Mice Lacking Phosphatidylinositol Transfer Protein-α Exhibit Spinocerebellar Degeneration, Intestinal and Hepatic Steatosis, and Hypoglycemia
2003; Elsevier BV; Volume: 278; Issue: 35 Linguagem: Inglês
10.1074/jbc.m303591200
ISSN1083-351X
AutoresJames G. Alb, Jorge D. Cortese, Scott E. Phillips, Roger L. Albin, Tim R. Nagy, Bruce A. Hamilton, Vytas A. Bankaitis,
Tópico(s)Endoplasmic Reticulum Stress and Disease
ResumoPhosphatidylinositol transfer proteins (PITPs) regulate the interface between lipid metabolism and cellular functions. We now report that ablation of PITPα function leads to aponecrotic spinocerebellar disease, hypoglycemia, and intestinal and hepatic steatosis in mice. The data indicate that hypoglycemia is in part associated with reduced proglucagon gene expression and glycogenolysis that result from pancreatic islet cell defects. The intestinal and hepatic steatosis results from the intracellular accumulation of neutral lipid and free fatty acid mass in these organs and suggests defective trafficking of triglycerides and diacylglycerols from the endoplasmic reticulum. We propose that deranged intestinal and hepatic lipid metabolism and defective proglucagon gene expression contribute to hypoglycemia in PITPα– / – mice, and that hypoglycemia is a significant contributing factor in the onset of spinocerebellar disease. Taken together, the data suggest an unanticipated role for PITPα in with glucose homeostasis and in mammalian endoplasmic reticulum functions that interface with transport of specific luminal lipid cargoes. Phosphatidylinositol transfer proteins (PITPs) regulate the interface between lipid metabolism and cellular functions. We now report that ablation of PITPα function leads to aponecrotic spinocerebellar disease, hypoglycemia, and intestinal and hepatic steatosis in mice. The data indicate that hypoglycemia is in part associated with reduced proglucagon gene expression and glycogenolysis that result from pancreatic islet cell defects. The intestinal and hepatic steatosis results from the intracellular accumulation of neutral lipid and free fatty acid mass in these organs and suggests defective trafficking of triglycerides and diacylglycerols from the endoplasmic reticulum. We propose that deranged intestinal and hepatic lipid metabolism and defective proglucagon gene expression contribute to hypoglycemia in PITPα– / – mice, and that hypoglycemia is a significant contributing factor in the onset of spinocerebellar disease. Taken together, the data suggest an unanticipated role for PITPα in with glucose homeostasis and in mammalian endoplasmic reticulum functions that interface with transport of specific luminal lipid cargoes. PITPs 1The abbreviations used are: PITPs, phosphatidylinositol transfer proteins; CHOP, CCAAT/enhancer-binding protein homology protein; CL, cardiolipin; CRD, chylomicron retention disease; DAG, diacylglycerol; ER, endoplasmic reticulum; ES cells, embryonic stem cells; FFA, free fatty acids; GFAP, glial fibrillary acidic protein; GM, gray matter; Glc-6-Pase, glucose-6-phosphatase; MEFs, murine embryonic fibroblasts; PIP, phosphoinositide; PtdCho, phosphatidylcholine; PtdIns, phosphatidylinositol; TG, triglyceride; TUNEL, terminal deoxynucleo-tidyltransferase-mediated dUTP nick-end-labeling; UPR, unfolded protein response; WM, white matter; CE, cholesteryl ester. mobilize PtdIns or PtdCho between membrane bilayers in vitro (1Cleves A.E. McGee T.P. Bankaitis V.A. Trends Cell Biol. 1991; 1: 30-34Abstract Full Text PDF PubMed Scopus (134) Google Scholar, 2Wirtz K.W.A. Annu. Rev. Biochem. 1991; 60: 73-99Crossref PubMed Scopus (380) Google Scholar). In vivo studies demonstrate that PITPs control the interface between membrane trafficking and lipid metabolic pathways in yeast (3Bankaitis V.A. Aitken J.R. Cleves A.E. Dowhan W. Nature. 