JEADV
DOI: 10.1111/jdv.13083
SHORT REPORT
Mycosis fungoides: association of KIR ligands and HLA-DQB1*05 with bad prognosis of the disease V. Brazzelli,1,* N. Rivetti,1 C. Badulli,2 A. Carugno,1 R. Cananzi,1 A. De Silvestri,3 M. Martinetti,2 G. Borroni1 1
Department of Clinical-Surgical, Diagnostic and Pediatric Science, Institute of Dermatology, Pavia, Italy Immunogenetic Laboratory, Immunohematology Service and Trasfusional Medicine, Pavia, Italy 3 Biometry and Statistics, Fondazione IRCCS Policlinico San Matteo, University of Pavia, Pavia, Italy *Correspondence: V. Brazzelli. E-mails: [email protected]; [email protected] 2
Abstract Background Mycosis fungoides (MF) is the most common cutaneous T-cell lymphoma. We previously reported that the prognosis of MF patients is not only related on clinical variables but it is also associated with peculiar HLA alleles. Until today, the association of HLA ligands for KIR with the prognosis of the disease has not yet been analysed. Objective We investigated the frequency of HLA ligands for killer cell Immunoglobulin-like receptors (KIRs) in MF patients, evaluating if the presence of particular HLA alleles that are ligands for KIR may have prognostic value. Methods The study includes 46 Caucasian MF patients that, between 1993 and 1997, underwent HLA genomic typing. All patients were diagnosed and followed up from 1977 to 2012 (mean follow-up of 11 years). Results MF patients have been divided into two groups (long survivors and dead patients). We noticed that the HLABw6/Bw6 specificity increased among the group of seven dead patients compared to the group of 39 long survivors (71.4% vs. 41.0%, P = ns, OR = 3.59), while in the long survivors group the HLA- Bw4/Bw4 specificity increased when compared to dead patients (23.0% vs. 0%, P = ns). Moreover, we observed that six of the seven dead patients had HLA-DQB1*05; the phenotypic frequency of this HLA allele, in dead and long survivors patients, was 85.7% and 23.0% respectively (P = 0.004; OR = 20). Conclusion Our observations suggest that the presence of the HLA-DQB1*05 alleles characterizes the patients with the poorest prognosis in MF. In addition, absence of the KIR-ligand epitope HLA-B Bw4 showed a trend of being more prominent in MF patients with the poorest prognosis. Received: 31 October 2014; Accepted: 27 January 2015
Conflicts of interest None disclosed.
Funding sources The research has been partially financially supported by the project number 556 2009–2012 of Current Research of IRCCS, Fondazione Policlinico San Matteo, Pavia.
Introduction Mycosis fungoides (MF) is the most common type of cutaneous T-cell lymphoma, with an annual incidence of 4.1 case per million.1 A retrospective study on 1422 MF patients provided a quantification of risk factors for disease progression,2 even if considering only clinical parameters. MF is characterized by clonal expansion of T-helper lymphocytes CD4+ in the skin, in fact the pioneer studies focused on HLA class II polymorphism. However, skin tumour infiltrates contain a considerable proportion of CD8+ and natural killer (NK) cells that are under the control of HLA class I molecules.3
JEADV 2015
In our recent study about the correlation between HLA polymorphisms and prognosis in MF,4 HLA-A*24 allele was found to be involved in the susceptibility of MF, while HLA-B*35 allele was typical of dead patients. Both these alleles belong to HLA class I. This evidence led us to consider another scenario in MF. We wondered whether the HLA class I heterodimers could be involved in the disease or in its fate. Recent literature shows that HLA class I molecules are ligands of NK cells inhibitory receptors.5,6 To prevent the killing of self cells, the NK lymphocytes express on their surface inhibitory receptors. Pathogen infected and tumour transformed cells, which miss surface HLA proteins that are ligands for killer cell
© 2015 European Academy of Dermatology and Venereology
Brazzelli et al.
2
Immunoglobulin-like receptors (KIRs), are the target of NK cytolytic activity. There are two main different inhibitory receptors for HLA class I proteins: the prototypic KIRs with ligands HLA-A3,-11,-23,-24,-32, HLA-B (Bw4) and HLA-C class I antigens, and CD94/NKG2A receptor, which recognizes HLA-E antigens.5,6 Capittini et al.7 showed, in healthy donors, the importance of HLA-A alleles that are ligands for KIR in balancing NK cells activity, demonstrating that they play a compensatory role when KIR-ligand epitope HLA-B Bw4 is absent. To evaluate if this condition is maintained in MF, we studied the frequency of HLA alleles that are ligands for KIR in a group of patients, followed up for 11 years. We observed that MF patients, and in particular those with the poorest prognoses, have a pattern of HLA-KIR ligands different from that of healthy controls.
