Main

After several local epidemics caused by coronaviruses (CoVs) in recent years, namely severe acute respiratory syndrome CoV (SARS-CoV) in 2002 and Middle East respiratory syndrome CoV (MERS-CoV) in 2012 and 2015, a novel virus, SARS-CoV-2, emerged at the end of 2019 and spread rapidly throughout the world to cause the pandemic known as COVID-191,2. 22 million cases of COVID-19 and over 780,000 deaths had been reported worldwide by mid-August 2020, with a mortality rate that remains elusive due to uncertainty about the true number of infections. Mechanistically, infection by either SARS-CoV or SARS-CoV-2 involves the action of the virus spike protein (S), which engages angiotensin-converting enzyme 2 (ACE2) as the entry receptor on the surface of target host cells3, and the cooption of the cellular serine protease TMPRSS2, a member of the type II transmembrane serine protease family of proteins, which are implicated in both cancer and viral infections4,5, for S protein priming6,7. The ACE2 receptor is highly expressed in lung alveolar type 2 cells, but is also present on endothelial and smooth muscle cells in various organs, including the heart, liver, kidney and digestive tract8. Although a substantial fraction of SARS-CoV-2-positive individuals are asymptomatic or paucisymptomatic carriers (the latter experiencing nonspecific symptoms similar to those of the common cold and sometimes gastroenteritis), severe COVID-19, which develops mostly in individuals with comorbidities, leads to acute respiratory failure, often associated with a cytokine storm, a prothrombotic immunopathology and profound lymphopenia, culminating in multiple organ dysfunction and death9,10,11,12.

Patients with cancer were considered more susceptible to SARS-CoV-2 infection than individuals without cancer not only because of age, given that cancer incidence is strongly linked to advancing age, but also because of the high prevalence of cancer risk factors also associated with COVID-19—in particular, thoracic computed tomography (CT) scan abnormalities and smoking, along with cancer-associated metabolic disorders such as diabetes and hypertension—as well as the side effects of chemotherapy that might aggravate COVID-19, including arterial hypertension, cardiomyopathy, systemic immunosuppression13 and accelerated cellular senescence14,15. Several reports have addressed the prevalence of patients with a clinical history of cancer in French, Chinese and Italian populations tested for SARS-CoV-2 infection16,17,18,19. One study of 84,246 consecutive individuals tested for SARS-CoV-2 from the Veneto region of Italy found that 5.7% had previously been diagnosed with cancer and, among those positive for SARS-CoV-2, 7.8% had a cancer diagnosis. Hence, the authors concluded that prevalence of cancer was not associated with risk of infection16. In contrast, two earlier studies reported an increased incidence of COVID-19 in patients diagnosed with cancer in China17,18. In the first report, patients with tumors were older and had a history of smoking or dyspnea or severe baseline CT scan manifestations, with 28% having been diagnosed with lung carcinoma17. In the second study, of 1,524 patients with cancer from a tertiary cancer center in China, the relative prevalence of COVID-19 was twice as high as in the general population18. Differences between the Italy-based studies versus China-based studies in the demographic profiles of their patient populations, such as a larger proportion of older males participating in the latter, may account for these apparent discrepancies.

The main risk factors of severe COVID-19 in the general population are gender (male/female sex ratio, 1.65:1), advanced age (median age >60), obesity and diseases such as congestive heart failure, coronary heart disease, diabetes, hypertension, hyperlipidemia and cancer9. Race and ethnicity are also associated with COVID-19 risk, with Black and Hispanic people being disproportionately affected compared with white people9. In contrast to the incidence, COVID-19 severity was found to increase when associated with cancer across studies and geographical sites, including France, China, the USA and Italy16,20,21. For instance, the study from the Veneto region of Italy reported a higher percentage of COVID-19-related hospitalizations (56% versus 34%) and deaths (14% versus 4%) among patients with a history of cancer than among those without such a history16.

