Hum. Reprod. Advance Access originally published online on September 30, 2005
Human Reproduction 2006 21(2):477-483; doi:10.1093/humrep/dei323
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
IVF within microfluidic channels requires lower total numbers and lower concentrations of sperm
1 Department of Urology, 2 Department of Biomedical Engineering, 3 Department of Macromolecular Science and Engineering, 4 Department of Obstetrics and Gynecology and 5 Department of Molecular and Integrated Physiology, University of Michigan, Ann Arbor, MI 48109, USA
6 To whom correspondence should be addressed at: 6428 Medical Sciences I, 1301 E Catherine St, Ann Arbor, MI 48109-0617, USA. E-mail: smithgd{at}med.umich.edu
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResultsDiscussionAcknowledgementsReferences |
|---|
BACKGROUND: Microfluidic technology has been utilized in numerous biological applications specifically for miniaturization and simplification of laboratory techniques. We sought to apply microfluidic technology to murine IVF. METHODS: Microfluidic devices measuring 500 µm wide, 180 µm deep, and 2.25 cm in length were designed and fabricated using poly(dimethylsiloxane) (PDMS). Controls were standard centre-well culture dishes with 500 µl of media, half of which also contained PDMS as a material control. Denuded mouse oocytes were placed into microchannels or centre-well dish controls in groups of 10, then co-incubated overnight with epididymal mouse sperm at various concentrations. Fertilization was assessed and Fisher’s exact test was used for statistical analysis (P < 0.05 significant). RESULTS: Fertilization rates between the two control groups (42%, no PDMS; 41%, with PDMS; not significant) were similar. Fertilization rates for denuded oocytes at standard mouse insemination sperm concentration (1°106 sperm/ml) was poorer in microchannels (12%) than controls (43%; P < 0.001). As insemination concentrations decreased, fertilization rates improved in microchannels with a plateau between 8°104 and 2°104 sperm/ml (4000–1000 total sperm). At these concentrations, combined fertilization rate for denuded oocytes was significantly higher in microchannels than centre-well dishes (27 versus 10%, respectively; P < 0.001), and was not significantly different from corresponding controls with a sperm concentration of 1°106 (37%; P = 0.06). CONCLUSIONS: Murine IVF can be conducted successfully within microfluidic channels. Lower total numbers and concentrations of sperm are required. Microfluidic devices may ultimately be useful in clinical IVF.
Key words: IVF/microfluidics/murine/sperm
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAcknowledgementsReferences |
|---|
Primary infertility affects up to 10–15% of couples, with roughly half of cases attributable to male factors such as oligozoospermia (Mosher and Bachrach, 1996
). Utilization of IVF and related medical procedures, categorized as assisted reproductive technologies, increases substantially each year. In 2001, there were 384 assisted reproduction clinics in the USA conducting a total of 107 587 treatment cycles with 40 687 babies born (Wright et al., 2004).
At its most basic level, IVF involves co-incubation of oocytes with an appropriate number of sperm in an easily controllable and physiologically amenable environment. Current techniques predominantly utilize culture dishes, culture tubes, or microdrops under oil. Insemination is based on concentration, where the total number of sperm needed for fertilization is dependent on system volume. Media volumes may range from 20 µl for microdrops to 1 ml in tubes and dishes, with a minimum of 500 000 total motile sperm generally considered suitable for standard IVF (Trounson and Gardner, 2000). Such techniques are often inadequate for severe oligozoospermic males. Alternatively, ICSI, which utilizes direct injection of a single sperm into an oocyte, has revolutionized the treatment of male factor infertility (Devroey and Van, 2004).
