Cytochrome oxidase activity of postsurgical bullous keratopathy endothelium

Proper corneal hydration is normally maintained by a pump-leak mechanism located on the lateral cell membranes of the endothelium (1,2). When endothelial cells become sufficiently dysfunctional in response to trauma, the cornea becomes irreversibly edematous and opaque. Previous studies in our laboratory have shown that mitochondrial cytochrome oxidase (CO) activity, an enzyme important in respiratory activities, is correlated with endothelial cell functional activity in corneas with Fuchs' endothelial dystrophy. In this study we investigated cytochrome oxidase activity in corneas with postsurgical bullous keratopathy. Corneas with aphakic and pseudophakic bullous keratopathy (ABK, PBK) were incubated in diaminobenzidine-cytochrome C media. Results showed that the staining pattern of ABK and PBK corneas was uniformly low compared to keratoconus corneas, and in contrast to previous studies on Fuchs' dystrophy corneas which demonstrated a regional staining pattern. This suggests that CO staining patterns correlate with functional activity and may be a useful technique to detect dysfunctional cells in various disease categories.     

Fuchs' corneal dystrophy. A clinicopathologic study of the variation in corneal edema

Corneal buttons from six patients with Fuchs' dystrophy had varying degrees of clinical edema measured in most cases by preoperative optical or ultrasonic pachymetry. These were sectioned in the operating room so that histologic correlations could be made. Histologically, marked thickening of Descemet's membrane and abnormal corneal endothelium corresponded to areas of severe clinical edema and were usually located in the central and paracentral regions. Descemet's membrane displayed multiple prominent guttata of varying size and shape, either facing the anterior chamber, or buried within multilaminar Descemet's membrane. In some corneas, aggregates of 10 nm fibrils were seen at the edges of guttata, corresponding to areas that stained for oxytalan fibrils. The endothelium was attenuated underlying the guttata. Clinical edema was not present unless accompanied by marked thickening of Descemet's membrane with multiple guttata and attenuation of corneal endothelium. The peripheral cornea was relatively clear clinically and showed minimal histologic changes.     

A COMPARISON OF CORNEAL ENDOTHELIAL MORPHOLOGY IN CORNEA GUTTATA, FUCHS' DYSTROPHY AND BULLOUS KERATOPATHY

Changes on corneal endothelial specular microscopy were compared in 14 patients with cornea guttata, 4 patients with Fuchs' corneal dystrophy and 19 patients with various forms of bullous keratopathy. The patients with cornea guttata showed well marked guttae 1 to 6 endothelial cells in diameter in the endothelial mosaic and in the relief mode while the endothelial mosaic was usually otherwise within normal limits. In 2 patients with Fuchs' dystrophy the endothelium could be examined, showing gross guttae but a few areas of relatively normal endothelial cells. The unaffected eye of 3 other patients snowed findings similar to cornea guttata, but with some reduction in endothelial cell count in 2 patients. The patients with bullous keratopathy fell into 2 groups, one with gross reduction in cell count in a markedly abnormal endothelial cell mosaic, the other a mixed group with moderate reduction in cell count and numerous guttae. Some miscellaneous cases included one of aphakic peripheral bullous keratopathy, one associated with cyclitis and aphakia and 2 with idiopathic non-surgical bullous keratopathy. We believe the corneal endothelium is not grossly abnormal away from the guttae in Fuchs' dystrophy, but the gross guttata formation determines the endothelial dysfunction.     

A comparison of corneal endothelial morphology in cornea guttata, Fuchs' dystrophy and bullous keratopathy

Changes on corneal endothelial specular microscopy were compared in 14 patients with cornea guttata, 4 patients with Fuchs' corneal dystrophy and 19 patients with various forms of bullous keratopathy. The patients with cornea guttata showed well marked guttae 1 to 6 endothelial cells in diameter in the endothelial mosaic and in the relief mode while the endothelial mosaic was usually otherwise within normal limits. In 2 patients with Fuchs' dystrophy the endothelium could be examined, showing gross guttae but a few areas of relatively normal endothelial cells. The unaffected eye of 3 other patients showed findings similar to cornea guttata, but with some reduction in endothelial cell count in 2 patients. The patients with bullous keratopathy fell into 2 groups, one with gross reduction in cell count in a markedly abnormal endothelial cell mosaic, the other a mixed group with moderate reduction in cell count and numerous guttae. Some miscellaneous cases included one of aphakic peripheral bullous keratopathy, one associated with cyclitis and aphakia and 2 with idiopathic non-surgical bullous keratopathy. We believe the corneal endothelium is not grossly abnormal away from the guttae in Fuchs' dystrophy, but the gross guttata formation determines the endothelial dysfunction.     

