Lipoprotein Lipase Deficiency Impairs Bone Marrow Myelopoiesis and Reduces Circulating Monocyte Levels

Abstract

Objectives

Tissue macrophages induce and perpetuate pro-inflammatory responses, thereby promoting metabolic and cardiovascular disease. Lipoprotein lipase (LpL), the rate-limiting enzyme in blood triglyceride (TG) catabolism, is expressed by macrophages in atherosclerotic plaques. We questioned whether LpL, which is also expressed in the bone marrow (BM), affects circulating white blood cells (WBCs) and BM proliferation, and modulates macrophage retention within the artery.

Approach and Results

We characterized blood and tissue leukocytes and inflammatory molecules in transgenic LpL-knockout mice rescued from lethal hypertriglyceridemia within 18 h of life by muscle-specific LpL expression (MCKL0 mice). LpL-deficient mice had ~40% reduction in blood WBC, neutrophils, and total and inflammatory monocytes (Ly6C/G^hi). LpL deficiency also significantly decreased expression of BM macrophage-associated markers (F4/80 and TNF-α), master transcription factors (PU.1 and C/EBPα) and colony-stimulating factors (CSFs) and their receptors (CSF-Rs), which are required for monocyte and monocyte precursor proliferation and differentiation. As a result, differentiation of macrophages from BM-derived monocyte progenitors and monocytes was decreased in MCKL0 mice. Further, while LpL deficiency was associated with reduced BM uptake and accumulation of TG-rich particles and M-CSF-M-CSF-R binding, TG lipolysis products (e.g., linoleic acid) stimulated expression of M-CSF and M-CSF-R in BM-derived macrophage precursor cells. Arterial macrophage numbers decreased after heparin-mediated LpL cell dissociation and by genetic knockout of arterial LpL. Reconstitution of LpL-expressing BM replenished aortic macrophage density.

Conclusions

LpL regulates peripheral leukocyte levels and affects BM monocyte progenitor differentiation and aortic macrophage accumulation.

INTRODUCTION

Although inflammation primarily serves to ameliorate injury or infection, tissue macrophage-mediated inflammatory responses promote atherosclerotic cardiovascular disease (CVD) and metabolic disorders. Accumulation of lipoproteins, especially LDL, in the aortic wall leads to an inflammatory cascade that attracts blood leukocytes and traps monocytes in the subendothelial space ¹. Infiltrated monocytes can differentiate to macrophages that phagocytose accumulated LDL and become foam cells, which contribute to fatty streaks that precede atherosclerotic plaque formation ². High white blood cell (WBC) levels thus directly accelerate atherosclerotic CVD ^{3, 4} and are also associated with obesity and type 2 diabetes, which also increase risk of CVD ^5–7. Further, hypercholesterolemia in mice promotes blood leukocytosis, increases in inflammatory monocyte (Ly6C^hi) levels, and macrophage infiltration to enhance progression of atherosclerotic lesions ⁸.

Tissue macrophages robustly express lipoprotein lipase (LpL), the rate-limiting enzyme in circulating triglyceride (TG) metabolism. Macrophages within atherosclerotic lesions are a major site of LpL expression ^{9, 10}. Loss-of-function studies show that macrophage-derived LpL increases atherosclerosis ^11–13, suggesting a direct role of LpL levels in mediating atherosclerotic lesion development. In our previous studies, aortic LpL levels in non-hypercholesterolemic mice positively correlated with aortic wall macrophage levels ^{14, 15}. Several mechanisms by which LpL may promote atherosclerosis have been proposed: a) LpL acts as a non-enzymatic bridging molecule that anchors atherogenic lipoproteins to the extracellular matrix in order to promote lipoprotein retention ^16–18; b) LpL hydrolyzes lipoproteins to produce pro-atherogenic, cholesterol-rich LDL ^{1, 19}; c) LpL-mediated lipolysis products serve as energy substrates (fatty acids, FAs) that increase macrophage inflammatory responses ²⁰; d) LpL functions as a monocyte adhesion molecule ^21–23; and e) LpL induces pro-inflammatory TNF-α expression and production by macrophages ²⁴. Atherogenic actions of LpL might thus lead to a “vicious cycle” whereby elevated macrophage levels increase LpL secretion, which promotes additional macrophage accumulation in the aortic wall. Recent genetic studies in humans have associated heterozygous LpL deficiency with early-onset coronary heart disease ²⁵, an effect that might reflect hypertriglyceridemia, increased postprandial lipemia, and low HDL cholesterol levels. The systemic and local effects of LpL on atherogenesis are still not fully defined.

