How COVID-19 vaccines and related proteins spread over time and body areas in mice and humans

All human tissue samples were processed using 2 mL microtubes prefilled with ceramic beads (Precellys lysing Kit, Bertin Technologies SAS, Montigny-le-Bretonneux, France) and 1 mL phosphate-buffered saline for tissue homogenisation on the Precellys 24 Tissue Homogeniser (Bertin Technologies SAS, Montigny-le-Bretonneux, France). Blood samples were diluted as needed using 0.9% sodium chloride (B. Braun SE, Hessen, Germany). For RNA/DNA extraction, 200 µL of the homogenised tissue lysate/ whole blood was transferred to the MagNA Pure96 (Roche Molecular Systems, CA, USA). Automated nucleic acid extraction was performed according to the manufacturer’s recommendations with whole process control (Roche Control Kit). To monitor the nucleic acid purification and detection processes and to prevent false-negative results, 20 µL of internal control RNA (IC-RNA; Roche Control Kit) were automatically added to each individual sample during the extraction procedure. The final elution volume amounted to 100 µL. BNT162b2-RNA-specific primers and probes were designed at the nucleotide positions in the S gene of SARS-CoV-2, in total alignment with the synthetic construct SARS_CoV_2_ectoCSPP gene (GenBank number: MT380725.1↗)²³. One-step reverse transcription PCR was run on the LightCycler 480 system (Roche Diagnostics International Ltd) by using the S-gene forward (5´-TCT ACC CAG GAC CTG TTC CTG C, 400 nM) and reverse primer (5´-GGG CAG CAC GGG GTT GTC GAA T, 400 nM) and probe (5´ FAM -TGG TGC CGG ACA CGT GGA TGG CGT-BHQ1, 100 nM), and a one-step RNA control kit (Roche Diagnostics International Ltd) as master mix. Briefly, the RT-qPCR master mix consisted of 0.1 µL RT Enzyme Solution, 4.0 µL RT-qPCR Reaction Mix, 1.0 µL RNA Process Control Detection Assay, 1.0 µL forward primer, 1.0 µL reverse primer, 0.5 µL probe, 7.4 µL PCR-grade water, and 5.0 µL RNA template. Thermal cycling conditions were as follows: reverse transcription at 50 °C for 10 min (slope 4.4); pre-incubation at 95 °C for 30 s (slope 4.4); amplification for 45 cycles consisting of denaturation at 95 °C for 5 s (slope 4.4) and annealing/extension at 60 °C for 30 s with single fluorescence acquisition (slope 2.2); followed by cooling at 40 °C for 30 s (slope 2.2). Fluorescence signals were detected in each amplification cycle in the FAM and Cy5 channels using the 3-colour hydrolysis probe filter setting, as recommended by Roche. Ct values were determined using the second derivative maximum method implemented in the LightCycler software (version 5.1). Detection of S-RNA was defined as a positive result by the presence of a sigmoidal amplification curve with a Ct value < 40 in the target channel (FAM). Negative samples were defined by the absence of a sigmoidal amplification curve in the target channel (FAM) in combination with a valid amplification curve in the internal control (IC) channel (Cy5) with Ct values between 30 and 35. Samples lacking an IC signal were classified as invalid. If sufficient material was available, invalid samples were reanalysed, starting the process anew from tissue disruption. Positive and negative controls were included in each RT-qPCR run. Amplification curves were independently reviewed by two trained operators. The samples were analysed once if a valid result (positive or negative) was obtained. For quantification, RNA was extracted from the commercially available BNT162b2 vaccine (BioNTech SE, Mainz, Germany), and absolute copy numbers were determined by droplet digital PCR (ddPCR) using the one-step RT-ddPCR Advanced Kit for probes (Bio-Rad, California, USA) on the Bio-Rad QX100 droplet digital PCR system. A standard curve was prepared in 1:10 dilution steps (n = 5) (Figure S1). BNT162b2 RNA copy numbers are given. Due to substantial inter-individual variability in postmortem tissue properties, tissue-specific limits of detection could not be experimentally determined. For tissue samples, an initial input corresponding to 10 mg (approximately 10⁶–10⁷ human cell equivalents) was targeted in order to avoid saturation of the extraction chemistry. As internal control RNA signals were considered acceptable up to 5 Ct below the expected value (corresponding to an under-quantification of approximately 1.5 log), we estimate that, under the applied conditions, the analytical sensitivity of the assay is on the order of 5 × 10⁴ copies per 10 mg human cell equivalents or per 1 mL of postmortem blood. This value represents an approximation rather than a formally validated tissue-specific LOD.

