Themes of liver transplantation


  • Potential conflict of interest: Dr. Fung advises Astellas and Bristol-Myers Squibb. He also received grants from Novartis.


Liver transplantation was the product of five interlocking themes. These began in 1958-1959 with canine studies of then theoretical hepatotrophic molecules in portal venous blood (Theme I) and with the contemporaneous parallel development of liver and multivisceral transplant models (Theme II). Further Theme I investigations showed that insulin was the principal, although not the only, portal hepatotrophic factor. In addition to resolving long-standing controversies about the pathophysiology of portacaval shunt, the hepatotrophic studies blazed new trails in the regulation of liver size, function, and regeneration. They also targeted inborn metabolic errors (e.g., familial hyperlipoproteinemia) whose palliation by portal diversion presaged definitive correction with liver replacement. Clinical use of the Theme II transplant models depended on multiple drug immunosuppression (Theme III, Immunology), guided by an empirical algorithm of pattern recognition and therapeutic response. Successful liver replacement was first accomplished in 1967 with azathioprine, prednisone, and antilymphoid globulin. With this regimen, the world's longest surviving liver recipient is now 40 years postoperative. Incremental improvements in survival outcome occurred (Theme IV) when azathioprine was replaced by cyclosporine (1979), which was replaced in turn by tacrolimus (1989). However, the biologic meaning of alloengraftment remained enigmatic until multilineage donor leukocyte microchimerism was discovered in 1992 in long-surviving organ recipients. Seminal mechanisms were then identified (clonal exhaustion-deletion and immune ignorance) that linked organ engraftment and the acquired tolerance of bone marrow transplantation and eventually clarified the relationship of transplantation immunology to the immunology of infections, neoplasms, and autoimmune disorders. With this insight, better strategies of immunosuppression have evolved. As liver and other kinds of organ transplantation became accepted as healthcare standards, the ethical, legal, equity, and the other humanism issues of Theme V have been resolved less conclusively than the medical-scientific problems of Themes I-IV. HEPATOLOGY 2010

The purpose of this contribution to the Master's Perspective Series is to describe in detail the provenance of liver replacement. In the absence until now of such an account, liver transplantation often has been characterized as a natural extension of renal transplantation. In reality, liver and kidney transplantation were codeveloped with the liver as the flagship organ, or alternatively the engine, for much of the time. In the process, the rising tide of organ transplantation altered the practice of hepatology, nephrology, and other organ-defined medical specialties; enriched multiple areas of basic and clinical science; and had pervasive ripple effects in law, public policy, ethics, and religion.

At first, liver transplantation was a fantasy. Transformation of the idea into a reality required essentially de novo development between 1957 and 1962 of five separate but interconnected themes: (I) metabolic interactions between intra-abdominal organs (hepatotrophic physiology), (II) the liver and multivisceral transplant models including donor organ procurement and preservation, (III) the immune system and its control with or without therapeutic immunosuppression, (IV) transplantation outcomes, and (V) humanism-associated issues (social, ethical, legal, public policy).

The five themes can be used to categorize all of the liver transplant milestones of the last half century1-71 as has been done by thematic color-coding and by numbers in Table 1. To help connect this history with the present and future, John Fung, a colleague of more than 25 years, was recruited as a collaborating author; fresh from his 5-year tenure as Co-Editor of HEPATOLOGY's sister journal, Liver Transplantation.

Table 1. Milestones of Liver Transplantation, Color-Coded According to Five Developmental Themes
  1. Theme colors = 1. Green: hepatotrophic physiology; 2. Red: transplant models; 3. Blue: immunology; 4. Pink: survival results; 5. Brown: humanism issues. With major co-themes, the text color is of the dominant one.

  2. Abbreviations: UW, University of Wisconsin; PTLD, posttransplant lymphoproliferative disorders.

inline image

DHHS, Department of Health and Human Services; GVH, graft-versus-host; HLA, human leukocyte antigen; HVG, host-versus-graft; NIH, National Institutes of Health; SRTR, Scientific Registry of Transplant Recipients; UCLA, University of California Los Angeles; UNOS, United Network for Organ Sharing.

