A hemoperfusion column selectively adsorbs LAP+ lymphocytes to improve anti-tumor immunity and survival of tumor-bearing rats

TL;DR

A new hemoperfusion column removes LAP+ immune cells, boosting anti-tumor immunity and survival in rats with cancer.

Contribution

A novel hemoperfusion column that selectively adsorbs LAP+ lymphocytes to enhance anti-tumor immunity in tumor-bearing rats.

Findings

DHP significantly reduced LAP+ Tregs in peripheral blood and tumor tissues.

Cytotoxic T-lymphocytes increased in tumor-bearing rats after DHP treatment.

Detached cells from the adsorbent negated the anti-tumor effect, highlighting their inhibitory role.

Abstract

Reducing the number of immunosuppressive cells in blood is a potential strategy for activating anti-tumor immunity, which provides a promising approach to cancer treatment. In this study, we developed an adsorbent designed to selectively target and adsorb lymphocytes expressing latency-associated peptide (LAP), which is abundantly expressed on the surface of CD4+ regulatory T cells (Tregs) and CD14+ monocytes. We investigated whether diethylenetriamine-conjugated polysulfone adsorbent-based direct hemoperfusion (DHP) enhances anti-tumor immunity in a rat cancer model with KDH-V liver cells. Our findings revealed that DHP significantly reduced LAP+ Tregs in both peripheral blood and tumor tissues in treated mice. Consequently, cytotoxic T-lymphocytes increased in tumor-bearing rats. The anti-tumor effect was negated by the addition of cells detached from the absorbent, indicating that these cells play a crucial role in inhibiting the observed therapeutic effect. The results suggest that depleting LAP+ immunosuppressive cells in blood can enhance anti-tumor immunity and improve survival of patients.

Figures

LAP+ cell adsorbent material and direct hemoperfusion.(A) Chemical reaction for preparation of the adsorbent. (B) Scheme of direct hemoperfusion. Blood was drawn from the femoral artery and passed through the column driven by a peristatic pump to return to a femoral vein.

Peripheral blood cell changes in rats carrying KDH-V cells.Unimmunized rats (n = 3) and rats immunized with irradiated KDH-V cells (n = 3) were subcutaneously injected with 3 × 105 KDH-V cells into the backs. Blood (0.5 mL) was collected from the subclavian vein every week. (A) Tumor growth size is shown as the average of the larger and smaller diameters. (B, C) White blood cell counts in unimmunized rats (B) and in immunized rats (C). (D) The granulocyte/lymphocyte ratio in peripheral blood of rats was calculated on the basis of results for (B) and (C). (E) Platelet counts in unimmunized and immunized rats. P values in comparison between the percentage on day 0 and the percentages on the other days were calculated using the paired Student’s t-test. * : P < 0.05. P values in parentheses are indicated as references because a blood sample could not be collected due to death of one rat.

Percentages of LAP+ cells in peripheral blood cells of rats carrying KDH-V cells. A representative flow cytometry profile is shown as a gating example. (A) White blood cells were gated as CD45-positive cells and then (B) lymphocytes (P2), monocytes (P3), and granulocytes (P4) were separated in an FSC/SSC plot. (C, D) CD4+ T cells and CD8+ T cells were identified among CD3+ cells and (E) B cells were identified as CD45RA+ cells among CD3- cells. The percentages of LAP+ cells in CD4+ T cells, CD8+ T cells, B cells, granulocytes, and monocytes were determined (E–I). The means and standard errors of percentages of LAP+ cells in each population were calculated in (J) unimmunized rats (n = 3) and (K) immunized rats (n = 3 on days 0 to 21 and n = 2 on day 34). P values in comparison between the percentage on day 0 and the percentages on the other days were calculated using the paired Student’s t-test. * : P < 0.05, : P < 0.01, *: P < 0.001.

Effects of the DHP column on LAP+ cells in peripheral blood of unimmunized rats carrying KDH-V cells. (A) DHP was performed 12 days after subcutaneous inoculation of KDH-V cells and blood was collected before and after DHP. (B) Means and standard deviations of total white blood cell, lymphocyte, and granulocyte counts before and after DHP (n = 3). (C) Means and standard deviations of platelet counts before and after DHP (n = 3). (D) Means and standard errors of the percentages of LAP+ cells in CD4+ T cells and CD8+ T cells (n = 3). P values were calculated using the paired Student’s t-test. * : P < 0.05.

LAP+ T cells in tumor tissues and cytotoxic T-lymphocyte activity before and after DHP. (A) DHP was performed 7 days after subcutaneous inoculation of KDH-V cells, and blood, tumor tissues, and spleens were collected 8 days after DHP. (B, C) Means and standard deviations of percentages of LAP+ cells in CD4+ T cells and CD8+ T cells in peripheral blood (B) and tumor tissues (C) were calculated (n = 2). (D) CTL activity of spleen cells against KDH-V cells was analyzed using a flow cytometer. The percentage of cytotoxicity was measured at different effector/target ratios (n = 3). (E, F) CTL activity against KDH-V cells of spleen cells from immunized rats was measured in two independent experiments (Exp. 1: averages and standard errors of quadruplicate culture, Exp. 2: averages and standard errors of 3 and 4 rats in the no treatment group and the DHP group, respectively). P values in comparison between the no treatment group and the DHP group were calculated using Student’s t-test. * : P < 0.05, : P < 0.01, *: P < 0.001. (G, H) Suppressive activity of cells adsorbed in the DHP column was determined by addition of detached cells in the CTL activity culture. Means and standard deviations of percent cytotoxicity were calculated in different numbers of detached cells (n = 3). P values in comparison between culture without detached cells (0%) and culture with detached cells (10% to 100%) were calculated using the unpaired Student’s t-test. * : P < 0.05.

