|Home | About | Journals | Submit | Contact Us | Français|
Monocyte chemoattractant protein-1 (MCP-1), encoded by the CCL2 gene, plays an important role in the initiation and progression of atherosclerosis. Ea.hy 926 endothelial cells grown under low folate conditions (LO cells) synthesize more MCP-1 mRNA and secrete more MCP-1 protein than folate-replete control cells (HI cells). We investigated the mechanisms underlying the modulation of MCP-1 expression by long-term “folate stress”.
CCL 2 transcription, assessed using promoter-reporter assays, is up-regulated in LO cells relative to HI cells, whereas MCP-1 mRNA stability is unchanged. This quantitative transcriptional bias under chronic low folate conditions is not attributable to differences in active NF-κB, but is associated with elevated levels of both total p38 and phospho-p38 that are detectable by Western immunoblotting. Transient, acute methotrexate-mediated folate depletion or exposure to high concentrations of homocysteine (Hcy) had no effect on MCP-1 synthesis by Ea.hy 926 cells. The p38 inhibitor SB-203580 abolished the excess MCP-1 production by LO cells. The quantitative transcriptional bias of CCL2 in LO cells was retained following massive induction by TNF-α.
During long-term folate stress, p38 is the primary determinant of CCL2 transcription. Long-term folate insufficiency “primes” Ea.hy 926 endothelial cells to have a quantitatively more vigorous response to cytokine-mediated inflammatory stress.
Homocysteine (Hcy) is a sulfur-containing amino acid formed during one-carbon metabolism (Supplemental Figure 1), the circulating concentration of which is elevated in hyperhomocysteinemia (HHcy). HHcy, generally defined as a circulating concentration of total Hcy between 13 and 100μmol/L, results from common functional polymorphisms in 5,10-methylenetetrahydrofolate reductase, cystathionine β synthase, methionine synthase, and other enzymes, as well as nutritional deficits and environmental factors. It is considered to be a graded, independent risk marker for atherothrombotic diseases [1–4]. Chronic low folate status is one of the most important dietary determinants for mild HHcy . However, folate can restore normal function to endothelium in which dysfunction has been experimentally induced by acute HHcy, even in the absence of any reduction in Hcy concentrations [6, 7]. Thus, it is unclear whether the risk of atherothrombotic diseases that is associated with a low folate/high Hcy phenotype long term is mediated directly via chronic folate insufficiency itself or indirectly via the attendant mild HHcy. To date, most cell culture studies have focused on the effects of extracellular applications of Hcy and have identified several potential mechanisms at the cellular level, including inflammation involving monocyte chemoattractant protein-1 (MCP-1) and IL-8 , oxidative stress , inhibition of glutathione peroxidase , and endoplasmic reticulum stress and unfolded protein response .
Endothelial dysfunction and consequent vascular injury is an early event in atherogenesis. MCP-1, a chemokine that is encoded by the CCL2 gene, mediates monocyte recruitment and entry into vessel walls at sites of injury , and plasma MCP-1 levels are elevated in rats made hyperhomocysteinemic by long-term diets rich in methionine and low in folate [13, 14].
We have previously established an in vitro model using the Ea.hy 926 endothelial cell line adapted to long-term low and high folate culture conditions . Unlike primary endothelial cells, Ea.hy 926 can be grown in bulk for many generations; however, although they have endothelial cell morphology and surface markers, they remain an approximate, rather than an exact, model. Ea.hy 926 cells grown in low folate medium have much lower intracellular folate concentrations than those grown in high folate medium but secrete only slightly more Hcy, making the model ideal for identifying phenotypic effects attributable to chronic folate insufficiency rather than to concomitant differences in Hcy concentration, or to acute applications of Hcy. In this model folate insufficiency is coincident with increased MCP-1 secretion, Ea.hy 926 cells grown under low folate conditions having two to three fold as much extracellular MCP-1 as cells grown under high folate conditions. The present study was undertaken to investigate the mechanisms underlying the modulation of MCP-1 expression by long-term “folate stress”, in the absence of Hcy-attributable effects.
Low folate adapted (LO) or control (HI) Ea.hy 926 cells were grown in Medium 199 (Invitrogen), containing 23 nM of folic acid, supplemented with 10% fetal calf serum, non-essential amino acids, gentamycin, penicillin G, and fungizone in the absence (LO) or presence (HI) of 9 μM of folic acid as described previously . Curcumin and SB-203580 were purchased from Sigma. TNF-α was a gift from AstraZeneca Pharmaceuticals.