We hypothesized that expression of LpL by WBCs and their precursors alters the circulating levels of monocytes as well as their entry into the aortic wall to promote atherosclerosis. We first tested this by examining whether LpL expression affects circulating WBC numbers and bone marrow (BM) myelopoiesis. Here we show that LpL deficiency was associated with reduced circulating WBC and monocyte levels, reduced BM colony-stimulating factor (CSF) and CSF-receptor (CSF-R) expression, and fewer macrophages in the aorta of high fat-fed mice.

METHODS

Materials and Methods are available in the online-only Data Supplement.

RESULTS

LpL deficiency alters blood leukocyte profiling

To assess whether LpL expression modulates circulating WBC levels and subtypes, LpL-deficient mice (TmLpL0) were generated by crossing floxed LpL mice with β-actin-driven tamoxifen-inducible-Cre transgenic mice ²⁶. On a chow diet, total WBCs, monocytes and percent monocytes/WBC (inset) were reduced in TmLpL0 mice compared to control floxed mice (Figure 1A). Because TmLpL0 mice are hypertriglyceridemic (~400 mg/dL, n=10), we questioned whether the alteration in blood leukocyte levels was caused by elevated blood TG levels rather than by LpL deficiency. Using an LpL deletion mouse model that expresses LpL only in skeletal muscle (to prevent neonatal death) and maintains normal cholesterol and TG levels ²⁷, denoted as MCKL0 mice ²⁸, we determined that this effect is indeed attributable to LpL deficiency (Figure 1B). The reduction in the percentages of monocytes/WBC (inset) in MCKL0 mice is consistent and comparable with percent monocytes/WBC in TMLpL0 mice, 40±0.62% vs. 36±1.81%, respectively (inset in Figures 1A and 1B). Other major leukocyte subsets, i.e., neutrophils and lymphocytes, were reduced by 30% and 60%, respectively (n=12–15, p<0.05). We also observed a marked shift in the ratio of Ly6C/G^hi (Gr-1^hi) to Gr-1^lo monocytes that resemble human inflammatory classical and anti-inflammatory non-classical monocytes, respectively, by flow cytometry. Compared to wild type (Wt) mice, blood monocytes in MCKL0 mice had a significantly greater non-classical to classical ratio (Figures 1C and 1D) comprising a smaller percentage of total WBCs (Figure 1D, inset). Both Wt and MCKL0 mice showed no signs of active inflammatory responses.

Figure 1.

LpL deficiency alters blood white cell (WBC) profiling. Total WBCs, blood monocytes (MO) and % MO/WBC (inset) in (A) floxed control and TmLpL0 mice (n=9 or 10) and in (B) Wt and MCKL0 mice (n=12–15) mice. *, p<0.001. (C) Representative flow cytometry plots and (D) quantification of MO subsets. Inset: quantification of blood total MO (CD45⁺CD115^hiGr-1⁺) by flow cytometry. *, p<0.001, classical vs. non-classical; **, p<0.05, MCKL0 vs. Wt, n=4, mean±SE.

LpL deletion modulates expression of growth factors required for BM hematopoiesis and myelopoiesis

We next studied whether changes in blood leukocyte profiling and decreases in total leukocytes in MCKL0 mice were associated with altered BM myelopoiesis. We observed decreased gene expression of macrophage-associated markers—F4/80, TNF-α, and mannose receptor (MR) —in unsorted BM cell homogenates of MCKL0 mice compared to Wt (Figure 2A). Expression of master regulatory transcription factors required for hematopoietic progenitor cell proliferation and differentiation, including PU.1 (p<0.05) and CEBP/α (p<0.05) (Figure 2B), was also reduced in BM of MCKL0 mice when compared with Wt mice. Reduced expression of PU.1 and C/EBPα may be attributable to the reduction in BM total lineage negative (lin⁻) stem and progenitor cells in MCKL0 mice when compared to Wt (supplemental Figures I a and I b). However, populations of BM progenitor cells at early hematopoietic branch points including hematopoietic stem and progenitor cells, common myeloid progenitors, and granulocyte-macrophage progenitors were not altered in MCKL0 compared to Wt mice (supplemental Figure I c).

Figure 2.

