Intratumoral administration of the antisecretory peptide AF16 cures murine gliomas and modulates macrophage functions

Glioblastoma has remained the deadliest primary brain tumor while its current therapy offers only modest survival prolongation. Immunotherapy has failed to record notable benefits in routine glioblastoma treatment. Conventionally, immunotherapy relies on T cells as tumor-killing agents; however, T cells are outnumbered by macrophages in glioblastoma microenvironment. In this study, we explore the effect of AF16, a peptide from the endogenous antisecretory factor protein, on the survival of glioma-bearing mice, the tumor size, and characteristics of the tumor microenvironment with specific focus on macrophages. We elucidate the effect of AF16 on the inflammation-related secretome of human and murine macrophages, as well as human glioblastoma cells. In our results, AF16 alone and in combination with temozolomide leads to cure in immunocompetent mice with orthotopic GL261 gliomas, as well as prolonged survival in immunocompromised mice. We recorded decreased tumor size and changes in infiltration of macrophages and T cells in the murine glioma microenvironment. Human and murine macrophages increased expression of proinflammatory markers in response to AF16 treatment and the same effect was seen in human primary glioblastoma cells. In summary, we present AF16 as an immunomodulatory factor stimulating pro-inflammatory macrophages with a potential to be implemented in glioblastoma treatment protocols.

www.nature.com/scientificreports/ AF16 is an active peptide of the larger antisecretory factor (AF) protein 19 . AF was discovered in the pituitary gland, cerebrum and intestinal mucosa but later found to be ubiquitously expressed by various tissues and immune cells, including macrophages [20][21][22] . Early on, studies described the anti-inflammatory effect of AF16 in a cholera-toxin colitis model 23 . When AF was blocked by monoclonal antibodies, T-cell mediated autoimmune encephalitis had worse outcomes in rats 24,25 . Moreover, AF16 was shown to reduce interstitial fluid pressure, edema and intracranial pressure after brain injury, autoimmune encephalitis, and in solid tumors [26][27][28][29] . A diet of AF-enriched egg yolk powder in xenografted glioblastoma models could improve survival, decrease tumor volumes and increase the effect of chemotherapy, tentatively by increasing uptake 30 . Importantly, exogenous AF administration has not been associated with any negative side effects 31 . However, the impact of AF on the immune aspect of the glioblastoma microenvironment in immunocompetent animals was not analyzed.
In this present study, we aim for the first time to investigate the functional effect of AF16 on glioblastoma survival in vivo and on the monocyte/macrophage population, with the emphasis on implications for the TME of glioblastoma. We also analyzed CD8 + T cells, M0 macrophages/microglia, and M1 and M2 macrophage populations in the glioblastoma TME, as well as selected inflammation-related molecules-cyclooxygenase 2 (COX-2) as a rate-limiting enzyme in the production of prostaglandins, e.g. PGE2 that is reportedly associated with glioblastoma immunosuppression and tumorigenesis 32 . Galectin 3 is associated with glioma progression and neo-angiogenesis but also with membrane damage, with galectin 3 being upregulated after temozolomide and radiotherapy treatment 33,34 . Phosphorylated Na + -K + -2Cl − cotransporter isoform 1 (pNKCC-1) is an activated ion pump essential for regulating cell volume during e.g. apoptosis or glioma cell migration and it has been shown to upregulate after TMZ treatment in glioblastoma models 35 . We have previously proved that intratumoral administration of cytostatics is safe, offers better survival with less toxic effects compared to the systemic route, and gives rise to long-lasting immunological memory (depending on the drug and model) 36 . When combined with peripheral vaccination with tumor cells, the therapeutic effect from the intratumoral drug delivery is enhanced. Based on our previous promising results and to achieve high local drug concentrations while circumventing the blood-brain barrier in glioblastoma, AF16 and TMZ were administered intratumorally.
To our knowledge, such a comprehensive study of the immune effects of AF16 in glioblastoma has not been conducted to date.

