There is currently some speculation as to whether the pancreas and liver-associated forms of
GPI-PLD represent two distinct sub-types of the GPI-PLD enzyme (Tsang et al., 1992).
Support for the existence of two isotypes came from the extraction of bovine, rat and human liver membrane GPI-PLD enzymes (Genbank accession numbers L11701 and LI 1702).
Chromatographic analysis of the enzyme fractions on Fractogei TSK-DEAE columns revealed two peaks of anchor-converting activity, termed PI and P2, which eluted at 50 - 80 mM and
1 2 0 - 170 mM sodium chloride concentration, respectively (Heller et a l, 1992). It appears that
PI represents a membrane-bound enzyme fraction, whereas P2 was only loosely associated with the membrane. Comparison of the PI and P2 forms against serum GPI-PLD indicated that the serum form eluted at 120 - 170 mM sodium chloride, suggesting that P2 was serum- derived.
Challenge to the theory that two forms of the enzyme exist comes from analysis of the genetic sequences of the two postulated GPI-PLD ‘isotypes’. Analysis revealed only 42 single-base changes between the two forms. This could be explained by genetic polymorphisms in the
Gpld gene, or by minor discrepancies in the sequencing technique (LeBoeuf et a l, 1988).
3.30.4 Secretion of the GPI-PLD protein
The study of GPI-PLD to date shows that, although the enzyme is found in high concentrations in serum, only a limited number of tissues are capable of producing the mRNA. It must be assumed, therefore, that the enzyme is secreted into the circulation, from where it would be available to all of the body’s tissues.
Evidence that GPI-PLD might be secreted from cells was gained from in vitro experiments,
using cell lines. For example, a plasmid containing the bovine GPI-PLD sequence (pBJ1682) was introduced into COS-1 cells. Expression of the transfected enzyme was confirmed by immunofluoresence of permeabilised cells, and by Western blot analysis, and enzyme activity was monitored by the degradation of radiolabelled mfVSG. Analysis of the cells lysates indicated that 48 hours after transfection, only a small fraction of the GPI-PLD was associated with the cells, and it was concluded that the majority o f the enzymatic activity was secreted
into the extracellular milieu (Scallon et a l, 1991).
Additional information, concerning the secretion of GPI-PLD, came from the study of the mouse pancreatic islet cell line, pTC3 (Deeg & Verchere, 1997), which were shown to be
phenotypically similar to normal P cells (d’Ambra et a l, 1990). The use of a cell line
permitted the study of specific pancreatic cells without the influence of contaminating cells. The cells were cultured in a GPI-PLD-free serum substitute, to remove contaminating serum- derived GPI-PLD.
GPI-PLD activity was demonstrated by the hydrolysis of a radiolabelled mfVSG substrate, and confirmation that the product of hydrolysis was phosphatidic acid. Secretion from PTC3 cells
was also shown to be stimulated by the addition of a number of insulin secretagogues, including glucose, phorbol myristic acid and isobutylmethylxanthine, and GPI-PLD release occurred in parallel with the release of insulin.
Two conclusions were made, regarding the release of GPI-PLD from the PTC3 cell line. Firstly, the overlap between the secretion of insulin and GPI-PLD suggested that there are common elements in the signal transduction pathways that mediate the release o f the two proteins, although the nature of these pathways was not established. Secondly, the release of GPI-PLD is almost certainly dependent on both a constitutive and a regulated pathway. Evidence for a regulated pathway was derived from study of the cellular and extracellular GPI- PLD activity. Stimulation of pTC3 cells promoted the decrease of the cellular GPI-PLD activity, with a concomitant increase in the activity detected in the extracellular medium. The possibility that GPI-PLD exists in the insulin secretory granule was provided by immuno- histochemistry, which showed a punctate staining pattern for both insulin and GPI-PLD. Evidence for a constitutive pathway came from inhibition of GPI-PLD secretion by the addition o f the protein synthesis inhibitor, cyclohexamide. This inhibition suggested that at least a
certain proportion of GPI-PLD is synthesised de novo. Both the regulated and the constitutive
theories, however, require further experimentation (Deeg & Verchere, 1997).
