Expression of Mouse Acute-Phase (SAA1.1) and Constitutive (SAA4) Serum Amyloid A Isotypes
Influence on Lipoprotein Profiles
Abstract—The serum amyloid A (SAA) family of proteins consists of inducible acute-phase members and a constitutive member that are minor apolipoproteins of normal high density lipoprotein (HDL). During inflammation, HDL cholesterol and apolipoprotein A-I (apoA-I) protein are decreased, and these changes are thought to be partly related to the increase in acute-phase SAA proteins that associate with the HDL particle to become the major apolipoprotein species. To determine the specific role of SAA in the alteration of HDL in the absence of a generalized acute-phase response, acute-phase Saa1.1 transgene expression was directed via an inducible mouse metallothionein promoter. Elevated levels of SAA1.1 (28±9 mg/dL) comparable to a moderate acute-phase response were achieved over a 5-day period. SAA association with the HDL particles at this concentration did not significantly alter the apoA-I or HDL cholesterol levels or change the lipoprotein profiles in the transgenic mice compared with wild-type mice. In addition, we used adenoviral vectors to increase the SAA expression to levels seen in a major acute-phase response. Injection of adenovirus expressing the mouse SAA1.1 protein resulted in high-level expression (72±8 mg/dL) but did not alter apoA-I levels. However, the SAA associated with the HDL particle gave rise to significantly larger HDL particles (≈10%). Adenoviral expression of the constitutive SAA4 protein resulted in an increase in HDL size (≈10%) and an increase in very low density lipoprotein levels (20-fold) and triglyceride levels (1.7-fold). These studies suggest that increases in acute-phase SAA proteins alone are insufficient to alter HDL cholesterol or apoA-I levels during inflammation. A role for constitutive SAA4 in HDL–very low density lipoprotein interactions should be considered.
- Received April 7, 1999.
- Accepted February 16, 2000.
Lipoprotein profiles are markedly altered during inflammatory processes.1 2 HDL cholesterol and apoA-I levels decrease considerably during an acute-phase reaction and are lower than normal in a chronic inflammatory state.3 The serum amyloid A (SAA) protein family of apolipoproteins has been implicated in the HDL decrease during inflammation.3 The mouse SAA family consists of 2 distinct subfamilies. The first subfamily constitutes the classic acute-phase SAA molecules (SAA1.1 and SAA2.1) that dramatically increase during inflammation and can even become the major apolipoprotein of HDL.4 5 Their association with normal HDL in vitro results in remodeling of the particle, yielding larger particles with a higher hydrated density and relatively less apoA-I.3 More recently, a second subfamily was discovered that is the major form of SAA in normal HDL (SAA4), where these molecules exist as minor apolipoproteins.6 7 Studies of the effect of the inflammatory subfamily of SAA on HDL have been confounded by the fact that the induction of these molecules also induces the totality of the acute-phase response. This response is associated with a variety of metabolic effects that could influence HDL levels in ways far beyond the effect mediated by SAA.2 It is of interest that Parks and Rudel8 reported that when an acute-phase response was induced in monkeys by chair restraint, the changes in the density distribution of HDL could not solely be explained by SAA induction. Data have indicated that there are other factors likely involved in altering the HDL particle.8
In the present study, we examined the influence of the acute-phase SAA proteins, major lipoproteins associated with HDL, and constitutive SAA4 on lipoprotein metabolism. The development of transgenic mice allowed for the overexpression of SAA in the absence of a generalized acute-phase response. A derivative of the mouse metallothionein (mMT) gene promoter that contains the 5′ and 3′ hypersensitive sites was used to drive SAA synthesis.9 This construct allows for reduction in the constitutive expression of the transgene and for more selective expression when the animals are exposed to zinc. We chose the Saa1.1 gene to generate transgenic animals because of its function as an acute-phase protein and amyloidogenic potential. In most mouse strains, SAA1.1 and SAA2.1 are the acute-phase SAA proteins, and during inflammation-associated amyloid formation, only SAA1.1 is selectively cleared into amyloid fibrils.10 In addition, we have developed adenoviral vectors that allow for high levels of SAA expression that are comparable to the levels produced after lipopolysaccharide (LPS) stimulation.11 12 For these studies, we used the acute-phase mouse Saa1.1 and constitutive Saa4 cDNAs for the construction of the viral vector to study the respective impact of the constitutive (SAA4) SAA subfamily compared with the acute-phase (SAA1.1) subfamily.