1990; 347: 561-562Crossref PubMed Scopus (438) Google Scholar, 4Cleves A.E. McGee T.P. Whitters E.A. Champion K.M. Aitken J.R. Dowhan W. Goebl M. Bankaitis V.A. Cell. 1991; 64: 789-800Abstract Full Text PDF PubMed Scopus (286) Google Scholar, 5Kearns B.G. Alb Jr., J.G. Bankaitis V.A. Trends Cell Biol. 1998; 8: 276-282Abstract Full Text Full Text PDF PubMed Scopus (76) Google Scholar, 6Yanagisawa L. Marchena J. Xie Z. Li X. Poon P.P. Singer R. Johnston G. Randazzo P.A. Bankaitis V.A. Mol. Biol. Cell. 2002; 13: 2193-2206Crossref PubMed Scopus (68) Google Scholar). By contrast, the physiological functions for mammalian PITPs, which are structurally unrelated to yeast PITPs (7Sha B. Phillips S.E. Bankaitis V.A. Luo M. Nature. 1998; 391: 506-510Crossref PubMed Scopus (232) Google Scholar, 8Yoder M.D. Thomas L.M. Tremblay J.M. Oliver R.L. Yarbrough L.R. Helmkamp Jr., G.M. J. Biol. Chem. 2001; 276: 9246-9252Abstract Full Text Full Text PDF PubMed Scopus (111) Google Scholar), are not understood at either the cellular or organismal levels. Mammals express at least three soluble PITPs: PITPα, PITPβ, and rdgBβ (9Dickeson S.K. Lim C.N. Schuyler G.T. Dalton T.P. Helmkamp Jr., G.M. Yarbrough L.R. J. Biol. Chem. 1989; 264: 16557-16564Abstract Full Text PDF PubMed Google Scholar, 10Tanaka S. Hosaka K. J. Biochem. (Tokyo). 1994; 115: 981-984Crossref PubMed Scopus (84) Google Scholar, 11Fullwood Y. dos Santos M. Hsuan J.J. J. Biol. Chem. 1999; 274: 31553-31558Abstract Full Text Full Text PDF PubMed Scopus (39) Google Scholar). PITPα and PITPβ share 77% primary sequence identity, are encoded by distinct genes, and exhibit biochemical differences. Yet both PITPα and PITPβ (and even yeast PITPs) function as soluble factors that stimulate various reconstitutions of PIP-dependent functions in permeabilized mammalian cells. These functions include regulated and constitutive membrane trafficking and phospholipase C-dependent signaling through G-protein-coupled receptors (12Hay J.C. Martin T.F.J. Nature. 1993; 366: 572-575Crossref PubMed Scopus (308) Google Scholar, 13Ohashi M. de Vries K.J. Frank R. Snoek G. Bankaitis V. Wirtz K. Huttner W.B. Nature. 1995; 377: 544-547Crossref PubMed Scopus (169) Google Scholar, 14Cunningham E. Tan S.K. Swigart P. Hsuan J. Bankaitis V.A. Cockcroft S. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 6589-6593Crossref PubMed Scopus (99) Google Scholar). Given the lack of PITP specificity in these assays, it remains unclear how faithful such reconstitutions are in reporting physiological functions for mammalian PITPs. Genetic studies are providing initial clues regarding PITP function in metazoans. An inherited form of light-enhanced retinal degeneration in Drosophila results from inactivation of a membrane-bound PITP (15Milligan S.C. Alb Jr., J.G. Elagina R.B. Bankaitis V.A. Hyde D.R. J. Cell Biol. 1997; 139: 351-363Crossref PubMed Scopus (126) Google Scholar). In mice, reduction of PITPα to 18% of wild-type levels is the basis for the vibrator neurodegenerative disorder (16Hamilton B.A. Smith D.J. Mueller K.L. Kerrebrock A.W. Bronson R.T. van Berkel V. Daly M.J. Kruglyak L. Reeve M.P. Nemhauser J.L. Hawkins T.L. Rubin E.M. Lander E.S. Neuron. 1997; 18: 711-722Abstract Full Text Full Text PDF PubMed Scopus (173) Google Scholar, 26Passonneau J. Lauderdale V. Anal. Biochem. 