Running a cardiotoxicity screen in six weeks

Someone has handed you a compound set and a date. The screen is cardiac safety, the answer is due in six weeks, and the first problem is not biology, it is that “cardiotoxicity screen” names at least four different experiments with different cells, different plates, different instruments and a ten-fold cost spread between them.

This page takes that request apart, tells you which of the four you are actually being asked for, prices the two that a laboratory can stand up inside six weeks using only figures published by named suppliers, and ends with a specification you can paste into an email.

The short version: if you own a multielectrode array reader, a 48-well proarrhythmia pilot is roughly $1,800 of MODELLED materials and is achievable in six weeks. If you do not own one, six weeks is not enough to buy one, and the correct move is to send the cells out or narrow the question to a plate-reader viability endpoint.

First, decide which of the four screens you are being asked for

The word “cardiotoxicity” covers two unrelated liabilities and two unrelated timescales. Getting this wrong is the most expensive mistake available, because it is made in week one and discovered in week five.

The reason this matters commercially: the first row needs a recombinant channel line and an instrument that has nothing to do with the other three rows, and the fourth row needs no specialist instrument at all. A request for “a cardiotoxicity screen” that is answered with the wrong row produces a technically clean dataset that does not address the liability anyone was worried about.

Ask the requester one question: what decision changes based on the result. If the answer is “whether we take the compound into a QT study”, it is row one or row two. If the answer is “whether we drop it from the series”, row four is usually sufficient and is a quarter of the cost.

Why the field converged on hiPSC-CM plus MEA

Two published multi-site studies are the reason this test system is the default for row two rather than a novelty. A cross-site reliability study run under the CiPA framework put the same compounds through hiPSC-derived cardiomyocyte MEA assays at multiple sites, and an international multi-site study assessed proarrhythmic potential across two commercial hiPSC-CM lines at ten sites. The CiPA initiative itself, described in the 2016 progress update, is the reason a mechanistic in vitro battery exists as an alternative to relying on hERG block alone.

The useful consequence for a buyer is not that the assay is validated for your purpose. It is that the seeding densities, plate formats and endpoint definitions are published, which means a supplier and a customer can write the same thing down. That is the precondition for a comparable quote.

What you are actually buying

Four line items, in descending order of how likely each is to delay you.

Cardiomyocytes: the published prices

These are list prices captured on 2026-09-01 and reproduced from our own price comparison work. They are not quotes.

Two findings drop out of that table before you have chosen a vendor.

Pack size beats vendor choice. The same BPS catalogue number is $405 per million at one million and $230 per million at five million, a 43 percent fall inside one supplier. If you compare Axol at $494 against BPS at $230 you are comparing pack sizes, not companies. Size the whole campaign, then price it. This is the single most consistent finding across our price work and it is covered in detail in the cardiomyocyte price comparison.

Kits are not a markup, they are a different SKU. Axol’s ventricular kit at $1,359 against $494 for cells alone is an $865 difference that buys the matched medium. If you have to buy that medium anyway, the kit is the comparable line; if you have a validated maintenance medium in house, it is not. A quote for cells and a quote for a kit are not the same quote, which is point three of the normalisation worksheet.

On licensing. BPS publishes the clearest statement we have found on any product page: the underlying iPSC technology is licensed from iPS Academia Japan, and commercial use of the cells is not allowed without an appropriate licence from iPS Academia Japan. If your screen supports a commercial programme rather than internal research, resolve that before you order, not after. See cell line licensing for commercial use.

Plates: the number nobody publishes

No multielectrode array vendor publishes a plate list price. What exists are transaction records. Dividing a US federal award obligation by the quantity stated in the award description gives $488.80 per 48-well MEA plate (NIH to Axion BioSystems, 2024-09-19, $24,440 for a stated 10 boxes at 5 plates per box) and $381.00 per plate on a 2021 award for a stated case of 50 CytoView MEA 48-white plates.