My Liverless Early Life

I was born in 1926 in the small town of LeMars, Iowa, and remained there uneventfully until joining the United States Navy directly from high school in 1944.72 (References 72 through 189 are available in the Supporting Information Material.) After the war's end, I remained “in training” for 14 consecutive years, beginning at Westminster College (Fulton, MO), and continuing in chronologic order at the university medical centers of Northwestern University, University of California Los Angeles (UCLA), Johns Hopkins, University of Miami, and again Northwestern. Tangible results from this period included Ph.D. and M.D. diplomas (Northwestern, 1952), board certificates in general and thoracic surgery, and a dozen publications of which the first five were in neuroscience.

The Neuroscience Venture.

My research on the brain stem circuitry of cats (and eventually monkeys) was started at Northwestern at the age of 23 years under the neurophysiology pioneer Horace W. Magoun and finished at UCLA after Magoun's recruitment there as one of the new school's founding chairpersons. Each of the five resulting publications73-77 generated 100 to 300 citations, and a figure from one75 was immortalized as the logo of the UCLA Brain Institute. However, the Ph.D. thesis from this research and completion of the Northwestern M.D. requirements marked the end of my neurophysiology career at the age of 26 years.

The science environment that existed 60 years ago at both Northwestern and UCLA was described in my long letter of response in 1991 to a request by a UCLA Brain Institute archivist (Supporting Information Appendix 1). As described in that letter, Magoun's influence cut deeply. He had no interest in, and very little tolerance for, research that did not have a clear mega-purpose. In our project, the global objective was to delineate with electrophysiologic technology the neural pathways serving the most fundamental elements of brain function: sleep versus wakefulness, cognition, and memory.

A Side Trip to Cardiac Physiology.

The Supporting Information Appendix 1 also contains a 1951 letter (discovered four decades later) from Magoun to Alfred Blalock, Chairman of Surgery at the Johns Hopkins Hospital, that undoubtedly contributed to my acceptance for surgical training at that great institution (1952-1956). After completing the first year in Baltimore, I put aside all clinical work for 18 months to develop a model of complete heart block in dogs, a complication being caused in patients by efforts to close atrial or ventricular septal defects.

With the technology adapted from my neurophysiology experience, I showed that low-voltage bipolar stimulation at any place on the ventricle was a safe and efficient treatment for the bradycardia of heart block. The cardiac pacemaking was promptly instituted clinically at Hopkins and elsewhere. Although the articles describing the experimental work78-80 also were frequently cited, my involvement in the subject of heart block now reached a dead end.

However, the youthful excursions were not wasted. What survived from my exposure to Magoun, and was evident in the heart block research, was the view that all biologic functions were products of a hierarchy of interacting systems and subsystems over which there were controls at multiple levels (i.e., regulatory brain equivalents). In this context, it was more important to learn how a given function was governed than to endlessly pursue details. The “big picture” approach (systems biology) would, in fact, be applied to liver transplantation, the third subject to which I directed concentrated attention.

The Succession of Themes

Anatomically-Influenced Physiologic Interactions Between Organs (Theme I).

While still at Johns Hopkins, I assisted Dr. Blalock in performing a splenorenal shunt in a patient with cirrhosis and insulin-dependent diabetes mellitus who then became insulin-free. The possibility that the portal diversion was responsible for the metabolic change seemed consistent with a then-current hypothesis that excessive degradation of endogenous insulin during its primary passage to the liver via the portal vein was the cause of some forms of diabetes.81 Testing elements of this hypothesis was not possible until after I moved to the new medical school of the University of Miami, Miami, FL, to complete my general surgery residency (1956-1958).

In Miami, I produced a colony of alloxan diabetic dogs, established the animals' steady-state insulin needs, and modified the liver's blood supply with portacaval shunt (Eck's fistula) or other alterations of the portal venous system.82,83 The objective of surgically ameliorating diabetes evaporated when the portal diversion procedures increased instead of decreased the insulin requirements.83 In addition, the hepatic atrophy and systemic morbidity caused by portacaval shunt in normal dogs84,85 appeared to be exaggerated in our diabetic animals.

Development of Liver Transplant Models (Theme II).

