We describe a protocol to measure transmigration by monocytes across human endothelial monolayers and their subsequent maturation into foam cells. This provides a versatile method to assess the atherogenic properties of monocytes isolated from people with different disease conditions and to evaluate factors in blood which may enhance this propensity.
Coronary artery disease (CAD) is a leading cause of morbidity and mortality worldwide. Atherosclerosis, a leading cause of CAD, is initiated by the transmigration of innate immune monocytes to inflammatory sites of deposited lipid called fatty streaks, which are present in arterial walls of medium to large arteries. The key pathogenic feature of lesions at this early stage of atherosclerosis is the maturation of monocytes which migrate into arteries to form foam cells or lipid-laden macrophages. Considerable evidence supports the hypothesis that risk of atherosclerosis is increased by chronic inflammatory conditions accompanying diseases such as rheumatoid arthritis and HIV, as well as general ageing, and that this risk is predicted by monocyte activation. While mouse models provide a good platform to investigate the role of monocytes in atherogenesis in vivo, they require genetic alteration of natural cholesterol metabolism and drastic alteration of normal mouse diets, and have limited suitability for the study of atherogenic influences of human comorbid diseases. This motivated us to develop a human in vitro model to measure the atherogenic potential of monocytes isolated from individuals with defined disease states. Currently, human in vitro models are limiting in that they evaluate monocyte transmigration and foam cell formation in isolation. Here we describe a protocol in which monocytes isolated from patient blood transmigrate across human endothelial cells into a type 1 collagen matrix, and their propensity to mature into foam cells in the presence or absence of exogenous lipid is measured. The protocol has been validated for the use of human monocytes purified from individuals with HIV infection and elderly HIV uninfected individuals. This model is versatile and allows monocyte transmigration and foam cell formation to be evaluated using either microscopy or flow cytometry as well as allowing the assessment of atherogenic factors present in serum or plasma.
Monocyte transmigration is a crucial step in the development of atherosclerotic plaque that may lead to thrombosis, stroke and myocardial infarction. Atherosclerotic plaques develop from fatty streaks, generally present at sites of low oscillatory blood flow in medium to large arteries, where deposited lipid contributes to endothelial activation and localized inflammation1. Monocytes are recruited to endothelial cells in fatty streaks via monocyte chemotactic proteins (such as CCL2) and transmigrate into the intima2. Following transmigration, monocytes may form atherogenic, lipid-laden macrophages called foam cells as a consequence of lipid uptake, lipid synthesis, down-regulation of cholesterol efflux or a combination of the above factors. Monocytes may also accumulate lipids in the circulation and have a 'foamy' phenotype, possibly predisposing cells for foam cell formation3,4. Foam cells are the defining feature of fatty streaks and early-stage atherosclerotic plaques and their formation is influenced by both lipid and inflammatory mediators5. Alternatively, monocytes have the ability to reverse transmigrate from the artery into the bloodstream6, thereby removing lipid from the intima and acting to maintain the health of the artery.
Determining the propensity of monocytes to transmigrate across arterial endothelium and form foam cells in the intima, or to reverse transmigrate and carry lipid out of the plaque, is a key requirement for understanding the role of monocyte activation in increasing atherosclerotic risk. Mouse models of CADs such as atherosclerosis are important in elucidating real-time in vivo information on fatty streak/atherosclerotic plaque development. However, these models require a genetic alteration of the natural cholesterol processing abilities of these animals usually coupled with drastic alterations in diet (such as the ApoE-/- Western-type diet model)7,8, thereby, inducing non-physiological accumulation of circulating lipid levels which drive plaque development. These models may have limited relevance to chronic inflammatory human conditions such as HIV infection which are not associated with increased circulating cholesterol or low-density lipoprotein (LDL) levels. Furthermore, differences in monocyte biology between humans and mice make the testing of immunological questions regarding the relevance of subpopulations of monocytes (such as intermediate monocytes (CD14++CD16+))9 difficult. This is important when studying the mechanisms driving cardiovascular disease as intermediate monocyte counts independently predict cardiovascular events10,11. While assays exist to sequentially measure either monocyte transmigration or foam cell formation in isolation, no in vitro assay has been validated for quantifying both aspects of early atherogenesis using the same cells from clinical cohorts. Transwell models utilize a modified Boyden two-chamber system whereby cells are loaded into the top chamber and transmigrate across a porous plastic barrier or cell monolayer into a lower chamber that typically contains media with chemoattractant12,13. Whilst widely used for analyzing leukocyte transmigration, these models do not generally incorporate a layer representing the intima, resulting in transmigrated cells migrating into solution, and do not allow for the measurement of foam cell formation or reverse transmigration of the same cells. Conversely, models of foam cell formation do not account for any transmigratory-induced changes to monocytes or effects of endothelial activation which is known to contribute to foam cell formation14. Furthermore, these systems induce foam cell formation from macrophages adhered to cell culture plates by the addition of saturating concentrations of exogenous oxidized low-density lipoprotein (oxLDL)15,16, a key inducer of foam cell formation. LDL used in these models is often oxidized by non-physiologically-relevant processes such as CuSO4 treatment17, therefore, questioning the physiological importance of studies using these models.
