SEARCH

SEARCH BY CITATION

Keywords:

  • blood vessels;
  • evolution of viviparity;
  • confocal microscopy;
  • embryonic development;
  • chorioallantois;
  • skinks

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. LITERATURE CITED

The evolution of viviparity requires modifications to multiple integrated physiological features to support embryonic development during pregnancy. Embryonic growth during pregnancy is dependent upon the capacity of the uterine vascular system to satisfy increasing embryonic oxygen demand throughout gestation. We tested the hypothesis that total surface area of uterine blood vessels increases in concert with embryonic growth, and hence its oxygen demand, during gestation. We used immunofluorescence and laser-scanning confocal microscopy to quantify uterine microvascular density and morphology during gestation in the oviparous skink Ctenotus taeniolatus and in Saiphos equalis, a skink species with prolonged egg retention. For C. taeniolatus, vessel density (Nv) and vessel length-density (Lv) in the embryonic hemisphere of the uterus is 23% and 17% less, respectively, than that of S. equalis and vascular surface-area does not differ as a function of embryo stage. For S. equalis, overall Nv, Lv, and vessel diameter (Dv), does not change during the first half of gestation but increases by 36% (Nv), 44% (Lv), and 60% (Dv) by near-term embryo stages late in gestation. The chorioallantoic membrane of S. equalis increases in absolute size but vascular density does not differ as a function of embryo stage. The marked increase in uterine vascular density during late gestation coincides with the phase of rapid growth in embryo mass and concomitant increase in metabolic rate. Expansion of the uterine vascular bed in concert with embryo size and metabolism is likely to be an important transitional step in the evolution of viviparity. Anat Rec, 293:829–838, 2010. © 2010 Wiley-Liss, Inc.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. LITERATURE CITED

Squamate reptiles (lizards, snakes, and amphisbaenians) exhibit a diverse range of reproductive life histories ranging from oviparous reproduction where eggs are laid at relatively early embryonic stages, to viviparous reproduction where embryonic development is completed within the female. Most squamate reptiles are oviparous, with oviposition typically occurring when embryos have achieved ∼25%–40% of their total development, which corresponds to stages 28–31 (Shine,1983; DeMarco,1993) of the 40 stage scheme of Dufaure and Hubert (1961). Approximately 20% of squamate species are viviparous, giving birth to precocial young at parturition (Andrews and Mathies,2000). Viviparity has evolved independently numerous times in squamate reptiles (Blackburn,1982; Shine,1985) and presumably occurs via selection for transitional stages of increasingly prolonged embryonic development in utero (Packard et al.,1977; Shine,1983,1985). Increases in duration of egg retention require a suite of integrated physiological features to both maintain the gravid state as well as support growth and differentiation of retained embryos (Guillette and Jones,1985; Andrews,1997; Mathies and Andrews,1999; Andrews and Mathies,2000).

Embryonic growth and development during gravidity/pregnancy is ultimately dependent upon oxygen supplied by the maternal uterine vascular system (Carter,2000; Falkowski et al.,2005). In lizards, oxygen availability is a primary factor that determines the amount of embryonic development that occurs in in utero (Andrews,2002; Parker et al.,2004; Parker and Andrews,2006). Oxygen diffuses from the uterine vascular bed to the embryonic blood supply in the vascularized extraembryonic membranes of the embryo. Embryonic growth increases throughout gestation, with an increase in embryonic mass and metabolic oxygen demand during the latter half of embryonic development (Thompson and Stewart,1997; Vleck and Hoyt,1991). The capacity of the uterus to match increasing metabolic demands of embryos as development progresses is therefore one of the major determinants of maximum embryonic stage at which squamate eggs are laid.

One of the most effective mechanisms for enhancing oxygen availability to embryos is by increasing the surface area for gas exchange via elaboration of the vascular bed of the uterus and extraembryonic membranes, particularly the chorioallantois. The proliferation of blood vessels from pre-existing vascular structures is called angiogenesis (Risau,1997). In addition to angiogenesis, increases in vessel diameter resulting in elevated uterine blood flow (Rosenfeld,1977; Reynolds and Redmer,2001) as well as enhanced oxygen binding affinity of embryonic blood (Grigg and Harlow,1981; Ingermann,1992) are also mechanisms by which oxygen demands of developing embryos are satisfied. In reptiles, increases in vascular surface area is associated with the evolution of complex placentation, but quantitative studies comparing uterine microvascular architecture in oviparous and viviparous species are rare (Guillette and Jones,1985; Masson and Guillette,1987). For example, in two closely related species of Sceloporus lizards, higher vascular density is associated with the viviparous species (S. bicanthalis) compared to the oviparous species (S. aeneus), although no information on the ontogeny of vascular proliferation associated with embryonic growth in utero is provided (Guillette and Jones,1985).