A connection of these studies to liver transplantation was made when C. Stuart Welch of Albany, NY, visited Miami in 1957 to give a lecture on the treatment of portal hypertension. During his talk, Welch made casual reference to a canine operation that he had reported in 19551 and more extensively a year later.86 In these articles, the term “liver transplantation” was used for the first time in the scientific literature. The Welch operation consisted of revascularization of an auxiliary liver allograft in the recipient's right paravertebral gutter with provision of portal venous inflow from the inferior vena cava (Fig. 1).

Figure 1.

Auxiliary liver homotransplantation in dogs (the Welch procedure). Note that the portal venous inflow of the extra liver is from the inferior vena caval bed while the native liver retains a normal blood supply. It was suspected from the beginning that this was a major flaw in the design of the procedure. [From: Ann Surg 1964;160:411-439.]

Recognizing that failure to provide the extra liver with a normal portal venous supply could handicap the allograft in the same way as the native livers were damaged in my nontransplant portal diversion models, I began the development of versatile transplant procedures to study the special qualities of splanchnic venous blood in dogs. One of the models was a method of total recipient hepatectomy, the unique feature of which was preservation of the retrohepatic inferior vena cava2 as in the first stage of today's piggy-back human liver transplantation. For liver allograft implantation, it was technically easier to simply remove this portion of the recipient vena cava and replace it with the comparable segment of the donor liver's vena cava into which all of the hepatic veins empty.3

Operative survival with the complete canine replacement operation (Fig. 2) was not accomplished until a few days after I moved to Northwestern in June 1958 for a final 12 months of cardiovascular surgical training that was expected to culminate in an academic practice in thoracic surgery. Instead, two steps were taken during the summer of 1958 that ensured pursuit of the liver research for at least 5 years beyond completion of the thoracic residency. The first step was the submission of a four-page NIH grant focused on metabolic studies in which liver replacement was one of the experimental models. The second step was my nomination by Northwestern for a John and Mary Markle Scholarship. Here, the emphasis was radically different.

Figure 2.

Complete liver replacement in the dog, circa 1958-1959. The fact that this was a canine rather than a human operation is evident only from the small multiple lobes of the allograft and the biliary drainage with cholecystoduodenostomy. In my first report,3 an “outflow block” syndrome resembling endotoxin shock was described if donor body weight was less than half that of the recipient (one cause of today's “small for size” syndrome).

Markle Scholar candidates were expected to identify an open-ended career objective. Ignoring advice to develop a “more realistic” project in the emerging field of open heart surgery, I proposed the life goal of clinical liver transplantation. In the autumn of 1958, I learned that the NIH grant would be fully funded for 5 years, and shortly thereafter that I had been selected as a Markle Scholar. The first phase of the canine liver project was nearly completed by the time I finished the thoracic residency and the dual revenue streams began on July 1, 1959. In addition, a second operation had been perfected in which the liver was transplanted as part of an allograft that contained all of the other intra-abdominal viscera (Fig. 3).6, 7

Figure 3.

(Bottom center): Multivisceral allograft transplanted in dogs in 19596, 7 and in humans for the first time three decades later.46 With removal of different organs from the common vascular stem, this original procedure has had many subsequent variations. (Bottom left): Liver-intestinal transplantation.47, 62 (Top): Cluster of upper abdominal organs.55 (Bottom right): Mid-gut organs except the liver. [From: Liver Transpl Surg 1998;4:1-14.]

The magnitude of the Markle proposal should have been intimidating, but it did not seem so at the time. The slate of liver transplantation was nearly blank in 1958, but what had to be done was transparent: make the operation biologically sound, make it practical, and find a way to prevent allograft rejection. I was not the only person to think that way. Although I did not learn of it until a year later, Francis D. (Franny) Moore had begun independent efforts to replace the dog liver during the summer of 1958 at the Peter Bent Brigham Hospital in Boston4, 5 that continued until the mid-1960s.87,88

Moore's transplant interests were not confined to the liver. This can be perceived most clearly by reading his book, Give and Take,89 and his autobiography, A Miracle and a Privilege,90 written four decades later. Epitomizing his ubiquitous presence, Moore presided as chief of surgery at the Brigham over the clinical renal transplant trials of Murray and Merrill that yielded the world's first example in any species of survival of an organ allograft for 1 year or longer.91 In this case, the kidney from a fraternal twin was transplanted to his irradiated brother on January 24, 1959, and functioned for the next 20 years without maintenance immunosuppression (Table 2).

