A randomized, placebo-controlled phase 3 trial of the PI3Kδ inhibitor leniolisib for activated PI3Kδ syndrome

Abstract

Activated phosphoinositide 3-kinase delta (PI3Kδ) syndrome (APDS) is an inborn error of immunity with clinical manifestations including infections, lymphoproliferation, autoimmunity, enteropathy, bronchiectasis, increased risk of lymphoma, and early mortality. Hyperactive PI3Kδ signaling causes APDS and is selectively targeted with leniolisib, an oral, small molecule inhibitor of PI3Kδ. Here, 31 patients with APDS aged ≥12 years were enrolled in a global, phase 3, triple-blinded trial and randomized 2:1 to receive 70 mg leniolisib or placebo twice daily for 12 weeks. Coprimary outcomes were differences from baseline in the index lymph node size and the percentage of naïve B cells in peripheral blood, assessed as proxies for immune dysregulation and deficiency. Both primary outcomes were met: the difference in the adjusted mean change (95% confidence interval [CI]) between leniolisib and placebo for lymph node size was -0.25 (-0.38, -0.12; P = .0006; N = 26) and for percentage of naïve B cells, was 37.30 (24.06, 50.54; P = .0002; N = 13). Leniolisib reduced spleen volume compared with placebo (adjusted mean difference in 3-dimensional volume [cm3], -186; 95% CI, -297 to -76.2; P = .0020) and improved key immune cell subsets. Fewer patients receiving leniolisib reported study treatment-related adverse events (AEs; mostly grades 1-2) than those receiving placebo (23.8% vs 30.0%). Overall, leniolisib was well tolerated and significant improvement over placebo was notable in the coprimary endpoints, reducing lymphadenopathy and increasing the percentage of naïve B cells, reflecting a favorable impact on the immune dysregulation and deficiency seen in patients with APDS. This trial was registered at www.clinicaltrials.gov as #NCT02435173.

Graphical abstract

Figure 1.

CONSORT participant flow diagram. CONSORT, Consolidated Standards of Reporting Trials; CT, computed tomography; MRI, magnetic resonance imaging.

Figure 2.

Coprimary endpoints. (A) Primary lymphadenopathy endpoint: least square mean of the change from baseline at D85 in the log₁₀-transformed sum of product diameters of the index lymph node lesions in the PD analysis set (leniolisib, n = 18; placebo, n = 8). One patient in the leniolisib group was excluded because of complete resolution of lesions by D85 (0 mm), therefore the log₁₀ transformation could not be computed. Compared with placebo, leniolisib reduced lymphadenopathy (P = .0006). (B) Primary naïve B-cell endpoint: least square mean of the change from baseline at D85 in the percentage of naïve B cells (CD19⁺CD27^-CD10^-) out of total B CD19⁺ cells in the B-PD analysis set (leniolisib, n = 8; placebo, n = 5). Compared with placebo, leniolisib increased naïve B-cell levels (P = .0002). Error bars are SEM. ∗∗∗P ≤ .001. CFB, change from baseline; SEM, standard error of mean.

Figure 3.

Changes in lymphoproliferation parameters. (A) Individual untransformed SPD of index lymph nodes (leniolisib, n = 19; placebo, n = 9). Reference range (≤1.5 × 1.5 cm) not shown as up to 6 lymph nodes may be counted per patient. (B) Individual spleen volumes (leniolisib, n = 20; placebo, n = 9). The gray box indicates the reference range for adults. All patients from the safety analysis set with measurements are included in (A) and (B), whereas efficacy analyses used the PD analysis set. (C-J) Radiographic renderings of spleen volume (C, D, G, H) and lymph node diameters (E, F, I, J) at screening and D85 from 2 patients in the leniolisib arm. The images in the top row (C-F) are from a 29-year-old female. The investigator diagnosed splenomegaly at screen (C) that was deemed absent by D85 (D). The index axillary lymph node lesion outlined in red decreased in size over the trial (E-F). Yellow and purple outlines in (F) indicate non-index lymph nodes. The images in the bottom row (G-J) are from a 17-year-old male. The investigator diagnosed splenomegaly at screen (G) that remained present by D85 despite the spleen decreasing in size (H). The index upper cervical lymph node lesion outlined in red (I-J) decreased in size over the trial.

Figure 4.

