A molecular circuit linking the BCR to the NAD biosynthetic enzyme NAMPT is an actionable target in Richter syndrome

Occurring in a minority of patients with chronic lymphocytic leukemia (CLL), Richter syndrome (RS) represents the transformation of the disease into an aggressive lymphoma.¹ In the great majority of cases, the histological presentation is that of a diffuse large B-cell lymphoma (DLBCL), even though cases with Hodgkin lymphoma histology have been described.² RS is, in most cases, clonally related to the preceding CLL phase, as inferred from genetic analysis of VDJ rearrangements.^3,4 Transformation of CLL into RS is correlated with the acquisition of novel genetic lesions, such as TP53 or CDKN2A mutations, typically associated with transformation of hematological malignancies from indolent to aggressive.^3,5,6 The prognosis of patients with RS, particularly of those with a lymphoma clonally related to the preceding CLL, is dismal, with a median survival of just a few months.⁷ Currently, there are no satisfactory therapeutic regimens for the treatment of these patients, even though several trials combining old and new drugs are ongoing. B-cell receptor (BCR) inhibitors, such as ibrutinib and acalabrutinib, have been initially used for patients with RS,^8,9 even though resistance rapidly develops. In addition, most patients with RS are likely to have been treated with BTK inhibitors during the CLL phase, limiting use of these drugs and raising the question of their role in the transformation and the development of resistance of RS cells.^10,11 BCL-2 inhibitor treatment has also been considered, usually in association with chemoimmunotherapy or with anti-CD20 antibodies.^12,13 A recent trial combining venetoclax with R-EPOCH achieved ∼50% complete responses, and a study of venetoclax plus R-CHOP is ongoing.¹⁴ In addition, it is important to explore new drugs or new combination strategies to achieve better and longer lasting responses.¹⁵ Our laboratory has extensively worked on this topic, exploiting xenograft models derived from patients with RS (RS-PDX) that were established from 4 different patients with RS and that maintain a tight resemblance to the primary tumor.¹⁶ We have analyzed the role of novel antibody drug conjugates targeting RS using tumor-specific antigens, such as ROR1,¹⁷ or B-cell restricted targets, such as CD37.¹⁸ We have also studied the effects of novel drug combinations, highlighting a synergy between the dual phosphoinositide 3-kinase γ/δ inhibitor duvelisib and BCL-2i.¹⁹ Even though safety issues on the use of duvelisib have limited initial enthusiasm toward the drug, PI3Ki are being evaluated in different combinations in clinical trials for patients with RS (see clinical trials NCT03534323 and NCT03884998).

Nicotinamide mononucleoside phosphoribosyl transferase (NAMPT), the rate-limiting enzyme in the synthesis of nicotinamide adenine dinucleotide (NAD) from nicotinamide (NAM), is the NAD biosynthetic enzyme that most rapidly responds to heightened cellular metabolic needs.^20,21 NAMPT is overexpressed in several human cancers and is an attractive tumor target.²² Even though inhibitors of NAMPT have been in clinical trials for many years now, they did not go beyond phase 2 due to significant toxicity and limited responses, attributed to compensatory NAD biosynthesis through alternative pathways.²³ Recently, OT-82, a small molecule with potent anticancer activity in hematological malignancies, was recognized as an effective NAMPT inhibitor, renewing interest in this target,^24-28 alone or in combination.^29,30 

This work originates from the observation that BCR cross-linking in CLL cells is followed by rapid NAMPT upregulation, suggesting that the 2 pathways are functionally linked.³¹ Here, we demonstrate the existence of a positive feedback loop involving the BCR and NAMPT and converging on AKT. The translational implication, demonstrated using in vitro cultures and RS-PDX models, is that the simultaneous block of PI3K and NAMPT has synergic effects, widening the spectrum of possible combination therapies for patients with RS.

The full list of reagents and antibodies used is in supplemental Methods.

