Introduction

The amniotic membrane (AM) is a thin membrane on the inner side of the fetal placenta, consisting of an epithelial coat and a layer of mesenchymal connective tissue. It possesses unique biological features, including anti-inflammatory, anti-bacterial, anti-viral, and immunological characteristics as well as anti-angiogenic and pro-apoptotic qualities. The AM is commonly used clinically for wound treatment and ocular surface reconstruction. In more recent years, hAECs isolated from the AM have been investigated as a form of regenerative medicine given their immunomodulatory, anti-tumorigenic, and anti-inflammatory properties. In this review article, we focus on the hAECs and their role as a therapeutic source. First, we discuss the historical use of the AM and highlight more recent developments in amnion-derived products. Next, we provide insight into hAECs themselves. This is followed by a summary of preclinical models using hAECs and current clinical trials investigating the safety of hAECs for cell therapy. Lastly, we describe tissue engineering using the AM as a bio-scaffold and preclinical studies exploring extracellular vesicles for cell-free therapy. This is summarized in Fig. 1.

Fig. 1
figure 1

The amniotic membrane is a source of cells, extracellular vesicles, and bio-scaffold materials

Full size image

History of the Amniotic Membrane in Medical and Surgical Applications

The human placenta has been used traditionally for over centuries in Chinese medicine. The Compendium of Materia Medica, a Chinese record of substances with medicinal properties published in 1593 by Li Shi-Zhen, devotes an entire section to the uses of the human placenta as a medicine [1]. More than 3 centuries later came the first documented use of the fetal membrane/amniotic membrane as a surgical material in skin transplantation in 1910 [2] and 1913 [3, 4]. The AM was employed in skin grafting and gave superior results when compared to xenografts or cadaveric coverings [2]. In 1913, intact amniotic membranes were applied to varicose ulcers, burns, scalds, and denudations of traumatic origin. Dr. Maximillian Stern found that the amniotic cellular tissue was taken up by the raw surface of the wound without remarkable adverse effects [3]. It would be another 20 years before another paper was published using the AM for wound repair or surgery. In the late 1930s, the AM was used for genital reconstructive surgery [5]. Then in the 1940s, it was used for the first time in ophthalmology to repair conjunctival defects [6] and ocular burns [7, 8]. Since then, the AM has been further investigated with studies and a large number of clinical trials reconfirming its successful use for skin injuries and in various clinical indications, including reconstructive surgery (genital and abdominal) [9,10,11,12,13,14,15,16], prevention of adhesions [10, 17, 18], ulcers [19,20,21], and burns [22, 23].

More Recent Developments with Amnion-Derived Products for Wound Healing

Ophthalmologists embraced the use of AMs increasingly from the 1990s [24], and they are now routinely used as a graft, spread over the ocular surface to treat epithelial defects or ulcers, or as a bandage to promote healing [25]. More recently, the commercial market for AM-based products for skin wounds has boomed. There are multiple companies offering human amnion-derived products for wound healing, ranging from sheets to cover wound beds to particulates used to “fill” tunneled wounds (summarized in Table 1). Primarily these products are applied as wound coverings (acute and chronic wounds, burns, pressure ulcers, diabetic ulcers, venous stasis ulcers). Additional uses in surgical procedures include extremity, vascular, spine, orthopedic, urological, colorectal, and general surgery. Investigative clinical trials run with these commercially available products focus primarily on the treatment of diabetic foot ulcers and venous leg ulcers. The successful use of the AM in wound healing and surgical applications is in part due to its low immunogenicity (see the section “Clinical Trials”). The AM does not require preconditioning, attributed to the low expression of classical major histocompatibility complex (MHC) classes I and II molecules and high expression of HLA-G from hAECs (see the section “Stem-like Cells from Amniotic Membranes”—“Physical Characteristics”) and EVs [26].

