A Putative Role for Platelet-Derived Growth Factor in Angiogenesis and Neuroprotection After Ischemic Stroke in Humans
Background and Purpose Growth factors control two important processes in infarcted tissue, ie, angiogenesis and gliosis. We recently reported that transforming growth factor-β1 (TGF-β1) might be involved in angiogenesis after ischemic stroke in humans; here we present data of an extensive study on platelet-derived growth factor (PDGF) and its receptors.
Methods We studied brain samples from patients who suffered from ischemic stroke for the expression of mRNA encoding PDGF-A, PDGF-B, and PDGF receptors (PDGF-R). Proteins were examined by Western blotting and immunohistochemistry using the antibodies to PDGF-AB, PDGF-BB, PDGF-Rα, and PDGF-Rβ.
Results At the mRNA level, PDGF-A and PDGF-B were expressed mainly in neurons in penumbra. PDGF-R mRNA was strongly expressed in some astrocytes but mainly in type III/IV neurons in infarct and penumbra. The least expression was seen in the contralateral hemisphere (P<.001). In contrast, both PDGF-AB and PDGF-BB immunoreactive products were present in most cell types: PDGF-Rα and PDGF-Rβ mainly on neurons, and PDGF-Rβ on some endothelial cells, with less staining of all the isoforms in the contralateral hemisphere. On Western blots, PDGF-AB and -BB were expressed more within white matter than gray matter of infarct/penumbra, whereas both isoforms of receptor were expressed mainly in gray matter compared with contralateral hemisphere. There was no or very weak expression of the receptor in white matter.
Conclusions PDGF proteins are highly expressed in white matter, suggesting that PDGF may exert its function in white matter participating either in regeneration of damaged axons or in glial scar formation. PDGF-BB and its receptor expressed on microvessel endothelial cells might be involved in angiogenesis after stroke. Thus, PDGF is likely to be angiogenic and neuroprotective in stroke.
Under hypoxic conditions, both infiltrating macrophages and host cells produce angiogenic factors and cytokines that directly or indirectly control new capillary growth.1 Early events involve activation of early response genes, c-fos and c-jun, which further regulate gliosis and angiogenesis in infarcted tissue. Fos expression colocalizes with the expression of basic fibroblast growth factor, being highest at the infarct periphery.2 The areas in which neurons tend to survive longer are the same as those demonstrated to be highly angiogenic.3 4 5 6 Thus, angiogenic factors might be neuroprotective and crucial determinants of neuronal survival after stroke. Not much is known about the expression of other growth factors after ischemia. We recently reported that TGF-β1 might be involved in angiogenesis after ischemic stroke in humans7 ; here we present data from an extensive study on PDGF and its receptors.
PDGF is produced by many cells, including neonatal rat smooth muscle cells, stimulated fibroblasts,8 cytotrophoblasts,9 transformed and neoplastic cells,10 activated monocytes/macrophages,11 and arterial endothelial cells.12 Some endothelial cells express PDGF-AA and PDGF-BB both in vivo and in vitro.13
PDGF can be either homodimeric (PDGF-AA, PDGF-BB) or heterodimeric (PDGF-AB). The mature portions of the A and B chains show approximately 60% conservation of amino acid sequences and complete conservation of the positions of eight cysteine residues. Different cells express different ratios of the A and B chains of PDGF.14
There are two distinct PDGF receptors: an α-receptor (PDGF-Rα) and a β-receptor (PDGF-Rβ).15 PDGF-Rα binds each of the three forms of PDGF dimers with high affinity, whereas PDGF-Rβ binds only PDGF-BB with high affinity and PDGF-AB with lower affinity.
PDGF appears to be ubiquitous in neurons throughout the central nervous system, and it is suggested to play an important role in nerve regeneration, mediation of glial cell proliferation, and differentiation.16 17
Both PDGF-A and -B chains are expressed in mammalian neurons. They induce neuronal cell differentiation and increase levels of the midsize neurofilament protein. This suggests that PDGF isoforms are involved in neuronal cell migration, growth, and differentiation in human brain development.18
PDGF-A transcripts are first expressed in brain during late embryogenesis in most neurons, well before the differentiation of most glia.19 This strongly implies that neurons express PDGF-A to attract and maintain supporting glial elements within the central nervous system.20
The PDGF-B chain containing protein is localized in neurons throughout the central nervous system of adult nonhuman primates such as Macaca nemestrina.16 Positive immunohistochemical staining reactions were observed to be restricted to neuronal cytoplasmic perinuclear regions and principal or secondary dendrites. The intensity of reaction varied with the location of the neurons. In contrast, blood vessels stained faintly, and there was no staining in glial cells.
