A Morphometric Study of the Endoplasmic Reticulum in Human

Gut: first published as 10.1136/gut.15.9.737 on 1 September 1974. Downloaded from

Gut, 1974, 15, 737-747

A morphometric study of the endoplasmic reticulum in human hepatocytes' Correlation between morphological and biochemical data in subjects under treatment with certain drugs

A. M. JtZtQUEL2, M. KOCH, AND F. ORLANDI From the Digestive Diseases and Cell Pharmacology Units, University of Ancona School of Medicine, the Centre for the Study of the Action of Drugs on the Liver, University of Camerino, and the Division of Gastroenterology, General Hospital, Ancona, Italy

SUMMARY The distribution of the endoplasmic reticulum in human hepatocytes is defined in quantitative terms using the techniques of morphometry. The subjects of the study are liver biopsies from normal, untreated subjects and patients being treated with various drugs. In contrast to rat hepatocytes, the amount of smooth endoplasmic reticulum (SER) in man exceeds that of the rough endoplasmic reticulum (RER) and accounts for 76.3 % of the total endoplasmic reticulum. This is to be taken into consideration in pharmacological or toxicological studies. In addition, two components of the SER have been identified: more prominent is the type 1 or vesicular which has a regular honeycomb pattern, made up ofcisternae with patent lumina and a mean width of 1500 A; the type 2, or non-vesicular, occurs in discrete foci of densely packed smooth membranes with a spacing of about 140 A. In subjects under short-term treatment with Benzodiazepin (diazepam) the RER re- http://gut.bmj.com/ mained unchanged but the SER membranes were significantly increased with a remarkable, two- to threefold increase of the SER type 2 in three out of four patients. A rise in incorporation of 14C-acetate into digitonin-precipitable sterols as measured in liver biopsy material was also noted in these three patients. The suggestion is made that the SER 2 represents the newly formed membranes whereas the SER 1 would represent 'adult' membranes. No changes were observed in two patients under short-term treatment with phenobarbital or Dilantin. on September 23, 2021 by guest. Protected copyright.

A few morphometric studies have already been con- the normal base-line data necessary for comparative ducted on animal liver in normal (Loud, 1968; purposes. However, such a study appears essential, Weibel, Staubli, Gnagi, and Hess, 1969; Hess, especially when dealing with toxicological or Weibel, and Preisig, 1973) or in experimental con- pharmacological studies. The subjectivity or the ditions (Wiener, Loud, Kimberg, and Spiro, 1968; aesthetic feelings of the observer often interfere Staubli, Hess, and Weibel, 1969; Hope, 1970; with an objective interpretation of findings and do Riede, Seebass, and Rohr, 1971a; Riede, Strassle, not permit a quantitative estimation of the changes Bianchi, and Rohr, 1971b; Dobbins, Rollins, Brooks, observed. In the present study, we have applied the and Fallon, 1972; Bolender and Weibel, 1973) but technique of Weibel, Kistler, and Scherle (1966) and quantitative analyses of human liver are still lack- Weibel et al (1969) to the study of human liver, ing. Ethical and practical considerations make it using the material obtained by needle biopsy in difficult to collect the material necessary for defining normal, untreated patients and in patients treated with various drugs. Our attention was especially Received for publication 15 May 1974. concentrated on drugs metabolized by microsomal 1Supported by a research grant n. 7300691/04 of the CNR, Rome. 'Please address all correspondence to: Dr A. M. Jz6quel, Division of enzymes, influencing bilirubin or cholesterol metab- Gastroenterology, Ospedale Regionale 60100 Ancona, Italy. olism. We attempted to correlate the biochemical 737 3 Gut: first published as 10.1136/gut.15.9.737 on 1 September 1974. Downloaded from

738 A. M. Jezequel, M. Koch, and F. Orlandi findings with eventual changes in the endoplasmic Electron microscopy reticulum. The biochemical data concerning the For electron microscopy, a small piece of tissue was effect of these drugs on hepatic biosynthesis of cut in 1 mm3 fragments and fixed in 1 % osmium cholesterol as measured on needle biopsies have been tetroxide buffered with Sorensen's phosphate buffer published elsewhere (Bamonti and Dini, 1973; pH 7.4. After dehydration with alcohol and propyl- Orlandi, Bamonti, Dini, Koch, and Jezequel, 1974). ene oxide, the fragments were embedded in a mix- ture of Epon and Araldite. The blocks obtained were Material and Methods cut on a LKB Ultratome III. Thick (1 micron) sections were stained with Azur II-Methylene blue SELECTION OF PATIENTS for examination with the light microscope. Ultra- thin sections were stained with lead citrate and ex- Group 1: Normal untreatedpatients amined with the electron microscope Philips EM Four biopsies of normal, untreated patients were 201. analysed. The selection procedure considered the clinical examination, laboratory findings (BSP re- Quantitative analysis tention at 45 min, SGOT, SGPT, serum bilirubin, For quantitative analysis, five blocks of each biopsy serum alkaline phosphatase). Special attention was material were chosen at random and electron micro- paid to the drug-taking history and only patients graphs from each section were taken on 35 mm film, 'drug-free' for at least three weeks were retained. The first at a magnification of4000 on the screen, ie, 1490 selection also took into account the light micro- on the film due to the reduction in the microscope. scopy findings. The randomization of the areas to be photographed was obtained by placing the screen on the lower Group 2: Treatedpatients without liver involvement left corner of a mesh of the copper grid (Weibel et al, Four patients with normal liver function tests and in 1969). If the angle was not suitable (technical im- need of sedative treatment were given Benzodiazepin perfections, extrahepatocytic space, section edges) (diazepam) according to the following scheme: thestage was movedto another corner. Subsequently, Pe. M. received 20 mg per day for two days; Br. P. the primary magnification was brought to 8600 on received 60 mg per day for three days; Ta. P. and the screen (3200 on film) and 24 500 (9100) without

