Efficient, long term production of monocyte-derived macrophages from human pluripotent stem cells under partly-defined and fully-defined conditions

Abstract

Human macrophages are specialised hosts for HIV-1, dengue virus, Leishmania and Mycobacterium tuberculosis. Yet macrophage research is hampered by lack of appropriate cell models for modelling infection by these human pathogens, because available myeloid cell lines are, by definition, not terminally differentiated like tissue macrophages. We describe here a method for deriving monocytes and macrophages from human Pluripotent Stem Cells which improves on previously published protocols in that it uses entirely defined, feeder- and serum-free culture conditions and produces very consistent, pure, high yields across both human Embryonic Stem Cell (hESC) and multiple human induced Pluripotent Stem Cell (hiPSC) lines over time periods of up to one year. Cumulatively, up to ∼3×10(7) monocytes can be harvested per 6-well plate. The monocytes produced are most closely similar to the major blood monocyte (CD14(+), CD16(low), CD163(+)). Differentiation with M-CSF produces macrophages that are highly phagocytic, HIV-1-infectable, and upon activation produce a pro-inflammatory cytokine profile similar to blood monocyte-derived macrophages. Macrophages are notoriously hard to genetically manipulate, as they recognise foreign nucleic acids; the lentivector system described here overcomes this, as pluripotent stem cells can be relatively simply genetically manipulated for efficient transgene expression in the differentiated cells, surmounting issues of transgene silencing. Overall, the method we describe here is an efficient, effective, scalable system for the reproducible production and genetic modification of human macrophages, facilitating the interrogation of human macrophage biology.

Figure 1. Protocols for the long-term production of PSC-MDM which are genetically modifiable.

A) Diagram outlining the defined and quick protocols for the differentiation of MDM from PSC. Using the defined protocol, PSC colonies are cultured on feeder-free Synthemax™ tissue culture plates and uniform EBs are generated using centrifugation of PSC in a 96-well plate (to isolate single EBs) or in a AggreWell™ plate. The defined EBs are formed in mTeSR™-1 medium and require the addition of three growth factors. Using the quick protocol, PSC colonies are cultured on inactivated feeder cells and EBs are generated using mechanical dissociation and culture in PSC media in ultra-low adherence plates. After 4 days of EB culture, EBs are transferred to regular tissue culture plates and differentiated in X-VIVO™ 15 media, supplemented with M-CSF and IL-3. Non-adherent PSC-MC can be harvested from the supernatant of cultures and plated in X-VIVO™ 15 media+M-CSF without serum, or in RPMI media +10% serum+M-CSF, to generate PSC-MDM. A pure population of genetically modified PSC-MDM can be obtained by transducing PSC with a self-inactivating lentiviral construct (in this case expressing RFP-IRES-Puromycin^R under control of the EF1alpha promoter). Scale bar, 200 µM. Detailed protocol described in materials and methods. B) Characteristics of the defined and quick protocol. C) Histogram represents RFP expression by PSC-MDM differentiated from PSC transduced with EF1a-RFP-IRES-Puromycin^R (black line) or control PSC-MDM (shaded gray).

Figure 2. Monocytopoiesis characterization in differentiation cultures.

A) Time-course analysis of cell surfaces marker expressed during monocytopoiesis. Adherent EBs were harvested at 12, 21 and 33 days after EB formation and stained for various cell surface markers. Immuno-fluorescence dot plot analysis of CD38 and CD34; Thy-1 and CD34; CD45 and CD14. Gates were determined by using the relevant isotype control antibodies. B) Forward Scatter (FSC) and Side Scatter (SSC) dot plot of harvested PSC-MC showing a gate around the homogenous cell population. C) Non-adherent PSC-MC were harvested from the supernatant of differentiation cultures and counted using a cell counter (Chemometec). The media were replaced for repeated PSC-MC harvests over a period of 52 week. Data represent the cumulative number of viable PSC-MC from 6 wells.

Figure 3. PSC-MC characterization compared to b-MC.

A) Morphology of MC. From left to right: Eosin-methylene blue staining of MC from cytospins, Scanning Electron Microscopy, and Transmission Electron Microscopy. Scale bar, 10 µM. D) Diameter of PSC-MC (n = 7) and b-MC (n = 3) were determined by an AO-DAPI stain (Chemometec). Bars represent the mean diameter ± SEM. C) Phenotype of MC. Surface expression of CD14, CD16, CD163, CD86 and MHC II were measured by flow cytometry. Histograms represent surface staining (black line) compared to the isotype control (shaded gray). D) The symbols reflect the relative ratio of the geometric mean fluorescence intensity (MFI) over the isotype control of independent experiments.