Despite ICSI’s widespread use, concerns remain for several technical and theoretical reasons. Up to 7% of oocytes may lyse secondary to injection and/or micromanipulation, which can be a serious issue in cases where a limited supply of oocytes is available for fertilization (Dumoulin et al., 2001; Rienzi et al., 2001). In addition, studies have demonstrated spindle damage and aneuploidy in early stage embryos following ICSI (Burrello et al., 2003; Lathi and Milki, 2004). Gamete micromanipulation requires a skilled and specially trained operator, and the process is labour intensive. On a theoretical level, ICSI bypasses all natural selection as sperm are artificially inserted into the oocyte to achieve fertilization (Tournaye, 2003). Ultimately, the long-term significance of such manipulation remains to be determined.
A new bioengineering technology, with ever-increasing applications, may hold significant promise for assisted reproduction. Microfluidics originated with attempts to miniaturize chemical and biological analysis devices in the laboratory and has now resulted in an ever-growing list of designs and uses (Kricka, 1998; Verpoorte, 2002). Current devices are often referred to as ‘laboratory-on-a-chip’, or micro-total analysis systems (µTAS), and function by allowing biological or chemical processes and interactions to occur as fluid flows within miniature channels and chambers. Recently, investigators have adapted microfluidic technology to assisted reproduction applications. Schuster et al. (2003) developed a microfluidic device utilizing parallel laminar flow streams, yielding an atraumatic method of obtaining motile sperm of increased normal morphology from unprocessed normal and poor quality semen. Beebe and co-workers have developed microfluidic techniques for cumulus cell removal from zygotes (Zeringue and Beebe, 2004), as well as a microchannel embryo culture system known as a Micro Embryo Culture Chip (MECC; Walters et al., 2004).
Microfluidics may be particularly suitable for IVF for a number of reasons (Suh et al., 2003). The microenvironment of a microchannel more closely resembles in vivo fertilization conditions than a culture dish or microdrop. Microfluidic channels allow for non-turbulent bathing of gametes with fresh media throughout insemination and co-incubation. Sperm–oocyte interactions occur in an active environment, rather than the static conditions present in a culture dish or droplet. In addition, via laminar flow, sperm can be predictably delivered to each oocyte within the microchannel, eliminating the randomness of sperm–oocyte interaction. In a culture dish, sperm may randomly travel in any direction, thereby relying on random sperm movement towards the oocytes; however, in a microchannel environment, sperm movement is limited by the direction of flow, allowing for active transport to the oocytes. Finally, microchannel environments utilize extremely small volumes of media, theoretically requiring fewer sperm to achieve insemination concentrations equal to standard IVF with larger volumes (Table I).
|
Based on these reasons, we sought to design, develop, and test a microfluidic system for in vitro microchannel insemination, co-incubation, and fertilization. Our primary goal was to establish feasibility of in vitro microchannel fertilization, using the mouse. Second, we hypothesized that microchannel IVF would utilize lower total numbers of sperm compared to standard IVF based on the small volumes of media present during co-incubation.
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAcknowledgementsReferences |
|---|
Design and development
Initial microfluidic systems were fabricated using materials and techniques common in microelectronics, the industry that inspired them; however, cost and complex production processes made this a less than ideal method (McDonald et al., 2000
; Boone et al., 2002). For our device, we utilized a polymer known as poly(dimethylsiloxane) (PDMS) which possesses numerous characteristics specifically suitable for biological use and easy production. It is non-toxic, transparent, insulating, and permeable to gases. In addition, from a fabrication standpoint PDMS permits sub-micrometre fidelity with moulding, cures at low temperatures, and easily seals reversibly or irreversibly to itself and a host of other materials (McDonald and Whitesides, 2002). Fabrication was done using techniques based on soft lithography (McDonald et al., 2000). The first step, rapid prototyping, was used to generate a master on a silicon wafer which served as the pattern mould for the actual