Table 1 Human amniotic tissue-derived commercial products for wound healing
Full size table

Denuded Human Amniotic Membranes Can Be Used as a Substrate for Growth of Other Stem Cells

Potential applications of the denuded AM as scaffolds have been explored in both animal and human studies to target tissues such as the eye, cartilage, peripheral nerve, and skin tissue engineering (TE). The extracellular matrix of the human amnion is an effective conduit for peripheral nerve regeneration, and the AM is a biodegradable scaffold with unique biochemical and mechanical characteristics for nerve regeneration [39, 40]. Miyamoto et al. showed that primate embryonic stem cells had undifferentiated growth when cultured on human amniotic epithelial feeder cells, while Ueno et al. utilized the denuded AM as a feeder layer to direct the neuronal differentiation of stem cells [41, 42]. Another group investigated the denuded AM as a carrier matrix to promote chondrocyte growth and support cartilage regeneration [43]. When epithelial and mesenchymal cells were seeded on an AM scaffold, the cells became highly interconnected and capable of penetrating the porous structure of the amnion scaffold. This observation has led to the suggestion of a novel approach for repair of prematurely ruptured fetal membranes, a leading indication for preterm delivery [44]. The seeding of epithelial cells on an AM scaffold is frequently utilized for ocular surface and skin reconstruction [45, 46]. The AM is able to accelerate re-epithelization due to the presence of a basement membrane and therefore is compatible as a unique biological skin substitute for treating deep dermal and full-thickness wounds [47]. Similarly, endothelial cells have been seeded on an AM scaffold for vascular TE purposes [48].

Stem-like Cells from Amniotic Membranes

Amniotic epithelial cells (hAECs) and amniotic mesenchymal stromal cells (hAMSCs) are the major stem and stem-like cells of the human amnion. No invasive technique is required to harvest hAEC and hAMSC as they are isolated from the epithelial layer of the amniotic membrane, which is usually discarded after birth along with the rest of the placenta. Specialized protocols have been developed for their isolation. Both hAEC and hAMSC possess unique properties that are vital in regenerative medicine including their multipotent differentiation potential, low immunogenicity [49], and anti-fibrotic and anti-inflammatory properties [50]. In vitro, hAMSC and hAEC have shown to develop into cells with mesodermal, ectodermal, and endodermal lineages [50]. hAMSCs are a multipotent stem cell population that express the classical MSC markers including CD90, CD44, CD73, and CD105 [50, 51] while hAECs are pluripotent and express different markers including OCT-4, NANOG, SOX-2, and TRA-1-60 [52,53,54]. Typically, a healthy term placenta will yield >10 times more hAEC compared to hAMSC [55]. In this review, we will focus on the therapeutic effects of hAEC. hAECs have been explored as a potential therapy for fibrotic and inflammation-based disorders including respiratory [56, 57, 58•, 59], gastroenterological [60•, 61, 62], neurological [63, 64], and cardiac conditions [65, 66].

Physical Characteristics

hAECs are small circular cells that have either a central or eccentric nucleus with one or two nucleoli and an abundant cytoplasm [53]. hAEC express embryonic stem cell markers SSEA-3 and SSEA-4 (stage-specific embryonic antigen 3 and 4) and tumor rejection antigens 1-60 (TRA 1-60) and 1-81 (TRA 1-81). The pluripotency of hAECs has been attributed to the expression of these markers as well as the expression of pluripotent stem cell transcription factors, Oct-4, Sox-2, Nanog, and Rex-1 [52,53,54]. The low immunogenicity of hAEC can be attributed to the expression of the class I human leukocyte antigen, HLA-G. The soluble form of HLA-G is known to induce the apoptosis of activated CD8+ T cells and modulate natural killer cell [67] and allo-cytotoxic T cell responses [68]. The membrane-bound form of HLA-G has also been shown to inhibit natural killer cell and T cell-mediated cytolysis [69, 70] and allo-specific CD4+ T cell proliferation [71, 72] and induce a Th2 response [73, 74]. Furthermore, hAECs express low levels of HLA class I and II molecules (i.e., HLA-A, B, C, and DR) which usually stimulate allogenic rejection [49, 75]. The immunomodulatory effects of hAEC are evident through their ability to suppress the T cell response, inhibit neutrophil and macrophage infiltration, and induce macrophage polarization [54, 56].

In culture, hAECs exist in three subpopulations—adherent, intermediate (loosely adherent), and free floating [76]. The adherent population consists of cells that grow in a single layer in culture with cobblestone epithelial morphology [77], the intermediate population consists of cells that are weakly adherent, and finally, there are a population of cells that remain free floating in culture [76]. All three subpopulations have been shown to have varying expression levels of OCT4, NANOG, and stem cell surface markers SSEA-4, TRA1-60, and TRA 1-81 [76]. hAECs reach replicative senescence after 6–10 passages in cell culture [78] due to the activation of epidermal growth factor receptor (EGFR) and cell-cell interactions at high cell densities in culture [78].

