The introduction of assisted reproduction, especially of IVF, during the 1980's, led to the development of a wide range of different sperm separation methods. Following the development of the classical swim-up method by Mahadevan & Baker [
4], more complicated techniques were developed to increase the number of motile spermatozoa even in severe andrological cases. On principle, these techniques can be differentiated in migration, density gradient centrifugation and filtration techniques. For all migration methods, the self-propelled movement of spermatozoa is an essential prerequisite, while for density gradient centrifugation and filtration techniques the methodology is based on a combination of the sperm cells' motility and their retention at phase borders and adherence to filtration matrices, respectively. The migration techniques can again be subdivided into swim-up, under-lay and migration-sedimentation methods [
8,
9]. For density gradient centrifugation, separation media like Ficoll
® [
10], Nycodenz [
11] and Percoll
® [
12,
13] including the products (IxaPrep
®, PureSperm
®, Isolate
®, SilSelect
®) have recently been introduced to replace Percoll
® [
14,
15]. The filtration methods like glass wool filtration [
16,
17] and filtration of spermatozoa on Sephadex beads [
18] and membranes [
19] are alternative techniques.
The ideal sperm separation technique should (i) be quick, easy and cost-effective, (ii) isolate as much motile spermatozoa as possible, (iii) not cause sperm damage or non-physiological alterations of the separated sperm cells, (iv) eliminate dead spermatozoa and other cells, including leukocytes and bacteria, (v) eliminate toxic or bioactive substances like decapacitation factors or reactive oxygen species (ROS), and (vi) allow processing of larger volumes of ejaculates. Since none of the methods available meets all these requirements, a variety of sperm separation techniques is mandatory in clinical practice to obtain an optimal yield of functionally competent spermatozoa for insemination purposes. Depending on the ejaculate quality, these methods have different efficiency and areas of use. In the conventional swim-up technique, functional spermatozoa can come into close cell-to-cell contact with defective sperm or leukocytes by centrifugation, thus causing massive oxidative damages of the sperm plasma membrane by ROS and consequently of sperm functions [
20]. Therefore, the quality of the ejaculates has direct consequences on the choice of a sperm separation method.
With regard to the possible risk of a seroconversion in women and in the offspring after performing ART with spermatozoa from a HIV-positive man initial concerns arose as HIV viruses reportedly can bind to and penetrate into spermatozoa [
21,
22]. Later research, however, showed that HIV-1 genomes found to an extend of 18% in seminal cells samples decreased to undetectable levels following a combined density gradient centrifugation with a subsequent swim-up [
23]. No seroconversion of the virus could be observed after fertilization of oocytes by IVF-ICSI [
24‐
26]. Other recent research [
27] showed that HIV RNA and DNA could be detected in separated spermatozoa even in treated patients. Thus, viral validation of separated spermatozoa is necessary and should be performed even in treated patients. Additionally, only HIV-tested maternal serum or commercially available serum albumin, which is HIV-free due to its processing, should be used as protein supplement for culture media.
In order to give an overview, only the four most common techniques classical swim-up, migration-sedimentation, density gradient centrifugation and glass wool filtration are discussed.
Swim-up procedure
Apart from a simple wash and subsequent resuspension of the male germ cells, the swim-up from a washed pellet is the oldest and most commonly used sperm separation method. Originally described by Mahadevan and Baker [
4], this method is still used largely in IVF laboratories around the world. Although its use among the male factor infertility group is very limited, the swim-up is still the standard technique for patients with normozoospermia and female infertility. Excellent fertilization rates were reported when these sperm preparations were used to inseminate human oocytes
in vitro. However, as the indications for IVF were expanded beyond simple tubal factor cases to couples with idiopathic infertility and, ultimately, to male factor cases, the problem of fertilization failure appeared [
28‐
30].
