BIOENGINEERED IN VITRO 3D MODEL OF HUMAN ATHEROSCLEROTIC PLAQUE

Abstract

The invention provides a method for the generation of a layered cellular 3 D microtissue aggregate, comprising the steps of contacting myeloid cells with a protein kinase C agonist, yielding primed myeloid cells; incubating the primed myeloid cells in the presence of LDL in a confined volume, particularly in a hanging drop culture; yielding a 3 D culture of myeloid cells; and incubating the 3 D culture together with fibroblasts in a hanging drop in the presence of LDL, yielding the layered cellular aggregate.

Claims

1. A method for the generation of a layered cellular aggregate, comprising the steps of a) providing a population of myeloid cells, wherein said myeloid cells are provided (i) ex-vivo from a patient by isolation from fresh blood using a double gradient centrifugation; or (ii) ex vivo as cell culture or cell line characterized by expression of monocyte/macrophage cell markers; b) in a differentiation-priming step, contacting said myeloid cells with a protein kinase C agonist, yielding primed myeloid cells; c) in a culture step, incubating said primed myeloid cells in the presence of low-density lipoprotein (LDL) in a confined volume, particularly in a hanging drop culture; yielding a 3-dimensional culture, particularly a sphere, of myeloid cells; d) in a co-culture step, incubating said 3-dimensional culture of myeloid cells together with fibroblasts in a confined volume, particularly in a hanging drop culture in the presence of LDL, yielding said layered cellular aggregate.

2. The method according to claim 1, wherein said protein kinase C agonist is a phorbol ester, particularly phorbol 12-myristate 13-acetate (PMA).

3. The method according to claim 1, wherein in said differentiation-priming step, said myeloid cells are contacted with 5-200 ng/ml PMA, particularly with 5-50 ng/ml PMA, more particularly with 10 ng/ml PMA, for 36-120 hours, particularly for 48-96 hours, more particularly for 72 hours.

4. The method according to claim 1, wherein subsequently to said differentiation-priming step, said primed myeloid cells are contacted with lipopolysaccharide (LPS).

5. The method according to claim 4, wherein subsequently to said differentiation-priming step, primed myeloid cells are contacted with 10-100 ng/ml LPS for 30 min-3 hours, in particular with 10 ng/ml LPS for 1 hour.

6. The method according to claim 1, wherein said myeloid cells are human myeloid cells.

7. The method according to claim 1, wherein said fibroblasts are human fibroblasts.

8. An in vitro engineered layered cellular aggregate, comprising an inner sphere comprising, particularly consisting essentially of, a plurality of myeloid cells, collagen, particularly collagen III and cholesterol, an outer lining comprising, particularly consisting essentially of, fibroblasts, wherein said outer lining substantially encases said inner sphere, and wherein said cellular aggregate has a diameter of 100 m-500 m, in particular approx. 250 m, and does not contain a pre-formed, cell-free scaffold.

9. The in vitro engineered cellular aggregate according to claim 8, wherein said plurality of myeloid cells comprises monocytes, macrophages and dendritic cells.

10. The in vitro engineered cellular aggregate according to claim 9, wherein said monocytes, macrophages and dendritic cells are each present at a defined ratio, wherein said defined ratio is 20%-40%, particularly 26%-36%, more particularly approximately 31% of monocytes; 25%-45%, particularly 29%-39%, more particularly approximately 34% of macrophages; 15%-45%, particularly 20%-40%, more particularly approximately 31% of dendritic cells.

11. The in vitro engineered cellular aggregate according to claim 8, wherein said plurality of myeloid cells is positive for the expression of a proinflammatory marker selected from the group comprising CXCL10, CCR7, IL23, PTGS1 and ALOX5.

12. The in vitro engineered cellular aggregate according to claim 8, wherein said plurality of myeloid cells is positive for the expression of a remodelling marker selected from the group comprising CCL17, CCL26, DC-SIGN, IL10 and SRB1.

13. A plurality, in particular a manifold of 8 or 12, more particularly 96 or 384, of a. the layered cellular aggregates according to claim 8, or of b. a layered cellular aggregate generated by the method of claim 1.

14. A method to assess the likelihood of a candidate compound to be effective in a treatment of atherosclerosis, comprising the steps of a. providing, particularly in a hanging drop culture, a layered cellular aggregate generated by the method of claim 1, or a layered cellular aggregate according to claim 8, or a precursor of the layered cellular aggregate obtainable by a method according to claim 1, wherein said precursor consists of said primed myeloid cells yielded by said differentiation-priming step, or said 3-dimensional culture of myeloid cells yielded by said culture step, or a co-culture of fibroblasts and said 3-dimensional culture of myeloid cells; b. contacting said layered cellular aggregate or said precursor with said candidate compound; and c. detecting an effect of said compound on said layered cellular aggregate or on the formation of said layered cellular aggregate, in particular a beneficial effect with regard to size of said layered cellular aggregate, cellular viability or cellular aggregation within said layered cellular aggregate; d. assigning to said candidate compound a high likelihood of being effective in a treatment of atherosclerosis if said effect is detected.

15. A method for identifying a biomarker of atherosclerosis, comprising the steps of a. providing, particularly in a hanging drop culture, a first layered cellular aggregate generated by the method of claim 1, or a first layered cellular aggregate according to claim 8, wherein the myeloid cells comprised in said first layered cellular aggregate have been provided from a patient suffering from monogenic familial hypercholesterolemia; b. comparing the transcriptome and/or proteome of said first layered cellular aggregate with the transcriptome and/or proteome of a control layered cellular aggregate; identifying a protein or mRNA that is unregulated or downregulated in said first layered cellular aggregate, thereby identifying said biomarker of atherosclerosis.

Description

BRIEF DESCRIPTION OF THE FIGURES

[0109] FIG. 1 shows the identification of different myeloid subsets using the vi-SNE workflow. Samples are stained and acquired at the FACS analyser. The resulting FCS files are processed to generate vi-SNE maps. Each cell recorded (single event) is positioned in a specific area of the high-dimensional space, here represented with commonly used FACS biaxial plots. Different myeloid subsets are positioned in separate regions of the high-dimensional space according to surface marker similarities. Distances between cells are representatives of cell proximity in high-dimensional rather than two-dimensional space. These myeloid subsets are automatically gated with using PhenoGraph and FlowSOM algorithms and identified in the vi-SNE map with different colours.

