Study of the SARS-CoV-2-specific immune T-cell responses in COVID-19-positive cancer patients.

BACKGROUND: Cancer patients are considered highly vulnerable to the COVID-19 pandemic. However, delaying cancer-specific therapies could have a deleterious effect on survival. The potential suppressive effects of chemotherapies or cancer-related microenvironment raised the question on how cancer patients' immune system responds to SARS-CoV-2 virus.
METHODS: We have started a prospective monocentric trial entitled COV-CREM (NCT04365322) in April 2020. The primary objective of the trial was to assess specific immune response's intensity and diversity to SARS-CoV-2 in infected patients.
RESULTS: In this study, we showed that cancer patients (28 solid tumours, 11 haematological malignancies) exposed to SARS-CoV-2 produced a high rate of specific antibodies, as observed in patients without a cancer history (n = 29). However, our results highlight a lack in the generation of T-cell responses against CoV-N, M and S proteins from the SARS-CoV-2 virus, suggesting that cancer patients failed to mount a protective T-cell immunity. Nevertheless, SARS-CoV-2 infection did not impair established immune memory since specific responses against common viruses were not hampered in cancer patients.
CONCLUSION: Given the severity and the unknown evolution of the ongoing COVID-19 pandemic, it is of fundamental importance to integrate cancer patients in vaccination programs.

Introduction

Today's fight against the SARS-CoV-2 virus generates a scientific emulation that brings together researchers from all horizons. Older age and medical comorbidities are the main risk factors for severe outcomes and higher mortality from COVID-19 infection [1]. This is why cancer patients are considered highly vulnerable [2,3]. Comorbidities and active cancer are the major factors for high risk COVID-19-related severe events [[4], [5], [6]]. In a SARS-CoV context, T-cells are critical in virus clearance by eliminating virus-infected cells both in mice models and human [[7], [8], [9]]. SARS-CoV viruses are composed of structural proteins namely N (nucleocapsid), M (membrane), E (envelope) and S (spike). Recently, Grifoni et al. reported the presence of SARS-CoV-2 specific T-cells, using peptides derived from these structural proteins, in convalescent patients who presented a mild illness form [9]. The supposed vulnerability of cancer patients in the COVID-19 pandemic raised the questions of how the immune system of these patients responds against the SARS-CoV-2. However, delaying treatment is not recommended during the COVID-19 pandemic. Cancer patients must receive antitumour treatment under a vigorous SARS-CoV-2 screening. Growing evidence sustains the rationale for the development of a vaccine against SARS-CoV-2 [7,8,10]. Thus, it is critical to understand how the immune system of the supposedly vulnerable cancer patients would respond against the SARS-CoV-2 to anticipate the impact of a possible vaccine on these patients.

Materials and methods

Patients

We have started a prospective monocentric trial entitled COV-CREM (NCT04365322) in April 2020. All patients were enrolled after the signature of informed consent in accordance with the French regulation. Three cohorts of patients were enrolled: i) patients with mild illness COVID-19 infection without cancer, whose clinical parameters have been described elsewhere [11], ii) patients with solid tumours and iii) patients with haematological malignancies. The analysed population was diagnosed with COVID-19 between March 10 and May 20. Patients with cancers were included if they presented a COVID-19 infection diagnosed based on RT-PCR performed on a nasopharyngeal swab or positive serology. The primary trial objective was to assess specific immune response's intensity and diversity to SARS-CoV-2 in infected patients. In this study, we presented the results of 28 patients with solid tumours and 11 patients with haematological malignancies. Results are comparatively analysed against immune analyses performed in healthy donor samples and patients treated for mild illness COVID-19 infection without a cancer history [11]. A detailed description of the methods can be found in the supplementary methods.

Immune responses analysis

Peripheral blood mononuclear cells (PBMCs) from patients and healthy donors were isolated by density centrifugation on Ficoll gradient (Eurobio). We assessed spontaneous T-cell responses against SARS-CoV-2 by using IFNγ ELISpot assay. Humoral responses were assessed by using ELISA assay to detect IgG antibodies to SARS-CoV-2. A detailed description of the laboratory analysis methods can be found in the supplementary methods.

