The amazing universe of hepatic microstructure


  • Valeer J. Desmet

    Corresponding author
    1. Liver Research Unit, Department of Morphology and Molecular Pathology, University of Leuven, Leuven, Belgium
    • Department of Pathology, University Hospital St Rafael, Minderbroedersstraat 12, B-3000 Leuven, Belgium
    Search for more papers by this author
    • fax: 32-16-33 65 48.

  • Potential conflict of interest: Nothing to report.


An informal review is presented by the author of his 50 years of involvement in practice and research in hepatopathology. Some background for the author's attitude and meandering pathway into his professional career serves as introduction to a short discussion of the main topics of his interest and expertise. Histogenesis of liver cancer was the theme of early work for a Ph.D. thesis, the results of which were lost into oblivion due to local rules and circumstances, but were rescued three decades later. His conclusions about the cells of origin of liver cancer remain concordant with the newer concepts in the field after nearly half a century. Studies in the field of chronic hepatitis became a long saga, involving the first classification of this syndrome by “the Gnomes” in 1968, histochemical investigations of viral antigens, lymphocyte subsets and adhesion molecules, and a quarter century later, the creation of a new classification presently in use. Cholestasis was a broadening field in diagnostic entities and involved the study of liver lesions, comprising pathways of bile regurgitation (including reversed secretory polarity of hepatocytes) and so-called ductular reaction. The latter topic has a high importance for the various roles it plays in modulating liver tissue of chronic cholestasis into biliary cirrhosis, and as the territory of hepatic progenitor cells, crucial for liver regeneration in adverse conditions and in development of liver cancer. Study of the embryology of intrahepatic bile ducts helped to clarify the strange appearance of the ducts in “ductal plate configuration” in several conditions, including some forms of biliary atresia with poor prognosis and all varieties of fibrocystic bile duct diseases with “ductal plate malformation” as the basic morphologic lesion. (HEPATOLOGY 2009;50:333–344.)

I feel honored for being asked to write a manuscript for HEPATOLOGY's series “Master's Perspective”. I am also surprised, because I still feel more like a student than a master, and a bit uneasy, because I never was the scientist who spent his life wrestling with a single problem, but rather was an amazed and naïve traveller in the wonderland of hepatopathology, trying hard to find my way, and, on occasion, searching a new road.

This is my story.

My birthplace is Passendale, a small village famous for having been the battleground in 1917 for one of the most devastating episodes in World War I (the battle of Passchendaele), which inspired the poem of Captain John McCrae, the young Canadian doctor who wrote “In Flanders Fields”. This poem was distilled from his feelings during battles in my region, which now is the location of the largest British war cemetery on the European continent (Fig. 1). Growing up in the shadow of the tombs of thousands of destroyed young lives cultivated in me a deep aversion to war, dispute, and hunt for power, and stimulated my respect for justice and agreement.

Figure 1.

Tyne Cot Cemetery, Passendale (Zonnebeke), Belgium. Out of several photographs, I selected this picture showing rows of tombs under an awesome sky, pregnant with looming, overhanging, somber clouds, that echo the still persisting threats of war and terror, notwithstanding the sacrifice of the thousands lying in these endless rows who gave their lives for freedom and for lasting peace. Courtesy Mike Sheil,

I went to medical school in Leuven, ending the first year with the highest degree, a rarity in those years. In the first week of the following year, I was fished up by Prof. Jozue Vandenbroucke who chased young talents for research. He sent me as “student-researcher” to Dr. Marc Verstraete, who led a recently started lab for blood coagulation studies. In spite of primitive conditions in terms of housing and equipment, I spent six marvelous years in this group of volunteers, enjoying the delight of finding something new and the flaming spirit of a team of enthusiastic youngsters with a charismatic leader. I graduated in 1957, a period when new winds had started blowing in histopathology. It was Vandenbroucke, by then chairman of Internal Medicine and apparently planning a “liver unit”, who finally influenced my shift into histopathology and lit the fire of interest in the liver, by providing me with a newly published book, “Liver: Structure and Function”. This text was written by two major gods of the medical Olympus from far-away America and who were totally unknown to me: Hans Popper and Fenton Schaffner.1 This was my first indirect contact with Hans Popper.2 Vandenbroucke sent me to Paris to study liver pathology with Louis Orcel at Clichy-based Hôpital Beaujon.