Survival of rats carrying KDH-V cells after DHP treatment that removed LAP+ cells. (A, B) KDH-V cells were subcutaneously inoculated in the backs of rats after immunization with irradiated KDH-V cells. Immunized rats were treated with DHP 5 to 7 days after tumor inoculation. Survival rates of DHP-treated rats (red), non-treated rats (blue), and rats treated with a control column (orange). (C, D) KDH-V cells were subcutaneously inoculated in the backs of rats without immunization with irradiated KDH-V cells. Immunized rats were treated with DHP 7 to 13 days after tumor inoculation. Survival rates of DHP-treated rats (red) and non-treated rats (blue). (E, F) KDH-V cells were subcutaneously inoculated in the backs of rats after immunization with irradiated KDH-V cells. Immunized rats were treated with gemcitabine (GEM) and/or DHP 5 to 11 days (when tumors were palpable) and 7 to 14 days after tumor inoculation, respectively. Survival rates of gemcitabine plus DHP-treated rats (red) and gemcitabine-treated rats (blue). P values were calculated using a log-rank test (SPSS software).

Full text

Introduction

The development of cancer treatments using new strategies is needed because there are still no complete cures for progressive cancers despite recent advances in cancer treatments. Immunotherapy using immune checkpoint inhibitors (ICIs) has recently been used as a treatment strategy in addition to surgical resection, chemotherapy, and radiotherapy. Inhibitors of PD-1/PD-L1 binding, which enhance T cell responses, have shown clinical efficacy for various tumors [1, 2]. Furthermore, anti-CTLA-4 antibody prolonged survival in the treatment of advanced melanoma [3]. However, not all cancer patients respond to treatment with ICIs and serious adverse events such as the development of autoimmune disease, which are called immune-related adverse effects (irAEs), are observed in some cases. Moreover, cancer recurrence despite continuous administration has been reported [4–6]. Thus, immune activation based on new methods is required for greater tumor reduction and improvement in the survival of patients.

Immune regulatory factors in tumor tissues and blood of hosts carrying tumors include immunosuppressive cytokines and cells. One of the immunosuppressive proteins is transforming growth factor (TGF)-β, which exists as a complex with latency-associated peptide (LAP) in plasma and on the cell surface of immunosuppressive cells. The expression of TGF-β in a tumor was shown to be positively correlated with poor prognosis in patients with advanced gastric cancer [7]. Immunosuppressive cells in hosts carrying tumors were identified in CD25^+ ^CD4^+^ T cells, some of which are regulatory T cells (Treg) expressing the transcription factor Foxp3, called conventional Treg cells (cTreg) [8, 9]. On the other hand, it was reported that CD4^+^ T cells expressing a complex of TGF-β and LAP on the cell surface (LAP^+^CD4^+^ T cells) have a regulatory function [10]. LAP^+^CD4^+ ^Foxp3^+^ T cells are a subgroup of Tregs and are enriched in blood and tumor tissues of patients with colorectal cancer [11]. LAP^+^CD4^+ ^Foxp3^-^ T cells in colon cancer tumors showed a 50-fold higher immunosuppressive function than did cTreg cells [12]. LAP^+^CD8^+^ T cells were also reported to be immunosuppressive cells, suggesting that LAP^+^ cells are broadly immunosuppressive [13].

Increases of LAP^+^ cells were reported in tumor tissues and peripheral blood of various cancer patients including patients with colon, liver, and pancreas cancers [14–16]. Not only T cells and B cells but also immature myeloid cells include LAP^+^ cells that exhibit immunosuppressive properties [17–19]. Therefore, we thought that removing immunosuppressive proteins and regulatory cells would activate anti-tumor immunity. This speculation is supported by the results of our previous study showing that treatment of tumor-bearing rats with a column that adsorbs TGF-β resulted in prolonged survival [20] and the results of another study showing that reduction of LAP^+^ cells by administration of an anti-LAP antibody activated anti-tumor immunity in tumor-bearing mice [21]. Based on the results of those studies, we thought that the removal of LAP^+^ cells, a representative type of TGF-β-producing cells, would be more effective than removal of TGF-β proteins in tumor therapy.

In the present study, we examined the increase in the number of LAP^+^ cells in rats inoculated with liver cancer cells, KDH-V cells, and the survival of the cancer-bearing rats after direct hemoperfusion (DHP) with a LAP^+^ cell adsorbent column. The adsorbent column reduced the percentage of LAP^+^ cells in peripheral blood and upregulated cytotoxic T-lymphocyte (CTL) responses against tumor cells, resulting in improvement in survival of the rats.

Materials and methods

Preparation of tumor-bearing rats

This study was carried out in strict accordance with the Guidelines for the Husbandry and Management of Laboratory Animals of the Research Center for Animal Life Science at Shiga University of Medical Science and in strict accordance with Fundamental Guidelines for Proper Conduct of Animal Experiments and Related Activities in Academic Research Institutions under the jurisdiction of the Ministry of Education, Culture, Sports, Science and Technology, Japan. The protocol was approved by the Shiga University of Medical Science Animal Experiment Committee (Permit numbers: 2015-5-12, 2017-3-14, 2020-4-12). All procedures were performed by institutionally licensed researchers under anesthesia with medetomidine (0.15 mg/kg), midazolam (2.0 mg/kg), and butorphanol tartrate (2.5 mg/kg) or sodium pentobarbital (200 mg/kg). The rats were monitored daily during the study to undergo veterinary examinations to help alleviate suffering. The rats were euthanized in a sevoflurane chamber at a humane endpoint when the tumor diameter reached 3 cm. The other rats were observed as long as 80 days after tumor inoculation and were thereafter euthanized for sample collection.

