Comparison and analysis of the immune landscape at the tumour invasion front in patients with pMMR/MSI-H and pMMR/MSS colorectal cancer.

Introduction
According to the GLOBOCAN database of the World Health Organization, colorectal cancer (CRC) is the third most common cancer globally and the second leading cause of cancer-related deaths worldwide. Data from 2023 show that there are approximately 1.93 million new cases and 935,000 deaths from CRC each year globally [1, 2]. The incidence and mortality rates of CRC are increasing annually, and it is projected that by 2040, the number of new cases will reach 3.2 million, and the number of deaths will reach 1.6 million [3].
The classification of CRC into proficient and deficient mismatch repair (pMMR/MSI-H and pMMR/MSS) subtypes has profound therapeutic implications, particularly in the era of immunotherapy [4]. The dMMR phenotype, characterised by a high tumour mutational burden and microsatellite instability, predicts a favourable response to immune checkpoint blockade, whereas pMMR tumours are largely refractory to such treatment [5–7]. This pivotal distinction has established immunotherapy as a standard for metastatic dMMR CRC and underscores the critical need to understand the underlying immune mechanisms driving this differential response [8, 9].

However, current understanding is predominantly derived from bulk genomic and transcriptomic analyses that lack spatial context. The tumour invasion front—a critical interface where tumour cells interact with the host stroma and immune system—is a key anatomical site determining tumour progression and treatment outcome [10]. We hypothesise that the immune activity at this frontier may fundamentally differ between pMMR/MSI-H and pMMR/MSS cancers, yet a detailed, spatially resolved comparison of the immune landscape in this specific region is lacking.
To address this gap, we employed high-plex multiplex immunofluorescence (IF) to perform a quantitative and spatial analysis of the immune microenvironment at the invasive margin of CRC specimens. Although bulk analyses have established transcriptional differences, they cannot reveal the spatial organisation of immune cells—a critical factor in anti-tumour efficacy as it dictates cell–cell interactions and functional capabilities [11, 12]. Multiplex IF is therefore necessary to decode the complex cellular interactions and functional states within the architecturally preserved tumour microenvironment, particularly at the invasion front where these interactions are most dynamic and clinically relevant [11, 13]. This study is the first to provide a comprehensive, spatially resolved profile of both adaptive and innate immune cells (including T-cell subsets and NK-cell subpopulations) specifically at the tumour invasion front in a cohort of patients with pMMR/MSI-H and pMMR/MSS CRC. Our objective was to identify unique immune features that could elucidate the mechanistic basis for the disparate immunotherapy responses and reveal potential therapeutic targets for overcoming resistance in pMMR disease.

Materials and methods

Study design and participants
This retrospective study analysed samples from 51 patients with CRC. The study spatially profiled the immune landscape at the tumour invasion front in patients with colorectal cancer, stratified by MMR status (32 pMMR vs. 19 dMMR). The design included the use of immunohistochemistry and molecular assays for MMR/MSI classification, coupled with multiplex immunofluorescence for high-plex, spatial analysis of immune cell subsets within the architecturally preserved tumour microenvironment. This study was conducted and reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines for cross-sectional studies. The detailed participant criteria are as follows.
The inclusion criteria were as follows: (1) first-time diagnosis during the visit; (2) clear pathological diagnosis of primary colorectal adenocarcinoma; (3) primary lesion located in the colorectum; (4) complete case records; and (5) availability of sufficient formalin-fixed paraffin-embedded (FFPE) tumour tissue samples for MMR testing and multiplex IF analysis. Patients across all tumour–node–metastasis stages (I–IV) were included.
The exclusion criteria were as follows: (1) other primary malignant tumours with colorectal metastasis; (2) patients with metastatic CRC at initial diagnosis; (3) concurrent infections or severe diseases of the heart, brain, liver or other organs (e.g. severe heart failure (NYHA class III/IV), liver cirrhosis (Child–Pugh class B/C) or chronic kidney disease (stage 4–5)); (4) haematological diseases (e.g. leukaemia, lymphoma); (5) history of autoimmune diseases or conditions requiring long-term immunosuppressive therapy; and (6) any other known diseases or conditions that could considerably alter immune status or interfere with the interpretation of the study results, as determined by the investigators. The ethics committee approved this study, and all patients provided written informed consent.