The purpose of our study is to test the hypothesis that density and total surface area of uterine blood vessels increases in concert with growth of the embryo, and hence its oxygen demand, during gravidity/pregnancy. The use of common techniques, such as vascular casts, to quantify microvascular architecture is constrained by the small size of most appropriate model reptile species. To overcome this limitation and meet our objective, we used indirect immunofluorescence and laser-scanning confocal microscopy to quantify uterine microvascular density and morphology in two species of Australian Sphenomorphous group skinks–Ctenotus taeniolatus and Saiphos equalis. These species were chosen because of their differing capacities to support embryonic development during egg retention. Ctenotus taeniolatus is a typical oviparous species which oviposits at a maximum embryo stage of 31 (S.L. Parker, unpublished data). Saiphos equalis oviposits eggs with relatively thin shells, but has the unusual capacity to retain eggs until embryonic development is complete (i.e., Stage 40; Smith and Shine,1997). Moreover, S. equalis has a simple placenta (sensu Weeks,1935) and therefore analyses of the vascular bed required for gas exchange are not confounded by substantial placentotrophy. On the basis of our knowledge of the developmental stage attained by embryos during gestation, we predict that density and total surface area of blood vessels increase during gestation in S. equalis, but not in C. taeniolatus.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. LITERATURE CITED

Collection and Husbandry of Lizards

Gravid females of C. taeniolatus (N = 10) were collected from The Royal National Park (34° 7′ S, 151° 03′ E) and S. equalis (N = 17) from populations in the vicinity of Sydney (33°, 59′ S, 150°, 03′ E) and from Riamukka State Forest (31°, 20′ S, 151°, 39 E), New South Wales, Australia. Females were held in cloth bags and transported within a few days of capture to an animal facility at The University of Sydney. Gravid females of C. taeniolatus were housed in glass aquaria and gravid females of S. equalis were housed in plastic containers until the appropriate reproductive stage was achieved. Daily photoperiod was provided by fluorescent room lighting (7.00–18.00 hr). The room in which females were housed was maintained at a constant temperature of 20°C. Glass aquaria were provided with a 40 W incandescent light suspended at one end of each container (9.00–17.00) as a heat source. Plastic enclosures containing gravid female S. equalis were heated on one end of the enclosure using electric heating tape. Females of both species were fed crickets dusted with vitamin mineral supplement every other day and provided with water daily.

Tissue Harvest and Fixation

Gravid females of C. taeniolatus were allocated to two reproductive stages (early: <Stage 25; N = 5, and mid: stages 26–30; N = 5) for sampling. Because C. taeniolatus typically lays eggs at a maximum embryo stage of ∼31, it was not possible to obtain uterine tissues at embryo stages beyond Stage 31. Gravid females of S. equalis were allocated to four reproductive stages (early: <Stage 25; N = 3, mid: Stages 26–34; N = 7, late: Stages 35–38; N = 3, and near term: Stages 39–40; N = 4 females) for sampling. The reproductive stages were categorized based upon major developmental events outlined in Dufaure and Hubert (1961) that describe the generalized temporal pattern of reptilian development (Andrews,2004). Briefly, “early” refers to the initial stages of development including neurulation until the start of organogenesis; “mid” refers to the majority of organogenesis and early embryonic growth phase including formation of limb buds; “late” refers to the late growth phase where in addition to growth in mass, scalation and pigmentation are completed; “near term” refers to the final portion of the late growth phase where embryonic differentiation is essentially complete but growth in embryo mass continues. At the appropriate reproductive stage, females were anesthetized with an intraperitoneal injection of 6 mg mL−1 sodium pentabarbitone, then euthanized by cervical dislocation. Both uteri with eggs were removed from the female and immediately fixed with eggs intact in 10% neutral buffered formalin for 24 hr. After fixation, uterine tissues were rinsed twice in distilled water and stored in 70% ethanol at 4°C. A single uterus with eggs from each female was allocated for vascular analyses.

