Building a blood-brain barrier model

The requirement usually arrives as a permeability question: does the compound get into the brain, or does the antibody, or the nanoparticle. Somebody has decided an in vitro barrier is needed and the first search returns a literature in which reported barrier tightness spans four orders of magnitude, several protocols contradict each other, and one widely used cell type has had its identity publicly disputed.

That is the actual state of the field rather than a bad search, and any page that hands you a single TEER target and a shopping list is hiding it from you.

This page is the sourcing route through it. The short version: an insert-based co-culture is the only BBB route you can assemble entirely from catalogue items with published prices, at roughly $2,000 to $3,000 of MODELLED materials for a 24-insert experiment. Everything tighter or more physiological is either a differentiation project or a quote-only product, and both take months.

Step 1, What are you actually measuring

Four different questions get called “a BBB model”, and the correct answer depends on which one you hold.

Almost every buyer asking for a BBB model needs row one or row two. Rows three and four are real and important, and they are a different budget and a different timeline.

The commercial consequence is direct: rows one and two can be assembled from catalogue cells and catalogue inserts. Rows three and four cannot be assembled from anything with a published price.

Step 2, The TEER question, answered honestly

Transendothelial electrical resistance is the standard barrier readout and it is the number you will be asked for. Here is the position we can defend.

TEER is not a property of your barrier alone. A published review of TEER measurement techniques for in vitro barrier model systems is explicit that the measurement depends on instrument, electrode configuration, chamber geometry, temperature, medium composition and the insert area used for normalisation. Two laboratories measuring the same monolayer will not report the same number unless they have matched all of that.

That is why the literature spans orders of magnitude, and why we will not print a single target value for you to aim at.

What we can give you is the shape of the field and the anchors that exist.

How to set a target without inventing one. Do not adopt a number from a paper. Instead:

1. Fix your measurement conditions first, instrument, electrode chamber, temperature, medium, insert area, and write them into the method.
2. Establish your own historical range on a positive control barrier over several runs.
3. Set your acceptance criterion as a floor relative to your own history, not to the literature.
4. Report paired pre- and post-experiment TEER on every insert, because the question is whether the barrier was intact when dosing started and still intact when it finished.

The paired measurement is the whole discipline. A permeability coefficient reported without the TEER values bracketing it is a number with no evidence that a barrier existed. This is the same logic that applies across barrier tissue generally, see intestinal and gut models, where TEER is treated as the inline acceptance test rather than as a characterisation figure.

Do not use TEER as your only barrier readout. Pair it with a paracellular tracer permeability measurement. TEER is sensitive to ionic permeability and can look acceptable when a monolayer has defects that a tracer will find.

Step 3, The identity controversy you must know about before choosing a route

This is the single most consequential piece of context for anyone about to commit months to an iPSC-derived BBB protocol, and it is routinely omitted from vendor material.

A widely adopted route derives brain microvascular endothelial-like cells from human pluripotent stem cells, described in a 2012 derivation paper and refined in a 2014 retinoic-acid-enhanced multicellular model and a 2016 methods paper. Those protocols produce cells with high barrier resistance and they have been used extensively.

In 2021 a paper argued that pluripotent stem cell-derived cells produced by that class of protocol are epithelial rather than endothelial in identity, and that acquiring vascular fate requires ETS transcription factors. The title states the position directly: pluripotent stem cell-derived epithelium misidentified as brain microvascular endothelium.

We are not going to adjudicate that. What we will say is what it means commercially:

• If your endpoint is paracellular permeability, high barrier resistance may be sufficient for your purpose regardless of how the cell’s lineage is finally classified, but you should expect a reviewer to raise it.
• If your endpoint depends on endothelial identity, transporter expression, immune cell adhesion and transmigration, endothelial signalling, angiogenic behaviour, the controversy goes to the heart of your claim and you need to read both positions before committing.
• If you are buying differentiated cells from a supplier, ask which protocol lineage the product comes from and what marker evidence the supplier publishes for endothelial identity. That is a fair, answerable question and a good supplier will have an answer.

Practical advice: do not start a months-long differentiation project on this route without reading both cited papers. They are short and the disagreement is legible to a non-specialist.

Step 4, What is actually purchasable, with prices

We checked every route on 2026-09-01. The honest position, split by evidence class.

Cells with published prices

The population doubling specification is the most under-read line in that table. Fifteen guaranteed population doublings against five is a threefold difference in how much material one vial becomes, and it will change your effective cost per experiment far more than the $57 difference between two vendors’ headline prices. See passage number and population doublings.

Inserts and permeable supports

Insert area is not well area, and this is where BBB seeding calculations go wrong. A 24-well insert is 0.33 cm² of growth area against 1.9 cm² for the well beneath it. Seeding an insert at a density calculated for the well over-seeds by nearly sixfold. The full table of insert areas and volumes is in vials to wells, and the format and membrane selection detail is in Transwell inserts and permeable supports.

Membrane choice is a design decision, not a stock decision. Pore size, pore density and membrane material affect both barrier formation and your ability to image through the membrane. Fix them before you order, because a specific membrane and pore-density combination not held in your territory adds weeks.

Chip and pre-formed routes

The bundle in row two is the cheapest deliberate BBB-specific purchase in the whole survey, and it is a genuine discount rather than a repackaging. If you are evaluating the chip route, it is the entry point.

The pre-formed BBB tissue route is quote only. We will not estimate a price for it. What we can tell you is the shape of the premium in the adjacent category: for the same vendor, an assay-ready angiogenesis plate pre-loaded with hTERT-immortalised HUVECs is $1,470.00 against $444.00 for the empty plate, an uplift of $1,026.00 or 231 percent. That uplift is the honest price of the cells, the matrix and the seeding labour you would otherwise do yourself. Whether it is good value is a question about your technician’s time and your seeding failure rate. See organ-on-chip platform costs.

