Review of the reproductive biology of amphipods and their endocrine regulation: Identification of mechanistic pathways for reproductive toxicants



The reproductive biology of amphipods is reviewed to update the knowledge of the male and female reproductive processes of oogenesis and spermatogenesis as well as the endocrine systems of amphipods with the aim of advancing studies of reproductive toxicology. The ovarian and reproduction cycles of female gammaridean amphipods are closely correlated with the molt cycle, which is under direct control by the steroid hormone 20-hydroxyecdysone. The ability of males to copulate and subsequently for females to ovulate is restricted to the early postmolt period of the females. New developments in our understanding of the molt cycle and the endocrine regulatory pathways for reproduction using genomics techniques on other crustacean species are also discussed. The arthropod sterol ponasterone A or xenobiotics such as the fungicide fenarimol have been shown to elicit endocrine disruption in some crustaceans by acting as an agonist for 20-hydroxyecdysone at the ecdysone receptor or by inhibiting the synthesis of 20-hydroxyecdysone, respectively, resulting in disruption of molting and reproduction. Recent studies suggest that cadmium can inhibit secondary vitellogenesis in amphipods. Experimental approaches for examining the metabolic pathways associated with ecdysteroid hormonal signaling or metabolism, exoskeleton maintenance and molting, and the regulation of vitellogenin in amphipods are discussed. This information should aid in the identification of useful biomarkers for reproductive toxicity. Environ. Toxicol. Chem. 2011;30:2647–2657. © 2011 SETAC


Amphipoda is an order of malacostracan crustaceans found in almost all aquatic environments, with approximately 7,000 species of amphipods described to date and placed into three or four suborders. One suborder, Gammaridea, contains more than 5,700 species, including all the freshwater and terrestrial species as well as many estuarine and marine species 1. Most species in the suborder Gammaridea are epibenthic amphipods and are an abundant and ecologically important component of aquatic benthic communities. They are sensitive to contaminants in their environment and are often the first organisms to disappear from contaminated sites 2–4. Because they meet many of the criteria used for selecting test organisms, they are commonly used in whole-sediment toxicity tests 5, 6. In particular, they are widely distributed, live in direct contact with sediment, are easy to handle and culture in the laboratory, and are tolerant to varying sediment physicochemical characteristics. A recent review provided an overview on the status of ecotoxicological testing with gammarid amphipods 7.

Empirical sediment quality guidelines (SQGs) have been adopted in many countries throughout the world. The empirically based SQGs were developed using large databases with matching measures of sediment chemistry and toxicity with field-collected samples 8, 9. These databases have also been used to evaluate the ability of SQGs to predict sediments to be either toxic or nontoxic in laboratory tests based primarily on the presence or absence of toxicity in 10-d amphipod tests or in benthic community assessments in numerous estuarine and marine environments 10. Assessments conducted with these databases in conjunction with field validation studies involving contaminated sediments suggest that SQGs are reasonable predictors primarily of acute effects or no effects on benthic organisms. However, several limitations have been observed regarding the use of SQGs in sediment quality assessments 10, 11. One concern is the capability of SQGs to predict adequately the presence or absence of chronic toxicity to sediment-dwelling organisms in field-collected sediments. Limitations are also associated with sediment toxicity tests using amphipods, because sublethal effects are not measured in traditional short-term standardized sediment toxicity tests 12, 13. Acute toxicity tests, typically 10 d or less, are more useful for identifying highly contaminated sediments through mortality. Species survival depends ultimately on the reproductive success and quality of offspring; thus, reproductive variables are sometimes considered more predictive indicators of toxicological effects at population and community levels that may be occurring at more slightly impacted sites in the field 14. Sublethal endpoints in laboratory-based sediment toxicity tests using amphipods have been found to be more sensitive to low concentrations of contaminants approaching the sediment quality guideline trigger values 15–19.

Additional research is needed to identify a reproduction-linked biomarker in amphipods that may be applied in the field, because laboratory-based toxicity tests are not necessarily predictive of longer term ecological effects 20, 21. Forbes et al. 22, in a study modeling effects of toxicants at the population level, demonstrated that impacts are likely equal to or less than effects on individual life-cycle traits, suggesting that risk assessments based on the latter are likely protective of population-level impacts. Sundelin et al. 20 have developed a method that allows for the detection of sediment contaminant effects on the reproduction of the deposit-feeding amphipod Monoporeia affinis using a number of different reproduction variables, including sexual maturation, fertilization success, embryo aberrations, and fecundity. This approach has allowed observation of reproductive disturbances of the amphipod M. affinis in the Baltic Sea and has been adopted by the International Council for the Exploration of the Sea as a recommended technique for biological effect monitoring. Despite a number of reports describing reduced fecundity in amphipods inhabiting polluted environments 14, 23–25, most studies have not reported on the mechanisms by which contaminants cause the aberrant reproduction. The present review evaluates the literature on the reproductive biology of amphipods and its endocrine regulation, with the aim of identifying physiological or biochemical pathways in amphipods vulnerable to xenobiotics.