RS-PDX models were established and maintained as described.¹⁶ 

To activate BCR signaling, RS cells were treated with plate-bound immunoglobulin M (IgM; 10 μg/mL; Southern Biotech, Birmingham, AL). Duvelisib (Selleckchem, Aurogene, Rome, Italy) was used at 5 μM for the indicated time points, whereas OT-82 was used at 0.625, 1.25, and 2.5 nM, according to the model.

Total RNA was extracted using the RNAse Plus Mini Kit (Qiagen, Hilden, Germany), retrotranscribed to complementary DNA using the High-Capacity cDNA Reverse Transcription Kit (ThermoFisher, Milan, Italy) and analyzed using the CFX384 Touch Real time detection system (Bio-Rad, Hercules, CA). All primers were from ThermoFisher. Gene expression levels were computed using the DDCT method.

Lysates were prepared as described¹⁷; proteins resolved by SDS-PAGE and transferred into a 0.2 μm nitrocellulose Trans-Blot Turbo Transfer membrane using the Trans-Blot Turbo Transfer System (all from Bio-Rad). The optical density of the bands was determined with the Image Lab 3.0 software (Bio-Rad).

NAD and adenosine triphosphate (ATP) levels were measured using bioluminescent assay (NAD/NADH-Glo Assay and CellTiter-Glo Luminescent Cell Viability Assay, Promega, Milan, Italy). Both NAD and ATP levels were normalized over total protein levels (RLU × μL/μg). Sirtuin activity was evaluated using a fluorimetric assay (EPI018, Sigma-Aldrich, Milan, Italy) and normalized on protein levels (pmol/min per μg).

Cell viability was determined by Annexin V-APC Apoptosis Kit (ThermoFisher). Samples were acquired by flow cytometry using FACSCelesta (BD Bioscience, Franklin Lakes, NJ) and analyzed with the FACSDiva (BD Bioscience) software.

Full details on immunoprecipitation and 2-dimensional (2D) gel electrophoresis are reported in supplemental Methods.

For SIRT1 and SIRT2 silencing, freshly isolated RS cells (6 × 10⁶) were transfected with validated siRNA for SIRT1 (sc-40986), SIRT2 (sc-40988), or scramble siRNA (sc-37007), from Santa Cruz Biotechnology (DBA Italia, Milan, Italy) using the NEPA21 Super Electroporator (Nepa Gene, Chiba, Japan). After transfection, RS cells were cultured in a complete medium and harvested at 24 hours for protein and RNA extraction or treated with duvelisib for 24 and 48 hours.

The Italian Ministry of Health approved all in vivo experiments (approval no. 571/2022-PR). RS cells (5 × 10⁶) were resuspended in RPMI-1640-Matrigel (1:1; Corning, Milan, Italy) and injected (double flank) into 8-week-old NOD/SCID/γ-chain^−/− (NSG) mice and left to engraft. When a palpable mass was apparent, animals were randomly assigned to 4 different groups and treated with vehicle (5% dimethyl sulfoxide, 90% PEG300, and 5% Tween 80 or 30% hydroxypropyl-b-cyclodextrin; all from Sigma-Aldrich), duvelisib (100 mg/kg), OT-82 (40 mg/kg), or their combination. Treatment was administered by oral gavage for 3 consecutive days, followed by 4-day off, for a total of 3 cycles. Four days before treatment initiation, animals were put on a niacin-free rodent diet (Envigo, Bresso, Italy). During the whole experiment, mice were monitored and tumor masses were regularly measured twice a week by caliper. Mice were euthanized 2 days after the last treatment.

A portion of tumor masses was formalin-fixed and paraffin-embedded for immunohistochemistry (IHC). To evaluate apoptosis, an anticleaved Caspase-3 antibody was used, followed by an antimouse HRP-conjugated antibody and 3,3′-diaminobenzidine (EnVision System, Dako, Glostrup, Denmark). Slides were analyzed using an AXIO Lab.A1 microscope (Zeiss, Oberkochen, Germany), equipped with a Canon EOS600D reflex camera, and the images acquired using the ZoomBrowserEX software (Canon, Tokyo, Japan).

