Functional morphology of suction discs and attachment performance of the Mediterranean medicinal leech (Hirudo verbana Carena)

Abstract

Medicinal leeches use their suction discs for locomotion, adhesion to the host and, in the case of the anterior disc, also for blood ingestion. The biomechanics of their suction-based adhesion systems has been little understood until now. We investigated the functional morphology of the anterior and posterior suckers of Hirudo verbana by using light and scanning electron microscopy. Furthermore, we analysed the adhesion qualitatively and quantitatively by conducting behavioural and mechanical experiments. Our high-speed video analyses provide new insights into the attachment and detachment processes and we present a detailed description of the leech locomotion cycle. Pull-off force measurements of the anterior and posterior suction organs on seven different substrates under both aerial and water-submersed conditions reveal a significant influence of the surrounding medium, the substrate surface roughness and the tested organ on attachment forces and tenacities.

1. Background

Medicinal leeches are often negatively connoted with their extensive medical misuse over decades which stigmatized them to be blood-sucking, disease-transmitting ectoparasites [1–4]. Today, these freshwater annelids experience an increasing importance in several medicinal fields, e.g. plastic surgery [5]. The majority of the medically used leeches worldwide belong to the species Hirudo verbana (Carena 1820) [6,7]. They also play an important role as model organisms in many biological research fields due to their physiologically and biochemically multifunctional saliva and their well-developed nervous and sensory systems [8–15].

Both the anterior and posterior suction disc participates in the animal's movement towards and attachment to the host, e.g. mammals and amphibians [16]. The anterior sucker additionally enables for feeding and consists of the ventral halves of the first four body segments that form a muscular and glandular suction cup (figure 1a) [17–20]. The central mouth opening is sealed by a velum behind which the retractable tripartite jaw apparatus is positioned [21,22]. The posterior suction disc consists of the last seven body segments and does not possess an orifice (figure 1b) [22,23]. A characteristic morphological feature of the posterior attachment organ is a joint-like constriction between itself and the rest of the body providing the leech with high moving flexibility [22,24].

Figure 1.

Morphology of medicinal leech suction discs. The figure shows light microscopy (LM) and scanning electron microscopy (SEM) images of the anterior (a,c,d) and posterior (b,e,f) suction discs of H. verbana. (a,b) LM images of attached anterior and posterior suckers. The retracted velum of the anterior sucker exposes the three jaws and the oral cavity. (c–f) SEM images of the inner surfaces of both suction organs showing furrows (black arrows) and gland-like structures (white arrows). (c,e) Marginal regions of inner sucker surfaces. (d,f) Central regions of inner sucker surfaces. (Online version in colour.)

Besides an undulatory swimming movement, leeches show two crawling modes, i.e. inchworm and vermiform crawling [20,25]. Both modes consist of the same sequence of anterior and posterior attachment and detachment processes that together form a typical movement cycle. In contrast with vermiform crawling, inchworm crawling is characterized by the formation of a body loop [18,20,26]. After the attachment of the anterior sucker to the substrate, a shift of the body's centre of gravity to anterior regions takes place, followed by a contraction wave on the body propagating in posterior direction. The posterior suction disc becomes detached and repositioned closer to the anterior sucker. This enables the release of the anterior sucker which ultimately reaches its next fixation point by an anteroposterior body elongation wave [18,20].

Information on the functional morphology and biomechanics of the medicinal leech suction organs is scarce at best. Gradwell [24] investigated the rhynchobdellid Placobdella parasitica and Lent & Dickinson [19] conducted the first experiments on the anterior attachment system of Hirudo medicinalis, but mainly focused on the termination of ingestion. Other studies focused on the sucker morphology in Branchiobdella or on the posterior attachment biomechanics of Whitmania pigra [27,28]. Here we studied the attachment abilities for both suction organs of the Mediterranean medicinal leech on several natural and artificial substrates and under aerial and underwater conditions. Furthermore, the general movement cycle as well as the single attachment and detachment processes of both suckers has been analysed.

2. Material and methods

2.1. Medicinal leeches

Adult leeches (body mass 0.9 ± 0.3 g, n = 40) were purchased from Biebertaler Blutegelzucht GmbH (Biebertal, Germany) and kept individually in small glass jars filled with rainwater (changed two to three times per week, sealed with perforated caps). The leeches were fed regularly according to Lent et al. [29].

2.2. Kinematical and functional–morphological analyses

2.2.1. Qualitative analyses of the suction organ kinematics and of the movement cycle

Single leeches were placed into a transparent water tank. By using a Motion Pro Y4 high-speed camera (IDT Inc., Tallahassee, FL, USA) equipped with a Makro-Planar T*2/100 ZF lens (Carl Zeiss AG, Oberkochen, Germany), the attachment and detachment processes of individual suckers as well as whole movement cycles were captured from frontal and lateral perspectives (framerate: 1000 fps, exposure time: 500 µs). The experimental set-up was illuminated by a Constellation 120 high-performance LED light source (IDT Inc.). The acquired videos were processed with the Motion Studio software (v. 2.10.05, IDT Inc.) and the total surface area of the suction organs were measured with ImageJ (v. 1.46r, US National Institutes of Health, Bethesda, MD, USA) both in retracted and extended conditions.

