Interleukin-1 genotype and outcome of unrelated donor bone marrow transplantation

Authors


Margaret L. MacMillan, MD, University of Minnesota, MMC 484, 420 Delaware Street SE, Minneapolis, MN 55455, USA. E-mail: macmi002@umn.edu

Abstract

Summary. The interleukin 1 (IL-1) gene family includes three members (IL-1-α, IL-1-β and IL-1Ra) that mediate immune and inflammatory responses through two specific cell surface receptors. Cytosine to thymine transitions at codons −889 and −511 in the IL-1-α and IL-1-β genes, respectively, and an 86-base pair repeat in the IL-1Ra are believed to influence gene transcription. We have genotyped these three polymorphisms in 90 donor/recipient pairs undergoing unrelated donor bone marrow transplantation (BMT) at the University of Minnesota. We found no association between the occurrence of acute GVHD and donor and/or recipient polymorphisms of any of the three IL-1 genes. The presence of at least one IL-1α− 889 T allele in the donor was associated with significantly improved survival in univariate analysis (survival at 1 year 40% C/C donor, 68% T/C donor, 75% T/T donor, P < 0·01). Multiple regression analysis showed that if the donor and recipient each possessed the IL-1α T allele there was significantly improved survival [relative risk (RR) 0·2, P < 0·01] and decreased treatment-related mortality (TRM; RR 0·2, P = 0·01). The presence of the IL-1β T allele in donor and recipient was also associated with improved survival (RR 0·2, P < 0·01) and decreased TRM (RR 0·1, P < 0·01). These data suggest that donor polymorphism in IL-1α and IL-1β might influence survival after unrelated donor BMT, but does not alter risk of GVHD.

Allogeneic donor bone marrow transplantation (BMT) offers a potential cure for a variety of malignant and non-malignant conditions. However, mortality and morbidity remain high from regimen-related toxicity, graft-versus-host disease (GVHD) and opportunistic infections (Kernan et al, 1993). Strategies to identify individuals who are likely to experience adverse outcomes after BMT might enable tailoring of preparative therapy, GVHD prophylaxis and infectious disease prophylaxis, leading to improved survival. In addition, identifying favourable prognostic factors in the donor may aid in selection and improve outcomes after BMT.

Evidence suggests that non-human leucocyte antigen (HLA) genetic variation in individuals influences risk of GVHD and BMT mortality, and that it might be possible to predict BMT outcome from a profile of recipient risk factors, including cytokine polymorphisms. Previous investigations have demonstrated a possible role for polymorphism in tumour necrosis factor-α (TNFα), interleukin 10 (IL-10) and members of the IL-1 family in outcome of sibling donor BMT (Middleton et al, 1998; Cavet et al, 1999; Lin et al, 2000; Cullup et al, 2001). However, there are no data that address this issue in unrelated donor BMT. IL-1 is one of the key cytokines in the initiation and maintenance of an inflammatory response. The IL-1 gene family is composed of three members [IL-1-α, IL-1-β and IL-1-receptor antagonist (IL-1Ra)] that mediate immune and inflammatory responses through two specific cell surface receptors. IL-1-α and IL-1-β are agonists and IL-1Ra is a competitive receptor antagonist. IL-1-β is an extracellular protein released from cells while IL-1-α is largely a regulator of intracellular events and mediator of local inflammation (Dinarello, 1996). Nearly all microbes and microbial products induce the expression of all three genes in humans (Dinarello, 1996).