We previously established a tumor-bearing rat model by inoculating syngeneic WKAH/Hkm rats with transplantable KDH-8 cells derived from chemically induced liver cancer developed at Hokkaido University [22], and the mean survival time of the rats was 55 days. Most of the KDH-8 cells are adherent to polystyrene dishes when cultured in RPMI1640 medium (Nacalai Tesque, Kyoto, Japan) supplemented with 10% fetal calf serum (FCS). After the floating cells were cultured repeatedly, completely floating cells were obtained and the cells were named KDH-V cells. KDH-V cells (3 × 10^5^ cells) were dispersed in phosphate-buffered saline (PBS) and injected subcutaneously into the backs of WKAH/Hkm rats to prepare tumor-bearing rats. Gemcitabine hydrochloride (1 mg) (Eli Lilly Japan K. K., Kobe, Japan) was injected into tumor tissues 5 to 11 days after tumor inoculation when the tumor was palpable. WKAH/Hkm rats were obtained from Japan SLC Inc. (Hamamatsu, Japan) and were reared at the Research Center of Animal Life Science, Shiga University of Medical Science.

Preparation of rats immunized with irradiated KDH-V cells

To induce KDH-V-specific immune responses in WKAH/Hkm rats, KDH-V cells were irradiated with 10,000 R of X-rays using an irradiator (MBR1520R; Hitachi Medico, Ltd., Tokyo, Japan) at the Central Research Laboratory, Shiga University of Medical Science. Irradiated KDH-V cells (X-KDH, 1 × 10^7^ cells) in 0.5 mL of PBS were inoculated subcutaneously in the backs of 4- to 8-week-old WKAH/Hkm rats. After 3 weeks or more, the immunized rats were subjected to the following experiments.

Analysis of blood cells

Blood (0.5 mL) was collected from the subclavian vein using a 23G × 1/4″ injection needle under anesthesia with medetomidine, midazolam, and butorphanol tartrate. The blood cell composition was measured using VetScan HMII (Abaxis Inc., Union City, CA).

For flow cytometric analysis, an antibody cocktail shown in S1 Table was added to 90 μL of whole blood cells and the cells were incubated at room temperature for 30 min, and then the cells in a washing buffer were centrifuged at 1,400 rpm for 5 min. Thereafter, red blood cells were lysed in FACS™ Lysing Solution (BD Biosciences, Franklin Lakes, NJ). More than 50,000 cells were analyzed with a CytoFlex S flow cytometer (Beckman Coulter, Inc., Brea, CA).

In vitro CTL assay

Effector cells were prepared from homogenized spleens after purification using Lymphocyte Separation Medium (density: 1.077) (Wako Pure Chemicals, Osaka, Japan). For target cells, KDH-V cells (1 × 10^7^ cells/mL in PBS) were labeled with carboxyfluorescein diacetate succinimidyl ester (CFDA-SE) (Dojindo Laboratories, Kumamoto, Japan) for 10 min at 37°C. For a CTL assay, effector cells and target cells were mixed and cultured at E/T ratios of 100, 50, 25 and 12.5 in 48-well plates for 22 h at 37°C (n = 4). After culture, propidium iodide (PI) was added and 1 × 10^5^ cells were acquired for the flow cytometer assay. PI-positive cells were recognized as dead KDH-V cells among the carboxyfluorescein succinimidyl ester (CFSE)-positive cells.

Detection of interferon-gamma (INF-γ) production using ELISpot

For an ELISpot assay, the IFN-γ EL-585 kit (R&D Systems, Minneapolis, MN) was used according to the manufacturer's instruction. Purified peripheral blood mononuclear cells (PBMC) were cultured for 20 h in ELISpot plates and then colored. IFN-γ-positive spots were counted using the ImmunoSpot S6 ULTRA Basic Analyzer (Cellular Technology Limited, Cleveland, OH) (n = 6 to 8 wells).

Preparation of hemoperfusion columns

According to the chemical formula shown in Fig 1A, chloro-acetamidomethylated polysulfone (CAMPS; DS =  0.5) was prepared by amidomethylation of polysulfone (Mn ~ 22,000; Sigma-Aldrich) with N-hydroxymethyl-2-chloroacetamide according to the following method. N-Hydroxymethyl-2-chloroacetamide (3.4 g = 0.027 mol) was dissolved in a mixture of nitrobenzene (30 mL) and sulfuric acid (60 mL) at 0 – 10°C with stirring. The resulting solution was immediately added slowly to a nitrobenzene solution (200 mL) containing polysulfone (22.1 g = 0.05 mol) (Sigma-Aldrich 428302) at 5 -15°C with intense stirring and then gentle stirring for 2 h at 20°C. The reaction mixture was poured into 3 L of cold methanol to precipitate the polymer.

LAP+ cell adsorbent material and direct hemoperfusion.(A) Chemical reaction for preparation of the adsorbent. (B) Scheme of direct hemoperfusion. Blood was drawn from the femoral artery and passed through the column driven by a peristatic pump to return to a femoral vein.

For purification, the obtained polymer was dissolved in 150 mL of N,N-dimethylacetamide (DMAc) and was mixed with chloroacetyl chloride (1 mL) with stirring at room temperature for 20 h. The resulting mixture was poured into distilled water. The obtained precipitate was immersed in 2 L of distilled water containing 3 g of sodium bicarbonate for 5 h. After the precipitate had been dried in a vacuum dryer at 50°C, the precipitate was extracted with methanol. The polymer was reprecipitated with DMAc-methanol, resulting in 22 g of the precipitate. The chemical structure was characterized by infrared (IR) spectra and proton nuclear magnetic resonance (^1^HNMR) spectra (S1 Fig). Substitution of the amidomethyl group was estimated to be 50% from an area ratio of methyl-hydrogen (3.80 ppm and 4.35 ppm) in a chloroacetoamidomethyl group to isopropylidene-hydrogen (1.66 ppm; singlet) in polysulfone.

