3.1 Isolation of primary intestinal cells of Eisenia fetida
In order to obtain a maximum number of vital cells from the intestinal tissue of E. fetida, a gentle yet efficient release procedure was established (Fig. 1). Based on previous experience with the release of primary cells from tissue, an enzyme-supported tissue disaggregation step was implemented in the protocol using collagenase II. Collagenase can effectively break down peptide bonds present in collagen, which is the main structural component in the extracellular matrix (ECM) (Rahman 2019; Ricard-Blum 2011).
Composition and in particular osmolality of the buffer receiving the released cells are of major importance for final cell yield and vitality. Here, two buffers were compared for the initial preparation of the intestinal tract as well as the enzymatic treatment. Both buffers, namely M-HBSS and LBSS, had previously been used to collect coelomocytes from E. fetida and other earthworm species (Diogène et al. 1997; Engelmann et al. 2004; Eyambe et al. 1991; Irizar et al. 2014; Stein and Cooper 1981). With 210 mOsmol/kg M-HBSS has a slightly higher osmolality than LBSS (171 mOsmol/kg), but both buffers should be in an acceptable range. Further, M-HBSS contains 5.5 mM glucose as potential C-source and 10 mM HEPES as additional buffering agent, whereas LBSS contains no C-source and only very low concentrations (0.4 mM) of phosphate as possible buffering agent (Table S1).
No statistically relevant difference in regard to cell viability could be found between the two buffer systems (Kruskal-Wallis-Test: χ2 = 4.42, p-value = 0.11) (Fig. 2A). The isolated cells had a cell size between 5.9 and 20.7 µm (median 11.9 µm, mean 12.6 ± 3.4 µm). The viability was > 70 % in all cases (n ≥ 4), which we considered sufficient for subsequent cultivations. However, there was a trend towards higher viabilities for cells isolated in M-HBSS (median 87.2% for M-HBSS (n = 11) vs. 79.8% for LBSS (n = 4)). The presence of glucose as a possible carbon source for the isolated cells in case of M-HBSS was most likely responsible for this small but noticeable effect. It is also possible that the higher buffering capacity of M-HBSS helped to stabilize the cells. The living cell yields (Fig. 2B) varied strongly between individual experiments, ranging from 11.8 x 106 to 93.7 x 106 living cells per gram of tissue. The buffer, either M-HBSS or LBSS, again had no significant influence on the average cell yield (Kruskal-Wallis-Test: χ2 = 0.68, p-value = 0.71), but the deviations were much more pronounced in case of LBSS. The more easily exhausted buffer capacity of the LBSS buffer may well have contributed to the low reproducibility of the protocol. Therefore, M-HBSS was used in further experiments.
Next, the impact of adding a mucolytic agent during the enzymatic digestion was tested, namely Guaiacol Glyceryl Ether (GGE). GGE is often used for coelomocyte isolation (Diogène et al. 1997; Engelmann et al. 2004; Eyambe et al. 1991) and dissolves in particular mucus-like tissue. When M-HBSS in the absence and presence of GGE (50.4 mM as proposed previously (Diogène et al. 1997)) was used in independent experiments (n ≥ 4), the results showed no significant influence of GGE on the isolation process, as neither viability (Dunn post-hoc comparison: p-value = 0.48) nor living cell yield (Dunn post-hoc comparison: p-value = 0.35) was affected (Fig. 2). Therefore, GGE was not used in subsequent experiments.
3.2 Cultivation of isolated primary intestinal cells from E. fetida
For the cultivation of the isolated primary cells, a suitable (basal) culture medium in terms of nutrients, osmolality, buffer system and capacity, as well as pH had to be identified. Moreover, in preliminary cultivation experiments with isolated primary cells, microbial contaminations of the cell culture were observed. This is not surprising, given that earthworms are known for their complex intestinal microbiome (Pass et al. 2015). As earthworms are soil feeders a high number of bacteria and fungi must be expected in the intestinal tract. Washing steps during the isolation were not sufficient to deplete the microbial burden, since presumably microbes were embedded and protected in a mucus layer. Better results were obtained after the addition of a complex mixture of antibiotics (penicillin, streptomycin, tetracycline, gentamycin) together with the antimycotic amphotericin B to the medium (hereafter referred to as PSTGA). PSTGA was supplemented during cell isolation and cultivation and securely prevented microbial contamination for at least 144 h.