Table 2. Characteristics of the First Successful Transplantation of Kidney Allografts with >6 Months Survival as of March 1963
PhysicianReferenceSiteDateDonor RelationshipGraft Survival
  • *

    Kuss and Hamburger described periodic administration of adrenal cortical steroids with these patients.

  • Patient death occurred at or shortly after listed time.

  • Patient underwent successful retransplantation in the 1970s; elected to French Parliament.

  • §

    First successful with drugs-only immunosuppression (no radiation).

Joseph E. Murray91Boston, MAJanuary 24, 1959Fraternal twin20 years
Jean Hamburger103Paris, FranceJune 29, 1959Fraternal twin26 years
Rene Kuss105Paris, FranceJune 22, 1960Unrelated18 months
Jean Hamburger*104Paris, FranceDecember 19, 1960Mother22 months
Rene Kuss*105Paris, FranceMarch 12, 1961Unrelated18 months
Jean Hamburger*104Paris, FranceFebruary 12, 1962Cousin15 years
Joseph E. Murray106,107Boston, MAApril 5, 1962Unrelated17 months§
Thomas E. Starzl8, 8108-109Denver, CO1962-1963First series: a mixture≥45 years

From my point of view, this faint signal that the genetic/immunologic barrier to organ alloengraftment might be surmountable made the liver transplant objective less distant. It seemed almost providential that the 5-year Markle Scholarship and NIH funding (1959-1964) for my liver project began a few months after the fraternal twin transplantation. The 5 years was equally split between Northwestern where I was elevated to a junior faculty position on July 1, 1959, and the University of Colorado where I was appointed Associate Professor of Surgery and Chief Surgeon at the Denver VA Hospital from November 1961.

The Immune System and Its Control (Theme III).

Until 1958-1960, the only organ allograft whose unmodified rejection had been thoroughly studied was the kidney. Rejection to death of our canine liver recipients usually occurred in 5-10 days.3 However, in rare outliers in which the biochemical indices of rejection improved spontaneously, the liver allograft's dominant histopathologic findings by 3 weeks were those of repair and regeneration.92 These were the first recorded exceptions to the existing dogma (based on skin graft research) that rejection, once started, was inexorable.

In the multivisceral grafts (Fig. 3), the pathology was subtly different. Rejection of the various organs, if they were part of the multivisceral graft, was less severe than when the organs were transplanted alone. Moreover, there was overt evidence in recipient tissues of a graft-versus-host (GVH) reaction, but without a skin rash or other manifestations of graft-versus-host disease (GVHD).7 The double immune reaction (host-versus-graft [HVG] and GVH) exposed by those experiments was shown a third of a century later to be a feature of alloengraftment and acquired tolerance no matter what the transplanted organ (see below).

Both my liver-alone and multivisceral transplant models were generally viewed as technical exercises of little if any scientific interest. One reason was the prevailing view that was concisely expressed in 1961 by the 1960 Nobel Laureate F. M. Burnet in a New England Journal of Medicine review titled, “The New Approach to Immunology”. The discouraging passage read: “… Much thought has been given to ways by which tissues or organs not genetically and antigenically identical with the recipient might be made to survive and function in the alien environment. On the whole, the present outlook is highly unfavorable to success”.93

I was poorly equipped to rebut this kind of opinion. My attempts in Chicago to use radiation therapy for canine liver transplantation in 1959-1960 failed miserably.94 During this bleak time, however, it was reported in a closely-spaced succession of articles that 6-mercaptopurine and/or its analogue, azathioprine, were immunosuppressive in nontransplant,95,96 rabbit skin graft,97,98 and canine kidney transplant models.99,100 The most extensive kidney transplant experiments were done by the 30-year-old English surgeon, Roy Calne101 who began his studies at the Royal Free Hospital in London in 1959 while still a registrar (resident). The work was continued in Boston with Joseph Murray after July 1960.102

In 1961, Calne visited our laboratory in Chicago and described his results. Shortly thereafter, I moved to Colorado, after making the decision to develop a human kidney transplant program there with drug immunosuppression as a forerunner for the liver objective. This would be a bold step because the renal center at the Brigham was the only one in the U.S. at the time with an active clinical transplant arm. After demonstrating in parallel canine kidney and liver transplant studies of azathioprine that advances with either organ would be applicable to the other, we concentrated our immunosuppression research on the simpler kidney model. Our most promising results were obtained by giving daily doses of azathioprine monotherapy before as well as after kidney transplantation, adding postoperative prednisone only when overt rejection developed.

