Characterization of isolated liver sinusoidal endothelial cells for liver bioengineering.

BACKGROUND: The liver sinusoidal capillaries play a pivotal role in liver regeneration, suggesting they may be beneficial in liver bioengineering. This study isolated mouse liver sinusoidal endothelial cells (LSECs) and determined their ability to form capillary networks in vitro and in vivo for liver tissue engineering purposes. METHODS AND RESULTS: In vitro LSECs were isolated from adult C57BL/6 mouse livers. Immunofluorescence labelling indicated they were LYVE-1+/CD32b+/FactorVIII+/CD31-. Scanning electron microscopy of LSECs revealed the presence of characteristic sieve plates at 2 days. LSECs formed tubes and sprouts in the tubulogenesis assay, similar to human microvascular endothelial cells (HMEC); and formed capillaries with lumens when implanted in a porous collagen scaffold in vitro. LSECs were able to form spheroids, and in the spheroid gel sandwich assay produced significantly increased numbers (p = 0.0011) of capillary-like sprouts at 24 h compared to HMEC spheroids. Supernatant from LSEC spheroids demonstrated significantly greater levels of vascular endothelial growth factor-A and C (VEGF-A, VEGF-C) and hepatocyte growth factor (HGF) compared to LSEC monolayers (p = 0.0167; p = 0.0017; and p < 0.0001, respectively), at 2 days, which was maintained to 4 days for HGF (p = 0.0017) and VEGF-A (p = 0.0051). In vivo isolated mouse LSECs were prepared as single cell suspensions of 500,000 cells, or as spheroids of 5000 cells (100 spheroids) and implanted in SCID mouse bilateral vascularized tissue engineering chambers for 2 weeks. Immunohistochemistry identified implanted LSECs forming LYVE-1+/CD31- vessels. In LSEC implanted constructs, overall lymphatic vessel growth was increased (not significantly), whilst host-derived CD31+ blood vessel growth increased significantly (p = 0.0127) compared to non-implanted controls. LSEC labelled with the fluorescent tag DiI prior to implantation formed capillaries in vivo and maintained LYVE-1 and CD32b markers to 2 weeks. CONCLUSION: Isolated mouse LSECs express a panel of vascular-related cell markers and demonstrate substantial vascular capillary-forming ability in vitro and in vivo. Their production of liver growth factors VEGF-A, VEGF-C and HGF enable these cells to exert a growth stimulus post-transplantation on the in vivo host-derived capillary bed, reinforcing their pro-regenerative capabilities for liver tissue engineering studies.

Hypoxic preconditioning of myoblasts implanted in a tissue engineering chamber significantly increases local angiogenesis via upregulation of myoblast vascular endothelial growth factor-A expression and downregulation of miRNA-1, miRNA-206 and angiopoietin-1.

Vascularization is a major hurdle for growing three-dimensional tissue engineered constructs. This study investigated the mechanisms involved in hypoxic preconditioning of primary rat myoblasts in vitro and their influence on local angiogenesis postimplantation. Primary rat myoblast cultures were exposed to 90 min hypoxia at <1% oxygen followed by normoxia for 24 h. Real time (RT) polymerase chain reaction evaluation indicated that 90 min hypoxia resulted in significant downregulation of miR-1 and miR-206 (p < 0.05) and angiopoietin-1 (p < 0.05) with upregulation of vascular endothelial growth factor-A (VEGF-A; p < 0.05). The miR-1 and angiopoietin-1 responses remained significantly downregulated after a 24 h rest phase. In addition, direct inhibition of miR-206 in L6 myoblasts caused a significant increase in VEGF-A expression (p < 0.05), further establishing that changes in VEGF-A expression are influenced by miR-206. Of the myogenic genes examined, MyoD was significantly upregulated, only after 24 h rest (p < 0.05). Preconditioned or control myoblasts were implanted with Matrigel™ into isolated bilateral tissue engineering chambers incorporating a flow-through epigastric vascular pedicle in severe combined immunodeficiency mice and the chamber tissue harvested 14 days later. Chambers implanted with preconditioned myoblasts had a significantly increased percentage volume of blood vessels (p = 0.0325) compared with chambers implanted with control myoblasts. Hypoxic preconditioned myoblasts promote vascularization of constructs via VEGF upregulation and downregulation of angiopoietin-1, miR-1 and miR-206. The relatively simple strategy of hypoxic preconditioning of implanted cells - including non-stem cell types - has broad, future applications in tissue engineering of skeletal muscle and other tissues, as a technique to significantly increase implant site angiogenesis.