Loss of LpL reduces bone marrow (BM) expression of growth factors essential for myelopoiesis and differentiation of MO to macrophages. (A) mRNA gene analyses of LpL (inset), macrophage marker (F4/80), macrophage-derived TNF-α and MR, (B) hematopoietic transcription factors (PU.1 and C/EBPα) and (C) G-CSF, GM-CSF and M-CSF and (D) their corresponding receptors in BM of Wt and MCKL0 mice. n=4–6, mean±SE. *, p<0.05.

These modifications in blood leukocyte profiling can also occur at the terminally differentiated stages of hematopoiesis. We thus examined whether LpL deficiency affects expression of growth factors that act on BM progenitor cells at lineage-committed stages, such as monocyte progenitors and monocytes. Strikingly, BM expression of all CSFs, including granulocyte-CSF (G-CSF), granulocyte macrophage-CSF (GM-CSF) and macrophage-CSF (M-CSF) (Figure 2C), as well as their corresponding receptors (G-CSF-R, GM-CSF-R and M-CSF-R, respectively), were substantially decreased in MCKL0 mice (Figure 2D).

Alterations in M-CSF and M–CSF-R expression relate to LpL-mediated TG metabolism

BM expression of LpL is associated with TG-rich lipoprotein catabolism, such as chylomicrons ²⁹. We questioned if LpL-mediated reductions in CSF/CSF-R expression that contribute to altered BM myelopoiesis and WBC levels reflect changes in cellular TG uptake. We speculated that LpL deficiency would reduce cellular TG uptake and metabolism, which would alter the expression of essential growth factors for myeloid cell differentiation and proliferation. To test this, we treated differentiating macrophages derived from Wt and MCKL0 mouse BM with linoleic acid (LA)-rich TG particles (soy oil-based lipid emulsion, Intralipid), a key pathway for delivering FAs to cells, labeled with [³H]cholesteryl ether ³⁰. As previously measured by our laboratory^30–32, we have established that TG mass essentially equals TG deposition since basal cellular TG levels are low and cellular TG synthesis is negligible during the 4h incubation (i.e., less than 1% of TG uptake). Our studies examining BM cell TG metabolism show that cellular TG uptake and accumulation were reduced by 22% and 50%, respectively, in the LpL-deficient cells (Figure 3A). The ratio of cellular TG deposition to TG uptake, which represents cellular TG utilization, was reduced by ~40% in BM-derived differentiating macrophages from MCKL0 mice compared to those in Wt mice (Figure 3A).

Figure 3.

LpL deletion affects cellular TG uptake and expression of M-CSF and M-CSF-R in BM-derived macrophages. Effects of LpL on (A) cellular TG uptake, TG deposition and TG utilization (ratio of TG deposition/uptake) in BM-derived differentiating macrophages. Results represent the mean±SE of experiments performed in triplicate (n=6). BM-derived macrophage precursor cells from MCKL0 (B) and Wt (C) mice (n=6) were treated with LpL (1 μg/ml), LpL+TG (100μg/ml) or LpL-mediated lipolysis products (LA, 150 μM in 1% BSA) for 3 h. Non-treated, NT. Changes in M-CSF, M-CSF-R and PPARγ (D) mRNA expression were assayed by RT-PCR. *, p<0.05.

We also examined whether the addition of exogenous LpL can rescue the defective TG uptake and metabolism. Addition of exogenous LpL increased TG uptake in Wt BM-derived macrophage precursor cells that already have adequate expression of LpL. The addition of exogenous LpL to the LpL-deficient cells increased cellular TG uptake, mass and utilization to levels similar to that in Wt cells (Figure 3A).

We next examined whether exogenous LpL or TG-derived metabolites, such as LA, can modulate M-CSF/M-CSF-R expression in BM-derived monocyte or macrophage precursor (BMMP) cells from MCKL0 mice. Differentiating monocytes or monocyte precursors were treated with LpL, LA-rich TG particles (Intralipid) in the presence of LpL and LA. The results show that the addition of exogenous LpL and TG particles had a small effect on M-CSF expression but had no effect on M-CSF-R expression. LA stimulated expression of both M-CSF and M-CSF-R (Figure 3B) (p<0.05) in MCKL0 BM-derived myeloid precursor cells. In Wt BMMP cells, while the addition of exogenous LpL alone had no effect on M-CSF and a small effect on M-CSF-R (p<0.05) expression, addition of TG in the presence of LpL increased M-CSF-R expression (ns) with a lesser effect on M-CSF levels. LA stimulated expression of M-CSF (Figure 3C).