By the time the incremental drug protocol was taken to the clinic in the autumn of 1962, six renal allograft recipients who were treated primarily or exclusively with the total body irradiation protocol of Murray's fraternal twin case (see earlier) had either passed or would soon reach the 1-year survival milestone, including two French patients to whom the donors were not genetically related (Table 2).91,103-105 In addition, Murray had transplanted a deceased donor allograft in Boston on April 5, 1962, under azathioprine-based immunosuppression.106,107 The kidney was destined to function for 17 months and become the world's first to survive for 1 year or more with a radiation-free (drugs-only) protocol. Enthusiasm generated by this last case was tempered, however, by the fact that the recipient was the only one of the first 10 in the Boston azathioprine series to survive longer than 6 months (details annotated in Starzl108).

Some members of our Denver team concluded from this sobering news that our accrual of more renal transplant cases would be a futile and embarrassing undertaking. My counter-argument was that our laboratory-based treatment strategy differed in many ways from the one used in the Boston protocol, including a role of prednisone equal in importance to that of azathioprine. The differences proved to be crucial. First in dogs, and then in human kidney recipients, the graded use of azathioprine and prednisone exposed the two features of the alloimmune response that provided the basis for the transplantation of all kinds of organs.

The two phenomena were capsulized in the title of a 1963 report of the first-ever series of successful kidney allotransplantations: “The Reversal of Rejection in Human Renal Homografts with Subsequent Development of Homograft Tolerance”.8 The principal evidence that the allografts (then called homografts) had somehow induced variable donor-specific tolerance was that the reversal of rejection frequently was succeeded by a time-related reduction, or in some cases elimination, of the need for maintenance immunosuppression. In fact, eight recipients in the 1962-1964 Colorado series of 64 still bear the world's longest functioning renal allografts, 45 or more years later.109 Six of the eight have been off all immunosuppression medications for 12-46 years.

Transplantation Outcomes with the Forerunner Kidney (Theme IV).

The >70% one-year patient and renal graft survival in our seminal Colorado series110,111 exceeded my own expectations and was not considered to be credible until David Hume in Richmond, VA, and others added their confirmatory experience. The worldwide reaction was remarkable. In the spring of 1963, there had been only three clinically active renal transplant centers in North America (Boston, Denver, and by now Richmond) and scarcely more in Europe. Only 1 year later, 50 new renal programs in the United States alone were either fully functional or were gearing up.

In reflecting back a dozen years later on the kidney transplant revolution of 1962-1964, I began my founding lecture for the American Society of Transplant Surgeons with the comments that: “From time to time, a news story appears about the birth of a husky, full-term baby, much to the amazement of the chagrined mother who had not realized that she was pregnant. Mother Surgery seemed to have been thus caught by surprise when clinical transplantation burst upon the scene in the early 1960s.”112

Issues of Humanism (Theme V).

Liver transplantation was swept up in the 1962-1964 kidney momentum. However, there were many reasons to be cautious, not the least of which were social, ethical, and legal concerns. Throughout 1962, I discussed these issues personally with key nonuniversity persons: the Colorado Governor (John Love), our U.S. Senator (Gordon Allot), the Denver Coroner, the Chief Justice of the Colorado Supreme Court, and clerical leaders. All ultimately expressed support. Resistance within the University was dealt with by the legendary medical school dean, Robert J. Glaser, and the University Chairman of Surgery, William R. Waddell.

Unprecedented technical challenges were expected. The liver replacement operation, which was difficult even under the optimal circumstances of the animal laboratory, predictably would be harder in recipients with portal hypertension and other pathophysiologic and anatomic changes of chronic liver disease. In the absence of artificial organ support, failure of the hepatic graft to promptly function would be tantamount to death. Finally, how could immediately life-supporting deceased donor livers be obtained in an era in which death was defined as the cessation of heartbeat and respiration?

