Treatment receipt and outcomes among lung cancer patients with depression.

AIMS: Among lung cancer patients, depression has been associated with increased mortality, although the mechanisms are unknown. We evaluated the association of depression with mortality and receipt of cancer therapies among depressed veterans with lung cancer. MATERIALS AND METHODS: A retrospective, cohort study of lung cancer patients in the Veterans Affairs-Northwest Health Network from 1995 to 2010. Depression was defined by ICD-9 coding within 24 months before lung cancer diagnosis. Multivariable Cox proportional analysis and logistic regression were used. RESULTS: In total, 3869 lung cancer patients were evaluated; 14% had a diagnosis of depression. A diagnosis of depression was associated with increased mortality among all stage lung cancer patients (hazard ratio = 1.14, 95% confidence interval: 1.03-1.27, P = 0.01). Among early-stage (I and II) non-small cell lung cancer (NSCLC) patients, the hazard ratio was 1.37 (95% confidence interval: 1.12-1.68, P = 0.003). There was no association of depression diagnosis with surgery (odds ratio = 0.83, 95% confidence interval: 0.56-1.22, P = 0.34) among early-stage NSCLC patients. A depression diagnosis was not associated with mortality (hazard ratio = 1.02, 95% confidence interval: 0.89-1.16, P = 0.78) or chemotherapy (odds ratio = 1.07, 95% confidence interval: 0.83-1.39, P = 0.59) or radiation (odds ratio = 1.04, 95% confidence interval: 0.81-1.34, P = 0.75) receipt among advanced-stage (III and IV) NSCLC patients. Increased utilisation of health services for depression was associated with increased mortality among depressed patients. CONCLUSIONS: Depression is associated with increased mortality in lung cancer patients and this association is higher among those with increased measures of depression care utilisation. Differences in lung cancer treatment receipt are probably not responsible for the observed mortality differences between depressed and non-depressed patients. Clinicians should recognise the significant effect of depression on lung cancer survival.

Cigarette smoke induces genetic instability in airway epithelial cells by suppressing FANCD2 expression.

Chromosomal abnormalities are commonly found in bronchogenic carcinoma cells, but the molecular causes of chromosomal instability (CIN) and their relationship to cigarette smoke has not been defined. Because the Fanconi anaemia (FA)/BRCA pathway is essential for maintenance of chromosomal stability, we tested the hypothesis that cigarette smoke suppresses that activity of this pathway. Here, we show that cigarette smoke condensate (CSC) inhibited translation of FANCD2 mRNA (but not FANCC or FANCG) in normal airway epithelial cells and that this suppression of FANCD2 expression was sufficient to induce both genetic instability and programmed cell death in the exposed cell population. Cigarette smoke condensate also suppressed FANCD2 function and induced CIN in bronchogenic carcinoma cells, but these cells were resistant to CSC-induced apoptosis relative to normal airway epithelial cells. We, therefore, suggest that CSC exerts pressure on airway epithelial cells that results in selection and emergence of genetically unstable somatic mutant clones that may have lost the capacity to effectively execute an apoptotic programme. Carcinogen-mediated suppression of FANCD2 gene expression provides a plausible molecular mechanism for CIN in bronchogenic carcinogenesis.

A human gene coding for a membrane-associated nucleic acid-binding protein.

Studies to clone a cell-surface DNA-binding protein involved in the binding and internalization of extracellular DNA have led to the isolation of a gene for a membrane-associated nucleic acid-binding protein (MNAB). The full-length cDNA is 4.3 kilobases with an open reading frame of 3576 base pairs encoding a protein of approximately 130 kDa (GenBank accession numbers and ). The MNAB gene is on human chromosome 9 with wide expression in normal tissues and tumor cells. A C3HC4 RING finger and a CCCH zinc finger have been identified in the amino-terminal half of the protein. MNAB bound DNA (K(D) approximately 4 nm) and mutagenesis of a single conserved amino acid in the zinc finger reduced DNA binding by 50%. A potential transmembrane domain exists near the carboxyl terminus. Antibodies against the amino-terminal half of the protein immunoprecipitated a protein of molecular mass approximately 150 kDa and reacted with cell surfaces. The MNAB protein is membrane-associated and primarily localized to the perinuclear space, probably to the endoplasmic reticulum or trans-Golgi network. Characterization of the MNAB protein as a cell-surface DNA-binding protein, critical in binding and internalization of extracellular DNA, awaits confirmation of its localization to cell surfaces.