Because of earlier reports on the role of peroxisome proliferator-activated receptor γ (PPARγ) in influencing lipid metabolism and production of M-CSF, we questioned if transcription of M-CSF is regulated by the PPARγ-mediated NF-κB signaling cascade. Our data show a consistent inverse relation between PPARγ and M-CSF in BMMP cells. Expression of M-CSF was downregulated, while PPARγ was upregulated, in the untreated BMMP cells in MCKL0 mice when compared to Wt. Further, when additional exogenous FA was supplied during BMMP differentiation, increases in M-CSF expression were associated with the suppression of PPARγ expression in both Wt and MCKL0 mice (Figure 3D).

LpL deletion attenuates BM-derived monocyte differentiation to macrophages

Because our data show that LpL deficiency reduced the expression of growth factors necessary for the differentiation of monocytes or monocyte precursor cells, we sought to directly assess whether the reduced expression of CSFs and CSF-Rs in BM or in local tissues modulates differentiation or proliferation of BM myeloid precursor cells. We isolated BM cells, including monocyte/macrophage precursors, from MCKL0 and Wt mice. There was no significant difference in total proliferated BM cells determined by hemocytometer (70~80 × 10⁶ cells/femur) between Wt and MCKL0 mice. Isolated BM myeloid precursor cells were cultured according to the standard protocol, including the addition of M-CSF to stimulate cell differentiation (See supplemental Methods and Materials). At day 1, similar numbers of monocytes, neutrophils and macrophages adhered to the culture dishes (Figures 4A and 4B inset). At day 7, quantification of harvested mature differentiated macrophages by flow cytometry revealed that monocyte differentiation to macrophages was decreased by ~40% under the MCKL0 background (Figures 4A and 4B). Thus, our findings (Figures 3A–B, 4A–B) suggest that LpL deficiency impairs monocyte/macrophage differentiation by causing reduced M-CSF/M-CSF-R expression and by attenuating cellular responsiveness to M-CSF.

Figure 4.

LpL deficiency reduces differentiation of BM-derived monocytes to macrophages and association of M-CSF and M-CSF-R. (A) Representative images, flow cytometry plots and (B) quantitation of macrophages differentiated from BM precursor cells at day 7 (inset, plated MO, neutrophils-NE and macrophages-MΦ at day 1) in Wt and MCKL0 mice. *, p<0.05 vs. Wt, n=6. (C) Cellular association of M-CSF and M-CSF-R was examined by flow cytometry in BM-derived MΦ precursor cells in Wt and MCKL0 mice. *, p<0.05. n=6. (D) Effects of sodium chlorate (NaClO₃) on cellular association of M-CSF and M-CSF. *, p<0.05. n=9, mean±SE.

LpL promotes binding of M-CSF to its receptor

We next sought to examine the potential mechanisms underlying the effects of LpL on modulating BM-derived macrophage precursor cellular responsiveness to M-CSF. Mammalian M-CSF is expressed as three subtypes – soluble glycoprotein, cell membrane-associated glycoprotein and matrix-anchored proteoglycan (PG) – that differentially stimulate target cells expressing M-CSF-R. It has been suggested that the density of monocyte-differentiated macrophages at the particular sites is controlled by local synthesis of the membrane-associated M-CSF and/or selective sequestration of the secreted PG form of M-CSF ³³. When we assessed this effect in vivo by sorting differentiating BM myeloid precursor cells in MCKL0 and Wt mice, we observed fewer M-CSF⁺/M-CSF-R⁺ cell populations in MCKL0 mice compared to Wt (Figure 4C). Potential interacting factors that may affect M-CSF and M-CSF-R cellular interaction were examined. These include heparin, a LpL cell dissociator, THL, a LpL lipolytic activity inhibitor, and sodium chlorate (NaClO₃), a proteoglycan synthesis inhibitor. Our data show that THL has no effects on M-CSF and M-CSF-R interaction, suggesting that the interaction of M-CSF and M-CSF-R is independent of LpL enzymatic activity (data not shown) and addition of heparin had no effects on cell association of M-CSF and M-CSF-R in BM (data not shown). On the other hand, cellular association of M-CSF and M-CSF-R was significantly reduced by NaClO₃ (Figure 4D), which supports the hypothesis that LpL plays a role in bridging BM M-CSF proteoglycan to cells to stimulate the maturation of monocyte precursor or monocytes. We thus propose that LpL modulates BM myelopoiesis by influencing M-CSF interaction with its receptor, which triggers a downstream signaling cascade to stimulate monocyte differentiation to macrophages.