Tissues were homogenised using 2 mL microtubes prefilled with ceramic beads (Precellys lysing Kit, Bertin Technologies SAS, Montigny-le-Bretonneux, France) and 1 mL TRIzol Reagent (Invitrogen) for tissue homogenisation on the Precellys 24 Tissue Homogeniser. Thereafter, RNA was solubilised and purified through phase separation with chloroform, followed by isopropanol precipitation. The RNA pellet was washed with 75% ethanol, air-dried, and resuspended in RNase-free water. Next, the High-Capacity cDNA Synthesis Kit (Applied Biosystems) was used for cDNA synthesis. The following primers were used to detect vaccine-derived mRNA (BNT162b2: forward: 5′-TACCAAGCTGAACGACCTGT-3′, reverse: 5′-TTGCTGTTCCAGGCAATCAC-3′; mRNA-1273: forward: 5′-CGATCTCTTGTAGATCTGTTCTC-3′, reverse: 5′-GGTGAACCAAGACGCAGTAT-3′)^24,25. Real-time PCR was performed using the Power SYBR^® Green PCR Master Mix (Thermo Fisher Scientific) on the StepOne Plus system (Applied Biosystems). Finally, relative expression was normalised to Gapdh and determined with the 2^−ΔΔCt method. In mice, an indicated RNA sequence was considered detectable when a value above or equal to 0.1 was measured, and undetectable when the value was below 0.1. In this case, the relative RNA expression was set to a detection limit of 0.0001. The threshold value was validated using positive and negative controls.

An aliquot of 50 µL of commercially available BNT162b2 was added to human HEK cells (confluency of 80%) to validate antibodies for the detection of vaccine-related spike protein (Figure S2B). Cells were fixed after 2 days in culture. Subsequently, cells were permeabilised with Triton-X100 at 0.2% in PBS for 5 min, blocked using MAXblock (ActiveMotif, CA, USA), and the BNT162b2 vaccine-related spike protein was stained overnight with different anti-spike antibodies. Specific binding was visualised by indirect detection using a fluorescently labelled secondary anti-mouse antibody, and two antibodies with the best signal-to-noise ratio were selected for all further experiments (anti-SARS-CoV-2 Spike (#GTX632604 and #GTX135356; GeneTex, Irvine, USA). Nuclei were counterstained with DAPI. Representative images were taken on a Leica SP5 Confocal microscope with a 40x objective. Alternatively, primary human muscle satellite cells (#P10976↗, InnoProt, Bizkaia, Spain) were incubated with 50 µL of commercially available BNT162b2, and spike protein was detected as described above after 48 h.

Vaccine mRNA was detected in paraffin sections with the RNAscope^® 2.5 HD Duplex Detection Kit (Advanced Cell Diagnostics, Newark, California, USA) following the manufacturer’s instructions. BNT162b2 mRNA (synthetic construct SARS_CoV_2_ectoCSPP gene; GenBank number: MT380725.1↗) was detected with a specific probe (RNAscope^® Probe - syn-SARSCoV-2-ectoCSPP-C1; #1117951-C1). A probe for the housekeeping gene GAPDH was used to assess RNA integrity (#442201-C1), and the RNAscope^® 3-plex Negative Control Probe (DapB; #320871) was used as a negative control (Figure S2B). RNAscope was performed on deltoid muscle tissue from two individuals with the highest vaccine mRNA levels, as determined by PCR, to assess the distribution of vaccine mRNA. Transcript levels of the housekeeping gene GAPDH were comparable between the two samples, although lower than those of the vaccine mRNA. This is not surprising given the high levels and stability-enhancing modifications of the injected mRNA.