Statistical analyses were performed using GraphPad v 9.0 (GraphPad Software, La Jolla, CA). The statistical test used is indicated in all figure legends, along with the P values.

NAD is a vital molecule for regulating redox reactions, as well as the activity of intracellular and extracellular NAD-dependent enzymes. Its biosynthesis is controlled through the activity of 4 rate-limiting enzymes, namely (1) NAMPT, which controls the biosynthetic pathway from NAM; (2) nicotinamide riboside kinase (NRK), from nicotinamide riboside; (3) nicotinate phosphoribosyltransferase (NAPRT), from niacin; and (4) quinolinic acid phosphoribosyltransferase from tryptophan, considered the only de novo pathway. By using RNA-sequencing data from a cohort of 14 primary RS samples derived from patients with newly diagnosed disease (and referred to as primary RS samples) and 4 RS-PDX models (RS9737, IP867/17, RS1316, and RS1050), we showed that NAMPT is invariably expressed in all samples, whereas NRK and NAPRT display generally lower (NRK) and more variable (NAPRT) levels of expression (Figure 1A). Quinolinic acid phosphoribosyltransferase constitutive expression was low to undetectable, except for 2 primary samples, supporting the notion that de novo biosynthesis does not occur in lymphocytes.³² RNA-sequencing data were confirmed in RS-PDX samples both at transcript and protein levels (Figure 1B-C), highlighting high expression of NAMPT in all 4 models. Intriguingly, RS9737 and RS1316 cells also showed a high constitutive expression of NAPRT, an enzyme associated to intrinsic resistance to treatment with NAMPT inhibitors in different cancer models.²³ 

We next investigated the effects of BCR activation on NAD biosynthetic enzyme expression in RS cells, using RS cells obtained from PDX mice and cultured for up to 4 days. Cross-linking of the BCR using plate-bound anti-IgM antibodies was followed by significant increase in NAMPT messenger RNA (mRNA) and protein levels after 24 hours of IgM stimulation (Figure 2A-B). Activation of the BCR pathway was confirmed by increased expression of both CCL3 and MYC, known BCR targets (supplemental Figure 1A-B).³³ In these conditions no statistically significant change in NMRK and NAPRT expression levels could be observed, confirming that NAMPT is the enzyme that most responds to environmental needs (supplemental Figure 2).

NAMPT upregulation was completely inhibited by treating RS cells with the PI3Kγ/δ inhibitor duvelisib, used at the dose of 5 μM for 24 hours (Figure 2A-B), which effectively blocks BCR signaling, while failing to induce apoptosis (not shown).¹⁹ As a further confirmation of the effects of BCR signaling inhibition on NAMPT mRNA and protein upregulation, we used the noncovalent and highly selective BTK inhibitor pirtobrutinib,³⁴ which inhibited BCR-driven NAMPT upregulation at the mRNA and protein levels (supplemental Figure 3A-B).

Consistent with NAMPT upregulation, cellular NAD levels were markedly increased 24 hours after BCR cross-linking compared with unstimulated cells in all 4 RS-PDX models (Figure 2C). Increased bioavailability of NAD levels fueled NAD-dependent sirtuins activity, as documented using a sirtuin activity assay (Figure 2D). Pretreatment of RS cells with duvelisib completely inhibited both NAD accumulation and sirtuin activity, demonstrating a specific linkage to BCR activation (Figure 2C-D). Inhibition of sirtuin activity was also documented after BCR blockade using pirtobrutinib (supplemental Figure 3C).

These findings highlight a molecular circuit connecting BCR activation to NAD biosynthesis through NAMPT upregulation, in turn, positively affecting the activity of NAD-dependent sirtuins.

