The role of nicotinic receptors in SARS-CoV-2 receptor ACE2 expression in intestinal epithelia

Background Recent evidence demonstrated that severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) propagates in intestinal epithelial cells expressing Angiotensin-Converting Enzyme 2 (ACE2), implying that these cells represent an important entry site for the viral infection. Nicotinic receptors (nAChRs) have been put forward as potential regulators of inflammation and of ACE2 expression. As vagus nerve stimulation (VNS) activates nAChRs, we aimed to investigate whether VNS can be instrumental in affecting intestinal epithelial ACE2 expression. Methods By using publicly available datasets we qualified epithelial ACE2 expression in human intestine, and assessed gene co-expression of ACE2 and SARS-CoV-2 priming Transmembrane Serine Protease 2 (TMPRSS2) with nAChRs in intestinal epithelial cells. Next, we investigated mouse and human ACE2 expression in intestinal tissues after chronic VNS via implanted devices. Results We show co-expression of ACE2 and TMPRSS2 with nAChRs and α7 nAChR in particular in intestinal stem cells, goblet cells, and enterocytes. However, VNS did not affect ACE2 expression in murine or human intestinal tissue, albeit in colitis setting. Conclusions ACE2 and TMPRSS2 are specifically expressed in epithelial cells of human intestine, and both are co-expressed with nAChRs. However, no evidence for regulation of ACE2 expression through VNS could be found. Hence, a therapeutic value of VNS with respect to SARS-CoV-2 infection risk through ACE2 receptor modulation in intestinal epithelia could not be established.


Background
Coronavirus disease 2019 , caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), is characterized by a vast release of cytokines. Aggravated by following sepsis, this was established to be the cause of death in 28% of the infected patients . At the time of writing this manuscript, no cure nor vaccine exists yet and hence, many studies investigating treatment options have been initiated. Amongst these, vagus nerve stimulation (VNS) has been put forward as potential therapy because of its ability to induce an anti-inflammatory effect through dampening systemic inflammatory responses. This 'cholinergic anti-inflammatory pathway' of the vagus nerve has been acknowledged for many years, in particular at the level of sepsis-related cytokines Tumor Necrosis Factor (TNF), Interleukin (IL)-1, and High Mobility Group Box 1 (HMBG1) (Tracey 2002). Therefore, since the COVID-19 outbreak late 2019, several trials (e.g. Clini-calTrials.gov; NCT04341415, NCT04368156, NCT04382391) have been originated to study the efficacy of electrical VNS as treatment for this hyperinflammatory disease (Staats et al. 2020;Fudim et al. 2020). However, clear evidence supporting the value of VNS in treating COVID-19 is lacking.
Thus far, it has been demonstrated that SARS-CoV-2 uses Angiotensin-Converting Enzyme 2 (ACE2) as the key receptor for entry and Transmembrane Serine Protease 2 (TMPRSS2) for viral spike protein priming ). In addition, increasing evidence suggests that nicotinic receptors (nAChRs) have a pivotal role in the pathogenesis of COVID-19. For instance, an unusually low prevalence of cigarette smoking (characterized by prolonged nAChR activation) was clinically observed in COVID-19 patients (Creamer et al. 2019;Petrilli et al. 2020). This was confirmed by various systematic reviews showing the protective effect of current or former smoking habit against COVID-19 hospitalization (Farsalinos et al. 2020;Team CC-R 2020).
Although ACE2 is abundantly present in lung alveolar epithelial and ciliated cells, the highest ACE2 expression in the human body was found in intestinal enterocytes (Hamming et al. 2004;Qi et al. 2020). Intriguingly, nAChRs are equally expressed by enterocytes in the intestinal epithelium (Richardson et al. 2003;Summers et al. 2003). The gastrointestinal tract was therefore introduced as target organ for SARS-CoV-2 infection with up to 34% of COVID-19 patients reporting digestive symptoms like diarrhea, nausea, and abdominal discomfort (Yang and Tu 2020). The virus can propagate through ACE2 expressing enterocytes, and viral RNA can be found in rectal swabs, even after negative nasopharyngeal testing (Lamers et al. 2020;Amirian 2020;Xiao et al. 2020).
The aim of this study is to delineate ACE2 expression in intestinal epithelial cells, and to evaluate the expression of nAChRs in relation to ACE2 in intestinal epithelium. Further, we address the potency of VNS in the reduction of SARS-CoV-2 transmission. To this end, we made use of intestinal tissues of mice and humans that underwent chronic electrical non-invasive VNS administered via implanted devices (Brinkman et al., manuscript in preparation) (Bonaz et al. 2016;Sinniger et al. 2020).

