|
Physician/Scientist
The Biochemical Defect in
Types A and B Niemann-Pick Disease
Sphingomyelin Chemistry and Metabolism
Sphingomyelin Chemistry
The underlying biochemical defect in Types A and B NPD is the deficient
activity of ASM, resulting in lysosomal accumulation of sphingomyelin and
secondary increases in the concentrations of cholesterol, and other
metabolically related lipids (e.g., bis- (monoacylglycero)phosphate, etc.).
Sphingomyelin is a phospholipid composed of a long chain base, generally
sphingenine (4-amino-2-octadecene-1,3-diol), a long chain fatty acid of
varying length, and a phosphocholine moiety. It is a common component of
the plasma membrane, subcellular organelles, ER, and mitochondria, and is
the major phospholipid of the myelin sheath and erythrocyte stroma.
Sphingomyelin composes from 5 to 20 percent of the total phospholipid in
most cell types and is primarily localized in the plasma membrane. First
discovered by Thudichum in 1884, it can be distinguished from other
phospholipids by a variety of unique chemical properties, including the
fact that it contains a 2:1 ratio of nitrogen to phosphorus.
Numerous procedures have been developed for the isolation of
sphingomyelin. Brain is the most frequent source, but lung, spleen, and
liver also have been used. Purified sphingomyelin is a white powder that
is only slightly soluble in cold alcohol or pyridine. It is completely
insoluble in acetone and diethyl ether, but can readily form emulsions in
water that exhibit birefringence. Other optical properties have been
noted, including a dextro-rotation in methyl alcohol/chloroform and
pyridine solutions. It has been suggested that sphingomyelin occurs as a
zwitterion with an isoelectric point of about 6.0.
The initial insights into the structure of sphingomyelin were derived
from studying its hydrolysis products, and it has long been
recognized that a fatty acid, two nitrogenous bases, choline and
sphingosine, and a phosphoric acid were essential components. The fatty
acid moiety may vary, but has been most frequently identified as lignoceric
acid. Palmitic and stearic acid moieties have also been found, as has the
unsaturated fatty acid, nervonic acid. Thannhauser and Boncoddo found that
brain sphingomyelin had a different fatty acid composition than
sphingomyelin isolated from other tissues. Stearic, lignoceric, and
nervonic acids composed nearly all of the fatty acid moieties in brain
sphingomyelins, while palmitic and lignoceric were the only fatty acids
present in lung and spleen sphingomyelins. Frankel et al. proposed
that brain sphingomyelin consists of salt-like complexes formed by the
condensation of several adjacent molecules of choline and phosphoric acid,
and that such long-chain polyaminophospholipids may play a central role in
producing nerve conductivity.
[top] [topics]
Sphingomyelin Synthesis
The major pathway for sphingomyelin synthesis is the enzymatic condensation
of ceramide and phosphocholine. The enzyme catalyzing this reaction,
phosphocholine ceramide transferase (i.e., sphingomyelin synthase), has
been identified in various tissues, including liver, kidney, spleen, and
brain. Sphingomyelin synthase transfers phosphocholine from
cytidine-5'-diphosphocholine (CDP-choline) to ceramide. Notably, there is
a stereochemical requirement necessary for this condensation reaction to
occur; that is, the sphingosine moiety of active ceramide must have the
trans configuration of the double bond, and the hydroxyl group on carbon 3
must have the threo relationship to the amino group on carbon 2. Since
most of the naturally occurring sphingosine is in the erythro form, an
isomerization must take place to allow the reaction to proceed. However,
the nature of this isomerization has not been satisfactorily determined.
In addition to the condensation pathway, other synthetic pathways for
sphingomyelin have been suggested, but not definitively proven. For
example, Funjino and Negishi have described a pathway in brain mitochondria
in which sphingosine acts as the acceptor for CDP-choline. The
sphingosylphosphocholine formed then serves as an acceptor for fatty acyl
CoA, thus forming sphingomyelin. In addition, the direct transfer of
phosphocholine from phosphatidylcholine to ceramide has been demonstrated
in various cell types, and there is some evidence to suggest that
phosphatidylcholine, as opposed to CDP-choline, may be the major source of
phosphocholine for sphingomyelin synthesis, as opposed to CDP-choline
(Merrill and Jones, 1990). The subcellular site of sphingomyelin synthesis has
not been well characterized; however, there is evidence for intracellular
synthesis and translocation of sphingomyelin to the plasma membrane via the
cis cisternae of the Golgi apparatus.