1974; 60: 405-412Crossref PubMed Scopus (627) Google Scholar). Gene ablation approaches suggest PITPβ plays an essential housekeeping function, whereas PITPα is nonessential for ES cell viability and is not a quantitatively significant factor in membrane trafficking, PIP metabolism, or growth factor signaling in ES cells (17Alb Jr., J.G. Phillips S.E. Rostand K. Cotlin L. Pinxteren J. Manning T. Guo S. York J.D. Sontheimer H. Collawn J.F. Bankaitis V.A. Mol. Biol. Cell. 2002; 13: 739-754Crossref PubMed Scopus (60) Google Scholar). In this report, we describe the consequences associated with ablation of PITPα function in the mouse. We find that PITPα, although dispensable for prenatal development, is required for neonatal survival. PITPα– / – neonates suffer from a severe spinocerebellar neurodegenerative disease and exhibit defects in dietary fat and α-tocopherol transport across the small intestine. This intestinal steatosis in some respects resembles CRD, a human disorder of unknown molecular etiology (18Levy E. Clin. Invest. Med. 1996; 19: 317-324PubMed Google Scholar, 19Aguglia U. Annesi G. Pasquinelli G. Spadafora P. Gambarella A. Annesi F. Pasqua A.A. Cavalcanti F. Crescibene L. Bagala A. Bono F. Oliveri R.L. Valentino P. Zappia M. Quattrone A. Ann. Neurol. 2000; 47: 260-264Crossref PubMed Scopus (36) Google Scholar). Liver steatosis is also prominent in the mutant mice, suggesting the possibility that PITPα nullizygosity also compromises lipoprotein assembly and/or neutral lipid secretion in hepatocytes. Finally, PITPα– / – mice are severely hypoglycemic. Our results suggest a novel and unanticipated role for PITPα in regulating cargo-specific lipid transport from the enterocyte and hepatocyte ER, endocrine pancreas function, and glycogen metabolism. Generation and Genotyping of PITPα– / – Mice—AB1-derived +/PITPαΔ1::neo* ES cells have been described previously (17Alb Jr., J.G. Phillips S.E. Rostand K. Cotlin L. Pinxteren J. Manning T. Guo S. York J.D. Sontheimer H. Collawn J.F. Bankaitis V.A. Mol. Biol. Cell. 2002; 13: 739-754Crossref PubMed Scopus (60) Google Scholar). The +/PITPα::neo/puro ES cells corresponded to the OST 1152 line (Lexicon Genetics OmniBank™ library (20Zambrowicz B.P. Friedrich G.A. Buxton E.C. Lilleberg S.L. Person C. Sands A.T. Nature. 1998; 392: 608-611Crossref PubMed Scopus (402) Google Scholar)). Mice were generated by injection of ES cells into C57BL/6 blastocysts, implantation of blastocysts into pseudopregnant foster mothers, and identification of male chimeric mice competent for germ line transmission of each allele. Genotypes for PITPαΔ1::neo* mice (JG line) were determined by using a three-primer PCR assay. We employed a primer specific for the homozygous PITPα+ / + genotype (AB-2; 5′-GCGAGGCATCACTCTTCCCCTC-3′), the heterozygous PITPα– / + genotype (AB-1B; 5′-CACCATCCCCCACGGTGACTG-3′), and the PITPα– / – genotype (PG-1; 5′-GAATGTGTGCGAGGCCAGAGG-3′) in a 33-cycle reaction (53 °C annealing temperature). Genotypes for PITPα::neo/puro mice (L1 line) were determined in two steps. First, a two primer assay that monitored Neo distinguishes PITPα+ / + from PITPα– / + and PITPα– / – mice. Primers GE-UP (5′-GGGCGCCCGGTTCTTT TTGTGA-3′) and GE-DO (5′-TTGGTGGTCGAATGGGCAGGTAGC-3′) were used in a 28-cycle reaction (60 °C annealing temperature). To distinguish PITPα– / + from PITPα– / – L1 mice, we resorted to immunoblot analyses of mouse brain using PITPα-specific serum (17Alb Jr., J.G. Phillips S.E. Rostand K. Cotlin L. Pinxteren J. Manning T. Guo S. York J.D. Sontheimer H. Collawn J.F. Bankaitis V.A. Mol. Biol. Cell. 2002; 13: 739-754Crossref PubMed Scopus (60) Google Scholar). Serum Analyses—Blood was collected from mice immediately after heart puncture and clotted, and serum was clarified by centrifugation. Serum glucose was determined by using either the Trinder assay (Sigma) or was measured by Antech Diagnostics (Farmingdale, NY). Insulin and β-hydroxybutyrate were determined using the Immunoassay System (Crystal Chem Inc.) and the β-hydroxybutyrate assay kit (Sigma), respectively. All other serum analyses were performed by Antech Diagnostics (Farmingdale, NY). Carcass Analyses—Carcass analyses were as described