iPSC-derived astrocytes, microglia and oligodendrocytes: what to buy and at what ratio

The short answer: human iPSC-derived astrocytes list from about $333 to $1,081 per million cells, microglia from about $333 to $1,043, and oligodendrocyte-like cells and OPCs from about $500 per million across the suppliers that publish prices. The most useful number on this page is not a price, though — it is the ratio.

Published multielectrode array work that achieves strong human network bursting uses a 2:1 neuron-to-astrocyte ratio, seeded at 90,000 neural progenitors to 45,000 astrocytes per well on a 24-well MEA plate, reaching a synchrony index near 0.9 and a mean firing rate around 13 Hz in about six weeks. No supplier publishes that ratio on a product page, and it is the single most common reason a neuronal culture that should have matured on an electrode array does not.

Why glia are usually a companion purchase

Astrocytes are rarely bought on their own. They are bought because a neuronal culture is not bursting, or because a reviewer asked why the model has no glia in it.

That framing matters for procurement, because it means the astrocyte order is constrained by the neuron order that already exists: same delivery week, compatible medium, compatible coating, compatible plate. Buying a good astrocyte from a different vendor with an incompatible medium is a worse outcome than buying a mediocre one that matches. Sort the compatibility first, then the price.

The price table

Supporting media, priced separately:

What the price table tells you

Pack size is again the biggest lever inside a supplier. iXCells astrocytes fall from $1,081 to $722 per million — a 33 percent reduction — between the one-million and two-million vial. bit.bio’s large ioMicroglia pack is exactly price-neutral per million against the small pack (both work out to 333 per million viable cell), which is unusual and means the large pack buys you convenience rather than savings.

bit.bio is roughly a third of the per-million price of the others, but the comparison is not clean: it sells in multi-vial packs, quotes viable cells explicitly, and publishes a single number valid in three currencies. A three-vial pack at 999 is a different purchase from a single vial at $1,081, and if you only need one million astrocytes, the cheaper per-million product is the more expensive order.

Disease and engineered lines diverge in an interesting way. bit.bio prices GFP, CRISPRko, CRISPRa and CRISPRi-ready microglia at the same 999 as wild type — engineered backgrounds are free relative to the base product — while charging a flat 1,298 for every disease line, whether it is an APOE variant, a TREM2 variant or a knockout. iXCells prices its Parkinson’s and PSEN2 astrocytes at $1,246 against $1,081 normal, a much smaller 15 percent premium.

Rodent astrocytes are the honest budget answer for method development. iXCells rat astrocytes are $482 per 0.5 million vial. Where species is not a constraint — qualifying an electrode plate, training a technician, optimising a seeding density — running rehearsals on rodent glia rather than $1,081 human vials is the largest single saving available in this category.

The co-culture ratios: what published work actually used

This is the part that is genuinely hard to find, so here it is with citations attached.

How to read that table

The 2:1 ratio is a starting point with real evidence behind it. Lemieux and colleagues state explicitly that NPCs were co-cultured with primary human astrocytes at a 2:1 ratio “in order to achieve optimal electrophysiological performance”, and reproduced the result across two independent MEA plates with NPCs from separate differentiation passages and astrocytes at different passage numbers, with no statistically significant difference in synchrony index, mean firing rate or number of active electrodes.

The absolute density matters as much as the ratio. 90,000 plus 45,000 in a 24-well MEA well is dense plating, applied as a 12 µL droplet onto the electrode field, incubated an hour, then flooded with 1 mL of medium. The authors note that alterations in cell density have been reported to affect parameters such as interburst interval, and attribute minor plate-to-plate variation to exactly this. A ratio without a density is not a protocol.

The coating is part of the recipe. In that study: 100 µL of 0.1 mg/mL poly-D-lysine on the electrode field at 37 °C for 4.5 hours, three water washes, overnight drying in a biosafety cabinet, then 20 µL droplets of 20 µg/mL laminin at 37 °C for at least an hour. If your co-culture is not attaching, the coating protocol is a more likely culprit than the ratio.

Primary versus iPSC-derived astrocytes is a live question. The Lemieux study used primary human fetal astrocytes and notes that earlier work found NGN2 neurons achieved high firing rates with rodent astrocytes but not with iPSC-derived astrocytes. That is a genuine caveat for anyone assuming an iPSC-derived astrocyte is a drop-in substitute for a primary one, and it is the most important thing on this page that no vendor will tell you. If your model must be fully human and fully iPSC-derived, budget for optimisation time rather than assuming the published ratio transfers.

