A novel definition and treatment of hyperinflammation in COVID-19 based on purinergic signalling

Hyperinflammation plays an important role in severe and critical COVID-19. Using inconsistent criteria, many researchers define hyperinflammation as a form of very severe inflammation with cytokine storm. Therefore, COVID-19 patients are treated with anti-inflammatory drugs. These drugs appear to be less efficacious than expected and are sometimes accompanied by serious adverse effects. SARS-CoV-2 promotes cellular ATP release. Increased levels of extracellular ATP activate the purinergic receptors of the immune cells initiating the physiologic pro-inflammatory immune response. Persisting viral infection drives the ATP release even further leading to the activation of the P2X7 purinergic receptors (P2X7Rs) and a severe yet physiologic inflammation. Disease progression promotes prolonged vigorous activation of the P2X7R causing cell death and uncontrolled ATP release leading to cytokine storm and desensitisation of all other purinergic receptors of the immune cells. This results in immune paralysis with co-infections or secondary infections. We refer to this pathologic condition as hyperinflammation. The readily available and affordable P2X7R antagonist lidocaine can abrogate hyperinflammation and restore the normal immune function. The issue is that the half-maximal effective concentration for P2X7R inhibition of lidocaine is much higher than the maximal tolerable plasma concentration where adverse effects start to develop. To overcome this, we selectively inhibit the P2X7Rs of the immune cells of the lymphatic system inducing clonal expansion of Tregs in local lymph nodes. Subsequently, these Tregs migrate throughout the body exerting anti-inflammatory activities suppressing systemic and (distant) local hyperinflammation. We illustrate this with six critically ill COVID-19 patients treated with lidocaine.


Introduction
Hyperinflammation and acute respiratory distress syndrome (ARDS) caused by coronavirus disease 2019 (COVID- 19) have become the world's number 1 challenge. The exponential pattern in the number of severe cases in the second and third waves of the SARS-CoV-2 pandemic has shown to reach nations' maximum ICU capacities in weeks rather than months after outbreak of the disease irrespective of rigorous population-based preventive measures. In a recently published systematic review, the case fatality rates in patients in the ICU across 7 countries vary between 14.9 and 66.7%, while the case fatality rates among those who required mechanical ventilation vary between 16.7 and 97.0% [1]. In addition, the case fatality rate in a cohort of 1035 critically ill COVID-19 patients requiring extracorporeal membrane oxygenation (ECMO, artificial lungs) is alarmingly high (37.4%) [2].
The clinical manifestations of severe COVID-19 consist of pneumonia with dyspnoea and hyperinflammation. Hyperinflammation is thought to be the basis of the development of severe and critical COVID-19 [3][4][5]. Currently, a clear-cut definition of hyperinflammation is lacking. Some authors describe the condition of hyperinflammation as a form of very severe inflammation with cytokine storm [6]. The criteria of hyperinflammation are not consistent and include clinical data and/or different combinations of the parameters of the activation of the pro-inflammatory response of the immune system (i.e. fever, rapid respiratory deterioration, cytokine, ferritin and/or CRP concentrations, changes in blood levels of several types of immune cells, etc., examples are presented in Table 1) [3, 4, 6-8, 9, 10-14]. In addition, the current definitions of hyperinflammatory syndrome do not provide an Table 1 Examples of the criteria of hyperinflammation. These criteria are not consistent and include different combinations of symptoms and laboratory parameters of the activation of the pro-inflammatory response of the immune system Author Year of publication Macrophage activation (ferritin concentration of 700 μg/l or more) Haematological dysfunction (neutrophil to lymphocyte ratio of 10 or more or both haemoglobin concentration of 9.2 g/dl or less and platelet count of 110 × 10 9 cells/L or less) Haematological dysfunction (neutrophil to lymphocyte ratio of 10 or more or both haemoglobin concentration of 9.2 g/dl or less and platelet count of 110 × 10 9 cells/L or less) Coagulopathy (D-dimer concentration of 1.5 μg/ml or more) Hepatic injury (lactate dehydrogenase concentration of 400 U/L or more, or an aspartate aminotransferase concentration of 100 U/L or more) Cytokinaemia (defined as an IL-6 concentration of 15 pg/ml or more, or a triglyceride concentration of 150 mg/dl or more, or a CRP concentration of 15 mg/dl or more) [7] Fajgenbaum DC and June CH 2020 Very severe inflammation with cytokine storm [6] Manson JJ et al. Cytokine storm, dysregulated macrophage activation, impaired natural killer cell response, lymphopenia, elevated absolute neutrophil count and neutrophil/lymphocyte ratio and increased levels of neutrophil extracellular traps (NETs) [4] Anka AU et al. 2021 Excessive secretion of pro-inflammatory cytokines and the recruitment of pro-inflammatory cells such as granulocytes and macrophages caused by tissue injury result in a snowballing of cytokine secretion leading to a systemic inflammatory response such as macrophage activation syndrome (MAS), secondary haemophagocytic lymphohistiocytosis (sHLH-cytokine storm) [11] Cardone MC et al.
explanation for the frequently observed co-infections or secondary infections in COVID-19 [15,16]. The results of non-randomised cohort studies with controls and of retrospective observational studies suggest that IL-1 receptor blockade (anakinra) [9,17], monoclonal antibodies against IL-6 receptors [18][19][20][21] and the combination of both drugs [22,23] may improve survival rate in at least a subgroup of patients with COVID-19. However, in prospective randomised controlled trials with the exception of one trial with tocilizumab and sarilumab in critically patients [24], anti-inflammatory therapy with anakinra [25] or tocilizumab [26][27][28][29][30][31][32] did not improve the outcome in moderate, severe and critically ill COVID-19. On December 10, 2020, an editorial commented that it is disappointing that nearly 10 months into the COVID-19 pandemic, a breakthrough treatment has not been identified [33]. Researchers of the US National Institute of Allergy and Infectious Diseases stated that although Remdesivir is effective to reduce time to recovery in hospitalised COVID-19 patients [34] and dexamethasone reduces mortality in critically ill COVID-19 patients [35], there is no treatment for early or mild infection [36]. Moreover, dexamethasone raises concerns because it increased the 28-day mortality in patients who did not receive respiratory support [35] and it dampens the "alarm phase" of the inflammation process including the capacity of detecting pathogens in mammals by the immune system [37]. In addition, administration of methylprednisolone (1 mg/kg/day intravenously) in COVID-19 reduced the blood levels of NK cells, CD4 + and CD8 + T-cells and increases the duration of throat viral RNA detectability indicating immune cell dysfunction [38]. Furthermore, targeted anti-viral medication has failed to treat COVID-19 effectively [39]. According to the World Health Organisation, after a record-breaking development, vaccine deployment is slow and has many challenges to overcome [40]. Vaccine hesitancy is relatively high [41] even among health care workers [42,43]. It could take more than a year to vaccinate enough people required to make an impact on SARS-CoV-2 spreading, while therapeutic measures that can immediately attenuate the course of SARS-CoV-2-related lung damage are promptly needed on a global scale. To make the matters worse, many scientists expect that SARS-CoV-2 may become endemic and is here to stay [44].
In this report, we developed a novel definition of hyperinflammation based on purinergic signalling. Subsequently, we describe our discovery of an old drug capable of attenuating hyperinflammation and illustrate this with six critically ill patients suffering from COVID-19. Finally, we present the future development of a new and more accessible administration route for this drug as shown in Fig. 1.

Purinergic signalling
In 1929 adenylic acid (identical to adenosine) was identified [45], and in the same year, the adenosine triphosphate (ATP) molecule was discovered and isolated [46]. Ten years later (1939), researchers contributed to the understanding of intracellular ATP as an intracellular energy transport molecule [47][48][49][50]. In 1948 and in 1959, it was reported that extracellular ATP has a different function than ATP within the cytoplasm [51,52]. The authors showed that extracellular ATP molecules have an intercellular signalling function. The intercellular signalling by nucleotides (ATP, ADP, UTP and UDP) and nucleoside (adenosine) is referred to as purinergic signalling. The purinergic co-transmission in neurons was discovered by Geoffrey Burnstock in 1972 [53]. It took over 20 years for the importance of purinergic signalling to be accepted [54,55]. Finally, researchers of the University of Ferrara first reported that the P2Z receptor (the former name of the P2X7R) plays an intriguing role in immunity, inflammation and cell death [56].
