Esophageal cancer is associated with the sixth highest cancer-related mortality rate and a median survival time of 29 months [
The interaction of the tumor and its associated immune compartment is supposed to play an important role in cancer progression [
Indoleamine 2,3-dioxygenase (IDO) is an intracellular enzyme affecting T cell activity and immune tolerance. IDO expression has been detected in immune cells, stromal cells, and cancer cells and revealed relevance in cancer development and progression [
The mechanism of IDO induction is not completely clear. However, it is widely accepted that interferons, particularly INF-
In the present retrospective study, we tested the hypothesis that the protein expression of the immune checkpoint IDO on immune cells is prognostic in a large cohort of EAC. Expression levels and spatial distribution of IDO, INF-
We analyzed formalin-fixed and paraffin-embedded (FFPE) material of 551 patients with esophageal adenocarcinomas that underwent primary surgical resection or resection after neoadjuvant therapy between 1999 and 2015 at the Department of General, Visceral and Cancer Surgery, University of Cologne, Germany. The standard surgical procedure consisted of a transthoracic en bloc esophagectomy with two-field lymphadenectomy (abdominal and mediastinal lymph nodes); reconstruction was done by formation of a gastric tube with intrathoracic esophagogastrostomy (Ivor-Lewis esophagectomy) [
Patient characteristics and IDO expression results on the test cohort (
IDO expression surface margin | IDO expression infiltration margin | |||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
Negative | Positive | Negative | Positive | |||||||||
No. | % | No. | % | No. | % | No. | % | No. | % | |||
Sex | ||||||||||||
Female | 16 | 10.1% | 5 | 31.3% | 11 | 68.7% | 0.430 | 5 | 33.3% | 10 | 66.7% | 0.783 |
Male | 143 | 89.9% | 62 | 43.4% | 81 | 56.6% | 58 | 41.1% | 83 | 58.9% | ||
Age group | ||||||||||||
<65 years | 70 | 44.3% | 31 | 44.3% | 39 | 55.7% | 0.746 | 31 | 45.6% | 37 | 54.4% | 0.248 |
>65 years | 88 | 55.7% | 36 | 77.0% | 52 | 23.0% | 31 | 35.6% | 56 | 64.4% | ||
Tumor stage | ||||||||||||
pT1 | 46 | 29.1% | 13 | 28.3% | 33 | 71.7% | 0.070 | 13 | 29.5% | 31 | 70.5% | 0.200 |
pT2 | 29 | 18.4% | 12 | 41.4% | 27 | 58.6% | 11 | 37.9% | 18 | 62.1% | ||
pT3 | 82 | 51.9% | 41 | 50.0% | 41 | 50.0% | 37 | 45.7% | 44 | 54.3% | ||
pT4 | 1 | 0.6% | 1 | 100% | 0 | 0.0% | 1 | 100% | 0 | 0.0% | ||
Lymph node metastasis | ||||||||||||
pN0 | 60 | 38.0 | 17 | 23.3% | 43 | 71.75 | 0.022 | 20 | 33.9% | 39 | 66.1% | 0.154 |
pN1 | 71 | 44.9 | 34 | 47.9% | 37 | 52.1% | 26 | 37.0% | 43 | 62.3% | ||
pN2 | 12 | 7.6 | 6 | 50.0% | 6 | 50.0% | 7 | 58.3% | 5 | 41.7% | ||
pN3 | 15 | 9.55 | 10 | 66.6% | 5 | 33.3% | 9 | 60.0% | 6 | 40.0% | ||
UICC stage | ||||||||||||
I | 41 | 26.1% | 21 | 51.2% | 20 | 48.8% | <0.001 | 23 | 57.6% | 17 | 42.5% | 0.045 |
II | 21 | 13.4% | 15 | 71.4% | 6 | 28.6% | 14 | 66.7% | 7 | 33.3% | ||
III | 75 | 47.8% | 69 | 92.0% | 6 | 8.0% | 60 | 78.9% | 16 | 21.1% | ||
IV | 20 | 12.7% | 17 | 85.9% | 3 | 15.0% | 17 | 85.0% | 3 | 15.0% |
Patient characteristics of the entire cohort, IDO expression results (
IDO expression single spot | |||||
---|---|---|---|---|---|
Negative | Positive | ||||
No. | % | No. | % | ||
Sex | |||||
Female | 31 | 53.4% | 27 | 46.6% | 0.331 |
Male | 204 | 46.6% | 234 | 53.4% | |
Age group | |||||
<65 years | 132 | 51.0% | 127 | 49.0% | 0.166 |
>65 years | 105 | 44.5% | 132 | 55.5% | |
pT1 | 13 | 26.8% | 52 | 73.2% | 0.004 |
