University of Groningen The interplay between CD4 T cells and tumor cells in Hodgkin lymphoma Veldman, Johanna

(1)

The interplay between CD4 T cells and tumor cells in Hodgkin lymphoma Veldman, Johanna

DOI:

10.33612/diss.191061669

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below.

Document Version

Publisher's PDF, also known as Version of record

Publication date:

2021

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA):

Veldman, J. (2021). The interplay between CD4 T cells and tumor cells in Hodgkin lymphoma: interactions and immune checkpoint blockade. University of Groningen. https://doi.org/10.33612/diss.191061669

Copyright

Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

The publication may also be distributed here under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license.

More information can be found on the University of Groningen website: https://www.rug.nl/library/open-access/self-archiving-pure/taverne- amendment.

Take-down policy

If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.

Download date: 16-10-2022

(2)

567326-L-bw-Veldman 567326-L-bw-Veldman 567326-L-bw-Veldman 567326-L-bw-Veldman Processed on: 7-10-2021 Processed on: 7-10-2021 Processed on: 7-10-2021

Processed on: 7-10-2021 PDF page: 165PDF page: 165PDF page: 165PDF page: 165

7

Summary, Discussion and Future

Perspectives

(3)

Summary and Discussion

Hodgkin lymphoma (HL) is a B cell malignancy characterized by a minority of tumor cells, called Hodgkin Reed-Sternberg (HRS) cells, surrounded by a heterogeneous inflammatory infiltrate usually containing many CD4+ T cells. These CD4+ T cells extensively interact with HRS cells and are emerging as crucial players in HL pathogenesis and treatment response. Of particular interest are the CD4+ T cells in direct contact with the HRS cells, called rosetting T cells. In this thesis we aimed to further characterize the rosetting T cells and their interaction with the HRS cells, with a specific focus on the role of CD4+ T cells in the highly successful immune checkpoint therapy in HL patients.

CD4+ T cells residing in proximity of HRS cells

In the majority of cases, HRS cells express all necessary components of the normal antigen presentation pathway [1]. Blockade of adhesion molecule pairs involved in immunological synapse formation decreased rosette formation, supporting a role of antigen presentation [2-4]. However, the specific role of HLA class II-T cell receptor (TCR) interactions, essential for formation of the mature immunological synapse, has never been studied in HL.

In Chapter 2 we have set-up a novel in vitro co-culture model to study the early interactions between CD4+ T cells and HRS cells. For this, HL cell lines were co-cultured with HLA class II matched peripheral blood mononuclear cells (PBMCs) and rosette formation was studied under normal conditions and after interfering with molecules that are key for immunological synapse formation (e.g. HLA class II-TCR, CD58-CD2 and CD54-CD11a). Using this model, we have shown that HLA class II-TCR interaction is important for T cell activation, while CD58-CD2 interactions are required for both rosette formation and T cell activation.

No role for CD54-CD11a interaction was seen in our model. Direct interaction between HLA class II-TCR associated CD4 and CD58-CD2 was confirmed in HL patient tissue using proximity ligation assays. Based on this data we suggest a model in which engagement of CD58 with CD2 results in T cell adhesion, thereby allowing HLA class II to bind to the TCR. This will then initiate mature immunological synapse formation, followed by signal amplification and contact stabilization through CD2-CD58 interactions altogether resulting in T cell activation.

Somatic mutations in CIITA and CD58, can lead to loss of HLA class II and CD58 cell- surface expression in HRS cells [5-8]. This suggests that there is a selective advantage for HRS cells that lack these molecules, possibly because the HRS cells progressed and became independent of interactions with the rosetting T cells. Indeed, HL patients that lost HLA class II often have extranodal disease and an adverse prognosis [9]. Moreover, it was recently

(4)

7

described that membranous expression of HLA class II by HRS cells was predictive for achieving complete remission (CR) and a prolonged progression free survival after immune checkpoint therapy in relapsed and refractory HL [10]. In line with this, it was shown that loss of CD58 on tumor cells was restricted to HL patients that experience relapse. Overall, loss of CD58 especially in primary cases was much less common than loss of HLA class II [7].

In our in vitro model, extensive rosetting of CD4+ T cells around HRS cells can be observed already within 30 minutes of co-culture. The interaction is dependent on immunological synapse formation and leads to T cell activation. Altogether this suggests that the response is polyclonal and that multiple antigenic peptides can be recognized through TCR-HLA class II interactions. This is in line with the high mutational burden in HL [11].

Although T cells are activated, only low levels of IL-2 were produced possibly due to incomplete activation, immune checkpoint interactions or production of suppressive cytokines by HRS cells [12,13]. It is currently unknown whether production of low IL-2 levels by T cells in the vicinity to HRS cells has a functional role in the pathogenesis. We hypothesize that low levels of IL-2 might provide pro-survival signals to the tumor cells, without inducing antitumor immune responses. It was already described that low levels of IL-2 can increase T cell survival, induce T cell tolerance and promote development of regulatory T cells (Treg), which is in line with the predominance of Treg in the HL microenvironment [14-17]. Together, this might lead to an immune suppressed microenvironment around the HRS cells.

It has remained difficult to characterize changes occurring in T cells after interaction with HRS cells both in vitro and in vivo, due to the difficulty of separating rosetting T cells from non-rosetting T cells [2]. Therefore, characterization of rosetting T cells was largely limited to immunohistochemistry. Interestingly, this revealed that CD4+ T cells in proximity of HRS cells lack CD26 [18,19]. CD26 is a cell-surface molecule with an important role in T cell co-stimulation and IL-2 production [20,21] and lack of this molecule might contribute to an incomplete activation and decreased IL-2 production as we have observed in our in vitro model.