The intracellular levels of ATP are high at millimolar concentrations (2-8 mM) [57], and the ATP concentrations in synaptic vesicles are even higher in the range of 5 to 100 mM [58]. In contrast, under normal resting conditions, the extracellular levels of ATP are quite low at nanomolar concentrations (<3 nM) [57,59]. Under specific conditions, ATP release can rise by more than 1000-fold [53, 57, 60, 61] and leads to a significant increase in the extracellular levels of ATP. The resulting significant increase in extracellular nucleotides and adenosine concentrations activates their purinergic receptors inducing certain cellular functions. Examples of such conditions are membrane depolarisation (i.e. sympathetic neuron endings) [53], mechanical stress (i.e. high mechanical power ventilation) [59][60][61][62][63], hypoxia [64], hyperosmosis, hypotonic and isotonic stress of endothelial cells [65][66][67][68], inflammation [69,70], surfactant release by alveolar epithelial type II cells [59][60][61], mucine release by airway smooth muscle cells [71], insulin release by pancreatic islet beta-cells [72,73], etc. There is an exception to this concept: Although a spontaneous ATP-induced inward Ca 2+ current through the P2X7R could not be detected below extracellular ATP levels of 200 μmol/ml [74], low tonic basal activation of P2X7R at nanomolar extracellular ATP concentrations promotes serum independent cellular proliferation [75], promotes closure of the wound area in scratch wound assay [76], protects from apoptosis [77], initiates anaerobic glycolysis independent of the oxygen contents [78], etc. (Table 2, rows 80-85). However, low tonic basal activation of the P2X7Rs by extracellular ATP does not cause a pro-inflammatory response of the immune system. Therefore, this topic is beyond the scope of this paper and will not be discussed here.
Clearance of the ATP molecule in order to avoid accumulation in the extracellular space is performed by enzymes attached to the outside of the cell membranes (ecto-enzymes) and by soluble enzymes excreted to the extracellular space ( Fig. 2) [57, [272][273][274][275]. A proportion of the enzymaticbreakdown product of ATP adenosine enters the cells via the equilibrative nucleoside transporters (ENT1 and ENT2) and concentrative nucleoside transporters (CNT1 and CNT2) (Fig.  2) [57, 60, 61]. The release and subsequently clearance of the extracellular nucleotides and adenosine cause fluctuation in the extracellular levels of ATP, other nucleotides and adenosine. These fluctuations in extracellular concentrations are indispensable for the receptor resensitisation after desensitisation following receptor activation as discussed below.

Purinergic signalling in inflammation and hyperinflammation
The purinergic control of cellular processes including the proinflammatory and anti-inflammatory responses of the immune system is depending on the activation and the desensitisation phenomenon of the nucleotides and adenosine receptors of the immune cells [74,[276][277][278][279][280][281][282]. Except for the P2X7R, all other purinergic receptors, i.e. P2XRs, P2YRs and P1 receptors (adenosine receptors-AdoRs), are subject to desensitisation [279][280][281][282][283]. In addition, a certain extent of desensitisation occurs after every activation, and this desensitisation requires time to return to the state of complete resensitisation [279,280]. The higher and the longer the stimulus of the activation, the higher the extent of desensitisation and the longer the recovery time to the state of complete resensitisation [278]. One of the P2 receptors, the P2X7 receptor, is not prone to desensitisation, and apart from the low tonic basal activation of this receptor at low nanomolar concentrations as mentioned above, the extracellular concentration of ATP required to activate this receptor is much higher. Activation of the P2X7R starts at 100 μM with an EC 50 of >1 mM [74,279,284].
Hyperinflammation is characterised by the activation of P2X7Rs and desensitisation of other P2 receptors and AdoRs As mentioned above, hyperinflammation starts when fluctuation of the extracellular nucleotides and adenosine no longer occurs and leads to prolonged activation of the P2X7Rs of the immune cells. Prolonged vigorous activation of the P2X7Rs leads to macropore formation and cytolysis with uncontrolled ATP release [222,223,227,299] (Table 2, rows 120-123) causing hyperinflammation with massive pro-inflammatory immune response, massive pro-inflammatory and antiinflammatory cytokine release: the cytokine storm (Fig. 3).
In the early phase of COVID-19, hyperinflammation may be confined to the site of viral entry (i.e. airway mucosa and conjunctivae) but as viral replication and viral spreading progress, systemic hyperinflammation devel ops. The upregulation of the expression of ectonucleotidases also leads to an increase in the concentrations of other nucleotides (i.e. ADP) and adenosine. These high extracellular concentrations of nucleotides and adenosine do not show concentration fluctuations required for the recovery (resensitisation) time from desensitisation causing a state of persistent desensitisation of all P2XRs, P2YRs [279,280,283,300,301] and AdoRs [282] with the exception of P2X7Rs. Consequently, the physiological function in the affected organs and inflamma tory response of the immune system are deactivated. This leads to the failure of organ function (i.e. ARDS in the lungs as we reported earlier [61]) and the immune system (immune paralysis) rendering the host susceptible to secondary co-infections (Fig.3). Sepsis-induced immunosuppression [302,303] or compensatory antiinflammatory response syndrome (CARS) in critically ill patients [304] was already raised by researchers in 1996 [305] and is a well-known phenomenon in critically ill patients [302]. Secondary bacterial infections occurred in 34.4% of 274 surviving elderly patients (age over 60 years) with COVID-19 and in 81.7% of 65 deceased patients [15]. In addition, it was found that 76 co-infections with other respiratory pathogens occurred in another cohort of 354 COVID-19 patients (16 of 115 mild cases (13.9%), 33 of 155 severe cases (21.3%) and 27 in 84 critical cases (32%)) [16]. In a meta-analysis involving 118 scientific reports on patients with COVID-19, co-infection with other pathogens at admission was observed in 19% and superinfection with other pathogens during admission in the hospital in 24% [306].

Control of hyperinflammation is annihilated by the downregulation of Tregs through the activation of P2X7R and the desensitisation of adenosine receptors
Tregs are key elements in the control of hyperinflammation [307]. Activation of AdoRA2As promotes the differentiation of naïve T-cells towards regulatory T-cells(Tregs) [112], increases the frequency of Tregs and the expression of CTLA-4 receptor and upregulates ecto-enzymes CD39 and CD73 expression accelerating adenosine generation from extracellular ATP [118] (Table 2, rows 25, 29, 30 and 33). This process is upset in case of desensitisation of AdoRs. In addition, activation of P2X7Rs inhibits the suppressive potential and stability of Tregs, inhibits the clonal expansion of Tregs, promotes Treg death, induces Treg depletion and reduces Treg IL-10 production ( Table 2, rows 86-88, 106 and 107). In COVID-19 patients, significant lower Treg frequencies [308][309][310], lower expression of forkhead box protein P3 (FoxP3), lower expression of transforming growth factor-β(TGF-β) and lower cytokine TGF-β secretion [309] are observed compared to healthy control. Additionally, a reduced proportion of specific SARS-CoV-2-reactive Tregs was reported [311]. The desensitisation of AdoRs and the activation of P2X7Rs may well be the underlying mechanism of the low Tregs frequency in severe and critically ill COVID-19.