Two months were enough to be caught by the fascination for histopathology “in toto”, and not only the hepatic aspect; with my mentor's permission, I continued my training, while still keeping a special interest in hepatic diseases. I spent an additional year at the Royal Postgraduate Medical School (Hammersmith Hospital) in London with Charles Harrison, where further training in pathology was helpful to survive the probing questions fired by Sheila Sherlock at postmortem demonstrations, and meeting Tony Pearse initiated interest in histochemistry. During those years, I carried the “Popper book” with me like a bible. The conviction grew that Popper's approach to liver pathology was the right and only way to practice histopathology: a continuous search for structural-functional relationships in health and disease. This concept was enforced by the marvellous results appearing at that time in the burgeoning fields of biochemistry, histochemistry, cell fractionation, and electron microscopy. The book did have some influence, since, on my return to Leuven in late 1959, I decided to start scientific work in no other area than experimental liver carcinogenesis.2 I used some enzyme-histochemical methods available at the time, hoping that the new techniques would help reveal some secrets of hepatic transformation into malignancy.

This topic is the first to be discussed. Having been a sort of scientific butterfly seduced by flowers all along my way and visiting a few of them, I present below a small bouquet gathered from along the road: hepatocarcinogenesis, chronic hepatitis, cholestasis, ductal plates, and metabolic disease. Most of these were challenges I met in diagnostic work, with the recurring question “what could be the reasons that this lesion looks like this?”

I. Histogenesis of Liver Cancer

I induced malignant liver tumors in rats using the carcinogen 3′-methyl-4-dimethyl-aminoazobenzene in a set-up which guarantees the rapid development of a wide variety of tumors: young male animals with a high-dose carcinogen in a low-protein diet.3, 4 The main aim was to identify the cell(s) of origin of the tumors. Some researchers had maintained that all tumors derived from hepatocytes,5 others thought that hepatocellular carcinomas (HCCs) originate from hepatocytes and cholangiocarcinomas (CC) from cholangiocytes,6 others let all tumor types derive from so-called cholangiofibrosis,7 and an occasional article suggested that mixed tumors might possibly derive from a more primitive cell type in the liver.8

My study yielded a variety of tumor types that merged imperceptibly from one into the other. Based on the morphology, histochemistry, and topography of successively developing lesions, and supported by a study by Elisabeth LeDuc outside the cancer field,9 my conclusion was that the cells of origin of liver cancer are: hepatocytes for well-differentiated HCCs, and “oval cells” for less differentiated HCCs, mixed HCC-CC, well-differentiated and poorly differentiated CCs, and some rare undifferentiated carcinomas (now termed progenitor cell tumors; Fig. 2). “Oval cells” are cell types activated to proliferate in early stages of intoxication with carcinogens.10 The nature of these cells had been a subject of debate: whether they are fibroblasts, endothelial cells, histiocytes, transformed hepatocytes, or biliary ductules. My study brought support for the latter interpretation, as also demonstrated in the meantime by Joe Grisham with studies using electron microscopy.11 In later years, they became known as hepatic epithelial progenitor cells. It is an irony of fate that nowadays a mesenchymal cell is a candidate progenitor again!12

Figure 2.

Schematic representation of the histogenesis of experimental liver tumors proposed in the thesis of V. Desmet in 1963.3 Hepatocytes are at the origin, through stages of precancerous lesions, of well-differentiated and moderately differentiated HCCs. Progenitor cells, again through precancerous lesions, are at the origin of less differentiated variants of HCC, of totally undifferentiated cancers, and of cholangiocarcinomas of different cell caliber and degree of differentiation. Old terminology of 1963: undifferentiated or oval cells equals progenitor cells; hepatoma equals HCC; adenocarcinoma equals bile duct carcinoma; anaplastic carcinoma equals progenitor cell tumor. This is the English version of Figure 94 in original thesis, and Figure 1 in Fausto.13

The thesis defense was followed by nomination as docent in histology and pathology and head of the local pathology department, resulting in a heavy overload of teaching and clinical routine with literally no time at all for writing or research during the first few years. My work on liver tumors had received a solemn funeral followed by oblivion! Thirty-one years later, Nelson Fausto heard about my hidden work. He saved it from obscurity by borrowing the concluding scheme of my forgottten thesis for his chapter “liver stem cells” in the third edition of “The Liver: Biology and Pathobiology”, edited by Irwin Arias et al.13 (Fig. 2). Alas, investigators in modern times are not reading chapters of a book, but only recent articles in the journals, or only abstracts of the same.