Cancer type, staging and specific therapies are additional risk factors for severe COVID-19 in this patient population. Patients with hematological, lung or breast cancer are more vulnerable than those with other cancers. In the Veneto study, breast and hematological cancers were associated with a higher risk of both hospitalization and death16. Lung cancer was associated with a fourfold risk of death due to SARS-CoV-2 infection16. A study reporting the clinical outcome of COVID-19 in 102 patients with lung cancer showed that the course of SARS-CoV-2 infection in these individuals was longer and more severe than that reported in the general US population21. About one-third of these patients experienced a relatively milder disease course and were treated as outpatients, two-thirds needed hospitalization and one-quarter died. Smoking status and chronic obstructive pulmonary disease (COPD) were the strongest determinants of severity21. 200 patients with COVID-19 and thoracic cancers from eight countries were identified and included in the TERAVOLT registry22. Univariate analyses identified age >65, current or former smoking, presence of any comorbidities and chemotherapy as being associated with increased risk of death. In line with the previous report21, however, multivariate analysis revealed only smoking history as associated with increased risk of death22. In hematological malignancies at older age, diagnosis of acute myeloid leukemia, indolent and aggressive non-Hodgkin lymphoma or plasma cell neoplasms, and severe or critical COVID-19 were associated with worse overall survival23. A retrospective single‐center analysis of 34 patients with hematological malignancies who developed COVID‐19 during follow‐up was conducted in Spain, reporting 11 deaths that could have been predicted by ECOG (Eastern Cooperative Oncology Group) status at disease onset24. Moreover, patients diagnosed with metastatic or stage IV carcinomas may be more susceptible to severe forms of COVID-19 than those with localized neoplasia25. Cancer treatments such as surgery, chemotherapy and immunotherapy have been reported to contribute to the severity of the COVID-19 among patients with cancer19,21,23,25,26,27. For example, a France-based study reported increased risk of deaths from COVID-19 in 178 patients with cancer with age >70, smoking status (current/former), ECOG score ≥2 at last follow-up visit, metastatic disease and use of cytotoxic chemotherapy in the past 3 months19. Separately, two studies reported that immunotherapy within the month before the first symptoms of COVID-19 developed was associated with increased severity and up to 30% death rates in 11 patients27 and 31 patients25, respectively. Age and receipt of immune-checkpoint inhibitor treatment remained significantly associated with COVID severity in a multivariate analysis of 563 patients with cancer25.

Conversely, pediatric patients with cancer seem to be relatively resistant to SARS-CoV-2 infection and severe COVID-19. Among 120 asymptomatic pediatric cancer patients, only 2.5% were positive for SARS-CoV-2, contrasting with a 14.7% rate in their asymptomatic caregivers18,19,28,29. This latter finding favors the possibility that age is a more important risk factor for SARS-CoV-2 infection than cancer.

In conclusion, as of August 2020, it is not clear whether cancer is an independent risk factor for severe COVID-19 or whether the observed cancer-associated risk depends on the peculiar demography and comorbidities of oncology patients30,31. However, given that the time elapsing between cancer diagnosis and SARS-CoV-2 infection was reported to affect the risk of death (in that the shorter the time elapsed, the higher the risk)16, we surmise that the role of therapies and co-medications may so far have been underestimated. Meta-analyses of well-described cancer cohorts, as well as prospective studies, are needed to fully disentangle the role played in COVID-19 severity by carcinogenesis and its clinical management, compared with age, sex and ECOG status.

Common risk factors for severe COVID-19 and cancer

In light of these intriguing data sets, it is important to elucidate the possible cause–effect relationship between severe COVID-19 and pre-existing pro-inflammatory and immunosuppressing conditions related to cancer (Fig. 1) and its treatments (Tables 1 and 2). In this section, we discuss the common risk factors between severe COVID-19 and cancer.

Fig. 1: Intertwined relationships between cancer and its comorbidities or therapies in relation to COVID-19 susceptibility.
figure 1

Aging, obesity, metabolic syndrome and exposure to carcinogens are predisposing factors for cancer. Aging, obesity and metabolic syndrome also represent comorbidities that influence susceptibility to and severity of SARS-CoV-2 infection. In patients with cancer, metastatic dissemination and poor ECOG performance status also favor COVID severity. Many genotoxic chemotherapies administered before SARS-CoV-2 viral infection ultimately enhance its severity, likely by inducing immunosuppression and cellular senescence in normal tissues, which in turn maintain local and systemic inflammation, but also through therapy-related adverse events that may include cardiovascular effects, asthenia and propensity to nosocomial infections. Immunosenescence and inflammaging, which are also promoted by aging and obesity, result in declining functions of the innate and adaptive immune systems, exacerbating overt inflammation and cancer dissemination and also increasing vulnerability to SARS-CoV-2 infection and risk of severe COVID-19.

Table 1 Drugs (by class) with anticancer effects or used for best supportive cancer care repurposed to become antivirals
Table 2 Anticancer drugs that potentially increase vulnerability to COVID-19

Aging, immunosenescence and inflammaging

Aging increases the incidences of both cancer and SARS-CoV-2 infection32, with potential key commonalities relating to immunosenescence and inflammaging (Fig. 1). Immunosenescence defines a status of declining immune system function associated with, or causing, quantitatively insufficient or qualitatively maladaptive responses to vaccination, infection and neoplasia, as well as an increased incidence of debilitating autoimmune diseases in the elderly population33,34. For example, levels of C-reactive protein are positively associated with senescent CD8+ T cells, plasmablasts and granulocytes in elderly people35. Notably, in patients with COVID-19, lower T cell counts are associated with clinical markers of inflammation, such as ferritin, D dimers and C-reactive protein, whereas high amounts of plasmablasts are associated with disease severity36. A recent immunophenotyping study of young and old individuals diagnosed or not diagnosed with COVID-19 confirmed that COVID-19 promotes age-induced immune cell polarization and gene expression related to inflammation and cellular senescence, and conversely, aging-associated dysregulated immune responses may at least partially account for vulnerability to COVID-19 in the elderly37. Moreover, the generation of naive T cells through thymopoiesis and their priming with novel antigens (such as tumor-specific neoantigens) or infectious agents (such as SARS-CoV-2) are compromised with aging38,39,40. This makes older people more vulnerable to both cancer and viral infections and less able to develop adaptive immune responses during SARS-CoV-2 infection or specific vaccination41,42, unless T cell cross-reactivities against seasonal coronaviruses are also shown to be similarly protective and endowed with reduced risk of antibody-dependent amplified breakthrough infection, compared with SARS-CoV-2-specific immune responses in the elderly43,44. This contrasts with the fact that higher IgG and IgM responses to the SARS-CoV-2 S and N proteins have been observed in elderly patients during the early phase of COVID-19, notwithstanding their higher viral loads than those of younger patients45.