2.2.2. Scanning electron microscopy

Leeches were anaesthetized prior to the preparation procedure using carbonized water. Suction disc samples were frozen with liquid nitrogen, cut, fixated in FAA (48 h) and dehydrated in a methanol sequence (70%, 90%, 100%, 24 h intervals). Samples were critical-point dried (BalTec CPD 030, Leica Microsystems GmbH, Wetzlar, Germany), fixed to aluminium stubs using conductive adhesive pads (G3347, Plano GmbH, Wetzlar, Germany) and gold-sputtered (Sputter Coater 108 auto, Cressington Scientific Instruments Ltd., Watford, UK). A Leica Leo 435 VP scanning electron microscope (Leica, Wiesbaden, Germany) was used.

2.3. Biomechanical analyses

2.3.1. Qualitative analysis of the attachment performance

Tests were carried out with an absorptive surface (GPP, glossy photo paper, Opti^®Photo Premium, Papyrus AB, Mölndal, Sweden) and a non-absorptive surface (HDPE, high-density polyethylene) in order to evaluate if fluid removal influences the attachment and movement performance. Single animals were placed onto a 10 × 10 cm plastic plate (equipped with the test surfaces) which was rotated stepwise from 0° to 180° in 45° intervals. At each inclination the ability to establish a successful contact with both suckers and to perform at least one complete locomotion cycle before falling off the substrate was recorded. To successfully pass a test, leeches had to adhere to or move on a given surface at a given inclination within three trials. If they did not succeed within the given amount of trials, they were randomly subjected to the next test.

2.3.2. Fabrication and characterization of test surface replicas

The attachment ability of H. verbana was tested experimentally on replicas of four natural and three artificial surfaces (figure 2). Adaxial leaf surfaces of Nymphaea sp. (water lily) and Caltha palustris (marsh marigold) (both collected in the Botanic Garden Freiburg), the surfaces of slate rock (collected near Holzmaden, SW-Germany) and human skin (from the second author's forearm) were chosen to represent natural surfaces. Two types of sandpaper with different grain sizes (Matador P280 and P80, Starcke GmbH & Co. KG, Melle, Germany) and acrylic glass (PMMA) represented the artificial surfaces. To measure only the influence of their structural properties on leech attachment, we replicated all test surfaces using a two-step procedure (electronic supplementary material, figure S1). Firstly, original surfaces were cast using polyvinyl siloxane (PVS, President light body, Colténe/Whaledent AG, Altstatten, Swiss). After hardening, surrounding PVS collars were modelled to the negatives. Complete negatives were filled with Toolcraft epoxy resin L and hardener S (both Conrad Electronics SE, Hirschau, Germany). The backsides of the replicas were flattened with a Teflon^® plate during overnight hardening. The surface roughness was characterized using an Olympus LEXT OLS4000 (Olympus Corp., Tokyo, Japan) by determining the arithmetic average height (S_a), root mean square height (S_q) and skewness (S_sk) (describing the vertical dimensions of surface irregularities) as well as the kurtosis (S_ku) and texture aspect ratio (S_tr) (providing spatial information) and the core void volume (V_vc) (a functional parameter normalized to the measurement area indicating the height of a fluid volume theoretically needed to fill the surface texture). Five rectangular sampling areas (1.2 × 1.2 mm) per replica were scanned (20×, fine mode).

Figure 2.

CLSM images of the seven replicated test substrates including surface profiles showing their different texturing. (a–d) Surface replicas of natural substrates (Nymphaea sp., C. palustris, human skin, slate rock). (e–g) Surface replicas of artificial substrates (fine and rough sandpaper, acrylic glass). (h) Surface profiles of all seven test surfaces (R1–R7).

2.3.3. Quantification of the attachment characteristics

In order to quantify the attachment forces of the anterior and posterior suckers, we modified a method used by Lent & Dickinson [19] (electronic supplementary material, figure S2). A small plastic bowl was glued to a 1 kg balance weight situated on a top-loading balance (PCB 10000-1, Kern&Sohn GmbH, Balingen-Frommern, Germany). The surface replicas were glued into the plastic bowls which allowed for their submersion in water. During tests the balance display was recorded continually with a USB camera (DigiMicro v. 2.0 Scale, dnt GmbH, Dietzenbach, Germany) with 25 fps.

Single leeches were held over this construction until they adhered to the test surface and then pulled off orthogonally, thereby decreasing the initial balance load (m_i). After each test, the video material was analysed using Virtual Dub (v. 1.10.4, Avery Lee) to find the minimum load (m_min). Maximum attachment forces (F_max) were calculated from the equation

with g being the Earth's standard acceleration due to gravity. Additionally, we calculated the tenacities (T) by normalizing the attachment forces according to the suction organ contact areas (SOCAs) measured in advance. The tenacity formula reads

In total, attachment forces and tenacities of 27 leeches were measured for both suckers on seven replicas in two different surrounding media (air and water). For tests in air, the leeches were gently dabbed dry with a paper towel.

2.4. Statistics

We used the statistics software GNU R v. 3.1.1 [30]. Considering both leech suckers within each experiment, we primarily had to work with longitudinal data. If not specifically indicated, all subsequently mentioned functions were already included in GNU R (references of all R packages are given in the electronic supplementary material, Methods S1).

