A step ahead toward precision medicine for chronic lymphocytic leukemia

The concept of precision medicine applied to human tumors implies the personalized tailoring of clinical management and treatment choices according to the status of an array of molecular biomarkers, in conjunction with other patient features.¹ In chronic lymphocytic leukemia (CLL), the extensive body of genetic data that have been accumulated in recent years has led to the identification of many new molecular biomarkers with prognostic value. However, only a few of these serve the role of true predictors for choosing the most appropriate treatment for any given patient.^1,2 The active search for molecular predictors in CLL is becoming increasingly more important in the current therapeutic landscape of the disease, that ranges from chemo-immunotherapy with both old and newer monoclonal antibodies (mAb) to chemo-free options based on B-cell receptor (BCR) inhibitors, targeting either Bruton tyrosine kinase or phosphatidyilinositol-3-kinase, and BCL2 inhibitors.^3,4

In this issue of Haematologica, Tausch et al. have analyzed the prognostic and, more importantly, the predictive role of a panel of gene mutations in the randomized, phase III COMPLEMENT1 trial comparing chlorambucil with ofatumumab-chlorambucil in treatment-naïve CLL patients not eligible for intensive therapy because of age or comorbidities.⁵ The COMPLEMENT 1 trial had documented that addition of the type 1 anti-CD20 mAb ofatumumab to chlorambucil leads to clinically significant improvement in progression-free survival (PFS) (22.4 months in the arm treated with ofatumumab chlorambucil vs. 13.1 months in the arm treated with single agent chlorambucil), with a manageable side effect profile.⁶ But whether ofatumumab provided an advantage to all molecular subgroups of CLL remains unexplored. Remarkably, in the genetic analysis performed by Tausch et al., mutations of NOTCH1 were seen to predict weak benefit from the addition of ofatumumab to the chlorambucil backbone.⁵ The NOTCH1 signaling pathway is a key feature in CLL growth and survival, and is deregulated by mutations in a sizable fraction of CLL⁷ (Figure 1). NOTCH1 mutations in CLL may target either the autoregulatory PEST domain, or the non-coding 3’- untranslated region (3’-UTR) sequence.⁷ In the context of the COMPLEMENT1 trial, the addition of ofatumumab to chlorambucil provided a significant benefit in PFS to NOTCH1 wild-type patients, whereas no statistically significant benefit was achieved in NOTCH1 mutated cases, including patients whose mutations disrupted the NOTCH1 PEST autoregulatory domain as well as patients with NOTCH1 mutations affecting the 3’-UTR of the gene.⁵

The refractoriness to ofatumumab imparted by NOTCH1 mutations is reminiscent of the refractoriness to another type 1 anti-CD20 mAb, namely rituximab, that had been observed in the CLL8 trial comparing fludarabine- cyclophosphamide with fludarabine-cyclophosphamide- rituximab (FCR) in young and fit CLL patients.⁸ In fact, in the CLL8 trial, rituximab failed to improve response and survival in patients carrying NOTCH1 mutations.⁸ The fact that NOTCH1 mutations behave as a predictor of reduced benefit from type 1 anti-CD20 mAb in two prospective, randomized trials with different anti-CD20 antibodies (ofatumumab in COMPLENT1; rituximab in CLL8), different chemotherapy backbones (chlorambucil in COMPLEMENT1; fludarabinecyclophosmide in CLL8), and different target CLL populations (patients not eligible to intensive therapy in COMPLEMENT1; patients eligible to fludarabine-containing regimens in CLL8) contributes further to the robustness of the predictive significance of NOTCH1 mutations in CLL treated with chemo-immunotherapy containing anti-CD20 type 1 mAb.^5,8

The obvious question is whether the novel, type 2 anti- CD20 mAb in use for CLL, namely obinutuzumab, may overcome the refractoriness imparted by NOTCH1 mutations to anti-CD20 therapy. This may be possible, since the glycoengineered type 2 anti-CD20 obinutuzumab exploits a different mode of action, based on enhanced antibody-dependent cell-mediated cytotoxicity and increased direct cell death compared to the type 1 anti- CD20 mAb rituximab and ofatumumab.⁹ Preliminary data seem to suggest that obinutuzumab might be able to overcome such refractoriness in the CLL11 trial comparing obinutuzumab-chlorambucil with rituximab-chlorambucil. ¹⁰

