A human cell model for dynamic testing of MR contrast agents

Diagnosis and intervention in the first stages of atherosclerosis could potentially offset economic and clinical costs of cardiovascular disease; however, there are no effective methods currently available for the detection of fatty streaks or early plaques. MR is an imaging modality suitable for use as a screening tool for early atherosclerosis (1). In principle, it should be possible to achieve useful images by employing MR contrast agents that promote signal enhancement in early lesions compared with the background obtained from surrounding healthy tissue (2–4). A physiologically relevant model system for contrast agent testing would facilitate the development of screening methods in a manner that can target eventual translation to clinical protocols.

Inflammation plays a key role in the progression of atherosclerotic diseases (5); thus, targeting components of the vascular inflammatory response for imaging is an interesting approach for development of noninvasive methods to detect early atherosclerosis. This report presents the results of an initial study in which a commercially available positive MR contrast agent was used to label human monocytes in a series of cell-based assays to determine the feasibility of detecting early atherosclerosis using labeled inflammatory cells. Results demonstrate proof-of-concept for this approach, and also illustrate the utility of employing an in vitro culture system to support the development and evaluation of novel MR contrast agents.

Materials and methods

Cell culture

The human acute monocytic leukemia cell line THP-1 (ATCC; Manassas VA, USA) and pooled human umbilical vein endothelial cells (HUVECs; Lonza, Walkersville MD, USA) were cultured according to the suppliers' directions. Both cell types are readily available, easy to culture, and useful for studying early atherosclerotic processes in vitro. Notably, THP-1 cells possess the ability to differentiate into macrophages (6), the inflammatory cells found in atherosclerotic plaques. HUVECs are commonly used to study endothelial physiology in vitro (7,8), including interactions with monocytes under flow conditions (9,10).

HUVECs were plated at ∼3500 cells/cm² on glass microscope slides precoated with 0.1% gelatin (BD Difco, Mississauga, ON, Canada) and grown to confluence prior to undertaking the monocyte contact experiments described.

Cell Gd labeling

THP-1 cells, at an initial density of 7 × 10⁵ cells/mL, were incubated for 17 h at 37°C and 5% CO₂ with the gadolinium (Gd) colloid Gado CELLTrack (BioPAL, Worcester, MA, USA) at a Gd concentration of 0.08 mg/mL in Roswell Park Memorial Institute (RPMI) medium (ATCC). Cell density and viability counts were made via trypan blue staining on a hemacytometer (Fisher Scientific, Ottawa, ON, Canada), and used to inform subsequent experiments.

In vitro Gd release

To determine the retention of the contrast agent over time within THP-1 cells, an in vitro release study was conducted. The release profile was established using six time points (10, 20, 40, 80, 160, and 320 min), from triplicate analyses. Following contrast agent incubation, 1 × 10⁶ cells were pelleted by centrifugation at 150× g in a swinging bucket rotor (Centra CL3R, Fisher Scientific) for 3 min, the supernatant removed, and the cells resuspended in 1.5 mL fresh prewarmed RPMI. For each time point, the cells were repelleted and 80% of the media (1.2 mL) removed by pipet for MR analysis. This volume was the maximum that could be collected while being sure not to disturb/remove any cells. Prewarmed fresh media (1.2 mL) was added to the samples and used to resuspend the pellet. This volume allowed the assumption of perfect sink mass transfer conditions (11) to evaluate contrast agent release. At the terminal time point, the cells were collected via centrifugation and resuspended in 300 µL 1% agarose (Bio-Rad, Mississauga, ON, Canada) in RPMI for MR assessment of contrast agent remaining associated with the cells. Supernatant time point samples were analyzed in a 96-well plate (300 µL/well), and release was assessed by determination of media MR relaxivity (R1).

THP-1 cell adhesion under flow to activated HUVECs in vitro

Similar to methods described in Reference 10, confluent HUVECs cultured on glass slides were treated with tumor necrosis factor-α (TNF-α; Sigma-Aldrich, Oakville, ON, Canada) at 7 ng/mL for 4 h to activate monocyte adhesion receptor expression (an inflammatory response relevant to athero sclerosis). These activated HUVECs were then transferred to a parallel-plate flow chamber (Figure 1) with a flow channel formed by a 4.6 cm × 1.2 cm cutout section from a 0.025 cm–thick silicone rubber gasket (Specialty Manufacturing, Saginaw, MI, USA) placed between the HUVEC slide and a ported acrylic top plate. The chamber was assembled with small clamps and connected to Norprene tubing (Cole Parmer, Montreal, QC, Canada), and flow was generated by a syringe pump (KD Scientific, Holliston, MA, USA). A heat lamp was used to maintain the experimental setup at ∼37°C.