Polymorphisms occur in each of the three IL-1 family genes which are found close together on the long arm of chromosome 2, distributed over 430 kb (Nicklin et al, 2002; Taylor et al, 2002). Polymorphisms at these three loci have been associated with increased risk for malignancy (Barber et al, 2000; El-Omar et al, 2000), mortality from infections (Read et al, 2000), chronic inflammatory diseases (Rider et al, 2000; Zhang et al, 2002) and autoimmune disorders (Pociot et al, 1992; Tarlow et al, 1994; McDowell et al, 1995; Rider et al, 2000; McGarry et al, 2001; Tazi-Ahnini et al, 2002; Zhang et al, 2002). As IL-1 is one of the cytokines implicated in the initiation and maintenance of GVHD (Ferrara & Deeg, 1991), and may modulate immune recovery after BMT (Weisdorf et al, 1994), we hypothesized that polymorphism of these genes may influence the risk of GVHD and the response to infection in recipients of allogeneic BMT. In this study, we prospectively genotyped unrelated bone marrow donors and recipients for polymorphism in the IL-1 family of genes [single nucleotide polymorphisms in IL-1α and Il-1β and variable number tandem repeat (VNTR) in IL-1Ra]. Genotypes were then correlated with occurrence of acute and chronic GVHD and survival.

Materials and methods

Patients

The study population included 90 consecutive donor/recipient pairs who received an unrelated donor BMT at the University of Minnesota. Patients and donors were typed for HLA-A, -B, using serological techniques, and for HLA-DRB1, using high-resolution molecular typing, to identify all World-Health-Organization-recognized specificities current at the time of BMT. Stem cells were HLA-A, -B, -DRB1 matched in 66 (73%) cases, and had a single antigen mismatch in 24 (27%) cases. Patient characteristics, including underlying disease, type of donor and GVHD prophylaxis are shown in Table I. Median age at time of BMT was 11·1 years (range 0·8–54·8 years). All transplant and GVHD protocols were reviewed and approved by the University of Minnesota Institutional Review Board. All patients and/or guardians gave informed consent.

Table I.  Patient demographics (n = 90).
Featuresn (%)
  • *

    Other malignancy: multiple myeloma, plasma cell leukaemia.

  • ALL, acute lymphoblastic leukaemia; AML, acute myelogenous leukaemia; CML, chronic myelogenous leukaemia; TLI, total lymphoid irradiation.

Age
 < 18 years54 (60)
 ≥ 18 years36 (40)
 median11·1 (0·8–54·8)
Sex
 Male52 (58)
 Female38 (42)
Year of transplant
 19981 (1)
 199924 (27)
 200040 (44)
 200125 (28)
Diagnosis
 Aplastic anaemia4 (4)
 Fanconi anaemia15 (17)
 Metabolic disorder30 (33)
 Haemoglobinopathy1 (1)
 ALL4 (4)
 AML14 (16)
 CML12 (13)
 Myelodysplastic syndrome3 (3)
 Hodgkin's lymphoma1 (1)
 Non-Hodgkin's lymphoma4 (4)
 Other malignancy*2 (2)
HLA matching
 A,B DRB1 matched66 (73)
 One locus mismatched24 (27)
Preparative therapy
 Cyclophosphamide and TBI31 (34)
 Other chemotherapy and TBI41 (46)
 Chemotherapy and TLI3 (3)
 Chemotherapy alone15 (17)
GVHD prophylaxis
 T-cell depletion (elutriation)38 (42)
 Methotrexate/cyclosporine20 (22)
 Cyclosporine + other32 (36)

Details of the preparative therapy and GVHD prophylaxis used have been previously reported (Ramsay et al, 1983; Kersey et al, 1987; McGlave et al, 1987, 2000; Kim et al, 1990; Wagner et al, 1996;Davies et al, 1997; Barker et al, 2001). A reduced-intensity preparative regimen was used in 11 (12%) patients. Patients received supportive care, which included ongoing prophylaxis for bacterial infections [250 mg penicillin VK (penicillin V potassium) b.i.d. and ciprofloxacillin 500 mg p.o. b.i.d], fungal infections (fluconazole 200 mg/d), Pneumocystis carinii pneumonia (trimethoprim-sulpha double strength b.i.d. every Monday and Tuesday) and cytomegalovirus (CMV) (800 mg acyclovir 10 mg/kg i.v. every 8 h followed by 800 mg p.o. five times per day). Children received the same prophylaxis, but appropriately dose adjusted for weight. Acute GVHD was diagnosed clinically with histological confirmation whenever possible. Symptoms of acute GVHD were graded by standard clinical criteria (Glucksberg et al, 1974; Weisdorf et al, 1990), modified to include upper gastrointestinal (GI) acute GVHD, according to the GVHD consensus conference (Przepiorka et al, 1995).

