https://orcid.org/0000-0002-8191-4044
Figures
Abstract
Background
Pancreatic ductal adenocarcinomas (PDAC) have heterogeneous tumor microenvironments relatively devoid of infiltrating immune cells. We aimed to quantitatively assess infiltrating CD3+ and CD8+ lymphocytes in a treatment-naïve patient cohort and assess associations with overall survival and microenvironment inflammatory proteins.
Methods
Tissue microarrays were immunohistochemically stained for CD3+ and CD8+ lymphocytes and quantitatively assessed using QuPath. Levels of inflammation-associated proteins were quantified by multiplexed, enzyme-linked immunosorbent assay panels on matching tumor and tissue samples.
Results
Our findings revealed a significant increase in both CD3+ and CD8+ lymphocytes populations in PDAC compared with non-PDAC tissue, except when comparing CD8+ percentages in PDAC versus intraductal papillary mucinous neoplasms (IPMN) (p = 0.5012). Patients with quantitatively assessed CD3+ low tumors (lower 50%) had shorter survival (median 273 days) compared to CD3+ high tumors (upper 50%) with a median overall survival of 642.5 days (p = 0.2184). Patients with quantitatively assessed CD8+ low tumors had significantly shorter survival (median 240 days) compared to CD8+ high tumors with a median overall survival of 1059 days (p = 0.0003). Of 41 proteins assessed in the inflammation assay, higher levels of IL-1B and IL-2 were significantly associated with decreased CD3+ infiltration (r = -0.3704, p = 0.0187, and r = -0.4275, p = 0.0074, respectively). Higher levels of IL-1B were also significantly associated with decreased CD8+ infiltration (r = -0.4299, p = 0.0045), but not IL-2 (r = -0.0078, p = 0.9616). Principal component analysis of the inflammatory analytes showed diverse inflammatory responses in PDAC.
Conclusion
In this work, we found a marked heterogeneity in infiltrating CD3+ and CD8+ lymphocytes and individual inflammatory responses in PDAC. Future mechanistic studies should explore personalized therapeutic strategies to target the immune and inflammatory components of the tumor microenvironment.
Citation: Tushoski-Alemán GW, Herremans KM, Underwood PW, Akki A, Riner AN, Trevino JG, et al. (2024) Infiltration of CD3+ and CD8+ lymphocytes in association with inflammation and survival in pancreatic cancer. PLoS ONE 19(2): e0297325. https://doi.org/10.1371/journal.pone.0297325
Editor: Kenji Fujiwara, Kimura Hospital, JAPAN
Received: August 27, 2023; Accepted: January 2, 2024; Published: February 12, 2024
Copyright: © 2024 Tushoski-Alemán et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the paper and its Supporting information files.
Funding: This research was funded by National Cancer Institute (https://www.cancer.gov/)(R01CA242003 to JGT, U54CA233444 to JGT, and U54CA233444-03S1 to ANR and JGT) and National Institute of Diabetes and Digestive and Kidney Diseases of the National Institutes of Health (https://www.niddk.nih.gov/) under award number 1U01DK108320 (SJH) and the Joseph and Ann Matella Fund for Pancreatic Cancer Research (https://www.uff.ufl.edu/giving-opportunities/023909-joseph-and-ann-matella-fund-for-pancreatic-cancer-research/) (JGT). Authors also receive salary support from the National Human Genome Research Institute of the National Institute of Health (https://www.genome.gov/) (T32 HG008958 to K.M.H., and A.N.R.). Authors are also supported by the Collaborative Alliance of Pancreas Education and Research (https://www.caper-pancreas.org/) (K.M.H. and A.N.R.). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Abbreviations: IPMN, Intraductal papillary mucinous neoplasm; PCA, Principal component analysis; PDAC, Pancreatic ductal adenocarcinoma; PNET, Pancreatic neuroendocrine tumor; TMA, Tissue microarray; TME, Tumor microenvironment
Background
Pancreatic ductal adenocarcinoma (PDAC) exhibits complex molecular and cellular heterogeneity, and clinical outcomes remain poor [1]. Compared to other tumor types, PDAC is considered an immune-cold cancer, escaping immune recognition by cytotoxic T-cells (CD8+), and other anti-tumor immune cell types. It is also resistant to immune checkpoint inhibition monotherapy [2,3]. Complex inflammatory processes have been shown to modulate PDAC progression [4,5]. For example, we and others have demonstrated that cancer associated fibroblasts (CAFs) impact the tumor microenvironment and foster tolerance [6,7]. Although PDAC is typically described as a non-immunogenic tumor, reprogramming the dense desmoplastic reaction, and resulting inflammatory response may offer therapeutic opportunities for patients.
The activation and recruitment of CD3+ and CD8+ lymphocytes are fundamental to mounting an effective tumor response in nearly all cancers. Various techniques have been employed to evaluate lymphocyte populations. Semi-quantitative approaches are routinely applied in clinical and diagnostic settings. Quantitative and automated approaches have been shown to enable standardized analysis of large datasets, enhanced reproducibility, and precision in anticipating immunotherapy response [8–10]. These approaches have not yet been fully assessed or compared in PDAC. Neoadjuvant therapy has been shown to alter the immune microenvironment of pancreatic cancers [11,12], and as such, a treatment naïve cohort may offer enhanced insight to true lymphocyte heterogeneity in PDAC. We hypothesize that a quantitative approach could offer improved accuracy in assessing lymphocyte infiltration for overall survival in pancreatic cancer patients. We further hypothesize that infiltrating T-cells are associated with inflammatory proteins in the tumor microenvironment. We anticipate that associations between infiltrating T-cells and inflammatory proteins in the tumor microenvironment may shed light on the underlying mechanisms of T-cell recruitment and activation to provide potential targets to improve therapeutic efficacy in PDAC.