Occasional coworkers again took up the topic of liver carcinogenesis. Mbowa Kalengayi studied liver cancer induced by aflatoxin B1; feeding of a minimal dose finally resulted in well-differentiated HCCs with very little oval cell proliferation in the early phase, whereas larger doses yielded faster development of a variety of less differentiated cancers after marked oval cell proliferation.14, 15 This result remained consistent with my earlier findings. Later work by Hiroshi Miyazaki and Tania Roskams on the choline-deficient acetylaminofluorene (CDAAF) model in the rat, compared with human disease, convincingly confirmed that OV6-positive progenitor cells can differentiate into hepatocytes and bile duct cells16 (Fig. 3). Marie-José Vanstapel discovered that cholangiocytes are marked by immunostaining for S-100 protein17 and for keratins using polyclonal antikeratin antiserum.18 Peter Van Eyken subsequently demonstrated by use of monoclonal antibodies against individual keratins that hepatocytes express keratin 8 (K8) and K18, whereas cholangiocytes express K7 and K19,19 in addition, thus providing the basis for cell lineage studies in embryonic development of intrahepatic bile ducts20, 21 and in HCCs.22, 23 During embryonic development, K19 is not only a marker of cholangiocytes, but of primitive hepatoblasts as well,24 and the liver cancer studies revealed that many tumors which were morphologically diagnosed as HCCs express bile duct–type K7 and/or K19, indicating a nonhepatocellular trait in HCCs.22, 24 This finding became of higher interest in later years. Inspired by microscopic slides revealing strongly K7-positive small cells in periportal regions, my coworker, electron microscopist Rita DeVos, searched for such cells at the ultrastructural level, and thus discovered a novel cell type hitherto unmentioned in the human liver, corresponding to liver progenitor cells differentiating into both hepatocytes and bile duct cells25 (Fig. 4). This manuscript was originally rejected; it emphasized the similarity to oval cells too much—a politically incorrect notion!—because oval cells were “rat pathology”. For sure, “All great truths begin as blasphemies.” (Bernard Shaw).

Figure 3.

Choline-deficient diet–acetyl aminofluorene (CDAAF) rat model of livercarcinogenesis, 8th day after cessation of 3 weeks treatment. OV6 immunostain on frozen section. Bile duct and ductule are intensely positive. Intermediate hepatocytes are characterized by weaker, mainly submembranous OV6 expression, and extended deeper into the lobule with time (Roskams et al.16). BD, bile duct; HA, hepatic artery; PT, portal tract; PV, portal vein. Courtesy Prof. Tania Roskams, Leuven, Belgium.

Figure 4.

Electron micrograph of liver parenchyma in chronic viral hepatitis C. A small progenitor cell surrounded by a basement membrane (large arrow) is situated next to three hepatocytes (H). The progenitor cell forms junctions (arrowhead) with the neighboring hepatocytes showing a bile canaliculus (BC). Note a bundle of cytokeratin filaments (small arrow) in the cytoplasm of the progenitor cell. Courtesy Prof. Rita DeVos, Leuven, Belgium.

Studies of my one-time coworker and now successor in hepatopathology Tania Roskams have clearly brought to center stage the oval cell equivalents of hepatic progenitor cells in human liver.26, 27 Further work by Roskams's team revealed that human HCCs, diagnosed according to classical criteria,28 express the biliary-type K7 and K19 in a considerable proportion of cases, suggesting a progenitor cell origin, and with K19 expression independently correlating with worse prognosis.29 Confirming studies led to international consensus for a new classification of primary liver cancers, with K19+ “mixed hepatobiliary carcinoma” representing a separate group besides HCC and CC.30 This proposal turns out to validate in humans my main conclusions from experimental studies half a century ago.3No great discovery was ever made without a bold guess.”(Isaac Newton).


ADPKD, autosomal dominant polycystic kidney disease; ALP, alkaline phosphatase; ARPKD, autosomal recessive polycystic kidney disease; ATPase, adenosine triphosphatase; BA, biliary atresia; CC, cholangiocarcinoma; CDAAF model, choline-deficient acetylaminofluorene model; CHF, congenital hepatic fibrosis; DPM, ductal plate malformation; EASL, European Association for the Study of the Liver; γ-GT, gamma-glutamyl transpeptidase; HAI, histology activity index; HBsAg, hepatitis B surface antigen; HCC, hepatocellular carcinoma; IASL, International Association for the Study of the Liver; K8, keratin 8; PBC, primary biliary cirrhosis; PiMZ, protease inhibitor type MZ; PLD, polycystic liver disease (without renal involvement); PMN, piecemeal necrosis; TfR, tranferrin receptor.