Immunofluorescence and Confocal Microscopy

Uterine vasculature was visualized using immnofluorescence and laser-scanning confocal microscopy on whole-mount tissue preparations. Immediately before processing for immunofluorescence microscopy, eggs were removed from the uterus and uterine tissues rinsed twice in 0.1 M phosphate buffered saline (PBS, pH 7.2) (10–15 min/rinse). Eggs removed from each uterus were dissected and embryos and extraembryonic membranes removed. Embryos were staged according to the Dufaure and Hubert (1961) staging system to determine the developmental stage at the time of sampling. After staging, embryos were dried to a constant mass at 40°C and weighed to the nearest 0.001 g. For eggs containing embryos at Stages > 31, the chorioallantoic membrane was removed and rinsed twice in 0.1 M PBS for immunofluorescence microscopy. Uterine and chorioallantoic tissues were permeabilised in a solution of 1% bovine serum albumin (BSA), 0.1% TritonX-100, 0.05% Tweene-20, in PBS (pH 7.2) overnight at 4°C. Following permeabilization, tissues were rinsed twice in PBS (10–15 min/rinse), then subjected to antigen retrieval using 20 μg mL−1 proteinase-K in TE buffer (pH 8.0) 3–4 hr at room temperature. Autofluorescence was quenched by incubating tissues in 0.1 M glycine in PBS for 1 hr at room temperature. Tissues were incubated with rabbit-anti-human von Willebrand Factor antibody (Dako, Carpinteria, CA) at a final concentration of 62 μg mL−1 in 1% BSA in PBS) for 24 hr at 4°C. Von Willebrand Factor is a glycoprotein that is constitutively expressed in endothelial cells and functions in vertebrate hemostasis (Sadler,1998). After incubation with primary antibody, tissues were rinsed twice (30 min/rinse) in PBS, then incubated in goat-anti-rabbit secondary antibody conjugated to FITC (Zymed Laboratories, San Francisco, CA) at a final concentration of 5 ng μL in PBS for 24 hr at 4°C. Uterine and chorioallantoic tissues were rinsed twice (30 min/rinse) in PBS then mounted in fluorescent mounting medium (Dako, Carpinteria, CA). Negative controls to detect non-specific binding of secondary antibody were treated in the same manner as experimental tissue sections except that the primary antibody was replaced with BSA/PBS.

Stacks of images were obtained from stained tissues using an Olympus Fluroview 1000 laser scanning confocal microscope with 488 nm laser excitation and 525–555 nm emission filter. Images were obtained using a 20× lens and with a pixel size of 1.24 × 1.24 μm. Microvascular density and morphology was sampled by obtaining images (6 fields of view/uterus, 635 × 635 μm per field of view at 200× magnification) along linear transects covering the surface of the uterus and chorioallantoic membrane. At least three images each were obtained from the embryonic and abembryonic regions of the uterus (Fig. 1). Image stacks were computationally reassembled to provide three-dimensional reconstructions of vasculature architecture of uterine and chorioallantoic tissues. An index of vessel density (Nv; number vessels 0.4 mm−2), vessel length-density (Lv; mm 0.4 mm−2), and vessel diameter (Dv) were measured using ImageJ image analysis software (National Institute of Health, Bethesda MD). Vessel density was measured by superimposing a grid (72 × 72 μm grid squares) over each field of view and counting the number of vessel segments that touched each vertical line on the grid. Vessels were marked and counted using the Cell Counter plug-in feature of ImageJ. Vessel length-density was determined by tracing vessels falling within in each field of view and measuring the total length of vessel segments using the NeuronJ plug-in feature of ImageJ. Mean vessel diameter per field of view was obtained by averaging the diameter measurements of five randomly selected vessels within each field of view in embryonic and abembryonic regions of the uterus.

thumbnail image

Figure 1. Representative diagram of a typical lizard uterus containing egg and embryo. Region of uterus above dashed line is the embryonic hemisphere, and the region of uterus below the dashed line is the abembryonic hemisphere. UT, uterus; CAM, chorioallantoic membrane; SM, shell membrane; BV, blood vessel.

Download figure to PowerPoint

Data Manipulation and Statistical Analysis

Statistical analyses were conducted using SAS Statistical Package version 9.1.2 (SAS Institute,2004). Contrasts of microvascular density and surface-area (Nv, Lv, and Dv) between species and among embryo stages were evaluated using a two-factor analysis of variance (ANOVA). For C. taeniolatus, the effect of reproductive condition on uterine Nv, Lv, and Dv was analyzed using a Student's t-test. For S. equalis, the effect of embryo stage on uterine Nv, Lv, and Dv was analyzed using a single factor ANOVA. Analyses of the effect of embryo stage on uterine vascular proliferation were based upon the average of data obtained from three fields of view for each embryonic and abembryonic region of the uterus. The effect of embryo stage on chorioallantoic vascular proliferation was based on the average of data from three fields of view obtained from each chorioallantoic membrane. Data are reported as mean ± SE unless otherwise noted and probability values less than 0.05 were considered statistically significant.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. LITERATURE CITED