Cell preparation for biochemical assays was described in Supplemental Methods.
The pCCL2-pt-luc vector, containing the 930 bp distal regulatory region upstream from the major transcriptional start site of the CCL2 gene cloned into the pGL3-basic vector (Promega) 5′ to the firefly luciferase reporter gene, was made according to Rovin et al . LO and HI cells were transfected as described in Supplemental Methods.
Isolation of total RNA, cDNA synthesis, and quantitative real-time RT-PCR for MCP-1 mRNA was performed as described previously .
MCP-1 concentrations in media were measured using an ELISA kit (PeproTech, Inc, Rocky Hill, NJ), with adjustment by protein content in the corresponding cell fraction.
Total cellular proteins (20 μg) were separated by SDS-PAGE. Western immunoblotting was performed using the following antibodies: phospho-p38 and p38 (Cell Signaling Technology), and GAPDH (Santa Cruz). The p38, phosphorylated p38 and GAPDH bands were detected using enhanced chemiluminescence reagents (GE Healthcare Amersham) and film. Bands were then digitized by scanning, and figures were assembled using PhotoShop.
Data were presented as the mean ± standard error of the mean and compared using the unpaired Student’s t test. Values of P<0.05 were considered significant. Results shown are representative of at least three independent experiments.
Acute treatment with Hcy has been reported to increase MCP-1 secretion in human aortic endothelial cells . To test if the elevation in MCP-1 secretion by Ea.hy 926 endothelial cells under conditions of chronic folate insufficiency previously reported by us  could be further modified by exposure to excess Hcy, LO and HI cell cultures were incubated for different periods of time with 0.5 mM DL-Hcy. Both cultures showed a time-dependent accumulation of secreted MCP-1 that, as expected, was greater in LO cells than HI cells (Figure 1A). Although extracellular Hcy caused a significant increase of intracellular Hcy levels (Figure 1B) in each culture at each time point (except the increase in HI cells at the 48 time point, which did not reach statistical significance), it did not result in increased secretion of MCP-1 or changes in intracellular folate derivative levels in either LO or HI cells (data not shown). Similar patterns for secreted MCP-1 and intracellular Hcy in LO and HI cells were seen after treatment with extracellular Hcy for up to 7 days (data not shown).
Hcy, and secreted MCP-1 Methotrexate (MTX), an inhibitor of dihydrofolate reductase , is known to acutely lower intracellular folate levels . To determine the effect, if any, of acute folate lowering on secreted MCP-1, LO and HI cells were treated with 0.5 μM MTX for 2 days. A substantial lowering of intracellular folate derivative levels in both LO (Figure 2A) and HI (Figure 2B) cells were not accompanied by markedly increased secreted or intracellular Hcy concentrations (Figures 2C and 2D). Furthermore, MTX treatment had no measurable effect upon MCP-1 secretion by either LO or HI cells (Figure 2E), and the higher MCP-1 levels in LO cells relative to HI cells remained unchanged.
We have previously reported  that MCP-1 mRNA levels are between 2 and 3 fold higher in LO cells than in HI cells. The elevated MCP-1 mRNA levels observed in LO cells relative to HI cells are likely attributable to either increased transcription or enhanced mRNA stability. The CCL2 transcriptional activities in LO and HI cell cultures were assessed after co-transfection of cells with the pCCL2-pt-luc and control pRenilla-TK reporter plasmids. LO cells had significantly higher levels of luciferase activity than HI cells (Figure 3A) reflecting enhanced levels of CCL2 transcription in the former. To test whether different MCP-1 mRNA degradation rates contribute to the difference in MCP-1 mRNA levels between LO and HI cells, confluent cultures of each were treated with actinomycin D, and total RNA was extracted at 0, 2, 4, 6, 9 and 24 hours and assayed by MCP-1 specific real-time PCR. No difference in MCP-1 mRNA half-life was observed between LO and HI cells (Figure 3B), indicating that the degradation kinetics are the same for both. The above findings indicate that the difference in MCP-1 mRNA concentrations between LO and HI cells is due to a transcriptional advantage in the former rather than to mechanisms involving transcript stabilization.