previously (21Nagy T.R. Gower B.A. Stetson M.H. Can. J. Zool. 1994; 72: 1726-1734Crossref Scopus (14) Google Scholar). Gastrointestinal tracts were removed (stomach, small and large intestine, and cecum) and carcasses weighed. Body water content of eviscerated carcasses was determined by drying to constant weight in a 60 °C oven and measuring differences between the pre- and post-drying carcass mass. Dried carcasses were minced, ground to a homogeneous mixture, and extracted with petroleum ether in a Soxhlet apparatus to determine fat mass and fat-free dry mass. Fat-free dry mass was burned overnight at 600 °C(>8 h) to determine eviscerated carcass ash. Histological Analysis—Mice were anesthetized with 1.25% Avertin and perfused with phosphate-buffered saline, 4% paraformaldehyde. Duodenum, ileum, cerebellum, pancreas, and spinal cord were harvested, flushed with fixative (duodenum and ileum only), dissected, and infused with fixative for 24 h. Samples were mounted in paraffin, and 5-μm-thick sections were stained and mounted. These sections were rehydrated by serial transfer from xylene to 50% EtOH, stained with hematoxylin/eosin, and mounted in Permount (Fisher). Where osmium staining was employed, duodenal and liver sections were stained in 5% potassium dichromate, 2% osmium tetroxide for 8 h prior to paraffin embedding (22Luna L. Manual of Histological Staining Methods of the Armed Forces Institute of Pathology. Armed Forces Institute of Pathlogy, Washington, DC1968: 143-145Google Scholar) and counterstained with toluidine blue O. Whole brains were extracted from mice perfused with 4% paraformaldehyde and 2.5% glutaraldehyde, washed for several days in phosphate-buffered saline, and mounted in paraffin. Brains were sliced in half along the sagittal plane; each half was embedded in paraffin; and 5- or 8-μm-thick slices were mounted onto treated slides. Sections were rehydrated to 50% EtOH. For visualization of Purkinje cells, slices were incubated with goat anti-calbindin antibodies (Santa Cruz Biotechnology; Santa Cruz, CA) and developed with the Vectastain ABC kit (Vector Laboratories, Burlingame, CA). After incubation in 2% osmium fumes for 10 min, slices were counterstained with toluidine blue O. For Oil Red O staining, livers were extracted from mice perfused with 4% paraformaldehyde, washed with phosphate-buffered saline, and frozen at –20 °C. Frozen livers were mounted, sectioned (8 μm), fixed to a histological slide, and placed in absolute propylene glycol (2 min). Slides were moved into Oil Red O solution (Newcomer Supply, Middleton, WI, catalog number 12722) for 1 h, differentiated in 85% propylene glycol (1 min), rinsed 2× in distilled water, counterstained with hematoxylin (10 s), and mounted in glycerin. Electron Microscopy—Mice were perfused with 4% formaldehyde, 2.5% glutaraldehyde. Biopsies from intestine, liver, and spinal cord were post-fixed with 1% OsO4, dehydrated with acetone, embedded in epoxy (23Mascorro J.A. Kirby G.S. Proceedings of the 44th Annual Meeting of the Electron Microscopy Society of America. 1986; : 222-223Crossref Google Scholar), sectioned (65 nm-thick), and stained with 4% uranyl acetate and Sato's lead mixture (24Sato T. J. Electron Microsc. 