[0110] FIG. 2 shows surface expression levels of key markers in different myeloid subsets. 15 cell populations were identified using the vi-SNE workflow. Each myeloid population retains a specific surface marker expression pattern indicated by the heat-map. The present heat-map represents the median fluorescence intensity of each marker in each myeloid subset and was computed using total of 112,000 events; n=7; 2,000 randomly selected events for each sample analysed (samples analysed for both b and t plaques: Mo, pMf/d, T2, T2L).

[0111] FIG. 3 shows density plots from blood-derived cells before and after differentiation-priming process. (a) Density distribution of recorded events before (Mo) and after (pMf/d) the differentiation-priming process. Areas of high event density are depicted in red and areas with low event density in blue. n=7; 2,000 randomly selected events for each sample analysed are shown (Mo and pMf/d). Therefore, in each vi-SNE maps 14,000 events are plotted. (b) Expression level plots. Each vi-SNE map shows 112,000 events; n=7; 2,000 randomly selected events for each sample analysed (samples analysed for both b and t plaques: Mo, pMf/d, T2, T2L).

[0112] FIG. 4 shows differentiation-priming effects on thp-1 cells. vi-SNE maps indicate different myeloid populations identified within thp-1 cells at the beginning (Mo) and at the end (pMf/d) of the differentiation-priming process. The part-of-whole graph shows the % of events and summarizes the vi-SNE results. n=7; 2,000 events per sample are reported. A total of 14,000 events are shown in each vi-SNE map.

[0113] FIG. 5 shows density plots from thp-1 cells before and after differentiation-priming process. (a) Density distribution of recorded events before (Mo) and after (pMf/d) the differentiation-priming process. Areas of high event density are depicted in red and areas with low event density in blue. In each vi-SNE maps 14,000 events are plotted. n=7; 2,000 randomly selected events for each sample analysed are shown (Mo and pMf/d). (b) Marker expression level plots.

[0114] FIG. 6 shows changes in gene expression profile of thp-1 cells upon differentiationpriming. The expression profile of pro-inflammatory (leftwhite panel) and remodelling (rightorange panel) genes is reported in fold change over Mo expression levels (where fold change=1). Data (n=5) were normalized on the averaged GAPDH and 18S expression.

[0115] FIG. 7 shows the myofibroblasts characterization. Myofibroblasts isolated from human umbilical vein express -smooth muscle actin (-SMA) and smooth muscle myosin heavy chain (SMMHC); scale bar 100 m.

[0116] FIG. 8 shows vi-SNE comparison between bioengineered b-plaques and t-plaques. (a) vi-SNE density plots. Areas of high event density are depicted in red and areas with low event density in blue. n=7; 2,000 events per sample were randomly down-sampled from each sample. A total of 14,000 events is shown in each vi-SNE map. (b) Marker expression level plots.

[0117] FIG. 9 shows density plots from myeloid populations isolated from native carotid plaques. (a) vi-SNE maps indicate the event density recorded in each myeloid population. Areas of high event density are depicted in red and areas with low event density in blue. PDC sub-populations are indicated in the vi-SNE map: type-, type- and type-. n=5. 2,000 events per samples are plotted for a total of 10,000 events. (b) Marker expression level plots.

[0118] FIG. 10 shows LDL effects on t-plaque myeloid populations. vi-SNE maps indicating different myeloid populations identified in LDL-rich (LDL) or LDL-free (w/o LDL) t-plaques. The part-of-whole graph summarizes the results depicted in the vi-SNE maps. n=7; 2,000 events per sample are reported for a total of 14,000 events in each vi-SNE map.

[0119] FIG. 11 shows LDL effects on myeloid populations isolated from t-plaques. (a) Event density maps of myeloid populations from LDL-rich (LDL) or LDL-free (w/o LDL) t-plaques. High-density (red) and low-density (blue) areas are shown. n=7; 2,000 events per sample are reported. 14,000 events are displayed in each vi-SNE map. (b) Marker expression level plots.

[0120] FIG. 12 shows the variation in gene expression profiles during t-plaque formation Heat-map indicating the expression levels of target genes of interest at T1 and T2 of t-plaque formation. Ps-plaques were biofabricated either in LDL-rich (T1L, T2L) or LDL-free (T1, T2) environment. Expression levels are reported in Ct over the mean expression of GAPDH and 18S housekeeping genes; n=5.

[0121] FIG. 13 shows the proposed hematopoietic differentiation model of the myeloid lineage. In the current model of myeloid differentiation, the common macrophage-dendritic cell precursor (MDP) derives from the common myeloid progenitor (MP) and gives rise to monocytes and to the common dendritic cell progenitor (CDP). The latter differentiates in plasmacytoid dendritic cells (PDC) and pre-classical dendritic cells (pre-cDC). Pre-cDC, PDC and monocytes circulate in the blood and can migrate within the plaque where they differentiate in classical dendritic cells (cDC), tissue resident plasmacytoid dendritic cells (PDC), macrophages or monocyte-derived dendritic cells (M-DC). In the proposed model a dendritic progenitor can circulate in the blood and differentiate in pre-cDC and PDC.

[0122] FIG. 14 shows a schematic view of the pseudo-plaque assembly including a step of cell conditioning and a step of hanging drop culture.

[0123] FIG. 15 shows confocal imaging of the atherosclerotic plaque. A) CD45* myeloid cells are embedded and surrounded by aSMA* fibroblasts (scale bar 100 m). B) Collagen deposition within the bioengineered plaque (scale bar 50 m). C) Intracellular cholesterol accumulation (scale bar 50 m).

[0124] FIG. 16 shows bioengineered plaques and human carotid plaques compared with qPCR. Pro-inflammatory and remodeling gene induction levels are measured and compared in the graph; p-value: *p=0.333; *19=0.002; ***p<0.001; test: RM two-way ANOVA with Sidak correction h plaque, t plaque, carotid artery n=5.