AF16 inhibits the TMZ-induced expansion of intratumoral macrophages and CD8 + T cells and increases galectin-3 and pNKCC-1 in the tumor microenvironment of murine GL261 gliomas.
For this histological experiment, we used brain tissue from a different cohort of mice where treatment was postponed assuring sufficient tumor area for analysis. Upon staining of the brain tissue, we recorded a significant increase in the tumor-infiltrating F4/80 + macrophage population in the TMZ group compared to the controls, while this increase was inhibited by combined AF16 + TMZ (Mann-Whitney U test, TMZ v control p = 0.029, TMZ v AF16 + TMZ p = 0.032, Fig. 3a). AF16 by itself did not induce a change in the F4/80 + cells intratumorally. In a similar way to F4/80, CD8 + cells were decreased by the combination of AF16 + TMZ compared to controls and TMZ alone (p = 0.029, Fig. 3b). After further staining for galectin-3, implicated in glioma progression, as well as in membrane damage and other cellular processes, we observed a significant increase of galectin-3 staining after AF16 + TMZ compared to controls (p = 0.029, Fig. 3c). A similar increase of pNKCC-1, an activated ion pump essential for cell electrolyte homeostasis, was also recorded in the AF16 + TMZ group compared to controls (p = 0.002, Fig. 3d). Staining using the M2 macrophage markers CD206 and COX-2, as well as the M1 macrophage/dendritic cell markers CD11c and MHC II did not show any significant differences between treatments ( Supplementary Fig. 1).
AF16 drives the secretome of murine M0 macrophages towards the M1 macrophage profile. Naïve RAW264.7 murine M0 macrophages were incubated with 2000 µg/ml AF16 to investigate the immunomodulatory properties of the molecule. This was compared with untreated RAW cells (M0-R) and RAW cells differentiated into M1 and M2 phenotypes. AF16 increased the majority of measured cytokines, including IL1β, IL2, IL6, TNFα, IL12p70, IL10 and KC/GRO to levels similar to M1 macrophages while their levels were reduced or unchanged in M2 macrophages, except for IL10 (Fig. 4). The analysis of collected supernatants showed that the M1 and M2 differentiation was successful based on the secretion of phenotype-specific cytokines compared to M0 macrophages (IL1β, IL6, TNFα increased in M1 and decreased in M2 cells; IL-10 increased substantially in M2 cells, Fig. 4).

AF16 modulates the secretion of proinflammatory cytokines by human M0 macrophages.
To investigate the effect of AF16 on cell line-derived and primary human macrophages, we treated cultures of M0 macrophages derived from the THP-1 cell line, as well as differentiated from primary human monocytes (see "Materials and methods" section) with 1 µg/ml and 100 µg/ml AF16 or 1 µM betamethasone as a control for inhibition of macrophage-derived inflammation. M0 macrophages were fully differentiated into M1 and M2 cells as controls for macrophage activation. In particular, an increase in IL1β from M0 macrophages after AF16 mirrors the increase in IL1β in M1 macrophages. M2 macrophages downregulated IL1β and betamethasone www.nature.com/scientificreports/ had negligible effect on IL1β levels in primary M0 macrophages (Fig. 5a). Furthermore, AF16 upregulated IL8 in supernatants of THP-1-derived macrophages but downregulated IL8 in primary macrophages while the rest of the cells upregulated IL8. Betamethasone downregulated IL8 from primary M0 macrophages (Fig. 5b). IL13 was upregulated in M0 macrophages after AF16, as well as in THP-1-derived M1 cells and downregulated in the rest of the cells and after betamethasone (Fig. 5c). IP10 was likewise upregulated in M0 macrophages and primary M1 cells after AF16 while M2 cells downregulated IP10. IP10 was undetectable in M1-T cells and BM had no effect on its production in primary M0 cells (Fig. 5d). TNFα was upregulated in M0 and primary M1 macrophages after AF16, downregulated in the other cell types and after betamethasone (Fig. 5g). MIP1β was upregulated by AF16 in THP-1-derived M0 macrophages, downregulated in primary M0 macrophages with levels undetectable in THP-1-derived M1 and decreased in M2-T cells. primary M1 and M2 cells upregulated MIP1β, while betamethasone treatment