Statistical analysis

Statistical analyses were performed using GraphPad Prism 6 software (San Diego, CA). The level of significance was set at p < 0.05 for all tests (∗p ≤ 0.05, ∗∗p ≤ 0.01, ∗∗∗p ≤ 0.001 and ∗∗∗∗p ≤ 0.0001). Variables were expressed as a median and interquartile range (IQR) and tested with the Mann–Whitney U test. Variables were expressed as frequencies and tested with Fisher's exact test.

Results

Cancer patients’ characteristics are reported in Table 1 and Table 2 . All enrolled patients presented a mild illness form of COVID-19. The viral load detected in nasal swabs was available for 16 patients. For the majority of them (88%), the viral load was high (mean 19.5 ct (15.29–36.76)). We notice that four patients were hospitalised for a preventive approach. The median age was 69. When diagnosed with COVID-19, 82% had a fever. At the time of infection, 61.5% of patients had a metastatic or advanced stage of cancer, of which 79.2% had ongoing anticancer therapies (excluding hormonotherapy) and thirteen (68.4%) displayed lymphopenia (lymphocyte count under 1G/L).

Table 1

Cancer patient's characteristics.

	With T cell responses		Without T cell response
	Solid tumours (n = 14)	Haematological malignancies (n = 4)	Solid tumours (n = 14)	Haematological malignancies (n = 7)
Patient characteristics n (%)
Age - median (year)	70	73	69	59
[Range]	[51–75]	[54–80]	[48–84]	[23–73]
<65	4 (28.6)	1 (25.0)	6 (42.9)	4 (57.1)
65–74	8 (57.1)	2 (50.0)	5 (35.7)	3 (42.9)
>75	2 (14.3)	1 (25.0)	3 (21.4)	0 (0)
Male	7 (50.0)	4 (100.0)	5 (35.7)	4 (57.1)
History of cancer n (%)
Metastatic or late stage	9 (64.3)	1 (25.0)	12 (85.7)	2 (28.6)
Active anticancer treatment during 4 weeks before COVID infection	10 (71%)	3 (75.0)	12 (85.7)	1 (50)
Chemotherapy	6 (42.9)	0 (0.0)	6 (42.9)	0 (0.0)
Chemotherapy + immunotherapy	0 (0.0)	1 (25.0)	1 (7.1)	0 (0.0)
Immunotherapy	2 (14.3)	1 (25.0)	1 (7.1)	0 (0.0)
Targeted therapy	1 (7.1)	1 (25.0)	4 (28.6)	1 (50)
Hormonotherapy	1 (7.1)	0 (0.0)	0 (0.0)	0 (0.0)
COVID-19 n (%)
Symptomatic	12 (85.7%)	4 (100.0)	14 (100.0)	7 (100.0)
Diagnostic of COVID-19
PCR	8 (57.1)	4 (100.0)	11 (78.6)	7 (100.0)
PCR+ >1 month	0 (0.0)	0 (0.0)	3 (21.4)	1 (14.3)
Serology	6 (42.9)	0 (0.0)	3 (21.4)	0 (0.0)
Time between COVID-19 infection and blood sample - median [range] in days	70.5 [48–109]	76.5 [49–92]	62.5 [41–85]	81.0 [48–119]
Immunological data
PNN median (mm3) [Range]	2 205 [1370–5250]	2 935 [1410–5980]	3 075 [902–9000]	5 550 [900–35 000]
Missing	4	0	0	1
LDH median (UI/I) [Range]	417 [207–464]	199 [177–235]	239 [61–341]	444 [206–750]
Missing	8	1	3	3
Lymphopenia at the moment of COVID-19 infection (G/L) n (%)	3 (42.8)	2 (50.0)	9 (64.3)	2 (28.6)
Median [range]	1 340 [600––2060]	1 505 [690–2320]	840 [300–1920]	1 300 [580–1 900]
Missing	7	0	0	1
CD3 (%) [range]	54.1 [26.4–89.4]	65 [14.7–94.5]	61.1 [31.7–87.5]	54.4 [5.2–91.1]
CD4 (%) [range]	51.3 [16.2–91.5]	31.1 [4.3–51.4]	52.2 [20.8–83.7]	46.7 [8.3–66.2]
CD8 (%) [range]	38.5 [7.1–69]	58.9 [39.4–92.3]	35.9 [13.8–63.3]	48.4 [24.7–90.3]
CD4/CD8 (%) [range]	2.4 [0.3–12.8]	0.6 [0.04–1.3]	2.0 [0.4–5.8]	1.3 [0.09–2.6]
CD19 (%) [range]	23.9 [1.5–58.3]	27.3 [1–87.8]	27.6 [2.9–75.1]	18.7 [0.1–59.1]
Missing	0	0	2	2

Table 2

Clinical characteristics of individual cancer patients.

