Beta-glucan contamination of pharmaceutical products: How much should we accept?

The immunomodulatory properties of beta-glucans have been the subject of considerable scientific study in the Western world for several decades. Non-clinical studies indicate that beta-glucans activate innate immunity as well as adaptive immunity, modulating both humoral and cell-mediated immune responses. Beta-glucans act on several immune receptors including Dectin-1 (considered the main receptor for beta-glucans [22]), complement receptor-3 (CR3) and toll-like receptors 2 and 6 (TLR-2/6) (reviewed in [23‐26]). Beta-glucans have been found to increase the anti-microbial activity of mononuclear cells and neutrophils, to enhance the functional activity of macrophages, to stimulate the proliferation of monocytes and macrophages and to stimulate the production of proinflammatory molecules such as complement components, interleukin (IL)-1α/β, TNF-α, IL-2, interferon (IFN)-γ, eicosanoids, IL-4 and IL-10. Consistent with these findings, beta-glucans have also been shown to protect against or ameliorate the effects of bacterial and other infections in animal models (reviewed in [23‐26]). The broad-ranging immunological effects of beta-glucans and the epidemiological associations between environmental beta-glucans exposure and respiratory allergies have also led to interest in their use in the prevention or treatment of allergic diseases. The ability of beta-glucans to rebalance a dysregulated Th1/Th2 equilibrium is thought to be beneficial in these conditions (reviewed in [27]).

It is thought that the immunological effects of orally administered beta-glucans are essentially the same as those of parenterally administered beta-glucans. Orally administered beta-glucans are absorbed through the gastrointestinal tract and taken up by tissue-resident macrophages. Here, they are fragmented, transported to the bone marrow and reticuloendothelial system and eventually released and taken up by other immune cells, leading to various immunological effects (reviewed in [23]). On the basis of these features, particulate beta-glucans have been evaluated as potential vaccines for invasive fungal diseases (reviewed in [28]) and beta-glucans particles have been proposed as a delivery system for oral vaccines, acting as both carrier and adjuvant [29].

It has been suggested that beta-glucans, administered alone, in combination with monoclonal antibodies, or as adjuvants alongside vaccines or other types of immunotherapy, could help reduce cancer growth. Beta-glucans may act either within tumour microenvironments (TMEs) or systemically, by activating or recruiting immune effector cells into tumours or by augmenting adaptive immune responses triggered by concurrent immunotherapy.

Beta-glucan administration has been reported to influence polarisation of tumour-associated macrophages away from an immunosuppressive and towards an activatory and tumouricidal phenotype (M1). This polarisation is thought to be mediated by engaging the cell surface C-type lectin receptor Dectin-1 and triggering the spleen tyrosine kinase (Syk)-Card9-Erk pathway via the ITAM domain [30]. This may result in production of inflammatory cytokines by macrophages and antigen-presenting cells, and tumour growth restriction through enhanced phagocytosis, antigen presentation and induction of Th1 and CTL responses. In one study, beta-glucan administration to mammary tumour-bearing mice triggered IL-12 production by macrophages which led to a switch from IL-4- to IFN-γ-producing Th1 cells [31]. Another study reported reversal of MDSC immunosuppressive phenotypes in response to particulate beta-glucan treatment. The beta-glucan promoted mature CD11c⁺ F4/80⁺ Ly6C^low MDSCs via a Dectin-1/NF-κB-dependent pathway, and this was associated with enhanced infiltration of dendritic cells, macrophages and CTLs in tumours [32].

Well-known immunosuppressive forces in TMEs can profoundly impair the maturation and antigen-presenting functions of DCs. TME-associated inflammation promotes IL-10- and TGFβ- secreting DC phenotypes, which supports accumulation of regulatory T cells. These effects have been shown to be reversed upon administration of beta-glucans and are accompanied by a reduction in regulatory T cells and enhanced tumour infiltration by mature DCs, macrophages and granulocytes in mouse models of cancer [33, 34]. Early data suggest potential enhancement of PD-L1 expression on mouse peritoneal macrophages by microparticulate beta-glucan, which could have a negative impact on T cell survival and activation [35]. However, potential effects in relation to cancer-associated immune surveillance, immunomodulation or treatment with anti-PD-1 checkpoint inhibitors are yet to be explored.