Aortic macrophage levels correspond to LpL expression

LpL actions described above led us to investigate in vivo the effects of LpL deletion on monocyte homing and macrophage accumulation in the aortic wall. For these effects, we proposed that LpL would need to be present at the surface of cells. To test this hypothesis, we utilized heparin to dissociate LpL from cell surfaces. We examined whether LpL affects aortic monocyte or macrophage levels through its cell-surface localization in the aortic lumen modulated by heparin. Wt mice were fed a chow or high saturated fat (HF) diet for 12 weeks in order to increase arterial LpL levels as previously described³⁴. Injection of heparin resulted in increased plasma free FA (FFA) levels (Figure 5A) and decreased plasma TG levels in both chow and HF feeding groups (Figure 5A). Plasma cholesterol levels were increased by HF diet but were not affected by heparin (Figure 5A). Aorta sections were stained with specific fluorescent markers for LpL, macrophages, and endothelial cells (ECs). Consistent with our previous reports ^{14, 34}, aortic immunofluorescence of ECs and smooth muscle cells was not altered by heparin treatment or HF-feeding, suggesting that these cells remained intact in the artery and their proliferation is not altered by heparin or HF diet (supplemental Figure II). On the other hand, HF-fed mice exhibited increased LpL and macrophage immunofluorescence in the aorta when compared with chow-fed mice (Figure 5B), with a major fraction of aortic LpL co-localizing with macrophages as indicated by yellow fluorescence in the merged images. Heparin treatment reduced this co-localization, especially in HF-fed mice (Figure 5C), indicating that a significant portion of macrophage-associated LpL is extracellular. Heparin treatment also led to a significant decrease in aortic macrophage number in HF-fed mice (Figure 5C), likely attributable to reduced LpL at cell surfaces, which would decrease macrophage levels in the aorta. This effect was rapid, manifesting by 10 min after heparin treatment, indicating that heparin treatment causes a rapid egress of these cells from the arterial wall.

Figure 5.

Effects of HF diets and heparin on plasma lipid profiles and arterial cell numbers. (A) At the end of chow or HF feeding, C57BL/6 mice were injected with saline (open bars) or heparin (filled bars) 10 min before sacrifice. Blood samples were assayed for (A) FFA, TG, and cholesterol (Chol). The results are expressed as mean±SE (n=10). (B) Representative confocal images and (C) quantitation of immunofluorescence (arbitrary units, AU) of aortic EC (blue), LpL (green), and macrophages (MΦs, red) in C57BL/6 mice fed chow or HF diets with or without heparin injection. L=lumen. Magnification is 40x. Scale bar: 50μm. n=3–4. #, (−) vs. (+) heparin; *, HF (−) vs. chow (−), p<0.05.

To confirm that these changes in aortic macrophage levels were due to the presence of aortic LpL, we examined aortic macrophage density and distribution in MCKL0 mice. Levels of aortic macrophage immunofluorescence were visually low in both chow-fed Wt and MCKL0 mice (Figure 6A), however, quantified macrophage levels were significantly reduced in MCKL0 mice when compared with Wt mice (Figure 6B). We next examined the role of LpL expressed specifically by macrophages in influencing aortic macrophage levels by comparing the number of aortic macrophages in MCKL0 mice transplanted with MCKL0 BM producing LpL-deficient macrophages (MCKL0→MCKL0) or Wt mouse BM producing LpL-expressing macrophages (Wt→MCKL0), as well as control Wt mice transplanted with Wt BM (Wt→Wt; control). Four weeks after transplantation, mice were placed on HF diets for 8 weeks. Although HF diets can increase MCKL0 mouse plasma cholesterol levels, their plasma TG levels remained in the normal range ²⁷. In the aorta, as expected, HF diets increased macrophage levels in Wt mice transplanted with Wt BM (HF Wt→Wt) when compared with the chow-fed group (Chow Wt→Wt) mice. Aortic LpL deficiency (HF MCKL0→MCKL0) reduced macrophage accumulation by >40% when compared with HF Wt→Wt. Reconstitution of Wt BM into MCKL0 mice (HF Wt→MCKL0) significantly increased aortic macrophages (p<0.05), and the levels of LpL and macrophages were comparable to those in Wt mice (Figure 6C). Quantification of macrophages in different aortic layers show that LpL deletion led to reduced macrophage counts primarily in the aortic media (Figure 6D). Aortic F4/80, CD/68, and chemoattractant MCP-1 mRNA levels were decreased in MCKL0→MCKL0 mice compared to Wt→MCKL0 (Supplemental Table I).

