The development of methods for the isolation of brain capillaries and brain capillary plasma membranes makes possible biochemical studies of the blood-brain.
Journal of Neuroscience Research 18:352-357 (1987)
Isolation of Intact Capillaries and Capillary Plasma Membranes From Frozen Human Brain W.M. Pardridge, J. Yang, J. Eisenberg, and W. W. Tourtellotte Department of Medicine, UCLA School of Medicine (W.M.P., J.Y., J.E.) and Neurology Service, Veterans Administration Wadsworth Medical Center W.T.), Los Angeles
w.
The development of methods for the isolation of brain capillaries and brain capillary plasma membranes makes possible biochemical studies of the blood-brain barrier (BBB), which is made up of brain capillaries. Studies aimed at assessing the role of the BBB in the pathogenesis of specific neurologic diseases, e.g., Alzheimer’s disease or multiple sclerosis, will necessitate the isolation of capillaries from brain involved with specific pathology. Such tissue is most readily available from banks containing frozen human brain. The present studies show that intact capillaries and capillary plasma membranes can be isolated from frozen human brain, including as little as five g of multiple sclerosis plaque tissue. Capillaries from frozen human brain are enriched in gamma-glutamyl transpeptidase, factor VIII antigen, and a 46K protein which has recently been shown to be a BBB-specific protein. These studies provide the basis for future biochemical studies of human brain microvessels in neurologic disease.
amino acid transport at the human BBB since the K, value for phenylalanine uptake by isolated human brain capillaries is nearly identical to that found for phenylalanine uptake by capillaries isolated from fresh rat brain [Choi and Pardridge, 19861. The extension of studies of the human brain capillary to neurologic problems requires isolating capillaries from postmortem brains of individuals with a particular neurologic disease such as multiple sclerosis or Alzheimer’s disease. The availability of adequate quantities of fresh autopsy human brain afflicted with a certain neurologic disease is oftentimes limited. However, human brain banks offer an important reservoir of autopsy brain tissue of particular neurologic diseases [Tourtellotte et al, 19841, and the question arises as to whether capillaries can be isolated in relatively intact form from frozen bank human brain material. The present study describes the isolation of capillaries from frozen human brain and provides data that these capillaries are relatively structurally and biochemically intact. Microvessels isolated from frozen human brain may provide a new model system for the investigation of the role Key words: blood-brain barrier, multiple sclerosis, of the BBB in the pathogenesis of neurologic disease. Alzheimer’s disease
INTRODUCTION The blood-brain barrier (BBB) is formed by epithelial-like tight junctions which are present in more than 99 of vertebrate brain capillaries [Brightman, 19771. Since brain capillaries can now be routinely isolated [Goldstein and Betz, 19831, the study of the biochemistry of the blood-brain barrier is now possible. Moreover, recent studies indicate that capillaries can also be isolated in high yield from fresh autopsy human brain, even from brains obtained from the pathologist up to 36 h postmortem [Pardridge et al, 1985bl. Capillaries isolated from fresh autopsy human brain have high activities of BBBspecific enzymes such as gamma-glutamyl transpeptidase or alkaline phosphatase, and have structurally intact insulin receptor [ Pardridge et al, 1985bl. Finally, recent studies show that capillaries isolated from fresh autopsy human brain are a useful model system for studying
0 1987 Alan R. Liss, Inc.
MATERIALS AND METHODS Patient Information Four frozen brain samples and two fresh brain specimens were obtained from six subjects with or without neurologic disease (Table I). Fresh autopsy brain was sectioned into 4-5-mrn coronal slices and snap frozen between two slabs pre-cooled in liquid nitrogen. The tissue was immediately placed in a -70°C freezer. When brain was frozen en bloc and then thawed for capillary isolation, the final yield and purity was unacceptable. Therefore, capillaries were isolated only from tissue quick-frozen as coronal slabs. The day prior to capillary isolation, the brain was removed from the freezer and allowed to thaw overnight Received January 30, 1987; revised March 26, 1987; accepted March 30, 1987. Address reprint requests to Dr. William M. Pardridge, Department of Medicine, UCLA School of Medicine, Los Angeles, CA 90024.