The potential administration of beta-glucans in combination with antibodies, vaccines and other immunotherapeutic or chemotherapeutic agents has been investigated in pre-clinical studies. Monoclonal antibodies, especially those which function via complement activation, have shown improved efficacy in animal models of cancer when co-administered with beta-glucans. These effects, most likely due to the ability of beta-glucans to bind complement receptor-3 (CR3) and to promote antibody-mediated complement activation against cancer cells, were also associated with recruitment of tumouricidal granulocytes to tumour sites [36, 37]. Consistent with these findings, encouraging efficacy was reported in a non-randomised phase II clinical study, in which the anti-EGFR complement-activating antibody cetuximab was administered in combination with a soluble β-1,3/1,6-glucan to patients with KRAS-mutant/EGFR signalling-resistant colorectal carcinomas [38]. Results of phase I/II trials of beta-glucans in combination with other immunotherapeutic agents also indicate good tolerability with promising signs of anti-tumour activity in patients with chronic lymphocytic leukaemia and neuroblastoma [39, 40].

Generally, non-clinical and clinical studies of the immunological effects of beta-glucans have not been designed to establish a dose–response relationship between beta-glucans and the effect of interest at low doses/concentrations, or to define the lowest dose/concentration at which immunological effects occur. However, based on a range of nonclinical studies, it seems unlikely that clinically significant immunological effects would occur with serum levels of 100 pg/mL or lower. One in vitro study showed that exposure to 100 pg/mL of beta-glucans for 48 h stimulated the production of IL-17 by dendritic cells in mixed lymphocyte reaction assays, but stimulation was most marked at 1 ng/mL (10 times higher than 100 pg/mL) and there was no effect on IFN-γ or IL-5 production at concentrations up to 100 µg/mL [41]. In a separate study, some induction of IL-1β and IL-12 by macrophages was detected at beta-glucans concentrations as low as 10 ng/mL (100 times higher than 100 pg/mL) [42]. However, production of these cytokines was much greater at higher concentrations, and for macrophages to produce TNF-α or IFN-γ, a beta-glucans concentration of at least 0.1 µg/mL (100 ng/mL) was required. In another study, an enhanced oxidative burst response and microbial killing by peripheral blood mononuclear cells were detected at beta-glucans levels of 100 ng/mL (0.1 μg/mL) or greater, but stimulation of the NF-κB-like DNA-binding protein by beta-glucans only occurred at concentrations of 370 ng/mL (0.37 μg/mL) and above [43].

A range of immunological effects have been demonstrated with beta-glucans concentrations between 0.2 and 10 µg/mL, including dendritic cell activation and maturation [44‐46], neutrophil chemotaxis [47], histamine release from basophils [48] and tumour cell killing by polymorphonuclear leucocytes [49]. For example, in one study, as little as 3 µg/mL of beta-glucans induced the production of intracellular and membrane-associated IL-1, but induction of secreted IL-1 required 25 µg/mL, with 50 µg/mL yielding maximal responses. Production of prostaglandin-E2 (PGE2) by glucan-activated human monocytes occurred at concentrations of as low as 12.5 µg/mL but plateaued at 25 µg/mL [45]. Examples of in vitro studies in which the immunological effects of beta-glucans have been evaluated at low concentrations are summarised in Table 1. Of note, many of the studies which looked for a dose–response effect also looked at duration of exposure. To see immunological effects at lower concentrations, relatively prolonged exposure to beta-glucans (12–72 h) may have been required.

Of note, in vitro studies in which the direct anti-tumour effects of beta-glucans have been evaluated suggest that higher concentrations of beta-glucans are required than those needed for immune modulation. For example, in an in vitro cytotoxicity analysis, beta-glucans concentrations up to 100 μg/mL did not directly affect the growth of colon 26-M3.1 cells [42]. In other studies, the proliferation of B16-F10 melanoma cells was reduced by 51 % after 48 h exposure to 750 μg/mL of beta-glucans [52]; proliferation of the gastric cancer cell line SGC-7901 was reduced in a dose-dependent manner at concentrations between 125 and 1000 μg/mL, with around 50 % inhibition with 400 μg/mL [53]; and proliferation of MCF-7 breast cancer cells was reduced in a dose-dependent manner at concentrations between 12.5 and 400 μg/mL, with 50 % inhibition at 400 μg/mL [54].