To test correlations, we performed Spearman's rank-order correlations (cor-function). The difference between the anterior and posterior increase of SOCA was analysed with the Wilcoxon test for paired data (wilcox.test-function). Furthermore, the data of the qualitative attachment ability tests was investigated using Fisher's exact test (fisher.test-function). Additionally, we compared specific SOCA groups (anterior, posterior, retracted, extended) with each other or with the leech body weights using a one-way repeated-measures ANOVA on rank transformed data (lme-function of nlme-package; anova-function). As post hoc analysis, we conducted pairwise comparisons with Bonferroni adjusted p-values (pairwise.t.test-function). The same statistical analysis was carried out to test for significant differences between the anterior and posterior attachment and detachment durations, as well as for the textural differences of the surface replicates according to specific roughness parameters. For all methods, the required assumptions were checked.

Finally, we performed a linear mixed-effects model approach for analysing the relationships between three independent variables (organ, surface and medium) and two-dependent variables (attachment force and tenacity). As fixed effects, we put organ, surface and medium into the model allowing all possible interaction effects. As random intercepts, we nested the factor organ inside the factor leech accounting for their biological link. In the full model, all possible interactions of organ, surface and medium inside the random slopes were allowed to cover the complete biology-based variation. Controlling the final models diagnostic plots revealed deviations from the homoscedasticity assumption, which is why we log-transformed the dependent variables using the natural logarithm and re-analysed the full models. Afterwards, all necessary assumptions were met. In general, model simplification was achieved by likelihood ratio tests of the full model with the effect in question against the model without this effect. Here, the crucial parameters were the p-values, the Akaike information criterion and the Bayesian information criterion which were used to simplify the models. A detailed description of the linear mixed-effects model analysis procedure is described in the electronic supplementary material, Methods S1. Average least-squares means were calculated for both final models (maximum attachment force model and tenacity model) to evaluate the results of this experiment.

3. Results

3.1. Qualitative analyses of the suction organ kinematics and of the movement cycle

3.1.1. Suction organs

The tested medicinal leeches brought their anterior and posterior suction organ into tight substrate contact during all observed attachment processes (figure 1a,b). When a leech was hindered in performing its normal locomotion cycle, the animal retracted the velum and simultaneously started a rhythmical pumping movement (figure 1a; electronic supplementary material, Video S9). SEM images revealed that the entire inner sucker surfaces are covered with differently pronounced furrows which gradually become less prominent from marginal to central regions (figure 1c–f). Furthermore, posterior suckers display a set of larger, radially arranged furrows that are less pronounced in anterior suckers (figure 1c,e). Mucus residues are visible in the furrows of both organs (figures 1d–f and 5). Moreover, gland-like structures are scattered across the entire inner sucker surfaces and are concentrated near the furrow rims (figure 1d,f).

Figure 5.

Intensive mucus production during leech attachment. (a) The picture shows H. verbana during upside-down crawling along a metal wire mesh. Intensive mucus production is visible in the head region (mucus thread) and on the leech back (drop-like mucus accumulation). The beads-on-a-string configuration of the thread suggests visco-elastic properties of the leech mucus (cf. [35]). (b,c) Scanning electron microscopy images showing secretory glands and mucus residues on the inner sucker surfaces of the anterior (scale bar is 90 µm) and posterior attachment organs (scale bar is 60 µm), respectively.

SOCA increases significantly during attachment (one-way repeated-measures ANOVA on rank transformed data, F_3,117 = 200.06, p < 0.001; table 1). Moreover, the animal's body weight correlates positively with the retracted anterior SOCA, the retracted posterior SOCA, the extended anterior SOCA and the extended posterior SOCA (electronic supplementary material, figure S3a) (Spearman's rank-order correlation: Spearman's ρ = 0.62, 0.73, 0.37 and 0.72, respectively). As to their medians, the anterior SOCA extends from 12.66 to 19.25 mm², whereas the posterior SOCA increases from 14.99 to 23.84 mm² during attachment (table 1). The resulting 59% posterior increase of SOCA is thereby significantly larger than the corresponding 52% anterior increase of SOCA (Wilcoxon Mann–Whitney signed rank test, W = 165, p < 0.001). Only the posterior increase of SOCA correlates positively with the body weight of H. verbana (electronic supplementary material, figure S3b; Spearman's rank-order correlation: Spearman's ρ = 0.03 and 0.42, respectively). Comparing the retracted and extended SOCAs, the posterior suckers are generally larger (one-way repeated-measures ANOVA on rank transformed data, F_3,117 = 200.06, p < 0.001; table 1). In addition, anterior retracted and posterior retracted SOCAs, as well as anterior extended and posterior extended SOCAs correlate positively with each other (electronic supplementary material, figure S3c; Spearman's rank-order correlation: Spearman's ρ = 0.76 and 0.67, respectively). The same holds for anterior retracted and anterior extended SOCAs, as well as for posterior retracted and posterior extended SOCAs (electronic supplementary material, figure S3d; Spearman's rank-order correlation: Spearman's ρ = 0.30 and 0.88, respectively). Regarding the attachment and detachment durations, only the posterior attachment process lasts significantly longer than its detachment process (one-way repeated-measures ANOVA on rank transformed data, F_3,27 = 4.87, p < 0.01; table 2).

Table 1.

SOCAs of H. verbana suckers before and after (retracted), as well as during (extended) surface attachment.

attachment organ	condition	SOCA (mm²)					increase of SOCA (mm²)
		median	IQR	Q1	Q3	n	median	IQR	Q1	Q3	n
anterior	retracted	12.66	2.91	11.21	14.12	40	5.72	5.13	4.08	9.21	40
	extended	19.25	5.58	16.76	22.33	40					40
posterior	retracted	14.99	5.45	13.62	19.07	40	8.61	2.82	7.04	9.86	40
	extended	23.84	6.98	21.22	28.19	40					40

Table 2.

