Sustained Cytotoxic Response of Peripheral Blood Mononuclear Cells from Unvaccinated Individuals Admitted to the ICU Due to Critical COVID-19 Is Essential to Avoid a Fatal Outcome

COVID-19, caused by SARS-CoV-2, ranges from asymptomatic infection to ARDS and fatal disease. Known risk factors (age, comorbidities) and immune dysregulation, including lymphopenia and functional exhaustion of cytotoxic lymphocytes, contribute to severity. Effective cytotoxic antiviral responses by NK, NKT, and CD8+ T cells are crucial for viral control. Interleukin-15 (IL-15) supports homeostasis and activation of cytotoxic lymphocytes and has shown promise in immunotherapy. This study asks whether the magnitude and sustainability of PBMC cytotoxic activity in unvaccinated individuals with critical COVID-19 admitted to the ICU is associated with survival, and whether ex vivo IL-15 can enhance cytotoxic responses, thereby informing potential immunomodulatory strategies.

Prior reports indicate that COVID-19 severity and mortality are linked to older age, male sex, and comorbidities (e.g., hypertension, diabetes). Lymphopenia is common in severe disease and correlates with prolonged hospitalization and mortality. Cytotoxic lymphocytes (CD8+ T cells, NK, NKT) can show functional exhaustion (e.g., elevated PD-1, TIGIT) in severe COVID-19, with reduced degranulation marker CD107a expression. IL-15 is essential for cytotoxic lymphocyte development, survival, and activation, and IL-15 or superagonists (e.g., N-803) expand NK and CD8+ T cells in clinical trials for cancer and HIV. However, monotherapy may be insufficient in exhausted states and responsiveness can be impaired if IL-15 receptor signaling is compromised. These findings motivated evaluating PBMC cytotoxicity over time in critical COVID-19 and testing IL-15 responsiveness ex vivo.

Design: Observational, longitudinal study of 23 unvaccinated adults with critical COVID-19 admitted to the ICU (Hospital Universitario Ramón y Cajal, Madrid, Spain) from October 2020 to April 2021, spanning Spain’s second and third pandemic waves. Peripheral blood was collected upon ICU admission (baseline, t = 0) and every 2 weeks up to 13 weeks or until outcome. Participants were grouped post hoc as Exitus (fatal outcome, n = 13) or Survival (discharged, n = 10). Ethics: Informed consent obtained; protocol CEI PI 32_2020-v2 approved by the Ethics Committees of Instituto de Salud Carlos III and the hospital, in accordance with the Helsinki Declaration. PBMC isolation: 5 mL EDTA blood processed by Ficoll-Hypaque density gradient; PBMCs cryopreserved. Viability assessed by microscopy and flow cytometry. Target cells: Vero E6 cells (ECACC 85020206) cultured in DMEM + 10% FCS, L-glutamine, penicillin/streptomycin. Pseudotyped SARS-CoV-2 infection and cytotoxicity assay: Vero E6 monolayers were infected for 48 h with single-cycle pseudotyped virus pNL4-3Δenv_SARS-CoV-2-SΔ19(G614)_Ren (100 ng p24 Gag/well), encoding SARS-CoV-2 spike (D614G) and Renilla luciferase. After washing, infected Vero E6 were co-cultured for 1 h at 1:1 ratio with participant PBMCs pretreated or not with IL-15 (13 ng/mL for 48 h; dose selected based on NK proliferation and memory CD8+ expansion data). Cytotoxicity was quantified by caspase-3/7 activity in detached Vero E6 using luminescence (Caspase-Glo 3/7). PBMCs from co-cultures were analyzed by flow cytometry to define effector populations and activation/exhaustion markers. Flow cytometry: Surface markers for phenotype/activation included CD3-PE, CD8-APC-H7, CD56-FITC, CD107a-PE-Cy7. Exhaustion markers included PD-1 (PD1-BV650) and TIGIT-AlexaFluor700; CD3-BV510, CD4-PE, CD8-APC-H7, CD56-FITC. Acquisition on BD LSRFortessa X-20; analysis with FlowJo v10.0.7. Key populations: NK (CD3−CD56+), NKT (CD3+CD56+), CD8+ T (CD3+CD8+). CD107a assessed as degranulation marker. RNAemia: Plasma RNA extracted (QIAamp MinElute, QIAcube). RT-qPCR targeting E and N genes per WHO guidance; positive if Cq < 45 cycles. Clinical/biochemical data: Demographics, comorbidities, treatments, signs/symptoms, ventilation needs, length of stay, and baseline biochemistry were collected. Statistics: GraphPad Prism 8.4.3. Between-group comparisons: Mann-Whitney U-test or one-way ANOVA with Tukey’s multiple comparisons as appropriate; within-group: Wilcoxon matched-pairs; Fisher’s exact test for categorical clinical characteristics. Two-tailed p < 0.05 significant. Cohen’s d computed (0.2 small, 0.5 medium, 0.8 large, ≥1.2 very large).

