Aberrant Cytoplasmic Expression of p16 Is a New Strong Prognostic Marker in Young MCL Patients Treated By Frontline Cytarabine-Based Immunochemotherapy, a Study from the Lysa Group, Nordic Lymphoma Group and the European MCL Network

Introduction

Despite therapeutic advances, optimal management of mantle cell lymphoma (MCL) in young high risk patients remains an unmet need. Progression within 24 months of starting chemotherapy (POD24) indicates poor prognosis but can't guide initial therapy. TP53 mutations (Eskelund 2017) and CDKN2A deletions (del) (Delfau-Larue 2015, Le Bris 2020) link to poor prognosis but don’t accurately identify high-risk MCL and are not available for all patients. High p53 expression (>50%) by immunohistochemistry (IHC) can surrogate TP53 mutations, indicating poor prognosis with 2.2 years median survival (Scheubeck 2023). We evaluated the prognostic value of p16^INK4A protein expression alone or in combination of p53 abnormalities.

Materials and methods

We assessed p16, p53, and BCL-1 expression by IHC in tumor biopsies from 321 young patients treated in four European clinical trials based on a high-dose ARAC regimen: EUMCL younger (R-DHAP arm) NCT00209222 (n=114), Nordic MCL2 ISRCTN 87866680 and MCL3 NCT00514475 (n=46), and LyMa NCT00921414 (n=161). We estimated PFS and OS by the Kaplan-Meier, compared groups using log-rank tests, and used Cox models for multivariate analyses. Genomic and p53 expression analyses were previously published (see above).

Results

As expected, Bcl-1 and p53 localized in the nucleus. In patients with CDKN2A biallelic del, p16 was not expressed. Some cases without biallelic del, unexpectedly overexpressed p16 in the cytosol. No CDKN2A or XPO mutation was identified to explain this mislocalization. In the Mino aggressive MCL cell line, which has no p16 genomic alteration, p16 was overexpressed mainly in the cytosol. We established the p16 cytoplasmic expression threshold at IHC>10% through ROC curve analysis based on relapse, using both training and validation sets. We then evaluated the prognostic value of p16 abnormalities (IHC>10% and/or genomic deletion) and p53 abnormalities (IHC>50% and/or deletion), both individually and in combination. p53 had overexpression in 31% and/or del in 17%. For p16, 13% had cytoplasmic expression and/or 26% had del. In 26% of cases both p16 and p53 abnormalities were observed defining cycle control double hit (ccDH). P16 high cytoplasmic expression was in 7% of low MIPI-c and 26% of high MIPI-c patients (p=.009). ccDH patients had significantly reduced response after induction, being present in 50% of non-responders vs 23% in responders (p=.052). POD24 patients presented more often with cytoplasmic p16 (30% vs 10%, p<.001) and high nuclear p53 expression (47% vs 28%, p=.011). Both markers showed reduced PFS (HR=1.9 [1.2-2.9], p=.006 for p16 and HR=1.7 [1.2-2.4], p=.006 for p53). Double expressor p16/p53 (DE) was also significantly associated with POD24 status (22% vs 5%, p<.001). Among responders (CR/CRu/PR at end of induction, n=287), only p16 expression and DE were associated with POD24 (33% vs 10%, p<.001 and 24% vs 5%, p=.001). Among patients with p53 abnormalities (n=106), p16 cytoplasmic expression was significantly associated with shorter median PFS (1.9 vs 7.6 years, log-rank test p=.007). Among 150 informative patients, ccDH patients were present among 59% in POD24 vs 17% in non-POD24 (p<.001). DE was associated with reduced median PFS in comparison with single or no expressor respectively (1.9, 6.6 and 10.2 y, log-rank test p=.0003) and median OS (3.9 y, not reached [NR] and NR, log-rank test p=.009). ccDH status was strongly associated with reduced median PFS (2.2 vs 8.7 y, HR=2.1 [1.3-3.4], p=.006) and median OS (3.9 vs 12.4 y, HR=2.3 [1.3-4.1], p=.006). Multivariate analysis (n=248), including the variables: MIPI-c, blastoid morphology, Ki67, DE and trial, with a backward selection confirmed the independent impact of DE and high MIPI-c on PFS (DE HR=2.9 [1.6-5.5] p=.0008; high MIPI-c HR=5.9 [3.1-11.2] p<.0001) and OS (DE HR=2.1 [1.1-4.4] p=.03; high MIPI-c HR=8.6 [4.2-17.6] p<.0001).

Conclusions

Cytoplasmic p16 expression (>10%), especially in combination with p53 overexpression (>50%) and MIPI-c, is a powerful prognostic marker in MCL treated with standard chemotherapy and more discriminative that p53 alone. To better identify poor prognosis patients, we propose a two-stage algorithm: to first identify DE p16/p53 patients, then perform gene copy number analysis (TP53 and CDKN2A) and mutational status (TP53). The mechanism and functional impact of cytoplasmic p16 expression require further exploration.