A defect in dendritic cell (DC) fitness with aging has also been reported46. Peripheral germinal center follicular helper T cells are diminished in older people after seasonal influenza vaccination. This post-immunization impairment in the differentiation of follicular helper T cells was recapitulated in 2-year-old mice compared with younger mice, was linked to impaired T cell priming by conventional DCs (cDC2s) and could be restored by topical application of a Toll-like receptor 7 (TLR7) agonist46. Notably, the age-dependent increase in susceptibility to coronaviruses is also associated with an impaired ability of lung DCs to migrate to mediastinal lymph nodes and prime SARS-CoV-specific CD8+ T cells47,48,49. Reduced numbers and impaired functions of DCs and T cells have been reported at the acute phase of severe COVID-19 in 17 patients50. Carcinogenesis is also associated with defective antigen presentation and DC functions, including DC loss and migration defects51,52, which paves the way toward decreased antiviral T cell responses.

In addition to sex and genetics, age is a major factor influencing interindividual differences in transcriptional responses to bacterial, fungal and viral challenges in human peripheral blood mononuclear cells53,54. In the absence of stimulation, the expression of 85% of all genes in mononuclear cells is directly affected by age. After viral stimulation, reduced innate immune responses (for example, type I interferons (IFNs)) are observed in samples from individuals more than 30 years of age. Age (and sex) are also important determinants of humoral immunity, with older individuals (and women) showing higher rates of seropositivity for most antigens53. The reduced ability of aged B cells to increase their metabolism, characterized by a strong reduction in oxidative phosphorylation after activation, contributes to the weakened antibody responses of the elderly to vaccination in general55. A pivotal study analyzing immune responses to the Pandemrix H1N1 influenza vaccine concluded that age was a fundamental component of interindividual variations in the early response (by day 1) to the vaccine, dominated by type I and II IFN fingerprints. Only late (day 7) parameters, such as the rise in transitional plasmablasts usually found in responders, were not influenced by age56.

Separate from such defects in innate and cognate immune responses, ‘inflammaging’ refers to a cytokine dysregulation associated with the age-dependent remodeling of the immune system, as well as to an inability to fine-tune systemic inflammation. Whereas acute, localized inflammation is required for tissue repair responses, systemic and chronic inflammation is harmful. Several common molecular pathways are associated with both aging and low-grade inflammation. For instance, changes in redox equilibrium, defects in the clearance of senescent cells, accumulation of cells with the senescence-associated secretory phenotype (SASP) and reduced autophagy are hallmarks of aging that activate the inflammasome platform, a key orchestrator of cellular inflammatory responses57. Interleukin 6 (IL-6), which has been referred to as the ‘gerontologist’s cytokine’58, is normally present at low levels in the blood, but is increased with aging or frailty (sarcopenia and muscle loss)59 and correlates with mortality60,61. IL-6 is involved in the pathogenesis of many chronic diseases, including cancer62,63. The IL-6–JAK–STAT3 pathway is hyperactivated in many types of cancer, driving the proliferation, survival and invasiveness of tumor cells and suppressing the antitumor immune response. Thus, strategies targeting this pathway have already received US Food and Drug Administration (FDA) approval to treat inflammatory conditions or myeloproliferative neoplasms and to manage certain adverse effects of chimeric antigen receptor–expressing T cells (CAR T cells)64. Given that IL-6 is therapeutically targeted by tocilizumab in the context of COVID-19 to reduce morbidity and mortality related to COVID-19 cytokine release syndrome45, it is conceivable—although it remains to be demonstrated—that inflammaging favors the development of severe SARS-CoV-2 infection. In light of the critical pathophysiological impact of IL-6 and the severity of COVID-19 in individuals with hematological malignancies, prospective controlled studies testing IL-6 receptor (IL-6R) blockade or JAK or STAT3 inhibitors are warranted in this particular subset of patients.