VNS in humans
VNS of the left cervical vagus nerve was performed in patients with active Crohn's disease continuously for 12 months as was described previously (Fig. 1b) (Bonaz et al. 2016;Sinniger et al. 2020). Ileal and colonic biopsies were collected during ileocolonoscopy prior to start of VNS, and 6 and 12 months after start of VNS. These biopsies were used for mRNA expression assessments. In total, 9 patients were included in the study. Of these, 2 patients were removed from the study after a 3-month follow-up because of worsened clinical state. After 1 year of VNS, 5 out of 7 patients were in clinical remission (assessed with the Crohn's Disease Activity Index, CDAI), and 6 in endoscopic remission (assessed with the Crohn's Disease Endoscopic Index of Severity, CDEIS).
The study was approved by the Institutional Ethics Review Board (Identifier 11-CHUG-28), registered at Clini-calTrials.gov (NCT01569503), and was conducted in accordance with the Helsinki Declaration and the Good Clinical Practice guidelines of the International Council.

Animals
Female C57BL/6NCrl inbred mice (10 weeks old) were obtained from Charles River Laboratories, Maastricht, the Netherlands and acclimated for 1 week before performing experiments. The animals were housed under specific pathogen-free conditions in the animal facility at the Amsterdam University Medical Centers, location AMC, Amsterdam, the Netherlands. Animals were maintained on a 12/12 light/dark cycle under constant temperature (20°C ± 2°C) and humidity (55%) conditions with ad libitum drinking water and chow. Mice were handled according to the guidelines of the Animal Research Ethics Committee of the University of Amsterdam. Experiments were conducted under a project (application number AVD118002017842; license holder number 11800) approved by the Dutch Central Animal Experiments Committee. Individual experiments were revised and approved by the Animal Research Ethics Committee of the University of Amsterdam. Prior to the experiments protocols were approved by this same committee. A total of 32 mice were used for this study. . The procedure was performed on mice under anesthesia with 2.5% isoflurane in 100% O 2 (flow 1 L/min). Preoperatively and 24 h postoperatively animals were administered 1 mg/kg meloxicam (Metacam; Boehringer, Ingelheim am Rein, Germany) and 5 mg/kg enrofloxacin (Baytril; Beyer Healthcare, Whippany, NJ, United States of America). Local anesthetics were administered through a lidocaine splash block (2%) before wound closure. Mice recovered for 10 days. At day 0, mice were allocated to the sham group when the impedance measurement of the implanted cuff was above 25 kΩ. Other mice were pair-matched based on weight, after which they were randomly allocated (1:1) to the sham or stim group. The sham group counted 14 mice and the stimulated group counted 16 mice.

VNS in mice
Bipolar stimulation was performed for 2 min 6 times daily for 12 days with the following parameters: 650 μAmp, 10 Hz, pulse width 200 μs (Fig. 1a) (Guyot et al. 2019). Proper stimulation was examined by observing behavioral changes of the mice (e.g. altered breathing, decreased movements). Also, before, during, and after the experiment, impedance measurements (using a single frequency of 1 kHz) with a Minirator MR Pro (NTI Audio, Essen, Germany) on the cuff electrodes were performed to ensure proper conduction. Further, the vagus nerves with cuffs from randomly selected mice were collected. After paraffin embedding, a hematoxylin-eosin (HE) staining was conducted to assess damage of the nerve. At the end of the study, mice were euthanized.

