Toll-like receptor genetic variations in bone marrow transplantation

The Toll-like receptor family mediates the innate immune system through recognizing the molecular patterns of microorganisms and self-components and leading the synthesis of the inflammatory mediators. We retrospectively examined whether or not genetic variations in toll-like receptor 1 (rs5743551, -7202GQ>A), toll-like receptor 2 (rs7656411, 22215G>T), and toll-like receptor 4 (rs11536889, +3725G>C) affected transplant outcomes in a cohort of 365 patients who underwent unrelated HLA-matched bone marrow transplantation (for hematologic malignancies through the Japan Marrow Donor Program. Only donor toll-like receptor 4 variation significantly improved the survival outcomes. A multivariate analysis showed that the donor toll-like receptor 4 +3725G/G genotype was significantly associated with a better 5-year progression-free survival and a lower 5-year transplant-related mortality than other variations. Furthermore, the donor toll-like receptor 4 +3725G/G genotype was associated with a significantly lower incidence of fatal infections than other variations. The validation study of 502 patients confirmed that the donor toll-like receptor 4 +3725G/G genotype was associated with better survival outcomes. Toll-like receptor4 genotyping in transplant donors may therefore be a useful tool for optimizing donor selection and evaluating pretransplantation risks.


INTRODUCTION
Allogeneic hematopoietic stem cell transplantation (SCT) is a potentially curative treatment for hematologic malignancies [1].Despite HLA matching and substantial improvements in supportive care, life-threatening complications, including severe infections, organ damage and graft-versus-host disease (GVHD), remain an enormous obstacle [2].There is growing evidence that non-HLA genetic variation involved in the immune response also represents a significant determinant of outcomes after SCT [3][4][5][6][7][8].
The toll-like receptor (TLR) family, the most important pattern recognition receptor family, plays a central role in sensing invading pathogens and tissue damage as danger signals, which leads to the massive release of inflammatory mediators such as pro-inflammatory cytokines, reactive oxygen species, antimicrobial peptides and acute-phase proteins into the bloodstream [9,10].TLRs are ubiquitously expressed on various cells, such as macrophages, dendritic cells, B cells, T cells, fibroblasts and epithelial cells.Previous reports [11][12][13] have suggested that TLRs contribute to the inflammatory processes after SCT, where subsequent translocation of bacterial components as well as release of endogenous danger molecules stimulate TLRs to trigger cytokine storm, leading to organ damage, which may vitiate the beneficial anti-microbial effect of TLRs.
Several TLR genes have been investigated in terms of the impact of their variation on outcome and susceptibility to infection and cancer.Among them, 3 functional variations [14][15][16] in the TLR1 (rs5743551, -7202A>G), TLR2 (rs7656411, 22215G>T) and TLR4 (rs11536889, +3725G>C) genes with high (>0.2) minor allele frequencies in the Asian population were retrospectively investigated to determine their associations with transplant outcomes in a cohort of patients who underwent unrelated HLA-matched bone marrow transplantation (BMT) for hematologic malignancies through the Japan Marrow Donor Program (JMDP).

Frequencies of TLR genotypes
The rs5743551 (-7202G>A) variation in the TLR1 gene, the rs7656411 (22215G>T) variation in the TLR2 gene and the rs11536889 (+3725G>C) variation in the TLR4 gene were genotyped in 365 recipients with hematologic malignancies and their unrelated donors in the discovery cohort (Table 1).The frequencies of A/A, A/G and G/G in the TLR1 -7202G>A genetic variant were 13%, 44% and 44% in the recipients and 8.8%, 41% and 50% in the donors (P=0.52),respectively.The frequencies of G/G, G/T and T/T in the TLR2 22215G>T genetic variant were 28%, 52% and 19% in the recipients and 36%, 44% and 20% in the donors (P=0.43),respectively.The frequencies of C/C, C/G and G/G in the TLR4 +3725G>C genetic variant were 7.4%, 41% and 52% in the recipients and 5.8%, 39% and 55% in the donors (P=0.85),respectively.