II. Chronic Hepatitis

In 1964, I collaborated with my clinical colleague Jan DeGroote, who was interested in noncirrhotic chronic hepatitis. We tried to subdivide the cases according to degree of inflammation, regarding those with “piecemeal necrosis” (PMN)31 (now termed interface hepatitis) as more severe, and hence more indicated for immunosuppressive treatment, because autoimmune hepatitis (lupoid hepatitis) was the only etiology known at the time.32, 33 This classification agreed with the “Habilitationsschrift” of Martin Schmid from Zürich.34 At the second meeting of the European Association for the Study of the Liver (EASL) in Göteborg, Sweden, in 1967, a slide seminar led by Peter Gedigk and Gerhard Korb from Marburg, Germany, gave rise to a babel of more than vivid discussions, not surprisingly when one considers that “chronic hepatitis” had been described in the literature by at least 47 terminologies.35 At DeGroote's proposal that a group of pathologists and clinicians should sit together and discuss chronic hepatitis and its terminology, a dozen colleagues enthusiastically agreed, and during dinner decided to skip the rest of the EASL meeting so they could start the job. Two full days' discussion ended with agreement to meet again next year, and to circulate a set of slides of so-called chronic hepatitis to all participants, who would send their diagnosis—in provisionally agreed terms—to Martin Schmid as host of next year's meeting. In Zürich, we concocted what became the standardized classification for the next 30 years36 and went straight to the World Congress of Gastroenterology in Prague, Czechoslovakia, to present the new classification, and to the subsequent meeting of the International Association for the Study of the Liver (IASL) in Karlovy Vary. This was where Sheila Sherlock, then president of IASL, baptized our group “the Gnomes of Zürich”, whom she claimed to be as influential in hepatological semantics as the real Gnomes of Zürich (the bankers) are in the world's financial affairs. The nickname would become, in later years, a label of prestige.2

Already at that time, evidence was growing that a newly discovered antigen,37 named “Australia antigen”, represented a hepatitis virus.38 In the early 1970s, my coworker Mukunda Ray studied liver biopsies by means of immunofluorescence for Australia antigen, then already known as hepatitis B surface antigen (HBsAg). At the IASL meeting in Versailles, France, in 1972, I happened to discuss with Gerald Klatskin a peculiar aspect of hepatocytes observed in some cases of chronic hepatitis, when shortly afterward Dr. Ray discovered HBsAg in those cells. Triumphantly, I was writing Dr. Klatskin that we found the meaning of those cells, when an article appeared describing them as “ground-glass hepatocytes”.39 That's how I learned to live with the disappointment of missing being the first to announce a discovery, consoling myself with the idea that at least I had not missed the finding.40, 41 In order to locate the viral antigens as targets of immune response, we were interested in demonstrating them in the membrane of hepatocytes, and later also at the ultrastructural level, as studied by Takashi Kojima42 who also worked on hepatitis Delta antigen.43, 44

For a while, there has been discrepancy between the writings of the Gnomes and those from Klatskin and his group from Yale University, New Haven, CT. The Gnomes considered PMN as a marker and driver of progression in chronic hepatitis, whereas the Yale group was emphasizing “bridging necrosis” as the pathway to cirrhosis.45–47 In later years, James Boyer bestowed me with the honor of inaugurating the Annual Gerald Klatskin Memorial Lectureship in Hepatology at Yale, a welcome opportunity to emphasize that both these lesions play their role in chronic viral hepatitis B: PMN as a marker of progressive variants of disease and bridging necrosis as a driver for faster evolution.48

Our interest in mechanisms of chronic liver inflammation stimulated year-long searches for actors in hepatic immunological reactions: antigen-presenting cells,49, 50 lymphocyte subsets,51, 52 human leukocyte antigen class I and II,53, 54 and intercellular adhesion molecule-1 (ICAM-1) expression55 in inflammatory parenchymal as well as bile duct diseases.56, 57 These investigations by Riccardo Volpes and Joost Van den Oord resulted in the concept of active immunomodulatory involvement of parenchymal and bile duct lining cells in intrahepatic immunological battles of cells and cytokines.58–60 This was only the beginning. The battleground of inflammation has been continuously completed by additional cytokines and chemokines with their receptors, regulatory lymphocytes, and other parties (e.g., see Guidotti and Chisari61).

