe respiratory disease, severe autoimmune disease, severe acute active infections or presence of neutralizing antibodies against the rAAV5 virus vector. The 8 patients enrolled maintained high urinary levels of ALA and PBG for more than eighteen months previous to MRI study (Supplementary Table). In total, 6 patients were receiving AIP-specific treatment in a scheduled manner for the control of chronic symptoms while 2 received therapy for acute attacks only (Supplementary Table). All the patients showed a severe condition with at least two hospitalizations during the previous year due to the need for hospital treatment (Supplementary Table). One patient suffered two episodes of tetraparesis and another, a paraparesis in the past. In both cases, the paresis occurred years before the neuroimaging analysis and patients had made a complete recovery by the time of the study. No epileptic seizures or clinical PRES were reported. None of the patients showed dyslipidaemia, one had diabetes mellitus type I and four presented chronic hypertension. Individualized BMI, courses of heme infusions and number of hospitalizations due to AIP symptomatology are indicated in the Supplementary Table. A PEPT2 genotyping analysis was performed as described (19) using DNA from peripheral blood cells. Finally, psychological symptoms were assessed using the Beck Depression Inventory II (BDI-II) and the Beck Anxiety Inventory (BAI) questionnaires.

For the structural neuroimaging analyses, sixteen sex- and age-matched healthy controls (two healthy subjects for each patient) were recruited in the Department of Internal Medicine at the Clínica Universidad de Navarra School of Medicine Hospital. The volunteers were Caucasians and had no neurologic nor psychiatric diseases. The exclusion criteria were systemic vascular disease or cardiovascular risk factors (hypertension and diabetes, hypercholesterolemia, obesity and tobacco smoking). Participants gave separate written informed consent after the nature and possible consequences of the studies were explained. Their participation and informed consent was approved by the Ethics Research Committee of the University of Navarra (CEI 2018-071).

Magnetic resonance imaging

Image acquisition

All images were acquired at the Clínica Universidad de Navarra using a Siemens Trio 3T MRI scanner (Erlangen, Germany). High-resolution structural T1-weighted images were obtained. The characteristics of the sequence were: 3D MPRAGE, TR = 1620 ms, TE = 3.09 ms, FA = 15°, voxel size = 1 × 1 × 1 mm³) and for fluid-attenuated inversion recovery images the parameters were: TR = 8150 ms, TE = 125 ms, FA = 15°.

Brain and ventricular volumes of the eight AIP patients and sixteen healthy sex- and age-matched controls were quantified in native space (i.e. original image space without normalizing it with an anatomic template) using Amira V5.2 software (Thermo Fisher Scientific, Thermo Electron LED GmbH Zweigniederlassung Osterode. Am Kalkberg, Osterode am Harz, Germany). To that aim, using the T1 Images each ventricular system was segmented by intensity thresholding. Subsequently, the resulting masks containing the tBV and ventricular volume were analyzed with the statistics tool to obtain the number of voxels. The volume of each structure was calculated in cubic centimeters and the tBV/ventricular volume ratio was obtained. Finally, individual brain and ventricle 3D reconstructions and visualization were performed from the obtained masks.

Brain perfusion measurements

Brain perfusion studies were performed in two women with AIP. Perfusion measurements were carried out using ASL-MRI, a technique that allows quantification of blood flow in physiological units of mL/min/100 g of brain tissue. Perfusion weighted images were obtained using the pulse ASL (PASL) technique. The PASL sequence consisted of a modified version of the flow-sensitive alternating inversion recovery (FAIR) technique with a saturation pulse applied after the global or slice-selective inversion (44). Label and control images were acquired using a 2D gradient-echo echo planar imaging sequence, with the following imaging parameters: in-plane resolution = 4 × 4 mm², FOV = 256 × 256 mm², slice thickness = 6 mm, 16 axial slices, TR/TE = 3500/17 ms, BW = 2895 Hz/pixel. Anatomical images (T1 weighted, T2 weighted and FLAIR) were also acquired in the same scan session. The preprocessing steps for the PASL images included motion correction, co-registration to anatomical images, subtraction of label and control to yield perfusion weighted images, and quantification of blood flow. Absolute quantification was performed using a single compartment model (45).

