Ageing, the major risk factor for severe COVID-19, results in the accumulation of a number of defects in the innate and adaptive immune systems. For example, the number of T and B lymphocytes, macrophages, granulocytes, and lymphatic follicles are significantly decreased in the elderly. Ageing macrophages and granulocytes adopt an enhanced inflammatory state secreting pro-inflammatory cytokines, and showing impaired phagocytosis, migration, and clearance, thereby compromising the ability of these cells to clear infections and damage [6]. Thus, the aged cells of the innate immune system generate a proinflammatory state (so called ‘inflammaging’) associated with reduced clearance of virus and virus-infected cells. This may also be associated with the accumulation of senescent cells in other tissues where they secrete a range of mediators known as the senescence-associated secretory phenotype (or SASP [7]). These mediators include many proinflammatory cytokines which may also contribute to inflammaging.

Despite the low-grade inflammation seen in the aged, the development of excessive numbers of terminally differentiated T cells (particularly CD28⁻ CD27⁻ CD45RA⁺ CD8⁺ T cells), with a paucity of naïve T cells has been observed (a condition known as immunosenescence) [8]. The relative lack of naïve T cells compromises the ability of the aged immune system to mount a defence against novel pathogens such as SARS-CoV-2. Increased numbers of senescent T cells are also associated with autoimmune disease, chronic viral infection (e.g., CMV or EBV) [9], as well as the reduced response to vaccines seen in the aged [10,11]. These cells show increased NK receptor (e.g., KLRG-1), granzyme B, and perforin expression and have lost antigen-specific cell killing but retain a strong nonspecific killing potential [12].

Intriguingly, senescent T cells are also associated with some of the underlying medical conditions which increase the risk of severe COVID-19 disease. In rodent models, senescent T cells can induce diabetes and obesity, while their clearance moderates the disease [13,14]. It is therefore tempting to speculate that one reason for the susceptibility of the aged, and those with chronic diseases such as rheumatoid arthritis and diabetes, is due to the higher prevalence of senescent cells in these patients, including those of the immune system [15]. Similarly, T cell immunosenescence is closely related to the development of cardiovascular disease [16], another chronic disease state associated with susceptibility to severe COVID-19 disease [17]. However the role, if any, of senescent T cells in the susceptibility of the aged to SARS-CoV-2 has yet to be proven.

In the normal course of a viral infection, the innate immune system reacts first through the interferon (IFN) system, driven partly by recognition by cells of virus-related Pathogen Associated Molecular Patterns (PAMPs) which trigger a range of Toll Like (and other) Receptors (TLRs), [18], resulting in the expression of many IFN-stimulated genes (ISGs) and the stimulation of the adaptive immune response. The SARS-CoV-2 virus encodes a number of proteins, including ORF3b and ORF6, which inhibit this IFN response. Presumably this inhibition of the initial IFN response allows the virus to replicate in the early stages of the disease. Consistent with this, patients that have defects in the IFN system (e.g., anti-IFN autoantibodies) tend to experience severe disease and delayed viral clearance [19,20,21,22]. IFN response genetic variants have also been identified which are associated with severe disease [23], including the antiviral restriction enzyme activators OAS1, 2 and 3, the antiviral receptor TLR7 [24] and the IFN receptor IFNAR2. Further analysis suggested that high expression of the JAK enzyme Tyk2 (and the chemokine receptor CCR2) or low expression of IFNAR2 is associated with life critical disease [25]. Since the Type 1 IFN response is low in severe SARS-CoV-2 infections, but with elevated chemokines, it has been suggested that defective anti-viral responses result in excessive SARS-CoV-2 replication, elevated chemokines with consequent innate immune activation, and a resultant cytokine-mediated hyperinflammation [26]. In addition, the relative lack of naïve T cells in the elderly (and those with underlying health conditions such as autoimmune disease) may further reduce the ability of the adaptive immune system to respond to this new infection. This may then be amplified by the pro-inflammatory innate immune cells, and the senescent T cells which, while not expressing CD28 and CD27, have acquired NK cell properties and secrete yet more cytokines [27].

Although the world has largely relied on vaccines for protection of our populations against SARS-CoV-2, there is still a need for therapies capable of reducing mortality in hospitalised patients, especially with the rapid evolution of new SARS-CoV-2 variants including the highly infectious Omicron variants. In addition, the resistance of a significant proportion of the population to being vaccinated, the escape of viruses from immune control, whether that be vaccine or infection induced, and the likelihood that other viruses with similar pathogenic mechanisms will be manifest in the future [28] means that there will be a need for COVID-19 therapeutics for some time to come.

Early in the pandemic there was an emphasis on testing already approved drugs in COVID-19 since this would be the fastest way of finding treatments which could be given to patients. The repurposing of anti-viral drugs such as hydroxychloroquine and Kaletra (lopinavir–ritonavir) was not a success with most agents not showing reproducible efficacy in clinical trials of hospitalised patients [29]. The anti-viral remdesivir was, however, approved in the USA for use as a result of the Adaptive Covid Treatment Trial-1 (ACTT-1), in which it reduced mortality and the duration of hospital stay [30]. It is now not recommended by the WHO for treating severely and critically ill patients but is for those at high risk of severe disease [31]. It has been difficult to prove that antiviral drugs are effective in COVID-19 due to the relatively mild symptomology during the incubation period, when the virus is most susceptible to such drugs. However, in many cases the viral load is already decreasing when patients come to hospital experiencing severe or critical disease, so immune modulators are required. This perhaps explains why the only drugs shown to reduce mortality in randomised clinical trials of hospitalised patients are anti-inflammatories [32]. The recent success of the antiviral Paxlovid was made possible by focusing on prospective patients at higher risk of hospitalisation due to underlying health conditions [33] and treating them in the early phase of the disease.