IL-1 genotypes

Peripheral blood was collected from donors and recipients prior to BMT, and DNA extracted using standard techniques. All analyses were performed blinded with respect to patient characteristics.

IL-1-α.  The 5′ flanking region of this gene contains a single base pair (bp) polymorphism at position − 889 (cytosine for thymine substitution). Polymerase chain reaction (PCR) amplification was performed using specific primers, essentially as described by Mansfield et al (1994). The PCR product was digested with Nco I, and the products analysed on a 12% polyacrylamide gel stained with ethidium bromide. Allele 1 (cytosine) yielded products of 83 and 16 bp; allele 2 (thymine) yielded a product of 99 bp.

IL-1-β. Position − 511 in the promoter region of this gene has a single base pair polymorphism (cytosine to thymine substitution, the cytosine allele completing an Ava I site). PCR amplification was performed essentially as described by Mansfield et al (1994). PCR amplification followed by Ava I digestion enabled the alleles to be identified on a 9% polyacrylamide gel stained with ethidium bromide. Allele 1 (cytosine) yielded products of 87 and 20 bp; allele 2 (thymine) yielded a product of 107 bp.

IL-1Ra.  Primers flanking the 86 bp VNTR in intron 2 of the IL-1Ra were used (5′ CTCAGCAACACTCCTAT 3′ and 5′ TCCTGGTCTGCAGGTAA 3′). PCR amplification was performed with a MgCl2 concentration of 2·0 mmol/l for 30 cycles of 1 min at 58°C and 1 min at 70°C. PCR products were analysed by electrophoresis through 2% agarose and alleles identified by size. Allele 1 was 410 bp, allele 2 was 240 bp and allele 3 was 500 bp.

Statistical analysis

Data regarding transplant patient characteristics, post-transplant complications and outcomes were prospectively collected by the Biostatistical Support Group using standardized collection procedures. The cumulative incidence of grade II–IV and grade III–IV acute GVHD and chronic GVHD were calculated by treating deaths from other causes as competing risks (Lin, 1997). Survival was estimated by the Kaplan–Meier method (Kaplan & Meier, 1958). Univariate comparisons were completed by using the log-rank test statistic. As previous data have indicated a possible ‘pro-inflammatory’ effect of IL-1α and Il-1β T alleles, donor/recipient pairs were grouped separately for IL-1α and Il-1β into four mutually exclusive categories for analyses of genotypes (Hurme & Santtila, 1998). These categories were pairs in which: recipient but not donor had at least one T allele, donor but not recipient had at least one T allele, both donor and recipient had at least one T allele, and neither donor nor recipient had a T allele (Table II). Similarly, allele 2 in the IL-1Ra VNTR has been associated with a ‘pro-inflammatory’ haplotype (Hurme & Santtila, 1998). For analyses, donor recipient pairs were categorized into four mutually exclusive categories as pairs in which: recipient but not donor had at least one A2 allele, donor but not recipient had at least one A2 allele, both donor and recipient had at least one A2 allele, and neither donor nor recipient had an A2 allele (Table II). To determine the independent effect of study variables, Cox regression was performed (Cox, 1972). Factors included in the models were recipient and donor age, year of transplant, recipient and donor sex, diagnosis (including standard risk versus high risk), CMV serology, GVHD prophylaxis [T-cell depletion (TCD) elutriation versus no TCD], conditioning regimen [total body irradiation (TBI) versus no TBI], HLA mismatch, CD34+ cell dose, and acute GVHD as a time-dependent variable. All variables were tested for violation of the proportional hazards assumption. As multiple comparisons were made in this study, only P-values ≤ 0·01 were considered statistically significant.