C57BL/6×C3H crosses were used in the generation of transgenic mice at the University of Kentucky Transgenic Mouse Facility. C57BL/6 mice were purchased from Jackson Laboratories, Bar Harbor, Me. Transgenic mice were backcrossed into C57BL/6 mice for 5 generations before use in these studies. Transgenic and nontransgenic mice were used, and these animals were siblings from transgenic×nontransgenic crosses. Animals were maintained on a Harlan Teklad rodent chow diet (No. 8664). Expression of the transgenic mouse Saa1.1 gene was achieved through feeding the mice 25 mmol/L ZnSO4 in their drinking water. Fresh drinking water was prepared daily. Adenoviruses were introduced into mice via tail vein injections. An acute-phase response was elicited by intraperitoneal injection of 50 μg LPS Escherichia coli (0111:B4, Difco Laboratories), and animals were euthanized after 24 hours. EDTA-anticoagulated blood was collected by cardiac puncture from Metofane (Schering Plough)-anesthetized animals in accordance with the Institutional Animal Care and Use Committee and Office for Protection from Research Risk guidelines.
The transgenic construct was prepared by isolation of the mouse Saa1.1 gene by using BglII and XbaI and ligation of the fragment into the NruI site of the MT-LCR plasmid.13 The MT-LCR plasmid contained the constitutive mMT promoter and the human growth hormone (hGH) 3′ untranslated region.9 In addition, the construct embodied the 5′ and 3′ mMT gene hypersensitive sites (mMT 5′ HS and mMT 3′ HS). The plasmid was linearized with SalI, and the Saa1.1-containing fragment was isolated via agarose gel electrophoresis with use of the Geneclean Kit (Bio 101, Inc). The fragment was injected into fertilized C57BL/6×C3H embryos, which were placed into pseudopregnant ICR females. Transgenic animals were identified initially by isolation of tail DNA as described previously and by Southern blot analysis performed on EcoRI-restricted DNA.14 The Saa1.1 transgene was detected by using a random primer–labeled BglII-XbaI restriction fragment. Subsequent analysis was performed by using polymerase chain reaction (PCR) primers against the Saa1.1 gene (5′-GGAAGCTTGGATGAA-GCTACTCACCAGCCTG-3′) and the hGH 3′ untranslated sequences (5′-GATGCAACTTAATTTTATTAGGACAAGGCTGG-3′), resulting in a 2-kbp fragment. As a positive control, the mouse heat-shock protein 70 was amplified (primer 1, 5′GCTAAGTGATC-ACCTGGATGCC-3′; primer 2, 5′-TCCCCTGGACTACCCA-ACTGTT-3′), resulting in a 476-bp fragment. PCR reactions were carried out as follows: 1× Perkin-Elmer PCR buffer, 0.3 mmol/L dNTPs, 0.75 mmol/L MgCl2, 2 ng each primer, and 1 U Taq polymerase. The reaction was carried out in a Perkin-Elmer 9600 Thermocycler at 95°C for 1 minute and then at 95°C (ramp 1 s, hold 15 s), 65°C (ramp 30 s, hold 15 s), and 72°C (ramp 1 s, hold 210 s) for 35 cycles, followed by 7 minutes at 72°C. PCR products were electrophoresed onto a 1% agarose gel in the presence of 0.1 μg/mL ethidium bromide for identification.
Temperature-sensitive adenoviral vectors were prepared by using a cytomegalovirus promoter and simian virus 40 poly(A) site as described previously.15 16 The vector contained 0 to 1 and 9 to 16 map units of the DNA sequence from adenovirus type 5. The Saa cDNAs (Saa1.1, Saa2.1, and Saa4) were PCR-amplified by using specific primers to amplify the coding regions and to insert unique cloning sites for ligation into the adenovirus transfer vector.17 18 19 Null adenoviruses, lacking the inserted sequence, were prepared as control viruses.
RNA Isolation and Analysis
Nontransgenic and transgenic mice were reared on water or 25 mmol/L ZnSO4 for 5 days and euthanized. Adenovirus-expressing animals were injected with virus, and after 3 days, livers were collected for RNA analysis. To compare the induction of the transgene, nontransgenic and transgenic mice were subjected to intraperitoneal injections of 50 μg LPS for 24 hours. Tissues were harvested and frozen in liquid nitrogen and stored at −80°C until processing. RNA was isolated from tissues as described previously, and 20 μg RNA was subjected to Northern blot analysis.20 Northern blots were probed with the Saa1.1 gene BamHI fragment radiolabeled by the random primer method.13 20 Resulting blots were subjected to autoradiography (same exposure time) and densitometric scanning by using a Personal Densitometer (Molecular Dynamics).
