Immune Cell Profiles in Patients Treated with Lenalidomide and Alternate Day Prednisolone Maintenance Post Upfront ASCT for Multiple Myeloma (LEOPARD Trial)

The LEOPARD trial evaluated lenalidomide and alternate day prednisolone (RAP) as post ASCT maintenance in newly diagnosed transplant eligible MM patients (TE NDMM). 60 patients were recruited. Estimated median potential follow-up was 44 months (IQR 26m – 52m). Median PFS from time of commencing RAP was 38.3m (95% CI, 25.8 to 54.8); median OS was not reached (71.4% of patients were alive at 36 months). Here we present the findings from correlative immune studies of this trial.

Aims:

To undertake mass cytometry (CyTOF) based immune profiling in patients with TE NDMM treated with RAP maintenance post ASCT.

Methods:

The LEOPARD trial was a phase II, multi centre, open label, single arm study of RAP maintenance after a single melphalan conditioned (200mg/m²) ASCT as part of up-front therapy. Patients were restaged at D+42 ASCT, and if eligible, were commenced on RAP maintenance (LEN 10mg daily increasing to 15mg daily after 8 weeks and alternate day prednisolone 50mg) within 8 weeks of D+0 of the ASCT. Therapy continued until toxicity/progression.

CyTOF was performed in sequential samples in two selected groups of patients: long runners (LR, n=7), defined as those with PFS > 36 months (median) and early relapsers (ER, n=8), defined as those who progressed/died before reaching the lower quartile of PFS. [All patients had peripheral blood collected at baseline (pre-ASCT), 6w post-ASCT and weeks 4, 8, 12, 20, 28 and 40 of RAP]. Cells were barcoded using the Cell-ID 20-Plex Pd barcoding kit (Fluidigm) followed by staining with sub-set/function defining antibodies (targeting myeloid, B, T and NK cells: CD45, CD3, CD19, CD5, CD1c, CD226. CD8, CD11c, CD16, CD127, CD138, CD123, NKG2A, TIGIT, TIM3, CD45RA. CD274, CD27, CD197, CD28, Ki67, CD66b, CD183, KLRG1, CD43, NKG2D, CD38, CD278/ICOS, CD25, HLA-DR, CD4, CD57, GramB, PD-1, CD14, CD56, CD11b, Tbet, CD33). Samples were acquired on the Helios instrument.

Supervised analysis was performed to determine differences in canonical immune cell populations. Unsupervised analysis was then performed: data were clustered in the VORTEX package. Significant differences in cluster frequency were assessed by Mann-Whitney test for statistical significance. Cluster phenotypes were determined and validated via multiple visualisation approaches.

Results:

Median age was 56yrs for LR versus 63yrs for ER. Median PFS for LR was 46.3m (38.4 – 51.5m) versus 10.2 m (2.1 – 21.3m) for ER.

Supervised analysis was performed on all samples, dichotomized into baseline and last time point sampled for each patient. At baseline, Ki67+CD8+ T cells, ICOS+CD8+ T cells, HLA-DR+CD4+ T cells and CD11c+ myeloid cells were enriched in LR compared to ER. At the last timepoint sampled, Ki67+CD8+ T cells and ICOS+CD8+ T cells were again enriched in LR compared to ER. Conversely, B-reg-like cells (CD19+CD5+CD43-) were enriched in ER compared to LR at the last timepoint sampled.

Unsupervised analysis was performed on all samples (all timepoints were pooled). Five clusters were significantly enriched in LR compared to ER. Four of these clusters represented activated/cytotoxic NK cells: CD56 dim, CD16-, NKG2A(CD159a)+, NKG2D(CD314)+, Granzyme B+ and CD38+, and additional expression of CD57 on one cluster; one cluster represented a mature myeloid population, with high expression of HLA-DR, CD11b and CD11c and low expression of CD33. One cluster was significantly enriched in ER compared to LR, representing activated neutrophils, with high expression of CD66b, CD11b and CD16.

The clusters that were enriched were then assessed longitudinally over all time points. There was no difference in the kinetics of these populations between groups.

Conclusions

Significant differences in both T-cell and NK cell populations were demonstrable at baseline in LR versus ER patients. Subsequently, durable responses to post-ASCT lenalidomide maintenance were associated with a cytotoxic, controlled immune response whereas early relapse was characterised by a more uncontrolled inflammatory response and the emergence of B-reg-like cells prior to relapse. We conclude that immune profiling at baseline and after initiation of therapy may help to predict a more sustained response to lenalidomide maintenance enabling pre-emptive tailored treatment decisions.