Molt cycle

Malacostracan crustaceans with high levels of fecundity and body growth have closely linked molting and oogenic cycles 26. In gravid females with embryos undergoing development in the external marsupial pouch, maturation of oocytes occurs concurrently within their ovaries in preparation for the next spawning. Furthermore, amphipods carry their broods, extending their reproductive cycle beyond spawning. In these species, the onset of molting is delayed until hatching and release of their juveniles 27, 28. Coordination of the molting of the rigid exoskeleton with the ovarian cycle also facilitates the movement of the newly ovulated oocytes through the oviducts into the marsupium, while the new exoskeleton is still flexible enough to allow their passage 29. The molt cycle stages for females have been described for the gammaridean amphipods Orchestia gammarella30 and Gammarus pulex31 as well as the terrestrial talitridean amphipod O. cavimana32. Lipid and protein are required to form the new cuticle 31. The principal energy reserve of gammaridean amphipods is lipid, and the main site for its storage is extensive adipose connective tissue in the head and thorax and in the abdomen surrounding the rectum. Martin 31 concluded that lipid and glycogen appearing in the epidermis are used in the formation of the new cuticle under the old exoskeleton.

Four major stages in the molt cycle, namely, postmolt, intermolt, premolt, and ecdysis, have been distinguished. The first stage, period A, is the early postmolt period that begins at the end of an exuviation of the old exoskeleton. The new cuticle is soft and flexible at this stage, and the coloring of the body is less than during other periods of the molt cycle. The second stage, period B, begins late postmolt with the gradual coloring of the antennae and legs. The calcareous concretions stored within the posterior caeca are then used to mineralize the new skeleton 33. The third stage, period C, is the intermolt stage. Consolidation continues during the first stage of this period by the secretion and calcification of the last postexuvial layers, with calcium carbonate making up to 40% of the crustacean cuticle. The fourth stage, period D, is the premolt stage. During this period, the new skeleton is progressively formed under the old one. Observation of the morphogenetic processes at the level of the propodite allows the subdivision of this period into four phases 32.

The cuticle of juveniles of the amphipod Parhyale hawaiensis consists of three functional layers: envelope, epicuticle, and procuticle outside the underlying epidermis layer. The outermost lipid and protein-containing envelope is involved in the control of water balance. The middle epicuticle, which is composed of a protein–catecholamine network, and the inner protein–chitin matrix, called the “procuticle,” together constitute the stiff, but elastic, exoskeleton 34. The cuticle of P. hawaiensis is produced by the epidermis during embryogenesis and is renewed during molting. The juvenile dorsal and ventral cuticles of P. hawaiensis are distinctly different. The dorsal epicuticle and procuticle are characterized by inclusions that are missing in these layers at ventral positions. Moreover, pore canals, which have been proposed to be mineralized or to be routes for mineralization, are present only in the dorsal cuticle. The ventral cuticle resembles the cuticle described for other crustaceans, including copepods, branchiopods, and decapods, whereas the dorsal cuticle seems to be a specific trait of amphipods and isopods 34.

Ovarian cycle in amphipods

The anatomical location and appearance of the ovary have been described in the marine amphipods O. gammarella35 and Echinogammarus marinus36 after isolation by dissection (Fig. 1). In some amphipods, such as Hyalella azteca and Gammarus fossarum, ovarian maturation can be observed through their translucent body when pigment is deposited in the oocytes 37, 38. The various stages of oogenesis in amphipods have been studied in O. gammarella39. Amphipods in general lay few eggs compared with many decapods, but the oocytes grow to a large size, which requires a large amount of the yolk protein vitellin 40. Culture conditions that provide animals with a continuous food supply containing omega-3 polyunsaturated fatty acids support high fecundity in the detritus-feeding amphipod Melita plumulosa41. On the contrary, seasonal limitation of essential omega-3 polyunsaturated fatty acids and essential amino acids can reduce the fecundity and population abundances of the deposit-feeding amphipods Leptocheirus plumulosus and Monoporeia affinis in the field 42, 43. Vitellogenesis in the amphipod O. gammarella, involves two phases, primary and secondary vitellogenesis 40. Oogenesis from oogonia to the end of primary vitellogenesis is a continuous process, and the primary vitellin originates from increased ribosomal activity when the oocyte enters primary vitellogenesis 39. The ovary is a paired, elongated organ, the transverse sections of which are identical from one end to the other. It possesses a germinative zone located next to the basal lamina. Cells formed by the division of the primary oogonia leave the germinative zone, enter meiotic prophase, and differentiate into oocytes. Oocytes of O. gammarella in primary vitellogenesis measure from 80 to 160 µm in diameter, and their germinal vesicle (egg nucleus) grows from 33 to 45 µm in diameter 39. At the end of primary vitellogenesis, the oocytes synchronously go into secondary vitellogenesis, in the beginning of a second intermolt period. Initially, fully grown primary follicles are surrounded by a permanent follicular tissue. A prominent feature of O. gammarella oocytes undergoing secondary vitellogenesis is the uptake by endocytosis of vitellogenin produced by the adipose tissue (see subsection Endocrine regulation of the ovarian cycle) that extends their diameter from 160 to 800 µm, and the diameter of the germinal vesicle grows from 45 to 60 µm 39. The resulting secondary follicles grow synchronously until ovulation. At the end of the intermolt, the oocyte is no longer attached to the follicular tissue because of the retraction of the macro- and microvilli 39. Ovulation transpires by means of its own movements during mating and following ecdysis 40. The breakdown of the germinal vesicle occurs some hours before ecdysis in O. gammarella44.

Figure 1.