3.30.5 Production of GPI-PLD by mast cells
As discussed, a number of studies have been performed to determine the production of GPI- PLD by mammalian tissues. One such study was concerned with human tissues, in which the researchers concluded that mast cells had the capacity to produce the GPI-PLD enzyme (Metz
et a l, 1992). The studies employed an anti-GPI-PLD monoclonal antibody, 612C, in immuno-
staining procedures using human tissues. Results indicated that 612C stained sparsely
distributed cells in the lung interstitium, gastric sub-mucosa and dermis. These cells
corresponded to cells that also stained with toluidine blue, a dye routinely used in the identification of mast cells granules. These studies, however, were designed to detect the protein, rather than the existence of GPI-PLD-specific mRNA.
Other studies, using RNA probes generated against bovine GPI-PLD, and screening of bovine tissues have provided evidence of production o f the enzyme in sub-populations of mast cells
found in the adrenal gland, lung and liver (Stadelman et a l, 1993). Using an alcian blue 8GX,
the GPI-PLD positive cells were identified as mast cells. To date, however, this is the only evidence that mast cells may produce the GPI-PLD enzyme, and these results are contradicted by the data collected during the course o f this thesis.
The observation that normal tissue mast cells may produce the GPI-PLD protein in situ lead us to believe that the GPI-PLD RNA would, likewise, be detected in the RBL-2H3 cell line. This was not the case. The experimental results collected during this thesis revealed that the RBL- 2H3 cell line does not produce the GPI-PLD RNA, although it does contain an enzymatically active GPI-PLD protein. It was, therefore, assumed that the protein was derived from the serum. This is an important observation. The aim of the research performed in this thesis was to determine the role of the GPI-PLD enzyme in the Type One Hypersensitivity cascade. It is, therefore, important to note that a potentially key enzyme in the allergic cascade is derived from the serum, rather than being created in an intracellular location.
3.30.6 Uptake of serum GPI-PLD by RBL-2H3 cells
The uptake of GPI-PLD from the extracellular culture medium into cells has been observed in one other cell line; the mouse neuroblastoma cell line, N2A, in which the uptake of bovine- serum GPI-PLD was studied. The results demonstrated that both intact GPI-PLD, and trypsin- treated fragments of the enzyme, were taken up in a concentration- and time-dependent way.
Cell-associated GPI-PLD activity also increased in a linear fashion (Hari et a l, 1997).
In addition, experimental studies have been performed, to determine the production of GPI- PLD in the placenta (personal communication with Ms S Deborde, Department o f Molecular Pathology, UCL). As was observed with the RBL-2H3 cell line, the RNA for the GPI-PLD was not detected in any of the placental components, using the RTR-PCR methodology. However, an active GPI-PLD protein was detected within the foetal syncitiotrophoblast, and it was assumed that the enzyme was taken up from the reserves of maternal serum GPI-PLD.
The observation that the RNA for GPI-PLD is detected in very few mammalian cell types, and that both RBL-2H3 cells and the placental tissues may derive their active GPI-PLD protein from the serum leaves a vital question unanswered. Why is the GPI-PLD protein not created intracellularly? Experimental evidence to answer this question was provided from studies in which the GPI-PLD protein was over-expressed in mammalian cells.
The GPI-PLD enzyme was over-expressed in macrophages, and the resultant effects on cellular signalling were determined. Results showed that the over-expression of GPI-PLD promoted the GPI-PLD-mediated cleavage of the GPI anchors in the early secretory pathway. The phosphatidic acid produced by GPI-PLD was rapidly converted into diacylglycerol (DAG),
through the action of an intracellular phosphatase (Singer et a l, 1997). The raised DAG levels
resulted in the translocation of intracellular PKCa to the surface membrane, ultimately promoting the phosphorylation, and activation, of PKCa (Tsujioka et al 1999). Furthermore,
PK C a was also translocated to the membrane of the endoplasmic reticulum, as determined by sucrose-density-gradient centrifugation and immunofluorescence microscopy.