Solid tumours

Patient number

Sex

Age

PS

Diabetes

BMI

HTA

Other comorbidities

Concurrent ongoing therapies

Cancer localisation

Metastatic disease Y/N

Active anticancer treatment during 4 weeks before COVID infection

Time between COVID-19 infection and blood sample in days

COVID-19 serology

At least one specific T-cell response (S, M or N)

CEF specificT-cell responses

03–01

M

72

1

NO

27.7

YES

NO

NO

Hepatocarcinoma

YES

Sorafenib

64,0

POS

NO

YES

03–02

F

53

0

NO

26.5

NO

NO

NO

Breast

YES

Trastuzumab Pertuzumab Hormonotherapy

72,0

POS

YES

YES

03–03

M

55

3

NO

23.8

NO

Arrhythmia

AmlodipineCordaronePerindopril

Pulmonary

NO

Durvalumab

85,0

POS

YES

YES

03–05

F

75

1

NO

26.5

YES

NO

NO

Breast

YES

Hormonotherapy

70,0

POS

YES

YES

03–06

M

62

3

NO

24.3

YES

Arrhythmia

SotalolApixaban

HNC

YES

NO

54,0

NEG

NO

YES

03–07

M

69

2

NO

19.8

YES

NO

NO

HNC

YES

Methotrexate

48,0

POS

YES

YES

03–10

F

69

0

NO

24.8

NO

Dyslipidemia

Pravastatine

Colorectal

YES

FOLFOX

63,0

POS

YES

YES

03–14

F

75

1

NO

21.5

NO

NO

NO

Breast

YES

Capecitabine Lapatinib

85,0

POS

NO

YES

03–15

M

73

1

NO

18

NO

NO

NO

Pulmonary

YES

Cisplatin Pemetrexed Prembrolizumab

58,0

POS

NO

NO

03–16

F

48

1

NO

24.4

NO

NO

NO

Breast

NO

Trastuzumab

61,0

POS

NO

NO

03–17

F

74

0

NO

18.6

NO

NO

NO

Glioblastoma

YES

NO

59,0

POS

NO

YES

03–18

F

62

0

NO

22.3

NO

NO

NO

Breast

YES

Tucatinib Capecitabine Trastuzumab

85,0

POS

NO

YES

03–19

M

67

1

NO

20.5

YES

NO

NO

Stomach

YES

FOLFIRI

67,0

POS

NO

YES

03–20

F

64

1

NO

21

NO

NO

NO

Ovarian

YES

Liposomal doxorubicin

60,0

POS

NO

NO

03–21

F

57

1

NO

40

YES

NO

NO

Ovarian

YES

Carboplatin Paclitaxel

83,0

POS

YES

YES

03–22

F

71

0

NO

22

NO

NO

NO

Breast

NO

NO

91,0

POS

YES

YES

03–23

M

72

2

NO

21

NO

NO

NO

Colorectal

YES

FOLFOX Bevacizumab

71,0

POS

YES

YES

03–24

M

70

0

NO

21

NO

NO

NO

Vesical

NO

NO

81,0

POS

YES

YES

03–25

M

74

1

NO

19.8

NO

NO

NO

HNC

NO

DCF

69,0

POS

YES

NO

03–26

M

84

1

NO

22.6

NO

NO

NO

Renal

YES

Nivolumab

65,0

POS

NO

NO

03–27

F

80

0

YES

26

NO

NO

NO

Hepatocarcinoma

YES

Sorafenib

56,0

POS

NO

YES

03–28

F

62

0

NO

19.3

NO

NO

NO

Breast

NO

Docetaxel

75,0

POS

NO

NO

03–29

F

70

0

NO

31

YES

NO

NO

Breast

YES