The methodology of this conventional swim-up is based on the active movement of spermatozoa from the pre-washed cell pellet into an overlaying medium. Typically, the incubation time is 60 minutes. This technique is distinguished by a very high percentage (>90%) of motile sperm, preferred enrichment of morphologically normal spermatozoa as well as the absence of other cells and debris. Considering that the efficiency of the technique is based on the surface of the cell pellet and the initial sperm motility in the ejaculate, the yield of motile spermatozoa is limited. Many layers of cells in the pellet may cause potentially motile spermatozoa in the lower levels of the pellet never to reach the interface with the culture medium layer. In addition, a significant decrease in the percentage of normally chromatin-condensed spermatozoa has been reported after the swim-up procedure [
31]. Another major disadvantage of this technique is the fact that for its use spermatozoa are pelleted, thus coming into close cell-to-cell contact with each other, cell debris and leukocytes, which are known to produce very high levels of reactive oxygen species (ROS) [
32]. Due to the extraordinary high amount of poly-unsaturated fatty acids in the sperm's plasma membranes [
33], these ROS cause lipid peroxidation and therefore a dramatic decrease in sperm functions, including motility [
34]. Overall, although many men's spermatozoa may not be impaired to the extent of inhibiting fertilization, some couples' chances of successful IVF will certainly be compromised. It is therefore not reasonable to continue and to use a technique, such as swim-up from pelleted semen with the inherent potential to cause irrevocable damage to spermatozoa prejudicial to a desired functional endpoint. Eventually, this knowledge led to the development of other more gentle sperm separation methods that also allow a higher recovery of motile and functional spermatozoa. The advantages and disadvantages of the conventional swim-up are summarized in table
1.
Table 1
Advantages and disadvantages of the conventional swim-up method.
- easy to perform | - restricted to ejaculates with high sperm count and motility |
- very cost-effective | |
- usually recovery of a very clean fraction of highly motile spermatozoa | - low yield |
| - spermatozoa can be massively damaged by reactive oxygen species |
| - significant decrease of the percentage of normally chromatin-condensed spermatozoa |
An attempt to overcome at least the problems caused by ROS, the "swim-up" can be performed directly from the liquefied semen. During this procedure, several aliquots of liquefied semen are taken from a sample and placed in tubes underneath an overlay of culture medium. Round-bottom tubes or 4-well dishes should be used to optimize the surface area of the interface between the semen layer and the culture medium. The tubes may also be prepared by gently layering culture medium over the liquefied semen. The placing of semen underneath the culture medium, however, provides a much cleaner interface zone. A maximum recovery is obtained by using multiple tubes with small volumes of semen per tube, thus maximizing the combined total interface area between semen and culture medium. Mortimer [
35] suggested the use of 250 μl semen and 500 to 600 μl culture medium per tube. After the incubation period, which is typically between 30 and 60 minutes, at 37°C, most of the upper culture medium layer is removed. This should be done with caution, working from the upper meniscus downwards, using a sterile pipette. Typically, 75 or 80% of the culture medium layer are removed and eventually combined, taking great care not to aspirate directly from the interface region. This procedure will also increase the total number of recovered spermatozoa, which can then also be used for ICSI [
36].
Other swim-up methods include the swim-up of spermatozoa in a specially supplemented medium. Such substances can be SpermSelect™, which is a highly purified preparation of hyaluronic acid (Pharmacia, Uppsala, Sweden) with an average molecular weight of 3,000 kDa that was used at a final concentration of 1 mg/ml in culture medium. Compared with the traditional swim-up from a washed pellet [
37], a swim-up directly from semen into a hyaluronic acid solution gave a significantly higher percentage of motile spermatozoa and, ultimately, the achievement of a higher pregnancy rate in a clinical IVF program [
38]. However, highly purified hyaluronic acid is expensive and it has been shown to increase the calcium-influx into spermatozoa and therefore induce acrosome reaction [
39]. Elevated local concentrations of hyaluronic acid in the cumulus oophorus have been shown to contribute to the acrosome reaction [
40]. Thus, it seems rather questionable whether this substance is favourable for IVF. In addition, whether these improved results in sperm motility were specifically due to the use of the hyaluronate or to the use of a method, which did not involve the initial pelletting of unselected spermatozoa, has not been ascertained. On the other hand, however, hyaluronic acid has been regarded as an effective alternative to test sperm penetration into human cervical mucus [
41‐
43].
Migration-sedimentation
A more sophisticated sperm separation technique is migration-sedimentation, which was developed by Tea et al. [
8]. Principally, this method is a swim-up technique combined with a sedimentation step. Special glass or plastic tubes with an inner cone are used. In contrast to the conventional swim-up procedure, spermatozoa swim up directly from liquefied semen into the supernatant medium and subsequently sediment in that inner cone within an hour's time. Thus, this method is a highly gentle technique, especially if compared with methods that require centrifugation steps before the sperm separation like the conventional swim-up. In the original version, a fraction of highly motile and functionally competent spermatozoa can be obtained. Unfortunately, the yield is very low and therefore the original method did not find wide acceptance for IVF. The advantages and disadvantages of this method are summarized in table
2. Recently, Zavos et al. [
9] proposed the use of a multi-chamber tube to retrieve functional spermatozoa for assisted reproductive techniques by means of a swim-up and sedimentation method. The assessment of its usefulness for IVF/ICSI, however, is still to be awaited.