Impact of Donors on hAEC Quality

A consequence of the poor proliferative capacity of hAECs necessitates ongoing tissue donation. As a result, it would be beneficial to assess donor-specific variations which potentially affect the therapeutic potency [50, 79]. The level of HLA-G in hAEC has been shown to vary between donors, with the greatest association with gestational age [54]. Larger studies to assess the impact of common maternal-fetal health complications and risk factors such as fetal growth restrictions, maternal smoking, or maternal obesity should be considered.

Preclinical Applications of hAEC

The therapeutic potential of hAEC has been evaluated in small and large animal models of cardiorespiratory diseases, brain injury and neurological disorders, liver disease, metabolic diseases, and autoimmune diseases (summarized in Table 2). hAECs have been shown to improve organ structure and function through their anti-inflammatory, anti-fibrotic, immunomodulatory, and regenerative effects as discussed earlier. Furthermore, the application of hAECs has progressed into clinical studies in diseases including bronchopulmonary dysplasia, chronic liver disease, and stroke.

Table 2 Preclinical applications of human amnion epithelial cells
Full size table

Clinical Trials

In recent years, multiple phase I clinical trials using allogenic hAECs have commenced (Fig. 2).

Fig. 2
figure 2

Summary diagram of phase I clinical trials currently underway using allogenic hAECs. The hAEC and their anti-inflammatory, anti-fibrotic, and immunomodulatory properties as well as their ability to support angiogenesis are being employed as a form of regenerative medicine. hAEC therapies are treating bronchopulmonary dysplasia in premature infants and liver disease (cirrhosis and liver fibrosis), acute ischemic stroke, and fistulising perianal Crohn’s disease in adults (18–85 years old)

Full size image

Bronchopulmonary dysplasia (BPD)

A first-in-human phase I safety study (ACTRN12614000174684; UTN: U1111-1151-8685) using allogenic hAECs in premature infants with BPD was successfully completed in 2018 [119]. Six premature infants with established severe BPD were enrolled to a single-center, open-label trial. Each infant received one million cells/kg by intravenous infusion. The study showed that allogeneic hAECs were well tolerated. A phase I dose escalation study was subsequently registered (ACTRN12618000920291) [120]. The study will include 24 infants, with the first 12 infants receiving a single infusion of 2 to 10 million/kg. Infants 13–18 will receive two infusions to achieve 20 million/kg and infants 19–24 will receive three infusions to total 30 million/kg. The study has recruited 14 infants to date.

Liver cirrhosis

In 2017, Lim et al. [121] published a study protocol for the first phase I trial evaluating the safety and tolerability of intravenously delivered allogenic hAEC in 12 patients with compensated liver cirrhosis (ACTRN12616000437460; UTN: U1111-1181-4339). This is a single-center, open-label, dose escalation of hAEC (0.5 to 3 million/kg) clinical trial with four cohorts of three patients each. The study has recruited 6 patients to date.

Ischemic stroke

hAECs are also currently evaluated in patients who have suffered an acute ischemic stroke (ACTRN12618000076279) [122]. Eligibility criteria for this trial include an ischemic stroke in the territory of the large main artery, within 24h of stroke onset, ineligible for clot retrieval and NIH stroke severity (NIHSS) scale between 6 and 15. This is a phase I, open-label, dose escalation 3+3 trial. Dosing starts at 2 million cells/kg with the final group receiving 32 million cells/kg. It is open for recruitment and has thus far treated 8 patients, with a target of 15 total.

Perianal fistulae

A phase I clinical trial (ACTRN12618001883202) was registered in 2018 to evaluate the safety of locally administered allogenic hAECs for the treatment of refractory perianal fistulising Crohn’s disease. This open-label study aims to recruit 10 adults with complex perianal Crohn’s fistulas who have previously failed one conventional treatment. The participants will receive a dose of 40 million cells/fistula, with up to 3 fistulas treated per participant (maximum 120 million cells). Four participants have been enrolled to date.