downregulated MIP-1β (Fig. 5e). MIP1α exceeded the upper detection limit in THP-1-derived M0 and M1 cells while it was downregulated in primary M0 cells both after AF16 and betamethasone. Both primary M1 and M2 cells increased MIP1α (Fig. 5f). In summary, AF-16 treatment increased the expression of IL1β, IL8, IL13, IP10, MIP1β and TNFα in the supernatants of primary and THP-1-derived macrophages. . Changes in expression of immune cell surface markers and immune-related molecules caused by AF16 and TMZ treatment compared to untreated animals. Representation of immunohistochemical staining of brain sections of GL261-bearing mice shows an increase of F4/80 + macrophages after TMZ treatment was abolished by addition of AF16 (TMZ compared to TMZ + AF16 p = 0.032 *) (a). A decrease in CD8 + T cells was recorded after the combination treatment with TMZ and AF16 compared to untreated animals and TMZ alone (CTRL compared to TMZ + AF16 p = 0.029 *; TMZ compared to TMZ + AF16 p = 0.029 *) (b). Increased expression of galectin-3 was seen after the combination treatment with TMZ and AF16 compared to untreated animals (CTRL compared to TMZ + AF16 p = 0.029 *) (c). Similarly, the combination of TMZ and AF16 also increased the expression of pNKCC1 compared to untreated animals (CTRL compared to TMZ + AF16 p = 0.002 **) (d

Discussion
In this paper, we investigated the role of AF16 as an adjuvant therapeutic agent in the context of glioblastoma TME, and its effect on the function of macrophages. Previous reports show that blockade of AF16 potentiates the immune response from T cells and exogenous administration of AF16 decreases inflammation in the gut tissue and in immune encephalitis in rats 22,24,25,29 but not in mice, despite treated animals having higher survival rates 37 . Others have shown that the parental molecule, antisecretory factor (AF), induced in mice by diet of specially processed cereals (SPC's), exerts direct antitumor effects with increased tumor apoptosis and reduced proliferation in a human xenograft glioblastoma model 30 . Moreover, SPC diet reduced the interstitial fluid pressure inside the tumor and increased doxorubicin and erlotinib uptake into tumor tissue 30 . Additionally, administration of AF16 decreased interstitial pressure in solid tumors 28 , as well as intracranial pressure and edema following brain injury 26,27 . It would be pertinent to break the chronic immunosuppression in the glioblastoma microenvironment that contains large numbers of tumor-associated macrophages [38][39][40] . Our results show that AF16 cured mice both alone and synergistically with TMZ in the immunocompetent mouse glioblastoma model, while also decreasing  The results from the in vitro treatment of human and murine M0 macrophages suggest that AF16 induces changes in the M0 secretome that mimic the M1 profile with several proinflammatory cytokines upregulated (IL-1β, IL-8, TNF-α, IP-10, MIP-1α and MIP-1β in human; and IL-1β, IL-2, IL-6, TNF-α and IL-12p70 in murine macrophages). Such AF16-induced secretion of proinflammatory cytokines by the increased M0 intratumoral macrophage population together with the effect of TMZ could be responsible for the in vivo survival benefit, tumor size decrease, and activation-induced cell death of the F4/80 + macrophages/ microglia and the CD8 + T cells 42,43 . We recorded discrepancies in levels of factors between THP-1-derived cells and primary human cells. This has been reported previously 44,45 and it reflects the inherent differences of Unexpectedly, local administration of AF16 combined with TMZ led to an increase in galectin-3 and pNKCC-1 expression intratumorally compared to untreated animals. Galectin-3 is an immunosuppressive molecule in the tumor microenvironment, and it is expressed abundantly in human macrophages and tumor cells under stress, i.e. cytostatic therapy 34,[46][47][48] . Under physiologic conditions, galectin-3 is associated with lysosomal membrane damage and autophagy, among other effects 33 . The increased expression of galectin-3 detected after AF16 and TMZ treatment could be explained by the cytotoxic effect of this therapy causing membrane damage and increased autophagy as part of