Liver model suppliers compared: spheroids, perfused chips, hepatocytes and cell lines

The short answer: a plate of 96 assay-ready human liver microtissues with 250 mL of matched medium lists at $4,280 from InSphero — roughly $45 per microtissue — and that is the only public price in the category. Everything else is quote-only, and no honest cross-supplier cost table can be built from what is published.

The second thing to know, and it reorders most shortlists: the four suppliers here are not selling the same kind of object. One sells finished tissue. One sells finished tissue in a different architecture. One sells an empty perfused plate that you fill with your own hepatocytes. One sells hepatocytes in a vial. If you have three quotes that all say “liver model”, establish which of those four you are being quoted before comparing any number.

The comparison

Read the table structurally, not row by row

Finished tissue that arrives already made. InSphero and MatTek HUREL. You buy a plate with living tissue in it and start at dosing. The commercial deal is that the supplier absorbs the aggregation risk and you accept their donor composition and their calendar.

A perfused scaffold plate you seed yourself. CN Bio’s PhysioMimix liver plates are consumables for an instrument. The plate arrives empty. This is a capability purchase, not a tissue purchase.

Hepatocytes as cells. bit.bio’s ioHepatocytes and HepaRG are vial products. Cheap per million by the standards of this category, and completely different from a microtissue.

The honest cost comparison between routes. A ready-to-use spheroid plate at $4,280 includes the cells. A perfused plate quote does not. At 400,000 to 600,000 primary human hepatocytes per well across twelve wells, a single Liver-12 plate consumes roughly 5 to 7 million primary human hepatocytes — a larger line item than most plate quotes, and one that appears nowhere in a plate price. Any comparison of the perfused route against the scaffold-free route that omits the cell input is not a comparison.

The InSphero price, unpacked

At $4,280 for one Akura 96 plate plus 250 mL of medium, with one microtissue per well, the arithmetic is roughly $45 per microtissue. Three things about that figure.

The medium is bundled and not optional. InSphero offers the product only as a bundle including the microtissues and the appropriate medium type at 250 mL. So $4,280 is the complete unit, not a tissue price with a medium premium on top, and there is no cheaper cells-only configuration to find.

The medium configuration is a specification choice, not packaging. TOX, AF and INF are separately catalogued at the same price and optimised for different application classes, from safety assessment through to efficacy testing. Ordering the wrong one is an experiment error rather than a cost error, and it is the field to get right in the first email.

Shipping is excluded and calculated separately against plate quantity and destination, with the supplier confirming order details and ship dates before the order is confirmed. Budget it as a real line, particularly outside the US and Europe.

The species panel pricing is worth noting for cross-species metabolism work. Dog at $9,140 and mouse at $9,680 are two-plate bundles, so roughly $4,570 and $4,840 per plate — a modest premium over human, and considerably less than building species-specific liver models in-house.

Note also that the 384-well human product is available on request with no public price. The one published price in this category covers exactly one format.

Donor pooling: what it buys and what it costs you

InSphero’s human product carries the clearest published pooling statement in the category: primary human hepatocytes pooled from ten donors, five male and five female, combined with non-parenchymal cells from a single donor.

What pooling buys. Broad human population response in a single assay. Interindividual variation in hepatic metabolism is large, and a ten-donor pool means a clearance or toxicity result reflects an average rather than one person’s enzyme complement. For safety assessment intended to generalise, that is the right design.

What pooling costs you. You cannot see the variation. If your question is about polymorphic metabolism, a susceptible subpopulation, or donor-to-donor variability in a toxicity signal, a pooled model averages away the phenomenon you are studying. MatTek catalogues HUREL Human alongside HUREL Human Pool precisely because both questions exist, and that choice is the single strongest reason to consider them over a pooled-only product.

The asymmetry that decides some experiments. In the InSphero human product the hepatocytes are pooled from ten donors but the non-parenchymal cells come from a single donor. If your endpoint is immune-mediated — Kupffer cell activation, cytokine release, immune-mediated liver injury — the immunological half of the model is single-donor whatever the pooling on the hepatocyte side.

That is not a criticism of the product; it is a fact about it that determines whether it answers your question, and there is published work directly on the point. Kermanizadeh and colleagues examined inter-individual Kupffer cell variability in the governance of hepatic toxicity in a 3D primary human liver microtissue model — the existence of that study is itself the argument for reading the non-parenchymal donor line on a datasheet rather than skipping it.