Step 5, MODELLED cost, insert-based co-culture

This is a model. Every unit price is real, sourced and dated. Every quantity is our assumption. It is not a quotation, and the largest single line, brain microvascular endothelial cells, has no published price and is shown as a gap rather than a guess.

The experiment: 24 inserts, endothelium on the apical side with astrocytes in the basolateral compartment, paired pre- and post-TEER, one tracer permeability endpoint.

Two honest observations about that table.

The plastic is trivial and the cells are everything. $186 of inserts and plate against roughly $1,400 of cells before the endothelium is even priced. Anyone optimising a BBB budget by shopping for cheaper inserts is optimising the wrong line.

The method development configuration is not a BBB model and we have labelled it as such. Running HUVEC on inserts is a legitimate and cheap way to commission your TEER instrument, establish your measurement conditions, and rehearse the insert handling before committing expensive brain endothelium. It is not a barrier model and no result from it should be presented as one. Do the rehearsal, then buy the real cells.

Step 6, The timeline

The insert route is three to six weeks and the differentiation route is months. If you are being asked for a barrier permeability result this quarter, that is the whole decision. See sourcing under a grant deadline.

One instrument question to settle in week zero: does your TEER meter have an electrode chamber matched to your insert diameter. Chamber geometry is one of the variables that makes TEER non-comparable between laboratories, and a chamber that does not match your insert makes your own measurements non-comparable between runs.

What breaks

Using the well area instead of the insert area. A 24-well insert is 0.33 cm²; the well beneath it is 1.9 cm². This single error over-seeds by nearly sixfold and it is the most common mistake in insert-based work.

Reporting permeability without paired TEER. Pre-dose TEER establishes the barrier existed. Post-dose TEER establishes it survived. A permeability coefficient without both is uninterpretable and a reviewer will say so.

Comparing your TEER to a published TEER. Different instrument, different chamber, different medium, different normalisation area. Build your own historical range and compare against that.

Treating HUVEC as a BBB model. They are generic endothelium. They are excellent for commissioning and rehearsal and they are not a brain barrier.

Buying on price per vial instead of per usable doubling. Fifteen guaranteed population doublings against five is a threefold difference in usable material. Check the guarantee, not the headline.

Ignoring the identity controversy on the iPSC route. If your claim depends on endothelial identity, you need a position on it before you generate the data, not after a reviewer asks.

Assuming pericytes are optional because they are hard to source. They may well be optional for your endpoint, but decide that on the biology, not on the fact that we could not find a published price for them. If the endpoint is a disease or inflammatory response, the non-endothelial cells are frequently where the response lives.

Buying an empty chip plate without pricing the seeding. The published assay-ready uplift in the adjacent vascular category is 231 percent, and that uplift is the honest cost of the cells, matrix and labour that an empty plate leaves on your bench. Price your own version of it before concluding the empty plate is cheaper.

The specification to send

The model

• Route: insert co-culture / chip with own cells / pre-formed perfused tissue / in-house differentiation
• Endpoint: paracellular permeability / efflux transporter substrate assessment / barrier response to a disease or inflammatory state / shear-dependent biology
• Cell composition required: endothelium alone, plus astrocytes, plus pericytes, state which and why
• Whether endothelial identity must be evidenced, and which markers you require

Cells

• Endothelial cell type: primary human brain microvascular / immortalised brain endothelial / iPSC-derived brain endothelial-like / HUVEC for method development
• If iPSC-derived: which protocol lineage, and what marker evidence for endothelial identity the supplier publishes
• Astrocytes: species (human iPSC-derived / primary human / rat) and quantity
• Pericytes: required or not; if required, ask for a quote rather than assuming a catalogue item
• Guaranteed population doublings and passage at shipment, ask explicitly, because it varies threefold between products
• Quantity, and whether the count is total at freeze or viable post-thaw
• Lot: entire quantity from one lot per cell type, reserved, in writing
• Media, supplements, growth factors and coating matrix: included or excluded, quoted separately

Format

• Insert format, growth area in cm², membrane material, pore size and pore density
• Companion plate, and confirmation the insert and plate are a matched pair
• Number of inserts including controls and at least one spare set
• TEER electrode chamber geometry matched to the insert diameter

Measurement and acceptance

• TEER instrument, electrode configuration, temperature and medium at measurement, stated
• Normalisation: the insert area used, stated explicitly
• Acceptance criterion expressed as a floor relative to your own historical range, not to a literature value
• Paired pre- and post-experiment TEER on every insert, as a recorded deliverable
• A paracellular tracer permeability measurement alongside TEER

If buying a pre-formed or assay-ready product

• Number of tissue models per plate and the guaranteed proportion available on arrival
• Barrier release criterion for the batch, the instrument it was measured on, and the value for your batch
• Production week and ship day, in writing
• Shelf life, and whether it includes transit
• Remedy if the batch arrives outside its release window, and the claim window in days

Commercial

• Use rights: internal research, fee-for-service, or commercial derivative
• Certificate of analysis and the QC panel included at this price
• Repeat-order price at your projected annual volume

Where we fit

We put that specification to named suppliers and normalise what comes back onto one basis, per insert, per chip, per usable population doubling, which in this category means pricing the seeding labour that an empty-plate quote leaves out and asking for the population doubling guarantee that a headline vial price hides.

The brain endothelial line is the item with no public price, so quote it first. HUVEC belongs in the rehearsal rather than the result. And if your endpoint does not actually require flow, the chip route is a cost you can drop.