HDL was isolated from plasma as described previously.21 Briefly, plasma density was adjusted to 1.09 g/mL with solid KBr and then centrifuged for 5.3 hours at 242 000g in a VTi90 rotor (Beckman Instruments) at 10°C. The infranatants containing HDL were collected, and the density was readjusted to 1.21 g/mL and centrifuged for 9.4 hours at 242 000g in the same rotor. HDL was collected from the top of the tube and extensively dialyzed against 150 mmol/L NaCl and 0.1% EDTA (pH 7.4). Alternatively, isolation of HDL from fast protein liquid chromatography (FPLC) fractions was achieved by pooling the peak fractions, adjusting the density to 1.21 g/mL, and centrifuging as above. The HDL collected was dialyzed against 15 mmol/L NaCl and 0.01% EDTA, pH 7.4. Samples were freeze-dried and resuspended in SDS sample buffer. Plasma samples (370 μL) were chromatographed by FPLC on 2 Superose 6 columns (Pharmacia LKB Biotechnology Inc) linked in series.22 23 Plasma from groups of 3 mice were pooled and centrifuged twice for 10 minutes at 10 000 rpm, first at room temperature and then at 4°C, before being applied to the column at a flow rate of 0.5 mL/min. The elution buffer was 150 mmol/L NaCl and 10 mmol/L Tris/HCl (pH 7.4), and 0.5 mL fractions were collected.
For electrofocusing, HDL samples (20 μg) were delipidated with 0.5 mL chloroform/methanol (2:1 [vol/vol]), and the delipidated apolipoproteins were suspended in 1% (wt/vol) SDS (Eastman Kodak Co), 7 mol/L urea, and 5% (vol/vol) 2-mercaptoethanol.24 Samples were electrofocused on ultrathin acrylamide gels containing 20% (vol/vol) ampholines at pH 3 to 10, 40% (vol/vol) ampholines at pH 4 to 6.5, and 40% (vol/vol) ampholines at pH 7 to 9.24
Apolipoproteins were quantified by electrophoresis of samples (plasma or HDL) on SDS gels and Western blot analysis. Plasma samples (5 μL) were separated on reducing SDS-PAGE (5% to 20% acrylamide) and electroblotted onto 0.45-μmol/L pore–sized nitrocellulose membranes (Schleicher and Schuell), whereas samples on electrofocused gels were pressure-blotted overnight at room temperature.24 After electroblotting or pressure blotting, the membranes were soaked in 25 mmol/L Tris/HCl (pH 8.3), 192 mmol/L glycine, and 15% (vol/vol) methanol and blocked for 16 hours at 4°C in 5% (wt/vol) nonfat dry milk in PBS containing 2% BSA. Proteins of interest were identified with appropriate antibodies and visualized by chemiluminescence (Amersham). Films were analyzed by densitometric scanning (Molecular Dynamics). A standard curve was obtained by simultaneous electrophoresis of purified apolipoproteins. Apolipoproteins were identified with 1 of the following antibodies: rabbit anti-mouse SAA (polyclonal antibody prepared against purified mouse acute-phase SAA), rabbit anti-mouse SAA4 (polyclonal antibody prepared against SAA4 peptide insert), or rabbit anti-mouse apoA-I (polyclonal, Calbiochem).
Lipoproteins were separated according to size by nondenaturing gradient gel electrophoresis (GGE) by use of 5% to 20% acrylamide gels for HDL, as described.24 Purified HDL from transgenic and nontransgenic mice was run in conjunction with molecular weight markers and visualized by Coomassie staining.
Lipid analyses were performed on FPLC fractions or plasma samples from animals that had been fasted for ≈7 hours. Total and free cholesterol levels as well as triglycerides and free fatty acids were determined enzymatically (WAKO Chemicals and Sigma Chemical Co). HDL cholesterol was measured enzymatically after precipitation of LDL and VLDL by heparin and manganese (WAKO Chemicals).
All data were presented as the mean±SEM. The statistical significance of the difference between the groups was estimated by ANOVA. A value of P<0.05 was considered significant.
Generation of mMT-SAA1.1 Transgenic Mice
To study the effects of SAA expression on HDL level, size, and composition, the Saa1.1 gene driven by the mMT promoter sequences was introduced into mice. Microinjection of the 22-kbp MT-Saa1.1 transgene into (B6×C3H)F2 fertilized eggs resulted in the generation of 5 (2 sets of injections) live founder mice, which survived to weaning. Southern blot analysis of tail DNA digested with specific restriction enzymes revealed that all of the founders carried intact copies of the transgene (data not shown). The number of transgene copies per cell for the various lines was estimated to range from ≈1 to 4.
The founder mice appeared to develop normally, and all lines were bred successfully. All lines were bred until a sufficient number of animals were available to determine the expression pattern of the Saa1.1 transgene. Of the founders, line 5 demonstrated the optimal conditions for testing the effects of SAA on HDL changes. Line 5 had undetectable levels of Saa1.1 RNA expression, and on treatment with ZnSO4, Saa1.1 expression increased significantly (Figure 1A⇓).