In this study, we assembled tissue-microarrays (TMAs) of treatment naïve PDAC tumor and non-PDAC controls. We quantitatively determined intratumoral CD3+ and CD8+ lymphocyte populations. The design of this study augments previous studies that have relied on semi-quantitative assessment and included neoadjuvant treated PDAC patients [13]. We performed multiplex analysis on matching tumors to assess relationships between immune cell infiltrations and inflammatory proteins in the tumor microenvironment. We also identified multiple inflammatory factors as indicators of high CD3+ or CD8+ involvement.
Methods
Patient cohort and tissue preparation
All studies were approved by the Institutional Review Board at the University of Florida (IRB201600873) and informed consent was obtained from all patients prior to any tissue collection. Patients were recruited into the study from May 2012 to January 2022. This study was conducted from January 2017 to March 2023. Authors had access to information that could identify individual patients after data collection. All tissue samples were obtained from our tissue bank at the University of Florida Department of Surgery. Only patients with pathology confirmed PDAC naïve to neoadjuvant chemotherapy or radiation were included in the cancer group. Tissue was also collected separately from patients with pancreatitis, pancreatic cyst, and intraductal papillary mucinous neoplasm (IPMN) to serve as non-PDAC controls. All surgical specimens were collected by sharp dissection on a back table in the operating room. Pancreatic tissues were also collected from patients with chronic pancreatitis or other benign conditions that should not impact the resected pancreas, i.e., duodenal adenoma. Tissue used in the study was collected from distinct and separate patients with the diseases shown in Table 1. Samples were then transported on ice in Dulbecco’s modified Eagle media supplemented with 10% fetal bovine serum (Lonza Group) and 1% antibiotic/antimycotic solution (Corning Inc.) to the lab immediately. Samples were flash frozen within 20 minutes of collection and stored at -80°C for downstream soluble protein analysis. Separately, matching FFPE blocks were requested from our pathology core for TMA assembly.
Tissue microarray assembly
Tissue was formalin-fixed, and paraffin embedded following resection. Representative formalin fixed paraffin-embedded blocks and tumor areas were selected by a pathologist (A.A) based on H&E-stained slides. Briefly, a hollow needle was used to remove 2 mm diameter tissue cores from regions of interest. These tissue cores were then inserted into a recipient paraffin block forming the TMA. Assembled TMAs for CD3+ and CD8+ lymphocytes originate from the same tumors and tissue. TMAs were randomly assembled in duplicate and subsequently assessed by a pathologist for quality assurance before subjecting to cutting 4-micron slides for subsequent immunohistochemical analyses.
Assessment of tissue microarray
Presence of CD3+ and CD8+ lymphocytes were evaluated by immunohistochemistry using anti-CD3 and anti-CD8 antibodies (Dako Omnis, Agilent). The project pathologist (A.A.) semi-quantitatively graded each core using traditional scoring methodology (scores ranging from 0–3). Scores between each TMA were calculated and compared to quantitative assessment. Slides of the stained TMAs were scanned using an Aperio Scanscope CS microscope and digital images were acquired for quantitative analysis. Quantification of positively stained cells was performed using the software package QuPath 0.3.2. [8,14] (see https://qupath.github.io/). Cores were only excluded if no analyzable tissue was present, or if PDAC was not confirmed for that core. Image type settings in QuPath were set to Brightfield (H-DAB). Total counted cells and positively stained cells (DAB) for each were tabulated using the positive cell count function in QuPath (analyze -> cell detection -> positive cell detection). Whole tissue cores were selected for analysis. Percentage of positive cells were calculated using positive cells / total cells *100 (%). TMAs were stained and assembled in duplicates, averages between two matching TMAs were used, when available. Quantification and grading were performed in a blinded fashion separate from survival analysis. A total of 99 individual cores in the PDAC group, and 36 non-PDAC cores in the CD3+ TMA were graded and quantitated across duplicates. In the CD8+ TMA, 90 individual PDAC cores and 37 non-PDAC cores were graded and quantitated across duplicates. The minor differences in sample sizes were due to quality-assurance (i.e. exclusion due to core integrity/damage across TMAs).
Soluble protein analysis
At the time of processing, tissue was thawed and weighed. Tissue was sharply divided into small pieces and placed into 2-mL Lysing Matrix D tubes (MP Biomedicals, Santa Ana, California, USA). For every 30mg of tissue, 500 uL of cell lysis buffer (Cell Signaling Technology, Danvers, Massachusetts) with Protease/Phosphatase Inhibitor (Cell Signalling Technology) was added. Samples were then placed on ice for 2 minutes each cycle. Lysates were collected and centrifuges at 13,000 relative centrifugal force for 10 minutes. Supernatants were collected and analyzed for total protein concentration. Homogenates were probed for 41 unique analytes using a commercially available multiplex analysis per the manufactures protocol (catalog no. HCYTMAG-60K-PX41; Millipore Sigma) and as described previously [15]. The multiplex assay was selected for its large inflammatory panel and used as exploratory analysis as previously described [15]. Data were acquired with the MAGPIX System (Luminex Corp) and analyzed using MILLIPLEX Analyst 5.1 (Millipore Sigma). Protein concentrations were normalized to total protein concentrations to yield individual analyte concentrations. Of the 41 analytes, those that provided non-informative data (no detection) were excluded. Samples from the TMA and in the soluble protein analysis were matched and originate from the same patient. These samples were processed separately (i.e., formalin fixation/paraffin embedding for the TMA, and sharp frozen for soluble protein analysis) but represent paired tissue after pathological confirmation.