Metabolic syndrome, cancer and COVID-19

Several meta-analyses have revealed an association between type 2 diabetes (T2D) and cancer, with the strongest relationship found for liver and pancreatic cancer, followed by endometrial cancer65. Similarly, morbidly obese individuals (body mass index ≥40 kg/m2) with T2D are more likely to become infected by SARS-CoV-2 and are at a higher risk of complications and death from COVID-1966. Interestingly, individuals with T2D were also at increased risk for SARS and MERS67. Relevant to this, insulin is a key hormonal enhancer of tumor metabolism and growth in obesity-associated insulin resistance68, and treatment of T2D during COVID-19 is being implemented to mitigate disease severity69.

Although human and mouse data analyses revealed that individuals with T2D have reduced ACE2 expression70, patients diagnosed with T2D also have elevated circulating levels of furin, a cellular protease that facilitates viral entry by cleaving the S1 and S2 domain of the SARS-CoV-2 spike protein (S)71. T2D inhibits neutrophil chemotaxis, phagocytosis and intracellular killing of microbes, resulting in impairments in adaptive immunity characterized by an initial delay in the activation of type 1 helper T cell (TH1 cell)–mediated immunity and a late hyperinflammatory response often observed in patients with diabetes72. This could explain the observed links between T2D and increased risk of adverse outcomes for COVID-19 and cancer, both of which depend on protective TH1/cytotoxic T cell type 1 (TH1/Tc1) immune responses.

In addition to facilitating virus entry, the metabolic syndrome may compromise the integrity of the intestinal barrier, a location of SARS-CoV-2 replication. In mice, hyperglycemia increases the permeability of this barrier through glucose transporter 2–dependent transcriptional reprogramming of intestinal epithelial cells and disruption of tight and adherens junctions73. Similarly, in humans, systemic influx of intestinal microbiome products correlates with failing glycemic control73. Given that both SARS-CoV and SARS-CoV-2 can be recovered from feces, infect intestinal epithelial cells and cause diarrhea74,75, hyperglycemia-induced dysfunction of the intestinal barrier might facilitate bacterial translocation, thus favoring systemic inflammation and immunosuppression. Interestingly, in a humanized mouse model of MERS-CoV infection on a high-fat diet, the course of infection was more severe and prolonged in male mice and was characterized by IL-17-producing helper T cell (TH17 cell) responses67, which are known to be pro-angiogenic and immunosuppressive in the course of cancer progression76. Notably, obese individuals contract more bacterial, viral and fungal infections than do lean counterparts, and respond relatively poorly to vaccination against influenza, hepatitis B, tetanus and rabies77. Obesity, alone or together with metabolic syndrome, induces defects in B cells similar to those associated with aging, contributing to systemic and B cell–intrinsic inflammation as well as to a surge in autoantibodies77.

Immunosuppression, lymphopenia, neutrophilia and interferon deficiency

Through their participation in immunosurveillance, lymphocytes control the incidence, progression and therapeutic response of cancers78. CD4+ and CD8+ T lymphocytes recognize tumor cells expressing immunodominant epitopes presented by major histocompatibility complex class II and class I, respectively. CD4+ lymphopenia, a hallmark of immunosuppressive viral infection, occurs in ~20% of patients with advanced pancreatic cancer, melanoma, non-Hodgkin’s lymphoma, breast cancer, sarcomas or hepatocellular carcinoma but is rare (~2%) in patients with localized disease79,80,81,82. In fact, cancer-associated lymphopenia, mostly affecting CD4+ T cell counts, has been reported to increase the risk of comorbidities (febrile neutropenia), resistance to a range of therapies and mortality across many cancer types79. Lymphopenia often accompanies cancer diagnosis, treatment or progression and is a side effect of chemotherapy and steroids. Radiotherapy also negatively impacts circulating lymphocyte counts83. An increased number of circulating neutrophils is often combined with decreased lymphocyte counts, resulting in a marked elevation of the neutrophil-to-lymphocyte ratio84. A high neutrophil-to-lymphocyte ratio is a poor prognostic marker and predicts short cancer-specific progression-free survival after blockade of programmed cell death protein 1 (PD-1)84, as well as severe COVID-1985. Beyond lymphopenia, a reduction in T cell receptor diversity and a functional impairment of other lymphoid and myeloid immune cells (such as natural killer (NK) cells, monocytes, DCs, and memory CD4+ and CD8+ T cells) have been detected in patients with localized primary tumors such as breast cancer, colon carcinoma and hepatocellular carcinoma79,86,87,88.

Given the critical role of T effector lymphocytes in eliminating virus-infected cells, an attenuated and functionally compromised T cell pool may pave the way toward the higher incidence and severity of COVID-19 in patients with cancer. Indeed, the outcome of COVID-19 may be determined by a ‘race’ between the cellular immune system that mounts a response to eliminate virus-infected cells and the immunosuppressive action of the pathogen89 (Fig. 2). Individuals who present with detectable memory B and T cell responses against seasonal coronaviruses may be able to mobilize a pool of effector T cells and mount neutralizing antibody responses that may, at least in part, prevent the thrombotic microangiopathy associated with SARS-CoV-2 viral endothelial infection90,91,92. Moreover, MERS-CoV and SARS-CoV-2 trigger apoptosis and necroptosis of T cells and reduce lymphopoietin IL-7 as an immunosuppressive strategy93.