Gi. M. received 40 mg per day for four days. The moving the centre of the picture. Six electron micro- http://gut.bmj.com/ needle biopsy was taken on the morning following graphs were taken at each level of magnification. A the cessation of therapy except in Ta. P. from whom total of 90 pictures was thus obtained for each it was taken at a four-day interval. biopsy. For the purpose of the present study, ie, One patient (Ca. V.) received 50 mg per day of the analysis of the endoplasmic reticulum, measure- diphenylhydantoin (Dilantin) for seven days. A ments were made on pictures taken at a magnifica- biopsy was taken at the beginning oftreatment and is tion of 9100. The films were examined on a project- included in group 1. Another biopsy was taken at ing screen with a secondary tenfold magnification on September 23, 2021 by guest. Protected copyright. the end of the seven-day period. upon which was superimposed a multipurpose test screen with 84 test lines of known length (Weibel et Group 3: Treatedpatient with hepatic dysfunction al, 1966, 1969). The evaluation of the membrane One patient with Gilbert's syndrome and serum bili- surface density Sv ofsmooth androughreticulumwas rubin of 2.06 mg% received 360 mg phenobarbital performed by countingthenumber ofintersections (I) per day for four days. A biopsy was taken before the ofmembrane profiles with the test lines and applying beginning of treatment and another one at the end of the formula Sv = 2 x 1/LT where LT is the length of therapy when the serum bilirubin had fallen to 0.60 the test lines enclosed in the cytoplasm. The results mg%. are expressed in u2 per 3 ofcytoplasm (Weibel et al, 1969). Estimation of the volumetric density (Vv) PROCESSING THE MATERIAL of the endoplasmic reticulum was given by the formula Vv = Pp where Pp is the fraction number of The hepatic biosynthesis ofcholesterol endpoints of the test lines enclosed within the pro- This was estimated on biopsy material as described files of the endoplasmic reticulum (Weibel et al, elsewhere (Orlandi et al, 1974). 1969). The mean width (d) of the cisternae was calculated from the volume to surface ratio (v/s) (Weibel Light microscopy et al, 1969) such as v/s = d/2 for the cisternac of the For light microscopy, a fragment of the biopsy was Golgi apparatus which are likened to plates and fixed in formaldehyde, embedded in paraffin, and v/s = d/4 for the rough and smooth endoplasmic stained with haematoxylin-eosin. reticulum which are likened to cylinders. Gut: first published as 10.1136/gut.15.9.737 on 1 September 1974. Downloaded from

A morphometric study of the endoplasmic reticulum in human hepatocytes 739

Fig 1 Low-power view ofparenchymal cells ofan untreated subject. In the cytoplasm, the SER is diffusely distributed and has a regular honeycomb pattern; only the type 1 or vesicular is present. Some RER cisternae are seen in areas free ofglycogen. Numerous mitochondria (M) contain crystalline formations seen as dense inclusions at this level ofmagnification. The Golgi apparatus (G) is seen next to bile canaliculi (bc) x 7500 http://gut.bmj.com/ on September 23, 2021 by guest. Protected copyright.

Results human liver (Ma and Biempica, 1971; Jezequel and Orlandi, 1972; Klion and Schaffner, 1968). We need ELECTRON MICROSCOPY only to insist on the relative paucity of the rough Group 1: normal untreatedpatients endoplasmic reticulum (RER), usually distributed Previous observations have already given a detailed along the nucleus, around or between mitochondria, account of the ultrastructural features of the normal and along the sinusoidal border of the hepatocyte. Gut: first published as 10.1136/gut.15.9.737 on 1 September 1974. Downloaded from

740) A. M. Jezequel, M. Koch, and F. Orlandi

Fig 2 Part ofa hepatocyte ofa patient (Pe. M.) treated with diazepam (20 mg/day/two days). The distribution of the SER is heterogeneous, due to the alternation of light and dense areas. The former (SER) contain large cisternae ofSER whereas in the latter (SER2) intricate membranes make up a dense network in which patent lumina are rarely seen. In this patient the incorporation of 14C-acetate into DPS was increased threefold. x 17 500

(with permission ofthe http://gut.bmj.com/ Israel Journal of Medical Science). on September 23, 2021 by guest. Protected copyright.

