Background
Microglia are highly specialized tissue resident macrophages within the brain. Their homeostatic functions include shaping neural circuits through promotion of neuronal growth and differentiation as well as synaptic pruning. Microglia have also been strongly implicated in a number of neurological diseases and injuries. Most recently, genetic studies have identified many genes that are highly expressed in microglia which are associated with altered risk of developing Alzheimer’s disease (AD), Parkinson’s disease, Frontotemporal Dementia, or Amyolateral Sclerosis [
1‐
4]. These new discoveries have placed microglia and neuroinflammation at the forefront of disease progression emphasizing the need for new model systems to enable the study of human microglia. Yet, microglia have proven to be difficult cells to study given that many differences exist between human and murine microglia [
5]. Additionally, there are significant challenges in isolating and culturing these cells [
6,
7]. Primary human microglia can be isolated in relatively limited numbers from postmortem brain tissue or following surgical resection of brain tumors or epileptic foci. However, given the considerable sensitivity of microglia to environmental changes, samples isolated from patients with neurological disease or following the agonal state prior to death, are likely to be activated and may differ depending on disease state, comorbidities, or cause of death. In order to study a more homeostatic human microglia and to utilize modern experimental manipulations such as CRISPR gene editing, many scientists have instead turned to induced pluripotent stem cells (iPSCs).
In the past 3 years, several labs, including our own, have developed various protocols for differentiating microglia-like cells from pluripotent stem cells [
8‐
13]. While the purity, yield, and reproducibility of these different approaches varies considerably, each of these methods produces myeloid cells that exhibit transcriptional profiles and many key functional or morphological characteristics of human microglia. However, the relatively complex nature of these protocols has made it challenging for labs new to stem cell culture or those lacking fluorescence-associated cell-sorting (FACS) core facilities to quickly adopt these approaches. We have therefore developed an appreciably simplified method (iPS-microglia 2.0) to produce both large numbers and highly purified cultures of human microglia. The resulting cells exhibit RNA transcript profiles that are nearly identical to iPS-microglia generated using our previously published protocol [
8], but provide a significantly increased yield at a reduced cost and omit the prior need for a hypoxic incubator and FACS capabilities, making the protocol more readily accessible for a wider variety of labs.
Discussion
In recent years, the importance of microglia in brain development, homeostasis, and disease has become increasingly clear. Because microglia have been implicated in many neurological diseases and injuries including neurodegeneration, traumatic brain injury, and developmental disorders, several groups have developed methods to try to make these cells more accessible for neurological research. Until recently, microglia could only be studied through brain biopsies, postmortem analysis, or in animal models. Although mouse models of neuroinflammation have been extremely useful in uncovering important new findings, many differences exist between human and murine microglia. For example, one recent study identified several co-regulated myeloid gene expression modules that occur in human AD, but do not occur in AD mouse models [
22]. Likewise, many differences exist between the murine and human complement system that is closely linked to neurodegenerative diseases including AD [
6]. At least two microglial-expressed AD risk genes, CR1 and MS4A4A, have no murine ortholog, further highlighting the challenges of studying the role of microglia in human disease using mouse models alone.
To study human microglia, some highly skilled groups have turned to human biopsy material. Working closely with neurosurgeons, these researchers have developed methods to isolate human microglia from brain tissue removed during a surgical resection of a brain tumor or intractable epileptic foci [
5,
7,
23]. Using this approach, researchers have uncovered exciting data and greatly advanced our understanding of the human microglial transcriptome. However, epileptic foci and tumor tissue induce neuroinflammation and despite best efforts to avoid isolating microglia from ‘diseased-effected tissue’ it is likely that microglia isolated from these patients exhibit considerable variation and alterations in activation state [
24,
25].
Another strategy for studying human microglia involves the isolation of microglia or their nuclei from postmortem brain tissue. Using this approach, researchers have uncovered important age-related differences in the human microglial transcriptome [
26]. Still, it remains unclear whether the agonal state that precedes death, inflammatory co-morbidities, or post-mortem delay might influence microglial gene expression. In the case of Alzheimer’s disease, most patients die from an accompanying infectious disease such as Pneumonia [
27,
28]. Interestingly, animal models of Pneumonia exhibit significant changes in brain microglial activation state [
29,
30]. Thus, it is likely that this and other common infectious co-morbidities can complicate the interpretation and analysis of postmortem-isolated human microglia.