We next evaluated the effects of NAMPT inhibitors in an ex vivo experimental setting. Briefly, RS cells, freshly purified from NSG animals, were plated in the presence of different doses of OT-82, with FK866 used as control. Drug escalation testing (from 0.625-10 nM) clearly showed a progressive decline in intracellular NAD levels, already highly significant at the lowest dose and at 24 hours (supplemental Figures 4A and 5A for OT-82 and FK866, respectively). Concomitantly, we observed a significant decrease in intracellular ATP levels. RS1050 and IP867/17 models showed a steeper decline in ATP and NAD levels than RS9737 and RS1316 (supplemental Figure 4B). These findings are indicative of the onset of apoptosis, which was already evident at 24 hours and reached maximal levels at 48 hours. With ∼20% residual live cells after 72 hours, RS1050 and IP867/17 cells showed a more complete apoptotic response than RS9737 and RS1316, in which the percentage of live cells at the same time point was in the range of 55% and 75%, respectively (supplemental Figure 4C-D). This striking difference was recapitulated by FK866, which also induced marked apoptosis selectively in RS1050 and IP867/17, whereas RS1316 and RS9737 remained mostly vital (supplemental Figure 5C-D). A likely explanation for the different behavior lies in the expression of NAPRT, which can compensate for NAMPT block and is higher in RS1316 and RS9737, while almost absent in the other 2 models (Figure 1B-C).

We next asked whether blockade of the BCR signaling pathway could synergize with OT-82 in inducing apoptosis of RS-PDX cells. To this purpose, we combined the sublethal dose of 5 μM duvelisib with 0.625 nM for RS1050 and IP867/17 and 1.2 nM for RS9737 and RS1316 (Figure 3A). Exposure to the combination of PI3Ki and NAMPTi for 48 hours and 72 hours strongly reduced cell viability compared with either drug used alone (Figure 3A-B for 72 hours; supplemental Figure 6A for 48 hours). On average, viability in combined treatment samples dropped to 50% after 48 hours (supplemental Figure 6A) and to 20% after 72 hours (Figure 3A), without differences based on NAPRT expression. Accordingly, by calculating the combination index³⁵ of duvelisib and OT-82 cotreatment, we highlighted a strong synergic effect (combination index <1) in all RS-PDX models (Figure 3A). Analysis of the apoptotic machinery demonstrated higher levels of cleaved Caspase 3 and PARP in the combination treatment compared with single treatments (Figure 3B; supplemental Figure 6B). These results were confirmed by showing that FK866 was synergic with duvelisib in inducing apoptosis, as determined using the Bliss Independence model (supplemental Figure 7). These findings were reproduced by inducing BCR blockade using pirtobrutinib, which effectively synergized with OT-82 in the induction of apoptosis (supplemental Figure 8).

Taken together, these results highlight a synergy between agents that block BCR signaling and NAMPT inhibitors in RS cells in inducing apoptosis, independently of the underlying genetic lesions and of the expression of NAPRT.

In examining potential convergence points to explain the synergy between PI3Ki and OT-82, we focused on AKT, a central effector in the BCR signaling pathway. To achieve full activation, AKT needs to be phosphorylated at serine 473 and threonine 398 residues, an event mediated by PI3K, and deacetylated at lysines 14 and 20 (K14/20), a second posttranslational modification regulated by sirtuins. AKT acetylation at these specific lysine residues suppresses its activity by preventing its PIP₃-mediated recruitment at the inner leaflet of the plasma membrane. SIRT-mediated deacetylation of K14/20 is therefore a necessary step for its activation.^36-38 Several studies have demonstrated how NAD depletion via NAMPT inhibition results in the decrease of PIP₃ binding and phosphorylated AKT levels.^37,39-42 For these reasons, we determined AKT activation levels after BCR cross-linking in the presence or absence of NAMPT and PI3K inhibitors.

As expected, upon BCR cross-linking, AKT was robustly phosphorylated both at Thr308 and Ser473 in all RS models, an event that was partially blocked by pretreatment with duvelisib and OT-82, whereas their combination achieved the highest effects (Figure 4A; supplemental Figure 9). Consistently, phosphorylation of GSK3b, which is directly downstream of AKT, was inhibited after combined treatment (Figure 4A; supplemental Figure 9).