Quantification of DNA binding to cell-surfaces by flow cytometry.

DNA binding to cell-surfaces has been documented in several studies. The interaction of DNA with cells has been shown to have therapeutic potential as a non-viral form of gene delivery and DNA vaccination. Recently, bacterial DNA binding and internalization has been demonstrated in some cells to trigger secretion of cytokines and cell activation. Previous studies to quantify DNA binding to cells have used radiolabeled DNA. Here we report a non-radioactive assay for quantification of cell-surface DNA binding based on the isoparametric analysis of flow cytometric data as described by Chatelier et al., Embo J., 5 (1986) 1181. This assay has the advantage over previously used procedures in not employing radioactive material and being able to discriminate viable from non-viable cells that bind DNA. With the importance of understanding the interaction of DNA with cells, this assay may have application for the identification and characterization of reagents designed to either enhance or inhibit DNA binding to cells.

Protein movement across cultured guinea pig trachea: specificity and effect of transcytosis inhibitors.

Airway surface liquid (ASL) is a complex fluid with solutes including electrolytes, lipids, mucins, and proteins. The proximal airways are absorptive for most solutes, including proteins. We investigated the process of protein movement across confluent primary cultures of guinea pig trachea grown on filters using fluorescent-labeled bovine serum albumin (BSA), ovalbumin (OA), and 70-kDa dextran (Dex). We found marked asymmetry of BSA and OA transepithelial flux, with apical-to-basolateral flux (JA-->B) 10 times greater than the opposite direction (JB-->A) for both proteins. The apparent permeability for Dex was the same as that for proteins in the basolateral-to-apical direction and showed no asymmetry. Increasing concentrations of unlabeled BSA, OA, or transferrin inhibited JA-->B for both BSA and OA without affecting Dex movement. Cooling reduced JA-->B for BSA without affecting JB-->A. Monensin and nocodazole each reduced JA-->B for BSA and OA without affecting JB-->A. Monensin eliminated all asymmetry for BSA movement. Brefeldin A did not affect JA-->B for either protein but did increase JB-->A for BSA. Treatment with the protease inhibitors increased JA-->B for BSA. Western immunoblotting demonstrated immunologically intact protein in the downstream compartment. We conclude that there is transcytosis of proteins across cultured trachea epithelium in the apical-to-basolateral direction, which is monensin sensitive, involves microtubules, is not dependent on proteolysis, and is not protein species specific. This process may be important for maintenance of the ASL, and defects in this process may contribute to the abnormally thickened airway secretion seen in airway diseases.

Diffusion of nonelectrolytes in the canine trachea: effect of tight junction.

We recently demonstrated through theoretical modeling that the exhaled ethanol (EtOH) profile from humans is consistent with a molecular diffusion coefficient (cm2/s) in the bronchial mucosa (Dti) that is only 8% of the diffusion coefficient in water (Dw; J. Appl. Physiol. 75: 2439-2449, 1993). Because of the small oil-water partition coefficient (lambda o:w) of EtOH (lambda o:w = 0.074), the reduced diffusion coefficient may be due, in part, to the epithelial tight junction in the paracellular pathway. We hypothesized that opening the tight junction would open an aqueous pathway and increase the diffusion coefficient of small (mol wt < 100) hydrophilic compounds. We mounted the mucosa from the membranous canine trachea in an Ussing-type diffusion cell and measured the diffusion coefficient of 2-ethoxyethanol (2-Ethx; lambda o:w = 0.042), EtOH, and methyl ethyl ketone (MEK; lambda o:w = 1.04) in the presence and absence of the epithelial tight junction. The tight junction was opened using a phosphate-buffered saline free of Ca2+ and Mg2+ with 0.5 mM ethylene glycol-bis (beta-aminoethyl ether)-N,N,N',N'-tetraacetic acid, and its integrity was assessed by measuring the transepithelial electrical resistance. Dti/Dw in the presence of Ca2+ and Mg2+ was 0.39, 0.34, and 0.39 for 2-Ethx, EtOH, and MEK, respectively, and increased 24.6, 11.7, and 1.11% in the absence of Ca2+ and Mg2+. We conclude that the effect of the tight junction on Dti increases with increasing water solubility but can account for only a small portion of the reduced Dti of EtOH as predicted by exhaled profiles.