Lateral view of the position of the neuroendocrine X organ/sinus complex, endocrine organs, and the reproductive system of a gammaridean amphipod in anatomical relation to some other visceral structures. Modifed after Charniaux-Cotton and Payen 40. ag = Androgenic gland; gn2 = second gnathopod; h = heart; i = intestine; mo = mandibular organ; nc = nerve cord; o = oostegite; os = ovigerous setae; ov = ovary; ovd = oviduct; osv = oocyte in secondary vitellogenesis; sv = seminal vesicles; t = testis; vd = vas deferens; Xo/sg = X organ/sinus complex; 5,7 = thoracic segments.

In other amphipods in which oogenesis has been studied, oocytes have been shown to develop similarly over two molt cycles 45–47. In Gammarus duebeni, maturation occurs over two maternal molt cycles, but two series of oocytes are always developing out of phase, so that each series of oocytes develops synchronously to reach an endpoint at the end of each cycle. During the first molt cycle, the cohort of oocytes reach a mean diameter of about 60 µm. They continue to grow in the second cycle, and at a size of about 90 µm at the commencement of molt stage C the deposition of yolk commences. They reach a size of 200 µm by the beginning of premolt and a final size of about 400 µm at ecdysis 45. Sheader 46 confirmed this pattern for oogenesis in G. duebeni in another study. During the first intermolt period of development, the oocytes did not exceed 100 µm in diameter. After molting and subsequent shedding of the batch of large oocytes, the small oocytes began to grow with rapid yolk accumulation. Growth of these oocytes continued until they reached final diameters of about 500 µm, prior to ovulation and transfer along the oviduct into the marsupium after mating 46. During development and until ovulation, the number of oocytes in the ovary decreased by cytolysis and resorption 46. In contrast, Geffard et al. 38 noted that the number of oocytes visualized externally undergoing secondary vitellogenesis in G. fossarum for the next reproductive cycle was similar to the number of embryos observed in the marsupium of the gravid females. In the continuous-breeding amphipod Melita plumulosa, oocytes also develop over two molt cycles, with oocytes in primary vitellogenesis measuring up to 50 µm in length and a final size of about 330 µm at ecdysis (Fig. 2). For most gammaridean species, if mating does not occur, the female does not ovulate, and the oocytes are resorbed following the molt 29, 47. In some species, the females are able to release eggs into their marsupium even if no copulation occurs, but these eggs never develop and disappear from the marsupium within a few days 48–50.

Figure 2.

Longitudinal histological sections stained with hematoxylin and eosin of gravid female Melita plumulosa amphipods prepared for the reproduction toxicity test as described previously 19 with all females synchronous at the same stage of their ovarian cycle and molt cycles. (Top) A female amphipod on day 1 of the test, postmolt <24 h after ovulation with embryos at stage 1 in the marsupial pouch and the ovary displaying a cohort of oocytes in primary vitellogenesis only. (Bottom) Female on day 7 of the test at late premolt stage of the molt cycle with embryos at stage six in the marsupial pouch and the ovary containing oocytes that are approaching completion of secondary vitellogenesis and a new cohort of oocytes in primary vitellogenesis. opv = Oocyte in primary vitellogenesis; osv = oocyte in secondary vitellogenesis; em1 = embryo at stage 1 of development; em6 = embryo at stage six of development. Scale bars = 100 µm.

Spermatogenesis in amphipods

The location and gross anatomical appearance of the testis, seminal vesicle, and vas deferens of the marine amphipod E. marinus after isolated by dissection has been described by Ford et al. 36. Various stages of sexual differentiation in the amphipod O. gammarella have been studied in detail by Charniaux-Cotton and Payen 35. Although females of gammarid amphipods are sexually receptive for only a brief period of the molt cycle, males are considered available for mating during most of their molt cycle 51. The G. pulex male has limited numbers of spermatozoa in its gonads, which has been estimated at 10,000 gametes per testis 52, which is in the same range reported for G. duebeni53. Lemaître et al. 52 found substantial sperm allocation to each reproductive event but also a relatively fast replenishment. Depletion of G. pulex spermatozoa from copulation had no impact on male reproduction, because males could engage in precopula within a few hours of a previous copulation 52. In contrast, G. duebeni males were shown to have reduced sperm counts after three consecutive matings that resulted in decreased female fecundity 53. Lower sperm counts have been reported in intersex males of the intertidal amphipod E. marinus54. In addition, intersex male C. volutator when paired with normal females were reported to fertilize fewer oocytes compared with normal males 55.

Reproductive behaviors of amphipods during mating

Female amphipods have been reported not to store spermatozoa, so a female must be accompanied by a male at each molt so that copulation and fertilization of the current brood will occur in a timely manner 49, 56. Precopulatory mate-guarding behavior (amplexus) is thought to have evolved in free-roaming epibenthic amphipods as a male competitive strategy in response to the brief period of female receptivity 27. Gammarid amphipod females typically produce several broods in succession during the warmer months in the temperate zone. Generally, amplexus begins toward the end of the female intermolt period and continues until the female molts. After her molt, a new marsupial brood chamber is formed under her ventral surface, and copulation occurs within minutes afterward, with the male injecting spermatozoa into the newly formed marsupial chamber while it is still flexible before it hardens together with the exoskeleton 27, 48. Then, within 1 h, ovulation occurs and subsequently fertilization is able to proceed externally in the marsupium 29. The male and female then separate. All the eggs of a gravid female brood are deposited in the female marsupial brood pouch simultaneously at ovulation 27, 57. After fertilization of the gametes, embryo development occurs in the marsupium, and the juveniles hatch and are released from the marsupium shortly before the female is due to molt again. The times of hatching and juvenile emergence are species specific. The female remains alone until a few days before her next molt, when amplexus is reinitiated and the cycle is repeated. Fecundity is significantly reduced if copulation is delayed for as short a time as 6 h in the amphipod crustacean Gammarus palustris29. Injections of female gammaridean amphipods with the crustacean molting hormone 20-hydroxyecdysone or ecdysterone hastened the onset of the molt and receptive sexual behavior 58. The female's behavior and its exoskeleton influence the male's behavior. When touched by a male, females either swam away or assumed one of three body postures during amplexus 59. One posture assumed by females is expressed only by newly molted females and always precedes copulation 59. Evidence for the importance of the exoskeleton as a stimulus for copulation was provided by the finding that males blinded with paint copulated with freeze-fixed animals but only with those that were newly molted females 59.