The rest of the cytoplasm was occupied by extensive ally some limited areas were occupied by densely areas of smooth endoplasmic reticulum (SER) packed, smooth membranes with a finely tubular associated with numerous rosettes ofglycogen (fig 1). appearance and a narrow lumen or an intricate The vesicles of the SER were usually round, small, network where cisternal lumina were not clearly and with a fairly regular, honeycomb appearance. recognizable. This is defined as type 2 or 'non- This is defined as SER type 1 or 'vesicular'. Occasion- vesicular' (fig 2). Gut: first published as 10.1136/gut.15.9.737 on 1 September 1974. Downloaded from

A morphometric study of the endoplasmic reticulum in human hepatocytes 741

Fig 3 High-power view ofan area ofproliferated SER type 2 in a patient (Ta. P.) treated with diazepam (40 ng/day/four days); the biopsy was taken four days after the end oftherapy. This picture illustrates the alternation ofthe areas with loosely arranged SER1 and the areas fitted with densely packed smooth membranes ofSER2. In the latter the rosettes ofglycogen are fewer than in the former. Although the biopsy was taken four days after drug withdrawal, the changes ofthe SER were still prominent in this patient and the incorporation of "4C-acetate showed a fivefold increase. x30 000 http://gut.bmj.com/ on September 23, 2021 by guest. Protected copyright.

The surface densities of the total ER, of the RER, ER,themeanvalue ofthe surface density expressed in and the SER is shown in fig 4 which gives the P2/Z3 of cytoplasm was 3-80 ± 0 52 of which 0-89 mean value for the pooled data from the four sub- ± 0-22, ie, about 23 7 %, belonged to the RER, and jects. In addition, we have mentioned the figures 2-91 ± 0-51, about 763 %, to the total SER (2-23 concerning the relationship of SER type 2 to the for type 1 and 0-68 for type 2). The Golgi apparatus total smooth endoplasmic reticulum. For the total occupied 0-23 + 017 ,2/ 3. Gut: first published as 10.1136/gut.15.9.737 on 1 September 1974. Downloaded from

742 A. M. Jezequel, M. Koch, and F. Orlandi

7 normal diazepam 0 * P(O.Ol Fig 4 Surface density of * endoplasmic reticulum in hepatocytes ofuntreatedpatients and patients treated with Benzodiazepin (diazepam). Each bar represents the mean values ofdata from . four subjects with the standard E deviation between subjects. 12 C. *

RE 1

CD.' 0

e ,U a I Sr rer golgi rer golgi http://gut.bmj.com/

The mean value for volumetric density of the Group 2a: patients with normal liver treated with total ER was 0.15. It was 0.04 for the RER, 0.092 Benzodiazepin (diazepam) for the SER (0.09 for type 1 and 0.002 for type 2), In the four patients of this group, ultrastructural and 0.007 for the Golgi apparatus (see fig 5). changes were observed at the level of the SER. In The evaluation of the mean width (d) of the ER patient Ta. P. the SER appeared globally increased cisternae gave the following figures: 1800 A for the in amount with an extensive development of SER on September 23, 2021 by guest. Protected copyright. RER cisternae, 1600 A for the SER type 1 whereas type 2 (fig 3). This sometimes gave to the hepatocyte it came to 120 A for the SER type 2 and 600 A for a'quilted' look. In Pe. M. and Br. P. the development the Golgi apparatus. This is close to the values of SER type 2 was less extensive and the hepatocytes found by direct measurement on the plates, which are took on a 'pepper and salt' appearance (Jezequel, 2000 A, 1500 A, 150 A, and 500 A respectively. Orlandi, and Tenconi, 1971) due to the alternation

Surface Density Normal Untreated Diazepam Br. P. Pe. M. Ta. P. Gi. M. Total ER 3-80 0.52 5.63 531 5 24 6-84 Total SER 2-91 ± 051 4-98 4.50 4.55 582 SER type 1 2-23 ± 0.49 3-29 3-11 2.56 5 06 SER type 2 0-68 0.35 1-69 1 39 1 99 0-76 RER 0-89 ± 022 0-65 0-81 0-69 1-02 Cholesterol synthesis 264 + 82.5 4257 822 1398 360 Table I Surface density ofendoplasmic reticulum and cholesterol synthesis in untreatedpatients and in patients receiving diazepaml 'In untreated patients, the mean values from four subjects ale reported with the standard deviation for the surface density; the mean value from 12 patients is given for cholesterol synthesis (I"C-acetate incoxporation into DPS). Individual values are given for the four treated patients. Gut: first published as 10.1136/gut.15.9.737 on 1 September 1974. Downloaded from

A morphometric study of the endoplasmic reticulum in human hepatocytes 743

normal diazepam Fig 5 Volume density of 0 endoplasmic reticulum in 0,20- hepatocytes ofuntreatedpatients 0 andpatients treated with Benzo- 0 diazepin (diazepam). Each bar 0. represents the mean values for 0.5 6S data from four subjects with the E standard deviation between *QF subjects.