Given the considerable challenges with isolation and study of postmortem or biopsied human microglia, several groups, including our own, developed protocols which utilize the power of stem cells to produce human microglia in vitro [
8]–[
13]
. These methods have begun to enable more detailed mechanistic studies of human microglia by allowing controlled experimental treatments, drug testing, and genetic manipulation. However, the currently existing protocols are relatively complicated and can be challenging to adopt, especially for groups with little prior stem cell experience. Thus, to address this challenge we developed and validated the greatly simplified and refined method presented here. In comparing this new method to our previously published differentiation protocol, we confirm that iPS-microglia 2.0 show highly similar RNA transcript profiles to iPS-microglia as well as primary fetal and adult microglia. In addition, iPS-microglia 2.0 remain distinct from blood monocytes and importantly display largely the same differentially expressed genes between microglia and monocytes as our previously published iPS-microglia.
To further investigate and characterize iPS-microglia 2.0 we functionally validated these cells by examining phagocytosis of three different substrates; Staphylococcus aureus, Zymosan A, and fibrillar beta-amyloid. While each substrate exhibited differential degrees of phagocytosis, these levels were equivalent between our previously described iPS-microglia and iPS-microglia 2.0. Lastly, to determine whether iPS-microglia 2.0 can also be used for in vivo studies, we transplanted microglia derived via both methods into xenotransplantation-compatible MITRG mice, confirming that engraftment, in vivo morphology, and marker expression was equivalent between iPS-microglia and iPS-microglia 2.0. Taken together, these functional and in vivo experiments further support the conclusion that microglia generated via these two methods are virtually identical.
In addition, we tested IDE1 as a small molecule agonist of TGFβ signaling cascades. To this end, we confirmed that substitution of TGFβ1 with IDE1 produced cells that are similar to iPS-microglia 2.0, and additionally highly similar to adult and fetal primary microglia. We have provided differential gene expression analysis to highlight the important differences between IDE- and TGFβ1-treated iPS-microglia 2.0, which researchers should consider when deciding whether to use TGFβ or cost-saving IDE1 for iPS-microglia generation.
Methods
Ethics Statement: All experiments were carried out according to human stem cell (hSCRO) and animal use (IACUC) protocols that were approved by the University of California, Irvine.
Find the complete catalog of materials and catalog numbers in Additional file
7: Table S5.
Simplified differentiation of iPSCs to HPCs
Improved and simplified differentiation of iPSCs to CD43+ primitive hematopoietic progenitor cells (HPCs) is achieved using Stem Cell Technologies STEMdiff™ Hematopoietic Kit (Catalog # 05310). On day − 1, feeder-free iPSCs that have been expanded in TeSR-E8 media are passaged with ReLeaSR (STEMCELL technologies) into mTeSR E8 medium with 0.5 μM Thiazovivin onto matrigel coated (1 mg/mL) 6-well plates (Corning Costar). Small aggregates of ~ 100 cells each are plated at 10–20 aggregates per cm2. The initial plating density is critical as higher density impairs mesoderm differentiation and lower density decreases yield. Thus one can plate iPSCs at 2–3 different densities and select the wells on day 0 that have optimal density to proceed with. When approximately two 100 cell colonies per cm2 have been achieved, replace TeSR-E8 medium with medium A (Basal medium plus Supplement A at 1:200 dilution, 2 mL per well of a 6-well). On day 2 (48 h after original media change), do not fully change media, but rather replace 50% medium A, 1 mL per well of a 6-well. On day 3, carefully remove all media by tilting the plate to one side and aspirating from the edge. Then add 2 mL/well medium B (Basal medium plus supplement B at 1:200). Without removing media, supplement with 1 mL/well of medium B on days 5, 7, 9. On day 10 and again on day and 12, non-adherent cells may be collected. To maintain purity, do not wash cells off the well, but merely remove medium with non-adherent cells carefully and centrifuge 300 x G 5 min. After centrifugation, replace conditioned medium back to each well and add 1 mL fresh medium B if further collection on day 12 will be completed.
FACS analysis has confirmed that these non-adherent cells represent highly pure populations (> 93%) of CD43
+ hematopoietic progenitor cells (Additional file
2: Figure S2). Importantly, simply collecting the floating cells is all that is required to isolate large numbers of highly purified CD43
+ cells. No FACS or MACS isolation is required as identical microglia are produced using any of these three methods (Additional file
1: Figure S1). However, because the cells are not being sorted for purity, the collection of non-adherent cells must be carefully completed. Do not spray medium over adherent cells to wash as this will loosen cells which are not CD43
+ and decrease culture purity.