Freeze-fracture label of functional and dysfunctional human corneal endothelium

Previous freeze-fracture results from our laboratory have shown a reduction in a population of intramembrane particles in the lateral endothelial membranes from dysfunctional human corneas. The size range of these intramembrane particles corresponds to that which has previously been reported for the glycoprotein enzyme Na, K-ATPase in enzyme enriched freeze-fractured membranes. In order to investigate glycoconjugate changes potentially related to the particle reduction, wheat germ agglutinin (WGA), which has been shown to bind to the sugar residues in the ATPase subunit, was used to label three types of corneas with dysfunctional endothelial cells (Fuchs' endothelial dystrophy, aphakic and pseudophakic bullous keratopathy) and two types of corneas (eye bank and keratoconus) with functional endothelium using the technique of thin section freeze-fracture label. Apical WGA labelling on all types of dysfunctional cells was shown to be drastically reduced in comparison to both types of functional corneal endothelial cells. Lateral membranes of dysfunctional cells, exposed by freeze-fracturing, also showed a great reduction in WGA labelling as compared to the fractured lateral membranes of functional cells. The differences observed in lectin labelling of lateral membranes may be related in part to the decreased intramembrane particle density observed in dysfunctional human corneal endothelial cells.     

Moderate Fuchs' endothelial dystrophy ATPase pump site density

The Na+, K+-ATPase pump site density on corneal endothelial cells from Fuchs' endothelial dystrophy corneas has been shown to be drastically decreased in end-stage disease (McCartney et al, Invest Ophthalmol Vis Sci 28:1955, 1987) and significantly increased in the early stages (Geroski et al, Ophthalmology 92:759, 1985) as compared to normal endothelium. In order to provide values for corneas between these two extremes, eye bank corneas from donors with no evidence of corneal edema but with guttata across the extent of the cornea were processed for autoradiography as well as immunohistochemistry. Pump site density was increased compared to end-stage disease but was less than values reported for either functional tissue or early stage disease. Similarly, immunohistochemistry results showed the amount of Na+, K+-ATPase antibody localization to be increased in respect to end-stage disease, but reduced as compared to functional tissue. These results suggest that pump site density on endothelial cells affected with Fuchs' endothelial dystrophy follows a gradual decline towards end-stage disease values as opposed to a sudden sharp deterioration after an initial increase.     

Keratopathy in pseudoexfoliation syndrome as a cause of corneal endothelial decompensation: a clinicopathologic study

PURPOSE: To provide clinical and histopathologic evidence of a distinct keratopathy as a potential cause of corneal edema in patients with pseudoexfoliation syndrome. DESIGN: Retrospective observational case series. PARTICIPANTS: Twenty-two patients with clinically diagnosed pseudoexfoliation syndrome undergoing penetrating keratoplasty for irreversible corneal endothelial decompensation. METHODS: The clinical and histopathologic findings of the corneal buttons are described compared with classic Fuchs' endothelial dystrophy. RESULTS: Clinically, the patients showed diffuse corneal edema, a pleomorphic and numerically reduced corneal endothelium, and retrocorneal flakes of pseudoexfoliation material in three cases. Histopathologically, all corneal buttons showed an abnormal diffuse, irregular thickening of Descemet's membrane and focal accumulations of locally produced pseudoexfoliation material onto or within Descemet's membrane in seven cases. The absence of typical guttata, a higher degree of fibroblastic transformation and melanin phagocytosis of endothelial cells, and a more pronounced endothelial cell loss distinguished the pseudoexfoliation specimens from specimens with classical Fuchs' dystrophy even in the absence of the pathognomonic pseudoexfoliation material. CONCLUSIONS: In patients with pseudoexfoliation syndrome, a distinct type of corneal endotheliopathy may occur, which can lead to an early corneal endothelial decompensation and which might have been previously misdiagnosed as an "atypical nonguttata Fuchs' endothelial dystrophy." This pseudoexfoliation keratopathy may potentiate the known complications in pseudoexfoliation eyes.     

