Follicular lymphoma (FL) is characterised by recurrent relapses. Its clinical behaviour suggests that FL cells progressively gain proliferation advantages and evade therapies, probably by accumulating genetic alterations. However, VDJ mutations showed that, instead of a linear evolution, some relapses may develop from minor subclones, different from those that dominated at time of diagnosis. Unfortunately, this technically demanding approach could only be applied to few patients with limited sets of sequential biopsies. To re-address this issue, we developed an original tumour and allele specific assay to analyze the mutations that accumulate on the switch μ on the der(14)t(14;18) upon class switch recombination (CSR) activation. These mutations were compared between biological samples obtained in different sites during follow up. Fifty patients were studied, and 169 biological samples analysed (86 lymph nodes (LN), 45 peripheral blood samples (PB), and 38 bone marrow samples (BM)). For 48 patients, clonal mutations were found in all samples. However, we observed frequent intra-clonal variations (in 25 LN; 17 BM and 19 PB) and frequent differences between LN and BM (9/17), LN and PB (9/21) and even PB and BM (9/18) obtained at short time intervals (<1 month). These results indicate that FL is often a heterogeneous disease where different t(14;18) positive subclones spread or proliferate asymmetrically in and between tumoral compartments. They also suggest that CSR remains active after the acquisition of the transformed phenotype. However, in spite of this apparent ongoing CSR activity, the comparison of serial LN biopsies did not show any significant accumulation of mutations over time: 36 paired LN, median interval between biopsies: 32 months, 370 vs 363 mutation in all 1st and 2nd LNs respectively, median 7.5 vs 7 mutations/LN, p=0.89. Surprisingly, a simple apparition of new mutations was only observed once. For all other pairs, either mutations remained identical (14 cases), or their evolutions suggested complex selection schemes, sometimes apparently from subclones from another tumoral compartment. Notably, when mutations were identical, the second biopsy had almost always been obtained in less than 3 years (13/14 pairs). This strikingly contrasts with paired biopsies obtained at intervals of 3 years or more in which significant differences were almost always seen (14/15 pairs). This correlation between the evolution of the mutations and the interval between biopsies thus suggests different progression patterns. Finally, in many cases, mutations observed in the first LNs were not found at the following relapses (12/36 pairs). This observation confirms that many relapses do not develop from the dominant subclones at diagnosis but from less mutated ones. In conclusion, our results suggest that there may be at least two different patterns of progression in FL. The first could be a direct and rapid evolution from the dominant subclone, explaining the stability of the mutations. The second, which could potentially happen after many years of clinical remission, could instead be indirect, from a clonally related and often earlier t(14;18) positive subclone probably arising from a different compartment.
Read full abstract