In Chapter 3, we further characterize the CD4+ T cells that reside in the proximity of HRS cells. For this, we sorted CD4+CD26- and CD4+CD26+ T cells from cHL lymph node cell suspensions and performed RNA sequencing and T cell receptor variable gene segment usage analysis. Our results show that the CD4+CD26- T cells, in the proximity of HRS cells, are antigen experienced and polyclonal with an exhausted phenotype. The gene expression profile of CD4+CD26- T cells showed features of both memory Treg and Tfh cells, well- characterized antigen experienced T cell subsets. This aberrant expression profile can be explained by expression of the transcription factors TOX and TOX2 that we identified as

(5)

being differentially expressed. TOX and TOX2 were both highly expressed in CD4+CD26- T cells and were frequently present in T cells directly rosetting around the HRS cells. HRS cells actively recruit antigen experienced memory Treg cells by secreting CCL17/TARC and CCL22/MDC [22-24]. As described in Chapter 2, these CD4+ T cells will then interact with HRS cells through TCR-HLA class II interactions. These interactions may very well not be permanent, but rather dynamic with CD4+ T cells recirculating into and out of the rosettes, while staying in the proximity of HRS cells. This hypothesis is strengthened by our live cell imaging data that shows attachment and subsequent release of CD4+ T cells to HRS cells (Figure 1). In this scenario, the CD4+ T cells are repeatedly in contact with HRS cells. We hypothesize that this repeated interaction will induce the distinct gene expression signature as observed in our RNA-seq data. In line with this, TOX and TOX2 are both induced by chronic or repeated antigen dependent stimulation of the TCR and both molecules are essential for the development of Tfh cells and the induction of an exhausted phenotype in CD4+ T cells [25-27]. TOX is known to induce several immune checkpoint molecules, including PD-1, and can also induce CXCL13 consistent with the genes upregulated in CD4+CD26- T cells [28,29]. In line with the exhausted phenotype, the CD4+CD26- T cells in HL are not clonally expanded, indicating lack of proliferation, and unresponsive to stimulation, as shown by no to low cytokine responses [18]. Reversing the exhaustion, by targeting the transcription factors TOX or TOX2, might be an interesting strategy for immunotherapy and more effective than targeting one or multiple immune checkpoint molecules.

In conclusion, our data presented in Chapters 2 and Chapter 3 show that antigen experienced CD4+ T cells are recruited by the HRS cells. These CD4+ T cells then repeatedly interact with HRS cells through HLA class II-TCR interactions, resulting in a low-level of activation, possibly induced by incomplete activation, ultimately leading to the induction of an exhausted state driven by TOX and TOX2 expression. These exhausted T cells survive and form an immune-suppressed tumor microenvironment (TME) around the HRS cells in which the HRS cells can survive and are protected from attack of more effective anti-tumor immune cells (Figure 2).

(6)

7

Figure 1. Live cell imaging to visualize rosette formation. The HL cell line L428 was co-cultured with peripheral blood mononuclear cells for 1 hour. Every 10 seconds an image was taken using a Deltavision Elite microscope.

The red cell interacts with the tumor cell twice (at 9.00 and 48.50) after which it is released again. The blue cell interacts with the tumor cell at 25.40 and moves away at 57.30, at that time point being however still attached.

(7)

Figure 2. Proposed model for early changes in CD4+ T cells based on our study. HRS cells secrete chemokines and cytokines which attract specific antigen experienced CD4+ T cells [1]. These CD4+ T cells interact with HRS cells through formation of the immunological synapse [2]. Repeated or prolonged interactions together with several inhibitory signals, derived from or induced by HRS cells, will result in incomplete activation of these T cells and prevent clonal expansion [3]. Moreover these interactions will induce the expression of the transcription factors TOX and TOX2 [4]. TOX and TOX2 will induce an exhausted phenotype within the CD4+ T cells, including the expression of PD-1 and CXCL13 [5]. This Figure was created using Servier Medical Art (http://smart.servier.com/).

Immune checkpoint molecules

In recent years, based on the high objective response rates (ORR) observed after PD- 1 blockade treatment in relapsed and refractory HL patients, many studies focused on the expression of immune checkpoint molecules on cells in the HL TME [30-34]. As immune checkpoint blockade (ICB) is usually well tolerated, this treatment has been moved to the first-line setting for HL patients, which generally presents in adolescents and young adults [35-37]. However, the mechanism of action of immune checkpoint inhibition in HL is not

(8)

7

completely clear yet and functional studies aiming to elucidate this are hampered by lack of in vitro and in vivo models.

In Chapter 4, we reviewed the literature on the plausible mechanism of action of ICB in HL, currently known immune evasion strategies and evaluated their possible contribution to resistance to ICB targeting PD-1. In solid malignancies CD8+ T cells are considered the main effector cells in PD-1 blockade responses, but in HL CD8+ T cells are usually not in close vicinity of HRS cells and HLA class I is frequently absent on HRS cells [8,38]. In contrast CD4+

T cells are abundant, always bound to HRS cells and expression of HLA class II on HRS cells was associated with favourable responses to ICB [10,32]. HRS cells possibly induce resistance to ICB through shaping of the microenvironment and lack of (neo)antigen presentation due to loss of HLA expression or low tumor mutational burden. In addition, in the HL TME there is active suppression of T cells through expression of multiple immune checkpoints, adenosine production or indoleamine 2,3-dioxygenase. In addition, tumor associated macrophages (TAMs) and NK cells might also play a pivotal role in the response to ICB treatment.

HRS cells often overexpress PD-L1 and PD-L2 due to a selective amplification of the 9p24.1 genomic region [39]. In addition, also TAMs express PD-L1 and PD-L2 [32,40]. PD-1 is expressed by cells in the TME such as T cells and NK cells [41,42]. Besides being expressed on the cell membrane, these molecules can also present as soluble molecules in the blood.