Here we describe an assay that quantifies monocyte transmigration and foam cell formation of the same cells which does not require the addition of exogenous oxLDL, thus better modelling the role of monocytes in foam cell formation. This model was originally developed by Professor William Muller (Northwestern University, Chicago)18, and has been further refined in our laboratory to assess ex vivo the atherogenicity of monocytes isolated under non-activating conditions from individuals with underlying inflammatory conditions accompanying diseases such as HIV infection19 as well as ageing20, that are associated with an increased risk of atherosclerosis. This model also provides a platform for answering basic biological questions regarding the propensity of different monocyte subsets to form foam cells20, the influence of endothelial activation by cytokines such as TNF on foam cell formation14, and the migratory properties of monocytes such as the depth and speed of transmigration in gels19. Furthermore, monocyte transmigration and foam cell formation can be quantified using standard microscopy, live cell imaging, flow cytometry and imaging flow cytometry, therefore, providing a versatile method to evaluate the role of monocytes in atherogenesis.
NOTE: All experiments using human biological samples were performed with ethics approval from the Alfred Hospital Human Ethics Committee, Melbourne. All experiments were performed in Class II Biosafety cabinets unless specified. "Prewarmed" refers to reagents warmed to 37 °C in a waterbath.
1. Preparation of Type I Fibrous Collagen Gels: Day 1
2. Expansion of Stored HUVEC: Day 1
3. Culturing HUVEC Monolayer on Collagen Gels: Day 5
4. Activation of HUVEC Monolayer and Isolation/Activation of Monocytes for Transmigration: Day 8
5. Transmigration of Primary Human Monocytes: Day 8
6. Quantitation of Foam Cells and Macrophages by Microscopy (Oil-Red O Stain)
7. Analysis of Transmigrated Cells by Flow Cytometry
Quantifying monocyte transmigration
Monocytes are added to the model as described in Figure 1, and six gels are prepared for each condition. Monocytes for 6 gels per donor (i.e., 5.0 x 104 monocytes per gel 6 gels = 3.0 x 105 monocytes per donor) are resuspended to a final volume of 600 µL of M199 media containing the required serum/isolated lipid. Cells (i.e, 100 µL per gel = 5.0 x 104 monocytes) are then aliquoted onto gels and incubated as above for 1 h to facilitate monocyte transmigration. Following transmigration, non-transmigrated cells are counted as above. In this example, 1.5 x 104 non-transmigrated cells were recovered. The percentage of forward transmigration was determined using the formulas described in step 5.4:
Following 48 h of incubation, 3.6 x 104 reverse transmigrated cells were recovered as above. The percentage of reverse transmigration was determined using the formula in step 5.8:
Forward and reverse transmigration may be compared between conditions using the appropriate statistical tests and monocytes prepared from multiple independent donors to determine whether the monocyte transmigration ability is altered between experimental conditions.