Palbociclib Hormonotherapy

85,0

POS

NO

NO

03–31

M

75

2

NO

20

NO

Ischemia Cardiopathy- Chronic Renal Failure

Sotalol Perindopril Metformin Lercanidipine

Pulmonary

YES

NO

70,0

NEG

YES

YES

03–34

F

64

1

NO

25

NO

NO

NO

Renal

YES

Nivolumab

101,0

NEG

YES

YES

03–40

M

74

1

YES

24.7

YES

NO

NO

Prostate

YES

Cabazitaxel

89

POS

YES

NO

03–42

F

51

2

NO

20

YES

NO

NO

Sarcoma

NO

NO

109

POS

YES

YES

03–44

F

58

1

NO

19

YES

NO

Hydrochlorothiazide

Pancreas

YES

Gemcitabine

41

POS

NO

YES

Haematological malignancies

Patient number

Sex

Age

PS

Diabetes

BMI

HTA

Other comorbidities

Concurrent ongoing therapies

Cancer localisation

Disease stage

Active anticancer treatment during 4 weeks before COVID infection

Time between COVID-19 infection and blood sample in months

COVID-19 serology

At least one T cell response (S, M or N)

CEF specific T-cell responses

03–12

M

73

1

NO

1.72

NO

NO

NO

DLBCL

Early

R–CHOP

49

NEG

YES

YES

03–13

M

23

1

NO

2.5

NO

NO

ValaciclovirTacrolimusPrednisone

ALL

Early

NO (allograft in October 2019)

48

NEG

NO

YES

03–32

M

80

1

NO

2

YES

NO

Sotalol

CLL

Early

R-Venetoclax

80

POS

YES

NO

03–35

F

48

2

YES

1.4

NO

Pulmonary transplantation

Tacrolimus

AML

Early

NO (Allograft in 2006)

48

NEG

NO

NO

03–36

M

54

1

NO

2

NO

NO

Tacrolimus

AML

Early

NO (Allograft in august 2019)

54

POS

NO

NO

03–37

M

54

2

NO

1.9

NO

NO

PrednisonePolyvalent Intravenous Immunoglobulin

AML

Early

NO (Allograft in 2007)

54

POS

YES

YES

03–38

M

72

1

NO

1.8

YES

Hypothyroidism

Levothyroxine

Multiple myeloma

Late

Isatuximab PomalidomideDexamethasone

72

POS

YES

YES

03–39

F

73

2

NO

1.7

NO

Stroke

Valaciclovir

AML

Late

Venetoclax

73

POS

NO

YES

03–41

M

67

1

NO

2.1

NO

Arrhythmia

Prednisone

Myelodysplastic syndrome

Late

NO

67

NEG

NO

YES

03–43

F

52

1

NO

2.2

NO

Hypothyroidism

Levothyroxine

CML

Early

Nilotinib

52

POS

NO

NO

03–46

M

66

1

NO

2.1

NO

NO

Valaciclovir Irbesartan

Bone marrow failure

Early

Cyclosporine

66

NEG

NO

YES

Abbreviation: ALL, Acute Lymphocytic Leukaemia; AML, Acute Myelogenous Leukaemia; CLL, Chronic Lymphocytic Leukaemia; CML, Chronic Myelogenous Leukaemia; DCF, Docetaxel Cisplatin Fluorouracil; DLBC, Diffuse Large B Cell Lymphoma; F, Female; HNC, Head and Neck Cancer; M, Male; PS, Performance Status.