device. A computer-assisted design (CAD) program was used to generate the structure and size of the microchannels, which were then printed onto a transparency using a high-resolution commercial image setter. This transparency acted as the photomask for subsequent photolithography, and allowed for generation of features and channels as small as 20 µm. Photolithography using the transparency was conducted on a photocurable epoxy bonded onto a silicon wafer at the desired channel height. Liquid pre-polymer PDMS was poured onto the wafer and allowed to cure, thereby generating a negative pattern of the master in PDMS. In other words, ridges on the silicon wafer form the microchannels within the PDMS device. Final processing involved affixing the PDMS device to additional PDMS or glass with plasma oxidation for 1 min using a Plasma Prep II Chamber (SPI Supplies; West Chester, PA, USA), which also made the channel surfaces hydrophilic resulting in improved fluid flow. Functionally, multiple variables were considered in construct of this microfluidic device for the purpose of in vitro co-incubation of sperm and oocytes with the ultimate goal of fertilization. First, an atraumatic and ergonomically simple method of loading media and gametes needed to be developed. Second, microchannel dimensions had to be carefully considered to allow for optimal flow without compromising the reliable delivery of sperm to oocytes. Dimensions excessively larger than a standard oocyte could result in a decreased chance of each sperm within the channel flow interacting with the oocyte. Conversely, excessively small channel dimensions could increase clogging and interruption of flow. In addition, a method of holding an oocyte in a specific position within the microchannel was necessary to prevent migration of the oocyte as sperm and media flow within the channel. Ideally, such a method had to be easy to manufacture and incorporate into a device. Finally, a reliable means of adjusting fluid flow was paramount for the manipulation of gametes once they were within the microchannel itself.
The final design (Figure 1) consisted of a funnelled inlet media well for media and gamete loading, a microchannel measuring 500 µm wide and 180 µm high, a three-dimensional barrier gate which permitted uninterrupted media and sperm flow without oocyte deformation or passage, and a negative pressure and gravity-driven flow system manipulated by fine silicone tubing sealed to the outlet well. Oocytes were placed within the device by manual loading via the funnelled inlet media well using a finely pulled glass pipette. They were then actively transported along the microchannel with gentle negative-pressure-driven flow using a mouth-pipette applied to the outlet tubing and ‘parked’ at the microchannel gate. This gate spanned the width of the microchannel, with each subdivision measuring 30 µm high by 30 µm wide, thereby obstructing movement of the oocyte. Sperm were then also loaded via the funnelled inlet media well. Gravity-driven flow using a siphon technique was initiated by mouth-pipette on the outlet tubing. This, combined with native sperm movement, allowed for delivery of sperm to oocytes for gamete interaction. Bulk flow rates were ~0.001 µl/s, followed by incubation under static conditions to maximize sperm residence time within the channel, using experience from previous sperm manipulation devices (Cho et al., 2003). Co-incubation therefore occurred within microchannels, and sperm had an opportunity to interact with oocytes as they travelled along the microchannel before passage through the gate. The volume of the microchannel itself was 2 µl, with a total system volume, including the inlet well, of 50 µl.
|
IVF testing
All animals used in these experiments were treated in accordance with approved protocols by the Animal Care and Use Committee of the University of Michigan. Oocytes were collected from 8 week old female B6C3F1 mice after intraperitoneal injection of 10 IU equine chorionic gonadotrophin on day 0 and 10 IU HCG 48 h later. Sixteen to 18 h following HCG injection, mice were killed by cranial/cervical dislocation. Each oviduct was surgically removed, and under magnification, the cumulus–oocyte masses were harvested into HEPES-buffered human tubal fluid (HTF-H; Irvine Scientific, Irvine, CA, USA) with 0.2% bovine serum albumin (BSA; Fisher Scientific, Pittsburgh, PA, USA; HTF-H/0.2% BSA). Oocytes were then individually denuded of cumulus cells by a brief incubation within 0.03% hyaluronidase (Sigma, St Louis, MO, USA) in HTF-H and gentle micropipette manipulation.