To validate the hypothesis of a sirtuin-dependent activation of AKT, we performed immunoprecipitation experiments, followed by western blotting using an anti–acetyl-lysine antibody. At the baseline and upon IgM cross-linking, no AKT acetylation could be observed, whereas treatment with OT-82 was followed by accumulation of acetylated AKT, even more pronounced in the presence of trichostatin-A and NAM, used to completely block sirtuin activity (Figure 4B).⁴³ Therefore, fully activated AKT appears to be both deacetylated and phosphorylated. Evidence of multiple posttranslational modifications was obtained using 2D IEF/SDS-PAGE electrophoresis and western blot for total AKT and demonstrating 4 AKT-immunoreactive spots: by using the acetyl-lysine blot signal, we identified the first one on the right as AKT without any posttranslational modifications, the second one to singly modified AKT (either acetylated or phosphorylated), and the third and fourth as doubly- and triply-modified AKT molecules. As shown by the western blot in Figure 4C, the levels of p-AKT (both p-Thr308 and p-Ser473) decreased upon treatment and, at the same time, proteome-wide acetylation levels increased, shown by the increasing acetyl-α-tubulin signal. Correspondingly, in the 2D blot, we observed a shift in intensity upon treatment toward the spots on the left, which corresponds to AKT, either at a single or double lysine residue, corresponding to its inactive form. Trichostatin-A and NAM were again used as controls because they lead to complete inhibition of sirtuins activity and maximal acetylation, as confirmed by the intensity of the dot corresponding to acetylated alpha tubulin (Figure 4D).

AKT acetylation is regulated through the activity of SIRT1, which deacetylates AKT at K14/20, and SIRT2, which deacetylates AKT exclusively at K20.^36-38,44 Both these enzymes are expressed by primary RS cells and PDX models, as verified by RNA sequencing, quantitative reverse transcription polymerase chain reaction, and western blotting, with SIRT2 being the dominant enzyme, both in primary samples and in RS-PDX models (Figure 5A). Treatment of RS cells with either SIRT1-specific (Selisistat, EX-527) or SIRT2-specific (Thiomyristoyl) inhibitors triggered a potent apoptotic response after 48 hours (supplemental Figure 10A) and 72 hours of exposure, which was further increased by the addition of duvelisib (Figure 5B). In line with these findings, genetic knockdown of either SIRT1 or SIRT2, obtained using siRNAs, markedly reduced cell viability, particularly in the presence of duvelisib (Figure 5C; supplemental Figure 10B). In all instances, the combination of the 2 drugs was significantly more effective in inducing apoptosis than each drug used alone (Figure 5B). Consistently, exposure to the combination of a SIRT inhibitor and duvelisib completely blocked AKT activation, as documented by western blot (Figure 5D). Overlapping results were obtained after genetic silencing of both enzymes (Figure 5E; supplemental Figure 10B-D).

The efficacy of the PI3Ki/NAMPTi drug combination was then validated in vivo using RS-PDX models. RS cells were subcutaneously injected in 8-week-old NSG mice and left to engraft until tumor masses were at least 0.2 cm³. Mice were then randomized in 4 groups and treated for 3 weeks with OT-82 (40 mg/kg), duvelisib (100 mg/kg), or their combination. A mixture of vehicles was used as the control. Treatments were administered for 3 consecutive days, leaving a 4-day washout period between cycles, as previously determined.^19,45 To increase the efficacy of OT-82 in vivo, 2 days before starting the treatment, regular mouse diet, supplemented with 115 mg/kg of niacin, was substituted with a custom diet containing 30 mg/kg of niacin, as previously described.⁴⁶ At the end of treatment, mice were euthanized and tumor masses recovered and measured. In line with in vitro results, it was clear that in all models, combination treatment was significantly more effective than individual treatments, as inferred by measuring volume and weight of the masses (Figure 6A for RS9737; supplemental Figures 11-13 for RS1316, RS1050 and IP867/17, respectively). We then performed flow cytometry–based apoptosis assays to show that the viability of RS-PDX cells was remarkably reduced in the masses treated with the drug combination (Figure 6B). NAD and ATP levels in the isolated RS-PDX cells showed drastic reduction after the combination of PI3Ki and NAMPTi, in line with the more robust apoptosis observed in these samples (Figure 6C). IHC assay for cleaved Caspase 3 on tumor masses confirmed a highly significant increase in staining of cleaved Caspase 3 in mice treated with the combination of the 2 drugs (Figure 6D). These in vivo results confirm how the combined use of these 2 drugs is much more effective than the individual treatments in reducing tumor progression.

