Skip to content
InterAcTec

We turn cell-cell interactions into a scalable drug-discovery readout.

  • 01Faster than imaging.
  • 02More mechanistic than endpoint cytotoxicity.
  • 03Scalable on existing flow-cytometry infrastructure.
Discuss your applicationExplore the platform

Built on Interact-omics, peer-reviewed in Nature Methods (2025) (opens in a new tab).

Schematic: an effector cell (teal, dotted membrane) in contact with a larger target cell (violet, dashed envelope). The contact zone between them is highlighted with a hatched lens.CONTACT ZONEEFFECTORTARGET
Schematic visualization

The question

Bring functional cell-cell engagement into focus.

Many modern therapies work by bringing cells together: T-cell engagers bridge T cells and target cells, CAR-T cells bind their targets, antigen-presenting cells instruct T cells. For a candidate, the decisive question is easy to ask and hard to measure at scale — which cells actually engage, how often, and how does that change under treatment?

Microscopy shows single contacts in fine detail, for comparatively few cells. High-plex tissue imaging and spatial transcriptomics map which cells sit next to each other in fixed tissue — proximity, not binding. Endpoint assays report the outcome, not the interaction that produced it. InterAcTec reads out the interaction itself — from a signal that standard flow cytometry normally throws away.

Effector cell

One cell, one event. Effector markers only.

Counted as a singlet

Target cell

One cell, one event. Target markers only.

Counted as a singlet

Engaged pair

Two cells pass the laser as one event: both marker sets, and a shifted forward-scatter ratio (FSC-A/FSC-H).

Usually gated out as a “doublet”

→ Resolved as an interacting pair

Schematic visualization · dashed line = laser interrogation pointMethod: Vonficht et al., Nature Methods 2025

The platform

From cell-cell engagement to a discovery readout.

  1. 01 · Encounter

    Two populations. One question.

    Effector and target cells share the sample. What matters for a candidate is which of them actually engage — and how that changes under treatment.

    Schematic: a field of effector cells (teal) and target cells (violet); one effector–target pair in the centre is about to meet.PROXIMITY ONLYENGAGED · CONTACT ZONEEFFECTOR MARKER →TARGET MARKER →BOTH MARKERS · ONE EVENTINTERACTION FREQUENCY BY CONDITIONILLUSTRATIVE DATA — NOT EXPERIMENTAL RESULTSHIGHLOWCONTROLCAND. ACAND. BCAND. C
    Schematic visualization

  2. 02 · Engage

    Proximity is not engagement.

    Cells can sit side by side by chance. The readout focuses on physically interacting pairs and on how their frequency shifts between conditions.

    Schematic: the central pair is in contact and the contact zone is highlighted, while a second pair elsewhere is merely adjacent without a contact zone.PROXIMITY ONLYENGAGED · CONTACT ZONEEFFECTOR MARKER →TARGET MARKER →BOTH MARKERS · ONE EVENTINTERACTION FREQUENCY BY CONDITIONILLUSTRATIVE DATA — NOT EXPERIMENTAL RESULTSHIGHLOWCONTROLCAND. ACAND. BCAND. C
    Schematic visualization

  3. 03 · Read out

    Every engaged pair becomes an event.

    A conjugate passes the cytometer as one event carrying both marker sets. Scatter signature and marker co-expression turn it into a countable, classifiable data point.

    Schematic: the engaged pair becomes a single event in a two-marker plot. Effector-only and target-only events lie along the axes; interacting events carry both markers.PROXIMITY ONLYENGAGED · CONTACT ZONEEFFECTOR MARKER →TARGET MARKER →BOTH MARKERS · ONE EVENTINTERACTION FREQUENCY BY CONDITIONILLUSTRATIVE DATA — NOT EXPERIMENTAL RESULTSHIGHLOWCONTROLCAND. ACAND. BCAND. C
    Schematic visualization

  4. 04 · Compare

    Conditions become comparable.

    Interaction frequencies per cell-type pair line up across compounds, concentrations, donors or time points — a readout you can rank and decide on.

