Characterising the Immunological Microenvironment in Newly Diagnosed Multiple Myeloma Bone Marrow By Time of Flight Cytometry Reveals Abnormalities in Antigen Presenting and Effector Lymphocyte Populations with Prognostic Significance

Background

Loss of immune surveillance is thought to contribute to disease progression and treatment resistance in a range of malignancies including multiple myeloma (MM). Understanding the degree and pattern of immunological abnormality present within the bone marrow microenvironment at the time of MM diagnosis is vital if we are to utilize emerging immunological therapeutic strategies successfully in MM. While immune checkpoint inhibition in the relapsed refractory setting has been disappointing, early intervention before immune subversion mechanisms have become established may enable more effective restoration of immunological disease control.

Method

Bone marrow samples from 18 patients with newly diagnosed, untreated, myeloma (NDMM) and 9 age matched bone marrow controls were labelled with metal-antibody conjugates and assessed by time of flight cytometry using the CyTOF platform. Expression of 36 protein targets including markers of proliferation, degranulation and cytokine production alongside phenotyping and viability markers were measured at the single cell level.

Differences in the abundance and function of distinct cellular populations were assessed using the unsupervised, automated CITRUS algorithm with the goal to identify novel cell populations.

Results

We observed a decrease in three key populations in NDMM; CD4 T cells with an effector phenotype (CD4^EF), CD8 T cells with an IL2 producing effector phenotype (CD8^EF), and dendritic cells. The expected elevation in malignant plasma cells was also seen and characterized. Interestingly CD8 T cells with a cytotoxic phenotype were not decreased.

The dendritic cell (DC) population contained two distinct subpopulations characterized by their level of CD16 expression. The CD16 positive population (DC¹⁶) expressed a range of activating receptors and cytokines while the CD16 negative population (DC^TOL) exhibited strong Ki67 expression. Within NDMM samples the DC¹⁶ population had stronger expression of PDL1 (p=0.0349) and loss of TIM3 (p=0.0122) and 2B4 (p=0.0148) compared to controls suggesting that this subset is less functionally active in NDMM. The DC^TOL population had a similar increase in PDL1 (p=0.0065) and loss of TIM3 (p=0.0165) but also had a shift towards CD107a (p=0.0014) and perforin (p=0.0196) expression, suggesting a tolerogenic role for this subset in myeloma.

Within the CD4^EF subset NDMM samples exhibited reduction in Ki67 (p=0.0054) compared to controls, suggesting that the decrease in population abundance might be due to loss of proliferation. Furthermore, the NDMM population had increased expression of TGFβ (p=0.0027) and FoxP3 (p=0.0409) suggesting that those cells that are present may have a regulatory role.

The CD8^EF population also showed reduction of Ki67 (p=0.0494) in NDMM compared to control samples. This was accompanied by a loss of DNAM1 (p=0.0096), suggesting a loss of co-stimulatory capacity, alongside elevations in TGFb (p=0.0022) suggesting a pro-tumor cytokine shift.

A broad spread of cellular abundance levels was noted in NDMM compared to controls which led us to investigate whether differences in cell population abundance was associated with survival. This was seen for the CD8^EF population, with higher abundance correlating with longer survival (r=0.6643, p=0.0026). Individuals with higher abundance of both the CD8^EF population and the DC¹⁶ population had a reduction in relapse and death in the first 36 months following diagnosis (p=0.0366).

Conclusions

This data demonstrates that even at the early time point of myeloma diagnosis there is evidence of both numerical and functional defects in key cell populations involved in antigen presentation and anti-tumor activity.

We propose that ineffective antigen presentation by PDL1 expressing DC populations results in poorly proliferative CD8 and CD4 effector populations with pro-tumor cytokine production. The high FoxP3 expressing CD4 population may also have a regulatory role.

This data highlights PDL1 as an important therapeutic target in NDMM where it may have a role in restoring immune surveillance at an early disease time point.

Harnessing emerging deep profiling technology to identify patterns of immunological change across multiple cellular subsets within one individual may enable us to identify immune signatures which predict outcome and response to treatment.