Science1 publisher3 min readPublished
Rare-earth probes catch HER3 receptors in stable pairs that cancer mutations loosen
Broad and MIT researchers watched single receptors on live cells for over 16 minutes and found that normal HER3 forms stable pairs. Cancer-linked HER3 mutations loosened those pairs, and the idea that this frees HER3 to drive signaling is still the team's working model.
The Scientist · Science desk
What happened
- Nanoparticles doped with ytterbium, erbium and thulium tracked single EGFR, HER2 and HER3 receptors on live cells at the same time, each in its own color.
- HER3 was meant to be the negative control, because its kinase domain is too weak to signal alone and the prevailing model treats it as a partner for other receptors.
- Across repeats with different labels and probes, wild-type HER3 pairs still far outlasted unstimulated EGFR pairs.
- EGFR's exon-19 deletion moves the opposite way, making EGFR pairs more stable, with signaling that tracks clinical aggressiveness.
- HER2 mutations only modestly stabilized HER2 pairs, consistent with HER2 cancers usually being driven by gene amplification.
Compiled by The ScientistSomething wrong?How this is made
Why it matters
- capability Researchers can measure how long receptor pairs last on a living cell by following the same molecules across thousands of frames, where dye-based tracking forces them to infer lifetimes from seconds of data.
- constraint Dimer-detection assays that use HER3 as an empty negative control now rest on a baseline that, in this work, forms stable pairs, so their calibration needs rechecking.
- decision Anyone weighing HER3 pair stability as a drug target first needs evidence that loosened mutant pairs raise signaling, a step the team's model so far proposes only as a possibility.
- exposure Labs that adopt the probe would buy heavy rare earths from a market in which China held 85% of oxide output in 2025 and has been widening its export controls.
The probe's advantage is duration. At 100-millisecond resolution, a 16-minute run gives at least 9,600 consecutive frames of the same molecules [1]. A conventional dye loses its signal within seconds [5]. Peng described the problem with short recordings this way: "If you're trying to understand a conversation between two people... if all of a sudden I just cut off this conversation, then you have to guess my answer based on our previous two-minute interaction, which is often impossible to do," he said [6]. Sixteen minutes was where they chose to stop. Peng said the probe can run for hours, and the team stopped once it had captured the long EGFR pairs it was after [7].
Peng's first reaction to the HER3 pairs was suspicion. "The first time we saw HER3 homodimers, we were really puzzled," Peng told GEN. "We thought that maybe this was some experimental artifact." [8] Swapping labels and probes is the obvious test of whether the tag itself is creating the pairs, and the result held through it [10].
The team's working model is that the stable pairs are a signaling-inactive pool that keeps HER3 from partnering other receptors [11]. Under that model, a mutation matters for what it releases. The HER3 mutations studied destabilize the pair, and according to GEN that potentially frees HER3 to form signaling-active heterodimers [13]. For EGFR, Peng said, the more stable the dimer, the more signaling it triggers [12]. For HER3 the model runs the other way, with a stable pair holding the receptor back [11].
The thing this doesn't tell you is whether a loosened HER3 pair actually raises signaling. For EGFR and HER2, the link between pair stability and disease rests on agreement with what the clinic already shows [12] [14]. That agreement is a correlation. GEN's account gives the direction of each effect but not lifetimes in seconds, the cell types used, or which HER3 mutations were tested. I think the imaging result is the firmer finding of the two. HER3 as a drug target still needs the signaling experiment, and GEN itself says only that the findings could help shape new cancer therapies [19].
The instrument is a lab build, with cameras, lenses and lasers bolted directly onto a metal breadboard [20]. Its chemistry depends on heavy rare earths [3]. Ytterbium and erbium are produced in far smaller volumes than light rare earths such as neodymium [15]. According to Benchmark Mineral Intelligence, China controlled 85% of rare earth oxide production in 2025, including an estimated 99% of dysprosium and terbium oxide [16]. Dysprosium oxide in North America averaged 4.4 times the Chinese price that year [17]. China's April 2025 export controls named seven elements [18]. None of the probe's three dopants was on that list [2], and GEN reports that China expanded the controls in October 2025 [18].
What to watch
- Cell experiments testing whether the destabilized HER3 mutant pairs actually increase heterodimer signaling, the step the drug-target idea depends on.
- Independent labs reproducing stable wild-type HER3 pairs with other single-molecule methods, and runs extended to the hours Peng says the probe allows.
- Whether China's October 2025 expansion of export controls, or later rules, covers ytterbium, erbium or thulium.