.E0E15° ois 0 0@ u-- 0 I 1 1 a in aX * totbl tal se r total ser er ser type2 rer golgi ser type2 rer golgi

of light, vesicular areas and dense, non-vesicular Group 2b: patient treated with Dilantin areas of the smooth membranes (fig 2). In patient No ultrastructural change was observed in the Gi. M. the vesicles of the SER appeared densely hepatic tissue at the end of the seven-day period of packed but the overall architecture of the reticulum treatment. There was no significant difference was normal. between the quantitative data obtained before and http://gut.bmj.com/ The inean values for the pooled data for evalua- at the end of treatment (table II). tion of the surface density were the following as expressed in 2/H3 of cytoplasm: 5.75 ± 0.77 for the Group 3 total ER, 0.79 ± 0.17 for the RER, and 4.96 ± The ultrastructure of the liver in the patient with 0-61 for the total SER (3.51 for SER type 1 and 1.45 Gilbert's syndrome appeared morphologically nor- for SER type 2) (fig 4). It can be seen that the total mal before treatment. The surface density of the ER SER comes to represent 88% and the RER 12% of was 4.57 U2/43; for the RER it was 1.12, and 3.45 on September 23, 2021 by guest. Protected copyright. the total endoplasmic reticulum. Individual values for the SER (2.66 for type 1 and 0.79 for type 2) are reported in table I. The application of Student's (table II). The Golgi apparatus was not identified. t test shows a significant difference (p < 0.01) After four days of treatment with 360 mg per day of between the surface densities of the SER as found in phenobarbital there was no change either in the untreated and in treated patients. ultrastructure of hepatocytes or in the surface In this group of patients the membranes of the density of the RER and SER (table II). Golgi apparatus were not identified in any of the pictures analysed. The evaluation of volumetric density as measured CHOLESTEROL BIOSYNTHESIS on the pooled data gave the following figures: 0.17 The normal range of cholesterol biosynthesis in for the total ER, 0.03 for the RER, and 0.135 for human liver biopsies has been estimated by incorpo- the SER (0-13 for type 1 and 0005 for type 2) ration of 14C-acetate into Digitonin-precipitable (fig 5). sterols (Orlandi et al, 1974). The patients studied The mean width of the ER cisternae was thus under group 1 (normal untreated) were part of a found to be 1600 A for the RER, 1500 A for SER larger series of 12 normal subjects where the mean type 1, and 140 A for SER type 2. These values are value of incorporation was found to be 264 ± 82.5 also comparable to those obtained by direct measure- nmoles per g of tissue in two hours of incubation. ments on the plates, which are 1800 A, 1400 A, and The values obtained under treatment with diazepam 150 A respectively. were greatly increased in three patients as reported Gut: first published as 10.1136/gut.15.9.737 on 1 September 1974. Downloaded from