Interestingly, these clinical trials are not the first to investigate allogeneic hAEC clinically. The very first study in 1981 [123] was conducted in London, England, where a layer of amniotic epithelial cells were transplanted into seven volunteers (six men and one woman, ages 28–80 years old) without immunosuppressive treatment. The aim of the study was to assess the immunogenicity and survival of the amniotic epithelial membrane implants, with the intention of using the implants for treating patients with enzyme defects. Specifically, Adinolfi et al. [124] showed that hAECs produce lysosomal enzymes that are capable of correcting in vitro the enzymatic defects of patients with Hurler’s syndrome. None of the volunteers showed signs of acute rejection and hAECs were present in biopsies for up to 7 weeks post-implantation.

Niemann-Pick disease

In 1987, an Italian team published a successful treatment of Niemann-Pick disease type B by implanting a suspension of allogeneic amnion membrane tissue [125]. The tissue suspension was injected into a subcutaneous thoracic pouch under the armpit. A total of six implants were placed at intervals of 1–4 months. Following the fifth and sixth implantations, serous secretion from the wound was noted. This was thought to be due to host-versus-graft rejection due to the presence of donor fibroblasts and macrophages in the tissue suspension. The investigators postulated that a pure suspension of amniotic epithelial cells might bypass the immune reaction. Notably, they commented that “In vitro culture and cryopreservation of epithelial cells, …, should allow us to store and inject high numbers of non-immunogenic cells, thereby avoiding graft rejection.” In 1992, the same team published their findings using repeated implantations of hAECs to treat Niemann-Pick disease in five patients over 4 years [126]. The hAEC treatment resulted in the abolishment of recurring infections, mainly of the respiratory tract, and improvements to the general conditions of the patients.

Looking into the Future

Amniotic Membrane as Bioscaffolds

The interest in human AM for TE is on the rise due to its non-xenogeneic origin, inexpensive, highly abundant source and their regenerative properties. TE is defined as the development of biological substitutes for the purpose of restoring, maintaining, or improving tissue function. The three major pillars of TE are scaffolds, cells, and growth factors. An important component of TE is the supporting matrix upon which cells and tissues grow, also known as the scaffold [127•]. The special structure and biological nature of the AM allows it to be an ideal candidate for the fabrication of TE scaffolds.

One of the oldest biomaterials used for scaffolds is the AM as it is easily obtained, processed, and transported. The extracellular matrix (ECM) components of the basement membrane of the AM create an almost native scaffold for cell seeding and the AM itself has important biological properties including anti-inflammatory, anti-microbial, anti-fibrotic, low immunogenicity and provide mechanical stability. Scaffolds are developed to support cell seeding in TE, promoting their differentiation and proliferation in the formation of implantable tissue. Design and selection of the biomaterials used for scaffolding is a critical step because successful cell seeding of the scaffold depends on the type and source of the living cells as well as the ECM components of the scaffold.

A major prerequisite for choosing a TE scaffold is its biocompatibility. Furthermore, its mechanical properties should include permeability, stability, elasticity, flexibility, plasticity, and resorbability at a rate congruent with tissue replacement. Scaffolds should also allow cell adhesion and the potential for delivery of biological agents. Immunocompatibility is another important feature of AM as a TE scaffold as it can bypass the immunological complications of xenogenic biomaterials. The presence or absence of certain ECM molecules within any basement membrane also influences adhesion and growth of the overlying cells as they detect and respond to the ECM including the composition, adhesive ligands, matrix stiffness, and spatial and topological organization of integrins [127•]. The AM is a scaffold that can be used either with the epithelial layer intact or denuded or decellularized. The spongy layer on the stromal portion of the amnion has an abundance of hydrated proteoglycans and glycoproteins. The AM has a mechanical response that is inherently dependent on the stage of pregnancy, described as “viscoelasticity.” This is a critical scaffold property in a majority of tissues. One measure of elasticity is Young’s modulus (the ratio of applied stress to strain) which is 3.6 MPa in preterm human AM (26–36 weeks) compared to 2.29 MPa at full term (36–40 weeks) [128]. Therefore, preterm AM is stiffer than term AM and these properties are of interest for matching mechanical integrity in TE application where one such example is that stiff scaffolds will lack the viscoelasticity of arteries.