The influence of nitric oxide synthase 2 on cutaneous wound angiogenesis.

BACKGROUND: Inducible nitric oxide synthase (nitric oxide synthase 2, NOS 2) inhibition significantly suppresses chronically ischaemic skin flap survival, possibly because of reduced angiogenesis. OBJECTIVES: To investigate the effect of genetic NOS 2 inhibition on cutaneous wound angiogenesis in two in vivo murine models. The impact of NOS 2 manipulation on vascular endothelial growth factor (VEGF)-A stimulated and fibroblast growth factor (FGF)-2 stimulated angiogenesis was also investigated in the Matrigel(®) plug assay. METHODS: (i) Matrigel plugs/incisional wounds: two groups of NOS 2-/- mice and two groups of wild-type (WT) mice had bilateral Matrigel plugs containing 500 ng mL(-1) VEGF-A or 1000 ng mL(-1) FGF-2 injected subcutaneously in the abdomen. A 2·5 cm long dorsal incisional skin wound was created and sutured closed in the same animals. Wounds and plugs were explored at 7 or 12 days. (ii) Excisional wounds: dorsal 0·5 × 1·0 cm excisional skin wounds were created in four groups (two NOS 2-/- and two WT) and explored at 7 or 14 days. Wounds and Matrigel plugs were examined histologically and morphometrically for determination of percentage vascular volume (PVV). RESULTS: The PVV in NOS 2-/- incisional wounds and excisional wounds was significantly less than in WT wounds (P = 0·05 and P < 0·001, respectively). The PVV was significantly less in VEGF-A stimulated Matrigel plugs compared with FGF-2 stimulated plugs in NOS 2-/- mice (P < 0·01), but not in WT mice. CONCLUSIONS: NOS 2 is significantly involved in angiogenic signalling in healing skin wounds, particularly within the first 7 days. However, Matrigel plug vascularization suggests that the role of NOS 2 in angiogenesis is related to VEGF-A but not FGF-2 stimulated angiogenesis.

The use of pimonidazole to characterise hypoxia in the internal environment of an in vivo tissue engineering chamber.

UNLABELLED: The distribution of hypoxic cells in an in vivo tissue engineering chamber was investigated up to 28 days post-implantation. METHODS: Arteriovenous loops were constructed and placed into bi-valved polycarbonate chambers containing 2 x 10(6) rat fibroblasts in basement membrane gel (BM gel). Chambers were inserted subcutaneously in the groin of male rats and harvested at 3 (n = 6), 7 (n = 6), 14 (n = 4) or 28 (n = 4) days. Ninety minutes before harvest, pimonidazole (60 mg/kg) was injected intraperitoneally. Chamber tissue was removed, immersion fixed, paraffin embedded, sectioned and stained immunohistochemically using hypoxyprobe-1 Mab that detects reduced pimonidazole adducts forming in cells, where pO2 < 10 mmHg. RESULTS: At 3 days a fibrin clot/BM gel framework filled the chamber. Seeded fibroblasts had largely died. The majority of 3 day chambers did not demonstrate tissue growth from the AV loop nor was pimonidazole binding present in these chambers. In one chamber in which tissue growth had occurred strong pimonidazole binding was evident within the new tissue. In four out of six 7 day chambers a broader proliferative zone existed extending up to 0.4 mm (approximately) from the AV loop endothelium which demonstrated intense pimonidazole binding. The two remaining 7 day chambers displayed even greater tissue growth (leading edge > 0.7 mm from the AV loop endothelium), but very weak or no pimonidazole binding. At 14 and 28 days the fibrin/BM gel matrix was replaced by mature vascularised connective tissue that did not bind pimonidazole. CONCLUSION: Employing a tissue engineering chamber, new tissue growth extending up to 0.4 mm from the AV loop endothelium (chambers < or = 7 days) demonstrated intense pimonidazole binding and, therefore, hypoxia. Tissue growth greater than 0.5 mm from the AV loop endothelium (7-28 days chambers) did not exhibit pimonidazole binding due to a significant increase in the number of new blood vessels and was, therefore, adequately oxygenated.