Table II.  IL-1 genotype frequencies.
Genotypen (%)
Donor IL-1α− 889 genotype
 C/C45 (50)
 T/C32 (36)
 T/T13 (14)
Donor IL-1β− 511 genotype
 C/C33 (37)
 T/C41 (46)
 T/T16 (18)
Donor IL-1Ra genotype
 A1/A149 (54)
 A1/A234 (38)
 A2/A25 (6)
 A2/A32 (2)
Recipient IL-1α− 889 genotype
 C/C44 (49)
 T/C39 (43)
 T/T5 (6)
 Missing1 (1)
Recipient IL-1β− 511 genotype
 C/C35 (39)
 T/C39 (43)
 T/T16 (18)
Recipient IL-1Ra genotype
 A1/A141 (46)
 A1/A237 (41)
 A1/A33 (3)
 A1/A41 (1)
 A2/A27 (8)
 A2/A31 (1)
IL-1α− 889 genotype
 Recipient had T25 (28)
 Donor had T25 (28)
 Both had T20 (22)
 Neither had T19 (21)
IL-1β− 511 genotype
 Recipient had T25 (28)
 Donor had T27 (30)
 Both had T30 (33)
 Neither had T8 (8)
IL-1Ra genotype
 Recipient had A227 (30)
 Donor had A223 (26)
 Both had A218 (20)
 Neither had A222 (24)

Results

IL-1 polymorphism frequencies

The allele frequencies of the three polymorphisms studied were similar to those previously reported (di Giovine et al, 1992; Pociot et al, 1992; Tarlow et al, 1993), as shown in Table II. Linkage disequilibrium was observed between the three genes, as reported in previous studies (Cox et al, 1998; Santtila et al, 1998; El-Omar et al, 2000; Hulkkonen et al, 2000). The IL-1β T allele was associated with the IL-1Ra A2 allele more frequently than expected by chance and, conversely, the IL-1β C allele was associated with IL-1Ra allele A1 (P < 0·01). The IL-1α T allele was associated with the IL-1Ra A1 allele (P = 0·01) and the T allele of IL-1β was significantly associated with the C allele of IL-1α (P < 0·01).

Clinical outcomes

Acute GVHD. Thirty-three recipients developed grades II–IV acute GVHD [grade II (n = 22), grade III (n = 3) and grade IV (n = 8)]. There was no association between the incidence of acute GVHD and donor and/or recipient polymorphism of any of the three IL-1 family genes studied (Table III). As the combination of IL-1β T allele and IL-1Ra A2 allele has been described as a ‘pro-inflammatory’ haplotype (Hurme & Santtila, 1998), recipients and donors carrying both these alleles were compared with those without and, again, no difference in GVHD frequencies was seen.

Table III.  Univariate analyses of genotype and frequency of grades II–IV GVHD.
LocusRate of acute GVHD at 100 d (95% CI)P-value
IL-1α− 889 genotype
 Recipient had T48% (28–68%) 
 Donor had T32% (14–50%) 
 Both had T25% (7–43%) 
 Neither had T42% (19–65%)0·20
IL-1β− 511 genotype
 Recipient had T24% (8–40%) 
 Donor had T44% (24–64%) 
 Both had T43% (25–61%) 
 Neither had T25% (0–53%)0·42
IL-1Ra genotype
 Recipient had A237% (19–55%) 
 Donor had A248% (27–69%) 
 Both had A233% (11–55%) 
 Neither had A227% (9–45%)0·55

Chronic GVHD.  Thirteen patients developed chronic GVHD. There was a trend towards a higher incidence of chronic GVHD if both donor and recipient lacked the IL-1α T allele. The cumulative incidence of chronic GVHD at 1 year for these patients was 29%[95% confidence interval (CI) 8–50%]vs 17% (95% CI 0–35%) if both donor and recipient had the T allele, 12% (95% CI 0–24%) if only the recipient had the T allele and 9% (95% CI 0–20%) if only the donor had the T allele (P = 0·09). In addition, there was a trend toward less chronic GVHD if both donor and recipient lacked the IL-1Ra allele A2. Cumulative incidence of chronic GVHD at 1 year for these patients was 5% (95% CI 0–13%) vs 11% (95% CI 0–25%) if both donor and recipient had the T allele, 14% (95% CI 0–28%) if only the recipient had the T allele and 33% (95% CI 13–53%) if only the donor had the A2 allele (P = 0·07). There was no association between the IL-1β polymorphism, or the IL-1β T allele, IL-1Ra A2 allele combination and the incidence of chronic GVHD.