Freeze fracture study of human corneal endothelial dysfunction

Intramembrane changes occurring in dysfunctional corneal endothelial cell membranes were examined using freeze-fracture and transmission electron microscopy techniques. Three categories of dysfunctional endothelium were examined: aphakic bullous keratopathy, pseudophakic bullous keratopathy and Fuchs' endothelial dystrophy. Keratoconus corneas and a donor eye bank eye were examined as normal controls. Four intramembrane changes were observed on replicas of freeze-fractured membranes in each category of dysfunctional endothelium. These were a marked reduction in intramembrane particle density on lateral membranes, altered apico-lateral junctional complexes, increased vesicle fusion sites on apical, lateral and basal membranes, and abnormal desmosome-like particle aggregates on the lateral endothelial cell membranes. The marked reduction in intramembrane particles on lateral membranes may be due to a change in the macromolecular components associated with pump dysfunction. The increase in membrane vesicle fusion sites and the breakdown in intercellular junctions may be associated with increased permeability during barrier dysfunction.     

Screening human donor corneas during organ culture for the presence of guttae

AIMS: To detect the presence of guttae by means of light microscopy during organ culture and to evaluate the influence of the presence of guttae in the donor tissue on transplantation outcome. METHODS: Donor corneas were investigated for the presence of guttae by means of light microscopy at the end of organ culture. Recipient corneal buttons from patients with severe Fuchs' dystrophy and donor corneas with advanced guttae were first studied by light microscopy and subsequently by transmission electron microscopy. Lastly, 168 consecutive donor corneas were evaluated for the presence of guttae and issued for transplantation. RESULTS: Corneal specimens with Fuchs' dystrophy displayed numerous round highly reflecting guttae at the level of the corneal endothelium. Donor corneas with advanced guttae showed less numerous guttae. Among 168 organ cultured donor corneas issued for transplantation, low density guttae were found in 43 (25.6%) corneas. The endothelial cell density and figure coefficient were significantly lower and organ culture time was significantly higher in the cornea guttata group than in the control group. The presence of grouped guttae significantly decreased the adjusted graft survival. The incidence of postoperative stage 3 cornea guttata was significantly higher when grouped guttae were found (5/6) than when no guttae or scattered guttae were found (8/101). CONCLUSION: Cornea guttata can be detected during organ culture by means of light microscopy. It is associated with a decrease in endothelial cell figure coefficient and cell density. The presence of grouped guttae is associated with poorer graft survival and more frequent stage 3 cornea guttata in the graft after transplantation.     

Confocal microscopy in cornea guttata and Fuchs'' endothelial dystrophy

AIMS: To report the appearances of cornea guttata and Fuchs' endothelial dystrophy from white light confocal microscopy. METHODS: Seven eyes of four consecutive patients with cornea guttata were prospectively examined. Of the seven eyes, three also had corneal oedema (Fuchs' dystrophy). In vivo white light tandem scanning confocal microscopy was performed in all eyes. Results were compared with non-contact specular microscopy. RESULTS: Specular microscopy was precluded by corneal oedema in one eye. In the remaining six eyes, it demonstrated typical changes including pleomorphism, polymegathism, and the presence of guttae appearing as dark bodies, some with a central bright reflex. In all seven eyes, confocal microscopy revealed the presence of round hyporeflective images with an occasional central highlight at the level of the endothelium. Changes in cell morphology and size were readily appreciated. CONCLUSION: By comparison with specular microscopy, the hyporeflective images with an occasional central highlight seen on confocal microscopy are consistent with the presence of guttae. Confocal microscopy may confirm the diagnosis of cornea guttata and Fuchs' endothelial dystrophy by demonstrating the presence of guttae. This technique is especially valuable in cases of corneal oedema, where specular microscopy may fail to visualise the endothelium. However, specular microscopy should remain the method of choice to evaluate the endothelium, principally because it is easier to use.     

