A literature search was performed to locate original research articles reporting the outcomes of mistletoe treatments for experimental tumors in animals. The search was limited to articles published from 1996 onwards, in English, in peer-reviewed journals included in the database PubMed (ncbi.nlm.nih.gov/pubmed/) and with relevant references cited. The aim of this was to facilitate access to the raw data for the interested readership. These criteria were adopted as a considerable number of studies are published in German and/or as books or meeting proceedings, and the peer-reviewed criterion ensures a priori the quality of articles. Several searches were performed using the following search terms: ‘(Viscum OR mistletoe) AND cancer’; ‘(Viscum OR mistletoe) AND tumor’; and ‘Viscum album AND tumor AND animal’. Only articles reporting in vivo animal models were considered for inclusion.

The parameters of analysis were as follows: i) Provenance and time of publication; ii) rationale underlying the study; iii) methods used [studied animal species, number analyzed, type of mistletoe/mistletoe-derived preparations used, treatment protocol, grafted tumor lineage (epithelial, mesenchymal, melanoma), blinding, random group allocation, controls (placebo, reference drug) and other combined therapies] iv) results (remission, improvement, worsening and survival) and v) mechanisms of action investigated (antiproliferative activity, cytotoxicity, immunomodulatory, tumor leukocyte infiltration, expression of specific phenotypic markers, and other tumor microenvironment aspects likely to correlate with the reported outcomes).

The data were entered in an ad hoc form and are presented in two tables. Table I lists the articles reporting in vivo studies with detailed descriptions of the substance used, dose and regimen, animal model tested, tumor cell lines, concomitant treatments and outcomes. Table II describes methodological aspects used to determine the quality of studies.

The literature search located 911 articles. Duplicates were excluded and the remaining studies were analyzed based on their titles and abstracts until 37 remained that met the inclusion criteria and were considered for the present analysis. These 37 studies were written in English, published in peer-reviewed journals from 1996 onward and included in PubMed (mean, 1.85 studies/year; range, 0–7). As mentioned above, only 1 similar review was located for comparison to the present study (29), which assessed 34 studies from 1938 to 2008 (mean, 0.48 studies/year) of which 21 had been published in peer-reviewed journals. The apparently favorable trend detected in the present review should be considered cautiously, as Kienle et al (29) considered breast and gynecological tumors only. However, according to those same authors ~50 articles/year are published on mistletoe, implying that the proportion of studies using in vivo animal models is rather low. Tables I and II describe the results corresponding to the analyzed parameters.

Viscum album (L.) is a parasitic plant whose composition may be influenced by the host plant or the harvest season (2). As a result of this, the commercially available preparations are typically made from mixtures of fresh leafy shoots harvested in both summer and winter, and fruits harvested in winter, when the concentration of the main active principles (lectins and viscotoxins) is highest (28). Only two studies included in the present review reported the harvest season, and this was reported as winter in both (27,33). A total of 32 studies used European mistletoe and 5 studies used Korean/Chinese mistletoe. The host species on which mistletoe was grown were reported as follows: 13 studies on apple, 4 on oak, 4 on pine, 4 on fir, 2 on poplar and 1 on plum (Prunus) trees. A total of 12 studies did not report the host species.

A total of 25 studies were published between 1996 and 2005 (mean, 2.5 studies/year) and 12 from 2006 to 2016 (mean, 1.2 studies/year). This is a ~50% decrease in the period 2006–2016 compared with the number of studies published in the previous decade. One hypothesis that could explain such a decrease in the number of studies is the trend towards reducing animal protocols in cancer research (34) due to their poor translational features, but this remains to be confirmed.

The majority of studies were conducted in Germany (20 in their entirety, >37 including multicenter studies), followed by Croatia, India and Korea (4 in each), Austria, Belarus, Brazil, Canada, China, France, Norway, Romania, Serbia and the UK (1 in each). Compared with the review by Kienle et al (29), in which studies originated in Germany, Switzerland, Austria, USA, India, Croatia and Serbia only, the distribution reported here represents a favorable trend, as it seems to denote a more global interest in the use of mistletoe.

Mice and rats were used in 32 and 5 studies, respectively. This finding is similar to the one reported by Kienle et al (mice, 28; rats, 6) (29); however, as shown herein, in recent decades mice have become the model used in almost all studies of antineoplastic drugs. This fact raises critical questions as to the translational aspects of such studies and the real possibility of extrapolating experimental findings to clinical practice with humans and other animal species (34,35). Such issues notwithstanding, veterinary clinical studies have demonstrated that mistletoe extracts elicit effective responses in various other animal species (36–39).