Table 2
Advantages and disadvantages of the original migration-sedimentation method according to Tea et al. (1984).
- usually very clean fraction of highly motile spermatozoa | - the original method is restricted to ejaculates of high sperm count and good motility |
- reactive oxygen species are reduced | - the original method has a very low recovery rate |
- very gentle separation method | |
| - special glass or plastic tubes are required |
| - tubes are more expensive and relatively sensitive |
| - for repeated use in IVF, glass and plastic tubes must be sterilized |
Sánchez et al. [
44] showed that after concentration of sperm cells in the ejaculate, even in cases with severe oligo- and/or asthenozoospermia a sufficient number of motile spermatozoa can be isolated for intracytoplasmic sperm injection (ICSI) after 2–3 hours of incubation. Compared with the density gradient centrifugation, these authors also demonstrated significantly better results for progressive motility, normal sperm morphology, chromatin condensation and reduction of the percentage of dead spermatozoa as determined by the eosin test. In addition, since spermatozoa isolated by this modified migration-sedimentation technique stick markedly less to the glass surface of the injection pipettes, this method has even an advantage over the density gradient centrifugation, which is normally used for these cases. In this regard, the side migration technique that was recently proposed by Hinting and Lunardhi [
45] is another interesting approach to obtain motile spermatozoa from very poor quality semen for ICSI as it also yielded better sperm quality.
Density gradient centrifugation
The typical methodology for the density gradient centrifugation comprised continuous [
46] or discontinuous gradients [
47]. With continuous gradients, there is a gradual increase in density from the top of the gradient to its bottom, whereas the layers of a discontinuous gradient show clear boundaries between each other. The ejaculate is placed on top of the density media with higher density and is then centrifuged for 15–30 minutes. During this procedure, all cells reach the semen sediment. However, highly motile spermatozoa move actively in the direction of the sedimentation gradient and can therefore penetrate the boundary quicker than poorly motile or immotile cells, thus, highly motile sperm cells are enriched in the soft pellet at the bottom.
A wide variety of methods using the principle of density gradient centrifugation to fractionate subpopulations of spermatozoa has been described in the literature. Ficoll
® has initially been used as gradient material for preparing spermatozoa [
48], but by far the most widely used substance for all methods of assisted reproduction (IUI, GIFT, IVF, ICSI, etc.) have been the polyvinylpyrrolidone (PVP)-coated silica particles Percoll
®. Normal sperm function in terms of sperm fertilizing ability as assessed in the zona pellucida-free hamster egg penetration test [
20,
49], as well as in human IVF [
13] and ICSI [
50] was observed. In October 1996, Percoll
® has been withdrawn from the market for clinical use in assisted reproduction [
51]. This was because of the risk of contaminations with endotoxins [
52‐
54], possible membrane alteration [
55,
56] and inflammatory responses that could be induced by the insemination of sperm populations contaminated with Percoll
®. In addition, Percoll
® adheres to the sperm membranes [
57] and might alter them by removing coating envelopes [
58]. Therefore, intensive washing of the spermatozoa after sperm separation with Percoll
® was recommended [
55]. This requires additional centrifugation and can again be detrimental to the spermatozoa because of the action of reactive oxygen species [
20].
Another commercial product known as Nycodenz (Nyegaard & Co., Oslo, Norway) was also used as a density gradient material for preparing human spermatozoa. Nycodenz is the same molecule, iohexol, as used in the X-ray contrast medium Omni-paque. Studies revealed a low incidence of adverse reactions during angiography [
11]. Both continuous and discontinuous Nycodenz gradients were evaluated, of which a four-layer discontinuous gradient was found to produce populations of highly motile spermatozoa with better yields and survival than either swim-up or Percoll
® gradients from oligozoospermic and asthenozoospermic semen samples [
11,
59]. Compared with the conventional swim-up procedure from a pelleted sperm population, the use of Nycodenz also seems to be superior regarding sperm penetration into zona-free hamster eggs [
60]. Thus, this technique has clearly great potential in the preparation of motile spermatozoa from poor quality semen for IVF use and warrants further investigation.