Survival analysis
Overall survival for our patient cohort was retrospectively obtained. Date of death was obtained through retrospective chart verification to determine survival time from surgical resection. Kaplan-Meier tests were applied, and the patient cohort was evenly split between the lower 50th percentile, and upper 50th percentile for CD3+ and CD8+ cell count percentages. Overall survival was also evaluated with semi-quantitative grading, categorizing the groups based on their grade (0, 1, 2, 3). Mantel-Cox survival comparisons were also made between quantification and grading for CD3+ and CD8+ patients.
Statistical analysis
All statistical analyses were performed in the software GraphPad Prism 9.4.1. Shapiro-Wilk tests of the lymphocyte count and inflammatory protein levels in each tumor was used to assess normality and distribution. Due to non-normally distributed data observed in both lymphocyte cell counts and soluble protein analysis, non-parametric tests were used. Mann-Whitney U test was used to compare cell counts and infiltrating lymphocyte levels across tissues, and Spearman correlation was used to assess associations. Multivariate regression was used to assess relationships between clinicopathologic characteristics (i.e., CD3+% and age) in addition to spearman correlations. Kaplan-Meier method was employed to evaluate differences in overall survival. Spearman correlation analyses were performed between the percent of CD3+/CD8+ cells and all 41 analytes in paired tumors or tissue. Data is presented as medians. A two-tailed P-value <0.05 is considered statistically significant.
Results
Quantitative analysis of intratumoral CD3+ and CD8+ lymphocytes in PDAC
The constructed TMA (prior to quality assurance) consisted of 59 treatment naïve pancreatic cancers and 37 non-PDAC controls. Clinicopathological characteristics are noted in Table 1. TMAs were immunohistochemically stained for CD3+ lymphocytes and CD8+ lymphocytes in duplicates. Our pathologist reviewed each TMA for quality assurance and semi-quantitatively graded the staining intensity in a blinded fashion. For quantitative assessment, cell counting was performed using the software package QuPath [8]. Following quality assurance, 56 PDAC tumors and 21 non-PDAC samples were analyzed in the CD3+ TMA, and 59 PDAC tumors and 24 non-PDAC samples were analyzed in the CD8+ TMA. The analyzable non-PDAC group included pancreatitis (n = 12), IPMN (n = 3), and other tissue (3 mucinous cystic neoplasm, 2 pancreatic neuroendocrine tumor, 1 pseudopapillary neoplasm) in the CD3 TMA. And the analyzable non-PDAC group for the CD8 TMA included pancreatitis (n = 13), IPMN (n = 4), and other (4 mucinous cystic neoplasm, 1 pancreatic neuroendocrine tumor, 1 squamoid cyst, and 1 pseudopapillary neoplasm).
PDAC displayed noteworthy heterogeneity in its CD3+ and CD8+ populations. Serial sections of tumor resulted in varying levels of CD3+ (Fig 1a) and CD8+ (Fig 1b) cells. To account for this, we created duplicate TMAs for each core, and, when possible, averaged the percent positive between the groups. We noted that the distribution of CD3+ and CD8+ lymphocytes in the PDAC tumors were distributed relatively evenly (not clustered). Non-PDAC pancreas maintained high cell density characteristic of normal pancreas, as opposed to the less dense, desmoplastic reactions that forms in tumors (Fig 2a and 2c). We observed significant increases in CD3+ and CD8+ populations (by percentages, and absolute values) in PDAC tumors compared to non-PDAC control tissues with the exception of PDAC vs. IPMN, however, this may be due to limited sample size of the IPMN cohort (Fig 2b and 2d). As expected, Spearman r correlation of CD3+ and CD8+ revealed a significant correlation (Fig 2e). Analysis of covariates (CD3/CD8 infiltration, age, sex, histologic grade) in a multi-variable linear regression revealed no correlation, and poor model fit (S1 Table). This has also been shown in previous studies [16]. Spearman correlation of individual covariates (age, sex, histologic grade) also showed no correlation (S1 Table).
Representative samples of immunohistochemical staining of tumor cores, slices originate from the same tumor: (A) CD3+ PDAC and subsequent quantification, and (B) CD8+ PDAC and subsequent quantification. Blue-colored cells represent cells labelled negative, red-colored cells represent cells considered positive.
(A) Cell count of CD3+ TMA, PDAC (n = 56, median = 17785), Pancreatitis (n = 12, median = 24219), IPMN (n = 3, mean = 23315), Other (n = 6, mean = 26582), Mann Whitney U tests of PDAC vs Pancreatitis (p = 0.0003); PDAC vs IPMN (p = 0.0041); PDAC vs Other (p = 0.0031). (B) Prevalence of CD3+ lymphocytes in PDAC (n = 56, median = 2.565%), Pancreatitis (n = 12, median = 0.2950%), IPMN (n = 3, median = 0.5628%), Other (n = 6, median = 0.1919%). Mann-Whitney test of PDAC vs Pancreatitis (p = <0.0001), PDAC vs IPMN (p = 0.0434), PDAC vs other (p = 0.0031). (C) Cell count of CD8+ TMA, PDAC (n = 59, median = 15569), Pancreatitis (n = 13, median = 26580), IPMN (n = 4, median = 26355), Other (n = 7, median = 29877). (D) Prevalence of CD8+ lymphocytes in PDAC (n = 59, median = 2.090%), Pancreatitis (n = 13, median = 0.6800%), IPMN (n = 4, median = 1.195%), Other (n = 7, median = 0.2200%). Mann-Whitney test of PDAC vs Pancreatitis (p = 0.0018), PDAC vs IPMN (p = 0.5012), PDAC vs Other (p = <0.0001). (E) Correlation of CD3+ and CD8+ lymphocyte percentages of PDAC and non-PDAC in the TMAs, r = 0.684 and p = <0.0001.