In persons undergoing intrastriatal embryonic mesencephalic implants, PDGF-B chain expression was increased in and around implants, mainly in astrocytes. This may have important implications for graft survival and function. Glial cells may utilize the elevated levels of PDGF-A and -B to proliferate, causing reactive gliosis. PDGF could lead to astrocyte and glial cell chemotaxis and proliferation.21 It is also possible that PDGF increases the survival of neurites and promotes their outgrowth.22 The latter is more likely because only PDGF-BB increases human astrocyte proliferation in vitro (J. Krupinski, M. Slevin, and P. Kumar, unpublished data, 1996). PDGF-AA and -BB treatment in an experimental rat model produces a significant increase in numbers of oligodendrocyte precursors, suggesting that PDGF may have an effect on recruitment into the oligodendrocyte precursor pool. PDGF subsequently increases the rate of oligodendrocyte precursor cell proliferation and the number of mature oligodendrocytes. Proteolipid protein and myelin basic protein mRNAs increase after addition of PDGF to cultured oligodendrocytes from rat cerebral white matter.23
Ischemia transiently increased mRNA expression of the PDGF-B chain but not the PDGF-A chain in injured neocortex.24 It was induced in neurons and later in macrophages and may have an important role in the healing process of injured brain. Takayama et al25 (1994) demonstrated that after experimental brain injury, shrunken neurons and numerous macrophages were present with enhanced PDGF-B chain–related immunoreactivity in the vicinity of the lesion. The distribution was closely related to neovascularization and astrogliosis.
Transcripts for PDGF-Rα and PDGF-Rβ subunits are expressed in neurons in nearly all regions of the brain.16 17 Brain capillary endothelial cells express functional PDGF-Rβ in normal brain of the nonhuman primate Macaca nemestrina; the PDGF released from injured neurons could contribute to a local angiogenic response.16
Although most macrovascular endothelial cells do not express PDGF receptors,13 PDGF-Rβ was demonstrated on endothelial cells of capillaries in rat brain22 and on proliferating capillaries in glioblastomas26 and wounds.27
In brain tumors, PDGF can be produced by the tumor cells themselves rather than as a result of angiogenesis after activation and derangement of the blood-brain barrier.28 PDGF-BB acts as a growth factor in meningiomas, initiating cell division via the β-receptor.29
PDGF has been shown to mediate interactions among glial cells in vitro.16 30 More recent evidence has indicated that PDGF may also be involved in controlling communication between neurons and glial cells and among neurons. The presence of receptors for PDGF on neurons of the developing nervous system is an essential piece of evidence. Ganglion cells are labeled with antibodies to PDGF-R only during the period of active process outgrowth. These findings suggest that PDGF is used as a mediator of intercellular signaling during neuronal development.
Materials and Methods
Human brains for in situ hybridization and Western blotting were obtained with permission from the ethics committee from the Department of Neuropathology, Jagiellonian University, Krakow, Poland (Table⇓). One normal human brain (from a patient who died in a road accident, aged 54 years) was obtained courtesy of Dr H. Reid, Department of Neuropathology, Manchester University (UK). Human fetal brain aged 16 weeks was obtained from Professor D. Donnai, Department of Medical Genetics, Manchester University (UK).
Competent cells were transformed, and plasmid was extracted and purified. The vector was cut by restriction endonucleases, and the insert was recovered from an agarose gel. A mouse PDGF-R cDNA (1161 bp) cloned in an ampicillin-resistant pUC9 plasmid vector, pmPDGFrec-HincII, was a kind gift of the late Professor H. Antoniades. PDGF-A and PDGF-B oligoprobes were purchased from R&D Systems. The probes were random prime labeled to specific activities of 1×106 to 1×108 cpm/mg using [35S]α-dCTP. Duplicate slides were treated with RNase A to estimate the contribution of nonspecific signal to the overall labeling.
