In situ detection of vaccine mRNA in the cytoplasm of hepatocytes during COVID-19 vaccine-related hepatitis

To the Editor:

We have read with high interest the article published in Journal of Hepatology by Boettler et al. ¹ where they report a case of acute hepatitis after the first dose of the BNT162b2 mRNA vaccine. The patient had an initial spontaneous recovery but relapsed after the second dose. Using imaging mass cytometry of the liver biopsy, the authors described an immune cell infiltrate predominated by CD8 T cells, which exhibited a panlobar distribution and contained spike-specific CD8 T-cell clones, pointing to the possibility of an autoimmune hepatitis (AIH)-like syndrome induced by the vaccine. There have been other reports of AIH-like hepatitis since the beginning of mRNA-based SARS-CoV-2 vaccination. Nevertheless, the incidence of AIH has not increased in 2021 during the COVID-19 vaccination period in Europe,² suggesting that triggering a bout of genuine AIH is unlikely the pathogenic mechanism of such vaccine-related events. Some authors have suggested molecular mimicry as a potential mechanism of liver damage³ although no similarity was found between soluble liver antigen and SARS-CoV-2 spike protein.⁴ Interestingly, most described cases of SARS-CoV-2 vaccine-related severe liver injury occurred after mRNA vaccines.⁵ Boettler et al. could not detect the spike protein in the liver by immunohistochemistry, a fact they attribute to the biopsy being performed 4 weeks after the peak of hepatitis. Thus, whether the final mechanism of hepatocyte injury is by antigenic mimicry or by a direct expression of the spike protein by vaccine-transduced hepatocytes remains unexplored.

Herein, we present a case of AIH-like hepatitis following SARS-CoV-2 vaccination wherein we could detect RNA encoding the spike protein within hepatocytes using highly sensitive and specific in situ hybridization (RNA-ISH).

A 67-year-old female without past medical history was admitted to the emergency room 12 days after the second dose of Pfizer-BioNTech (BNT162b2), presenting abdominal pain, fatigue and jaundice. Liver tests showed aspartate aminotransferase 1,201 IU/L, alanine aminotransferase 1,618 IU/L, alkaline phosphatase 211 IU/L, gamma-glutamyltransferase 71 IU/L, total bilirrubin 9.56 mg/dl, direct bilirrubin 9.08 mg/dl, international normalized ratio 0.9 and albumin 4.23 mg/dl. Antinuclear antibody with HEp-2 substrate (1:80) and anti-liver kidney microsomal antibody (1:40) were only mildly positive. Laboratory tests were negative for hepatitis A, B, C and E viruses, cytomegalovirus and Epstein-Barr virus. PCR for the detection of N and E genes of SARS-CoV-2 was negative. A liver ultrasound was normal. Liver biopsy showed chronic portal and interface hepatitis with lymphocyte and plasma cell infiltration. Considering that the hepatic biosynthetic function was preserved, we decided to withold the initiation of corticosteroids. Liver tests progressively improved over the next three months until complete recovery with no treatment.

In situ detection of SARS-CoV-2 mRNA transcripts in FFPE tissue sections was carried out using the RNAscope assay (Cat No. 848561, Advanced Cell Diagnostics, Abingdon, UK) coupled to quantitative immunofluorescence. The probe spanned 20 nucleotides within the Spike region between nucleotides 21631 and 23303 of the SARS-CoV-2 isolate Wuhan-Hu-1 genome (NCBI reference sequence: NC_045512.2). Whole slide image analysis and SARS-CoV-2 mRNA quantification was performed using ImageJ software version 1.52c (NIH, Bethesda, MD, USA). Briefly, the fluorescence signal level of SARS-CoV-2 mRNA was measured in the cellular compartment given by an expansion of each detected nucleus which creates an approximation of the full cell area. Two cases of AIH unrelated to COVID-19 and one post-mortem liver tissue from an individual diagnosed with severe COVID-19 were included as control tissues. The liver tropism of SARS-CoV-2 has been extensively demonstrated in both biopsies and post-mortem tissue analyses.⁶ The level of SARS-CoV-2 mRNA in the liver tissue of our patient with vaccine-related hepatitis was similar to the one found in the liver post-mortem biopsy obtained immediately after death from an individual with severe COVID-19. No SARS-CoV-2 mRNA transcripts were detected in AIH unrelated to COVID-19 (Fig. 1A-D).

Fig. 1.

In situ SARS-CoV-2 mRNA measurement using quantitative fluorescence and patient's biochemical tests.

SARS-CoV-2 mRNA transcripts (yellow channel) were detected using in situ hybridization in (A) the liver of an individual with hepatitis after the second dose of the Pfizer-BioNTech (BNT162b2) vaccine, (B) a post-mortem liver tissue from an individual diagnosed with severe COVID-19 (as a positive control). (C and D) No SARS-CoV-2 mRNA transcripts were detected in the liver tissues from individuals with autoimmune hepatitis unrelated to COVID-19. Nuclei are highlighted with blue. Scale bars represent 200 μm (A-D). (E) Patient’s course of AST, ALT, total bilirubin, ALP and GGT activity. ALP, alkaline phosphatase; ALT, alanine aminotransferase; AST, aspartate aminotransferase; GGT, gamma glutamyltransferase.

In line with the case reported by Boettler et al.,¹ our results suggest that lipid nanoparticles bearing mRNA molecules encoding SARS-CoV-2 proteins can reach hepatocytes under certain circumstances and deliver mRNA in high quantities that could be used by the translational machinery of the cells to produce spike. These peptides could then be presented through the MHC class I antigen presentation machinery and promote their recognition by previously sensitized CD8 T-cell clones. In our case, like in others described recently,^{1 , 4} hepatitis ocurred after the second dose of the vaccine, suggesting that previous exposure could enhance the severity of hepatocyte targeting by cytotoxic T lymphocytes. To the best of our knowledge this is the first report on in situ hybridization of vaccine mRNA in hepatocyte cytoplasm using commercially-available in situ RNA hybridization probes.

Another important teaching point is that these very rare cases of acute hepatitis after mRNA vaccines may resolve spontaneously and, may not always require the use of steroids (Fig. 1E). In the more severe cases, a rapid steroid tappering and avoiding long-acting immunosuppressants would likely be safe and effective, in contrast with the usual approach to AIH. Whether the duration of the expression of the spike protein by mRNA vaccine-trasduced hepatocytes could be related to the duration or the intensity of the liver damage or the relapse during or after steroid tappering are unanswered questions that deserve further investigation.

Finally, these findings should be taken into account in clinical trials of cancer vaccines using lipid nanoparticle-packed mRNA. The expression of tumor neoantigens by hepatocytes could modify the response to vaccines and perhaps trigger similar cases of liver injury.

Financial support

JA’s research is funded by Agencia Estatal de Salud (AES, FIS PI20-01663) and by Fundacion Echebano (Pamplona, Spain).

Conflict of interest

None of the authors have any conflict of interest regarding this manuscript.

Please refer to the accompanying ICMJE disclosure forms for further details.

Authors’ contributions

L. M-N. followed the case patient, captured clinical and lab data and helped selecting the control patients for histological analyses, drafting the manuscript. C. A. performed the RNA- In situ hybridization and the quantification, revision of the manuscript. B.S. diagnosed and followed the patient during her disease, revision of the manuscript. J.A. conception, design of the histological test, interpretation of the data, design of the figure and writing of the manuscript.

Supplementary data

The following are the supplementary data to this article: