Characterization of circulating immune cells and correlation with Tie2/Angiopoietins level in well differentiated neuroendocrine gastroenteropancreatic tumors: a cross-sectional analysis

The immune environment represents a new, but little explored, tool for understanding neuroendocrine neoplasms (NENs) behavior. An immunosuppressed microenvironment is hypothesized to promote NENs progression. A missing profiling of circulating leukocyte and peripheral blood mononuclear cells (PBMCs) subpopulations would open new perspectives in the still limited diagnostic-therapeutic management of NENs. A cross-sectional case-control pilot study was performed recruiting 30 consecutive subjects: 15 patients naïve to treatment, with histologically proven gastroenteropancreatic (GEP) neuroendocrine tumors (NETs) and 15 healthy controls, matched for age and sex. PBMCs subpopulations were studied by flow cytometry. Soluble Tie2 (sTie2), Angiopoietin-1 (Ang-1), Angiopoietin-2 (Ang-2) were evaluated by ELISA. Immune cell profiling revealed a significant lower CD3−CD56+ natural killer (NK) cell count in NETs vs controls (p = 0.04). NK subset analysis showed a reduced relative count of CD56+CD16+ NK cells (p =0.002) in NETs vs controls. Patients with NET showed a higher percentage of CD14+CD16++ non-classical monocytes (p = 0.01), and a lower percentage of CD14+CD16+ intermediate monocytes (p = 0.04). A decrease in percentage (p = 0.004) of CD4+ T-helper lymphocytes was found in NET patients. Evaluation of cellular and serum angiopoietin pathway mediators revealed in NET patients a higher relative count of Tie2-expressing monocytes (TEMs) (p < 0.001), and high levels of Ang-1 (p = 0.003) and Ang-2 (p = 0.002). Patients with GEP-NET presented an immunosuppressed environment characterized by a low count of cytotoxic NK cells, a high count of anti-inflammatory non-classical monocytes, and a low count of T-helper lymphocytes. Higher levels of TEMs and angiopoietins suggest a crosstalk between innate immunity and angiogenic pathways in NETs.


Introduction
Neuroendocrine tumors (NETs) are well-differentiated neoplasms characterized by a low proliferation rate with high variability in terms of age and site of onset, biological features, clinical presentation and management [1,2]. Gastroenteropancreatic (GEP) NETs may exhibit an unpredictable behavior ranging from indolent to highly aggressive forms [1]. The exact mechanisms involved in NETs aggressiveness and metastatization have not been fully understood. Immune tumor microenvironment plays a key role in tumors' growth and metastatic spread [3,4]; moreover, peripheral blood mononuclear cells (PBMCs) have been demonstrated to play a prominent role in cancer defense mechanisms [5,6]. There is pre-clinical evidence that, in the heterogeneous field of NETs, tumor progression is promoted by an immunosuppressed microenvironment created by a plethora of tissue-infiltrating immune cells [7], so that tumor immune infiltrate can predict poor prognosis [8,9]. Moreover, immune microenvironment features changes are associated with G1/G2 to G3 transition [10]. Changes in circulating leukocytes and PBMCs subpopulations can mirror the local alteration of the microenvironment [11]; thus, flow cytometric analysis may provide important insights into the immune status of the patient by providing information about the numbers and phenotypes of the immune cells. In NETs, leukocytes subpopulations and PBMCs are not completely investigated even if immunological alterations could represent a signal of neoplastic spread [7,12]. Indeed, tumor-infiltrating myeloid cells are involved in key processes during tumor growth and spread [13,14]. Tie2-expressing monocytes (TEMs), a subset of monocytes that express the angiopoietin receptor, play an important role in tumor angiogenesis [6]. Indeed, TEMs are selectively recruited to tumor microenvironment and promote angiogenesis in a paracrine manner, accounting for most of the proangiogenic activity of myeloid cells in tumors [15]. Tie2 is a tyrosine kinase receptor that regulates vascular quiescence [16]; the ligands of Tie2, Angiopoietin-1 (Ang-1) and -2 (Ang-2), regulate its activation status; particularly Ang-2 can be produced by tumor cells promoting angiogenesis and inflammatory response, and increasing endothelial permeability and tumor cell transmigration [17,18]. Furthermore, Ang-2 regulates Tie2 expression of TEMs markedly enhancing their innate proangiogenic activity [19]. There is evidence that angiopoietin pathway promotes NET progression, and its activation is correlated with tumor burden [20][21][22]. In detail, serum levels of Ang-2 and Ang-1 [20,21,23] together with the soluble fraction of the Tie2 receptor [21,24] and TEMs [21] are elevated in patients with NET. Starting from the available evidence we performed a pilot study to perform a multiparametric peripheral immune cells analysis in patients with GEP-NET, which has never been disclosed; moreover, we investigated the crosstalk between immune and angiopoietins profile.