Taken together, these results place GPI-PLD in the signalling cascade initiated through the
generation of diacylglycerol. Although the DAG is a by-product of GPI-PLD mediated
catalysis, being generated from the action of a phosphatase on phosphatidic acid, nevertheless DAG plays a pivotal role in the activation of PKCa. Through the activation of protein kinase C a number of intracellular signalling pathways are likely to be affected.
Firstly, the role of PKC in the IgE-mediated degranulation of mast cells was discussed in Section 1.3.33.6. The potential exists that the activation of mast cells might be as a result of the GPI-PLD-mediated generation of DAG, and the resultant activation of PKC. Through the activation of the PKC enzyme, mast cells could undergo degranulation, without the initial IgE- FceRl cross linking event. Clearly, the non-specific activation of mast cells, with the release of allergic mediators, would be an unwanted event.
Secondly, it was recently suggested that PKC regulates the translocation o f newly synthesized
proteins, including docking protein a, across the endoplasmic reticulum membrane (Gruss et
al., 1999). Indeed, the phosphorylation of microsomal proteins by PKC enzymes improved their
translocation efficiency in vitro. Overall, it appears likely that the PKC enzymes are involved
in the early stages of cellular secretion.
This evidence suggests that, if the GPI-PLD were present in an intracellular location in the mast cells, it may promote the non-specific activation of the cells. The evidence provided from these experimental studies, therefore, support the results generated in this chapter. That is, it is preferential for the mast cells to take up the GPI-PLD protein from the serum, rather than creating the protein intracellularly.
3.30.7 Intracellular Location of GPI-PLD
The intracellular location of GPI-PLD was studied through the use of labelled anti-GPI-PLD antibodies, which were employed in immuno-histochemical analysis of a variety of cell lines. A punctate staining pattern was observed in a variety of myeloid cell lines, including human myeloblasts and mouse macrophages (Xie & Low, 1994), and in a pancreatic |3TC3 cell line (Deeg & Verchere., 1997). This punctate pattern was taken to mean that the enzyme is concentrated in limited intracellular locations, such as secretory vesicles (Xie & Low, 1994).
Secondly, centrifugation techniques were employed to separate washed rat liver tissue into different cellular fractions, according to the density of the fraction. Using this technique, a nuclear fraction, heavy mitochondrial fraction, light mitochondrial or lysosomal fraction, microsomal fraction and supernatant were purified. Analysis of fraction-specific marker enzymes determined the purity of the fraction, and the GPI-PLD activity in each fraction was then quantified.
The results indicated that the majority o f enzyme activity was detected in the light mitochondrial or lysosomal fraction, suggesting a lysosomal or peroxisomal localisation for GPI-PLD. However, the subcellular distribution pattern o f GPI-PLD showed more similarities with the lysosomal marker enzymes, including (3-N-acetyl-D-glucosaminidase, rather than the
peroxisomal marker enzyme, catalase (Hari et al., 1996). The observation that P-N-acetyl-D-
glucosaminidase may be co-localised with GPI-PLD, combined with the observation that P-N- acetyl-D-glucosaminidase is released concomitantly with histamine, as discussed in Section 2.8.3, raises the possibility GPI-PLD may be located in the same granules as histamine. The relevance o f this observation in the study of mast cell degranulation certainly warrants further investigation.
The third line of evidence for the intracellular location of the GPI-PLD enzyme was provided from the mouse neuroblastoma cell line, N2A, in which the uptake of purified bovine-serum GPI-PLD was previously demonstrated (Section 3.30.6). Chloroquine is a weak base, which is
rapidly taken up by cells. It accumulates primarily in the lysosomes (de Duve et al., 1966) and
serves to neutralise these compartments. As a consequence of the reduced acidity of
lysosomes, the activity of the lysosomal proteolytic enzyme is inhibited (Wibo & Poole, 1974), resulting in reduced levels of protein degradation. Experimentally, the inclusion of chloroquine in the uptake of radiolabelled GPI-PLD into N2A cells resulted in a reduced degradation of the GPI-PLD, as detected by SDS-PAGE. The conclusion was made, therefore, that GPI-PLD was located in the lyosomal compartment o f the cell, in which it would be in contact with
chloroquine (Hari et al., 1997).