Matrices for follicle culture
Synthetic and biologic matrices for the support of follicle growth and maturation have been studied in several animal models as well as in humans. Table
1 presents various 3-D systems and matrices that have been applied to the culture of rodent pre-antral follicles. Table
2 summarizes 3-D culture models for follicles from larger animals, primates and humans.
Table 1
Studies examining matrices for 3-D culture of pre-antral ovarian follicles from rodents
Jin et al. | 2010 | In situ culture followed by encapsulation of follicles in 0.25% ALG-alginate or FA-fibrin alginate | Mouse | In situ: 4 days Alginate: 12 days | NA | Alginate: 69.6% FA: 75% | Alginate: 75% FA: 86% (from follicles with antrum formation) | Alginate: 61.3 ± 2.4 FA: 88 ± 8.7 (from oocytes undergoing GVBD) | Formation of 2 cell embryos Alginate: 33% FA: 54 ± % |
Xu et al. | 2009 | Fresh follicles (Foll) and cryopreserved (Cryo-Foll) and crypreserved ovarian tissue (Cryo-OV) ALG-alginate (0.25%) | Mouse | 12 days | 100-130 μm | Fresh-Foll 78% Cryo-Foll 74% Cryo-OV 72% | Fresh-Foll: 83% Cryo-Foll 69% Cryo-OV: 92% | Fresh-Foll:59% Cryo-Foll: 64% Cryo-Ov 68% | Gap junction protein, connexin expression also studied-down regulated after cryopreservation |
Shikanov et al. | 2009 | ALG (0.25%) FA fibrin alginate | Mouse | 12 days | 100-130 μm | ALG:78% FA:77-81% | ALG:88% FA: 72-88% | ALG:67% FA:75-82% | NA |
West et al. | 2007 | ALG (0.7%, 1.5%, 3%) | Mouse | 8-12 days | 100-130 μm 150-180 μm | 100-130 μm 15-42% 150-180 um 83-91% | 100-130 μm 31-66% 150-180 um 46-91% | NA | Low matrix stiffness increase growth, antrum, GVBD and E2 |
Kreeger et al. | 2006 | Tested alginate with ECM 2 versus multi-layer follicles ALG alginate alone (1.5%) CI collagen I FN fibronectin RGD peptides CIV collagen IV LN laminin | Mouse | 8 days | 100-130 μm versus 150-180 μm | ALG 64% vs 69% CI 65% vs 67% FN 70% vs 72% RGD 72% vs 62% CIV 72% vs 48% LN 63% vs 61% | ALG 12% vs 38% CI 18% vs 25% FN 23% vs 17% RGD 13% vs 36% CIV 20% vs 36% LN 29% vs 11% | ALG 40% CI 44% FN 71% RGD 65% CIV 50% LN 71% *Multi-layer follicles | Transition to secondary Follicle promoted by CI and RGD MII formation promoted by: FN, RGD, LN |
Xu et al. | 2006 | ALG (1.5%) | Mouse | 8 days | 150-180 μm | 93% | 82% | 71% | Live births of pups |
Xu et al. | 2006 | ALG (0.25%, 0.5%,1%.1.5%) | Mouse | 12 days | 100-130 μm | 74-85% | 78-88% | 56-67% | Decreasing 2-cell and blast with increasing % ALG |
Heise et al. | 2005 | ALG (1%) Encapsulation with or without FSH in gel medium | Rat | 72 hrs | 150-160 um | NA | NA | NA | Inclusion FSH with ALG and culture medium, Follicle diameter increased by 33% |
Mousset-Simeon et al. | 2005 | Microdrops under oil Agar Millicell-CM membrane insert | Mouse | 12 days | 100-130 μm | Microdrops 72% Membrane 46% Agar 30% | Microdrops 63% Membrane 33% Agar 26% | Microdrops 53%, Membrane 56% Agar 13% | 2-D Microdrop high survival. Maturation rate similar to 3-D on membrane |
Kreeger et al. | 2005 | Compared effects two versus multi-layered follicles ALG(1.5%)-Collagen I matrix FSH 5-50 mIU/ml | Mouse | 8 days | 100-130 μm 150-180 μm | 66-77% 30-72% | 21-27% 9-43% | Multi-layer follicles 40-78% | Hormone secretion (E2 and progesterone) in multi-layer follicles FSH dependent |