[top] [topics]
Sphingomyelin Degradation
The catabolism of sphingomyelin has been intensively studied. To date,
four distinct mammalian sphingomyelinase activities have been described: 1)
an intracellular, cation-independent activity with an acidic pH optimum
and lysosomal localization (i.e., ASM), 2) a Mg2+-dependent neutral
sphingomyelinase located at the cell surface, 3) a Mg2+-independent
neutral sphingomyelinase, and 4) a secreted, Zn2+-dependent enzyme that
also has an acidic pH optimum (see below). In addition,
sphingomyelinases have been identified in various microorganisms, including
Pseudomonas aeruginosa, Staphylococcus aureus, Bacillus cereus, and
Caenorhabditis elegans (Lin et al. 1998). For the purpose of this review,
only the human acid sphingomyelinase activities, which are deficient in
Types A and B NPD, will be discussed in detail.
[top] [topics]
Acid Sphingomyelinase
The existence of a sphingomyelin cleaving enzyme was demonstrated in 1940
by the pioneering work of Thannhauser and Reichel. During the subsequent
25 years, similar enzymatic activities were demonstrated in various
tissues, including liver, brain, and kidney. However, it was Kanfer et
al. who first substantially purified a sphingomyelinase activity
from rat liver in 1966 and characterized its physical and kinetic
properties. Shortly thereafter, Brady and coworkers demonstrated the
deficiency of this enzymatic activity in biopsied livers from patients with
Type A NPD.
Subsequently, human ASM (EC 3.1.4.12) was purified from various sources,
including placenta, brain, and urine. Although its central role in the
pathophysiology of NPD has been known for nearly 30 years, until recently
the physicokinetic properties of this important hydrolytic enzyme were not
well characterized. For example, the molecular weight estimates for ASM
varied from about 70 to over 300 kDa. In addition, the catalytically
active form of the enzyme has been described as a monomer, dimer, and
tetramer, and conflicting reports have appeared concerning its substrate
specificity and kinetic properties. Undoubtedly, some of this variation
results from the fact that different purification procedures and enzyme sources
have been used to isolate ASM, and that the purified enzyme aggregates
readily. Furthermore, a variety of radioactive, colorimetric, and
fluorescent substrates have been employed to determine the enzyme's kinetic
properties, providing an additional source of variation.
The isolation of the full-length cDNA encoding human ASM has helped to
clarify some of these inconsistencies and provide further insights into the
biology of ASM (see "Molecular Genetics of Acid
Sphingomyelinase").
Based on the full-length cDNA sequence, the calculated molecular weight of
the ASM polypeptide is about 64 kDa, and there are six potential
N-glycosylation sites. Thus, it can be predicted that the glycosylated ASM
monomer is likely to be about 72 to 74 kDa, consistent with the more recent
molecular weight estimates. Site-directed mutagenesis studies have further
revealed that five of the six glycosylation sites in human ASM are
utilized, and that the two C-terminal sites are the most important for
enzyme maturation and activity (Ferlinz 1995).
Limited studies have been performed on the kinetic properties of ASM. In contrast to the neutral sphingomyelinase, it had
been reported that ASM does not require divalent cations for catalytic
activity since chelators such as EDTA do not
inhibit the enzymatic activity. However, more recent data have shown that
ASM is, in fact, a zinc metalloprotein (see "Relationship Between
Intracellular and Secretory ASM" below). In vitro, ASM activity is clearly
dependent on the presence of detergents, and Yedgar and Gatt 111 have shown
that when the Triton X-100:sphingomyelin ratios are 4:1 or greater, the
reaction displays regular Michaelis-Menten kinetics. Km values ranging
from about 10 to 500 ?M have been reported dependent on the substrate and
assay system.
