Sample Pre-treatment: FFPE¶

Applicable samples: Formalin-Fixed Paraffin-Embedded (FFPE)
Version: v4.0
Last modified: 2026-09-08

What's next?

After completing pre-treatment, proceed to the unified RCA Protocol (Hybridization → Ligation → RCA → Post-fix).

Convention

In the following steps, "Jar" denotes a step performed in a separate EasyDip™ Slide Staining Jar previously decontaminated with RNaseZap®. Unless otherwise indicated, all steps are at room temperature (RT).

1. Baking (~16 h)¶

Goal: Adhere section to slide and stabilize tissue before downstream steps.

Attach: One coverslip with a mounted FFPE section onto a microscope slide, taping the edge close to the sample area where the sample name is written.
Bake: 56°C for 16 h.

2. Deparaffinization (~1 h)¶

Goal: Remove paraffin and hydrate tissue through graded solvents.

Jar #1: Dewaxing agent (or xylene), 10 min x 2.
Jar #2: 100% Ethanol, 10 min.
Jar #3: Methanol--acetic acid 3:1 (v/v), 5 min.
Jar #2: 100% Ethanol, 5 min.
Jar #4: 85% Ethanol, 5 min.
Jar #5: 70% Ethanol, 5 min.
Jar #6: NFW, 3 min.

3. RNA Retrieval (~1 h)¶

Goal: Antigen/epitope retrieval to restore RNA accessibility.

Jar #7: Sodium citrate 0.01 M pH 6 + RVC 1:20 (v/v) (= 10 mM final), pre-warmed at 80°C; incubate 45 min at 80°C in a water bath.

Do not omit the RVC

Ribonucleoside vanadyl complex (NEB S1402S), diluted 1:20 (v/v) into the citrate buffer. This is 45 minutes in a hot, low-salt aqueous buffer — the longest, wettest, warmest window in the entire workflow, and the one where any surviving RNase does the most damage. Carry-over is not a concern: the NFW wash and the full ethanol series follow immediately.

The 200 mM stock diluted 1:20 gives 10 mM final — exactly NEB's stated concentration for maximum ribonuclease protection, so the dilution in the source protocol is not arbitrary.

First use: thaw the sealed vial at 65 °C until it is a clear green-black solution, then aliquot into minimal-headspace vials, seal tightly and freeze immediately at −20 °C.

RVC was specified in the source protocol and in three of our own lab records but was lost when this page was written (restored in v2.0). See the Changelog.

Jar #6: NFW, 5 min.
Jar #5: 70% Ethanol, 5 min.
Jar #4: 85% Ethanol, 5 min.
Jar #2: 100% Ethanol, 5 min.
Air-dry tissue sections at RT.

Note: At this point, sections can be stored dry for several days at +4°C.

Assemble Chamber: Cover each tissue section with a PDMS chamber.

4. Rehydration (~20 min)¶

Goal: Rehydrate tissue for aqueous-based steps.

Incubate: 100% Ethanol, 5 min.
Incubate: 85% Ethanol, 5 min.
Incubate: 70% Ethanol, 5 min.
Incubate: NFW, 5 min.

5. Re-fixation (~15 min)¶

Goal: Restore the crosslinks the 80 °C retrieval deliberately removed, before the protease acts.

Incubate: 4% PFA (fresh) in 1X PBS, 10 min at RT.
Wash: 1X PBS, 2 x 2 min.

Why re-fix tissue that was already formalin-fixed

By this point the section is deliberately less crosslinked than the block was. Heat retrieval works by hydrolysing methylene bridges — that is its mechanism, and it is why probe access is restored. What comes out of Jar #7 is a partially de-crosslinked section.

RNA in FFPE is held by protein–RNA crosslinks and physical entrapment; retrieval degrades both, and the protease in the next step degrades more. Re-fixing here re-anchors what is left before it has to survive ~16 h at 45 °C, formamide stringency washes and an overnight RCA.