P2X7R antagonist restores the reduced Tregs population and Tregs function in hyperinflammation
As stated above, infected cells release ATP into the extracellular space. Obviously, the P2X7R antagonist blocks the activation of the P2X7Rs. Because a significant proportion of the ATP release to the extracellular space is mediated by the P2X7R (Table 2, rows 77-79), P2X7R a ntagonism combined with the upregulated ATP hydrolysing activity of the ectoenzymes results in the decrease of the extracellular ATP concentrations. This can potentially abrogate hyperinflammation Fig. 2 Clearance of extracellular ATP and adenosine by ectonucleotidases and soluble extracellular nucleotidases . This process is indispensable to enable receptors to recover from desensitisation following receptor activation (resensitisation, see text under the heading "Purinergic signalling in inflammation and hyperinflammation" for explanation). CD39,Ecto-nucleoside triphosphate diphosphohydrolase 1-3 (ENTPD 1-3); CD73, Ecto-5′-nucleotidase (5'-NT); NPP, nucleotide pyrophosphatase/phosphodiesterase; TNAP, tissue nonspecific alkaline phosphatase; ADA, adenosine deaminase; ADK, adenosine kinase; HGPRT, hypoxanthine-guanine phosphoribosyltransferase; ATP, adenosine triphosphate; ADP, adenosine diphosphate; AMP, adenosine monophosphate; ADO, adenosine; ENTs, equilibrative nucleoside transporters; CNTs, concentrative nucleoside transporters Fig. 3 A schematic presentation of the activation of the purinergic receptors of the immune cells causing a pro-inflammatory response leading to hyperinflammation. Viral infection drives the controlled cellular release of ATP molecules. Increased extracellular nucleotides levels activate P2XRs and P2YRs. Upregulation of the extracellular ATP hydrolysing enzymes as depicted in Fig. 2 results in the increase of extracellular adenosine levels followed by the activation of the adenosine receptors (AdoRs). These processes initiate the physiologic pro-inflammatory response of the immune system. The green line at the bottom of the graph represents the extracellular ATP levels. The ascending part is caused by the ATP release, and the descending part results from the clearance of ATP by the extracellular or membrane-bound ATP hydrolysing enzymes. As the disease progresses and extracellular ATP levels increase above 1 mM, the P2X7R is additionally and effectively activated leading to a severe immune response. Except for P2X7Rs, all these receptors are known to be subject to desensitisation. Desensitisation of a receptor is defined as being unresponsive to activation by the ligand, resulting in (near) zero transmembrane signal transduction. A certain extent of desensitisation occurs after every activation, and this desensitisation requires time to return to the state of complete resensitisation. Increasing intensity and duration of the activation stimuli leads to increasing extent of desensitisation and duration of the recovery time to the state of complete resensitisation (represented by brown boxes with increasing size at the bottom of the graph). Severe viral infection can increase the controlled ATP release beyond the capacity of the extracellular enzymes to clear ATP and adenosine molecules. This causes a sustained high extracellular ATP and adenosine levels preventing the purinergic receptors from recovering from the state of desensitisation. The capacity to clear invading microorganisms diminishes leading to immune paralysis. In addition, prolonged high extracellular levels of ATP and activation of the P2X7R lead to macropore formation and cell death with uncontrolled release of ATP. In turn, this leads to vigorous activation of the P2X7R of the immune cells promoting massive production of cytokines ending in a cytokine storm and hyperinflammation and the concomitant immune paralysis. Moreover, P2X7R inhibition promotes the cell-autonomous conversion of CD4+ T cells into Tregs after stimulation of their T-cell receptors (TCRs) [190]. In addition, P2X7R knock-out mice, mimicking the state of complete P2X7R inhibition, show an increase in tissue Tregs, prevent Tregs death and the Tregs produce more IL-10 and TGF-β [191]. Experimental inhibition of P2X7Rs restores the Tregs levels and function ( Table 2, rows 86-88, 106 and 107) [190][191][192]. Inhibition of the P2X7R or P2X7R knock-out can attenuate severe inflammation in abdominal sepsis [312] and in acute lung injury [313,314]. Apparently, amelioration of hyperinflammation by P2X7R inhibition is based on the increased activation and clonal expansion of the anti-inflammatory Tregs population ( Table 2, rows 86-88, 106 and 107).
Some authors proposed that the P2X7R is an ideal candida te to target in COVID-19-associated severe pneumonia [298,315], and others suggested that hyperactivation of the P2X7R plays a key role in the neuropathology of COVID-19 and that P2X7R antagonism may prevent or treat neurological manifestations of COVID-19 [316].

Lidocaine is a P2X7R antagonist
In 2015 it was discovered that lidocaine is a P2X7R antagonist [74], and therefore, lidocaine can potentially reduce the clinical symptoms of hyperinflammation significantly. In experimental sepsis, lidocaine improves organ failure [317][318][319] and survival [317]. In septic patients, lidocaine reduces neutrophil recruitment by the mitigation of chemokine-induced arrest and transepithelial neutrophil migration [320]. Neutrophil recruitment is an important facilitating process in the pathogenesis of multiple organ failure [320] and hyperinflammation in COVID-19 [321][322][323][324]. In patients with skin lesions from atopic dermatitis, lidocaine increases the proportion of Tregs and upregulates the FoxP3 expression [325]. In addition, lidocaine increases the IL-10 levels in mechanically ventilated mice [326] and decreases the TNF-α in BAL, plasma and lung samples in pigs undergoing surgery for lung resection [327].
The P2X7R antagonist dose-response relationship of lidocaine is presented in Fig.4. The IC 50 for the inhibition of the P2X7R by lidocaine is about 66.07 μg/ml (0.28 mM) [74] where IC 50 is defined as the required extracellular concentrations of the receptor antagonist to reach an inhibitory effect of halfway between maximal activation and maximal inhibition (half-maximal inhibitory concentration). The main issue is that the IC 50 for P2X7R inhibition is much higher than the maximal tolerable plasma concentration for mammals. The maximal tolerable plasma concentration in humans is about 4.7 μg/ml (0.02 mM); this corresponds with an IC 10 or lower (<10% inhibitory concentration, Fig.4). Above this lidocaine plasma concentration, adverse effects in increasing severity occur as presented in Table 3 [328,329]. Thus, systemic lidocaine plasma concentrations of >4.7 μg/ml must be avoided [328,329]. Caveat: The inhibitory concentrations of lidocaine for P2X7R as presented in Fig.4 are not corrected for the series resistance (in the range of 1-3 MΩ) of the used whole-cell voltage clamp method with two puller microelectrodes [74]. One should bear in mind that after correction for series resistance, the reported inhibitory concentration values including IC 50 are expected to be higher [330].
VGSCs conduct sodium ions inward and are essential for the transduction of sensory stimuli, the generation of the action potential and the release of neurotransmitters from sensory neuron terminals. Lidocaine inhibition of VGSCs can effectively reduce pain signalling [340]. In addition, VGSCs are present on dendritic cells (maintain chemokine-induced migration) [341], macrophages (regulate phagocytosis and endosomal pH during LPS-mediated endosomal acidification) [342], microglia (regulate phagocytosis cytokine release ad migration) [343], neutrophils (regulate attachment, transmigration and chemotaxis) [344] and T-cells (regulate positive selection of CD4 + T cells) [345]. However, until date no relevant data have been published suggesting that other VGSC antagonists (such as HYP-17 [346], A-803467 [347,348], PF-05089771 [349], phenytoin [350] or tetrodotoxin [351,352]) may substitute non-steroidal anti-inflammatory drugs let alone may suppress COVID-19-related hyperinflammation [353]. A plausible reason is that during hyperinflammation-including hyperinflammation in COVID-19-the cytokine levels (i.e. IL-1β [354], IL-6, IL-10 [355,356] and IL-12 [357]) are high. Reportedly, IL-1β [358] and IL-6 [359] inhibit sodium currents of VGSCs, and IL-10 downregulates the expression of VCSCs [360]. Moreover, activation of the P2X7R reduced the density and currents of VGSCs [361]. Therefore, we do not consider the inhibitory properties of lidocaine on VGSCs to be relevant for the treatment of hyperinflammation in COVID-19. At first glance, the downregulation of the expression of TLR 2 [338] and TLR 4 [318] is an important antiinflammatory mechanism directly induced by lidocaine. But at a closer look, it appeared that activation of P2X7R by the agonist cathelicidin (LL-37) leads to the upregulation of the protein expression of TLR2, TLR3, TLR4 and TLR 5 [288]. This is in line with the MyD88 (myeloid differentiation primary-response protein 88)-dependent activation of NF-κB (nuclear factor kappa-light-chain-enhancer of activated B cells) following the activation of the P2X7R by BzATP [362,363]. The MyD88-dependent activation of NF-κB is part of the TLR4/NF-κB pathway. Therefore, it is unsurprising that the inhibition of P2X7R by its antagonists (Brilliant Blue G, A-438079 and A-740003) neutralises the abovementioned P2X7R-induced upregulation of TLRs [362]. Consequently, we argue that lidocaine inhibits inflammation directly by blocking P2X7Rs independent from the neutralisation of the P2X7R-induced upregulated TLR2 and TLR4.
The subpopulation of NMDA receptors present on the peripheral neurons are involved in nociception, and their number increases during inflammation contributing to the sensitisation of peripheral nerves to nociceptive stimuli. NMDA receptor antagonists have anaesthetic-like effects [364]. In addition, NMDA receptor antagonist can prevent hypoxic neuronal death, IL-1β and TNFα release [365], reduce the activation of inflammatory experimental colitis [366] and suppress glial pro-inflammatory cytokine expression [367]. Moreover, the NMDA receptor antagonist memantine can increase IL-10 inhibition: 3.00 mM (702.90 μg/ml); and 98% inhibition: 10.00 mM (2343.00 μg/ml), respectively. The usual plasma concentrations in clinical settings are indicated by the green box, and the targeted concentrations in the lymph nodes are indicated by the magenta box. Note that the maximal tolerable plasma levels for human (about 4.7 μg/ml-0.02 mmol/L) are much lower than the required extracellular concentrations of lidocaine to effectively inhibit the P2X7R. Source: Okura D, et al. [74] production in BCR/CD40-activatedB-cells [368]. Lidocaine inhibits NMDA receptors [339,369,370], and thus the antiinflammatory properties of lidocaine could be attributed to the inhibition of NMDA receptors. However, it has been reported that the anti-inflammatory effect in T-cell functions (inhibition of antigen-specific T-cell proliferation, T-cell cytotoxicity, Tcell migration towards chemokines and decrease in IL-2 and IFN-γ production by Th1 effector cells in favour of IL-10 and IL-13 production by Th2 cells) of the NMDA receptor antagonist ifenprodil is effective both in wild-type and in NMDA receptor (GluN1) knockout mice [371]. Moreover, it was found that KN-62, an inhibitor of Ca2+/calmodulin-dependent kinase type II and a potent P2X7R antagonist, provides neuroprotection against NMDA-induced cell death [372]. Therefore, we argue that the anti-inflammatory properties of NMDA receptor antagonists (including lidocaine) should be attributed to the inhibition of P2X7Rs rather than to the inhibition of NMDA receptors.