Attachment and detachment durations of H. verbanas anterior and posterior suckers.

attachment organ	condition	time (s)
		median	IQR	Q1	Q3	n
anterior	attachment	1.00	0.28	0.95	1.24	10
	detachment	0.96	0.34	0.77	1.11	10
posterior	attachment	1.12	0.28	1.09	1.37	10
	detachment	0.60	0.57	0.54	1.11	10

3.1.2. Attachment and detachment processes

A characteristic series of individual attachment and detachment events forms the typical locomotion cycle of the leech (figure 3). Before initial contact, a large portion of the contact surface the leech is going to adhere to is covered by the upper and lower margins of its suction cup (figure 3a,l; arrows). When being pulled to lateral positions, these sucker margins gradually uncover the central sucker region which allows for a development of full SOCA during anterior attachment (figure 3a–e). Simultaneously, the central part of the suction organ extends to facilitate the initial contact to the substrate (figure 3a,f). During contact formation, the SOCA gradually increases (figure 3b–g). In reverse, the SOCA decreases during detachment. Additionally, the pressure compensation in the sucker, a lower margin shear-off movement and an upward tilt of the upper margins support the anterior sucker detachment (figure 3h–l, arrows). After complete detachment, the sucker displays its initial configuration (figure 3l).

Figure 3.

Suction disc kinematics. Anterior attachment (a–e, frontal; f–g, lateral) and detachment (h–l, frontal; m–n, lateral) processes. Arrows mark the upper (black) and lower (white) margins of the suction organ. Posterior attachment (o–s, lateral; t–u, frontal) and detachment (v–z, lateral; zi–zii, frontal) processes. Images are frames from the electronic supplementary material, Videos S1–S8. Original brightness and contrast have been increased to improve the image clarity.

Prior to attachment, the posterior sucker displays a slight extension of its central part (figure 3o–p). Subsequently, the sucker is pressed against the substrate leading to flattening and increase of SOCA (figure 3q–s, t–u). The posterior detachment process differs from the anterior one. The SOCA decreases by a furling movement of the suction disc margins that is combined with a shearing resulting from the leech locomotion (figure 3v–z, zi–ii).

3.2. Qualitative analysis of the attachment performance

Statistically, the leech attachment performances on GPP and HDPE do not differ (Fisher's exact test, p > 0.05; table 3). Two general trends can be noted. Firstly, a higher surface inclination (greater than 90°) entails an increase in unsuccessful attachment attempts (table 3). Secondly, this trend is more characteristic for leeches trying to attach to the absorptive GPP. Considering the qualitative locomotion data, only two of five surface inclinations (0° and 180°) displayed significant influences on adhesion (Fisher's exact test, p < 0.05; table 3). In both cases, the movement performance on GPP was poorer compared with the locomotion on HDPE. However, in each subtest complete locomotion cycles were observed for at least some leeches, demonstrating a general movement potential on both substrates at each tested inclination. In general, the locomotion tests also showed more unsuccessful movement attempts on surfaces with higher surface inclination, a trend which was more pronounced on GPP (table 3).

Table 3.

Qualitative attachment and locomotion analysis on absorptive (GPP) and non-absorptive (HDPE) substrates.

		surface inclination (°)
		0		45		90		135		180
attachment	surface	yes	no	yes	no	yes	no	yes	no	yes	no
	HDPE	28	0	28	0	25	3	23	5	24	4
	GPP	28	0	28	0	21	7	19	9	17	11
Fisher's exact test statistic	p-value	p > 0.05		p > 0.05		p > 0.05		p > 0.05		p > 0.05

		surface inclination (°)
		0		45		90		135		180
locomotion	surface	yes	no	yes	no	yes	no	yes	no	yes	no
	HDPE	28	0	26	2	19	9	10	18	17	11
	GPP	23	5	23	5	12	16	5	23	4	24
Fisher's exact test statistic	p-value	p < 0.05		p > 0.05		p > 0.05		p > 0.05		p < 0.001

3.3. Quantification of the attachment characteristics

The results of the surface characterization are shown in the electronic supplementary material, table S1. The seven test substrates did not significantly differ in their S_ku and S_tr values (one-way repeated-measures ANOVA on rank transformed data, S_ku, F_6,24 = 1.92, p > 0.05, S_tr, F_6,24 = 1.20, p > 0.05). Regarding the S_a, S_q and V_vc parameters, pairwise comparisons of the test surfaces predominantly displayed significant differences with only three exceptions (one-way repeated-measures ANOVA on rank transformed data, S_a, F_6,24 = 86.96, p < 0.001, S_q, F_6,24 = 88.50, p < 0.001, V_vc, F_6,24 = 88.48, p < 0.001). The replicas of rough sandpaper versus human skin, of fine sandpaper versus C. palustris and of slate rock versus Nymphaea sp. did not differ in height roughness and core void volume. The acrylic glass replica was smoother and had a smaller core void volume than any other tested surface (p < 0.05). Considering the S_sk roughness parameter, only the values of rough sandpaper and human skin varied significantly from each other (one-way repeated-measures ANOVA on rank transformed data, S_sk, F_6,24 = 2.73, p < 0.05).

