In the last decade, the phenomenon of honey bee colony death has attracted the interest of scientists, beekeepers and the general public alike. Increased annual honey bee colony losses have been associated with many factors, including viruses, bacteria, fungi, mites and trypanosomatids (Higes et al., 2006; vanEngelsdorp et al., 2009; Cox-Foster et al., 2007; Dainat et al., 2012; Francis et al., 2013; Ravoet et al., 2013; Cepero et al., 2014). Trypanosomatids are flagellated protozoan parasites and they are the direct descendants of the first eukaryotes that appeared on the earth ~1.5 billion years ago (Knoll, 1992). Several genera of the Trypanosomatidae cause disease in humans, domestic animals, plants (Galanti et al., 1998) and insects (Schmid-Hempel & Tognazzo, 2010). Indeed, the interest in insect trypanosomatids, especially those affecting bees, has increased among researchers due to their possible involvement in colony collapse disorder (CCD). Until now honey bees were thought to be infected by only a single trypanosomatid species, Crithidia mellificae, the trypanosomatid first detected in Australia during the diagnostic examination of Apis mellifera apiaries (Langridge & McGhee, 1967). The original experimental infection with C. mellificae suggested it had no pathological effects on this hymenopteran host, which explains why this organism has been largely ignored for more than 40 years.

However, a relationship between C. mellificae infection and colony losses was recently proposed in the USA (Runckel et al., 2011) and Belgium (Ravoet et al., 2013). Indeed, C. mellificae was 6.15-fold more abundant in CCD-affected colonies than in unaffected colonies in a colony-cohort study, and its detection was strongly associated with the presence of N. ceranae (Runckel et al., 2011). In addition, a high prevalence of C. mellificae (70.5%) was reported in Belgium (Ravoet et al., 2013), and a correlation was detected between C. mellificae incidence in July and winter mortality. This study also confirmed the link between C. mellificae and N. ceranae previously described in the USA (Runckel et al., 2011), suggesting that co-infection by these pathogens has a negative synergistic impact on colony longevity (Ravoet et al., 2013). More recently, a holistic screening of professional apiaries with typical honey bee depopulation symptoms in Spain (Cepero et al., 2014) revealed a higher prevalence of N. ceranae with only a moderate presence of trypanosomatids. Genetic analyses of the 18S rDNA and GAPDH sequences of these trypanosomatids proved that they were phylogenetically distant from those of the reference ATCC30254 C. mellificae strain, providing the first hard evidence of honey bee infection by a different lineage (Cepero et al., 2014). The existence of this taxon, which was later named Lotmaria passim (Schwarz et al., 2015), had been previously suggested in the USA (Cornman et al., 2012). It is also worth noting that all the sequences attributed to C. mellificae until then, including its draft genome assembly (Runckel et al., 2014), were more closely related to this new taxon than to the “true” C. mellificae (Cepero et al., 2014).

Together, these studies inspired renewed interest in this pathogen and its relationship with the highly prevalent microsporidium N. ceranae as one of the possible causes of bee colony collapse in different parts of the world. Given the discrepancies regarding the pathological effect of C. mellificae in honey bees (Langridge & McGhee, 1967; Ravoet et al., 2013), we evaluated the effect of experimentally infecting bee colonies under laboratory conditions with the C. mellificae ATCC30254 reference strain, with or without N. ceranae. In this way we were able to specifically assess the influence of C. mellificae on the bees’ life expectancy and determine whether a synergistic effect of this pathogen with N. ceranae could be detected.

To perform these experiments, the C. mellificae reference strain ATCC30254 was first cultured in ATCC medium 355 as recommended, and it was further sub-cultured in solid medium as described by Popp & Lattorff (2011). Visible, isolated colonies were selected and resuspended in milliQ water (PCR-quality) to obtain trypanosomatid spheroid forms for experimental infection (Schwarz & Evans, 2013). The DNA from those colonies was extracted as described previously in Cepero et al. (2014) and the trypanosomatid species was confirmed by PCR (Meeus et al., 2010) and subsequent sequencing (Cepero et al., 2014). Freshly purified N. ceranae spores for experimental infection were obtained as described elsewhere (Martín-Hernández et al., 2011, 2012; Higes et al., 2013), counting both the N. ceranae spores and C. mellificae spheroid forms in a haemocytometer. DNA was extracted from an aliquot of Nosema spores to confirm the species by PCR (Martín-Hernández et al., 2012). Nosema-free and trypanosomatid-free honey bees were obtained as described previously (Higes et al., 2013). Emerging worker bees from three different colonies were removed carefully, confined to cages in groups of 25 and kept in the incubator for 5 days at 27ºC, the optimum temperature for C. mellificae growth (Langridge & McGhee, 1967). Sixteen replicate cages were prepared, 12 of which were used to analyse survival: three cages infected with fresh N. ceranae spores; three with fresh C. mellificae spheroid forms; three co-infected with N. ceranae and C. mellificae; and three that remained uninfected (used as controls).