The number of animals included in each study ranged from 5 to 20 (median, 8), excluding 4 studies that induced autochthonous urinary bladder carcinoma (22,40–42), which included a large numbers of animals in the control and treated groups (range, 20–100). The possible reason for this may be the notably high procedure-related mortality, as is discussed below. These findings agree with those of Kienle et al (n=5 to 20) (29) and can be rated adequate according to the current standards (43); however, it should be noted that none of the studies presented a justification for the sample size used.

There was an even distribution of tumor types in the studies, including the following: 14 in melanoma cell lines, 15 in epithelial cell lines and 13 in mesenchymal cell lines. Only 7 studies reported negative or no effects relative to models of melanoma (44–46) or epithelial tumors (40–42,45,47). The toxic effects reported were rare and unrelated to the type of tumor. One study reported skin inflammation and necrosis, mainly peritumoral (26) in mice bearing melanoma, whereas 2 studies revealed weight loss in rats and nephrotoxicity in mice bearing epithelial tumors (22,47).

A total of 23 studies used whole plant extracts, 13 used lectins alone, 4 used recombinant lectin I, 3 used triterpenes and 1 used a high and low molecular fraction (> and <30 kDa; total number is >37 as some studies tested more than one substance either alone or in combination). These findings differ from the ones reported by Kienle et al (29): Whole plant extracts=24; lectins=2; and recombinant lectins=2. There is a clear tendency towards increased use of isolated active components, lectins in particular, and the previous interest in other isolated proteins and polysaccharides (4 each) appears to have been replaced by lipophilic components.

A total of 36 studies tested mistletoe administered via various systemic routes (intraperitoneal, oral, intranasal, intravenous, subcutaneous) and 5 at the local tumor site. Three studies pretreated the tumor cells and 4 did not report the route used. These results show a tendency towards systemic administration compared to the results of the study by Kienle et al (29), in which the frequency of systemic and local administration was similar (17 vs. 15 studies, respectively).

A total of 3 studies compared the subcutaneous (SC) and intraperitoneal (IP) routes of administration. In the Braun et al study (48), a reduction in the number of lung and liver colonies of RAW-117 P (liver) and L-1 (lung) lymphosarcoma was obtained only with IP administration of aqueous extract standardized for ML-II/III; however, in a later study they used the SC route only and obtained less positive results (49). Lenartz et al (50) reported a reduction in F98 glioma volume for both the SC and IP routes, but tumor necrosis was observed in the animals treated with IP administration only. This supports the hypothesis that IP administration leads to better pharmacokinetics of the main active compounds. Nevertheless, Burger et al (45) compared the SC and IP routes in F9 testicular carcinoma and did not find a difference in the antitumor effect.

Yoon et al (51) compared a variety of dosages (1–100 µg) of aqueous extract of Korean mistletoe via various routes-systemic (intravenous, subcutaneous) and mucosal (oral and intranasal). Systemic administration induced a dramatic inhibition of colon carcinoma even at a very low dose (2 µg). Furthermore, mucosal routes were associated with metastasis inhibition at a low dose and significantly, no antitumor effect was detected with higher doses via the intranasal route. One study (52) used lectin-rich extract (Iscador Qu) to investigate the effects of pretreatment on tumor cells, systemic and local treatment. Pretreatment mitigated tumor growth, indicating that the antitumor effects mistletoe extract were transferred in vivo to growing tumors; however, this effect was not due to reduced cell proliferation or increased cell death. While intratumor treatment significantly reduced tumor growth, systemic administration had no significant effect.

Taken together, the results are highly variable and divergent, and no definite conclusions can be drawn. With that proviso, there do appear to be indications that local administration may be more effective (see below).

Only 5 of the 37 studies included reported having randomly allocated the animals to the experimental groups (40,53–56), whereas in the Kienle et al (29) review, 6 of 31 studies reported randomization. Randomization is a crucial step in any experimental protocol, particularly in the case of complementary medicine as the risk of false positive or false negative results is very high (57). It may therefore be beneficial if a methodological guideline were formulated, as it has been for homeopathic studies (57). In the present review, as well as in the Kienle et al review (29), none of the included studies reported whether blinding was performed or not. It has been suggests that this is not reported as it is a standard procedure (29). However, unlike in clinical studies, the use of blinded protocols in pharmacological basic research is very rare. For example, in the present review, searching the key works ‘animal model’ and ‘blind protocol’ in PubMed only identified 94 articles. This is another point to be considered in a putative guideline.

All 37 studies included control groups, as follows: 26 studies used a placebo-saline/distilled water (5), vehicle (18) or culture medium (1); 12 compared to untreated (10) and/or healthy mice (3). A total of 5 studies compared mistletoe with standard chemotherapy/adjuvant drugs (polysaccharide K, cytarabine, doxorubicin, cyclophosphamide, 5-fluorouracil) alone or in combination (25,45,53,58,59); and 1 study compared with interleukin-2 alone and in combination (47).

