High levels of NK Cells in graft are associated with reduced febrile neutropenia after haploidentical peripheral blood stem cell transplantation

Abstract

Objective

To investigate the impact of natural killer (NK) cells in the graft on the outcome following haploidentical peripheral blood stem cell transplantation (haplo-PBSCT).

Methods

We retrospectively collected data from patients who had undergone haplo-PBSCT at our centre from January 2019 to November 2021. The percentage of NK cells in stem cell grafts was detected by flow cytometry. Based on the median (range) count of NK cells (1.8 [0.4–6.0] × 10⁸/kg), patients were separated into high and low NK cell count groups.

Results

Data from 96 patients were analysed. Patients were evenly distributed (48 in each group) into high and low NK cell count groups. There was no significant difference in neutrophil and platelet recovery between the two groups. However, the rates of febrile neutropenia, bacterial infection, and invasive fungal disease (IFD) were significantly reduced in the high compared with the low NK cell count group. There was no significant difference in rates of cytomegalovirus (CMV) and Epstein–Barr virus (EBV) infections between groups. There was no significant difference between groups in grades II and above acute graft versus host disease (GVHD), progression-free survival (PFS), or overall survival (OS).

Conclusion

A high number of NK cells in the graft may reduce febrile neutropenia, IFD and bacterial infection following haplo-PBSCT but it has no significant effect on aGVHD, PFS, or OS.

Introduction

Haploidentical peripheral blood stem cell transplant (haplo-PBSCT) is an effective treatment for haematologic malignancies. Studies have established that numbers of CD3⁺ and CD34⁺ cells in the graft have a predictive value on the outcomes of haplo-PBSCT.^1–5 However, opinions differ on the impact of natural killer (NK) cells in the graft tissue.^1–5 The purpose of this retrospective study was to investigate the impact of NK cells on the outcome following haplo-PBSCT in a cohort of patients who had undergone haplo-PBSCT for haematologic malignancies at our centre from January 2019 to November 2021.

Methods

Study population

We retrospectively reviewed data from consecutive patients who had undergone haplo-PBSCT for haematologic malignancies at the First Affiliated Hospital of Xi'an Jiaotong University (Xi'an, China) from January 2019 to November 2021. Definitions of acute myeloid leukaemia (AML), acute lymphoblastic leukaemia (ALL), chronic myeloid leukaemia (CML), myelodysplastic syndrome (MDS) and acute mixed leukaemia (ML) were made according to 2017 European LeukemiaNet (ELN) guidelines. ⁶ Patients with non-haematologic malignancies, such as aplastic anemia, were excluded.

All patients received modified busulfan and cyclophosphamide (BuCy2) conditioning regimen and a dose of 10 mg/kg of anti-human thymocyte immunoglobulin (ATG, Sanofi). Patients also received cyclosporin A, short-term methotrexate, and mycophenolate mofetil to prevent graft-versus-host disease (GVHD). Cyclosporin A was administered at 3 mg/kg/day from Day 9 to 6 months post-transplantation. Methotrexate was administered on Days 1, 3, 6, and 11 post-transplantation at 15, 10, 10, and 10 mg/m², respectively. Oral mycophenolate mofetil was administered at 1 g/day from Day −1 to Day +30 and tapered to 0.5 g/day from Day +31 to Day +45. In addition, norfloxacin and ganciclovir were administered from Days −9 to −1 to prevent bacterial and viral infections. Acyclovir was initiated from the day of neutrophil recovery to 6 months post-transplantation to prevent viral infections and posaconazole or voriconazole were initiated from Days-9 to 3 months post-transplantation to prevent fungal infections. Trimoxazole was administered from the beginning of the conditioning period to the day of Cyclosporine A was stopped to prevent Pneumocystis carinii.

The reporting of this study conforms to STROBE guidelines, ⁷ and it was approved by the Ethics Committee of the First Affiliated Hospital of Xi'an Jiaotong University. Written/verbal consent was not required because it was a retrospective study and patient data were anonymized prior to analysis.