Loret de Mola et al. | 2004 | Collagen treated membrane Collagen gel encapsulation | Mouse | 10 days | 118 μm | Membrane 55% Collagen gel 15% | NA | Membrane 17% Collagen gel 19% *Based on number of recovered eggs | Follicle size larger in collagen gel but maturation and rate of 2-cell formation not enhanced |
Adam et al. | 2004 | Microdrops under oil Millicell-CM membrane insert | Mouse | 6 days | 150-174 μm 175-200 μm | Microdrops 77% Membrane 83% | NA | Membrane 79% | Membrane insert Fert rate 75%, blast rate 48% |
Pangas et al. | 2003 | Alginate | Mouse | 10 days | 82 μm | 68% | NA | 40% | TEM indicate follicles in ALG maintained ultrastructure |
Gomes et al. | 1999 | Collagen gel encapsulation | Mouse | 6 days | 135 μm | NA | NA | NA | Follicle volume and response to FSH increased with 3-D culture in collagen |
Nayudu et al. | 1992 | Millicell-CM membrane insert | Mouse | 6-7 days 3-5 days | 125-150 μm 150-180 μm | NA | NA | NA | FSH stimulated growth, antrum formation, E2 dose response to FSH levels |
Torrance et al. | 1989 | Collagen gel encapsulation | Mouse | 14 days | 20-95 μm | 36% | NA | NA | Growth to multi-laminar stage but no antrum formation |
Table 2
Summary of 3-D culture studies with follicles from human, primate and large domestic animal species
Amorim et al. | 2009 | Alginate (ALG)1% | Human | 7 days | 34-52 μm | 44--70 μm | 90% | NA | Alginate culture system supported growth of Isolated follicles from frozen-thawed ovary |
Xu et al. | 2009 | Alginate 0.5% Matrigel embedded | Human | 30 days | ~175 μm | 715 μm | NA | 75% | Both 3-D systems supported growth of isolated human follicles |
Xu et al. | 2009 | Alginate (ALG) 0.25% versus 0.5% | Rhesus monkey | 30 days | 100-300 μm | 20 vs 78% | 60 vs 78% | Yes | Higher ALG better survival and growth. LH addition with FSH negative effect on survival and P4 secretion |
Itoh et al. | 2002 | Collagen gel | Cow | 13 days | 145-170 μm | 304 μm | NA | Yes | Serum-free culture. Insulin, FSH and LH together induced earlier antrum formation |
Abir et al. | 2001 | Collagen gel | Human | 24 hours | 35-45 μm | 70 μm | NA | NA | Collagen matrix supported growth of fully isolated follicles but not tissue slice with partially isolated follicles |
Hovatta et al. | 1999 | In situ and partially isolated follicles Millicell + Matrigel | Human | ~28 days total | NA | NA | NA | No | Tissue slices better less oocyte extrusion than collagenese isolated. Four weeks to reach secondary stage |
Yamamoto et al. | 1999 | Collagen gel | Cow | 14 days | 500-700 μm | NA | 37% | Yes | MII 27%, 42% fertilization, 4% blastocyst One live birth. |
Hiraoi et al. | 1994 | Collagen gel | Pig | 16 days | 220-300 μm | NA | NA | Yes | 40% MII formation in oocytes ≥110 μm No MII from oocytes <110 μm. Oocytes capable of being fertilized |
Roy and Treacy | 1993 | Agar | Human | 5 days | 90-220 μm | NA | NA | Yes | FSH induced antrum formation, hormone secretion. No FSH, no E2 secretion |
All of the matrices adopted for 3-D culture essentially permit spherical growth of the follicle, preserving the physical integrity of granulosa cell and oocyte's interaction. Nayadu et al [
54] accomplished this using a Millicell hydrophobic insert. The non-tissue culture treated surface prevented granulosa cell migration that could disrupt follicle architecture.