The objectives of the studies exhibit clear temporal trends. Studies conducted between 1996 and 2005 sought to establish the antitumor effect of mistletoe and its possible mechanisms, cytotoxicity and immunomodulation in particular (21,46,54–56,60–68), eventually with comparison to or in combination with standard chemotherapeutic agents. By the end of the 1990 s, the focus of studies shifted towards the study of lectins, in particular ML-I (21,50), also including Korean mistletoe (46,51,53,66). Between 2000 and 2001, novel studies concerning recombinant lectin were beginning to be published (22–24); however, the specific effects of lectins II and III began to be explored later (48,49). The overall profile of studies performed in the past decade is notably different to previous studies, tending to investigate biologically active principles (33,53,59,69) including lipophilic ones, in particular triterpenes (25–27). A shift can be also noticed in investigations of the putative mechanism of action, with a stronger focus on genomics (44,52,69) and oxidative stress (27,28).

A total of 12 of 37 studies (~30%) specifically investigated side effects as variables of interest, including mortality (45,53,55,56), clinical morbidity (26,53,55), body weight (22,23,25,53,55,56,59), food intake (23,33,53,55,56), water intake (33), histology on necropsy (25,26,33,55), behavior (26,33,68), vitality score (33), physiological response (33), hematological parameters (59), overall appearance (68) and agility (23). The majority of studies reported no significant toxic findings, the exception being a study by Strüh et al (26) in which the highest tested dose of solubilized triterpenes administered SC peritumorally resulted in skin necrosis and apathy, which led to 3/13–14 dropouts. Furthermore, animals in all the treated groups exhibited skin inflammation and peritumoral inflammatory infiltrating cells. In the Seifert et al study (59), mistletoe grown on pine trees proved to be more toxic compared to that grown on fir. In the Burger et al study (45) mistletoe was demonstrated to be considerably less toxic than doxorubicin, but showed no differences compared with 5-fluorouracil.

It is worth noting that in the studies using N-methyl-N- nitrosourea (NMU)-induced autochthonous bladder urinary carcinoma the tumor-induction procedure and complications caused high mortality, ranging from 20 to 50% (22,41,42). In a later study, Kunze et al (40) employed N-butyl-N-(4-hydroxybutyl)-nitrosamine (BBN) as a tumor inducer; although the procedure-related mortality was significantly lower (0 to 11%), the number of animals that developed tumors was also low, as is discussed below.

A total of 32/37 studies reported some antitumor effect, 3 of which had mixed results. Facina et al (44) did not detect any changes in tumor volume due to mistletoe treatment, however they did demonstrate evidence of genetic regulation of apoptosis. Burger et al (45) reported a dose-dependent reduction in tumor growth for 3 of the tested cell lines, but no effect on the other 2 lines tested. In Park et al study (46), mistletoe was demonstrated to have no effect on tumor growth; however, it did induce dose-dependent inhibition of metastasis and led to a slight increase in survival rates. Two studies reported no antitumor effect (41,42) and 2 reported an overt negative effect in the treated groups (40,47). Finally, 1 study assessed immunomodulatory activity only, with no investigation of antitumor effects (62); the results from this study indicated that treatment with mistletoe restored the immune response, however prolonged application also had a dose-dependent immunosuppressing effect on normal cells. The relationship between the concentration of active molecules and the balance between cytotoxicity and immunomodulatory effects requires further investigation; the combination of both factors in the same protocol may be a useful tool to treat aggressive tumors with mistletoe components.

A total of 21 studies assessed tumor growth inhibition. Of these, 17 found a reduction in tumor size, 7 of them in a dose-dependent manner and with eventual inversion of the effect at a higher dose (33), and one of these studies (45) indicated a reduction in tumor size in only 3 out of the 5 tested cell lines. In 1 study, remission was reported when mistletoe was administered inside the tumor, whereas there was no such effect when it was administered via the systemic route (52). Two studies reported an unexpected increase in tumor size in the treated groups (40,47). Future studies investigating the underlying molecular mechanisms may explain these paradoxical results.

Survival was analyzed in a total of 12 studies, and was higher in the treated groups in all cases, most of them in all studied doses but dose-dependence was reported in 2 studies (24,56).