Uterine Microvascular Architecture in C. taeniolatus

In C. taeniolatus, there is no significant variation in overall Nv, Lv, or Dv during the period of gravidity (Table 1; Figs. 2A,C,E, 3A,C,E,G). There is correspondingly little change in embryo mass during the period of egg retention, with an overall increase in dry mass of 0.37 mg from early developmental stages (<Stage 25) to that at oviposition when embryos reach stages 27–31 (Fig. 2G ). Vessel density and surface area, however, are higher in the embryonic hemisphere of the uterus than in the abembryonic hemisphere (Tables 1 and 2; Fig. 3A,C,E,G). When embryonic and abembryonic hemispheres of the uterus are analyzed separately, no differences in Nv, Lv, or Dv as a function of embryonic stage is detected (Table 2). Correspondingly, there is little apparent change in vessel morphology indicative of angiogenesis (Fig. 3A,C,E,G).

Table 1. Statistical tests (2-way ANOVAs) and results of contrasts for overall uterine vascular density (Nv), length-density (Lv), and vessel diameter (Dv) as a function of embryonic stage and region (embryonic vs. abembryonic) of the uterus for Ctenotus taeniolatus and Saiphos equalis
 ContrastResults
  1. Significant results are in bold.

C. taeniolatus
Nv  
 Embryonic stageF1, 16 = 2.4, P = 0.14
 LocationF1, 16 = 29.5, P < 0.001embryonic > abembryonic
 Stage x locationF1, 16 = 0.81, P = 0.38
Lv  
 Embryonic stageF1, 16 = 0.78, P = 0.390
 LocationF1, 16 = 24.4, P < 0.001embryonic > abembryonic
 Stage x locationF1, 16 = 1.6, P = 0.218
Dv  
 Embryonic stageF1, 16 = 0.62, P = 0.443
 LocationF1, 16 = 0.42, P = 0.562
 Stage x locationF1, 16 = 0.18, P = 0.674
S. equalis
Nv  
 Embryonic stageF3, 24 = 8.4, P < 0.001early = mid = late < near term
 LocationF1, 24 = 84.2, P < 0.001embryonic > abembryonic
 Stage x locationF3, 24 = 0.39, P = 0.764
Lv  
 Embryonic stageF3, 24 = 13.6, P < 0.001early = mid = late < near term
 LocationF1, 24 = 106.6, P < 0.001embryonic > abembryonic
 Stage x locationF3, 24 = 0.56, P = 0.644
Dv  
 Embryonic stageF3, 24 = 6.7, P = 0.002early = mid < late = near term
 LocationF1, 24 = 6.9, P = 0.015embryonic > abembryonic
 Stage x locationF3, 24 = 1.08, P = 0.377
thumbnail image

Figure 2. Mean values, SE for uterine blood vessel density, length-density, vessel diameter, and embryo dry mass as a function of embryo stage in gravid female Ctenotus taeniolatus (A, C, E, G) and Saiphos equalis (B, D, F, H). (A) uterine numerical blood vessel density, (C) uterine blood vessel length density, (E) uterine blood vessel diameter, and (G) embryo dry mass as a function of embryo stage in gravid female C. taeniolatus. (B) uterine numerical blood vessel density, (D) uterine blood vessel length density, (F) uterine blood vessel diameter, and (H) embryo dry mass as a function of embryo stage in gravid female Saiphos equalis.

Download figure to PowerPoint

Table 2. Mean values, SE (number of females), and results of statistical tests (Student's t-tests and ANOVAs) for uterine vessel density (Nv), vessel length-density (Lv) and vessel diameter (Dv) sampled at early and mid embryonic stages for Ctenotus taeniolatus, and early, mid, late and near-term embryonic stages for Saiphos equalis
 EarlyMidLateNear termResults
  1. Statistically significant results are in bold. Within each row, values with different letters are significantly different from one another (P < 0.05).