[0125] FIG. 17 shows the product production pipeline. Thp1 monocytes are cultured in vitro up to the desired concentration (about 2 mio cells for one 96 well plate). The STEP1 (differentiation-priming) lasts about 3 days (72 h+1 h). At this point the cells are detached and inserted in a robotic system for STEP2. Hanging drops of 10 l are generated automatically on the lid of a 96 or 384 well plate. The bottom of the 96/384 well plate is automatically filled with 100 l 1x PBS to create a humid environment necessary to keep the drop volume constant. The plates are kept at 37 C. and 5% CO.sub.2 for 48 h. At this point myofibroblasts are automatically added to the forming bioengineered spheroids and the plates are be kept at 37 C. and 5% CO.sub.2 for additional 48 h after one week of processing, the product is ready to be delivered to the customer.

[0126] FIG. 18 shows scanning electron microscopy imaging of the ps-plaque. (a) SEM image of a ps-plaque. Scale bar 40 m. (b) Detail of the ps-plaque external surface (myofibroblast layer). Scale bar 4 m. (c) Inner architecture of the ps-plaque. Scale bar 4 m. With a detail (d, d1) showing macrophage/dendritic cells surfaces in contact with LDL cholesterol. Scale bar 10 m.

[0127] FIG. 19 shows transmission electron microscopy imaging of the ps-plaque. (a) TEM image of a ps-plaque showing accumulation of structured cholesterol crystals at the external plaque surface. Scale bar 40 m. (b) Detail of the ps-plaque external surface. Apoptotic myofibroblast in contact with a structured cholesterol crystal. Scale bar 4 m.

[0128] FIG. 20 shows biofabricated human atherosclerotic plaque. vi-SNE maps indicating the cell populations at the beginning (Mo) and at the end (pMf/d) of the differentiation-priming process.

[0129] FIG. 21 shows a comparison between biofabricated plaques and human carotid plaques: vi-SNE maps from CD45+ populations isolated from blood-derived plaques (b-plaques), thp1 plaques (t-plaques) and carotid plaques.

[0130] FIG. 22 shows a comparison between biofabricated plaques and human carotid plaques: The stacked-bar chart summarizes the percentage of events for each population.

[0131] FIG. 23 shows LDL effects on myeloid cells isolated from b-plaques. (a) vi-SNE maps indicating myeloid populations identified within b-plaques biofabricated either using LDL-enriched (LDL) or LDL-free (w/o LDL) medium. The stacked-bar chart summarizes the results from the vi-SNE maps and reports the percentage of events recorded in each population. (b) Event density distribution vi-SNE maps show high-density (red) and low-density of events (blue) areas. (c) Marker expression level vi-SNE plots.

[0132] FIG. 24 shows a transcript analysis of pro- and anti-inflammatory gene targets in b-plaques and LDL effects on cell viability and plaque dimension. (a) Heat-map indicating the expression levels of target genes of interest at T1 and T2 of b-plaque formation in LDL presence (T1L, T2L) or absence (T1, T2). Expression levels are reported in Ct over the mean of the housekeeping genes GAPDH and 18S; n=5 (b) ATP levels, indicated in relative luminescence units (RLU), were measured and compared in b- and t-plaques; n=5. (c) Plaque circular cross-section area (mm.sup.2) as indicator of plaque dimension; n=5. (d-e) The ps-plaque necrotic area was measured at the circular cross-section at the great circle of the spheroid, and indicated as percentage over alive cells. Living cells (green) are stained with calcein while dead cells (red) with Eth-1.

[0133] Table 1 shows myeloid populations in ps-plaques and native carotid plaques.

[0134] List 1 shows a primer table. Gene of interest, forward (FW) and reverse (RV) primer sequences are listed.

EXAMPLES

Example 1: Methods

[0135] Isolation of Myeloid Cells from Blood.

[0136] Myeloid cells were isolated from fresh human blood using a double gradient centrifugation. The blood was provided by the Zurich blood bank (Blutspende ZrichNr. 6676) and maintained at room temperature in slow rocking motion until processing. First, 20 ml of blood from each donor were diluted 1:2 with 1PBS (Sigma) at room temperature and layered onto a Ficoll solution (1.077 g/ml, Sigma). Samples were then centrifuged at 400 g for 30 min without break. Second, a 46% iso-osmotic Percoll gradient was performed to separate the lymphocytes from the PBMCs as previously described (Menck, K., J Vis Exp, e51554, doi:10.3791/51554 (2014)). Briefly, the buffy coat was re-suspended in 20 ml of xVivo15 chemically defined medium (Lonza) without red phenol and carefully layered on top of a Percoll solution prepared with 50% RPMI medium with red phenol (Sigma), 46% Percoll (GE Healthcare) and 4% 1PBS (Sigma). The second gradient was centrifuged at 550 g for 30 min without break and the white cell ring at the interphase was collected for further processing.

[0137] Myofibroblasts Isolation.

[0138] Human umbilical vein myofibroblasts (HUVM) were isolated from human umbilical cords. The tissues were processed in accordance to the ethical permit released by the Kantonale Ethikkommission Zrich (KEK-Stv-21-2006). Briefly, umbilical cords were stored after labor at 4 C. in DMEM medium (Sigma) prepared with 10% FBS (Gibco), 1% GlutaMax (Gibco) and 1% Penn/Strep (Gibco) for maximum 2 h prior to processing. The umbilical vein was carefully extracted from the umbilical cord and the inner lumen was flashed twice with 1PBS. The adventitia layer was peeled off with the help of forceps and scalpel. The intima layer was removed by incubating the inner lumen for 30 min in a 1 mg/ml collagenase/dispase (Roche) solution in 1PBS. The remaining endothelial cells were washed out from the lumen with 1PBS. The remaining media layer was minced into small pieces of approximately 2 mm length and let adhere for 10 min on the bottom of a petri dish. The tunica media fragments were then covered in DMEM medium and maintained at 37 C., 5% CO.sub.2 and 95% humidity. The medium was replaced every 48 h. After about 20 days myofibroblasts sprouting from the minced pieces reached about 80% confluence and were ready for sub-culturing.

[0139] Cell Culture.