Among patients with cancers, 14 out of 28 patients with solid tumours (50.0%) and 4 out of 11 patients with haematological malignancies (36.4%) exhibited T-cell responses against at least one of the SARS-CoV-2 proteins (CoV–S, CoV-M or CoV–N) (Fig. 1 A). In line with a potential cross-reactivity with previous coronavirus exposition, recognition of CoV–S, M and N proteins by T-cells was observed in respectively 33.3%, 33.3% and 14.3% of the healthy donors tested (Fig. 1B). Furthermore, SARS-CoV-2 elicited specific immune responses against CoV-M and CoV–N in 58.6 and 65.5% of the patients without a cancer history (p ≤ 0.05 and p ≤ 0.01 respectively) (Fig. 1B, middle and right panel). The T-cell immune responses’ intensity against SARS-CoV-2 proteins was investigated. No statistically significant differences were observed between the responses of patients without a cancer history and both cohorts of cancer patients. This provides evidence that immune responses during SARS-CoV-2 infection do not sustain antigen memory T-cells detectable in the PBMCs. There were no significant frequency differences in CoV–S T-cell responses between patients with solid tumour (42.8%) or haematological malignancies (27.3%), patients without a cancer history (37.9%) or healthy donors (33.3%) (Fig. 1B, left panel). Specific CoV-M and N T-cell response frequencies tend to be lower in patients with haematological malignancies compared to patients without a cancer history (27.3% versus 58.6% (p = 0.15) and 36.4% versus 65.5% (p = 0.15) for SARS-CoV-M and N respectively) (Fig. 1B middle and right panel). Specific CoV–N T-cell response frequencies were statistically lower in patients with solid tumours compared to patients without a cancer history (35.7% versus 58.6%, p = 0.0348). The one against CoV-M tends to be lower (39.3% versus 58.6%, p = 0.12) (Fig. 1B). Interestingly, two cancer patients without peripheral SARS-CoV-2 specific T-cells had prolonged virus RNA detection after symptoms resolution, while no prolonged RNA production was reported in patients with specific adaptive immunity. The SARS-CoV-2 specific T-cells absence was not correlated to clinical determinants. Similarly, lymphopenia and lymphocyte subset distribution did not account for the lack of SARS-CoV-2-specific immunity (Table 1).

Fig. 1

SARS-CoV-2 specific T-cell responses were increased in COVID-19 patients without a cancer history compared to COVID-19 patients with solid or haematological malignancies. PBMCs from 21 healthy donors and COVID-19 patients without a cancer history (n = 29), solid tumour (n = 28) or haematological malignancies were analysed for SARS-CoV-2 and antiviral-specific T-cell responses by IFNγ ELISpot assay. A. Frequency of patients with specific responses to at least one SARS-CoV-2 protein. B. Frequencies (%) of positive SARS-CoV-2 specific immune responses in healthy donors and COVID-19 patients (Mann–Whitney test, ∗p > 0.05). C. Frequency (%) of positive antiviral memory CD8 T-cell responses in healthy donors and COVID-19 patients. D. The median serology index (S/CO) in COVID-19 patients without a cancer history, solid tumour or haematological malignancies (Mann–Whitney test, ∗p > 0.05, ∗∗p > 0.01). E. Heatmap showing the positivity or the negativity of the serology index, T-cell immune responses to SARS-CoV-S, M and N proteins for each patient included in the study (online Morpheus software). Healthy donors' population is represented by light gray, COVID-19 patients without a cancer history by black, COVID-19 patients with solid tumour by blue and haematological malignancies by violet. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

The other immunological issue for cancer patients is the potential risk that SARS-CoV-2 virus-mediated infection dampens pre-existing memory T-cell repertoire. To investigate this issue, we concomitantly measured the reactivity against common viruses using IFNγ ELISpot assay. Median intensities and frequencies were similar across all patient groups (Fig. 1C). This demonstrated that the recent history of COVID-19 did not hamper viral memory T-cell pool in patients with or without cancers. It suggests that the lack of SARS-CoV-2 specific T-cells is related to a weak immune adaptive response initiation rather than a global T-cell dysfunction.

We then analysed the SARS-CoV-2 humoral responses. We compared the serology index between our patients’ cohorts (Fig. 1D). The median serology index in patients without a cancer history was 5.6 S/CO [IQR: 2.9–6.6] and it was statistically lower compared to that of patients with solid tumours 6.9 S/CO [IQR: 5.0–8.3] (p = 0.016). In contrast, patients with haematological malignancies had a median serology index equal to 3.25 S/CO [IQR 0.1–6.3] that was statistically lower when compared to both patients with solid tumours and without a cancer history (p = 0.0017 and p = 0.052). A heatmap was generated, using the online Morpheus software (https://software.broadinstitute.org/morpheus/), to correlate all immunological parameters investigated (Fig. 1E). These results unravel that SARS-CoV-2 infection is associated with a lack of adaptive T-cell immunity in cancer patients. In addition, a discrepancy was unravelled, in the context of cancer history, between SARS-CoV-2 seroconversion and the presence of memory T-cells.