Our attempts to identify a suitable basal medium were based on published medium formulations (Battaglia and Davoli 1997; Bilej et al. 1990; Engelmann et al. 2004; Gong et al. 2014; Irizar et al. 2014; Roch et al. 1975; Toupin et al. 1977), mostly L-15 and SDM (standard compositions are shown in Table S5) and are summarized in Table 1. Cultivation experiments with L-15 medium were performed at room temperature without additional CO2, whereas in case of the SDM based formulations 37 °C and an atmosphere containing 5% CO2 were used, as described in literature (Gong et al. 2014; Irizar et al. 2014).
Table 1 Investigated mediac used in cultivation experiments
|
Basal
medium (% v/v)
|
FCSb
(% v/v)
|
Cell culture grade water (% v/v)
|
Galactose (g/L)
|
Lactalbumin hydrolysate (g/L)
|
Osmolalityd (mOsmol/kg)
|
L-15-10
|
90
|
10
|
-
|
-
|
-
|
339 ± 4
|
L-15-10-GLA
|
90
|
10
|
-
|
1.3
|
4.5
|
345 ± 3
|
SDMa-13
|
87
|
13
|
-
|
-
|
-
|
326 ± 9
|
SDM-13-GLA
|
87
|
13
|
-
|
|
4.5
|
344 ± 2
|
Hansen S-301
|
22
|
13
|
65
|
1.3
|
4.5
|
157 ± 5
|
aSDM: Schneider´s Drosophila Medium.
bFCS: Fetal calf serum.
cAll media were supplemented with PSTGA. When indicated, media were in addition supplemented with galactose and lactalbumin hydrolysate (GLA) using a concentrated stock solution (200 g/L). L-15 media were supplemented with 4 mM L-glutamine.
dOsmolality is shown as mean ± SD (n=5).
Cell number and viability were analyzed over 144 h of cultivation (data not shown). While no increase in cell number was observed in any of the media formulations over the cultivation time, cell viability remained > 90 %, showing that the cells could be kept alive for at least 144 h. As no medium was clearly superior to the others, cultivation in L-15 medium at room temperature was chosen as the simplest approach for further experiments. L-15 medium allows cell cultivation without the sodium carbonate / carbon dioxide buffering system as it utilizes free base amino acids (L-arginine, L-histidine, L-cysteine) as buffering agents (Leibovitz 1963). Cultivation at room temperature accommodates E. fetida primary cells, since the natural habitat of E. fetida is the soil, and laboratory cultures are commonly kept at temperatures between 15 and 20 °C (Miles 1963; Presley et al. 1996; Tripathi and Bhardwaj 2004). For coelomocytes from the congeneric E. hortensis, an increase in temperature above 25 °C is known to significantly increase cell death rates (Fuller-Espie et al. 2015). Strikingly, this was not observed here for the intestinal cells, which survived well when cultivated in SDM based media at 37 °C.
The addition of HEPES did not influence cell number or viability (data not shown), but had a beneficial effect on the reproducibility of the experiments, presumably due to the higher buffering capacity. Therefore, the medium was supplemented with HEPES in the subsequent experiments.
Although cell numbers varied slightly during cultivation, significant proliferation was never observed, including for the SDM based preparations, which had previously been proposed to support proliferation of intestinal cells from another earthworm species (Gong et al. 2014). Proliferation is not necessary for ecotoxicology experiments, but may be useful in other types of research. The observed lack of proliferation could be explained by a lack of specific growth factors for E. fetida cells in the basal culture medium. For primary cells, specific growth factors are typically supplied via blood serum or cell homogenate. Here, an investigative three-step adaption into that direction was performed comprising 1) a supplement screening involving worm filtrate (WF), but also fetal calf serum (FCS) as a standard cell culture media additive, 2) an adaption of the osmolality, and finally 3) a verification of the appropriate conditions by measuring the metabolic activity.
First, we recorded cell number and viability at different concentrations and combinations of FCS and WF in cultivation experiments (Fig. 3A). L-15 medium without additive yielded a slight reduction of cellular viability over time, albeit never dropping below 80 %. Both FCS or WF appeared to slightly improve cell viability. Since there was no clear difference between WF and FCS as additive, we chose a mixture of 10% FCS and 10% WF as starting composition for the osmolality adaption, evaluating different dilutions of L-15 medium for cultivation, corresponding to final osmolalities between 308 and 381 mOsmol/kg (Table S6). In the subsequent cultivation experiments, there was no clear influence of the medium dilutions on cell number or viability (data not shown). For the common earthworm Aporrectodea caliginosa, a broad range of body fluid osmolality from 175 to 684 mOsmol/kg was measured for different dehydration states suggesting a high tolerance of that species against a broad range of osmolalities (Bayley et al. 2010). Therefore, L-15-60% (10% v/v FCS and WF, 60% v/v L-15, 20% v/v cell culture grade water) was chosen for further experiments, as its osmolality of approximately 310 mOsmol/kg was considered to be closest to characteristic osmolalities of terrestrial animals (Stankiewicz and Plytycz 1998).