Figure 1.NOTCH1 signaling pathway and effects of NOTCH1 mutations on CLL susceptibility to anti-CD20 mAb. In the context of a wild-type NOTCH1 gene (left panel), ligands (DLL -1, -3, -4 belonging to the Delta-like family or JAGGED -1, -2 belonging to the Serrate family) expressed by stromal cells and by antigen presenting cells (APC) bind to the extracellular portion of the NOTCH1 receptor on CLL cells. Ligand-receptor binding triggers sequential cleavages of the NOTCH1 receptor mediated by the ADAM10 metalloprotease and the S3 γ-secretase. As a consequence, the IntraCellular NOTCH1 (ICN) domain is free to translocate to the nucleus, where it interacts with RBPJ and other co-activators to induce transcription of target genes promoting cell growth and survival and other cellular programs. The signaling cascade is terminated by ubiquitinylation of the NOTCH1 autoregulatory PEST domain, that is mediated by the FBW7 complex and leads to ICN degradation in the proteasome. In CLL cells with wild type NOTCH1 genes, type 1 anti-CD20 antibodies (rituximab, ofatumumab) induce cell death in vitro and, in vivo contribute to better patient outcomes in patients treated with chemo-immunotherapy. NOTCH-1 mutations occur in a sizeable fraction of CLL (right panel), upregulate NOTCH1 signaling and lead to increased expression of target genes. Most mutations in CLL disrupt the PEST domain, reducing proteasomal degradation of ICN and stabilizing ligand-triggered NOTCH1 signaling. Type 1 anti-CD20 mAb are less efficacious against NOTCH1 mutated CLL cells both in vivo and in vitro. The exact mechanism of anti-CD20 refractoriness associated with NOTCH1 mutations is not fully understood, but has been suggested to be linked, at least in part, to downregulation of CD20 expression.

Guidelines for CLL still recommend chemoimmunotherapy as a therapeutic option despite the advent of BCR and BCL2 inhibitors.¹¹ In this context, knowledge of NOTCH1 mutation status might be important in clinical decision-making whenever a chemoimmunotherapy regimen containing an anti-CD20 mAb is being offered to patients. The evidence acquired so far on anti-CD20 refractoriness and NOTCH1 mutations would support the concept that, in the presence of a mutated NOTCH1 gene, the use of a chemo-immunotherapy regimen containing a type 1 anti-CD20 mAb may not be the most appropriate choice and might be replaced by one of the many other therapeutic options that are currently available for CLL.^5,7,11 Recommendations by guidelines on this specific issue are desirable at this stage.

The use of anti-CD20 mAbs in CLL is not limited to chemo-immunotherapy regimens both in treatmentnaïve and in relapsed/refractory patients. For example, the MURANO trial has shown the superiority of venetoclax- rituximab compared to bendamustine-rituximab in relapsed/refractory CLL.¹² The CLL14 trial has documented that venetoclax-obinutuzumab associates with longer PFS compared to chlorambucil-obinutuzumab in treatment- naïve CLL.¹³ The iLLUMINATE trial has shown the advantage of ibrutinib-obinutuzumab over chlorambucil obinutuzumab as first-line therapy.¹⁴ Ibrutinib-rituximab is superior to chemo- immunotherapy in an Eastern Cooperative Oncology Group (ECOG) trial devoted to treatment-naïve CLL.¹⁵ At present, it is not known whether the reduced efficacy of type 1 anti-CD20 mAbs observed in NOTCH1 mutated patients treated with chemo-immunotherapy would also be a feature of novel chemo-free regimens based on BCR or BCL2 inhibitors in combination with an anti-CD20 mAb.