Short-term treatment of vascular smooth muscle cells and THP-1 monocytes with Hcy has been reported to induce MCP-1 secretion via activation of the NF-κB pathway [20, 21]. To test if the NF-κB pathway is involved in increased MCP-1 secretion in Ea.hy 926 endothelial cells subjected to chronic folate insufficiency, the NF-κB-luc reporter was transfected into LO and HI cells. No difference in NF-κB activity was observed between LO and HI cells (Figure 4A). LO and HI cells were also treated with 50 μM of curcumin, a known inhibitor of NF-κB pathway, for 3 hours (exposure to curcumin for 24 hours leads to cell death, data not shown). MCP-1 secretion, examined 24 hours after removal of curcumin, was not depressed; rather, it was modestly increased in both LO and HI cells (Figure 4B). To confirm that CCL2 transcription could be induced by pro-inflammatory NF-κB-dependent mechanisms, LO and HI cells were treated with TNF-α, reported by others to activate the NF-κB pathway and MCP-1 expression in Ea.hy 926 cells , in the absence and presence of curcumin. Although MCP-1 mRNA levels were dramatically elevated in both LO and HI cells after treatment with TNF-α the ratio of MCP-1 mRNA between LO and HI cells remained similar to those observed in the absence of TNF-α stimulation (Figure 4C). Analogous qualitative and quantitative responses were seen in MCP-1 protein levels (data not shown). As expected, curcumin ablated approximately 75% of the TNF-α dependent MCP-1 mRNA elevation in both LO and HI cells; again the ratio of MCP-1 mRNA between the two cultures remained similar to those observed at baseline. These experiments suggest that NF-κB is not involved in regulating the elevated secretion of MCP-1 by LO cells relative to HI cells in the absence of an overt inflammatory stimulus.
The p38 inhibitor SB-203580 has been reported to lower MCP-1 secretion stimulated by Hcy treatment in monocytes  and endothelial cells . To test if p38 is involved in the elevated MCP-1 secretion attributable to chronic folate insufficiency, LO and HI cell cultures were treated with 20 μM of SB-203580 in DMSO carrier or with carrier alone for 24 hours. SB-203580 significantly decreased MCP-1 secretion by LO cells but not by HI cells (Figure 5A). Furthermore, SB-203580 significantly reduced MCP-1 mRNA levels and pCCL2-pt-luc reporter activity in LO cells but not in HI cells (Figures 5B and 5C). Therefore p38 is involved in the increased MCP-1 transcription, synthesis and export that are attributable to low folate status. It has been reported by others that the increased levels of phosphorylated components in the p38 MAPK pathway might be involved in Hcy-induced MCP-1 secretion in endothelial cells [24, 25]. To determine the relative levels of total and activated p38 in both LO and HI cells, Western immunoblotting was performed with antibodies to total p38 and to the phosphorylated form. LO cells expressed higher levels of both total p38 and phospho-p38 levels than HI cells (Figure 5D), suggesting that low folate status modulates baseline MCP-1 expression in Ea.hy 926 cells through p38.
MCP-1 is one of the inflammatory chemokines that is upregulated during early atherogenesis and is expressed by endothelial cells in developing atherosclerotic lesions . In the general population, low folate status is a common determinant of HHcy , which in turn is regarded as a graded, independent risk marker for atherothrombotic disease [1–4]. While most research has investigated the effects of Hcy on the pathogenic processes underlying atherosclerosis, we investigated the possible mechanism by which chronic folate insufficiency induces endothelial cells to produce elevated amounts of MCP-1. Our results suggest that chronic folate insufficiency leads to augmented p38 expression, which in turn, enhances CCL2 gene transcription. This results in increased MCP-1 biosynthesis that is followed by export of MCP-1 from the cells.
In contrast to reports from others in which acute Hcy treatments of several different cell types induce elevated MCP-1 secretion [8, 21, 23], the present study showed that acute Hcy treatment of Ea.hy 926 endothelial cells had no effect on MCP-1 secretion for up to 48 hours post-treatment, regardless of whether cells were maintained under low or high folate conditions (Figure 1A). This result was observed even though the cells treated with Hcy had higher intracellular Hcy than control cells (Figure 1B). Furthermore, short-term MTX treatment, which lowered intracellular folate derivative concentrations in both LO and HI cells, did not affect MCP-1 secretion (Figure 2E). Taken together, these results suggest that the elevated MCP-1 export by Ea.hy 926 cells adapted to chronic folate insufficiency cannot be re-capitulated by acute stimuli (i.e. short-term increase of intracellular Hcy mediated by the application of extracellular Hcy, or rapid folate lowering). Rather, it appears to be the result of signal transduction pathways that are imposed by chronic folate insufficiency. The differences between our results and those reported by others may reflect MCP-1 regulatory mechanisms that differ between Ea.hy 926 cells and monocyte/macrophage cells [21, 23] and between Ea.hy 926 cells and endothelial cells of aortic origin .