All subjects were informed about the procedures and objectives of the study and blood samples were obtained only after collection of informed consent. The Declaration of Helsinki protocols were followed. The study was approved by the Ethics Committee of the Foundation IRCCS Policlinico San Matteo, Pavia. DNA extraction and HLA class I and class II molecular typing
DNA extraction and HLA class I and class II molecular typing techniques are the same as described in a recent work published by our group.4 Statistical analysis
Materials and methods
The distribution of HLA-A,-B,-C alleles/ligands for KIR were analysed using the chi-squared test and Fisher test. All calculations were made by Stata 12 (StataCorp, College Station, TX, USA).
Population study
Results
The study includes 46 Caucasian patients already included in a recent work published by our group.4 Patients were divided into two groups: the first one composed of 39 patients, defined as “long survivors” (31 have shown no clinical recurrence of disease after therapy; eight still have MF in patches resistant to skin directed therapy, or relapsing few months after the clinical remission); the second one composed of seven patients who died for the clinical worsening of the disease.4 All patients were in stage IA or IB when they entered the study. Differently from patients of the first group (“long survivors”), who maintained the same stage over the follow-up time, the patients of the second group progressed to more advanced stages (stages IIB and IVA), and they died because of a clinical evolution of the disease, with the development of nodules and tumours, ulceration of lesions and lymph node involvement without visceral involvement. Most of them died because of sepsis. As regards the therapeutic regiments, “long survivors” patients received PUVA therapy as first-line therapy. Nine patients received narrowband UVB therapy because of comorbidity which contraindicated PUVA therapy or because of the high UVA cumulative dosage. Six patients also received RePUVA therapy because PUVA therapy alone was not sufficient to determine a good control of the disease. The second group of patients, who have died, were initially treated with narrowband UVB or PUVA therapy or Re-PUVA therapy but, for the extension and the poor control of the disease, they were also treated with interferon-a, radiotherapy and chemotherapy, with poor response and death due to clinical worsening of the disease. The control group consisted in 6364 healthy donors, ethnically matched with the patients.
JEADV 2015
HLA class I KIR ligands
The frequency of HLA-A ligands for KIR3DL2 (HLA-A3 and A11), was lower in patients than in controls (19.57% vs. 31.79%, P = 0.03). On the contrary, the frequency of HLA-A ligands for KIR3DL1 (HLA-A23, -24, -32 and HLA-Bw4 ligand motif alleles) was higher in patients compared to controls (41.3% vs. 35.92%, P = ns). HLA-C alleles were grouped according to their KIR ligand motif.5 No statistically significant difference was observed between MF and controls (C1: 45.6% vs. 49.4%, P = ns; C2: 54.4% vs. 50.6%, P = ns). HLA-KIR ligand-receptor pairs
HLA-A*03 and *A11 are the ligands for KIR3DL2 NK receptor, and HLA-A*23, A*24 and A*32 are the ligands for KIR3DL1 NK receptor. Both MF patients and controls were subdivided in three groups according to the Bw4 ligand motif (Bw4/Bw4, Bw4/Bw6 and Bw6/Bw6). The frequencies of HLA-A ligands for KIR3DL2 and KIR3DL1 receptors, with and without the HLA-B Bw4 ligand motif, are reported in Table 1. In Table 2 the distribution of HLA-C1 and C2 ligands for KIR in the three HLA-B groups are reported. HLA alleles and MF prognosis
Subdividing patients into two groups, “long survivors” and dead patients, we noticed that five of the seven patients who died, lacked the HLA-Bw4 KIR ligand motif being Bw6/Bw6 homozygous (71.4% in dead patients vs. 41.0% in long survivors, P = ns; OR = 3.59).
© 2015 European Academy of Dermatology and Venereology
Mycosis fungoides: HLA ligands for KIR and prognosis
Table 1 Distribution of HLA-A ligands for KIR3DL2 and KIR3DL1 in the three HLA-B cross reactive groups (Bw4/Bw4, Bw4/Bw6, Bw6/Bw6) in MF patients and in controls Bw4/Bw4
Bw4/Bw6
Bw6/Bw6
KIR3DL2 NK receptor ligands: HLA-A3 and A11 MF Patients*
11.11%
Long survival patients
11.11%
7.14%
31.25%
0.00%
50.00%
20.00%
23.13%
30.20%
39.09%
Dead patients Controls*
12.5%
28.57%
KIR3DL1 NK receptor ligands: HLA-A 23,24,32 MF Patients†
44.44%
37.5%
42.86%
Long survival patients
44.44%
42.86%
43.75%
0.00%
0.00%
40.00%
32.34%
34.95%
39.13%
Dead patients Controls†
*As regards the distribution of HLA-A ligands for KIR3DL2, the two samples showed the same ascending trend although, in MF patients, the frequencies of these ligands were always lower than in controls. †As regards the distribution of HLA-A ligands for KIR3DL1, the two samples did not show the same trend. In fact, while in controls the frequency of HLA-A ligands for KIR3DL1 increased from Bw4/Bw4 to Bw6/Bw6, in MF it remained quite stable in the three subgroups although always higher than in controls.
