Investigating heredity in cutaneous T-cell lymphoma in a unique cohort of Danish twins

Cutaneous T-cell lymphomas (CTCLs) are uncommon but potentially fatal malignancies. The most prevalent clinical forms of CTCL are mycosis fungoides (MF) and the more aggressive leukemic variant, Sézary syndrome (SS). Although the etiology is largely unknown, some lines of evidence indicate that genetic factors and heredity play a role in CTCL. Thus, independent studies reported on strong linkage disequilibrium between MF/SS and specific HLAclass II allotypes in Caucasians and Ashkenazi, indicating the existence of a significant genetic predisposition to CTCL. Moreover, examples of CTCL occurring conjointly in monozygotic twin pairs have been reported further suggesting a possible relevance of genetic factors in the CTCL etiology. However, genetic studies in multi-generation families and larger cohorts of twins have never been conducted. Accordingly, we have taken advantage of the Danish Twin Register and other nationwide population-based registers, to study heritability, predisposition to infectious diseases, comorbidity, hospitalizations and mortality for a 30+-year period in a cohort of 42 twins with CTCL (case twins) and their 42 co-twins, 420 ageand sex-matched twin controls (case controls) and their 420 co-twin controls. The 42 twin pairs comprised 13 monozygotic and 27 dizygotic twin pairs, whereas two twin pairs were of unknown zygosity. Patient characteristics of the CTCL case twin cohort are shown in Supplementary Table 1. Female–male ratio was 1:1.8 and the average age at time of the CTCL diagnosis was 53 years (range: 5–85 years) showing that case twins with CTCL did not differ in terms of age at onset and female–male ratio from what has been described in other cohorts of Caucasian patients with CTCL. Likewise, the mortality was 2.7-fold increased in the case twins compared with the case– controls, adjusted HR 2.65 (95% confidence interval (CI) 1.66–4.24) (Table 1), further indicating that the CTCL cases represented a typical CTCL cohort. Surprisingly, all twin pairs were discordant for CTCL, that is, none of the co-twins were diagnosed with CTCL. Importantly, all co-twins were monitored from birth until 1 June 2015 or death and none were lost to follow-up. On average, the co-twins that were alive at the time of case diagnosis were monitored for 20 years (range: 3–40 years) after the case twins were initially diagnosed with CTCL. As none of the co-twins were diagnosed with CTCL within a period of minimum 3 years and up to a maximum of 40 years after the corresponding case twin was diagnosed with CTCL, our findings indicated that the complete absence of CTCL in the co-twins was not a result of a short observation time. Although none of the co-twins developed CTCL, we examined whether they were diagnosed with other hematological malignancies or had an increased frequency of cancer in general. Importantly, none of the co-twins were diagnosed with nonHodgkinor Hodgkin lymphomas (Table 2, upper part). Moreover, the frequency of breast cancer, cancers in the respiratory organs and other malignancies was similar among co-twins and co-twin controls (Table 2, upper part, second column versus forth column and data not shown) indicating that the risk of lymphoma and other cancers in co-twins was similar to that of the control population. The frequency of cancer other than lymphoma was comparable in case twins and co-twins (Table 2, upper part, first versus second column). As CTCL patients have an increased risk of infectious diseases such as pneumonia and sepsis (reviewed in Willerslev-Olsen et al.), we compared the frequency of infectious diseases in caseand co-twins and the corresponding control cohorts. As expected, the frequency of pneumonia and sepsis was significantly higher for the CTCL cases (36% and 17%, respectively) than for the case–controls (17% and 4%, respectively, Table 2, upper part), supporting the notion that CTCL patients carry an increased risk of infections due to an impaired immune defense (reviewed in Girardi et al. and Willerslev-Olsen et al.). Restricting the analysis to monozygotic twins showed the same picture, that is, a higher frequency of pneumonia and sepsis among CTCL cases than case–controls (Table 2, lower part). In contrast, the frequency of pneumonia and sepsis in co-twins was 14% and 2%, respectively, which was similar to the frequency seen in co-twin controls (12% and 5%, respectively, Table 2, upper part, column two versus column four). The frequency in co-twins of other infectious diseases was also similar to the frequency in the controls (data not shown) indicating that co-twins—unlike their CTCL case twins—did not display an increased risk of retracting infectious diseases and chronic infections. In support, the frequency of hospitalization was not increased in the co-twins when compared with the co-twin control population (Table 2, upper part, column two versus column four). The frequency of common diseases such as ischemic heart disease, hypertension and chronic obstructive pulmonary disease was also similar in cotwins and co-twin controls (Table 2, upper part, column two versus column four), indicating that the overall morbidity and disease spectrum in co-twins was very similar to that seen in the control population and distinctly different from their CTCL case twins. Indeed, we found no difference in mortality between co-twins and co-twin controls, adjusted HR 1.08 (95% CI 0.65–1.82) (Table 1),