During the 1980s, we participated in the search for markers of non-A/non-B hepatitis (subsequently hepatitis C). In the absence of a reliable serological test, we used electron microscopy, looking for ultrastructural changes specific for the elusive virus.62 So-called “nuclear particles” had been considered a serious candidate, but screening of dozens of biopsies and controls let us conclude that these nuclear changes had no relation to a virus.63, 64 A plausible candidate was found several years later.65

With the discoveries of other members of the hepatitis alphabet (A, C, D, E, G), of drug-induced hepatitis, and continuing studies on autoimmune liver disease, the need was growing for a new nomenclature that reflected etiology, severity, and stage of the disease. In 1993, IASL president Juan Rodés asked me to convene with a very limited budget a small group for discussing a new classification of chronic hepatitis to be presented at the society's next meeting in Cancun. I wrote a working text and sent it for discussion to two pathologists (Michael Gerber, Peter Scheuer) and two clinicians (Jay Hoofnagle, Michael Manns); I succeeded in getting them to Leuven in October 1993: a small committee for a small meeting, in a small room, in a small hotel, in a small city (Fig. 5). Fortunately, my partners agreed with the basics of the submitted discussion text. The new classification was presented in Cancun and was published 2 weeks later.66 At the 25th anniversary of HEPATOLOGY in 2006, it was the journal's most cited article.67

Figure 5.

IASL committee for classification of chronic hepatitis, October 1993, Congreshotel, Begijnhof, Leuven, Belgium. From left to right: Jay Hoofnagle, the late Michael Gerber, Michael Manns, the late Peter Scheuer, Valeer Desmet.

In 1981, Knodell et al. published a “Histology Activity Index” (HAI), which generated a numerical score for liver biopsies from patients with asymptomatic chronic active hepatitis, and provided endpoints for statistical analysis of serial changes in liver histology.68 A revised scoring system, inspired by Knodell's HAI but which tried to avoid its deficiencies, was published by the Gnomes in 1995.69 With the prevailing emphasis on evidence-based medicine,70 several proposals were published in the meantime, among them the now most popular Metavir system proposed by 10 French hepatopathologists.71 Like all scoring systems, the Metavir staging was built for semiquantitating progression rather than regression of fibrosis. Furthermore, in any system, it is notoriously difficult to write professor-proof and error-proof definitions of the lesions to be scored. Regrettably, this caused occasionally ill-founded clinical conclusions about reversal of cirrhosis,72 ignoring its macronodular and incomplete septal forms,73 and hence eliciting my firm protest,74, 75 although I do agree that in selected circumstances cirrhosis may reverse.76 This keeps in mind the old adage: “Παντα ρϵι, oυδϵν μϵνϵι” (“Panta rhei, ouden menei”) (Everything is in flow, nothing remains the same.) (Heraclitus of Ephesus, ca. 535-475 B.C.).

III. Cholestasis

Cholestasis entails intriguing liver changes, but two alterations became a focus of more interest: reversed polarity of hepatocytes and ductular reaction.

Reversed Secretory Polarity of Parenchymal Cells in Cholestasis.

Although Albert Coons had started his early studies on fluorescently labeled antibodies in the 1940s,77 the blooming of immunohistochemistry in diagnostic histopathology and research started mainly after reliable reagents became commercially available in the mid-1970s. In the late 1960s and early 1970s, we studied cholestatic liver tissue using enzyme histochemical techniques, including those for gamma-glutamyl transpeptidase (γ-GT), adenosine triphosphatase (ATPase), and nonspecific alkaline phosphatase (ALP).

Obstructive cholestasis in the rat,78, 79 in the adult lamprey after metamorphosis,80 and in humans81 revealed opposite changes in canalicular and sinusoidal membranes of hepatocytes: canalicular disappearance of ATPase was associated with its appearance at the sinusoidal pole, and γ-GT and ALP, normally expressed in canaliculi and faintly on sinusoidal membranes, revealed increased membranous expression at both sides of the cell, suggesting that in these conditions an active secretory activity normally taking place at the canaliculus was shifted—at least in part—toward the opposite side, implying bile secretory mechanisms at the sinusoidal membrane. Ultrastructural investigations revealed disappearance of microvilli in canalicular membranes and their increase in sinusoidal ones, compatible with the histochemical changes. Alternative basolateral secretion of bile constituents seemed insufficient for avoiding overload of bile constituents, because a cholestatic picture still develops with lysosomes containing cell debris and bilirubin pigment.82 Renal tubular cells in rats after bile duct ligation show changes similar to those in hepatocytes.83