A variety of optically clear gels have also been applied towards follicle culture in different animal models. Follicles have either been completely encapsulated to create a 3-D environment or grown on a gel membrane with medium bathing both surfaces to simulate 3-D culture. Gels that have been used for tissue engineering include hydrogels like agar/agarose, calcium alginate, and hyaluronan, all from naturally derived polymers, as well as synthetic polymers such as poly (ethylene glycol) and poly (vinyl alcohol) (reviewed in [
21,
55]). Gels containing collagen alone as well as compounds containing collagen in combination with ECM proteins have also been applied to
in vitro follicle growth. The physical characteristics of each of these matrices permit physical expansion of the follicular unit during growth.
Hydrogels contain polymers that cross-link or self-assemble into hydrophilic structures. The 3-D cross-linking is what gives the gel its stiffness. The temperature and conditions for this cross-linking can be a critical factor in determining subsequent development of the follicle. For instance agar, derived from seaweed, requires exposure to elevated non-physiologic temperatures for melting before the cross-linking or gelling step, potentially damaging the follicle [
56]. Higher rates of atresia were observed in follicles grown on agar as compared to those placed in microdrop culture or in 3-D culture on a hydrophobic membrane insert [
26]. In contrast, Huanmin et al. (2000) described active follicular growth and antrum formation with caprine follicles embedded within agar [
57]. Their data did however show that secondary follicles survived better than primary follicles in this 3-D agar culture system. Agar embedding has also been applied to human and hamster pre-antral follicles [
58,
59]. Follicles were biologically competent, secreting steroids and synthesizing DNA. Low melting point agarose may be a better matrix for follicle embedding, permitting encapsulation at temperatures more conducive to continued cell growth.
Collagen, a protein found in fibers of connective tissue, is rich in glycine and proline and can be hydrolyzed in to a gel by boiling. This biomaterial has been widely applied to follicular culture. Eppig and colleagues used collagen membrane inserts as substrata in an attempt to simulate 3-D follicle culture. The membrane inserts with follicles were suspended in wells, and follicles were exposed to culture medium from below as well as above [
13,
60]. The biomaterial was not tissue culture treated. It did however allow follicle attachment but minimized granulosa cell migration.
In vitro follicle maturation resulted in the formation of metaphase II oocytes, with the capability of producing live young after
in vitro fertilization, growth and transfer to foster mothers. Despite this achievement, follicle growth on collagen treated membranes had limited potential in terms of maintaining spheroid follicle structure and follicles were susceptible to flattening over time in culture and to premature oocyte ovulation.
To create a more spatially uniform 3-D culture system, follicles have also been embedded in collagen gel [
61‐
64]. Spontaneous follicle disruption as a result of discontinuous or distorted basal lamina and granulosa cell migration was decreased in the 3-D collagen system compared to control 2-D culture systems [
62]. Follicle growth rate has also been reported to be superior [
63]. Granulosa-cell oocyte complexes embedded in collagen matrix remained rounded and compacted with neuronal-like outgrowths towards the oocytes [
64]. Two limitations of the collagen gel have however been noted. The collagen gel is susceptible to shrinkage over time, affecting the gel's natural properties as well as reducing visibility during microscopic assessment [
61]. Also, follicle extraction from the collagen requires enzymatic digestion of the gel, with the potential for subsequent damage to the oocyte [
65].