Although the function of these soluble molecules is currently unknown, several studies have described both tissue and plasma markers involved in the PD-1/PD-L axis as biomarkers with clinical value in classical HL (cHL). In Chapter 5, we studied whether plasma levels of proteins in the PD-1/PD-L axis are representative for the expression pattern as observed in cHL tissue.

Soluble (s)PD-L1 levels were found to be the most promising biomarker based on increased expression in plasma of patients compared to controls and its correlation with several clinical parameters. Moreover, high PD-L1 expression in the TME contributed to high sPD-L1 levels in the blood, while PD-L1 expression by the tumor did not. However, the total percentage of PD-L1 in the tissue did not directly correlate to sPD-L1 in the blood. In addition, we were able to confirm that PD-1 expression both in the tissue and plasma was positively influenced by VEGF.

Although the ORRs are high in HL, the CR rates upon PD-1 blockade are scarce and the progression free survival period is limited [43]. When PD-1 blockade is going to be used in the first-line setting it will be crucial to identify patients that will respond to treatment at an early stage. Based on the results of our study we suggest to monitor sPD-L1 levels during ICB treatment and compare the levels with the predictive value of PD-L1 expression status in tissue biopsies obtained before start of treatment. In our non-ICB treated cohort, we were

(9)

not able to assess the effect of sPD-L1 levels on outcome after standard first-line treatment due to the small number of relapses. However, in metastatic melanoma and non-small cell lung cancer, it was described that high pre-treatment sPD-L1 levels were associated with an increased likelihood of progressive disease after PD-1 or CTLA-4 blockade treatment [44,45].

Contradictory results have been described for the predictive value of changes in sPD-L1 levels after ICB. In patients with lung and gastric cancer, reduction of soluble PD-L1 significantly correlated with tumor regression after PD-1 blockade treatment [46].

Interestingly in melanoma, both PD-1 and CTLA-4 blockade often resulted in increased sPD- L1 levels [45]. Patients with a long-term or delayed increase of sPD-L1 had more favorable outcomes, whereas a temporary short increase directly after treatment was associated with progressive disease and shorter survival. Most favorable responses were obtained in patients with low to moderate sPD-L1 levels before treatment. The authors suggested that the increase of sPD-L1 indirectly reflects antitumor immune responses, as secretion of sPD- L1 can result from cytokine induction, cellular stress, cell injury and tumor cell death. In cHL, no data on sPD-L1 after ICB are available. After conventional first-line treatment, we observed a drop in sPD-L1 levels in HL (data not shown).

To explain these discrepant results an important consideration is whether sPD-L1 is derived from the tumor cells or from TAMs. In case sPD-L1 is mainly tumor cell derived, increased sPD-L1 levels can indicate tumor cell death and thus effective antitumor immune responses [45]. When sPD-L1 is mainly TAM derived, ICB might result in decreased PD-L1 levels on TAMs and thereby decreased sPD-L1 levels, resulting in less inhibition of T and NK cells and indirectly tumor cell killing. In the HL TME, the majority of PD-L1 is expressed by the abundant TAMs [32]. In addition, we showed that high PD-L1 expression in the TME and not on the tumor cells was associated with high sPD-L1 levels. The PD-L1+ TAMs are in the vicinity of PD-L1+ HRS cells and are enriched for contacts with PD-1+ T cells, resulting in increased immunosuppression in the tumor cell areas [32]. Shortly after start of ICB, a decrease in PD-L1+ TAMs was seen, while total TAM content did not change, suggesting PD- L1 downregulation [47]. In line with this, at relapse during ICB also a decrease of PD-L1+

TAMs was observed with however an increase in PD-1+ T cells [48]. Altogether, this suggests that a decrease of sPD-L1 in HL after ICB, as a result of a decrease of PD-L1 expression on TAMs, will probably be a favourable prognostic factor. Upon ICB relapse, the increased PD-1 on T cells in the vicinity of HRS cells, might compensate for the decreased PD-L1 on TAMs, resulting in increased interactions between PD-1+ T cells and PD-L1+ HRS cells thereby restoring the PD-1-PD-L axis.

Besides expression of PD-1, T cells within the HL TME express several other immune checkpoint molecules [30,31,33,34]. To improve CR upon ICB, new immune checkpoint

(10)

7

blockers are being developed and combination therapies are being investigated. To be able to select the most promising combination approaches an in vitro model to reliably predict efficiencies of these combined treatment strategies would be of major help. In Chapter 6, we aimed to set-up a long-term in vitro model that mimics HL biology and enables us to study the effect of ICB on CD4+ T cell functionality. Our model includes a two-step co-culture approach of PBMCs with irradiated HL cell lines. In the first co-culture step of 7 days, expression of PD-1 and CTLA-4 on CD4+ T cells was induced, although upregulation was faster and more pronounced with HLA class II unmatched compared to matched PBMCs. In the second co-culture step of 4 days, ICB treatment was added and effects on T cell activation, proliferation and cytotoxic potential were determined. Nivolumab treatment consistently increased T cell activation and proliferation in co-cultures with a PD-L1+ HL cell line using both matched and unmatched PBMCs. Ipilimumab was not effective in co-cultures with matched PBMCs, possibly due to the low percentage of CD4+CTLA-4+ T cells after the first co-culture step. Interestingly, combination treatment in co-cultures with unmatched PBMCs increased T cell activation to a greater extent than nivolumab single-treatment, while no additional effect on proliferation was observed. Despite effects on T cell activation and proliferation nivolumab treatment did not significantly increase tumor cell killing.

Interestingly, we do see an increased cytotoxic potential in the CD4+ T cell compartment mainly after the second co-culture step, which was even more prominent after nivolumab treatment.