Selective inhibition of the P2X7Rs of the immune cells in the lymphatics avoids exceeding the maximal tolerable plasma concentration of lidocaine and inhibits hyperinflammation in two stages
As mentioned above, the main issue is that the IC 50 for P2X7R inhibition is much higher than the maximal tolerable plasma concentration for mammals because P2X7Rs are indispensable for normal physiological functions (i.e. in the central nervous system [373], the peripheral nervous system [374] and in the lungs [60, 61]). Therefore, intravenous or oral administration aimed at achieving an effective concentration of lidocaine to inhibit P2X7Rs in serum and in target organs will hamper organ functions and is potentially dangerous.
The lymphatic system is populated exclusively by trafficking immune cells, i.e. naïve T cells, activated T cells, B cells [375], dendritic cells [376], monocytes [377], macrophages [378], neutrophils [379], mast cells [380], eosinophils [381] and basophils [382]. We postulate that selective inhibition of the P2X7Rs of the immune cells of the lymphatic system by lidocaine suppresses hyperinflammation in two stages: stage 1, the selective inhibition of the P2X7Rs of the immune cells residing in the lymph nodes induces clonal expansion of Tregs in these lymph nodes; stage 2, subsequently, these Tregs migrate throughout the body exerting anti-inflammatory activities reducing systemic and (distant) local hyperinflammation (Fig. 1).
The endothelium of the dermal capillaries of the skin belongs to the structural type "continuous endothelium" [383]. Although capillary walls can transport substances from blood to tissue, the absorption of substances from tissue to blood is, if any, extremely low [384]. Apparently, specialised initial lymphatics harbouring one-way valve leaflets capable of absorbing fluid and molecules from the interstitium are localised in the dermis. The absorbed lymph fluid is then propelled forward in the lymphatic network by collecting lymphatic vessels harbouring a rhythmic contracting muscle layer [385]. This system brings fluids and particles into the lymph nodes where numerous immune processes take place. The administration route to target the lymphatic system in a domestic swine model is illustrated by the subcutaneous or intradermal injection of compounds (isosulfan blue, fluorescein and radioactive technetium-99 isotope-Tc 99 ) and by tracing the extent and the transit time of the distribution of these compounds using whole body scintigraphy in pigs [386]. The absorption of intradermal application of radioactive Tc 99 into the lymph nodes is 10 times faster than after deep subcutaneous application and leads to higher concentrations in the lymph nodes related to these lymphatic vessels [386]. Radionuclide lymphoscintigraphy with molecules of different sizes after intradermal and subcutaneous injections showed that smaller particles (i.e. 99mTc-dextran and 99mTc-human serum albumin) migrate more rapidly towards the lymphatic vessels and lymphatic nodes than larger particles (i.e. radiocolloids of larger molecular size) [387]. The rate of clearance of 99m Tc-pertechnetate and 99m TcDTPA after subcutaneous and intradermal administration in the back of the hand in humans is 1 %/min and 8 to 10 %/min, respectively [387].
The additional advantage is that the plasma concentrations of subcutaneously administered lidocaine are much lower than intravenously administered lidocaine. Intravenous administration of 2 mg/kg lidocaine in cats is almost immediately followed by a peak plasma concentration of 3.6 μg/mL [388]. In contrast, the achieved mean peak plasma concentrations after the subcutaneous administration of 30 mg/kg, 20 mg/kg and 10 mg/kg lidocaine are much lower: 1.69, 1.07 and 0.77 μg/mL, respectively [389]. Note that the applied subcutaneous dose [389] is 15, 10 and 5 times higher than the intravenous dose, respectively [388,389]. Reportedly, the difference in the plasma concentrations after intravenous and subcutaneous administration of lidocaine is caused by the fact that, in contrast to the intravenous administration, a large proportion of the subcutaneously administered lidocaine is drained into the lymphatic system [390][391][392]. Obviously, this slows down the release of lidocaine to the venous blood. This is confirmed for bevacizumab in mice [390], for trastuzumab in rats [391] and for docetaxel in rats by [392].
As stated above, lymphatic absorption after intradermal administration is much higher than after deep subcutaneous administration [386,387]. Practically, the intradermal infusion with lidocaine is not an accepted administration route for lidocaine. Therefore, we argue that a subdermal administration of lidocaine using a catheter inserted just beneath the dermis (subdermal infusion, Fig. 5) will result in higher concentrations of lidocaine in the draining local lymph nodes than a deep subcutaneous or intravenous infusion as depicted in the schematic presentation of the putative distribution of lidocaine in Fig. 6.
In summary, by means of the subdermal administration of lidocaine, we can ensure high concentrations of lidocaine in the local lymph nodes enabling an effective inhibition of the P2X7R of the immune cells while keeping the lidocaine plasma concentrations <4.7 μg/ml(stage 1a and 1b in Fig. 1). The induced Tregs clonal expansion in these local lymph nodes produces Tregs which migrate t h r o u g h o u t t h e b o d y c o n t r o l l i n g t h e o n g o i n g hyperinflammation (stage 2 in Fig. 1). Obviously, the subdermal administration route may also apply to other P2X7R antagonists.
Three other P2X7R antagonists have been tested in human: CE-224,535 500 (Pfizer), AZD-9056(Astra-Zeneca) and JNJ-54175446 (Johnson and Johnson). A phase IIa study with CE-224,535 in patients with rheumatoid arthritis not responding adequately to methotrexate was recently reported [393]. Patients in the treatment arm received oral CE-224,535 500 mg twice/day for 12 weeks. Although the safety and tolerability for the compound were acceptable, CE-224,535 was not effective in this group of patients. The results of a phase II study with AZD-9056 in patients with active rheumatoid arthritis despite treatment with methotrexate or sulphasalazine was published. The treatment arm consists of oral AZD-9056 100 or 400 mg/day for 6 months [394]. The AZD-9056 used in this trial is non-lipophilic as indicated by the fact that this compound cannot penetrate the blood-brain barrier [395]. The authors conceded that "AZD-9056 does not have significant efficacy in the treatment of RA, and the P2X7 receptor does not appear to be a therapeutically useful target in RA" [394]. Recently, a randomised, placebo controlled, sequential-group, single-centre ascending dose phase I study was reported. The patients in the 5 treatment arms received 0.5, 2.5, 10, 50, 150 and 300 mg JNJ-54175446, respectively. The authors reported dose-dependent plasma levels, no serious adverse events, ex vivo attenuation of lipopolysaccharide-induced IL-1β release in peripheral blood and confirmation of passive brain penetration of JNJ-54175446 [396]. The approach of the P2X7R antagonist therapy of the above-mentioned authors is quite different from ours: While these authors directly targeted the diseased organs via the gut absorption of the drug, we target the immune cells in local lymph nodes inducing an anti-inflammatory immune response which in turn targets the diseased organs (Fig. 1). This is illustrated by the following study concerning a placebo-controlled, multicentre, doubleblind phase IIa study in patients with moderately to severely active Crohn's disease. The patients in the treatment arm received oral AZD-9056 200 mg/day for 28 days. The authors found a significant improvement in the Crohn's Disease Activity Index (CDAI) at day 28 [397]. In contrast to the skin, the endothelium of the mucosal capillaries of the mouth and the gastrointestinal tract are fenestrated allowing molecules to pass from the submucosal tissue into the capillaries [383]. Unlike the failure of the treatment of rheumatoid arthritis described above, the successful treatment of gut inflammation here can be attributed to the absorption of non-lipophilic oral AZD-9056 by the mucosa-associated lymphoid tissue (MALT). This is the inductive site of the mucosal immune system consisting of mesenteric lymph nodes, Peyer's patches and isolated lymph follicles [398,399]. Although lymphatic transport to the lymph nodes of the non-lipophilic oral AZD-9056 is limited [400,401], AZD-9056 inhibits P2X7Rs of the local T-cells via absorption by the inductive sites of MALT. This induces a local anti-inflammatory immune response executed by the effector sites of MALT consisting of lamina propria lymphocytes and intraepithelial lymphocytes [398,399].