Despite all these advantages, there are some potential challenges that need to be addressed when applying any biologically derived material to TE uses. For example, transmission of infectious diseases is always a risk when using human tissues; therefore, precautions and safety criteria must be adhered to. Another issue with TE scaffolds is the possibility of invoking an inflammatory reaction upon implantation (i.e., foreign body reaction or immune rejection). AM has been reported to downregulate TGF-β and its receptor expression by fibroblasts and in doing so, reduces the risk of fibrosis and inflammation [127•]. The physical difficulty in handling and placement of thin AM sheets has limited their use in routine clinical care. In TE applications where AM function as a cell delivery matrix, strategies to improve AM bio-stability are often utilized as it could take longer for transplanted cells to home to the target site [47]. There is a plethora of investigations that focus on enhancing AM usage as a TE scaffold by employing surface modification targeted for various TE applications. When AM is expected to support the in vivo viability of the transplanted cells; therefore, treating AM with crosslinking agents may improve the bio-stability and mechanical strength of the scaffold. One such example is the study by Gobinathan et al. which showed that genipin-crosslinked AM has better bio-stability and the slowest degradation rate compared to decellularized and native AM [129].

Urethral Reconstruction

Xenografts of the urethra made with denuded human AM have been used to minimize potential rejection and maximize biocompatibility. Denuded human AM seeded with rabbit urethral epithelial cells were subcutaneously implanted in a rabbit model of urethral injury with resolution of urethral defects in 1 month. The cell-seeded denuded AM grafts were intact without obvious infiltration of inflammatory cells compared to intact AM patches [130].

Ocular Regeneration

The application of AM in ocular disorders is often limited by its relatively rapid degradation and resorption in vivo. To overcome this, crosslinking of AM has been used to reinforce the biomaterial structures. In particular, the fabrication of photo-crosslinked AM using UV irradiation has been developed as a scaffold for limbal stem cell culture. These physically crosslinked AM matrices exhibited negligible cytotoxicity to the corneal epithelial cells irrespective of the irradiation time and maintained the undifferentiated cell phenotype [131].

Wound Healing

Solubilized AM combined with a hyaluronic acid (HA-SAM) hydrogel was developed to provide a wound treatment that is easy to produce, store, and apply to wounds. Using murine and porcine models of full-thickness wound healing, HA-SAM significantly accelerated wound closure through re-epithelialization and prevented wound contraction [132, 133] and conformed to a non-uniform wound shape. A major benefit of using a hydrogel is its potential to match the rate of growth factor release to specific wound types, e.g., fast release hydrogel for acute burns and slow release hydrogel for diabetic wounds. Moreover, an aseptically processed human amnion and chorion allograft (AmnioBand) has been shown to be superior for wound healing in patients with diabetic foot ulcers compared to FDA-approved engineered skin substitutes (Apligraf). AmnioBand brings even greater value for the patients in terms of the healing efficacy endpoints, graft cost, and graft wastage [27].

Skin Regeneration

The poor mechanical and handling characteristics of AM have led to the development of a 3D skin substitute by reinforcing an AM scaffold with biodegradable polymer [134]. Silver nanoparticles incorporated with poly-[Lactide-co-Glycolide-co-Caprolactone] terpolymer (PLGC) and fibrin coating are used to reinforce the AM and to deliver bioactive molecules to the wound site. This combination scaffold has excellent biocompatibility and the addition of PLGC-silver nanoparticles is expected to provide excellent mechanical properties and potential benefit in treating infectious wounds due to their anti-microbial activity. A fibrin sealant can also act as a hemostat to minimize bleeding upon applying to the wound site; hence, a combined scaffold has potential use for dermal regeneration with better clinical handling.

Cartilage Regeneration

The collagen-rich ECM of AM has been investigated as a potential scaffold for cartilage regeneration. Fabrication of hybrid denuded AM-chitosan hydrogels has been proposed for articular cartilage TE as a rich source of collagen and the study has demonstrated that these hydrogels had a higher elastic modulus than chitosan or collagen hydrogels [135].

hAEC-Derived EVs

In the last decade, significant strides have been made in the regenerative medicine sector with a number of cellular therapies attaining market approval. While there are reports of stem cell-based therapies showing benefit in preclinical disease models, there is increasing evidence that the cells themselves are not critical to the functional outcome. Instead, stem cells serve as bio-factories releasing bioactive products including extracellular vesicles (EVs) (Fig. 3) and growth factors. EVs are naturally occurring nanoparticles (70–120nm) shed by virtually all cell types. Cells selectively package bioactive materials (e.g., proteins, miRNAs) into their exosomal cargo, which then serve as signaling packets to allow intercellular communication [136, 137]. Indeed, the so-called paracrine effects of stem cells are increasingly attributed to EV release and targeting [138, 139].