Table 2 Distribution of HLA-C1 and C2 ligands for KIR in the three HLA-B cross reactive groups (Bw4/Bw4, Bw4/Bw6, Bw6/Bw6) in MF patients and in controls Bw4/Bw4
Bw4/Bw6
Bw6/Bw6
Distribution of HLA-C1 ligands for KIR MF Patients*
88.89%
75.00%
52.38%
Long survival patients
88.89%
71.43%
43.75%
0.00%
100.00%
54.76%
65.04%
73.40%
Dead patients Controls*
80%
Distribution of HLA-C2 ligands for KIR MF Patients†
77.78%
68.75%
80.95%
Long survival patients
77.78%
64.29%
93.75%
0.00%
100.00%
40.00%
72.92%
67.80%
63.84%
Dead patients Controls†
*The frequency of HLA-C1 ligands for KIR decreased progressively in patients moving from Bw4/Bw4 to Bw6/Bw6 groups while, in controls, it shows an opposite trend. The percentage of people having the functional Bw4Bw4-C1C1 block was significantly higher in patients than in controls: 88.89% vs. 54.76%; P = 0.04; OR 6.61. Consequently, the frequency of carriers of the Bw6Bw6-C1C1 pair was significantly reduced in MF than in controls: 52.38% vs. 73.40% (P = 0.031; OR 0.4). †Considering the frequency of HLA-C2 KIR ligand-receptor pairs, we noticed a progressive decrease from Bw4/Bw4 to Bw6/Bw6 groups in controls. An opposite trend was observed in MF where an excess of Bw6Bw6-C2C2 pair was evident although statistically not significant: 80.95% vs. 63.84% (P = n.s.; OR = 2.41).
We noticed that patients with the poorest prognosis carried the HLA-DQB1*05, when compared with long survivors (85.7% vs. 23.0%; P = 0.004; OR = 20).
JEADV 2015
3
Discussion In Literature several works8,9 investigated the correlation between clinical variables and long term outcome of MF patients. Since the 1980s a growing number of studies10–12 analysed the possible link between HLA genes and susceptibility to MF, but even today the data are conflicting. Many authors reported that specific HLA Class I ligands/KIR combination contribute to the development of diseases such as diabetes and skin diseases,13,14 but the impact of HLA alleles/ ligands for KIR has never been studied in MF. HLA class I affects the activity of NK cells. The regulation of NK-mediated cytolysis could be modified by the combination of KIR receptors and HLA ligands.15 For instance, people with HLA-Bw4 allele and both group 1 and group 2 HLA-C alleles have three inhibitory KIR/HLA interactions, whereas an individual without HLA-Bw4 alleles and homozygous at the HLA-C locus has only one inhibitory KIR/HLA interaction. Starting from the hypothesis of Capittini et al.7 who showed in healthy people that, in balancing NK cells activity, HLA-A ligands for KIR play a compensatory role when HLA-B Bw4 KIR-ligand epitope is absent, we wondered whether this condition is also represented in MF patients. One of the HLA class I alleles (ligands for KIR) that regulate antitumour NK toxicity interacting with KIRs is HLA-A24. HLA-A*24 allele frequency was higher, in a statistically significant manner, in MF than in controls, as previously reported in our study.4 HLA-A24 molecule, together with HLA-A23 and HLA-A32 proteins, is a KIR3DL1 ligand. We detected the presence of HLA-A ligands for KIR3DL1 in the 41.3% of MF patients but their frequency did not increase from HLA-B Bw4/Bw4 to the HLA-B Bw4/Bw6 and the HLA-B Bw6/Bw6 groups as it is observed in the control group. We also analysed the HLA ligands for KIR3DL2. As expected, their frequency increased moving from Bw4/Bw4 to Bw6/Bw6 groups. In MF we observed that, when the frequency of HLA-B Bw4 ligands for KIR decreased, only the frequency of HLA-A ligands for KIR3DL2 increased in a functional compensatory way even if only the 19.57% of the patients carried these alleles, compared to the 31.79% of the controls. Analysing HLA-C ligands, we noticed an opposite trend between MF and controls. In particular, moving from Bw4/Bw4 to Bw6/Bw6 groups, we observed in MF patients a decrease in the frequency of the C1 alleles and a stability of the C2. In controls there was an increase in the C1 alleles and a decrease in the C2 alleles. Our data show that MF patients have a pattern of HLA alleles/ligands for KIR different from the control group, in which HLA-A always plays a compensatory role when KIR-ligand epitope HLA-B Bw4 is absent. In fact, in MF not only the HLA-A ligands but also the HLA-C1 and -C2 ligands, the most powerful
© 2015 European Academy of Dermatology and Venereology
Brazzelli et al.
