Page 49 - Haematologica Vol. 110 - January 2025

PERSPECTIVE ARTICLE
long-lived effector memory cells that were already present in untreated patients at the time of CML diagnosis. Since the group of CML patients with clonal expansions had a better prognosis,65 it has been suggested that following dasatinib therapy, LGL may mediate therapeutic activity against leukemia, either attacking CML stem cells or by eliminating residual CML cells, thus ultimately favoring the maintenance of responses. The evidence of clonotypes of cytomegalovirus-specific CD8 T cells68 and the finding that patients with large granular lymphocytosis more easily reached a major molecular response64 further point to the protective role of these clonal expansions.
Examples in the settings mentioned above support the hypothesis that these clonal reactive populations originate in response to discrete events, sometimes beneficial (such as the response to pathogens or graft-versus-leukemia activity), sometimes ineffective. They undergo malignant transformation only in exceptional cases. In fact, chronic antigen-driven stimulation of the immune system may trigger a polyclonal cytotoxic T lymphocyte response that subsequently evolves, eventually leading to the lymph- oproliferative disorder, particularly upon acquisition of mutations. According to this interpretation, Awada et al.69 demonstrated STAT3 mutations in two out of 13 (15%) pa- tients with T-LGLL after solid organ transplantation and allogeneic HSCT. These mutations, by conferring oncogen- ic properties to affected cells, solidify their autonomous proliferation, leading to progression towards the full-blown disease. Taken together, these observations recapitulate the stepwise model of cancer progression, in which a series of events (whether mutational or related to microenviron- ment) drives clonal expansions with progressively more disordered phenotypes.
The work-up of T-cell clones of uncertain significance
T-CUS, either concurrently with, or independently of other hematologic/non-hematologic diseases, is a premalignant condition that has the potential to evolve into T-LGLL or other T-cell malignancies. Similarly to other previously men- tioned precursor states, T-CUS is preferentially observed in older individuals. The risk of evolution to the full-blown malignancy is estimated to be approximately 1% per year, but this figure will require more extensive evaluation. This means that most individuals will never experience a life-threatening disease.
In terms of pathogenesis, the lack in T-cell clonal expansions of a common antigen specificity,21 of shared TCR clonotypes70 as well as the lack of common TCRA and TCRB clonotypes in CD8+ TCRaβ in LGLL even among HLA-matched individuals,71 reinforce the concept of a widespread cellular activation. Emerging evidence indicates that the inciting event may be universal to the entire immune system. In fact, recent
G. Semenzato et al. studies in T-LGLL highlighted a considerable overlap be-
tween leukemic and non-leukemic parts of TCR repertoires via possible common triggers.33,72 The antigen-driven clo- notypes in T-LGLL patients occurred concomitantly with non-antigen-driven clones and neither shared T-LGLL clo- notypes nor T-LGLL clonotypes targeting known antigens were detected.70,72 In contrast with T-LGLL, it is likely that in T-CUS, clonal T cells do not progress towards pervasive clone overgrowth or suppression of hematopoiesis. Instead, they are maintained in a steady-state equilibrium, possibly mediated by the microenvironment.
The prevalent phenotype in T-CUS is CD3+CD8+CD57+CD28– CD27– and, to some extent, positive to NK receptors (p58 molecules, the killer immunoglobulin-like receptors), a pat- tern which is consistent with fully differentiated effector/ memory T cells with low proliferative and high cytotoxic activity. However, based on the surface immunophenotype, T-CUS encompasses a spectrum of phenotypic variants, all competent in cytotoxicity, including CD8 T-CUS, CD4 T-CUS, and gd T-CUS. Data available in T-CUS on gain-of-function STAT mutations, which are regarded as the hallmark of T-LGLL, are still scanty.1-3,54,62 Also in terms of microenvi- ronmental features, including non-clonal T lymphocytes, macrophages and other immunocompetent cells, only a few detailed accounts are available, mainly related to the overt malignancy.72,73
At this time, following appropriate analysis of the history of putative conditions mounted or perpetuated by highly specific polarized immune responses to strong antigen- ic stimuli that might discern reactive lymphocytosis, the definition of T-CUS, incidental T-cell clones detected in other conditions and T-LGLL must take into account the following items:
- Persistent clonality. Evidence of persistent clonality over time, demonstrated as reported above, is a prerequisite. - Clone size. In T-CUS the size of the involved clone is, by definition, smaller than the threshold required to establish a diagnosis of T-LGLL, this set threshold acting as an artificial watershed. This threshold makes a tentative distinction be- tween the end of the physiological range and the beginning of disease. However, as in all other precursor states, the threshold by itself does not help separate individuals at risk of progression from those bound to remain in a pre-malig- nant condition. For the time being, the arbitrary LGL count less than 0.5x109/L distinguishes T-CUS from T-LGLL1,2,7,15,28 but this criterion might evolve. T-CUS and indolent T-LGLL likely represent a biological continuum, making the dis- trinction between the two entities challenging due to the presence of a gray zone (Figure 1). Sometimes the LGL count exceeds 0.5x109/L and for this reason observation for at least 6 months is mandatory to ensure that LGL expan- sion is not a transient/benign reactive proliferation. A less pronounced Vβ skewing in T-CUS than in indolent T-LGLL might aid in differentiating T-CUS from T-LGLL. However, this distinction is largely a matter of semantics as the two
Haematologica | 110 January 2025
41