Discussion

To date, few studies have analysed specific SARS-CoV-2 cellular and humoral responses in cancer patients. This study, based on patients with solid tumours treated mostly at a metastatic stage, showed a high rate of SARS-CoV-2 antibody production. The specific T-cell response frequency in cancer patients was similar to the one in healthy donors and significantly lower than the frequency in COVID-19 patients without a cancer history underlying the weakness of SARS-CoV-2 immunogenicity in cancer patients. The presence of SARS-CoV-2 specific T-cells in healthy donors might be explained by the sequence homology between structural proteins from various coronaviruses suggesting the existence of cross-reactive memory T-cells [12,13]. Indeed, high degrees of similarities between coronaviruses and SARS-CoV-2 concerning S, M and N structural proteins have been recently described [9,14].

This study is intended to be descriptive but not comparative. The small cohort of cancer patients studied is heterogeneous in terms of cancer type and anticancer treatment at the time of COVID-19 infection, which make it difficult to statistically explain confounding factors of non-response. However, both cohorts of patients were comparable in terms of COVID-19 severity and included patients with mild disease. The T-cell responses against common viruses (Cytomegalo-, Epstein–Barr and Flu-virus) were used as a control of memory T-cell responses. The level of T-cell responses against CEF was similar between the two cohorts excluding a possible abnormality of the peripheral memory T-cells compartment. Nevertheless, the effects of the previous line of anticancer treatment and lymphopenia at the time of the infection remain to be investigated in a larger cohort of patients.

The impairment of adaptive immunity due to immune-suppressing therapies might alter cancer patients’ outcome following COVID-19. Patients with haematological malignancies have usually long-lasting immunodeficiency because of the malignancy itself but also because of anticancer treatments such as rituximab or haematopoietic stem-cell transplantations [15]. The first report on COVID-19 in chronic lymphocytic leukaemia (CLL) patients receiving venetoclax showed a substantial adaptive immune deficiency that might lead to an impairment of T-cell responses and more COVID-19 severity in infected CLL patients [16]. The cases of severe COVID-19 have also been reported in patients treated with anti-CD20 targeted therapy supporting the impact of B-cell depletion thereby the incapacity to produce neutralising antibodies on the COVID-19 course [17,18]. Nevertheless, in a recent multicentric study including 177 patients with lymphoma, active treatment and the number of prior lines did not show significant differences in mortality due to the infection [19]. Furthermore, in a study including 423 cases of patients with solid tumours, the active anticancer treatment did not show a significant association with the COVID-19 severity [20]. The potential effects of immunosuppressive treatments on the severity of COVID-19 disease and outcome need future investigations in a larger cohort of patients.

The induction of anti-SARS-CoV-2 protective immunity by vaccine in cancer patients is an important medical need. However, the heterogeneity of SARS-CoV-2 specific-T-cell responses among cancer patients emphasises the interest to investigate anti-SARS-CoV-2 immune responses following vaccination in this population.

To summarise, this study raises three major concerns to be considered for further research programs: i) adaptive T-cell immunity targeting SARS-CoV-2 is weak in cancer patients compared to common viruses; ii) specific immunoglobulin monitoring is not sufficient to characterise anti-SARS-CoV-2 immunity and assays to monitor specific T-cell responses should be developed, and iii) SARS-CoV-2 infection does not alter the common virus's memory T-cell responses level in cancer patients. Even if those first data need to be confirmed in a larger cohort, the present results support the promotion of specific SARS-CoV-2 vaccination programs for cancer patients given the ongoing COVID-19 pandemic severity.

Authors’ contribution

L. Mansi, L. Spehner and M. Kroemer had full access to all data in the study and take responsibility for the integrity of the data and accuracy of data analysis. Study concept and design: M. Kroemer, K. Bouiller, L. Mansi, L. Spehner, E. Daguindeau, C. Borg and O. Adotévi. Acquisition, analysis or interpretation of data: all authors. Drafting of the manuscript: L. Mansi, M. Kroemer, L. Spehner and C. Borg. Statistical analysis: L. Mansi and L. Spehner. Study supervision: L. Mansi, C. Borg, O. Adotévi and M. Kroemer. We are grateful to the Department of Oncological Clinical Research (A. Martin, J. Meier, A. Sauget and D. Berthod) and the nurses team (D. Bourdenet, M. Pagnot and A. Terzibachian).

Fundings

No funding directly supported this work.

Conflict of interest statement

The authors declare no conflict of interest.