4
KIR inhibitors did not increase in frequency in the absence of HLA-B Bw4 ligand. Until today, the association of HLA alleles that are ligands for KIR (HLA ligands for KIR) with the prognosis of the disease has not yet been analysed. We evaluated if the absence or the presence of particular HLA ligands for KIR could be involved in the prognosis, taking advantage from 11 years of mean follow-up. Interestingly, subdividing all the patients in two groups (long survivors and dead patients), we noticed that the HLA-Bw6/Bw6 specificity increased among the group of seven dead patients compared to long survivors (71.4% vs. 41.0%, P = ns, OR = 3.59), while in the long survivors group, the HLA- Bw4/ Bw4 specificity increased when compared to dead patients (23.0% vs. 0.0%, P = ns). Regarding the class II alleles, among the patients, we observed that six of the seven dead patients had HLA-DQB1*05; the phenotypic frequency of this HLA allele, in dead and long survivors group of patients, was respectively 85.7% and 23.0% (P = 0.004; OR = 20). This observation partially agrees with a work of Jackow et al.11 in which the HLA-DQB1*05 allele was found to be significantly increased in patients with Sezary syndrome and MF. In conclusion, on the basis of these observations, we observed that the presence of the HLA-DQB1*05 allele distinguish the MF patients with the poorest prognosis. Moreover, even if the number of MF dead patients is small and the statistical significance is not reached (P = ns, OR = 3.59), also the absence of the KIRligand epitope HLA-B Bw4 could distinguish MF patients with poorest prognosis. This is a preliminary retrospective study with interesting and encouraging results. For this reason, we are already recruiting other MF patients to confirm our data.
Acknowledgements The authors are grateful to Geraldine Mc Crossan and Umberto Tosi for their helpful language revision of the draft.
JEADV 2015
References 1 Bradford PT, Devesa SS, Anderson WF, Toro JR. Cutaneous lymphoma incidence patterns in the United States: a population-based study of 3884 cases. Blood 2009; 113: 5064–5073. 2 Quaglino P, Pimpinelli N, Berti E et al. Mycosis fungoides: disease evolution of the “lion queen” revisited. G Ital Dermatol Venereol 2012; 147: 523–531. 3 Zinzani PL, Ferreri AJ, Cerroni L. Mycosis fungoides. Crit Rev Oncol Hematol 2008; 65: 172–182. 4 Brazzelli V, Rivetti N, Badulli C et al. Immunogenetic factors in mycosis fungoides: can the HLA system influence the susceptibility and prognosis of the disease? Long-term follow-up study of 46 patients. J Eur Acad Dermatol Venereol 2014; 28: 1732–1737. 5 Bari R, Leung M, Turner VE et al. Molecular determinant-based typing of KIR alleles and KIR ligands. Clin Immunol 2011; 138: 274–281. 6 Thielens A, Vivier E, Romagne F. NK cell MHC class I specific receptors (KIR): from biology to clinical intervention. Curr Opin Immunol 2012; 24: 239–245. 7 Capittini C, Tinelli C, Guarene M et al. Possible KIR-driven genetic pressure on the genesis and maintenance of specific HLA-A, B haplotypes as functional genetic blocks. Genes Immun 2012; 13: 452–457. 8 Kim YH, Liu HL, Mraz-Gernhard S, Varghese A, Hoppe RT. Long-term outcome of 525 patients with mycosis fungoides and Sezary syndrome: clinical prognostic factors and risk for disease progression. Arch Dermatol 2003; 139: 857–866. 9 Benton EC, Crichton S, Talpur R et al. A cutaneous lymphoma international prognostic index (CLIPi) for mycosis fungoides and Sezary syndrome. Eur J Cancer 2013; 49: 2859–2868. 10 Safai B, Myskowski PL, Dupont B, Pollack MS. Association of HLA-DR5 with mycosis fungoides. J Invest Dermatol 1983; 80: 395–397. 11 Jackow CM, McHam JB, Friss A, Alvear J, Reveille JR, Duvic M. HLADR5 and HLADQB1*03 class II alleles are associated with cutaneous T-cell lymphoma. J Invest Dermatol 1996; 107: 373–376. 12 Hodak E, Lapidoth M, Kohn K et al. Mycosis fundoides: HLA class II associations among Ashkenazi and non-Ashkenazi Jewish patients. Br J Dermatol 2001; 145: 974–980. 13 van der Slik AR, Koeleman BP, Verduijn W, Bruining GJ, Roep BO, Giphart MJ. KIR in type 1 diabetes: disparate distribution of activating and inhibitory natural killer cell receptors in patients versus HLA-matched control subjects. Diabetes 2003; 52: 2639–2642. 14 Matusiak Ł, Białynicki-Birula R, Szepietowski JC. Emerging role for the killer-cell immunoglobulin-like receptors genotype, in the susceptibility of skin diseases. J Dermatol Sci 2013; 71: 3–11. 15 Biassoni R, Ugolotti E, De Maria A. NK cell receptors and their interactions with MHC. Curr Pharm Des 2009; 15: 3301–3310.