Experimental studies in AIP mice

Compound heterozygote T1/T2 (AIP) mice showed 30% of normal PBGD activity and a slightly elevated urine ALA and PBG levels (46). T1 and T2 mice with heterozygous mutations do not develop the disease but the biallelic model replicates the drug-precipitated biochemical abnormalities of acute porphyria in humans (18). To overload the deficient enzymatic step and biochemically imitate severe human AIP; half of the AIP animals received a total of 16 phenobarbital challenges (doses of 75, 80, 85 and 90 mg/kg/day administered intraperitoneally), induced twice a month for 8 months. In order to evaluate the role of hepatic PBGD deficiency in CNS involvement, two additional AIP mouse groups (with and without phenobarbital challenges) received a single dose of 5 × 10¹² gc/kg of rAAV2/5-PBGD vector in order to enhance hepatic PBGD activity. Liver-targeted gene therapy was performed in 8- to 12-week-old mice, one month before the start of the phenobarbital challenges. Experimental protocols were performed according to European Council Guidelines and approved by the local Animal Ethics Committee.

MRI studies in the brain of AIP mice

Ventricular size was measured by brain MRI in a 1-year-old wild-type and AIP mice treated or not with liver gene therapy. MRI studies were carried out in a Bruker Biospec 70/20 scanner using a combination of a linear coil (for transmission) with a mouse head phase array coil (for reception). Animals were anesthetized with sevoflurane (5% for induction and 3% for maintenance) and placed in an MRI-adapted stereotaxic holder with a water circulating blanket to maintain body temperature. Respiration and body temperature were continuously monitored. As an anatomical reference, we acquired a T2-weighted axial sequence (TR = 4200 ms; TE, 33 ms; α = 180°; FOV = 1.8 × 1.8 cm; matrix = 256 × 256; slice thickness = 0.5 mm, number of slices = 28).

Cerebral blood flow (CBF) measurements were made in the same 1-year-old animals using ASL (47,48). CBF maps were calculated using a FAIR-EPI sequence using selective and global inversion (TR: 15 000 ms, TE: 16.4 ms, α = 90°; FOV = 1.8 × 1.8 cm; matrix = 80 × 64; slice thickness = 1 mm, inversion recovery time (TIR) = 25 ms, number of TIR values = 22, number of slices = 1). The value of CBF was calculated according to the formula:

$$\mathrm{CBF}=\lambda \left(\frac{1}{T_{1,\mathrm{sel}}}-\frac{1}{T_{1,\mathrm{nosel}}}\right),$$

where λ represents the blood/brain distribution coefficient (0.9 mL/g) (49).

Structural studies in the brain of AIP mice measured by brain histology

Adult female mice received a total of three phenobarbital challenges on alternate weeks. Two weeks after the last challenge, animals were perfused with 4% paraformaldehyde. Then, brains were immersed in para-formaldehyde overnight and, finally, in a mixture of 20% Glycerol +2% dimethyl sulfoxide for 48 h. Classical histological analyses were performed with Nissl stain for cells or Fibers. The hippocampus sections of mouse brain slices were immunostained for EDNRB (GenBank accession number NM_007904; Abcam, Cambridge, UK, ref: ab117529) and glial fibrillary acidic protein (GFAP; DakoCytomation, Glostrup, Denmark, ref: Z0334).

Measurement of plasma Endothelin-1 and gene expression of endothelin receptor type B and aminolevulinate synthase