Study design and participants
In this cross-sectional case-control pilot study we enrolled 30 subjects: 15 patients with well differentiated G1 and G2 GEP NET (study group) and 15 age and sex-matched healthy volunteers (control group). The study was conducted at a referral center in Italy between September 2019 and November 2021. The primary outcome of the study was the quantification of PBMCs subpopulations compared with healthy controls. The main secondary outcome was the evaluation of cellular and circulating angiopoietin pathway compared with healthy controls. The inclusion criteria for patients with well differentiated GEP NET were: age between 18 and 80 years, histologically proven diagnosis of G1 or G2 GEP NETs, patients must be naïve to treatment. The exclusion criteria for both groups included: active nonneuroendocrine neoplasms; significant respiratory, hepatobiliary, or pancreatic diseases, severe chronic kidney disease stages 4 to 5, infections, surgery, or trauma requiring hospitalization within 2 months before study enrollment and any blood or rheumatic disorders in the last 5 years. All patients and matched controls provided written informed consent prior to study participation. The study was approved by the local review board of Sapienza University of Rome (protocol 0645/2020, reference No. 5917), and was conducted in accordance with the Declaration of Helsinki. This study adhered to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines for reporting.

Blood sample analysis
Blood samples were collected from patients and controls at 8 am after overnight fasting. The absolute white blood cell count in the peripheral blood was determined using the Sysmex optical hematology analyzer (Roche) and was reported as the number of cells per microliter of whole blood. PBMCs were isolated from fresh whole blood using a Ficoll-Paque density gradient for cytometry analyses. The samples were analyzed using the CytoFLEX S flow cytometer (Beckman Coulter). Biexponential analysis was performed using CytExpert (Beckman Coulter) and FlowJo V10 (TreeStar) software. Monocytes were identified as CD3 − CD14 + cells and were further divided into classical (CD14 ++ CD16 − ), intermediate (CD14 + CD16 + ), and nonclassical (CD14 + CD16 ++ ) monocytes [25]. T lymphocytes were identified as CD19 − CD3 + cells after lymphocyte gating and were subsequently analyzed for surface expression of CD4 and CD8. Natural killer (NK) cells were identified as CD14 − CD19 − CD3 − CD56 + cells. CD3 − CD56 + NK cells were further analyzed for surface expression of CD16. NK cells were further divided into two subsets based on CD56 density: CD56 low cells and CD56 high cells. Monocytes were also analyzed for Tie2 surface expression. Absolute cell counts were obtained by multiplying subpopulation percentages with absolute monocyte and lymphocyte counts, as determined using the Sysmex optical hematology analyzer. Serum angiopoietin pathway mediators were evaluated by ELISA, using the following kits: human Ang-1 ELISA Kit (Thermo Fisher Scientific, Waltham, MA, USA); human Ang-2 ELISA Kit (Thermo Fisher Scientific, Waltham, MA, USA); human soluble Tie2 (sTie2) ELISA Kit (Abcam, Cambridge UK). All analyses were performed in the same laboratory, according to manufacturer's instructions.