Fig. 2: A race between viral replication and effective innate and cognate immune responses in cancer.
figure 2

Cancer, at an advanced or metastatic stage, may compromise the delicate equilibrium between viral replication and appropriate innate (for example, type I and III IFN) and cognate immune responses (for example, memory TH1 responses and antibody-secreting cell cross-reactivity with other beta coronaviruses) in the lung alveolar epithelium and mediastinal lymph nodes. The SARS-CoV-2 virus infects endothelial cells of the lung alveolar capillaries, inducing microthrombi and severe endothelial injury. The failure of the immune system to control early viral replication and to prevent endothelial injury may lead to a marked release of chemokines, cytokines and/or alarmins and a viral sepsis initiated or maintained by pulmonary or medullary hematopoiesis. Gr, granulocyte; HCoV, human coronavirus; IFN, interferon; IL, interleukin; ISG, interferon-stimulating gene; ORF, open reading frame; RGD, Arg‐Gly‐Asp.

If people with cancer start this ‘race’ with a handicap due to pre-existing T cell defects, this makes them particularly susceptible to COVID-19-associated severe pneumonia or systemic organ failure93. Conversely, it is unclear whether cancer-associated immunosuppressive cells—such as regulatory T cells, TH17 cells, myeloid-derived suppressor cells and ‘exhausted’ PD-1 ligand–positive T cells—might mitigate lung inflammation caused by SARS-CoV-279,94,95,96,97. Longitudinal high-dimensional immunomonitoring will help delineate favorable and deleterious cancer-associated immune factors for each stage of COVID-19 infection.

Type I and II IFN responses are intertwined and essential for long-term protective anticancer and antiviral immune responses98,99,100. Defective type I IFN responses by conventional or plasmacytoid DCs during natural immunosurveillance, or by tumor cells after chemotherapy or radiotherapy, are associated with tumor progression101,102,103,104. Intestinal dysbiosis, as well as tumor-intrinsic genetic defects, account for aberrant IFN-α/β receptor (IFNAR) signaling, preventing full efficacy of immunotherapy in patients with tumors101. RNA viruses engage pattern-recognition receptors (e.g., TLR3, TLR7 and TLR8, RIG-I, MDA5) on antigen-presenting cells, culminating in the induction of type I IFNs and an endoplasmic reticulum stress response that shuts down viral protein translation while igniting innate effector cells, such as NK cells or DCs105,106. Elegant studies have revealed the beneficial role of type I IFN signaling at early steps of infection and the protective impact of type II IFNs released by airway memory CD4+ T cells, or intranasal instillations of recombinant IFNγ, against SARS-CoV infection in mice107. Separately, treatment with the TLR3 agonist poly(I:C) protected mice against highly pathogenic coronavirus species, including group 2c (MERS-like) coronaviruses108,109. Similarly, TLR signaling through the TRIF adaptor protein protected mice from lethal SARS-CoV disease110. However, in stark contrast to other beta coronaviruses, SARS-CoV-2 fails to induce—or perhaps subverts—a type I IFN response in immune cells111. Thus, in response to SARS-CoV-2, the host does not increase the production of type I and type III IFNs but instead produces high levels of inflammatory chemokines and cytokines, particularly IL-6, thus stimulating emergency hematopoiesis and attracting granulocytes and monocytes to lung lesions111,112,113 (Fig. 2). As a failing immune response enables sustained viral replication, a positive feedback loop becomes established, eventually allowing the virus to prevail. These findings have been substantiated by in vitro studies in which normal lung parenchyma from five healthy human donors was infected with SARS-CoV versus SARS-CoV-2. In contrast to SARS-CoV, SARS-CoV-2 was highly replicative and poorly immunogenic, failing to trigger a type I IFN response and TH1 chemokine release114.

However, the timing of activation of the type I IFN pathway is key for antiviral and antitumor immune responses. For example, in separate mouse-based studies, delayed IFNβ treatment failed to effectively inhibit virus replication of SARS-CoV-2115 and MERS-CoV116, increased infiltration and activation of monocytes, macrophages and neutrophils in the lungs115,116 and enhanced the expression of pro-inflammatory cytokines, resulting in fatal pneumonia during SARS-CoV infection116,117 and in tumor resistance to PD-1 blockade115,116,117,118.

Altogether, these results imply that cancer and COVID-19 may be concomitantly aggravated by comorbidities such as aging, metabolic disorders, and innate and cognate immunosuppression (Fig. 1). These comorbidities may also compromise the efficacy of immune-based anticancer and antiviral therapies.

