Over 480 million infections and 6 million deaths have been recorded since the novel coronavirus disease 2019 pandemic began over 2 years ago. Coronavirus disease 2019 is principally a respiratory illness caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), which infects the cells of the airway and can also replicate in the gastrointestinal tract. Notably, over half of those with SARS-CoV-2 infections shed SARS-CoV-2 genomes in their stool,1 and infectious viruses can be isolated from their fecal samples,2 suggesting that the virus also infects and replicates in the intestinal tract.

Intestinal organoids are versatile tools used to study viral infections ex vivo, and the clinical isolates of the SARS-CoV-2 outbreak strain can infect human and bat intestinal organoids.3,4 Newly emerged variants, such as delta and omicron, have an increased capacity to infect the respiratory tract, and omicron preferentially infects the upper airway and spares the lower respiratory tract.5 However, the replication preferences of these variants in the intestinal tract are currently unknown. Hence, we hypothesized that the omicron variant may have altered infective potential in the intestinal tract.

We previously reported on an induced pluripotent stem cell–derived apical-out intestinal organoid (mini-gut) containing different types of intestinal cells, including epithelial, goblet, Paneth, and enteroendocrine cells derived from 3 germ layers to form a 3-dimensional arrangement mimicking the intestinal tissue without introducing Matrigel or another extracellular matrix6 (Figure 1A and B, Supplementary Figure 1). The epithelial cells were arranged on the surface of the mini-gut and stably expressed the SARS-CoV-2 entry receptor, angiotensin converting enzyme 2 (ACE2), and the facilitating enzyme, transmembrane serine protease 2 (TMPRSS2), on the cell surface (Figure 1C), suggesting susceptibility of the mini-gut to SARS-CoV-2 infection. We thus investigated if human mini-gut organoids could be infected with the prototype Wuhan strain and currently circulating delta and omicron (BA.1 and BA.2) SARS-CoV-2 strains to evaluate the proliferation efficiency of these virus strains in the gut model. By serial measurement of viral RNA and titer in the culture supernatants, viral replication peaked 4 days after infection, and the replication efficiency of the delta strain was approximately 4–6 times higher than that of the Wuhan strain. In contrast, the replication capacity of the omicron strains was extremely low in the mini-gut model (Figure 1D and E, Supplementary Figure 2A). Interestingly, this trend was seen in the intestinal epithelial Caco-2 cells, where viral shedding of the omicron strain was severely reduced compared with the Wuhan and delta strains. In contrast, the omicron strain showed replication capability almost equivalent to the Wuhan strain in lung epithelial Calu-3 cells and African green monkey kidney epithelial VeroE6 cells, irrespective of exogenous TMPRSS2 expression (Figure 1F and G, Supplementary Figure 2B and C). These results imply that the infection efficiency of omicron is selectively reduced in intestinal epithelial cells.

Levels of a cell death marker, lactate dehydrogenase (LDH), were significantly increased in the culture supernatants from the mini-guts infected with the Wuhan and delta strains, whereas the LDH levels in the supernatants of omicron-infected mini-guts were comparable with that of the uninfected control (Figure 1H). Similarly, the high mobility group box 1 (HMGB1), a representative of damage-associated molecular patterns that are secreted during cell destruction, was detectable in the mini-guts infected by both the Wuhan and delta strains but not in the omicron-infected mini-guts (Figure 1I). Simultaneously, higher levels of various cytokines were detected in delta-infected mini-guts than in the omicron-infected mini-guts (Supplementary Figure 2D). HMGB1 is secreted on viral infection and induces inflammatory responses by stimulating immune cells to produce pro-inflammatory cytokines. By failing to secrete HMGB1 and possibly other damage-associated molecular patterns, it can be presumed that omicron infection might not induce inflammatory responses in the intestines, whereas the delta strain could do so.

Immunostaining using SARS-CoV-2 nucleocapsid antibodies was performed to visualize the presence of the virus in the mini-gut. The Wuhan strain–infected mini-guts retained distinct cellular morphology, and the infected cells were sporadically distributed among the organoid tissue, whereas delta-infected mini-guts exhibited clustering of virus-infected cells (Figure 1J and K), presumably because of the enhanced fusogenicity (membrane fusion activity) of this strain (Supplementary Figure 2E). Virus-infected cells were extremely scarce in the omicron infection model (Figure 1J). Of note, viral antigens were restricted to the luminal side of the intestinal epithelium and were not observed in the basal lamina and submucosal tissues, suggesting that SARS-CoV-2 infection is unlikely to intrude into the lamina propria. Interestingly, cell-surface expression of ACE2 was prominently downregulated in virus-infected cells (Figure 1L).

Intestinal organoids can also be used to test drug susceptibility of SARS-CoV-2,7 but there are a few reports yet on these tests. Therefore, we evaluated the effect of a neutralizing antibody cocktail (imdevimab and casirivimab) to reduce viral cell-to-cell transmission. Immunofluorescence analysis showed that the antibody cocktail treatment reduced viral spread to adjacent cells, even 2 days after infection with the delta strain, thereby suppressing the formation of infected cell clusters (Supplementary Figure 2F and G). Although it may not be practical to administer antibody drugs to the intestinal tract, our observations indicate that neutralizing antibodies secreted into the intestinal mucosa can inhibit the lateral spreading of progeny viruses.

Studies have shown differences in the anatomic site preferences between omicron and other SARS-CoV-2 variants in the respiratory tract,5 and we have revealed a similar phenomenon with respect to the intestinal tract. Currently, it is still elusive as to why the replication efficacy of the omicron variant is markedly reduced in gut organoids. The spike protein of the omicron variant is substantially mutated compared with the conventional strain (Supplementary Figure 2H), which may result in changes in its receptor binding, proteolytic cleavage, and stability in intestinal cells or tissues.8 Further analysis is needed to confirm the precise molecular mechanisms involved. Moreover, the low replication ability and cytotoxicity properties of the omicron variant in certain tissues and contrastingly high replication ability in other tissues must be assessed further.