Skin penetration studies
The ability of infective hookworm larvae to enter their host has been intriguing to researchers for many years. Early experiments were performed by Goodey in 1922 [18]. Since then, migration behaviour of hookworm larvae has been subject of several studies to analyse the mechanism of skin penetration [12], to evaluate the impact of protease inhibitors and antisera [13, 19–21], or as a parameter of fitness [22].
The present study aimed to establish an in vitro migration model as a reliable system to achieve reproducible and comparable results. Migrated larvae are often calculated as percentage of the original number of larvae, i.e. the number of larvae that had been calculated to be pipetteted. Practically, pipetting of exactly 300 larvae is not possible. To take these variations into account, in the present study the migration ratio was defined as migrated larvae as percentage of the number of larvae recovered from both compartments. This calculation excludes the possibility of migration ratios lower than 0% and higher than 100% and minimizes the effect of slightly variable absolute numbers of larvae. This is especially important to allow comparison of migration ratios from experiments using different lots of larvae suspensions. As this calculation omits larvae that started migration but did not succeed to completely migrate through the skin, the "skin lag" was introduced as additional parameter. It expresses the difference between the sum of larvae recovered from both compartments and the theoretically used total number of 300 larvae as percentage of 300 larvae. A higher skin lag compared to the respective control assay under standard conditions within the same experiment indicates that larvae for example remained in the skin and therefore could not be counted. The skin lag should be used to compare the results of parallel setups rather than from experiments using different lots of larvae suspensions.
Furthermore, the models as well as the experimental conditions vary between the different research groups which may lead to non-comparable results as these parameters might influence the migration. Therefore, the possible impact of different factors on the results of in vitro migration experiments was evaluated.
The use of thawed skin instead of fresh skin has the advantage of a higher practicability and flexibility. Furthermore, freezing allows using skin of one individual more efficiently, for example if there is a limit in the number of available PERL chambers, space in the incubator or simply time. The use of frozen skin also allows to obtain skin opportunistically from clinics or pathology, and to store it, so that for animal welfare reasons no purpose-collected skin from animals was needed. Consequently, comparison experiments started with this parameter. From pharmacological trials analysing transdermal absorption, different results regarding the comparability of fresh and thawed skin are known. According to Franz [15] and Harrison et al.[23], freezing has no effect on skin permeability. In contrast to that Ahlstrom et al.[24] observed differences in permeability for hydrocortisone between fresh and thawed canine skin, with a higher permeability in thawed skin. Nevertheless these authors also recommend the use of thawed skin, since even if it is a little more permissive than fresh skin, it still serves as limiting factor for diffusion and penetration of pharmaceutics. Although freezing might influence the chemical and physical status of the skin as a barrier, it could be assumed that the migration of hookworm larvae might not be affected to the same extent as the permeability for chemicals. The migration ratios in the present study were nearly identical for the use of fresh and thawed skin, and consequently no statistically significant difference was detected. These results are consistent with results presented by Matthews [25], who also observed no difference in the migratory behaviour of infective larvae of A. tubaeforme when using fresh or thawed skin. However, in contrast to migration ratios the skin lag in the present experiments was significantly lower when thawed skin was used. This indicates that some larvae remain in the fresh skin, e.g. after being trapped by still active immune cells, or hindered by structures, which might be destroyed by freezing and thawing. Thus, thawed skin might be a more artificial system than fresh skin. However, using the migration ratio as calculation basis for PERL chamber migration experiments, results are comparable. Because of practicability, flexibility as well as animal welfare aspects, thawed skin was used as standard parameter for all further experiments.
Another suggested important parameter concerning larval movement or skin penetration was the incubation temperature. The temperature-depending activity of hookworm larvae and other larvae is a commonly known phenomenon [9, 26, 27]. In the present study larvae also migrated in higher numbers through canine skin, when the temperature was closer to the temperature of the mammalian host. The highest skin lag in the present experiments was observed at 22°C. This observation is most likely due to the fact, that larvae had obviously started migration but apparently were not able to completely pass through the skin, at least not in the given time of 12 h. Most of the larvae incubated at 7°C had not started migration at all after 12 h, so the majority of them could be recovered from the donor compartment, which resulted in a lower skin lag than in the experiment at 22°C. For the experiments at 32°C and 37°C, most of the larvae were recovered from the acceptor chamber, explaining the low skin lags in these experiments. Although there was no statistically significant difference between the migration and skin lag at 32°C and 37°C, an incubation temperature of 37°C was chosen as a standard parameter since this temperature is comparable with the natural host's temperature.