    Illustrative chart, not experimental results: interacting-event frequencies arranged in columns for a control and three candidates.PROXIMITY ONLYENGAGED · CONTACT ZONEEFFECTOR MARKER →TARGET MARKER →BOTH MARKERS · ONE EVENTINTERACTION FREQUENCY BY CONDITIONILLUSTRATIVE DATA — NOT EXPERIMENTAL RESULTSHIGHLOWCONTROLCAND. ACAND. BCAND. C
    Illustrative data — not experimental results

Why InterAcTec

More interactions, more cell types, less effort.

How the readout compares with imaging — from imaging flow cytometry and nanowell microscopy to high-plex tissue imaging and spatial transcriptomics — with sequencing-based mapping and with endpoint assays. In numbers from the published record.

  • 7×faster cell acquisition than imaging flow cytometry35,000 vs 5,000 cells per second
  • ~170×more cell-cell interactions analysed in one experiment than in a published PIC-seq experiment414,564 vs 2,389 (droplet pairing: ~2,000)
  • 3 hof live co-culture before the flow readout — spatial transcriptomics needs up to ~6 days of prep and run0.5–3 h vs 2–3 d prep + < 3 d run (Xenium)
  • 52cell-type pairs resolved in one experiment, on a standard flow cytometervs one value per well in an endpoint kill assay

Figures compare published instrument specifications and reported experiment sizes; they are not a head-to-head benchmark. Details and sources below.

01

Faster than imaging.

Every imaging method — imaging flow cytometry, live-cell and nanowell microscopy, high-plex tissue imaging, spatial transcriptomics — has to capture, segment and classify cells in pictures. The interaction readout reads engaged pairs as events in a standard flow run: no image processing, far more cells per second, and live samples instead of fixed sections.

7–17×higher maximum acquisition rate than imaging flow cytometry

3 h vs 6 dCo-culture before the flow readout: 0.5–3 hours. Spatial transcriptomics (Xenium): 2–3 days of sample prep plus up to 3 days of instrument run — up to ~6 days before data.

Maximum acquisition rate (cells per second, linear scale)
  • Full-spectrum flow cytometer (InterAcTec readout)35,000
  • Imaging flow cytometer, 20× objective5,000
  • Imaging flow cytometer, 40× objective2,000
  • High-plex tissue imaging (PhenoCycler-Fusion)~1,700

    “1 million cells in 10 minutes”, per imaging pass

Manufacturer maximum rates; practical rates depend on sample and panel. Tissue imaging repeats its pass for every marker cycle. Sources: Cytek Aurora specification (35,000 events/s); Cytek Amnis ImageStreamX Mk II specification; Akoya PhenoCycler-Fusion specification sheet.

Time before data, as reported (hours, linear scale)
  • InterAcTec readout0.5–3 h

    Co-culture in the published CAR-T experiments, then a standard flow run

  • Nanowell time-lapse microscopy (TIMING)6 h

    Imaging window — producing 1–2 TB of video to analyse

  • Droplet co-encapsulation killing assay10 h

    Imaging window

  • High-plex tissue imaging (PhenoCycler-Fusion)6–26 h

    Instrument run, depending on markers and tissue size

  • Spatial transcriptomics (Xenium, 480 genes)up to ~6 d

    2–3 days sample prep, then a run of under 3 days

Different methods report different steps, so this is not a head-to-head benchmark. Staining and acquisition add to the InterAcTec time; image analysis adds to the imaging methods. Sources: Vonficht et al., Nature Methods 2025; Romain et al., J Clin Invest 2022 (CAR-T products in nanowells); Lu et al., Bioinformatics 2019 (TIMING 2.0); NK-cell killing in droplets, Scientific Reports 2021; PhenoCycler-Fusion protocol, STAR Protocols 2024; 10x Genomics Xenium Analyzer specification.

Endpoint cytotoxicity

One outcome per well: did target cells die?

Interaction readout

Illustrative data — not experimental results

Illustrative matrix: interaction frequency for three cell-type pairs at four time points. Not experimental results.
Cell-type pairt1t2t3t4
Effector · Targetlowmediumhighhigh
Effector · Bystanderlowlowmediumlow
Bystander · Targetlowlowlowlow

Which pairs engage, how often, and when — plus the signalling state of engaged cells.