[0140] HUVM were cultured in DMEM medium with 10% FBS and 1% GlutaMax and the medium was replaced every 48-72 h. For sub-culturing, HUVM were detached using trypsin 0.5% (Sigma) for 4 min and seeded at a cell density of 4,000 cells/cm.sup.2. HUVM were expanded up to passage 5 prior to use for the experiments in this study. Human monocytic leukaemia cell line (thp-1) isolated from the peripheral blood of a 1-year-old human male with acute monocytic leukemia, were purchased from Sigma. Thp-1 cells were cultured in suspension in xVivo15 medium and the medium was replaced every 2-3 days. Thp-1 cells were seeded at a density of about 100,000 cells/ml and sub-cultured at a density of 800,000 cells/ml.

[0141] Ps-Plaque Biofabrication.

[0142] The pseudo-plaque production pipeline encompasses three steps: differentiation, priming and hanging-drop. First, fresh blood-derived myeloid cells or thp-1 cells were seeded onto petri dishes for 72 h and differentiated in chemically defined xVivo15 medium with 10% FBS in order to achieve a macrophage/dendritic cell phenotype. To induce thp-1 differentiation 10 ng/ml of phorbol 12-myristate 13-acetate (PMA, Sigma) were added to the culture medium. Second, a priming step was performed to obtain heterogeneous macrophage/dendritic cell populations with both pro-inflammatory and remodelling phenotypes. For this purpose the differentiated cells were rinsed in 1PBS and treated for 1 h in xVivo15 medium with 10% FCS and 10 ng/ml lipopolysaccharide (LPS, Sigma). Finally, the primed cells were transferred in hanging-drop culture. Briefly, adhesive myeloid-derived cells were mechanically detached by 20 min incubation in 0.05 mM EDTA (Life Technologies) in 1PBS at 4 C. and gentle scraping. Cells were re-suspended at a cell density of 2.410.sup.6 cells/ml in presence of LDL 50 g/ml (LEE Biosolutions) in xVivo15 medium with 10% FBS. Droplets of 10 l were pipetted on the lead of a 10 cm diameter petri dish and kept in hanging-drop culture for 48 h. To the core of myeloid-derived cells assembled during the 48 h incubation, an external layer of HUVM was added. HUVM were prepared at a cell density of 410.sup.5 cells/ml in DMEM medium with or without 50 g/ml LDL and 10 l of the cell suspension were carefully added to each pre-existing drop and cultured in hanging-drop for further 48 h.

[0143] Flow Cytometry.

[0144] Biopsies of carotid branches were obtained from patients undergoing carotid endarterectomy and shunting, secondary to vascular stenosis (Ethik Kommission der Universitat Witten/HerdeckeNr. 79/2012). Carotid plaques and biofabricated ps-plaques were digested with 1 mg/ml collagenase/dispase solution in 1PBS for 15 min at 37 C. Cells were gently pipetted through a cell strainer with the mesh size of 40 m (Falcon) and incubated for 5 min at 4 C. with magnetic beads coated with anti CD45 antibodies, according to the provider instructions (MACS Miltenyi Biotec). CD45+ cells were magnetically sorted and stained with Zombie Aqua fixable viability kit (BioLegend) for 5 min and fixed over night at 4 C. in a 1% Paraformaldehyde (PFA, Sigma) solution in 1PBS. The single cell suspension was stained for 15 min at room temperature in FACS buffer prepared with 5% FCS and 0.01% NaN.sub.3 (Sigma) in 1PBS with an optimized FACS antibody panel including: CD14-PerCP (#325631, Biolegend), CD16-Alexa 700 (#360717, Biolegend), CD11b-Alexa594 (#101254, Bioloegend), CD11c-PE-Cy5 (#301609, Biolegend), CD36-BV605 (#563518, Becton Dickinson) and SRA-1-PE (# REA460, MACS Miltenyi Biotec). Each antibody was previously titrated to establish the optimal working concentration. Samples were acquired using LSRFortessa analyser (Becton Dickinson) and signal compensation was performed using OneComp eBeads (eBioscience).

[0145] VI-SNE Workflow.

[0146] The FCS files obtained from the FACS analysis were pre-processed using the software Flowjo (FlowJo, LLC). First, cell populations of interest were gated according to forward and side scatter (FSC and SSC) parameters. Second, singlets were gated and Zombie Aqua dye negative events, representing the alive population of interest, were exported for further processing. Data post-processing was performed using the R platform and the Cytofkit package. Briefly, pre-processed FCS files from each sample were loaded onto Cytofkit, randomly down-sampled to 2,000 events (ceil; n=2,000) and computed using t-Distributed Stochastic Neighbor Embedding (t-SNE) algorithm (Van Der Maaten, L, J Mach Learn Res 9, 26 (2008)). Each event recorded was positioned in a specific location of the high-dimensional space. The output was a vi-SNE biaxial plot where distances between events are representatives of cell proximity in high-dimensional rather than two-dimensional space. The proximity between events is based on similarities in surface marker expression levels. Different myeloid subsets were positioned in separate regions in high-dimensional space according to surface marker similarities. Automatic gating of myeloid subsets was performed through a preliminary clustering step with PhenoGraph algorithm (k=42) and a following metaclustering step with FlowSOM algorithm (k=10).

[0147] Immunofluorescence and Immunohistochemistry.