Tracheal vascular and smooth muscle responses to air temperature and humidity in dogs.

Twenty-five dogs were anesthetized, paralyzed, and artificially ventilated. Their cranial tracheal arteries were perfused bilaterally with blood at constant flow, and the perfusion pressures (Patr) were measured. Tracheal smooth muscle function was assessed by recording changes in external diameter (delta Dtr). The perfused segment of the trachea was exposed to air at a constant unidirectional airflow of 25 l/min. Group 1 (n = 6) was exposed to cold dry air, ambient room air, and hot dry and hot humid air, each for 10 min with exposures starting from zero flow. The tracheal vascular responses to all four conditions were small vasodilations (delta Patr from -2 to -6%) followed by recovery or small vasoconstrictions. In group 2 (n = 19), exposures to cold dry and hot humid air were preceded and followed by body-temperature fully humidified air. Cold dry air caused a sustained vasodilation (delta Patr -9.0 +/- 1.1%), and hot humid air usually caused a biphasic response: a vasoconstriction (delta Patr 4.4 +/- 1.0%) followed by a vasodilation (delta Patr -5.7 +/- 1.9%). The warm humid air after cold dry air or hot humid air caused a further vasodilation, which lasted a short time after cold dry air (delta Patr -3.7 +/- 0.4%) but greater than 10 min after hot humid air (delta Patr -13.8 +/- 1.4%). In both groups, all exposures that cooled the trachea (cold dry air, ambient room air, and hot dry air) caused smooth muscle contraction, and hot humid air that warmed the trachea caused relaxation.

Laser-Doppler measurement of tracheal mucosal blood flow: comparison with tracheal arterial flow.

Blood flow in the tracheal mucosa (Qm) has been measured by laser-Doppler flowmetry in anesthetized sheep and dogs. The values have been compared with tracheal arterial inflow (Qtr) by use of an electromagnetic flow probe and with tracheal arterial perfusion pressure (Ptr) produced by mechanical perfusion. Changes in blood flow were caused by injections of methacholine, phenylephrine, and histamine into the perfusion circuit. These interventions produced a range of measurements for each animal. Correlations of Qm against Qtr were significant in two of five animals (R = 0.03-0.93); correlations of Qm against Ptr were significant in two of four animals (R = 0.56-0.96). Percent changes in Qtr were generally much larger than those of Qm, and there was considerable variability between Qm and either Qtr or Ptr. Qm reflected the same vascular changes as Ptr or Qtr in 28 interventions and showed an opposing change in 4 cases. In 11 interventions, changes measured by Ptr or Qtr were not reflected by any changes in Qm. Thus qualitative changes in tracheal perfusion measured with these methods were usually the same; quantitatively the three methods showed great differences. These differences may reflect different regulatory mechanisms in various components of the tracheal vasculature or different technical aspects of the methods used.

Lung mechanical effects on the bronchial circulation.

The bronchial circulation, like the pulmonary circulation, is subjected to the mechanical effects of lung inflation with each breath. Lung inflation, in all species studied, decreases the bronchial blood flow. This is due to changes in both lung volume and transpulmonary pressure. With this reduction of flow there is also a relative increase in the proportion of bronchial blood flow drained from the lung through systemic veins. These changes are probably due to purely mechanical effects on the systemic-to-pulmonary anastomotic bronchial blood vessels with both compression and stretching of the vessels.

Prostaglandins alter methacholine-induced secretion in ferret in vitro trachea.

The ferret trachea was mounted in an organ bath containing Krebs-Henseleit solution with additional bovine serum albumin (BSA). Tracheal secretions were collected and analyzed for albumin, and lysozyme, a specific marker of serous cell secretion. The total secretion volume and output and concentrations of albumin and lysozyme were calculated. Secretion was stimulated with methacholine (20 microM) (Mch), and the effects of the prostaglandins PGD2, PGE1, and PGF2 alpha on methacholine-induced secretion were studied. All responses were dose dependent. PGF2 alpha at 10(-5) M increased the volume of Mch-stimulated secretion twofold, the lysozyme output sixfold, and concentration over threefold, while decreasing the albumin transport by one-half. PGD2 at 10(-5) M reduced Mch-induced secretion volume to 75% control, increased albumin transport to 135%, without affecting lysozyme secretion. PGE1 at 10(-5) M increased Mch-stimulated albumin transport and concentration over twofold, decreased lysozyme release to less than one-third of control, and had no effect on secretion volume. PGE1 caused the albumin concentration to exceed that of the outer bath, indicating active transport. We conclude that prostaglandins selectively alter tracheal secretion induced by cholinergic stimuli.