pT2 | 27 | 49.1% | 28 | 50.9% | |
pT3 | 179 | 51.1% | 171 | 48.9% | |
pT4 | 9 | 50.0% | 9 | 50.0% | |
pN0 | 74 | 38.1% | 120 | 61.9% | <0.001 |
pN1 | 79 | 44.6% | 98 | 55.4% | |
pN2 | 40 | 64.5% | 22 | 35.5% | |
pN3 | 40 | 65.6% | 21 | 34.4% | |
Neoadjuvant treatment | |||||
Yes | 90 | 41.7% | 126 | 58.3% | 0.029 |
No | 145 | 51.8% | 135 | 48.2% | |
UICC stage | |||||
I | 34 | 33.7% | 67 | 66.3% | <0.001 |
II | 44 | 38.9% | 69 | 61.1% | |
III | 118 | 57.3% | 88 | 42.7% | |
IV | 37 | 50.0% | 37 | 50.0% |
Immunohistochemistry and RNA BaseScope analyses were performed on tissue microarrays. Construction of the TMA was described previously [
In the first step, we analyzed a test cohort of 165 EAC. Therefore, we built a tissue microarray (TMA) with multiple tumor spots according to the suggestions of the international immunooncology working group for assessing tumor-infiltrating lymphocytes (TILs) on solid tumors [
The immunohistochemical and RNA BaseScope data were statistically correlated with survival and molecular data like
Immunohistochemistry (IHC) was performed on TMA slides. For IDO the rabbit IgG monoclonal antibody (D5J4E; dilution 1 : 400; Cell Signaling Technology, USA) and for CD3 the rabbit monoclonal antibody (SP7; dilution 1 : 50; Thermo Fisher Scientific, USA) were used. All immunohistochemical stainings were performed using the Leica BOND-MAX stainer (Leica Biosystems, Germany) according to the protocol of the manufacturers. The evaluation of immunohistochemical expression was assessed manually by two pathologists independently (PL and HL). Discrepant results were resolved by consensus review.
For IDO, the expression in <1% lymphocytes was defined as negative (
For CD3, CD3 expression in <3 lymphocytes/mm2 was evaluated as negative, >3–50 lymphocytes/mm2 were assessed as low positive, and >50 lymphocytes/mm2 were defined as highly positive considering peritumoral and intratumoral distribution.
Concerning the multispot TMA, four spots of tumor surface and invasive margin each were examined. We built the average of the scores and matched the four samples to one category based on limit values: 0-0.49 = negative and 0.5-2 = positive.
Multiplex immunohistochemistry staining was performed on a Ventana Discovery Ultra automatic staining system using TMA slides. The following primary monoclonal antibodies were used: IDO, Cell Signaling; mouse CD8 clone C8/144B, mouse CD68 clone PG-M1 (both Dako/Agilent, USA), and rabbit CD4 clone 4B12 (Roche, Switzerland, ready to use). After conjugation with an antibody-bound enzyme (horseradish peroxidase or alcalic phosphatase), detection was carried out using DISCOVERY Silver kit (IDO), DISCOVERY Yellow kit (CD68), DISCOVERY Teal kit (CD4), and DISCOVERY Purple Kit (CD8; all Ventana/Roche, Switzerland). Counterstaining was done with hematoxylin and bluing reagent.
The RNA BaseScope assay was performed according to the manufacturer’s instruction. In brief, paraffin-embedded TMA blocks were cut into 5
Clinical data were collected prospectively according to a standardized protocol. SPSS Statistics for Mac (Version 21, SPSS) was used for statistical analysis. Interdependence between stainings and clinical data were calculated using the chi-squared and Fisher’s exact tests and displayed by cross-tables. Survival curves were plotted using the Kaplan-Meier method and analyzed using the log-rank test. Univariate and multivariate analyses were performed for prognostic factors of overall survival using the Cox regression model. All tests were two sided.