Epithelial cell, keratocyte, and endothelial cell apoptosis in Fuchs' dystrophy and in pseudophakic bullous keratopathy

PURPOSE: To elucidate the pathomechanism of Fuchs' dystrophy and pseudophakic bullous keratopathy (PBK) by examining cell apoptosis in different corneal layers. METHODS: The authors studied corneal buttons obtained from 21 eyes following central penetrating keratoplasty: 14 corneal buttons (13 patients, age 70.8+/-10.0 years) with Fuchs' dystrophy, and 7 buttons (7 patients, age 69.6+/-10.2 years) with PBK. Four buttons from enucleated eyes with choroidal melanoma served as controls. Histologic changes were examined using light microscopy with hematoxylin-eosin (HE) staining. The average numbers of apoptotic cells per field of view (125x magnification) in separate samples of the epithelial, stromal, and endothelial layers were determined using the TUNEL (terminal deoxyribonucleotidyl transferase-mediated dUTP-digoxigenin nick-end labeling) assay. RESULTS: In 11 of the Fuchs' dystrophy corneas and 2 of the PBK corneas, apoptotic activity was detected. In the control corneas no apoptotic activity was found. Compared to the controls there was a statistically significant difference in the mean (normalized) apoptotic cell numbers for all three layers (p=0.01 in each case) in the Fuchs' dystrophy corneas, and for the stromal layer (p<0.01) in PBK corneas. The apoptotic cell numbers for the epithelial and endothelial layers of the latter were higher, but the difference was not statistically significant (p=0.07, 0.07). CONCLUSIONS: Apoptosis may play a role in the pathomechanism of Fuchs' dystrophy and in keratocyte death in corneas with PBK.     

A correlative electron microscopic and freeze-fracture examination of cat corneal endothelial wound repair

In an attempt to create a model for sustained corneal edema in humans, the present study has examined wounded cat corneal endothelium. Small central (7 mm) wounds or large 90 percent debridement wounds were created with an olive tipped cannula and corneas sampled from 1 to 75 days post-wounding were processed for light and transmission electron microscopy and freeze-fracture. In small wounds, wound closure was complete by 14 days and corneal edema was absent. During wound closure, leading edge cell membranes had decreased intramembrane particles, numerous vesicle fusion sites and lacked cell junctions. Endothelium behind the wound margin was multilayered with fragmented cell junctions. After wound closure, endothelium returned to the morphology of non-wounded endothelium except that an abnormal posterior collagenous layer (PCL) was present. Wound closure was greatly retarded in large wounds and corneas remained edematous at 75 days. The morphology of the endothelial cells was similar to that in small wounds except for the presence of large multinucleated cells and a thicker PCL. These large wound findings are similar to those observed in chronically stressed dysfunctional human corneal endothelium and in this animal model may represent a similar response to injury.     

Central corneal thickness and intraocular pressure measures in human corneas with endothelial guttata: an observational quality control study

Purpose: The aim of the present study was to assess central corneal thickness (CCT) and intraocular pressure (IOP) in eyes where the corneas were affected by different degrees of severity of endothelial pseudo‐guttata or guttata. Methods: In a prospective, case series observational study, non‐contact tonometry and non‐contact specular microscopy (NCSM) with pachymetry for central corneal thickness measures were undertaken as routine procedures on predominantly older patients without a history of corneal problems or contact lens wear. For those showing any signs of corneal endothelial abnormalities, images of the central cornea endothelium were further processed to measure the area (as a percentage) occupied by the guttata. Results: Abnormal endothelial images were obtained from 43 patients (seven with bilateral changes) with an average age of 67.5 years. Between 1.5 and 54.9 per cent of the endothelial images were affected by guttata, which were assigned grade 1 (20 eyes), grade 2 (18 eyes) or grade 3 (11 eyes). When assessed by grade, the central corneal thickness increased and the measured IOP decreased as the guttata became more numerous and confluent. Regression analyses revealed only a weak association between central corneal thickness (p = 0.044, r = 0.149) or the measured IOP (p = 0.090, r = ‐0.244) and the extent of the guttata (percentage). With the apparently contrasting IOP and central corneal thickness effects, no significant IOP‐CCT relationship was noted (p ≥ 0.268, r ≤ 0.160). Conclusions: Where corneas have mild‐to‐modest non‐dystrophic endothelial guttata, there may be a less predictable effect of corneal thickness on the outcome of tonometry.     