Next, cellular vitality, i.e. metabolic activity, is an equally important indicator for cell cultivation and toxicity testing. Using L-15-60% as basis the impact of the FCS/WF supplement on cellular vitality was investigated using the MTT assay as analytical tool (Fig. 3B). Cells cultivated in 10% FCS + 10% WF or 20% FCS reached the highest mitochondrial metabolic activity, i.e. 150% compared to that in L-15-60% without FCS/WF supplementation. L-15-60%, supplemented with 10% FCS, 10% WF, as well as 4 mM L-glutamine, 25 mM HEPES, and PSTGA was therefore chosen as standard culture medium for E. fetida cells.
Cell seeding density is a critical factor for primary cell cultivation, as a sufficient number of cells is needed for cell-cell interactions as well as for the production of autocrine growth factors. However, in the case of primary cells, proliferation often stops once confluency is reached during cultivation. Therefore, the effect of seeding cell densities was analyzed microscopically between 0.053 x 106 cells/cm2 and 0.421 x 106 cells/cm2 in 24 well plates with 1 mL culture medium (Fig. 4).
In these experiments, cell seeding densities between 0.158 x 106 cells/cm2 and 0.210 x 106 cells/cm2 were optimal. Lower seeding densities led to large, uncovered areas, whereas higher cell densities resulted in an increased number of floating cell aggregates after 48 h. This seeding density is in accordance with previously published results using P. aspergillum primary epithelial cells (Gong et al. 2014). Inter alia, the occurrence of floating cellular aggregates indicated that the cells failed to properly adhere to the cell culture plate. Adherence is an important factor for growth of intestinal cells considering that these cells are derived from epithelial tissue. To promote adherence of the isolated cells to the culture plate, different coating strategies were evaluated based on standard cell culture coating materials, in particular poly-L-lysine, gelatin (porcine), collagen type I (human) and collagen type II (bovine) (Davidenko et al. 2016; Harnett et al. 2007; Liberio et al. 2014). None of the treatments improved cell adhesion. We currently assume that the lack of suitable adhesion factors is a major contribution to the lack of proliferation observed for the intestinal cells.
3.3 Ecotoxicity testing using primary intestinal cells of E. fetida
Even in the absence of proliferation, highly viable, metabolically active primary cells present an excellent basis for a study of acute toxic effects of common ecotoxins on the cellular level. For a demonstration, we examined the influence of known environmental pollutants, namely Ag nanoparticles and metal ions (Cu2+, Cd2+) on the metabolic activity of the isolated cells (Fig. 5).
Exposure to Ag nanoparticles led to a slight, but significant decrease in metabolic activity (LMTreatment F9 = 4,897; p < 0.001) in a concentration-dependent manner compared to the negative control (cells in culture medium) (Tukey post-hoc comparison: p-value48 h = 0.130, p-value96 h = 0.084) (Fig. 5A). Cells exposed to the citrate buffer used to suspend the Ag nanoparticles (“citrate control”, amount corresponding to 6 µg/mL Ag nanoparticles) also showed a reduced metabolic activity, but the effect was clearly enhanced in presence of the 3 and 6 µg/mL Ag nanoparticles. On the organismic level, Ag nanoparticles show only slight to negligible effects on traditional ecotoxicological endpoint markers like growth, mortality and reproduction of E. fetida (Kwak et al. 2014; Shoults-Wilson et al. 2011). Kwak et al. (2014) demonstrated the importance of the nanoparticular material, since Ag nanoparticles showed a slight effect on E. andrei while the exposure to pure Ag ions showed no effect at all. On the other hand, isolated coelomocytes of E. fetida cultivated in RPMI-1640 medium showed an LC50 value of 6 µg/mL (Garcia-Velasco et al. 2019). Coelomocytes cultivated in L-15 medium, on the other hand, showed a much higher resistance to Ag nanoparticles (LC50 > 100 µg/mL) (Garcia-Velasco et al. 2019). Differences between the studies may stem from differences in particle size as well as different cell types used and therefore non-comparable cellular reactions. In the end only additional studies on the cellular level can elucidate the mechanistic basis for the observed toxic effects.