Fig. 3
figure 3

Amnion epithelial cell-derived extracellular vesicles. Electron microscopy showing cup-shaped morphology of EVs

Full size image

EVs have several advantages over cell therapies. Unlike their cells of origin, EVs are non-replicative, non-living bio-stable nanoparticles that do not require complex storage, transport, and handling [140]. The manufacturing, formulation, and clinical delivery of EV therapeutics are significantly simpler compared to cell therapies, thereby representing a more cost-effective form of regenerative medicine. EVs are enclosed by lipid bilayers and this protects and stabilizes their bioactive cargo compared to the direct delivery of growth factors, nucleic acids, or cells alone. EVs are easy to isolate and remain stable over long periods of time without the need for liquid nitrogen storage. There is therefore a potential to deliver regenerative medicine with a simplified cold chain and significantly lower cost of goods.

The topology of EVs is similar to that of cells, with lipid bilayer membranes decorated with extracellular receptors and ligands as well as cytoplasmic proteins and RNAs contained within. Membrane proteins such as tetraspanins and integrins are central to the identity and function of EVs, and these proteins could also be used to identify the cell type and influence EV uptake by recipient cells.

EVs from hAECs contain a myriad of growth and signaling factors that have immunomodulatory properties and can regulate cell differentiation. Recently, we have shown that hAECs release EVs that have similar regenerative properties to the cells themselves in the setting of experimental stroke and lung and liver fibrosis models [60•, 61, 141, 142•, 143]. We reported that hAEC-EVs exhibit anti-fibrotic properties by decreasing the number of activated hepatic stellate cells resulting in reduced collagen deposition. In addition, we have shown through potency assays that hAEC-EVs exert their anti-inflammatory effect by reducing neutrophil myeloperoxidase activity, suppressing CD3/CD28 activated T cell proliferation, increasing macrophage phagocytic activity, and shifting their polarization state [141, 142•]. These EVs are enriched with various microRNAs such as miR-27a, miR-23a, miR-203a, miR-34a, miR-150, and miR-194 which exert anti-fibrotic properties. The combination of the anti-fibrotic drug serelaxin and hAEC-EV in a model of experimental lung fibrosis demonstrated broader protection compared to pirfenidone, the standard of care. The therapeutic efficacy of hAEC-EV in treating basement membrane-induced fibrosis and related airway dysfunction can be enhanced by the co-administration of serelaxin [85].

Zhao et al. demonstrated that local injection of hAEC-EV could reduce collagen deposition in the rat full-thickness skin wound model [144•]. Another study also reported that hAEC-EV promoted proliferation and migration of fibroblasts and therefore could play an effective role in promoting scarless wound healing [145]. In a study by Zhang et al., the therapeutic potential of hAEC-EV in restoring ovarian functions following chemotherapy was investigated. They demonstrated an increased number of follicles and improved ovarian function in a murine premature ovarian failure model upon hAEC-EV transplantation [146].

Overall, EVs from hAECs contain cargo consistent with their biological properties to potentiate tissue regeneration, participate in immune modulation, and function as potential alternatives to stem cell therapy. As such, their untapped potential as cell-free therapeutics and the further possibility to bioengineer EVs to mediate specific biological functions, facilitate EV uptake, and EV targeting warrant future research exploration.

Conclusions

While hAECs have proven promising as a cell-based therapy, they remain a challenging treatment to integrate into the healthcare system due to their production cost and the challenges of delivering to smaller hospitals and remote communities. In order to overcome these geographical and socioeconomic barriers, the focus has shifted to investigating cell-free regenerative medicines that utilize the amniotic cells’ innate capacity to secrete EVs. Clinical application of EVs should be considered in the near future as numerous animal studies have shown the therapeutic potential of EVs as a cell-free form of regenerative medicine.