Figure 6.

Loss of LpL reduces macrophage levels in the arterial wall. (A) Representative confocal images and (B) quantification of macrophage (MΦ) immunofluorescence of Wt and MCKL0 mouse artery sections stained for EC and MΦ (n=4). *, p<0.05, vs. Wt. (C) Representative confocal images and (D) immunofluorescence (arbitrary units, AU) of MΦs in specific aortic layers quantitated for chow-fed Wt→Wt, HF Wt→Wt, MCKL0→MCKL0 and Wt→MCKL0 mouse artery sections stained for EC and MΦ (n=4–5). Scale bar: 50 μm.*, vs. intima; #, vs. chow Wt→Wt; †, vs. HF Wt→Wt; ‡, MCKL0→MCKL0 vs. Wt→MCKL0, p<0.05.

Therefore, our data in the aorta indicates that effects of LpL on blood monocyte levels and properties can contribute to modulations in immune cell responses in an important tissue involved in CVD, the artery.

DISCUSSION

The role of LpL in vascular disease is unclear, as macrophage-specific LpL deficiency reduces atherosclerotic CVD in animal models but partial LpL deficiency in humans increases CVD risk ^{11, 13, 25, 35}. These data suggest that the role of macrophage-LpL in vascular pathology is cell-specific. Indeed, we found that loss of LpL reduced the response of BM-derived macrophages to M-CSF. We went on to show that this is due to changes in both expression of M-CSF and in M-CSF interaction with monocyte and macrophage precursor cells. Aside from its action as a lipolytic enzyme, LpL has a number of anchoring functions that increase lipoprotein association with extracellular matrix ¹⁷ and attachment of cells to endothelial surfaces ²¹. We now provide evidence for a new and unique action of LpL that captures and promotes M-CSF activities. In addition, while investigating the effects of loss of LpL in arteries with heparin, we discovered that heparin also reduces arterial macrophage content, likely by inhibiting LpL-mediated macrophage retention in the aortic wall.

Although LpL is highly expressed in BM, few studies have assessed the relationship between LpL and BM hematopoiesis or blood leukocyte parameters. Our findings indicate that transgenic (MCKL0) LpL loss-of-function mouse models are associated with substantial reductions in blood leukocytes and, specifically, monocytes. We found that LpL deficiency led to reduced expression of master regulatory transcription factors (PU.1 and C/EBPα) and growth factors (CSFs and CSF-Rs) critical for BM myeloid cell proliferation and differentiation. Although the total numbers of lineage negative progenitors cells were reduced in MCKL0 mice (supplemental Figure I b), the reduced myeloid cell differentiation was only observed at the later stage of hematopoiesis—monocyte progenitors to monocytes or monocytes to macrophages. Further studies are needed to investigate whether LpL deficiency also alters BM myeloid cell proliferation.

We suggest that the observed decreases in blood monocytes and tissue macrophages in MCKL0 mice result from lower levels of CSFs, such as M-CSF, which catalyze the differentiation of both monocyte progenitors and monocytes. M-CSF is produced in a variety of tissues with expression levels and functions depending on local microenvironmental stimuli. We suggest that the density of macrophages in particular tissues is controlled by local synthesis of membrane-associated M-CSF and/or selective sequestration of the secreted PG M-CSF. For example, the Stanley group has reported that a supplement of recombinant human M-CSF restored macrophage density in some tissues, such as liver or spleen, in CSF-deficient osteopetrotic mice but was unable to completely correct phenotypic defects³³. Transgenic expression of membrane-bound CSF alone in these mice restored macrophage numbers in some tissues but did not rescue the loss of overall BM hematopoiesis, implicating the requirement of circulating M-CSF and/or PG M-CSF for BM hematopoiesis. In our MCKL0 mouse model, isolated BM had reduced M-CSF expression levels, presumably representative of membrane-bound M-CSF. When MCKL0-derived BM monocytes were stimulated with exogenous M-CSF (equivalent to secreted M-CSF), these cells differentiated to macrophages at a substantially slower rate than those isolated from Wt mice (Figures 4A and 4B), suggesting impaired M-CSF action in these LpL-deficient BM-derived monocytes. In fact, while LpL deficiency reduced BM M-CSF protein levels (ns), plasma M-CSF levels were similar between WT and MCKL0 mice (supplemental Figure III). However, this lack of response may also be attributable to the reduced expression of M-CSF-R in both plasma (~ 46% reduction, p<0.05) and BM cells (Figure 3B). Our preliminary studies show that the addition of exogenous LpL has little or no effects on BMMP cell differentiation (data not shown).