This work identifies, to the best of our knowledge, for the first time a possible therapeutic role for the NAMPT inhibitor OT-82 in PDX models of RS. NAMPT, the rate-limiting enzyme in the biosynthesis of NAD from NAM, has been an attractive tumor target for many years, considering its central role in supplying NAD upon increased metabolic needs of cells. However, the first inhibitor introduced in clinical trials, FK866, was unsuccessful because of both toxicity issues and lack of efficacy. Subsequent studies attributed lack of efficacy to expression of other NAD biosynthetic enzymes able to bypass NAMPT block and reconstitute NAD levels. Interest in NAMPT inhibitors, particularly for hematological cancers, was recently refueled by the introduction of the novel inhibitor OT-82, which has shown both tissue and cancer specificity in hematological tumors and no adverse effects in murine models of disease.^25,45 

In this work, we studied the effects of OT-82 in PDX models of RS,^16,17 which remains an orphan disease with no effective therapies. Analysis of the expression data, both in primary RS samples and RS-PDX models, of the 4 different NAD biosynthetic enzymes showed that NAMPT is significantly more expressed than the other enzymes, although its expression levels are heterogeneous. Importantly, upon BCR cross-linking, we observed a marked increase in NAMPT mRNA and protein levels in all RS-PDX models, confirming that this is the enzyme responsible for increased NAD production in the presence of anabolic signals. This behavior is unique to NAMPT because the expression of NAPRT and NMRK does not appear to be modulated by the activation or inhibition of BCR signaling in RS cells.^47,48 This finding highlights the existence of a connection between BCR signaling and transcriptional NAMPT upregulation, even though identification of the exact transcription factor involved is still lacking (Figure 7). Many of the effects stemming from NAMPT upregulation and the subsequent increase in NAD levels are mediated by the increase in the deacetylating activity of sirtuins on their protein targets.⁴⁸ From this perspective, NAMPT is considered the major regulator of sirtuin activity; not only does its activity generate NMN to start the recycling of NAM back to NAD, but it also acts as a scavenger of NAM itself, which acts as a sirtuin inhibitor. Sirtuins influence cellular responses to genomic instability by regulating cell cycle, DNA repair, cell survival, and apoptosis, thus having critical roles in cancer progression and metastasis.^49-53 Specifically, SIRT1 can induce chromatin silencing and regulates a wide variety of biological processes and cellular functions by deacetylating a large number of nonhistonic proteins, including AKT.⁴⁴ SIRT2 is also a key regulator of multiple physiological pathways. In the cytoplasm, in which it is predominantly localized, SIRT2 acts on a variety of substrates, including AKT.³⁸ Several different studies have highlighted the central role of SIRT1 and SIRT2 in hematological malignancies, and overexpression of SIRT1 and SIRT2 has also been reported in CLL and DLBCL, with elevated levels being associated with poor prognosis in CLL, AML, CML, and DLBCL.^54-58 Promising results have also been obtained from studies investigating the therapeutic potential of sirtuin inhibitors both in solid and hematological tumors. The SIRT1 inhibitor selisistat exhibited a strong antiproliferative effect in multiple leukemia cell lines,⁵⁹ whereas the antitumor activity of the SIRT1/2 inhibitor cambinol has been demonstrated on BCL6-expressing Burkitt lymphoma cells and xenografts.⁶⁰ Sirtinol treatment induced growth arrest and apoptosis in human breast cancer cells, lung cancer cells, and leukemic cells; and led to increased sensitivity to chemotherapy in prostate cancer cell lines.^41,61,62 Lastly, the SIRT2 inhibitor thiomyristoyl strongly reduced the viability of leukemic cell lines compared with nonhematopoietic cancer cell lines.