ATPase pump site density in human dysfunctional corneal endothelium

Proper corneal hydration is maintained by a Na, K-ATPase pump located in the lateral membranes of the endothelial cells. In dysfunctional corneas this pumping action appears to break down as the corneas become edematous. In order to provide quantitative and qualitative data on the Na, K-ATPase pump site density on dysfunctional and functional human corneal endothelial cells, the present study has employed both autoradiographic and histochemical techniques. Computer-assisted morphometrics and statistical analysis showed that there was a significant reduction (P less than 0.001) in 3H-ouabain binding, and thus ATPase pump sites, in the three types of corneas (Fuchs' endothelial dystrophy, aphakic and pseudophakic bullous keratopathy) with dysfunctional endothelia as compared to both types of corneas (eye bank, keratoconus) with functional endothelial cells. There were no significant differences amongst the dysfunctional types or between the two functional types of corneal endothelial cells in respect to density of silver grains. Histochemical staining for ATPase showed less p-nitro-phenylphosphatase histochemical reaction product present on dysfunctional endothelial lateral membranes than in the functional cells.     

Two-dimensional gel electrophoretic comparison of endothelial cell-Descemet's membrane proteins in Fuchs' dystrophy and normal corneas.

Two-dimensional polyacrylamide gel electrophoresis was used to compare the proteins isolated from the combined corneal endothelial cell-Descemet's membrane complex of three pairs of corneas with Fuchs' dystrophy with three pairs of normal corneas. Normal or Fuchs' dystrophy endothelium and Descemet's membrane was documented by pathologic analysis of each cornea. Proteins were separated by isoelectric point in the first dimension and molecular weight in the second dimension. Over 300 proteins were resolved from each sample, and similar patterns were noted in both groups. No consistent differences were detected between the corneas with Fuchs' dystrophy and the normal corneas. Allelic variations of some proteins were detected in both groups.     

Gas stress test for assessment of corneal endothelial function

PURPOSE: This study was undertaken to evaluate the endothelial pump function by monitoring both corneal swelling response under hypoxia and dehydration response following hypoxia in vivo. METHODS: Humidified nitrogen gas was used to obtain corneal swelling, and humidified gas mixed with oxygen and nitrogen was used for corneal dehydration. First, in 6 young volunteers, we investigated the most suitable oxygen level for evaluating pump function by changing oxygen levels. Then, with the optimal oxygen level, we attempted to evaluate pump function in 53 normal subjects, 5 Fuchs' dystrophy patients, and 3 iridocorneal endothelial syndrome (ICE) patients. RESULTS: Swelling rate showed similar values regardless of age, but both dehydration rate and swelling rate plus dehydration rate decreased with aging. The swelling rate of 5 guttata corneas was significantly higher than that of age-matched control corneas. In contrast, dehydration rate markedly decreased in guttata corneas, while the swelling rate plus dehydration rate of guttata corneas was comparable to that of age-matched corneas. In the 3 ICE corneas, however, swelling rate, dehydration rate, and swelling rate plus dehydration rate were markedly lower than those of both the fellow corneas and the age-matched control corneas. CONCLUSION: These observations lead us to conclude that in order to evaluate pump function, it is necessary to monitor not only dehydration response following hypoxia but also swelling response under hypoxia.     

Growth of human corneal endothelium on altered Descemet's membrane.

To determine whether Descemet's membrane (DM), which is altered by disease, interferes with endothelial cell growth, healthy human corneal endothelial cells were seeded onto DM from normal corneas and diseased corneal buttons from patients with Fuchs' endothelial dystrophy and pseudophakic bullous keratopathy (PBK). DM was first peeled off the corneal specimens and the endothelial cells removed by trypsinization. A suspension of first-passage corneal endothelial cells (2,000 cells/microliters; obtained from donor eye bank eyes and grown in Dulbecco's minimal essential medium with 10% fetal bovine serum and 1.5% chondroitin sulfate) were seeded on DM. Epidermal growth factor (10 ng/ml) and insulin (1 U/ml) were added to the medium after seeding cells on the DM. The cells attached and flattened within 1 hour and reached confluency in 1 week on normal DM. Cells grown on DM from corneas of patients with Fuchs' endothelial dystrophy also flattened and grew to confluency in 1 week. Cells grown on DM from corneas of patients with PBK did not grow to confluency. Further studies with bovine and rabbit corneal endothelial cells showed similar growth pattern to human cells. These data indicate that DM from corneas of patients with Fuchs' dystrophy does not interfere with the growth of corneal endothelial cells but that DM from corneas of patients with PBK does interfere with cell growth.     