[0148] Myofibroblasts were fixed for 20 min in 4% PFA in 1PBS and maintained in 1PBS at 4 C. until further processing and not more than 7 days. Cells were stained with the primary antibodies anti-alpha smooth muscle actin (aSMA, # ab7817, Abcam) and anti-smooth muscle myosin heavy chain (SMMHC, # ab53219, Abcam) overnight at 4 C. and with secondary antibodies (anti-mouse #715-605-151, Jackson Immuno Research; anti-rabbit # A11008, Life Technologies) and phalloidin (# A12381, Life technologies) for 1 h at 37 C. Nuclei were counterstained with DAPI and the slides were mounted in Vectaschield (Vector Laboratories). The ps-plaques were carefully washed in 1PBS and fixed in PFA as described above. Plaques were dehydrated overnight in a solution of 25% sucrose (Sigma) in 1PBS, embedded in OCT matrix (CellPath) and stored at 20 C. Slices of 5 m were cut, rehydrated in 1PBS for 15 min and stained with primary antibodies: anti-Collagen type Ill (# ab7778, Abcam), anti-aSMA and anti-CD45-PeCy5 (#304009, BioLegend) overnight at 4 C. Secondary antibody staining was performed (anti-mouse, 715-545-151, Jackson Immuno Research; anti-rabbit # A11008, Life Technologies) for 1 h at 37 C. For the Filippin sections were quenched for 10 min with 1.5 mg/ml glycine (Sigma) in 1PBS prior to addition of 250 g/ml Filippin III dye (Sigma) at room temperature for 2 h. Sections were washed 3 times in 1PBS and nuclei were counterstained with propidium iodide 1 mg/ml (BioLegend) for 5 min. Slides were mounted in Vectaschield. Images were acquired in grey scale with the confocal microscope (Leica SP8). Image post-processing, specifically the choice of appropriate pseudo-colours, was performed using ImageJ.

[0149] RT-qPCR.

[0150] Total RNA was extracted using the GenElute Mammalian Total RNA Kit (Sigma), following the manufacturer's instructions. Reverse transcription was performed for each sample in a 20 l reaction mixture containing 1 g of RNA, 1PCR buffer, 5 mM MgCl.sub.2, 10 mM of each dNTP, 0.625 M oligo d(T).sub.16, 1.875 M random hexamers, 20 U RNase inhibitor and 50 U MuLV reverse transcriptase (all from Life Technologies). The conditions for the reverse transcription were the following: 25 C. for 10 min, 42 C. for 1 h, followed by 99 C. for 5 min. The resulting cDNA was amplified in duplicate by quantitative real-time PCR in 10 l reaction mixture with 200 nM of each specific primer (List 1) and 1Fast Syber Green qPCR MasterMix (Applied Biosystems). For the amplification reaction, StudioQuant 7 was used (Applied Biosystem). The amplification program was set as follows: 95 C. for 5 min, followed by 40 cycles at 95 C. for 10 sec, 60 C. for 15 sec, 72 C. for 20 sec. GAPDH and 18S served as housekeeping genes and their amplification data were averaged and used for sample normalization. The software Excel (Microsoft) was used for the comparative quantification analysis.

[0151] Ps-Plaque Viability Assay.

[0152] Cell viability within the plaque was measured using CellTiter-Glo 3D Cell Viability Assay (Promega). Briefly, the biofabricated plaques were washed in 1PBS and dispensed in a opaque-walled 96 well plate (Costar). Each ps-plaque (1 plaque/well) was dispensed in 15 l of 1PBS. Equal volume of CellTiter-Glo 3D Reagent was added to each well for a final volume of 30 l. Luminescence was measured after a 30 min of incubation at room temperature with SPECTRAmax Gemini-XS (Bucher biotech) and ATP levels were reported in relative luminescence units (RLU).

[0153] Quantification of Ps-Plaque Area and Necrotic Area.

[0154] For the measure of the plaque necrotic area, every plaque was stained for 40 min in a solution of calcein (5 M) and eth-1 (15 M) from the LIVE/DEAD Viability/Cytotoxicity Kit, for mammalian cells (Life Technologies). Ps-plaques were imaged using an inverted microscope (Leica, DM IL LED) and post-processed in ImageJ. Briefly, images underwent colour 2D Parallel iterative deconvolution using the WPL method (Max number of iteration=5; Max number of threads.sup.2=4). The results of the point of spread function obtained from the deconvolution were normalized and the green and red channels were thresholded with the MaxEntropy setting. The ps-plaque necrotic area was measured as necrotic area over alive area and indicated as percentage. Plaque dimension was measured using the bright field images of the plaque circular cross section. First, the image was converted to 8-bit format and thresholded with the MaxEntropy method. Second, the area of the particles was analysed from objects with a dimension larger than 1,000 px in order to exclude debris or single cells not belonging to the bioengineered plaque. Plaque area was reported in mm.sup.2.

[0155] Statistical Analysis.

[0156] vi-SNE cluster counts and PCR comparative quantitations were analysed using multiple comparison analysis. First, Gaussian distribution of the data was confirmed with Shapiro-Wilk normality test. Second, repeated measures (RM) two-way ANOVA with Tukey's multiple comparison test was applied. Luminescence, plaque and necrotic area were analysed with paired t-test. All statistical analyses were performed with GraphPad Prism Version 6, GraphPad Software, San Diego, Calif., USA). Significance was accepted at p<0.05. All data are presented as means.d.

Example 2: Differentiation-Priming Strategy Promotes Population Redistribution in Cells from Myeloid Origin

[0157] A two-step bioengineering method for the assembly of the ps-plaque was established (FIG. 14). Myeloid cells isolated from freshly drawn blood and thp-1 cells were differentiated towards macrophage/dendritic phenotype and primed with LPS to obtain a mixed population of pro-inflammatory and remodelling cell populations. The success of the differentiation-priming strategy was verified using a fluorescence-activated cell sorting (FACS) and the results were computed using dimensionality reduction and clustering algorithms, PhenoGraph and FlowSOM respectively (Amir el, A. D., Nat Biotechnol 31, 545-552, doi:10.1038/nbt.2594 (2013)). With this technique the inventors identified 15 cell populations in the multidimensional space that they classified according to the differential expression levels of key surface markers (FIG. 20; FIG. 1; FIG. 2 and Table 1). In samples isolated from the blood the inventors identified 4 over-represented populations: classical monocytes (Mod), macrophages (WM, pre-classical dendritic cells (pre-cDC) and an unknown myeloid progenitor population (Prog) (FIG. 20; FIG. 2). Each population at the end of the differentiation-priming process was monitored and a significant decrease in the unknown myeloid progenitors (p=0.005) observed, coupled with a significant increase in pre-cDC (p=0.05) (FIG. 20). Additionally, the inventors observed a priming-induced increase in CD11c surface levels within the pre-cDC population and in the myeloid progenitors (FIG. 3). In cell samples from untreated thp-1, the inventors observed an initial population distribution similar to the one found in blood samples. When they applied the priming process to thp-1 monocytes, the inventors observed a significant reduction in the myeloid progenitors count (p<0.001). The latter was concomitant with a decrease in pre-cDC count (p<0.001) and the appearance of plasmacytoid dendritic cells (PDC) (FIG. 4, FIG. 5). Furthermore, the inventors observed that the priming process triggered the proliferation of classical monocytes (p<0.001) (FIG. 4). The inventors then analysed the expression levels of pro-inflammatory and remodelling gene targets. In blood derived cells they observed induction of CXCL10 (p<0.001), CCL17 (p<0.05), DC-SIGN (p<0.001) and SRB1 (p<0.001) upon treatment, indicating the overall stronger induction of remodelling over pro-inflammatory genes (FIG. 6). When they analysed the changes in thp-1 cells gene expression levels upon differentiation-priming, the inventors observed induction of CXCL10 (p=0.03), PTGS1 (p=0.05) and IL10 (p=0.01), indicating pro-inflammatory gene up-regulation over remodelling genes (FIG. 6).