Treat both as upper bounds for that buyer on that date rather than as list prices: the obligation may include shipping or unstated line items, and federal buyers may hold negotiated pricing. The full derivation, and the equivalent figures for other platforms, are in what MEA plates and consumables actually cost.

For a plain 96-well imaging or viability endpoint, the plate is ordinary tissue-culture plastic and is not a schedule risk. We do not publish a price for it because we could not verify one from a named source in a form that survives distributor and territory variation.

How many cells: the arithmetic

Seeding densities here are the suppliers’ own published figures for their own products.

Note that both formats land on approximately 50,000 cells per well by completely different routes, one by droplet volume and concentration, one by area and density. That is a coincidence of these two particular formats, not a rule, and it does not transfer to 24-well or 384-well. The general method, with growth areas for every format and the correction factors that sit between the vial label and the well, is in vials to wells.

The correction that catches most people: check whether the vial is labelled as total cells at freeze or as viable cells post-thaw. If it is a freeze-basis count with an 80 percent viability specification, your usable count is 80 percent of the label and the vial count above is wrong by one or two vials.

MODELLED cost: two routes

This is a model. Every price is real and sourced. Every quantity is our assumption. It is not a quotation and no supplier has quoted against it.

Route A, 48-well MEA proarrhythmia pilot

Twelve compounds at three concentrations plus vehicle and positive control, one plate.

The comparison that matters is the cell line. The same 2.76 million cells bought as three Axol ax2508 vials is 3 × $494 = $1,482 against $1,150 for one BPS 5M vial, 29 percent more, plus three thaw events instead of one. Buying the larger vial is usually right when the whole requirement fits inside it, and usually wrong when it does not, because you cannot refreeze the remainder.

Route B, 96-well viability and high-content plate

Ten compounds at eight concentrations in duplicate, one full plate.

Look at the two pack constructions. $1,555 buys exactly enough for one plate with 240,000 cells of margin. $2,300 buys enough for two plates and costs 48 percent more. If there is any chance of a repeat, and on a first-time assay there always is, the second construction is the cheaper one per usable plate, at $1,150 per plate against $1,555.

What the model deliberately does not tell you

It does not tell you the instrument cost, because MEA vendors do not publish one and we will not estimate it. It does not include media, because the figure varies by supplier and configuration and we could not verify a single number. And it does not include your own time, which on a first-time assay is the largest line in the whole exercise.

The six-week timeline, and where it actually breaks

The binding constraint is not the cells. Catalogue cardiomyocytes for normal donor lines run one to three weeks, which fits. MEA plates run two to eight weeks and are frequently made to order, which does not fit reliably. If your plate is not confirmed in stock in week zero, the six-week schedule is aspirational.

The second constraint is the beat-stabilisation window. Cardiomyocytes thawed onto an array need days to form a syncytium and settle to a stable spontaneous rate. Dosing into a drifting baseline generates a compound effect that is partly the culture growing up. Budget baseline recording days explicitly rather than treating them as slack.

The third is analysis, and it is almost always underestimated. MEA field potential data is not a single number per well. If nobody has committed to an analysis plan before dosing, the week-six report becomes a week-nine report.

Assay quality: the number to agree before you start

Whatever the readout, agree the acceptance criterion for the assay itself before the plate is dosed. The standard statistic is Z-prime, defined by Zhang, Chung and Oldenburg in 1999 as a dimensionless measure of separation between positive and negative controls that accounts for both signal window and variability. It is the reason a screening plate needs its controls laid out before the compounds, not after.

Two practical consequences for a first-time cardiac screen:

• A positive control belongs on every plate. Without a compound known to move your endpoint in your hands, a null result is uninterpretable, you cannot distinguish “no effect” from “assay did not work today”.
• Vehicle concentration is a design parameter, not a detail. Fix the final vehicle percentage across every well including controls, and put a vehicle-only region on the plate. A dose-response curve with rising vehicle along it is not a dose-response curve.

What breaks that nobody warns you about

Lot boundaries running through conditions. Two lots of the same catalogue number can differ enough to break a comparison. If the campaign needs more cells than one lot can supply, split the lot boundary across replicates rather than across compounds, and record which wells came from which lot. That is recoverable. Discovering it afterwards is not.

