Phenylbutyrate up-regulates the adrenoleukodystrophy-related gene as a nonclassical peroxisome proliferator

Abstract

X-linked adrenoleukodystrophy (X-ALD) is a demyelinating disease due to mutations in the ABCD1 (ALD) gene, encoding a peroxisomal ATP-binding cassette transporter (ALDP). Overexpression of adrenoleukodystrophy-related protein, an ALDP homologue encoded by the ABCD2 (adrenoleukodystrophy-related) gene, can compensate for ALDP deficiency. 4-Phenylbutyrate (PBA) has been shown to induce both ABCD2 expression and peroxisome proliferation in human fibroblasts. We show that peroxisome proliferation with unusual shapes and clusters occurred in liver of PBA-treated rodents in a PPARalpha-independent way. PBA activated Abcd2 in cultured glial cells, making PBA a candidate drug for therapy of X-ALD. The Abcd2 induction observed was partially PPARalpha independent in hepatocytes and totally independent in fibroblasts. We demonstrate that a GC box and a CCAAT box of the Abcd2 promoter are the key elements of the PBA-dependent Abcd2 induction, histone deacetylase (HDAC)1 being recruited by the GC box. Thus, PBA is a nonclassical peroxisome proliferator inducing pleiotropic effects, including effects at the peroxisomal level mainly through HDAC inhibition.

Figure 1.

Expression of the Abcd2 and Aox genes is up-regulated by PBA in liver. (A) RNA samples from liver and brain of a control rat and from a rat treated per os with PBA (dose ×1.5) for 9 wk were analyzed by RT-PCR in duplicates. (B) RNA was extracted from liver of control rats (n = 10) and rats treated with PBA (dose ×1) for 6–7 wk (n = 6) or PBA (dose ×1.5) for 6–9 wk (n = 4). Samples were individually analyzed twice by RT-PCR. The intensities for Abcd2 and Aox mRNA were quantified by digital imaging and normalized using glyceraldehyde-phosphate dehydrogenase (Gapdh) intensities. Data (means ± SD) are presented as fold induction in relation to the values for control rats (taken equal to 1 and not depicted in the figure). (C) 2 mg sodium ciprofibrate (Cipro) or 20 mg PBA in 100 μl of water was injected into the cerebral fourth ventricle (100 μl/h). Control rats received equiosmolar amounts of NaCl. RNA was extracted from the whole brain and analyzed individually by Northern blotting (n = 4 for each group of rats).

Figure 2.

PBA induces peroxisome proliferation in liver. (A–D) Peroxisomes were stained for catalase activity. Liver of a control rat (A) and a rat treated per os with PBA (dose ×1.5) for 6 wk (B). Note peroxisome clusters and a phi-shaped peroxisome (arrowhead) in B. (C) Nearly touching peroxisomes suggesting a fission or budding process (Roels, 1991) in liver of a rat treated with PBA (dose ×1) for 6 wk. (D) Four phi-shaped peroxisomes in liver of a rat treated as in C. Unstained nucleoids are clearly visible in several peroxisomes. (E) Immunogold localization of AOX in a phi-shaped peroxisome, adjacent to an unlabeled mitochondrion, in liver of a rat treated as in C. Bars: (A, B, D, and E) 1 μm; (C) 100 nm.

Figure 3.

PBA modifies peroxisome distribution in ependymal cells of the fourth ventricle. Peroxisomes were stained for catalase activity. Peroxisomes in cells of a control rat (A) or a rat treated per os with dose ×1 of PBA for 6 wk (B). Peroxisomes are small and few when compared with liver (Fig. 2). Peroxisome clusters as shown in B were more frequently observed in treated animals. Cilium is seen in A. Bars, 200 nm.

Figure 4.

Peroxisomal genes are inducible by PBA and butyrate in hepatocytes. Rat primary hepatocytes were cultured in the presence of different concentrations of PBA, butyrate, and ciprofibrate for 72 h. Total RNA was analyzed by RT-PCR for Abcd2 and Aox expression. Results are presented as fold induction in relation to the control values for cells cultured in the absence of drug. Control data, taken equal to 1, are not reported in the figure. Data represent means ± SD of two independent experiments.

Figure 5.

Abcd2 is inducible by PBA and butyrate in glial cells. Differentiated CG-4 oligodendrocytes, pure primary astrocytes, and mixed primary cultures of oligodendrocytes and astrocytes were treated with different concentrations of PBA (A) and butyrate (B) for 72 h. Gene expression was analyzed by RT-PCR. Results are means ± SD of three independent experiments and presented as in Fig. 4.

Figure 6.

PBA causes Abcd2 induction and peroxisome proliferation in PPARα −/− cells. Fibroblasts (A) and hepatocytes (B) were prepared from fetal and adult liver, respectively, of PPARα +/+ and PPARα −/− mice and exposed to PBA or fenofibrate (Feno) for 72 h. Gene expression was analyzed by RT-PCR. Data are means ± SD of two to five independent experiments and presented as in Fig. 4. Untreated (C) or 2.5 mM PBA-treated (D) hepatocytes from PPARα −/− mice were examined for the number of peroxisomes by LM after anti-catalase protein A–gold immunolabeling with silver enhancement. Bars, 20 μm.

Figure 7.

TSA, a potent HDAC inhibitor, induces Abcd2 expression. Rat primary hepatocytes were treated with different doses of TSA and for different times. Abcd2 expression was analyzed by RT-PCR. Results from three independent experiments are presented as fold induction of the control values (equal to 1) for untreated cells.

Figure 8.

The GC and CCAAT boxes control basal expression and allow full induction of the basic Abcd2 promoter by PBA and butyrate. COS-7 cells were cotransfected with constructs containing normal or mutated Abcd2 promoter upstream of the luciferase reporter gene and pCMV-βGal to allow normalization of the luciferase activity. Mutations of the GC and CCAAT boxes and deletions are indicated in the schematic diagram. The cells were cultured with or without different doses of PBA or butyrate. (A) Analysis of the dose-dependent effect of PBA and butyrate. (B) Analysis of the role of the GC boxes. (C) Analysis of the role of the CCAAT box and of 5′UTR sequence. In each experiment, results are presented in relation to the luciferase activity of untreated p277-transfected cells taken equal to 1. Fold induction showing the effect of the treatment for each construct is indicated. Data are shown as means ± SD of three to six independent experiments performed in triplicate wells.

Figure 9.

Sp1/Sp3 and NF-Y bind to the GC1 and CCAAT boxes and HDAC1 is recruited through the GC1 box. (A) EMSA was performed using the ³²P-labeled GC1 or CCAAT oligonucleotides with 10 μg of nuclear extracts (NE) from rat liver. Competition experiments were performed with a 50-fold molar excess of unlabeled oligonucleotides. Supershifts were performed using specific antibodies or serum for control. Retarded complexes are indicated. (B) DAPA was performed with the same nuclear extracts using the biotinylated normal or mutated GC1 oligonucleotide. Complexed proteins were resolved by SDS-PAGE and revealed by Western blot using an anti-HDAC1 antibody.