In addition to ACE2, another membrane protein called TMPRSS2 is also known to be involved in coronavirus infection.

The research team introduced the coronavirus NL63 into the channel lined with intestinal cells and observed what happened. The Intestine Chip showed signs of infection: the layer of gut cells became “leaky” as the connections between them were compromised by the virus. To try to cure the infection, the researchers then administered nafamostat, a short-acting anticoagulant drug, into the channel lined with blood vessel cells.  True to form, nafamostat significantly reduced the amount of virus present in the Intestine Chip 24 hours after infection, though it did not restore the integrity of the connections between the cells.

The team then tried the same experiment using remdesivir. To their surprise, they found that remdesivir didn’t reduce the amount of virus in the Intestine Chip, and it also damaged the cells in the blood vessel channel, causing them to detach almost completely from the channel wall.

“We were surprised that remdesivir displayed such clear toxicity to the vascular tissue in the Intestine Chip,” said co-first author Girija Goyal, a senior research scientist at the Wyss Institute.  “GI symptoms have been previously reported in clinical trials of remdesivir, and this model now gives us a window into the underlying causes of those symptoms. It could also help us better understand the efficacy and toxicity of other similar drugs.”

A more complete picture of human gut health

Having established that their Intestine Chip could successfully model interactions between viruses, drugs, and the gut, the team tested a variety of other drugs that are taken orally including toremifene, nelfinavir, clofazimine, and fenofibrate, all of which have been shown to inhibit infection by SARS-CoV-2 and other viruses in vitro. Of those, only toremifene showed similar efficacy to nafamostat in reducing NL63 viral load.

Because the immune system interacts with both pathogens and drugs via the inflammatory response, the researchers then introduced a mixture of human immune cells called ​​peripheral blood mononuclear cells (PBMCs) into the blood vessel channel of the Intestine Chip to study this process. They found that more PBMCs attached themselves to the blood vessel wall in chips that had been infected with NL63 than in uninfected chips, and that the blood vessel cells were damaged as a result. They also observed that NL63 infection caused the secretion of multiple inflammatory cytokines that signal the body to recruit immune cells to the infection site.

Pre-treating the Intestine Chip with nafamostat prior to the introduction of the virus and PBMCs did reduce the secretion of some cytokines, but it did not mitigate the blood vessel damage, nor did it suppress the inflammatory response completely. Nafamostat pre-treatment did, however, increase the production of an antimicrobial protein called Lipocalin-2, implying that this type of protein could play a role in the cellular response to coronavirus infections.

This research was supported by the Defense Advanced Research Projects Agency (DARPA) under Cooperative Agreement HR0011-20-2-0-040, the National Institutes of Health (UH3-HL141797), Bill and Melinda Gates Foundation, and Wyss Institute for Biologically Inspired Engineering at Harvard University.