CoChat

Exploring the Impact of In Vitro-Transcribed mRNA Impurities on Cellular Responses

Open access

The DOI matches its Crossref record; 48 of the 48 references it cites were found in a registry or an index; none of the 31 sources its claims cite is checked for support yet.

Abstract

Advances in mRNA technology have enabled mRNA-based therapies to enter a new era of medicine. Such therapies benefit from a single, standardized in vitro transcription (IVT) manufacturing process applicable to a wide range of targets. This process includes several downstream purification steps, which aim to eliminate impurities that potentially affect safety and efficacy. However, it is not fully understood which impurities are the most critical; hence, some efforts are still needed to establish the correlation between RNA impurities and their role in limiting therapeutic efficacy. To study this relationship, we produced in vitro-transcribed mRNAs using several bacteriophage T7 RNA polymerases, including one wild-type and four engineered variants. Important attributes of the mRNA such as integrity, purity, and functional activity were then measured using advanced physicochemical and cellular assays. For impurities including abortive transcripts, mRNAs containing partial poly(A) tails, and double-stranded (ds)RNA byproducts, structure-function relationships have been established by tracking cellular responses (i.e., protein expression, reactogenicity) in multiple cell models. By varying the T7 RNA polymerase, different levels of sense-antisense dsRNA byproducts were measured by mass photometry, contributing directly to immunological reactogenicity in bone marrow-derived dendritic cells. T7 RNA polymerase differences with regard to short (<20 nucleotides) 3'-loopback dsRNA byproducts were also further investigated using native mass spectrometry by precisely resolving these impurities at the nucleotide level. Overall, this study highlights the importance of developing sensitive and advanced analytical methods to characterize IVT mRNA impurities and understand their interaction with cellular machinery in order to ensure quality control of RNA-based therapies.

Overview

CoChat Summary

TL;DR

Using five T7 RNA polymerases (one wild-type and four engineered mutants), this study links IVT mRNA impurities—abortive transcripts, partial poly(A)/3′-heterogeneity, and dsRNA byproducts—to protein expression and innate immune activation in human and mouse cell models. Engineered T7 RNAPs reduced abortive transcripts and dsRNA/HMW species, improving translation and lowering reactogenicity, while mass photometry and native MS resolved distinct impurity classes.

Key findings

  • T7 RNAP choice changed mRNA purity: abortive-transcript prepeak levels by CGE and IP-RP HPLC were highest for T_WT (12.9 and 11.7%) and lowest for T_E3 (3.6 and 1.5%); postpeak 3′-loopback dsRNA was similar across assays at close to 3.5%.
  • Protein expression correlated with purity: Pearson r values were 0.82 for main peak and 0.84 for prepeak. In CD8+ T cells at day 1, T_WT mRNA gave the lowest zsGreen expression (35–65%) versus mutants (80–95%); T_E2 had the lowest HMW level (4%) and highest expression (≈92% at day 1), a 2-fold improvement over T_WT. eGFP_160 poly(A) gave 60% positive cells at day 1 versus 52% and 44% for 110 and 90 adenosine residues; IP-RP purification improved purity by 25% and expression to >95% at day 1, 60% at day 5, and 15% at day 8.
  • Reactogenicity tracked dsRNA/HMW impurities: T_WT showed the highest dsRNA by ELISA (≈13–30 AU) and the highest BMDC activation, T_E1 and T_E2 were least activating, and T_E3/T_E4 were intermediate; HMW levels by MP/mCE positively correlated with BMDC activation. ELISA did not detect dsRNA for m1Ψ-modified mRNAs except T_WT, while MP showed similar levels to unmodified mRNAs.
  • Native MS resolved short impurities: for mRNA_T (theoretical mass 120.5 kDa), T_WT showed n+1 and n+2 ratios of 0.60 and 0.15, while T_E1 was below 0.05 for both; 3′-loopback species differed by 2517 and 5074 Da, corresponding to 8 and 16 nucleotides, and represented 30–50% of total mRNA species for T_WT, T_E2, and T_E4 but below 5% for T_E1 and T_E3.