Statistical analyses
Continuous variables are reported as mean ± standard deviation or as median (interquartile range) when appropriate. Normally distributed variables were assessed using the Shapiro-Wilk test. Homoscedasticity and homogeneity of variances were assessed by visual inspection and with Levene's test. Differences between patients with well differentiated GEP NET and age-and sex-matched controls were evaluated using the t test for normally distributed variables and using the nonparametric Mann-Whitney test for nonnormally distributed variables. Differences between the binomial proportions of the two independent groups of a dichotomous-dependent variable were assessed for homogeneity using the chi-square test or Fisher's exact test, as appropriate. Correlations between variables were estimated using the Pearson correlation for normally distributed variables and using the Spearman correlation for nonnormally distributed variables. All statistical analyses were performed with SPSS Statistics version 27.0 (IBM SPSS Statistics Inc., Chicago, IL, USA) and GraphPad Prism for macOS (version 9.0, GraphPad Software, LLC).

Results
We screened 16 consecutive patients with well differentiated GEP NET. One patient was excluded because of synchronous myeloproliferative neoplasm. Thus, 4 women and 11 men patients with well differentiated GEP NET were enrolled in addition to 15 age-and sex-matched controls. The mean age of the patients was 60.3 ± 9.7 years. G1 NETs were 46.7%, G2 were 53.3%. Locally advanced or metastatic disease represented the 80%. The control group consisted of 9 women and 6 men with a mean age of 52.7 ± 15.8.
The baseline characteristics of the study population are summarized in Table 1.
Immune cell profiling revealed a lower CD3 − CD56 + NK cell count in patients with NET   Total CD3 + T lymphocyte count was not significantly different between the study groups. However, a decrease in percentage (55.4 ± 8.1% vs 63.9 ± 6.6%; p = 0.004) and in absolute count (554 ± 307 cells/µl vs 820 ± 285 cells/µl; p = 0.02) of CD4 + T helper lymphocytes were found in NET patients.
Immune profile assessment of the study population is summarized in Table 2 and Fig. 1.
Angiopoietin profile assessment of the study population is summarized in Table 3 and Fig. 2.
To better investigate whether immune cells and angiopoietins levels are related in patients with GEP-NET we performed a correlation analysis, which is shown in Table 4. We found a direct correlation between total CD3 − CD56 + NK cells and CD4 + CD8 + T lymphocytes. CD56 + CD16 + NK cells were directly correlated with intermediate monocytes and non-classical monocytes, and inversely with classical monocytes. CD56 high NK cells were inversely correlated with CD4 + CD8 + T lymphocytes and TEMs. CD56 low NK cells were directly correlated with CD4 + CD8 + T lymphocytes. Moreover, we found a direct correlation between CD56 low NK cells and TEMs. Total monocytes were directly correlated with CD4 + CD8 + T lymphocytes and inversely with CD4 + T lymphocytes. There was a direct correlation between intermediate monocytes and CD4 + CD8 + T lymphocytes. Nonclassical monocytes were directly correlated with total CD3 + T lymphocytes. Finally, there was a direct correlation between CD4 + CD8 + T lymphocytes and TEMs.
No Both cellular and serum angiopoietin pathway mediators did not correlate with tumor grade, primary tumor site, and presence of distant metastases. However, Ang-1 levels were inversely correlated with Ki67 (r s = −0.52, p = 0.048).

Discussion
The present original pilot study describes, to the best of our knowledge, for the first time the complete immunophenotyping Data are expressed as means ± standard deviation c Statistically significant difference between the groups and its correlation with angiopoietin pathway of patients with well-differentiated GEP-NET. Our assessment of immune profile highlights that patients with GEP-NET show a downregulated immune profile in terms of a low count of cytotoxic NK cells, intermediate monocytes and T helper lymphocytes, and a high count of non-classical monocytes. Angiopoietin cellular profile shows a high count of TEMs endowed with a proangiogenic activity and elevated serum angiopoietins levels. In these patients TEM could represent the key link for a crosstalk between immune system and angiogenesis. This last evidence has never been demonstrated in the literature. Below we discuss our results based on the analyzed PBMCs subpopulations.