Inducible nitric oxide synthase (iNOS) activity promotes ischaemic skin flap survival.

We have examined the role of nitric oxide (NO) in a model of functional angiogenesis in which survival of a skin flap depends entirely on angiogenesis to provide an arterial blood supply to maintain tissue viability. The different effects of nitric oxide synthase (NOS) inhibitors on rat skin flap survival appeared to be explained on the basis of their NOS isoform selectivity. Skin flap survival was decreased by iNOS-selective (inducible NOS) inhibitors, S-methyl-isothiourea, aminoguanidine and aminoethylthiorea; unaffected by the non-selective inhibitor nitro-imino-L-ornithine; and enhanced by the cNOS (constitutive NOS, that is endothelial NOS (eNOS) and neuronal NOS (nNOS)) inhibitor, nitro-L-arginine methyl ester. Skin flap survival was reduced in mice with targeted disruption of the iNOS gene (iNOS knockout mice), and the administration of nitro-L-arginine methyl ester significantly increased flap survival in iNOS knockout mice (P<0.05). iNOS immunoreactivity was identified in mast cells in the angiogenic region. Immunoreactive vascular endothelial growth factor (VEGF) and basic fibroblast growth factor were also localized to mast cells. The combination of interferon-gamma and tumour necrosis factor-alpha induced NO production and increased VEGF levels in mast cells cultured from bone marrow of wild-type, but not iNOS KO mice. The increased tissue survival associated with the capacity for iNOS expression may be related to iNOS-dependent enhancement of VEGF levels and an ensuing angiogenic response. Our results provide both pharmacological and genetic evidence that iNOS activity promotes survival of ischaemic tissue.

Late occlusion of microvascular vein grafts in replantation.

Two cases are described in which patients presented 16 and 17 years, respectively, after complete or incomplete amputation/replantation of the arm. In case 1, the patient complained of coldness, pain, and tingling in the replanted arm in the previous 24 hours and noticed that his fingers had gone white. Arteriography and subsequent surgery revealed obliteration of the vein graft (inserted in the distal brachial artery) by neointimal thickening and atherosclerotic plaque, which was confirmed in a subsequent morphologic examination. In case 2, the patient presented with discomfort and a pulsatile swelling on the inner aspect of his upper arm. Arteriography and surgery revealed an aneurysm in the previously inserted vein graft in the brachial artery, with some atherosclerotic degeneration. Both vein grafts were successfully replaced with a fresh autologous vein graft and the patients remain well several years later. The 2 cases suggest that as part of replantation surgery of a limb, it is essential to maintain postoperative clinical monitoring for signs of graft degeneration in all patients with long-term vein graft insertion.

Long-term studies of cold-stored rabbit femoral artery and vein autografts.

UNLABELLED: In previous studies we have shown that 80-100% of rabbit femoral vascular autografts cold-stored at 4 degrees C for 3 weeks remain patent 3 weeks after reinsertion in the femoral artery. The present study reports the effect on graft patency of increasing either the period of cold storage prior to reinsertion or the duration of reperfusion to 6 months. Rabbit femoral blood vessels were cold-stored (CS) at 4 degrees C for varying periods. CS autografts were reinserted into the contralateral leg for 3 weeks or 6 months. Graft patency was determined and grafts examined by histological, immunohistochemical and electron microscopical techniques. Six months after reinsertion patency of 4-week CS arterial and 1-week CS venous grafts was 40% and 27% respectively, very much lower than the 80-100% seen after 3 weeks reperfusion. Arterial grafts CS for 6 months had a patency rate of 70% after 3 weeks reperfusion but 0% after 6 months. Morphological examination suggests that the delayed failure of cold-stored vascular grafts is caused by thrombus superimposed on intimal hyperplasia within the graft. CONCLUSIONS: Cold-stored vascular grafts are useful prostheses when only 3-4 weeks graft patency is required. They are not suitable for use when long-term graft patency is needed.

A morphological study of the long-term repair process in experimentally stretched but unruptured arteries and veins.