Inducible mRNA Expression of MT-Saa1.1 in Transgenic Mice
To examine the expression pattern of the Saa1.1 transgene, RNA was prepared from various tissues and investigated by Northern blot hybridization by using the Saa1.1 cDNA probe. The results obtained from founder mouse line 5 are shown in Figure 1A⇑ (see Figure⇑ I and Table⇓ I published online at http://atvb.ahajournals.org). We examined the expression of the transgene in the liver, because that is where Saa1.1 is expressed during the acute phase. Figure 1A⇑ illustrates the expression of Saa in nontransgenic mice; control and ZnSO4-treated mice had no detectable levels of Saa, whereas LPS-treated mice had high levels of Saa mRNA. In the transgenic mice under control conditions, no Saa was expressed, but with ZnSO4, there was a moderate induction of the Saa1.1 transgene. Injection of the transgenic mice with LPS resulted in a large increase in Saa mRNA, which was even higher than that detected in the nontransgenic mice. This was probably due to the inducibility of the mMT promoter by LPS.9 If the induction of Saa in LPS-treated nontransgenic mice is considered to be 100±15% (n=6), the ZnSO4-treated transgenic mice had a 40±20% (n=6) induction of Saa1.1, whereas LPS treatment of transgenic mice resulted in a 126±17% induction of the Saa gene.
Next, we examined the overall expression pattern of the Saa1.1 transgene and endogenous Saa genes in control and ZnSO4-treated mice (data not shown; see Table⇑ I, published online at http://atvb.ahajournals.org). In nontransgenic mice in the absence or presence of ZnSO4 there was no detectable Saa mRNA in any tissue. LPS injection of nontransgenic mice showed the typical induction of Saa mRNA in the liver, with a small induction in the heart, lung, kidney, and spleen. In the transgenic mice under uninduced conditions, Saa1.1 mRNA was detectable only in the brain. On induction with ZnSO4, there was a significant increase in Saa1.1 mRNA in the liver, heart, and intestine and, to a much lesser degree, in the lung. The brain expression of transgenic Saa1.1 appeared to be constitutive and maintained at low levels.
Protein Expression in Transgenic Mice
Isoelectric focusing (IEF) and Western blot analysis of purified HDL showed the specificity of SAA1.1 induction (Figure 1B⇑). Western blot of ZnSO4-treated transgenic mice showed the induction of SAA1.1 (Figure 1B⇑, lane 1). Compared with the level in acute-phase nontransgenic mice (lane 3), the level of SAA1.1 protein in the transgenic mice is analogous to a modest acute-phase response (28±9 mg/dL, Table 1⇑) and is composed of only a single isotype, namely, SAA1.1. The levels of SAA1.1 in the transgenic mice varied from 20 to 37 mg/dL. Shorter or longer treatment of mice with ZnSO4 did not show increased levels of SAA1.1 expression (data not shown). On the basis of quantitative Western blot analysis from FPLC fractions, >95% of the SAA1.1 protein was associated with the HDL fraction (96±7%, data not shown).
Adenoviral Expression of Mouse SAA proteins
To study the effects of SAA proteins on HDL, we needed to express SAA at levels comparable to an acute-phase response. High-level expression of mouse SAA1.1 and SAA4 proteins was achieved by adenovirus-mediated expression in C57BL/6 mice. The Saa1.1 and Saa4 cDNAs were cloned into a second-generation (temperature-sensitive) adenoviral vector and injected via the tail vein into mice to examine the expression of SAA proteins. Approximately 95% of the virus infected the liver of these mice and allowed for tissue-specific expression (data not shown). Figure 2A⇓ shows the expression of the mouse SAA1.1 on HDL 3 days after injection of 1×1011 particles of adenovirus SAA1.1 (lanes 1 to 5). SAA levels peaked at 24 to 48 hours after injection and remained constant throughout the experimental protocol. The level of SAA1.1 reached 72±8 mg/dL (on the basis of quantitative Western blots, Table 1⇑), which was comparable to the level observed in mice when an acute-phase response was induced by LPS (lane 6). In addition, injection of SAA1.1 and SAA2.1 viruses resulted in a significant increase in SAA protein (86±10 mg/dL, Figure 2B⇓). Control adenovirus–injected mice (null) were used to demonstrate the specific changes in SAA expression (lane 7). SAA4 adenoviral protein expression showed a pattern of expression similar to SAA1.1 protein expression (ie, HDL-associated), and the levels of SAA4 reached ≈92±8 mg/dL (Figure 2C⇓).