Methods

  • IVT reactions used five T7 RNAPs (one WT, four mutants) with PCR or linearized DNA templates encoding zsGreen, mScarlet, eGFP constructs with 90/110/160-adenosine poly(A) tails, and a smaller mRNA_T construct.
  • Cellular models included human monocyte-derived dendritic cells, human CD8+ T cells, and mouse FLT3L-derived bone marrow-derived dendritic cells; DCs received mRNA-LPX complexes (DOTMA/DOPE at a 1.3:2 charge ratio), and moDCs received zsGreen mRNA by electroporation.
  • Analytics included IP-RP HPLC/UV, CGE-LIF, mass photometry, native mass spectrometry, dsRNA ELISA, mCE, flow cytometry, and Luminex; purity/impurity measures were correlated with protein expression and immune activation, including Pearson correlation.

Limitations

  • Stated in the paper: the exact mechanisms behind 3′-heterogeneity and 3′-loopback transcription remain unknown; native MS of large mRNAs is limited by long size, adduct formation, and inherent heterogeneity, with insufficient resolution for zsGreen intact mass and a need for further instrument development.
  • Stated in the paper: the J2 dsRNA ELISA may be insufficiently specific or sensitive for m1Ψ-modified mRNA, or m1Ψ may alter mRNA conformation; CGE cannot distinguish short 3′-loopback dsRNA, mCE showed limited resolution, and the J2 antibody cannot bind dsRNA <40 bp.
  • Evident from the text: functional comparisons used small samples in places (e.g., N = 2 donors for some expression data; N = 6 in 3 experiments for BMDC cytokine data), and the paper notes mouse-to-mouse variability, species-related differences between human and mouse expression hierarchies, and possible sequence-specific effects between zsGreen and mScarlet.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1021/acs.analchem.4c04162