There have also have been a number of reports suggesting that one or
more sphingolipid activator proteins (SAPs) may influence the activity of
ASM, thus performing a similiar function in vivo that detergents such as
Triton X-100 play in vitro. However, recent data obtained from knock-out
mice in which the SAP precursor protein was disrupted by gene targeting
have shown that the ASM activity was normal or near normal in various
tissues from these animals, and that no sphingomyelin was accumulated
(Suzuki and Vanier). The earlier results may be explained by the fact that
the ASM polypeptide contains a domain in the N-terminal portion that has
high homology to the SAP proteins. Thus, it may be hypothesized that ASM
does not require exogenous SAP proteins in order to carry out its catalytic
function because it is able to do so using its own "SAP-like" sequences.
This "SAP-like" domain is separated from the remainder of the protein by a
proline-rich (hinge) region, perhaps linking it to a catalytic region.
Various inhibitors of ASM activity have been identified, including
5'-adenosine monophosphate (5'-AMP), tricyclic antidepressant drugs such as
midalcipran, cationic amphiphilic drugs, and nephrotoxic aminoglycosides
such as gentamicin. The data using 5'-AMP suggested that the inhibition
was noncompetitive and reversible, and that the inhibitory mechanism
involved the combined effect of both the phosphate group and the purine
ring. Octylglucoside is also a potent inhibitor of ASM activity, and
octyl-Sepharose columns have been employed in many of the purification
schemes. From these and other results, investigators have concluded that a
carboxyl group (perhaps donated from an aspartate or a glutamate residue)
and a protonated histidine are involved in sphingomyelin binding to the
enzyme. Although the active site specificity involves both the ceramide
and diacylglycerol moieties, the presence of the phosphodiester linkage is
the only absolute requirement.
Apo C-III also has been shown to stimulate ASM activity in vitro. This
is a particularly intriguing finding in light of reports that
sphingomyelinase is responsible for the aggregation and uptake of LDL
particles into cultured cells (Tabas). Many of the biochemical properties of
Apo C-III are similar to those of the SAP proteins (e.g., molecular weight
of about 9.5 kDa; pI in the range of 4.4 to 4.6), and it has been suggested
that the mechanism of ASM activation by the "SAP-like domain" and Apo C-III
may be similar.
Polyclonal108,118 and monoclonal119 antibodies have been raised against
various purified ASM preparations and used to study the biosynthesis of ASM
in normal and NPD fibroblasts. After labeling the cells for 16 h with
[35S]methionine, Jobb et al. found that a single polypeptide of
about 110 kDa could be immunoprecipitated from cells derived from normal
individuals and that the 110-kDa polypeptide was partially processed into
an 84-kDa species by 72 h. ASM also was secreted by I-cell disease
fibroblasts, suggesting that the enzyme is trafficked to lysosomes by the
mannose 6-phosphate receptor-mediated pathway. Recent data using
recombinant ASM has suggested that the mannose 6-phosphate uptake system
accounts for ~50% of the enzyme uptake by cultured skin fibroblasts (He et
al. 1999) (see "Overexpression of Human ASM in Chinese Hamster Ovary
Cells"). More recent processing studies using cultured skin
fibroblasts also have shown that ASM is synthesized as a 75 kDa precursor
that is processed into several smaller molecular weight species (e.g., 72,
67 and 57 kDa) within lysosomes, and that small amounts of the precursor
form were secreted into the culture media, along with the 57 kDa form.
[top] [topics]
Other Mammalian Sphingomyelinases
In addition to the intracellular ASM, at least three other
sphingomyelinases have been described in humans. Neutral sphingomyelinase
is a membrane-bound enzyme with a pH optimum of about 7.5. This enzyme,
which
requires Mg2+ for activity, is found in most mammalian tissues and is the
predominant sphingomyelinase in the brain. Neutral sphingomyelinase levels
are normal in Types A and B NPD patients, suggesting that this polypeptide
is the product of a distinct gene. Recently, a cDNA clone has been
obtained that encodes an enzyme with many of the properties attributed to
the Mg2+-dependent neutral sphingomyelinase (Tomuk et al. 1998). A
Mg2+-independent, neutral sphingomyelinase activity also has been
reported, as well as a Zn2+-dependent acidic sphingomyelinase in human and
bovine serum. This latter enzyme is encoded by the same gene as the
intracellular ASM, and similarly, is deficient in Types A and B NPD
patients (see "Relationship Between Intracellular and
Secreted ASM").