Evaluating foam cell formation
Brightfield microscopy
Oxidized lipoproteins are known to promote monocyte-derived foam cell formation in vitro in comparison to unoxidized lipids15. Here we confirm that the treatment of monocytes with oxLDL, commonly used to induce foam cell formation in conventional foam cell induction models, enhances monocyte-derived foam cell formation in this model of transendothelial migration and foam cell formation in comparison to native LDL. Following monocyte transmigration and foam cell formation in gels incubated with M199 containing 50 µg/mL native or CuSO4-oxidized LDL (see step 5.1, Figure 1), cells were analyzed by microscopy. Examples of foam cells and macrophages observed in representative experiments are shown in Figure 2. Cells were counted in three different fields per gel, by focusing at the HUVEC monolayer and progressively focusing deeper into the gel. The cell counts are recorded for each donor as shown in Table 2 for a single donor, to compare the percentage of foam cells formed in response to oxidized and native LDL. Aggregate data from n = 12 independent counts are shown in Figure 4 to confirm that incubation of monocytes with oxLDL in this model induces more foam cell formation than conditions where monocytes are incubated with native LDL or M199 media alone.
Flow cytometry
To determine the phenotype of migrated cells, the cells are extracted from digested gels and labeled using a standard flow cytometry panel. Cells were labeled with a live/dead cell marker and antibodies specific for CD45 and other surface phenotype markers using standard flow cytometry protocols. Compensation controls must be also prepared for full compensation as per standard flow cytometry techniques. Migrated cells are gated as shown in Figure 3 in plots representative of 10 independent experiments. Live cells are gated using live/dead cell viability assays (Figure 3B) and doublets are excluded by single cell analysis (Figure 3C). Cells are then gated as CD45+ in order to exclude HUVEC (as these cells are CD45–, Figure 3D). The major population is then gated by forward scatter/side scatter discrimination (Figure 3E) and the mean fluorescence intensity (MFI) of receptors of interest are determined in comparison to either fluorescence minus one (FMO) or isotype control (Figure 3F) to determine expression (ΔMFI) or percentage of positive cells.
Figure 1: An in vitro model of monocyte transmigration and foam cell formation. (A) PBMC, (B) total monocytes or (C) FACS-sorted monocyte subsets are added to type I fibrous collagen gels formed in 96-well plates and overlaid with a monolayer of activated primary human umbilical vein endothelial cells (HUVEC), whose integrity may be assessed by silver staining (D) and allowed to (E) transmigrate for 1 h in the presence of serum or lipid containing media. Non-transmigrated cells are counted (F) and gels are incubated for a further 48 h with the same serum/lipid containing media. Following incubation, (G) reverse migrated cells are counted and the percentage of foam cells vs. macrophages in the gel is determined by (H) phase contrast microscopy following staining with Oil-Red O or the phenotype of migrated cells is determined by (I) flow cytometry. Please click here to view a larger version of this figure.
Figure 2: Examples of foam cells and macrophages in an in vitro model of monocyte transmigration and foam cell formation. Representative examples of cells scored as foam cells (white arrows) and macrophages (black arrows) after Giemsa staining to visualize cells and Oil-Red O staining to visualize lipid droplets as determined by phase contrast microscopy of extracted gels using an inverted microscope at 40X magnification. Scale bar = 20 µm. Please click here to view a larger version of this figure.
Figure 3: Gating strategy of transmigrated cells by flow cytometry. Transmigrated cells are phenotypically characterized by flow cytometry following extraction of cells from gels. Cells are gated by (A) FSC/SSC, (B) live cells, (C) single cells, (D) CD45+, (E) FSC/SSC and (F) expression of receptor compared to isotype or fluorescence minus one (FMO) control. Expression of receptors of interest are defined as mean fluorescence intensity (MFI) and expressed with respect to levels of isotype. ΔMFI = Stain (MFI) minus isotype/FMO control (MFI). Please click here to view a larger version of this figure.
Figure 4: Effect of exogenously added oxLDL and LDL on monocyte-derived foam cell formation. Foam cell formation is determined under conditions where monocytes from a single donor are incubated with media (M199) or 50 µg/mL unoxidized low-density lipoprotein (LDL) or copper (II) sulphate oxidized LDL (oxLDL) added to gels post-transmigration. Counts for 12 distinct sites are shown. Foam cells are expressed as the percentage of total migrated cells. Median and interquartile ranges are shown in bar graphs and comparisons were made using non-parametric Mann-Whitney U tests. ***p <0.001, ****p <0.0001. Please click here to view a larger version of this figure.