As many avulsion amputations are incomplete and the vessels remain intact, the immediate pathology and long-term repair process (to 3 months post-injury) of experimentally stretched but unruptured rabbit femoral arteries and veins were examined. In stretched arteries, circumferential skip lesions involving endothelium, internal elastic lamina (IEL) and media occurred frequently and often up to 3 cm from the point of stretch. Medial smooth muscle cells (SMC) were significantly damaged or lost at lesions. Macrophages and neutrophils were found in lesions 1-4 days post-injury. Between 2-4 days, lesions were covered by endothelium and synthetic state SMC appeared in the media. At 1 week, a thin neointima (which persisted to 3 months) covered many lesions. The media at lesions gradually filled with SMC but generally remained disorganised even at 3 months post-injury. Stretching caused tears in vein walls, particularly close to the point of injury. There was no evidence of venous damage or repair in specimens examined 3 weeks and 3 months post-injury.

The effect of cold ischemia on the patency of microvascular repair following arterial avulsion injury.

UNLABELLED: Avulsion injuries have a poor prognosis for survival in clinical replantation surgery. Arterial thrombosis is the most significant factor contributing to avulsion replant failure, and severe arterial damage has been observed with this injury. However, patency rates of experimentally avulsed arteries repaired immediately are much higher than in the clinical situation. This paper evaluates the effect of an added component--ischemia--on the patency of experimentally avulsed arteries. All avulsions seen clinically are subject to some degree of ischemia prior to replantation. Ninety rabbits had both femoral arteries avulsed under general anesthesia. A 6.5-cm graft was harvested from the left distal femoral artery. In 20 rabbits (group 1: 0 hours of ischemia) the graft was immediately inserted into the defect in the right femoral artery. Sixty rabbits (20 grafts per group) had their grafts stored at 4 degrees C for either 10 hours (group 2), 18 hours (group 3), or 24 hours (group 4) and reinserted into the right femoral artery in a second operation. Patency was assessed 3 weeks after reinsertion. Groups 1 and 2 maintained high patency rates (85 percent); however, group 3 (70 percent) and group 4 (45 percent) had lower patency rates than group 1, with a significant difference between groups 1 and 4 (p < 0.01). In a fifth group (10 grafts), avulsed 24-hour ischemic grafts were hydrodilated prior to reinsertion. The patency rate of this group increased significantly (90 percent) compared with group 4 (p < 0.005). CONCLUSION: These experiments suggest that a combination of avulsion injury and ischemia time is responsible for the poor clinical results of avulsion replantations.

Effects of cold storage on the subsequent structure and function of microvenous autografts.

A study was undertaken to evaluate the structure and patency of cold stored rabbit femoral veins following reinsertion as autografts for 3 weeks. The periods of cold ischaemic storage were 1, 2, 3 and 4 weeks (n = 10/gp) and 6 and 10 weeks (n = 6/gp). All rabbits were subject to 3 operations under general anaesthesia. In the first, a 4 cm segment of left femoral vein was harvested and stored at 4 degrees C for the specified ischaemic interval. Following storage the graft was microsurgically reinserted at a second operation into the right femoral artery of the donor rabbit. Three weeks later, graft patency was assessed by surgical exploration and the graft processed for light and electron microscopy. Patency rates remained over 80% after 3 weeks in all groups except the 10 week storage group where only 1 of 6 (17%) grafts was patent at 3 weeks. In all groups normal vein structure was absent, being replaced by thin walled necrotic areas or by neointimal ingrowth. The excellent patency rates achieved indicate it is possible to cold preserve extra lengths of vein grafts harvested at initial operation for reuse should regrafting be necessary.

Cold stored femoral vessels as microvascular allografts: a preliminary study.

Cold stored femoral arteries or veins have been reinserted successfully as autografts into rabbits. The present study examines whether grafting with cold stored vascular allografts is equally successful. Rabbit femoral arteries and veins were stored at 4 degrees C for 4 weeks before insertion as allografts into unrelated animals. Three weeks after insertion into the femoral artery all venous allografts and 80% of arterial allografts were patent, but patency of both graft types declined over the next few weeks. A small number of cold stored venous allografts when inserted into the femoral vein occluded within 3 weeks. No histological evidence of rejection was apparent. The findings suggest that cold stored vascular allografts could be used successfully as an arterial "prosthesis" to support free flaps where relatively short term patency is required until the flap can establish sufficient peripheral inset to survive in its own right. This technique could be applied when autologous veins are not available or not justified.

The functional and structural effects of hypothermic storage on ischaemic arterial grafts.