With our in vitro model we can measure the effect of ICB on CD4+ T cells. Although our model needs additional optimization and adjustment steps this is an important approach to gain better understanding of the effects of ICB. Especially the treatment with nivolumab was effective in our model, although not much tumor cell killing was observed. Ipilimumab might be used in combination with nivolumab in case CTLA-4 expression is high in the cHL TME. It was described that CTLA-4+ T cells are abundant in the cHL TME and outnumber PD- 1+ T cells [34]. These CTLA-4+ T cells are usually not PD-1+ and are more frequently in direct contact with HRS cells compared to PD-1+ T cells [31,34]. Moreover, CTLA-4+ T cells are still present and more often in direct contact with HRS cells in patients that relapse after PD-1 blockade treatment. Altogether this suggests that treatment with ipilimumab in combination with nivolumab or at relapse after nivolumab treatment might be an interesting option.

However, no studies have been conducted to directly compare the efficiency of nivolumab or ipilimumab single with combined treatment in cHL patients. Recently, a study tested the efficacy of brentuximab vedotin (BV) with nivolumab, with ipilimumab or both in relapsed and refractory HL [49]. The ORR and CR were 89% and 61% in the BV/nivolumab group, 76%

and 57% in the BV/ipilimumab group and 82% and 73% in the BV/combined ICB group.

Median progression free survival was 1.2 years in the BV/ipilimumab group and was not

(11)

reached in the other two groups. More adverse events were observed in the groups receiving treatment that included ipilimumab. These results suggest that nivolumab mono- therapy might be more effective than ipilimumab single-treatment in cHL patients, but these results need to be validated. In addition, ipilimumab treatment might benefit a subset of patients, for example patients that are unresponsive to or relapse after PD-1 treatment.

Future perspectives

The role of HLA class II and antigenic peptide presentation

A big limitation of current studies in HL is the difficulty to study HL biology and treatment effects in vitro and translate this to patient samples. Using a novel in vitro model, we showed that T cell rosetting is established by formation of the immunological synapse and activation of rosetting T cells critically depends on TCR-HLA class II and CD2-CD58 interactions. An important next step would be to identify if indeed antigens presented by HLA class II are recognized during this process. To do this an HLA-DM knockout cell line can be generated and the effect on T cell rosetting and T cell activation can be subsequently studied. HLA-DM is required to displace the CLIP protein from the peptide binding groove of HLA class II to make it accessible for antigen loading [50,51]. If HLA-DM is not present, CLIP will be retained and there will be no proper antigen presentation. Lack of HLA-DM and retention of CLIP has been described in HL patients as a mechanism of immune escape [8]. If this approach indeed confirms the importance of antigenic peptides, it would be interesting to identify the neoantigens presented. In HL cell lines, neoantigen candidates can be identified by eluting HLA class II bound peptides followed by an HLA peptidomics approach [52,53]. Unfortunately this approach cannot be taken in HL patient cell suspensions, due to the scarcity of HRS cells present and the high amount of input needed for this technique.

However in patients, two recently described approaches to isolate HRS cells using enzymatically digested and mechanically disrupted FFPE material could possibly be used [54- 56]. To obtain enriched HRS cell populations the first approach flow sorted HRS cells (resulting in 40-90% enrichment), whereas the second approach made use of an image- based cell sorting technology (DEPArray^TM technology) (resulting in recovery of pure HRS cells). After isolation, bulk or single cell genomic profiling approaches can be used to identify somatic non-synonymous mutations [56,57] used for in silico peptide HLA binding predictions. Once neoantigen candidates are identified, neoantigen specific T cells can be isolated, expanded and tested for their reactivity against their target peptide [58]. Ultimately the TCR of these neoantigen specific T cells can be characterized and used for TCR gene transfer or CAR T cell therapy to induce tumor cell regression. As HL has a high mutational

(12)

7

burden and polyclonal T cells were identified in the TME, probably multiple neoantigens will be identified [11,57]. It would be interesting to see if shared neoantigens/neoantigens derived from the same genes are identified among different HL patients and/or cell lines.

Besides characterization of antigenic peptides, the involvement of HLA class II in TME composition could be studied. It would be interesting to see if HL cases with or without HLA class II expression differ in the presence of cell types, expression profile of cells in general and T cells in particular and whether there are changes upon relapse. Recently described techniques such as multiparametric immunofluorescence or single cell RNA sequencing technologies might aid in uncovering these remaining questions. Using a combination of these techniques it was recently shown that in HLA class II negative cases LAG-3+ T cells are common in close proximity to HRS cells [31], whereas in HLA class II positive cases PD-1+ and CTLA-4+ T cells are likely more dominant [32,34]. Interestingly, lack of HLA class II was described as predictor of unfavourable outcome after PD-1 blockade treatment [10]. These findings are important and have implications for treatment choices.

Expanding the long-term in vitro HL-PBMC model

To increase treatment success and improve treatment outcomes in HL patients it is important to understand the mechanism of action of current treatment options. The model proposed in this thesis might be suitable to gain further insight (Figure 3). One of the remaining questions is how well the model reflects the in vivo situation. To establish this, it is important to identify whether the CD4+ T cells in the model share characteristics with CD4+

T cells residing in the cHL TME. The proteins showing a CD4+CD26- T cell specific expression pattern as identified by us in the study described in Chapter 3 might provide a starting point for future studies. One question that remains to be answered is whether co-culture of PBMCs with HRS cells in our model results in induction of TOX and TOX2 in CD4+ T cells and how long it takes to induce these molecules? In addition, it would be interesting to follow CD26 dynamics in our model. Will CD4+CD26+ T cells become CD26- and which factors are responsible for this change? These outstanding questions could be answered using our model. Moreover, it will be crucial to identify and characterize the specific CD4+ T cell subset that responds to ICB treatment (e.g. TOX+ or TOX2+ CD4+ T cells). Single cell RNA sequencing can be used to further characterize CD4+ T cells in treated in vitro co-cultures or in responding patients compared to untreated co-cultures or non-responding patients.