© 2015 European Academy of Dermatology and Venereology
DOI: 10.1111/jdv.13083
SHORT REPORT
Mycosis fungoides: association of KIR ligands and HLA-DQB1*05 with bad prognosis of the disease V. Brazzelli,1,* N. Rivetti,1 C. Badulli,2 A. Carugno,1 R. Cananzi,1 A. De Silvestri,3 M. Martinetti,2 G. Borroni1 1
Department of Clinical-Surgical, Diagnostic and Pediatric Science, Institute of Dermatology, Pavia, Italy Immunogenetic Laboratory, Immunohematology Service and Trasfusional Medicine, Pavia, Italy 3 Biometry and Statistics, Fondazione IRCCS Policlinico San Matteo, University of Pavia, Pavia, Italy *Correspondence: V. Brazzelli. E-mails: [email protected]; [email protected] 2
Abstract Background Mycosis fungoides (MF) is the most common cutaneous T-cell lymphoma. We previously reported that the prognosis of MF patients is not only related on clinical variables but it is also associated with peculiar HLA alleles. Until today, the association of HLA ligands for KIR with the prognosis of the disease has not yet been analysed. Objective We investigated the frequency of HLA ligands for killer cell Immunoglobulin-like receptors (KIRs) in MF patients, evaluating if the presence of particular HLA alleles that are ligands for KIR may have prognostic value. Methods The study includes 46 Caucasian MF patients that, between 1993 and 1997, underwent HLA genomic typing. All patients were diagnosed and followed up from 1977 to 2012 (mean follow-up of 11 years). Results MF patients have been divided into two groups (long survivors and dead patients). We noticed that the HLABw6/Bw6 specificity increased among the group of seven dead patients compared to the group of 39 long survivors (71.4% vs. 41.0%, P = ns, OR = 3.59), while in the long survivors group the HLA- Bw4/Bw4 specificity increased when compared to dead patients (23.0% vs. 0%, P = ns). Moreover, we observed that six of the seven dead patients had HLA-DQB1*05; the phenotypic frequency of this HLA allele, in dead and long survivors patients, was 85.7% and 23.0% respectively (P = 0.004; OR = 20). Conclusion Our observations suggest that the presence of the HLA-DQB1*05 alleles characterizes the patients with the poorest prognosis in MF. In addition, absence of the KIR-ligand epitope HLA-B Bw4 showed a trend of being more prominent in MF patients with the poorest prognosis. Received: 31 October 2014; Accepted: 27 January 2015
Conflicts of interest None disclosed.
Funding sources The research has been partially financially supported by the project number 556 2009–2012 of Current Research of IRCCS, Fondazione Policlinico San Matteo, Pavia.
Introduction Mycosis fungoides (MF) is the most common type of cutaneous T-cell lymphoma, with an annual incidence of 4.1 case per million.1 A retrospective study on 1422 MF patients provided a quantification of risk factors for disease progression,2 even if considering only clinical parameters. MF is characterized by clonal expansion of T-helper lymphocytes CD4+ in the skin, in fact the pioneer studies focused on HLA class II polymorphism. However, skin tumour infiltrates contain a considerable proportion of CD8+ and natural killer (NK) cells that are under the control of HLA class I molecules.3
JEADV 2015
In our recent study about the correlation between HLA polymorphisms and prognosis in MF,4 HLA-A*24 allele was found to be involved in the susceptibility of MF, while HLA-B*35 allele was typical of dead patients. Both these alleles belong to HLA class I. This evidence led us to consider another scenario in MF. We wondered whether the HLA class I heterodimers could be involved in the disease or in its fate. Recent literature shows that HLA class I molecules are ligands of NK cells inhibitory receptors.5,6 To prevent the killing of self cells, the NK lymphocytes express on their surface inhibitory receptors. Pathogen infected and tumour transformed cells, which miss surface HLA proteins that are ligands for killer cell
© 2015 European Academy of Dermatology and Venereology
Brazzelli et al.
2
Immunoglobulin-like receptors (KIRs), are the target of NK cytolytic activity. There are two main different inhibitory receptors for HLA class I proteins: the prototypic KIRs with ligands HLA-A3,-11,-23,-24,-32, HLA-B (Bw4) and HLA-C class I antigens, and CD94/NKG2A receptor, which recognizes HLA-E antigens.5,6 Capittini et al.7 showed, in healthy donors, the importance of HLA-A alleles that are ligands for KIR in balancing NK cells activity, demonstrating that they play a compensatory role when KIR-ligand epitope HLA-B Bw4 is absent. To evaluate if this condition is maintained in MF, we studied the frequency of HLA alleles that are ligands for KIR in a group of patients, followed up for 11 years. We observed that MF patients, and in particular those with the poorest prognoses, have a pattern of HLA-KIR ligands different from that of healthy controls.
