Clinical features and genomic landscape of myeloproliferative neoplasm (MPN) patients with autoimmune and inflammatory diseases (AID)

extended genomic pro ﬁ ling in the context of MPN remain

AID diagnosis was based on recommended international criteria specific to each AID (including both serology and clinical manifestations in all cases) [8,9]. AID subtypes were classified as inflammatory arthritis, connective tissue disease, inflammatory dermatosis, systemic vasculitis, or organ-specific AID, and categorized as unclassified or incomplete when the AID did not meet diagnosis criteria or when data were missing respectively. All cases have been reviewed by two expert internists from our center (AM and DE), and response to treatment was evaluated based on validated criteria (detailed in Table S2). Patients with interferonalpha-induced AID were excluded from this study. AID and MPN diagnoses were considered concomitant if the interval between the 2 diagnoses was less than 6 months.
The median age of the whole cohort was 51.6 IQR [40.4-63.2] years. Overall, a total of 95 patients (6%) had an associated diagnosis of AID and were compared to the remaining 1446 MPN patients without AID (Table 1). There was a predominance of female patients (62 (65%) versus 773 (54%) within the AID group (p = 0.03)). MPN subtypes, MPN driver mutations, and the occurrence of constitutional or microvascular symptoms at MPN diagnosis did not differ between the 2 groups (Table 1).
A total of 103 diagnoses of AID were reported in 95 patients, including 8 patients that presented at least 2 different AID. The AID diagnosis predated MPN diagnosis in 41/103 (40%) cases with a median of 7.0 IQR [3.9-13.9] years, while it was considered concomitant in 13/103 (13%) cases and posterior in 35 (34%) cases with a median interval of time of 5.4 IQR [2.6-10.4] years (Fig. S2A). AID diagnoses mainly included organ-specific diseases (n = 48, mostly autoimmune dysthyroidism (n = 33)), followed by inflammatory arthritis (n = 13), connective tissue diseases (n = 9), inflammatory dermatosis (n = 9) and systemic vasculitis (n = 6) ( Fig. S2B and Table S3). Only one patient was diagnosed with VEXAS syndrome (a 67-year-old male patient with JAK2-mutated PV and concomitant TET2, NFE2 and UBA1 mutations), while the remaining 18 AID were unclassified. No significant difference between the different AID categories was observed regarding MPN-related characteristics, subtypes repartition or driver mutations (Table S4). Exhaustive AID clinical manifestations and biological features are listed in Table S5.
The median follow-up of the whole cohort was 8.3 IQR [3.7-14.3] years representing a total of 12,790 person-years. The median number of AID treatment lines was 1 IQR [0-4] in our cohort of MPN patients with associated AID (missing data for 30/103 AID) (Table S9), and response evaluation was gathered only for 15/73 (15%) cases (13 complete responses, 1 partial response, 1 steroid dependance). Ten (11%) and 122 (8%) patients died in the AID and control groups respectively during the follow-up period. The association of MPN with AID did not significantly impact overall survival (OS) (median 34.9 IQR [30.2-not estimable] years and not reached in the AID and control groups respectively, p = 0.72) (Fig. 1A). Moreover, the association with AID did not significantly impact MDS/AML transformation-free survival (p = 0.37), or secondary myelofibrosis-free survival in PV and ET patients (p = 0.91) (Fig. 1B). In MPN subtypes subgroup analysis, AID association did not impact patient's outcome (Fig. S4).
Our epidemiological results confirm those from a previous smaller series of MPN patients (n = 345) in which autoimmune manifestations were present in 8% of patients (including 21/34 (62%) female) [10,11]. Despite variations in the general population in industrialized countries, several studies estimated the global prevalence of AID around 3-5%, and a gender imbalance with a female to male ratio ranging from 1:1 to 1:10 [12]. We report in this study a similar prevalence, suggesting that MPN are not associated with an increased occurrence of AID. Moreover, through rigorous reviewing of all AID diagnoses by internist experts from our center, which allowed us to exclude MPN treatment-induced AID (notably with interferon-alpha) and nonimmune hypothyroidism (mostly iatrogenic), we describe here the full spectrum of MPN-associated AID. On the one hand, AID was not associated with a particular MPN subset or hematological phenotype, while on the other hand, the spectrum of associated AID, which mostly included organ-specific diseases as autoimmune dysthyroidisms, was similar to the global distribution of AID in non MPN patients [12]. Additionally, only 1 patient out of 1541 in our cohort had VEXAS syndrome, underlying the scarcity of UBA1 mutations in the context of Ph-MPN. Altogether, these results substantially differ from MDS, in which (i) the panel of AID includes inflammatory disorders with systemic involvement, mainly inflammatory arthritis, systemic vasculitis, connective tissue diseases and inflammatory dermatosis [13,14] and (ii) the prevalence of both UBA1-mutated (causative of VEXAS syndrome) [15] and non-mutated [9] systemic inflammatory disorders is higher, suggesting a non-fortuitous association. We also describe the extended genomic landscape of MPN patients with associated AID. In both with and without AID cohorts, the majority of our MPN patients had additional mutations. Of note, the proportion of HMR mutations did not differ between MPN patients with and without AID, regardless of MPN subtypes. This is consistent with the similar prognostic profile observed in these 2 groups, given the well-characterized prognostic weight of additional somatic mutations in the context of MPN [2,7]. TET2 mutations, which have been linked to AID in the context of clonal hematopoiesis [16] and MDS [14], although highly prevalent, were not significantly enriched in our cohort of MPN patients with associated AID. Finally, the long-term follow-up of our cohort allowed us to characterize late occurring events. The association of AID with MPN did not significantly alter patients' outcome, whether it was regarding OS, or disease progression to MDS/AML or secondary MF in patients, suggesting that the hematological malignancy remains the main therapeutic challenge.
In conclusion, our study depicts in detail the clinical characteristics and the mutational landscape of AID in the context of MPN. Overall, we found a low prevalence of AID, without any specific association with MPN subtype. The presence of AID did not impact MPN patients' prognosis.