Six COVID-19-induced ARDS patients
From April 2020 until end of July 2020, two of the authors of this report (AS and TK) have successfully treated six critically ill patients with COVID-19 admitted to the ICU of the Showa University in Tokyo, Japan, with lidocaine. The lidocaine treatment was based on off-label use. The Medical Ethical Committee of the Showa University, School of Medicine, Tokyo, approved the collection, analysis and publication of patients on mechanical ventilation admitted to the ICU (protocol number 3313). The administration was initially intravenously in the two first patients, followed by subdermally (a superficially inserted subcutaneous catheter as illustrated in Fig. 5). In the other four patients, only the subdermal administration was further applied. The concentration of the intravenous lidocaine infusion solution is 20 mg/ml (2%), the route for continuous administration of lidocaine commonly used in daily practice. The dose for intravenous administration is 0.6 mg/kg/h as recommended earlier [402]. Due to the limited efficacy of intravenous lidocaine and based on the hypothesis of selectively targeting the inhibition of the P2X7Rs of the immune cells, the infusion in both patients was converted to subdermal infusion of 1.0 mg/kg/h (dosage as reported by Japanese researchers [403]) after 7 and 6 days, respectively. The time course of clinical parameters of these six patients  6 Schematic presentation of the putative distribution of intravenous, oral, transmucosal (i.e. in the oral cavity) and subdermal administered lidocaine. Administration of hydrophilic lidocaine (lidocaine HCL) through a (central) venous catheter or by oral intake results in concentration gradients with the highest value in the venous blood and the lowest value in the lymph nodes. The reason is that by the time lidocaine reaches the lymph nodes, the drug is massively diluted and may never reach the effective concentration required to adequately inhibit the P2X7Rs of the immune system. In contrast, after subdermal injection of hydrophilic lidocaine, apart from a minimal absorption by the dermal capillaries, almost all the lidocaine is absorbed by the lymphatic system via the initial lymphatics. Because the fluid in the afferent collecting lymphatics originates from the interstitial fluid of the tissues, dilution of the concentration of lidocaine occurs. This fluid is then drained into the local lymph nodes. The extent of the dilution of lidocaine in the targeted lymph nodes is far less drastic compared to the (central) venous administration of the drug. We postulate that with continuous subdermal infusion, we can achieve concentrations of lidocaine in the lymph nodes sufficient to effectively inhibit the P2X7Rs of the immune cells. Theoretically, similar results may be expected from transmucosal and transdermal administration of lipophilic lidocaine base with a high concentration. Obviously, the subdermal, transmucosal and transdermal administration routes may also apply to other P2X7R antagonists is presented in Figures 7, 8, 9, 10. In about 20% of the inserted subdermal cannulae, local subdermal indurations were observed. Whenever this occurred, the infusion cannula was removed and replaced with a new cannula at a different location. The first patient (Fig.7), a 63-year-old male (75 kg, 168 cm), developed fever and nausea on March 27, 2020, and 3 days later, he started to cough and became dyspnoeic. After 5 days, the PCR SARS-Cov-2 test was positive, and he was admitted to the hospital with SARS-Cov-2-induced ARDS. Co-morbidities include COPD and smoking 60 cigarettes per day for more than 40 years. About 40 years earlier, the patient suffered from pneumothorax. On day 3, the patient deteriorated and was intubated and mechanically ventilated due to poor blood gases. No haemodynamic instability was observed. The CT scan showed bilateral ground glass opacities compatible with ARDS. On day 5, the patient was transferred to the ICU of the university hospital because of further respiratory deterioration. The patient received favipiravir for 14 days after admission; the patient did not receive dexamethasone. Prone position mechanical ventilation was initiated due to the progression of the respiratory disease with an extremely low PaO 2 /FiO 2 ratio of 63.3 mm Hg (severe ARDS according to the Berlin definition. The Berlin definition of ARDS includes severe PaO 2 /FiO 2 ratio ≤100 mm Hg, moderate PaO 2 /FiO 2 >100 to 200 mm Hg, mild PaO2/FiO2 >200 to 300 mm Hg, no ARDS PaO2/FiO2 >300 mm Hg [404]). The initial ventilator settings include APRV, P high 27 cm H 2 O, T high 7.0 s, P low 0 cm H 2 O and T low 0.32 s. The Fig. 7 Patient 1, the first of the six cases with severe COVID-19 treated with subdermal lidocaine in the ICU of the Showa University, Tokyo, Japan. A 63-year-old male with COVID-19-induced ARDS, was admitted to the hospital. The CT scan showed bilateral ground glass opacities. Co-morbidities: COPD, smoking 60 cigarettes per day for more than 40 years. About 40 years before admission, the patient suffered from pneumothorax. After admission the clinical condition deteriorated requiring an ICU admission and mechanical ventilation on day 4. On day 11, continuous intravenous lidocaine of 0.6 mg/kg/h was initiated, but the patient's condition kept worsening with high pulmonary artery pressures and reduced aeration of the lung. On day 19, the continuous intravenous lidocaine of 0.6 mg/kg/h was changed to continuous subdermal lidocaine of 1 mg/kg/h. This was followed by improvement of the clinical condition, and on day 20, the aeration of the lung was improved, but the pulmonary artery pressures remained high. Despite this the P/F ratio was gradually improving, and ECMO weaning was done on day 50. No new ECG changes were observed during treatment with lidocaine. Blood metHb were within the normal range (0.3-0.8%). On day 99, he was weaned from the mechanical ventilator and was discharged from the ICU on day 121. CT scan on day 146 showed reduced ground glass opacities in both lungs and some interstitial change in upper and middle fields of the lung and improvement of the pneumothorax. The patient was discharged from the hospital on day 187, he went home, and he could walk but needed extra oxygen supply of 2L/min. Nine months after admission, the patient is doing well and has returned to work. The patient visited the hospital 3 months after discharge: He only uses oxygen 1 L/min to go shopping and during physical training (out-patient rehabilitation). He talked to the treating intensivist without requiring oxygen and had no shortness of breath or tachypnoea. The red-coloured labels of the legends refer to graph plots using the (left) primary Y-axis, and the black-coloured labels of the legends refer to graph plots using the (right) secondary Y-axis PaCO 2 was normal. The echocardiographic estimated pulmonary arterial systolic pressure (PASP) was 80 mm Hg. The Krebs von Lungen 6 (KL-6, a marker for lung fibrosis [405]) plasma level was highly elevated (1299 U/mL; normal value <425 U/mL), CRP was also high (40.4 mg/L; normal value <10 mg/L), and albumin was 2.2 g/dl. The white blood cell count, platelet count and urine production were normal. On day 4, the chest X-ray was not improved. On day 6, the PaO 2 /FiO 2 ratio was slightly increased but remained low at 103 mm Hg, and the chest X-ray showed progression of the ARDS. ECMO was initiated due to exhausted ventilatory strategy. On day 9, the PaO 2 /FiO 2 ratio improved but remained low at around 153 mm Hg, but the CRP declined to around 21.8 mg/L. The patient was put on muscle relaxants. The patient's ARDS status had improved from severe to moderate ARDS. From day 10 until day 30, the ferritin levels were well >1000 ng/ml (>100 μg/dl, normal values <300 ng/ml). From day 11 until day 62, D-Dimer was very high reaching 121.9 nM/L day 14. On day 11, no improvement of the blood gases was observed, and it was decided to treat the patient with continuous intravenous lidocaine 0.6 mg/kg/h. The CRP showed a progressive decline from 19 (on day 12) to 12.8 (on day 16) and 7.4 (on day 19), but the PaO 2 /FiO 2 ratio remained poor at around 90 mm Hg (severe ARDS according to the Berlin criteria) and the chest X-ray image on day 15, 3 days after the initiation of the intravenous lidocaine infusion, deteriorated dramatically. The lidocaine plasma concentrations were 3.4 μg/ml on day 13 and 5.4 μg/ml on day 14. On day 19, the continuous intravenous lidocaine infusion was replaced by continuous subdermal lidocaine infusion of 1 mg/kg/h. Although the PaO 2 /FiO 2 ratio remained unchanged on day 20 (1 day after the switch to the continuous subdermal lidocaine), the chest X-ray improved clearly. On day 21, the lidocaine plasma concentration was 2.6 μg/ml, and albumin was 2.5 g/dl. From day 22, the PaO 2 /FiO 2 ratio was gradually improving reaching 151 mm Hg on day 34 The CT scan showed bilateral ground glass opacities. Co-morbidity: Asthma. After admission the patient's condition was deteriorating. On day 5, continuous intravenous lidocaine of 0.6 mg/kg/h was initiated, but the clinical condition and the P/F ratio kept worsening. On day 11, the intravenous lidocaine of 0.6 mg/kg/h was changed to continuous subdermal lidocaine of 1 mg/kg/h. A few days later, this was followed by improvement of the clinical condition and the P/F