Transcriptional control of M-CSF-R has been extensively studied ³⁶. Master regulators of hematopoiesis, such as PU.1, C/EBPα and NFκB, control M-CSF-R expression. Inflammatory stimuli, such as endotoxin LPS, also induce M-CSF-R expression. On the other hand, transcriptional control of M-CSF is still not understood. Our data show that, although exogenous LpL slightly increased M-CSF expression in MCKL0 BM myeloid precursor cells, there were no effects on M-CSF-R expression. Can increased cellular FA uptake stimulate M-CSF or M-CSF-R expression in vivo? It has been shown that peripheral monocyte overexpression of scavenger receptor type A (SR-A) and CD36, which enhance FA uptake, positively correlated with serum M-CSF levels in humans ³⁷. Therefore, when BM-derived myeloid precursor cells were treated with FAs, LA increased M-CSF expression. Thus, we suggest that LpL is important for M-CSF expression and required for M-CSF cellular interactions that cooperatively regulate monocyte differentiation to macrophages. On the other hand, the PPARs, which modulate lipid metabolism, cellular proliferation and inflammatory responses, are widely expressed in many tissues, including in monocytes. Some FAs are direct ligands of PPARs ³⁸ and important regulators of the expression of different PPARs ³⁹. Among the PPAR family, PPARγ is a negative regulator of M-CSF expression and production through transrepression of NFκB ⁴⁰. PPARγ overexpression and treatment with its ligand, pioglitazone, decreased M-CSF gene expression and protein levels ⁴⁰. Our data show a consistent inverse relation between PPARγ and M-CSF in BMMP cells; expression of M-CSF was downregulated, while PPARγ was upregulated in MCKL0 mice. The addition of exogenous LpL and FA during BMMP differentiation increased PPARγ expression, which linked to a decrease in M-CSF expression in both Wt and MCKL0 mice. Our results are consistent with our earlier report ²⁷ that MCKL0 mice had greater increases in expression of all PPARs in adipose tissues. It has been postulated that increased PPAR expression is induced by FAs through de novo lipogenesis ^{27, 41}.

Because there are several positively charged binding motifs in LpL, we hypothesize that LpL anchors M-CSF by binding to the negatively charged residues on M-CSF. Studies from the Chait group show that macrophage-secreted M-CSF contains PG domains that can bind lipoproteins, such as LDL ⁴². Acetylation or copper oxidation, which instill a net positive charge upon LDL, resulted in a loss of M-CSF binding ⁴². Based on the known ability of LpL to bind PGs or LDL, we propose that LpL sequesters secreted circulating M-CSF at the cell surface and directs it to the M-CSF-R. In fact, our data on assessing whether exogenous LpL affects binding of M-CSF to its receptor by immunobinding assays show that M-CSF association with M-CSF-R decreased with increasing amounts of LpL in a solid-phase assay (supplemental Figure IV). In the absence of LpL, cellular association of M-CSF and M-CSF-R was significantly reduced. We have examined potential interacting factors that may affect M-CSF and M-CSF-R cellular interaction. The results indicate that M-CSF and M-CSF-R interaction is independent of LpL enzymatic activity. Consistent with the negative results from the immunobinding assay (data not shown), the addition of heparin has little or no effects on cell association of M-CSF and M-CSF-R in BM. On the other hand, cellular association of M-CSF and M-CSF-R was significantly reduced by NaClO₃, which supports our hypothesis that LpL plays a role in bridging BM M-CSF proteoglycan to cells to catalyze the maturation of monocyte precursor or monocytes. We have previously reported that LpL bound to cell-surface PGs also bound LDL and facilitated LDL delivery to the LDL receptor ⁴³. Thus, similar to lipoproteins with anchoring functions, dimeric LpL might capture circulating M-CSF and increase its proximity to its receptor. Lack of LpL thus impairs the efficacy of M-CSF in stimulating myelopoiesis.