At this point, HPCs may be frozen at 2–4 million cells per mL in BamBanker (Wako). If frozen, HPCs should be thawed directly into microglial differentiation medium with cytokines (below) and plated onto Matrigel-coated plates at 10,000 cells per cm2. We typically find that viability post-thaw is between 70-95%, with improved viability when greater densities of HPCs are thawed together.
Updated differentiation of CD43+ HPCs to iPS-microglia 2.0
Volumes specified for 35 mm well (1 well of a 6-well plate).
On day 0 of iPS-microglia differentiation, plate HPCs at 10,000 cells per cm2 onto 1 mg/mL Matrigel-coated plates (100,000 per 35 mm well). Plate cells into iPS-microglia medium at 2 mL per 35 mm well: DMEM/F12, 2X insulin-transferrin-selenite, 2X B27, 0.5X N2, 1X glutamax, 1X non-essential amino acids, 400 μM monothioglycerol, 5 μg/mL insulin. Immediately before use, microglial medium should be supplemented with 100 ng/mL IL-34, 50 ng/mL TGFβ1, and 25 ng/mL M-CSF (Peprotech) taken from single-use frozen aliquots (important: do not freeze/thaw these cytokines as it will significantly impair differentiation and yield as well as induce activation. It is crucial to thaw cytokines immediately before use). Throughout the differentiation of HPCs to microglia, these cells will predominantly grow non-adherently. On days 2, 4, 6, 8, and 10, add 1 mL fresh media plus freshly thawed tri-cytokine cocktail. Cytokines are diluted to the concentrations listed above before adding to conditioned medium. Do not fully remove media during the microglial differentiation as the cells secrete paracrine cytokine signals and will not properly differentiate upon removal of those. On day 12, collect 6 mL media from each 35 mm well leaving 1 mL conditioned medium on the plate. Centrifuge non-adherent cells in removed medium for 5 min at 300 x G. Aspirate medium and resuspend non-adherent cells in 1 mL fresh medium plus tri-cytokine cocktail per 35 mm well and add back to the same well which contains the 1 mL conditioned medium. Continue to supplement media (1 mL) on days 14, 16, 18, 20, 22, and 24. On day 25, centrifuge cells leaving 1 mL conditioned media per 35 mm well as on day 12. On day 25, cells should be resuspended in microglia media plus 100 ng/mL IL-34, 50 ng/mL TGFβ1, 25 ng/mL M-CSF, 100 ng/mL CD200 and 100 ng/mL CX3CL1 to further mature microglia and ensure homeostasis. On day 27, feed cells with microglia media with five cytokine cocktail (1 mL per well). On day 28 cells collected for RNA sequencing or use for transplantation or functional assays. If necessary, cells can be maintained for 1–2 additional weeks via media supplementation as above, although longer-term culture is not advised.
Isolating RNA
Total RNA was isolated from cells using RNeasy Mini kit (Qiagen). Approximately 1 million iPS-microglia cells were lysed in RLT buffer and RNA was isolated per manufacturer’s instructions with DNAse treatment (10 min) and increased spin times to maximize yield (16,000 x G for 1.5 min). RNA integrity was measured using the Bioanalyzer Agilent 2100. All libraries were prepared from samples with RNA integrity values ≥9.7. 500 ng RNA per sample was used to create RNA-seq libraries through the Illumina TruSeq mRNA stranded protocol. Each sample was then sequenced in the Illumina HiSeq 4000 platform.
RNA sequencing analysis
RNA sequencing read integrity was verified using FastQC. BBDuk was used to trim adapters and filter out poor quality reads [
31]. Reads were aligned to the GRCh.38.12 human transcriptome using Kallisto [
32]. Lowly expressed genes (expression count summed over all samples < 10) were removed before differential expression analysis. Differential Expression of TPM was calculated using DESeq2 [
33]. An FDR cutoff of 0.001 and fold change of at least 2 was used to determine differentially expressed genes (Additional file
3: Table S1, Additional file
4: Table S2, Additional file
5: Table S3, Additional file
6: Table S4). Visualizations were constructed in part using R in addition to the Genialis visual informatics platform (app.
genialis.com) [
34]. Gene ontology analysis was performed using EnrichR.