Reagent | [Stock] | [Final] | Volume for 60 gels (µL) |
NaOH | 100 mM | 35.7 mM | 1071 |
M199 | 10x | 0.71x | 213 |
AcCOOH | 20 mM | 4.58 mM | 687 |
Collagen | 5 mg/mL | 1.71 mg/mL | 1029 |
Total volume | – | – | 3000 |
Table 1: Type I fibrous collagen gel preparation
LDL | oxLDL | |||||||
Gel | Count | Foam cells1 | Migrated cells1 | Foam cells (%)2 | Foam cells1 | Migrated cells1 | Foam cells (%)2 | |
1 | 1 | 5 | 44 | 11.4 | 26 | 55 | 47.3 | |
2 | 5 | 27 | 18.5 | 10 | 23 | 43.5 | ||
3 | 11 | 52 | 21.2 | 19 | 39 | 48.7 | ||
2 | 1 | 5 | 34 | 14.7 | 9 | 31 | 29 | |
2 | 14 | 50 | 28 | 32 | 55 | 58.2 | ||
3 | 5 | 37 | 13.5 | 19 | 37 | 51.4 | ||
3 | 1 | 10 | 37 | 27 | 28 | 52 | 53.8 | |
2 | 7 | 33 | 21.2 | 11 | 31 | 35.5 | ||
3 | 7 | 38 | 18.4 | 28 | 53 | 52.8 | ||
4 | 1 | 9 | 44 | 20.5 | 14 | 33 | 42.4 | |
2 | 7 | 33 | 21.2 | 22 | 47 | 46.8 | ||
3 | 11 | 50 | 22 | 11 | 30 | 36.7 | ||
Median foam cells (%) | 20.8 | 47 | ||||||
Average foam cells (%) | 19.8 | 45.5 |
Table 2: Raw cell counts from comparison of monocyte-derived foam cells following LDL vs oxLDL treatments. 1Cell counts per field of view in gel. 2Percentage of foam cells = foam cell counts/total migrated cell counts x 100. *Data are representative of one experiment using monocytes from a single donor
The protocol described here offers a versatile and physiologically relevant method for assessing the atherogenicity of monocytes from human clinical cohorts, by combining both monocyte transmigration and foam cell formation. This model offers advantages over alternative methods of foam cell formation as it takes into account the effect of monocyte transmigration on foam cell formation and allows the measurement of reverse transmigration6 in addition to the inherent propensity of monocytes to mature into foam cells in the presence or absence of exogenous factors. Furthermore, the role of endothelial activation is also taken into account as we have shown that endothelial activation upregulates oxidation of lipid species that is associated with foam cell formation14. Finally, the phenotype of monocytes that undergo egress (a property associated with plaque regression) can be quantified and characterized by microscopy and flow cytometry.
Due to the three-dimensional nature of transmigration into gels, identifying and scoring the transmigrated cells as foam cells or macrophages can be difficult as cells transmigrate to different levels in the gel19. To aid the identification of foam cells, fresh Oil-Red O stain must always be used. It is also of note that different ex vivo disease states or in vitro conditions may alter monocyte transmigration in the gel. We have used live cell imaging to identify that monocytes from HIV-infected individuals tend to transmigrate to shallower points in gels and at different speeds than those from uninfected individuals19. Furthermore, foam cell formation is associated with less monocyte motility in the gel so cells tend to move in circles at a particular z-section in comparison to cells from HIV-uninfected individuals which tend to migrate in a relatively straight line towards the bottom of the gel. These findings raise the possibility of a similar phenomenon occurring in experiments using human patient samples from other disease states associated with increased monocyte atherogenicity.
When performing flow cytometry-based analyses it is common to observe a significant proportion of dead cells in the recovered cell population when assessing viability using live/dead cell markers (Figure 4B). These cells are predominantly CD45– HUVEC that are damaged during the collagen digestion/maceration step required to extract monocyte-derived cells from the collagen matrix. This process is particularly harsh on the HUVEC monolayer, but is not detrimental to phenotyping the transmigrated cells.