In Situ Hybridization
This was carried out as described previously for TGF-β1.7 Grains were counted using a Cambridge Seescan Image Analysis System. At least 300 cells were counted per tissue section, ie, approximately 50 cells of each type.
Goat anti-human polyclonal antibody to PDGF-AB was purchased from Collaborative Biomedical Products, and rabbit anti-human polyclonal antibodies to PDGF-BB, PDGF-Rα, and PDGF-Rβ were purchased from Santa-Cruz. All antibodies were used at a dilution of 1:100 in PBS, pH 7.6. After dewaxing and rehydration, sections were treated with 1% saponin in PBS, pH 7.6, for 30 minutes. Blocking of endogenous peroxidase activity was carried out in 0.3% H2O2 for 5 minutes followed by incubation in 1% BSA for 20 minutes. The secondary antibody was goat anti-chicken peroxidase conjugated (DAKO) used at a dilution of 1:200 in PBS for 45 minutes. Peroxidase activity was visualized using diaminobenzidine (0.006% in dH2O) for 10 minutes followed by dehydration, clearing in xylene, and permanent mounting in DPX.
Protein extraction, determination, and blotting on PVDF membranes were carried out as described previously for TGF-β1.7 The protein concentration of each sample was determined using an assay developed by Bradford31 to ensure loading of equal amounts during Western blotting. Briefly, proteins were transferred to PVDF membrane (Immobilin), prewetted in methanol, and equilibrated in Towbin buffer containing 192 mmol/L glycine, 25 mmol/L Tris-HCl, 1% SDS, and 20% methanol, pH 8.3, using a Hoefer electroblotting apparatus (1 hour, 0.8 mA/cm2 gel). After blotting, membranes were blocked with blocking buffer (7.5% Marvel in TBS-Tween, pH 7.4) for 1 to 24 hours; membranes were then stained with primary antibody in blocking buffer at a dilution of 1:100 for 1 to 24 hours. The same antibodies were used as for immunohistochemistry. Membranes were subsequently rinsed in blocking buffer (5 3-minute washes) and then stained with the appropriate peroxidase-conjugated secondary antibody in blocking buffer for 1 hour, after which they were rinsed in TBS-Tween, pH 7.4 (5 3-minute washes), and the proteins were visualized using an ECL kit (Amersham).
Semiquantitative Protein Estimation on Blots
The relative amounts of different proteins were estimated directly from x-ray film by densitometric scanning using an LKB 2202 ultrascan laser densitometer. In every case, stroke tissue proteins were compared with those of contralateral hemisphere (used as controls), and the percentage was calculated (relative to a control of 100%).
PDGF Protein Expression in Tissue Sections
PDGF-AB immunoreactive products were present in both penumbra and infarct. All cell types expressed PDGF-AB. There was stronger staining in infiltrating monocytes/macrophages and glial cells in the vicinity of proliferating blood vessels (Fig 1a⇓ through 1c). PDGF-AB was also present in the control contralateral hemisphere, but the staining was weaker (Fig 1c⇓).
Western Blotting of PDGF
There was remarkable consistency among three membranes obtained by Western blotting, indicating that the percentage values were accurate for each tissue studied.
In 6 of 9 patients, there was more PDGF-AB in infarct/penumbra tissue than in contralateral hemispheres: approximately 263%, 410%, 528%, 35%, 75%, 82%, 108%, 185%, and 158% (Fig 2a⇓). When samples were taken from areas of penumbra alone in 2 patients, there was no change in PDGF-AB expression compared with infarct (Fig 2b⇓ and 2c⇓). The main differences between control and affected brain tissues were in white matter: 255% and 186% in infarct and penumbra in patient 32 and 585% and 1021% in infarct and penumbra in patient 49, respectively.