Cites

48 of 48 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Figure 1 Impact of mRNA purity on translation efficiency. Purity determination of zsGreen mRNAs produced using five different T7 RNAPs (one wild-type, T_WT; four engineered, T_E) by CGE-LIF (a, top) and IP-RP HPLC/UV (a, bottom) methods. (b) For each RNAP T7 polymerase (each represented by a different color), two batches of mRNA have been synthesized. Prepeak, main peak, and postpeak regions are represented by dotted, plained, and stripped histograms, respectively, for both CGE-LIF (b, top) and IP-RP HPLC/UV (b, bottom). (c) The zsGreen protein expression levels obtained from Hu moDCs (gMFI) are represented with histograms (N = 2 donors). (d) Representation of correlation between the IP-RP HPLC purity (prepeak %, solid triangles, and main peak %, open triangles) according to the zsGreen protein expression level (gMFI). (e) IP-RP HPLC purity (%) for eGFP mRNA samples with different poly(A) tail lengths (90, 110, and 160 adenosine residues) are represented with histograms (prepeak, main peak, and postpeak regions are represented by dotted, plained, and stripped histograms, respectively). Each eGFP mRNA sample is represented by a color. Panel (f) represents the eGFP protein expression (%) for the mRNA with different poly(A) tail sizes over time (from 0 to 14 days). Histograms show mean ± SEM for N = 2. Correlations are defined by using the Pearson correlation coefficient (r).
Figure 1 Impact of mRNA purity on translation efficiency. Purity determination of zsGreen mRNAs produced using five different T7 RNAPs (one wild-type, T_WT; four engineered, T_E) by CGE-LIF (a, top) and IP-RP HPLC/UV (a, bottom) methods. (b) For each RNAP T7 polymerase (each represented by a different color), two batches of mRNA have been synthesized. Prepeak, main peak, and postpeak regions are represented by dotted, plained, and stripped histograms, respectively, for both CGE-LIF (b, top) and IP-RP HPLC/UV (b, bottom). (c) The zsGreen protein expression levels obtained from Hu moDCs (gMFI) are represented with histograms (N = 2 donors). (d) Representation of correlation between the IP-RP HPLC purity (prepeak %, solid triangles, and main peak %, open triangles) according to the zsGreen protein expression level (gMFI). (e) IP-RP HPLC purity (%) for eGFP mRNA samples with different poly(A) tail lengths (90, 110, and 160 adenosine residues) are represented with histograms (prepeak, main peak, and postpeak regions are represented by dotted, plained, and stripped histograms, respectively). Each eGFP mRNA sample is represented by a color. Panel (f) represents the eGFP protein expression (%) for the mRNA with different poly(A) tail sizes over time (from 0 to 14 days). Histograms show mean ± SEM for N = 2. Correlations are defined by using the Pearson correlation coefficient (r).CC BY-NC-ND · PubMed Central
Figure 2 Analysis of HMWs and their impact on protein expression. (a) Mass photometry profiles are obtained for zsGreen mRNA synthesized using the five T7 RNAPs, including a stress T_WT sample (70 °C, 10 min, black-stripped profile). (b) HMW species (%) are represented for MP (two panels on the left) and mCE (two panels on the right) and for two mRNA sequences: zsGreen and mScarlet (top and bottom panels, respectively). (c, d) Human CD8+ T cells (N = 2 donors: donor#1 is represented on the left and donor#2 on the right) have been transfected with heated (70 °C, 10 min) or nonheated mRNA samples (T_WT, T_E1, T_E2, T_E3, and T_E4). The levels of zsGreen protein expression (% positive cells) are reported over time (from day 1 to day 15). Histograms show mean ± SEM for N = 2.
Figure 2 Analysis of HMWs and their impact on protein expression. (a) Mass photometry profiles are obtained for zsGreen mRNA synthesized using the five T7 RNAPs, including a stress T_WT sample (70 °C, 10 min, black-stripped profile). (b) HMW species (%) are represented for MP (two panels on the left) and mCE (two panels on the right) and for two mRNA sequences: zsGreen and mScarlet (top and bottom panels, respectively). (c, d) Human CD8+ T cells (N = 2 donors: donor#1 is represented on the left and donor#2 on the right) have been transfected with heated (70 °C, 10 min) or nonheated mRNA samples (T_WT, T_E1, T_E2, T_E3, and T_E4). The levels of zsGreen protein expression (% positive cells) are reported over time (from day 1 to day 15). Histograms show mean ± SEM for N = 2.CC BY-NC-ND · PubMed Central
Figure 3 Functional analysis of the different mRNAs derived from different T7 RNAP using the mRNA-LPX complex in BMDCs. (a) Experimental schema showing BMDC, mRNA-LPX preparation, and analytic measurements. See Figure S6 for the gating strategy. (b) Protein expression of the cell surface activation marker CD86 measured on cDC1 BMDCs by flow cytometry (gMFI denotes the geometric mean fluorescence intensity). PBS indicates nontreated BMDC control. The small molecule TLR7/8 agonist R848 was included as a positive control for innate immune activation via an RNA-sensing pattern recognition receptor. (c) Cytokines measured in the cell culture supernatant by Luminex. N = 6 in 3 independent experiments; data points with shared symbols indicate mRNA technical replicates. (d) Scatter plots demonstrating the correlation between innate immune activation indicated by CD86 expression on BMDCs (x-axis) and HMW purity measured by MP or mCE (y-axis). Top, zsGreen; bottom, mScarlet. Representative data from 2 of 3 experiments. (e) Median fluorescence intensity (MFI) of zsGreen (top) or mScarlet (bottom) among fluorescent-protein-expressing cDC2 BMDCs.
Figure 3 Functional analysis of the different mRNAs derived from different T7 RNAP using the mRNA-LPX complex in BMDCs. (a) Experimental schema showing BMDC, mRNA-LPX preparation, and analytic measurements. See Figure S6 for the gating strategy. (b) Protein expression of the cell surface activation marker CD86 measured on cDC1 BMDCs by flow cytometry (gMFI denotes the geometric mean fluorescence intensity). PBS indicates nontreated BMDC control. The small molecule TLR7/8 agonist R848 was included as a positive control for innate immune activation via an RNA-sensing pattern recognition receptor. (c) Cytokines measured in the cell culture supernatant by Luminex. N = 6 in 3 independent experiments; data points with shared symbols indicate mRNA technical replicates. (d) Scatter plots demonstrating the correlation between innate immune activation indicated by CD86 expression on BMDCs (x-axis) and HMW purity measured by MP or mCE (y-axis). Top, zsGreen; bottom, mScarlet. Representative data from 2 of 3 experiments. (e) Median fluorescence intensity (MFI) of zsGreen (top) or mScarlet (bottom) among fluorescent-protein-expressing cDC2 BMDCs.CC BY-NC-ND · PubMed Central
Figure 4 Determination of 3′-heterogeneity and 3′-loopback dsRNA byproducts by native mass spectrometry. (a) Superposition of the deconvoluted spectra of mRNA_T without poly(A) synthesized by 5 different T7 polymerases, showing the presence of n + 1/n + 2 3′-heterogeneity transcript in blue and the presence of 3′-loopback dsRNA in orange with a mass difference of 2517 and 5074 Da corresponding to 8 and 16 nucleotide residues. (b) 3′-Heterogeneity transcripts (AU) are represented for the five different mRNA (n, n + 1, and n + 2 are represented as white, gray, and black dots, respectively). (c) 3′-Loopback dsRNA (AU) are represented (open triangles) for the five different mRNAs.
Figure 4 Determination of 3′-heterogeneity and 3′-loopback dsRNA byproducts by native mass spectrometry. (a) Superposition of the deconvoluted spectra of mRNA_T without poly(A) synthesized by 5 different T7 polymerases, showing the presence of n + 1/n + 2 3′-heterogeneity transcript in blue and the presence of 3′-loopback dsRNA in orange with a mass difference of 2517 and 5074 Da corresponding to 8 and 16 nucleotide residues. (b) 3′-Heterogeneity transcripts (AU) are represented for the five different mRNA (n, n + 1, and n + 2 are represented as white, gray, and black dots, respectively). (c) 3′-Loopback dsRNA (AU) are represented (open triangles) for the five different mRNAs.CC BY-NC-ND · PubMed Central