In contrast to particles, whose size and surface coverage are known to have a significant effect on toxicity, hindering the direct comparison of published results, metal ions are considered to be more standardizable toxins. The response of the cells to copper and cadmium ions are summarized in Fig. 5B. The results show a significant decrease in metabolic activity for all tested concentrations (LMTreatment F8 = 5.027, p < 0.001), except 800 µg/mL CdCl2 after 48 h exposure. After 96 h the cells seemed to recover to some extent, but the metabolic activity was still low compared to the controls. A more pronounced effect had been expected in particular for the respective higher metal ion concentrations, i.e. 400 µg/mL CuCl2 and 800 µg/mL CdCl2, since these are already in the range of the median lethal concentration for E. fetida in soil, namely 500 – 700 mg/kg soil for copper and 600 – 1800 mg/kg soil for cadmium (Bernard et al. 2015; Neuhauser et al. 1985). Interestingly, coelomocytes isolated from E. fetida showed a similar response at least to changing cadmium doses, where the viability decreased in the presence of 100 µg/mL followed by an increase in viability at concentrations of 500 µg/mL (Irizar et al. 2015). The discrepancy between organismic effects and the reactions described here suggest different toxicological mechanisms on different levels of biological complexity.
3.4 Effects of MP particles on primary intestinal cells isolated from E. fetida
Finally, the influence of MP particles was investigated on the cells as an example of a new and increasingly important environmental pollutant with particular relevance for unspecific soil feeders such as E. fetida. PS microparticles were chosen as representatives of non-biodegradable commodity plastics, whereas PLA microparticles were chosen as example for a biodegradable polymer. The possibility of a cellular uptake of the particles was also studied, using particle sizes between 0.2 and 3 µm.
Surface properties and in particular a biomolecular corona on the particle’s surface have recently been suggested as decisive for cell particle interaction and uptake (Ramsperger et al. 2020). Therefore, once the MP particles are added to the cells in the protein-rich culture medium the formation of a protein corona is likely. ζ-potentials and size distribution measured of MP particles incubated at various conditions are summarized in Table 2.
Table 2 ζ-potential and size distribution measured for the MP particles used in this study
|
Nominal particle size
(µm)
|
ζ-potential
(mV)
|
Measured particle size
(µm)
|
|
|
KCl
|
Culture medium
|
KCl
|
Culture medium
|
PS
|
0.2
|
-47.4 ± 0.3
|
-25.3 ± 0.0
|
0.2 ± 0.006
|
0.2 ± 0.008
|
0.5
|
-52.8 ± 0.2
|
-27.6 ± 0.0
|
0.5 ± 0.008
|
0.6 ± 0.04
|
2
|
-76.7 ± 0.3
|
-28.8 ± 0.0
|
1.8 ± 0.03
|
1.5 ± 0.04
|
3
|
-78.9 ± 0.3
|
-29.3 ± 0.2
|
3.1 ± 0.08
|
3.3 ± 0.1
|
PLA
|
0.5
|
-1.1± 0.0
|
-1.3 ± 0.4
|
0.6 ± 0.02
|
1.1 ± 0.1
|
2
|
-3.9 ± 0.3
|
-11.2 ± 1.1
|
1.5 ± 0.2
|
2.4 ± 1.2
|
Incubation in culture medium led to a reduction of the ζ-potential. Even particles with significantly different ζ-potential before incubation showed similar ones after incubation in culture medium independent of particle diameter. This indicates the development of a similar protein corona on the surface of all investigated PS particles. Pristine PLA particles, on the other hand, initially showed a small negative ζ-potential, which was slightly increased in case of the 2 µm particles after incubation. As expected, the PLA particles were colloidally instable due to the low ζ-potential, and thereby the size of the PLA particles nearly doubled after incubation in the culture medium and the size distribution became wider.
Within the 24 exposure experiments (Fig. 6), only cells incubated with high concentrations of 2 µm PS particles showed a significantly reduced metabolic activity after 48 h of incubation (LMTreatment F24 = 2.291, p = 0.004, Tukey post-hoc comparison: p-value2 µm PS = 0.025). In the presence of 0.5 µm PLA at the high concentration after 48 h the cells even showed a significantly higher metabolic activity (LMTreatment F24 = 2.291, p = 0.004, Tukey post-hoc comparison: p-value0.5 µm PLA = 0.006).
Finally, confocal microscopy showed no apparent uptake of fluorescent MP particles of any type and size by the cells. While particles ≤ 0.5 µm showed some tendency for attachment to the cellular membranes, no signs of uptake or attachment were seen for larger particles (Fig. 7). These findings are in line with recent results for murine epithelial cell lines, where also no uptake of particles > 0.2 µm was observed. This might also explain the observed low effect of the particles on the metabolic activity of the cells. However, since the smallest investigated particles (0.2 µm) seem to attach to cells, possible secondary or cumulative effects cannot be excluded. Organismic effects, like tissue damage or the inflammation of the gut tissue as shown previously (Jiang et al. 2020) might derive from particles which are not taken up by cells but persist in extracellular spaces in the tissue.