The effect of KIR/HLA incompatibilities on patient outcomes in the setting of haploidentical hematopoietic stem cell transplant with post-transplant cyclophosphamide

In many hematological malignancies, allogeneic hematopoietic stem cell transplant (allo-HSCT) can be curative. However, in cases where a human leukocyte antigen (HLA)-identical family donor is unavailable, and an unrelated donor has not been located, it is possible to conduct a HLA-haploidentical HSCT from a family donor. This approach risks severe graft-versus-host disease (GvHD). To circumvent this, T-cell depletion is conducted which, despite preventing GvHD, can lead to higher relapse rates and high rates of mortality from infection due to immunodeficiency.

Natural killer (NK) cell cytotoxicity is regulated by the recognition between killer cell Ig-like receptors (KIR) and their respective HLA. A lack of this bonding leads to NK cells destroying damaged or allogeneic cells. Using post-transplant cyclophosphamide (PTCy) to deplete donor cells is one strategy that has been adopted in patients undergoing haploidentical HSCTs.

In order to assess the impact of an incompatibility between KIR/HLA (KIR/HLA inc.) in T-replete haploidentical allo-HSCT with PTCy on patient outcomes and NK cell reconstitution, Catherine Willem, Nantes, France, and colleagues conducted a retrospective study in 51 donor/recipient (D/R) pairs. The results were recently published in The Journal of Immunology.¹

Study design and patient characteristics

• Prospective study using the PROMISE database of the European Society for Blood and Marrow Transplantation (EBMT) (N = 51)
• Median age of patients: 57 years (29–70)
• Majority had myeloid diseases (31/51)
• Conditioning regimen:
  • Baltimore (fludarabine-based) regimen (n = 25)
  • Clo-Baltimore (clofarabine-based) regimen (n = 26)
• Haploidentical allo-HSCT conducted with peripheral blood stem cells (PBSC)
• GvHD prophylaxis:
  • PTCy (50 mg/kg/d) on days +3 and +4
  • Cyclosporine A and mycophenolate mofetyl from day +5
• Definitions:
  • Overall survival (OS): time from day 0 of allo-HSCT to death or last follow-up
  • Disease-free survival (DFS): time from day 0 of allo-HSCT to time without evidence of relapse/disease progression as of date of death/last follow-up

Outcomes and GvHD rates of whole cohort

• Rates of GvHD in all patients:
  • Grade 2–4 acute GvHD (aGvHD): 57% (n = 29)
  • Grade 3–4 aGvHD: 16% (n = 8)
  • Median time to onset of GvHD post-transplant: 29 days (17–121)
  • Moderate/severe chronic GvHD (cGvHD): 20% (n = 10)
• As of April 2018:
  • Median follow-up: 29.1 months (12.4–50)
  • Of the survivors, 31% (n = 16) relapsed and 37% (n = 19) died
    • Causes of death: relapse (n = 11), infection (n = 4), GvHD (n = 3) and dermatomyositis (n = 1)
  • Two-year OS: 61.6% (95% CI, 49–76)
  • Two-year DFS: 50.3% (95% CI, 38–66)
  • When comparing Baltimore versus Clo-Baltimore and lymphoid versus myeloid malignancies no statistically significant differences were observed between groups in terms of grade 2–4 acute GvHD (aGvHD) or moderate/severe GvHD incidence, relapse, death, 2-year DFS and 2-year OS
  • A lower incidence of relapse was observed in patients developing grade 2–4 aGvHD

Impact of KIR/HLA inc. on outcome

• Genetic profiling of the 51 D/R pairs was conducted
• Inhibitory (inh) KIR/HLA inc. (inh. KIR/HLA inc.) defined by the presence of KIR2DL1, KIR2DL2/3 and KIR3DL1 in donor with an absence of the respective conjugate HLA ligands in the recipient
• The effect of inh. KIR/HLA inc. on aGvHD rate, relapse rate, two-year DFS and two-year OS is shown in Table 1

Table 1: The impact of inh. KIR/HLA inc. on key measures of outcome

NK cell analysis

• In 34 D/R pairs, the kinetics of NK cell reconstitution were analyzed by immunophenotyping
  • A depletion in NK cells from day 5–20 was observed
    • Proliferative NK cells were targeted by the early PTCy administration
  • NK cell recovery associations:
    • At day 25, patients on the Baltimore regimen had a higher NK cell recovery versus the Clo-Baltimore regimen (P = 0.0001)
    • Not associated with GvHD, relapse or inh. KIR/HLA inc.
    • inh. KIR/HLA inc. impacted KIR NK cell reconstitution post-transplant
  • NKp46⁺ 2B4⁺ subsets indicating NK cell activation and differentiation appeared from day 30
  • Higher representation of NKp46⁺ 2B4⁺ cells in patients correlated with development of GvHD at day 25, P = 0.03
  • Lower representation of NKp46⁺ 2B4⁺ cells up to day 25 in patients with relapse (P = 0.03)
  • Patients developing GvHD had fewer undifferentiated 2B4⁺ NKp46^- cells at day 20 (P = 0.04)

Conclusion

In this specific setting, of RIC PBSC, haploidentical HSCT with PTCy, the presence of inh. KIR/HLA inc. were associated with improved clinical outcomes, survival rates and lower relapse rates. The authors proposed a model, showing the impact of inh. KIR/HLA inc. on GvHD and relapse in this setting.

The model shows:

• In the presence of KIR/HLA inc.: alloreactive NK cells are activated and proliferate and are quickly targeted by PTCy. This activation promotes differentiation of NK cells associated with GvHD and a lower relapse rate
• In the absence of KIR/HLA inc.: quiescent NK cells are not targeted or eliminated by PTCy and are therefore inefficient at limiting relapse

Being able to predict NK cell alloreactivity may prove to be a significant consideration when choosing a donor for a haploidentical transplant and in predicting a patient’s outcome.