Figure 4. PSC-MDM characterization compared to b-MC.

A) Morphology of MDM. From left to right: Representative brightfield image (scale bare 200 µM), Scanning Electron Microscopy and Transmission Electron Microscopy (scale bar, 10 µM) D) Diameter of PSC-MDM (n = 3) and b-MDM (n = 6) were determined by an AO-DAPI stain (Chemometec). Bars represent the mean diameter ± SEM. C) Phenotype of MC. Surface expression of CD14, CD16, CD163, CD86 and MHC II were measured by flow cytometry. Histograms represent surface staining (black line) compared to the isotype control (shaded gray). D) The symbols reflect the relative ratio of the geometric mean fluorescence intensity (MFI) over the isotype control of independent experiments.

Figure 5. PSC-MDM activation.

A) Cytokine production and surface expression of CD206 from resting (top), classically activated (middle) and alternatively activated (bottom) PSC-MDM. Histograms represent surface staining (black line) compared to the isotype control (shaded gray).

Figure 6. PSC-MDM pathogen interactions.

A) Phagocytosis of fluorescent zymosan particles. Representative histrogram showing percentage of zymosan-postive cells. B) Surface expression of the HIV-1 (co-) receptors CD4, CCR5 and CXCR4 were measured by flow cytometry. Histograms represent surface staining (black line) compared to the isotype control (shaded gray). C) Reverse transcription of HIV-1 was measured by detecting late (pol) products by q-PCR after 30 h of infection. Symbols represent the relative mean number of copies of HIV-1 DNA ± SEM of independent experiments (n = 3), normalised to the number of cells using a β-actin control. D) HIV-1 replication was measured by p24 ELISA. Symbols represent the mean p24 concentration at different time points ± SEM of independent experiments (n = 7).

Figure 7. Phenotype of human Pluripotent Stem Cell lines.

A) PluriTest analysis of Illumina HT12v4 transcriptome array data shows iPS-OX1-18, iPS-OX1-19 and iPS-OX1-23 to cluster with pluripotent stem cells (red cloud) and not with partly- or differentiated cells (blue clouds). Each circle represents one hiPSC line. B) Expression of human pluripotent stem cell markers. Surface expression of TRA-1-60 and SSEA-4, as well as total expression of Nanog and Oct 3/4, were measured by flow cytometry. Histograms represent surface staining (black line) compared to the isotype control (shaded gray).

Figure 8. Further phenotype of human Pluripotent Stem Cell lines.

Pluripotent stem cells were harvested after adaptation to matrigel/mTeSR™ and after undirected differentiation for 16d as embryoid bodies (10,000 cells per EB) and expression of transcripts was assessed by qPCR. A) Silencing of transgenes in iPS cell lines (normalised to NHDF transduced with Yamanaka reprogramming retroviruses 5d before harvest, ‘NHDF+Y’, relative to actin). B) Expression of endogenous version of reprogramming genes. C) Expression of additional endogenous pluripotency-associated genes. D) Expression of Endoderm-associated genes. E) Expression of mesoderm-associated genes (BRACHYURY and MSX1) and Ectoderm-associated genes (PAX6 and MAP2). B–E normalised to corresponding HUES2 transcripts (relative to actin).

Figure 9. CytoSNP interrogation of human Pluripotent Stem Cell lines.

KaryoStudio Detected Regions for NHDF-OX1 (parental fibroblasts) and derived hiPSC lines iPS-OX1-18, iPS-OX1-19 and iPS-OX1-23. The chromosomes which contained regions detected by KaryoStudio as deviating from expected (indicated in ‘Found Region’ column by green for amplification, orange for deletion) are shown, as are the X and Y chromosomes (which, being single copy, are Called despite being the expected copy number). Where an amplification has been detected, the SNP B allele frequency can be seen to show a corresponding small triploid region and the Smoothed Log R Ratio shows a small spike. Chromosomes that are not shown here were those (the majority) that had no regions detected by KaryoStudio as deviating from expected, and so are considered karyotypically normal. Note that the small changes detected here are well below the level of detection in other karyotype methods. The genes that these affect are listed in Table 1.