Effects of antigen on tracheal circulation and smooth muscle in sheep of different ages.

The effects of Ascaris suum antigen on tracheal circulation and tracheal smooth muscle tone were compared in two groups of sheep: the first group was 1 yr old (14 sheep) and the second 5 yr old (8 sheep). Cranial tracheal arteries of anesthetized and paralyzed sheep were perfused at constant flow with monitoring of perfusion pressure. Tracheal smooth muscle tone was assessed by measuring changes in the external diameter of the cranial trachea. Close-arterial injection of antigen (1-20 micrograms) in young sheep produced dose-dependent vasodilation (6.1-15.5% fall in perfusion pressure) and smooth muscle contraction (0.06-0.28 mm reduction in tracheal diam). In old sheep, antigen (1-20 micrograms) produced vasoconstriction (4.1-16.8%) but no smooth muscle response. The smooth muscle contraction in young sheep was blocked by mepyramine (2 mg/kg iv) suggesting mediation by release of histamine. The vasodilation in young sheep and the vasoconstriction in old sheep were reduced by indomethacin (5 mg/kg iv), and the residual response was further reduced by FPL 55712 (2 mg/kg iv), suggesting mediation by both cyclooxygenase products and leukotrienes. Thus antigen given in the tracheal vasculature releases a mixture of inflammatory mediators. This mixture of mediators or their actions on the tracheal vasculature and smooth muscle may depend on the age of the sheep.

Pulmonary artery infusion of prostacyclin increases lobar bronchial blood flow.

Intrapulmonary systemic to pulmonary bronchial blood flow [Qbr (s-p)] decreases with administration of cyclooxygenase inhibitors. This effect may be due to a decrease in the production of vasodilating prostaglandins and reflect either a decrease in the total intrapulmonary bronchial blood flow (Qbr), or a redistribution of the intrapulmonary systemic venous return. In nine open chested dogs the left lower lobe (LLL) was isolated and perfused in situ. Blood flow to the extrapulmonary airways (Qep), and Qbr were measured by the reference flow technique. Qbr (s-p) was measured as the overflow from the closed LLL perfusion circuit. After ibuprofen, PG-I2 was infused into the LLL PA and the Qbr (s-p) was continuously monitored. Qbr, and Qep were measured before and after ibuprofen, and during and after the PG-I2 infusion. The upstream pressure for Qbr (s-p) was estimated with and without PG-I2 infusion. After ibuprofen the Qep, Qbr, and Qbr (s-p) fell to 45, 22, and 17%, respectively, of the pre-ibuprofen values (P less than 0.05). PG-I2 increased the Qbr (s-p) and Qbr (P less than 0.05), while Qep was unchanged. During all experimental conditions the simultaneous measurements of Qbr and Qbr (s-p) were not different from each other (P less than 0.001). The upstream pressure for Qbr (s-p) increased from 30 to 50 cm H2O (P less than 0.05). Intralobar bronchial blood flow is drained almost entirely through the pulmonary circulation, and PG-I2 in the LLL pulmonary circulation increases systemic blood flow to the LLL, probably acting at the level of a systemic arteriole.

Cold and hyperosmolar fluids in canine trachea: vascular and smooth muscle tone and albumin flux.