Effects of SAA on Lipoproteins
To understand the effects of SAA on changes in the HDL particle, plasma and HDL samples were subject to Western blot, lipid, and FPLC analysis. Table 1⇑ shows the results of the quantitative Western analysis. As described above, the SAA1.1 transgenic mice treated with zinc had levels of the SAA1.1 protein at 28±9 mg/dL compared with levels between 0.5 and 0.8 mg/dL in nontransgenic (with or without zinc) or transgenic mice treated with water, respectively. In addition, injection of mice with LPS resulted in a dramatic increase in SAA in the plasma (108±8 mg/dL). With the SAA1.1 adenoviral vector, the level of plasma SAA1.1 (72±8 mg/dL) was comparable to an acute-phase response. Measurement of apoA-I levels in transgenic mice showed that levels did not change during induction of the transgenic SAA1.1 protein (Table 1⇑). The apoA-I levels were essentially identical between the nontransgenic and transgenic mice in the absence or presence of ZnSO4. Animals injected with adenoviral vector SAA1.1 also showed no detectable change in apoA-I levels, whereas mice injected with LPS (50 μg IP), compared with nontransgenic controls, had a significantly lower concentration of apoA-I protein (76±14 mg/dL and 126±4 mg/dL, respectively).
The effects of SAA1.1 expression on plasma lipids in the C57BL/6 mice are shown in Table 2⇓. The levels of HDL cholesterol were not altered in either the transgenic mice or the mice injected with the adenoviral vectors expressing the SAA1.1 protein. A similar result was observed in the plasma cholesterol levels in transgenic and adenovirus-injected mice. Elevated SAA1.1 levels did not significantly alter plasma triglycerides or free fatty acids.
To further analyze the lipoproteins in the transgenic and adenovirus-injected mice, plasma was collected from control and transgenic mice, control mice on zinc, transgenic mice on zinc (5 days after induction), and adenovirus-injected mice (3 days after injection) and separated by gel filtration on Superose 6B columns (FPLC). The FPLC profiles for the nontransgenic and transgenic mice were essentially identical for the different mice and different treatments (data not shown). Lipoproteins from SAA1.1 transgenic and nontransgenic animals were analyzed by GGE to compare the HDL particle sizes (see Figure⇑ II published online at http://atvb. ahajournals.org). HDL purified by ultracentrifugation did not show any size difference between nontransgenic and transgenic mice in the absence or presence of zinc (all HDL particles were ≈5.3 to 5.4 nm). These results were consistent with the FPLC data. Only in mice injected with LPS (50 μg) was there a difference in the size of the HDL particle (≈6.0 nm). These particles showed a significant increase in HDL size.
The SAA1.1 protein showed a modest change in FPLC profile (Figure 3B⇓) but no change in plasma lipids (Table 2⇑) in the adenovirus-injected animals compared with the null-injected animals (Figure 3A⇓). In addition, there was a significant increase in the size of the SAA1.1 HDL particle (10%), as determined by GGE analysis (Figure 4⇓, lane 1 compared with controls [lanes 3 and 4]). In addition to the SAA1.1 protein expression, the SAA4 virus showed an unexpected change in plasma lipoprotein size (10%, Figure 3C⇓ and Figure 4⇓, lane 2) and composition (Table 2⇑). The SAA4 adenovirus–injected animals showed a 1.7-fold increase in plasma triglyceride levels (Table 2⇑) compared with all other animals. This increase in triglyceride levels was a result of an increase in VLDL levels (20-fold) in the SAA4 adenovirus–injected animals, as indicated in the FPLC profiles (in Figure 3⇓, compare panel C with panel A). The shift in the size of the HDL particle was comparable to the size change seen in LPS-treated animals (Figure 3D⇓ and Figure 4⇓, lane 5). Lower doses of SAA4 virus (0.25×1011 particles per mouse) resulted in reduced levels of SAA4 protein (23±6 mg/dL) and still demonstrated a significant increase in triglyceride (74±4 mg/dL, n=10; P=0.05 compared with null injection) and VLDL levels as well as a small shift in the HDL profile (data not shown).
Recently, we had shown that human SAA1 expression in the apo A-I transgenic mice did not have an effect on HDL composition.25 To determine whether the expression of SAA2.1 and SAA1.1 would affect HDL cholesterol and apoA-I levels, we injected mice with the SAA2.1 and SAA1.1 adenoviral vectors (1×1011 particles [total] per mouse, Figure 2B⇑). Although elevated expression of SAA2.1 and SAA1.1 protein was detected and was comparable to an acute-phase response, no changes in HDL cholesterol or apoA-I were encountered (Table 2⇑).