Regurgitation of bile via vesicle-mediated transport in the sinusoidal direction was demonstrated in cholestatic hepatocytes in several laboratories. By use of the freeze-cleave replica technique, structural changes suggesting increased permeability were detected in hepatocellular tight junctions (blood–bile barriers).84 The heterogeneous appearance of canaliculi in cholestatic liver tissue was explained, at least in part, by new development of canaliculi attempting to rescue bile secretory function85 along a mechanism analogous to canalicular development studied by Chris Peeters in fetal liver.86–88 Altogether, the findings allowed one to summarize that in prolonged cholestasis new canaliculi develop, and bile regurgitates through several pathways: directly from canaliculi through leaky tight junctions, from inside the cell by reversed vesicle-mediated transcytosis, and by reversed transmembrane secretion at the sinusoidal side89, 90 (Fig. 6). The latter prediction of “reversed secretory polarity” became remarkably elucidated by the astonishingly complex, but masterfully regulated, batteries of ATP-binding cassette transporters in sinusoidal and canalicular membranes of hepatocytes, in apical membranes of ileal enterocytes, and in renal tubular cells.91, 92 Truly, “Nothing is as simple as it seems.” (Albert Einstein).

Figure 6.

Scheme depicting three pathways of bile regurgitation: (1) Intercellular escape through leaky tight junctions (red arrows); (2) reversed vesicle-mediated transcytosis (green arrows); and (3) reversed secretory polarity of the hepatocytes (sinusoidal cell membrane activity: blue arrows). Modified after Desmet.89

Ductular Reaction.

Three types of ductular reaction are described in liver histopathology.93 Type I corresponds to elongation of pre-existing ductules by proliferation of their lining cells and is mainly seen in acute bile duct obstruction. Type II refers to ductular metaplasia of liver cell plates, predominantly observed in chronic cholestatic conditions like primary biliary cirrhosis (PBC) (Fig. 7). Type III consists of activation and proliferation of hepatic progenitor cells, appearing as periportal ductular structures in the case of (sub)massive hepatocellular necrosis, and representing an alternative (attempt at) parenchymal regeneration when hepatocellular regeneratory capacity is insufficient, which is primarily the case in chronic liver diseases.94

Figure 7.

Keratin 7 immunostain in ductular reaction type 2 in sclerosing cholangitis, illustrating ductular metaplasia of hepatocytes. Ductules are intensely positive. Hepatocytes start expressing K7 in their submembranous area, and subsequently throughout their cytoplasm, while the cell volume is shrinking and acquiring a cholangiocytic phenotype.

Besides providing a means for early detection of periportal cholatestasis, the studies of Peter Van Eyken using K7 and K19 immunostaining elegantly contributed to interpreting type II ductular reaction as ductular metaplasia of hepatocytes by revealing a gradual transition from negative hepatocytes into positive ductular cells, starting with expression of K7 and later K19, a sequence opposite to what is seen during embryonic development of intrahepatic bile ducts (Fig. 7).95, 96 The early, transient expression of the extracellular matrix component tenascin showed that ductular reaction represents a pacemaker for progressive fibrosis in chronic cholestasis, analogous to the fibrosis-inducing role of interface hepatitis in chronic inflammation.97, 98 The study by Dr. Vanstapel and colleagues on S100 protein in ductules,99 and the results by Dr. Roskams and colleagues on ductular expression of neuroendocrine markers100, 101 (Fig. 8) focused more on type III ductular reaction, emphasizing the existence in human liver of progenitor cells100–102 with a multidrug-resistant phenotype, which is of high importance in regeneration and carcinogenesis.103, 104 Further investigations by Dr. Roskams's team revealed expression of neuroendocrine markers in coproliferating hepatic stellate cells as well, leading to the concept of a neuroendocrine compartment in the liver and its neuroregulation.105

Figure 8.

Positive staining for the neuroendocrine marker chromogranin A in ductules (vertical green arrows) and intermediate hepatocytes (horizontal yellow arrows). Alcoholic cirrhosis, frozen section, chromogranin A immunostain. Courtesy Prof. Tania Roskams, Leuven, Belgium.