The effects of hypothermic ischaemia on blood vessels are unknown. This study aimed to determine the 3 week patency rate and the pathology of 9 experimental groups of hypothermically stored ischaemic arteries and one control group in a rabbit femoral artery model. Ischaemia times were 0 h, 24 h, 1, 2, 4, 6, 8 and 10 weeks (Groups 1-8). Patency was over 80% in all groups after 3 weeks reinsertion. Following reinsertion control grafts maintained normal arterial structure, but cellular degeneration had occurred in all ischaemic grafts and appeared complete after 4 weeks ischaemia. The graft connective tissue framework frequently remained intact. Repair was evident in central graft regions after 2 weeks ischaemia and 3 weeks reinsertion, but occurred only adjacent to the anastomosis in 4-10 week ischaemic arteries. Four week ischaemic arteries (Groups 9 and 10) reinserted for 6 and 12 weeks respectively exhibited near complete repair but patency dropped to 60% in the 12 week group.

Hydrostatic distention and pharmacological treatment of epinephrine-induced microarterial spasm.

The efficacy of using hydrodistention and vasodilatory drugs to relieve spasm in arteries was investigated. The femoral arteries of 64 rabbits (divided into five groups) were placed in spasm by the topical application of epinephrine (1 mg/ml). In group A (controls) vasospasm was induced without further treatment. In group B vasospasm was induced, and the arteries were hydrodistended with normal saline. In three additional groups vasospasm was induced and either 10% lidocaine hydrochloride (group C), verapamil hydrochloride (group D), or chlorpromazine hydrochloride (group E) was applied. Thirty minutes later the vessels were hydrodistended. All vessels were measured at set intervals, and some specimens were retained for light and electron microscopy. All methods of spasm relief were successful, although to varying degrees. Hydrodistention alone produced the widest dilation for the longest time. Lidocaine was the most successful drug treatment alone. Verapamil followed by hydrodistention was the most successful combination regimen, but did not produce better results than hydrodistention alone. Hydrodistention alone produced significant arterial wall damage, resulting in permanent structural modifications. Prior vasodilatory drug treatment reduced but did not eliminate hydrodistention damage. Although hydrodistention dilates for longer periods than vasodilatory drugs, the arterial wall damage associated with hydrodistention indicates that it should be used only when all other methods of vasodilation have failed.

Comparative study of vascularized and nonvascularized tendon grafts for reconstruction of flexor tendons in zone 2: an experimental study in primates.

Vascularized tendon grafts were compared with nonvascularized tendon grafts in a primate experimental model. In four monkeys, seven vascularized extensor hallucis longus grafts were placed in the foot's digital fibroosseous canal and these were compared with eight nonvascularized tendon grafts similarly placed in the opposite extremities. The juncture techniques and postoperative protocols were identical for both tendon groups. All tendons were explored 5 months after insertion. The vascular pedicle was patent in all vascularized tendons. Three tendon ruptures occurred in nonvascularized tendons and only one rupture occurred in a vascularized tendon. The digits with vascularized tendons demonstrated a significantly better simulated total active motion (117 degrees versus 128 degrees, p less than 0.05) than digits with nonvascularized tendons. The difference was even more significant when a localized tenolysis of the proximal juncture of the tendon graft was performed (215 degrees versus 138 degrees, p less than 0.01). This study supports the concept that vascularized tendon grafts may be advantageous in scarred tendon beds.

The significance of resection length on the patency rate, and the histopathology, of experimentally avulsed and microsurgically repaired blood vessels.

The significance of resection length on patency rate, and the histopathology, of microsurgically repaired avulsed blood vessels was examined at 3 weeks in two groups of experimentally avulsed rabbit femoral arteries repaired by different surgeons and in a single series of avulsed and repaired veins. All veins were patent 3 weeks after avulsion and microsurgical repair. Histopathology indicated that the vast majority of lesions in veins were removed at resection. Surgeon A achieved 75% patent arteries and Surgeon B achieved 100% arterial patency (resecting 3.7 mm more on average than Surgeon A). Histopathology revealed numerous luminal circumferential lesions remained in the avulsed artery wall following resection. These lesions were sites of smooth muscle cell proliferation and neointima formation. Although similar arterial damage occurs in human avulsion, considerably lower patency rates are achieved for human arterial avulsion repair than were reported in this experimental study. Factors in addition to vessel wall damage must therefore be involved in thrombosis and occlusion of repaired avulsed arteries. Such factors might be lengthy ischaemia time and severe spasm.