We have studied the effects of liquids of various osmolalities and temperatures on the tracheal vasculature, smooth muscle tone, and transepithelial albumin flux. In 10 anesthetized dogs a 10- to 13-cm length of cervical trachea was cannulated to allow instillation of fluids into its lumen. The cranial tracheal arteries were perfused at constant flow, with monitoring of the perfusion pressures (Ptr) and the external tracheal diameter (Dtr). Control fluid was Krebs-Henseleit solution (KH) with NaCl added to result in a 325-mosM solution (isotonic). Hypertonic solutions were KH with NaCl (warm hypertonic) or glucose (hypertonic glucose) added to result in a 800-mosM solution. All solutions were at 38 degrees C, with isotonic and the hypertonic NaCl solutions also given at 18 degrees C (cold isotonic and cold hypertonic). Fluorescent labeled albumin was given intravenously, and the change in fluorescence in the fluid was measured during each 15-min period. Changing from warm isotonic to cold isotonic decreased Dtr and Ptr. Changing from warm isotonic to warm hypertonic or hypertonic glucose decreased Ptr with no change in Dtr. The cold hypertonic responses were not different from cold isotonic responses. Warm hypertonic solution increased albumin flux into the tracheal lumen over a 15-min period to three times that of the control period, persisting for 15 min after replacement with warm isotonic solution. Cooling induces a vasodilation and smooth muscle contraction of the trachea, whereas hypertonic solutions result in vasodilation and, if osmolality is increased with NaCl, an increase in albumin flux into the tracheal lumen.

Bronchial circulation and cyclooxygenase products in acute lung injury.

The role of cyclooxygenase products in the response of the bronchial circulation to acute lung injury was examined in 30 dogs. By use of an open-chest preparation the left lower lobe (LLL) pulmonary circulation was isolated, continuously weighed, and perfused in situ. The anastomotic bronchial blood flow [Qbr(s-p)] was measured as the rate of increase in the volume of the LLL-perfusion circuit. Four groups of dogs were studied. In group A, six dogs received cyclooxygenase inhibition (COI) with either indomethacin (2 mg/kg) or ibuprofen (10 mg/kg). In group B (n = 10) lung injury caused by airway instillation of glucose (15 mg) with glucose oxidase (500 micrograms/kg) (G/GO) or LLL pulmonary arterial infusion of alpha-napthyl thiourea (ANTU, 2 mg/kg). Group C (n = 10) received COI, and 30 min later injury was induced as above with either ANTU or G/GO. Group D (n = 4) received COI immediately after anesthesia; then, 30 min after completion of the surgical preparation, injury was induced with ANTU or G/GO. After COI, Qbr(s-p) decreased to 35 +/- 9% of the basal values (P less than 0.05). After administration of ANTU or G/GO, Qbr(s-p) increased irrespective of whether COI was present. 6-Ketoprostaglandin F1 alpha (6-keto-PGF1 alpha) and thromboxane B2 (TxB2) were measured by radioimmunoassay in the LLL pulmonary artery and systemic venous blood, demonstrating an increase in 6-keto-PGF1 alpha due to surgical preparation and confirming complete COI in those animals receiving COI immediately after anesthesia. These findings demonstrate that 1) the bronchial circulation is capable of a sevenfold increase in flow in response to acute lung injury.(ABSTRACT TRUNCATED AT 250 WORDS)

Upstream pressure for systemic to pulmonary flow from bronchial circulation in dogs.

Systemic to pulmonary flow from bronchial circulation, important in perfusing potentially ischemic regions distal to pulmonary vascular obstructions, depends on driving pressure between an upstream site in intrathoracic systemic arterial network and pulmonary vascular bed. The reported increase of pulmonary infarctions in heart failure may be due to a reduction of this driving pressure. We measured upstream element for driving pressure for systemic to pulmonary flow from bronchial circulation by raising pulmonary venous pressure (Ppv) until the systemic to pulmonary flow from bronchial circulation ceased. We assumed that this was the same as upstream pressure when there was flow. Systemic to pulmonary flow from bronchial circulation was measured in left lower lobes (LLL) of 21 anesthetized open-chest dogs from volume of blood that overflowed from pump-perfused (90-110 ml/min) pulmonary vascular circuit of LLL and was corrected by any changes of LLL fluid volume (wt). Systemic to pulmonary flow from bronchial circulation upstream pressure was linearly related to systemic arterial pressure (slope = 0.24, R = 0.845). Increasing Ppv caused a progressive reduction of systemic to pulmonary flow from bronchial circulation, which stopped when Ppv was 44 +/- 6 cmH2O and pulmonary arterial pressure was 46 +/- 7 cmH2O. A further increase in Ppv reversed systemic to pulmonary flow from bronchial circulation with blood flowing back into the dog. When net systemic to pulmonary flow from bronchial circulation by the overflow and weight change technique was zero a small bidirectional flow (3.7 +/- 2.9 ml.min-1 X 100 g dry lobe wt-1) was detected by dispersion of tagged red blood cells that had been injected.(ABSTRACT TRUNCATED AT 250 WORDS)

Temperature dependence of intraparenchymal bronchial blood flow.