Differentially expressed genes can be further studied in cHL tissue for their abundance and localization relative to HRS cells before, shortly after start of treatment and at relapse following treatment with ICB. Also, molecules on HRS cells that might influence treatment success can be studied in the model by generating knockout or overexpression cell lines.

(13)

Taken together, identifying characteristics of CD4+ T cells that will respond to ICB together with molecules on HRS cells that are necessary for responses, will aid in patient selection and personalized treatment.

An important discrepancy in our model compared to patients treated with ICB is the limited tumor cell killing after treatment. Although we were able to measure T cell activation and proliferation in response to nivolumab treatment, we were not able to detect a significant increase in tumor cell killing using an LDH cytotoxicity assay kit that has been successfully used before in diffuse large B cell lymphoma (DLBCL) [59]. In patients HRS cells disappear from the tissue days after start of nivolumab treatment [47] and therefore the exact mechanism of HRS cell killing will require additional study. It is possible that HRS cell death after ICB is not a direct result of cytotoxic immune cells, but merely the result of loss of T cell help. The HL cell lines used in our model are not dependent on T cell help and can survive without T cells being present. This might explain the lack of tumor cell killing after ICB in the in vitro model. If however direct cytotoxicity is necessary for HRS cell killing after ICB, the low amount of cytotoxic CD4+ T cells or the lack of other important cell types in our model might provide an alternative explanation. In our model we used total PBMCs and only see a limited increase in CD4+GranzymeB+ and CD4+TIA-1+ cells after the first co-culture step. Interestingly, a study by Tanijiri et al. (2007) described that HL cell lines were able to induce high numbers of CD4+ cytotoxic T lymphocytes from naïve CD4+ T cells after 8 days of co-culture without additional stimulation [60]. These induced CD4+ T cells were able to produce IFN-γ, expressed high amounts of granzyme B and TIA-1 and had a marked killing ability. It would be worthwhile to decipher whether using specific CD4+ T cell subsets, such as naïve CD4+ T cells or memory Treg, in the co-culture instead of total PBMCs would result in a faster increase in cytotoxic CD4+ T cells and subsequent HRS cell killing.

Next to CD4+ T cells, other cell types lacking in our model might be relevant for tumor cell killing. Although we started with total PBMCs which include monocytes, we did not add additional stimulation for other cell types, which might lead to depletion of monocytes and lack of TAMs during our co-cultures. TAMs can be highly relevant, as one of the earliest changes after PD-1 blockade treatment was strong reduction of PD-L1+ TAMs residing close to the HRS cells while the total TAM content did not change [47]. Another study showed that loss or suppression of PD-L1 expression on macrophages led to a decreased expression of tumor promoting M2 macrophage markers and an increased expression of anti-tumor M1 macrophage markers [61]. Similarly, PD-L1 blockade treatment also resulted in a decrease of M2-like and an increase in M1-like markers on TAMs [62]^. These changes in the macrophage department resulted in a pro-inflammatory TME, increased T cell proliferation, T cell cytotoxic potential and tumor reduction [61,62].

Altogether, this indicates that PD-1-PD-L1 blockade could remodel the macrophage

(14)

7

compartment by affecting the M1/M2 polarization, resulting in a pro-inflammatory and immunostimulatory phenotype necessary for tumor cell killing. PD-1 blockade will also directly disturb interactions between PD-L1+ TAMs and PD-1+ T cells or PD-1+ NK cells contributing to their activation [42]. If the M1-like TAMs that appear after ICB are newly differentiated or transformed from M2-like TAMs is currently unclear and needs to be studied. To investigate the effect of TAMs in our model, we can add in vitro M1 or M2 polarized macrophages [63] to the long-term HL model on day 7 together with the start of treatment. It would be interesting to see if M2 macrophages polarize to M1 macrophages after PD-1 blockade and whether addition of macrophages indeed results in increased tumor cell killing.

Figure 3. Long-term in vitro model to study effects of immune checkpoint blockade in Hodgkin lymphoma. In the current model PBMCs and irradiated (IR) HL cell lines are co-cultured for 7 days in a ratio of 10:1. On day 7 new IR HL cell lines or viable non-IR HL cell lines and IC inhibitors are added. Read outs are between day 8 and day 11. White boxes indicate additional steps that can be taken to improve the model.

New treatment options and translation of the model

After additional optimization of the model and characterization of the cell types/molecules important for treatment success, additional new (combination) treatments can be tested. These include, among others ICB targeting LAG-3 or TIM-3, which are both frequently expressed in TME of HL [33]. In addition, small molecule inhibitors of TOX represent an interesting novel treatment option [64]. Testing different treatments alone or combined, will result in the identification of effective and synergistic combinations which will aid in the design of future clinical trials with the most promising treatment combinations. To

(15)

get even closer to the in vivo situation the model might benefit from 3D co-culture systems described for follicular lymphoma (FL) and ovarian cancer [65,66]. In these models it is possible to co-culture primary tumor cells or cell lines with patient derived tumor infiltrating lymphocytes and TAMs for up to 15 days. Ultimately, this model might not only be suitable for HL, but also for other malignancies in which similar cell types as we characterized in HL can be found, such as FL and DLBCL [67,68]. Patients with FL and DLBCL can often not be cured, but treatment with ICB showed some beneficial effects. Although the effects were limited to a small subgroup of patients, responses were often durable and in FL, patients that responded had higher numbers of infiltrating T cells [69-71]. Knowledge obtained in the HL model might be translated to the these lymphomas by identifying whether the specific CD4+ T cell subsets responding to ICB in HL are also present and this might allow identification of specific subsets of lymphoma patients that are likely to respond to ICB.

Using this model will likely spare efforts and costs in health care and limit unnecessary patient participation on less promising treatment designs.