Example 3: Defining the Gravity-Guided Biofabrication of Human Atherosclerotic Plaques

[0158] Primed cells were detached from the petri dish and cultured in hanging drop for 48 h to foster cell aggregation. LDL was added to the culture medium to mimic the atherosclerotic niche composition. At the end of the 48 h incubation, myofibroblasts isolated from the human umbilical vein (HUVM) were added to the hanging drop to establish a co-culture system (FIG. 14; FIG. 7). After additional 48 h the inventors observed the -SMA+ HUVM cells integrating within the pre-existing myeloid CD45+ cell aggregates. Moreover, they observed the formation of a thin fibrotic layer around the bioengineered spheroid (FIG. 1d), the assembly of collagen clumps within the ps-plaque and intra-plaque accumulation of lipid aggregates (FIG. 15b).

Example 4: vi-SNE Analysis Reveals Plasmacytoid and Activated Dendritic Cells as Main Myeloid Components in Human Fibroatheroma

[0159] To corroborate the ps-plaque model the inventors conducted a comparison study between bioengineered and human atherosclerotic plaques isolated from patients that underwent carotid endarterectomy. First, they sorted CD45+ populations from bioengineered blood-derived ps-plaques (b-plaques), thp1-derived ps-plaques (t-plaques) and human carotid plaques. Second, they analysed and compared the population distribution within the samples using flow cytometry. The inventors observed large similarities in population distribution within b-plaques and t-plaques. In detail, they found that the main cell populations are classical monocytes, macrophages, activated dendritic cells and plasmacytoid dendritic cells (FIG. 2a,b; FIG. 8). When they analysed CD45+ cells from carotid plaques, they identified PDC and aDC populations as main myeloid plaque components (FIG. 21, FIG. 22; FIG. 9). The inventors further investigated the event density distribution within the PDC populations in human carotid plaques. The inventors identified 3 major areas of the vi-SNE map corresponding to peculiar PDC phenotypes that they classified as , and (FIG. 9). PDC type- represents a relatively small cluster with phagocytic and lipoprotein clearance predisposition due to high surface levels of scavenger receptors CD36, SRA-1 and CD14. PDC type-8 is a larger cell cluster characterized by CD16high, indicating a possible involvement in pro-inflammatory reactions. PDC type- appears to be exclusively specialized in lipid and lipoprotein uptake, provided the predominant surface expression levels of CD36 (FIG. 9). Interestingly, in both bioengineered plaque models (b- and t-plaques) the inventors identified PDC Type-8 (FIG. 8). Finally, they analysed the gene expression profile of the CD45+ populations in ps-plaques and carotid plaques. The inventors reported a significant down-regulation of pro-inflammatory and remodelling gene targets in carotid plaques compared to bioengineered plaques (FIG. 16a-b).

Example 5: Low-Density Lipoprotein Promotes the Differentiation of a Precursor Myeloid Population in Biofabricated Plaques

[0160] The inventors investigated the effects of LDL on the differentiation of myeloid (CD45+) subpopulations isolated from b- and t-plaques. To do so, they biofabricated ps-plaques using either the established protocol based on LDL-enriched medium or using LDL-free medium. They applied the vi-SNE workflow to compare the respective cell populations. In b-plaques they observed a reduced count of precursors in LDL-enriched versus LDL-free controls (p<0.001, FIG. 23a) suggesting differentiation triggered by LDL. The difference in precursor counts can also be appreciated in the respective density plots (FIG. 23b). Additionally, they compared vi-SNE density plots from ps-plaques biofabricated in LDL-rich and LDL-free medium. To investigate variations in LDL triggered surface antigen expression within each population they overlapped the density plots with the marker expression level plot. The inventors observed an LDL dependent density shift in aDC towards vi-SNE areas with CDllchigh, CD16high and CD36high expression levels (FIG. 23b,c). In t-plaques, they observed an LDL dependent density shift of M1 towards CD36high and CDllchigh areas of the vi-SNE map and of Mocl towards a CD36high vi-SNE area (FIG. 10; FIG. 11).

Example 6: The Hanging-Drop Environment Allows the Establishment of a Pro-Inflammatory Niche

[0161] To uncover possible transcriptional effects exerted by LDL on key target genes, the inventors investigated the expression profiles of the myeloid component during two steps of the ps-plaque biofabrication: (i) after 48 h in hanging-drop (T1) and (ii) at the end of the hanging-drop process (T2). They compared T1 and T2 from ps-plaques produced in LDL-free or LDL-enriched environments. Surprisingly, despite the induction of dendritic cell-specific intercellular adhesion molecule DC-SIGN (p<0.001 versus p=0.003, FIG. 24a) they found any LDL-dependent significant transcriptional change in neither b-plaques nor t-plaques (FIG. 24a; FIG. 12). On the other hand, they found that the hanging-drop process had, per se, a major influence on the gene expression levels by directly or indirectly promoting the establishment of a pro-inflammatory environment. In detail, in b-plaques the inventors observed a significant down-regulation of CCL26 (p<0.001) and up-regulation of key pro-inflammatory genes CXCL10, CCR7 and IL23 (p<0.001) during the transition from T1 to T2 and independently from the presence of LDL (FIG. 24a). They observed an indirect pro-inflammatory effect in t-plaques, exerted through the down-regulation of the anti-inflammatory cytokines CCL26, MO and CCL17 (p<0.001, FIG. 12).