The precise molecular mechanism through which NOTCH1 mutations confer resistance to anti-CD20 type 1 mAb remains, to a certain extent, elusive (Figure 1). Though the biological relationship between NOTCH1 mutation expression and CD20 cell surface expression was not a specific focus of the report by Tausch et al., measuring CD20 levels by flow cytometry in the COMPLEMENT1 trial population failed to reveal differences between NOTCH1 mutated and wild-type cases.⁵ Conversely, in a wide CLL series of almost 700 cases, CLL cells from cases harboring mutations of the NOTCH1 PEST domain showed lower CD20 expression compared to NOTCH1 wild-type cases.¹⁶ Reduced surface expression of CD20 appears to be a feature also of CLL cases harboring a different type of NOTCH1 mutations affecting the 3’-UTR of the gene.¹⁷ Lower CD20 expression on the cell surface of CLL cells has been shown to be coupled to lower mRNA levels of the MS4A1 gene that encodes the CD20 antigen.¹⁶ As a consequence, cell lysis induced by anti-CD20 type 1 antibodies, namely rituximab and ofatumumab, appears to be also lower in NOTCH1 mutated cases compared to CLL without this genetic lesion.¹⁶ Consistent with these observations, pharmacological inhibition of the NOTCH1 protein or siRNA silencing of the NOTCH1 gene have been shown to induce upregulation of the CD20 molecule on CLL cells.¹⁶ It is well known that several epigenetic and transcription factors regulate expression of the MS4A1 gene and of the CD20 antigen.¹⁸ Interestingly, mutations of the NOTCH1 intracellular domain lead to accumulation of mutated NOTCH1 in the nucleus and may alter the fine epigenetic regulation of MS4A1 and CD20 expression through interactions with the RBPJ transcription factor that is involved in the NOTCH1 signaling pathway.^16,18

Overall, the biological relationship between NOTCH1 signaling, its deregulation by mutations and expression of CD20 requires further investigation, ideally in study designs aimed at comparing different type 1 and type 2 anti-CD20 mAb in order to understand not only the mechanisms of resistance, but also the strategies to overcome such refractoriness. It should also be considered that NOTCH1 belongs to a molecular pathway and that mutations in B-cell malignancies may also target other players of the pathway.⁶ Because these genetic alterations either potentiate positive signals or compromise negative regulators of NOTCH1, it would be interesting to understand whether alterations of other NOTCH1 pathway genes, in addition to NOTCH1 itself, might have an effect on anti-CD20 mAb response in vitro and in vivo.

The clinical management and therapeutic landscape of CLL have changed substantially over the last few years and continue to evolve. The availability of a variety of treatment options, ranging from chemo-immunotherapy to molecular inhibitors of the BCR and BCL2 pathways, has generated the need to search for robust biomarkers that may assist clinicians in choosing the most suitable and sustainable treatment strategy for every patient. Guidelines recommend TP53 disruption and IGHV mutation status as molecular predictors and these are commonly used when choosing treatment.¹⁹ Tausch et al. now consolidate NOTCH1 mutation status as a novel potential biomarker for optimizing anti-CD20 treatment when a chemo-immunotherapy option is offered to patients.⁵ Other predictive biomarkers are also emerging, and include loss of function mutations of BIRC3, that deregulate the NFkB pathway and confer resistance or reduced efficacy with chemo-immunotherapy regimens,^20,21 as well as use of specific stereotyped BCR subsets, in particular subset #2, as observed in the correlative analysis of multicentric clinical trials.²²

Step by step, precision medicine is becoming a solid reality in the field of CLL for the benefit of patients and to optimize allocation of resources in clinical practice. At present, the available therapeutic options for CLL that are recommended by guidelines have not always been subjected to rigorous and multiple head-to-head prospective comparisons, thus leaving several unanswered questions when physicians and patients need to make a treatment choice. Choosing wisely, based on robust molecular predictors, coupled to the patient’s fitness and comorbidities, might represent a viable and clinically meaningful strategy for achieving the best therapeutic outcome for the individual patient and to satisfy the need to optimize resources for the patient community.

Footnotes

Correspondence

GIANLUCA GAIDANO - gianluca.gaidano@uniupo.it