Identification of patients with pMMR/MSI-H and pMMR/MSS CRC
Immunohistochemical detection of MLH1, PMS2, MSH2 and MSH6 was used to determine MMR function in patients with CRC. Tumour tissue samples and normal intestinal mucosa samples taken 5 cm away from the tumour were collected as controls. The tissues were processed through standard procedures, including dehydration, clarification and embedding to produce paraffin sections with a thickness of 5 μm. During immunohistochemical staining, the sections were first dewaxed and rehydrated. Heat-induced or enzyme-mediated antigen retrieval was then performed to expose antigens in the tissues. The sections were blocked with BSA blocking solution to reduce nonspecific binding. Subsequently, specific antibodies for MLH1, PMS2, MSH2 and MSH6 were added dropwise, and the sections were incubated in a humidified chamber for 1–2 h. After incubation, unbound primary antibodies were washed away with phosphate-buffered saline, and the corresponding secondary antibodies were added and incubated again. Finally, 3,3′-diaminobenzidine chromogenic reagent was applied, and the colour development was observed under a microscope, with the reaction terminated promptly.
When interpreting the results, the expression of MLH1, PMS2, MSH2 and MSH6 proteins in the nuclei of tumour cells should be examined under a microscope. If all four proteins are fully expressed in the nuclei of tumour cells, the case is classified as pMMR. If one or more proteins are absent in the nuclei of tumour cells, the case is classified as dMMR. During interpretation, positive expression of these proteins in normal epithelial cells, lymphocytes and stromal cells surrounding the tumour serves as an internal control to ensure accuracy. Throughout the experimental process, it is essential to strictly adhere to standard operating procedures to maintain sample quality, reagent selection and consistency of staining conditions. Results should be interpreted by experienced pathologists, and verification by a second reviewer may be necessary to ensure reliability and accuracy.

Multiplex fluorescence staining
Multiplex IF staining was performed on FFPE tissue sections using a standardised multiplex fluorescent immunohistochemistry protocol based on tyramide signal amplification (TSA). Briefly, sections were deparaffinised, rehydrated and subjected to antigen retrieval in citrate buffer (pH 6.0). Endogenous peroxidase activity was blocked, and sections were incubated with a protein-blocking solution.
A sequential staining protocol was employed using the following primary antibodies: CD4 (25229, CST, MA, USA), CD8α (98941, CST, MA, USA), CK-Pan (4545, CST, MA, USA), IFN-γ (8455, CST, MA, USA), CD16 (24326, CST, MA, USA) and CD56 (99746, CST, MA, USA). For each cycle, sections were incubated with a primary antibody followed by a horseradish peroxidase-conjugated secondary antibody and then the corresponding TSA fluorophore. After each staining cycle, antibody elution was performed using microwave heating in antigen retrieval buffer to remove bound antibodies before proceeding to the next cycle. This process was repeated sequentially for all markers.
The following TSA reagents were used for detection: iF440-TSA (G1250, Servicebio, Wuhan, China), iF488-TSA (G1231, Servicebio, Wuhan, China), iF546-TSA (G1251, Servicebio, Wuhan, China) and iF555-TSA (G1233, Servicebio, Wuhan, China). After completion of all staining cycles, slides were mounted with an anti-fade mounting medium.
Image acquisition was performed using a Vectra/PhenoImager HT system (Akoya Biosciences), and subsequent multispectral image analysis, including cell segmentation and phenotyping, was conducted using inForm or HALO image analysis software.

Data statistics and analysis
Statistical analysis was performed using R version 4.3.2 and SPSS version 26.0 software. For measurement data following a normal distribution, the mean ± standard deviation (x ± s) was used to represent the data, and t-tests were conducted. For measurement data not following a normal distribution, the median (interquartile range) (M [P25, P75]) was used, and the Mann–Whitney U test or Wilcoxon rank-sum test was applied. Count data were expressed as percentages and analysed using the chi-square (χ2) test or Fisher’s exact test. A p-value less than 0.05 was considered statistically significant.

Results

Patient information
A total of 51 patients with CRC were included in this study, with 19 (37.3%) patients in the dMMR group and 32 (62.7%) in the pMMR group. There were 30 (58.8%) men and 21 (41.2%) women. The median age was similar between the dMMR group (73 years) and the pMMR group (70 years), with no significant difference. The average tumour diameter was 5.72 cm in the dMMR group and 4.23 cm in the pMMR group. In the dMMR group, most patients (89.5%) presented with localised disease (stages I–III), whereas only two patients (10.5%) had advanced stage IV disease. Similarly, in the pMMR group, 29 patients (90.6%) had localised disease (stages I–III), and 3 patients (9.4%) were diagnosed with stage IV cancer. There were no significant differences between the two groups in terms of clinical or pathological characteristics (p > 0.05) (Table 1).

Analysis of T-cell landscape at the tumour invasion front in patients with pMMR/MSI-H and pMMR/MSS CRC
Multiplex IF staining was employed to analyse the immune cell landscape and microenvironment at the tumour invasion front in CRC tissues from patients in both the pMMR/MSI-H and pMMR/MSS groups. The results indicated that the proportion of CD8⁺ T cells at the tumour invasion front was significantly higher in patients with dMMR CRC than in those with pMMR CRC (26.84% ± 3.17% vs. 6.29% ± 1.62%, p < 0.001). Conversely, the proportion of CD4⁺ T cells was significantly lower in patients with dMMR CRC than in those with pMMR CRC (19.02% ± 2.81% vs. 37.71% ± 3.52%, p < 0.001) (Fig. 1A, B). In terms of spatial distribution, in the dMMR samples, CD8⁺ T cells were more concentrated within a distance of 5–15 μm from the tumour tissue, whereas in the pMMR samples, the distribution of CD8⁺ T cells was more uniform (Fig. 1C). Regarding CD4⁺ T cells, there was no significant difference in their distribution between the two groups.