regulated death of tumor cells and senescent immune cells. NKCC-1 and its active phosphorylated form (pNKCC-1) have recently been implicated to play an important role in macrophage activation 49,50 as well as in glioma invasion 51 . Ilkhanizadeh et al. reported that diet of specially processed cereals, which induce host production of AF, decreased the expression of pNKCC-1 on human glioblastoma cells growing in immune-depleted mice 30 . There is, however, no mention of the effect on the tumor-infiltrating macrophages. Given that we demonstrated that AF16 caused a surge in proinflammatory cytokines in human and murine macrophages, this activation of the M0 population can be the cause of the pNKCC-1 increase we saw in mice receiving the combination of AF16 and TMZ. After staining of brain tissue for COX-2, a key enzyme in the production of prostaglandins, we did not see a difference between treatment groups, implying that prostaglandin signaling is not affected by AF16.
In addition to macrophage modulation, we wanted to elucidate the AF16-induced effect on the inflammatory secretome of primary human glioblastoma cells. Among the proteins which expression was increased by AF16 on human glioblastoma cells were proinflammatory chemoattractants of granulocytes, T cells and monocytes/ macrophages (CCL3, CCL4, CCL7, CCL8, CCL20), T cell-stimulatory signals (CD40, TNFSF14, TNFrSF9), as well as metalloproteinases and other enzymes (MMP-1, MMP-10, uPA) or other molecules (IL8, TGFα). Several of these factors have been associated with glioblastoma growth and progression 52 , while their expression was also increased by irradiation (MMP1, CXCL6, IL8, MMP10, CXCL1). The proteins expressed by human glioblastoma cells that were decreased by AF16 have in the majority of cases been associated with lower T cell and higher myeloid cell tumor infiltration (LIF, IL6, CSF1), glioblastoma progression and poor prognosis (OPG, CDCP1,  [53][54][55] , while some factors (CCL2, Flt3L, CXCL10) were described as proinflammatory and potentially tumor-suppressive 56 . Irradiation had negligible effect on factors in the "AF16decreased" group, except on CXCL5, which it increased. Overall, the human glioblastoma secretome analysis of AF16-induced changes showed a complex picture of proinflammatory and tumor-promoting factors that were modulated by AF16, with several tumor-promoting factors decreased, which could contribute to the favorable survival in vivo. Of note, TMZ treatment induced only minor changes in the inflammatory secretome of human glioblastoma cells in our experiment. Furthermore, analysis of fresh-frozen surgical samples of human glioblastoma and Grade II astrocytoma reveals that several proteins with higher expression in glioblastoma compared to astrocytoma were downregulated by AF16 treatment (4EBP1, CCL2, CXCL10, LIF, CD40, OPG, IL6, FGF5) while correspondingly, some factors with higher expression in astrocytoma compared to glioblastoma were also upregulated by AF16 in glioblastoma (CCL3, CCL4, TGFα). This could be evidence that AF16 modifies the TME of glioblastoma to become less pro-malignant while acknowledging that several glioblastoma-associated proteins were upregulated by AF16 treatment (uPA, IL8, etc.). In summary, we propose AF16 as a novel immunomodulatory molecule that can cure immunocompetent mice and prolong overall survival in immunodeficient mice with high-grade glioma. It also stimulates the production of proinflammatory cytokines in human and murine M0 macrophages. In this regard, AF16 mimicked the cytokine pattern induced by M1 differentiation and exerted opposite effect than betamethasone on several factors. Moreover, AF16 modulated several proteins associated with glioblastoma progression and inflammation, secreted by primary human glioblastoma cells. Our trial study is limited by the low number of human cell and tissue donors, as well as lack of deeper mechanistic investigations. With further research in this area warranted, AF16 as an active moiety of the ubiquitous antisecretory factor protein could be an attractive candidate for therapy without known side effects 31 in conditions with a dysregulated immune response, i.e. glioblastoma.