A working starting protocol

For a 24-well MEA plate with human iPSC-derived neurons, if you have no other information:

1. Coat with poly-D-lysine then laminin, and dry properly between steps.
2. Seed at 2:1 neurons to astrocytes, around 90,000 plus 45,000 per well, as a small high-density droplet on the electrode field.
3. Expect spontaneous firing around 11 DIV with no network activity yet. This is normal and is the point at which most people conclude the experiment has failed.
4. Expect the mean firing rate to climb after roughly 17 DIV, and network bursting and synchrony to develop over the following weeks.
5. Half medium changes every 2 to 4 days.
6. Judge the culture at six weeks, not at two.

Then optimise the ratio for your own cells. Treat 2:1 as the condition to beat, not the answer.

Microglia are a different purchase entirely

Astrocytes are bought to make neurons work. Microglia are bought because they are the experiment.

The catalogue reflects this. bit.bio’s microglia range is the deepest disease panel in the glial category — APOE 4/4 and 4/3 for late-onset Alzheimer’s, TREM2 R47H homozygous and heterozygous, and GPNMB and P2RY12 knockouts — all at a flat 1,298 per two-vial pack, alongside GFP and three CRISPR-ready backgrounds at the wild-type price of 999. Sex is specified as a product attribute rather than a footnote, with separate male and female SKUs at both pack sizes.

Two practical points:

• Microglia media is expensive and separate. Axol prices its microglia media and supplement kit at $515 against $1,043 for the cells — the medium is half the cost of the cells again. The bundled kit at $1,324 saves $234 against buying both separately.
• Microglial phenotype is medium-dependent to an unusual degree. These cells change state readily. Whatever medium the supplier characterised them in is the medium in which their published phenotype holds, and substituting it is not a cost saving.

Oligodendrocytes and OPCs: what is available, and what “-like” means

bit.bio is the only supplier in our set publishing prices for both oligodendrocyte-like cells (io1028S) and OPC-like cells (io1100S), both at 999 per two-vial pack of more than one million viable cells each, with a CRISPRko-ready oligodendrocyte variant at the same price.

Take the “-like” suffix seriously. It is doing real work in that product name. Full myelinating maturity in vitro is difficult and slow, and a supplier calling a product “oligodendrocyte-like” is being accurate rather than modest. If your endpoint is myelination of axons rather than expression of oligodendrocyte markers, ask specifically what myelination evidence exists for the product, at what day, in what co-culture configuration.

The OPC product is the appropriate purchase if your experiment is about the progenitor state — proliferation, differentiation trajectory, or response to differentiation cues — rather than about mature myelinating cells.

The compatibility problem, stated plainly

The recurring failure in this category is not quality and it is not price. It is that neurons and glia are routinely bought from different vendors with incompatible media, and nobody sells the pair as a validated co-culture package with a stated ratio.

Before you place a glial order, check:

• Medium. Can both cell types survive in one medium, or does the protocol require a blend? If a blend, in what proportion, and did either vendor characterise their cells in it?
• Coating. Poly-D-lysine, poly-ornithine, laminin, or a vendor-specific substrate — do the two products want the same one?
• Delivery week. Both cell types need to arrive in a window that fits one seeding day. Two vials a month apart is two experiments.
• Counting convention. A 2:1 ratio calculated from total-cells-at-freeze on one side and viable-post-thaw on the other is not a 2:1 ratio.
• Attachment order. Some protocols seed astrocytes first and let them establish; others seed both together as a master mix, as the Lemieux protocol does. These give different results and the choice should be deliberate.

Where to go next

If the destination is an electrode array, MEA-ready neurons covers the checks that decide whether a neuron product will actually burst, and the post-thaw viability page explains why the cell counts on the two sides of your ratio may not mean the same thing. The cortical neuron sourcing page is the companion purchase for most astrocyte orders.

How we can help

We quote neurons and glia as one order rather than two, which is the only way the media and coating compatibility gets checked before the cells are on a plane. We will normalise the counts to one convention so your seeding ratio is the ratio you intended, align the delivery weeks so both arrive for one seeding day, and ask each supplier the question they answer readily but never publish: what neuron-to-glia ratio and seeding density did they use to generate the data on their own datasheet.

Tell us the neuron subtype, the glial type, the plate format and the target seeding date.