Survival and treatment-related mortality (TRM).  The Kaplan–Meier estimate of 1-year survival for the entire cohort of 90 patients was 55% (95% CI 44–66%). IL-1α genotype of the donor significantly influenced survival [survival at 1 year 40% C/C donor (95% CI 26–54%), 68% T/C donor (95% CI 50–86%), 75% T/T donor (95% CI 50–100%), P < 0·01](Fig 1). The causes of death for patients with the IL-1α C/C genotype were GVHD (n = 10), relapse (n = 9), infection (n = 5), multisystem organ failure (n = 2) and adult respiratory distress syndrome (n = 1). For patients with the T/C IL-1α genotype, the causes of death included GVHD (n = 3), relapse (n = 3) and infection (n = 4). GVHD was the cause of death for all five patients with the T/T IL-1α genotype.

Figure 1.

Probability of survival stratified by donor IL-1α genotype.

Survival was not significantly affected by recipient IL-1α genotype [survival at 1 year 50% C/C recipient (95% CI 34–66%), 58% T/C recipient (95% CI 42–74%), 75% T/T recipient (95% CI 32–100%), P = 0·42](Fig 2). There was a trend towards improved survival and reduced TRM in patients where the donor had at least one IL-1β T allele (TRM 31% if donor and recipient had T allele vs 88% if neither had T allele, P = 0·02). There was no association between polymorphism in the IL-1Ra VNTR in donor or recipient and survival or non-relapse mortality (data not shown). In addition, no significant associations were noted between donor or recipient having the so-called ‘pro-inflammatory’ combination of an IL-1β T allele and IL-1Ra allele A2 and outcomes after BMT (data not shown).

Figure 2.

Probability of survival stratified by recipient IL-1α genotype.

Multiple regression analysis was used to look at the independent effects of IL-1α, IL-1β and IL-1Ra on mortality (Table IV). The T alleles of IL-1α and IL-1β each showed an independent effect on mortality, with the presence of a T allele in the donor showing the strongest association with reduced mortality and improved survival. IL-1Ra did not confound the results of IL-1α and IL-1β nor did it show an independent association with mortality. Mortality was increased in patients with grades III and IV GVHD. The inclusion of GVHD in the model did not significantly alter the effect of IL-1α and IL-1β polymorphisms.

Table IV.  Mortality at 1 year: Cox regression analysis.
FactorRelative risk (95% CI)P-value
  • *

    Reference category.

Il-1α− 889 genotype
 Neither had T*1 
 Recipient had T0·8 (0·4–1·9)0·64
 Donor had T0·3 (0·1–0·9)0·03
 Both had T0·2 (0·05–0·6)< 0·01
IL-1β− 511 genotype
 Neither had T*1 
 Recipient had T0·3 (0·1–0·8)0·02
 Donor had T0·1 (0·05–0·4)< 0·01
 Both had T0·2 (0·06–0·5)< 0·01
Acute GVHD
 Grade 0–II*1 
 Grade III–IV2·9 (1·2–7·2)0·02

Multiple regression analysis was also used to look at the independent effect of IL-1α, IL-1β and IL-1Ra on TRM (Table V). Again, the T alleles of IL-1α and IL-1β each showed an independent effect on TRM, with the presence of a T allele in the donor showing the strongest association with reduced TRM. TRM was increased in patients with grades III and IV GVHD. The inclusion of acute GVHD in the model did not significantly alter the effect of IL-1α and IL-1β polymorphisms.

Table V.  Treatment-related mortality: Cox regression analysis.
FactorRelative risk (95% CI)P-value
  • *

    Reference category.