This model differs from others in that no artificially oxidized exogenous lipid is required in the culture media to drive monocyte-derived foam cell formation. Instead, foam cell formation in this assay is influenced by the serum present in the culture media, allowing for the discrete effects of soluble factors from clinical samples to be evaluated. As such, we have shown that incubation of control monocytes with serum from either young HIV-infected19 or elderly HIV-uninfected20 individuals promotes foam cell formation, indicating that soluble factors can independently promote foam cell formation. As the assay is influenced by soluble components, a single batch of pooled human serum must be used for experiments aimed at determining the inherent difference of atherogenic properties of monocytes derived from different individuals in order to standardize control conditions. This assay may also be used to evaluate the role of specific lipid species from clinical samples (Figure 3). For example, we have found that incubation of monocytes with media containing isolated lipid species such as high-density lipoproteins from individuals with known lipid dysfunction24 also promotes foam cell formation. In this case, monocytes from a single healthy individual or group of individuals can be used in the presence of serum or serum factors derived from test subjects. Therefore, this model allows for specific questions to be asked regarding the discrete effects of both soluble and cellular components on foam cell formation.
This assay utilizes HUVEC as a model of coronary artery endothelium due to the practical difficulty in obtaining large numbers of low passage number primary coronary artery cells. However, we have compared outcomes from assays using both human coronary artery endothelial cells or HUVEC and observed little difference in the monocyte transmigration or foam cell formation (data not shown), indicating that the use of HUVECs in this system is an acceptable substitute. Manual counting of foam cells is a limiting factor in this model as it may introduce operator bias. Therefore, all samples mounted on slides must be blinded to the operator in order to remove the potential of bias. We have, however, compared foam cell formation as measured by manual counting and by imaging flow cytometry and found that these methods give similar results19. A key limitation of this model is that monocyte adhesion and transmigration occur in the absence of shear forces associated with physiological blood flow in vivo. We hypothesize that shear flow will predominantly affect monocyte transmigration and not subsequent foam cell formation within the matrix; however, this must be considered when interpreting results from this assay. This assay also does not model the influence of smooth muscle cells in monocyte-derived foam cell formation which is a physiological limitation of the system; however, this is consistent in all conditions allowing the discrete atherogenic potential of monocytes to be compared between different disease states.
In summary, this model provides a versatile and physiologically relevant method for quantifying monocyte transmigration and foam cell formation from human samples in different disease states ex vivo. This model has further applications for evaluating the propensity of monocytes to form foam cells in conditions where monocyte atherogenicity is associated with increased risk of CAD such as obesity25, diabetes26 and chronic kidney disease27. Therefore, this model may also be optimized for use in disease states other than atherosclerosis where cellular transmigration may influence disease pathogenesis.
The authors have nothing to disclose.
The authors gratefully acknowledge the work of Prof. William Muller and Dr. Clare Westhorpe for their key role in development of earlier iterations of this model. The authors would also like to thank the AMREP Flow Cytometry core for the sorting of monocyte subsets and the Alfred Hospital Infectious Disease Unit clinical research nurses for the recruitment of HIV+ individuals for some studies. The authors gratefully acknowledge the contribution to this work of the Victoria Operational Infrastructure Support Program received by the Burnet Institute. TAA is supported by an RMIT University Vice-Chancellor's Postdoctoral Fellowship. This work was supported by NHMRC project grant 1108792 awarded to AJ and AH. TK is supported by NIH grants NIH K08AI08272, NIH/NCATS Grant # UL1TR000124.