Immunohistochemical stainings were performed using a Ventana BenchMark XT autostainer (Roche Diagnostics Deutschland GmbH, Mannheim, Germany) as previously described²⁶. The specificity of anti-SARS-CoV-2 spike antibodies had been thoroughly tested previously²⁶. In addition, we performed double staining of the most specific anti-SARS-CoV-2 spike protein antibody (#GTX632604) with an anti-SARS-CoV-2 nucleocapsid protein directed antibody (#40143-T62; SinoBiologica) in tissue of a patient with fatal COVID-19 to confirm specificity and proper sub-cellular localisation of spike protein signal (Figure S2A). Moreover, in every round, negative controls were included using the contralateral muscle and/or applying secondary-antibody-only and isotype control antibodies on tissues (Mouse IgG1 Isotype Control, #14-4714-82; eBiosciences, Inc., SanDiego, USA) (Figure S2D + E). Antibodies used in our study were directed against SARS-CoV-2 Spike (#GTX632604 and #GTX135356; GeneTex, Irvine, USA) or HLA-DR/DP/DQ (clone CR3/43, #M0775; Dako/ Agilent, Santa Clara, USA), CD68 (#M0876; Dako/ Agilent), CD8 (clone SP16; #ab101500; Abcam, Cambridge, UK), and CD3 (clone SP7, #ZYT-RBK024-05; Biozol, Germany). Representative images were taken with a Zeiss Axioscope 5 microscope and Axiocam 208 color camera. Immunofluorescence staining was performed as previously described²⁷. Antibodies used for protein-specific detection were SARS-CoV-2 Spike and HLA-DR/DP/DQ or IBA1 (#HS-234308; Synaptic Systems, Göttingen, Germany) or CD14 (#ab133335; Abcam). Detection was performed using fluorescently labelled secondary anti-rabbit, -guinea pig, and -mouse antibodies, respectively. The tyramide signal amplification (TSA) method was used to perform double staining of spike protein and HLA-DR/DP/DQ or CD56 according to the manufacturer’s recommendations (TSA Plus Fluorescein; #NEL741001KT; Akoya Biosciences, Marlborough, USA). Representative images were taken on a Leica SP5 Confocal microscope (provided by UMIF; UKE).

Abundance of spike⁺-cells was assessed in five different fields of view (each 273000 µm²) per mouse muscle in a semi-quantitative manner. Values for scoring were defined as follows: 0 (no spike⁺-cells detectable), 1 (single positive cells with a total of 1–5 spike⁺-cells detectable only within one field of view), 2 (> 5 spike⁺-cells detectable sometimes in small groups in several fields of view), 3 (numerous spike⁺-cells detectable with several cluster of positive cells within different fields of view). Scoring was performed blinded to the experimental groups by an experienced pathologist.

Continuous variables were summarised using mean (SD) and median (IQR), as appropriate. Continuous variables were inspected for approximate normality using histograms. Categorical variables were summarised as numbers (%). Scatter plots were used to illustrate the data. Means and 95% confidence intervals are shown. GraphPad Prism software version 9.1.1 (GraphPad Software, CA, USA) was used for data illustration. Data curation and analysis were conducted using STATA/MP, Version 17.0 (StataCorp, Texas, USA).

The development of mRNA-based vaccines has been a game-changer in the fight against COVID-19. However, as with any novel technology, knowledge gaps remain. To obtain a temporal impression of vaccine kinetics at the vaccination site and in various organs, we intramuscularly injected mice with either BNT162b2 or mRNA-1273 into the right medial gastrocnemius muscle. High BNT162b2 mRNA levels were detectable within the injected muscle tissue by qPCR one day after vaccination. Its abundance decreased over the following days, and by seven days after the initial vaccination, BNT162b2 vaccine mRNA was no longer detectable (Fig. 1A). After vaccine administration, BNT162b2 mRNA was detected widely in the peripheral organs, with pronounced abundance observed in the spleen. However, its levels decreased below the detection limit by days 3 to 4 (Fig. 1B). Following mRNA-1273 vaccination, mRNA kinetics in the right medial gastrocnemius muscle were highest one day after injection, subsequently decreasing day by day (Fig. 1A). Meanwhile, mRNA abundance in peripheral organs was primarily confined to the spleen, remaining consistently high until day 7 (Fig. 1B). To investigate if the vaccine-derived mRNA is translated into protein at the injection site, we assessed the abundance of spike protein expression in muscle tissue by immunohistochemistry (IHC). In accordance with BNT162b2 mRNA abundance, spike protein expression in muscle tissue at the vaccination site was detected and peaked as early as day one (Fig. 1C). Spike protein-positive cell detection decreased during the following time points, and could only be detected in one sample on day 4, but not at later time points post injection of mRNA vaccine (Fig. 1C). A similar spike protein kinetic was detected in the muscle tissue after vaccination with mRNA-1273 (Fig. 1D). When investigating spike protein expression after BNT162b2 or mRNA-1273 vaccination via IHC, we found that fibroblasts and satellite cells were a significant source of the spike protein (Fig. 1E, Figure S3A). In summary, similar kinetics of vaccine-related mRNA and spike protein abundance in the right medial gastrocnemius muscle were observed in mice following vaccination with both mRNA-based vaccines investigated, BNT162b2 and mRNA-1273. Besides fibroblasts, satellite cells presented a source for spike protein expression in the injected muscle in mice.