Microvessels From Frozen Human Brain
353
TABLE I. Tissue Weights, Autolysis Time, Patient Diagnosis Number Frozen 1 2 3 4
Fresh 5 6
Tissue (8)
Auto1ysis (h)
Agelsex
Diagnosis
Cortex (56) Cortex (79) Cortex (73) Plaque1 periplaque ( 5 )
7 19 46 10
72lM 771F 60lM 50lM
Alzheimer’s disease Parkinson’s disease Pulmonary embolism Multiple sclerosis
Cortex (83) White matter (12)
36 20
59lM 50lM
Esophageal cancer Multiple sclerosis
at 4°C. Capillaries were isolated with a mechanical homogenization technique [Goldstein and Betz, 19831, using modifications recently described in detail [Pardridge et al, 1985al. Briefly, a homogenate of brain tissue is obtained using a mechanical homogenization technique and the fraction containing the brain vessels was isolated as the pellet of a 13% dextran flotation at 5,800g. The pellet was then passed over a 210-pm nylon mesh sieve to remove the large vessels, and the fraction passing through the sieve contained microvessels, red cells, and nuclei. The microvessels were separated from red cells and nuclei with a glass bead column [Goldstein and Betz, 19831. At this point, the capillaries were free of other cellular elements. Unlike the capillaries isolated from fresh brain tissue, capillary fractions from frozen brain tissue often contain large globules of proteinaceous material. These were separated from the microvessels using a 45 % Percoll density centrifugation as described recently [Pardridge et al, 1985al. The capillary fraction was then frozen in 300 mM sucrose, 0.02 M Tris (pH = 7.4), and 2 mM dithiothreitol in a liquid nitrogen tank at -70°C. Capillaries can be recovered with a 50-70% yield from the cryopreservation procedure, and the cryopreservation of brain capillaries has been shown not to alter brain capillary insulin receptor activity or brain capillary protein phosphorylation activity [Pardridge et al, 1985b, c]. The yield of purified microvessels averaged 2.3 & 0.7 mg capillary protein per 10 g cortex (mean f SE), which is comparable to the yields for capillary isolation using fresh laboratory animal brain [Goldstein and Betz, 1983; Pardridge et al, 1985al.
membrane cytoskeletal structure by sonication, and these two fractions are separated by 5-min 25,OOOg centrifugation, which results in formation of a dense basement membrane pellet and a supernatant of brain capillary plasma membranes. This plasma membrane fraction has been shown in previous studies to be enriched in gammaglutamyl transpeptidase, alkaline phosphatase, insulin receptor activity, and to contain the majority of the microvessel protein phosphorylation activity [Pardridge et al, 1985b,c]. In addition, an antiserum against this plasma membrane fraction using bovine brain capillaries has been prepared, and this antiserum selectively illuminates the lateral margins of isolated brain capillary endothelia [Pardridge et al, 19861.
E1ectrophoresis The plasma membranes were separated by sodium dodecylsulfate polyacrylamide gel electrophoresis (SDSPAGE) using 6-15% gradient gels and the method of Laemmli [1970]. These methods have been described in detail from this laboratory recently [Pardridge et al, 1985c, 19861. Oftentimes the amount of plasma membranes from the frozen brain specimens were of limited quantities. For example, only 5-10 g of multiple sclerosis specimens were usually available (Table I). Therefore, the Kodavue (Eastman-Kodak, Rochester, NY) staining technique, which is more sensitive than Coomasie blue, was employed [Yudelson, 19841. Approximately 40 pg of plasma membranes were spotted per well, and after electrophoresis the slab gels were fixed in methanol for 2 hr followed by a change in methanol and fixing for an additional 1 Yz hr. The presensitization and sensitization Isolation of Capillary Plasma Membranes steps were followed per the manufacturer instructions When capillary plasma membranes were isolated, using 1 minute durations for each step. this procedure was performed on the same day as the isolation of the capillaries. The method is from Lidensky Gamma-GlutamYl TransPePtidase Histochemistry and Drewes [1983] as described recently [Pardridge et Brain capillaries obtained from frozen human brain al, 1985bl. Briefly, the method involves hypotonic lysis were adsorbed to a glass slide with centrifugation in a of the microvessels and recovery of the microvessel cytospin at 300 rpm for 5 min [Pardridge et al, 19861. ghosts as the pellet of a 15,OOOg centrifugation. The The capillaries were fixed in acetone for 5 min at 4°C basement membranes are separated from the plasma and gamma-glutamyl transpeptidase was assayed using
354
Pardridge et a1
Microvessels From Frozen Human Brain
Fig. 1. A: Capillaries from frozen human brain are shown to be positive for gamma-glutamyl transpeptidase, a blood-brain barrier (BBB)-specific enzyme. These capillaries were isolated from control human brain (Table I, No. 3). B: Capillaries isolated from frozen human brain are positive for human factor VIII antigen using an immunoperoxidase technique and a
355
mouse monoclonal antibody to human factor VIII at a dilution of 1:50. These capillaries were isolated from the cortex of brain from a subject with Parkinson’s disease (Table I, No. 2). C: Toluidine blue stain of capillaries isolated from 5 g of multiple sclerosis plaque/periplaque tissue (Table I, No. 4). Magnification. X40-100.