The natural scaffolding upon which cells are organized
in vivo, known as the extracellular matrix (ECM), is composed of collagen, along with laminin and fibronectin. ECM has been shown to play an important role in regulating cell behavior, differentiation and secretory activity (reviewed in [
30]). One commercially available ECM tested for follicle growth is matrigel [
49,
66,
67]. This ECM product is derived from the Engelbreth-Holm-Swarm (EHS) mouse sarcoma. Matrigel is composed of collagen IV, laminin, fibronectin, entactin, heparin sulfate proteoglycans, and a variety of growth factors such as EGF, FGF, IGF-1, PDGF and TGF-β [
68,
69]. Murine pre-antral follicles in 3-D culture in matrigel exhibited higher growth and survival rates than those in conventional culture [
28]. Hovatta et al demonstrated higher survival of follicles in frozen-thawed human ovarian tissue placed in culture on matrigel coated inserts [
29,
49]. Autocrine and paracrine signaling by ECM molecules and associated growth factors likely affect folliculogenesis. The interactions between ECM proteins and follicles from different animal models needs to be further studied. The source and type of ECM could also play a role in regulating follicle growth during 3-D culture. The size of ECM molecules can present problems and an alternative solution has been to adsorb known sequences of matrix peptides, such as RGD (Arg-Gly-Asp) or laminin-derived peptide sequences on to synthetic matrices (reviewed in [
30]).
To date the most widely applied system for follicle encapsulation and 3-D culture has been alginate produced by brown algae [
39,
45,
46,
48,
67,
70‐
73]. Alginate in the presence of calcium crosslinks to form a hydrogel. This property facilitates encapsulation of follicles under physiologic conditions. Pangas et al. (2003) first applied this system to the 3-D culture of granulosa-cell oocyte complexes (GOC) from 12-day old mouse pre-antral follicles [
70]. GOCs were embedded in alginate beads ranging in size from 0.5 to 1 mm in diameter. Light microscopic and TEM ultra-structure studies suggested that the alginate did not interfere with oocyte or granulosa cell growth development over a 10 day culture interval. Moreover, oocytes recovered from the encapsulated GOCs were able to resume meiosis, undergo fertilization and produce viable offspring [
39]. This 3-D system has also been applied to secondary follicles. Follicles embedded in alginate hydrogels responded to FSH stimulation in a dose-dependent fashion, secreting estradiol and progesterone [
71]. Alginate matrix stiffness and density can affect secondary follicle expansion, hormone production and oocyte maturation [
45,
46]. Non-human primate follicles have also been successfully cultured in calcium alginate gels for up to 30 days [
48]. The encapsulated monkey pre-antral follicles secreted estrogen, progesterone and androstenedione and responded to FSH in the culture milieu. Interestingly, follicles cultured in 0.5% alginate performed better than those in 0.25% alginate, suggesting that primate follicles may require more physical support. One concern however is that denser matrices could potentially limit access to hormones and other nutrients. Heise et al. (2005) reported inhibited delivery of FSH to micro-encapsulated follicles [
36]. Follicle diameters increased with inclusion of FSH in the hydrogel but still did not reach that observed in un-encapsulated controls. Clearly, the physical attributes of the 3-D matrix selected for follicle culture needs to be tailored towards the species and follicle stage being cultured.
In humans, pre-antral follicle growth
in vitro offers an avenue through which cryopreserved ovarian tissue can be utilized without the need for transplantation. Human follicles isolated from fresh or cryopreserved ovarian tissue have been successfully cultivated in calcium alginate hydrogels but functionality needs to be further characterized [
72]. Initial data with frozen mouse ovarian tissue certainly suggests that meiotically competent oocytes can be recovered after
in vitro maturation of isolated follicles in this 3-D culture system [
73].
To further simulate the
in vivo environment, ECM molecules have been combined with calcium alginate to construct synthetic ECM matrices for 3D culture [
74]. The adhesion peptide sequence arginine-glycine-aspartic acid (RGD) common to ECM proteins has been synthetically created and coupled to calcium alginate to construct such a synthetic matrix for follicle growth. Hormone secretion by follicles was directly related to adhesion peptide concentration and a three-fold increase in progesterone and estradiol secretion could be induced by adjusting matrix parameters. In a separate study, these investigators combined calcium alginate with additional ECM components such as collagen I, collagen IV, laminin and fibronectin [
75]. Matrix effect on growth from two-layered to multi-layered follicles as well as oocyte maturation to metaphase II was compared. Transition to the multi-layered, secondary follicle was enhanced in alginate matrices with RGD or collagen I. Final maturation of oocytes and resumption of meiosis was promoted by presence of fibronectin, laminin or RGD peptide.