Other replacement products for Percoll
® that were introduced into the market from the mid nineties and more commonly used in assisted reproduction are IxaPrep
® (MediCult, Copenhagen, Denmark), SilSelect
® (FertiPro N.V., Beernem, Belgium), PureSperm
® (NidaCon Laboratories AB, Gothenburg, Sweden) or ISolate
® (Irvine Scientific, Santa Ana, CA, USA). In contrast to Percoll
®, which is a PVP-coated silica that can have deleterious effects on sperm membranes [
56], all these replacement products contain silane-coated silica particles, are adjusted for the osmolarity with polysucrose and have very low toxicity. All these replacement products are non-irritating and are approved for human
in vivo use. The results of sperm preparation using these new products compared with Percoll
® regarding recovery rate, motility, viability, normal sperm morphology and velocity parameters like VAP or VCL vary considerably among different working groups. While Claassens et al. [
61] and Söderlund and Lundin [
62] did not find differences in the recovery rate between Percoll
® and PureSperm
®, Chen and Bongso [
63], depending on the number of layers included for the density gradient, reported significantly higher values for PureSperm
®. For IxaPrep
® it is even more confusing because Yang et al. [
64] found no difference to Percoll
®, while Makkar et al. [
14] found the replacement substance more effective. On the other hand, McCann and Chantler [
65] as well as Ding et al. [
66] found Percoll
® superior. These authors attribute the better sperm quality obtained after the IxaPrep
® preparation to a significantly decreased production of nitric oxide, which is regarded as sperm toxicant that reduces motility [
67]. This could be due to an activation of guanylyl cyclase, thus increasing cGMP production, which inhibits sperm motility [
68]. On the other hand, nitric oxide is also known to be a physiologic mediator for vasodilatation, immunosuppression, neurotransmission and cytotoxicity [
69‐
72].
Regarding the other parameters such as motility, viability, normal sperm morphology or velocity parameters like VAP, the data currently available also vary considerably among different working groups. The reason for this can be attributed to the different conditions of the sperm separation, e.g. volume of semen to be separated, g-force, centrifugation time or the number of layers of the gradient, and reflect the important role of the methodology. Moreover, this also shows that these data cannot be directly compared. Overall, the Percoll®-replacement products are good and reasonable alternatives, and this not only for the fact that Percoll® is no longer allowed to be used for clinical purposes in assisted reproduction.
Glass wool filtration
During glass wool filtration, which has already been described by Paulson & Polakoski in 1977 [
17], motile spermatozoa are separated from immotile sperm cells by means of densely packed glass wool fibres. The principle of this sperm separation technique lies in both the self-propelled movement of the spermatozoa and the filtration effect of the glass wool. The success of this method is directly linked to the kind of glass wool used [
73]. Thus, factors like the chemical nature of the glass (i.e. borate glass, silicate glass or quartz glass), the surface structure and charge of the glass wool, thickness of the glass wool fibres or the pore size of the filter have to be taken into consideration. In clinical practice, the glass wool, code number 112, from Manville Fiber Glass Corp. (Denver, CO, USA) or SpermFertil
® columns from Mello (Holzhausen, Germany) have been tested extensively. Potential risks of the technique such as damages of the spermatozoa or the occurrence of glass wool fragments in the filtrate essentially depend on the kind of glass wool used and on the intensity of the washing prior to the filtration.
Compared with the swim-up or migration-sedimentation, glass wool filtration, just as density gradient centrifugation, is a technique that uses the whole volume of the ejaculate and therefore yields a significantly higher total number of motile spermatozoa [
31,
74]. Thus, it can also be used for patients with oligo- and/or asthenozoospermia [
74]. Like density gradient centrifugation, glass wool filtration also provides the advantage that the sperm separation can directly be performed from the ejaculate. Only after the separation of the functional spermatozoa from the immotile ones, leukocytes and debris, a centrifugation step will be necessary to remove the seminal plasma. This is an important aspect as this procedure reduces cellular damage by reactive oxygen species. The advantages and disadvantages of this method are summarized in table
4.
Table 4
Advantages and disadvantages of the glass wool filtration.
- simple to perform | - a bit more expensive |
- normally, recovery of spermatozoa with good motility | - the filtrate is not as clean as it is with other sperm separation methods |
- spermatozoa from ejaculates with a very low sperm density can be separated | - remnants of debris are still present |
- good yield | |
- leukocytes are eliminated to a large extent | |
- reactive oxygen species are significantly reduced | |
By means of glass wool filtration, it is even possible to prepare motile spermatozoa from patients with retrograde ejaculation (Henkel et al., unpublished). In these cases, the procedure includes adjustment of the osmolarity of the patient's urine to values of about 350 mOsmol/kg by drinking water. Prior to the ejaculation, the patients are requested to urinate most of the urine in the bladder. The small amount of antegrade-produced ejaculate is collected in a plastic beaker, while the retrograde fraction of the ejaculate needs to be urinated immediately into a jar with 50 ml culture medium containing human serum albumin to dilute the urine, Finally, the urine/medium mixture has to be centrifuged, resuspended in 3 to 4 ml of fresh medium and filtrated on the glass wool column. As constituents of the urine can damage the spermatozoa, a speedy work-up of such ejaculates is mandatory.