744 A. M. JezequeI, M. Koch, and F. Orlandi.

Dilantin Phenobarbital Untreated 50 mgld/7d Untreated 360 mg/d/4d Total ER 4-43 ± 0-82 3-64 + 0-65 4 57 ± 082 458 ± 049 Total SER 3.57 ± 1-13 2-81 ± 0 79 3145 10 3-44 ± 084 SER type 1 2-96 ± 1.01 1-64 ± 0.53 2-66 i 037 2-99 ± 072 SER type 2 0-61 ± 0.10 1-17 ± 0.53 0 79 ± 050 0.45 i 020 RER 0-86 ± 0.37 0-83 ± 0.50 1-12 i 020 1*14 t 0.75 Cholesterol synthesis 134 144 - _ Table II Surface density ofendoplasmic reticulum and cholesterol synthesis before and after treatment with phenobarbital or dilantin' 'The quantitative analysis of liver biopsy material was performed in each of these two patients before and at the end of treatment. The mean values obtained from analysis offive blocks of each biopsy with the standard deviation between blocks are given. in table I. In the patient treated with Dilantin, the calculated values of the width of cisternae and values values remained almost identical before and after obtained by direct measurement. treatment (table II). In the patient treated with The development of the SER surface in man is to phenobarbital the cholesterol biosynthesis was not be taken into consideration when dealing with patho- measured. logical conditions: an increase in the SER has been described in various situations or after treatment Discussion with certain drugs, and physiopathological considerations have been made without taking into Although limited in several ways, the present study account this fundamental feature. indicates that in the normal human hepatocyte, the On the other hand, it appears that in human hepa- surface density of the SER exceeds that of the RER. tocytes the SER does not have a uniform appearance, This was already suggested by subjective observa- due to the presence of a 'non-vesicular' component. tions. However, morphometric analyses give a better In some instances, it is difficult to draw aline between term of comparison with other animal and species. vesicular non-vesicularcomponents, especially in http://gut.bmj.com/ Weibel et al (1969) found a surface density of 8.1 the case of small, packed vesicles. For the present 2/g3 of cytoplasm for the RER and 60 pU2J,3 for study, counts were made on the pictures at a final the SER in the rat, whereas Loud (1968) found mean magnification of 91 000. In these pictures we con- values of 3.32 2/U3and2.20u2/4U3respectively.Inthe sidered as non-vesicular the areas of the reticulum dog, the values were 2.48 m2/cm3 and 3.73 m2/cm3 where the smooth membranes made an intricate net- (Hess et al, 1973). The discrepancies between the two work sometimes finely tubular, but where the sharp former studies have been explained by differences in lumina of vesicles were not recognizable (ie, a dis- counting procedures (Weibel et al, 1969). In any tance between membranes ofless than 0.02 ,u). In this on September 23, 2021 by guest. Protected copyright. event, it appears that in the rat the surface of the SER regard, it is important that the sections are kept at is less than that of the RER whereas the contrary is fairly uniform thickness to avoid images of super- true in dog as in man. The values we obtained for the imposition. In our series, the thickness was kept total SER surface density in man are intermediary be- around 600 A. The effect of section thickness has tween those in rats (Loud, 1968) and in dogs (Hess been previously discussed (Loud, 1968; Weibel et al, et al, 1973). The RER surface density appears much 1969). The membranous structures become difficult less than in any of the studies previously mentioned. to identify or are 'lost' when they are tilted at an The volumetric density of the SER andRERissimilar angle of more than 600 from the direction of the in rat (Weibel et al, 1969) and in man but the volume electron beam; this effect is least on thinnest sections to surface ratio and consequently the mean width of but it is thought that counts of intersections should the RER or SER comes to values higher than in rat be increased by about 50 % to correct for membranes as calculated by Weibel et al (1969). Moreover, present but not visible (Loud, 1968). Although such instead of plates as in rat hepatocytes (Weibel et al, an effect should particularly interfere with the counts 1969) we considered the RER cisternae as cylinders. of SER type 2 components, no correction factor has Evidence that this represents the true configuration been applied since the precise extent of error was ofRER in needle biopsies ofhuman liver comes from unknown. In any case. if this might affect the abso- the appearance of the section of the lumen of RER lute measurements, it should not change the terms cistemae-irregularly shaped but most often roughly of comparison. circular or oval-and by the agreement between The participation ofa fine network (or SER type 2) Gut: first published as 10.1136/gut.15.9.737 on 1 September 1974. Downloaded from

A morphometric study of the endoplasmic reticulum in human hepatocytes 745 in the composition of the SER in human liver has not membranes. Further studies would be needed to been previously recognized. The variable configura- metabolism (Marcucci et al, 1970). In rat, it has been tion of the SER is confirmed by the evaluation of claimed that diazepam reduces sleeping time and the ratio v/s and of d: this shows a difference of increases liver weight (Heubel, 1969) but no changes more than 1 to 10 between the SER type 1 and type 2, in liver weight, hepatic cholesterol synthesis, cyto- meaning that the SER type 2 is a system ofextremely chrome P-450 levels, or hepatic ultrastructure have fine tubules. followed the administration of diazepam (10 mg/kg In previous morphometric studies, the authors ip) for six days to rats or guinea-pigs (unpublished have counted the Golgi apparatus as part of the observations). In man, diazepam seems to affect the SER. This is amply justified since it is likely that bilirubin levels in newborn and interfere with pheno- many of the cisternae, especially when cut in a barbital metabolism (Heubel, 1969; Heubel and perpendicular plane, are undistinguishable from the Muhlberger, 1972). In addition, and contrary to SER. However, we preferred to count separately rats or guinea-pigs, the administration of diazepam the Golgi elements when these were clearly recog- to human subjects is followed by ultrastructural nizable as part of the system of horse-shoe-shaped changes in hepatocytes together with increased bags, often with dilated ends. This distinction was hepatic cholesterol synthesis, as shown in the present motivated by the observation of various pathological study. Then it appears likely that the parent drug conditions where the Golgi apparatus appears pro- or its metabolites are not handled in man as in minent, and the necessity of measuring this param- laboratory animals. eter. It can be seen that there are great variations In patients treated with diazepam, the pattern of aroundthemeanvaluewhich aredue to the configura- proliferation of smooth membranes is similar to tion of the Golgi apparatus itself. For this reason, that observed after treatment with rifampicin only when the figures are significantly higher than (Jezequel et al, 1971), a drug known as an inducer of in untreated patients can they be taken into account microsomal enzymes in human liver (Schoene, in pathological situations. Low or even zero values Fleischmann, and Remmer, 1972a). A direct esti- are due either to the sampling or-when there is an mation of drug-metabolizing enzymes in biopsies intense proliferation of SER membranes-to the of diazepam-treated patients would be of great difficulty of identifying the elements of the Golgi interest. apparatus. A relationship between the increase in cholesterol In rat liver, some structural differences have been synthesis and increase of microsomal enzymes has http://gut.bmj.com/ observed in the various lobular regions, the SER been demonstrated in various instances in vivo or being more abundant in the cells immediately in vitro after treatment with phenobarbital (Jones adjacent to a central vein (Loud, 1968). Others andArmstrong,1965 ;Orrenius,Dass,andGnosspelius, (Weibel et al, 1969; Hess et al, 1973) have made no 1969). The unchanged lipid composition of micro- attempt to localize the hepatocytes, and, as in the somal membranes in phenobarbital-treated rats present study, the values given represent 'the averages would be maintained by a decreased breakdown of

