Specific polymorphisms of cytokine genes are associated with different risks to develop single-system or multi-system childhood Langerhans cell histiocytosis

Authors


Dr Maurizio Aricò, U.O. Onco-Ematologia Pediatrica, Ospedale dei Bambini ‘G. Di Cristina’, Via Benedettini 1, 90134 Palermo, Italy.
E-mail: arico@ospedalecivicopa.org

Summary

Cytokines and chemokines determine mobilisation of Langerhans cells and their dysregulation is implicated in the pathogenesis of Langerhans cell histiocytosis (LCH). Twenty point mutations of 12 different cytokine genes were studied in 41 Italian children, 15 with single-system (SS) and 26 with multi-system disease. The allele and genotype distributions of interleukin-4 (IL-4) and interferon-γ (IFNγ) were significantly different in patients vs. 140 controls (P = 0.007, and P = 0.018). Older children with single-system disease shared the ‘anti-inflammatory profile’ determined by the intermediate producer genotype IFNγ +874A/T (P = 0.029) and the high-producer genotypes IL-4 –590C/T and T/T (P = 0.029). Our findings suggest that specific cytokine gene variants affect susceptibility to LCH and its clinical heterogeneity.

Langerhans cell histiocytosis (LCH) is a rare disorder (3.5–7 cases/million/year) that affects all age groups (Aricòet al, 2003) with dysregulated growth, activity and trafficking of Langerhans cells (LC). The clinical manifestations and course are highly variable, ranging from a solitary bone lesion to a disseminated disease with multiple-organ involvement and high mortality rate, despite aggressive treatment (Aricò & Egeler, 1998).

Familial clustering together with the propensity to develop tumour and the association with specific immune response genes, support the hypothesis that LCH is genetically determined (Aricò & Danesino, 2001; Beverley et al, 2005).

Langerhans cells reside in the epidermis, serving as sentinels of the immune system and are characterised by the CD1a-antigen (Cresswell et al, 2005). The ability of LC to migrate from the epidermis to regional lymph nodes is of pivotal importance for the induction of cutaneous immune response. There is increasing evidence that both cytokines and chemokines determine the mobilisation of LC and that dysregulation in this pathway is implicated in the pathogenesis of LCH (Annels et al, 2003).

We investigated whether functional variants of 12 different cytokine genes (20 point mutations) could be implicated in the predisposition to LCH and whether the clinical heterogeneity parallels a genetic heterogeneity.

Patients and methods

Forty-one consecutive Italian patients, all with biopsy-proven LCH, diagnosed according to current criteria established by the Histiocyte Society (Broadbent et al, 1989) were studied. The pattern of disease manifestations was classified into single-system (SS) and multi-system (MS).

The patients, 25 males and 16 females, were diagnosed at a median age of 6.1 years (range, 2 months to 17 years); 15 had SS disease (median age, 7.5 years) and 26 had MS disease (median age, 5.3 years); 16 were diagnosed with LCH when <2 years of age, 25 were >2 years old at diagnosis.

One hundred and forty healthy Italian subjects were tested as controls. Typing was performed at the genomic level on DNA extracted from peripheral blood samples using a salting out method.

Special kits were used, set up at the Heidelberg University (Brown Deer, WI, USA). The polymerase chain reaction single-stranded polymorphism technique was employed, as described elsewhere (Uboldi de Capei et al, 2003). The following biallelic single nucleotide polymorphisms (SNPs) were defined: interleukin (IL)-1α−889C/T, IL-1β (−511C/T; +3962T/C), IL-1R pst1 1970C/T, IL-1RA mspa111100T/C, IL-4Ra +1902G/A, IL-12 −1188C/A, +874A/T, tumour necrosis factor-α (−308G/A; −238G/A), IL-2 (−330T/G; +160G/T), IL-4 (−1098T/G; −590T/C; −33 T/C), IL-6 (−174G/C; nt565G/A), IL-10 (−1082G/A; −819C/T;-592C/A).