For coating of the polymer on non-woven fabric fibers made of polyethylene terephthalate (PET), 1 g of CAMPS (DS = 0.5) was dissolved in 600 mL of dimethylformamide (DMF) and then 60 g of the PET-non-woven fabric fibers were added. Most of the DMAc was removed in a vacuum rotary evaporator. The obtained wet fibers were dried at 50°C in a vacuum dryer for preparation of CAMPS-coated fibers.

To introduce diethylenetriamine as a functional group, CAMPS-coated fibers (50.0 g) were immersed in 500 mL of 1% diethylenetriamine-dimethyl sulfoxide (DMSO) solution and were heated for 4 h at 50°C. The resulting fibers were washed in distilled water and dried to obtain diethylenetriamine bound fibers (50.8 g; the intended adsorbent). For preparation of a control absorbent, CAMPS-coated fibers (16.6 g) were heated for 4 h at 50°C in DMSO solution (180 mL) comprising diethylenetriamine (1 g: 9.7 mmol) and N-acetyl-L-leucine (2.08 g: 10 mmol). The resulting fibers were washed with distilled water and dried to yield a control adsorbent (17.2 g).

A cylindrical column with an inner diameter of 1 cm and a length of 2 cm was filled with 0.25 g of the resulting adsorbent for DHP. The columns were sterilized by an autoclave before DHP.

Procedure for DHP

WKAH/Hkm rats were anesthetized by subcutaneous injection of sodium pentobarbital (200 mg/kg) (Fig 1B). The groin skin on one side was incised to expose the femoral artery and vein, and then two 24 G × 3/4″ Surflo F&F needles (Terumo, Tokyo, Japan) were inserted into the artery and vein to connect to an X1-50 extension tube (Top Ltd., Tokyo, Japan). Saline containing 200 units of heparin (2 mL, AY Pharma, Tokyo, Japan) was administered through the vein. Femoral arterial blood was led to a three-way stopcock, a 2 × 4 mmφ silicon tube that was set in an MP2000 microtube pump (Tokyo Rikakikai Co., Ltd., Tokyo, Japan), the column, an air trap, and a three-way stopcock to return the blood to the femoral vein. Immediately before extracorporeal circulation, the circuit was washed with 40 mL of saline containing 200 units of heparin. DHP was performed for 1 h. During the circulation, heparin was continuously infused at a rate of 100 units/h using a mini-syringe pump (Terumo Corp., Tokyo, Japan). After DHP, the catheter was removed from the artery and 4 mL of saline was flushed to return the blood. The skin was sutured after returning the blood and removing the catheter from the vein.

Statistical analysis

Statistical analysis was performed using a two-tailed Student’s t-test for paired comparisons in Figs 2 and 3 and unpaired comparisons in Figs 4 and 5. Statistical analysis of survival was performed using a log-rank test in an IBM SSPS Statistics. Significance was determined at P < 0.05 (*P < 0.05, **P < 0.01, ***P < 0.001).

Peripheral blood cell changes in rats carrying KDH-V cells.Unimmunized rats (n = 3) and rats immunized with irradiated KDH-V cells (n = 3) were subcutaneously injected with 3 × 105 KDH-V cells into the backs. Blood (0.5 mL) was collected from the subclavian vein every week. (A) Tumor growth size is shown as the average of the larger and smaller diameters. (B, C) White blood cell counts in unimmunized rats (B) and in immunized rats (C). (D) The granulocyte/lymphocyte ratio in peripheral blood of rats was calculated on the basis of results for (B) and (C). (E) Platelet counts in unimmunized and immunized rats. P values in comparison between the percentage on day 0 and the percentages on the other days were calculated using the paired Student’s t-test. * : P < 0.05. P values in parentheses are indicated as references because a blood sample could not be collected due to death of one rat.

*Percentages of LAP+ cells in peripheral blood cells of rats carrying KDH-V cells. A representative flow cytometry profile is shown as a gating example. (A) White blood cells were gated as CD45-positive cells and then (B) lymphocytes (P2), monocytes (P3), and granulocytes (P4) were separated in an FSC/SSC plot. (C, D) CD4+ T cells and CD8+ T cells were identified among CD3+ cells and (E) B cells were identified as CD45RA+ cells among CD3- cells. The percentages of LAP+ cells in CD4+ T cells, CD8+ T cells, B cells, granulocytes, and monocytes were determined (E–I). The means and standard errors of percentages of LAP+ cells in each population were calculated in (J) unimmunized rats (n = 3) and (K) immunized rats (n = 3 on days 0 to 21 and n = 2 on day 34). P values in comparison between the percentage on day 0 and the percentages on the other days were calculated using the paired Student’s t-test. * : P < 0.05, **: P < 0.01, **: P < 0.001.

Effects of the DHP column on LAP+ cells in peripheral blood of unimmunized rats carrying KDH-V cells. (A) DHP was performed 12 days after subcutaneous inoculation of KDH-V cells and blood was collected before and after DHP. (B) Means and standard deviations of total white blood cell, lymphocyte, and granulocyte counts before and after DHP (n = 3). (C) Means and standard deviations of platelet counts before and after DHP (n = 3). (D) Means and standard errors of the percentages of LAP+ cells in CD4+ T cells and CD8+ T cells (n = 3). P values were calculated using the paired Student’s t-test. * : P < 0.05.