Immunohistochemical localization of ATPase in human dysfunctional corneal endothelium

Previous light microscopic autoradiographic results from our laboratory have shown a quantitative decrease in ATPase pump site density in dysfunctional corneal endothelium. In order to develop antibody techniques to correlate these findings with electron microscopic localization of pump sites, three types of corneas with dysfunctional endothelium (Fuchs' endothelial dystrophy, aphakic and pseudophakic bullous keratopathy) and two types of corneas with functional endothelium (age matched eye bank and keratoconus) were obtained at the time of transplant surgery. Corneas were fixed, frozen and cryostat sections were incubated in a rabbit kidney ATPase primary antibody followed by incubation in an HRP-labeled secondary antibody. Functional endothelia showed dense antibody labeling along the lateral cell membranes and there was a substantial reduction in labeling along the lateral membranes of dysfunctional endothelia. Positive tissue controls of rabbit kidney showed HRP reaction product in the convoluted tubules. Control tissue, incubated in either non-immune serum or primary antibody preabsorbed with ATPase, confirmed the specificity by having diminished or no reaction product. These results provide immunohistochemical confirmation of our autoradiographic data showing a quantitative reduction in pump site density on dysfunctional corneal endothelium.     

Clinical and specular micrographic findings in corneal guttata and Fuchs' corneal endothelial dystrophy

Focal excrescenses of altered basement membrane material, known as corneal guttata, are a common occurrence in the elderly. A small proportion of those who develop this condition exhibit signs of corneal endothelial dysfunction and progress to Fuchs' corneal endothelial dystrophy. The relationship between the distribution and relative density of guttata and the morphological appearance of the corneal endothelium is as yet poorly understood. In this study, 40 eyes of 20 caucasian patients (5 male and 15 female) identified as having central corneal endothelial guttata were examined (mean age 76 years). Ten eyes noted as having a previous history of ocular trauma, surgery or anterior segment inflammation were excluded from the study. Clinical and specular micrographic findings are compared. Clinical examination confirmed that guttata were most densely distributed in the central cornea. Examination of all four mid-peripheral quadrants identified the superior cornea as being least severely affected in the majority of cases. In 26 of the 30 eyes examined there was evidence of pigmentary dusting of the corneal endothelium. The severity of guttata did not appear to correlate with the density or distribution of the pigment. In cases of moderate to advanced disease, where guttata had become confluent and thus obscured the view of the corneal endothelium, the clinical differentiation of pigment, guttata and diffuse scarring became difficult. Specular microscopy confirmed the accuracy of the clinical five-point scale used to quantify the distribution and density of guttata in the study. Quantitative analysis of corneal endothelial cells from the mid-peripheral cornea indicated that the severity of guttata did not necessarily correlate with a reduced cell count. Central cell counts proved elusive as reflections from the irregular surface created by confluent guttata failed to produce specular images of the underlying endothelium. The authors conclude that the specular microscope is a useful diagnostic tool in the differentiation of guttata from other abnormalities of the endothelium and Descemet's membrane. Specular micrographic imaging also enables the examiner to quantify the distribution and density of guttata. The relationship between the severity of corneal guttata, endothelial cell counts and pigmentation is not well defined and further study of the progression of the disease process is required.     

The pathologically changed Descemet membrane. Cultivation of human corneal endothelium on transformed membrane]

Corneal endothelial diseases are connected with structural and biochemical changes of Descemet's membrane (DM). Little is known about the possible effects of these changed basement membranes. In the present study we investigated the influence of changed DM on endothelium. Human corneal endothelial cells were cultured on DM of healthy corneas as well as corneas with Fuchs' endothelial dystrophy or bullous keratopathy, and cell growth and morphology were compared. On healthy DM, cells formed a confluent monolayer within 4 days. Cells cultured on Fuchs' DM showed a similar pattern. The polygonal cell pattern was interrupted merely in the area of the guttae, which were covered by cell processes alone. Culturing cells proved to be difficult on DM from corneas with bullous keratopathy. Even though the cells attached to the DM, they did not spread but became spindle shaped and rarely formed intercellular contacts. Even after 7 days no confluent monolayer was established. These results indicate that some though not all pathological DM changes interfere with cell growth.