To examine the local spatiotemporal distribution of mRNA vaccines in humans, we first measured BNT162b2 mRNA by qPCR and spike protein abundance by IHC in the muscle tissue at the injection site of 11 patients who died shortly after receiving BNT162b2 vaccination. The median age of the patients was 87 years (IQR: 78–93), with 64% (n = 7) being female. The median time between death and cooling at 4°C was 5 days (range: 1–9; IQR: 3–6). None of the corpses showed any signs of putrefaction, neither macroscopically nor histologically. Vaccination-related and clinical baseline characteristics are provided in Supplementary Table 1. Importantly, none of the patients showed vaccine-related gross organ pathologies or death in a direct causative relationship to the vaccination (Supplementary Table 1). mRNA (#1: N/A, #02: 2.09 × 10⁷ and #04: 3.50 × 10⁴ copies/10 mg tissue) was detected by qPCR in the injection-side deltoid muscle of two of those 11 patients and in 2 of those 11 patients by in-situ hybridisation, one to three days after they received the vaccination (Fig. 2A). Spike protein expression was found in the deltoid muscles of two patients vaccinated the day before (Fig. 2A). Next, we wanted to investigate whether vaccine mRNA and spike protein could also be detected in major organs or compartments after BNT162b2 vaccination in humans. Interestingly, BNT162b2 mRNA levels in whole blood samples were below the detection limit under the tested conditions in all but one patient, consistent with minimal transient haematogenous dissemination observed in mice. Furthermore, neither mRNA (detected by qPCR) nor spike protein (detected by IHC) could be detected under the tested conditions in any of the major organs examined across all 11 investigated patients (Fig. 2B).

To examine the distribution of BNT162b2 mRNA in muscle tissues of vaccinated patients, we used in-situ hybridisation (ISH). ISH revealed the widespread distribution of BNT162b2-related mRNA in the deltoid muscle of two cases; both individuals had died one day post-vaccination (Fig. 3A). BNT162b2 RNA abundance was detected in a small number of cells in the interstitial space of the skeletal muscle (Fig. 3A). To further assess the cellular targeting of BNT162b2 in human muscle tissue, we performed IHC to detect spike protein. Contrary to previous animal data and recent assumptions^29,28, spike protein expression in the muscle tissue of the patients studied here did not colocalise with antigen-presenting cells in immunofluorescence and confocal microscopy, but rather with fibroblasts and CD56⁺ satellite cells (Figs. 3B and 4A). We then double-stained for macrophages (IBA1) and spike protein-expressing cells but could not detect colocalisation of both populations (Fig. 3C). Interestingly, a similar vaccine-related expression pattern of spike protein, primarily in fibroblasts and satellite cells, was observed in the deltoid muscle of an mRNA-1273-vaccinated patient two days post-vaccination (Supplementary Table 1; Fig. 3B + C). To explore the susceptibility of human muscle satellite cells in more detail, we incubated the commercially available BNT162b2 vaccine with primary human muscle satellite cells in vitro (Figure S3B). Of note, these cells were highly susceptible to uptake of vaccine mRNA and expressed abundant amounts of vaccine-related spike protein (Figure S3B). In the patient vaccinated with mRNA-1273, the injection site beneath the skin was also sampled. Cross-sectioning and staining with an anti-spike antibody revealed that spike protein⁺-cells could also be detected at the injection site. As in the muscle tissue, the majority of spike protein⁺-cells could be identified as fibroblasts according to shape within connective and adipose tissue (Figure S5).

Taken together, BNT162b2 and mRNA-1273-related mRNA and spike protein can be detected at the vaccination site of selected patients within a few days after vaccination, with satellite cells and fibroblasts being a potential human source of spike protein expression limited to a local distribution in the vaccinated muscle in our patient cohort.