Semi-quantitative and quantitative assessments and overall survival
We then compared quantitative analysis to semi-quantitative pathological grading to observe patterns between systematic scoring and quantitative techniques. As above, the TMAs were subject to evaluation by an independent pathologist and graded according to standard pathology guidelines (score 0–3). Individual cores were assessed, and duplicate TMA cores were used individually (as opposed to averaged values) to appropriately, and directly, assess grading and quantification conversion. Fig 3A and 3C visually demonstrate the impact of conversion from integer semi-quantification to continuous variable data from the digital quantification. To assess the prognostic value of CD3+ and CD8+ infiltration and compare differences in grading and quantification, we performed Kaplan-Meier tests. We split the cohort at the median, thus CD3+ or CD8+ low tumors from the bottom 50th percentile for cell count versus CD3+ or CD8+ high tumors from the top 50th percentile group. Survival by quantitative grading was split into three groups, 0, 1, 2. No tumors were graded as 3. In our survival comparison, grades of 0 were compared with the tumors with “low” infiltration, and grades of 1 or 2 were compared with tumors with “high” infiltration.
(A) Comparison of semiquantitative grading of cores with quantitative assessment for CD3+ staining (blue lines represent a grade of 0, red lines represent a grade of 1, green lines represent a grade of 2). In the CD3+ TMA, 99 individual cores in the PDAC group, and 36 cores in the non-PDAC group were graded and quantitated. (B) CD3+ overall survival comparison of quantitative methods vs grading method (CD3+ low = solid red, CD3+ high = solid blue, 0 = dashed red line, 1 = dashed blue line). CD3+ low tumors had a median survival of 273 days, and a median survival of 642.5 days in CD3+ high tumors (p = 0.2184). CD3+ cores graded 0 demonstrated a median survival of 252 days, CD3+ cores graded 1 had a median survival of 778 days (p = 0.0017). In comparing grading vs. quantitative technique, no statistically significant difference was found; CD3 low vs. 0 (p = 0.1907), and CD3 high vs. 1 (p = 0.7003). (C) Comparison of semi-quantitative grading of cores converted to quantitative assessment for CD8+ staining. In the CD8+ TMA, 90 individual cores and 37 non-PDAC cores were graded and quantitated. (D) CD8+ overall survival comparison of quantitative methods vs grading method. CD8+ low tumors had a median survival of 240 days, CD8+ high tumors had a median survival of 1059 days (p = 0.0003). CD8+ tumors with grade 0 vs 1 had a median survival of 240 days and 642 days (p = 0.0057), and grade 1 vs 2 had a median survival of 642 and 951 days (p = 0.7158). Grading vs. quantitative technique showed no significant differences, CD8+ low vs 0 (p = 0.7982), CD8+ high vs 1 (p = 0.6887), CD8+ high vs 2 (p = 0.5882).
We did not observe a significant difference in overall survival between CD3+ low and high groups through quantitative technique, with a median survival of 273 days in CD3+ low tumors, and 642.5 days in CD3+ high tumors (p = 0.2184). Pathological grading for CD3+ cores demonstrated a statistical significance between 0 and 1 grades. CD3+ cores graded 0 demonstrated a median survival of 252 days, CD3+ cores graded 1 had a median survival of 778 days (p = 0.0017) (Fig 3b). No statistically significant differences in survival were found in grading vs. quantitative technique, CD3 low vs. 0 (p = 0.1907), and CD3 high vs. 1 (p = 0.7003) No cores in the CD3+ group were graded 2.
We observed a significant difference between CD8+ low and CD8+ high tumors, with a median survival of 240 days, and 1059 days, respectively (p = 0.0003). CD8+ tumors with grades 0, 1, 2, also demonstrated statistically significance, with a median overall survival of 240 days, 642 days, and 951 days, respectively (p = 0.0156) (Fig 3d). Comparison of grading and quantitative technique found no significant differences between CD8+ low vs. 0 (p = 0.7982), CD8+ high vs. 1 (p = 0.6887), and CD8+ high vs 2 (p = 0.5882). Taken together, this data shows how a quantitative approach may provide more precise, continuous data that does not necessarily result in superior prognostication in this relatively small sample set.
Identification of an inflammatory protein signature in PDAC with increased CD3+ and CD8+ infiltration
Inflammatory signaling mediates the activation and recruitment of cells to the tumor site [17,18]. We hypothesized that CD3+ high tumors and CD8+ high tumors had distinct tumor microenvironment inflammatory milieus. To test this, we performed a multiplex analysis of 41 inflammatory analytes with subsequent Spearman correlation analyses to find associations between these analytes and levels of CD3+ (Fig 4a) or CD8+ (Fig 5a) infiltration in PDAC. We performed a principal component analysis (PCA) of CD3+ PDAC, pancreatitis, IPMN and other miscellaneous tissue. Principal component analysis (PCA) used 31 analytes, as data from 10 analytes (G-CSF, IL-9, IL-1B, IL-2, IL-3, IL-4, IL-5, MIP-1A, RANTES, and TNFB) were non-informative (Figs 4d and 5c). PCA visually demonstrates the diversity of inflammatory signatures and notable heterogeneity in PDAC, with some PDAC showing overlap between IPMN and non-PDAC signatures. Out of 41 inflammatory proteins analyzed (Table 2) We found higher levels of IL-1B and IL-2 to be significantly correlated with lower CD3+ infiltration, r = -0.3704, -0,4275, with p = 0.0187, 0.0074, respectively (Fig 4b and 4c).