We measured the attachment forces of the anterior and posterior suckers on various surface replicas both under aerial and water-submersed conditions (electronic supplementary material, figure S2). In order to not only account for single effects of suction organs, surrounding media and test surfaces, but also for their interaction, we adopted a linear mixed-effects model approach (see the electronic supplementary material, Methods S1, for detailed descriptions). For each applied dataset (measured or log-transformed data) and each respective attachment parameter (maximum attachment force or tenacity), we observed the same tendencies for given test conditions (table 4).

Table 4.

Attachment characteristics according to experimentally determined and modelled data. Only significant factors (organ, medium, surface and surface:medium) are shown. The measured data are indicated as medians and interquartile ranges (IQR) (n = 27). The results of the linear mixed-effects models are depicted by least-squares means, standard errors (s.e.), degrees of freedom (d.f.) and lower and upper confidence limits (CL). Except for the degrees of freedom, all modelled data are shown by non-dimensional, logarithmized values (natural logarithm). A, anterior; P, posterior; D, dry; W, wet; R1, Nymphaea sp.; R2, C. palustris; R3, human skin; R4, slate rock; R5, fine sandpaper; R6, rough sandpaper, R7, acrylic glass.

factor	condition	measured data		modelled data
		median	IQR	LS mean	s.e.	d.f.	lower CL	upper CL
maximum attachment force (Fmax)
organ		(mN)		(/)			(/)
	A	547.62	515.10	6.12	0.05	49.73	6.02	6.23
	P	447.21	399.11	5.98	0.05	49.73	5.87	6.09
medium
	D	390.93	351.61	5.76	0.05	37.62	5.67	5.86
	W	630.29	545.37	6.34	0.05	37.62	6.24	6.43
surface replica
	R1	612.63	430.79	6.38	0.06	103.35	6.26	6.51
	R2	615.57	329.00	6.37	0.06	103.35	6.25	6.49
	R3	251.94	440.51	5.41	0.06	103.35	5.29	5.53
	R4	631.53	499.34	6.40	0.06	103.35	6.28	6.52
	R5	466.95	297.77	6.11	0.06	103.35	5.99	6.23
	R6	176.59	180.27	5.16	0.06	103.35	5.04	5.28
	R7	671.38	483.51	6.52	0.06	103.35	6.40	6.64
surface:medium
	R1:D	466.57	272.39	6.15	0.08	225.86	6.00	6.30
	R2:D	497.24	267.73	6.20	0.08	225.86	6.05	6.35
	R3:D	82.41	130.72	4.62	0.08	225.86	4.47	4.77
	R4:D	451.26	332.41	6.17	0.08	225.86	6.02	6.32
	R5:D	399.44	219.71	5.98	0.08	225.86	5.83	6.13
	R6:D	146.37	133.42	4.92	0.08	225.86	4.77	5.07
	R7:D	552.80	268.82	6.32	0.08	225.86	6.17	6.47
	R1:W	763.82	458.48	6.62	0.08	225.86	6.47	6.77
	R2:W	704.85	419.48	6.55	0.08	225.86	6.40	6.70
	R3:W	521.06	418.89	6.21	0.08	225.86	6.06	6.36
	R4:W	840.23	525.99	6.63	0.08	225.86	6.48	6.78
	R5:W	541.55	334.51	6.25	0.08	225.86	6.10	6.40
	R6:W	242.23	190.00	5.40	0.08	225.86	5.25	5.55
	R7:W	896.64	735.75	6.72	0.08	225.86	6.57	6.87
tenacity (T)
organ		(mN mm−²)		(/)			(/)
	A	27.70	24.79	3.14	0.04	56.08	3.05	3.23
	P	15.38	11.13	2.60	0.04	56.08	2.52	2.67
medium
	D	16.04	14.30	2.58	0.03	49.62	2.51	2.65
	W	26.70	24.74	3.15	0.03	49.62	3.09	3.23
surface replica
	R1	22.31	24.63	3.20	0.05	218.03	3.10	3.31
	R2	23.66	16.06	3.19	0.05	218.03	3.09	3.29
	R3	10.35	18.73	2.23	0.05	218.03	2.13	2.33
	R4	25.70	20.40	3.22	0.05	218.03	3.12	3.32
	R5	18.08	11.46	2.93	0.05	218.03	2.83	3.03
	R6	8.44	7.05	1.98	0.05	218.03	1.88	2.08
	R7	30.14	23.24	3.34	0.05	218.03	3.23	3.44
surface:medium
	R1:D	18.26	9.01	2.97	0.07	459.21	2.84	3.11
	R2:D	20.59	9.48	3.02	0.07	459.21	2.88	3.15
	R3:D	3.65	5.03	1.43	0.07	459.21	1.30	1.57
	R4:D	20.49	12.51	2.99	0.07	459.21	2.85	3.12
	R5:D	16.35	10.07	2.79	0.07	459.21	2.66	2.93
	R6:D	6.31	6.18	1.74	0.07	459.21	1.60	1.87
	R7:D	22.96	17.12	3.13	0.07	459.21	3.00	3.27
	R1:W	35.64	29.25	3.44	0.07	459.21	3.30	3.57
	R2:W	29.21	18.42	3.37	0.07	459.21	3.23	3.50
	R3:W	21.96	18.18	3.03	0.07	459.21	2.89	3.17
	R4:W	32.66	26.17	3.45	0.07	459.21	3.32	3.59
	R5:W	21.98	14.33	3.07	0.07	459.21	2.93	3.20
	R6:W	9.28	8.76	2.22	0.07	459.21	2.08	2.35
	R7:W	38.11	27.61	3.54	0.07	459.21	3.40	3.67

