Cellular Composition and Contribution of Tertiary Lymphoid Structures to Tumor Immune Infiltration and Modulation by Radiation Therapy

Abstract

Immune-based anti-cancer strategies combined with radiation therapy (RT) are actively being investigated but many questions remain, such as the ideal treatment scheme and whether a potent immune response can be generated both locally and systemically. In this context, tumor-associated tertiary lymphoid structures (TLS) have become a subject of research. While TLS are present in several types of cancer with strong similarities, they are especially relevant in medullary breast carcinoma (MBC). This suggests that MBC patients are ideally suited for investigating this question and may benefit from adapted therapeutic options. As RT is a corner-stone of MBC treatment, investigating interactions between RT and TLS composition is also clinically relevant. We thus first characterized the lymphoid structures associated with MBC in a patient case report and demonstrated that they closely resemble the TLS observed in a genetical mouse model. In this model, we quantitatively and qualitatively investigated the cellular composition of the tumor-associated TLS. Finally, we investigated TLS regulation after hypo-fractionated RT and showed that RT induced their acute and transient depletion, followed by a restoration phase. This study is the first work to bring a comprehensive and timely characterization of tumor-associated TLS in basal conditions and after RT. It highlights cellular targets (i.e., Tregs) that could be selectively modulated in subsequent studies to optimize anti-tumor immune response. The study of TLS modulation is worth further investigation in the context of RT and personalized medicine.

Keywords: KP model; medullary breast carcinoma; microenvironment; radiation therapy; tertiary lymphoid structures.

Figure 1

Immune infiltrate in MBC (patient 1, P1) and basal breast carcinoma (patient 2, P2). Serial histological samples of P1 and P2 were stained with routine markers. Stainings showed the following repartition: estrogen receptor [(A–K) ER: P1−, P2−], progesterone receptor [(B–L) PR: P1−, P2−], human epidermal growth factor 2 [(C–M) HER2: P1−, P2+], Ecadherin [(D–N) Ecadh: P1+, P2+], P1−, P2−], and Ki67 [(E–O): P1 30% in tumor and tertiary lymphoid structures (TLS); P2 5% in tumor and 0% in immune aggregates]. Additional staining with non-routine markers was performed. They allowed characterization of TLS in P1 but not in P2: B cells [CD20 (F–P) P1 germinal center, P2 scarce cells], CD4 T cells [CD4 (G–Q) P1: present in all compartments; P2: non organized], CD8 T cells [CD8 (H–R) P1: restricted to the TLS periphery and invading the tumor; P2: rare in the aggregates and absent in the tumor]. IF recapitulating the key immune components, including PD1, FoxP3 CD8, and tumor cytokeratine+ are shown in (I) and (S). PD1+ cells are absent in P2. (T) Intra-tumoral number of CD8 and CD4 cells in P1 and P2.

Figure 1

Immune infiltrate in MBC (patient 1, P1) and basal breast carcinoma (patient 2, P2). Serial histological samples of P1 and P2 were stained with routine markers. Stainings showed the following repartition: estrogen receptor [(A–K) ER: P1−, P2−], progesterone receptor [(B–L) PR: P1−, P2−], human epidermal growth factor 2 [(C–M) HER2: P1−, P2+], Ecadherin [(D–N) Ecadh: P1+, P2+], P1−, P2−], and Ki67 [(E–O): P1 30% in tumor and tertiary lymphoid structures (TLS); P2 5% in tumor and 0% in immune aggregates]. Additional staining with non-routine markers was performed. They allowed characterization of TLS in P1 but not in P2: B cells [CD20 (F–P) P1 germinal center, P2 scarce cells], CD4 T cells [CD4 (G–Q) P1: present in all compartments; P2: non organized], CD8 T cells [CD8 (H–R) P1: restricted to the TLS periphery and invading the tumor; P2: rare in the aggregates and absent in the tumor]. IF recapitulating the key immune components, including PD1, FoxP3 CD8, and tumor cytokeratine+ are shown in (I) and (S). PD1+ cells are absent in P2. (T) Intra-tumoral number of CD8 and CD4 cells in P1 and P2.

Figure 2

Tertiary lymphoid structures (TLS) appear after tumor onset are maintained during disease progression and share common features with TLS in MBC. (A) Full KP lung section H&E stained. TLS are shown by red arrows, they are located around bronchus. Analysis on serial sections is required for proper TLS detection that can be present in one section and absent in 50 µm further. Green arrow showed the area of macro-dissected tumor used for flow cytometry. (B) KP lungs were sampled at 7, (timing of tumor occurrence), 15, and 20 weeks (final stage) post-cre installation. Tumor and TLS quantification were performed based on H&E staining similar to (A). TLS appeared after tumors’ onset. (C) Representative images are shown and illustrate common properties of TLS from KP and MBC such as the presence of germinal center and organizational features (1), 30% of proliferation (2). Additional stainings were performed. They show expression of CXCL13 chemokine (3), down left corner positive signal comes from the tumor (T), presence of Prox1⁺ Lyve1⁺ CD31⁺ lymphatic vessels (4). Picture 5 shows presence of aligned CXCR4⁺ cells in between TLS and tumor T. Picture 6 shows TLS engulfed by the tumor. Common features of KP versus MBC TLS are summarized in table (D).

Figure 3

Presence of tertiary lymphoid structures (TLS) near tumors is correlated with tumor apoptosis, independently from quantitative B or T cell infiltration. Cleaved Caspase3 co-stainings with CD45 (immune), EpCap (tumor), CD3 (T cells) and B220 (B cells) apoptotic cells are mostly tumour cells (A). Control tumors with or without TLS were stained to assess wether the immune contexture correlates with apoptosis (White bands delineate tumors) (B). Quantification per unit surface and T/B cell ratio is shown is (C).

Figure 4

Radiation therapy (RT) increases tertiary lymphoid structures (TLS) apoptosis and their transient depletion. Apoptosis quantification measured by cleaved caspase 3 in TLS in control of KP TLS, 2 h and 14 days post-RT. TLS size was measured with Zen Zeiss software (A). Quantification per surface unit is showed in (B). The prevalence of CD3+ T cells was assessed in TLS and tumours to evaluate intrinsic response to RT of those cells in different compartments (C).

Figure 5

Histological and flow cytometry analysis correlation highlights granulocytes as biomarker of intra-tumor immune compartment and B cells as biomarker of tertiary lymphoid structures (TLS). Macro-dissected tumor H&E shows presence of TLS in samples. Granulocytes are almost absent in TLS and very abundant in tumors (A). CD68 and CD11c+ cells that include macrophages, monocytes, and DCs, but not granulocytes, are present in both compartments but display different morphological features (B). Tumor versus TLS quantification for CD68, CD11c, or granulocytes is shown in (C). Granulocytes are histologically much more prevalent than B cells in the tumor compartment (D). Size histological comparison of TLS and tumors is shown in (E). Strong prevalence of B cells obtained by Flow Cytometry analysis suggests TLS are also analysed with this method (F).

Figure 6

Flow cytometry analysis confirms transient depletion of tertiary lymphoid structures (TLS) following RT. (A) The B cells/granulocyte ratio confirms with histological analysis of the TLS depletion (B). Tumor+ TLS lymphoid and granulocytes variations are presented in % of whole immune cells analyzed after macrodissection (C). T/B cells, CD8/Tregs, and CD4/CD8 ratio (D). Numeration of CD103 positive cells amongst CD8 cells (E).