derived from a true random sample' (Weibel et al, phospholipids and increased rate of cholesterol on September 23, 2021 by guest. Protected copyright. 1969). synthesis (Eriksson and Dallner, 1973). The role of In patients receiving diazepam, there was a marked cytochrome P-450 in cholesterol synthesis is still a increase in the total smooth endoplasmic reticulum. matter for discussion (Atkin, Palmer, English, This increase is evident in fig 4 concerning the Morgan, Cawthorne, and Green, 1972). Increased pooled data. However, when looking at the indivi- availability of the microsomal enzyme hydroxy- dual figures (table I) it appears that in one patient methyl-glutaryl CoA reductase-a rate-limiting step (Gi. M.) the increase in the SER involved the vesi- in the hepatic cholesterol synthesis (Dietschy and cular component exclusively, the values of the non- Wilson, 1970)-due to a non-specific enzymatic vesicular (type 2) component being in the normal induction, might also be responsible for the increased range, in contrast to the other three patients. synthesis of cholesterol. That the increase of the The increase in SER membranes follows the surface of total SER membranes is not correlated administration of a number of substances known to with increased cholesterol synthesis in man is shown induce microsomal enzymes (Remmer and Merker, by the patient Gi. M. (see table I). However, in the 1965; Stenger, 1970), although this is not a specific same patient the values for the SER type 2 were in feature of enzymatic induction (Stenger, 1970; the lower normal range, in contrast to the other three Jezequel and Orlandi, 1972; Jezequel, 1974). patients. It is then possible that the increase of the Diazepam is a highly lipid-soluble drug, metabolized vesicular component results from a decreased catab- by microsomes (Marcucci, Fanelli, Mussini, and olism of 'adult' membranes without increased Garattini, 1969) with notable species differences in synthesis, whereas the SER type 2 would be the site Gut: first published as 10.1136/gut.15.9.737 on 1 September 1974. Downloaded from