In a study on the male mating behavior in G. pulex, males were found to invest a large quantity of spermatozoa in each reproductive event relative to their sperm reserve, but the sperm count recovered relatively rapidly within the period of precopula duration in either wild populations or amphipods in the laboratory. The males that had recently mated were more likely to pair again than those males that had not mated recently, independently of reserves of spermatozoa in the testes, male size, or energy storage 52. In Eogammarus oclairi, several parameters associated with population structure affect the male investment in mate guarding. The male reproductive response was found to be dependent on male–male interactions. Males respond to the relative size of males, the ratio of sexually competing males to females that are ready to mate (operational sex ratio), and the density of males in experimental populations. Males guard longer when females are a limiting resource or when the rate of encounter with other males is higher. Mean guarding time varied between a few hours when the operational sex ratio (male:female) was 1:3 to >7 d when the sex ratio was 2:1. It was also affected by male density, increasing as male density increased, suggesting that the stimulus to guard depends on the intensity of male–male competition 60.

Other amphipod species exhibit different pairing behaviors for mating. Mature adults of the tube-building amphipods Ampelisca vadorum and A. abdita leave their individual tubes in the substratum and swim freely in the water, and the male seeks out the female in synchrony with lunar periodicity. Copulation probably occurs at that time 61. In other tube-building amphipods, such as Microdeutopus gryllotalpa and Corophium volutator, females generally reside alone in individual tubes at the beginning of their intermolt periods. However, a female will permit a male to enter her tube toward the end of the intermolt. Copulation occurs shortly after the female molts, and the male then leaves the tube 62–64. Corophium volutator females release gender-specific, waterborne pheromones that guide males 65.

Spermatophores, sperm transfer, and sperm storage by females

The fertilization process in amphipods in general is poorly understood. Fertilization is external, which could result in substantial mixing of spermatozoa from multiple males, but the guarding behavior of gammaridean amphipods during amplexus would limit this occurence. During copulation in G. palustris, the male holds the female perpendicularly, intromission occurs two or three times, and then the pair separates 49. After copulation, the female then ovulates and deposits her oocytes into the marsupium a few minutes after copulation while the exoskeleton is still fairly flexible to allow their passage through the oviducts 29. Eggs and spermatozoa are considered to meet in the marsupium chamber, formed externally beneath the abdomen prior to fertilization, but the site of fertilization in amphipods has not been clearly demonstrated for most species. The sequence of events during fertilization has also been described for the tube-dwelling amphipod Lembos websteri50. Males enter the tube of a receptive female, for only as long as mating necessitates, after release of her brood and a recent molting have occurred. The male and female take a tail-to-tail position with their ventral surfaces opposed. The urosome of the male is inserted into the female brood pouch, and spermatozoa are discharged 50. The female then lays eggs immediately after copulation. The eggs stick together in separate masses from each ovary initially within a gelatinous matrix.

In decapod crustaceans, such as shrimp, crayfish, and crabs, the immotile spermatozoa are typically surrounded by a protective and adhesive membrane structure referred to as a spermatophore, when transferred to the female 66, 67. Hastings 68 reported breeding males of the amphipod Ampelisca brevicornis extruding spermatophores from their papillae. Histological examination of these animals showed the vas deferens to be packed with ripe spermatophores. Reger and Fain-Maurel 69 described tubules surrounding bundles of spermatozoa in the vas deferens of the amphipods O. gammarella, Talitrus saltator, and Gammarus montaigui, but whether they function as spermatophores is unknown. Sundelin and Eriksson 70 presented images of spermatophores containing spermatozoa from the deposit-feeding amphipod Monoporeia affinis. Amphipod spermatozoa have been described to have a very long flagellum that display no movement if isolated in solution 54, 71 but slow movement when in close proximity to eggs removed shortly after discharge into the marsupium 72, 73. Female amphipods have been commonly reported not to store spermatozoa across the molt 49, 56. However, we observed that isolated female M. plumulosa, which have a continuous 7-d ovarian cycle, produce male and female juveniles, presumably using stored spermatozoa from a mating in a previous reproductive cycle (R.M. Mann and R.V. Hyne, Centre for Ecotoxicology, Lidcombe, NSW, Australia, unpublished data). Eight of twenty gravid females initially isolated produced a second additional brood of viable juveniles once either two or three weeks after releasing their first brood, which was three or four weeks after initially being isolated from males. Asexual reproduction has been reported to occur in Corophium bonnellii50, 74, when no males have been observed in field populations examined. Another amphipod species, Stegocephalus inflatus, has been shown to be a hermaphrodite, beginning adult life as a male and becoming female after one or two molts 75.