the method of DeBault and Cancilla [1980], which em- RESULTS ploys gamma-glutamyl-4-methoxy-2-naphthylamideas a Capillaries from frozen ..uman brain were isolated substrate. free of adjoining brain tissue and were strongly positive for gamma-glutamyl transpeptidase as shown in Figure Factor VIII Immunoperoxidase 1A. The capillaries were also positive for factor VIII Capillaries obtained from frozen human brain were antigen, and a representative sample is shown in Figure adsorbed to a glass slide in a cytospin and fixed in 10% 1B; the mouse ascites control gave no reaction with the formalin in Tris-buffered saline (TBS: 0.1 M Tris, 0.15 capillaries (not shown). Capillaries could be isolated from M NaC1, pH = 7.4) at 4°C for 10 min. The avidin/ as little as 5 g of multiple sclerosis plaque tissue, as biotinhmmunoperoxidase technique was used [Pardridge shown in Figure 1C for control brain. The capillaries et al, 19861 employing a kit obtained from Vector Labs isolated from Alzheimer’s disease or Parkinson’s disease (Burlingame, CA). The primary antiserum was a 1:50 brain were also strongly positive for gamma-glutamyl dilution of a mouse monoclonal antibody to human factor transpeptidase activity and the enzyme activity was comVIII and mouse ascites was used as a control (Cappel parable to the control microvessels shown in Figure 1A. Labs, Westchester, PA). The secondary antibody was The Kodavue plasma membrane protein band patused at a 1:20 dilution and was a biotinylated horse terns after protein separation on a 6- 15 % gradient SDS antimouse antiserum. polyacrylamide gel is shown in Figure 2. This study
356
Pardridge et a1
PARK
and the studies indicate that despite many hours of the postmortem condition and despite freezing of the brain AD MS NL tissue, the capillary plasma membranes are reasonably frozen ,-fresh, frozen rat bovine intact, based on the protein profiles after separation on polyacrylamide gels (Fig. 2). The intactness of the brain capillaries isotated from frozen brain material is in accor200k dance with the structural intactness of capillaries isolated from fresh autopsy human brain, even up to 36 hr postmortem [Pardridge et al, 1985bl. Recent studies have 116k shown that capillaries can be isolated from rabbit brain - 92k after placing the corpse at 4°C for up to 40 hr [Choi and Pardridge, 19861. These capillaries are isolated in high 66k yield after 40 hr of autolysis time and the band patterns of the capillary proteins after separation on SDS-PAGE gradient gels are nearly identical when capillaries iso- 45k lated after either 0 or 42 hr of autolysis in the rabbit are compared. The brain capillary appears to be remarkably resistant to the effects of ischemia and the clinical ana- 31k logue of this may be the negative brain scans seen for up to a week after an ischemic insult to brain, suggesting structural intactness of the BBB during ischemia [Schall - 21k and Quinn, 19711. Many biochemical properties of the human brain capillary will be lost during the first few - 14k hours after death. For example, adenvl cvclase is not detectable in capillaries isolated from fceshhuman brain Fig. 2. Kodavue stain of capillary plasma membranes from after 20 hr of autolysis (W.M. Pardridge and J. Eisenfresh or frozen human brain, fresh rat brain, or fresh bovine cortex. The capillary plasma membrane fraction was separated berg, unpublished data). Nevertheless, several other on SDS-PAGE 6-15% gradient gels. Capillaries were isolated Properties are intact after many hours of autolysis, and from two fresh human brains; control cortex (Table I, No. 5) these include gamma-glutamyl transPePtidase, alkaline and uninvolved multiple sclerosis white matter (Table I, No. phosphatase, insulin receptor [Pardridge et al, 1985b1, 6) and from two specimens of frozen human brain (Alz- phenylalanine transport [Choi and Pardridge, 19861, and heimer’s disease cortex, and Parkinson’s disease cortex; Table the 46K protein [Pardridge et al, 19861 (Fig. 2). The 46K I , NOS. 1 and 2 ) . Abbreviations used: AD = Alzheimer’s protein has recently been shown to be specific to brain disease; MS = multiple sclerosis; NL = normal; PARK = capillaries as this protein is not found in a capillary-free Parkinson’s disease. preparation of brain synaptosomes [Pardridge et al, 19861. Moreover, an antiserum that is largely directed to of capilthe 46K protein selectively illuminates brain capillaries the plasma membrane band laries from two fresh human brains and from two frozen using immunoperoxidase studies but does not illuminate human brains with the band patterns of capillaries iso- capillaries in other bovine tissues such as heart, ludney,
\1
-
-
-
lated from fresh rat of fresh bovine brain. The principal proteins are a 46K protein and a family of low molecular weight proteins at 1 4 - 1 8 ~ . addition, a 2JoK and a 34K doublet are also prominent in the capillaries isolated from frozen human brain, as well as from fresh bovine brain. Capillaries were isolated from cortex in all of the specimens shown in ~i~~~~2, except for the fresh multisclerosis specimen which was uninvolved white matter (No. 6, Table I).