The test cohort comprised 165 patients with EAC that underwent surgical resection. There was a male preponderance with 149 male (90.4%) and 16 female (9.6%) patients with a median age of 65.1 years (range 33-85 years) at the time of operation. To confirm our results, a single-spot TMA with additional 386 patients was analyzed, resulting in 551 patients in total. The median follow-up for the entire cohort was 57.7 months with a calculated 5-year survival rate of 26.6%. 159 (96.3%) of the samples were evaluable (Table
UICC stage adjusted survival for the entire patient cohort (
IDO immunostaining was localized in the cytoplasm/membrane of tumor-infiltrating lymphocytes and cancer cells (Figures
Immunohistochemistry of IDO and INF-
On the single-spot TMA, we found IDO expression on TILs in 261 patients (52.6%). Again, a strong correlation between IDO-positive samples and early tumor stages (pT1/2) (
To evaluate which subtypes of T cells expressed IDO, we performed multiplex immunohistochemistry staining on two exemplary TMAs. We correlated IDO-positive cases with the expression of CD4, CD8, and CD68 semiquantitatively (Figure
Multicolor immunohistochemistry for IDO (black signals), CD4 (teal/blue signals), CD8 (purple signals), and CD68 (yellow signals): (a) coexpression of IDO and CD4 (
The expression of CD3 was evaluated for multispot (test cohort) and single-spot TMA. CD3 distribution was predominantly seen peritumorally (
IDO expression on TILs positively correlated with the amount CD3-positive T cells within the tumor (
High levels of CD3-positive TILs are associated with an improved overall survival (OS) compared to CD3-poor tumors considering the single-spot TMA of 551 patients (496 patients analyzable;
(a) High amounts of CD3-positive T cells are associated with an improved OS in esophageal adenocarcinoma. (b) Tumors with IDO-positive TILs show better median overall survival of 47.7 months in IDO-positive tumors compared to a median OS of 22.7 months for IDO-negative tumors. (c) In the pT1/2 group, patients with IDO-positive expression have a calculated average OS of 142.1 months (median not reached) compared to an average OS of 37.1 months (median OS 30.5 months),
Within the test cohort, correlation between INF-
In the test cohort of 165 patients, no statistically significant overall survival difference was detectable for IDO-expressing tumors, although a correlation with nodal-negative tumors and low UICC stages in IDO-positive samples was seen. However, there was a trend towards improved overall survival (OS) in patients with IDO expression.
On the single-spot TMA, tumors with IDO-expressing TILs showed an improved median OS (47.7 months (95% confidence interval (CI) 20.9-73.8 months)) compared to IDO-negative patients (median OS 22.7 months (95% CI 18.8-26.6 months),
Subgroup analyses revealed a particularly pronounced difference in OS in the group of pT1/2 stage patients. Within this group of lower tumor stages, IDO-positive patients reached a calculated average OS of 142.1 months (median not reached, average 95% CI 115.2-168.9 months) compared to an average OS of 37.1 months (95% CI 23.6-50.7 months, median OS 30.5 months (95% CI 19.9-41.1 months,
In a multivariate Cox regression analysis, IDO expression on TILs and the histopathological parameters pT and pN stages were independent prognostic markers (Table
Multivariate Cox regression model; HR = hazard ratio.
Hazard ratio | 95% confidence interval | |||
---|---|---|---|---|
Lower | Upper | |||
Sex (male vs. female) | 1.557 | 0.863 | 2.807 | 0.141 |
Age group (<65 vs. >65 years) | 1.351 | 1.01 | 1.807 | 0.043 |
Tumor stage (pT1/2 vs. pT3/4) | 1.429 | 0.916 | 2.229 | 0.116 |
Lymph node metastasis (pN0 vs. pN+) | 2.987 | 2.105 | 4.239 | 0.022 |
CD3 (low vs. high) | 0.666 | 0.459 | 0.966 | 0.032 |
IDO on TILs (negative vs. positive) | 0.729 | 0.537 | 0.991 | 0.044 |
216 (43.5%) patients received neoadjuvant therapy, whereas 280 (56.4%) patients primarily underwent surgical resection. The prognostic impact of IDO on OS was independent on whether neoadjuvant treatment was administered or not. IDO expression remains a positive prognostic marker in both patient groups with primary surgery and surgery with neoadjuvant treatment. Patients who underwent primary surgery without any kind of neoadjuvant treatment with the presence of IDO expression showed a median OS of 104.6 months (95% CI 50.1–159.2 months) while IDO-negative patients showed an OS of 25.4 months (95% CI 13.5–37.3 months,
Here, we report the expression of the immune checkpoint protein IDO on tumor-associated inflammatory cells in a large and well-characterized cohort of 551 patients with EAC. We evaluated the level of heterogeneity and distribution of IDO-positive TILs within the tumor. IDO expression on TILs was a strong and statistically independent prognostic biomarker for an improved overall survival in EAC. IDO expression correlated significantly with low UICC stages (I/II) and nodal-negative status (pN-). Furthermore, we found a strong correlation with INF-
For the test cohort of 165 patients, we built a multispot TMA considering two different tumor localizations (surface and infiltration margins) proving low heterogeneity within the 4 spots of one tumor localization and a consistent expression pattern between surface and infiltration margins, respectively. Thus, the expression of IDO in randomly taken EAC samples by endoscopic tumor biopsy is most likely representing overall tumor IDO expression. Furthermore, the absence of significant heterogeneity was one reason to evaluate IDO expression on a single-spot TMA with 386 additional patients (551 in total) to confirm our results.