SAA proteins are the most inducible acute-phase reactants in the mouse. Two of these isotypes, SAA2.1 and SAA1.1, are induced in approximately equal quantities after an inflammatory response. These 2 isotypes differ in only 9 of 103 amino acids. SAA1.1 seems to be metabolically the most dynamic, because it is more rapidly cleared than SAA2.1 and, consequently, is deposited in amyloid fibrils.10 The constitutive SAA, SAA4, is found in a subset of HDL and presumably plays a role in the normal functioning of the HDL particle.26 Thus, we chose to overexpress the SAA1.1 and SAA4 isoforms for analysis. Previous studies of the effect of SAA on HDL metabolism have been confounded by the fact that the induction of an acute-phase response could influence metabolism in a variety of different ways.2 Because the acute-phase response was associated with a high level of expression of SAA and because of the in vitro data indicating that SAA can displace apoA-I from HDL, it was proposed that the reductions in HDL cholesterol and apoA-I protein were a direct result of SAA expression.3 To determine the significance of SAA expression on altered HDL cholesterol and apoA-I levels in vivo, we generated mice transgenic for the mouse Saa1.1 gene and adenoviral vectors that express the SAA1.1 and SAA4 proteins.
In these experiments, Saa1.1 was expressed mainly in the liver of transgenic mice obtained by using a transgenic construct and by liver direct gene transfer of adenoviral vectors.12 16 Transgene expression was accomplished by use of the mMT promoter, which contains the 5′ and 3′ hypersensitive sites, which allows for zinc-inducible expression of the inserted gene.9 Induction of the transgene was attained by the addition of ZnSO4 to the drinking water of the mice for 5 days before experimentation. The SAA1.1 transgenic protein was induced at a moderate level compared with LPS induction (28 mg/dL versus ≈100 mg/dL, respectively). Variability in SAA1.1 expression could be due to the amount of water consumed by the transgenic animals and the time of experimentation after water intake. To obtain a consistent higher level of expression, we used adenoviral vector–mediated gene transfer of the Saa1.1 gene. Adenoviral vectors were delivered via tail vein injections and resulted in constitutive high levels of SAA1.1 expression in mice (>70 mg/dL). Our results also showed that the expression of the inserted sequence coding for the SAA protein was preferentially associated with mouse HDL particles (see Figure⇑ III, published online at http://atvb.ahajournals.org).
The major finding of the present study was that the moderate to high level of expression of mouse SAA1.1 over a period of time did not result in any significant changes in HDL cholesterol or decrease in apoA-I concentrations. Data provided evidence that these effects in the acute-phase response were due to factors in addition to SAA induction. These results are consistent with a recent study reporting that expression of SAA2.2 (known also as the CE/J isoform and similar to SAA2.1; it is nonamyloidogenic) did not alter HDL cholesterol or apoA-I levels.25 In addition, in the present study, expression of SAA2.1 and SAA1.1 in the same animal at levels comparable to levels achieved with LPS induction did not alter HDL composition. These results appear to be in conflict with previous in vitro data indicating that when increasing amounts of purified SAA were added to human HDL in vitro, increasing amounts of apoA-I were displaced from the particles.3 The limitation of this experimental setting is now more obvious. The HDL used in the in vitro experiments was restricted in obtaining additional lipid from sources such as other cells or other lipoprotein particles, as may occur in vivo. The present study places the HDL in an environment in which it can interact with other lipoproteins that can serve as a source of lipid supply. The possibility exists that HDL may be able to accommodate the addition of SAA by acquiring additional lipid and thus preventing the displacement of apoA-I. Our transgenic mice did not express SAA at levels that mimic the most severe of acute-phase responses, such as that the response induced by LPS injection.27 However, the adenoviral vectors increased SAA levels to levels comparable to those found with a major acute-phase response and still did not significantly alter HDL. We cannot exclude the possibility that induction of SAA at ≈100 mg/dL will not result in apoA-I displacement and that this could lead to decreased apoA-I levels. However, our SAA induction was at least 20- to 50-fold higher than that induced by a high fat diet; thus, one can state with some confidence that the decreases in HDL seen when mice are fed an atherogenic diet are unlikely mediated by SAA.28
Previous studies using SAA adenoviral vectors demonstrated a lack of change in HDL cholesterol and apoA-I levels in mice.25 However, no increase in the size of the HDL was detected when either the mouse or human SAA viruses were used. The levels of SAA were comparable to levels in our transgenic animals, which did not show changes in HDL size. These studies used first-generation viruses, which were limited in time and level of expression compared with the second-generation viruses. The present results showed higher levels of SAA expression and increased size of SAA-HDL. The increased size of HDL was most likely due to the higher levels of SAA present on the HDL particle and changes in lipid to accommodate the SAA. Even though adenoviral SAA levels were comparable to those after LPS injection, no change in HDL cholesterol or apoA-I was detected.