Inhibition of primary or metastatic tumor formation was assessed by 1 (56) and 12 studies, respectively. Tumorigenesis was inhibited in all studies with the exception of Timoshenko et al (47). Based on these findings, the authors of the present study suggest that the action of mistletoe active principles in the tumor microenvironment should be investigated further, as the onset of the metastasis-formation process depends on it. Additionally, the variability of results among the different studies may also be attributed to the fact that most protocols were based on transplantable tumors, in which tumorigenesis cannot be evaluated, since tumor cells are inoculated directly in the animals, being immunogenic stimuli. The cases that are most illustrative of this are the 4 studies that assessed the effect of mistletoe lectin on autochthonous urinary bladder carcinoma. Kunze et al (41,42) and Elsässer-Beile et al (22) induced tumors using NMU, and Kunze et al (40) induced tumors with BBN. In all 3 studies by Kunze et al, a lectin isolated from unspecified mistletoe was used. Of these, the 2 studies that induced tumors with NMU had a high procedure-related mortality, and no significant difference was found between the control and treated groups. In contrast, the study that used BBN had a very low procedure-mortality rate, but also a low number of induced tumors in the control and treated groups (n=6 and 4, 10.2 and 6.7%, respectively, in the 6-month experiment; n=16 and 13, 25.8 and 19.7%, respectively, in the 15-month experiment). In this case, at 15 months, the tumor size was twice larger, on average, in the lectin-treated group. Elsässer-Beile et al (22) used recombinant lectin starting 8 or 14 weeks following tumor induction and reported significantly lower rates of atypical hyperplasia and neoplastic transformation in all treated groups compared with the control. The main difference between these two groups of experiments is that Kunze et al used the subcutaneous route for long periods of time (6 and 15 months), whereas Elsässer-Beile et al (22) applied shorter 6-week treatments via the local route of administration (intravesical instillation). It could therefore be that the divergence in these results is due to the methodological differences between the studies. Mengs et al (54) used the intravesical route to administer a lectin-rich aqueous mistletoe extract to mice implanted with urinary bladder carcinoma MB49 cells in a short 4-week course of treatment; the results of this study indicated a dose-dependent statistically significant tumor growth inhibition in the treated groups.

Several putative mechanisms for the anticancer action of mistletoe were tested, including tumor cell viability [reduced, (27,53,63)] and number [reduced, (37)]; tumor cell turnover [increased, (70)]; antiproliferative effect [4 studies, no effect in 1 (26)]; proapoptotic action (increased in 8 studies); increased tumor necrosis (6 studies); increased macrophage cytotoxicity (66); reduction of mitotic figures in tumor cells (65,70); cycle cell regulation [deregulated in one study (53), and no effect in another (46)]. Furthermore, 1 study reported degradation of specific miRNAs, this being an indirect effect (affection of precursors) (71) and 2 studies reported a reduction in oxidative stress (27,58).

Reduced tumor invasiveness was reported by 1 study (65), whereas in another (70), 11% of the treated animals exhibited no histological evidence of viable tumors on day 11 after the onset of treatment. Reduced angiogenesis was identified in 2 studies (26,46); however, in 1 study, the vascular count increased when high doses were used (33). The possible underlying antineoplastic mechanisms require further investigation using specific molecular methods in animal models, as developing an effective mistletoe anticancer agent may on the addition of different mechanisms, possibly associated with a mixture of active molecules in different concentrations and their interactions with the tumor microenvironment. Few studies have been performed from this perspective; a total of 5 studies analyzed the role of local cell infiltrates in tumor development, with an increase in absolute numbers of infiltrating leukocytes being found in 3 studies (23,63,70) and no significant effect found in 2 studies (40,41).

For the immunomodulatory action of mistletoe, induction of the Th-1 response was assessed and detected in 1 study (44) and cytokines were assessed in 2, of which 1 reported no effect (22) and the other reported upregulation of IL-12 only (67). An increase in natural killer cell activity was detected in 2 studies (51,66). Macrophage stimulation was reported in 1 study (55) and 2 studies found an increase in white cell count in the peripheral blood, but not of granulocytes (48,49). Upregulation of the number of thymocytes (49), activation of spleen cells (64), lymphocyte stimulation (CD4⁺ and CD8⁺, but only anti CD8⁺ activity was enhanced) (55) and finally stimulation of cytotoxic population (63) were also reported.

In conclusion, the majority of studies investigating the effects of mistletoe in animal cancer models analyzed in the present review reported positive outcomes in terms of the inhibition of tumor formation and growth. However, there is considerable methodological heterogeneity among studies, which precludes performing conclusive comparisons and explains the variation in results. A great deal of research is still required with standardization of mistletoe preparation, dose, concentration, regime of administration, length of treatment, targeted cancer type, and other aspects. Similarly, while it is universally assumed that mistletoe has cytotoxic and immunomodulatory effects, there is a wide variation among studies in terms of the mechanisms assessed. The present review found little or no continuity within and among research groups, with little progress being made in the past 20 years of studies with experimental animal models. Based on these results, the present authors suggest that the community of interested researchers should develop a guideline for reporting in vivo mistletoe cancer treatment experiments.