1968; 17: 158-159PubMed Google Scholar). Samples were viewed at 80 kV in a Phillips Tecnai 12 microscope (FEI Co., Eindhoven, The Netherlands) and imaged with a Gatan MultiScan model 794 digital camera (Gatan, Pleasanton, CA). Epoxy-embedded samples were sectioned for histological analysis (2-μm thickness) and stained with 1% toluidine blue O in 1% sodium borate. Digital images were collected with a SPOT RT digital camera (Diagnostic Instruments, Inc., Sterling Heights, MI), using Plan Fluor Nikon objectives mounted in a Nikon Eclipse E400 microscope (Nikon Inc., Melville, NY). Morphometric analysis of spinal cord sections was carried out with Scion Image software (Scion Corp., Frederick, MD). α-Tocopherol Analyses—Brain α-tocopherol was extracted as described (25Jishage K. Arita M. Igarashi K. Iwata T. Watanabe M. Ogawa M. Ueda O. Kamada N. Inoue K. Arai H. Suzuki H. J. Biol. Chem. 2001; 276: 1669-1672Abstract Full Text Full Text PDF PubMed Scopus (162) Google Scholar), injected onto a 250 × 4.6 mm Phenomenex ODS 20 reversed phase C18 column (5-μm particle diameter), and eluted with methanolic 0.05% ammonium acetate using a flow rate of 1.5 ml/min. The high pressure liquid chromatography system consisted of a PerkinElmer Life Sciences model LC200 gradient pump, an AS 200 Autosampler, and an LC 295 programmable UV-visible light detector set at 292 nm. Liver Glycogen Analysis—Quantification of liver glycogen was by the method of Passonneau and Lauderdale (26Passonneau J. Lauderdale V. Anal. Biochem. 1974; 60: 405-412Crossref PubMed Scopus (627) Google Scholar). Glycogen was extracted from acidified liver homogenates and hydrolyzed to glucose with amyloglucosidase (Sigma). Glucose was determined by a glucose oxidase-coupled Trinder assay (Sigma). Adenylate Nucleotide Analysis—ATP and ATP/ADP ratios were measured using the ApoGlow™ kit (BioWhittaker Molecular Applications; Rockville, MD). Samples (0.5 mg tissue) were collected and rapidly frozen. Tissues were thawed and homogenized in the presence of a nucleotide-releasing mixture supplied by the manufacturer. Acid-extracted ATP was assayed with a luciferase-based system with picomolar sensitivity. ADP was converted to ATP and measured after a 5-min incubation at 22 °C. Pancreatic Histology—Whole pancreas from mice perfused with 4% paraformaldehyde was embedded in paraffin and serially sectioned (thickness = 5 μm). Islet numbers were assessed by sequential observation of hematoxylin/eosin-stained sections. Care was exercised to separate islets from patches of connecting ducts and intervening connective tissue, and not to re-score the same islet in successive sections. Islet-like, encapsulated structures larger than 100 μm that were detected in at least three consecutive sections were scored as islets. The total pancreatic area studied was similar in all sections. PITPα– / – Mice Develop to Term—Ablation of PITPα function in the mouse was achieved using two independent null alleles. First, a homologous recombination vector was constructed where exons 8–10 of the PITPα gene are replaced with a neo cassette (Fig. 1A). This mutation (PITPαΔ::neo*) deletes PITPα residues 162–257, a region critical for PITPα function (17Alb Jr., J.G. Phillips S.E. Rostand K. Cotlin L. Pinxteren J. Manning T. Guo S. York J.D. Sontheimer H. Collawn J.F. Bankaitis V.A. Mol. Biol. Cell. 2002; 13: 739-754Crossref PubMed Scopus (60) Google Scholar). Second, survey of the Lexicon Genetics OmniBank™ gene trap library (see "Experimental Procedures") identified an insertion mutation in the PITPα structural gene (PITPα::neo/puro). This allele is genetically similar (although not identical) to PITPαΔ::neo* as it also truncates PITPα after residue 162 (Fig. 1B). Mice were derived from each targeted ES cell line. PITPα– / + offspring are phenotypically normal and fertile, and mice homozygous for either of these two mutations exhibit indistinguishable phenotypes. The phenotypic data presented below were obtained from both PITPαΔ1::neo* and PITPα::neo/puro homozygous animals. Intercrosses with heterozygous mice carrying either PITPαΔ1::neo* or the PITPα::neo/puro allele yielded genotypes consistent with a fully penetrant autosomal recessive mutation. From a dedicated pool of 408 live births, 89 PITPα– / – progeny were recovered (Fig. 1C). The genotypic distribution of 106 PITPα+ / +, 