Versican (VCAN) Proteolysis Predicts T-Cell Infiltration in Myeloma Bone Marrow Post- Autologous Stem Cell Transplant (ASCT)

Background/ Hypothesis:

Versican (VCAN), a large matrix proteoglycan, accumulates in myeloma lesions (Hope et al., Blood 123(21): 3305-3315, 2014). VCAN sustains tumor-promoting inflammation but also generates immunosuppression through IL-6 and IL-10-mediated antigen-presenting cell dysfunction (Tang et al., Cell Rep 13(12): 2851-2864, 2015). VCAN proteolysis generates bioactive fragments (VCAN-derived matrikines, e.g., versikine) that oppose the functions of the parent macromolecule, intact VCAN (Hope et al., Blood 128(5): 680-685, 2016). VCAN-matrikines promote "T-cell inflammation" through co-ordinated effects on antigen-presenting and myeloid regulatory cells. In particular, we have shown that VCAN-matrikines promote Batf3-lineage dendritic cell (DC) differentiation from Flt3L-mobilized bone marrow progenitors and induce IL-12 but not IL-10 production by myeloid cells (Hope et al., J Immunol, in press, 2017). Batf3-DC constitute an intratumorally sparse DC subset with critical roles in tumor antigen cross-presentation and effector T-cell homing. Recently, VCAN-producing macrophages were shown to expand post-ASCT in myeloma and correlate with minimal residual disease (MRD) persistence (Arana, Zabaleta et al., Blood 128(22): 482-482, 2016). However, it is unknown whether VCAN proteolysis predicts T-cell infiltration post-ASCT for myeloma at a timepoint (Day 90-100 post-ASCT) when MRD is routinely quantitated and immunotherapies to control MRD are often instituted.

Specific Aims:

1. To detect VCAN proteolysis in paraffin-embedded bone marrow tissue obtained post-ASCT for myeloma

2. To investigate the relationship between T-cell infiltration and VCAN proteolysis in post-ASCT myeloma bone marrow

Study Design:

Bone marrow core biopsies from 40 patients who underwent ASCT for myeloma were analyzed at day 90-100 post-ASCT. ADAMTS protease-mediated VCAN proteolysis was analyzed by immunohistochemistry (IHC) through detection of neo-epitope DPEAAE as previously shown (Hope et al., J Immunol, in press, 2017). DPEAAE staining intensity was scored in the following scale: 0=none, +=patchy, ++= diffuse /weak, +++= diffuse/ strong. T-cell infiltration was scored in the following scale: rare, dispersed and intense. Clusters of T-cells (>5 CD8+ cells clustered together) were independently scored. Granzyme B staining status by IHC was scored using a scale similar to CD8+ status.

Results:

We found that prior de-calcification procedures did not prevent optimal evaluation and scoring of VCAN proteolysis status by IHC. Intense CD8+ T-cell infiltration and/or CD8+ clustering were only seen in the presence of intense (++ or +++) VCAN proteolysis. 14/14 marrows classified as displaying intense CD8+ infiltration plus CD8+ clusters were also scored as VCAN proteolysis ++/+++. 17/17 marrows with intense CD8+ infiltration were scored as VCAN-proteolysis ++/+++. 18/18 marrows with at least one CD8+ cluster detected in the specimen were VCAN proteolysis ++/+++. 67% of marrows (n=12/18) with +++ VCAN proteolysis had intense CD8+ infiltration. 83% (n=15/18) of marrows with +++ VCAN proteolysis had at least one CD8+ cluster detected. Intense staining for granzyme B was only seen in the presence of ++/+++ VCAN proteolysis (n=10/10).

Conclusions/ Future implications:

Intense VCAN proteolysis correlated with evidence of adaptive immune activation post-ASCT. T-cell infiltration in myeloma bone marrow post-ASCT was exclusively seen in the presence of intense VCAN proteolysis. Mechanistically, we hypothesize that T-cell infiltration is the result of enhanced Batf3-DC density in VCAN proteolysis-rich myeloma tumor microenvironments. We are currently investigating VCAN proteolysis-detection as an immune biomarker to predict efficacy of anti- myeloma immunotherapies as well as in relationship with MRD status post ASCT. There is a dearth of immune biomarkers that can be analyzed on standard de-calcified, paraffin-embedded bone marrow core specimens. For example, PD-L1 detection is problematic using de-calcified bone marrow specimens. By contrast, detection of VCAN proteolysis was unaffected by the de-calcification process. VCAN-matrikines (e.g., versikine) can also be therapeutically used to boost Batf3-DC whose paucity is often the limiting factor for efficacy of checkpoint inhibition and other immunotherapies.