Analysis of NK cell landscape at the tumour invasion front in patients with pMMR/MSI-H and pMMR/MSS CRC
Results from multiplex IF staining showed that there was no significant difference in the total number of NK cells at the tumour invasion front between the two groups. However, the proportion of CD56 bright⁺ cells was significantly higher in patients with dMMR CRC than in those with pMMR CRC (6.69% ± 1.04% vs. 1.93% ± 0.48%, p < 0.001) (Fig. 2A, B).

Discussion
The findings of our study reveal a distinct immune landscape at the tumour invasion front between dMMR and pMMR CRCs. Specifically, dMMR tumours exhibited a considerably higher density of CD8⁺ T cells and a greater proportion of CD56 bright⁺ NK cells, along with a lower proportion of CD4⁺ T cells, than pMMR tumours. Furthermore, CD8⁺ T cells in dMMR cases showed a more concentrated spatial distribution within 5–15 μm of the tumour margin. These pronounced differences in both composition and spatial organisation of immune infiltrates provide a spatial immunological basis for the enhanced response to immunotherapy observed in patients with dMMR CRC in clinical settings.
The substantially higher density of CD8⁺ T cells at the invasion front of dMMR tumours is consistent with the high neoantigen burden characteristic of dMMR/MSI-H CRCs, which is known to foster an immunogenic microenvironment [14, 15]. Our spatial analysis extends this paradigm by demonstrating that these activated cytotoxic T cells are not only more abundant but also positioned in immediate proximity to cancer cells, a spatial configuration likely critical for effective tumour cell recognition and killing. This contrasts sharply with the immune-excluded phenotype often described in pMMR/MSS CRC, where T cells are retained in the stroma [16, 17]. The mechanisms underlying this differential recruitment and positioning warrant further investigation but may involve enhanced chemokine signalling driven by the constitutive interferon signature reported in dMMR tumours [18, 19].
A novel and intriguing finding of our study is the specific enrichment of CD56 bright NK cells at the dMMR invasion front. Although the total NK cell content did not differ, the shift towards the CD56 bright immunoregulatory subset reveals a potential and previously underappreciated role for innate immunity in shaping the adaptive immune response in dMMR CRC. Cytokines such as IFN-γ, TNF-α and GM-CSF are potently produced by CD56 bright NK cells [20–22]. Therefore, we speculate that these cytokines might contribute to the observed CD8⁺ T cell abundance and activation through several potential pathways: IFN-γ can enhance antigen presentation and directly boost CD8⁺ T cell function [23, 24]; TNF-α can promote T cell recruitment and survival [25]; and GM-CSF can support dendritic cell maturation, thereby potentiating T cell priming [26]. Thus, the colocalisation of CD56 bright NK cells and CD8⁺ T cells at the invasion front suggests a potential synergistic interplay between innate and adaptive immunity, which could foster a potent antitumour milieu in dMMR CRC. This spatial association indicates a possible mechanism whereby innate immune cells directly contribute to the efficacy of adaptive immunotherapy in dMMR CRC, moving beyond the traditional focus solely on T cells. However, this hypothesis remains speculative and requires functional validation through co-culture experiments or spatial transcriptomics to confirm causal relationships and specific cytokine dependencies.
However, this study has certain limitations. First, the number of patients with CRC included is relatively small. In future research, more patient cases should be collected to fully validate the experimental conclusions. Second, the reasons behind the increased proportions of CD8+ T cells and CD56 bright NK cells at the tumour invasion front in patients with dMMR CRC require further investigation. Additionally, no correlation could be established between the immune landscape and clinical outcomes such as survival or therapy response, as follow-up data were not available for this cohort. Future studies with long-term clinical data are warranted to determine whether the spatial features identified here possess prognostic or predictive value. Finally, although informative, our multiplex IF panel did not include functional exhaustion markers (e.g. PD-1, TIM-3) on T cells or more specific activation markers on NK cells. Incorporating these markers in future work could provide deeper insights into the functional states of spatially positioned immune cells.

Conclusion
In conclusion, our study suggests a markedly more immunogenic microenvironment at the tumour invasion front of dMMR CRC, characterised by enhanced CD8+ T cell and CD56 bright NK cell infiltration compared with pMMR tumours. These findings provide a new perspective for understanding the mechanisms underlying the differential immunotherapy responses between pMMR/MSI-H and pMMR/MSS CRC and may serve as an important theoretical basis for optimising immunotherapy strategies for pMMR CRC. However, given the small sample size, these results should be considered preliminary. Further validation in larger cohorts is needed to confirm these observations and explore their clinical applicability.

Supplementary Information
Below is the link to the electronic supplementary material.