Materials and methods
Cell medium and culture, cytokines and antibodies. Unless otherwise stated, all culturing media used were RPMI 1640 supplemented with 2 mM L-Glutamine, 1 mM sodium pyruvate, 10 mM HEPES, 50 μg/ ml gentamicin (Invitrogen AB, Sweden) and 10% FBS (Biochrom AB, Germany), henceforth referred to as R10 medium. All cells were kept in a humidified incubator at 37ºC in 5% CO 2 atmosphere. Dividing cells were passaged regularly to avoid overgrowth and were kept in culture for less than 20 passages before analysis. All handling of cells occurred under sterile conditions. The following cytokines were used for macrophage differen-

Convection-enhanced delivery (CED) of AF16 in mouse glioblastoma model in immunocompetent and immunodeficient animals.
All animal experiments were conducted in accordance with the ARRIVE guidelines. Animal procedures were approved by the Ethical Committee for Animal Research in Lund-Malmö, permit numbers M151-15 and 14006/2019, and were performed in accordance with the practices of the Swedish Board of Animal and European Union Animal Rights and Ethics Directives. Mice were observed daily, and individuals were euthanized immediately by cervical dislocation when neurological signs of tumor growth occurred (inactivity, tremors, hunched posture, and weight loss). The primary endpoint was overall survival (OS) and animals that reached day 150 after tumor inoculation were considered cured. C57BL/6 mice (8-10 weeksold females, n = 62, purchased from Taconic Bioscience A/S, Denmark) and NOD scid mice (8-10 weeks-old females NOD-Prkdc scid , n = 18, bred at the BMC in-house breeding core facility, Lund University, Sweden) and were each stereotactically injected with cells of the mouse glioma cell line GL261 (5000 cells/3 µl injection volume, cells kindly provided by Dr. G Safrany, "Frédéric Joliot-Curie" NRIRR, Hungary) into the right cerebral hemisphere under general anesthesia, according to an established protocol 57,58 . Briefly, on day 7 post-injection, pumps for intratumoral delivery (3-day mini-osmotic pumps Alzet® model 1003D, fill volume 100 μl, pumping rate 1 μl/h (DURECT™ Corporation, USA)) were filled with 300 µg/72 µl AF16 solution (NOD-Prkdc scid n = 5, C57BL/6 n = 16)  were each stereotactically injected with the GL261 glioma cell line on day 12 and treated with AF16 (n = 5), TMZ (n = 5) and the combination of both (n = 5) as described above with one group left without treatment as a control (n = 4). One animal died immediately after tumor cell inoculation likely due to surgical complications and was therefore excluded from analysis. As soon as the first animal developed neurological symptoms, the whole cohort was euthanized and fresh cerebra were harvested. Immediately after harvest, the brain tissue was frozen and fixed in cooled isopentane (− 55 °C, VWR International AB) and kept thereafter at − 80 °C until further analysis. Due to small tumor volumes, one animal each in TMZ and TMZ + AF16 groups could not be analyzed for all markers. Before staining, brain tissue was sectioned into 6 µm-thick sections on a cryostat (Leica, Germany). Then, sections were fixed with acetone for 10 min at room temperature, rehydrated with PBS, blocked for 20 min with 5% goat serum (Jackson ImmunoResearch, USA) and stained with antibodies before being mounted in DAPI-containing mounting medium (ProLong™ Gold antifade, Invitrogen). Tumor area was measured in the largest diameter in serial sections from different treatment groups. Areas stained with F4/80, CD206, CD11c, MHC II, CD8, COX-2, galectin-3 and pNKCC-1 were measured with the cellSens Dimension software on the fluorescent microscope BX-53 (both Olympus LRI Instrument, AB). Negative control sections were used for validation where the primary antibody was omitted.