DSS treatment
All animals were treated with dextran sodium sulphate (DSS) to induce acute colitis. 2.25% (w/v) DSS (TdB Consultancy, Uppsala, Sweden) was added to the drinking water for 5 consecutive days. Drinking water with fresh DSS solution was replaced daily. Following DSS-treatment, mice received normal drinking water for 7 days, adding up to a total experiment length of 12 days. During the study, bodyweight and behavior were monitored daily. VNS was started at the same day as DSS treatment. mRNA was extracted from frozen ileal and colonic tissue after lysis in ISOLATE II RNA Lysis Buffer RLY and isolation according to the Bioline ISOLATE II RNA mini kit (both GC biotech b.v. Alphen a/d Rijn, the Netherlands) and cDNA was synthesized by use of the Revertaid first strand cDNA synthesis kit (ThermoFisher Scientific, Landsmeer, the Netherlands). qPCR was performed using SensiFAST SYBR No-ROX (GC biotech b.v.) on a CFX96 Touch™ Real-Time PCR Detection System (Bio-Rad Laboratories B.V., Veenendaal, the Netherlands) and expression levels were analyzed using LinRegPCR software (Ruijter et al. 2009). Expression levels were normalized for reference genes Eef2, Nono, and Gapdh for mouse and ACTB, PSMB6, and RPLP0 for human after stability assessment with geNorm ( Vandesompele et al. 2002). Primers (all Sigma-Aldrich Chemie N.V., Zwyndrecht, the Netherlands) are listed in Table 1.

Statistical analysis
Graphs were made with GraphPad Prism 8.0 (GraphPad Software, La Jolla, CA, USA) and a Mann-Whitney U test (2 groups) or a Kruskal-Wallis test (> 2 groups) was used to check for statistical significance. In all tests, P < 0.05 was accepted as an indication for statistical significance. All data are expressed as median with interquartile range plus the individual data points.
Notably, ACE2 is clearly expressed in goblet cells, although it has been demonstrated that SARS-CoV and SARS-CoV-2 cannot infect goblet cells in both airway and intestinal epithelia (Lamers et al. 2020). Moreover, the nAChR that was mostly co-expressed with ACE2 and TMPRSS2 was α7 nAChR (encoded by CHRNA7). This is an essential regulator of inflammation by inhibiting cytokine release through the aforementioned cholinergic anti-inflammatory pathway (Wang et al. 2003). This is relevant because Russo and Leung reported on the role of α7 nAChR in SARS-CoV-2 and demonstrated that smoking results in an upregulation of α7 nAChR leading to an increase of ACE2 (Leung et al. 2020a;Russo et al. 2020;Leung et al. 2020b).