Neuroendocrine control of molting and reproduction

Crustacean hormones are either neuropeptides secreted by the X organ/sinus gland complex or a steroidal, glycopeptide, or terpenoid hormone secreted by one of the three main endocrine tissues: the Y organ, the androgenic gland or ovary, or the mandibular organ, respectively (Fig. 3). The X organ/sinus gland complex is the main neuroendocrine organ of crustaceans. A variety of neuropeptides 76 is synthesized in the X organ and transported to the sinus gland, where these neuropeptides are stored and released into the hemolymph. The anatomical relationship of the sinus gland as a separate storage structure connected to the X organ by a dense nervous tissue was first described by Bliss and Welsh 77. Both organs are paired, and they serve as a key endocrine junction of central neurosecretory signals. The X organ is part of the medulla terminalis in the eyestalk in decapods and projects axons to the sinus gland, which is located in the eyestalks. The neuropeptides are produced in the eyestalks of decapods, but in crustaceans with no differentiated eyestalks, including amphipods, the X organ/sinus gland complex is located in homologous structures of the central nervous system in the head 35, 78–80. The peptide hormones released from the X organ have diverse effects on growth, molting, and reproduction and have been studied extensively over the last two decades, initially by eyestalk removal or effects from transplantation or from the addition of organ homogenates (for reviews, see 81–83).

Figure 3.

Schematic representation of the major endocrine organs, their secreted hormones, and their target tissues involved in male and female reproduction in amphipods. Stimulatory effects are indicated with solid arrows, and inhibitory effects are indicated with dashed arrows. The X organ/sinus gland complex downregulates the ovary and androgenic gland by releasing gonad/vitellogenesis-inhibiting hormone (GIH/VIH), the Y organ by releasing molt-inhibiting hormone (MIH), and the mandibular organ by releasing mandibular organ-inhibitory hormone (MOIH). The X organ/sinus gland complex also upregulates the ovary and androgenic gland by releasing vitellogenesis/gonad-stimulating hormone (V/GSH). The androgenic gland produces and secretes androgenic gland hormone (AGH), which acts on the testis to promote male sex development and male reproduction. The Y organ produces and secretes the prohormone α-ecdysone (α-E), leading to the active steroid 20-hydroxyecdyone (20-HE) that initiates formation of the new cuticle. Low titers of 20-hydroxyecdyone and secretion of V/GSH by ovarian follicles in secondary vitellogenesis act on adipose tissue to initiate production of vitellogenin to be taken up by the oocytes after transport via the hemolymph. The mandibular organ produces and secretes methyl farnesoate (MF), which possibly up-regulates the Y organ.

Information concerning these neuropeptides has been accumulating from many species, although the number of these neuropeptides in a single species is still uncertain 76. Currently, the many members of this gene family can be grouped into either the crustacean hyperglycemic hormone (CHH) or the molt-inhibiting hormone (MIH) subtypes 76, 84. All CHH molecules are classified as subtype I, and the MIH, mandibular organ-inhibiting hormone (MOIH), and gonad (or vitellogenesis)-inhibiting hormone (GIH/VIH) molecules are classified as subtype II 76.

Y organ and regulation of molt cycle and vitellogenesis

In many malacostracan crustaceans, including amphipods, spawning must be preceded by a molt. Regulation of the molting cycle in crustaceans involves two endocrine organs: the X organ/sinus gland complex and the Y organs, a pair of epithelioid glands located in the cephalothorax first described by Gabe 85. The Y organs are nonneural endocrine organs that produce and secrete ecdysone from cholesterol 86. In the case of crustaceans, ecdysone is peripherally transformed into the active hormone 20-hydroxyecdysone. The synthesis and secretion of ecdysone by the Y organs are negatively regulated by the neuropeptide MIH 81 and may be positively regulated by the crustacean juvenile hormone methyl farnesoate 26. In the course of the molting process, the ecdysteroid titer increases continuously and reaches its maximum shortly before apolysis, the separation of the old cuticle from the epidermis. During the molt, proteases and chitinases are synthesized by epidermal cells and accumulate in the molting fluid between the epidermis and the old cuticle 87.

The role of the Y organ in the initiation of molting and vitellogenesis in amphipods was first demonstrated in O. gammarella, in which vitellogenin appears in a low concentration in the postmolt stage. Its concentration increases through the intermolt and maintains a steady level until the D2 premolt stage and then diminishes before exuviation of the old exoskeleton during ecdysis. This fluctuation in vitellogenin level is paralleled by similar fluctuation in the level of hemolymph ecdysteroids 88, 89. Ecdysteroids have also been analyzed during the 46-d molt cycle of adult males of the terrestrial crustacean Orchestia cavimana90. During the molt cycle of O. cavimana, minimal ecdysteroid concentrations occurred immediately after ecdysis, and a very sharp peak concentration occurred just prior to molting at the end of Dl molt stage, producing a 230-fold increase compared with the middle of intermolt 90.

Ecdysteroids also perform key roles in crustacean reproduction 26. The importance of ecdysteroids in vitellogenin synthesis in O. gammarella was revealed by the Y organectomy that resulted in the depression of vitellogenin synthesis and the consequent retardation of ovarian growth 91. Embryonic ecdysteroids were present in early stage I embryos of the amphipod Leptocheirus plumulosus and increased with advancing embryonic development 92. This pattern supports other observations that generation of an embryonic Y organ can supplement maternally derived ecdysteroids 92.