For adhesion experiments, THP-1 cells (Gado CELLTrack–labeled or unlabeled) were resuspended in 15 mL R PMI and flowed over activated HU V ECs at 0.25 mL/min. One-hour experiments used 4.5 × 106 cells (3.0 × 10⁵ cells/mL), and 15-min studies used 1.8 × 10⁷ cells (1.2 × 10⁶ cells/mL). A field of view (FOV) sized640µmlong(directionofflow)by480µm was visualized at 10× on an inverted microscope (Nikon Instruments, Irvine, CA, USA) with a CCD camera (Dage, Michigan City, IN, USA). Following flow runs, all slides were transferred to a 10 cm × 10 cm square Petri dish and embedded in a 1% agarose-PBS gel to stabilize and preserve the cells prior to MR imaging.

Monocytes in a conformation mimicking endothelial localization

Both Gado CELLTrack–labeled and unlabeled THP-1 cells were collected by centrifugation and resuspended in 300 µL 1% agarose-RPMI solution at 37°C. Gaskets from flow chambers were placed across glass slides supporting HUVEC monolayers, and the THP-1/agarose suspensions were plated onto the HUVEC slides, inside the gasket. Following agarose solidification, the gaskets were removed, and the slides transferred to a Petri dish and embedded in agarose, as described in the previous paragraph for cell stabilization and preservation prior to MR imaging.

MR imaging

MR imaging was performed on a 3 Tesla MR scanner (Signa VH/i; GE Healthcare, Waukesha WI, USA), using an 8-channel head coil. For microplate analyses, T1 measurements were taken using an inversion recovery spin-echo pulse sequence with an echo time (TE) of 22 ms (minimum TE for this pulse sequence), repetition time (TR) of 5000 ms, and inversion time (TI) values of 50 ms, 100 ms, 150 ms, 200 ms, 250 ms, 300 ms, 400 ms, 500 ms, 750 ms, 1000 ms, 1500 ms, 1750 ms, 2000 ms, 2500 ms, 3000 ms, and 4000 ms. One signal average, a FOV of 16 cm × 12 cm, a slice thickness of 3 mm, and a reconstruction matrix of 512 × 192 were used. For the glass slide experiments, a fast spin-echo inversion recovery pulse sequence with the following parameters was used: TE = 7 ms, TR = 5000 ms, and TI = 1500 ms, one signal average, FOV 16 cm × 16 cm, slice thickness = 5 mm, and a reconstruction matrix of 256 × 192. ImageJ software (NIH, Bethesda MD, USA) was used for signal intensity measurements on all MR images, and linear regression analysis was used to calculate the R1 parameter values with Matlab (Mathworks, Natick,MA, USA).

Results and discussion

To determine the feasibility and appropriate timing for subsequent experiments and imaging with Gado CELLTrack, in vitro assays were conducted to evaluate the release profile and labeling potential of this positively charged nanoparticle in THP-1 monocytes. To accomplish this, the amount of contrast agent released from labeled cells into media samples was determined by MR relaxation rate (R1) at times between 10 min and 320 min post-labeling. As shown in Figure 2, the sharp drop in R1 between 10 and 20 min indicates that 85% of the total release occurred within this inter val. However, after 320 min, pelleted cells still had significantly higher relaxation (12%) than the same number of unlabeled cells—4.021 × 10⁻⁴ versus 3.605 × 10⁻⁴ ms⁻¹. At 320 min, the relaxation rate of media was equivalent to that of unlabeled cells (3.605 × 10 ⁻⁴ ms⁻¹), indicating that no additional agent had been released since the penultimate time point and agent release was completed. Information provided by the manufacturer characterized Gado CELLTrack as a 30-nm positively charged nanoparticle. Cell permeability to ions results in a net difference of electric charge on opposite sides of the cell membrane. In quiescent cells, the difference in electric charge across the membrane is called the resting membrane potential. Studies conducted on cultured human monocytes have reported a resting membrane potential in the range of -30 mV to -40 mV (12), which favors the retention of positively charged molecules within the cell. Therefore, the positive charge (and possibly, the size) of Gado CELLTrack could account for the small amount of retention observed in MR measurements. While this agent was determined to not be optimal for future studies focused upon translation of monocyte labeling with MR contrast agent to animals or humans, the data does suggest some form of loading within the THP-1 cells, making it suitable for the proof-of-concept work presented here.