With the concept of hepatic progenitor cells in ductules and the canals of Hering106—fundamental in basic physiology and in pathology—my early fascination about “ductular reaction” (as termed by Hans Popper in 1957107) has been fully justified. The proposal by Gershom Zajicek of a “streaming liver”, with a progenitor cell–fed lineage of hepatocytes migrating during their lifespan from periportal to centrolobular zones in normal liver,108 was criticized by later studies using genetic tagging of hepatocytes with β-galactosidase.109

However, my critical analysis of articles on both sides of the argument concluded that “streaming” of the normal liver is not finally excluded,110 thus joining Galileo Galilei's “E pur si muove!” (And yet it moves!).

IV. Ductal Plate Fantasies

During a 20-year period, I had some problems diagnosing biliary atresia (BA). My problems were not merely the same as those of everyone in this admittedly difficult diagnosis, but that I simply did not understand some published illustrations, described as bile duct proliferation (e.g., figure 6 in Stowens111). A presentation by the Danish pathologist Mogens Jørgensen at the 1971 meeting of IASL near Hamlet's Kronborg Castle in Helsingør, Denmark, and later published in his thesis “The Ductal Plate Malformation”112 brought relief. This was the reason I suggested that Peter Van Eyken in the 1980s should study the embryonic development of intrahepatic bile ducts with our recently discovered histochemical marker of “biliary”-type keratins.19 I recognized ductal plate malformation (DPM), which describes lack of remodeling or wrong remodeling of embryonic ductal plates, as the puzzling lesion in some cases of BA. I am grateful to Francesco Callea (Rome, Italy) for giving me the opportunity to study a fascinating series of liver specimens from his department (then in Brescia), derived from porto-enterostomies performed by Dr. Guido Caccia. This material provided the basis for my conviction of the existence of a subgroup of patients with BA characterized by DPM and worse prognosis (“early severe BA”). I published this as an invited chapter in a textbook.113 Since then, DPM in BA has been confirmed114, 115 and shown to be a sign of faster evolution and of worse prognosis.116

Extensive study of the literature convinced me that DPM was the basic morphological lesion in all congenital diseases of intrahepatic bile ducts, including a significant minority of cases of BA. After presentation of these thoughts at a meeting organized by Nicholas LaRusso before the annual American Association for the Study of Liver Diseases meeting in Chicago in 1990, Paul Berk (then editor of HEPATOLOGY) asked me to provide a review on this subject. When I finally succeeded squeezing this commitment in between my overload of work and could submit my manuscript in early 1992,117 one reviewer criticized it into pieces for rejection. Victim of my promise to a colleague, I had no other choice but to write a rebuttal longer than the manuscript itself, finding consolation in Cicero's “Damnant quod non intelligunt.” (They condemn what they do not understand.) (Marcus Tullius Cicero; 106 v.C. - 43 v.C.) I proposed that DPM affects different levels of the intrahepatic biliary tree, thus characterizing various anatomo-clinical entities: Caroli's disease and Caroli's syndrome with DPM of larger hilar ducts; congenital hepatic fibrosis (CHF) and autosomal recessive polycystic kidney disease (ARPKD) characterized by DPM at the level of interlobular ducts; and autosomal dominant polycystic kidney disease (ADPKD), as well as Von Meyenburg complexes revealing DPM in finer, more peripheral ducts117, 118 (Fig. 9). Polycystic liver disease (PLD) without renal involvement was only discovered some years later. The so-called central dot sign, characteristic for Caroli's disease and syndrome, corresponds to macroscopic DPM in larger bile ducts visualized by modern imaging techniques.119–121 CHF, associated with ARPKD, can be explained by DPM of interlobular ducts at birth, followed by progressive destructive cholangitis of variable duration and intensity, resulting in a biliary-type fibrosis with variable amounts of bile duct structures remaining in a ductal plate configuration. The dilated renal collecting tubules are subject to similar epithelial involution associated with fibrosis. This hypothesis explains most clinical features in CHF.122Logic brings you from A to B; fantasy brings you everywhere.” (Albert Einstein).

Figure 9.