Sperm were collected from 8 week old male B6C3F1 mice killed by cranial/cervical dislocation. Each epididymis was surgically removed, the cauda portion was isolated then gently minced with fine scissors into HTF/2% BSA. After allowing sperm to swim out into media during a 20 min incubation, media were separated from epididymal tissue. Sperm for insemination were then obtained after a 1 h swim-up in a humidified 5% CO2 incubator at 37°C. Sperm concentration and percentage motility were obtained with a Makler counting chamber, using the average of two measurements. Dilution to various required concentrations was performed as needed.
Insemination, co-incubation and fertilization were performed using HTF/2% BSA in all cases. Microfluidic devices were pre-filled with HTF/2% BSA media, irradiated under UV light for 15 min, and equilibrated overnight in a humidified 5% CO2 incubator at 37°C. Standard insemination controls consisted of centre-well culture dishes pre-filled with HTF/2% BSA and exposed to UV light for 15 min, and equilibrated overnight in a humidified 5% CO2 incubator at 37°C side-by-side with our microfluidic devices. In addition, a device material control group was included in initial phases of testing. These were identically and concurrently prepared standard insemination centre-well culture dishes with a 3 mm cube of PDMS placed in the centre media wells.
Freshly harvested denuded metaphase II oocytes were loaded into microchannel devices and culture dishes using finely pulled glass pipettes, adding 10 oocytes per device or dish. Within microfluidic devices, oocytes were allowed to flow into the channel by gravity, then were actively transported to the three-dimensional barrier gate. Swim-up sperm were then added to either culture dishes or the inlet loading well of the microfluidic devices at the desired insemination concentration. Co-incubation of sperm and oocytes was conducted overnight in a humidified 5% CO2 incubator at 37°C. The total media volume for microfluidic insemination was 50 µl compared to a control centre-well insemination total volume of 500 µl. Fertilization was assessed 18–20 h after insemination and was strictly defined as the occurrence of early cleavage with development to the 2-cell stage. Fertilization rates derived from various treatment regimens were statstically compared using Fisher’s exact test, with P < 0.05 considered significantly different.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAcknowledgementsReferences |
|---|
All initial experiments included the device material control group in order to ensure that PDMS had no effects which confounded standard mouse IVF. No significant difference was seen in fertilization rates between the two control groups. For standard insemination at 1x106 sperm/ml (Hogan et al., 1994
) without PDMS exposure, the fertilization rate of denuded oocytes was 42% (n = 173) compared to 41% (n = 157) for standard insemination with exposure to the 3 mm cube of PDMS throughout co-incubation. Thereafter, we eliminated the PDMS material control group to maximize comparison testing between standard and microfluidic insemination. At the standard mouse sperm insemination concentrations of 1x106 sperm/ml, we achieved very poor fertilization rates of denuded oocytes with the microfluidic device (12%, n = 67) compared to centre-well culture dishes performed concurrently (43%, n = 97; Figure 2). This difference was statistically significant (P < 0.001) and was apparent early in experiments, contrary to our initial hypothesis. We therefore were required to formulate a new hypothesis. While examining microchannel co-incubation under the microscope during this phase of testing, we were encouraged by a subjective belief that increased sperm–oocyte interaction occurred compared to our centre-well controls. We postulated that this resulted in a relative increase in local sperm numbers in the vicinity of the oocyte despite a stable insemination concentration. In order to compensate for the relative increase in gametes, one could increase the dimensions of the channel; however, the corollary of this would be to keep channel characteristics the same, but decrease the concentration of sperm. We subsequently hypothesized that as insemination concentration is decreased for microfluidic co-incubation, fertilization rates improve compared to standard insemination. A secondary hypothesis was that fertilization rates with microfluidic insemination at low concentration would be similar to standard insemination at standard sperm concentrations, therefore ultimately requiring fewer sperm for fertilization.