Plasma EDN1 (UniProtKB-Q6FH53) measurements in human and mice were made using a quantitative enzyme-linked immunosorbent assay (R&D systems, Abingdon, UK; ref: DET100). A gene-specific SYBR-Green I-based polymerase chain reaction (PCR) assays were designed for murine EDNRB (HGNC: 3180) and hepatic ALAS1 (EC 2.3.1.37, HGNC: 396) messenger RNA. The relative transcript level was determined using primers annealing specific complementary DNA sequences of EDNRB (forward primer: 5′-AAGAATGCCCAAGAGAAAAC-3′, reverse primer: 5′-AAAAAGGAAGGAAGGAAAATC-3′, product length of 239 bp) and ALAS1 (forward primer: 5´-CAAAGAAACCCCTCCAGCCAATGA-3′, reverse primer: 5´-GCTGTGTGCCGTCTGGAGTCTGTG-3′, product length of 104 bp). A real-time PCR experiment was run on an iQ5 real-time PCR detection system (Bio-Rad, Hercules, CA, USA). PCR product was confirmed with agarose gel electrophoresis and melting curve analysis. The amount of gene transcripts was calculated as the n-fold difference relative to the control mouse using the β-actin gene as an internal control: 5′-CGCGTCCACCCGCGAG-3′ (forward), and 5′-CCTGGTGCCTAGGGCG-3′ (reverse, product length: 194 bp). Results are expressed according to the formula 2^{ΔCt(Actin) − ΔCt(EDNRB)}, where ΔCt represents the difference in threshold cycle between genes.

Statistical analysis

The results were plotted as mean ± standard deviation. The normality of the brain volumes and ventricular system from human subjects were examined using Shapiro–Wilk tests and the Levene test. Ratios between individual brain and ventricular volumes were calculated for both AIP and healthy controls. The Mann–Whitney test was applied using the JASP program 0.8.1.2 for Windows, for group comparison. Statistically significant differences were established with a P-value < 0.01.

The statistical analyses of experimental studies were performed using GraphPad Prism® 5 (GraphPad Software, Inc., La Jolla, CA). Before the analysis, data were transformed using the formula Log (1 + x) to normalize variances. Comparisons between two groups were analyzed by Student t-tests. In the case of comparisons involving more than two groups, data were analyzed using two-way analysis of variance, and pairwise comparisons were made using Bonferroni’s Multiple Comparison Tests. The statistical significance of differences was tested with two-tailed t-tests and P values of <0.05 indicated statistically significant differences.

Acknowledgments

The authors gratefully acknowledge Elvira Roda and Alberto Rico for their excellent work in the histologic preparation of the experimental samples and Pauline Harper and Eliane Sardh from the Porphyria Center Sweden, Karolinska University Hospital, Stockholm (Sweden) for supplying plasma samples from patients with porphyria before and after kidney and liver transplantation. Original T1 and T2 mouse strains were provided by Prof. U.A. Meyer (Biozentrum of University of Basel, Switzerland).

Conflict of Interest statement.None declared.

Funding

The Spanish Institute of Health Carlos III (Fondo de Investigación en Salud) (cofunded by European Regional Development Fund grant numbers PI15/01951 and PI18/00860); the Spanish Fundación Mutua Madrileña de Investigación Médica; the Spanish Fundación Eugenio Rodriguez Pascual; European Regional Development Fund/Ministerio de Ciencia; Innovación y Universidades – Agencia Estatal de Investigación [BFU2017-82407-R and RTI2018-094494-B-C22]; the Spanish Fundación Federación Española de Enfermedades Raras para la Investigación de Enfermedades Raras, the Department of Health of the Government of Navarra [046-2017_NAB7 and 0011-1383-2019-000006 (PI031)]; the European Commission 7^th Framework Programme; AIPGENE (Grant 261506); the Comunidad de Madrid [S2017/BMD-3867 RENIM-CM] (co-financed by European Structural and Investment Fund); the Instituto de Salud Carlos III (ISCIII, to the Centro Nacional de Investigaciones Cardiovasculares (CNIC)); the Ministerio de Ciencia, Innovación y Universidades (MCNU); and the Pro CNIC Foundation; Severo Ochoa Center of Excellence [SEV-2015-0505].

Declaration

The financial sponsors had no role in the analysis or the development of conclusions. The investigators are solely responsible for the content and the decision to submit the manuscript for publication.

Author notes

The authors wish it to be known that, in their opinion, the first 5 authors should be regarded as joint first authors: D.J., E.O.L., L.C., M.A. F-S and X.M.

These authors jointly supervised this work: J.L.L., D.D., G.G-A., M.A.P., M.D. and A.F.