(16)

7 References

1. Delabie J, Chan WC, Weisenburger DD, De Wolf-Peeters C. The antigen-presenting cell function of Reed- Sternberg cells. Leuk Lymphoma. 1995;18(1-2):35-40.

2. Fromm JR, Kussick SJ, Wood BL. Identification and purification of classical Hodgkin cells from lymph nodes by flow cytometry and flow cytometric cell sorting. Am J Clin Pathol. 2006;126(5):764-780.

3. Rengstl B, Schmid F, Weiser C, et al. Tumor-infiltrating HLA-matched CD4(+) T cells retargeted against Hodgkin and Reed-Sternberg cells. Oncoimmunology. 2016;5(6):e1160186.

4. Sanders ME, Makgoba MW, Sussman EH, Luce GE, Cossman J, Shaw S. Molecular pathways of adhesion in spontaneous rosetting of T-lymphocytes to the Hodgkin's cell line L428. Cancer Res. 1988;48(1):37-40.

5. Schneider M, Schneider S, Zuhlke-Jenisch R, et al. Alterations of the CD58 gene in classical Hodgkin lymphoma. Genes Chromosomes Cancer. 2015;54(10):638-645.

6. Steidl C, Shah SP, Woolcock BW, et al. MHC class II transactivator CIITA is a recurrent gene fusion partner in lymphoid cancers. Nature. 2011;471(7338):377-381.

7. Abdul Razak FR, Diepstra A, Visser L, van den Berg A. CD58 mutations are common in Hodgkin lymphoma cell lines and loss of CD58 expression in tumor cells occurs in Hodgkin lymphoma patients who relapse. Genes Immun. 2016;17(6):363-366.

8. Nijland M, Veenstra RN, Visser L, et al. HLA dependent immune escape mechanisms in B-cell lymphomas:

Implications for immune checkpoint inhibitor therapy? Oncoimmunology. 2017;6(4):e1295202.

9. Diepstra A, van Imhoff GW, Karim-Kos HE, et al. HLA class II expression by Hodgkin Reed-Sternberg cells is an independent prognostic factor in classical Hodgkin's lymphoma. J Clin Oncol. 2007;25(21):3101-3108.

10. Roemer MGM, Redd RA, Cader FZ, et al. Major Histocompatibility Complex Class II and Programmed Death Ligand 1 Expression Predict Outcome After Programmed Death 1 Blockade in Classic Hodgkin Lymphoma. J Clin Oncol. 2018;36(10):942-950.

11. Liang WS, Vergilio JA, Salhia B, et al. Comprehensive Genomic Profiling of Hodgkin Lymphoma Reveals Recurrently Mutated Genes and Increased Mutation Burden. Oncologist. 2019;24(2):219-228.

12. Liu Y, Sattarzadeh A, Diepstra A, Visser L, van den Berg A. The microenvironment in classical Hodgkin lymphoma: an actively shaped and essential tumor component. Semin Cancer Biol. 2014;24:15-22.

13. Veldman J, Visser L, Berg AVD, Diepstra A. Primary and acquired resistance mechanisms to immune checkpoint inhibition in Hodgkin lymphoma. Cancer Treat Rev. 2020;82:101931.

14. Cader FZ, Schackmann RCJ, Hu X, et al. Mass cytometry of Hodgkin lymphoma reveals a CD4(+) regulatory T- cell-rich and exhausted T-effector microenvironment. Blood. 2018;132(8):825-836.

15. Marshall NA, Christie LE, Munro LR, et al. Immunosuppressive regulatory T cells are abundant in the reactive lymphocytes of Hodgkin lymphoma. Blood. 2004;103(5):1755-1762.

16. Ye C, Brand D, Zheng SG. Targeting IL-2: an unexpected effect in treating immunological diseases. Signal Transduct Target Ther. 2018;3:2-017-0002-5. eCollection 2018.

17. Zheng SG, Wang J, Wang P, Gray JD, Horwitz DA. IL-2 is essential for TGF-beta to convert naive CD4+CD25- cells to CD25+Foxp3+ regulatory T cells and for expansion of these cells. J Immunol. 2007;178(4):2018-2027.

18. Ma Y, Visser L, Blokzijl T, et al. The CD4+CD26- T-cell population in classical Hodgkin's lymphoma displays a distinctive regulatory T-cell profile. Lab Invest. 2008;88(5):482-490.

19. Poppema S. Immunology of Hodgkin's disease. Baillieres Clin Haematol. 1996;9(3):447-457.

20. Fleischer B. CD26: a surface protease involved in T-cell activation. Immunol Today. 1994;15(4):180-184.

21. Klemann C, Wagner L, Stephan M, von Horsten S. Cut to the chase: a review of CD26/dipeptidyl peptidase- 4's (DPP4) entanglement in the immune system. Clin Exp Immunol. 2016;185(1):1-21.

22. Iellem A, Mariani M, Lang R, et al. Unique chemotactic response profile and specific expression of chemokine receptors CCR4 and CCR8 by CD4(+)CD25(+) regulatory T cells. J Exp Med. 2001;194(6):847-853.

23. Ishida T, Ishii T, Inagaki A, et al. Specific recruitment of CC chemokine receptor 4-positive regulatory T cells in Hodgkin lymphoma fosters immune privilege. Cancer Res. 2006;66(11):5716-5722.

(17)

24. Niens M, Visser L, Nolte IM, et al. Serum chemokine levels in Hodgkin lymphoma patients: highly increased levels of CCL17 and CCL22. Br J Haematol. 2008;140(5):527-536.

25. Alfei F, Kanev K, Hofmann M, et al. TOX reinforces the phenotype and longevity of exhausted T cells in chronic viral infection. Nature. 2019;571(7764):265-269.

26. Balanca CC, Salvioni A, Scarlata CM, et al. PD-1 blockade restores helper activity of tumor-infiltrating, exhausted PD-1hiCD39+ CD4 T cells. JCI Insight. 2021;6(2):10.1172/jci.insight.142513.