NK cells
To date, data about NET microenvironment showed that numerous immune cells including NK cells, B and T lymphocytes, macrophages, mast cells, and dendritic cells, infiltrate NETs, thus creating an immunosuppressed microenvironment favorable for tumor progression [7]. NK cells have been shown to have reduced cytolytic activity in patients with GEP-NET [26]; moreover, NK cell activity has been related to disease status, decreasing in patients with disease progression and increasing in patients responding to therapy [26]. However, circulating subpopulations of NK cells have not yet been studied within NETs. NK cells can be divided based on CD56 surface density into CD56 low NK cells, with higher cytotoxic activity, and CD56 high , with lower cytotoxic activity [27]. In fact, CD56 low NK cells contain higher amounts of perforin, granzymes and cytolytic granules [28]. CD56 high NK cells are responsible for higher cytokine production than CD56 low NK cells [27]. CD56 high NK cells are more immature forms than CD56 low [29] and differ in CD56 low under the influence of the surrounding microenvironment [30]. Beyond their innate cytotoxic activity, thanks to the expression of the CD16 receptor, NK cells can activate antibody-dependent cell-mediated cytotoxicity (ADCC) and, therefore, interface with adaptive immunity [31].  In our study, we found that NET patients have a significant reduction of peripheral total NK cells. Moreover, subset analysis showed a reduced percentage and absolute count of CD16 + NK cells, and a reduced absolute count of CD56 low NK. The depletion of NK cells may be caused by a protracted activation and subsequent downregulation due to the natural history of GEP-NETs. Indeed, given the pathological features of GEP-NETs a long host-tumor interaction can be established in these patients, thus a prolonged immune activation could be hypothesized. Together with functional data [7,26], our evidence suggest that the so called NK exhaustion, which has been sustained in different type of cancer [32], may also be present in GEP-NETs representing a mechanism of tumor immune escape leading to tumor growth and progression.

T helper lymphocytes
Along with innate immune response, adaptive immunity plays a relevant role in cancer control [33]. Indeed, CD3 + CD4 + T helper lymphocytes can exert a direct cytolytic action or modulate the tumor microenvironment. Furthermore, they can increase grade and extent of CD3 + CD8 + cytotoxic T lymphocytes responses [34]. In patients with GEP-NET, the presence of a tumor infiltrate of CD3 + CD4 + T helper lymphocytes, and CD3 + CD8 + cytotoxic T lymphocytes has been demonstrated [35,36]. In a cohort of patients with well-differentiated NET of pancreatic origin, tumor lymphocyte infiltrate did not appear to be associated with tumor grade or other clinical-pathological variables. However, progressive activation of the immune system during the progression and accumulation of tumor mutations has been hypothesized since most patients with liver metastases (97%) had some degree of CD3 + infiltration [36].
In our study, the mean values of the circulating subpopulation of CD3 + CD4 + T helper lymphocytes were significantly lower than in a control population. This data seems to confirm the abovementioned findings. In fact, exhaustion of CD3 + CD4 + T helper lymphocytes could represent a marker of chronic immune activation [37]. In this view our results would show that peripheral immune cytotoxic response reflects a microenvironment supporting tumor survival. This data became particularly interesting if we focus on the strict relation between immune system and angiopoietin pathway, exerting a key role in NET maintenance and progression.

Monocytes
Myelomonocytic cells are known to facilitate angiogenesis, tumor progression and metastasis [38]. Monocytes exerts a complex role in cancer, having both pro-tumoral and antitumoral activities [39]. Human monocytes can be divided in three main populations: classical, non-classical, and intermediate [25]. Having a cytotoxic activity, classical monocytes are recruited into tumoral tissue, within the tumor microenvironment their antitumoral function may be hampered, indeed they could differentiate into tumor-associated macrophages (TAM) which facilitate tumorigenesis by   promoting immune suppression [39]. TAM have been detected in both primary tumor tissue and metastases of patients with NET [7]. The tumor infiltrate of TAM has been positively correlated with the degree, stage, and proliferative activity of the tumor [40]. Intermediate monocytes are the main population responsible for T lymphocyte activation, while their role in cancer is not clear [41]. Nonclassical monocytes seem to have a leading role in creating a pro-tumoral microenvironment by stimulating tumor angiogenesis and suppressing T lymphocytes function [39]. However, some evidence reports their role as anti-tumoral cells given their patrolling role as scavengers of tumor cells and debris [39]. Moreover, non-classical monocytes can recruit NK in tumoral tissue and inhibit regulatory T cells [39]. Data on circulating monocytes in NETs are scarce. In our study, patients with GEP-NET showed an increased number of non-classical monocytes and a reduced number of intermediate monocytes compared to the control population. The imbalance between non-classical monocytes and classical monocytes aside from favoring tumor growth and progression could also increase tumor angiogenesis, whereas the depletion of intermediate monocytes could indirectly promote a pro-tumoral microenvironment reducing T cells activity.