This is standard practice in every published padlock/RCA FFPE protocol: 3.7% formaldehyde 10 min (Mignardi 2015), 4% PFA 10 min (Jiang 2023; Rao 2023), 3.7% PFA (van den Heuvel 2018) — all before the protease. 3.7% formaldehyde and 4% PFA are the same fixative: 37% formaldehyde stock diluted 1:10, versus paraformaldehyde depolymerised to ~4% formaldehyde. We use 4% PFA throughout, as in Pre-treatment: FF §2. Bersani 2023 places it after the protease instead, anchoring what survived rather than protecting what is about to be digested; both are defensible and it is an arm of Permeabilization Optimization.

A re-fixed section is harder to digest

Added in v2.0. If signal drops after adopting this step, the pepsin time likely needs to go up, not the fixation removed — this is one of the reasons §6 runs for 30 min. Run the FFPE ladder rather than guessing.

6. Digestion (~40 min)¶

Goal: Enhanced permeabilization for FFPE tissue (longer than FF).

Incubate: 0.1 mg/mL Pepsin (in 0.1 M HCl) at 37°C for 30 min.
Wash: PBS-Tween 0.05% wash 3 x 2 min.

Judge the digestion at the microscope, not by the clock

Adequate digestion looks like clear bare nuclei with intact nuclear membranes and translucent perinuclear spaces. Glassy perinuclear space means under-digested; soft nuclei with a disrupted envelope and enlarged perinuclear space means over-digested. Criteria and the full pilot design: Permeabilization Optimization.

New in v4.0 — the time tripled. Do not carry a v3.0 time forward

v3.0 ran this step for 10 min, which put total exposure at roughly a third of every published FFPE padlock/RCA condition. That was accepted at the time as a deliberate one-variable step, with the page stating that under-digestion was the expected failure mode and that the fix would be to extend the time. This version is that extension. The concentration, acid and temperature are unchanged — only the clock moved.

A section digested for 10 min and one digested for 30 min are not comparable. Check which version a run followed before pooling it with another.

Where 0.1 mg/mL × 30 min comes from

Four independent FFPE padlock/RCA protocols converge on 0.1 mg/mL, and three of the four run it for 30 min. Exposure below is enzyme activity × time, in Sigma units (ΔA₂₈₀ 0.001/min, pH 2.0, 37 °C, haemoglobin — see Pepsin Solution Preparation):

Protocol	Pepsin	Time	≈ U·min/mL
Jiang 2023, clinical FFPE (P7012, ≥2,500 U/mg)	0.1 mg/mL	30 min	~7,500
Carow 2023 (grade not stated)	0.1 mg/mL	20 min	~6,400
Mignardi 2015, FFPE lung tumour	0.1 mg/mL	30 min	~9,600
Bersani 2023 (Roche)	0.1 mg/mL	30 min	—
Ours, v2.1 and earlier	0.4 mg/mL	10 min	~12,800
Ours, v3.0	0.1 mg/mL	10 min	~3,200 — below all of them
Ours, v4.0	0.1 mg/mL	30 min	~9,600

The new condition is bounded on both sides by conditions we have evidence about: it is at the published mode, and it is below the ~12,800 of v2.1 — which the 2026-09 re-attribution found was never demonstrated to over-digest (see the Changelog). Under-digestion was the real risk, and it is the one this closes.

Why the concentration stays at 0.1 mg/mL rather than going back to 0.4. Concentration and time are not interchangeable for tissue digestion: a high concentration for a short time is diffusion-limited, so the section surface and edges digest while the interior lags; a lower concentration for longer lets the enzyme equilibrate through the section before much cleavage happens. On a 5-10 µm section the effect is smaller than on a block, but it is the likely reason the field settled where it did.

Caveat worth knowing: that convergence is in mass, not activity — only one of the four papers names a pepsin grade. Our P6887 (>=3,200 U/mg) at 0.1 mg/mL is at least 1.28x the activity of the one that does (Jiang, P7012, >=2,500 U/mg), so ~9,600 is a floor. Confirm at the microscope with the criteria above rather than trusting the arithmetic; if this section type looks over-digested at 30 min, 20 min (Carow) is the next step down, not a concentration change.

Autofluorescence quenching has moved

The 1% NaBH₄ pre-quench that used to sit between rehydration and digestion was removed in v2.0. Quenching now runs after RCA, per Autofluorescence Quenching. See the Changelog for why.