27. Yao C, Sun HW, Lacey NE, et al. Single-cell RNA-seq reveals TOX as a key regulator of CD8(+) T cell persistence in chronic infection. Nat Immunol. 2019;20(7):890-901.

28. Wang X, He Q, Shen H, et al. TOX promotes the exhaustion of antitumor CD8(+) T cells by preventing PD1 degradation in hepatocellular carcinoma. J Hepatol. 2019;71(4):731-741.

29. Yoshitomi H, Kobayashi S, Miyagawa-Hayashino A, et al. Human Sox4 facilitates the development of CXCL13-producing helper T cells in inflammatory environments. Nat Commun. 2018;9(1):3762-018-06187-0.

30. Annibali O, Bianchi A, Grifoni A, et al. A novel scoring system for TIGIT expression in classic Hodgkin lymphoma. Sci Rep. 2021;11(1):7059-021-86655-8.

31. Aoki T, Chong LC, Takata K, et al. Single-Cell Transcriptome Analysis Reveals Disease-Defining T-cell Subsets in the Tumor Microenvironment of Classic Hodgkin Lymphoma. Cancer Discov. 2020;10(3):406-421.

32. Carey CD, Gusenleitner D, Lipschitz M, et al. Topological analysis reveals a PD-L1-associated microenvironmental niche for Reed-Sternberg cells in Hodgkin lymphoma. Blood. 2017;130(22):2420-2430.

33. El Halabi L, Adam J, Gravelle P, et al. Expression of the Immune Checkpoint Regulators LAG-3 and TIM-3 in Classical Hodgkin Lymphoma. Clin Lymphoma Myeloma Leuk. 2021;21(4):257-266.e3.

34. Patel SS, Weirather JL, Lipschitz M, et al. The microenvironmental niche in classic Hodgkin lymphoma is enriched for CTLA-4-positive T cells that are PD-1-negative. Blood. 2019;134(23):2059-2069.

35. Brockelmann PJ, Goergen H, Keller U, et al. Efficacy of Nivolumab and AVD in Early-Stage Unfavorable Classic Hodgkin Lymphoma: The Randomized Phase 2 German Hodgkin Study Group NIVAHL Trial. JAMA Oncol.

2020;6(6):872-880.

36. Ramchandren R, Domingo-Domenech E, Rueda A, et al. Nivolumab for Newly Diagnosed Advanced-Stage Classic Hodgkin Lymphoma: Safety and Efficacy in the Phase II CheckMate 205 Study. J Clin Oncol.

2019;37(23):1997-2007.

37. Voltin CA, Mettler J, van Heek L, et al. Early Response to First-Line Anti-PD-1 Treatment in Hodgkin Lymphoma: A PET-Based Analysis from the Prospective, Randomized Phase II NIVAHL Trial. Clin Cancer Res.

2021;27(2):402-407.

38. Tumeh PC, Harview CL, Yearley JH, et al. PD-1 blockade induces responses by inhibiting adaptive immune resistance. Nature. 2014;515(7528):568-571.

39. Green MR, Monti S, Rodig SJ, et al. Integrative analysis reveals selective 9p24.1 amplification, increased PD- 1 ligand expression, and further induction via JAK2 in nodular sclerosing Hodgkin lymphoma and primary mediastinal large B-cell lymphoma. Blood. 2010;116(17):3268-3277.

40. Hollander P, Kamper P, Smedby KE, et al. High proportions of PD-1(+) and PD-L1(+) leukocytes in classical Hodgkin lymphoma microenvironment are associated with inferior outcome. Blood Adv. 2017;1(18):1427-1439.

41. Ferrarini I, Rigo A, Visco C, Krampera M, Vinante F. The Evolving Knowledge on T and NK Cells in Classic Hodgkin Lymphoma: Insights into Novel Subsets Populating the Immune Microenvironment. Cancers (Basel).

2020;12(12):10.3390/cancers12123757.

42. Vari F, Arpon D, Keane C, et al. Immune evasion via PD-1/PD-L1 on NK cells and monocyte/macrophages is more prominent in Hodgkin lymphoma than DLBCL. Blood. 2018;131(16):1809-1819.

43. Merryman RW, Armand P, Wright KT, Rodig SJ. Checkpoint blockade in Hodgkin and non-Hodgkin lymphoma. Blood Adv. 2017;1(26):2643-2654.

44. Okuma Y, Wakui H, Utsumi H, et al. Soluble Programmed Cell Death Ligand 1 as a Novel Biomarker for Nivolumab Therapy for Non-Small-cell Lung Cancer. Clin Lung Cancer. 2018;19(5):410-417.e1.

45. Zhou J, Mahoney KM, Giobbie-Hurder A, et al. Soluble PD-L1 as a Biomarker in Malignant Melanoma Treated with Checkpoint Blockade. Cancer Immunol Res. 2017;5(6):480-492.

(18)

7

46. Ando K, Hamada K, Watanabe M, et al. Plasma Levels of Soluble PD-L1 Correlate With Tumor Regression in Patients With Lung and Gastric Cancer Treated With Immune Checkpoint Inhibitors. Anticancer Res.

2019;39(9):5195-5201.

47. Reinke S, Brockelmann PJ, Iaccarino I, et al. Tumor and microenvironment response but no cytotoxic T-cell activation in classic Hodgkin lymphoma treated with anti-PD1. Blood. 2020;136(25):2851-2863.

48. Sasse S, Reddemann K, Diepstra A, et al. Programmed cell death protein-1 (PD-1)-expression in the microenvironment of classical Hodgkin lymphoma at relapse during anti-PD-1-treatment. Haematologica.

2019;104(1):e21-e24.