02

More mechanistic than endpoint cytotoxicity.

A kill assay tells you whether target cells died. The interaction readout shows who engaged whom, how often and when — including T-cell receptor signalling (phospho-CD247) inside the engaged cells.

52 vs 1cell-type pairs resolved in one experiment, versus one outcome per well

52 pairs: LCMV infection experiment. Sources: Vonficht et al., Nature Methods 2025.

03

Scalable on existing flow-cytometry infrastructure.

No sorter, no sequencing, no reporter mice, no imaging platform. The readout runs on multicolour fluorescence flow cytometers — at costs the authors put orders of magnitude below single-cell genomics.

~170×more interacting cells in one experiment than in a published PIC-seq experiment

Cell-cell interactions analysed individually in one experiment (interacting cells or pairs, linear scale)
  • Interact-omics, LCMV infection time course414,564
  • PIC-seq, T cell–dendritic cell co-culture2,389
  • Droplet co-encapsulation, NK cells and targets~2,000
  • Nanowell microscopy, clinical CAR-T products1,589

Experiment sizes as reported in each publication. Nanowell arrays can hold up to 200,000 wells per experiment; tissue-imaging methods infer proximity rather than count bound pairs, so they are not shown. Sources: Vonficht et al., Nature Methods 2025; Giladi et al., Nature Biotechnology 2020 (PIC-seq); NK-cell killing in droplets, Scientific Reports 2021; Romain et al., J Clin Invest 2022 (CAR-T products in nanowells).

  1. 01

    Your samples

    Co-cultures, PBMCs, bone marrow and other suspensions

  2. 02

    Your flow cytometer

    Multicolour fluorescence instruments

  3. 03

    Your data

    New runs — or existing datasets acquired to the guidelines

  4. 04

    Interaction mapping

    PICtR analysis framework

  5. 05

    Interaction readout

    Frequencies per cell-type pair and condition

Existing infrastructure Added by InterAcTec

Side by side

Contact, proximity or outcome.

Methods that sound alike answer different questions. Every approach has its place — InterAcTec is built for the questions that need the actual engagement, at scale, on existing infrastructure.

  • Physical engagement

    Cells that are bound to each other are detected as a unit — the event a cell-bridging drug is meant to change. Measured in live samples, so it can be followed under treatment.

    InterAcTec · imaging flow cytometry · nanowell microscopy · PIC-seq

  • Proximity in fixed tissue

    High-plex imaging and spatial transcriptomics show which cells sit next to each other in a fixed section. Interactions are inferred from distance or ligand–receptor co-expression, not observed as binding.

    PhenoCycler · IMC · MIBI · COMET · Xenium · CosMx · MERSCOPE

  • Outcome only

    One value per well — whether target cells died or a cytokine was released — without showing which cells engaged to produce it.

    Endpoint cytotoxicity · plate-based release assays

Schematic

Spatial methods infer interactions from proximity or ligand–receptor co-expression; transient interactions in blood cannot be studied with them. Sources: Armingol et al., Nature Reviews Genetics 2021; Vonficht et al., Nature Methods 2025.

Where each method stands.

  • Strength
  • Partial
  • Limitation

  • InterAcTec

    Interact-omics

    What is measured
    Physically engaged cells, across all cell-type pairs
    Sample
    Live suspensions: co-cultures, blood, bone marrow
    Scale, as reported
    >34 M cells and ~415,000 interactions in one experiment
    Time to data
    0.5–3 h co-culture, then a flow run
    Instrument
    Standard flow cytometer

Imaging

  • Imaging flow cytometry

    e.g. ImageStreamX

    What is measured
    Contacts seen in images
    Sample
    Suspensions
    Scale, as reported
    Up to 5,000 cells/s (20×), 2,000 (40×)
    Time to data
    Same-day run, plus image analysis
    Instrument
    Dedicated imaging cytometer