Criteria for biomaterial evaluation
Increasing follicular diameter is typically used as a measure of follicle maturation. During
in vitro growth, especially in traditional 2-D culture systems where there is granulosa cell expansion, an increase in horizontal diameter of the follicle does not necessarily correlate to overall follicular growth [
62]. With 3-D culture the biomaterial presents equal counter-forces in all directions, minimizing flattening and allowing equal growth along all axes. Follicle volume as well as diameter should therefore be taken into account when comparing different substrata.
Another outcome measure indicative of follicle functionality and growth is antrum formation. This accumulation of fluid within the follicle complex has been shown to vary with 2- versus 3-D culture systems, as well as the biomaterial used for follicle encapsulation. The shear elastic modulus and diffusion characteristics of the biomaterial must be carefully balanced. Torrance et al. (1989) noted no antrum formation in follicles cultured in collagen, despite an apparent increase in follicular diameter over the 14 day culture interval. It was suggested that the double gelling of the collagen during follicle encapsulation allowed just enough flexibility for some granulosa cell proliferation, but that the overall high shear elastic modulus (increased stiffness) inhibited antrum formation [
61]. Interestingly, this was not observed when follicles were individually cultured in collagen microbeads [
76].
A relationship between decreased gel stiffness and greater antrum formation was also observed with calcium alginate hydrogel when tested at concentrations of 3%, 1.5% and 0.7% [
45]. The study of Xu et al. (2006) most clearly illustrates the opposing influences of the rigidity of the biomaterial at high gel concentration and its interference with diffusion and optimal growth [
46]. Oocytes obtained from follicles encapsulated in 0.25% alginate had a higher developmental capacity than those cultured in 1.5% alginate.
In vitro maturation and fertilization of oocytes in 0.25% vs 1.5% calcium alginate were significantly higher (41% vs 5%, respectively). Moreover oocytes derived from the stiffer gel were clearly impaired and unable to undergo
in vitro blastulation [
46].
Interestingly, follicles from primates showed the opposite relationship between gel rigidity and follicle growth. Follicle survival and diameter were increased with culture in 0.5% calcium alginate as compared to 0.25% [
48]. Ovarian stroma of primates is more rigid than that found in rodents and it has been suggested that perhaps primate as well as human follicles may require a stiffer biomaterial to optimize
in vitro culture and growth. The 100% survival rate and 75% antrum formation observed with human secondary follicles grown in 3-D culture in 0.5% calcium alginate matrix further support this supposition [
67].
Non-gel culture systems
Despite the aforementioned benefits of follicle encapsulation as a model for 3-D culture, there are also difficulties. The process of encapsulation as well as the removal of follicles from the gel can be problematic, sometimes resulting in loss of healthy follicles [
61,
63,
65].
Alternatives methods for 3-D culture of follicles that do not involve encapsulation have therefore also been explored. Suspension culture of follicles in orbiting test tubes [
36,
37,
77], rotating-wall vessels [
77], and roller bottle systems [
23] can maintain the 3-D morphology of the follicles without encapsulation. Unfortunately these systems have not been extremely effective. The rate of rotation necessary to keep the follicles from descending to the bottom of the vessels imposes shear stress on the follicles causing follicle degeneration [
77]. Moreover, the only way to negate this effect was to encapsulate the follicles before subjecting them to suspension culture with rotation [
37,
77].
Suspension culture in rotating systems with its accompanying shear stress resulted in more follicle loss than that observed with embedding and removal of follicles from gels. Follicle survival with culture in a rotating-wall culture vessel was only 9% [
77] as compared to the 15% observed after embedding and removal from collagen gel culture [
63]. With marsupial follicles, survival rate in the roller culture system was higher; nearly 49%, but follicles exhibited no antrum formation [
23].