49. Diefenbach CS, Hong F, Ambinder RF, et al. Ipilimumab, nivolumab, and brentuximab vedotin combination therapies in patients with relapsed or refractory Hodgkin lymphoma: phase 1 results of an open-label, multicentre, phase 1/2 trial. Lancet Haematol. 2020;7(9):e660-e670.

50. Denzin LK and Cresswell P. HLA-DM induces CLIP dissociation from MHC class II alpha beta dimers and facilitates peptide loading. Cell. 1995;82(1):155-165.

51. Sherman MA, Weber DA, Jensen PE. DM enhances peptide binding to class II MHC by release of invariant chain-derived peptide. Immunity. 1995;3(2):197-205.

52. Hombrink P, Hassan C, Kester MG, et al. Discovery of T cell epitopes implementing HLA-peptidomics into a reverse immunology approach. J Immunol. 2013;190(8):3869-3877.

53. Kampstra ASB, van Heemst J, Janssen GM, et al. Ligandomes obtained from different HLA-class II-molecules are homologous for N- and C-terminal residues outside the peptide-binding cleft. Immunogenetics. 2019;71(8- 9):519-530.

54. Bolognesi C, Forcato C, Buson G, et al. Digital Sorting of Pure Cell Populations Enables Unambiguous Genetic Analysis of Heterogeneous Formalin-Fixed Paraffin-Embedded Tumors by Next Generation Sequencing. Sci Rep.

2016;6:20944.

55. Juskevicius D, Jucker D, Dietsche T, et al. Novel cell enrichment technique for robust genetic analysis of archival classical Hodgkin lymphoma tissues. Lab Invest. 2018;98(11):1487-1499.

56. Mangano C, Ferrarini A, Forcato C, et al. Precise detection of genomic imbalances at single-cell resolution reveals intra-patient heterogeneity in Hodgkin's lymphoma. Blood Cancer J. 2019;9(12):92-019-0256-y.

57. Wienand K, Chapuy B, Stewart C, et al. Genomic analyses of flow-sorted Hodgkin Reed-Sternberg cells reveal complementary mechanisms of immune evasion. Blood Adv. 2019;3(23):4065-4080.

58. van der Lee DI, Reijmers RM, Honders MW, et al. Mutated nucleophosmin 1 as immunotherapy target in acute myeloid leukemia. J Clin Invest. 2019;129(2):774-785.

59. Zhang R, Lyu C, Lu W, Pu Y, Jiang Y, Deng Q. Synergistic effect of programmed death-1 inhibitor and programmed death-1 ligand-1 inhibitor combined with chemotherapeutic drugs on DLBCL cell lines in vitro and in vivo. Am J Cancer Res. 2020;10(9):2800-2812.

60. Tanijiri T, Shimizu T, Uehira K, et al. Hodgkin's reed-sternberg cell line (KM-H2) promotes a bidirectional differentiation of CD4+CD25+Foxp3+ T cells and CD4+ cytotoxic T lymphocytes from CD4+ naive T cells. J Leukoc Biol. 2007;82(3):576-584.

61. Zhang Y, Du W, Chen Z, Xiang C. Upregulation of PD-L1 by SPP1 mediates macrophage polarization and facilitates immune escape in lung adenocarcinoma. Exp Cell Res. 2017;359(2):449-457.

62. Xiong H, Mittman S, Rodriguez R, et al. Anti-PD-L1 Treatment Results in Functional Remodeling of the Macrophage Compartment. Cancer Res. 2019;79(7):1493-1506.

63. Zarif JC, Hernandez JR, Verdone JE, Campbell SP, Drake CG, Pienta KJ. A phased strategy to differentiate human CD14+monocytes into classically and alternatively activated macrophages and dendritic cells.

BioTechniques. 2016;61(1):33-41.

64. Agrawal V, Su M, Huang Y, Hsing M, Cherkasov A, Zhou Y. Computer-Aided Discovery of Small Molecule Inhibitors of Thymocyte Selection-Associated High Mobility Group Box Protein (TOX) as Potential Therapeutics for Cutaneous T-Cell Lymphomas. Molecules. 2019;24(19):10.3390/molecules24193459.

65. Long L, Yin M, Min W. 3D Co-culture System of Tumor-associated Macrophages and Ovarian Cancer Cells.

Bio Protoc. 2018;8(8):10.21769/BioProtoc.2815.

(19)

66. Wagar LE, Sworder B, Khodadoust MS, Davis MM, Alizadeh AA. Follicular Lymphoma Organoids for Investigating the Tumor Microenvironment. Blood. 2019;134 (Supplement_1):2799.

67. Maestre L, Garcia-Garcia JF, Jimenez S, et al. High-mobility group box (TOX) antibody a useful tool for the identification of B and T cell subpopulations. PLoS One. 2020;15(2):e0229743.

68. Roider T, Seufert J, Uvarovskii A, et al. Dissecting intratumour heterogeneity of nodal B-cell lymphomas at the transcriptional, genetic and drug-response levels. Nat Cell Biol. 2020;22(7):896-906.

69. Armand P, Janssens A, Gritti G, et al. Efficacy and safety results from CheckMate 140, a phase 2 study of nivolumab for relapsed/refractory follicular lymphoma. Blood. 2021;137(5):637-645.

70. Ansell SM, Minnema MC, Johnson P, et al. Nivolumab for Relapsed/Refractory Diffuse Large B-Cell Lymphoma in Patients Ineligible for or Having Failed Autologous Transplantation: A Single-Arm, Phase II Study. J Clin Oncol. 2019;37(6):481-489.

71. Xie W, Medeiros LJ, Li S, Yin CC, Khoury JD, Xu J. PD-1/PD-L1 Pathway and Its Blockade in Patients with Classic Hodgkin Lymphoma and Non-Hodgkin Large-Cell Lymphomas. Curr Hematol Malig Rep. 2020;15(4):372- 381.

(20)

7

(21)