Stem Cell Mobilization and Collection

Donors received twice-daily subcutaneous injections of 5 μg/kg granulocyte colony-stimulating factor (G-CSF). Hematopoietic stem cells were collected on Day 4 of mobilization if the total mononuclear cell (MNC) count was >6 × 10⁸/kg (patient weight), and the total CD34⁺ cell count was >4 × 10⁶/kg (patient weight). The CD34⁺ cells, CD3⁺CD4⁺ T cells, CD3⁺CD8⁺ T cells, NK cells, Treg cells, and B cells in the collection were detected by flow cytometry.

Infectious complications

Febrile neutropenia was defined as a neutrophil count <500/μl and fever (i.e., axillary temperature of >8°C for ≥1 hour). Serum levels of procalcitonin, β-(1-3) glucan, galactomannan (GM) antigen, cytomegalovirus (CMV), and Epstein-Barr virus (EBV) DNA were measured once weekly until the neutrophil count had recovered. All patients with febrile neutropenia underwent a microbiological examination of their blood for the presence of microorganisms. Serum GM antigen was considered positive if the GM index was ≥0.5. CMV or EB viremia were defined as a CMV and EBV DNA load detected by molecular assay of >500 copies/ml even if asymptomatic.

All patients with febrile neutropenia were given broad-spectrum antibiotics such as carbapenem (1g q8h) or piperacillin/tazobactam (4.5g q8h) which were discontinued if the patient had a normal temperature for seven consecutive days or until neutrophil recovery. If these broad-spectrum antibiotics were ineffective or the body temperature increased during treatment, IV antifungals such as voriconazole (200 mg q12h), caspofungin (50 mg qd), or amphotericin-B liposomes (3 mg/kg qd) were added until neutrophil recovery. Thereafter, oral antifungals were used. The definition of IFD was based on the European Organization for Research and Treatment of Cancer/Mycoses Study Group (EORTC/MSG) 2008 criteria. ⁸ Patients were classified as having proven, probable, possible, or no invasive pulmonary aspergillosis (IPA).

Neutrophil and Platelet Engraftment and GVHD

Neutrophil recovery was assessed as the first of three successive days with an absolute neutrophil count of ≥500/μl after the posttransplant nadir. Platelet recovery was assessed as the first of three consecutive days with a platelet count of ≥20,000/μl in the absence of platelet transfusion for seven consecutive days. Acute GVHD was diagnosed and graded according to published guidelines. ⁹

Survival analysis

Relapse was recorded as disease recurrence. Progression-free survival (PFS) was calculated from transplant day to the date the disease relapsed. Overall survival (OS) was calculated from disease onset to day of death or last follow-up visit.

Statistical Analysis

Statistical analyses were performed using IBM SPSS Statistics for Windows, Version 26.0. (Armonk, NY: IBM Corp). A P-value <0.05 was considered to indicate statistical significance. Continuous variables were expressed as mean ± SD or medians and ranges, and differences were analysed using Student's t-test and one-way ANOVA (data with normal distributions) and Wilcoxon rank sum test (data with skewed distributions). Categorical variables were compared using χ² test. Survival analysis was performed using Kaplan Meier curves.

Results

Patients’ Characteristics

In total, 96 patients were identified (Table 1). Most patients were male and over 18 years of age. Overall, there were 48 cases of AML, 30 of acute ALL, two of CML, 11 of MDS and 5 of acute ML. The median (range) NK cell count in the grafts was 1.8 (0.4–6.0) × 10⁸/kg. Patients were separated into two groups relative to the median value; the high NK cell count group (48 cases) had NK cell counts ≥1.8 × 10⁸/kg and the low NK cell count group (48 cases) had values <1.8 × 10⁸/kg. The baseline characteristics of patients in the two groups were comparable (Table 1).

Table 1.

Patients’ Baseline Characteristics.