  2. Embryonic, embryonic hemisphere; Abembryonic, abembryonic hemisphere of uterus.

C. taeniolatus
Nv (vessels × 0.4 mm−2)
 Embryonic130.4 ± 4.2 (5)144.5 ± 4.7 (5)t8 = -2.23, P = 0.06
 Abembryonic104.3 ± 6.5 (5)108 ± 7.1 (5)t8 = -0.38, P = 0.71
Lv (mm × 0.4 mm−2)
 Embryonic14.0 ± 0.76 (5)15.4 ± 0.83 (5)t8 = -1.4, P = 0.21
 Abembryonic11.3 ± 0.65 (5)11.1 ± 0.55 (5)t8 = 0.44, P = 0.75
Dv (μm)
 Embryonic11.1 ± 0.60 (5)10.9 ± 0.65 (5)t8 = 0.27, P = 0.79
 Abembryonic11.8 ± 0.88 (5)11.0 ± 0.45 (5)t8 = 0.81, P = 0.44
S. equalis
Nv (vessels × 0.4 mm−2)
 Embryonic180 ± 17.5 (3) a176 ± 10.2 (7) a200 ± 6.3 (3) a, b242 ± 17.4 (4) bF3, 13 = 5.15, P = 0.014
 Abembryonic110 ± 18.6 (3) a98 ± 6.4 (7) a111 ± 0.4 (3) a, b145. ± 11.7 (4) bF3, 13 = 3.90, P = 0.04
Lv (mm × 0.4 mm−2)
 Embryonic17.4 ± 1.24 (3) a18.2 ± 1.1 (7) a21.6 ± 1.6 (3) a, b25.4 ± 1.43 (4) bF3, 13 = 8.21, P = 0.002
 Abembryonic10.0 ± 1.45 (3) a9.62 ± 1.4 (7) a11.4 ± 1.04 (3) a, b15.5 ± 1.03 (4) bF3, 13 = 6.21, P = 0.010
Dv (μm)
 Embryonic8.8 ± 0.89 (3) a11.6 ± 1.2 (7) a, b16 ± 1.2 (3) b16 ± 1.71 (4) bF3, 13 = 6.13, P = 0.008
 Abembryonic8.9 ± 0.46 (3)9.9 ± 0.83 (7)11.1 ± 1.3 (4)12.5 ± 2.0 (4)F3, 13 = 1.49, P = 0.268
thumbnail image

Figure 3. Immunofluorescent confocal micrographs of uterine microvasculature of Ctenotus taeniolatus (A, C, E, G) and Saiphos equalis (B, D, F, H). A, C. microvasculature from embryonic hemisphere (A) and abembryonic hemisphere (C) of C. taeniolatus uterus containing a newly ovulated embryo (<Stage 25). E, G. microvasculature from embryonic hemisphere (E) and abembryonic hemisphere (G) of C. taeniolatus uterus containing a mid stage embryo (Stage 28). B, D. microvasculature from embryonic hemisphere (B) and abembryonic hemisphere (D) of S. equalis uterus containing a newly ovulated embryo (<Stage 25). F, H. microvasculature from embryonic hemisphere (F), and abembryonic hemisphere (H) of S. equalis uterus containing a near term embryo (Stage 39). Arrows indicate invaginations of vessels indicative of intussusceptive angiogenesis. Scale bar is 50 μm.

Download figure to PowerPoint

Uterine Microvascular Architecture in S. equalis

In S. equalis, overall uterine Nv, Lv, and Dv varies among embryo stage and vessel density increases in parallel with growth in embryo mass (Table 1; Fig. 2B,D,F,H). Numerical blood vessel density, Lv, and Dv vary spatially between embryonic and abembryonic hemispheres of the uterus with the highest vascular density associated with the embryonic hemisphere of the egg, in the region of the uterus apposed to the highly vascularized chorioallantoic membrane of the embryo (Tables 1 and 2; Fig. 3B,D,F,H). On average, uterine Nv and Lv are 73% and 80% higher, respectively, over the course of gravidity in the embryonic hemisphere than in the abembryonic hemisphere. Uterine Nv and Lv in both embryonic and abembryonic hemispheres remain relatively constant prior to Stage 30 when embryonic growth in mass is relatively small. Vessel density and surface area increase markedly, however, late in gestation with Nv and Lv significantly higher at near-term embryo stages (i.e., Stages 39–40) than at early, mid and late embryo stages (Table 2; Fig. 2B,D,F,H). For example, Nv and Lv in the embryonic hemisphere increases by ∼36% (Nv) and 44% (Lv), respectively, at near-term embryo stages compared to that of early and middle embryo stages. Similarly, Dv in the embryonic hemisphere varies as a function of embryo stage (Table 2). Vessel diameter in the embryonic hemisphere increases from an average diameter of 10 μm at early and mid-embryo stages to ∼16 μm during late and near term embryo stages (total increase of 60%). Diameter of vessels in the abembryonic hemisphere does not differ as a function of embryo stage during gravidity. There are notable changes in structure and morphology of microvessels during gestation, indicated by the appearance of numerous small invaginations on the surface of vascular endothelial cells (Fig. 3B). These features are present during all reproductive stages sampled, and are consistent with intussusceptive angiogenesis (splitting of vessels through formation of cellular transluminal pillars; Burri and Djonov,2002).