Plate-edge evaporation. Outer wells of a multiwell plate lose more medium over a multi-day culture than interior wells. On a screen this produces a systematic position artefact that looks exactly like a weak compound effect. Either leave the perimeter to buffer, or randomise compound position across the plate, and say which you did in the method.

The instrument recording schedule. A single MEA reader recording 48 wells does so on a timetable. If two experiments in the building want the same instrument in the same week, that is a scheduling conflict discovered at the worst possible moment. Book it in week zero.

Single-use plates. MEA plates are designed for one cell plating. A repeat run is a new plate, at a new lead time. Buy the repeat plate with the first one.

If you do not own the instrument

Six weeks is not long enough to procure a multielectrode array system. Two honest routes exist.

Send the assay out. Published core-facility fee schedules exist and give a shape for what this costs: the Boston Children’s Hospital IDDRC Human Neuron Unit publishes MEA recording at $49.00 internal / $64.00 external academic / $84.00 external industry per recording, high content imaging capture at $162.00 / $211.00 / $276.00 per plate, and a project planning fee of $290.00 / $377.00 / $493.00 per new project. Those are that facility’s published rates for its own customer classes, not a market price, and the unit “per recording” needs pinning down before you can build a total from it. They are useful as a sanity check on a commercial quote.

Narrow the question. Row four of the first table, viability, ATP, LDH, troponin release, high-content imaging, needs a plate reader or an imager, both of which most laboratories already own. It answers a genuinely different question from the MEA route, and it answers it inside six weeks with equipment on site. If the decision the requester needs is “drop it or keep it”, that is frequently sufficient.

The wrong route is to buy an instrument on a six-week deadline. Capital procurement, install, training and a first-pass method development do not compress, and you will spend the money and still miss the date.

The specification to send

Paste this, fill it in, and send the identical document to every supplier. The fields are the ones that make two quotes comparable rather than the ones that make you sound expert.

Cells

• Cell type: human iPSC-derived cardiomyocytes; subtype ventricular / atrial / not constrained
• Quantity: [total cells], and state whether that is total at freeze or viable post-thaw
• Pack size: quote at 1 × 10⁶ and at 5 × 10⁶, and give the price per million at each
• Lot: entire quantity from one lot, reserved, stated on the order confirmation
• Donor: [constrained / unconstrained]; state sex and age if constrained
• Purity specification: cardiac troponin T positive percentage, by the method the supplier uses
• Post-thaw viability: specification floor, method (trypan blue or AO/PI) and timepoint
• Media and supplement: included / excluded; if excluded, quote them as separate lines
• Coating substrate: included / excluded; name the substrate

Format and delivery

• Delivery format: cryopreserved vials on dry ice
• Requested delivery date and requested ship day; no Friday deliveries
• Shipping: quoted as a separate line, with mode and incoterm

Assay context to state, so the supplier can tell you if it is a poor fit

• Plate format and platform: [e.g. 48-well MEA, named platform and plate catalogue number]
• Intended seeding density and route (droplet spotting or area-based)
• Assay window: days post-thaw at which dosing occurs
• Endpoint: field potential duration / beat rate / contractile amplitude / viability

Commercial terms

• Use rights required: internal research only, or fee-for-service, or commercial derivative
• Certificate of analysis and the QC panel included at this price
• Remedy on a lot failing the viability specification, and the claim window in days
• Repeat-order price for the same specification in three months

Ask for the price at both pack sizes even if you think you know which one you want. In our recorded data that single question moves the per-million price more than the choice of supplier does.

Where we fit

We put that specification to named suppliers, normalise what comes back onto one unit and one counting convention, and tell you where the quotes genuinely differ, including whether the media is in or out and whether the whole quantity is available from one lot before you raise a purchase order.

Sometimes the right answer is the cheaper adjacent readout, or, on a six-week deadline, sending the assay out rather than buying an instrument you will still be commissioning when the report is due. Our fee is the same either way.