Decreased myelopoiesis induced by selective LpL deletion may be beneficial in some scenarios. The M-CSF/M-CSF-R signaling pathway is activated via the concerted action of other growth factors, such as GM-CSF/GM-CSF-R. While M-CSF expression is generally constant, GM-CSF expression increases in response to inflammatory stimuli ⁴⁴. Myeloid lineage populations display differential responses to M-CSF and GM-CSF, albeit through unknown mechanisms. Interestingly, therapeutic antagonism of M-CSF and GM-CSF has been shown to ameliorate some inflammatory/autoimmune diseases, such as rheumatoid arthritis ⁴⁴.

Atherosclerosis-prone disease models, such as apoE- or LDL receptor (LDLR)-knockout (KO) mice, display reduced atherosclerosis if macrophage-derived LpL is deleted. ApoE-KO mice lacking LpL expression in macrophages after cre/loxP gene targeting had fewer atherosclerotic lesions than apoE-KO with LpL-expressing macrophages after a HF, high-cholesterol diet ¹³. Further, Babaev et al. reported that irradiated LDLR−/− mice transplanted with fetal liver cells from LpL-deficient mice had smaller atherosclerotic lesions when compared to LDLR−/− mice reconstituted with LpL-expressing fetal liver cells ³⁵. In our models, LpL deficiency not only reduced aortic macrophage levels, but also reduced monocyte counts substantially. This makes it difficult to determine whether the reduction of atherosclerotic lesions observed in other studies is due to the lack of local LpL expression in arterial MΦs or due to changes in altered BM and circulating leukocyte profiles.

LpL deficiency in the arterial wall is associated with fewer aortic macrophages even prior to atherosclerotic development ¹⁴. Associations between LpL and aortic macrophages were examined through BM transplantation techniques that restore LpL to arterial cells, and the use of heparin, which mediates dissociation of LpL from the cell surface. We expect that heparin would further decrease aortic macrophages in LpL-deficient MCKL0 mice; however, the decreases would be difficult to detect since the levels of aortic macrophage in MCKL0 mice are already low. Heparin has been proposed to have anti-inflammatory effects through its binding of various chemokines and cytokines ⁴⁵, prevention of pro-inflammatory mediators from interacting with their respective receptors ^{46, 47}, and modulation of TNFα and NFκB activity ^48–50. Although the ability of heparin to induce LpL cellular dissociation is well established, the phenomenon of “heparin-releasable macrophages” in an acute setting has not been previously reported. The interaction of LpL with heparin, and likely a number of other molecules, is due to regions of positively charged heparin binding motifs in LpL. Others have reported that heparin interferes with the adhesion of leukocytes to the endothelium through its binding to L- and P-selectin ⁵¹. In contrast to other studies, we report an acute effect (~10 min) of heparin on dissociation of arterial macrophages and these effects on macrophage levels are correlated with macrophage-associated heparin-releasable LpL. Further, we have established the close relation between LpL and macrophages in atherogenesis in LDLR−/− mice in our earlier reports ¹⁵. We speculate that the heparin would also affect arterial macrophage abundance in LDLR−/− mice. Nonetheless, the effects of heparin are complex in modulating function and activity of a number of proteins that are important in regulating inflammatory responses acutely and chronically. This is one limitation of the current studies.

In summary, LpL deficiency decreased circulating levels of total leukocytes, neutrophils, and total and inflammatory monocytes. These effects were partly mediated through CSFs and CSF-Rs, resulting in significant decreases in blood monocytes and BM-derived monocyte differentiation to macrophages. Loss of LpL in the arterial wall was also associated with reduced aortic macrophages in non-atherosclerotic mice. LpL can now be recognized as not only a key player in directing arterial lipid deposition, but also as a potential regulator of blood and peripheral inflammatory responses.

HIGHLIGHTS.

• Loss of LpL reduces circulating WBC and monocyte levels.

• Loss of LpL decreases bone marrow expression of master regulators includes PU.1, C/EBPα and colony-stimulating factors and their receptors that are essential for myeloid cell differentiation.

• LpL deficiency markedly reduces aortic macrophage counts. Reconstitution of LpL-expressing bone marrow cells was able to replenish aortic macrophage density.

• We propose that LpL, in addition to being an important molecule for triglyceride metabolism and anchoring of arterial LDL and other lipoproteins, also contributes to regulation of blood monocyte levels and monocyte differentiation to macrophages.

• Our findings provide novel insights into the regulation of inflammatory responses and may have therapeutic applications for a number of metabolic and blood cell diseases.

ABBREVIATIONS

BM

bone marrow

CSF

colony-stimulating factor

LpL

lipoprotein lipase

WBC

white blood cell