These questions and issues mandated consideration of the less draconian auxiliary hepatic transplant operation of Welch that might allow recipient survival, even if the graft failed. This option was undermined when the rapid atrophy of auxiliary livers that previously had been ascribed to rejection in unmodified dogs,86,113 was shown to be equally severe in animals in which rejection was prevented with azathioprine.11 The die was cast for the liver replacement (orthotopic) option.

The First Human Liver Transplantations

Liver replacement was carried out in seven deceased donor liver recipients between March 1963 and January 1964: five in Denver (cases 1-4 and 6), one in Boston (case 5 by Moore's team), and one in Paris (case 7) (Table 3).10, 1188,114 All seven patients died, two during the operation and the other five after 6.5-23 days. Neither primary nonfunction nor uncontrolled rejection of the grafts were lethal factors in any of the failures.

Table 3. First Recipients of Replacement Livers
AgeDateCity (Ref)Liver DiseaseSurvival (Days)Main Cause of Death
3Mar-1-1963Denver (10)Biliary Atresia0Intraoperative bleeding
48May-5-1963Denver (10)Hepatoma, Cirrhosis22Pulmonary emboli, sepsis
68Jun-3-1963Denver (10)Duct Cell Carcinoma7.5Pulmonary emboli
52Jul-10-1963Denver (11)Hepatoma, cirrhosis6.5GI bleeding, pulmonary emboli/edema, liver failure
58Sep-16-1963Boston (88)Colon metastases11Pneumonitis, hepatic abscesses, failure
29Oct-4-1963Denver (11)Hepatoma23Sepsis, bile peritonitis, pulmonary emobli
75Jan-?-1964Paris (114)Colon metastases0Intraoperative hemorrhage

At autopsy of the four Denver patients who survived the operation, pulmonary emboli were found that apparently had originated in the bypass tubing used to decompress the blocked systemic and splanchnic venous beds during the removal and replacement of the native liver. Ironically, the bypass which had been an essential component of the canine operation, is not mandatory in most human recipients, or even in dogs if venous collateralization is encouraged by bile duct ligation a month in advance.115

By the time our fourth and fifth liver recipients were reported to the American Surgical Association in April 1964,11 all clinical liver transplant activity had ceased in what would be a voluntary 3.5-year worldwide moratorium. The self-imposed decision to stop did little to quiet polite but unmistakably disapproving discussions of an operation that had come to be perceived as too difficult to ever be tried again.

The Moratorium

In effect, it now would be necessary to return to ground zero and reexamine all five of the themes of Table 1. The central assumption of Theme I had been that portal venous blood contained hepatotrophic molecules. The hypothesis was consistent with our results in 1958-1960 in nonimmunosuppressed canine recipients of replacement livers,3 and especially with the acute atrophy of Welch's auxiliary grafts in azathioprine-treated dogs (see above, and Starzl et al.11). The possibility was now explored of providing the auxiliary allografts with direct access to the portal molecules.116

But what were the hepatotrophic factors? Using double liver fragment nontransplant models derived from Welch's auxiliary liver operation (Fig. 4), it was proved during and after the moratorium that insulin is the principal (although not the only) hepatotrophic molecule in portal blood; that insulin is avidly removed by the liver; and that its primary passage through the hepatic microvasculature is crucial for the maintenance of liver size, ultrastructure, function, and the capacity for regeneration.27, 28116-122 When other molecules subsequently were identified that had insulin-like or diametrically opposite effects (Table 4), hepatotrophic physiology blossomed into multiple research areas of metabolism and regenerative medicine.123,124

Figure 4.

The double liver models that led to progressively precise identification of the hepatotrophic factors that influenced liver size, ultrastructure, function, and the capacity for regeneration: (A) Welch's index operation of auxiliary liver allotransplantation (see also Fig. 1); (B) nontransplant split liver model that differentiated the effect on the liver of systemic venous (vena caval) versus splanchnic (portal) blood; (C) separation with the double liver fragment model of the qualities of venous blood from the upper and lower abdominal viscera; and (D) selective infusion of candidate hepatotrophic molecules into one or the other of two liver fragments, both of which had an arterial supply only. [From: Liver Transpl Surg 1998;4:1-14.]