|
Microfluidic insemination was therefore conducted at a number of sperm concentrations. An effect plateau between 8 ° 104 and 2 °104 sperm/ml (4000 to 1000 total sperm) was found (Figure 3, Table II). At these concentrations, the combined fertilization rate of denuded oocytes was significantly higher in microchannels than in centre-well dishes performed concurrently (27%, n = 312 versus 10%, n = 88, respectively; P < 0.001; Figure 4). In addition, the fertilization rate was similar to standard insemination controls at a sperm concentration of 1 ° 106 performed concurrently (37%, n = 102; P = 0.06; Figure 4).
|
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAcknowledgementsReferences |
|---|
In order for this device to contribute as a usable IVF technology, it is imperative that materials used in its construction and materials in contact with gametes during co-incubation are proven not to be deleterious to fertilization and early embryo development. Although other authors had shown that human sperm survival (Schuster et al., 2003
) and porcine and murine embryo culture (Raty et al., 2004; Walters et al., 2004) were not adversely affected by PDMS, this fact had not yet been definitively demonstrated and published specifically in regards to mouse gametes and IVF. Fertilization rates between our centre-well culture dish control groups with and without the 3 mm PDMS cube were nearly identical. This confirms that murine IVF rates are not influenced by the presence of PDMS. It should also be stressed that the cubes of PDMS used for control testing were taken through the entire device manufacturing process, including rapid prototyping and plasma oxidation, side-by-side with the microfluidic devices. Therefore, materials and manufacturing processes used for this microfluidic insemination device do not influence fertilization rates compared to standard centre-well culture dish controls.
Centre-well dishes with 500 µl of media remain the most straightforward method to conduct murine IVF, and could be considered the gold standard (Hogan et al., 1994). We did not perform experiments using any other methods or techniques (i.e. microdrops under oil) in order to compare microfluidic IVF with a simple and universally accepted control technique for mouse IVF. The corollary is that in experienced hands, specialized methods of mouse IVF may produce improved results compared with our more straightforward, universal technique (Goossens et al., 2003). The intent of our experiment, however, was not to illustrate and compare outcomes across a range of techniques, but rather to illustrate the feasibility of microfluidic IVF, the potential benefits of microfluidic IVF, and provide some insight and hypotheses into future directions for improvement and refinement of microfluidics in their application to IVF.
Results of the microfluidic insemination device at standard insemination concentrations were discouraging when compared with centre-well controls. We wish to propose several theories in regards to this poor fertilization with standard concentrations and microfluidic insemination. For traditional IVF, optimal in vitro insemination concentrations were established within a limited range of media volume. We believed that this optimal concentration would be constant regardless of co-incubation format. In microfluidic insemination and co-incubation, however, the small media volumes, active fluid transport and implementation of gates which alter channel geometry may shift this optimum by influencing the chemistry and physics of the insemination environment. For example, gametes use large amounts of metabolic substrates and create many byproducts (Dietl and Rauth, 1989; Hickman et al., 2002). Delivery, diffusion and dissipation of these metabolic substrates would theoretically be altered in a constant flow microscale environment, perhaps prohibitively so in the setting of relative increases in sperm–oocyte interaction. Additionally, dimensions of the microchannel result in a limited amount of space around the oocyte in two of three dimensions compared to centre-well dishes. Similar theories regarding a microchannel environment have been advanced in regards to embryo culture in microchannels (Beebe et al., 2002). Finally, sperm-oocyte interaction is increased compared with static systems due to physical factors such as active convective transport of sperm to the oocyte and by partial blockage of sperm passage past the grate, which may increase local sperm concentration in vicinity of the oocyte.