Cancer therapeutics with possible pro-COVID-19 effects

Tumor control achieved by oncological treatments is counterbalanced by the cardiovascular toxicities of cytotoxic agents, which often cause premature discontinuation of an effective therapy or undermine overall survival. An association between cancer and cardiovascular diseases, as well as a direct relationship between hypertension and cancer incidence and mortality, have been documented119. Thus, arterial hypertension (AHT) is both the most common comorbidity of cancer and a frequent adverse effect of antineoplastic therapies120. Pre-existing AHT is known to increase the risk of other cardiac adverse events due to oncologic treatments, in particular heart failure121. Many antineoplastic treatments (of which a non-exhaustive list is provided in Tables 1 and 2), particularly small molecules or antibodies targeting the growth factor VEGF or its receptor VEGF-R2 and tyrosine kinase inhibitors, cause AHT, compromising the long-term outcome of chemotherapy120. All these side effects theoretically complicate the prognosis of COVID-1925. However, and notwithstanding the high serum VEGF levels found in patients suffering from severe SARS-CoV-2, clinical trials in China are currently assessing the effect of targeting VEGF with bevacizumab in COVID-19 (NCT04275414) in preventing immunothrombosis.

Cytotoxic drugs administered at ablative or non-myeloablative dosages stimulate bone marrow progenitors and the exodus of both immature and mature cells, including granulocytes, monocytes and platelets. As previously reported, the resulting leukocytosis and production of the cytokine G-CSF could facilitate the differentiation of myeloid-derived suppressor cells and modulate neutrophil-to-lymphocyte ratios, paving the way to tumor progression, metastasis and poor clinical outcome122.

Genotoxic chemotherapies induce cellular senescence in normal tissues, where they promote local and systemic inflammation that causes or exacerbates the debilitating effects of chemotherapy15. In mice, ablating senescent cells reduces many side effects of cytotoxic agents, including cancer recurrence, cardiac dysfunction and myelosuppression15. Moreover, the risk of chemotherapy-induced asthenia is higher in people with increased expression of a senescence marker in T cells before chemotherapy15. Accordingly, after and during chemotherapy, patients with cancer are particularly susceptible to severe COVID-1919.

Radiation-induced pulmonary fibrosis (RIPF) is a common complication of thoracic radiotherapy for lung and breast cancer, observed in 16–28% of patients123. RIPF leads to irreversible destruction of lung architecture and disruption of gas exchange. Because the pathophysiology of RIPF features epithelial cell dysfunction and senescence, pro-inflammatory cytokine release and dysfunction of innate and adaptive immunity, it is not surprising that patients with lung cancer may have an elevated susceptibility to severe COVID-1921,124. Nevertheless, with appropriate dosing and timing, radiotherapy may be beneficial against acute respiratory distress syndrome, with single-fraction radiation or short courses of radiation being recommended125,126.

An association of checkpoint inhibitor–based immunotherapy with the aggravation of COVID-19, including increased hospitalization and severe respiratory conditions, was first reported in 31 patients25. This negative prognostic link was independent of age, cancer type and other comorbid conditions or coadministered medications such as steroids25. In this case, immune-checkpoint inhibitors may have exacerbated immune-related pneumonitis or T cell cytokine release, as previously discussed127. Given that many therapeutic actions currently used in oncology may increase the risk of severe SARS-CoV-2 infection, current guidelines related to cancer care during the COVID-19 crisis advise the postponement of all non-mandatory cancer therapies128.

In recent years we have accumulated an unprecedented understanding of the molecular pathways and immune-tolerance mechanisms governing the incidence and severity of human neoplasia, leading to a large swath of targeted anticancer therapies and immunotherapies. Despite their specificity, however, small-molecule inhibitors and antibody-based therapies induce both on- and off-target effects—the latter including immune-related pneumonitis and diabetes, among other conditions—that could increase the susceptibility of patients with cancer to COVID-19 (Fig. 3, Tables 1 and 2).

Fig. 3: Side effects of cancer therapeutics that represent aggravating comorbidities of COVID-19.
figure 3

Major side effects triggered by the main classes of compounds in the oncological armamentarium, including conventional therapies (such as cytotoxic chemotherapy, hormone therapy and radiotherapy), targeted therapies (such as TKI and mTOR inhibitors) and immunotherapies (such as immune-checkpoint inhibitors and CAR T cells), that can exacerbate COVID-19. On- or off-target unwanted effects of the drugs listed in each rectangle are indicated in uppercase letters. See also Tables 1 and 2. BiTE, bispecific T cell engagers; EGFR, epidermal growth factor receptor; HER2, human epidermal growth factor receptor 2; LH-RH, luteinizing hormone–releasing hormone; mTOR, mammalian target of rapamycin; TKI, tyrosine kinase inhibitor.

Cancer therapeutics with potential anti-COVID-19 effects

The quest for safe agents capable of inhibiting SARS-CoV-2 infection and replication has been intense over the past several months, spurring screening campaigns aimed at drug repurposing129. We have identified a number of antiviral drugs with potential antineoplastic properties and, reciprocally, anticancer agents with potential antiviral effects (Tables 1 and 2). In this section we discuss some of these therapeutic agents.