(A) Heatmap of 41 analytes in PDAC (n = 55) and non-PDAC tissue (Pancreatitis (n = 12), IPMN (n = 3), Other (n = 6)), expression levels were normalized to protein concentration (B, C): Spearman correlation analysis of IL-1B (r = -0.3704, p = 0.0187) and IL-2 (r = -0.4275, p = 0.0074) show decreased concentrations were significantly associated with increased CD3+ infiltration. (D) Principal component analysis plot of inflammatory signatures in different tissues, PCA was performed using 31 analytes, as 10 analytes (G-CSF, IL-9,IL-1B,IL-2,IL-3,IL-4,IL-5,MIP-1A,RANTES and TNFB) had limited detection across our patient cohort. PDAC (blue) is noted to be heterogenous in its inflammatory profile compared to non-PDAC tissue (Pancreatitis = red, Other = purple, IPMN = green).
(A) Expression of 41 analytes in PDAC (n = 58) and non-PDAC (Pancreatitis (n = 13), IPMN (n = 4), Other (n = 7)) expression is normalized to protein concentration (B) Spearman correlation analyses showing increased expression of IL-1B significantly associated with decreased CD8+ infiltration (r = -0.4299, p = 0.0045). (C) Principal component analysis plot of inflammatory signatures in different tissues, PCA was performed using 31 analytes, as 10 analytes (G-CSF, IL-9,IL-1B,IL-2,IL-3,IL-4,IL-5,MIP-1A,RANTES and TNFB) had limited detection across our patient cohort. PDAC (blue) is noted to be heterogenous in its inflammatory profile compared to non-PDAC tissue (Pancreatitis = red, Other = purple, IPMN = green).
We next aimed to identify inflammatory proteins associated with CD8+ infiltration. PCA was performed using 31 analytes as mentioned above. The PCA similarly shows high variability in the presence of inflammatory proteins in tumors, and some overlap between non-PDAC (Fig 5c). Of the 41 inflammatory proteins assessed (Table 3), only IL-1B demonstrated statistical significance and was associated with decreased CD8+ infiltration (r = -0.4299, p = 0.0045) (Fig 5b).
Discussion
In this work, we digitally measured intratumoral CD3+ and CD8+ lymphocytes and confirmed its effectiveness in the quantitative assessment of T-cell infiltration, which is predictive of overall survival in PDAC. PDAC demonstrated increased CD3+ and CD8+ infiltration compared to non-PDAC controls, and substantial heterogeneity. We found a significant increase in overall survival with high levels of CD8+ infiltration, but not CD3+ infiltration. We further demonstrated correlations between inflammatory protein signatures and levels of CD3+ and CD8+ infiltration.
Previous studies have used a range of approaches to evaluate the presence of CD3+ and CD8+ lymphocytes in tumors, including both quantitative and semiquantitative methods [19]. In this study, we contributed a standardized and reproducible method for quantifying CD3+ and CD8+ T-cells within a treatment-naïve patient cohort with PDAC. Semi-quantitative grading and quantitative approach through QuPath demonstrated similar outcomes in both prevalence of lymphocyte infiltration, and prognosis. The limited sample size in tumor or tissue graded 2 or 3 may limit the generalizability of these findings, and this analysis should be reproduced in tumors with expectantly higher lymphocyte infiltration, such as melanoma or microsatellite instable solid tumors. Through the identification of patients with objectively higher levels of CD8+ infiltration, we may be able to more effectively select patients who are most likely to benefit from immune-based therapies, such as checkpoint inhibitors or adoptive cell transfer.
While PDAC has been shown to be resistant to current immunotherapeutics, recruitment and activation of T-cells into the tumor site may still be a viable approach [20,21]. Individual inflammatory molecules have been leveraged to guide immune cells into the tumor-microenvironment [22,23], however, a multifaceted approach to modifying inflammation, involving multiple inflammatory molecules, may result in improved immunogenicity in pancreatic cancer [24]. In this study, we established connections between T-cells, effector T-cells and inflammatory molecules in order to identify potential candidates that could aid in the immune recognition of PDAC.
The recruitment and activation of immune cells into the tumor microenvironment is fundamental to the success of immunotherapy. Our findings, as well as others, illustrated the importance of tumor-derived cytokines in lymphocyte recruitment [25–28]. Factors secreted by cancer cells, such as GM-CSF and G-CSF have been reported to be associated with the recruitment of leukocytes into the TME [26,29]. Additionally, cytokines have been identified as mediators of therapeutic resistance in PDAC [30]. These cytokines remodel ambient cell populations to create a tumor permissive environment. Infiltration of T-cells in solid tumors have been associated with superior survival, however, in PDAC, populations of intratumor T-cells are scarce and are often unable to mount an effective anti-tumor immune response [31]. IL-1B has been shown to promote tumor survival, proliferation, and metastatic potential [32]. IL-1B also influences the tumor microenvironment by enhancing desmoplasia and immune suppression in pancreas and breast cancer [32–34]. IL-2 is a cytokine responsible for the growth, proliferation, and survival of T-cells. IL-2 has been shown to play a vital role in the regulation of CD8+ T-cells to maintain their reactivity against tumor cells [35,36]. Our data suggests that IL-1B may play a more dominant role in regulating T-cell recruitment into the tumor microenvironment. We posit that IL-2‘s role is more nuanced, potentially contributing to the activation and exhaustion of T-cells or effecting other T-cell subtypes and immune cells [37]. A more precise method for remodeling the tumor microenvironment may be achievable though spatial analysis of the microenvironment with respect to the secretion of inflammatory proteins and lymphocyte involvement. We previously reported differences in the transcriptome of cancer-associated fibroblasts based on proximity to tumor [38], these changes may be mediated by surrounding inflammatory molecules. In immune-hot solid tumors that respond well to immune checkpoint inhibitors, such as melanoma, T-cell and other immune cells that are recruited into the tumor-microenvironment are essential to mounting an effective immune response [39]. We speculate that there are inflammatory characteristics of CD3+ or CD8+ abundant tumors that can be exploited with tumors with less immunogenicity.