Therefore, we only present the significant single and interaction effects of the log-transformed tenacity values here. First, medicinal leeches created significantly higher tenacity values with their anterior suction organs (t₂₈ = 8.95, p < 0.001, table 4, electronic supplementary material, table S2). Secondly, water-submersed suction organs adhered more strongly to the given substrates than dry-dabbed suckers under aerial conditions (t₇₁₅ = −16.69, p < 0.001, table 4, electronic supplementary material, table S2). Thirdly, the structural properties of the replicated contact surfaces (R1–R7) also influenced the attachment performances. The seven tested replicas can be divided into four significantly different groups (table 4). The first group is formed by the smooth surface replicas of R1, R2, R4 and R7, on which the leeches displayed the highest tenacities (table 4). Slightly lower tenacity values were observed on R5. Higher roughness (on R3 and R6) led to further reductions in tenacity (table 4). The same tendencies are observed for nearly all S_a, S_q and V_vc roughness parameters with only few exceptions. Specifically, R2 and R5 as well as R3 and R6 did not significantly differ in their roughness parameters but in their tenacity values (R2 versus R5: t₇₁₅ = 4.05, p < 0.01; R3 versus R6: t₇₁₅ = 3.93, p < 0.01, table 4, electronic supplementary material, table S2). In summary, the attachment performance of H. verbana decreased with increasing substrate structuring. As derived from our model, the interaction of the factors medium and surface significantly influenced the tenacity values on all test substrates (table 4). Only on R5 the attachment ability did not change significantly. On R3, this effect was approximately four times higher than on the other replicas.

4. Discussion

4.1. Performance of attachment organs during locomotion

Our high-speed videos reveal that larger leeches possess larger SOCAs. In both types of suckers, SOCA increases during the attachment process. The observed morphological changes presumably lead to an increase of the maximum attachment force and enhance leech attachment. The differences found between the two attachment organs (the posterior sucker differing from the anterior sucker by a slightly but significantly larger increase of SOCA, a positive correlation of SOCA with body weight and generally larger dimensions) can be interpreted as adaptations to their different biological functions. Whereas the posterior sucker exclusively provides attachment, the anterior sucker additionally enables food ingestion [17,19,29]. Therefore, we assume that the anterior sucker morphology is subject to a compromise between optimal fixation and ingestion capabilities. However, both suckers are capable of securely anchoring the parasite to the host although leech body mass increases 10-fold during blood meals [19,29].

Except for the posterior attachment process, that lasted significantly longer than the posterior detachment, all other processes did not vary significantly from each other concerning duration (table 2). We conclude that this prolonged attachment process of the sucker further emphasizes its crucial role during surface contact formation and securing of attachment.

Moreover, we suggest that the metameric segmentation of their body enables leeches to precisely coordinate the single attachment and detachment processes. According to its innervation (head ganglion, tail ganglion and 21 body segment ganglia), the leech body can be subdivided in 23 functional units, each possessing its own ganglion and influencing the hydroskeleton by adjusting its local musculature. H. verbana possesses four muscle types. Centrally located longitudinal muscles antagonize with peripheral circular muscles, both types together being responsible for the adjustment of the leech body. Two layers of oblique muscles are present at an intermediate position. Dorsoventral muscles play a crucial role in body flattening during swimming [22,25,31–33]. Although the specific sucker anatomy of H. verbana remains unknown, we speculate that muscle tissue distribution as well as their synergistic interactions strongly influence medicinal leech attachment and locomotion, as already shown for W. pigra [28]. Therefore, we interpret their involvement in the locomotion cycle according to our high-speed video data and their anatomical set-up in normal body segments.

At the beginning of the locomotion cycle, the leech is posteriorly attached to the substrate (figure 4a). Prior to attachment of the anterior sucker, the leech ‘scans’ its environment for an appropriate contact area (figure 4b) during which an anteroposterior elongation wave sequentially propagates over the body segments. This results from a relaxation of longitudinal and a simultaneous contraction of circular muscles (figure 4b,c). Simultaneously, the leech everts central parts of its anterior suction organ by contracting oblique and circular muscles and relaxing longitudinal muscles in the head unit (figure 3a). We hypothesize that the oblique muscles are inserted between marginal and central sections of the anterior sucker. Then the everted regions form the initial substrate contact (figure 4c). A following increase of SOCA simultaneously takes place with the forward movement of the animal. The relaxation of the oblique muscles in the head unit potentially leads to a partial reversion of the inner suction area, resulting in a further increase of SOCA. Finally, the suction effect enhances the contact formation between the head unit and the substrate, leading to the final increase of the SOCA. Anatomically, this is achieved by a contraction of centrally located longitudinal muscles simultaneously with a relaxation of the circular muscles in the head unit leading to a passive aside-pushing of the upper and lower marginal parts of the anterior sucker (figure 3f–g). By this the installation of the animal's anterior fixpoint is complete (figure 4d). Together with the suction force generation two events occur: a posteroanterior shift of the body's centre of gravity and an anteroposterior contraction wave of the body segments caused by the reverse muscle activity compared with the one during the elongation wave.

Figure 4.