Mass of embryos changes relatively little during the first half of gestation, whereas, embryo mass increases exponentially after embryo stage 35 (Fig. 2H). For example, average embryo dry mass increases by ∼4.5 mg between Stages 33 and 37 compared to an increase of ∼14 mg between Stages 37 and 40.

The Chorioallantois of S. equalis

Vessel density, Lv, and Dv of the chorioallantoic membrane of S. equalis does not vary significantly during gestation. The chorioallantoic membrane becomes fully formed at approximately embryo Stage 30 and increases in absolute size during development until it eventually covers the majority of the egg's surface area by embryo stage 40. From mid to near-term embryo stages, overall chorioallantoic Nv and Lv, and Dv average 180 ± 6.5 vessels 0.4 mm−2 (Nv), 19 ± 0.4 mm−2 (Lv), and 13 ± 1.6 μm (Dv), respectively. Structural changes indicative of intussusceptive angiogenesis are present on the surface of vascular endothelial cells of the chorioallantoic membrane (Fig. 4).

thumbnail image

Figure 4. Chorioallantoic microvasculature of Stage 36 Saiphos equalis embryo. Scale bar is 50 μm.

Download figure to PowerPoint

Interspecific Comparisons of Uterine Microvascular Architecture

Overall uterine Nv and Lv at early and mid embryo stages is lower in C. taeniolatus than in S. equalis, and Nv and Lv is higher in the embryonic region of the uterus than in the abembryonic region irrespective of species (Table 3, Fig. 3). Thus, there is a significant interaction between species differences in vessel surface area of embryonic versus abembryonic hemispheres of the uterus (Table 3). In C. taeniolatus, uterine Nv and Lv in the embryonic hemisphere is ∼23% and 17% less, respectively, than that of S. equalis (Table 2). In contrast, uterine Nv and Lv in the abembryonic hemisphere does not differ between species. Contrasts of overall vessel diameter at early and mid embryonic stages do not differ between species and the main effect of region of the uterus (embryonic vs. abembryonic) is not significant (Table 3).

Table 3. Interspecific contrasts (2-way ANOVAs) of overall vessel density (Nv), length-density (Lv), and vessel diameter (Dv) as a function of location (embryonic vs. abembryonic hemisphere) of the uterus in Ctenotus taeniolatus (C.t.) and Saiphos equalis (S.e.)
 ContrastResults
Nv  
 SpeciesF1, 48 = 18.5, P < 0.001C.t. < S.e.
 LocationF1, 48 = 58.0, P < 0.001embryonic > abembryonic
 Species x locationF1, 48 = 12.6, P = 0.001 
Lv  
 SpeciesF1, 48 = 11.7, P = 0.001C.t. < S.e.
 LocationF1, 48 = 55.7, P < 0.001embryonic > abembryonic
 Species x locationF1, 48 = 11.3, P = 0.002 
Dv  
 SpeciesF1, 48 = 0.3, P = 0.603
 LocationF1, 48 = 1.7, P = 0.194
 Species x locationF1, 48 = 3.6, P = 0.062

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. LITERATURE CITED

Hypoxic conditions limit embryonic growth in a variety of animal taxa including insects (Woods and Hill,2004), amphibians (Seymour et al.,2000), turtles (Kam,1992), lizards (Andrews,2002; Parker et al.,2004; Parker and Andrews,2006), and birds (Black and Snyder,1980). The association between embryonic growth in utero and uterine microvascular proliferation demonstrates a fundamental link between developmental processes and physiological features that provide oxygen to embryos during development. In C. taeniolatus, embryos remain small throughout the period of egg-retention, with little increase in embryo mass (total dry mass increase of ∼0.4 mg; Fig. 2G). Because size of C. taeniolatus embryos remains very small before oviposition, embryonic oxygen consumption is presumably also relatively low (Thompson and Stewart,1997) and the uterine vascular bed is able to accommodate embryonic oxygen demand throughout the period of egg retention. Similarly, before Stage 30, embryos of S. equalis remain small and the uterine vascular bed undergoes relatively little change in vascular density and surface area (Fig. 2B,D,F,H). Beyond Stage 35, however, embryonic mass increases rapidly and the increased growth is associated with substantial rise in vascular proliferation. The marked increase in uterine vessel surface area and vessel diameter late in gestation may be a response to the rapidly increasing metabolic demands of embryos.