As previously mentioned, standard concentrations of 1x106 sperm/ml may result in a relative increase in sperm–oocyte interactions compared to centre-well controls. In order to compensate for this relative increase, one could increase the dimensions of the channel; however, the corollary of this would be to keep channel characteristics the same, but decrease the concentration of sperm. This line of thinking led to our secondary hypothesis and what we feel is the most intriguing development to emerge from testing of this microfluidic IVF device. Use of decreased insemination concentrations, combined with the low volumes of the microchannel, led to a significant overall decrease in total sperm numbers required for fertilization. As stated in the original hypothesis, based on mathematic principles, decreasing volume with a constant concentration decreases the actual number of sperm. Factoring a decreased concentration to the reduction from decreased volume allows for an even lower actual number of sperm utilized (Table III).
|
What factors present within the microfluidic IVF device would allow for fertilization at these decreased concentrations? We propose that the very characteristics which originally inspired interest in microfluidic IVF can be used to explain these results. As previously hypothesized, the microenvironment of a microchannel may more closely resemble in vivo conditions of fertilization compared to a culture dish or microdrop. This environment may therefore bestow conditions more suitable for efficient sperm–oocyte interaction, as well as increase the potential for this interaction. It has been demonstrated that only a few hundred sperm eventually reach the ampulla of the oviduct for fertilization in humans (Settlage et al., 1973; Ahlgren, 1975). By decreasing the sperm concentration and allowing interaction in the microchannel environment, optimization of sperm and oocyte co-factors may result. Microfluidic channels also allow for non-turbulent bathing of gametes with fresh media throughout insemination and co-incubation, further simulating in vivo conditions (Beebe et al., 2002).
Most importantly, we propose that fluid flow and active transport of gametes, as well as microscale dimensions and the presence of a barrier gate, may work to increase the local sperm concentration around the oocyte compared with the actual insemination concentration leading to a relative increase in sperm–oocyte interaction compared with centre-well controls. In effect, the microchannel device itself may serve as a sperm-concentrating device, funnelling sperm to the oocyte, unlike the static interactive environment of a culture dish in traditional IVF. Innate physical characteristics of this microfluidic channel combined with principles of laminar flow result in predictable delivery of sperm to oocytes, reducing the randomness of sperm–oocyte interactions present in a culture dish. As a corollary, it may be the randomness of sperm movement in culture which necessitates increased insemination concentrations for standard IVF versus microfluidic IVF. These studies have demonstrated that actual numbers of sperm per oocyte can be sharply reduced by reducing dependence on random interaction. This characteristic of microfluidics, in conjunction with small volumes, ultimately may prove decisive in success of this device by improving efficiency of co-incubation.
Previous investigators have attempted, with some success, to reduce the volume of insemination medium with various low volume vessels, although none has gained widespread acceptance. Van der Ven et al. (1989) tested the use of sterile, non-heparinized haematocrit capillary tubes (75 mm length, 0.9 mm inner diameter) for IVF in humans. Normozoospermic samples were used with standard culture tubes as controls. Volumes of 5–10 µl containing a range of 500–4000 sperm per oocyte were used in these capillary tubes. Overall fertilization rates between controls and capillary tubes was similar (78 and 66%, respectively), although slightly lower for sperm totals of 500–1000 (56%) compared to 2000–4000 total sperm (79%). Ranoux and Seibel (1990) used embryo cryopreservation straws in volumes up to 200 µl (Ranoux et al., 1988) with 2000–4000 motile sperm. Results compared favourably with controls, with 167 of 322 oocytes (51.8%) fertilized using the straw technique.
As with any new technology, design plays an important role. Improvements in loading methods, which can allow for visualization under magnification without adjustment of the microscope focus, should improve outcomes. Addition of oocytes and sperm to the device under a microscope, which requires time outside of the humidified 5% CO2 environment, has significant deleterious effects on gamete health and survival. Reduction of this time currently requires a significant learning curve. Lastly, the implementation of more sophisticated but less operator-reliant mechanisms for fluid flow should improve efficiency. Current studies are focused on a variation of the gravity-driven, horizontally oriented reservoir pumping system (Zhu et al., 2004) from the microfluidic sperm sorter developed by Cho et al. (2003) and its application to microfluidic insemination.