Transplant outcomes according to TLR genotypes
Univariate analyses (Tables 2 to 4) showed that only the donor TLR4 genotype significantly improved the survival outcomes, with the donor TLR4 +3725G/G genotype associated with a better 5-year PFS than the C/C or C/G genotype (62% vs. 43%, P=0.0068; Figure 1A).Five years was set as the study timepoint according to the median follow-up period among the survivors (970 days; range, 125 to 4798 days).The donor TLR4 +3725G/G genotype also exhibited a trend toward a better 5-year OS (54% vs. 41%, P=0.055; Figure 1B) and a lower 5-year TRM (23% vs. 34%, P=0.080; Figure 1C) but did not reduce the 5-year relapse rate (15% vs. 23%, P=0.12; Figure 1D).The decrease in the number of analyzed cases regarding the transplant outcomes was due to the lack of data on the survival time and date of relapse in some cases.An analysis by mean imputation for missing data also showed that the donor TLR4 +3725G/G genotype was associated with a better 5-year PFS (52% vs. 39%, P=0.034) and a trend toward a better 5-year OS (56% vs. 43%, P=0.051) compared to other genotypes (Supplementary Table 1), indicating that the presence of the missing data did not markedly impair the results.After adjusting for clinical factors in the multivariate model (Table 5), the donor TLR4 +3725G/G genotype remained associated with a better 5-year PFS (HR, 0.57; 95% CI, 0.40-0.82;P=0.0022) and also showed tendencies toward a better 5-year OS (HR, 0.75; 95% CI, 0.55-1.0;P=0.061) and 5-year TRM (HR, 0.63; 95% CI, 0.40-1.0;P=0.057) than other variations.No other genotypes of TLR1, TLR2 or TLR4 significantly influenced the survival outcomes in multivariate analyses.
When the main causes of TRM were analyzed according to the TLR4 +3725G>C genotype, the donor TLR4 +3725G/G genotype was associated with a significantly lower incidence of fatal infections than other genotypes (P=0.047; Figure 2).The donor TLR4 +3725G/G genotype resulted in one fifth of the cumulative incidence of fatal infection (0.7% vs. 4.6%; P=0.11; Figure 3), although there were no statistical differences.

The validation cohort study
The characteristics of the patients in the validation cohort were similar to those of the patients in the discovery cohort with the exception that the donors in the validation cohort were HLA-DPB1 allele-mismatched.The donor TLR4 +3725G/G genotype was associated with a better 5-year OS (58% vs. 51%, P=0.032; Figure 4B; Table 6) as well as a trend toward better 5-year PFS (52% vs. 47%, P=0.089; Figure 4A).According to a multivariate analysis, the donor TLR4 +3725G/G genotype remained associated with a significantly better 5-year OS (HR, 0.75; 95% CI, 0.56-0.99;P=0.043; Table 7) and moderately better 5-year PFS (HR, 0.81; 95% CI, 0.62-1.1;P=0.12).

DISCUSSION
The discovery cohort study, which consisted of recipient and donor pairs for which the HLA-A, -B, -C, Pneumonia, cytomegalovirus 2