Uterine vascular density varies during the reproductive cycle in both oviparous and viviparous species of lizards, with highest vascular densities associated with gravidity/pregnancy (Guillette and Jones,1985; Masson and Guillette,1987). For example, uterine vascular densities in the oviparous lizards Plestiodon obsoletus, Sceloporus undulatus, and Crotaphytus collaris are higher in gravid females compared to vitellogenic and nonreproductive females. In these three species, vascular density of gravid females ranges from ∼20 to 35 vessels mm−2 (Masson and Guillette,1987). Overall vascular densities observed in our study, however, are 4–9 times higher than that reported by Masson and Guillette (1987). Phylogenetic differences between skink and phrynosomatid lineages may provide one explanation for the large difference in vascular density values observed in our study versus that of Masson and Guillette (1987). Alternatively, the lower density values reported by Masson and Guillette (1987) may be attributable to the ability to resolve microvasculature using immunofluorescence and confocal microscopy at relatively high magnification (200×) versus ink the vascular casts used by Masson and Guillette (1987) at relatively low (20×) magnification.

Overall uterine vascular surface area at early to mid embryo stages is lower in C. taeniolatus than in S. equalis (Tables 2 and 3, Figs. 2 and 3). In addition to having a lower overall vascular surface area, very little angiogenesis occurs during the period of egg retention in C. taeniolatus, whereas angiogenesis occurs throughout gestation in S. equalis. In both species, angiogenesis occurs primarily by intussusception; we did not observe vessel proliferation by other angiogenic processes such as sprouting. Intussusception is a common form of angiogenesis that facilitates rapid increase in vessel surface area as well as playing a role in remodeling of vascular structures (Patan et al.,1992; Burri and Djonov,2002). In addition to intussusceptive angiogenesis, there is a general increase in length and tortuosity of blood vessels in the uterus of S. equalis as gestation progresses, suggesting that vessel lengthening also contributes to the increased vascular surface area at advanced embryo stages (Fig. 3B,F).

In addition to the increase in overall Nv and Lv, Dv also increases significantly during late gestation in S. equalis, (Table 2, Figs. 2F, 3B,F). Increased diameter of microvessels results in an elevated volume of blood transported to uterine tissues, which facilitates greater oxygen availability to the growing embryo. Increasing vessel diameter and the associated rise in uterine blood flow late in gestation is a typical pattern observed in placental mammals (Rosenfeld,1977; Reynolds and Redmer,2001; Lang et al.,2003; Borowicz et al.,2007). In sheep, for example, uterine blood flow increases from ∼23 mL min−1 in nonpregnant females to ∼1,300 mL min−1 by the end of gestation in pregnant females (Rosenfeld,1977). Our observations in S. equalis likely reflect a common pattern in viviparous species where expansion of the uterine vascular bed matches the metabolic demands of the embryo as embryonic oxygen requirement increases.

Regional Patterns of Microvascular Architecture

In both species, Nv and Lv of blood vessels is higher in the embryonic hemisphere of the uterus than in the abembryonic hemisphere (Tables 1 and 2, Fig. 3). Moreover, a greater amount of intussusceptive angiogenesis occurs in vessels in the embryonic hemisphere than in the abembryonic hemisphere. The highly vascularized embryonic hemisphere of the uterus is closely apposed to the chorioallantoic membrane of the embryo where it mediates respiratory gas exchange between maternal and embryonic circulation (Packard et al.,1977; Yaron,1985). The primary function of the less vascularized abembryonic hemisphere, however, is less clear. The ultrastructure and morphology of uterine epithelial cells in some species of viviparous skinks indicates that the abembryonic region of the uterus of functions in nutrient transport and not gas exchange (Thompson et al.,2006; Biazik et al.,2009). Because C. taeniolatus and S. equalis are lethicotrophic the vast majority of embryonic nutrients are obtained from yolk reserves rather than from the maternal circulation. Consequently, if uterine transport does occur in the abembryonic hemisphere of these species, it would most likely be limited to relatively small compounds such as water or inorganic ions.

The overall morphology of the chorioallantoic microvasculature of S. equalis is similar to that of uterine vasculature (Figs. 3B,D,F,H, Fig. 4), but Nv, Lv, and Dv in the chorioallantois do not increase during the rapid growth phase late in gestation (Stages 35–40). Average chorioallantoic vascular density values obtained in our study (180 ± 6.5 vessels 0.4 mm−2) are similar to those reported in alligator (184 ± 18.4 mm−2), tortoise (129 ± 6.8 mm−2), and chicken (213 ± 7.4 mm−2) (Corona and Warburton,2000). Although vessel density does not increase significantly, the surface area of the chorioallantoic membrane increases in absolute size as it grows over the inner surface of the eggshell. The combination of both increasing vascular density of the uterus and expansion of the chorioallantoic membrane function together to meet elevated embryonic oxygen demand during late gestation. Conversely, in Stage 30 embryos of C. taeniolatus, the newly formed chorioallantoic membrane is extremely small and thus embryonic gas exchange prior to Stage 30 likely occurs by simple diffusion or via the vascularized yolk sac membrane.