DISCUSSION The present studies describe for the first time the isolation of brain capillaries and capillary plasma membrane fractions from bank frozen human brain. Capillaries can be isolated from as little as 5 g of starting material,
liver LPardridge et 19861. The present results showing that intact capillaries can be isolated from brains of subjects with neurologic disease, such as Alzheimer’s disease, Parkinson’s disease, or multiple sclerosis, indicates brain capillary biochemistry in these diseases may now be to investigation. There is evidence that both Alzheimer’s disease and multiple sclerosis may involve the BBB in the early pathogenetic events of these diseases. For example, Alzheimer’s disease is characterized in some cases by extensive congophilic angiopathy involving primarily small arterioles [Wong et al, 19851. Isolation of brain microvessels, therefore, may provide one strategy for purifying the cerebrovascular amyloid precursor protein in Alzheimer’s disease. With regard to multiple sclerosis, an early event in this disease is the formation of perivasOr
a17
Microvessels From Frozen Human Brain
cular cuffs of circulating lymphocytes around small venules in white matter [Adams, 19771. The presentation of a foreign or autoantigen to the circulating lymphocytes by the brain microvessel endothelial cell may be a possible initial event in the immune response in multiple sclerosis. If so, the isolation of microvessels from plaque tissue (Fig. 1C) may provide a new strategy for the biochemical investigation of the cerebral microvasculature in the initiation of the immune response in multiple sclerosis. For example, recent attempts [Tuskada et al, 19851 to find endothelial autoantigens in multiple sclerosis may have been unsuccessful because brain microvessels from multiple sclerosis tissue were not used.
357
DeBault LE, Cancilla PA (1980): Gamma-glutamyl transpeptidasc in isolated brain endothelial cells: Induction by glial cells in vitro. Science 207:653-655. Goldstein G, Betz AL (1983): Recent advances in understanding brain capillary function. Ann Neurol 14:389-395. Laemmli UK (1970): Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227:680-685. Lidinsky WA, Drewes LR (1983): Characterization of the blood-brain barrier: Protein composition of the capillary endothelial cell membranes. J Neurochem 41: 1341-1348. Pardridge WM, Eisenberg J, Yamada T (1985a): Rapid sequestration and degradation of somatostatin analogues by isolated brain microvessels. J Neurochem 44: 1178-1184. Pardridge WM, Eisenberg J, Yang J (198%): Human blood-brain barrier insulin receptor. J Neurochem 44: 1771-1778. Pardridge WM, Yang J, Eisenberg J (1985~):Blood-brain barrier protein phosphorylation and dephosphorylation. J Neurochem 45: 1 141- 1147. ACKNOWLEDGMENTS Pardridge WM, Yang J, Eisenberg J, Mietus LJ (1986): Antibodies to blood-brain barrier bind selectively to brain capillary endotheDawn Brown skillfully prepared the manuscript. lial lateral membranes and to a 46K protein. J Cercb Blood Tissue specimens were obtained from the National NeuFlow Metab 6:203-211. rological Research Bank (VAMC, Wadsworth, CA Schall GL, Quinn JL 111 (1971): Brain Scanning. In Blahd WH (ed): 90073), which is sponsored by NINCDWNIMH, NMSS, “Nuclear Medicine.” New York: McGraw-Hill Book Co., pp HD Foundation, and the Veterans Administration. This 262-264. work was supported in part by the California Department Tourtellotte WW, ltabashi HH, Rosario I, Berman K (1984): The National Neurological Research Bank. A collection of cryopreof Health Services and by the Gustafson Fund. served human neurological specimens for neuroscientists. Ann NY Acad Sci 436:513-516. Tsukada N, Behan WMH, Behan PO (1985): Search for autoantibodREFERENCES ies to endothelial and smooth muscle cells in patients with multiple sclerosis. Acta Neuropathol (Berl) 66:134- 139. Adams CWM (1977): Pathology of multiple sclerosis: Progression of the lesion. Br Med Bull 33: 15-20. Wong CW, Quaranta V, Glenner GG (1985): Neuritic plaques and cerebrovascular amyloid in Alzheimer disease are antigenically Brightman MW (1977): Morphology of blood-brain interfaces. Exp related. Proc Natl Acad Sci USA 82:8729-8732. Eye Res [Suppl] 25: 1-25. Choi T, Pardridge WM (1986): Phenylalanine transport at the human Yudelson J (1984): The visualization of protein electrophoregrams by blood-brain barrier. Studies in isolated human brain capillarphysical development. Bio Techniques 2:42-47. ies. J Biol Chem 261:6536-6541.
