Background: Donor grafts with more naive T cells and plasmacytoid dendritic cells were associated with improved overall survival after unrelated donor bone marrow, but not peripheral blood stem cell (PBSC) transplants (Waller, E JCO 2014). Here we present results on influence of innate and adaptive immune subsets in G-CSF mobilized allografts on incidence of acute GVHD (aGVHD) and chronic GVHD (cGVHD) in 238 patients (pts).

Methods: We analyzed the absolute numbers and percentages of T, NK, NKT and B cells along with an extensive immunophenotypic characterization of their activation status in consecutive PBSC allografts obtained from sibling and unrelated donors between 2010 - 2014 and studied their association with the incidence of aGVHD and cGVHD. Wilcoxon rank sum tests were used to screen differential marker expression between those who did vs. did not develop aGVHD and similarly for cGVHD. Significant markers were evaluated in the multivariable (m.v.) setting along with known prognostic factors, including: recipient age, related vs. unrelated donor, female donor vs. not, Anti-thymocyte globulin (ATG) use (yes vs. no), and Reduced-intensity conditioning (RIC) vs. not. Cutpoints for markers were generated using recursive partitioning algorithms and evaluated in m.v. models.

Results: Of the 238 alloSCT pts evaluated, most (71%) had unrelated donors, 64% received ATG, where most pts with unrelated donors received ATG (83%), and 78% received RIC. The incidence of aGVHD and cGVHD was 58% and 38% respectively. A total of 107 pts had grade II-IV aGVHD reported (71 II, 28 III, 8 IV), and 92 of 192 evaluable for cGVHD (at least 100 days of f/u) had reported cGVHD. Median follow-up in living pts was 21 months (range: 1.4 to 41.1 months). Table 1 shows dichotomized markers most influential on aGVHD. Higher absolute numbers of T cells, activated T cells, CD8+ cells, CD8+ cells expressing IL-7 receptor and CD27 were associated with higher incidence of aGVHD. Higher number of Stage 4 NK cells expressing stem cell factor receptor, and T-regs were associated with a lower incidence of aGVHD. Similar analyses were done for cGVHD (Table 2). Higher absolute numbers of activated T lymphocytes, activated B lymphocytes, KIR expressing CD3+ cells, CD8+ lymphocytes and activated NK cells were associated with higher incidence of cGVHD. When the percent of these makers in relation to total lymphocytes was evaluated regarding association with aGVHD, higher percent of T-regs (OR: 0.204, p=0.0018), effector memory T cells (OR: 0.45, p=0.024) and NKG2D positive NK cells (OR: 0.38, p=0.0008) conferred protection from aGVHD . Similar analysis for cGVHD showed higher percent of naïve CD4+ T cells conferred protection from cGVHD (OR: 0.44; p=0.0062) while higher percent of CD8+ cells (OR: 3.93; p=0.0032) and activated NK cells (OR: 2.08; p=0.024) was associated with cGVHD.

Conclusions: These results show a protective role of donor T-regs, CD4+ T cells and Stage 4 NK cells from aGVHD. Additionally, higher content of activated T cells, CD8+ cells and B lymphocytes are associated with higher incidence of cGVHD. Higher content of activated NK cells seems to protect from aGVHD, but not from cGVHD. Updated results including multivariable analyses will be presented. These findings showing the influence of specific subsets in the allograft on aGVHD and cGVHD may provide opportunities for therapeutic interventions for graft engineering or pharmacologic methods for targeting specific immune subsets to decrease incidence of aGVHD and cGVHD.

Table 1

Univariate model results for aGVHD with dichotomized markers using cutpoints:

Marker Absolute 
OR p-value 
CD3+/CD5616- (T lymphocytes) 3.07 0.0013 
CD3+/HLA DR+ (Activated T lymphocytes) 3.26 0.012 
CD8+/CD45RA- (CD 8+ lymphocytes) 2.56 0.012 
CD8+/CD27+ (Effector Memory CD8 cells) 3.25 0.0082 
CD8+/CD127+ ( CD8 cells expressing IL-7 receptor) 2.92 0.073
 
CD4+/CD25+/CD127-(T regs) 0.43 0.057 
CD3-/CD16-/CD56+/CD117+ (Stage 4 NK cells expressing Stem cell factor receptor) 0.12 0.0007 
Marker Absolute 
OR p-value 
CD3+/CD5616- (T lymphocytes) 3.07 0.0013 
CD3+/HLA DR+ (Activated T lymphocytes) 3.26 0.012 
CD8+/CD45RA- (CD 8+ lymphocytes) 2.56 0.012 
CD8+/CD27+ (Effector Memory CD8 cells) 3.25 0.0082 
CD8+/CD127+ ( CD8 cells expressing IL-7 receptor) 2.92 0.073
 
CD4+/CD25+/CD127-(T regs) 0.43 0.057 
CD3-/CD16-/CD56+/CD117+ (Stage 4 NK cells expressing Stem cell factor receptor) 0.12 0.0007 

Table 2

Univariate model results for cGVHD with dichotomized markers using cutpoints:

Marker Absolute 
OR p-value 
CD3+/HLA DR+ (Activated T lymphocytes) 4.41 <0.0001 
CD19+/CD80+ (Activated B lymphocytes) 2.44 0.0036 
CD3+/CD69+/CD158b+ (Activated T lymphocytes expressing KIR) 2.81 0.017 
CD3-/CD5616+/CD69+ (Activated NK cells) 2.74 0.0024 
CD3-/CD5616+/CD314+ (NKG2D positive NK cells) 2.33 0.008 
Marker Absolute 
OR p-value 
CD3+/HLA DR+ (Activated T lymphocytes) 4.41 <0.0001 
CD19+/CD80+ (Activated B lymphocytes) 2.44 0.0036 
CD3+/CD69+/CD158b+ (Activated T lymphocytes expressing KIR) 2.81 0.017 
CD3-/CD5616+/CD69+ (Activated NK cells) 2.74 0.0024 
CD3-/CD5616+/CD314+ (NKG2D positive NK cells) 2.33 0.008 

Disclosures

Blum:Celgene: Consultancy.

Author notes

*

Asterisk with author names denotes non-ASH members.

Sign in via your Institution