Locomotion cycle of H. verbana. (a) Starting position. (b) Searching (including elongation wave propagation). (c) Start of anterior attachment process. (d) Start of posterior detachment and end of anterior attachment processes. (e) Posteroanterior shift of the body's centre of gravity and end of posterior detachment process (including contraction wave propagation). (f) Loop formation (exclusive for inchworm crawling). (g) Start of posterior attachment and anterior detachment processes. (h) Anteroposterior shift of the body's centre of gravity and end of posterior attachment process. (i) End of anterior detachment process and completion of locomotion cycle. The asterisk marks the body's centre of gravity. Dotted lines subdivide the leech body in head unit, 21 body segment units and tail unit. The central compass indicates the anterior (A), dorsal (D), posterior (P) and ventral (V) directions.

Both events facilitate the following posterior suction disc detachment characterized by the relaxation of longitudinal muscles and by contraction of circular muscles in the ‘tail unit’ (figure 4d–e). This fine-tuned muscle activity leads to a furling movement and a cancellation of the negative pressure generation inducing the SOCA decrease (figure 3zi–zii). Additionally, the propagation of the contraction wave causes tensile forces on the posterior sucker ultimately shearing it off the substrate. Subsequently, the propagation of the contraction wave leads to an overall shortening of the leech and to the relocation of the posterior attachment organ close to the anterior organ (figure 4f). The loop formation in the inchworm crawling type can be explained by an asymmetrical contraction pattern of the longitudinal muscles in the body segments. The asymmetry itself is probably due to intensity variations or time delays of single contraction events. Directly after detachment, the slightly everted posterior disc is ready for new contact formation. The ‘tail unit’ is pushed against the substrate initiating organ flattening which is supported by relaxation of circular and contraction of longitudinal muscles (figure 4g). Additionally, SOCA increases due to suction force generation. Simultaneously, the leech shifts its body's centre of gravity back to the posterior end, hereby completing the posterior attachment process (figure 4h). For finishing the first loop of the locomotion cycle, H. verbana releases its anterior organ. For that, circular muscles in the head unit contract while at the same time longitudinal muscles elongate to achieve both the SOCA decrease and the cancellation of negative pressure generation. As the ventral halves of the first four body segments form the anterior sucker [21], the musculature of the fifth body segment directly surrounds the attachment area. We conclude that the aforementioned lower margin of H. verbana's front sucker in fact belongs to the ventral half of the fifth body segment. Furthermore, the contraction of circular muscles in this segment results in a shear movement which detaches the ventral halves of the first four body segments (figures 3h–k and 4g,h). The dorsal halves of these segments forming the upper sucker margin are then passively moved to their initial positions. Possibly, they support the general anterior detachment process by an upward tilting movement (figure 3k–l). The release of the anterior sucker may be assisted by a transient, asymmetric and time-delayed contraction of the dorsal and ventral longitudinal muscles in the front part of the body. Finally, H. verbana can continue its cyclic locomotion with the search for another contact zone (figure 4i).

4.2. Overall attachment performance

The suckers of H. verbana attach successfully to a broad range of surfaces differing in roughness, texture and shaping (namely stones, plants, skin, fur, etc.) with a firmer attachment underwater than in air. Our qualitative attachment experiments at varying inclinations display a positive influence of wetted suction organs on both the attachment and the locomotion. These results could imply an involvement of other functional principles in addition to the primary suction adhesion during leech attachment and locomotion. Moreover, our pull-off tests clearly show that the surrounding medium, the surface roughness, the inspected suction organ and the interactions between the surrounding media as well as the test surfaces significantly influence the attachment.

We also observed that F_max is higher in the anterior sucker. From an ecological perspective, this may be due to the fact that the anterior attachment strength during food intake is crucial. The functional morphology responsible for the different attachment abilities of the anterior and posterior suction organs remains unclear. Possibly, the observed pumping behaviour involving both the jaw apparatus and the pharyngeal cavity, may explain these differences by causing a higher underpressure in the anterior sucker. When not being able to continue its usual locomotion cycle, leeches significantly increased the volume of the suction chamber by opening the velum. Simultaneously, the jaw apparatus was moved periodically back and forth (figure 1a; electronic supplementary material, Video S9). This behaviour is similar to the suction mechanism of the larval net-winged midges of Hapalothrix lugubris (Blephariceridae) [34]. Both the volume increase and the pumping movement probably enhance anterior attachment forces. During unhindered attachment, none of the afore-mentioned reactions were notable. Therefore, we consider this phenomenon as an adaptive behaviour used to maintain the anterior attachment while handling a stress situation (when being pulled off the test substrate).

The finding that leeches adhere significantly less firmly on all test surfaces at aerial conditions than underwater correlates to their typical habitats with wet or water-submersed conditions. We assume that the presence of water or a mucus layer covering the suction area favours attachment by compensating small surface irregularities. The presence of mucus residues, gland-like structures and a furrow network, probably enhancing mucus secretion and distribution (cf. [28]), support this hypothesis (figures 1c–f and figure 5). Remarkably, no hairy structures were found on the sucker surfaces which reportedly seal underwater suction discs in other animals [36–38]. Together with the results of our qualitative attachment experiments this supports the hypothesis that wet adhesion is an important factor in leech attachment. Moreover, leeches are able to crawl upside-down along metal wire meshes, where other attachment principles than suction must apply, e.g. physical clamping or adhesives (note the visible mucus in figure 5, with threads showing a beads-on-a-string morphology possibly indicating visco-elastic properties (cf. [35]).