In addition to these enzymes, another acidic human sphingomyelinase has
been purified to apparent homogeneity from placenta. Despite the fact that
the molecular weight of this purified enzyme was consistent with that of
ASM purified from urine, none of the amino acid sequences obtained from the
microsequenced proteolytic fragments were the same. Thus, this
sphingomyelinase is a unique protein, distinct from ASM. The existence of
another phospholipase that can cleave sphingomyelin at an acidic pH may
explain some of the difficulties in developing accurate heterozygote
testing for Types A and B NPD by enzymatic assay; however, the existence of
this enzyme still awaits confirmation.
[top] [topics]
Relationship Between Intracellular and Secreted ASM
For many years it has been known that Types A and B NPD result from the
deficient activity of lysosomal ASM. This enzyme has been studied in
detail, and many of its basic biochemical properties have been investigated.
As noted above, it had been generally assumed that in contrast to the Mg2+-dependent netural sphingomyelinase, ASM activity was cation-independent.
This notion was based primarily on the fact that treatment of the enzyme
with metal chelators such as EDTA did not affect enzymatic activity. In
1989, an acidic sphinigomyelinase activity was reported in human and bovine
serum that had the unique property of being Zn2+-dependent, in contrast to
intracellular ASM. However, beyond this initial intriguing report, this
zinc-dependent activity was not studied further until 1996, when Schissel
et al. (Schissel et al. 1998) reported that many different cell lines
secreted a Zn2+-dependent ASM with properties identical to those described
for the Zn2+-dependent enzyme in serum. Quite surprisingly, cells from
patients with Types A or B NPD were deficient in this enzymatic activity,
as well as intracellular ASM. Moreover, CHO cells engineered to
overexpress the human ASM cDNA overexpressed both the Zn2+-dependent, secreted form of ASM and the intracellular form that did
not require zinc for catalytic activity.
Thus, the same primary ASM mRNA leads to the production of two distinct
forms of ASM, one secreted from cells and requiring zinc for activity, and
the other intracellular (presumably lysosomal) and not
requiring zinc for activation. While the biological function of the
secreted ASM remains unknown, the implications for NPD are profound since
mutations in the ASM gene lead to the deficiency of both enzymatic
activities. Thus, it is possible that some aspects of the NPD pathology
may, in fact, not be a result of intracellular accumulation of sphingomyelin,
but rather related to the function of the secreted enzyme. Further
clarification of the unique biological roles of the two ASM forms and their
implications for NPD will be an obvious area of future research. Most
recently, it has been shown that the intracellular ASM, described as a
cation-independent enzyme, is in fact inactivated by the zinc specific
chelator, 1, 10- phenanthroline (Schissel et al. 1998). Thus, the main
difference between the secreted and intracellular ASM activities is not
their zinc requirements (both, in fact, require zinc cations for enzymatic
activity), but rather in their intracellular trafficking and relative
exposure to intracellular pools of zinc.
[top] [topics]
Overexpression of Human ASM in Chinese Hamster Ovary Cells
One of the important recent developments that will have a major impact on
the treatment of NPD is the overexpression of human ASM in CHO cells. As
noted above, one of the difficulties encountered in the study of ASM
previously was the small amounts of enzyme available from natural sources.
Moreover, the enzyme obtained from tissue homogenates was prone to
aggregation and difficult to purify. These limitations have now been
overcome by overexpressing the human ASM cDNA in CHO cells using a
DHFR-based amplification/expression vector system. Notably, ASM activity
in the media of the stably transfected cells was ~700-fold greater than in
the parental cells and required zinc cations for maximal activity. A rapid
purification protocol wash has been developed to isolate the recombinant
human enzyme from the culture media of these overexpressing cells,
and the yield of this procedure is 5
to 10 mg of homogenous enzyme/liter of culture media. The physicokinetic
properties of this enzyme have been extensively studied and appear
identical to those of the native enzyme obtained from non-recombinant human
tissues (He et al. 1999). Thus, there is now an excellent source
of purified human ASM available for clinical evaluation in the NPD
knock-out mouse model (see "Animal Models"), as well as for
crystallization and other structural studies.
[top] [topics]
|