*LAP+ T cells in tumor tissues and cytotoxic T-lymphocyte activity before and after DHP. (A) DHP was performed 7 days after subcutaneous inoculation of KDH-V cells, and blood, tumor tissues, and spleens were collected 8 days after DHP. (B, C) Means and standard deviations of percentages of LAP+ cells in CD4+ T cells and CD8+ T cells in peripheral blood (B) and tumor tissues (C) were calculated (n = 2). (D) CTL activity of spleen cells against KDH-V cells was analyzed using a flow cytometer. The percentage of cytotoxicity was measured at different effector/target ratios (n = 3). (E, F) CTL activity against KDH-V cells of spleen cells from immunized rats was measured in two independent experiments (Exp. 1: averages and standard errors of quadruplicate culture, Exp. 2: averages and standard errors of 3 and 4 rats in the no treatment group and the DHP group, respectively). P values in comparison between the no treatment group and the DHP group were calculated using Student’s t-test. * : P < 0.05, **: P < 0.01, **: P < 0.001. (G, H) Suppressive activity of cells adsorbed in the DHP column was determined by addition of detached cells in the CTL activity culture. Means and standard deviations of percent cytotoxicity were calculated in different numbers of detached cells (n = 3). P values in comparison between culture without detached cells (0%) and culture with detached cells (10% to 100%) were calculated using the unpaired Student’s t-test. * : P < 0.05.

Results

Changes in white blood cells in a KDH-V rat tumor model

We established a rat tumor model to evaluate the efficacy of treatment with a hemoperfusion column. After 3 × 10^5^ KDH-V cells were subcutaneously inoculated in the backs of naïve rats (unimmunized rats) and rats immunized previously with the irradiated KDH-V cells (immunized rats), the tumor became palpable on day 7 and thereafter the tumor gradually increased in size over time (Fig 2A, S2 and S3 Tables). Tumor sizes reached the humane endpoint 20 days and 35 days after tumor inoculation in unimmunized and immunized rats, respectively. Thus, immunization with irradiated tumor cells prolonged the survival time of the rats inoculated with KDH-V cells.

Peripheral blood cells in rats bearing KDH-V cells were examined. In unimmunized rats, the mean white blood cell count did not change until day 9 and decreased on day 13. The mean lymphocyte count decreased significantly on day 13, whereas the mean granulocyte count increased significantly on day 20 (Fig 2B, S4 Table). On the other hand, in immunized rats, the mean white blood cell counts gradually decreased until day 14 and increased on days 21 and 34. Moreover, the mean lymphocyte counts in immunized rats gradually decreased, whereas the mean granulocyte counts increased after day 21 (Fig 2C, S5 Table). Granulocyte/lymphocyte ratios increased on day 20 and day 34 in unimmunized rats and immunized rats, respectively, but differences were not significant (Fig 2D). The mean platelet counts in unimmunized rats decreased significantly on day 9 and increased on day 13, whereas the mean platelet counts in immunized rats increased on day 34 (Fig 2E). Thus, the number of peripheral white blood cells in the tumor-bearing rats changed with tumor growth.

LAP-positive cells in peripheral blood of rats carrying KDH-V cells

We determined the percentage of cells expressing LAP (LAP^+^ cells) on the cell surface by flow cytometry. Lymphocytes (P2), monocytes (P3), and granulocytes (P4) in CD45^+^ cells (P1) were defined in FSC/SSC dot plots (Fig 3A, B). Among CD3^+^ cells (P5) (Fig 3C), CD4^+^ T and CD8^+^ T cells were identified by their respective surface antigens (Fig 3D). CD45RA^+^ and CD3^-^ cells were considered as B cells (Fig 3E). Thereafter, the percentages of LAP^+^ cells (red dots) were determined in comparison with control staining (blue dots) (Fig 3E-I). The percentages of LAP^+^ cells in granulocyte marker-positive cells in the P4 and P3 fractions were defined as granulocytes and monocytes, respectively (Fig 3H, I).

We examined changes in LAP^+^ cells in the tumor-bearing rats. The mean percentages of LAP^+^ cells in CD4^+^ T cells, CD8^+^ T cells, and granulocytes decreased on day 9 (4.2%, 7.3%, and 22.2%, respectively) compared to the percentages on day 0 before tumor inoculation (21.1%, 18.5%, and 58.2%, respectively) and the mean percentages of LAP^+^ cells in CD4^+^ T cells and CD8^+^ T cells increased on day 20 in unimmunized rats (23.0% and 20.1%, respectively) (Fig 3J, S6 Table). The mean percentages of LAP^+^ B cells increased on day 9 (20.5%) and day 20 (30.7%) compared to the mean percentage on day 0 (9.4%), whereas the mean percentages of monocytes, which were mostly positive for LAP before tumor inoculation, decreased on day 9 (51.0%) and day 20 (64.0%) compared to the mean percentage on day 0 (99.4%).

The percentages of LAP^+^ cells in immunized rats showed a different tendency from that in unimmunized rats. The mean percentages of LAP^+^ cells in CD4^+^ T cells, CD8^+^ T cells, B cells, and granulocytes significantly increased on day 7 after tumor inoculation (29.7%, 32.9%, 35.1%, and 50.4%, respectively) and they decreased on day 34 (16.4%, 16.9%, 24.7%, and 10.9%, respectively) (Fig 3K, S7 Table). On the other hand, the percentage of LAP^+^ cells in monocytes gradually decreased to 71.5% on day 34. Thus, the percentages of LAP^+^ cells in lymphocytes and granulocytes showed increases at early time points but decreased in the final stage after tumor inoculation in immunized rats.