In general, attachment forces and tenacities are negatively correlated to surface roughness (figure 6). Primarily, significantly rougher substrates (according to the S_a, S_q and V_vc) lead to significantly poorer attachment performances. However, two exceptions contradict this general tendency: H. verbana adhered significantly stronger to the replica of C. palustris compared with the replica of fine sandpaper (table 4) and higher attachment forces and tenacities were achieved on the rough sandpaper replica compared with the human skin replica. In both cases, the S_a, S_q and V_vc parameters do not significantly vary between the respective test substrates, but the S_sk parameters indicate major differences in their surface texture. The human skin replica possesses numerous deep and narrow cavities, whereas the rough sandpaper substrate exhibits large grains with huge interspaces leading to statistically different S_sk values. Similarly, although not significant, the C. palustris replica predominantly possesses small cavities, whereas the fine sandpaper substrate is characterized by medium-sized grains and interspaces. The remaining roughness parameters (S_ku and S_tr) do not vary significantly between the test substrates and we conclude that they do not markedly influence the attachment properties. In addition to the afore-mentioned roughness effects, we assume that other factors (e.g. friction, van der Waals forces, actual contact area or spacing of surface irregularities) further influence the overall attachment performance (figure 6).

Figure 6.

Parameters influencing the attachment performance of H. verbana. Increasing surface contact areas favour leech attachment, whereas increasing surface roughness as well as increasing interspaces between surface irregularities lower attachment performance.

Finally, we found an interrelation between the test parameters surrounding medium and surface roughness for all tested substrates. Considering the surface roughness results, the surrounding medium either significantly enhanced (wet condition) or reduced (dry condition) the average values of attachment force and tenacity. We conclude that smooth surfaces displaying only small irregularities lead to stronger attachment due to large contact areas. Nevertheless, the presence or absence of fluid (e.g. water or mucus) slightly increased or decreased the ability of the leeches to compensate for small surface texturing. On the other hand, strongly textured substrates (e.g. rough sandpaper) hinder strong attachment. Surface irregularities and large interspaces impede the functionality of suction adhesion systems due to a reduced underpressure generation resulting from an imperfect sealing of the suction chamber. Medium effects could slightly influence the leeches' attachment characteristics on such surfaces by potentially sealing small irregularities while simultaneously increasing the contact area. But neither the addition of fluid, nor the high sucker deformability sufficiently increased the underpressure and the contact area to obtain high attachment performances.

In contrast with that the interaction effect of medium and surface roughness on the human skin replica was approximately four times as large as on any other tested substrate. We hypothesize that the presence of an incompressible fluid or mucus can fill the irregularities of the human skin replica, thereby immensely improving the suction chamber sealing, the underpressure generation and the contact area formation.

Surprisingly, the attachment ability of H. verbana was not significantly different on the fine sandpaper replica according to the given interaction effect. Probably the contact area reduction in combination with large interspaces between single grains lowered the attachment forces and tenacity values compared with those of smoother substrates. Furthermore, we assume that the interspaces of this test surface were deep and wide enough to minimize possible beneficial influences of both the sucker morphology and the interaction effect of medium and surface roughness.

We conclude that leech attachment is mainly due to negative pressure generation (i.e. suction), although both upside-down crawling along metal wire meshes and reduced attachment performances under dry conditions indicate that leeches possibly employ additional attachment principle(s) (figure 5a). Here mechanical interlocking, capillary and viscous forces or van der Waals forces (especially, in the case of dry conditions) could contribute to adhesion [39]. For example, higher friction forces especially between the furrowed sucker margins and rough surfaces could improve the leech attachment by counteracting their inward movement in pull-off measurements [36]. Moreover, biological adhesives or glues could play a supportive role in the temporary attachment of the anterior and posterior suction organs in a similar way as in limpets [40–42], or the involvement of a duo-gland system using both attaching and releasing agents [43].

Compared with recent studies on the arhynchobdellid leech W. pigra, the maximum attachment forces of H. verbana posterior suckers were approximately 10 times smaller [28], but analytically still high enough to even assure the attachment of freshly fed leeches to a given substrate. Nevertheless, our results show attachment forces and tenacities comparable with those of other aquatic animals employing suction adhesion like clingfish [36,37,44], although the suction organs of, e.g. Gobiesox maeandricus equally function across a broader spectrum of different surface textures compared with H. verbana [36]. None of the so far mentioned suction adhesion systems is as strong, efficient and reliable as the suction system of Octopodidae [38,45–47]. Further research on the attachment properties of the Mediterranean medicinal leech is required ranging from physical (e.g. influences of pull-off speed, water temperature and detachment angles), chemical (e.g. influences of surface properties and biological adhesives) and biological (e.g. influences of metabolic rates and stress levels) perspectives. Furthermore, depth effects and the presence of biofilms on the test substrates may alter the attachment ability [37,41,42,48]. The relevance of the leech attachment system functioning as possible role model for novel biomimetic applications could then be evaluated.

Ethics

The authors assure that all leeches were treated in accordance with the German animal welfare and ethical guidelines. Further details on the animal keeping conditions can be taken from the Methods section.

Data accessibility

The datasets supporting this article have been uploaded as part of the electronic supplementary material.

Authors' contributions

T.K. designed the study, collected the data, carried out the data analysis and the statistical analyses and drafted the manuscript. L.E. participated in the data collection and assisted the data analysis. F.G. participated in the design and coordination of the study and helped to draft the manuscript. T.S. participated in the design of the study and the statistical analyses, helped to draft the manuscript and coordinated the study. S.P. participated in the design and coordination of the study, helped to interpret the data and assisted in drafting the manuscript. All authors gave final approval for publication.

Competing interests

We have no competing interests.

Funding

Master-project paid for by institutional money.