The species-specific products invert the pooling instructively: the beagle dog model uses single-donor hepatocytes with single-donor non-parenchymal cells, while the C57BL/6 mouse model uses pooled donors on both sides. That reflects the underlying biology — an inbred mouse strain is genetically uniform in a way a dog colony is not — and it is a reminder to check pooling per product rather than per supplier.

Longevity: the specification that decides which assays you can run

This is the axis on which these models differ most usefully, and two suppliers state it clearly.

InSphero: up to 4 weeks. High viability and consistent albumin secretion maintained for up to four weeks, with stable cytochrome P450 activity over an extended period. Four weeks is what makes repeated-dose studies, long-term liver safety work and chronic exposure designs possible in a spheroid format at all. It also creates scheduling flexibility on arrival — the plate does not force you to start on the day it lands.

CN Bio: at least 4 weeks under perfusion. Different mechanism — continuous perfusion providing oxygen and nutrients maintains cell phenotype and function — but a comparable stated window. The trade is that you supply and seed the cells.

MatTek HUREL: framed as phenotypic stability rather than a day count. The published protocol guidance is more useful than the marketing claim: serum-free dosing medium is recommended for assays up to 7 days, and serum-containing maintenance medium is recommended for long-term assays beyond 7 days. That distinction is a real protocol constraint and it means the two media supplied with a HUREL plate are not interchangeable. The maintenance medium is also what is recommended for the media change on plate arrival — using the dosing medium at that step is a protocol error, not a preference.

What to take from this. If your assay is a 24-to-72-hour cytotoxicity or clearance measurement, longevity is not a differentiator and you should buy on price, availability and donor composition. If your assay runs past a week — repeated dosing, chronic steatosis, slow-onset toxicity, oligonucleotide work — longevity is the gating specification and it narrows the field immediately.

What the literature establishes about spheroid liver models generally

None of the following ranks one supplier against another. All of it is context that should shape what you expect from any of them.

Spheroid hepatocyte models are an established DILI platform, not an emerging one. Bell and colleagues characterised primary human hepatocyte spheroids as a model system for drug-induced liver injury, liver function and disease, and that work underpins much of what this category sells. Proctor and colleagues published on the utility of spherical human liver microtissues for prediction of clinical drug-induced liver injury. If you need to justify the model class in a regulatory-adjacent document, that literature exists and is citable.

Higher-density formats are being validated. A published 384-well physiological liver microtissue microplate system for preclinical hepatotoxicity assessment of therapeutic small molecules demonstrates that the format InSphero sells on request has literature behind it, which is relevant if throughput rather than per-plate price is your constraint.

The non-parenchymal composition is a live variable, not a detail. The Kupffer cell variability work cited above is the clearest demonstration that the single-donor non-parenchymal half of a pooled model can drive the result.

The production calendar is the actual constraint

For finished-tissue routes, the binding constraint is a date rather than a price. Liver microtissues are grown on a weekly production calendar and they do not exist until they are made.

InSphero publishes a production schedule with order-by dates and expected delivery windows, and states that unless otherwise indicated, requests received by the listed order-by date ship on the following Tuesday to US customers or Wednesday to European customers by express delivery, with worldwide delivery within 14 days of order.

The published operational detail is unusually specific and worth acting on:

• Ordering commits you to accepting delivery within the stated expected-delivery window. This is not a shipment you can defer once placed.
• Availability on a desired shipment date must be verified with a representative, and the supplier may adjust the date depending on capacity in a given week.
• Some destinations are excluded from certain weeks, and species-specific microtissues to certain countries require separate deadline confirmation.
• Ordering through Science Exchange and Scientist.com is supported, which matters if your institution routes procurement through a marketplace.

InSphero ships in a temperature- and humidity-controlled system and publishes a specific commitment: all plates shipped had microtissues in every well destined to be filled, confirmed by optical scans of each plate, and at least 95 percent of microtissues are guaranteed available for use on arrival. That figure is checkable and rare in this market, and it tells you what to do on arrival — count the wells, because counting is the acceptance test and the supplier has evidence on their side of it.

Ordering three weeks before you need the tissue is not caution. It is the normal procedure.

A note on plate material, because it changes your data

CN Bio publishes something most vendors leave unstated: its multi-chip consumable plates are made of cyclic olefin copolymer, an amorphous thermoplastic, chosen because it is less adsorbent than commonly used polydimethylsiloxane, so experiments are less prone to drug loss from non-specific binding.

Small-molecule absorption into PDMS is a documented problem in microfluidic culture and it silently changes effective dose. If you are comparing a perfused platform against a static plate, or comparing two perfused platforms, ask what the wetted material is. A supplier that publishes it is telling you something useful, and it is a legitimate technical differentiator rather than a marketing line.