There was higher expression in infarct/penumbra than in contralateral controls in all patients studied: approximately 1720%, 275%, 140%, 290%, 500%, 213%, 180%, and 323% (Fig 3a⇓). In 2 patients, the penumbra was studied separately, and it contained less PDGF-BB than control contralateral hemisphere (67% and 63% in the first patient and 40% in the second) (Fig 3b⇓ and 3c⇓). Gray matter and white matter samples taken from infarcts, penumbras, and contralateral control areas were studied in the same 2 patients. The highest differences were in white matter between infarcts and contralateral hemispheres: approximately 470% and 1830% in patients 32 and 49, respectively. In patient 49, the white matter of the penumbra showed an increase to 167% and 533%. However, in patient 32, the white matter of the penumbra showed a decrease to 40%.
In Situ Hybridization for PDGF mRNA
There was weak PDGF-A mRNA transcription overall. Only neurons within penumbra contained a significant signal (more than 5 grains per cell), which was higher than the control contralateral hemisphere (P<.0001, Spearman’s rank correlation test). In infarcts, neurons expressed fewer grains than in control. In areas of active gliosis within the penumbra, astrocytes expressed more grains than in the other areas. PDGF-A mRNA was found in only a few large blood vessels of infarct or penumbra (Fig 4a⇓ through 4c, Fig 5a⇓).
Only type I/II neurons in penumbra had higher expression than in normal contralateral hemisphere (P<.001, Spearman’s rank correlation test). In the border zones between infarct and penumbra, astrocytes produced more grains than controls. Endothelial cells in infarcts and penumbra had higher median grain numbers than controls, but grains were found only in some blood vessels and then mostly in single endothelial cells. If the latter were surrounded by microglia/macrophages or lymphocytes, the number of grains was smaller (Fig 4d⇑ through 4f, Fig 5b⇑).
PDGF-R Protein Expression in Tissue Sections
Most neurons, but only some gigantic astrocytes, were stained with antibody to PDGF-Rα in stroke tissue (Fig 1d⇑). In contralateral normal hemisphere, there was a similar pattern of staining but in fewer cells (Fig 1e⇑). White matter was virtually negative in both infarcted and contralateral hemisphere (Fig 1f⇑).
There was more staining obtained with antibody to PDGF-Rβ in infarcted/penumbra tissue (Fig 1g⇑). It was localized to some endothelial cells, reactive astrocytes, and fewer neurons compared with PDGF-Rα staining. In contralateral control hemispheres, staining was only present in neurons (Fig 1h⇑). Some blood vessels in the white matter of infarcted hemispheres were positive but not in contralateral hemispheres (Fig 1i⇑).
Western Blotting of PDGF-R
In all 3 patients studied (patients 49, 50, and 51), there was more PDGF-Rα in stroke and penumbra compared with the contralateral hemisphere. This increase in gray matter was to 115%, 182%, and 147% in infarct and to 142%, 176%, and 145% in penumbra, respectively. In the white matter of patients 49 and 50, there was a decrease in PDGF-Rα to 85% and 70% in infarct and to 55% and 72% in penumbra, respectively (Fig 6a⇓ and 6b⇓). In patient 51, there was an increase in white matter to 181% and 122% in infarct and penumbra, respectively (Fig 6c⇓). In normal human brain, PDGF-Rα was mainly expressed in gray matter, with less expression in white matter. In human fetal brain, expression was reserved to gray matter (Fig 6d⇓).
In all patients studied (patients 49, 50, and 51), there was more PDGF-Rβ in infarct and penumbra compared with the contralateral hemisphere. This increase was observed in gray matter to 120%, 298%, and 151% in stroke and to 124%, 296%, and 183% in penumbra, respectively (Fig 7a⇓ through 7c). In the white matter of patients 49 and 50, there was a decrease in PDGF-Rβ to 68% and 81% in infarct, respectively. In the penumbra, patient 49 showed a decrease to 96%, and patient 50 showed an increase to 178% (Fig 7a⇓ and 7b⇓). In patient 51, PDGF-Rβ could not be detected in control white matter. In normal human brain, PDGF-Rβ was mainly expressed in gray matter, with less expression in white matter. In human fetal brain, expression was reserved to gray matter (Fig 7d⇓). Thus, the pattern of expression of PDGF-Rα and PDGF-Rβ protein was the same.