Univariate statistical analysis was performed using the EPI5 statistical package. Chi square analysis and Fisher's exact tests were used to compare frequencies of cytokine alleles, haplotypes and genotypes in patients versus the controls. A logistic regression model was further applied.

Results

Table I summarises all the significant associations found at the allelic level (two alleles per patient were studied). The only allele with a significantly different frequency in the entire group of patients was IL-4 −590T.

Table I.  Cytokine alleles with statistically significant deviations from the controls.
CytokineAlleleControls (n = 280) n (%)LCH (n = 82) n (%)SS (n = 32) n (%)MS (n = 50) n (%)Age ≤ 2 years (n = 32) n (%)Age>2 years (n = 50) n (%)
  1. IL, interleukin; SS, single-system; LCH, Langerhans cell histiocytosis.

IL-4 −590C248 (88.6)62 (75.6)24 (75.0)38 (76.0)25 (78.1)37 (74.0)
T32 (11.4)20 (24.4)8 (25.0)12 (24.0)7 (21.9)13 (26.0)
  P = 0.003P = 0.029P = 0.016P = nsP = 0.056
IL-4 −33C242 (86.4)66 (80.5)23 (71.9)43 (86.0)27 (84.4)39 (78.0)
T38 (13.6)16 (19.5)9 (28.1)7 (14.0)5 (15.6)11 (22.0)
  P = nsP = 0.029P = nsP = nsP = ns
IL-1β +3962C208 (74.3)56 (68.3)19 (59.4)37 (74.0)26 (81.3)30 (60.0)
T72 (25.7)26 (31.7)13 (40.613 (26.0)6 (18.7)20 (40.0)
  P = nsP = nsP = nsP = nsP = 0.037
IL-1R +970C187 (66.8)47 (57.3)22 (68.8)25 (50.0)20 (62.5)27 (54.0)
T93 (33.2)35 (42.7)10 (31.2)25 (50.0)12 (37.5)23 (46.0)
  P = nsP = nsP = 0.022P = nsP = ns

Two additional alleles had a different distribution only in specific subgroups: IL-1β +3962T (exon 5) was more frequent among patients >2 years of age and the allele IL-1R +970T (5′UTR untranslated region), was more frequent among patients with MS disease.

Table II summarises all the significant associations found at the genotypic level. Functionally, the patients may be considered high producers of IL-4 (as per IL-4 –590C/T and T/T genotypes) and intermediate producers of interferon γ (IFNγ) (as per IFNγ +874A/T genotype). The first profile was more evident in the SS and >2 years subgroups, while the latter was more evident in the SS and ≤ 2 years.

Table II.  Cytokine genotypes with statistically significant deviations from the controls.
CytokineGenotypeControls (n = 140) n (%)LCH (n = 41) n (%)SS (n = 16) n (%)MS (n = 25) n (%)Age ≤ 2 years (n = 16) n (%)Age >2 years (n = 25) n (%)
  1. IL, interleukin; SS, single-system; IFN, interferon; LCH, Langerhans cell histiocytosis.

IL-4 −590CC110 (78.6)22 (53.7)8 (50.0)14 (56.0)9 (56.3)13 (52.0)
CT28 (20.0)18 (43.9)8 (50.0)10 (40.0)7 (43.7)11 (44.0)
TT2 (1.4)1 (2.4)0 (0)1 (4.0)0 (0)1 (4.0)
  P = 0.007P = 0.025P = nsP = nsP = 0.019
IFNγ +874AA42 (30.0)6 (15.0)0 (0)6 (25.0)2 (12.5)4 (17.0)
AT66 (47.1)29 (72.5)12 (75.0)17 (70.8)13 (81.3)16 (66.0)
TT32 (2.9)5 (12.5)4 (25.0)1 (4.2)1 (6.2)4 (17.0)
  P = 0.018P = 0.029P = 0.049P = 0.035P = ns
IL-1β +3962CC76 (54.3)21 (51.2)6 (37.5)15 (60.0)11 (68.7)10 (40.0)
CT56 (40.0)14 (34.2)7 (43.8)7 (28.0)4 (25.1)10 (40.0)
TT8 (5.7)6 (14.6)3 (18.7)3 (12.0)1 (6.2)5 (20.0)
  P = nsP = nsP = nsP = nsP = 0.042
IL-IR +970CC60 (42.9)12 (29.2)8 (50.0)4 (16.0)5 (31.3)7 (28.0)
CT67 (47.8)23 (56.2)6 (37.5)17 (68.0)10 (62.5)13 (52.0)
TT13 (9.3)6 (14.6)2 (12.5)4 (16.0)1 (6.2)5 (20.0)
  P = nsP = nsP = 0.038P = nsP = ns