All subjects were informed about the procedures and objectives of the study and blood samples were obtained only after collection of informed consent. The Declaration of Helsinki protocols were followed. The study was approved by the Ethics Committee of the Foundation IRCCS Policlinico San Matteo, Pavia. DNA extraction and HLA class I and class II molecular typing
DNA extraction and HLA class I and class II molecular typing techniques are the same as described in a recent work published by our group.4 Statistical analysis
Materials and methods
The distribution of HLA-A,-B,-C alleles/ligands for KIR were analysed using the chi-squared test and Fisher test. All calculations were made by Stata 12 (StataCorp, College Station, TX, USA).
Population study
Results
The study includes 46 Caucasian patients already included in a recent work published by our group.4 Patients were divided into two groups: the first one composed of 39 patients, defined as “long survivors” (31 have shown no clinical recurrence of disease after therapy; eight still have MF in patches resistant to skin directed therapy, or relapsing few months after the clinical remission); the second one composed of seven patients who died for the clinical worsening of the disease.4 All patients were in stage IA or IB when they entered the study. Differently from patients of the first group (“long survivors”), who maintained the same stage over the follow-up time, the patients of the second group progressed to more advanced stages (stages IIB and IVA), and they died because of a clinical evolution of the disease, with the development of nodules and tumours, ulceration of lesions and lymph node involvement without visceral involvement. Most of them died because of sepsis. As regards the therapeutic regiments, “long survivors” patients received PUVA therapy as first-line therapy. Nine patients received narrowband UVB therapy because of comorbidity which contraindicated PUVA therapy or because of the high UVA cumulative dosage. Six patients also received RePUVA therapy because PUVA therapy alone was not sufficient to determine a good control of the disease. The second group of patients, who have died, were initially treated with narrowband UVB or PUVA therapy or Re-PUVA therapy but, for the extension and the poor control of the disease, they were also treated with interferon-a, radiotherapy and chemotherapy, with poor response and death due to clinical worsening of the disease. The control group consisted in 6364 healthy donors, ethnically matched with the patients.
JEADV 2015
HLA class I KIR ligands
The frequency of HLA-A ligands for KIR3DL2 (HLA-A3 and A11), was lower in patients than in controls (19.57% vs. 31.79%, P = 0.03). On the contrary, the frequency of HLA-A ligands for KIR3DL1 (HLA-A23, -24, -32 and HLA-Bw4 ligand motif alleles) was higher in patients compared to controls (41.3% vs. 35.92%, P = ns). HLA-C alleles were grouped according to their KIR ligand motif.5 No statistically significant difference was observed between MF and controls (C1: 45.6% vs. 49.4%, P = ns; C2: 54.4% vs. 50.6%, P = ns). HLA-KIR ligand-receptor pairs
HLA-A*03 and *A11 are the ligands for KIR3DL2 NK receptor, and HLA-A*23, A*24 and A*32 are the ligands for KIR3DL1 NK receptor. Both MF patients and controls were subdivided in three groups according to the Bw4 ligand motif (Bw4/Bw4, Bw4/Bw6 and Bw6/Bw6). The frequencies of HLA-A ligands for KIR3DL2 and KIR3DL1 receptors, with and without the HLA-B Bw4 ligand motif, are reported in Table 1. In Table 2 the distribution of HLA-C1 and C2 ligands for KIR in the three HLA-B groups are reported. HLA alleles and MF prognosis
Subdividing patients into two groups, “long survivors” and dead patients, we noticed that five of the seven patients who died, lacked the HLA-Bw4 KIR ligand motif being Bw6/Bw6 homozygous (71.4% in dead patients vs. 41.0% in long survivors, P = ns; OR = 3.59).
© 2015 European Academy of Dermatology and Venereology
Mycosis fungoides: HLA ligands for KIR and prognosis
Table 1 Distribution of HLA-A ligands for KIR3DL2 and KIR3DL1 in the three HLA-B cross reactive groups (Bw4/Bw4, Bw4/Bw6, Bw6/Bw6) in MF patients and in controls Bw4/Bw4
Bw4/Bw6
Bw6/Bw6
KIR3DL2 NK receptor ligands: HLA-A3 and A11 MF Patients*
11.11%
Long survival patients
11.11%
7.14%
31.25%
0.00%
50.00%
20.00%
23.13%
30.20%
39.09%
Dead patients Controls*
12.5%
28.57%
KIR3DL1 NK receptor ligands: HLA-A 23,24,32 MF Patients†
44.44%
37.5%
42.86%
Long survival patients
44.44%
42.86%
43.75%
0.00%
0.00%
40.00%
32.34%
34.95%
39.13%
Dead patients Controls†
*As regards the distribution of HLA-A ligands for KIR3DL2, the two samples showed the same ascending trend although, in MF patients, the frequencies of these ligands were always lower than in controls. †As regards the distribution of HLA-A ligands for KIR3DL1, the two samples did not show the same trend. In fact, while in controls the frequency of HLA-A ligands for KIR3DL1 increased from Bw4/Bw4 to Bw6/Bw6, in MF it remained quite stable in the three subgroups although always higher than in controls.