Schematic representation of the intrahepatic biliary tree and the different congenital fibrocystic bile duct diseases characterized by ductal plate malformation (DPM), described in Desmet.117 The location of the diseases at different levels of the tree indicates the approximate size of the intrahepatic bile ducts affected by DPM in a particular disorder. The entities on the left are characterized by mild or marked dilatation of the bile duct structures, whereas those listed on the right often reveal a variable degree of involution of the ductal plate remnants associated with fibrosis. The scheme indicates that DPM is a basic morphological characteristic of all mentioned congenital bile duct diseases. Redrawn according to Desmet.117

Further fantasies with ductal plates may be worthwhile.123

V. Metabolic Disease

The original work of some coworkers concerned metabolic disease. Francesco Callea studied α1-antitrypsin deficiency. He described the characteristic immunohistochemical picture of the liver in heterozygous PiMZ individuals in conditions of clinical stimulation: the “recruitment-secretory block” phenomenon.124 Use of a monoclonal antibody specific for the Z mutant of α1-antitrypsin allowed distinction of PiZ gene carriers from PiM-like subjects in the absence of serum protein analysis.125 He described a new type of ground-glass hepatocyte due to intracellular storage of fibrinogen in fibrinogen deficiency,126 thus contributing to the group of endoplasmic reticulum storage disorders.127 Raf Sciot investigated the immunohistochemical expression of hepatic transferrin receptor (TfR) in experimental iron overload,128 in genetic hemochromatosis,129, 130 and in secondary iron overload,131 revealing decreased expression and even disappearance of TfR with increasing siderosis levels. Instead, HCC and hepatoblastoma show markedly increased TfR in comparison with surrounding liver tissue, reflecting their malignant nature.132, 133

VI. The Future

During the past nearly half century of my wanderings through hepatopathology, liver needle biopsy and diagnostic histopathology were crucial in fostering the dramatic progress in hepatology. It is astounding to compare the knowledge of today with that of 1963, reminding “Gentlemen, I must tell you that half of what you have been taught is wrong, and we do not know which half.” (Dean Edsall, Harvard Medical School, Graduation Address, 1920).

My work in diagnosis and research and its advancement was nourished all the time by progress in histological technologies, including electron microscopy and (immuno)histochemistry, and fertilized by innovations in surrounding fields, including liver imaging, virology, immunology, liver surgery, and transplantation, but most of all in later years by giant leaps in the new field of molecular biology.

Some fear that this new world is heralding the end of diagnostic histopathology. I rather think this will not be the case. During all my active life, pathology was challenged and has changed, in pace with progress in medicine and biology. Pathologists adapted by absorbing novelties and specializing in particular domains of methodology or higher expertise in organ or disease pathology. The same is apt to happen in the future at a deeper level, including familiarization with molecular techniques. Indeed, this is already happening today. Diagnostic histopathology has every reason to survive as an endless source of information, thanks to the steadily growing power of immunohistochemistry and hybridohistochemistry for localizing molecules in tissues, cells, and organelles, a heritage we owe to pioneers in creating labeled76 and monoclonal134 antibodies and promoters of in situ hybridization,135, 136 who designed and propagated magic bullets that recognize targets with predefined specificity.137 In coming years, further progress in technology, deeper differentiation in specialty fields, and tighter teamwork will prevail, but sound knowledge of pathology and the serendipity of probing minds will certainly remain essential assets.

In retrospect, perhaps I made mistakes by not publishing some thoughts in journals, instead of in more hidden sites such as theses, chapters in books, and meeting reports. But why should I regret this? After all, “I did it my way.” (“My Way”, Frank Sinatra).

In ending, I dedicate my bunch of reminiscences to all who made it possible. I thank the “Nationaal Fonds voor Wetenschappelijk Onderzoek” and the “Fonds voor Geneeskundig Wetenschappelijk Onderzoek” for supporting my research activities. I am grateful to my parents for their lessons in life; to my mentors, especially Marc Verstraete, Jozue Vandenbroucke, and Hans Popper, who put me on the track; to my colleagues, hepatologists Jan DeGroote and Johan Fevery; to my coworkers, those mentioned in the references list, and all the one-time members of my department; my secretary for 25 years Marjan Weckx; and Rita DeVos for help with illustrations. My past and present fellow Gnomes deserve my gratitude because I learned so much from them. I reserve a special thanks for Tania Roskams and Bo VanDamme for giving me the opportunity to continue being fascinated by progress in hepatopathology since my compulsory retirement at the age of 65 in 1996. Last, but not least, I lack the words to rightly thank my family—my wife, three daughters, and two sons—for the environment, support, and peace they have provided me. Without them, I could not have done what I did, could not have gone where I went, and could not have dreamt all that I have been dreaming.