Table 4. Hepatotrophic Factors Revealed by 1994 with Portal Diversion, Double Liver Fragment, or Partial Hepatectomy Models
  • It is noteworthy that numerous humoral and cellular mechanisms involved in liver size homeostasis and regeneration (not shown here) are the same as those involved in immunologic responsiveness (rejection) and unresponsiveness (tolerance). Annotated in Francavilla et al.124

  • *

    Mitogenic in tissue culture.

  • Inhibitory in tissue culture.

  • Annotated in HEPATOLOGY 1994;20:747–757.24

 Growth factors:
  Augmenter of liver regeneration (ALR)
  Insulin-like growth factor II (IGF II)
  Transforming growth factor α (TGF-α)*
  Hepatocyte growth factor (HGF)*
  FK binding protein12 (FKBP12)
 Growth factors:
  Transforming growth factorβ (TGFβ)

Although the moratorium studies did not support reconsideration of auxiliary liver transplant trials, they added a new dimension to the operation of portacaval shunt, which had been used primarily to treat complications of portal hypertension. With the demonstration of the profound effects of portal diversion on protein, carbohydrate,119 and lipid metabolism,121 portacaval shunt was used to favorably alter the course of three categories of inheritable metabolic disorders: glycogen storage diseases,125,126 familial hyperlipoproteinemia,127,128 and alpha-1-antitrypsin deficiency.129,130 The dramatic amelioration of the pathophysiology of these diverse conditions (e.g., hyperlipoproteinemia, Fig. 5) presaged their definitive correction with liver replacement (see next section).

Figure 5.

The dramatic effect of portacaval shunt on serum cholesterol concentration in a child with homozygous familial hyperlipoproteinemia. These observations24 and canine studies of lipid synthesis with the models shown in Fig. 425 suggested that the liver was the principal site of cholesterol homeostasis. Although we considered familial hyperlipoproteinemia to be a candidate disease for liver replacement since the mid-1970s, this was not accomplished until February 14, 1984,44, 45 by which time more evidence that this was an appropriate step was obtained in New York, NY; Bethesda, MD; and Dallas, TX. Interactions over more than a dozen years between experts in cholesterol metabolism in these cities and the author (T.E.S.) are described in the chapter “The Little Drummer Girls” of The Puzzle People.128 [From: Lancet 1973;2:940-944.]

Themes II (the surgical operations) and III (immunology) were pursued with both kidney and liver canine transplant models. These efforts included the construction and testing of equipment with which livers could be preserved for 1 or 2 days,131 the experimental development and clinical introduction of antilymphoid globulin,13,132 and the demonstration that immunosuppression-aided organ tolerance was more frequently induced by the liver than by the kidney.12 In addition, studies of our burgeoning human kidney recipient population clarified the role of human leukocyte antigen (HLA) matching in all kinds of organ transplantation.14

Activity also had intensified on the humanism issues (Theme V). The agenda items at medical ethics conferences in 1966-196715, 16 included human experimentation, living organ donation, informed consent, and the equitable allocation of organs. The most definitive consequence of these discussions was an evolving consensus that the end of life was more appropriately defined by brain death than by the previous criteria of cessation of heart beat and respiration.18

The Liver Transplant Beachhead

Despite these accomplishments, confidence about our impending liver trial was nowhere near the level that had existed during the run-up to the 1963 attempts. The legacy of doubt from the earlier failures was cancelled by a critical new factor. This was the arrival in 1966 of Carl Groth, a 32-year-old Fulbright Fellow from Stockholm who joined all of the thematic developments and became a key member of both the donor and recipient teams. With Groth's leadership, multiple examples of prolonged human liver recipient survival were produced in 1967 (Fig. 6), using triple-drug immunosuppression (azathioprine, prednisone, and antilymphoid globulin).17

Figure 6.

The first three human recipients with prolonged survival following liver replacement in July and August 1967. The adult, Carl Groth, was a Swedish surgeon-in-training whose tenures in Denver as a Fulbright Scholar (1966-1968) and faculty member (1970-1971) were near the beginning of his Olympian career. After returning to Stockholm to occupy a Chair in transplantation surgery created for him at the Karolinska Institute, Groth developed the multiorgan transplant program that produced the first liver transplantations in Sweden. His numerous honors include the King's Medal of his country and the Medawar Prize, the highest distinction of the International Transplantation Society.