Where each supplier is genuinely strongest

InSphero is strongest on price transparency, scheduling discipline and the availability guarantee. They are the only supplier here with a public price, the only one publishing a production schedule with order-by dates, and the only one attaching a numeric availability commitment to what arrives. The pooling statement is the clearest in the category. Their weakness is inflexibility: a fixed format at one microtissue per well, one pooled human composition, a bundled medium you must choose correctly in advance, a calendar that constrains your start date, and no public price on the 384-well format.

MatTek HUREL is strongest on species breadth and on the single-donor option. Seven animal species plus human and pooled human, with ZooPlates for cross-species panels, is a capability nobody else here matches, and the availability of single-donor human material alongside the pool means both the generalising and the variability question can be asked on one architecture. The USB temperature recorder shipped with each plate, writing a PDF with no software required, is the best cold-chain documentation practice in this comparison. Their weakness is price opacity — nothing is published — and the dosing-versus-maintenance medium distinction is a genuine protocol trap for a first-time user.

CN Bio is strongest where perfusion is the point. Continuous perfusion through a scaffold with microchannels, an enclosed recirculating system per well, at least four weeks of maintained function, and a plate stated compatible with primary cells, iPSCs, immortalised lines, circulating immune cells, precision-cut tissue slices and organoids — that last list is the widest cell-source flexibility in this comparison, and it is the route if you want to run your own cells or your own immune co-culture. Publishing the COC plate material is a mark of technical seriousness. Their weakness is the total cost of entry: an instrument, the plate, and 5 to 7 million primary human hepatocytes per Liver-12 plate, priced as three separate lines, none of them public.

bit.bio ioHepatocytes are strongest on cost per cell and consistency. At roughly 67 per million viable cells they are inexpensive by the standards of iPSC-derived material, and the deterministic manufacturing removes the lot-to-lot variability of primary hepatocytes. Their weakness is that they are cells, not a model — you build the tissue, you take the aggregation risk, and the single iPSC background means no donor variability at all, which is either the feature or the disqualifier depending on your question.

HepaRG is strongest as the inexpensive, inexhaustible workhorse. A differentiated human hepatic cell line with a long publication record removes donor sourcing entirely and makes method development, assay qualification and technician training affordable in a way primary hepatocytes never are. Its weakness is that it is a cell line: it carries one genotype, its metabolic enzyme expression profile is not that of pooled primary hepatocytes, and for a claim intended to generalise across the human population it is not a substitute for pooled primary material at any price. We could not retrieve a public price for it, which limits what this comparison can say about the economics.

Choosing

Specification checklist for a liver model quote

• Model class first: pre-formed scaffold-free spheroid, self-assembling co-culture, perfused scaffold you seed, or hepatocytes as cells. Nothing else is comparable until this is settled.
• Species, and for animal models, the strain.
• Donor pooling on the hepatocyte side and on the non-parenchymal side, separately.
• Which non-parenchymal cells are included — Kupffer, endothelial, stellate — and whether the composition is guaranteed or indicative.
• Longevity stated as days with a named functional readout, not as an adjective.
• Medium configuration, and whether it is bundled or separate. For InSphero, TOX versus AF versus INF is a specification field.
• Medium volume supplied, against your feeding schedule and study duration.
• Plate format: 96-well, 384-well, 12-well perfused, 48-well perfused.
• Microtissues per well and the availability guarantee on arrival.
• Production schedule and order-by date for your target start, confirmed with a named representative rather than read off a published table.
• Shipping cost and route to your specific address, and whether your destination has restricted weeks.
• Wetted plate material, for anything perfused or anything where compound loss matters.
• For the perfused route: the instrument, the plate, and the cell requirement per well, priced as three lines.

Where to go next

Buying liver microtissues and spheroids has the full sourcing detail including QC on arrival and what a purchase does not include. 3D tissue model suppliers compared puts liver alongside skin, gut and airway with formats and price bands. Production calendars and shelf life explains why the date, not the price, is usually the binding constraint in this category.

How we can help

The failure mode here is a calendar collision — a study start date that does not line up with a production week, or a plate that arrives before the assay reagents do. We check the published production schedule against your target start before anything is committed, confirm the medium configuration matches the assay class, get shipping quoted to your actual address rather than assumed, and put the perfused and scaffold-free routes side by side on a cost-per-well basis including the cell input — which is the comparison that decides the question and the one no vendor will run for you.

Tell us the species, the study duration, the readout and the target start date.