Improvement of anti-tumor immunity in tumor-bearing rats by DHP

Since removal of TGF-β using a DHP fiber column prolonged survival in a KDH-8 rat cancer model in our previous study (20), we tried to remove LAP^+^ cells, which are representative TGF-β-producing cells, in the present study. After coating the surface of the PET fiber with chloroacetamidomethylated polysulfone, a diethylenetriamine group was introduced to provide functional groups to prepare an adsorbent, which was packed in a column (Fig 1). Twelve days after inoculation of KDH-V cells into unimmunized rats, tumor-bearing rats were subjected to DHP (Fig 4A). Blood cell counts before and after circulation showed a slight increase in the lymphocyte count (Fig 4B, S8 Table), whereas granulocytes and platelets decreased to 45% and 36% of the counts before circulation, respectively (Fig 4B and C). The levels of reduction of LAP^+^ cells by DHP were 46% for CD4^+^ T cells and 50% for CD8^+^ T cells (Fig 4D, S9 Table).

To confirm whether the anti-tumor activity was increased by the DHP, 4 unimmunized rats were inoculated with KDH-V cells, and then 2 rats underwent DHP to examine blood cells, tumor-infiltrating lymphocytes (TIL), and CTLs in the spleen (Fig 5A). In the blood 8 days after DHP treatment (day 15), the mean percentages of LAP^+^ cells in CD4^+^ T cells and CD8^+^ T cells of DHP-treated rats (21% and 19%) were 33% and 32% lower than those in untreated rats (31% and 28%), respectively (Fig 5B, S10 Table). In addition, the percentages of LAP^+^ cells in CD4^+^ T cells, CD8^+^ T cells, and granulocytes in tumor tissues (TIL) were 52%, 62%, and 37% lower in DHP-treated rats than in untreated rats, respectively (Fig 5C, S10 Table). Moreover, CTL activity of the splenocytes against KDH-V cells was higher in DHP-treated rats than in untreated rats (n = 2) (Fig 5D, S11 Table). Furthermore, after the immunized rats underwent hemoperfusion with a column on day 11 after tumor inoculation, the CTL activity was examined 3 days after DHP treatment (day 14) (Fig 5E). The CTL activity against KDH-V cells was higher in splenocytes from DHP-treated immunized rats than in those from untreated immunized rats in two independent experiments (Fig 5F, S12 and S13 Tables). Thus, CTL activity specific for KDH-V cells was upregulated after DHP treatment in both unimmunized and immunized rats.

In addition, the immunosuppressive function of cells adsorbed to the column was examined. The adsorbed cells were detached from the column with PBS after hemoperfusion on day 11 after tumor inoculation. When the collected cells were mixed with splenocytes from the same unimmunized rat (Fig 5G), the CTL activity decreased depending on the number of adsorbed cells added (Fig 5H, S14 Table). This result indicates that the DHP column removes immunosuppressive cells from rats carrying a tumor.

We further examined the duration of immune upregulation after the DHP by an ELISpot assay. After immunized rats had been inoculated with KDH-V cells into their backs, they were subjected to DHP on day 12 or day 14. Peripheral blood was collected from the rats on day 17 (S2A Fig). IFN-γ-producing cells specific for the tumor antigen were detected in an ELISpot assay. The number of IFN-γ-producing spots was larger in blood of rats that had undergone DHP than in blood of untreated rats on day 5 and day 3 after the DHP (S2B Fig, S15 Table). Similar results were obtained in rats on day 1 and day 3 after DHP (S2C-H Fig, S16–S18 Tables). Thus, the upregulation of INF-γ-producing cells continued for at least 5 days after DHP.

Survival time of rats carrying KDH-V cells after DHP treatment

Finally, we examined the effects of the LAP^+^ cell adsorbent column on survival of rats carrying KDH-V cells. Immunized rats were inoculated with KDH-V cells and then DHP treatment was performed for 60 min on days 5–7 (Fig 6A). Rats in the DHP-treated group survived significantly longer than did rats in the non-treated group and the control-treated group (log-rank test, P < 0.001) (Fig 6B). On the other hand, in unimmunized rats, the DHP treatment did not extend the survival period (Fig 6C, D). These results suggest that the LAP^+^ cell adsorbent column upregulates the tumor-specific CTL response to reduce tumor progression. To examine an additive effect with alternative treatment, gemcitabine, a nucleoside analogue, was injected into subcutaneous tumors of rats and the DHP treatment was performed (Fig 6E, F). A combination of the DHP treatment with administration of gemcitabine showed significant effects on survival time of rats compared to the gemcitabine treatment, although the survival rates and survival times of rats treated with gemcitabine were lower and shorter than those of rats without the gemcitabine treatment (Fig 6B). These results indicate that the DHP treatment has effective in the rats carrying immunity specific for tumor antigens.

Survival of rats carrying KDH-V cells after DHP treatment that removed LAP+ cells. (A, B) KDH-V cells were subcutaneously inoculated in the backs of rats after immunization with irradiated KDH-V cells. Immunized rats were treated with DHP 5 to 7 days after tumor inoculation. Survival rates of DHP-treated rats (red), non-treated rats (blue), and rats treated with a control column (orange). (C, D) KDH-V cells were subcutaneously inoculated in the backs of rats without immunization with irradiated KDH-V cells. Immunized rats were treated with DHP 7 to 13 days after tumor inoculation. Survival rates of DHP-treated rats (red) and non-treated rats (blue). (E, F) KDH-V cells were subcutaneously inoculated in the backs of rats after immunization with irradiated KDH-V cells. Immunized rats were treated with gemcitabine (GEM) and/or DHP 5 to 11 days (when tumors were palpable) and 7 to 14 days after tumor inoculation, respectively. Survival rates of gemcitabine plus DHP-treated rats (red) and gemcitabine-treated rats (blue). P values were calculated using a log-rank test (SPSS software).