Phagocytosis assay
Phagocytic activity of iPS-microglia was examined using the the Amnis Imagestream (Millipore) to combine immunofluorescence and flow cytometry. iPS-microglia or iPS-microglia 2.0 were treated with either 1 μg/mL pHrodo tagged zymosan A beads, 20 μg/mL S. aureus, or 2 μg/mL fluorescent beta-amyloid (Anaspec). After allowing 1 h at 37 degrees for phagocytosis, microglia were resuspended in cold FACS buffer (DPBS, 1% BSA, 0.5 mM EDTA) and stained for 30 min at 4 degrees with 1:100 anti-CD45 (Biolegend, clone HI30) and Zombie-violet live/dead stain. 10,000 events were captured for each sample which were gated for in focus, live cells before analysis. IDEAS software was used to generate masks of internalized signal (substrate within CD45) and percent of cells with internalized substrates were calculated as well as mean fluorescent intensity which remained constant for each cell type.
Animals
All animal procedures were conducted in accordance with the guidelines set forth by the National Institutes of Health and the University of California, Irvine Institutional Animal Care and Use Committee. The MITRG mouse was purchased from Jackson Laboratories (stock #017711); briefly, this strain was developed on a BALB/c background containing two knockouts alleles: Rag2− (Rag2tm1.1Flv) and il2γc− (Il2rgtm1.1Flv); and three humanized knock-in alleles: hCSF-1 (Csf1tm1(CSF1)Flv), h-IL-3/GM-CSF (Csf2/Il3tm1.1(CSF2,IL3)Flv), and hTPO (Thpotm1.1(TPO)Flv). All mice were age and sex matched and group housed on a 12 h/12 h light/dark cycle with food and water ad libitum.
Adult intracranial transplants
All mouse surgeries and use were performed in strict accordance with approved NIH and AALAC-certified institutional guidelines. Direct intracranial injections of iPS-microglia into the cortex and hippocampus were performed on adult MITRG mice. Briefly, adult mice (2–3 months old) were anesthetized under continuous isoflurane and secured to a stereotaxic frame (Kopf). Using a 30-guage needle affixed to a 10 μL Hamilton syringe, mice received 2 μL of mature iPS-microglia suspended in sterile 1X DPBS at 50,000 cells/μL at each injection site. Transplantation was conducted bilaterally in the cortex and hippocampus at the following coordinates relative to bregma: anteroposterior, − 2.06 mm; dorsoventral, − 1.75 mm (hippocampus), − 0.95 mm; mediolateral, ±1.75 mm. Cells were injected at a rate of 50,000/30s with 4 min in between injections. The needle was cleaned with consecutive washes of PBS, 70% (vol/vol) ethanol, and PBS in between hemispheres and animals. Animals were allowed to recover on heating pads before being placed in their home cages and received 2 mg/mL Acetaminophen (Mapap) diluted in water for five days. Animals were perfused 2 months following surgery with 1X PBS followed by 4% paraformaldehyde, entire brains were removed for immunohistochemistry and confocal microscopy.
Immunohistochemistry and confocal microscopy
Fixed half brains were first cryoprotected in a 30% sucrose and 0.05% NaN3 solution in 1X PBS for 72 h. Tissue was then sectioned into 40 μm thick slices on a freezing microtome (Leica SM 2010R), and stored in 0.05% NaN3 solution in 1X PBS as free floating wells. For staining, tissue was blocked for 1 h in 1X PBS, 0.2% Triton X-100, and 10% goat serum. Immediately following blocking, sections were placed in primary antibodies diluted in 1X PBS and 1% goat serum and incubated overnight on a shaker at 4 °C. Sections were labeled with combinations of anti-Ku80 (1:250; Abcam ab79220), anti-Iba1 (1:200; Wako 019–19,741), anti-P2RY12 (1:200; Sigma HPA014518) and mounted with DAPI Fluoromount (SouthernBiotech). Immunofluorescent sections were then visualized and captured using an Olympus FX1200 confocal microscope. Images represent confocal Z-stack taken with equivalent laser and detection settings.
Acknowledgements
The authors would like to thank Eric Martinez for his help with cell culture and RNA isolation The authors would like to thank the Genomics core at UCI particularly Melanie Oakes, Christina Lin, and Seung-Ah Chung. We would also like to thank Dr. Jennifer Atwood of UCI’s Institute for Immunology for her assistance with FACS. Finally, the authors would like to thank Amrita Choudhury from Genialis for her excellent assistance in RNA-sequencing analysis.