Pneumocystis pneumonia 1
Brain abscess, fungal 1 Sepsis, bacterial 2 TNC, total number of nucleated cells harvested.www.impactjournals.com/oncotarget-DRB1, -DQB1, and -DPB1 alleles were completely matched, revealed that the donor G/G genotype at rs11536889 (+3725G>C) of the TLR4 gene predicts significantly better 5-year PFS than other genotypes in patients with hematological malignancies receiving unrelated BMT.The beneficial effects of the donor TLR4 +3725G/G genotype were seen on TRM and death attributable to infections but were not evident with respect to GVHD, suggesting that the donor TLR4 +3725G/G genotype helps prevent fatal infections.The validation cohort study, which consisted of recipient and donor pairs with one or two HLA-DPB1 mismatched alleles, showed consistent results with the donor TLR4 +3725G/G genotype associated with better survival outcomes.The mechanisms through which the donor TLR4 +3725G/G genotype exerts its beneficial effects remain to be determined.A recent report [16] demonstrated that the TLR4 +3725G>C variation located in its 3'-untranslated region was functional, and monocytes from TLR4 +3725G/G subjects expressed lower levels Evidence of TLR4 activation having a negative influence on outcomes following severe infection has been demonstrated in previous studies using mouse models [17][18][19], showing that blockade of TLR4 inhibited systemic inflammatory responses in sepsis, which reduced the disease severity and lethality.Although one may deduce from these findings that the putatively hypoactive TLR4 +3725G/G genotype increases susceptibility to infectious diseases, the TLR +3725G/G genotype was associated with a lower risk of developing sepsis in China [20], suggesting this to be unlikely.This hypothesis may also be supported by the association of the TLR +3725G/G genotype with a lower rate of Gram-negative infection in sepsis in Germany [21], and with a lower susceptibility to severe gastric atrophy related to Helicobacter pylori infection [22,23], hepatitis B recurrence after liver transplantation [24] and advanced periodontitis [25,26] in the Asian populations.Of note, the TLR +3725G/G genotype was beneficially associated with a lower risk of prostate cancer in Sweden [27] and a lower risk of developing chemotherapy-induced neutropenia in children with ALL in Netherland [28], although their etiological relevance to the current findings are unclear.In the present study, the TLR +3725G>C variation did not significantly influence engraftment and relapse rates after SCT.Bochud, et al. [29] reported that donor TLR4 haplotype S4 among haplotypes S1 to S4 was associated     with a risk of invasive aspergillosis after allogeneic hematopoietic stem cell transplantation.However, because haplotype S4 is absent in Asian populations [30,31], which is defined by the TLR4 variations rs4986790 [D299G] and rs4986791 [T399I], but not by TLR4 variation rs11536889 that was investigated in the present study, we decided not to investigate the potential link between each haplotype and the transplant outcomes.One major limitation associated with the present study is that the detailed information on the infections, including the types, severity, treatments, and therapeutic appropriateness, was beyond the scope of the current study and the only the information on the causes of death (such as fatal infections) was available.
In conclusion, the findings of the present data suggested that the donor TLR4 +3725G>C variations predicted better survival outcomes after SCT than other genotypes.Therefore, TLR4 +3725G>C genotyping in donors may be a valuable tool for selecting donors and evaluating pretransplantation risks that, combined with other currently known risk factors, can form the basis for carrying out suitable tailoring of transplantation strategies.Considering the plausible functional roles of these variations, they may be candidates for future

Patients
The patients underwent BMT through the JMDP with T cell-replete marrow between May 1993 and November 2005 (Table 1).No patients had a history of any prior transplantation.The final clinical data analyses of these patients were completed by November 25, 2009.The conditioning regimen varied according to the underlying disease and the condition of the patient.The combination of cyclophosphamide (CY) combined with total body irradiation (TBI) was mainly used for the myeloablative conditioning (MAC) regimen, whereas the combination of fludarabine and melphalan or busulfan was mainly used for the reduced-intensity conditioning (RIC) regimen [32].Cyclosporine or tacrolimus with short-term methotrexate was used for GVHD prophylaxis [33,34].No patients received anti-T cell therapy, such as antithymocyte globulin or ex vivo T cell depletion, in this study.All patients and donors gave their informed consent at the time of transplantation to take part in molecular studies of this nature, in accordance with the Declaration of Helsinki.This project was approved by the Institutional Review Board of Aichi Medical University School of Medicine and the JMDP.All methods were performed in accordance with the approved guidelines and regulations.
For the validation study, the TRL4 rs11536889 variations were imputed using the data from the 1000 Genomes Project 36, as described in our previous study [35].The cohort included 502 patients with hematological malignancies and their unrelated donors who were HLA-A, HLA-B, HLA-C, HLA-DRB1, and HLA-DQB1 allele-matched but who were mismatched by one or two HLA-DPB1 alleles.The diagnoses included AML (n=149; 30%), ALL (n=152; 30%), MDS (n=58; 12%), ML (n=61; 12%) and CML (n=82; 16%).HLA 12/12-allele-matched transplants were incorporated into the discovery cohort to eliminate the impact of HLA-allele mismatch on transplant outcomes, and all 365 available pairs were analyzed in the discovery cohort study.In conducting the validation study, in order to minimize the influence of HLA on the transplant outcomes, all 502 transplants from HLA-DPB1-mismatched donors were incorporated into the validation cohort.