Discussion

We have newly developed a column that adsorbs LAP^+^ cells in peripheral blood and that extends the survival period of tumor-bearing rats by DHP. The cancer model rat used in the present study showed a high percentage of LAP^+^ T cells in peripheral blood. After treatment with the LAP^+^ cell-adsorbing column, reduction of LAP^+^ T cells was observed in tumor tissues of the cancer rats as well as in peripheral blood. The CTL activity of splenocytes in the tumor-bearing rats treated with DHP was higher than that in the tumor-bearing rats without DHP. The cells adsorbed in the column were immunosuppressive because they inhibited CTL activity dose-dependently in the cell culture. Thus, reduction of LAP^+^ cells with DHP is thought to have the potential for cancer immunotherapy.

In the present study, reduction of LAP^+^ cells in peripheral blood activated an anti-tumor immune response and extended the survival period of rats carrying the liver cancer cells. In our previous study, we revealed that removal of the immunosuppressive protein TGF-β by using a TGF-β-adsorbing column improved the survival of tumor-bearing rats [20]. Therefore, we further developed the column to remove cells producing TGF-β, i.e., LAP^+^ cells, because we postulated that the removal of LAP^+^ cells would be more effective than removal of TGF-β for enhancing anti-tumor acquired immunity. The percentages of LAP^+^ cells in blood of the column-treated rats were lower than those in blood of the untreated rats 8 days after DHP. Moreover, the CTL activity of splenocytes in the column-treated rats was higher than that in the untreated rats 8 days after DHP. Therefore, the CTL activation effect was sustained for at least 8 days after DHP treatment.

The effects of DHP on amelioration of the survival period and survival rate were significant in immunized rats but not in unimmunized rats. This result indicates that LAP^+^ T cells affect acquired immune responses including a CTL response. Priming and expansion of CTLs specific for tumor antigens follows activation of innate immunity and antigen presentation. Therefore, immunization with irradiated KDH-V cells might allow the priming of CTL responses specific for the tumor antigen and reduction of LAP^+^ cells might help the expansion and effector function of CTLs, effects of which were not seen in the unimmunized rats without priming of CTLs specific for the tumor antigen. Because the KDH-8 tumor growth, which might allow the priming of CTLs, was slower than that of KDH-V, it was thought that DHP absorbing TGF-β showed anti-tumor effects in the previous study [20]. Restoration of CTL activity specific for tumor antigens by immune checkpoint inhibitors indicates that the CTL responses are suppressed by tumor progression in human cancer patients [1, 2]. Therefore, it is thought that the present preimmunization model before inoculation with a rapid-growing tumor represents CTL responses that are primed by tumor cells in the early stage and are suppressed by tumor growth such as human cancers.

The removal of LAP^+^ cells using DHP has advantages in comparison with other immunotherapies. Administration of an anti-LAP antibody improved anti-tumor immunity of mice in a previous study [21] in which the anti-LAP antibody was injected every three days for one month after tumor implantation for elongation of survival time. In addition, the effect of antibody therapy was gradually attenuated after repeated administration as reported in ICIs [4–6]. On the other hand, the hemoperfusion therapy using the DHP column was effective for improving the survival of tumor-bearing rats with only one DHP treatment even against a gemcitabine-resistant tumor, and enhancement of the CTL response by the column was effective for at least 8 days. These are advantages of hemoperfusion therapy in anti-tumor therapy.

Therapeutic effects of an anti-CCR8 antibody to remove Tregs have been reported in mouse tumor models [23–26]. CCR8 is preferentially expressed on the cell surface of Tregs that have distributed in tumor tissues but not in peripheral blood. Therefore, it is thought that depletion of Tregs in tumor tissues elicits antitumor immunity in local tumor tissues. On the other hand, the DHP column treatment to deplete LAP^+^ T cells in peripheral blood induced a decrease in the percentage of LAP^+^ T cells not only in peripheral blood but also in tumor tissues in addition to upregulation of CTL responses specific for tumor cells in spleen cells. Therefore, the mechanism for activation of CTLs specific for tumor cells might not be the same in the case of an anti-CCR8 antibody for depletion of local Tregs and the case of a DHP column for depletion of blood LAP^+^ T cells, for example priming of CTLs in the secondary lymphoid organs.

Repeated column treatment may be considered to further improve the efficacy of the hemoperfusion therapy. One of the concerns arising from hemoperfusion therapy is reduction of platelets, which might induce a bleeding tendency. In our additional study, the platelet counts after DHP were above 300,000/μL and they returned to the basal level 5 to 7 days after hemoperfusion therapy in the unimmunized rat model (S3 Fig). Therefore, bleeding due to a low platelet count is thought to be unlikely if the interval between hemoperfusion treatments is 7 days.

Another concern of the DHP column is activation of autoimmune responses by depletion of LAP^+^ T cells including Tregs. Autoantibody responses and autoimmune gastritis were observed in mice 13 days after systemic Treg depletion but not in mice treated with anti-CCR8 antibody that depleted Tregs in tumor tissues [25]. Although the DHP column in the present study reduced LAP^+^ T cells in peripheral blood, no inflammatory symptom such as diarrhea was not noticed until 75 days after DHP. Further observation and analysis of autoimmune responses should be required in the DHP-treated rats.

