Venetoclax durable response in adult relapsed/refractory Philadelphia-negative acute lymphoblastic leukemia with JAK/STAT pathway alterations

Abstract

High-risk relapsed/refractory adult Philadelphia-negative (Ph-) B-cell acute lymphoblastic leukemia (B-ALL) is a great challenge due to limited possibilities to achieve and maintain a complete response. This also applies to cases with extramedullary (EM) involvement that have poor outcomes and no accepted standard therapeutic approaches. The incidence of EM localization in relapsed/refractory B-ALL is poorly investigated: data on patients treated with blinatumomab reported a 40% rate. Some responses were reported in EM patients with relapsed/refractory B-ALL treated with inotuzumab ozogamicin or CAR-T. However, molecular mechanisms of response or refractoriness are usually investigated neither at the medullary nor at EM sites. In the complex scenario of pluri-relapsed/refractory B-ALL patients, new target therapies are needed. Our analysis started with the case of an adult pluri-relapsed Ph- B-ALL patient, poorly sensitive to inotuzumab ozogamicin, donor lymphocyte infusions, and blinatumomab in EM disease, who achieved a durable/complete response after treatment with the BCL2-inhibitor venetoclax. The molecular characterization of medullary and EM samples revealed a tyrosine kinase domain JAK1 mutation in the bone marrow and EM samples at relapse. By comparing the expression level of BCL2- and JAK/STAT pathway-related genes between the patient samples, 136 adult JAK1 ^wt B-ALL, and 15 healthy controls, we identified differentially expressed genes, including LIFR, MTOR, SOCS1/2, and BCL2/BCL2L1, that are variably modulated at diverse time points and might explain the prolonged response to venetoclax (particularly in the EM site, which was only partially affected by previous therapies). Our results suggest that the deep molecular characterization of both medullary and EM samples is fundamental to identifying effective and personalized targeted therapies.

Keywords: JAK/STAT; Philadelphia-negative cells; acute lymphoblastic leukemia; extramedullary; venetoclax (BCL-2 inhibitor).

FIGURE 1

(A) Timeline of patient’s leukemia medical history (Dg, diagnosis; ALLO, allogeneic transplantation; MRD, minimal residual disease; CR, complete response; Rel, relapse; DLI, donor lymphocyte infusion; RT, radiotherapy; BLINA, blinatunomab; INO, inotuzomab ozogamicin; PET PR, partial response; VEN, venetoclax; “Stars” indicate the four analyzed time points (1: extramedullary relapse #1—Dec/2018; 2: third relapse #2—Jan/2019; 3: post-inotuzomab ozogamicin hematological remission #4—Jun/2019; 4: post-venetoclax remission—Jul/2021). (B) Whole-body (upper panel) and axial scan (lower panel) 18F-FDG PET at four time points: at relapse, in December 2018 (i); after six cycles of inotuzumab ozogamicin followed by radiotherapy, in November 2019 (ii); after 6 months of treatment with venetoclax, in July 2020 (iii), and after 27 months of treatment with venetoclax and four DLIs, in April 2022 (iv). (C) 18F-FDG PET images (axial scan) at four time points showing i) inguinal pelvic mass at relapse - SUV 9.5 (December 2018); ii) partial resolution—SUV 2.9/Deuville 3 (November 2019) reached after treatment with inotuzumab ozogamicin and radiotherapy; iii) complete resolution—SUV 1.8 (July 2020—Deuville 1) after 6 months of venetoclax, and iv) after 27 months of continuative treatment with venetoclax and four DLIs—SUV 1.4 (April 2022—Deuville 1).

FIGURE 2

(A) Characteristics and analyses of patient’s sample at each time point. BM, bone marrow; EM, extra-medullary; FFPE, formalin-fixed paraffin-embedded; Ino, inotuzumab ozogamicin; MNCs, mononuclear cells; MRD, minimal residual disease; mut, mutation; na, not available; Neg, negative; PB, peripheral blood; Pos, positive;VEN, venetoclax; wt, wild type; and y, yes. (B) CRLF2 expression analysis in the patients’ samples and comparison groups. For the CRLF2 gene, swarm plots of log-transformed TPM are displayed. Each plot is a combination of one group of samples with one time point sample. The time point sample is highlighted with a bigger dot, and its color changes, depending on its z-score value (green if |z|>1.96; otherwise, red). The horizontal black lines mark the median. Ctrls, healthy donor group (n = 15); Ph+, BCR-ABL1-positive group (n = 31); TN, triple-negative group (n = 105). (C) JAK1 mutation in the patient’s sample. JAK1 protein diagram with the site of the V651M somatic variant (NM_002227:exon14:c.G1951A:p.V651M; Human hg19) within the tyrosine kinase domain and DNA Sanger sequencing chromatograms of the JAK1 V651 position at different time points (1: extramedullary relapse- Dec/2018; 2: third relapse—Jan/2019; 3: post-inotuzomab ozogamicin hematological remission—Jun/2019; 4: post-venetoclax remission—Jul/2021). * site of heterozygous mutation (V651M); #V651 wild-type site.

FIGURE 3

STAT3, LIFR, STAT5B, MPL, MTOR, MYC, TP53, IL13, PIK3CB, and differential expression analysis between the patient’s samples and comparison groups. For each gene, multiple swarm plots of log-transformed TPM are displayed. Each plot is a combination of one group of samples with one time point sample. The time point sample is highlighted with a bigger dot, and its color changes, depending on its z-score value (green if |z|>1.96; otherwise, red). The horizontal black lines mark the median. Ctrls, healthy donor group (n = 15); Ph+, BCR-ABL1-positive group (n = 31); TN, triple-negative group (n = 105).

FIGURE 4

Protein–protein interaction network of the top differentially expressed genes between the patient’s samples and the Ctrl cohort. Edges represent protein–protein associations. Confidence ≥0.700; maximum number of interactors ≤20. Edge confidence: high (0.700) and highest (0.900) (see https://string-db.org/cgi/network).