Intraoral mucosal reconstruction with microvascular free jejunal autografts: an experimental study.

This experimental study investigated the problem of covering bare bone adequately in the oral cavity to provide a stable, functionally acceptable reconstruction. In eight dogs, intraoral mucosal defects were created whilst preserving the mandibular arch. The defects were reconstructed with a microvascular jejunal patch. Six animals had their reconstructions stressed postoperatively by the administration of irradiation and by feeding solid food. Two control dogs did not receive irradiation. Grafts were assessed clinically and histologically for 6 months. Rapid mucosal healing occurred in seven of eight dogs. The grafts conformed well to the mandibular contour and were tolerant of postoperative radiotherapy and the chewing of solid food. Structural integrity of seven grafts was maintained although subtotal villous atrophy occurred in irradiated grafts and to a lesser extent in control grafts. In one animal whose graft mucosa sloughed, the wound was re-epithelialized from the adjacent buccal mucosa. Microvascular jejunal patches therefore provided a durable, functionally adequate reconstruction of the mouth floor.

The long-term fate of microvenous autografts.

Changes in length, external diameter, wall thickness, and morphology were examined in 60 intraarterial vein grafts and 30 intravenous vein grafts inserted in rabbit femoral vessels. Half of each graft type was explored at 6 or 12 months, giving four experimental groups. The overall patency for intraarterial grafts at exploration was 98 percent and for intravenous grafts 100 percent. In comparison with the initial graft length resected, all four groups were significantly shorter at the completion of anastomosis, and three of the four groups also were significantly shorter at exploration. The overall loss in length of grafts varied between 26 and 30 percent of original length. External diameter was significantly increased (from between 133 and 201 percent) in all four groups at exploration compared to the normal femoral vein. Intravenous grafts maintained normal vein morphology to 12 months. Intraarterial grafts were modified by the ingrowth of smooth-muscle cells from the recipient artery, thereby creating a neointima that significantly thickened their walls at both 6 and 12 months.

The use of long synthetic microvascular grafts to vascularise free flaps in rabbits.

A 5 cm length of 2 mm internal diameter (i.d.) synthetic, expanded polytetrafluoroethylene (PTFE, or Gore-Tex) vascular graft was used to connect 25 rabbit inferior epigastric flaps to the contralateral femoral vessels. In 15 animals an expanded PTFE graft connected the opposite femoral artery to the flap while the ipsilateral venous drainage remained intact. In the remaining 10 animals an expanded PTFE graft was used to replace the venous drainage of the flap and connected to the opposite femoral vein while the ipsilateral femoral artery supplied the flap. Flap survival and graft patency were evaluated over 3 weeks. Ten of 15 flaps with intra-arterial grafts survived at 3 weeks (67%). Only 27% (4/15) of their supplying grafts remained patent for 3 weeks, although 67% (10/15) were patent at 10 days. All 10 flaps, where expanded PTFE grafts replaced venous outflow, failed within 36 hours. At exploration these grafts were thrombosed or collapsed. In conclusion, currently available 2 mm (i.d.) expanded PTFE vascular graft cannot maintain patency in a low blood flow circulation supplying an isolated free flap.

Changes in area and thickness of expanded and unexpanded axial pattern skin flaps in pigs.

The pig buttock flap model was used firstly, to compare changes in expanded axial skin flap area with control flaps prior to flap elevation and 3 or 4 months post-flap inset and, secondly, to compare the thickness of expanded and control flaps at elevation and for 3 to 6 months post-flap inset. Following 5 weeks' expansion and prior to elevation, the expanded tissue had gained a significant 63.3% mean increase in area compared with the control tissue (p less than 0.01). Immediately post elevation and inset, the expanded flaps were still significantly larger than the control flaps by a mean 29.8% (p less than 0.01) but had lost 56% of the original area gained. Little change in area occurred in the 3 months post-flap inset as the expanded flaps were still a mean 23.4% larger than the control (p less than 0.01). Dermal and cellular non-keratinised epidermal layers thickened markedly in expanded skin compared to control skin. Following elevation and inset of the flaps, both dermis and epidermis thickened in expanded and control flaps.