Most publications relating to the use of lentinan are in Japanese (usually with an English abstract), and these indicate that lentinan can be given via the i.v. (most commonly), intraperitoneal, intrapleural [70‐72] and intraarterial routes [73]. The commonest dose/schedule appears to be 2 mg i.v. weekly, but i.v. doses as high as 10 mg have been given (see [6] for example). In addition, there are several ongoing Japanese phase II trials in patients with advanced gastric cancer (Trial identifiers in the World Health Organisation meta-register of clinical trials: JPRN-UMIN000010724, JPRN-UMIN000008590, JPRN-UMIN000007726 and JPRN-UMIN000001913). In these trials, where specified, lentinan is given i.v. at a dose of 2 mg weekly in combination with chemotherapy (TS-1, with or without cisplatin). A randomised phase III study of TS-1 alone versus TS-1 plus lentinan in advanced or recurrent gastric cancer appears to be complete but not published in the English literature. In this study, lentinan was given i.v. weekly but the dose is not specified.

The pharmacokinetics of lentinan given i.v. over 2 h at doses of 1, 2 or 4 mg, have been evaluated in healthy volunteers and gastric cancer patients [74]. After 1 mg of lentinan, plasma concentrations reached a maximum at the end of infusion (51–73 ng/mL) and decreased gradually thereafter. In healthy volunteers, lentinan concentrations 24 h after administration were (mean ± SD) 71 ± 21 ng/mL after a 4 mg dose, and 53 ± 11 ng/mL after a 2 mg dose. In three patients given 1 mg, levels were around 10–20 ng/mL after 24 h. Thereafter, levels declined slowly over approximately 7 days until they reached the detection limit. The pharmacokinetics of i.v. lentinan in humans were similar to those in the rat. Lentinan was found to be stable in human plasma, and the decline in plasma levels was thought to be due to uptake or degradation in cells, probably Kupffer cells in the liver.

In phase I/II placebo-controlled studies, HIV-positive patients were given 2, 5 or 10 mg of lentinan (or placebo) i.v. over 10 min once a week for 8 weeks, or 1 or 5 mg of lentinan i.v. over 30 min twice a week for 12 weeks [6]. Side effects were mainly mild, especially when the infusion was carried out over a 30-min period. When the infusion was over a 30-min period, there were no severe side effects and only four dropouts due to toxicity or patient preference. However, when administration was over a 10-min period, severe side effects occurred (one case each of anaphylactoid reaction, back pain, leg pain, depression, rigor, fever, chills, granulocytopenia and elevated liver enzymes) and four patients discontinued therapy as a result. Other investigators have also linked the side effects of lentinan (such as “oppression in the anterior chest” and dryness of the throat) to rapid infusions [7].

An individual patient-based meta-analysis of randomised trials of lentinan in gastric cancer included 650 patients from five trials of chemotherapy (combinations of S-1, mitomycin C and cisplatin), with or without lentinan (2 mg i.v. weekly) [75]. The concurrent use of lentinan with chemotherapy was found to significantly prolong overall survival compared with chemotherapy alone (hazard ratio 0.80; 95 % confidence intervals [CI] 0.68–0.95; stratified log rank p = 0.011). No major differences in haematological or non-haematological adverse events were reported for the two treatment regimens, although leucopenia was reported in 5.6 % of patients receiving chemotherapy plus lentinan versus 1.6 % of patients receiving chemotherapy alone. A meta-analysis has also been published of randomised chemotherapy trials, with or without lentinan, in non-small cell lung cancer (NSCLC) [76]. All the trials were conducted in China and published in Chinese. The addition of lentinan to chemotherapy resulted in higher response rates (relative risk [RR] = 1.31, 95 % CI 1.14–1.52) with less frequent Grade 3–4 gastrointestinal toxicity (RR = 0.54, 95 % CI 0.43–0.68) and Grade 3–4 granulocytopenia (RR = 0.65, 95 % CI 0.51–0.70) than with chemotherapy alone.

A randomised phase II trial published since the gastric cancer meta-analysis included 78 patients with advanced gastric cancer who received S1-based chemotherapy as first line treatment, with or without lentinan (2 mg i.v. every 2 or 3 weeks) [77]. Median overall survival was significantly longer in the lentinan group than in the chemotherapy alone group (689 days [95 % CI 431–2339 days] versus 565 days [95 % CI 323–662 days]; p = 0.0406). The most frequently observed severe (Grade 3–4) toxicity was neutropenia (chemo-immunotherapy 50 %, chemotherapy alone 45 %) but Grade 3 febrile neutropenia was only observed in one patient in each group. There were essentially no differences in the incidence or severity of adverse effects between patients who did and those who did not receive lentinan. Administration of lentinan was observed to suppress the granulocyte:lymphocyte ratio.