Recently, it has been shown that the group II nonpancreatic phospholipase (sPLA2) enzymes are classic acute-phase reactants and are elevated during inflammatory conditions.29 In addition, sPLA2 can act on lipoprotein particles to alter their structure and composition by hydrolyzing the phospholipids at the sn-2 position.30 Transgenic mice that express the human sPLA2 enzyme demonstrated reduced levels of HDL particles, and these particles had reduced levels of phospholipid and were also significantly reduced in size.31 SAA has been shown to be a cofactor for sPLA2, even when associated with the HDL particle.32 The interpretation of these results is that the modified nature of the acute-phase HDL particle will lead to altered metabolism of not only the particle but also the turnover of the lipoproteins associated with HDL. Even though the C57BL/6 mice used in the present study lack the sPLA2 enzyme because of a mutation in the gene, hepatic lipase may be able to substitute for sPLA2 in these animals.33 34 These studies indicate that SAA alone is insufficient to affect HDL cholesterol and apoA-I levels but may participate in a more global effect to modify the HDL particle.
The effects of SAA4 on lipoprotein profiles were more profound than those detected for SAA1.1. The increase in triglyceride levels, specifically on VLDL, raises some interesting questions. These alterations may be the result of viral effects in the in vivo situation. We have seen similar results with higher doses of viruses; however, at the concentrations used in the present study, this is not likely the case. In addition, the SAA1.1 and null control viruses did not show the same effects on triglycerides as seen with the SAA4 virus. The level of SAA4 found in the adenovirus-injected mice vastly exceeds the amount of protein detected in normal or acute-phase mice, so the changes in triglyceride and VLDL levels are somewhat exaggerated.6 Using less virus gave rise to lower levels of SAA4, yet triglyceride and VLDL levels were still elevated. This suggests that SAA4 at physiological levels may carry out the function demonstrated in the present study. These data could suggest the possible interactions between HDL and VLDL that may be mediated by SAA4. Because SAA4 is found predominantly on a subset of HDL particles, this suggests that the SAA4 may be involved in HDL-VLDL lipid transport.
It is possible that SAA could alter the function of the HDL particle beyond merely influencing of the concentration of apoA-I levels. Increased levels of SAA may alter the metabolism and influence the clearance of HDL particles by impacting HDL–scavenger receptor BI interactions.35 The transgenic mice and adenoviral vectors that we have developed may potentially aid in the investigation of the role of SAA in vascular lipid and amyloid deposition.
This work was supported by National Institutes of Health grants NS-32221 and AG-12981 (M.S.K.) and AG-10886 (F.C.d.D.) and the Sanders-Brown Center on Aging. We would like to thank Amy R King, Darin Cecil, and John Cranfill for their technical assistance on RNA preparations and Connie Gerardot for her generation of the transgenic construct.
Hoffman J, Benditt EP. Changes in high-density lipoprotein content following endotoxin administration in the mouse: formation of serum amyloid protein-rich subfractions. J Biol Chem. 1982;257:10510–10517.
Coetzee GA, Strachan AF, van der Westhuyzen DR, Hoppe H, Jeenah MS, de Beer FC. Serum amyloid A-containing human high-density lipoprotein 3: density, size and apolipoprotein composition. J Biol Chem. 1986;261:9644–9651.
Strachan AF, de Beer FC, Coetzee GA, Hoppe HC, Jeenah MS, van der Westhuyzen DR. Characteristics of apo-SAA-containing HDL3 in humans. In: H. Peeters, ed. Protides of the Biological Fluids. Brussels, Belgium: Pergamon Press; 1986;34:359–362.
Small DM. Determination of HDL structure. In: Catapano AL, Bernini F, Cossini A, eds. High-Density Lipoproteins: Physiopathology and Clinical Relevance. New York, NY: Raven Press; 1993;1:30–50.
De Beer MC, Kindy MS, Lane WS, de Beer FC. Mouse serum amyloid A protein (SAA5): structure and expression. J Biol Chem. 1994;269:4661–4667.
Parks J, Rudel L. Alteration of high density lipoprotein subfraction distribution with induction of serum amyloid A protein (SAA) in the nonhuman primate. J Lipid Res. 1985;26:82–91.
Palmiter RD, Sandgren EP, Koeller DM, Brinster RL. Distal regulatory elements from the mouse metallothionein locus stimulate gene expression in transgenic mice. Mol Cell Biol. 1993;13:5266–5278.
Webb NR, de Beer MC, van der Westhuyzen DR, Kindy MS, Banka CL, Tsukamoto K, Rader DL, de Beer FC. Adenoviral vector-mediated overexpression of serum amyloid A in apoA-I-deficient mice. J Lipid Res. 1997;38:1583–1590.