Table 2 Distribution of HLA-C1 and C2 ligands for KIR in the three HLA-B cross reactive groups (Bw4/Bw4, Bw4/Bw6, Bw6/Bw6) in MF patients and in controls Bw4/Bw4
Bw4/Bw6
Bw6/Bw6
Distribution of HLA-C1 ligands for KIR MF Patients*
88.89%
75.00%
52.38%
Long survival patients
88.89%
71.43%
43.75%
0.00%
100.00%
54.76%
65.04%
73.40%
Dead patients Controls*
80%
Distribution of HLA-C2 ligands for KIR MF Patients†
77.78%
68.75%
80.95%
Long survival patients
77.78%
64.29%
93.75%
0.00%
100.00%
40.00%
72.92%
67.80%
63.84%
Dead patients Controls†
*The frequency of HLA-C1 ligands for KIR decreased progressively in patients moving from Bw4/Bw4 to Bw6/Bw6 groups while, in controls, it shows an opposite trend. The percentage of people having the functional Bw4Bw4-C1C1 block was significantly higher in patients than in controls: 88.89% vs. 54.76%; P = 0.04; OR 6.61. Consequently, the frequency of carriers of the Bw6Bw6-C1C1 pair was significantly reduced in MF than in controls: 52.38% vs. 73.40% (P = 0.031; OR 0.4). †Considering the frequency of HLA-C2 KIR ligand-receptor pairs, we noticed a progressive decrease from Bw4/Bw4 to Bw6/Bw6 groups in controls. An opposite trend was observed in MF where an excess of Bw6Bw6-C2C2 pair was evident although statistically not significant: 80.95% vs. 63.84% (P = n.s.; OR = 2.41).
We noticed that patients with the poorest prognosis carried the HLA-DQB1*05, when compared with long survivors (85.7% vs. 23.0%; P = 0.004; OR = 20).
JEADV 2015
3
Discussion In Literature several works8,9 investigated the correlation between clinical variables and long term outcome of MF patients. Since the 1980s a growing number of studies10–12 analysed the possible link between HLA genes and susceptibility to MF, but even today the data are conflicting. Many authors reported that specific HLA Class I ligands/KIR combination contribute to the development of diseases such as diabetes and skin diseases,13,14 but the impact of HLA alleles/ ligands for KIR has never been studied in MF. HLA class I affects the activity of NK cells. The regulation of NK-mediated cytolysis could be modified by the combination of KIR receptors and HLA ligands.15 For instance, people with HLA-Bw4 allele and both group 1 and group 2 HLA-C alleles have three inhibitory KIR/HLA interactions, whereas an individual without HLA-Bw4 alleles and homozygous at the HLA-C locus has only one inhibitory KIR/HLA interaction. Starting from the hypothesis of Capittini et al.7 who showed in healthy people that, in balancing NK cells activity, HLA-A ligands for KIR play a compensatory role when HLA-B Bw4 KIR-ligand epitope is absent, we wondered whether this condition is also represented in MF patients. One of the HLA class I alleles (ligands for KIR) that regulate antitumour NK toxicity interacting with KIRs is HLA-A24. HLA-A*24 allele frequency was higher, in a statistically significant manner, in MF than in controls, as previously reported in our study.4 HLA-A24 molecule, together with HLA-A23 and HLA-A32 proteins, is a KIR3DL1 ligand. We detected the presence of HLA-A ligands for KIR3DL1 in the 41.3% of MF patients but their frequency did not increase from HLA-B Bw4/Bw4 to the HLA-B Bw4/Bw6 and the HLA-B Bw6/Bw6 groups as it is observed in the control group. We also analysed the HLA ligands for KIR3DL2. As expected, their frequency increased moving from Bw4/Bw4 to Bw6/Bw6 groups. In MF we observed that, when the frequency of HLA-B Bw4 ligands for KIR decreased, only the frequency of HLA-A ligands for KIR3DL2 increased in a functional compensatory way even if only the 19.57% of the patients carried these alleles, compared to the 31.79% of the controls. Analysing HLA-C ligands, we noticed an opposite trend between MF and controls. In particular, moving from Bw4/Bw4 to Bw6/Bw6 groups, we observed in MF patients a decrease in the frequency of the C1 alleles and a stability of the C2. In controls there was an increase in the C1 alleles and a decrease in the C2 alleles. Our data show that MF patients have a pattern of HLA alleles/ligands for KIR different from the control group, in which HLA-A always plays a compensatory role when KIR-ligand epitope HLA-B Bw4 is absent. In fact, in MF not only the HLA-A ligands but also the HLA-C1 and -C2 ligands, the most powerful
© 2015 European Academy of Dermatology and Venereology
Brazzelli et al.