213 PITPα– / +, 89 PITPα– / – corresponds to a 1.000:2.009:0.840 ratio that approximates closely the 1:2:1 ratio predicted by Mendel's rules. Correct gene targeting in the progeny was verified by PCR genotyping and immunoblotting of brain extracts with a specific PITPα antiserum (Fig. 1D). Antibodies directed against the PITPα N terminus failed to detect a truncation product in mice homozygous for PITPαΔ1::neo* or PITPα::neo/puro, suggesting that both alleles represent null mutations. Finally, we find that relative PITPβ levels are unchanged in PITPα+ / + and PITPα– / – brain (Fig. 1D), indicating that PITPα– / – mice do not activate compensatory processes that increase PITPβ expression. Neonatal Mortality of PITPα– / – Mice—PITPα– / – mice failed to thrive and died at a very young age. In a sample pool of 57 PITPα– / – mice, 40% died within 48 h after birth (Fig. 2A). These early P0 and P1 deaths were not characterized by obvious external abnormalities. Moreover, postmortem analyses revealed that the stomachs of the expired PITPα– / – mice contained copious quantities of milk, indicating that mutant animals had nursed. Of the PITPα– / – progeny that survived past P1, a steady incidence of mortality was observed between P2 and P11. Almost all mice expired by P11, and only one PITPα– / – mouse lived to P14 (Fig. 2A). PITPα– / – progeny that survived past P1 were initially indistinguishable from PITPα+ / + and PITPα– / + littermates in size, external morphology, and behavior. By P4, however, two phenotypes rapidly asserted themselves. First, most PITPα– / – mice were moribund and exhibited little spontaneous movement. These mice did respond to touch, however. By contrast, ∼10% of the PITPα– / – mice experienced spontaneous seizures. All PITPα– / – mice were severely ataxic and were generally incapable of maintaining themselves upright. We also observed coarse action tremors upon limb extension in PITPα– / – mice. Second, these mutant mice failed to thrive. Although the PITPα– / – mice gained body mass, they did so slowly (Fig. 2B). By P10, surviving PITPα– / – mice were 2.5-fold less massive than their PITPα+ / + and PITPα– / + littermates (Fig. 2C). PITPα– / – Mice Suckle Effectively—Several lines of evidence indicate that the failure of PITPα– / – mice to thrive is not the simple consequence of neurological defects. We observed PITPα– / – mice in the act of suckling, and postmortem analyses indicated both copious quantities of milk in stomachs of these mice (Fig. 2D) and substantial amounts of digested matter throughout the PITPα– / – intestinal tract (not shown). The suckling competence of PITPα– / – mice notwithstanding, the mutant animals exhibited pathologically low body fat levels. This reduction is obvious when the subcutaneous fat pads of PITPα+ / + and PITPα– / – mice are compared. Whereas PITPα+ / + controls exhibit large axillary and inguinal fat pads, these structures are absent from PITPα– / – animals (Fig. 2D). Chemical analyses of eviscerated carcasses quantified these differences; PITPα+ / + and PITPα– / – mice exhibited total body fat contents of 15.0 ± 1.1 and 4.0 ± 0.5% of total body mass, respectively (Fig. 2E). Significantly, the relative water contents of PITPα+ / + and PITPα– / – carcasses are similar (Fig. 2E), indicating that PITPα– / – mice are not dehydrated, as would be expected if there were substantial suckling defects. The relative fat-free dry mass contents and carcass bone ash contents are also comparable (Fig. 2E). Consistent with those measurements, we find that organ/total body mass ratios in PITPα– / – mice for brain, liver, and other major organs are also proportional to total body mass (not shown). Elevated Apoptosis and Purkinje Cell Defects in PITPα– / – Cerebellum—Nissl staining does not reveal