Previous studies suggested that bronchial vascular resistance, like that of the skin, changes with the temperature of the surrounding tissue. To investigate this phenomenon, we recorded anastomotic (systemic to pulmonary) (Qbrs-p) and total (Qbr) bronchial blood flow over a temperature range centered on normal. In 7 open-chested dogs the in situ left lower lobe (LLL) was separately ventilated (30 degrees C, 5% CO2 in humidified air) and was suspended in a fabric net from a strain gauge for continuous recording of weight. The pulmonary circulation of the LLL was pump-perfused at 255 +/- 69 ml/min in a closed circuit with temperature set at 30, 33, 36, 39 and 42 degrees C. Qbrs-p was measured as overflow from the LLL vascular circuit corrected for LLL weight changes. Qbr, tracheal, mid-esophageal and coronary flow were measured with 15 mu radiolabelled microspheres injected in the left atrium. The animal's core temperature and that of the humidified air around the LLL were held constant. Qbr and Qbrs-p were equal and reached a peak at 36 degrees C with lower levels of flow at higher and lower temperatures. Esophageal, tracheal and coronary flow and cardiac output did not change nor did pressures in the systemic and LLL pulmonary artery and in the LLL airways. An intralobar change in temperature above or below 36 degrees C decreases only the lobar bronchial blood flow and does not influence blood flow to other nearby tissues including those vascularized by the bronchial circulation.

Temperature alters bronchial blood flow response to pulmonary arterial obstruction.

Lobar bronchial blood flow has been reported to increase and decrease acutely after pulmonary arterial obstruction (PAO). Because bronchial blood flow (Qbr) to the trachea and bronchi is influenced by inspired air temperature, we investigated whether temperature differences could explain these disparate results. In 10 open-chested dogs the left lower lobe (LLL) was isolated and perfused in situ with autologous blood at a controlled temperature with an independent vascular circuit. The abdomen and the chest of the dog were enclosed in a Plexiglas box in which air was fully humidified and temperature could be regulated. Qbr, determined by the reference flow technique using 16 micron microspheres, was measured before and 30 min after onset of PAO with the air in the box being either at 27 or 39 degrees C and with warmed LLL blood (37 degrees C) in the latter condition. Anastomotic bronchial blood flow [Qbr(s-p), determined as overflow from the closed LLL vascular circuit and measured in ml X min-1 X 100 g dry lung wt-1 X 100 Torr mean systemic pressure-1] was measured continuously at both temperatures. Both before and after PAO, Qbr and Qbr(s-p) were closely correlated: Qbr (ml/min) = 1.12 + 0.978Qbr(s-p); R = 0.912. This was true regardless of the presence or the absence of pulmonary flow, showing that the distribution of bronchial blood flow between the anastomotic and the nonanastomotic portion does not change acutely during PAO. When the air in the box was 27 degrees C, Qbr(s-p) was 19.5 +/- 5.2 (SE) and increased to 38.6 +/- 8.1 with PAO (P less than 0.007).(ABSTRACT TRUNCATED AT 250 WORDS)

Birefringence measurements of structural inhomogeneities in Rana pipiens rod outer segments.

The birefringence (deltan) of Rana pipiens rod outer segments (ROS) reveals microstructure inhomogeneities not seen with other techniques. In the basal 20-30-micron length of the ROS there is a nearly linear axial gradient in deltan of approximately equal to -2 x 10(-5)/micron. No consistent deltan gradients are found in the distal 20-30 micron of the ROS. Using glycerol imbibition to separate the intrinsic and form birefringence components, we found that the basal deltan gradient was principally due to a gradient of the intrinsic birefringence component. The disk membrane volume fraction decreases uniformly from the basal end to the distal end, while the disk membrane refractive index increases. The contributions of these changes to the form birefringence approximately cancel, so that the form component is fairly uniform along the ROS axis. Because its axial distance from the inner segment is a measure of the time since a disk membrane was formed, these gradients may reflect a disk membrane aging process. Occasionally a highly birefringent, 2-micron-wide band is seen at the basal end ot the ROS, possibly where the envelope membrane folds to form new disk membranes.