Example 7: Low-Density Lipoprotein Enhances Cell Death in Ps-Plaques Biofabricated with Primed Blood Cells

[0162] To further explore the effects of LDL on the ps-plaque model the inventors conducted a bivalent analysis. First, they investigated the cell viability within the ps-plaque. They measured and compared the ATP levels produced by the biofabricated plaques in LDL-enriched and LDL-free medium using a luminescence-based ATP assay. The inventors did not find any significant LDL-dependent differences in ATP levels in either b- or t-plaques (FIG. 24b). However, they observed a general tendency of lower ATP levels in plaques produced in LDL-rich environments. The inventors then measured the necrotic area at the circular cross-section with a calcein-ethidium based cell viability assay. They found that the necrotic area at the cross-section of b-plaques fabricated in presence of LDL was significantly larger in comparison to their LDL-free counterparts, suggesting an LDL-dependent necrotic effect (p<0.001, FIG. 24d,e). Second, they investigated differences in plaque dimensions to verify possible effects of LDL on cell proliferation. The inventors used the circular cross-sectional area of the ps-plaque as an indicator of plaque size. They found no difference in circular cross-section areaconsequently in sizeof LDL-enriched vs LDL-free plaques (FIG. 24c).

DISCUSSION

[0163] With the biofabrication of the ps-plaque the inventors aimed at replicating cellular architecture and extracellular microenvironment of a human atherosclerotic plaque to close an open modelling gap in the field of atherosclerosis research. It has been described that the fibroatheroma cellular composition is mainly characterized by macrophages and dendritic cells retaining pro-inflammatory and remodelling abilities. To achieve plaque cell populations as similar as possible to human atherosclerotic plaque phenotypes, the inventors established a differentiation-priming protocol based on a mild LPS stimulation of cultured adhesive myeloid cells. To visualize and quantify the effects of this procedure on cell population remodelling they used the vi-SNE workflow. With this strategy they identified a total of 15 cell populations, differently distributed among samples. The inventors were able to classify these populations according to the prevalence of specific surface markers. They were also able to track intra-population density shifts and changes in numbers of events. The sensitivity of the vi-SNE analysis allowed the identification of under-represented myeloid populations, otherwise difficult to identify with commonly used flow cytometry analysis tools. Thanks to the vi-SNE workflow the inventors identified in blood-derived myeloid samples both plasmacytoid dendritic cells (PDC) and pre-classical dendritic cells (pre-cDCs). Interestingly, although thp-1 cells and blood-derived myeloid populations share remarkable similarities concerning population distribution, PDC are almost absent from thp-1 samples before the differentiation-priming treatment, emphasizing the differences between the thp-1 cell line and their physiological counterparts previously discussed by Bosshart and Heinzelmann. Moreover, the vi-SNE analysis reported a yet unidentified myeloid population in both thp-1 and blood derived samples. The inventors observed a significant decrease of this population upon differentiation-priming treatment in both blood-derived and thp-1 samples. The decrease was concomitant to a significant increase in pre-cDC count in blood-derived samples and to an increase in PDC in thp-1 samples. Based on the current myeloid differentiation map and on their observations the inventors propose that the yet unidentified population could be classified as a circulating common precursor of pre-cDC and PDC, differentiating from the common dendritic cell precursors located in the bone marrow (FIG. 13). Further investigation on this cell population could improve understanding and redesigning of the myeloid differentiation map. Primed cells and myofibroblasts were used for ps-plaque biofabrication and helped to generate a stratified cell-spheroid with myofibroblasts located at the periphery and a compact, collagenous and lipid-rich core of CD45+ cells. The inventors sorted and compared the CD45+ populations derived from ps-plaques and native human carotid plaques using the vi-SNE workflow. They found that PDC and activated dendritic cells (aDC) are the main plaque myeloid component of thin-cap stage atherosclerotic plaque. This finding is per se surprising provided that macrophages and macrophage-derived foam cells are thought to be the main cellular component of atherosclerotic lesions, at least in early developmental stages, as discussed by Moore et al., Cell, 2011, and Randolph et al., Circ Res, 2014. Bonanno et al., Cyometry, 2000, previously analysed the cell component of human carotid plaques using flow cytometry showing that about 17% of the lesion (considering cells of lymphoid and myeloid origin and smooth muscle cells) was constituted by CD68+ cells. The inventors also reported that about 40% of the cells within the plaque expressed MHC class II molecules (HLA-DR+) suggesting that they could act as antigen-presenting cells. It was also proven that early-committed immature DCs are positive for CD68 and HLA-DR markers supporting the idea that the cells analysed by Bonanno et al., Cytometry, 2000, might have been in part PDC and activated dendritic cells. Additionally, it is known that PDC aggravate atherosclerotic lesion formation and their depletion reduces aortic plaque growth by 46% in Apoe.sup./ mice. PDC are also able to uptake oxidized LDL (ox-LDL) ex vivo, and promote PDC-driven antigen-specific T-cell proliferation. Finally, it was reported that PDC function and cytokine release is impaired in patients suffering from coronary artery disease. Taken together, these discoveries are in line with our findings and might change the scenario of future atherosclerosis treatments. Within the PDC population the inventors identified 3 overrepresented sub-populations that they named type-, type- and type-. The inventors observed that these subpopulations retain some degree of specialization due to differential marker expression levels. This difference could be the basis of a differential contribution to plaque maturation. For instance, PDC type- display surface marker expression levels (CD36.sup.high, CD14.sup.high and SRA-1.sup.high) of a specialized scavenger population. With the ps-plaque model the inventors were able to investigate LDL effects on intra-plaque population remodelling and cell viability. In detail, they monitored LDL-dependent event density shift within the PDC population towards PDC type- phenotype. This shift was not concomitant with the increase in the PDC count implying a PDC polarization towards CD36high and CD16high vi-SNE regions and indicating a possible LDL-triggered acquisition from PDC of scavenger and pro-inflammatory phenotype. Additionally, the inventors found that LDL presence during ps-plaque formation significantly decreased the count of dendritic precursors in both b- and t-plaques and triggered the polarization of aDCs towards CD11c.sup.high CD16.sup.high and CD36.sup.high levels. These findings are supported by the previous observations that LDL and mildly oxidized LDL affect DC maturation and promote pro-inflammatory function. Moreover, the inventors observed an LDL-dependent decrease of plaque cell viability in b-plaques but not in t-plaques. It was shown that LDL and ox-LDL accumulate in the cytoplasm of the phagocyte and ultimately contribute to a deregulation of lipid metabolism by activating the unfolded protein response (UPR), leading to cell death. The non-significant decrease in cell viability observed in t-plaques might be due to intrinsic differences in population counts among ps-plaque types. In fact, t-plaques show higher intermediate monocyte counts when compared to b-plaques. Furthermore, b-plaques display a larger population of activated dendritic cells compared to t-plaques. In summary, t-plaques are constituted by a more immature cellular milieu compared to b-plaques. For this reason the inventors hypothesize that populations within t-plaques would require more time to develop towards a death-susceptible stage in presence of LDL. Finally, the inventors investigated time-dependent effects of LDL on the expression profile of myeloid cells within the plaque. They found no significant difference in transcript levels of selected pro-inflammatory and remodelling target genes comparing LDL-rich and LDL-free plaques. LDL effects on myeloid cell transcriptome were previously investigated by exposing the cells directly in contact with modified forms of LDL and not by directly testing native lipoproteins. Though the latter might retain slower time of action at the transcriptomic level compared to its modified counterparts, as previously observed on human smooth muscle cells. Interestingly, the inventors detected time-dependent gene induction leading to pro-inflammatory cell phenotype independent to LDL treatment. The latter was either prompted by direct up-regulation of pro-inflammatory target genes in b-plaques (CXCL10, CCR7, IL23, PTGS1) or indirectly triggered by down-regulation of anti-inflammatory genes in t-plaques (CCL17, CCL26, IL10). It was recently shown that three-dimensional spheroid cultures of adipose-derived mesenchymal stem cells (MSC) enhance protein levels of the anti-inflammatory tumor necrosis factor-alpha stimulated gene/protein 6 (TGS-6). On the other hand, the study conducted by Bartosh et al., Proc Natl Acad Sci USA, 2010, did not include any test to verify the possible concomitant release of pro-inflammatory proteins, leaving an unanswered question open for further investigations. In conclusion, the ps-plaque is assembled with myeloid cell populations that are shared with human native plaques. These cells are embedded in a collagenous and lipid-rich extracellular matrix surrounded by a fibrotic layer. To the inventors' knowledge the ps-plaque can be considered the in vitro model closer to human fibroatheroma available up to date.