The heterogeneity of the pancreatic tumor microenvironment has posed challenges for both treatment and study [40–42]. In our study, we aimed to minimize variability by only including treatment naïve PDAC patients. Yet, we still observed significant heterogeneity in both the CD3+ and CD8+ populations and in the levels of inflammatory molecules across samples. We reported limited detection of certain cytokines (G-CSF, IL-9, IL-1B, IL-2, IL-3, IL-4, IL-5, MIP-1A, RANTES, and TNFB), though these cytokines may still be useful for further investigation. While our exploratory analysis found significant associations between IL-1B and IL-2 and lymphocyte infiltration, there is a need for additional mechanistic study to exploit the inflammatory axis to induce immune response to tumor. Future studies with larger validation cohorts that are more representative of the patient population could greatly improve our understanding of inflammation in PDAC and other solid tumor malignancies. Assessing the location of lymphocytes and inflammatory proteins (i.e., central tumor or tumor margin) may also prove useful to increase infiltration and penetration of lymphocytes in solid tumors. The relatively small sample size, and the fact that the patient cohort used in our analysis consisted largely of T-stage III tumors, may limit generalizability of our findings. A larger sample size, from various stages of tumor progression that assess whole tumor may provide greater statistical power, reduce the risk of type II errors, and allow for additional subgroup analyses.
Cancer cells employ pro/anti-inflammatory cytokines and chemokines to suppress or promote an immune-rich microenvironment [43–45]. The presence or absence of these cytokines can dictate local cell populations and ultimately the host’s response to malignancy. Here, we found that increased levels of IL-1B and IL-2 were associated with lowered CD3+ T-cell infiltration, and increased IL-1B was associated with decreased CD8+ T-cell populations. Solid-tumor malignancies, such as PDAC, have a particularly inflexible scaffolding in the form of stroma that add an additional physical barrier to limit immune-cell involvement and sensitivity to traditional therapeutics [46,47]. Ongoing clinical trials are utilizing cytokines as vehicles to increase tumor immunogenicity and remodeling the tumor microenvironment [48]. Our study combines advances in intratumoral immune cell quantification and characterization of inflammatory components to advance our understandings of the PDAC tumor microenvironment.
Supporting information
S1 Table. Comparisons of infiltrating CD3+ and CD8+ lymphocytes and covariate analysis.
https://doi.org/10.1371/journal.pone.0297325.s001
(XLSX)
S2 Table. Dataset of patient cohort, CD3+ and CD8+ lymphocyte quantification, and inflammatory assay.
https://doi.org/10.1371/journal.pone.0297325.s002
(XLSX)
Acknowledgments
We would like to thank Dr. Dongtao Ann Fu, Cathy Yuping Sun, and the University of Florida Molecular Pathology Core (RRID:SCR_016601) for assistance with TMA assembly, staining, and scanning. We would also like to acknowledge the contributions of our patients and their families for enabling us to advance our understanding of pancreatic cancer.
References
- 1. Siegel RL, Miller KD, Fuchs HE, Jemal A. Cancer statistics, 2022. CA Cancer J Clin. 2022;72: 7–33. pmid:35020204
- 2. Delitto D, Wallet SM, Hughes SJ. Targeting tumor tolerance: A new hope for pancreatic cancer therapy? Pharmacol Ther. 2016;166: 9–29. pmid:27343757
- 3. Somani VK, Zhang D, Dodhiawala PB, Lander VE, Liu X, Kang L-I, et al. IRAK4 Signaling Drives Resistance to Checkpoint Immunotherapy in Pancreatic Ductal Adenocarcinoma. Gastroenterology. 2022;162: 2047–2062. pmid:35271824
- 4. Purohit A, Saxena S, Varney M, Prajapati DR, Kozel JA, Lazenby A, et al. Host Cxcr2-Dependent Regulation of Pancreatic Cancer Growth, Angiogenesis, and Metastasis. Am J Pathol. 2021;191: 759–771. pmid:33453178
- 5. Lafaro KJ, Melstrom LG. The Paradoxical Web of Pancreatic Cancer Tumor Microenvironment. Am J Pathol. 2019;189: 44–57. pmid:30558722
- 6. Erez N, Truitt M, Olson P, Hanahan D. Cancer-Associated Fibroblasts Are Activated in Incipient Neoplasia to Orchestrate Tumor-Promoting Inflammation in an NF-κB-Dependent Manner. Cancer Cell. 2010;17: 135–147. pmid:20138012
- 7. Delitto D, Delitto AE, DiVita BB, Pham K, Han S, Hartlage ER, et al. Human Pancreatic Cancer Cells Induce a MyD88-Dependent Stromal Response to Promote a Tumor-Tolerant Immune Microenvironment. Cancer Res. 2017;77: 672–683. pmid:27864347
- 8. Bankhead P, Loughrey MB, Fernández JA, Dombrowski Y, McArt DG, Dunne PD, et al. QuPath: Open source software for digital pathology image analysis. Sci Rep. 2017;7: 16878. pmid:29203879