obvious defects in development or morphology of PITPα– / – cerebrum, thalamus, hippocampus, or cerebellum. Normal cellularity was observed in all regions examined (data not shown). However, PITPα insufficiencies evoked functional defects in the cerebellum. These defects were apparent at several levels. Whereas TUNEL staining showed a sparse and random distribution of apoptotic foci in wild-type brain, apoptosis was more prevalent throughout PITPα– / – cerebellum. This is particularly evident in the external granule layer of mutant cerebellum (Fig. 3A). Second, we find Purkinje cell defects in PITPα– / – animals. Purkinje cells normally align themselves into a sharply defined monolayer that lies between the molecular and external granular layers of the cerebellum (Fig. 3B). These cells elaborate well developed apical dendritic stalks which arborize into luxurious branches that penetrate into the molecular layer of the cerebellum. PITPα– / – Purkinje cells, while retaining normal flask-shaped cell body morphologies, exhibit either abbreviated apical dendritic stalks or no obvious stalks at all (Fig. 3B). Moreover, the dendritic branches emanating from the abbreviated apical stalks are less arborized than those of PITPα+ / + Purkinje cells. We also observed cases where these cells exhibit defects in spatial alignment, reside off of the defined layer, or are otherwise misoriented. Degenerative Disease in PITPα– / – Cerebellum—PITPα– / – cerebellum suffered major degenerative insult as judged by the extent of reactive gliosis. Wild-type cerebellum shows low levels of GFAP, a specific marker for activated astrocytes (Fig. 3C). By contrast, PITPα– / – cerebellum exhibits a dramatic staining for GFAP, and reactive gliosis is especially prominent in the white matter trunk of this organ (Fig. 3C). Even in these relatively less affected areas, GFAP immunohistochemistry suggests some 10% of the cells represent activated astrocytes that are frequently seen to be enveloping neuron cell bodies (Fig. 3D). Reactive gliosis is not observed throughout the mutant brain. Whereas PITPα– / – cerebellum and brain stem are heavily infiltrated with activated glial cells, and the subthalamic region is also involved, significant gliosis is not observed in the neocortex, striatum, hippocampal formation, and other forebrain regions (not shown). Finally, electron microscopy reveals ER defects in cerebellar neurons of PITPα– / – mice. These defects manifest themselves as significant vacuolations of the smooth ER region, even though adjacent regions of rough ER retain normal morphology (not shown). Such vacuolations are not observed in smooth ER of PITPα+ / + cerebellar neurons. Neurodegeneration and Inflammation in PITPα– / – Spinal Cord—We consistently observed WM deficits in PITPα– / – cervical, thoracic, and lumbar spinal cord. Whereas ventral WM is well developed in PITPα+ / + cervical spinal cord, the corresponding WM regions are thin and sparse in PITPα– / – animals (Fig. 4A). WM to GM area ratios in cervical spinal cord are 0.45 ± 0.1 for PITPα+ / + and 0.32 ± 0.02 for PITPα– / – mice, respectively (p < 0.01; n = 6). In addition to the WM deficits, evidence of inflammation pervades all regions of the PITPa – / – spinal cord. Toluidine blue O-staining reveals areas of abnormally heavy staining in the ventral horn where motor neuron cell bodies reside, suggesting cell damage in these areas (Fig. 4A). Moreover, whereas the WM/GM interface of PITPα+ / + cervical spinal cord presents normal vascular structures, neurons, and accessory cells, corresponding regions of the PITPα– / – spinal cord exhibit densely stained cells and damaged neurons (Fig. 4B). These pathologies encompass a spectrum of cell death events. These range from occasional apoptotic neurons (condensed nuclei, fragmented cytoplasm; not shown) to large numbers of neurons undergoing aponecrotic processes (Fig. 4B). Aponecrosis is a form of cell death associated with reductions in cellular ATP and increases in ADP levels (27Formigli L. Papucci L. Tani A. Schiavone N. Tempestini A. Orlandini G.E. Capaccioli S. Zecchi Orlandini S. J. Cell. Physiol. 