The development of premature atherosclerosis has been recognized since 1976 as a clinical problem in patients with systemic lupus erythematosus (SLE) (1). Numerous subsequent studies have confirmed these initial observations, with reported frequencies of coronary artery disease (CAD) ranging from 2% to 19.8% in clinical prospective studies and from 10% to 54% in autopsy studies (2). Questions raised by these observations are many and focus on the identification of risk factors in this SLE population. As in other populations, traditional risk factors that are predictive of CAD in patients with SLE include age, male sex, elevated homocysteine levels, obesity, hypertension, diabetes mellitus, and creatinine levels (2). However, as reported by Esdaile et al, even when these traditional risk factors were taken into account in a group of SLE patients, the risk of myocardial infarction was still increased 8.3-fold because of the disease process itself (3). In addition, although chronic steroid intake has been blamed for an increased risk of CAD, Petri concluded, after careful study and use of longitudinal regression analysis, that the effects of steroids are indirect and affect the magnitude of 3 traditional risk factors: elevated cholesterol level, hypertension, and obesity (2).
In studies of SLE, a specific pattern of dyslipidemia has been noted in both pediatric and adult cases (4, 5). This pattern, which is best appreciated in patients with active disease who are untreated, consists of elevated levels of very low-density lipoproteins (VLDL) and triglycerides, and decreased levels of high-density lipoprotein (HDL) cholesterol and apolipoprotein A1 (Apo A1). As stated above, this pattern was noted in both untreated pediatric (4) and adult (5) cases of active SLE. In 2 case reports, extremely high triglyceride levels were associated with decreased lipoprotein lipase (LPL) activity in the plasma (6, 7). In 1 case, the patient's triglyceride level decreased from >2,100 mg/dl to 95 mg/dl with steroid treatment (6), suggesting a role for autoimmunity in patients with very high triglyceride levels. Other investigators have reported higher total cholesterol (201 mg/dl) and higher serum triglyceride levels (134 mg/dl) in SLE patients compared with those in matched controls (168 mg/dl and 74 mg/dl, respectively) (8).
These data stimulated further analysis by Borba et al, who evaluated clearance of artificial chylomicrons labeled with tritiated triglyceride (3H-TG) and cholesteryl ester (14C-CE) in 7 untreated female patients with active SLE and 7 age-matched female controls (9). The SLE group demonstrated a marked decrease of the 3H-TG fractional catabolic rate (FCR) (mean ± SD 0.024 ± 0.014 versus 0.049 ± 0.010 minute−1; P = 0.04) and of 14C-CE FCR (0.005 ± 0.008 versus 0.022 ± 0.008 minute−1; P = 0.007) compared with the control group. In addition, a striking reduction of LPL activity, measured as area under the curve of the remaining triglycerides determined after thin-layer chromatography, was observed in the SLE patients. These results suggested that both lipolytic catabolism of chylomicrons and removal of chylomicron remnants by the liver are impaired in SLE. Such inhibition of lipolysis could be one of the factors contributing to the elevated triglyceride and VLDL levels observed in untreated patients with active SLE. This inhibition would presumably implicate decreased activity of the enzyme LPL as a major factor responsible for the catabolism of chylomicrons.
These data stimulated our efforts to determine whether immunologic responses directed to LPL are present in SLE patients, and if there is any relationship of autoantibody to LPL and triglyceride levels in patients with SLE.
- Top of page
- PATIENTS AND METHODS
Sera obtained from groups of patients with SLE, rheumatoid arthritis (RA), polymyositis and dermatomyositis, Sjögren's syndrome, and from normal, unrelated control subjects were assessed for their ability to bind LPL-coated plates. The results are listed in Table 2. As shown, antibodies to LPL are not disease-specific, occurring as frequently in patients with polymyositis and progressive systemic sclerosis as in those with SLE. Although there appear to be many positive results in the normal population, details of the assay must be explored to elucidate this finding. For each 20 disease sera assayed, a panel of 4 normal human sera were run on each plate. Positive results were defined as those with a value exceeding the mean ± 2 SD of the OD units from the panel of 4. The group of unrelated control subjects (n = 75) were matched with patients for age, race, and sex. To illustrate the quantitative differences in anti-LPL activity between the disease groups and controls, the percentages of sera with anti-LPL values >0.6 OD units are listed in Table 3.