We could validate our findings obtained from patients vaccinated with mRNA-based vaccines, when we investigated the deltoid muscle of a patient vaccinated with Ad26.COV2.S (Johnson & Johnson), an adenovirus-based vaccine encoding SARS-CoV-2 spike protein. We also included a second anti-spike antibody for detection to further validate our study (Figure S5). Indeed, a strong and even more widespread expression of spike protein compared to mRNA vaccination could be detected after vaccination with Ad26.COV2.S five days post vaccination (Fig. 3B; Figure S5). Moreover, in accordance with our findings from BNT162b2-vaccinated patients, the majority of spike protein-positive cells could be identified as spike protein⁺ fibroblasts (Fig. 3B; Figure S5) but also CD56⁺ satellite cells (Fig. 4B). Interestingly, we also found the engulfment of Ad26.COV2.S spike protein-positive cells by IBA1⁺ or CD14⁺ cells (Fig. 3C + 4B). When investigating vaccine cell targeting and expression of spike protein with immunofluorescence, spike protein colocalised with a subset of HLA-DR/DP/DQ⁺ cells in the muscle after vaccination with Ad26.COV2.S (Fig. 4B).

In line with local spike protein expression at the vaccination site, we also detected a local inflammatory response consisting of HLA-DR/DP/DQ⁺ antigen-presenting cells, CD68⁺ macrophages, but not CD8⁺ cytotoxic cells, infiltrating the muscle upon mRNA or Ad26.COV2.S vaccination (Fig. 5A). However, local inflammation with an abundance of CD3⁺ T cells at the vaccination site was observed (Fig. 5B). This immune infiltration was missing in the contralateral un-vaccinated control muscle (Fig. 5B, Figure S2D). Taken together, vaccination with mRNA-based BNT162b2, mRNA-1273, and vector-based Ad26.COV2.S leads to a local spike protein expression in human muscle tissue mainly by fibroblasts, satellite cells, and partly antigen-presenting cells and causes a sterile inflammatory response in the muscle.

To elicit an effective immune response by preventing the early degradation of mRNA and facilitating its uptake, mRNA is encapsulated in specific lipid nanoparticles, termed mRNA-lipid nanoparticles (mRNA-LNP)³⁰. Early kinetic studies demonstrated that diverse application routes can lead to a local and systemic protein expression of the mRNA-encoded protein in mice³¹. In line with these previous findings, a report using radiolabelled mRNA-LNP detected lymphogenic spread of such mRNA upon intramuscular injection, not only in the muscle itself, but also to the paraaortic lymph node compartments during the first days of vaccination¹⁶. In a recent study, mRNA could be detected in muscle, draining lymph nodes, spleen, and liver of vaccinated rats and non-human primates upon mRNA-LNP vaccination³². These findings are in line with our results, which also demonstrate mRNA detection in muscle, liver and spleen in experimental mice after vaccination. More importantly, this and our study both demonstrate that mRNA detectability in the primary injection site recedes 7 days after vaccination³². Our study furthermore expands the organ mapping after mRNA-LNP vaccination of mice and demonstrates that after 7 days, mRNA can only be detected in the spleen of mice being vaccinated with mRNA-1273, but not in any other organ. Since both vaccines BNT162b2 and mRNA-1273 were administered to the mouse muscle in identical volumes, these differences in spatial and temporal organ distribution might be attributed to subtle differences in mRNA vaccine manufacturing such as LNP composition, nucleotide usage, cellular uptake preferences, or stability.

In humans, biodistribution studies of vaccine-derived mRNA remain scarce. A longitudinal study demonstrated that mRNA could not be detected any longer in the blood of healthy volunteers with a mean age of 42 years (range 24–70) for 28 days after mRNA-based vaccination³³. Furthermore, another study could detect spike protein and its encoding mRNA in the draining axillary lymph nodes of vaccinated individuals for up to 60 days³⁴. In our study, however, detection of mRNA in humans after vaccination was limited to the local injection site, while we could not detect mRNA in draining lymph nodes or distant organs. The median age of individuals in this study was 87 years, and many were likely affected by immunosenescence and multimorbidity. Thus, impaired immune responses may impact the long-term presence of mRNA and spike protein in this patient cohort. Interestingly, the majority of vaccine-derived spike protein, both after mRNA- and vector-based vaccination, could be detected in fibroblasts in vaccinated humans and in mice. This is in line with a recent report using single cell analysis of the injection site, which also mainly detected vaccine-related mRNA in fibroblasts¹⁹. Besides fibroblasts, we could identify satellite cells in the muscle tissue as a target of vaccine-derived spike protein expression. The susceptibility of satellite cells for BNT162b2 uptake and spike protein expression was also confirmed in human satellite cells in vitro. Although we did not perform quantitative experiments here, these cells still expressed considerably more spike protein than HEK-cells when cultivated with the same amount of mRNA vaccine for the same time. Taken together, our study increases the understanding of the biodistribution of vaccine-derived mRNA by showing that in mice, early mRNA expression is predominantly present at the injection site, the blood and the lymphatic organs such as the spleen, and decreases significantly 7 days after injection. In recently deceased and mRNA-vaccinated humans, mRNA is not detectable outside of the local injection site. Importantly, the limitations of our study also need to be addressed to avoid overinterpretation of our published data. Firstly, the mice were vaccinated with a higher dosage of the mRNA vaccine per kilogram of body weight than humans normally receive. This could potentially affect the time period over which the vaccine mRNA can be detected, and increase the number of organs in which it can be detected. Secondly, the mRNA-LNP was optimised for humans, meaning that not all of the findings derived from our mouse data may be applicable to human pharmacodynamics. Third, while vaccine-associated spike mRNA and protein was measured in experimental animals without measurable delay in organ processing, our data from human vaccines were derived from post mortem tissue samples of patients who died in a temporal association with a vaccination. Unfortunately, the stability of mRNA vaccines might be affected by the latter in an organ-dependent manner, thus, our measurements might rather underestimate abundance of vaccine related spike mRNA or protein. Lastly, although based on a unique set of human tissues, the case numbers are still small and might limit the generalisability of our data.