ratio. No new ECG changes were observed during treatment with lidocaine. Blood metHb were within the normal range (0.1-0.6%). The patient was discharged from the ICU on day 30 home on day 37. At 3 months after admission, the patient is doing well. The red coloured labels of the legends refer to graph plots using the (left) primary Y-axis, and the black-coloured labels of the legends refer to graph plots using the (right) secondary Y-axis (moderate ARDS). The KL-6 on day 22 dropped to 458 U/L (this is only slightly above the normal value of <450 U/l). On day 31, the CRP was low at 1 mg/L, and the lidocaine plasma concentration was 1.2 μg/ml. The muscle relaxants were discontinued. Albumin was 2.3 g/dl. On day 33, the chest Xray was further improved, and the CRP remained low at 5.5 mg/L. The patient was awake and could communicate with the nurses. On day 38, the lidocaine plasma level was 2.3. On day 43, the PaO 2 /FiO 2 ratio was increased to 214 mm Hg. According to the Berlin definition of ARDS [404], the patient's ARDS status had changed from moderate to mild. Albumin was 2.8 g/dl. On day 50, the patient was weaned from ECMO. On day 51, the patient underwent tracheotomy. Because the clinical condition of the patient was stabilised with a low CRP of 6.3 mg/L on day 55, the continuous subdermal lidocaine was discontinued on day 57. On day 69, he developed pneumothorax requiring pleural drainage. On day 99, he was weaned from the mechanical ventilator and was discharged from the ICU on day 121. No new ECG changes were observed during treatment with lidocaine. Blood methaemoglobin (metHb) were within the normal range (0.3-0.8%). CT scan on day 146 showed reduced ground glass opacities in both lungs, some interstitial change in upper and middle fields of the lung and improvement of the pneumothorax. The patient left the hospital on day 187, he went home, and he could walk but needed extra oxygen supply of 2L/min. Nine months after admission, the patient is doing well and has returned to work. The patient visited the hospital 3 months after discharge: He only uses oxygen 1 L/ min to go shopping and during physical training (out-patient Fig. 9 Left graph: Patient 3. A 59-year-old male with respiratory distress and bilateral ground glass opacities on the CT scan. Co-morbidity: Obesity, diabetes mellitus and gout. No new ECG changes were observed during treatment with lidocaine. Blood metHb were within the normal range (0.1-0.4%). The patient was discharged from the ICU on day 8 and was discharged home on day 20. After 3 months, he is doing well. Right graph: Patient 4. A 51-year-old male with fever, dyspnoea and cough due to COVID-19. The CT scan showed bilateral ground glass opacities. Co-morbidity: none. No new ECG changes were observed during treatment with lidocaine. Blood metHb were within the normal range (0.1-0.3%). The patient was discharged from the ICU on day 8 and was disc harged home on day 28. At 3 months, he is doing well and has returned to work. The red-coloured labels of the legends refer to graph plots using the (left) primary Y-axis, and the black-coloured labels of the legends refer to graph plots using the (right) secondary Y-axis rehabilitation). He talked to the treating intensivist without requiring oxygen and had no shortness of breath or tachypnoea.
The second patient (Fig.8) is a 68-year-old male (75 kg, 164 cm) with SARS-Cov-2-induced ARDS and positive SARS-Cov-2 PCR test admitted to the university hospital. Co-morbidity is asthma. The CT scan showed bilateral ground glass opacities. Haemodynamically the patient was stable. The patient received tocilizumab on day 8 and favipiravir for 14 days; he did not receive dexamethasone. On day 2, the respiratory conditions deteriorated, and the PaO 2 /FiO 2 ratio is 118 mm Hg (moderate ARDS according to the Berlin ARDS definition [404]). The patient was intubated and required mechanical ventilation. The initial ventilator settings include pressure control, peak inspiratory pressure 28 cm H 2 O, PEEP 13 cm H 2 O and respiratory rate 30/min. CRP was 10.6 mg/L, and KL-6 was 486 U/ml. White blood cell count, platelet count and urine production were normal. The ferritin levels remained >1000 ng/ml (100 μg/dl) during the entire ICU stay. Albumin was 2.9 g/dl. In the following 3 days, the PaO 2 /FiO 2 ratio improved to around 150 mm Hg. The PaO 2 /FiO 2 ratio dropped from 152 on day 5 to 84 mm Hg on day 6. CRP was increased to 22.9, and the KL-6 was increased to 762 U/ml. The patient was put in prone position and given muscle relaxants. Continuous intravenous lidocaine of 0.6 ml/kg/h was started. Albumin was 1.8 g/dl. On day 7, the PaO 2 /FiO 2 ratio increased to 128 mm Hg, CRP dropped to 10.3 mg/mL and the lidocaine plasma concentration was 2.2 μg/ml. From day 3 until discharge from the ICU, D-dimer values were elevated reaching 75 nM/L on day 14. On day The CT scan showed bilateral ground glass opacities. Co-morbidity: Fatty liver. No new ECG changes were observed during treatment with lidocaine. Blood metHb were within the normal range (0.1-0.3%). On day 14, the patient was discharged from the ICU. On day 20, the patient was discharged home and is doing well at 3 months after admission. Right graph: Patien ts 6. A 59-year-old male with fever, dyspnoea and cough due to COVID-19. CT scan showed bilateral ground glass opacities. Co-morbidity: Hypertension on medication. No new ECG changes were observed during treatment with lidocaine. MetHb were within the normal range (0.1-0.3%). On day 13, the patient was discharged from the ICU. He was discharged from the hospital on day 20, and a t 3 months after admission, he is doing well, played golf regularly and has returned to work. The red-coloured labels of the legends refer to graph plots using the (left) primary Y-axis, and the black-coloured labels of the legends refer to graph plots using the (right) secondary Y-axis 8, although the PaO2/FiO2 ratio improved from 84 to 125 mm Hg, the mechanical ventilatory strategies were exhausted, and the patient was put on ECMO. The KL-6 was increased to 845 U/L, and lidocaine plasma level was 2.9 μg/ml. The PaO 2 / FiO 2 ratio improved to 238 mm Hg on day 9, but on day 10, a sharp drop of the PaO 2 /FiO 2 ratio to 60 mm Hg was observed, and CRP was 2.0 mg/ml. The patient's ARDS status had changed from moderate to severe according to the Berlin ARDS criteria [404]. Lidocaine treatment was switched from continuous intravenous to continuous subdermal (dosage: 1 mg/kg/h). On day 14, the lidocaine plasma level was 2.7 μg/ml. KL-6 dropped to 549 U/l. On day 17, the clinical condition of the patient was improving, and the PaO 2 /FiO 2 ratio reached 158 mm Hg. The patient was weaned from ECMO. The PaO 2 /FiO 2 ratio improved further reaching 291 mm Hg on day 21, and the patient's ARDS status has changed from moderate to mild ARDS [404]. On day 22, mechanical ventilation was discontinued, and the patient was extubated. The patient was orientated, and no signs of confusion were detected. CT scan on day 25 showed persistent ground glass opacities in both lungs, some pulmonary effusion (right >left), and no signs of vascular thrombosis. In addition, no signs of deep venous thrombosis were found in the lower extremities. Lidocaine treatment was continued until discharge from the ICU on day 30. No new ECG changes were observed during treatment with lidocaine. Blood metHb were within the normal range (0.1-0.6%). The patient was discharged home on day 37. At 3 months after admission, the patient is doing well.
The third patient (Fig. 9left), a 59-year-old male (109 kg, 170 cm), was admitted to the university hospital with respiratory distress and bilateral ground glass opacities on the CT scan with a positive SARS-CoV-2 test. Comorbidities include obesity (BMI 37.7 kg/m 2 ), diabetes mellitus and gout. The patient required immediate intubation and mechanical ventilation. The patient received tocilizumab on day 3 and favipiravir for 15 days and did not receive dexamethasone. The initial ventilator settings are pressure control, peak inspiratory pressure 30 cm H 2 O, PEEP 15 cm H 2 O and respiratory rate 25/min. The PaO 2 /FiO 2 ratio on admission was 160 mm Hg (moderate ARDS according to the Berlin definition [404]), CRP was 39.3 mg/L and KL-6 was 294 U/ml. White blood cell count was increased (13.10 −9 /L) and platelet count and urine production were normal. Albumin was 2.1 g/dl. Haemodynamic parameters were stable. On the admission day, continuous subdermal lidocaine was started at 1 mg/kg/h. On day 2, the PaO 2 /FiO 2 ratio improved to 283 mm Hg, and the patient's ARDS status had changed from moderate to mild ARDS. CRP was 41 mg/L, KL-6 was 268 U/L and the lidocaine plasma level was 3.7 μg/ml. Albumin was 1.7 g/dl. On day 4, the PaO 2 /FiO 2 ratio was 302 mm Hg, and the patient's ARDS status had changed from mild ARDS to no ARDS according to the Berlin ARDS criteria. On day 5, the PaO 2 /FiO 2 ratio was improved further to 328 mm Hg, and CRP dropped to 16.4, and the patient was extubated. The patient was orientated, no signs of confusion were detected. The patient was discharged from the ICU on day 8; CRP was 2.3 mg/ml. Albumin was 2.5 g/dl. No new ECG changes were observed during treatment with lidocaine. Blood metHb were within the normal range (0.1-0.4%). The patient was discharged home on day 20. After 3 months, he is doing well.