Interferon-based therapies

Recombinant IFNγ and IFNα2b have been widely used against cancer, alone or combined with other cancer treatment modalities130. Because SARS-CoV-2 compromises the paradigmatic type 1 interferon antiviral response111,113,114, IFN administration—either locally by vapor inhalation or systemically, and alone or in combination with ribavirin106,131, lopinavir/ritonavir, remdesivir132 or hydroxychloroquine—has arisen as a promising treatment approach against COVID-19131 (Tables 1 and 2). IFNγ as an inhaled aerosol has been found to be effective against tuberculosis in a controlled clinical trial and to be safe and ameliorate pulmonary function in a phase 1 clinical trial of patients with idiopathic pulmonary fibrosis133. Thus, inhaled IFNγ has been proposed as a treatment against COPD, tuberculosis and idiopathic pulmonary fibrosis133. It will be interesting to follow SARS-CoV-2 viral loads, virulence and the expression of type I IFN–related genes in the blood and lungs of patients with cancer over the course of their cytotoxic therapies, to estimate the suitability of IFN supplementation in treating COVID-19. Alternatively, another way to increase the systemic circulating levels of type I IFNs would be to use pattern-recognition-receptor agonists. For instance, subcutaneous administrations of a TLR9 agonist increases circulating levels of type I IFNs and decreases hepatitis C viral loads107.

Immune-checkpoint blockers

On theoretical grounds, immunotherapy with immune-checkpoint blockers might simultaneously boost cytotoxic T lymphocyte (CTL) immune responses134 against virus-infected and neoplastic cells. At this stage, however, there is little information on the role of T cells during the severe phase of COVID-19. Indeed, T cells also might cause lung immunopathology, meaning that their therapeutic reactivation would exacerbate the disease135. In contrast, it is also possible that CTL exhaustion accompanying respiratory virus–induced pneumonitis might render such cells unable to eliminate inflammatory myeloid cells, in which case immune-checkpoint blockers might have anti-inflammatory effects. Indeed, virus-induced secondary hemophagocytic lymphohistiofibrosis (HLH) has been reported to result from inappropriate and prolonged macrophages activation10,136. Supporting the view that T cell exhaustion might be causal in the HLH syndrome, a recent clinical study showed a remarkable efficacy of anti-PD-1 antibodies in treating HLH induced by Epstein–Barr virus in four out of six patients treated137. However, although a retrospective analysis of 62 patients with lung cancer adjusted for gender and smoking status showed the non-inferiority of anti-PD-1 treatment in cases of concomitant COVID-19 and non-small-cell lung cancer138, the contrary effect has been reported recently across various malignancies, as discussed above25.

IL-6–JAK–STAT3 blockade

As outlined above, overt inflammation accompanied by elevated plasma IL-6 levels is associated with severe COVID-1910 and disseminated malignancies64. Despite the absence of a causative link between the manifestation of a cytokine release syndrome and COVID-19 severity, investigators have attempted to interfere with this inflammatory cascade. Tocilizumab, an anti–human IL-6R monoclonal antibody, was recently approved to treat severe SARS-CoV-2-related pulmonary complications by the National Health Commission of the People’s Republic of China. Observational studies conducted in China in patients with severe COVID-19 receiving tocilizumab showed an improvement of clinical and radiological outcome139 and attenuation of the hyperactivated inflammatory immune responses140 as well as restoration of a robust T cell-associated adaptive immunity141 in these patients142,143. Meta-analyses of studies of IL-6R blockade to treat rheumatoid arthritis revealed a drastic reduction of neutrophil counts in chronically treated patients144. Such neutrophil depletion might alleviate lung inflammation in severe COVID-19, as emergency hematopoiesis and bone marrow exodus of immature neutrophils and pre-neutrophils may predict the switch between moderate and severe cases112. However, given that the first randomized trial assessing the efficacy of IL-6R blockade in overt pneumonia (the COVACTA phase 3 trial, NCT04320615) did not reach its primary endpoint145, stratification of patients may be necessary to reach clinical significance.

Androgen-deprivation therapy

Androgen-receptor signaling enhances TMPRSS2 expression in non-prostatic tissues, including lungs146, potentially contributing to the increased vulnerability of male individuals to SARS-CoV-2. Androgen-deprivation therapy (ADT) decreases the levels of TMPRSS2 and may be used to counter the severity of SARS-CoV-2 infection in male patients, potentially in combination with other inhibitors of viral entry or replication. Accordingly, in one study, patients with prostate cancer receiving ADT had a significantly lower risk of severe COVID-19 than patients who did not receive ADT147. An even greater difference was found when comparing patients with prostate cancer receiving ADT to patients with other types of malignancies147. Hence, ADT with luteinizing hormone–releasing hormone (LH-RH) modulators or AR inhibitors (Tables 1 and 2) may be worth evaluating as interceptive medications against COVID-19 in prophylactic therapy148, as well as therapeutic measures for high-risk male patients147.