Furthermore, standard PERL chamber experiments were conducted without additional CO2 because the presence of 5% CO2 led to a decreased migration ratio. Maybe this is due to the fact that under natural conditions the environment at the start of migration is ambient air containing 0.039% CO2 only. Interestingly, a high percentage of percutaneously migrated larvae used in subsequent feeding assays died within 48 h in the presence of additional CO2 whereas the other larval populations remained unaffected.
Concerning incubation time, the larvae obviously invaded the skin very quickly, but it took several hours before the majority of larvae were detected within the acceptor chamber. After 12 h incubation, migration ratios reached a maximum. Longer incubation periods did not result in higher migration or lower skin lag. From in vivo trials with beagles it is known that larvae of A. caninum need more time for skin penetration and migration than larvae of A. braziliense[28]. Williamson et al.[13] described that in their in vitro assays up to 100% of the A. caninum larvae had penetrated canine skin within 30 min and therefore could not be recovered from the skin surface. Such high a penetration was not achieved in the present experiments, and also not after pre-incubation at 37°C as performed by Williamson et al[13], which led to even higher numbers of remaining larvae (data not shown).. Thus, the larvae might have been activated before they were actually placed onto the skin. Kopp et al.[22] reported that in their experiments even 84.1% of the larvae successfully traversed canine skin within 2 h in contrast to up to 12 hours in the present setup. This rapid penetration and migration can be explained by the use of abdominal skin of very young puppies, generally about 6 weeks of age, and thoroughly removed subcutaneous tissue (S. Kopp, personal communication). Therefore, the barrier the larvae had to penetrate was very thin, whereas in the present study the skin was from dogs, which were at least several months old, and only loose subcutaneous tissue was removed. Another possible explanation could be the used A. caninum isolates, but three different isolates tested in the present study behaved comparably (data not shown). However, the present study revealed with approximately 86% penetrated A. caninum larvae within 1 h a quick invasion into the skin.
The orientation of the skin also influences larval migration and statistically significant differences were observed by testing this parameter. With the epidermal side upwards, 90.3% of the recovered larvae had completely migrated through the skin, in contrast to 81.0% when the dermal side was presented. The higher migration ratio with the epidermal side on top is not surprising since it mimics the natural conditions and larvae are attracted by the hair follicle system during migration [28]. The result that the tested acceptor media do not influence migration was expected since larvae most probably do not recognize which medium is beneath the skin, which functions as diffusion and penetration barrier [15].
Exsheathment and initiation of feeding of the third-stage larvae are assumed to be first steps in the development to parasitic stages [16, 29, 30]. The serum-stimulation method is commonly accepted as a model for the start of parasitic development [31]. However, Hawdon et al.[30] showed that feeding is not necessarily needed for development of hookworms, especially not after oral infection. Unfortunately, there are currently no data available on the percentage of feeding of A. caninum larvae and the time-course of the putative resumption of feeding after percutaneous infection in vivo. Regarding the purpose of feeding, Hawdon et al.[30] hypothesise that feeding is necessary after skin invasion, since the larvae have to cross different tissues (skin, lung etc.) on their way to the small intestine and therefore need much more time and energy than after being orally ingested. Therefore, the serum-stimulation resulting in the resumption of feeding presumably mimics percutaneous rather than oral infection in vivo, but lacks the migration step. In the study presented here the pmL3 -CO2 reached maximum feeding levels later than the ssL3 +CO2. This unequal time course may be caused by different stimuli or a different sequence of stimuli. But both, percutaneous migration and serum-stimulation induce exsheathment and feeding, two characteristics currently viewed as first steps of development into parasitic stages. However, since A. caninum larvae will usually not come directly into contact with dog serum in vivo before penetrating the skin, it might be assumed that the serum-stimulation omits important stimuli triggering further development. Thus, serum-stimulation alone may not reflect the whole truth. Compared to natural conditions, the PERL chamber percutaneous migration might be more adequate for the examination of molecular mechanisms and changes during the stage conversion towards parasitism. And indeed, larvae migrated through PERL chambers seem to be different from the infective and serum-stimulated larvae. So in current studies on gene transcription patterns in infective, percutaneously migrated, and serum-stimulated hookworm larvae, the obtained preliminary data show good evidence of different transcriptional regulations between these populations.