  • Live-cell & nanowell microscopy

    e.g. TIMING, Beacon

    What is measured
    Contact and killing kinetics of single pairs
    Sample
    Fluorescently labelled cells in wells or nanowells
    Scale, as reported
    1,589 CAR-T cells in a clinical study; 500–60,000 cells per Beacon run
    Time to data
    6 h of imaging, 1–2 TB of video per experiment
    Instrument
    Automated microscope or optofluidic system

  • Droplet & nanovial assays

    e.g. droplet pairing, nanovials

    What is measured
    Killing or secretion of single cells in droplets or on antigen-coated particles
    Sample
    Cells encapsulated or loaded into particles
    Scale, as reported
    ~2,000 cell pairs per droplet experiment; nanovials: >1 M events sorted in <1 h
    Time to data
    10 h of droplet imaging; nanovials read out by sorting
    Instrument
    Microfluidics or sorter, plus consumables

  • High-plex tissue imaging

    e.g. PhenoCycler, IMC, MIBI, COMET

    What is measured
    Spatial neighbourhoods — proximity, not binding
    Sample
    Fixed tissue sections (e.g. FFPE)
    Scale, as reported
    Up to 100+ markers; ~380,000 cells in a published dataset
    Time to data
    6–26 h per run; 40-plex in <24 h
    Instrument
    Dedicated imaging platform

  • Spatial transcriptomics

    e.g. Xenium, CosMx, MERSCOPE

    What is measured
    Proximity and ligand–receptor co-expression (inferred)
    Sample
    Fixed tissue sections
    Scale, as reported
    480–5,000 genes per panel
    Time to data
    2–3 days prep, then <3 to <6 days per run
    Instrument
    Dedicated platform

Sequencing & reporters

  • Sequencing of interacting cells

    PIC-seq

    What is measured
    Transcriptomes of sorted interacting pairs
    Sample
    Sorted suspensions
    Scale, as reported
    2,389 interacting cells reported
    Time to data
    Sorting, library prep and sequencing
    Instrument
    Sorter + sequencer

  • Reporter-based labelling

    LIPSTIC

    What is measured
    Labelled contacts for a pre-defined receptor pair
    Sample
    Engineered mice only
    Scale, as reported
    Flow readout
    Time to data
    Requires breeding reporter lines
    Instrument
    Flow + reporter mouse lines

Outcome assays

  • Endpoint cytotoxicity

    e.g. release or kill assays

    What is measured
    Target-cell death only — one value per well
    Sample
    Co-cultures
    Scale, as reported
    Plate-based, many wells
    Time to data
    Depends on assay
    Instrument
    Plate reader

Summary based on the cited publications and specifications; examples are named for orientation, not as endorsements. Sources: Vonficht et al., Nature Methods 2025; Cytek Amnis ImageStreamX Mk II specification; Romain et al., J Clin Invest 2022 (CAR-T products in nanowells); Lu et al., Bioinformatics 2019 (TIMING 2.0); Bruker Beacon optofluidic system specification; NK-cell killing in droplets, Scientific Reports 2021; de Rutte et al., ACS Nano 2022 (nanovials); Akoya PhenoCycler-Fusion specification sheet; PhenoCycler-Fusion protocol, STAR Protocols 2024; Rivest et al., Scientific Reports 2023 (COMET seqIF); 10x Genomics Xenium Analyzer specification; Giladi et al., Nature Biotechnology 2020 (PIC-seq).

How it fits

A new readout. A familiar workflow.

  1. 01

    Prepare

    Any cell suspension compatible with flow cytometry — co-cultures, PBMCs, bone marrow or other liquid samples.

  2. 02

    Acquire

    Multicolour acquisition on a fluorescence flow cytometer, following the framework’s acquisition guidelines.

  3. 03

    Map

    Interacting cells are identified by scatter ratio and co-expression of mutually exclusive markers, and assigned to a cell-type pair.

  4. 04

    Compare

    Interaction frequencies per cell-type pair are compared across conditions, donors or time points.

Panel design, controls and analysis scope are defined together for each application.

Applications

Where engagement is the mechanism.