Honey bees (Apis mellifera iberiensis) were infected as described elsewhere (Higes et al., 2013), administering 1 µL of a sucrose solution containing 50,000 N. ceranae spores or 50,000 C. mellificae spheroid forms. The co-infected bees received 1 µL of each pathogen solution, while uninfected control bees were fed with 2 µL of the 50% sucrose solution alone. After inducing infection, the bees in each cage were fed ad libitum with 50% sugar syrup + 2% Promotor L (Calier Lab). Infected and uninfected bees were kept at 27 °C in separate incubators (Memmert ® Mod. IPP500, 0.1 °C), and the replicate cages were observed daily. Cumulative honey bee mortality was analysed using a Kaplan-Meier Log-Rank survival analysis and the slope of the survival function was analysed according to the Mantel-Cox test with the aid of IBM Statistics V22.0.

The mortality of the workers infected with N. ceranaeand N. ceranae+C. mellificae was significantly higher than that of those that received the C. mellificae treatment alone and that of the control worker bees. Bee survival at 30 days post-infection (dpi) was 77.8%, 64.5%, 0% and 0% for the uninfected control honey bees, and the honey bees from cages infected with C. mellificae, N. ceranae and a mixture of both parasites, respectively (Fig. 1). The mortality of the workers infected with N. ceranae and N. ceranae+C. mellificae was significantly higher than that of those that received C. mellificae alone and that of the control worker bees (Kaplan-Meier Long-Rank test, p<0.001). Conversely, the survival of the control and the C. mellificae infected bees was significantly higher than that of the other two groups (Kaplan-Meier Long-Rank test, p<0.001). The slopes of the survival functions were statistically different according to the Mantel-Cox test (p<0.05). The evolution of control worker bee survival was similar and not significantly different from that of those infected with C. mellificae alone (p>0.05). Likewise, the N. ceranae and mixed infection groups also displayed a parallel evolution in terms of survival (p>0.05), which differed considerably from the other two groups from 20 dpi onwards. The highest mortality rates of worker honey bees occurred in the cages infected with N. ceranae, irrespective of the presence of C. mellificae.

Since the first report of N. ceranae infecting of A. mellifera bees there has been some controversy about the effects of these microsporidia. However, most of the studies published in recent years have confirmed that N. ceranae has a pathogenic effect on this host, expressed as a shortening of the honey bee (workers) lifespan (Higes et al., 2013). Significantly, the longevity of infected bees varies strongly among these studies, probably due to the many factors that influence the development and the course of the infection. For example, this is the case of the temperature, which must be taken into consideration in laboratory studies since N. ceranae has been proposed to be eurithermal (Higes et al., 2010). In this study, the survival of honey bees following experimental infection with N. ceranae is consistent with that seen previously (reviewed by Higes et al., 2013; Williams et al., 2014; Huang et al., 2015; Milbrath et al., 2015), confirming that bees infected with N. ceranae have a shorter life expectancy than uninfected bees. The incubation temperature used may explain the distinct evolution in the survival of infected honey bees relative to other laboratory experiments performed in our laboratory (Martín-Hernandez et al., 2009).

Previous co-infection experiments with N. ceranae and C. mellificae under laboratory conditions have shown to modify the bees’s immune response in a distinct manner to that produced when bees were exposed only to one pathogen or another. Mixed C. mellificae and N. ceranae infection altered the repertoire of systemic antimicrobial peptides and reduced cellular immunity at the level of the immune response, although the mortality of the infected bees was not recorded (Schwarz & Evans, 2013). An epidemiological link between the presence of C. mellificae and bee colony collapse has been suggested in field conditions, either alone or in association with N. ceranae (Runckel et al., 2011; Ravoet et al., 2013), although no information regarding worker bee mortality was reported. The effect of the infection with C. mellificae spheroid forms on honey bee survival suggests that such cellular stage does not alter the insect’s longevity, and does not significantly modify the mortality of bees co-infected with N. ceranae when compared to that of the bees infected with the microsporidium alone. These results are consistent with those of Langridge & McGhee (1967) suggesting that C. mellificae has no pathological effect on honey bees. However, further research is needed to determine the development of this trypanosomatid in different culture media, investigate its life cycle and the possible variability of its pathogenic capacity. The finding of a new trypanosomatid taxon in collapsed colonies under field conditions raises new questions about the role of C. mellificae in honey bee losses, an issue that should be addressed through appropriate experimental infection and epidemiological analyses.