⁶³ Together, these and other studies encourage the use of SIRT1 and SIRT2 inhibitors for solid and hematological cancers. However, the translation of in vitro and in vivo results into small molecule sirtuin inhibitors with the potential to be tested in clinical trials has been hampered by a variety of difficulties: the lack of isoform specificity, limited potency and bioavailability, and poor pharmacokinetics and pharmacodynamics of the candidates advanced to the trial stage.^64,65 In fact, no sirtuin inhibitor has been approved for the treatment of solid or hematological cancers. Some of the possible explanations are that sirtuins could be redundant in their actions, with significant overlaps in substrate specificity; or that, because some of their cellular roles are not dependent on enzymatic activity, enzymatic inhibition alone exhibits only moderate efficacy. It is reasonable to conceive that reducing NAD levels, by inhibiting NAMPT activity, could have a polypharmacological effect on the activity of the whole sirtuin family. In this study, we demonstrated the synergy arising from BCR signaling inhibitors, both at the BTK and PI3K level, combined with OT-82–mediated NAMPT inhibition (Figure 7). The dual treatment induces strong apoptosis in vitro in all the RS-PDX models tested. We also confirmed the same finding in vivo, observing a striking tumor regression in all RS-PDX models, independent of their genetic background. Importantly, we show that combined administration of duvelisib and OT-82 converges on AKT activation through the modulation of SIRT1/2 activity obtained by decreasing NAD levels. We also highlight the critical role that SIRT1 and SIRT2 play in RS cells, especially when they are targeted in combination with BCR signaling inhibition, either by pharmacological inhibition of their activities or by silencing their expression using siRNA. It must be noted that, although we corroborated the role of AKT in the synergy established between OT-82 and duvelisib, we cannot in principle exclude that the depletion of NAD through NAMPT inhibition could result in additional and yet to be defined effects. For example, SIRT1 and SIRT-2 has been reported to induce c-Myc overexpression, promoting tumor growth.^66-69 

In conclusion, our data emphasize the crucial nature of NAMPT as the main NAD biosynthetic enzyme and prove how NAD levels, especially in the context of BCR inhibition, are a potential metabolic Achilles heel of RS, which could be exploited therapeutically. NAMPT and BCR inhibitors cooperate in turning off central proneoplastic cellular pathways that promote RS proliferation and survival and their coadministration should be further investigated and potentially translated to clinical applications. Further studies are needed to determine whether these mechanistic observations may be clinically useful.

The authors thank Andrea Iannello and Elisabetta Bolli for technical help with the in vivo work. The authors also thank Oncotartis for providing OT-82 for part of the experiments.

This work was supported by the Associazione Italiana per la Ricerca sul Cancro (AIRC IG-23095 [S.D.] and MFAG 23107 [T.V.]) and by the Innovative Training Networks (ITN) INTEGRATA program (grant agreement 813284) (S.D.).

Contribution: V.G.M. planned and performed experiments and contributed to manuscript writing; A.F. planned and performed experiments and together with S.D. organized text and figures; N.V., M.M., M.R., N.A.P., and C.M. performed experiments; D.G.E. and T.V. planned and critically reviewed experiments; and S.D. planned and interpreted experiments and wrote the manuscript.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Silvia Deaglio, Laboratory of Functional Genomics, Department of Medical Sciences, University of Turin, via Nizza 52, 10126 Turin, Italy; email: silvia.deaglio@unito.it.