Endocrine regulation of the ovarian cycle

Amphipods are gonochorists, and the function of mature gonads of females and males is regulated by different hormones 78. Ovarian oogenesis is a major reproductive process, during which oocytes grow as a result of vitellogenin production and its accumulation in the form of yolk protein (vitellin) in the oocytes 40. The ovarian and reproduction cycles in females of gammaridean amphipods are also closely correlated with molting and male copulation, because it is known that the ability of females to ovulate and mate is restricted to the early postmolt period 26, 29, 58. Also, embryogenesis is linked to the molting cycle, because the first molting of females carrying their fertilized eggs after copulation occurs only after the embryos have fully hatched 29.

The ovarian cycle is regulated both negatively and positively by CHH/MIH/GIH gene family neuropeptides secreted by the X organ/sinus gland complex. In amphipods, neurosecretory cells of the median part of the protocerebrum secrete a neurohormone inducing secondary folliculogenesis and triggering secondary vitellogenesis. The neurohormones controlling secondary vitellogenesis in amphipods have not yet been isolated, and their mode of action remains unclear 40. The adipose tissue has been shown to be the site of vitellogenin synthesis in the amphipod O. gammarella. Immunoenzymatic techniques and injection of tritiated leucine indicate that the fat body of O. gammarella contains and releases vitellogenin only when the ovaries are in secondary vitellogenesis 93, 94. In O. gammarella, secondary folliculogenesis requires a low titer of ecdysteroids, which characterizes postecdysis. After secondary folliculogenesis, the secondary follicle cells become endocrine and secrete a vitellogenin-stimulating ovarian hormone (V/GSH) 40, 95. Junéra et al. 95 found that, if females of O. gammarella were bilaterally ovariectomized, synthesis of vitellogenin completely ceased 5 to 8 d after the operation and was restored by an ovarian transplant.

Androgenic gland hormone and regulation of spermatogenesis

The discovery of the role of the androgenic gland in the amphipod O. gammarella by Charniaux-Cotton 96 paved the way for studies on the function of this gland. The androgenic gland is a paired structure, one of the pair being attached to each vas deferens in most orders of Malacostraca 35. The androgenic gland hormone (AGH) regulates sexual differentiation of the male reproductive system in malacostracan crustaceans, such as amphipods, and also its functioning and the development of the male secondary sexual characteristics 81. In males, the primordial androgenic glands develop and synthesize the AGH, which induces male sexual differentiation. In females, however, the primordial androgenic glands do not develop, and female sexual differentiation is induced spontaneously (for reviews, see 35, 97). Reproduction of mature crustacean males involves the androgenic gland. The testes are directly controlled by the AGH secreted from the androgenic gland, with gonad-stimulating hormone (GSH) and gonad-inhibiting hormone (GIH) influencing the secretion of AGH, instead of directly acting on the testes. The androgenic hormone in the isopod Armadillidium vulgare consists of two peptide chains, A and B, composed of 29 and 44 amino acids, respectively, linked by two disulfide bridges and carrying one N-glycosylation 98. Mass spectrometry analysis provided evidence for a structural homologue of AGH in the androgenic glands of the amphipods G. duebeni99 and E. marinus36 that likely corresponded to the glycosylated A-chain peptide of AGH characterized in A. vulgare.

Mandibular organ

The paired mandibular organ is located more anteriorly above the Y organ at the base of the tendon associated with the posterior abductor muscle of the mandibles and was first described by Le Roux 100. It was found subsequently to be the site where the terpenoid methyl farnesoate (MF) is synthesized 101. Methyl farnesoate is produced by mandibular organs of numerous crustaceans and is the unepoxidated form of the insect juvenile hormone (JHIII), which has been well studied and characterized 102. The synthetic activity of MO is negatively regulated by the neuropeptide mandibular organ-inhibiting hormone (MOIH), as evidenced by several eyestalk ablation studies in the shore crab 103. The functions of MF in crustaceans are elusive; a major difficulty in investigating the roles of MF has been the anatomical location of the MO, which has made its surgical removal difficult, so it has not been possible to perform ablation studies to investigate MF function. Experimental evidence suggests that major functions of MF in crustaceans are in the positive regulation of the molt cycle and enhancing reproductive maturation 102, 104. Methyl farnesoate has been implicated as a morphogen in postembryonic larval stages of crustacean development 104 and in reproductive maturation in decapods, in which it increases production of vitellogenin and oocytes 104, 105. Methyl farnesoate also stimulates gonadal maturation of both male and female decapods 104, 106. Olmstead and LeBlanc 107 also demonstrated that MF is a male sex determinant in daphnids.


Environmental contaminants can have unanticipated endocrine activity and interfere with hormonally regulated physiological processes. Evidence from field studies indicates that an endocrine disruption of sexual differentiation has been occurring in some amphipod populations. Studies on the deposit-feeding amphipod M. affinis showed that male sexual development and the development of olfactory sensilla on the antennules was delayed or disrupted when the animals were exposed to sediment from a lake presumably receiving endocrine disruptors from a refuse dump 108. The incidence of intersexuality was found to be significantly increased in the marine/estuarine amphipod E. marinus from polluted sites in eastern Scotland 109. Androgenic glands and a putative androgenic gland hormone (AGH) peptide were present in males. However, in both normal and intersex females, the androgenic glands were present only in a rudimentary form, and the AGH peptide was not detected. Intersex males were found to possess abnormal glands that appeared hypertrophied, and AGH was not detected. These data suggested that the intersex phenotype was manifested via perturbations of AGH 36. However, the underlying cause of the higher incidences of intersexuality in amphipods collected from the polluted sites was not determined.