Gel preparation reagents | |||
NaOH | Sigma-Aldrich | 221465-500G | 0.1 M NaOH diluted in H20 |
10X M199 | Sigma-Aldrich | M0650 | |
AcCOOH | Sigma-Aldrich | 695092-100ML | 20 mM Acetic acid diluted in H20 |
Cultrex Bovine Collagen I | R&D Systems | 3442-050-01 | Type I Fibrous Collagen |
Name | Company | Catalog Number | Yorumlar |
Cell culture | |||
M199 | Life Technologies | 11150-059 | M199 media containing Earle's salts, L-glutamine and 2.2 g/L Sodium Bicarbonate. Media supplemented with 100 µg/mL L-glutamine and 100 U/mL penicillin/streptomycin |
M20 | Supplemented M199 containing 20% heat-inactivated pooled or donor serum. Individual lipid species such as LDL can be added to M199. |
||
HUVEC | Primary human umbilical cord endothelial cells (HUVEC) can be isolated from umbilical cords donated with informed consent and ethics approval. Isolated HUVEC may also be purchased commercially. | ||
Human coronary artery endothelial cells | Primary human coronary artery endothelial cells can be isolated from arteries donated with informed consent and ethics approval. Isolated cells may also be purchased commercially. | ||
EDTA | BDH Merck | 10093.5V | Ethylenediaminetetraacetic acid (EDTA) – 0.5 M, pH 8.0 |
EGTA | Sigma-Aldrich | E3889-500G | Ethylene glycol-bis(β-aminoethyl ether)-N,N,N',N'-tetraacetic acid (EGTA) – 1 mM, pH 8.0 |
0.05% trypsin EDTA | Gibco | 25300-054 | 0.05% trypsin/0.53 EDTA (1X) |
L-glutamine | Gibco | 25030-081 | L-glutamine (200 mM) |
Penicillin/streptomycin | Gibco | 15140-122 | Penicillin/streptomycin (10,000 units/mL Penicillin and 10,000 µg/mL Streptomycin) |
New born calf serum | Gibco | 16010-142 | New born calf serum: New Zealand origin |
Fibronectin | Sigma-Aldrich | F1056-1MG | Fibronectin – 50 µg/mL aliquots prepared and stored |
PBS (1X) | Gibco | 14200-075 | Dulbecco's Phosphate Buffered Saline (10X): Dilute to 1X with sterile H20 |
0.1% TNF | Gibco | PHC3015 | Recombinant human TNF – Reconstituted in H20 and stored in 10 µg/mL aliquots |
Low-density lipoprotein | Merck Millipore | LP2-2MG | Low-density lipoprotein (LDL) |
Name | Company | Catalog Number | Yorumlar |
Microscopy | |||
1 or 2% formaldehyde | Polysciences | 4018 | 1 or 2% formaldehyde diluted with sterile H20 |
50% and 78% methanol | Ajax Finechem | 318-2.5L GL | 50% or 78% v/v methanol, diluted with H20 |
Oil Red O stain | Sigma-Aldrich | O0625-25G | Dilute to 2 mg/mL in 22% 1M NaOH and 78% (v/v) methanol |
Microscope slides | Mikro-Glass | S41104AMK | Twin frosted 45 degree ground edge microscope slides (25 X 76 mm) |
Cover slips | Menzel-Gläser | MENCS224015GP | 22 x 40 mm #1.5 size glass cover slips |
Double-sided tape | 3M Scotch | 4011 | Super strength exterior mounting tape (25.4 mm x 1.51 m) |
Giemsa stain | Merck Millipore | 1.09204.0500 | Giemsa's azur eosin methylene blue solution (dilute stock 1:10 in H20) |
Hole punch | Hand-held single hole punch (6.35 mm punch) | ||
Name | Company | Catalog Number | Yorumlar |
Flow cytometry | |||
Collagenase D | Roche Diagnostics | 11088858001 | Collagenase D diluted in M199 media to 1 mg/mL |
35 µm nylon mesh capped polystyrene FACS tubes | BD Biosciences | 352235 | 35 µm nylon mesh capped polystyrene FACS tubes |
Live/Dead Fixable Yellow Dead Cell stain | Life Technologies | L34959 | Live/Dead Fixable Yellow Dead Cell stain |
FACS wash | Prepare by mixing 1 X PBS–, 2 mM EDTA and 1% New born calf serum | ||
Name | Company | Catalog Number | Yorumlar |
Plasticware | |||
96 well plate | Nunclon | 167008 | Delta surface flat-bottomed 96 well plate |
10 cm Petri Dish | TPP | 93100 | Sterile 10 cm Petri dish |
1.5 mL Eppendorf tubes | Eppendorf | 0030 125.150 | Eppendorf tubes |
Transfer pipette | Samco Scientific | 222-20S | Sterile transfer pipette (1 mL, large bulb) |