746 A. M. J.ze'quel, M. Koch, and F. Orland of active synthesis and represent the newly formed cholesterol biogenesis andcatabolism. Biochem. J., 128, 237-242. Bamonti, F., and Dini, M. (1973). Hepatic biosynthesis ofcholesterol assess the validity of this hypothesis. in man. Normal range and changes by drugs inducing hepatic In all four patients, no correlation was found microsomal enzymes. Rendic. Gastroent., 5, 52-53. Bolender, R. P., and Weibel, E. R. (1973). A morphometric study of between the daily intake of diazepam or the length the removal ofphenobarbital-induced membranes from hepato- of treatment on the one hand and the surface in- cytes after cessation of treatment. J. Cell Biol., 56, 746-761. the SER or the changes in cholesterol Crigler, J. F., Jr., and Gold, N. I. (1969). Effect of sodium pheno- crease of barbital on bilirubin metabolism in an infant with congenital synthesis on the other hand. It must be noted that non hemolytic hyperbilirubinemia and kernicterus. J. clin. four days after withdrawal of the drug, the surface Invest., 48, 42-55. Dietschy, J. M., and Wilson, J. D. (1970). Regulation of cholesterol of SER components and the cholesterol synthesis metabolism. Part I. New Engl. J. Med., 282, 1128-1138. were still elevated in the patient Ta. P. This is in Dobbins, W. O., III, Rollins, E. L., Brooks, S. G., and Fallon, H. J. (1972). A quantitative morphological analysis of ethanol effect keeping with quantitative studies by Bolender and upon rat liver. Gastroenterology, 62, 1020-1033. Weibel (1973) who showed that ER membranes in Eriksson, L. C., and Dallner, G. (1973). Influence of phenobarbital rats persist at the induced treatment on the turnover of rat liver microsomal lipids. phenobarbital-treated FEBS Letters, 29, 351-354. level on the third day after cessation of treatment Gotze, H., Sidiropoulos, D., Hess, F. A., and Berthelot, P. (1972). and fall to control levels on the fifth day. Das Crigler-Najjar-Syndrom: klinische, biochemische, mor- phologische und therapeutische Aspekte. Helv. paedial. Acta, It is known that both phenobarbital and phenyl- 27, 335-351. hydantoin provoke an increase in smooth mem- Herdson, P. B., Garvin, P. J., and Jennings, R. B. (1964). Reversible Garvin, and biological and fine structural changes produced in rat liver by a branes in rat hepatocytes (Herdson, thiohydantoin compound. Lab. Invest., 13, 1014-1031. Jennings, 1964; Remmer and Merker, 1965) but Hess, F. A., Weibel, E. R., and Preisig, R. (1973). Morphometry of studies conducted on animals given high doses of a dog liver. Normal base-line data. Virchows Arch. Abt. B Zell- path., 12, 303-317. drug cannot be simply extrapolated to patients given Heubel, F. (1969). Interferenz von Diazepam und Pentobarbital an therapeutic doses. In our patient with Gilbert's der ratte und am Menschen. Naunyn-Schmiedebergs Arch. Pharmakol. exp. Path., 264, 246-247. syndrome, the administration of phenobarbital was Heubel, F., and Muhlberger, G. (1972). Senkung des Neugeborenen- followed by a fall in serum bilirubin but no apparent Bilirubinspiegels durch Diazepam. Int. J. Clin. Pharmacol., 6, change in the amount of smooth membranes. After 332-341. Hope, J. (1970). Stereological analysis of the ultrastructure of liver Dilantin, there was neither an increase in cholesterol parenchymal cells during pregnancy and lactation. J. Ultrastruct. synthesis nor an increase in smooth membranes. Res., 33, 292-305. Jez6quel, A. M. (1974). Ultrastructural basis of drug metabolism. (The apparent increase of the SER type 2 is not Israel J. med. Sci., 10, 380-385. significant.) It may come as unexpected that no Jez6quel, A. M., and Orlandi, F. (1972). Fine morphology of the liver administration of as a tool in clinical pharmacology. In Liver and Druigs, edited by morphological change follow the F. Orlandi and A. M. Jez6quel, pp. 145-192. Academic Press, both phenobarbital and Dilantin, known to produce London and New York. http://gut.bmj.com/ remarkable changes in SER in animals. The absence Jlz6quel, A. M., Orlandi, F., and Tenconi, L. T. (1971). Changes ofthe smooth endoplasmic reticulum induced by rifampicin in of changes in human hepatocytes after short-term human and gtuinea-pig hepatocytes. Gut, 12, 984-987. administration ofphenobarbital in patients with nor- Jones, A. L., and Armstrong, D. T. (1965). Increased cholesterol biosynthesis following phenobarbital-induced hypertrophy of mal liver resulted from previous observations made agranular endoplasmic reticulum in liver. Proc. Soc. exp. Biol., on a non-quantitative basis (Jezequel and Orlandi, (N.Y.) 119, 1136-1139. and also from morphometric Klion, F. M., and Schaffner, F. (1968). Ultrastructure of human liver 1972; Jezequel, 1974) In ofLiver Diseases, an infant with the syndrome in health and disease. Laboratory Diagnosis studies in Crigler-Najjar edited by F. W. Sunderman and F. W. Sunderman, Jr., on September 23, 2021 by guest. Protected copyright. (Gotze, Sidiropoulos, Hess, and Berthelot, 1972). pp. 469-488. Hilger, London. Loud, A. V. (1968). A quantitative stereological description of the This appears in contradiction to other reports ultrastructure of normal rat liver parenchymal cells. J. Cell. (Whelton, Krustev, and Billing, 1968; Crigler and Biol., 37, 27-46. and Ma, H. M., and Biempica, L. (1971). The normal human liver cell: Gold, 1969). The fact that the normal amount cytochemical and ultrastructural studies. Amer. J. Path., 62, distribution of the SER before treatment have not 353-390. been taken into consideration may explain Marcucci, F., Fanelli, R., Mussini, E., and Garattini, S. (1969). The always metabolism of Diazepam by liver microsomal enzymes of rats apparent discrepancies. In addition, further studies and mice. Europ. J. Pharmacol., 7, 307-313. are needed to assess the effect of drug dosage, length Marcucci, F., Fanelli, R., Mussini, E., and Garattini, S. (1970). Further studies on species differences in Diazepam metabolism. of treatment, and individual variations in the liver Europ. J. Pharmacol., 9, 253-256. response to these therapeutic agents. Orlandi, F., Bamonti, F., Dini, M., Koch, M., and J&z6quel, A. M. (1974). Hepatic cholesterol synthesis in man: Effect of Diaze- Thanks are due to Dr E. R. Weibel for his critical pam and other drugs. Europ. J. clin. Invest., in press Orrenius, S., Dass, M., and Gnosspelius, Y. (1969). Overall bio- review of the manuscript. chemical effects of drug induction on liver microsomes. In The technical assistance of Miss M. L. Chiodi is Microsomes and Drug Oxidations, edited by J. R. Gillette et al, pp. 251-277. Academic Press, New York and London. gratefully acknowledged. Remmer, H., and Merker, M. J. (1965). Effect of drugs on the forma- References tion of smooth endoplasmic reticulum and drug-metabolizing enzymes. Ann. N. Y. Acad. Sci., 123, 79-96. Atkin, S. D., Palmer, E. D., English, P. D., Morgan, B., Cawthorne, Riede, U. N., Seebass, C., and Rohr, H. P. (1971a). Ultrastrukturell- M. A., and Green, J. (1972). The role of cytochrome P.450 in morphometrische Untersuchungen an der Leberparenchym- Gut: first published as 10.1136/gut.15.9.737 on 1 September 1974. Downloaded from