TABLE-US-00001 TABLE 1 Myeloid populations in ps-plaques and native carotid plaques. Plaque % b- % t- % % nat. % components Abbreviation plaque % total plaque total plaque total Monocytes Classical Mocl 24.92 26.31 34.64 35.14 0.2 0.23 Intermediate Moin 0.97 0.44 0.03 Non classical Monc 0.42 0.06 0.00 Macrophages Derived from MOcl M1 26.66 29.59 38.06 39.18 0.04 0.23 Derived from MOcl M2 1.38 0.11 0.00 Derived from MOin M3 0.27 0.21 0.05 Activated aM 1.28 0.8 0.14 Dendritic Plamacytoid PDC 5.37 43.83 4.37 24.86 86.68 99.54 cells Differentiated PDC Diff PDC 3.85 2.76 4.21 Pre-classical Pre-cDC 5.12 0.31 0.06 Classical cDC 1.6 0.42 1.01 Activated aDC 0.93 0.40 2.27 Activated SRA-1.sup.high aDCSRA-1 6.58 4.18 5.15 Progenitors Prog 20.38 12.42 0.16 debris 0.27 0.82 0.00 0.00

[0164] List 1 shows primers. Gene of interest, forward (FW) and reverse (RV) primer sequences are listed.

TABLE-US-00002 SEQIDNO01:CXCL10for GCAAGCCAATDTGTCCACG SEQIDNO02:CXCL10rev ACADTCCTTGCTAACTGCDTCAG SEQIDNO03:CCR7for GAAAGTCCAGAAACTGDCCCACCTGC SEQIDNO04:CCR7rev CCCCTCTGAAGAACCGAACCACTCCD SEQIDNO05:CCL17for CCAGGGATGCCATCGTDDGTAACTGTGC SEQIDNO06:CCL17rev CCTCACTGTCCCTCTTCTTCGTCCCTGGAA SEQIDNO07:CCL26for GCCTGADTGCAGCATCATGATGG SEQIDNO08:CCL26rev CGGATGACAADCAGCTGAGTCAC SEQIDNO09:DC-SIGNfor TCGAGGATACAAGAGCDAGCA SEQIDNO10:DC-SIGNrev AAGGAGCCCAGCCAAGAG SEQIDNO11:IL10for CTGTGAAAACAAGAGCAACCC SEQIDNO12:IL10rev GAAGCTTCTGDCCCTCCC SEQIDNO13:IL23for GCAGATTCCAAGCCTCAGTC SEQIDNO14:IL23rev DCAACATATGCAGGTCCCA SEQIDNO15:PTGS1for CCCCAGTGAATCCCTGDGD SEQIDNO16:PTGS1rev AAGGTGGCADGACAAACTCC SEQIDNO17:ALOX5for CCCCGACDTGAGAAAATCT SEQIDNO18:ALOX5rev GGCTGCACTCTACCATCTCC SEQIDNO19:SRB1for TCCTCACDCCTCAACCCTG SEQIDNO20:SRB1rev TCCCAGTDGTCCAATGCC SEQIDNO21:GAPDHfor GTCACTGGTGGACCTGACCT SEQIDNO22:GAPDHrev ACCTGGTGCTCAGTGTAGCC SEQIDNO23:18Sfor CCCGGGGAGGTAGTGACGAAAAAT SEQIDNO24:18Srev GCCCGCTCCCAAGATCCAACTAC