4
KIR inhibitors did not increase in frequency in the absence of HLA-B Bw4 ligand. Until today, the association of HLA alleles that are ligands for KIR (HLA ligands for KIR) with the prognosis of the disease has not yet been analysed. We evaluated if the absence or the presence of particular HLA ligands for KIR could be involved in the prognosis, taking advantage from 11 years of mean follow-up. Interestingly, subdividing all the patients in two groups (long survivors and dead patients), we noticed that the HLA-Bw6/Bw6 specificity increased among the group of seven dead patients compared to long survivors (71.4% vs. 41.0%, P = ns, OR = 3.59), while in the long survivors group, the HLA- Bw4/ Bw4 specificity increased when compared to dead patients (23.0% vs. 0.0%, P = ns). Regarding the class II alleles, among the patients, we observed that six of the seven dead patients had HLA-DQB1*05; the phenotypic frequency of this HLA allele, in dead and long survivors group of patients, was respectively 85.7% and 23.0% (P = 0.004; OR = 20). This observation partially agrees with a work of Jackow et al.11 in which the HLA-DQB1*05 allele was found to be significantly increased in patients with Sezary syndrome and MF. In conclusion, on the basis of these observations, we observed that the presence of the HLA-DQB1*05 allele distinguish the MF patients with the poorest prognosis. Moreover, even if the number of MF dead patients is small and the statistical significance is not reached (P = ns, OR = 3.59), also the absence of the KIRligand epitope HLA-B Bw4 could distinguish MF patients with poorest prognosis. This is a preliminary retrospective study with interesting and encouraging results. For this reason, we are already recruiting other MF patients to confirm our data.
Acknowledgements The authors are grateful to Geraldine Mc Crossan and Umberto Tosi for their helpful language revision of the draft.
JEADV 2015
References 1 Bradford PT, Devesa SS, Anderson WF, Toro JR. Cutaneous lymphoma incidence patterns in the United States: a population-based study of 3884 cases. Blood 2009; 113: 5064–5073. 2 Quaglino P, Pimpinelli N, Berti E et al. Mycosis fungoides: disease evolution of the “lion queen” revisited. G Ital Dermatol Venereol 2012; 147: 523–531. 3 Zinzani PL, Ferreri AJ, Cerroni L. Mycosis fungoides. Crit Rev Oncol Hematol 2008; 65: 172–182. 4 Brazzelli V, Rivetti N, Badulli C et al. Immunogenetic factors in mycosis fungoides: can the HLA system influence the susceptibility and prognosis of the disease? Long-term follow-up study of 46 patients. J Eur Acad Dermatol Venereol 2014; 28: 1732–1737. 5 Bari R, Leung M, Turner VE et al. Molecular determinant-based typing of KIR alleles and KIR ligands. Clin Immunol 2011; 138: 274–281. 6 Thielens A, Vivier E, Romagne F. NK cell MHC class I specific receptors (KIR): from biology to clinical intervention. Curr Opin Immunol 2012; 24: 239–245. 7 Capittini C, Tinelli C, Guarene M et al. Possible KIR-driven genetic pressure on the genesis and maintenance of specific HLA-A, B haplotypes as functional genetic blocks. Genes Immun 2012; 13: 452–457. 8 Kim YH, Liu HL, Mraz-Gernhard S, Varghese A, Hoppe RT. Long-term outcome of 525 patients with mycosis fungoides and Sezary syndrome: clinical prognostic factors and risk for disease progression. Arch Dermatol 2003; 139: 857–866. 9 Benton EC, Crichton S, Talpur R et al. A cutaneous lymphoma international prognostic index (CLIPi) for mycosis fungoides and Sezary syndrome. Eur J Cancer 2013; 49: 2859–2868. 10 Safai B, Myskowski PL, Dupont B, Pollack MS. Association of HLA-DR5 with mycosis fungoides. J Invest Dermatol 1983; 80: 395–397. 11 Jackow CM, McHam JB, Friss A, Alvear J, Reveille JR, Duvic M. HLADR5 and HLADQB1*03 class II alleles are associated with cutaneous T-cell lymphoma. J Invest Dermatol 1996; 107: 373–376. 12 Hodak E, Lapidoth M, Kohn K et al. Mycosis fundoides: HLA class II associations among Ashkenazi and non-Ashkenazi Jewish patients. Br J Dermatol 2001; 145: 974–980. 13 van der Slik AR, Koeleman BP, Verduijn W, Bruining GJ, Roep BO, Giphart MJ. KIR in type 1 diabetes: disparate distribution of activating and inhibitory natural killer cell receptors in patients versus HLA-matched control subjects. Diabetes 2003; 52: 2639–2642. 14 Matusiak Ł, Białynicki-Birula R, Szepietowski JC. Emerging role for the killer-cell immunoglobulin-like receptors genotype, in the susceptibility of skin diseases. J Dermatol Sci 2013; 71: 3–11. 15 Biassoni R, Ugolotti E, De Maria A. NK cell receptors and their interactions with MHC. Curr Pharm Des 2009; 15: 3301–3310.
© 2015 European Academy of Dermatology and Venereology