A morphometric study of the endoplasmic reticulm in human hepatocytes 747

zelle der Ratte nach Hemmung der proteinsynthese durch Stenger, R. J. (1970). Organelle pathology of the liver: the endoplasmic Cycloheximid. Virchows Arch. Abt. B. Zellpath., 9, 16-27. reticulum. Gastroenterology, 58, 554-574. Riede, U. N., Strassle, H., Bianchi, L., and Rohr, H. P. (1971b). Weibel, E. R., Kistler, G. S., and Scherle, W. F. (1966). Practical Ultrastructural-morphometric analysis of rat liver cell after stereological methods for morphometric cytology. J. Cell Biol., orotic acid administration. Exp. molec. Path., 15, 271-276. 30, 23-38. Schoene, B., Fleischmann, R. A., and Remmer, H. (1972a). The Weibel, E. R., Staubli, W., Gnagi, H. R., and Hess, F. A. (1969). cytochrome P-450 and the activities of various microsomal Correlated morphometric and biochemical studies on the liver enzymes in human liver. Proc. europ. Soc. Drug. Toxic, 13, cell. I. Morphometric model, stereologic methods and normal 249-253. morphometric data for rat liver. J. Cell. Blol., 42, 68-91. Schoene, B., Fleischmann, R. A., Remmer, H., and Von Oldershausen, Whelton, M. J., Krustev, L. P., and Billing, B. H. (1968). Reduction H. F. (1972b). Determination of drug-metabolizing enzymes in in serum bilirubin by phenobarbital in adult unconjugated needle biopsies of human liver. Europ. J. Clin. Pharmacol.,4, hyperbilirubinemia. Amer. J. Med., 45, 160-164. 65-73. Wiener, J., Loud, A. V., Kimberg, D. V., and Spiro, D. (1968). A Staubli, W., Hess, R., and Weibel, E. R. (1969). Correlated morpho- quantitative description of cortisone-induced alterations in the metric and biochemical studies on the liver cell. II. Effects of ultrastructure of rat liver parenchymal cells. J. Cell Biol., 37, phenobarbital on rat hepatocytes. J. Cell Biol., 42, 92-112. 47-61.

The August 1974 Issue THE AUGUST 1974 ISSUE CONTAINS THE FOLLOWING PAPERS The effect of stimulation on the myoelectrical Occurrence of cholesterol crystals in human bile activity of the rectosigmoid in man I. TAYLOR, WILLEM VAN DER LINDEN AND FUMIO NAKAYAMA H. L. DUTHIE, R. SMALLWOOD, B. H. BROWN, AND D. LINKENS Alkaline phosphatase in auxiliary liver transplanta- tion J. JERSKY A comparison of stool fluid and stool dialysate obtained in vivo MICHAEL J. TARLOW AND HAZEL Clinical and laboratory study of postvagotomy BUCHAN, AND http://gut.bmj.com/ THOM diarrhoea G. G. BROWNING, K. A. C. MACKAY Incidence and significance of faecal hydroxystearic acid in alimentary disease H. S. WIGGINS, JOY R. Histopathological analysis of benign polyps in PEARSON, J. G. WALKER, R. I. RUSSELL, AND T. D. patients with carcinoma of the colon and rectum KELLOCK G. EKELUND AND C. LINDSTROM The Progress Report Mitochondrial antibodies (AMA) effect of calcium on gastric acid and gastrin on September 23, 2021 by guest. Protected copyright. secretion in antrectomized subjects J. CHRISTIANSEN, DEBORAH DONIACH AND GEOFFREY J. WALKER J. F. REHFELD, AND F. STADIL Progress report Electrical activity of gastrointestinal A duodenal role in gastrin release J. R. HAYES, smooth muscle H. L. DUTHIE JOY ARDILL, T. L. KENNEDY, AND K. D. BUCHANAN Notes and activities

Copies are still available and may be obtained from the PUBLISHING MANAGER, BRITISH MEDICAL ASSOCIATION, TAVISTOCK SQUARE, LONDON, WC1H 9jR, price, 87ip