AUTHORS
Giacinto Luca Pedone, Paolo Ricchi2, Ribersani Michela3, Francesco Arcioni Sr, Hematologist4, Ceglie Teresa5, Filippo Mazzi6, Antonella Sau7, Maria Pia Solfrizzi, MD, PhD8, Valentini Caterina Giovanna9, Anna Marcello10, Paola Bianchi Sr, PhD11, Marta Bortolotti Jr12, Francesco Passamonti, MD13, Wilma Barcellini, MD14, Bruno Fattizzo, MD15
AFFILIATIONS
1 Fondazione IRCCS Ca' Granda Ospedale Maggiore Policlinico, Milan, Italy, University of Milan, Milan, Italy
2 AORN Cardarelli, Naples, Italy
3 Policlinico Umberto I University Hospital, Rome, Italy
4 Pediatric Onco-Hematology Unit with Bone Marrow Transplantation, Perugia Hospital, Italy
5 San Luigi Gonzaga Hospital, Orbassano, Italy
6 Azienda Ospedaliera Universitaria Integrata Verona, Verona, Italy
7 Pescara Hospital, Pesacara, Italy
8 A. Perrino Hospital, Brindisi, Italy
9 Fondazione Policlinico Universitario A. Gemelli IRCCS, Rome, Italy
10 Fondazione IRCCS Ca' Granda Ospedale Maggiore Policlinico, Milan, Italy
11 Fondazione IRCCS Ca' Granda Ospedale Maggiore Policlinico, Milan, Italy
12 Fondazione IRCCS Ca' Granda Ospedale Maggiore Policlinico, Milan, Italy
13 Fondazione IRCCS Ca' Granda Ospedale Maggiore Policlinico, Milan, Italy, University of Milan, Milan, Italy
14 Fondazione IRCCS Ca' Granda Ospedale Maggiore Policlinico, Milan, Italy
15 Fondazione IRCCS Ca' Granda Ospedale Maggiore Policlinico, Milan, Italy, University of Milan, Milan, Italy
ABSTRACT
Introduction
Pyruvate kinase deficiency (PKD) is the most frequent congenital non-spherocytic hemolytic anemia, due to biallelic PKLR mutations reducing erythrocyte ATP production and lifespan. Mitapivat, an oral allosteric pyruvate kinase (PK) activator, is the first approved disease-modifying therapy. The ACTIVATE and ACTIVATE-T registration trials required at least a missense PKLR allele as inclusion criteria, leaving outcomes across the full genotypic spectrum underexplored. We analyzed the genetic asset of PKD patients treated with mitapivat in a multicenter real-world study.
Methods
We retrospectively studied 30 consecutive adults with PKD given mitapivat through patient-named use programs at 9 Italian centers (5 mg BID, escalated to 20 and 50 mg BID). Diagnosis was based on reduced PK activity and/or biallelic PKLR variants (NM_000298.6); genotype was available in 26. Alleles were classified as missense (M) or non-missense (N: nonsense, frameshift, splice-site), genotypes as M/M, M/N and N/N, and their localization on PK domains was assessed. Response was defined as hemoglobin (Hb) rise ≥1.5 g/dL at ≥2 consecutive visits or ≥33% transfusion-burden reduction (TBR) within 48 weeks; complete response required normal sex-specific Hb and normalized hemolysis. Missense variants were mapped onto the mitapivat-bound PK tetramer (PDB 8XFD, Gemmi 0.7.5).
Results
Median age was 45 (IQR 29-57) years, baseline Hb 9.3 (8.5-10.2) g/dL; 10 (33%) were regularly transfused, 20 (67%) splenectomized. The 26 genotyped patients carried 24 distinct variants across 52 alleles (15 missense, 9 non-missense); 17 were compound heterozygous, 9 homozygous. Genotypes were M/M 16 (62%), M/N 6 (23%), N/N 4 (15%). Alleles clustered in the C domain (30/52, 58%), the C-terminal regulatory region involved in tetramer stability and mitapivat binding, and the A domain (15/52, 29%), the largest central barrel-motif domain binding the substrate.
Over a median follow-up of 19 (10-29) months 20/30 (67%) responded: 18 (60%) with an Hb rise ≥1.5 g/dL and 7/10 (70%) transfused patients with TBR, of whom 6 (60%) became transfusion independent. In responders, Hb increased from 9.4 to 11.8 g/dL (+2.2, p<0.0001), with reticulocytes (229 to 101 x10^9/L), LDH (233 to 201 U/L, p=0.013) and indirect bilirubin (2.6 to 0.7 mg/dL, p<0.0001) falling. Responses were rapid (3 patients at 5 mg, 8 by 2 months at 20 mg) and durable (18/20 at 12 months, 90%). Ten patients (33%) discontinued, 8 for inefficacy.
Considering allele distribution: 13/16 (81%) M/M vs 2/6 (33%) M/N and 1/4 (25%) N/N (p=0.032) responded, with attenuated median Hb improvements of 2.5, 1.0 and 0.6 g/dL (p=0.08). The presence of at least one non-missense allele reduced response to 3/10 (30%) vs 13/16 (81%, p=0.01). At the variant level, c.1456C>T p.(Arg486Trp), prevalent in Southern Europe, was the commonest variant (16/52 alleles, 31%) and predicted response (11/12 carriers, 92%, p=0.047), as did c.1151C>T p.(Thr384Met) (3/3). Conversely, only 1/3 carriers of the Northern Europe and the US prevalent c.1529G>A p.(Arg510Gln) variant responded, and neither carrier of the nonsense c.721G>T p.(Glu241Ter) did. Notably, the single N/N responder was homozygous for the C-terminal truncating c.1528C>T p.(Arg510Ter) and achieved TBR without an Hb rise, suggesting a residual activatable enzyme (only 2.6% of lost transcript).
No variants lay within the mitapivat-binding pocket, with identical distances in responders and non-responders (median 10.3 vs 10.3 Å, p=0.94). Distance to the nearest adjacent subunit was shorter in responders (5.9 vs 10.6 Å, p=0.010), at both the A-A′ (7.8 vs 10.6 Å, p=0.010) and C-C′ (5.9 vs 17.4 Å, p=0.023) interfaces, and persisted excluding p.(Arg486Trp) carriers (p=0.066).
Conclusions
Mitapivat exceeded trial efficacy in routine clinical practice, with response driven by genotype at both the genotype and individual allele levels. Missense allele load and specific substitutions predicted therapeutic benefit. The high overall response reflects the high prevalence of the Southern European p.(Arg486Trp) allele spectrum. Because non-missense genotypes, including a biallelic C-terminal truncation, can still respond, genotype should guide counseling on response probability rather than serve as a rigid selection criterion for mitapivat treatment. Finally, structural mapping suggested subunit-interface integrity, rather than drug-pocket proximity, as the determinant of mitapivat rescuable PK enzyme.
DISCLOSURES
Giacinto Luca Pedone: Alexion Pharmaceuticals - Honoraria; Roche - Honoraria; Novartis - Honoraria; Recordati Rare Diseases - Honoraria