The fourth patient (Fig. 9right) is a 51-year-old male (68 kg, 175 cm). Ten days before admission, he developed fever and 2 days before admission dyspnoea and coughing. On the day of admission, the PCR SARS-CoV-2 test was positive. The CT scan showed bilateral ground glass opacities. Comorbidity is none. The patient was intubated and put on mechanical ventilation on admission. The patient received favipiravir for 14 days; he did not receive dexamethasone. On day 3, he was transferred to the university hospital because of deterioration of pulmonary condition. The initial ventilator settings include pressure control, peak inspiratory pressure 24 cm H 2 O, PEEP 12 cm H 2 O and respiratory rate 15/min. The haemodynamic conditions were stable. White blood cell count and platelet count were normal. Albumin was 2.6 g/dl. Continuous subdermal lidocaine was started immediately. On day 3, the PaO 2 /FiO 2 ratio was 214 (moderate ARDS according to the Berlin definition [404]). KL-6 was 177 U/L, and CRP was 17.4 mg/L. On day 5, the PaO 2 /FiO 2 ratio was increased to 382 (the patient's ARDS status had changed from mild ARDS to no ARDS), and lidocaine plasma concentration was 5.2 μg/ml. CRP was 27.3mg/L. Lidocaine plasma levels on day 3 and 4 were 3.4 and 4.2 μg/ml, respectively. KL-6 was 163 U/L. The patient was extubated. The patient was orientated, and no signs of confusion were detected. The patient was discharged from the ICU on day 8, and the CRP was 9.3 mg/L. No new ECG changes were observed during treatment with lidocaine. Blood metHb were within the normal range (0.1-0.3%). He was discharged home on day 28. At 3 months, he is doing well and has returned to work.
The fifth patient (Fig. 10left) is a 58-year-old male (80 kg, 175 cm). Nine days before admission, he developed a sore throat. A day later, he developed fever. Two days before admission, he started coughing and was dyspnoeic. On the day of admission, the PCR SARS-Cov-2 test was positive. Comorbidity includes fatty liver. The CT scan showed bilateral ground glass opacities. The patient was initially admitted to the hospital ward. The patient received tocilizumab on day 7 and favipiravir for 10 days; dexamethasone was not prescribed. On day 3, the patient deteriorated and had to be intubated and put on mechanical ventilation. On day 4, the patient was transferred to the university hospital due to deterioration of the pulmonary condition. The initial ventilator settings include pressure control, peak inspiratory pressure 27 cm H2O, PEEP 12 cm H2O and respiratory rate 25/min. PaO 2 /FiO 2 ratio was 188 (moderate ARDS according to the Berlin definition). Haemodynamic parameters were stable, and CRP was 12.9 mg/ml. White blood cell count was increased (14.4.10 9 /L), but platelet count was normal. KL-6 was 330 U/L. Continuous subdermal lidocaine was started at 1 mg/kg/h at arrival at the ICU of the university hospital. Albumin was 2.8 g/dl. On day 5, the PaO 2 /FiO 2 ratio was unchanged, CRP was 10.4 mg/L and the lidocaine plasma level was 4 μg/ml. On day 6, the lidocaine plasma level was 3.2 μg/ml. KL-6 remained stable at 400 U/L. Albumin was 2.3 g/dl. On day 10, the respiratory insufficiency had cleared; although the PaO2/FiO2 ratio remained 184, the CRP dropped to 2.4 mg/L, and KL-6 was 322 U/L. The patient was extubated, and he was orientated; no signs of confusion were detected. On day 14, the patient was discharged from the ICU. No new ECG changes were observed during treatment with lidocaine. Blood metHb were within the normal range (0.1-0.3%). On day 20, the patient was discharged home and is doing well at 3 months after admission.
The sixth patient (Fig. 10right) is a 59-year-old male (65 kg, 175 cm) with fever, dyspnoea and cough due to COVID-19. CT scan showed bilateral ground glass opacities. Co-morbidity includes hypertension on medication. The patient was admitted to the general ward. KL-6 233 U/L, white blood cell count and platelet count were normal. Albumin was 3.6 g/dl. On day 3, there is a deterioration of the respiratory function necessitating a transfer to the ICU and mechanical ventilation. Tocilizumab was given on day 4. The patient received favipiravir for 11 days, and the patient did not receive dexamethasone. The initial ventilator settings include pressure control, peak inspiratory pressure 22 cm H 2 O, PEEP 10 cm H 2 O and respiratory rate 20/min. Continuous subdermal lidocaine of 1 mg/kg/h was initiated after admission to the ICU. Haemodynamic parameters were stable. CRP was 6.3 mg/L, and KL-6 was 263 U/L. On day 4, a progressive respiratory failure occurred requiring intubation and mechanical ventilation. PaO 2 /FiO 2 ratio was 218 mm Hg; the haemodynamic parameters remained stable. CRP was 6.3 mg/L, and the white blood count and platelet count were normal. Lidocaine plasma level was 4.6 μg/ml. On day 5, the PaO 2 /FiO 2 ratio dropped further to 164 mm Hg. Lidocaine plasma level was 3.4 μg/ml. Albumin was 3.2 g/dl. On day 9, the clinical condition of the patient improved. The ventilator settings could be decreased, the PaO 2 /FiO 2 ratio remained 207 mm Hg during the weaning period, and CRP was 0.7 mg/L. On day 10, the patient was extubated, he was orientated, and no signs of confusion were detected. On day 13, the patient was discharged from the ICU. No new ECG changes were observed during treatment with lidocaine.
Blood metHb were within the normal range (0.1-0.3%). He was discharged from the hospital on day 20, and at 3 months after admission, he is doing well, played golf regularly and has returned to work.

Additional 14 patients with COVID-19-induced ARDS
From July 2020 until beginning of December 2020, 14 additional critically ill patients with COVID-19-induced ARDS requiring mechanical ventilation were treated in the ICU of the Showa University with continuous subdermal lidocaine infusion (1 mg/kg/h) plus intravenous or oral dexamethasone (6 mg/day) as reported earlier [35]. Of these 20 patients, 19 survived, but an 87-year-old female patient died of invasive aspergillosis. No other patient developed secondary coinfections (unpublished data, personal communication by AS and TK).

Discussion of the clinical cases
After completing the novel definition of hyperinflammation, we developed a new approach to target the lymphatic system with continuous subdermal administration of lidocaine. This is meant to increase the anti-hyperinflammatory effect of lidocaine while avoiding toxic plasma levels. We described the treatment of six critically ill patients with COVID-19 with lidocaine. Two patients required mechanical ventilation and ECMO, and four patients were treated with mechanical ventilation. As mentioned under the heading "Introduction", the case fatality rates of patients requiring mechanical ventilation and/or ECMO are alarmingly high [1,2]. Patient 1 and patient 2 were older than 60 years. Additionally, patient 1 had COPD and had smoked 60 cigarettes per day for more than 40 years. Patient 3 suffered from obesity and diabetes mellitus. These are serious prognostic factors for bad outcome COVID-19 [406,407]. Patient 1 and patient 2 were initially treated with continuous intravenous lidocaine through a central venous line. In both patients, the pulmonary conditions deteriorated after the initiation of intravenous lidocaine: Patient 1 who was already on ECMO showed progressive pulmonary deterioration on the chest X-rays, and patient 2 deteriorated further necessitating the initiation of ECMO therapy. Remarkably, the pulmonary conditions of both patients improved within 48 h after the switch from intravenous to subdermal continuous lidocaine. The lidocaine plasma levels remained around 5 μg/ml. To our knowledge, these six cases represent the first observations of the promising treatment of critically ill COVID-19 patients with lidocaine targeting P2X7Rs of the immune cells in the lymphatics. All patients recovered completely from their illness. None of the patients showed the feared side effect of cardiac arrhythmia and methaemoglobinaemia during lidocaine therapy. Our findings suggest that continuous subdermal lidocaine infusion at the rate of 1 mg/kg/h has the potential to mitigate hyperinflammation and ARDS in critically ill COVID-19-patients. Obviously, although all six patients appeared to respond positively to the treatment and no severe adverse effects were observed, no final conclusions can be made on the efficacy of lidocaine in critically ill COVID-19 patients.