Only after the emergence of COVID-19 and the widespread use of mRNA-LNPs against SARS-CoV-2 have the immune responses upon mRNA-LNP vaccination been systematically addressed³⁵. To date, and supported by a wide array of studies, the systemic mechanisms of mRNA vaccination are considered relatively well understood^25,36. Nevertheless, the local immune response at the injection site remains incompletely characterised. Even before the COVID-19 pandemic, a first study highlighted a direct local immune response in the vaccinated muscle upon mRNA-LNP vaccination, as increased chemokines and cytokines, such as IL-6, TNF, and CXCL9, could be detected³⁷. Such chemokines might play a role in recruiting cells of the innate immune system to the local injection site, as another study observed an early influx of innate inflammatory cells, such as monocytes and dendritic cells, in the injected muscle of rhesus macaques that had been vaccinated with mRNA-LNP for influenza H10 hemagglutinin³⁸. This study also demonstrated an early interferon-based local and systemic response signature in these animals. Newer studies underline the finding that upon mRNA-LNP injection, an influx of innate and adaptive immune cells occurs at the injection site¹⁹. In line with these previous reports, we could equally detect an infiltration of antigen-presenting cells, macrophages, and T cells in the injected muscle, as determined by IHC. However, we could not detect expression of spike protein by these cells. However, limitations in detection methods pose challenges to fully understand the mechanisms of mRNA vaccines. Moreover, immunohistochemistry can only detect substantial amounts of protein and may fail to identify soluble spike proteins or small quantities of protein. Interestingly, we could detect engulfment of spike-positive cells by IBA1⁺- and CD14⁺-cells in the muscle tissue of one patient five days after vaccination with the vector-based Ad26.COV2.S vaccine.

As outlined above and suggested by multiple studies, the local immune response at the site of injection might contribute significantly to the efficacy of mRNA-LNP vaccination³⁹. One important factor of this local immune response is the target cells of mRNA-LNP that facilitate the translation of the mRNA-encoded immunogenic protein. Undoubtedly, professional antigen-presenting cells, such as dendritic cells, can function in such a manner¹⁷. Some studies demonstrate that further infiltrating immune cells, such as macrophages, are capable of taking up mRNA-LNP and producing the mRNA-encoded protein at the site of the injection and in the draining lymph nodes^25,39. Nonetheless, different studies also found other potential target cells of mRNA-LNP over the years. For example, one study could also detect mRNA-positive immune cells such as monocytes, T, and B cells at the local site of vaccination¹⁶. After a challenge with intradermally applied mRNA-LNP in human skin explants, another study found that adipocytes were the first and most abundant cell type to express the mRNA-encoded eGFP, underlining the notion that non-immune cells can also be potential target cells for mRNA vaccines⁴⁰. This finding is in line with our study, as we have mainly detected mRNA transcripts at the injection site in non-immune cells such as fibroblasts and satellite cells. Additionally, a recent study corroborates our results, as they could detect highly enriched mRNA after vaccination with mRNA-LNP in fibroblasts at the injection site as well¹⁹. Mechanistically, the authors found that local mRNA-containing fibroblasts produced increased amounts of IFN-β that specifically enhanced antigen-specific cellular immune responses.