The first repurposed drugs to be approved for COVID-19 were the broadly acting immune-suppressive corticosteroids, which had modest beneficial effects on mortality and have been widely used throughout the world since mid-2020. The greatest effect of low dose dexamethasone on COVID-19 mortality was seen in the RECOVERY trial in patients receiving oxygen support or invasive mechanical ventilation. In patients with mild disease an increase in mortality was seen, perhaps partly as a consequence of suppression of the endogenous anti-viral immune response. This indicates that the timing of steroid (and other immune suppressant) administration may be critical [34]. There followed a very large number of clinical trials assessing various immune modulators on hospitalised patients with varying degrees of success. Amongst the agents tested were inhibitors of the JAK enzymes which mediate cytokine receptor signalling.

The JAKs are tyrosine kinases activated by over 50 different cytokines via their receptors including IFNα, IFNγ, TNFα, IL-1β, IL-2, IL-6, and IL12. Activated JAKs phosphorylate intracellular surfaces of the receptors, promoting Signal Transducer and Activator of Transcription (STAT) binding and subsequent activation again through phosphorylation. The activated STAT proteins then regulate gene transcription in the nucleus. The four JAK enzymes (JAK1, JAK2, JAK3, and TYK2) activate seven STAT family members, the outcomes largely depending on the specific combination of JAKs and STATs activated by any given receptor [34]. In addition, a number of growth factors and related molecules use the JAK/STAT pathways including leptin, erythropoietin, thrombopoietin, and granulocyte–macrophage colony-stimulating factor (GM-CSF) [35].

JAK inhibitors including baricitinib, tofacitinib, peficitinib, upadacitinib, and filgotinib have been approved for the treatment of autoimmune diseases such as rheumatoid arthritis where they act as disease modifiers [35]. These approved inhibitors are all competitive with ATP showing varying amounts of selectivity between the JAKs based on in vitro assays. However, as discussed by Tanaka et al. [35], the in vitro selectivity does not always translate to predictable differences in cell-based cytokine signalling. JAK inhibitors such as ruxolitinib have also been developed for the treatment of myeloproliferative diseases which can be driven by mutations in JAK2 [36], and in which thrombosis is a major cause of mortality [37].

The entry of viruses into cells is mediated by a number of routes including Clathrin-Mediated Endocytosis (CME), caveolin-mediated endocytosis, macropinocytosis, and some other poorly described non clathrin- or caveolin-mediated mechanisms [38]. By far the best understood of these pathways is CME, which mediates the internalisation of many ligands with their receptors including the EGFR, transferrin, and low-density lipoprotein receptors [39] as well as membrane proteins such as the SARS-CoV-2 receptor ACE2. This pathway involves the assembly of the clathrin polyhedral lattice beneath the membrane and the association of tetrameric adaptor proteins, which bind to the sorting signals expressed on the cytoplasmic side of the virus receptor proteins. Many viruses have been shown to use the clathrin pathway, although different mechanisms may be used by a given virus in different cells, and multiple mechanisms may be exploited within a given cell type [39,40]. The dependency of SARS-CoV-2 on TMPRSS2 cleavage of the spike protein has suggested that fusion of the virus membrane with the plasma membrane was a major route of infection, bypassing the requirement for endocytosis. However, most coronaviruses are endocytosed prior to infection and SARS-CoV-2 has now been shown to be internalised through CME after binding to ACE2 [41,42]. The ACE2 receptor expresses a PDZ binding motif on the cytoplasmic side of the plasma membrane which serves as an endocytic motif for the clathrin adaptor subunit AP2. The AP2 adaptor is required for optimal formation of the clathrin lattice, although there are a number of other clathrin adaptors (e.g., NUMB and EPS15) which can mediate CME depending on the internalisation signals expressed by different membrane proteins. ACE2 is not the sole SARS-CoV-2 receptor, others such as some integrins [43], neuropilin 1 [44], and mGluR2 [45] have been implicated as receptors for SARS-CoV-2 and also contain PDZ binding domains.

Perhaps the best-characterised regulators of CME are the Numb-Associated Kinases (NAKs), AP2-associated protein kinase 1 (AAK1) and cyclin G-associated kinase (GAK). These kinases stimulate cargo recruitment [46], as well as the assembly and internalisation of the clathrin/AP2/cargo complex [47]. AAK1 and GAK are also involved in the further intracellular trafficking of the endocytic vesicle including removal of the clathrin coat [48,49,50]. As far as viruses are concerned, AAK1 and GAK may also regulate the infectivity of viruses that depend on AP2, such as the flaviviruses (e.g., Dengue and West Nile viruses), HCV, HIV, and Ebola virus [51,52,53,54].