Researchers from Lima, Peru, reported the treatment of 28 (three mild, 21 moderate and four severe) COVID-19 patients with 0.5% lidocaine HCL solution with an intravenous dose of 1 mg/kg once a day for 2 days and 2% lidocaine HCL solution with a subcutaneous dose of 1 mg/kg once a day for 2 days [408]. The authors aimed at the improvement of pain, cough, respiratory rate and oxygen saturation. They found improvement in most patients. In severe cases, this treatment did not improve the oxygen saturation. As expected, treatment with a low daily dose of lidocaine once per day for a total treatment duration of 2 days could not adequately inhibit the P2X7Rinduced hyperinflammation in COVID-19.
Recently, a group of researchers from Strasbourg, France, announced a study entitled: "Impact of intravenous lidocaine on clinical outcomes of patients with ARDS during COVID-19 pandemia (LidoCovid): A structured summary of a study protocol for a randomised controlled trial" (ClinicalTrials.gov Identifier: NCT04609865) [409].
Lately, an extraordinary treatment of COVID-19 ARDS was reported [410]. The authors performed lung transplantations in three critically ill COVID-19 ARDS patients: a 28-year-old female, a 62-year-old male and a 43-year-old male. The first patient underwent lung transplantation after weeks on veno-venous ECMO support with elevated pulmonary arterial pressures and severe secondary Serratia marcescens pneumonia. The second patient underwent lung transplantation after 100 days on v e n o -v e n o u s E C M O s u p p o r t c o m p l i c a t e d b y Pseudomonas aeruginosa pneumonia, haemothorax and empyema, while the third patient after 90 days on the mechanical ventilator. This patient suffered from many complications: asystolic cardiac arrest, heparin-induced thrombocytopenia, a left frontal lobe infarct of the cerebral cortex, Serratia marcescens-mediated pneumonia with bacteraemia, acute kidney injury, a left haemothorax requiring thoracotomy and lung decortication, a right pneumothorax requiring tube thoracostomy, hypernatremia associated with seizures and malnutrition. Before lung transplantation, the patient developed increasing clinical signs of pulmonary fibrosis and severe pulmonary hypertension. The first two patients are reported to have achieved independence in daily life activities several months after lung transplantation. Three months after lung transplantation, t h e t h i r d p a t i e n t m a d e i m p r o v e m e n t s i n t h e neurocognitive status and muscular strength at an inpatient rehabilitation centre.
Far less drastic is our proposed treatment of hyperinflammation in COVID-19-induced ARDS with lidocaine, an old drug that is readily available to hospitals all over the world at a low cost. In November 1948, Xylocaine was approved by the Food and Drug Administration (FDA) in the USA [411]. Lidocaine is used as a local anaesthetic [411], treatment of chronic neuropathic pain [412] but also for the prophylaxis or treatment of ventricular arrhythmia [328,329]. Recently, intravenous lidocaine has been administered as general anaesthetic replacing opioids in the perioperative settings [413]. Potentially, lidocaine, as a P2X7R antagonist, can abrogate hyperinflammation, can restore the capacity of the immune system to combat secondary co-infections and can improve the clinical condition in critically ill COVID-19 patients. Despite several in vitro [326,327,414,415], animal studies [319,[416][417][418][419][420] and patient cohorts [408,421] on the anti-inflammatory properties of lidocaine, completed clinical trials which deliver a proof of concept (i.e. a randomised controlled trial) have not yet been performed. We postulate that because the maximal tolerable plasma concentration of lidocaine is much lower than the required extracellular concentration to effectively inhibit P2X7Rs, intravenous systemic administration of lidocaine simply cannot not be used to effectively treat hyperinflammation. This is a plausible reason why 5 years after the discovery of lidocaine as a P2X7R inhibitor (published in 2015) [74] the drug is still not used as an antihyperinflammatory treatment in clinical practice.

Concluding remarks
As stated in the introduction, therapeutic measures that can immediately attenuate the course of SARS-CoV-2-related lung damage are promptly needed on a global scale. In contrast to the investigational P2X7R antagonists described above, continuous subdermal infusion of 2% lidocaine solution to primarily deposit lidocaine into the lymphatics is readily available and can be used in the daily practice immediately and, in principle, even outside the ICU and is very well affordable. Therefore, this therapy deserves to be investigated in larger placebo controlled randomised clinical studies with COVID-19 patients.

Future development
However, our experience with subdermal administration of lidocaine in the ICU made clear that this method may not be routinely suitable outside hospital settings. Needless to say that high complexity and high-cost treatments (requiring highly skilled nurses and infusion pump equipment) are inaccessible to low-income COVID-19 patients in developing countries. Also, as the severity and case fatality rate of COVID-19 increase with age [406], the case fatality rate in elderly patients in nursing homes is strikingly high, and many residents have poor access to medical care [422]. This encouraged us to explore alternative uncomplicated methods of lidocaine administration accessible to everyone, particularly elderly COVID-19 patients and COVID-19 patients in developing countries.
Recently, researchers stated in their article on targeting the P2X7R in COVID-19 that the P2X7R antagonists for human use are available only in oral form and that this might be an inefficient route of drug delivery [298]. We found a solution to this problem. Permeability of the skin and mucous membrane to water, drugs, etc. is said to be dependent on the site of the administration [423,424]. For example, the permeability constant of the floor of the mouth (sublingual mucosa), lateral border of the tongue and buccal mucosa for tritium-labelled water is 22, 17 and 13 times as high as human skin, respectively [423]. We argue that this also applies to lidocaine. As mentioned above, the endothelium of the mucosal capillaries of the mouth and the gastrointestinal tract belong to the structural type "fenestrated endothelium" allowing molecules to pass from the submucosal tissue into the capillaries [383]. Lidocaine hydrochloride is highly soluble in water (solubility of 680 mg/ml in water) [425] and therefore will mainly be absorbed by the submucosal capillary [426] and the inductive sites of MALT [398,399]. In contrast, the highly lipophilic lidocaine base (solubility of 4 mg/ml in water, 760 mg/ml in 95% ethanol and 790 mg/ml in chloroform) [425] is preferably absorbed by the local initial lymphatics in the submucosal tissue [426,427]. In addition, the lymphatic drainage of the floor of the mouth is extensive, involving many lymph nodes [428][429][430][431].
We estimate that with a sublingual administration of lipophilic lidocaine base (Fig. 1), we may reach the IC 50 of the P2X7Rs in the draining lymph nodes to control systemic hyperinflammation and avoid toxic lidocaine plasma levels ( Figs. 4 and 6). Obviously, such solution may also apply to other P2X7R antagonists. We stress that sublingual and buccal administration of lipophilic lidocaine is different from oral administration of lidocaine. Oral administration of lidocaine is aimed at the resorption of the drug in the gastrointestinal tract (Fig.6).
There are other methods of targeting the immune cells in the lymphatics, i.e. transdermal administration of lipophilic P2X7R antagonist with skin penetration enhancers (i.e. alpha-terpineol [432], ethanol [433] and lipid based nanoformulations [434]), intravenous administration of a P2X7R antagonist using nano-sized drug delivery systems [435], liposomes or polymer micelles [436] and oral administration of a P2X7R antagonist using delivery systems for intestinal lymphatic drug transport such as chylomicrons [437].
Acknowledgements We dedicate this report to the memory of Professor Geoffrey Burnstock, the discoverer of purinergic signalling. This report makes clear that purinergic signalling is indispensable to see through the complicated pathophysiology of hyperinflammation in COVID-19.
We would like to thank Coosje van der Pol, PhD, for her valuable and constructive editing suggestions during the writing of this research work.
Author contribution Conception, design and data handling: DH, AS, CvK, EPK and TK Substantial contributions to the conception and design of the manuscript: DH, AS, CvK and EPK Acquisition, analysis or interpretation of data for the manuscript: DH, AS, EPK and TK Drafting and revising the manuscript: DH, AS, CvK, PJvdS, EPK and TK Drafting the manuscript: DH, AS, CvK, PJvdS, EPK and TK Revising the manuscript critically for intellectual content: DH, AS, CvK, PJvdS, EPK and TK Final approval of the version to be published: DH, AS, CvK, PJvdS, EPK and TK Agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved: DH, AS, CvK, PJvdS, EPK and TK Data Availability The datasets generated during and/or analysed during the current study are not publicly available due to privacy reasons but are available from the corresponding author on reasonable request.
Code availability Not applicable.

Declarations
Ethics approval, consent to participate and consent for publication The Medical Ethical Committee of the Showa University, School of Medicine, Tokyo, approved the collection, analysis and publication of patients on mechanical ventilation admitted to the ICU (protocol number 3313).