Changes in B-cell parameters. (A) Mean naïve B-cell percentages (CD19⁺CD27^-CD10^-) over time. n values for baseline, D29, D57, and D85 for each group are as follows: leniolisib, 20, 19, 19, and 16; placebo, 10, 9, 9, and 10. (B) Mean transitional B-cell percentages (CD19⁺CD27^-CD10⁺) over time. n values for baseline, D29, D57, and D85 for each group are as follows: leniolisib, 19, 20, 20, and 17; placebo, 10, 8, 9, and 9. (C) Mean plasmablast percentage (CD19⁺CD27⁺CD38⁺⁺) over time. n values for baseline, D29, D57, and D85 for each group are as follows: leniolisib, 20, 20, 19, and 16; and placebo, 10, 9, 9, and 10. (D) Mean serum IgM level over time. n values for baseline, D29, D57, and D85 for each group are as follows: leniolisib, 21, 20, 21, and 21; and placebo, 10, 10, 10, and 10. (E) Mean switched (CD19⁺CD27⁺IgD^-) and nonswitched (CD19⁺CD27⁺IgD⁺) memory B-cell percentages over time. n values for baseline, D29, D57, and D85 for each group are as follows: leniolisib, switched, 20, 19, 19, and 16; placebo, switched, 10, 8, 9, and 10; leniolisib, nonswitched, 20, 19, 20, and 17; and placebo, nonswitched, 10, 9, 9, and 10. Baseline was calculated as the average of D−1 and D1. All patients from the safety analysis set with measurements are included in the figures, whereas efficacy analysis of (A) used the B-PD analysis set. Gray boxes indicate reference range from literature (A-C,E) or laboratory (D).^, In (E), gray box indicates reference range for switched memory B cells and dotted line indicates reference range for nonswitched memory B cells. Error bars are SEM.

Figure 5.

Changes in T-cell parameters. (A) Mean CD4⁺ and CD8⁺ senescent T-cell percentages (CD57⁺) over time. n values for baseline, D29, D57, and D85 for each group are as follows: leniolisib, CD4⁺, 18, 15, 16, and 15; placebo, CD4⁺, 9, 9, 7, and 8; leniolisib, CD8⁺, 19, 16, 19, and 16; and placebo, CD8⁺, 10, 9, 7, and 8. (B) Mean CD4⁺ and CD8⁺ PD-1⁺ T-cell percentages over time. n values for baseline, D29, D57, and D85 for each group are as follows: leniolisib, CD4⁺, 19, 17, 20, and 16; placebo, CD4⁺, 10, 9, 7, and 8; leniolisib, CD8⁺, 17, 14, 16, and 16; and placebo, CD8⁺, 10, 9, 7, and 7. (C) Mean CD4⁺ and CD8⁺ T-cell percentages over time. n values for baseline, D29, D57, and D85 for each group are as follows: leniolisib, CD4⁺, 19, 17, 20, and 16; placebo, CD4⁺, 10, 9, 7, and 8; leniolisib, CD8⁺, 19, 17, 20, and 16; and placebo, CD8⁺, 10, 9, 7, and 8. (D) Mean naïve (CD45RA⁺CD62L⁺) CD4⁺ and CD8⁺ T-cell percentages over time. n values for baseline, D29, D57, and D85 for each group are as follows: leniolisib, CD4⁺, 18, 17, 20, and 16; placebo, CD4⁺, 10, 9, 8, and 7; leniolisib, CD8⁺, 19, 17, 19, and 17; and placebo, CD8⁺, 10, 9, 7, and 8. Note that reference ranges are greater than graph axis, CD4⁺, 49.3% to 72.0%; CD8⁺, 48.6% to 87.5%. (E) Mean CD4⁺ and CD8⁺ terminally differentiated effector memory T-cell percentages (CD45RA⁺CD62L^-) over time. n values for baseline, D29, D57, and D85 for each group are as follows: leniolisib, CD4⁺, 18, 17, 19, and 16; placebo, CD4⁺, 10, 9, 7, and 8; leniolisib, CD8⁺, 19, 17, 20, and 16; and placebo, CD8⁺, 10, 9, 7, and 8. (F) Mean serum CXCL13 level over time. n values for baseline, D29, D57, and D85 for each group are as follows: leniolisib, 21, 20, 21, and 21; and placebo, 10, 10, 10, and 9. Baseline was calculated as the average of D−1 and 1. All patients from the safety analysis set with measurements are included in the figures. Reference ranges are from personal communication (A-B; Manish Butte, University of California, Los Angeles, written communication, 29 March 2022), the literature (C-E), or laboratory (F). In (A-E), gray boxes indicate reference ranges for CD4⁺ cells, and dotted lines indicate reference ranges for CD8⁺ cells. In (F), gray box indicates reference range. Error bars are SEM.

Figure 6.

Changes in baseline cytopenias. (A) Hemoglobin levels over time for individual male patients. Light red indicates patient whose lower limit of normal was 113 g/L, which was exceeded by D85. (B) Hemoglobin levels over time for individual female patients. (C) Platelet levels over time for individual patients. (D) Lymphocyte levels over time for individual patients. No patients in the placebo group had baseline lymphopenia. Baseline was calculated as the average of D−1 and D1. All patients from the safety analysis set with measurements are included in the figures. Gray boxes indicate reference range. In (A-B), asterisks indicate patients receiving iron. HGB, hemoglobin.