Each application below was demonstrated in the peer-reviewed study. We scope new applications together with you.

01

T-cell engagers & bispecific antibodies

Research question
Does the molecule bridge T cells and target cells — and how efficiently over time?
Readout
Frequency of T cell–target cell interacting pairs versus control, per time point.
Decision value
Compare candidates, formats or concentrations on engagement itself.

Demonstrated · Blinatumomab (CD3×CD19) in PBMCs and patient bone marrow (Nature Methods, 2025)

02

CAR-T cells

Research question
Do CAR-T cells engage their intended targets, and which other cells do they contact?
Readout
Pairs of CAR-T and target cells, alongside all other interacting cell-type pairs in the sample.
Decision value
Compare constructs or donors on engagement before downstream functional assays.

Demonstrated · Anti-CD19 CAR-T cells with B-cell targets (Nature Methods, 2025)

03

Antigen-specific T-cell responses

Research question
Are T cells engaging antigen-presenting cells in response to a stimulus?
Readout
T cell–APC interaction frequencies across stimulated and control conditions.
Decision value
Read out immune activation at the level of the cellular contact.

Demonstrated · OT-II T cells with splenocytes; CytoStim-stimulated PBMCs (Nature Methods, 2025)

Working on a different modality?

Discuss your application

Beyond discovery

The same readout in patient samples.

Interaction signatures can be measured in clinical material and mined from existing cytometry data — for translational research and biomarker exploration. These are research findings, not a validated diagnostic.

B-ALL · 42 patients

Engagement tracked treatment response

In bone-marrow samples from paediatric patients treated with blinatumomab, drug-induced T cell–B cell interactions were stronger in good responders, and high T cell–myeloid interactions at baseline were associated with therapy failure.

Source · Nature Methods, 2025

IBD · 31 donors

Interaction landscape in blood

PBMCs from healthy controls (n=11), ulcerative colitis (n=9) and Crohn’s disease (n=11): 29.9 million cells analysed, 362,102 interacting cells, with disease-specific differences in selected interaction pairs.

InterAcTec case study — download

JIA · re-analysis

Insights from an existing dataset

Publicly available spectral cytometry data from juvenile idiopathic arthritis re-analysed for interactions: 7.8 million cells, comparing healthy donors, inactive and active disease, blood and synovial fluid.

InterAcTec case study — download

Evidence

A peer-reviewed foundation.

Interact-omics maps cellular landscapes and interactions across immune cell types from cytometry data. The framework and its analysis toolkit, PICtR, are described and validated — including against imaging flow cytometry — in Nature Methods.

Vonficht D, et al. Ultra-high-scale cytometry-based cellular interaction mapping. Nature Methods 22, 1887–1899 (2025).
Read the publication (opens in a new tab)

Demonstrated across four fields

  • Oncology

    42patients with B-ALL

    • T-cell engager (blinatumomab) and anti-CD19 CAR-T cell engagement of B cells
    • Drug-induced T–B interactions were stronger in good responders

    Source · Nature Methods, 2025

  • Immunology

    Allimmune cell types in one panel

    • Antigen-specific T cell–APC engagement
    • TCR signalling (phospho-CD247) measured inside interacting T cells

    Source · Nature Methods, 2025

  • Autoimmunity

    37.8 Mcells across two case studies

    • Juvenile idiopathic arthritis: blood vs synovial fluid, active vs inactive disease
    • Inflammatory bowel disease: ulcerative colitis and Crohn’s disease vs controls

    InterAcTec case studies — download

  • Infectious disease

    ~415,000interactions, 52 cell-type pairs

    • Viral infection (LCMV) time course across lymph node, spleen and bone marrow
    • 34.4 million cells in 36 samples

    Source · Nature Methods, 2025

Contact

Let’s discuss your discovery workflow.

Tell us about your modality, your cells and the decision you need to make. We’ll assess together whether an interaction readout fits.

  • Engagement readouts for T-cell engagers, CAR-T cells and other cell-bridging modalities
  • Re-analysis of existing cytometry datasets
  • Your details are only used to answer your request
or email hello@interactec.bio