PDGF-R mRNA Expression
The highest expression was around types III and IV ischemic neurons32 in infarcted areas. Proliferating endothelial cells produced a lot of grains, both in penumbra and infarct. Macrophages in the vicinity of blood vessels contained more grains than macrophages elsewhere. Astrocytes, especially large astrocytes in the penumbra, expressed more grains than macrophages. Generally, white blood cells and oligodendrocytes did not express PDGF-R mRNA (Fig 4g⇑ through 4i, Fig 5c⇑).
This study demonstrated an increase in PDGF both at the mRNA and protein levels in stroke infarcts and penumbras compared with contralateral hemispheres. PDGF-A mRNA was expressed mainly in neurons and larger blood vessels in the areas of penumbra. PDGF-B mRNA was expressed in type I/II neurons32 of penumbra and some microvessel endothelial cells in stroke/penumbra. Astrocytes expressed both PDGF chains in some areas, mainly in white matter and in a limited number of cells. The highest differences shown by Western blotting were for PDGF-B chain, and they were attributable to an increase within infarcts rather than in penumbras. The most pronounced increase was in tissue extracts from white matter compared with gray matter.
The highest grain counts for PDGF-R mRNA were in type III/IV neurons within ischemic infarcts. At the protein level, there was higher expression of both PDGF-Rα and PDGF-Rβ in infarcts and penumbras than in contralateral normal-looking hemispheres. This was mainly localized in neurons in gray matter. PDGF-Rβ was also expressed in some microvessels.
Our results demonstrated greater protein expression of PDGF in white matter as opposed to gray matter. Thus, PDGF released in brain may contribute to nerve regeneration and to glial cell proliferation that leads to gliosis and scarring. Both glial cells and neurons were found to synthesize PDGF.16 17 Glia secrete molecules important for neuronal function, and in turn glia may be directed toward and supported by the neuronal elements they serve.
It is also possible that neurons play a part in the general response to injury. For example, the signals responsible for astrocyte proliferation at the site of an injury are not known. The neuronal secretion of PDGF may be ideally suited to stimulate migration and division of additional glial cells in proximity to the neuron. It is important that neuronal expression of PDGF in the central nervous system is ubiquitous, which is consistent with its general function as a growth/differentiation factor for glia.16
The previously reported localization of PDGF-B chain in axons and probable terminals in the spinal cord, brain stem, hypothalamus, and pituitary suggests a possible neuroregulatory role for PDGF.16 In our study, PDGF-B chain was expressed more in white matter of stroke/penumbra than in contralateral hemisphere. A number of effects of PDGF, including induction of c-fos in neurons,22 stimulation of tyrosine phosphorylation,33 inhibition of gap-junctional communication,34 and increase in amino acid transport, are consistent with possible involvement in synaptic transmission.
Immunohistochemical studies demonstrated that activated/proliferating endothelial cells expressed PDGF-AB. They were surrounded by glial cells and cell infiltrates, which were also strongly stained. Other authors have suggested that it is PDGF-BB that participates in angiogenesis.35 36 PDGF-BB can elicit endothelial cell proliferation and the formation of endothelial cords/tubes in vitro via direct action on PDGF-Rβ that are expressed on proliferating but not quiescent endothelial cells.37 In our studies, PDGF-Rβ but not PDGF-Rα was expressed on some of the endothelial cells in stroke infarcts and penumbras. Angiogenic endothelial cells did not express the PDGF-B chain, whereas quiescent nonangiogenic endothelial cells did express PDGF-B chain transcripts. This suggests that angiogenic endothelial cells are not the source of PDGF-B chain. It is possible that PDGF-BB plays a part in early recruitment of endothelial cells before the action of other cytokines, such as TGF-β and interleukin-1 or -6. The latter may switch off PDGF-BB in proliferating endothelial cells. PDGF may control in vivo angiogenesis via the action of other cells.38 Any direct action of PDGF in angiogenesis would require the expression of PDGF receptors on neovascular endothelium. Raines et al13 (1990) demonstrated that quiescent endothelial cells in vitro originating from microvessels do not express PDGF receptors. However, PDGF-B receptors were demonstrated on endothelial cells of capillaries in the rat brain,22 in glioblastomas,26 and in wounds.27 In our studies, ischemic endothelial cells expressed mRNA for PDGF-R, but mainly in areas where they were surrounded by either infiltrating cells or glia. This may confirm that indeed endothelial cell expression of PDGF-R is secondary and is induced by other cells. PDGF-Rβ might be of particular interest as a marker for a population of phenotypically different endothelial cells undergoing angiogenesis, although angiogenic endothelial cells themselves are not the source of PDGF-BB.