Data management and statistical analyses
Data were collected by the JMDP using a standardized report form.Follow-up reports were submitted at 100 days, 1 year and then annually after transplantation.Only recipients were routinely measured for the pretransplantation cytomegalovirus (CMV) serostatus.The time to neutrophil engraftment was defined as the first of 3 consecutive days with an absolute neutrophil count of more than 0.5×10 9 /L.Acute GVHD developing within the first 100 days posttransplantation was diagnosed and graded based on the established criteria [36].A classification of chronic GVHD observed in patients who survived beyond day 100 was based on the Seattle criteria [37].The overall survival (OS) was calculated from the date of transplantation to the date of death from any cause.Disease relapse was defined as the number of days from transplantation to disease relapse or progression.The transplant-related mortality (TRM) was defined as death due to any cause other than relapse or disease progression.The PFS was defined as survival without disease relapse or progression.Any patients who were alive at the last follow-up date were censored.Data regarding the clinical and microbiological characteristics of infections, postmortem changes, prophylaxis against infections and therapy for GVHD given on an institutional basis were not considered in this study.www.impactjournals.com/oncotargetAll of the statistical analyses were carried out using the EZR software package [38].The probabilities of OS and PFS were calculated using the Kaplan-Meier method, and comparisons between groups were performed via the log-rank test.The occurrence of TRM, disease relapse, acute GVHD and chronic GVHD were compared using the Gray test [39] and analyzed using the cumulative incidence analysis [40], considering relapse, death without disease relapse, death without acute GVHD, death without chronic GVHD and death without engraftment as respective competing risks.The cumulative incidence of fatal infection was analyzed using the Gray test, while considering fatal infections without disease relapse and relapse or death without fatal infection as competing risks.A multivariate Cox model was constructed for the OS and PFS, and a Fine-Gray competing risk regression model was constructed for TRM, relapse, grades 2-4 acute GVHD, grades 3-4 acute GVHD and chronic GVHD using stepwise selection at a significance level of 5% to evaluate the hazard ratio (HR) associated with the TLR1, TLR2 and TLR4 genotypes.Recipient age at the time of BMT, sex, pretransplantation CMV serostatus, disease characteristics (i.e.disease type, disease lineage and disease risk at transplantation), donor characteristics (i.e.age, sex compatibility and ABO compatibility), transplant characteristics (i.e.MAC or RIC and total number of nucleated cells harvested per recipient weight) and year of transplantation were used as covariates.Although these covariates were all available for the discovery study, some of these covariates were not available for the validation study (Table 1).The median was used as the cut-off point for continuous variables.The chi-squared and Mann-Whitney tests were used to compare the results of two groups.The linkage disequilibrium (LD) structure among the single nucleotide polymorphisms in the TLR1, TLR2 and TLR4 gene was determined using HAPLOVIEW, version 4.2 [41,42].For all analyses, P<0.05 was considered statistically significant.

Figure 1 :
Figure 1: The Kaplan-Meier analysis of the progression-free survival rates (A) and the overall survival rates (B), and the estimated incidence curves of transplant-related mortality (C) and disease relapse (D) after transplantation according to the donor TLR4 genotype in the discovery cohort.The solid lines represent the donor G/G genotype, and the dashed lines represent the donor C/C or C/G genotype.

Figure 2 :
Figure 2: The main causes of death after transplantation according to the recipient (A) and donor TLR4 genotype (B) in the discovery cohort.The asterisk denotes P <0.05.

Figure 3 :
Figure 3: The cumulative incidence of fatal infection after transplantation according to the donor TLR4 genotype in the discovery cohort.The solid lines represent the donor G/G genotype, and the dashed lines represent the donor C/C or C/G genotype.

Figure 4 :
Figure 4: The Kaplan-Meier analysis of the progression-free survival rates (A) and the overall survival rates (B) after transplantation according to the donor TLR4 genotype in the validation cohort.The solid lines represent the donor G/G genotype, and the dashed lines represent the donor C/C or C/G genotype.

Table 5 : The results of a multivariate analysis regarding the association between TLR1, TLR2 and TLR4 variations and clinical outcomes after transplantation in the discovery cohort
TNC, total number of nucleated cells harvested.Underlined and bold results regarding the genotype represent P <0.05.