Table 2. Prevalence of anti-LPL in disease sera*
| ||No. tested||No. (%) positive|
|Systemic lupus erythematosus||105||49 (46.7)|
|Progressive systemic sclerosis||31||13 (41.9)|
|Rheumatoid arthritis||80||10 (12.5)|
|Sjögren's syndrome||30||3 (10.0)|
Table 3. No. (%) of sera with anti-LPL >0.60 OD units*
|Disease group||Anti-LPL >0.6 OD units|
|Systemic lupus erythematosus (n = 105)||20 (19.1)|
|Polymyositis (n = 31)||5 (16.1)|
|Progressive systemic sclerosis (n = 30)||5 (16.7)|
|Rheumatoid arthritis (n = 80)||6 (7.5)|
|Sjögren's syndrome (n = 30)||2 (6.6)|
|Normal (n = 75)||1 (1.3)|
As shown, the number of sera with anti-LPL levels >0.6 OD units is very low in the normal population and clearly distinguishes patients with SLE, polymyositis, and progressive systemic sclerosis from controls. Although the levels in patients with RA and Sjögren's syndrome are higher than those in controls, the difference is not statistically significant.
We used 2 approaches to address the issue of whether any part of the antibody activity with LPL resides in the anti-dsDNA antibodies, which are so characteristic of SLE—particularly active disease. First, we tested the ability of dsDNA to inhibit the reactivity of sera positive for anti-LPL activity. Table 4 lists 10 SLE sera reactive with LPL and the inhibition of a 1:100 dilution of serum by dsDNA (100 μg/ml and 10 μg/ml) in their reactivity with LPL-coated plates. From 35% to 70% of the LPL activity was inhibited by dsDNA at a final concentration of 100 μg/ml. This suggests that part of the anti-LPL binding activity resides in the anti-dsDNA population. Conversely, for sera that were anti-Sm–positive, anti-Ro–positive, or anti–U1 RNP–positive, attempts to inhibit reactivity with LPL by affinity-purified Sm, Ro, and U1 RNP, respectively, were unsuccessful, indicating no relationship of these antibody specificities to inhibition of LPL activity. However, normal sera with elevated values for anti-LPL were inhibited in their binding by dsDNA in a manner quite similar to that of SLE sera with anti-LPL activity.
Table 4. Inhibition of SLE sera reactivity with LPL by dsDNA*
|Patient||% inhibition by dsDNA|
|100 μg/ml†||10 μg/ml‡|
Experiments were then designed to test directly the possibility that at least part of the anti-LPL activity of SLE sera is attributable to cross-reactivity with anti-dsDNA antibodies. Affinity-purified anti-dsDNA antibodies were prepared from sera obtained from 5 SLE patients, and all 5 at concentrations of 25 μg/ml reacted with LPL-coated plates, yielding OD values of 0.3–0.7 in the standard assay at 1-hour incubation with the substrate. Moreover, the reactivity was blocked 75–95% by 100 μg/ml of dsDNA. Even more impressive were the reactivities of human monoclonal anti-dsDNA antibodies with LPL-coated plates, as shown in Table 5.
Table 5. Reactivity of monoclonal anti-dsDNA with LPL*
|Human monoclonal antibody||Concentration||OD value at 1 hour||% blocked by 100 μg/ml dsDNA|
Four of the 5 human monoclonal anti-dsDNA antibodies at concentrations of 1–3 μg/ml reacted strongly with LPL-coated plates, and these reactions were efficiently blocked by dsDNA. Monoclonal anti-dsDNA RH14, even at a concentration of 15 μg/ml, reacted weakly and was not blocked by dsDNA.
Clearly, at least part of the antibody activity directed to LPL resides in the anti-dsDNA antibody population. This reactivity of anti-dsDNA antibodies is not attributable to contamination of the LPL preparation with dsDNA, because treatment of LPL with DNase has no effect on the reactivity of anti-dsDNA antibodies with LPL. Comparable treatment of dsDNA solutions by DNase ablated its ability to coat plates for an anti-dsDNA reaction. The ability of dsDNA to block, on average, only 50.8% of the anti-LPL activity of SLE sera suggests that there is an antibody to LPL that is independent of anti-dsDNA antibodies and is of comparable importance. Direct evidence for the existence of IgG with anti-LPL activity other than anti-dsDNA was obtained by passing SLE IgG from anti-dsDNA–positive sera over a DNA cellulose column. In 4 of 4 instances, effluent IgG from DNA cellulose columns devoid of anti-dsDNA activity bound LPL-coated plates. Such data suggest the existence of at least 2 types of anti-LPL in SLE sera, one being anti-dsDNA and the other IgG without dsDNA-binding activity. Not surprisingly, anti-LPL and anti-dsDNA titers frequently cofluctuated, reflecting the cross-reactivity of anti-dsDNA antibodies with LPL.
Finally, 26 SLE sera had elevated levels of anti-LPL but normal levels of anti-dsDNA; conversely, 9 SLE sera were reactive with dsDNA by ELISA and negative for anti-LPL. Thus, several lines of evidence indicate that there are 2 types of anti-LPL, one of which is specific for LPL and nonreactive with dsDNA, and a second type that reacts with both LPL and dsDNA.
Ig isotypes of anti-LPL activity.