Il-1α− 889 genotype
 Neither had T*1 
 Recipient had T1·0 (0·4–2·5)0·99
 Donor had T0·4 (0·1–1·1)0·07
 Both had T0·2 (0·05–0·7)0·01
IL-1 β− 511 genotype
 Neither had T*1 
 Recipient had T0·3 (0·1–1·1)0·01
 Donor had T0·1 (0·04–0·4)< 0·01
 Both had T0·1 (0·05–0·5)< 0·01
Acute GVHD
 Grade 0–II*1 
 Grade III–IV2·6 (1·0–7·7)0·06

Discussion

GVHD occurs in 20–40% of recipients of HLA-matched sibling donor grafts, indicating that factors other than HLA match are also important in the initiation of GVHD. A number of previous studies have suggested that polymorphism in cytokine genes influences susceptibility to post-BMT complications. Middleton et al (1998) studied recipients of sibling donor BMT, and demonstrated that the homozygous d3 genotype of the TNF-α microsatellite and the presence of IL-10 alleles with greater numbers of dinucleotide repeats were preferentially associated with grade III/IV GVHD. The association of severe GVHD with IL-10 alleles with greater numbers of dinucleotide repeats was confirmed by the same authors in a larger population from another institution (Cavet et al, 1999). Associations between IL-6 and interferon-γ polymorphism and susceptibility to GVHD have also been shown in sibling donor BMT recipients (Cavet et al, 2001; Socie et al, 2001). These data suggest that non-HLA genetic variation between individuals influences risk of GVHD and BMT mortality, and that it might be possible to predict BMT outcome from a profile of donor or recipient risk factors, including cytokine polymorphisms.

IL-1 is an inflammatory cytokine implicated in both acute and chronic inflammatory diseases. Two functionally similar molecules, IL-1α and Il-1β, are encoded by separate genes (IL1A and IL1B) (Nicklin et al, 2002). IL-1β is primarily an extracellular protein, while IL-1α is chiefly an intracellular mediator of local inflammation (Dinarello, 1996). The third gene of the family (IL1RN) encodes IL-1 receptor antagonist (IL-1Ra), an anti-inflammatory non-signalling molecule that competes for receptor binding with IL-1α and IL-1β. The overall contribution of IL-1 to the pro-inflammatory response depends on the balance between these three molecules. The occurrence of GVHD after BMT is associated with release of pro-inflammatory cytokines, often initially triggered by the conditioning regimen used for transplant preparation. Previous animal studies have shown that neutralization of IL-1 can significantly decrease GVHD mortality rates (Hill et al, 1999). We hypothesized that differences in the ability of BMT donors or recipients to produce IL-1 might influence the occurrence of post-BMT complications. Two previous studies have examined the IL-1 polymorphism and outcome of sibling donor BMT. Cullup et al (2001) studied the IL-1β and IL-1Ra polymorphism and showed modest evidence for an association between donor IL-1Ra genotype and incidence of acute GVHD. In a larger study, Lin et al (2000) reported little association between donor or recipient IL-1β and IL-1Ra genotypes and frequency of GVHD. Neither study investigated IL-1α polymorphism, and no previous studies have examined unrelated donor BMT recipients.

In this study, we chose to examine three polymorphisms, one in each of the genes IL1A, IL1B and IL1RN, that might influence outcome of BMT. These polymorphisms were selected for study because they have been shown in in-vitro studies to influence protein production, and in epidemiological studies to be associated with acute or chronic inflammatory disease. The 5′ flanking region of the IL-1-α gene contains a single base pair polymorphism at position − 889 (cytosine for thymine substitution) in the transcriptional regulatory region of the IL-1-α gene. The T/T genotype creates the consensus site for a novel transcription factor (Skn-1) and is associated with a significant increase in promoter activity compared with the C/C genotype. In our study we found that the donor IL-1α genotype had a significant effect on the outcome of BMT, with survival significantly improved in donor/recipient pairs in which the donor had at least one T allele at this locus. Contrary to our expectation, there was no association of donor or recipient IL-1α genotype with occurrence of acute or chronic GVHD. In multivariate analysis, after adjustment for known BMT risk factors, donor IL-1α genotype remained a significant predictor of survival, independent of the occurrence of GVHD.