To the best of our knowledge, we are the first to evaluate IDO protein expression in EAC. Rosenberg et al. analyzed IDO mRNA expression using TCGA data of squamous cell cancer and adenocarcinomas of the esophagus correlating other checkpoint markers like PD-L1 and CTLA4 [
Within our cohort, IDO protein expression on TILs was an independent prognostic biomarker within all tumor stages indicating a tremendous survival benefit especially in pT1/2 tumor stages. Since we did not detect heterogeneity in the expression pattern, high levels of IDO expression on TILs in endoscopic biopsies could probably define a subgroup of patients with a favorable prognosis, possibly with no further benefit of radiochemotherapy in this group. This could be particularly interesting for patients in advanced tumor stages with an extremely reduced expectancy of life avoiding an aggressive therapy concept with reduced benefit. Neoadjuvant treatment did not influence the prognostic effect of IDO on overall survival in this cohort of EAC. In a mouse model, IDO was shown to increase after radiotherapy [
Physiologically, IDO protein catalyzes the elimination of the essential amino acid tryptophan [
However, and in opposite to these findings, we clearly demonstrate not only a favorable impact of IDO expression on TILs on overall survival in a large cohort of 551 patients with EAC but also IDO to be an independent marker for prognosis. This is absolutely in common with former research, e.g., a recent study by Patil et al. [
Apart from IDO expression on inflammatory cells, we additionally found expression on tumor cells in 9.2%. IDO expression has been previously described in various cell types, including endothelial cells, mesenchymal stromal cells, fibroblasts, and various myeloid-derived antigen-presenting cells such as DCs and macrophages, as well as tumor cells [
Still, open questions concerning the interaction of IDO and cancer growth need to be answered. For example, the significant decrease of IDO expression in advanced tumor stages remains cryptic. We assume that there is an interaction between immune checkpoint expression of the tumor microenvironment and invasive tumor growth. However, it is not clear whether overexpression of immune regulatory proteins negatively influences local tumor growth, or changes in tumor biology in advanced tumor stages might be responsible for a decreased level of checkpoints. Furthermore, the downstream metabolites of the tryptophan catabolism are known to influence IDO expression and activity and need to be considered in targeted therapy concepts [
Our study has strengths and limitations. We analyzed two independent and well-characterized cohorts of EAC. Furthermore, we identified an expression of IDO on inflammatory and tumor cells and discussed potential pro- and antitumoral effects. However, the study is retrospective and a selection bias cannot be excluded. We were not able to include patients who received neoadjuvant treatment and showed a complete tumor response or those with advanced tumors that were not eligible for surgical therapy. Beyond that, functional conclusions concerning biological mechanisms of IDO are not feasible on formalin-fixed material; therefore, future studies have to clarify the potential of IDO as a predictive biomarker in EAC.
In summary, our study describes the rate of IDO expression on TILs in EAC and demonstrates a strong and statistically independent positive prognostic effect in a very large group of EACs for the first time. IDO expression correlates significantly with low UICC stages (I/II) and negative lymph node status. However, prospective studies need to confirm our results. Since we find a favorable effect of IDO expression on overall survival in EAC, we assume that IDO interaction with tumor cells might be more complex than anticipated. Particularly, the IDO-expressing cell type as well as the metabolites of tryptophan catabolism might influence the effectiveness of future clinical trials investigating antibody-based IDO blockade in EAC.
All data of the study are available whenever requested.
All procedures performed in the current study involving human tumor specimens were in accordance with ethical standards of the local research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards. This retrospective study was performed according to the criteria of the ethics committee of the University Hospital of Cologne.
The funding exclusively provides financial support for student research assistance and had no influence on the conducted study itself.
All authors declare that they have no conflict of interest.
Heike Loeser, Max Kraemer, Philipp Lohneis, and Alexander Quaas contributed equally to this work.
The authors would like to thank Wiebke Jeske and Uschi Zenz for their excellent technical assistance. Max Kraemer is supported by the Koeln Fortune Program/Faculty of Medicine, Universität zu Köln, grant number 410/2018.