- 9. Duan K, Jang G-H, Grant RC, Wilson JM, Notta F, O’Kane GM, et al. The value of GATA6 immunohistochemistry and computer-assisted diagnosis to predict clinical outcome in advanced pancreatic cancer. Sci Rep. 2021;11: 14951. pmid:34294813
- 10. Miles GJ, Powley I, Mohammed S, Howells L, Pringle Jhoward, Hammonds T, et al. Evaluating and comparing immunostaining and computational methods for spatial profiling of drug response in patient-derived explants. Laboratory Investigation. 2021;101: 396–407. pmid:33318618
- 11. Dias Costa A, Väyrynen SA, Chawla A, Zhang J, Väyrynen JP, Lau MC, et al. Neoadjuvant Chemotherapy Is Associated with Altered Immune Cell Infiltration and an Anti-Tumorigenic Microenvironment in Resected Pancreatic Cancer. Clinical Cancer Research. 2022;28: 5167–5179. pmid:36129461
- 12. Li K, Tandurella JA, Gai J, Zhu Q, Lim SJ, Thomas DL, et al. Multi-omic analyses of changes in the tumor microenvironment of pancreatic adenocarcinoma following neoadjuvant treatment with anti-PD-1 therapy. Cancer Cell. 2022;40: 1374–1391.e7. pmid:36306792
- 13. Golesworthy B, Wang Y, Tanti A, Pacis A, Romero JM, Cuggia A, et al. Intra-Tumoral CD8+ T-Cell Infiltration and PD-L1 Positivity in Homologous Recombination Deficient Pancreatic Ductal Adenocarcinoma. Front Oncol. 2022;12. pmid:35547873
- 14. Loughrey MB, Bankhead P, Coleman HG, Hagan RS, Craig S, McCorry AMB, et al. Validation of the systematic scoring of immunohistochemically stained tumour tissue microarrays using QuPath digital image analysis. Histopathology. 2018;73: 327–338. pmid:29575153
- 15. Underwood PW, Gerber MH, Nguyen K, Delitto D, Han S, Thomas RM, et al. Protein Signatures and Tissue Diagnosis of Pancreatic Cancer. J Am Coll Surg. 2020;230: 26–36e1. pmid:31672677
- 16. Kiryu S, Ito Z, Suka M, Bito T, Kan S, Uchiyama K, et al. Prognostic value of immune factors in the tumor microenvironment of patients with pancreatic ductal adenocarcinoma. BMC Cancer. 2021;21: 1197. pmid:34758773
- 17. Saxena S, Singh RK. Chemokines orchestrate tumor cells and the microenvironment to achieve metastatic heterogeneity. Cancer and Metastasis Reviews. 2021;40: 447–476. pmid:33959849
- 18. Foeng J, Comerford I, McColl SR. Harnessing the chemokine system to home CAR-T cells into solid tumors. Cell Rep Med. 2022;3: 100543. pmid:35492880
- 19. Moratin J, Mock A, Obradovic S, Metzger K, Flechtenmacher C, Zaoui K, et al. Digital Pathology Scoring of Immunohistochemical Staining Reliably Identifies Prognostic Markers and Anatomical Associations in a Large Cohort of Oral Cancers. Front Oncol. 2021;11. pmid:34395287
- 20. Raj D, Yang M-H, Rodgers D, Hampton EN, Begum J, Mustafa A, et al. Switchable CAR-T cells mediate remission in metastatic pancreatic ductal adenocarcinoma. Gut. 2019;68: 1052–1064. pmid:30121627
- 21. Bruchard M, Geindreau M, Perrichet A, Truntzer C, Ballot E, Boidot R, et al. Recruitment and activation of type 3 innate lymphoid cells promote antitumor immune responses. Nat Immunol. 2022;23: 262–274. pmid:35102345
- 22. Bellone G, Turletti A, Artusio E, Mareschi K, Carbone A, Tibaudi D, et al. Tumor-Associated Transforming Growth Factor-β and Interleukin-10 Contribute to a Systemic Th2 Immune Phenotype in Pancreatic Carcinoma Patients. Am J Pathol. 1999;155: 537–547. pmid:10433946
- 23. Lutz V, Hellmund VM, Picard FSR, Raifer H, Ruckenbrod T, Klein M, et al. IL18 Receptor Signaling Regulates Tumor-Reactive CD8+ T-cell Exhaustion via Activation of the IL2/STAT5/mTOR Pathway in a Pancreatic Cancer Model. Cancer Immunol Res. 2023;11: 421–434. pmid:36758176
- 24. Tähtinen S, Kaikkonen S, Merisalo-Soikkeli M, Grönberg-Vähä-Koskela S, Kanerva A, Parviainen S, et al. Favorable Alteration of Tumor Microenvironment by Immunomodulatory Cytokines for Efficient T-Cell Therapy in Solid Tumors. PLoS One. 2015;10: e0131242. pmid:26107883
- 25. Marigo I, Bosio E, Solito S, Mesa C, Fernandez A, Dolcetti L, et al. Tumor-Induced Tolerance and Immune Suppression Depend on the C/EBPβ Transcription Factor. Immunity. 2010;32: 790–802. pmid:20605485
- 26. Bayne LJ, Beatty GL, Jhala N, Clark CE, Rhim AD, Stanger BZ, et al. Tumor-Derived Granulocyte-Macrophage Colony-Stimulating Factor Regulates Myeloid Inflammation and T Cell Immunity in Pancreatic Cancer. Cancer Cell. 2012;21: 822–835. pmid:22698406
- 27. Khajah M, Millen B, Cara DC, Waterhouse C, McCafferty D-M. Granulocyte-macrophage colony-stimulating factor (GM-CSF): a chemoattractive agent for murine leukocytes in vivo. J Leukoc Biol. 2011;89: 945–953. pmid:21393420