The link between immune system and angiopoietin pathway
Despite the high vascularization is a well-known feature of NET [42], the available data on the levels of circulating angiogenic factors in NETs are scarce and inconsistent. Tumor angiogenesis represents a key mechanism in the biology of these neoplasms since it is implicated both in local tumor growth, favoring tumor metabolism [43], and in metastatization of cancer cells. The newly formed vessels, in fact, thanks to their greater permeability, favor cellular extravasation [44].
Cancer cells can produce several angiogenic factors, including the Ang family and their receptors. Ang-2 acts as a Tie2 receptor antagonist on endothelial cells, promoting angiogenesis and antagonizing Ang1-mediated blood vessel stability [45,46]. Serum levels of Ang-2 and Ang-1 [20,21,23] together with the soluble fraction of the Tie2 receptor [21,24] are higher in patients with NET compared to controls. Some authors found that in patients with GEP-NET, show high levels of TEMs which could be recruited in the tumor microenvironment, where they could cooperate in tumor growth and the development of metastases [21]. Our study corroborated this hypothesis, demonstrating that patients with GEP-NET have higher plasma levels of TEMs compared to a control population. Moreover, we confirmed that both Ang-1 and Ang-2 levels are higher in GEP-NET patients compared to controls, and we found that Ang-1 levels are inversely correlated with Ki67, in line with the Ang-1 role exerted in vessel stability [45,46]. In vitro, Ang-2 stimulates TEMs to express several proangiogenic genes, interleukin (IL)-10, an immunosuppressive cytokine, and CCL17, a chemokine for regulatory T cells [19]. Indeed, TEMs suppresses T cell proliferation via Ang-2-stimulated release of IL-10; moreover, IL-10 increases CD4 + to CD8 + T cells ratio and promotes the expansion of FoxP3 + regulatory T cells [47], which are known to be elevated in midgut neuroendocrine neoplasms [35] and represents an independent prognostic factor [9]. Therefore, we could speculate that TEMs in GEP-NETs represent the key link for a crosstalk between immune system and angiogenesis.
This study has some limitations. The study cohort is small and included patients with pancreatic and intestinal well differentiated NET (G1 and G2), and some of the analyses performed are exploratory and therefore the results need be validated in larger studies. Furthermore, the study lacks from histological data which could have been useful to confirm the data obtained from peripheral blood. Nevertheless, several studies in the literature analyzed tissue microenvironment in patients with NET but circulating profile of host with GEP-NET as never been explored until now. Finally, our study lacks functional data from PBMCs.
In conclusion, our study highlighted the crucial role of the immune system in the interaction between host and GEP-NET. We presented the features of GEP-NET circulating immune environment that are summarized as following: (i) a low count of cytotoxic NK cells; (ii) a low count of T helper lymphocytes (iii) a low count of T cellstimulating intermediate monocytes; (iv) a high count of non-classical pro-tumoral monocytes; (v) TEMs involvement in tumor angiogenesis.
Overall, the data of the present original pilot study led to a better understanding of the host's immune system response to GEP-NET occurrence. The circulating immune profile shows a downregulated immune environment. This result could mirror known data on the tissue immune microenvironment, which plays a key role in the maintenance and progression of NET.
Further prospecting investigations are needed to understand if the immune and angiogenic circulating profile could be representative of the NET immune microenvironment, and if they could represent biomarkers for treatment and prognosis.
Funding The analyses were supported by the ministerial research project PRIN 2017Z3N3YC.

Compliance with ethical standards
Conflict of interest The authors declare no competing interests.
Ethical approval This study was performed in line with the principles of the Declaration of Helsinki. The study was approved by the local review board of Sapienza University of Rome (protocol 0645/2020, reference No. 5917).
Consent to participate All patients gave their written consent to sample collection and their use data for the purposes of research.
Consent for publication Consent for publication has been obtained from the patients.