Endocrine disruption resulting from chemical contaminants can take place with different modes of action. Endocrine-disrupting chemicals may interfere with hormonal signaling through directly binding to a hormone nuclear receptor known as the ecdysone receptor and acting as an agonist or antagonist 81, 82, 110. Ponasterone A and RH 5849 are two compounds that elicit 20-hydroxyecdysone-like activity in crustaceans 111. Ponasterone A is a phytosterol that was first isolated from plants and has high ecdysteroid activity in crustaceans and insects 112. Alternatively, some chemicals can elicit antihormonal effects by interfering with synthesis of the hormone. Inhibitors of cytochrome P450 enzymes have been shown to interfere with steroid hormone synthesis, resulting in reduced hormone levels and associated endocrine toxicity 110. Several studies have examined the effects of xenobiotics on the enzymatic pathway responsible for ecdysteroid metabolism in crustaceans 111, 113. Fenarimol, a fungicide that inhibits synthesis of ecdysteroid in arthropods 81, negatively impacted the deposit-feeding amphipod M. affinis, with decreased male mating and fertilization rates 114.

A wide range of contaminants has been implicated as endocrine-disrupting chemicals for amphipod species, including petroleum hydrocarbons and heavy metals 82, 115. Borowsky et al. 23, 24 found that reproductive females of the epibenthic amphipod Melita nitida developed abnormal oostegite setae and that their fecundity was significantly lower when exposed to low concentrations of polycyclic aromatic hydrocarbons associated with petroleum-contaminated sediment. Sundelin and Eriksson 70 reported high frequencies of malformed embryos in the deposit-feeding amphipod M. affinis, in which high concentrations of polycyclic aromatic hydrocarbons were measured in settling particulate matter and bottom sediments. An abnormal structure of oocytes in vitellogenesis was reported for the amphipod G. pulex sampled at river sites where the natural steroids 17β-estradiol and estrone had been detected, together with male fish producing vitellogenin, a typical biomarker of exposure to estrogenic substances 116. However, the cause of the pathologies in G. pulex is uncertain because, in the laboratory, crustaceans (crabs and shrimps) did not produce vitellin in response to estrogens 117. Gefford et al. 38 recently developed a reproductive toxicity test using the amphipod G. fossarum, encompassing molting, follicle growth, and embryonic development, to provide a better understanding of the effects of substances disrupting these processes. Cadmium was found to inhibit secondary vitellogenesis, and nonylphenol had a specific concentration-dependent effect on embryonic development; however, the mechanistic pathway by which these toxicants caused the toxicity was not determined.


Information obtained from the identification of the mechanistic pathways by which environmental contaminants cause reproductive toxicity in amphipods could be used to distinguish suitable biomarkers for evaluating reproductive toxicity in amphipods. It is evident from this review that disruption by environmental contaminants of two critical physiological processes in amphipods, namely, molting and secondary vitellogenesis, would cause reproductive toxicity. As discussed previously, the ovarian and reproduction cycles of female gammaridean amphipods are closely correlated with molting. The ability of males to copulate and subsequently of females to ovulate is restricted to the early postmolt period of the females; hence, fecundity must be preceded by a molt 48, 58. Molting in crustaceans is regulated by a multihormonal system but is under immediate control of the steroid hormone 20-hydroxyecdysone produced in the Y organs. An increase in 20-hydroxyecdysone concentrations mediates a cascade of gene-regulatory events mediating molting and reproduction. The ecdysone receptor is a ligand-dependent transcription factor, and it activates transcription of target genes by forming a heterodimer with another nuclear receptor, the ultraspiracle protein 118. Binding to the DNA helix is mediated by the nuclear receptor that encompasses two zinc-finger polypeptides, where Zn complexes four cysteine amino acids, thereby facilitating the folding of a protein domain involved in the DNA–protein interaction 119. Displacement of zinc in the DNA binding protein domain by Pb and Cd has been proposed as a mechanism causing toxicity of these metals 120, 121, but whether these metals cause reproductive toxicity in amphipods by this mechanism has not been determined. Other nuclear receptor genes have been identified in the Daphnia pulex genome, but little is known of the function of the nuclear receptors of crustaceans 122. Some of these nuclear receptors possess ligand-binding sites and may also serve as targets for disruption by environmental contaminants. Gagné et al. 123 provided evidence that contaminants disturbed exoskeleton integrity and vitellogenesis in the Gammarus sp. collected from polluted sites in the field.

A major component of the crustacean cuticle is chitin, a polymer of β-1,4-linked sugar derivatives, predominantly N-acetyl-β-glucosamine (NAG). Degradation of cuticular chitin requires two chitinolytic enzymes from the epidermis, chitinase and N-acetyl-β-glucosaminidase. Chitinase, an endo-splitting enzyme, cleaves the chitin polymer into oligomer NAGs, which are further broken down to monomer NAGs by the exo-splitting chitobiase or N-acetyl-β-glucosaminidase 124. Chitobiase activity was found to correlate well with the profiles of ecdysteroids during the molting cycle of the crab Uca pugilator125 and to peak within 6 h prior to ecdysis in Daphnia magna124. In addition, injection of the molting hormone 20-hydroxyecdsyone resulted in a significant increase in epidermal chitobiase activity 126. Very little is known about the direct impacts of environmental chemicals on chitobiase activity in amphipods. Gagné et al. 123 found that gammarid amphipods from contaminated sites contained significantly higher levels of acid-soluble proteins in their exoskeletons and a decreased proportion of chitin.