2000; 182: 41-49Crossref PubMed Scopus (318) Google Scholar, 28Sperandio S. de Belle I. Bredesen D.E. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 14376-14381Crossref PubMed Scopus (795) Google Scholar). Analyses from six PITPα– / – mice indicate that 18 ± 1% of the total motor neuron cell bodies in the ventral horn exhibit properties of aponecrosis. These include pericytoplasmic vacuolation, reduced cytoplasmic contents, and cytoplasmic proliferation of irregular electron-translucent vesicles (Fig. 4B). We did not observe aponecrotic motor neurons in the ventral horn of PITPα+ / + spinal cord. Extensive vacuolation and membrane blebbing is prevalent in cells that line the vasculature of PITPα– / – spinal cord (not shown). Axons surrounding capillaries and small vessels in the GM are enlarged and often damaged. This is unlikely to represent a perfusion artifact, as this phenotype was not recorded in any of the spinal cord sections analyzed from 14 PITPα+ / + mice. Toluidine blue O-staining reveals the presence of cells containing an abundance of purple intracellular granules in the perivascular matrix (Fig. 4C, left panel). This obvious purple metachromasia is diagnostic of inflammatory mast cells (29Galli S.J. Dvorak A.M. Dvorak H.F. Prog. Allergy. 1984; 34: 1-141PubMed Google Scholar, 30Bebo Jr., B.F. Yong T. Orr E.L. Linthicum D.S. J. Neurosci. Res. 1996; 45: 340-348Crossref PubMed Scopus (58) Google Scholar). Accordingly, we find extratissular macrophages in perivascular tissue or even in the vessels themselves (Fig. 4C, right panel), suggesting the blood/brain barrier of PITPα– / – mice is itself compromised. Finally, spinocerebellar injury in PITPα– / – mice also includes processes that resemble those of other myelin-related central nervous system inflammatory disorders (31Pender M.P. Stanley G.P. Yoong G. Nguyen K.B. Acta Neuropathol. 1990; 80: 172-183Crossref PubMed Scopus (48) Google Scholar). First, macroscopic swelling in the dorsal spinal columns is apparent. The non-neuronal area (neuropil) comprises 54 ± 3% in PITPα+ / + spinal cord (n = 6; p = 0.01) and 70 ± 6% of total area in PITPα– / – spinal cord, respectively. Degenerative processes in neuropil, typified by swollen axons lacking organelles and cytoskeletal filaments, are also obvious (not shown). Second, demyelination is scored in both WM and GM areas of PITPα– / – mice and especially in the dorsal spinal columns. Supporting cells that may be remodeling myelin are also observed (Fig. 4D). The fragility of the mutant central nervous system notwithstanding, a variety of PITPα– / – neurons are amenable to primary culture. Cultured cortical neurons, cerebellar granule cells, and spinal cord dorsal root ganglion neurons from PITPα– / – mice are not more fragile than their PITPα+ / + counterparts in any obvious way. Indeed, titration of nerve growth factor concentrations from 50 to 5 ng/ml revealed no differences in the thresholds of trophic factor required to sustain viability of PITPα– / – versus PITPα+ / + dorsal root ganglion neurons in culture (not shown). PITPα– / – cerebellar granule cells are similarly robust. These data suggest a significant cause of neuronal injury in PITPα– / – mice is a hostile physiological environment, rather than some overriding cell autonomous defect. Lipid Dysregulation in PITPα– / – Enterocytes—Failure of PITPα– / – mice to thrive suggests a malabsorption disorder. Histolog
Referência(s)