Chronic electrical VNS does not alter intestinal ACE2 expression in mice and humans
As a positive correlation of ACE2 and CHRNA7 was previously observed, and chronic cholinergic stimulation typically leads to upregulation of nAChRs, we investigated whether chronic VNS affected the expression of ACE2 (Leung et al. 2020b;Melroy-Greif et al. 2016).
Accordingly, intestinal ileal tissues of mice that were subjected to electrical chronic VNS, applied via an implantable device, were obtained and analyzed for Ace2 mRNA expression. Adequate VNS was confirmed by behavioral changes of the mice, impedance measurements, and HE stainings not showing any damage of the vagus nerve caused by the cuff electrodes (Fig. 3). Our data and stimulated animals (n = 16). Levels were corrected for reference genes and epithelial markers E-cadherin (Cdh1) and Villin (Vil1) to obtain epithelial fraction. b Relative mRNA expression of ACE2 in human ileum (left) and colon (right) samples of patients treated with VNS. Levels were corrected for reference genes and epithelial markers E-cadherin (CDH1) and Villin (VIL1) to obtain epithelial fraction. mRNA levels of ACE2 were assessed in biopsies collected prior to VNS, and 6 and 12 months after start VNS. Non-responders (red) are patients with no clinical and endoscopic remission. ACE2: Angiotensin-Converting Enzyme 2; VNS: vagus nerve stimulation show no effect on Ace2 expression levels. Since samples contained bulk RNA of intestine, values were corrected for epithelial markers Cdh1 and Vil1 to determine the epithelial fraction. Also after correction, no differences were observed (Fig. 4a). Next, we assessed ACE2 mRNA expression in human intestinal tissues of patients treated with chronic VNS as was described previously (Bonaz et al. 2016;Sinniger et al. 2020). In line with the mouse data, no differential expression was observed between samples collected prior to VNS and 6 and 12 months after start VNS. Correction for CDH1 and VIL1, markers that have shown to be stable under inflammatory conditions (data not shown), did not alter the results (Fig. 4b). Though not significant, a trend towards higher levels of ACE2 at 12 months VNS could be detected. Noticeably, ACE2 expression levels in colon samples were relatively low when compared with ileal samples, which is corroborated by earlier literature showing low basal expression levels of ACE2 in colonic tissue (Qi et al. 2020).
It should be mentioned that in this study we made use of already available tissues. Thus, stimulation parameters such as duration of stimulation, pulse width, and frequency were not optimized for the current research question. In addition, both VNS experiments have been performed in diseased subjects; mice were exposed to DSS inducing colitis, and all humans had active Crohn's disease. ACE2 has been shown to be upregulated in inflammatory bowel diseases (IBD), and in ileum in particular (Hashimoto et al. 2012;Bangma et al. 2020;Krzysztof et al. 2020). For that reason, in our experiments, basal ACE2 expression levels might have been higher when compared with healthy subjects because of the colitis background of the tissues, possibly mitigating the results. The considerable divergence between ileal and colonic expression levels substantiates this confounding effect. Though, our analyses did not show significant differences between uninflamed and inflamed samples (Fig. 4b).
Data on the role of α7 nAChRs in colitis are conflicting. Sun et al. demonstrated increased CHRNA7 expression upon 2,4,6-trinitrobenzenesulfonic acid (TNBS)induced colitis, whereas Baird et al. found decreased CHRNA7 expression in IBD (Baird et al. 2016;Sun et al. 2007). Further, in an experimental colitis model, selective α7 nAChR agonists worsened disease (Snoek et al. 2010). These data imply that colitis is associated with CHRNA7 expression, again modifying the ACE2 expression levels perchance confusing the current results.
Importantly, when in this manuscript we focus on nicotinic receptors and α7 nAChR specifically, it is important to note that other receptors might play a role in regulating ACE2 as well since VNS unequivocally activates other acetylcholine-binding receptor types. This was not examined in this study.
In addition, we show that VNS does not alter intestinal ACE2 mRNA expression levels. Protein expression levels, which might be of more importance in this context, have not been assessed. Hence, an effect on SARS-CoV-2 infection could still be observed. This is substantiated by Lamers et al. who demonstrated that SARS-CoV-2 infection of gut enterocytes is independent from the ACE2 expression level (Lamers et al. 2020). Conversely, since SARS-CoV-2 is known to downregulate ACE2 expression, it is plausible that VNS could restore ACE2 levels in infected patients (Glowacka et al. 2010;Kuba et al. 2005). To this end, VNS studies including SARS-CoV-2 infected patients should be performed. Regardless of the ACE2 expression level, the antagonizing effect of VNS on the devastating systemic cytokine storm might strengthen this intervention as potential treatment for COVID-19. This is greatly underlined by the positive outcomes of the RECOVERY trial studying the effect of low-dose steroid treatment with dexamethasone on SARS-CoV-2 (RECOVERY Collaborative Group et al. 2020). Alike VNS, dexamethasone is acknowledged as inhibitor of the production of pro-inflammatory cytokines (Horton and Remick 2010;Nyhlen et al. 2004).
Concerning the effect of VNS on systemic cytokine responses, a critical note is that these responses depend on the stimulation parameters such as frequency, amplitude, and pulse width as was demonstrated by Tsaava et al., and recently reviewed (Tsaava et al. 2020;Bonaz 2020). Although low frequency stimulation (as was used in our studies) was not examined, stimulation with high amplitudes resulted in an increase in IL-6 and decrease in IL-5 and TNF-α. When refining these parameters in VNS, the subsequent positive regulation of cytokines could substantiate the role of VNS in treating COVID-19.
Finally, granting the potential role for VNS in SARS-CoV-2 infection, one must be aware that parasympathetic neuronal activity through the vagus nerve can induce bronchoconstriction, an absolutely undesirable adverse effect in COVID-19 patients (Belmonte 2005). Even though results on this matter are conflicting, application of VNS in these patients must be performed with great caution (Miner et al. 2012;Steyn et al. 2013). To elucidate and overcome this issue, future studies should focus on more specific stimulation techniques exclusively targeting the cytokine-producing organ of interest.

Conclusions
The implication of VNS as treatment for SARS-CoV-2 infection has gained high interest. Here, we show that despite gene co-expression of nAChRs, ACE2, and TMPRSS2 in the intestinal epithelium, chronic noninvasive electrical VNS does not significantly alter intestinal ACE2 expression at a transcriptional level. As bio-electronic neuromodulatory techniques continue to evolve as treatment for controlling inflammation, and with the current SARS-CoV-2 outbreak for COVID-19 in particular, further experimental investigations are needed to shed light on the therapeutic potential of VNS for SARS-CoV-2 infection control.