In addition to the separation of spermatozoa, glass wool filtration has been shown to eliminate leukocytes to an extent of up to 90% [
75]. Since leukocytes are frequent even in normal ejaculates [
76] and produce 100-times more ROS than spermatozoa [
32], this effect significantly contributes to a reduction of free radicals in the ejaculate [
73,
75]. This is of paramount importance for the functionality of spermatozoa because the male germ cells are particularly susceptible to oxidation by ROS because of their extraordinary high content of polyunsaturated fatty acids in their plasma membrane [
33,
77,
78].
Another clinically interesting aspect related to glass wool filtration is chromatin condensation, which has repeatedly been shown to be predictive of fertilization
in vitro [
79‐
81]. Glass wool filtration [
31] like the density gradient centrifugation with PureSperm
® [
82] or the migration-sedimentation technique [
44] significantly selects normally chromatin-condensed spermatozoa, while conventional swim-up or Percoll
®-centrifugation decrease this sperm parameter. As human sperm chromatin condensation follows a seasonal rhythm, which even shows a shift of about half a year on the southern hemisphere [
83], this might have a clinical impact on the results in IVF. Should a patient be examined in winter when the quality of sperm chromatin condensation is high [
83] and referred to IVF in summer when the percentage of normally chromatin-condensed spermatozoa is significantly lower, IVF for this patient might fail. Thus, for these patients a sperm separation by means of glass wool filtration, PureSperm
® or migration-sedimentation might be beneficial.
Sephadex columns
In the early nineties sperm separation by means of Sephadex beads emerged [
18] and a commercial sperm separation kit based on this principle (SpermPrep
®) has become available. Compared to migration-sedimentation and swim-up from pelleted semen it produced significantly higher yields [
86]. Moreover, morphologically normal sperm cells could be enriched in the filtrate after SpermPrep
® separation as well as significantly higher pregnancy rates for intrauterine insemination as compared with the conventional swim-up method [
87]. In a comparison between SpermPrep
® method and Percoll
® centrifugation, Percoll
® separated spermatozoa showed a significantly higher percentage of normally chromatin-condensed and morphologically normal spermatozoa [
88]. However, the fertilization rates reported by these authors were similar. López et al. [
89] used a prepacked PD-10 column containing Sephadex G-25 particles (Pharmacia Biotechnology, Uppsala, Sweden), which is normally used to desalt proteins in solutions, to separate human spermatozoa and compared the results with the SpermPrep
® method and Percoll
® centrifugation. The PD-10 column and density gradient centrifugation in Percoll
® yielded a comparable number of spermatozoa and showed similar percentages of morphologically normal spermatozoa after sperm separation. On the contrary, the SpermPrep
® method resulted in significantly lower values of sperm count and morphology.
Transmembrane migration
Another alternative sperm separation technique that was also developed in the late eighties is migration/filtration of motile spermatozoa through a Nuclepore membrane filter. These filters are unusual because their pores are cylindrical and at right angles to the plane of the membrane [
90]. The spermatozoa, therefore, have straight channels to swim through the membrane. Unfortunately, these membranes had a very low ratio of the total cross-sectional area of the pores to the overall membrane area. Consequently, the yield is extremely low. Primarily, this method was used for testing the motility of sperm populations treated with various pharmacological agents, but not as a preparation method for assisted reproduction [
91].
Another approach of separating viable human spermatozoa by means of membranes was undertaken by Agarwal et al. [
19] using a membrane which has been developed for selective removal of leucocytes (L4 membrane). Besides a significant increase of motility, ejaculates filtered through this membrane have been shown to contain fewer leukocytes. This fact is, of course, of importance in those cases that have increased numbers of leucocytes in the ejaculate as a result of infections. Moreover, this membrane seems to be selective for spermatozoa with normal membrane integrity [
92,
93] and sperm producing low amounts of reactive oxygen species [
94]. However, despite these advantages of the membrane it has never come into practical clinical use for human assisted reproduction.