Other small molecules

Based on its capacity to inhibit autophagy, hydroxychloroquine (HCQ) is being evaluated in clinical trials to treat autophagy-dependent cancers149. In spite of initial reports that HCQ might reduce the duration of SARS-CoV-2 infection and the severity of COVID-19150, later reports failed to confirm significant effects of the drug151,152, suggesting that the simultaneous treatment of cancer and COVID-19 with HCQ is not advisable. Nonetheless, there may be other small molecules with dual anticancer and anti-SARS-CoV-2 properties. For example, inhibitors of translation elongation factors may have such a dual activity because both cancer cells and coronaviruses rely on cap-dependent mRNA translation. A recent study that screened a large library of compounds for their capacity to inhibit the replication of SARS-CoV-2 identified inhibitors of elongation factor 1A (eEF1A) and eukaryotic initiation factor 4A (eIF4A)153 that are currently being evaluated to treat multiple myeloma and KRAS-mutated cancers, respectively153,154(Tables 1 and 2). A separate report screened a library of FDA-approved drugs for inhibitors of coronavirus replication and identified Abelson (ABL) kinase inhibitors, including the anticancer drug imatinib mesylate, as inhibitors of both SARS-CoV and MERS-CoV in vitro155, and a later study confirmed that imatinib mesylate inhibits SARS-CoV-2 replication in cultured cells as well156. Although no in vivo data are yet available, the use of imatinib mesylate to treat COVID-19 is an active area of research (Tables 1 and 2). It should be noted that imatinib mesylate has the capacity to stimulate T and NK lymphocyte–mediated anticancer responses, suggesting that it has immunostimulatory effects157. Whether it can also stimulate immune responses against SARS-CoV-2 remains to be investigated. Other small molecules (including JAK1/2 inhibitors158,159 and anti-CD26 antibody160,161,162; see Tables 1 and 2) currently in use for hematological malignancies to alleviate graft-versus-host disease, cytokine storms or overt inflammatory responses may also be of interest against severe COVID-19, as discussed elsewhere159,163. Many additional antiviral and tumoricidal therapeutics or interceptive strategies (such as the anti-inflammatory use of low-dose thoracic irradiation164,165,166,167,168,169,170,171 or prophylactic vitamin D172,173,174,175,176,177,178,179,180) listed in Tables 1 and 2 are currently under investigation.

Conclusions and future directions

The current COVID-19 crisis has a lesser impact on healthy and fit children and adolescents, while claiming its deadly toll among all other segments of the population: the sick, the unfit and the elderly, including patients with cancer. Malignant disease predisposes to severe COVID-19 for multiple reasons, primarily because (i) patients with cancer fall into general at-risk categories because of their average advanced age, predisposing factors such as obesity and smoking, and comorbidities such as T2D and hypertension; (ii) cancer intrinsically has negative effects on patients’ general health status; and (iii) antineoplastic therapies such as surgery, chemotherapy and radiotherapy may debilitate the immune system and cause immunosenescence and inflammaging. However, whether cancer per se is an independent risk factor for severe COVID-19 remains to be elucidated. It should also be kept in mind that during the COVID-19 outbreak, morbidity and mortality of patients with cancer may have been substantially affected not only by the viral disease itself but also by the extreme pressure exerted by COVID-19 on the healthcare system, which led to the postponement of cancer treatments and the allocation of scarce resources, such as intensive care beds and ventilators, to patients with better prognoses181,182,183. During the present COVID-19 pandemic, oncology departments are frequently confronted with the challenge of treating patients with both cancer and COVID-19, raising a strong argument for exploring therapeutic strategies that could simultaneously improve both diseases. Several drugs that have direct inhibitory effects on SARS-CoV-2 replication in vitro184,185,186,187,188 (and that still require further characterization in clinical trials; Table 3) are also known for their potential anticancer effects, supporting the idea that such agents, including imatinib mesylate and inhibitors of cap-dependent translation, might have a dual therapeutic activity against cancer and COVID-19. Given the uncertainties about the benefits of PD-1 and/or PD-L1 or IL-6R blockade25,145, other possibilities are being investigated, such as passive transfer of neutralizing anti-SARS-CoV-2 antibodies for frail patients at a moderate to severe stage of COVID-19189. Finally, active vaccination will be an option for patients at high risk of developing severe COVID-19 but still capable of mounting protective antiviral T cell responses41. However, candidate vaccines will have to undergo large-scale phase 3 clinical trials to assess their effectiveness and safety, so the earliest regulatory approvals and roll-outs are not expected before late 2020 or early 2021. All these possibilities await urgent investigation to allow clinical oncologists to navigate between cancer and COVID-19 in full compliance with the Hippocratic Oath: primum non nocere—first, do no harm.

Table 3 Antiviral drugs proposed against SARS-CoV-2 infection displaying antitumor effects