We have previously reported that MCP-1 mRNA and secreted protein levels are higher in LO cells than in HI cells . Steady-state levels of mRNA reflect the relative balance between transcription and degradation. The CCL2 transcriptional activity and MCP-1 mRNA degradation experiments (Figures 3A and 3B) suggested that the difference in steady-state MCP-1 mRNA levels mandated by chronic folate insufficiency is likely to result from differential activity of transcription signaling pathways.
The NF-κB pathway plays an important role in atherogenesis  and regulates the expression of many atherosclerosis-related inflammatory cytokines and chemokines . Acute treatment with Hcy can induce elevated MCP-1 secretion via the NF-κB pathway in human umbilical vein endothelial cells, THP-1 cells, and human vascular smooth muscle cells [20, 29, 30]. However, no previous studies have examined the effects of chronic folate insufficiency on NF-κB activation. The data from our experiments using NF-κB promoter assays showed that the levels of activated NF-κB do not differ between Ea.hy 926 cells adapted to growth under low or high folate conditions (Figure 4A). In addition, curcumin, an inhibitor of NF-κB, did not lower the MCP-1 secretion by folate depleted LO cells (Figure 4B). These results suggest that NF-κB in Ea.hy 926 endothelial cells is not responsive to chronic folate stress.
The inflammatory cytokine TNF-α has been reported by others to stimulate MCP-1 expression via the NF-κB pathway in Ea.hy 926 cells grown in regular medium , prompting an evaluation of whether folate depleted LO cells retain the ability to respond to TNF-α by activating NF-κB. Treatment of both LO and HI cells with TNF-α massively induced MCP-1 mRNA (Figure 4C) and protein (data not shown). Despite the magnitude of induction, the differential expression of MCP-1 in LO and HI cells was retained. Furthermore, even after treatment with the NF-κB inhibitor curcumin, which largely ablates the MCP-1 induction driven by NF-κB, the residual MCP-1 expression exhibited the same ratio in favor of LO cells. This indicates that the two processes governing responses to chronic folate stress and inflammation are distinct. The folate stress dependent effect does not appear to be modified by the quantitatively greater, transient, upregulation mediated by TNF-α, which further supports the proposition that the chronic folate stress dependent effect is not mediated via NF-κB.
The p38s are a group of stress-activated protein kinases  and represent rational candidates for mediating folate stress-induced MCP-1 secretion in endothelial cells. SB-203580, a potent p38 inhibitor, consistently lowered MCP-1 mRNA and protein levels as well as CCL2 transcriptional activity in chronically folate depleted cells, indicating that p38 is involved in the MCP-1 elevation observed under long term folate stress. Activation of p38 may be assessed by examining p38 phosphorylation . Therefore we studied the effect of chronic folate insufficiency on p38 activity. Our results showed that LO cells have higher total p38 as well as phosphorylated p38 than HI cells, suggesting that an elevated p38 level mandated by chronic folate stress is accompanied by a proportionally increased amount of phosphorylated p38. Acute treatments of endothelial cells with Hcy [24, 25], thrombin , and interaction with activated platelet  elevate phosphorylated p38 without any change in total p38. The results presented here show that Ea.hy 926 endothelial cells grown in chronic low folate condition acquire higher levels of baseline p38 which appears to contribute to the elevated baseline MCP-1 transcription and MCP-1 protein secretion observed in cells under folate stress. Therefore, our study suggests that under chronic folate stress, endothelial cells are “primed” via elevated baseline p38 level for increased MCP-1 production in the absence of other environmental challenges.
In the presence of an inflammatory stimulus, NF-κB is responsible for the quantitatively much larger CCL2/MCP-1 response. The ultimate level of induction likely depends on a combination of inputs from chronic folate stress and acute inflammatory stimuli that upregulate transcription via p38 and NF-κB, respectively. The p38 mediated “priming” of cells under chronic folate stress introduces an additional regulatory mechanism governing MCP-1 production that is likely to be particularly relevant in individuals who are nutritionally compromised. The potential biological and clinical implications of this observation warrant further studies.
In conclusion, p38 is the primary determinant of CCL2 transcription during long-term folate stress. Folate stress “primes” Ea.hy 926 endothelial cells to have a quantitatively more vigorous response to cytokine-mediated inflammatory stress. Transient, acute methotrexate-mediated folate depletion or exposure to high concentrations of homocysteine had no effect on MCP-1 synthesis by Ea.hy 926 cells.
This work was supported by National Institutes of Health grants AR47663, HD039195, P20RR020741 and ES013508, and by grant 4100038714 from the Pennsylvania Department of Health
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.