	Low NK Cell Group	High NK Cell Group	Statistical significance
No. patients	48	48
Age, y	31 (10–66)	33 (12–63)	ns
<18	5 (10)	6 (13)	ns
≥18	43 (90)	42 (88)
Sex
Male	28 (58)	30 (63)	ns
Female	20 (42)	18 (38)
Donor Age, y	41 (9–59)	33 (9–58)	ns
ABO blood type of donor and recipient			ns
ABO compatibility	23	27
ABO Major incompatibility	10	9
ABO Minor incompatibility	8	7
ABO Major and minor incompatibility	7	5
Diagnosis			ns
AML	24 (50.00)	24 (50.00)
ALL	16 (33.33)	14 (29.17)
CML	1 (2.08)	1 (2.08)
MDS	5 (10.42)	6 (12.50)
ML	2 (4.17)	3 (6.25)
Disease Status Pretransplant
CR	41 (85)	42 (88)	ns
NCR	7 (15)	6 (13)
Risk Stratification*			ns
High-risk	20 (42)	18 (38)
Low and intermediate-risk	28 (58)	30 (63)
Mononuclear cells (×108/kg)	9.99 (5.89–16.45)	10.01 (5.6–15.23)	ns
CD34+ cells (×106/kg)	9.03 (3.01–20.9)	8.98 (2.98–19.63)	ns
CD3+ cells (×106/kg)	3.50 (0.37–6.02)	3.51 (0.46–6.11)	ns

Data are expressed as, median (range), n, or n (%)

Abbreviations: ALL, acute lymphoblastic leukaemia; AML, acute myeloid leukaemia; CML, chronic myeloid leukaemia; CR, complete remission; MDS, myelodysplastic syndrome; ML acute mixed leukaemia; NCR, not complete remission; NK, natural killer.

^*Risk Stratification according to 2017 European LeukemiaNet (ELN) guidelines. ⁶

Infectious complications

The incidence of febrile neutropenia was higher in the low NK cell count group (37/48, 77%) compared with the high NK cell count group (27/48, 56%; P  =  0.03) (Table 2). Within 100 days post-transplantation, 48 patients had a bacterial infection. Sites of the infection were as follows: upper respiratory tract (44%); lung (38%); blood (8%); skin and soft tissue (6%); urinary tract (6%); gastrointestinal tract (4%); central nervous system (2%). Some patients had more than one site of infection. Of the four bloodstream infections, three were gram-negative and one was gram-positive. From samples where bacteria were detected the following pathogens were identified: Escherichia coli (54%); Staphylococcus epidermidis (6%); Klebsiella pneumoniae (4%); Pseudomonas aeruginosa (4%); Enterococcus faecalis (2%); Staphylococcus aureus (2%).

Table 2.

Infectious complications.

	Low NK Cell Count Group	High NK Cell Count Group	Statistical significance
No. patients	48	48
Febrile Neutropenia			P = 0.03
Yes	37	27
No	11	21
Duration of Fever (Days)	3.1 ± 0.3	2.8 ± 0.4	P = 0.049
Bacterial Infection			P = 0.041
Yes	29	19
No	19	29
Invasive Fungal Disease			ns
Proven	5	3
Probable	10	3
Possible	9	3
CMV viremia			ns
Yes	8	7
No	40	41
EBV Viremia			ns
Yes	1	2
No	47	46

Data are expressed as, mean ± standard deviation, or n.

Abbreviations: CMV, cytomegalovirus; EBV, Epstein-Barr virus.

The cumulative incidence of bacterial infection after transplantation was higher in the low NK cell count group (29/48, 60%) compared with the high NK cell count group (19/48, 40%; P = 0.041). Of the 33 patients that had evidence of fungal infections, 21 had Aspergillus infection and 12 had Candida infection. The lungs were the main site of fungal infection. The incidence of proven and probable IFD was higher in the low NK cell count group (15/48, 31%) compared with the high NK cell group (6/48, 13%; P = 0.026) (Table 2). There was no difference between low and high NK cell count groups in the prevalence of CMV or EBV viremia (Table 2).

Neutrophil and Platelet Engraftment and GVHD

The mean time to neutrophil engraftment was 10 days (95% CIs; 10–11) in the high NK cell count group and 10 days (95% CI; 9–11) in the low NK cell count group. The mean time to platelet engraftment was 11 days (95% CI; 10–12) in the high NK cell count group and 12 days (95% CI; 10–13) in the low NK cell count group. There was no statistically significant difference between the two groups in neutrophil and platelet recovery times.