Among the cytokine genotypes, logistic regression defined a hierarchy of risk to the disease: IL-4 –33 T/T was the most implicated [odds ratio (OR) = 5.63, P = 0.035, 95% confidence interval (CI) 1.13–28.2], followed by the IFNγ +874A/T (OR = 3.00, P = 0.006, 95% CI 1.38–6.58).

Discussion

This study explored the hypothesis that specific polymorphisms of genes involved in the production of cytokines may be significantly associated with LCH or with its clinical variants.

Our findings support the concept that peculiar cytokine genes polymorphisms contribute to an intrinsic genetic propensity to develop LCH and its manifestations. Although the number of patients investigated remains limited (LCH is a very rare disease) the main outcome of this study was that patients with MS or SS have significantly different genetic characteristics. This finding may be considered to be in keeping with the observation that SS and MS disease are really different clinically. In fact, while SS is a benign disease, with no effect on survival, MS LCH patients have 20% risk of mortality (Aricò & Egeler, 1998; Beverley et al, 2005; Jubran et al, 2005).

Thus, the concept that different patterns of disease expression may reside primarily on the patient's intrinsic characteristics lends further support to genetic investigations. Two recent studies have investigated the human leucocyte antigen (HLA) system in patients with LCH. McClain et al (2003) showed that patients presenting with single bone disease had a particularly high frequency of the HLA-DR4 allele and, in this group, every Caucasian patient had either the HLA-Cw7 or DR4 allele. In another independent study of Nordic patients, those with SS disease more often had the allele HLA-DRB1*03 (Bernstrand et al, 2003). Altogether these data suggest an immunogenetic heterogeneity.

The two major findings of this study are: (i) the predominance of IL-4 high producer genotypes (IL-4 −590T/C and IL-4–33T/T) and IFNγ +874A/T intermediate secretor genotype and (ii) the prominent association between these anti-inflammatory profiles and LCH with lower clinical aggressiveness and later onset.

Interleukin-4 downregulates IFNγ production from T-helper (Th)1 lymphocytes and, together with IL-10, exerts a protective role in stress conditions. In patients with rheumatoid arthritis for instance, the phenotype high producer of IL-4 seems to protect from severe joint destruction; among patients undergoing heart transplant, those who received the graft from a donor high-producer of IL-4 have a lower incidence of rejection episodes (Bijlsma et al, 2002). Our results of a positive association of the genotypes resulting in high IL-4 gene transcription (namely IL-4 −590T/C and IL-4 −33T/T) with the mild form of LCH are in keeping with these observations. A possible explanation is that the IL-4 high producer genotype could favour a Th cell type 2-mediated antibody response to pathogenic stimulation able to circumscribe the disease, thus avoiding a rapid spread.

In conclusion, this study substantiates the hypothesis that cytokine genetic variants, identified by individual SNPs or specific genotypes of IL-4 and IFNγ regulatory pathways, may confer susceptibility to – or protection from – multi-system LCH, the most severe form of a disease that may be absolutely benign, which is associated with a 20% fatality rate.

Acknowledgments

The authors are grateful to Associazione Italiana Ricerca Istiocitosi (AIRI) for generous support and to all the participants in the CSS AIEOP Istiocitosi for their cooperation.

Ancillary