As described earlier, a 500 ng dose of beta-glucans could in theory be co-administered with a single dose of our product. In comparison, the maximum dose of i.v. lentinan used in routine practice or clinical trials is 10 mg, representing a safety margin of 1:20,000. From the study published by Yajima et al. [74], a 1 mg dose of lentinan given over 2 h resulted in a maximum plasma concentration of 73 ng/mL in patients with gastric cancer, indicating initial rapid clearance (since without any clearance the plasma concentration should have been around 200 ng/mL, based on a 5 L estimated blood volume). Extrapolating from these data suggests that a 500 ng dose of beta-glucans given over 2 h would only increase plasma levels to approximately 37 pg/mL, well within normal physiological levels of 17 pg/mL ± 34 pg/mL.

Imprime PGG^® has been evaluated in a range of clinical trials [78]. These indicate that Imprime PGG^® can be given i.v. at doses of 4 mg/kg weekly, and as single doses up to 6 mg/kg (for a 70 kg adult, this would equate to a 420 mg dose). Typical systemic clearance of beta-glucan in healthy subjects and cancer patients treated with Imprime PGG^® and cetuximab (with or without chemotherapy) were 0.491, 0.565, and 0.690 L/h, respectively [79]. The effective half-life of beta-glucan ranged from 19.5 to 27.3 h. In one study, mean area under the curve over 24 h of beta-glucan in Cycle 1 was similar to that in Cycle 3 (362 and 383 µg.hr/mL, respectively) [80]. Likewise, mean peak concentrations of beta-glucan in Cycles 1 and 3 were similar (44.3 and 47.8 µg/mL, respectively). Minimal accumulation of beta-glucan was observed with trough concentrations on Day 1 Cycle 2 and Day 1 Cycle 3 of 0.0680 and 0.117 μg/mL, respectively.

Detailed safety data are available from a randomised phase II study (n = 89) of paclitaxel, carboplatin and bevacizumab with or without Imprime PGG^® (4 mg/kg weekly) in patients with previously untreated NSCLC [81]. Events potentially associated with hypersensitivity reactions occurred more frequently with Imprime PGG^®, including one Grade 3 anaphylactic reaction. As a result, the authors recommend premedication with low dose corticosteroids and anti-histamines prior to Imprime PGG^® administration. Of interest, overall and Grade 3–4 infections were less frequently reported in the Imprime PGG^® group compared to control (47.5 % vs 63.3 % overall; 5.1 % vs 10.0 % Grade 3–4). Immune-mediated adverse events (e.g. immune-mediated hepatitis or endocrinopathies) that are reported with T cell modulators (such as ipilimumab) were not observed with Imprime PGG^®. Higher response rates, longer duration of response and improved survival were also reported for patients in the Imprime PGG^® group, although the study was not powered to demonstrate statistically significant improvements in these parameters.

Similar results were reported for a randomised phase II (n = 90) trial of Imprime PGG^® in combination with paclitaxel, carboplatin and cetuximab in NSCLC [82]. Overall (all grades) and Grade 3–4 adverse events potentially associated with hypersensitivity or infusion reactions were not increased in the Imprime PGG^® treatment arm in this study, but infections were less frequently reported (an absolute difference of 5 % between the two treatment groups). Immune-mediated adverse events (e.g. endocrinopathies) were not observed with Imprime PGG^®. Higher response rates were reported in patients who received Imprime PGG^® in addition to standard therapy, especially in patients with higher levels of pre-existing anti-beta-glucans antibody levels, although time-to-event outcomes were balanced between the two groups.

Ongoing trials of Imprime PGG^® include a large (n = 795) randomised phase III trial in combination with cetuximab in patients with advanced KRAS wild-type colorectal cancer (NCT01309126). This trial started in April 2011 and is expected to complete in 2016. Other ongoing trials include a phase I/II trial of Imprime PGG^® in combination with rituximab in patients with indolent non-Hodgkin lymphoma (NCT02086175). A phase I/II trial of Imprime PGG^® in combination with an antibody and gemcitabine in pancreatic cancer was terminated early due to a drug recall (drug not specified) (NCT02132403).

The usual dose of Imprime PGG^® used in clinical trials is 4 mg/kg (280 mg for a 70 kg adult), 560,000 times higher than the maximum dose of beta-glucans (500 ng) that could be theoretically be administered with our product.