Other non-gel approaches have included serial culture of follicles in new wells each day to prevent attachment [
78] and flattening, or culture in simple microdrop under an oil overlay [
79]. Inverted microdrop suspension culture has also been tested as a means to maintain the 3-D architecture of follicles [
23,
80]. Follicles are placed in microdrops under oil on the bottom of a tissue culture plate and then hung upside down during culture. Oil is ideally suited as a biomaterial for micro-culture environments, allowing maintenance of pH and temperature around the follicle and free gas exchange [
81]. However, its hydrophobic properties could potentially allow the escape of lipid soluble follicle secretions and growth factors in to the oil layer, ultimately hindering growth [
82]. It should however be noted that while inverted suspension culture yielded survival rates similar to that observed with alginate gels, the meiotic maturation rate was only 10% [
23], far less than that what has been achieved with gel encapsulation of follicles [
26,
39,
70]. Handling large numbers of follicles in inverted suspension culture would also be a delicate and labor intensive process. This method would be especially unsuitable for follicles from the human ovary, which might require as long as three months of culture.
Microfluidic culture
The final aspect of follicle culture that needs some attention is the development of culture vessels or systems that maximize diffusion of nutrients and gases through the selected biomaterial while allowing retention of the delicate micro-environment of the follicle and the concentration of essential trophic factors around the oocyte. To accurately mimic the
in vivo ovarian environment, fluid flow across the encapsulated follicle is vital. Also, within the ovarian environment follicles are grown in close proximity of each other, allowing sharing and concentration of secreted factors. The logistics of co-culturing numerous encapsulated follicles can perhaps be aided by the use of microfluidics that allow precise control and manipulation of fluids using microchannels. Microchannels increase the surface area-to-volume (SAV) ratio, implementing laminar fluid flow [
83](reviewed in [
84]).
Diffusion across biomaterials has been shown to be influenced by not only the biomaterial and its concentration but also by its shape or presentation. Encapsulating in microbeads of gel may allow more uniform diffusion across all surfaces as compared to culture with follicles embedded in a single continuous layer of gel. Survival and antrum formation by cultured pre-antral buffalo follicles was demonstrated to be better after culture in collagen microbeads as compared to a continuous layer of collagen matrix [
76]. Tiny microbeads containing follicles in a biomatrix, combined with a system of microchannels could be used to create a network of individual follicles sharing nutrients. A dynamic medium exchange could therefore be applied to follicle culture in a manner that avoids the shear stress observed with rotating culture systems and preserves a "co-culture" atmosphere.
A variety of microfluidic culture systems have been described. Cell immobilization with continuous media flow is the common goal. This can be accomplished with microposts on the culture surface to entrap cells and create a matrix support while still allowing laminar flow of fluid to pass by [
85,
86] or by entrapping cells between walls of PDMS with continuous flow of culture medium above the cells [
83]. Microwells can also be used as architectural supports in microfluidic systems and act as nests for cells to culture in while fluid is exchanged above or below [
87,
88]. Microfluidics in combination with valves and micro-scale pumps provide the option of continuous media flow in ways similar to that seen
in vivo[
89,
90]. Microfluidics thus permits dynamic culture conditions and medium flow without disturbing the cell itself.
Application of microfluidics to the field of reproductive biology has gained much attention. It has been applied to sperm sorting [
91,
92], oocyte handling and fertilization [
93‐
96] and embryo culture [
83,
97‐
101]. Follicle culture in microfluidic devices needs to be explored. This type of system may be ideal for providing the 3-D environment necessary for maintaining follicle architecture over long intervals in culture, allowing adequate oxygenation and nutrient exchange and at the same time permitting sequestration of autocrine/paracrine factors within the vicinity of the growing follicle. The ideal microfluidic model would allow monitoring and harvest of individual follicle but also a sharing of the microenvironment to attain the benefit of "co-culture".