- 28. Herremans KM, Szymkiewicz DD, Riner AN, Bohan RP, Tushoski GW, Davidson AM, et al. The interleukin-1 axis and the tumor immune microenvironment in pancreatic ductal adenocarcinoma. Neoplasia. 2022;28: 100789. pmid:35395492
- 29. Pylayeva-Gupta Y, Lee KE, Hajdu CH, Miller G, Bar-Sagi D. Oncogenic Kras-Induced GM-CSF Production Promotes the Development of Pancreatic Neoplasia. Cancer Cell. 2012;21: 836–847. pmid:22698407
- 30. McAndrews KM, Chen Y, Darpolor JK, Zheng X, Yang S, Carstens JL, et al. Identification of Functional Heterogeneity of Carcinoma-Associated Fibroblasts with Distinct IL6-Mediated Therapy Resistance in Pancreatic Cancer. Cancer Discov. 2022;12: 1580–1597. pmid:35348629
- 31. Clark CE, Hingorani SR, Mick R, Combs C, Tuveson DA, Vonderheide RH. Dynamics of the Immune Reaction to Pancreatic Cancer from Inception to Invasion. Cancer Res. 2007;67: 9518–9527. pmid:17909062
- 32. Das S, Shapiro B, Vucic EA, Vogt S, Bar-Sagi D. Tumor Cell–Derived IL1β Promotes Desmoplasia and Immune Suppression in Pancreatic Cancer. Cancer Res. 2020;80: 1088–1101. pmid:31915130
- 33. Eyre R, Alférez DG, Santiago-Gómez A, Spence K, McConnell JC, Hart C, et al. Microenvironmental IL1β promotes breast cancer metastatic colonization in the bone via activation of Wnt signalling. Nat Commun. 2019;10. pmid:31676788
- 34. Caronni N, La Terza F, Vittoria FM, Barbiera G, Mezzanzanica L, Cuzzola V, et al. IL-1β+ macrophages fuel pathogenic inflammation in pancreatic cancer. Nature. 2023;623: 415–422. pmid:37914939
- 35. Shah P, Forget MA, Frank ML, Jiang P, Sakellariou-Thompson D, Federico L, et al. Combined IL-2, agonistic CD3 and 4-1BB stimulation preserve clonotype hierarchy in propagated non-small cell lung cancer tumor-infiltrating lymphocytes. J Immunother Cancer. 2022;10. pmid:35110355
- 36. Laporte KM, Hernandez R, Santos Savio A, Malek TR. Robust IL-2-dependent antitumor immunotherapy requires targeting the high-affinity IL-2R on tumor-specific CD8 + T cells. J Immunother Cancer. 2023;11. pmid:37270181
- 37. Liu Y, Zhou N, Zhou L, Wang J, Zhou Y, Zhang T, et al. IL-2 regulates tumor-reactive CD8+ T cell exhaustion by activating the aryl hydrocarbon receptor. Nat Immunol. 2021;22: 358–369. pmid:33432230
- 38. Han S, Fu D, Tushoski GW, Meng L, Herremans KM, Riner AN, et al. Single-cell profiling of microenvironment components by spatial localization in pancreatic ductal adenocarcinoma. Theranostics. 2022;12: 4980–4992. pmid:35836806
- 39. Blando J, Sharma A, Higa MG, Zhao H, Vence L, Yadav SS, et al. Comparison of immune infiltrates in melanoma and pancreatic cancer highlights VISTA as a potential target in pancreatic cancer. Proceedings of the National Academy of Sciences. 2019;116: 1692–1697. pmid:30635425
- 40. Connor AA, Gallinger S. Pancreatic cancer evolution and heterogeneity: integrating omics and clinical data. Nat Rev Cancer. 2022;22: 131–142. pmid:34789870
- 41. Raghavan S, Winter PS, Navia AW, Williams HL, DenAdel A, Lowder KE, et al. Microenvironment drives cell state, plasticity, and drug response in pancreatic cancer. Cell. 2021;184: 6119–6137.e26. pmid:34890551
- 42. Herremans KM, Underwood PW, Riner AN, Neal DW, Tushoski-Alemán GW, Forsmark CE, et al. A protein-based machine learning approach to the identification of inflammatory subtypes in pancreatic ductal adenocarcinoma. Pancreatology. 2023;23: 615–621. pmid:37391359
- 43. Lippitz BE. Cytokine patterns in patients with cancer: a systematic review. Lancet Oncol. 2013;14: e218–e228. pmid:23639322
- 44. De Boeck A, Ahn BY, D’Mello C, Lun X, Menon S V., Alshehri MM, et al. Glioma-derived IL-33 orchestrates an inflammatory brain tumor microenvironment that accelerates glioma progression. Nat Commun. 2020;11: 4997. pmid:33020472
- 45. Garcia Garcia CJ, Huang Y, Fuentes NR, Turner MC, Monberg ME, Lin D, et al. Stromal HIF2 Regulates Immune Suppression in the Pancreatic Cancer Microenvironment. Gastroenterology. 2022;162: 2018–2031. pmid:35216965
- 46. Jiang H, Hegde S, Knolhoff BL, Zhu Y, Herndon JM, Meyer MA, et al. Targeting focal adhesion kinase renders pancreatic cancers responsive to checkpoint immunotherapy. Nat Med. 2016;22: 851–860. pmid:27376576
- 47. Sharbeen G, McCarroll JA, Akerman A, Kopecky C, Youkhana J, Kokkinos J, et al. Cancer-Associated Fibroblasts in Pancreatic Ductal Adenocarcinoma Determine Response to SLC7A11 Inhibition. Cancer Res. 2021;81: 3461–3479. pmid:33980655
- 48. Hecht JR, Lonardi S, Bendell J, Sim H-W, Macarulla T, Lopez CD, et al. Randomized Phase III Study of FOLFOX Alone or With Pegilodecakin as Second-Line Therapy in Patients With Metastatic Pancreatic Cancer That Progressed After Gemcitabine (SEQUOIA). Journal of Clinical Oncology. 2021;39: 1108–1118. pmid:33555926