- Cohort: 23 unvaccinated ICU patients with critical COVID-19; Exitus (n=13) vs Survival (n=10). No significant between-group differences in age (median 65 vs 63 years), sex (73.9% men overall), comorbidities (65.2% ≥1; dyslipidemia 52.2%, hypertension 39.1%, diabetes 26.1%), treatments (corticosteroids/antibiotics/anticoagulants 82.6% each), or clinical presentation (pneumonia 100%, cough/dyspnea/fever 87%). Invasive ventilation: Exitus 92.3%, Survival 100%. - Length of stay: Hospital median 59.0 days (IQR 36.0–100.0) Exitus vs 73.0 (44.5–90.0) Survival; ICU median 49.0 (25.5–82.0) Exitus vs 44.0 (25.0–66.5) Survival; no significant differences. - Lymphocytes: Exitus showed lymphopenia (<1000 cells/mm3) at admission only; by Week 2 all above threshold. Survival group had higher lymphocyte counts at baseline (1.54-fold; p=0.0318; Cohen’s d=1.00) and Week 4 (1.71-fold; p=0.0030; Cohen’s d=3.17). Survival counts rose to Week 6 then declined toward Week 10; Exitus remained lower overall. - PBMC cytotoxicity (against SARS-CoV-2–pseudotyped Vero E6): Exitus remained low and unchanged throughout; significantly lower than Survival at admission (2.69-fold lower; p=0.0234; Cohen’s d=1.12) and Week 4 (5.58-fold lower; p=0.0290; Cohen’s d=1.28). In Survival, cytotoxicity rose steadily to Week 6 then declined. - IL-15 responsiveness: Exitus PBMC cytotoxicity did not significantly improve with IL-15 at any time point. Survival PBMC cytotoxicity increased with IL-15 by a mean 2.30-fold across samples; at Week 4, IL-15–induced cytotoxicity increased 6.18-fold (p=0.0303; Cohen’s d=1.05). - NK cells (CD3−CD56+): No significant differences in frequencies between groups; CD107a degranulation in NK cells increased over time in Survival but did not reach statistical significance; IL-15 had no significant effect on NK CD107a. - NKT cells (CD3+CD56+): From Week 4, both groups increased; at Week 6 Exitus had 4.17-fold higher levels (p=0.0571; Cohen’s d=3.15; not significant). CD107a expression unchanged between groups and with IL-15. - CD8+ T cells (CD3+CD8+): Higher at baseline in Exitus (2.07-fold; p=0.0409; Cohen’s d=0.94) and remained elevated vs Survival. CD107a on CD8+ T cells decreased at Week 4 in Exitus (1.53-fold; p=0.0313) and stayed low; IL-15 produced a non-significant slight benefit. In Survival, CD107a dipped at Week 4 then increased thereafter (not significant). - Exhaustion markers: At Week 4, Exitus showed higher PD-1 on CD4+ T cells (1.72-fold; p=0.0095; Cohen’s d=6.87) and CD8+ T cells (2.92-fold; p=0.0317; Cohen’s d=2.07) vs Survival. NKT TIGIT increased over time in Exitus; at Week 2, 1.76-fold higher than Survival (p=0.0357; Cohen’s d=0.33). No significant NK differences for PD-1/TIGIT. Exhaustion analyzed up to Week 6 due to sample availability. - RNAemia: No plasma SARS-CoV-2 RNA detected at baseline in any participant (Cq < 45 threshold not reached). Overall: Sustained PBMC cytotoxicity and IL-15 responsiveness were associated with survival, while persistent low cytotoxicity and increased exhaustion characterized fatal outcomes.

Despite similar demographics, comorbidities, treatments, and lengths of stay, survivors exhibited higher lymphocyte counts, increasing PBMC cytotoxicity through Week 6, and robust ex vivo responsiveness to IL-15. In contrast, fatal cases showed persistently low PBMC cytotoxicity that did not improve with IL-15 and exhibited elevated exhaustion markers (PD-1 on CD4+/CD8+ T cells, TIGIT on NKT cells), consistent with dysfunctional cytotoxic responses. Although CD8+ T cell counts were higher in the Exitus group, their functional degranulation (CD107a) was reduced, underscoring that functional competence rather than abundance determined effective antiviral activity. NK and NKT compartments did not show significant quantitative differences overall, though trends toward increased activation in survivors were observed. These findings support the hypothesis that sustained, functional cytotoxic responses in peripheral blood are essential for clearing infection and avoiding fatal outcomes in critical COVID-19. They also suggest that IL-15–based immunomodulation could enhance cytotoxicity in responsive patients, whereas profound immune exhaustion and potential IL-15 receptor pathway impairment may necessitate combination strategies (e.g., checkpoint inhibition, ADCC-enhancing antibodies) to restore function.

The study demonstrates that impaired antiviral cytotoxic capacity of PBMCs is a critical determinant of fatal outcomes in unvaccinated ICU patients with critical COVID-19, independent of common clinical risk factors. Survivors showed increasing cytotoxicity and responsiveness to IL-15, while Exitus patients exhibited persistent dysfunction and elevated exhaustion markers. Monitoring exhaustion markers and ex vivo responsiveness to IL-15 may help identify patients at risk of poor outcomes and guide immunomodulatory interventions. Future research should include larger cohorts and evaluate optimized IL-15 dosing regimens combined with other immunomodulatory agents (e.g., immune checkpoint inhibitors or monoclonal antibodies) to overcome exhaustion and enhance antiviral cytotoxicity.

- Limited sample size (n=23) restricts generalizability; nonetheless, several comparisons achieved statistical significance with large/very large effect sizes. - Exhaustion marker analyses were limited to baseline through Week 6 due to sample availability beyond this time. - Ex vivo IL-15 was tested at a single dose and duration; in vivo responses and optimal regimens were not assessed. - All participants were unvaccinated; results may differ in vaccinated or variant-specific contexts.