It is thought that removal of LAP^+^ cells can be applied for patients with various cancers. LAP^+^ T cells, which include some Foxp3^+^ T cells, increase in patients with colorectal cancer, non-muscle-invasive and muscle-invasive urinary bladder cancer, and other cancers [14–16,27–29]. In addition, the percentage of Foxp3^+^ Tregs in patients with non-muscle-invasive urinary bladder cancer was shown to be correlated with risk stratification and recurrent-free survival [27]. The percentage of LAP^+^CD4^+^ T cells in peripheral blood of hepatocellular carcinoma patients was reported to be correlated with tumor size [29]. These results suggest that LAP^+^CD4^+^ T cells and Foxp3^+^ Tregs migrate from peripheral blood to tumor tissues or circulate in the tissues. Therefore, the removal of LAP^+^ cells could be applied for treatment of patients with various cancers who have a high percentage of LAP^+^ cells in peripheral blood. To prove the efficacy of DHP treatment in other tumor models, development of which is a limitation in the present study, we are establishing a rat pancreatic cancer model.

In the present study, we showed the efficacy of hemoperfusion column therapy in a rat cancer model. The removal of LAP^+^ cells could be applied to various tumors since an increase of LAP^+^ cells has been reported in other cancers [13–16]. Since the column removes regulatory lymphocytes, it consequently has an ameliorative effect on immune responses regardless of the type of cancer. Thus, removal of LAP^+^ cells from cancer patients may be an alternative immune therapy to ICIs.

Supporting information

S1 Table Combination of antibodies to detect surface antigens. (TIFF)

S2 Table Diameters of tumors in unimmunized rats in Fig 2A.(TIFF)

S3 Table Diameters of tumors in immunized rats in Fig 2A.(TIFF)

S4 Table Blood cell counts in the peripheral blood of unimmunized rats in Fig 2B, D, and E.(TIFF)

S5 Table Blood cell counts in the peripheral blood of immunized rats in Fig 2 C, D, and E.NT: not tested due to difficulty to collect blood samples.(TIFF)

S6 Table Percentages of LAP^+^ cells in the peripheral blood of unimmunized rats in Fig 3J.(TIFF)

S7 Table Percentages of LAP ^+^ cells in the peripheral blood of preimmunized rats in Fig 3K.NT: not tested due to difficulty to collect blood samples.(TIFF)

S8 Table Blood cell counts in the peripheral bloods in Fig 4B and C.(TIFF)

S9 Table Percentages of LAP^+^ cells in the peripheral blood in Fig 4D.(TIFF)

S10 Table Percentages of LAP ^ +^ cells in the peripheral blood and tumor tissues in Fig 5B and C.(TIFF)

S11 Table Percent cytotoxicity of spleen cells of unimmunized rats in Fig 5D.(TIFF)

S12 Table Percent cytotoxicity of spleen cells of immunized rats in Fig 5F Exp. 1.(TIFF)

S13 Table Percent cytotoxicity of spleen cells of immunized rats in Fig 5F Exp. 2.(TIFF)

S14 Table Percent inhibition of cytotoxicity in the presence of cells detached from the columns in Fig 5H.(TIFF)

S15 Table Number of IFN-γ positive spots in S1B Fig.(TIFF)

S16 Table Number of IFN-γ positive spots in S1D Fig.(TIFF)

S17 Table Number of IFN-γ positive spots in S1F Fig.(TIFF)

S18 Table Number of IFN-γ positive spots in S1H Fig.(TIFF)

S1 Fig IR and NMR spectrum analysis of chemical structure of absorbents. (A) IR spectrum (KBr tablet); A Prestige-21 Fourier transform infrared spectrophotometer (Shimadzu Corporation, Kyoto, Japan) was used in the Central Research Laboratory in the Shiga University of Medical Science. Peaks of polysulfone skeleton were detected at 2968, 1579, 1487, and 1248 cm^-1^ (arrow heads). The carbonyl (C = 0) stretch of amidomethyl group was detected at 1671 cm^-1^ (arrow). (B) NMR spectrum (Chloroform-D); A JNM-ECZ400S NMR spectrometer (JEOL Ltd., Tokyo, Japan) was used in the Central Research Laboratory in the Shiga University of Medical Science. Levels of isopropylidene hydrogen of polysulfone skeleton (6H) (a), a peak of CH_2_ in chloroacetyl group (b), peaks of CH_2_ in an aminobenzyl group (c), and aromatic hydrogen (left peaks) were 1.66 ppm (singlet), 3.8 ppm (singlet), 4.35 ppm (doublet), and 6.8 – 7.9 ppm, respectively.(PDF)

S2 Fig Duration of the effects of LAP ^+^ cell depletion on IFN-γ production. Irradiated KDH-V cells (X-KDH-V) were inoculated into rats 21 days before inoculation of KDH-V cells. (A, B) DHP treatment was performed 12 or 14 days after inoculation of KDH-V cells. Blood cells were collected on day 17. (C, D) The DHP treatment was performed 11 days or 12 days after inoculation of KDH-V cells. Blood was collected on day 12. (E, F) DHP treatment was performed 11 days after inoculation of KDH-V cells. Blood was collected on day 14. (G, H) DHP treatment was performed 16 days after inoculation of KDH-V cells. Blood was collected on day 21. (B, D, F, H) The number of IFN-γ-producing cells in peripheral blood cells was determined by an ELISpot assay. Means and standard errors of 6 or 8 wells of culture are shown. P values in indicated comparisons were calculated using Student’s t-test. * : P < 0.05, **: P < 0.01, ***: P < 0.001.(TIFF)

S3 Fig Platelet count after DHP. Column treatment was performed twice in three cancer rats. Squares and circles show platelet counts before and after DHP, respectively. The platelet counts decreased immediately after DHP (arrows). The platelet counts 5 to 7 days after the first DHP were higher than those before the first DHP.(TIFF)