Next step: RCA Protocol (Hybridization → RCA)

Changelog¶

v4.0 -- 2026-09-08 -- Digestion time raised from 10 to 30 min. Pepsin concentration (0.1 mg/mL), acid (0.1 M HCl) and temperature (37 °C) are unchanged and were re-checked against the literature rather than assumed. MAJOR: exposure triples, from ~3,200 to ~9,600 U·min/mL, so sections either side are not comparable.

What prompted it — a corrected attribution, not a new experiment. v3.0 cut the concentration fourfold because a fresh-frozen run that had been over-digested (0.5 mg/mL instead of 0.4) came out with markedly fewer signal dots, and over-digestion was taken as the mechanism behind the lab's low-yield runs generally. That inference no longer holds. Re-running BZ09 with reduced digestion did not recover the signal; what those runs had in common was an RCA primer synthesised without its 3' phosphorothioate, which lets phi29's proofreading exonuclease destroy free primer overnight and leaves fewer circles primed. The digestion step was not the variable that moved. See the internal note The RCA Primer Needs a 3' Phosphorothioate (lab site → Research).

So v3.0 cut a step that was not the problem, and cut it too far. At 10 min this page sat at roughly a third of every published FFPE padlock/RCA condition — a fact v3.0 stated plainly and accepted as a deliberate one-variable move, on the understanding that the time would be extended if signal dropped. This release is that extension, made on evidence rather than on a failure.

Why 30 min and not 20. Of the four independent FFPE padlock/RCA protocols that converge on 0.1 mg/mL, three run it for 30 min (Mignardi 2015, Bersani 2023, Jiang 2023) and one for 20 (Carow 2023). 30 min puts us at ~9,600 U·min/mL — level with Mignardi, above Jiang's ~7,500, and still below the ~12,800 of v2.1, the condition the re-attribution above found was never shown to over-digest. Both bounds are therefore anchored to something we have evidence about. The §5 re-fixation, which makes a section harder to digest, argues the same direction. 20 min is the step down if a section type looks over-digested at 30; the concentration is not the knob.

The acid and the concentration were re-checked, not carried over. 0.1 M HCl is what the padlock/RCA literature uses without exception — Mignardi, the Ke lab, El-Heliebi, van den Heuvel, Bersani, Carow, Lawrence, Sankowski, Andersson, Kalhor, Roberts & Bassett. The 0.01 M that once appeared on Pepsin Solution Preparation came from an smFISH lineage (Annaratone 2017) and was resolved out in v2.0. 0.1 mg/mL is where all four FFPE padlock/RCA protocols land. Neither needed to move; only the clock did.

Confirm the new condition at the microscope rather than assuming it — the §6 morphology criteria exist for this. Background: the internal evidence review Permeabilization / Pepsin Digestion (lab site → Research), which now carries a correction banner for the same reason.

v3.0 -- 2026-08-18 -- Digestion concentration cut from 0.4 to 0.1 mg/mL. Acid (0.1 M HCl), temperature (37 °C) and time (10 min) unchanged. MAJOR: total exposure is a quarter of v2.1, so data either side is not comparable.

Why 0.1 mg/mL. Four independent FFPE padlock/RCA protocols use it — Mignardi 2015, Bersani 2023, Jiang 2023, Carow 2023. We were at 4x that, delivered for a third of their time. Concentration and time are not interchangeable: high-concentration, short-time digestion is diffusion-limited and over-digests the section surface while the interior lags, whereas lower concentration for longer equilibrates first.

Why the time did not move with it. Those papers pair 0.1 mg/mL with 20-30 min, so 10 min leaves us at about a third of the published exposure, compounded by the §5 re-fixation. That is accepted deliberately: changing one variable keeps the result interpretable, which is the rule stated in Pepsin Solution Preparation and applied throughout this family. Under-digestion is the expected failure mode. If signal drops, extend the time — 20 then 30 min — rather than restoring the concentration.

Comparison against the source protocol, now exact. The supplementary SOP of Annaratone et al. names the same catalogue number we use (Sigma P6887) and states its steps run at room temperature unless otherwise stated, with no exception on the pepsin step. The source is therefore 0.25 mg/mL for 15 min at RT, while v2.1 was 0.4 mg/mL for 10 min at 37 °C — within 7% on mass x time, with the entire real difference being a temperature change that was never recorded as a decision. 37 °C is pepsin's optimum and the temperature its activity unit is defined at.