After establishing that Gado CELLTrack would allow detection of labeled monocytes by MR imaging, studies to contact THP-1 cells and HUVECs in a basic tissue model of inflammation in early atherosclerosis were undertaken. Following 1 h of uninterrupted THP-1 flow over a TNF-α–activated HUVEC monolayer in vitro, the average number of adherent cells per 0.3 mm2 FOV in a representative microscopy image (Figure 3A) was 566 ± 23. This extrapolated to 1.9 × 10⁵ ± 7.7 × 10³ cells/cm² across the flow path. For the 5.5 cm² flow channel employed, this correlated to a total of approximately 1.0 × 10⁶ attached cells. 1.0 × 10⁵ to 1.0 × 10⁶ THP-1 cells were previously discovered to be optimal for T1-weighted MR detection of labeled cells in a 96-well plate (data not shown). Based on this information, and the relatively large area covered by 1.0 × 10⁶ adherent THP-1 cells (Figure 3A), we determined that it would be reasonable to attempt detection of this number of adherent THP-1 cells in the MR scanner. However, the contrast release profile indicated that at 1 h post-labeling, a considerable loss of contrast agent had occurred (Figure 2). Since 1 h of flow was required to reach the desired adherent THP-1 cell density of 1.0 × 10⁶ cells/flow path, and the THP-1 cells at that time would be essentially unlabeled, we decided to use a mimic of THP-1/HUVEC adhesion to establish initial proof-of concept. Thus, an immobilized cell system representative of THP-1 cell adherence to activated HUVECs was chosen as the first configuration to be imaged via MR.

To perform experiments investigating whether a signal could be detected from monocytes in contact with endothelial cells, Gado CELLTrack–labeled and unlabeled THP-1 cells were suspended in agarose and plated at 1.0 × 10⁶ cells/flow path on top of a HUVEC monolayer. This synthetic construct mimicked the amount of THP-1 cell adhesion to activated HUVECs observed after 1 h of flow (cell configuration shown in Figure 3B). For comparison, Figure 3, A and C shows adherent THP-1 cells following flow experiments, and Figure 3D has no THP-1 cells, only HUVECs. A T1-weighted MR image of the slides from this experiment is presented in Figure 4A, and includes labeled T HP-1/HUVEC, un labeled THP-1/HUVEC, and a HUVEC-only control. This image clearly shows the signal generated by the labeled THP-1 cells in agarose plated on top of HUVECs, and confirms that neither unlabeled THP-1 cells nor HUVECs alone generated signal enhancement. This data supplies proof-of concept that MR scanning of 1.0 × 10⁶ THP-1 cells distributed across an area of 5.5 cm² (182,000 cells/cm²) provides a specific signal for labeled monocytes in contact with endothelial cells.

Based on the successful imaging of labeled immobilized THP-1 cells overlayed on HUVECs, a dynamic adhesion experiment with labeled THP-1 cells was run in the flow chamber. To minimize contrast agent loss, the experimental time was shortened from 60 to 15 min of flow. To provide the same total number of THP-1 cells for HUVEC contact in this shortened experiment as in the 1-h study, the density of labeled THP-1 cells in the flow media was increased to 1.2 × 10⁶ cells/mL (previously 3.0 × 10⁵). Figure 3C illustrates the number of adherent monocytes (190 ± 64 per field of view, 6.3 × 10⁴ ± 2.1 × 10⁴ cells/cm²) that was achieved following 15 min of flow; a 3-fold reduction from the 1-h experiment. However, upon MR imaging, an obvious signal was recovered from the labeled THP-1 cells (Figure 4B), confirming the result obtained from the immobilized cell configuration, and suggesting that the MR approach may be capable of distinguishing an even smaller number of labeled cells. Additionally, this result demonstrates the feasibility of using a dynamic in vitro model to develop imaging strategies that target inflammatory responses in vivo, while accounting for target cell interactions.

Results of this study show that the concept of labeling inflammatory cells with a positive MR contrast agent for detection of early inflammatory processes is feasible and specific in both a synthetic and physiologically based model of monocyte/endothelial inter action. Additionally, the use of a simple in vitro system for the development and evaluation of contrast agents was shown to be an effective tool. While the contrast agent in this study was nonspecific and not well suited for follow-up work, our results suggest that molecular agents specifically targeted to monocytes should provide an efficient way to identify inflammatory responses by MR imaging. This is particularly important for the early detection of atherosclerosis, since initial lesions are clinically silent (13) and the first manifestation of disease is often an acute, fatal event such as myocardial infarction or stroke (14). Development of an MR-based screening method would allow noninvasive evaluation of patients, and would likely enable early intervention in sub-clinical cases.

A method of translating our findings from in vitro work to in vivo studies is to extract monocytes from a model athero sclerotic animal, label them in vitro, then reintroduce them into the animal and image their recruitment/localization to fatty streaks. In theory, the monocytes would home to the site of inflammation and differentiate into macrophages while retaining the label. However, based upon the results presented here, the use of a targeted contrast agent may be required. By employing in vitro systems to support the development of these agents, it is likely that the path to feasible compounds will be shortened, in terms of time and money, compared with methods that rely strictly on animal models to validate performance.