Verification

Claims the paper makes

0 Valid · 0 Not valid · 20 Not checked · 86 Own results and interpretations

57 more claims

Create a free account to see all 106 claims and to verify the 30 references still open.

See all claims — free account
  • It may state two separate claims as one. 33
  • Probably not a substantial claim. 12

Create a free account to see these claims.

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

3 of 3

Citations

Cites

Cited by

See all 22

Metadata

Authors
Camperi, J., Roper, B., Freund, E., Leylek, R., Nissenbaum, A., Galan, C., Caothien, R., Hu, Z., Ko, P., Lee, A., Chatla, K., Ayalew, L., Yang, F., Lippold, S., & Guilbaud, A.
Affiliations
Genentech (US)
Corresponding author
  • Camperi, Julien
  • Guilbaud, Axel
Venue
Analytical Chemistry
Published
2024-10-24
Volume
96
Issue
44
Pages
17789-17799
ISSN
  • 0003-2700 (ISSN-L)
  • 1520-6882
OpenAlex
W4403737274
Cited by
22
References
46
Citations by year
2026: 12 · 2025: 10
FWCI
2.15
Type
journal article
Language
en
Open access
Locations
  • Analytical Chemistry (publishedVersion, cc-by-nc-nd)
  • PubMed Central (publishedVersion, cc-by-nc-nd)
  • Europe PMC (PubMed Central) (publishedVersion, cc-by-nc-nd)
  • PubMed (publishedVersion)
Topics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › RNA Interference and Gene Delivery
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › RNA and protein synthesis mechanisms
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Molecular Biology Techniques and Applications
Keywords
in vitro transcribed mRNA, mRNA therapeutics, mRNA quality control, structure-function relationship, native mass spectrometry, mass photometry
MeSH terms
Animals, Humans, DNA-Directed RNA Polymerases, RNA, Messenger, Transcription, Genetic, Viral Proteins

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

A language model drew the claims from the paper on 2026-11-21, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-11-21 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.