All the data presented thus far represent IgG anti-LPL activity. To assess anti-LPL activity in the IgM and IgA classes, isotype-specific activity was assessed with appropriate conjugates. A sample of 89 SLE sera that had previously been studied with an anti-IgG reagent were also studied with anti-IgM and anti-IgA reagents. Using the anti-IgG reagent, 44 (49.4%) of 89 sera were positive. When the anti-IgM and anti-IgA reagents were used, 29.2% and 18.0% of sera, respectively, were positive. Of the total sera, 54 (60.7%) were positive for antibody of 1 or more isotypes.
Presence of LPL in immune complexes.
The technique for measuring immune complexes containing LPL was described in Patients and Methods. Using 5 SLE sera with high anti-LPL activity in ELISA and 5 normal sera with negligible anti-LPL activity in ELISA, precipitates from SLE sera had 0.023–0.170 OD differences at 400 nm compared with precipitates from normal sera run in parallel manner. These data suggest that SLE sera have LPL bound to IgG in immune complexes.
Relationship of lipid levels and anti-LPL.
The existence of an immune response to LPL raises the question of whether there is an association of such antibodies and an effect on lipid levels. This was investigated by studying fasting lipid levels and the level of anti-LPL as measured by the OD of a 1:100 dilution of serum on LPL-coated plates in 85 SLE patients. The data are shown in Table 6 and illustrate the correlation of the lipid levels with the OD achieved in the anti-LPL assay.
Table 6. Relationship of anti-LPL activity and serum lipids in SLE*
|Lipid||Pearson's correlation coefficient||P|
As seen, highly significant relationships exist between anti-LPL levels and total triglycerides, triglyceride LDL, Apo B, and Apo E levels. Notably, no relationships are seen between anti-LPL levels and total cholesterol, triglyceride HDL, or Apo A1 levels. A further refinement of these data can be achieved by partitioning the SLE patients into 2 groups: those with elevated levels of anti-LPL (>0.17 OD units), and those with normal levels of anti-LPL (<0.17 OD units). These data are shown in Table 7 and, as can be seen, the correlation coefficients for triglycerides, LDL triglycerides, Apo B, and Apo E in the anti-LPL–positive group are higher than those in the total group of SLE patients. There are no statistically significant positive correlations between lipid levels and anti-LPL in the SLE patients with OD values <0.17 units. Interestingly, in the anti-LPL–positive group, there is a positive correlation coefficient of 0.46501 and a P value of 0.0192 between total cholesterol and LPL activity that was not apparent in the total SLE patient group, as shown in Table 6. There is no immediately obvious explanation that would relate decreased LPL activity with elevations of serum cholesterol, but the finding poses an important challenge for understanding these relationships.
Table 7. Lipid levels in SLE patients with anti-LPL values <0.17 and >0.17 OD units*
|Lipid||Anti-LPL <0.17 OD units (n = 47)||Anti-LPL >0.17 OD units (n = 25)|
|Pearson's correlation coefficient||P||Pearson's correlation coefficient||P|
The data in Table 7 further strengthen the possible mechanistic role of antibody to LPL in the hyperlipidemia seen in SLE patients.
- Top of page
- PATIENTS AND METHODS
These observations raise the possibility that antibodies to LPL could be a factor leading to elevated levels of triglycerides, Apo B, and Apo E. Studies on the ability of IgG with anti-LPL activity to inhibit LPL enzyme activity are underway but are thus far unrevealing. Conversely, early preliminary measurements of serum LPL activity show an inverse relationship to anti-LPL activity, suggesting either an inhibitory effect of antibody on LPL activity or the facilitation by antibody of enhanced turnover of the enzyme, either or both of which could be responsible for lower serum LPL activity. Only detailed studies that are underway can reveal the mechanisms of these observed relationships.
One must address the issue of the role of hyperlipidemia in the development and/or progression of CAD. Although all investigators agree that elevated cholesterol levels increase the risk of CAD, the data have not been as clear for elevated triglycerides and the risk of CAD. In 1991, a review of the literature revealed that triglyceride levels were frequently associated with CAD in univariate analysis (14). However, the same studies showed that the relationship between triglycerides and CAD less frequently survived multivariate analysis. Subsequent studies found more compelling data, even after multivariate analysis, relating elevated triglyceride levels and the risk of CAD (15, 16). Moreover, it has been reported that the elevated triglyceride level characteristic of familial LPL deficiency is associated with premature atherosclerosis (17). It has also been reported that mutations in the LPL gene that are not associated with complete LPL deficiency nonetheless increase the risk of ischemic heart disease (18) and progression of CAD (19). Finally, it has been reported that LPL activity is inversely related to the severity of angina pectoris in patients with CAD (20).
It is becoming clear that elevated triglyceride levels can contribute to the risk and/or progression of CAD, and that decreases of LPL activity can play a significant role in such triglyceride elevations. In any case, the presence of a positive correlation between triglyceride levels and anti-LPL activity supports a continuing analysis of this correlation and its relationship to premature myocardial infarction and stroke in patients with SLE.