Changes in the concentrations of vitellogenin, the energy-rich egg-yolk protein precursor, has been proposed as a biomarker to characterize the maturation stage of oocytes in female crustaceans and as an indicator of reproductive impairment 127, 128. Studies reporting the induction or regulation of vitellogenin in amphipods are limited. Xuereb et al. 129 investigated the expression of the vitellogenin gene in a freshwater amphipod, G. fossarum, using calibrated real-time reverse transcription–polymerase chain reaction. Females expressed from 200 to 700 times more vitellogenin transcripts than males, depending on the female reproductive stage. Female amphipods showed significant elevation of vitellogenin mRNA levels at the end of the intermolt phase and at the beginning of the premolt phase. The impact of experimental or environmental exposure to chemicals on female vitellogenin gene expression has not been assessed. In another study using a liquid chromatography–tandem mass spectrometry-based assay to quantify vitellogenin concentrations in G. fossarum, vitellogenin synthesis was inhibited after exposure to methyl farnesoate 130.

One approach for the development of new biomarkers for reproductive toxicity is the application of the new molecular technologies that are providing insights into mechanisms of toxic action of chemicals and are likely to be useful tools in biomarker discovery. DNA microarrays are diagnostic tools that allow gene transcript expression changes in genomes to be assessed in response to environmental stimuli. Identification of specific gene functions of amphipods for development of gene transcript biomarkers of chemical exposure is presently at an early stage of development, and the only aquatic invertebrate whose genome has been sequenced is D. pulex ( A limited number of expressed sequence tag studies for a few other crustaceans have yielded some cDNA libraries for the amphipod Gammarus pulex, representing a range of developmental stages and phases of the molt cycle for both sexes together with sex-specific subtractive suppression hybridization libraries 131. A microarray has recently been constructed from expressed sequence tags from Hyalella azteca [] for the evaluation of gene expression signatures after exposure to different sediment pollutants and to correlate changes with whole animal toxicity.

Another approach to identify a biomarker for monitoring the effects of environmental contaminants on amphipod reproduction is to use proteomics to measure changes in protein expression of the females during their reproductive cycle. The use of proteomics to detect the effects of contaminants on physiological mechanisms is a relatively new and sensitive approach that can assist in understanding the impacts of xenobiotics on aquatic organisms 132–134. Proteomics has much to offer even in species used in ecotoxicological studies that are poorly represented in sequence databases. Protein spots can be matched by searching sequence databases for similar but not necessarily identical proteins. To date, the proteomic studies with aquatic invertebrates that have not been sequenced have identified on average only 15 proteins, highlighting the problem with matching amino acid sequences across species 134. Recently, a few studies have appeared using proteomics in ecotoxicological research on crustacean species with nonsequenced genomes, including three amphipod species 135–137. Leroy et al. 135 detected more than 560 protein spots on two-dimensional polyacrylamide gel electrophoresis and found a total of 21 proteins exhibiting significant expression differences in G. pulex exposed to two polychlorinated biphenyl congeners, with 14 of these proteins identified by mass spectrometry. In a study comparing the brains of two amphipod species, G. pulex and G. insensibilis, infected with parasites using proteomic tools, Ponton et al. 136 detected over 500 protein spots in each species using image analysis software and identified over 30 proteins. Ralston-Hooper et al. 137 reported on the use of proteomics in an interspecies comparison and demonstrated that the amphipods H. azteca and Diporeia spp. responded with similar proteomic profiles when exposed to atrazine and its metabolite desethylatrazine.


Amphipods are generally recognized as influential test organisms for many sediment toxicity assessments, and amphipod toxicity test results can correlate positively with gradients of sediment contamination and an impairment of the in situ benthic community 138–140. Reproduction rarely is used in ecotoxicological testing but has been shown to be reduced significantly and generally more sensitive compared with survival 15, 141 and length 16. A more basic understanding of mechanisms of reproductive toxicity is required to support extrapolation of what is being measured in reproduction toxicity tests to the ecological responses in the environment. If the xenobiotic affects fecundity, the focus of future research should be on understanding the mode of action of the contaminant by which it is affecting reproduction and to distinguish the toxicant effect from other noncontaminant stressors. Because other factors such the nutrient status can also affect amphipod fecundity, a mechanistic understanding will facilitate more accurate predictions of the effects of xenobiotics on field populations. The recent literature suggests that metabolic pathways associated with ecdysteroid hormonal signaling or metabolism, exoskeleton maintenance and molting, and the regulation of vitellogenin have potential as biomarkers of reproductive toxicity to amphipods caused by contaminants. However, more research is needed to determine the specificity of these biomarkers, the external/internal factors influencing the response, and the pathways affected by the contaminants.


The author is grateful to R. Mann and Y. Kobayashi (Office of Environment and Heritage) and M. Byrne (University of Sydney) for their help in preparing the histology sections and the digital images. The anonymous referees and corresponding editor are also thanked for thoughtful comments that improved the manuscript.