In conclusion, we have demonstrated that mouse IVF can be conducted successfully within microfluidic channels. Not only are lower total numbers of sperm required due to use of reduced media volumes, we have also demonstrated murine fertilization within microchannels using lower insemination concentrations, further decreasing sperm requirements. We continue to develop design improvements which will result in increased efficiency and ease of use. Such microfluidic devices should ultimately be useful in clinical IVF, not only for oligozoospermic patients but potentially as a replacement for standard insemination.
Design, development, and testing of new technology for use in biological applications can be a rigorous and complex process. The device described in this publication is in the first stages of infancy, but we are certainly encouraged and excited about the significant promise it has shown thus far. Device materials and manufacturing techniques have been reliable; ergonomically the device allows for loading and manipulation of gametes, and although our initial hypothesis was incorrect, it is encouraging that a reduction in total sperm numbers requisite for fertilization was achieved with this microfluidic technology. At present, proof-of-concept and biological possibilities of microchannel IVF have been demonstrated. Experimental support for our hypotheses and further work and modifications, from a bioengineering standpoint, offers significant promise.
|
Acknowledgements |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAcknowledgementsReferences |
|---|
The authors would like to thank Dr Carrie Cosola-Smith for critical review of this manuscript. Portions of this research have been supported by grants from the National Institutes of Health (HD 049607–01; S.T. and G.D.S.) and College of Engineering Technology Development Fund (GAP fund; S.T. and G.D.S.).
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAcknowledgementsReferences |
|---|
Ahlgren M (1975) Sperm transport to and survival in the human fallopian tube. Gynecol Invest 6,206–214.[Web of Science][Medline]
Beebe DJ, Wheeler M, Zeringue H, Walters E and Raty S (2002) Microfluidic technology for assisted reproduction. Theriogenology 57,125–135.[CrossRef][Web of Science][Medline]
Boone TD, Fan ZH, Hooper HH, Ricco AJ, Tan H and Williams SJ (2002) Plastic advances microfluidic devices. Analyt Chem 74,78A–86A.[Medline]
Burrello N, Vicari E, Shin P, Agarwal A, De PA, Grazioso C, D’Agata R and Calogero AE (2003) Lower sperm aneuploidy frequency is associated with high pregnancy rates in ICSI programmes. Hum Reprod 18,1371–1376.
Cho B, Schuster TG, Zhu X, Chang D, Smith GD and Takayama S (2003) Passively driven integrated microfluidic system for separation of motile sperm. Analyt Chem 75,1671–1675.[Medline]
Devroey P and Van SA (2004) A review of ten years experience of ICSI. Hum Reprod Update 10,19–28.
Dietl JA and Rauth G (1989) Molecular aspects of mammalian fertilization. Hum Reprod 4,869–875.
Dumoulin JM, Coonen E, Bras M, Bergers-Janssen JM, Ignoul-Vanvuchelen RC, van Wissen LC, Geraedts JP and Evers JL (2001) Embryo development and chromosomal anomalies after ICSI: effect of the injection procedure. Hum Reprod 16,306–312.
Goossens E, Frederick V, De Block G, Van Steirteghem AC and Tournaye H (2003) Reproductive capacity of sperm obtained after germ cell transplantation in a mouse model. Hum Reprod 18,1874–1880.
Hickman DL, Beebe DJ, Rodriguez-Zas SL and Wheeler MB (2002) Comparison of static and dynamic medium environments for culturing of pre-implantation mouse embryos. Comp Med 52,122–1226.[Web of Science][Medline]
Hogan B, Beddington R, Costantini F and Lacy E (1994) Manipulating the Mouse Embryo: A Laboratory Manual. 2nd edn, Cold Spring Harbor Laboratory Press, New York.
Kricka LJ (1998) Miniaturization of analytical systems. Clin Chem 44,2008–2014.
Lathi RB and Milki AA (2004) Rate of aneuploidy in miscarriages following in vitro fertilization and intracytoplasmic sperm injection. Fertil Steril 81,1270–1272.[CrossRef][Web of Science][Medline]
McDonald JC and Whitesides GM (2002) Poly(dimethylsiloxane) as a material for fabricating microfluidic devices. Accounts Chem Res 35,491–499.
