Tertiary lymphoid structures predict survival and response to neoadjuvant therapy in locally advanced rectal cancer

Abstract

Tertiary lymphoid structure (TLS) contributes to the anti-tumor immune response, and predicts the prognosis of colorectal cancer patients. However, the potential impact of TLS in shaping the immune status of rectal adenocarcinoma, and the intrinsic relationship between TLS and neoadjuvant therapies (neoTx) remain unclear. We performed hematoxylin-eosin staining, immunohistochemical and biomolecular analyses to investigate TLS and tumor-infiltrating lymphocytes (TILs) in 221 neoTx-treated and 242 treatment-naïve locally advanced rectal cancer (LARC) patients. High TLS density was significantly associated with the absence of vascular invasion, a lower neutrophil-to-lymphocyte ratio, increased TLS maturity, a longer recurrence-free survival (RFS) (hazard ratio [HR] 0.2985 95% confidence interval [CI] 0.1894-0.4706, p < 0.0001) and enhanced infiltration of adaptive immune cells. Biomolecular analysis showed that high TLS-score was strongly associated with more infiltration of immune cells and increased activation of immune-related pathways. TLS⁺ tumors in pre-treatment specimens were associated with a higher proportion of good respond (62.5% vs. 29.8%, p < 0.0002) and pathological complete remission (pCR) (40.0% vs. 11.1%, p < 0.0001), and significantly increased RFS (HR 0.3574 95%CI 0.1489-0.8578 p = 0.0213) compared with TLS^- tumors in the neoTx cohort, which was confirmed in GSE119409 and GSE150082. Further studies showed that neoTx significantly reduced TLS density and maturity, and abolished the prognostic value of TLS. Our study illustrates that TLS may have a key role in mediating the T-cell-inflamed tumor microenvironment, which also provides a new direction for neoTx, especially neoadjuvant immunotherapy, in LRAC patients.

Fig. 1. The expression of TLS and its correlates with improved survival in LARC patients.

TLSs are defined as dense aggregates of B cells (CD20, b) with an adjacent T-cell zone (CD3, a) with dendritic cells (DCs, CD11c, c) and lacking the surrounding capsule (scale bars, 200 μm). Germinal centers (GCs) were confirmed to contain strictly confined regions of CD21 follicular dendritic cells (d) in a field of larger CD23 B cells (e) surrounded by PNAd vessels (f) (scale bars, 200 μm). g Distribution of TLS density and TLS maturation category among 242 rectal cancer patients. h The relationship between TLS maturity and TLS density. i RFS was compared in patients with high and low TLS density by Kaplan–Meier survival curve. j RFS was compared in patients with mTLS⁺ and mTLS^- by Kaplan–Meier survival curve. In the graphs, the plotted values encompass the range from the minimum to the maximum, with the inclusion of the median.

Fig. 2. The role of TLSs in the TIME.

a TIL expression in primary resected rectal tissue was categorized into NT patients with low and high TLS density (scale bars, 600 μm and 200 μm). b CD4⁺ T cells, CD8⁺ T cells, CD20⁺ B cells and CD45RO⁺ cells were compared in NT patients with high and low TLS density. c Gene set enrichment analysis with hallmark gene sets in the NT cohort (30 high vs. 30 low TLS density). *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. In the graphs, the plotted values encompass the range from the minimum to the maximum, with the inclusion of the median.

Fig. 3. The interrelationship between TLS and neoTx.

a, b The expression ratio of TLS in pre-neoTx specimens was compared in responders and non-responders, pCR and ipCR. Analysis was performed by Fisher’s exact test. c RFS was compared in patients with TLS⁺ and TLS⁻ tumors by Kaplan–Meier survival curve. d, e 21 TLS-scores were compared in responders and non-responders from GSE119409 (n = 56, P = 0.0186) and GSE150082 (n = 39, P = 0.0402). R, responders; NR, non-responders. TLS density (f) and TLS maturity (g) were compared in post-neoTx with residual tumors and NT patients. RFS was compared by the Kaplan–Meier survival curve in patients with high and low TLS density (h) and mTLS⁺ and mTLS⁻ (i) in the post-neoTx with residual tumors cohort. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. In the graphs, the plotted values encompass the range from the minimum to the maximum, with the inclusion of the median.