Kindy MS, King AR, Yu J, Gerardot C, Whitley J, de Beer FC. Adenoviral expression of murine serum amyloid A proteins to study amyloid fibrillogenesis. Biochem J. 1998;332:721–728.
Yamamoto KI, Goto N, Kosaka J, Shiroo M, Yeul YD, Migita S. Structural diversity of murine serum amyloid A genes: evolutionary implications. J Immunol. 1987;139:309–315.
Hogan B, Consantini F, Lacy E. Manipulating the Mouse Embryo: A Laboratory Manual. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory; 1986.
Kozarsky KF, McKinley DR, Austin LL, Raper SE, Stratford-Perricaudet LD, Wilson JM. In vivo correction of LDL receptor deficiency in the Watanabe heritable hyperlipidemic rabbit with recombinant adenoviruses. J Biol Chem. 1994;269:13695–13701.
De Beer MC, de Beer FC, McCubbin WD, Kay CM, Kindy MS. Structural prerequisites for serum amyloid A fibril formation. J Biol Chem. 1993;268:20606–20612.
Yamamoto KI, Shiroo M, Migita S. Diverse gene expression for isotypes of murine serum amyloid protein during acute-phase reaction. Science. 1986;232:227–231.
De Beer MC, de Beer FC, Beach CM, Carreras I, Sipe JD. Mouse serum amyloid A protein: complete amino acid sequence and mRNA analysis of a new isoform. Biochem J. 1992;283:673–678.
Hedrich CC, Castellani LW, Warden CH, Puppione DL, Lusis AJ. Influence of mouse apolipoprotein A-II on plasma lipoproteins in transgenic mice. J Biol Chem. 1993;268:20676–20682.
Liu MS, Jirik FR, LeBoeuf RC, Henderson H, Castellani LW, Lusis AJ, Ma Y, Forsythe IJ, Zhang H, Kirk E, Brunzell JD, Hayden M. Alteration in lipid profiles in plasma of transgenic mice expressing human lipoprotein lipase. J Biol Chem. 1994;269:11417–11424.
Strachan AF, de Beer FC, van der Westhuyzen DR, Coetzee GA. Identification of three isoform patterns of human serum amyloid A protein. Biochem J. 1988;250:203–207.
Hosoai H, Webb NR, Glick JM, Tietge UJF, Purdom MS, de Beer FC, Rader DJ. Expression of serum amyloid A protein in the absence of the acute-phase response does not reduce HDL cholesterol or apoA-I levels in human apoA-I transgenic mice. J Lipid Res. 1999;40:648–653.
De Beer MC, Yuan T, Kindy MS, Asztalos BF, Roheim PS, de Beer FC. Characterization of constitutive human serum amyloid A protein (SAA4) as an apolipoprotein. J Lipid Res. 1995;36:526–534.
Meek RL, Benditt EP. Amyloid A gene family expression in different mouse tissues. J Exp Med. 1986;164:2006–2017.
Liao F, Andalibi A, de Beer FC, Fogelman AM, Lusis AJ. Genetic control of inflammatory gene induction and NF-κB-like transcription. J Clin Invest. 1993;91:2572–2579.
Pruzanski W, Vadas P, Browning J. Secretory non-pancreatic group II phospholipase A2: role in physiologic and inflammatory processes. J Lipid Mediat Cell Signal. 1993;8:161–167.
Pruzanski W, de Beer FC, de Beer MC, Stefanski E, Vadas P. Serum amyloid A protein enhances the activity of secretory phospholipase A2. Biochem J. 1995;309:461–466.
De Beer FC, de Beer MC, van der Westhuysen DR, Castellani LW, Lusis AJ, Swanson ME, Grass DS. Secretory non-pancreatic phospholipase A2: influence on lipoprotein metabolism. J Lipid Res. 1997;38:2232–2239.
Pruzanski W, de Beer FC, de Beer MC, Stefanski E, Vadas P. Serum amyloid A protein enhances the activity of secretory non-pancreatic phospholipase A2. Biochem J. 1995;309:461–464.
Kennedy BP, Payette P, Mudgett J, Vadas P, Pruzanski W, Kwan M, Tang C, Rancourt DE, Cromlish WA. A natural disruption of the secretory group II phospholipase A2 gene in inbred mouse strains. J Biol Chem. 1995;270:22378–22385.
Bamberger M, Glick JM, Rothblat GH. Hepatic lipase stimulates the uptake of high-density lipoprotein cholesterol by hepatoma cells. J Lipid Res. 1983;24:869–876.
Acton S, Rigotti A, Landschulz KT, Xu S, Hobbs HH, Krieger M. Identification of scavenger receptor SR-BI as a high-density lipoprotein receptor. Science. 1996;271:518–520.