Position − 511 in the promoter region of the IL-1-β gene has a single base pair polymorphism (cytosine to thymine substitution). This polymorphism is in almost complete (99·5%) linkage disequilibrium with another polymorphic cytosine to thymine transition at position − 31 of the IL-1-β gene; the presence of a T or a C at − 511 reliably predicts the presence of a T or a C at position − 31 (Cullup et al, 2001). The − 31 polymorphism involves a TATA sequence in the IL-1β promoter. The T allele is associated with increased binding of transcription initiation factors. The reported association of the − 511 T allele with some inflammatory disorders likely reflects increased IL-1β production in carriers of this allele who will also be carriers of the − 31 T allele (Dinarello, 1996; Lin et al, 2000; Cullup et al, 2001; Taylor et al, 2002). In the present study, we found that the presence of at least one − 511 IL-1β T allele in the donor was associated with a fivefold reduction in mortality that was independent of the occurrence of GVHD, similar to, but independent of, our findings with the IL-1α genotype. Of note, in all the analyses performed, the effect of donor genotype on outcome was more notable than the effect of recipient genotype, and no effect of IL-1 polymorphism on risk of acute GVHD was seen.

It is important to note that IL-1α genotype has been shown to influence IL-1β production, therefore, our observations regarding IL-1α and IL-1β genotype might indicate that donor cells that can produce high levels of IL-1β are associated with reduced TRM and better survival after BMT. Analysis of serum IL-1β levels in 400 normal donors showed a significant relationship between IL-1α− 889 genotype and production of IL-1β (Hulkkonen et al, 2000). Individuals homozygous for the IL-1α− 889 T allele produced significantly more IL-1β than IL-1α C/C homozygotes or heterozygotes, who both produced similar levels. These data suggest that the products of the various loci of the IL-1 gene cluster interact in the regulation of production of the effector molecules or, perhaps less likely, the presence of an unidentified locus associated with high IL-1β production in linkage disequilibrium with these loci. In the same study, analysis of IL-1β genotype showed a stepwise but non-significant increase in IL-1β production between C/C homozygotes, C/T heterozygotes and T/T homozygotes. Notably, there was evidence for an interaction between IL-1α and IL-1β genotype, with markedly (around threefold) increased IL-1β production in individuals with an IL-1α T/T genotype who also had at least one IL-1β T allele. If our hypothesis that donor cells that produce high levels of IL-1β are associated with reduced TRM and improved survival is true, we would expect outcomes to be particularly favourable in recipients receiving cells from a donor with at least one IL-1β T allele and an IL-1α T/T genotype, as they would produce the highest protein levels. Seven recipients in our study had such a donor and all had survived to 1 year post BMT.

Additional studies will be needed to confirm these findings in other and larger datasets. The mechanism for reduced TRM while GVHD is unaltered is unclear. IL-1 has been implicated in the response to infection, and it is possible that patients with more favourable IL-1 genotypes have a reduced frequency of infection or are more likely to survive a particular infection (Read et al, 2000). Examination of data regarding cause of death in patients who died in the present study showed that numbers were too small to support or refute this (data not shown).

In summary, the data reported in this study should be interpreted with caution, as patient numbers are limited, transplant strategies were heterogeneous and the findings differ from our original hypothesis that IL-1 polymorphism would influence risk of GVHD. Despite these limitations, the data support a larger investigation of donor IL-1 genotype and the outcome of unrelated donor BMT. Donor genotype appears to be more important than recipient genotype and, if these findings are confirmed, IL-1 genotype could potentially be included in donor selection strategies.

Acknowledgments

This work was supported in part by a grant from the Children's Cancer Research Fund and the Irvine McQuarrie Research Scholar Award.

Ancillary