Confirm the new condition at the microscope rather than assuming it: the §6 morphology criteria exist for this. Background: the internal evidence review Permeabilization / Pepsin Digestion (lab site → Research).

v2.1 -- 2026-08-18 -- Follows RCA Protocol v6.0 and RCA: FFPE v3.0: this page now hands off to hybridization rather than to blocking, and the FFPE-specific RNase H that used to be added downstream is gone. MINOR — nothing performed on this page changed.
v2.0 -- 2026-08-18 -- Three changes: RVC restored, NaBH₄ pre-quench removed, re-fixation added. Digestion acid stated as 0.1 M. MAJOR — the step list changed, data either side is not directly comparable, and anyone trained on v1.0 needs to be told.

1. RVC restored to the retrieval buffer. Ribonucleoside vanadyl complex (NEB S1402S) at 1:20 (v/v) is specified in the source protocol and was transcribed into three of our own lab records (20241014_ffpe-protocol-2024, 20241015_ffpe-cubic-l, 20241222_prism-protocol-updated), all reading Sodium citrate 0.01 M pH 6 pre-warmed at 80 °C (better + RVC 1:20). It was dropped when those records were promoted into this page — an omission, not a decision. It protects the longest, hottest, wettest window in the workflow (45 min at 80 °C in a low-salt aqueous buffer).

2. The §5 NaBH₄ pre-quench is removed. Autofluorescence quenching belongs after RCA, which is what Autofluorescence Quenching v1.1 has said all along — this page was the outlier. Four reasons: the post-amplification fix re-introduces the aldehydes a pre-quench removed; borohydride reduces Schiff bases to secondary amines irreversibly, cementing exactly the crosslinks the retrieval just spent 45 min undoing; the hydrogen evolution it causes risks section detachment at the start of a 24 h+ on-slide workflow; and quenching is most effective close to imaging. The source protocol also runs NaBH₄ after the pepsin, for 15 min — we had it before, for 30. NaBH₄ has no positive effect on rolony formation; it only changes background.

3. Re-fixation added before the digestion — 4% PFA, 10 min, RT. Every published padlock/RCA FFPE protocol crosslinks at this point (Mignardi 2015, Jiang 2023, Rao 2023, van den Heuvel 2018 before the protease; Bersani 2023 after) and we did not. It is needed because of the heat retrieval: 45 min at 80 °C hydrolyses methylene bridges by design, so the section entering the protease step is less crosslinked than the block was. Note this is separate from the post-RCA fix in RCA Protocol, which anchors rolonies; this one anchors RNA.

4. The digestion acid is 0.1 M HCl, matching every protocol page, the padlock/RCA literature without exception, and the FF acid-only step. reagents/pepsin.md had specified 0.01 M since v2.0, so runs that followed the reagent page used pH ≈ 2 and are a comparison boundary against runs after this version.

Also: pepsin concentrations are now stated in mg/mL, not %(w/v) — 0.04% is 0.4 mg/mL, 0.01% is 0.1 mg/mL. No concentration changed. See Pepsin Solution Preparation v3.0.

Provenance, now recorded. This pre-treatment traces to Annaratone et al., Oncotarget 2017;8(11):18680–18698, DOI 10.18632/oncotarget.15727 (Crosetto and Bienko labs), corroborated by Asim et al., Nat Commun 2017, DOI 10.1038/s41467-017-00393-y. One paragraph there accounts for four of this page's steps — methanol–acetic acid 3:1, citrate 80 °C/45 min with RVC, pepsin-in-HCl, and 1% NaBH₄. Its readout is smFISH, not padlock/RCA, so none of it was validated against SplintR or phi29 accessibility. Evidence and the remaining open questions: the internal evidence review Permeabilization / Pepsin Digestion (lab site → Research). - v1.0 -- 2026-01-29 -- First FFPE pre-treatment release: baking, deparaffinization, RNA retrieval, rehydration, NaBH₄ pre-quench and pepsin digestion, split out so it feeds either the sample-agnostic RCA Protocol or RCA: FFPE. Note the Section 5 NaBH₄ step is a pre-treatment quench and is distinct from the post-RCA Autofluorescence Quenching protocol.

Comments

Ask / Feedback