Mapping of inflammation-associated secretome of primary human brain tumors.  (ATCC TIB-71, LGC Standards, UK) were cultured in R10 medium. RAW264.7 cells correspond to M0 stage of macrophage maturation (M0-R) and they were incubated for 24 h with 100 ng/ml concentration of LPS and 20 ng/ml concentration of murine IFN-γ to obtain M1 macrophages (M1-R) or 20 ng/ml concentration of murine IL-10 to obtain M2 macrophages (M2-R).
Maturation of M0, M1 and M2 macrophages from primary human monocytes. Primary human macrophages were differentiated by a slight modification of the protocol described by Zarif et al. 59 . Briefly, peripheral blood mononuclear cells (PBMCs) from healthy donors (n = 2) were purchased from the local blood donation bank and separated by density centrifugation of the buffy coat on Ficoll-Paque™ (GE Healthcare Life Sciences, Sweden). Next, highly purified classical monocytes were obtained with the Classical Monocytes MACS Isolation Kit (Miltenyi Biotec, Germany) following the manufacturer's instructions. After purification, classical monocytes were incubated with 20 ng/ml M-CSF for 6 days to obtain M0 macrophages with medium changed on Day 1 and Day 3. To obtain finally differentiated M1 and M2 cells (M1, M2), M0 cells were washed with R10 at Day 6 and cultured for another 24 h in either a M1 cytokine cocktail (20 ng/ml GM-CSF, 20 ng/ml IFN-γ, 20 ng/ml IL-6 and 20 ng/ml LPS) or a M2 cytokine cocktail (20 ng/ml M-CSF, 20 ng/ml IL-4, 20 ng/ml IL-6), respectively.
MesoScale discovery multiplex plate assay. Macrophages (M0) were incubated with 1 µg/ml and 100 µg/ml AF16 or 1 µM betamethasone for 24 h. Murine M0 macrophages (M0-R) were incubated with 2000 µg/ml AF16 for 24 h. Supernatants were collected from culture plates under sterile conditions and centrifuged at 12,000 rpm for 10 min to separate any solid particles. The final supernatants were then removed and frozen at − 20 °C until analysis. The supernatants were analyzed in duplicates with MesoScale V-PLEX® (MesoScale Discovery, USA) human Proinflammatory Panel 1 (IFNγ, IL1β, IL2, IL4, IL6, IL8, IL10 IL1β, IL2, IL4, IL5, IL6, IL10, IL12p70, KC/GRO, TNFα) according to the manufacturer's instructions. When analyzing data, only the range between Lower Level of Detection (LLOD) and Upper Level of Quantification (ULOQ) was considered and when values were below the LLOD or above the ULOQ, they were replaced by the LLOD/ULOQ values. In total, cells from two different human donors and one representative THP-1 experiment are shown. Data are shown in pg/ml on a logarithmic scale. Factors where change was induced in both donors are included (all raw data included in Supplementary Table 1).

Confirmation of proteomics data with mRNA expression in the TCGA database. To validate
previous results in a large brain tumor dataset, low grade glioma and glioblastoma level 3-preprocessed mRNA expression data were obtained from The Cancer Genome Atlas (TCGA) database at the GDS website: https:// portal. gdc. cancer. gov/. The R/Bioconductor package, TCGAbiolinks, was used for database access.
Data analysis. All data was analyzed and visualized using the Prism 8 Software (GraphPad, USA) or the free statistical software R (https://www.R-project.org/ ). Animal survival data were visualized with Kaplan-Meier curves and the treatment groups were compared with log rank tests with median survival. Non-parametric Mann-Whitney U-test was used to analyze the tumor area and immunohistochemical data from mouse brain sections.