Regulation of Angiogenesis and the Evolution of Viviparity

According to the most widely accepted model, viviparous reproduction evolves through gradual increases in the proportion of embryonic development occurring within the uterus of the gravid female (Packard et al.,1977; Shine,1983,1985). This evolutionary transition putatively requires at least three major events: 1. the capacity to retain eggs in utero for increasingly prolonged periods of time, 2. the thinning or complete loss of the eggshell, and 3. the capacity to support embryogenesis during prolonged egg retention (Packard et al.,1977; Andrews and Mathies,2000). Several species of lizards are capable of prolonged egg retention, however, embryonic development in most of these species is arrested at approximately the limb bud stage (Dufaure and Hubert embryo stage 30) of embryonic development. Hypoxia is likely the proximate factor that is responsible for this developmental arrest (Andrews,2002; Parker et al.,2004; Parker and Andrews,2006). In Sceloporus lizards for example, the maximum stage attained by embryos retained in utero is determined by in utero oxygen availability (Parker et al.,2004; Parker and Andrews,2006). The capacity of the uterine vascular bed to increase in concert with embryo size and metabolism is likely to be an important transitional step that must occur more or less simultaneously with increases in the duration of egg retention during the evolution of reptilian viviparity.

Despite the fundamental role of uterine angiogenesis in gestation, the underlying mechanisms regulating uterine vascular proliferation during the squamate reproductive cycle are not known. Uterine angiogenesis in reptiles is presumably under primary control of the ovarian steroid hormones oestrogen and progesterone (Yaron,1972). These hormones act via their receptors to initiate gene expression in target cells. In mammals, progesterone in particular, is associated with angiogenesis (Walter et al.,2005; Girling et al.,2007). In reptiles, the role of progesterone has not been fully elucidated but is likely to be responsible for oviductal hypertrophy (Guillette and Jones,1985) and in delaying oviposition/parturition (Guillette et al.,1991). One of the key local inducers of uterine angiogenesis that is under control of progesterone is vascular endothelial growth factor (VEGF) (Hyder,1999; Walter et al.2005). In mammals, VEGF is expressed in the uterus and is selectively mitogenic for endothelial cells and also results in increased vascular permeability (Ferrara and Davis-Smyth,1997; Ferrara,1999). Given the central role of VEGF in stimulating uterine vascular proliferation in the mammalian reproductive cycle, it is likely that this protein or a homologue also plays a similar role in the regulation of uterine angiogenesis in reptiles.

In addition to endocrine influences, uterine angiogenesis could also be locally stimulated by hypoxia caused by the growing embryo (Guillette and Jones,1985). Under this scenario, a high oxygen diffusion gradient between maternal and embryonic blood supply would deplete oxygen content in uterine tissues and therefore stimulate angiogenesis in areas of low tissue PO2. Low tissue PO2 is a potent stimulator of angiogenesis, and in vertebrates the physiological response to low tissue PO2 is largely mediated by hypoxia-inducible factors (HIFs; Semenza,1999). Hypoxia-inducible factors are transcription factors that induce expression of genes including VEGF, which collectively initiate a variety of responses, including angiogenesis, to restore tissue PO2 to homeostatic levels (Forsythe et al.,1996). Hypoxia-inducible factors are present in many vertebrate classes (Semenza,1999; Powell and Hahn,2002) and are known angiogenic inducers in uterine and placental tissues of mammals (Rajakumar and Conrad2000; Borowicz et al.,2007). Identifying molecular mechanisms regulating both maternal and embryonic contributions of placental angiogenesis in squamate reptiles will help to reveal fundamental genetic changes associated with the evolution of viviparity.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. LITERATURE CITED

Animals were collected with New South Wales Scientific Collecting License number S10693 and the research was approved by the University of Sydney Animal Ethics Committee (number L04/9-2006/2/4461). Technical assistance was received from Renee Whan and the staff of the Electron Microscopy Unit, The University of Sydney. Assistance in the field and laboratory was received from J. Herbert, J. Biazik, B. Murphy, D. Pike, J. Stewart, L. Lindsay, L. Venuto, L. Young, and M. Barthet.

LITERATURE CITED

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. LITERATURE CITED