McDonald JC, Duffy DC, Anderson JR, Chiu DT, Wu H, Schueller OJ and Whitesides GM (2000) Fabrication of microfluidic systems in poly(dimethylsiloxane). Electrophoresis 21,27–40.[CrossRef][Web of Science][Medline]
Mosher WD and Bachrach CA (1996) Understanding U.S. fertility: continuity and change in the National Survey of Family Growth, 1988–1995. Fam Plann Perspect 28,4–12.[CrossRef][Web of Science][Medline]
Ranoux C and Seibel MM (1990) New techniques in fertilization: intravaginal culture and microvolume straw. J In Vitro Fertil Embryo Transfer 7,6–8.[CrossRef][Web of Science][Medline]
Ranoux C, Poirot C, Foulot H, Dubuisson JB, Aubriot FX, Chevallier O and Cardone V (1988) Human egg fertilization in capillary tubes. J In Vitro Fertil Embryo Transfer 5,49–50.[CrossRef][Medline]
Raty S, Walters EM, Davis J, Zeringue H, Beebe DJ, Rodriguez-Zas SL and Wheeler MB (2004) Embryonic development in the mouse is enhanced via microchannel culture. Lab on a Chip 4,186–190.[CrossRef][Web of Science][Medline]
Rienzi L, Greco E, Ubaldi F, Iacobelli M, Martinez F and Tesarik J (2001) Laser-assisted intracytoplasmic sperm injection. Fertil Steril 76,1045–1047.[CrossRef][Web of Science][Medline]
Schuster TG, Cho B, Keller LM, Takayama S and Smith GD (2003) Isolation of motile sperm from semen samples using microfluidics. Reprod Biomed Online 7,75–81.[Medline]
Settlage DS, Motoshima M and Tredway DR (1973) Sperm transport from the external cervical os to the fallopian tubes in women: a time and quantitation study. Fertil Steril 24,655–661.[Web of Science][Medline]
Suh RS, Phadke N, Ohl DA, Takayama S and Smith GD (2003) Rethinking gamete/embryo isolation and culture with microfluidics. Hum Reprod Update 9,451–461.
Tournaye H (2003) ICSI: a technique too far? Int J Androl 26,63–69.[CrossRef][Web of Science][Medline]
Trounson AO and Gardner DK (2000) Handbook of In Vitro Fertilization. 2nd edn, CRC Press LLC, Boca Raton, FL.
Van der Ven HH, Hoebbel K, Al Hasani S, Diedrich K and Krebs D (1989) Fertilization of human oocytes in capillary tubes with very small numbers of spermatozoa. Hum Reprod 4,72–76.
Verpoorte E (2002) Microfluidic chips for clinical and forensic analysis. Electrophoresis 23,677–712.[CrossRef][Web of Science][Medline]
Walters EM, Clark SG, Beebe DJ and Wheeler MB (2004) Mammalian embryo culture in a microfluidic device. Methods Mol Biol 254,375–382.[Medline]
Wright VC, Schieve LA, Reynolds MA, Jeng G and Kissin D (2004) Assisted reproductive technology surveillance–United States, 2001. MMWR Surv Summ/CDC 53,1–20.
Zeringue HC and Beebe DJ (2004) Microfluidic removal of cumulus cells from mammalian zygotes. Methods Mol Biol 254,365–374.[Medline]
Zhu X, Chu LY, Chueh B-H, Shen M, Hazarika B, Phadke N and Takayama S (2004) Arrays of horizontally-oriented mini-reservoirs generate steady microfluidic flows for continuous perfusion cell culture and gradient generation. Analyst 129,1026–1031.[CrossRef][Medline]
Submitted on March 20, 2005; resubmitted on July 25, 2005; accepted on August 17, 2005.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||