In the low NK cell count group, 17 patients had acute GVHD of grades II and above and seven patients had acute GVHD of grades III and IV. In the high NK cell count group, 14 patients had acute GVHD of grades II and above and seven patients had acute GVHD of grades III and IV. There was no statistically significant difference between the two groups in the cumulative incidence of grades II and above acute GVHD.

Survival analysis

Although the numbers of relapsed patients in the low and high NK cell count group were 4 and 11, respectively, the difference between groups was not statistically significant. For PFS, the 1-year, 2-year, and 3-year figures in the low NK cell count group were 80%, 78%, and 79%, respectively. For the high NK cell count group, the 1-year, 2-year, and 3-year figures were 83%, 81%, and 81%, respectively. For OS, the 1-year, 2-year, and 3-year figures in the low NK cell count group were 82%, 76%, and 76%, respectively. For the high NK cell count, the 1-year, 2-year, and 3-year figures were 83%, 81%, and 81%, respectively. There were no statistically significant differences between the two groups in PFS or OS (Figure 1).

Figure 1.

Progression-free survival (PFS) and overall survival (OS) for high and low natural killer (NK) cell count groups. PFS was calculated from transplant day to the date the disease relapsed and OS was calculated from disease onset to the day of death or last follow-up visit. There were no statistically significant differences between the two groups in PFS or OS.

Discussion

NK cells are cytotoxic components of innate lymphoid cells (ILC). ¹⁰ Research has shown that NK cells kill bacteria and parasites in addition to tumour and virus-infected cells. ¹¹ Studies have shown that NK cells exhibit in vitro activity against both Aspergillus and non-Aspergillus molds.^12–15 We found that the rates of febrile neutropenia, bacterial infection, and IFD were significantly reduced in the high NK cell count group compared with the low NK cell count group. Our findings suggest that NK cells may play an important role in antibacterial and antifungal activity in vivo.

A previous study in patients with acute myeloblastic leukaemia who received a busulfan-fludarabine-ATG conditioning regimen for allogeneic transplant, reported that NK cell count in the graft was significantly associated with OS but not with neutrophil or platelet engraftment. ¹⁶ In our study, we found no significant difference in neutrophil or platelet recovery times or OS between high and low NK cell count groups. Differences between studies may be related to differences in methodology and the impact of other factors such as CD34⁺ cells, bone marrow microenvironment, infection, etc.

Indeed, the effects of NK cells on posttransplant OS, PFS and GVHD, remain controversial. For example, one study reported that transplants from matched-related and unrelated donors with high contents of CD3⁺, natural killer-like T cells (NKT), and CD16-CD56⁺ subpopulations in the peripheral blood stem cell transplant (PBSCT) graft were associated with poor immunologic recovery and compromised event-free survival (50% vs. 80%) due to both increased relapse incidence and non-relapse mortality. ¹⁷ The investigators also found that the significant independent predictor of moderate and severe chronic GVHD was the high prevalence of iNKT-positive, Vβ11-positive, and double-positive cells in the grafts. ¹⁷ Another study reported that in human leukocyte antigens (HLA)-matched sibling transplants, a dose of <1.5 × 10⁷/kg NK/T cells was associated with disease-free survival (DFS), and a dose of <3 × 10⁷/kg NK cells was associated with OS. ¹⁸ Moreover, a study in patients with aplastic anaemia who received haplo-hematopoietic stem cell transplantation (HSCT), reported that the incidence of grades II to IV acute GVHD may be related to the number of CD34⁺ cells. ¹⁹ In this present study, we found no difference between low and high NK cell groups in the cumulative incidence of grades II and above acute GVHD, PFS, or OS. Again, differences in findings may be due to differences in transplant types, GVHD prevention protocols, or different numbers and subsets of infused NK cells. In addition, the number of NK cells in our low NK cell group (i.e., <1.8 × 10⁸/kg) was higher than previously reported in the literature. ²⁰

Our study had several limitations. For example, subsets of NK cells were not analysed. In addition, the use of ATG may have impacted the number and function of NK cells in graft tissue after the transplant. Therefore, more research on different transplantation models is required to substantiate our results. Furthermore, different NK cell counts have been obtained from other centres and so it is difficult to define the threshold value that plays the most effective role after transplantation. More studies are required before implementing these research findings into clinical practice.