The higher abundance of PDGF-R protein in the gray matter suggests that this growth factor might be important in remodeling and neuronal plasticity after ischemia. Interestingly, it is mainly white matter that contains high amounts of PDGF-AB and PDGF-BB protein; thus, although mRNA is produced in neuronal cell bodies in gray matter, PDGF may exert its function in white matter, participating either in regeneration of damaged axons or in glial scar formation. If the latter process could be controlled, this might decrease scar size after ischemia and increase the possibility for formation of new contacts between partially damaged neurons.
Selected Abbreviations and Acronyms
|PDGF||=||platelet-derived growth factor|
|TGF||=||transforming growth factor|
- Received June 12, 1996.
- Revision received October 4, 1996.
- Accepted October 10, 1996.
- Copyright © 1997 by American Heart Association
Krupinski J, Kaluza J, Kumar P, Kumar S, Wang J. Prognostic value of blood vessel density in ischaemic stroke. Lancet. 1993;342:742. Letter.
Krupinski J, Kaluza J, Kumar P, Kumar S, Wang J. Role of angiogenesis in patients with cerebral ischemic stroke. Stroke. 1994;25:1794-1798.
Krupinski J, Kumar P, Kaluza J, Kumar S. Increased expression of TGF-β1 in brain tissue after ischemic stroke in humans. Stroke. 1996;27:852-857.
Pech M, Rao C, Robbins K, Aaronson S. Functional identification of regulatory elements within the promoter region of platelet-derived growth factor II. Moll Cell Biol. 1989;9:396-405.
Ross R, Masuda J, Raines E, Gown A, Katsuda S, Sasahara M, Malden L, Masuko H, Sato H. Localization of PDGF-B protein in macrophages in all phases of atherogenesis. Science. 1990;248:1009-1012.
DiCorleto P, Bowen-Pope D. Cultured endothelial cells produce a platelet-derived growth factor like protein. Proc Natl Acad Sci U S A. 1983;80:1919-1923.
Matsui T, Heideman M, Miki T, Popescu N, La Rochelle W, Kraus M, Pierce J, Aaronson S. Isolation of a novel receptor cDNA establishes the existence of two PDGF receptor genes. Science. 1989;243:800-804.
Raff M, Hart I, Richardson W, Bolsover S. PDGF and intracellular signaling in the timing of oligodendrocyte differentiation. J Cell Biol. 1989;109:3411-3417.
Pomeroy S, Purves D. Neuron/glia relationships observed over intervals of several months in living mice. J Cell Biol. 1988;107:1167-1175.
Hermanson M, Funa K, Hartman M, Claesson-Welsh L, Heldin C, Westermark B, Nister M. Platelet-derived growth factor and its receptors in human glioma tissue: expression of messenger RNA and protein suggests the presence of autocrine and paracrine loops. Cancer Res. 1992;52:3213-3219.
Reuterdahl C, Sundberg C, Rubin K, Funa K, Gerdin B. Tissue localization of beta receptors for platelet-derived growth factor and platelet-derived growth factor B chain during wound repair in humans. J Clin Invest. 1993;91:2065-2075.
Eke A, Conger C, Anderson M, Garcia J. Histological assessment of neurons in rat models of cerebral ischemia. Stroke. 1990;21:299-304.
Pierce J, Heidaran M, Yu J, Jensen R, Aaronson S. A deletion in the extracellular domain of the alpha platelet-derived growth factor (PDGF) receptor differentially impairs PDGF-AA and PDGF-BB binding affinities. J Biol Chem. 1992;267:2884-2887.
Battegay E, Rupp J, Iruela-Arispe L, Sage E, Pech M. PDGF-BB modulates endothelial proliferation and angiogenesis in vitro via PDGF-BB receptors. J Cell Biol. 1994;125:917-928.