The first Denver successes were bolstered by the opening in 1968 of a second clinical liver program by Roy Calne in Cambridge, England,133 following preclinical studies in outbred pigs.21,134 The early trials were described in my 1969 book titled, “Experience in Hepatic Transplantation”,22 based on our first 25 human liver replacements and eight performed elsewhere (four by Calne). Collateral support was provided with the use of the same immunosuppression regimen for the first successful human heart, lung, and pancreas transplantations (Table 5).135-137 However, the promise of the nonrenal procedures, and even of deceased donor kidney transplantation, was unfulfilled for the next 12 years because of immunosuppression-related morbidity and mortality.

Table 5. The Domino Effect in 1968-1969 of the 1967 First Successful Human Liver Transplantations
OrganCityDatePhysician/ SurgeonReference
  • *

    Patient died after 10 months; all others in table lived >1 year with functioning graft. The first >1-year survival of isolated lung recipients was not reported until 1987.

  • Kidney and pancreas allografts in a uremic patient.

HeartCape Town1/2/68Barnard135

Half or more of the liver recipients treated during this time died within the first post-transplant year. The most encouraging observation was that many patients who survived to this milestone were quietly compiling years of good health thereafter (Fig. 7).64,155 Despite deepening suspicion that progress in the whole field of organ transplantation had permanently stalled, the new French and German liver teams of Henri Bismuth and Rudolf Pichlmayr joined the Denver-Cambridge (England) alliance in the early 1970s, followed later in the decade by the Dutch group of Rudi Krom. Much of the medical-scientific, logistic, and administrative framework of hepatic transplantation that exists today was developed by the five mutually supportive liver centers during the frustrating period between 1969 and 1979.

Figure 7.

World's longest surviving liver recipient whose 40th post-transplant anniversary took place January 22, 2010. The primary disease diagnosis was biliary atresia, but the right lobe of her excised liver contained an incidental 2.7-cm × 1.8-cm hepatoma. The serum alpha-fetoprotein level was 6000 ng/mL at 1 post-transplant month, trace-present at 4 months, and undetectable since.189 The patient's companion, now a retired United States Marine, is her husband of many years. The statue behind them is Roberto Clemente (1934-1972), the greatest baseball right fielder of all time who was killed bringing food by air flight to victims of the catastrophic Nicaraguan earthquake.

Most of the indications for liver transplant candidacy were obvious, including inheritable disorders with a definitive biochemical explanation (e.g., Wilson's disease23). The acid test of liver transplantation ultimately would help elucidate the mechanisms or pathophysiology of less well-understood inborn errors: e.g., the three diseases that were palliated by portacaval shunt (see above). Four patients with alpha-1-antitrypsin deficiency underwent liver transplantation between 1973-1977.138,139 Liver replacement for treatment of glycogen storage disorders,140,141 hyperlipoproteinemia,44, 45 and a growing panoply of other metabolic diseases awaited better immunosuppression.

The Liver Avalanche

Improvements in therapy were heralded in 1979 by Roy Calne's report of cyclosporine-based immunosuppression in 34 patients, including two liver recipients.33 The side effects of cyclosporine precluded its use as a single agent. However, when it was substituted for azathioprine in our two-drug or three-drug therapeutic algorithm that included dose-maneuverable prednisone,34 cyclosporine's full potential was realized. Kidney recipients were the first to be treated, with liver recipients close behind. Eleven of our first 12 liver recipients treated in Colorado with cyclosporine-based immunosuppression during 1979-1980 survived for >1 year.35

More experience in 1981-1982 (now in Pittsburgh) was confirmatory. In December 1981, these findings were reported to C. Everett Koop, the U.S. Surgeon General, who initiated a Consensus Development Conference for liver transplantation that would include input from the European centers. Before the conference, I prepared a summary of our experience for presentation on November 1, 1982, at the American Association for the Study of Liver Diseases, and publication in HEPATOLOGY the same month.36 An updated version was presented to the Consensus Development Conference on June 20-23, 1983.

The consensus committee concluded that liver transplantation had become a “clinical service” as opposed to an experimental procedure.38 The resulting worldwide stampede to develop liver transplant centers was even more dramatic than that of kidney transplantation 20 years earlier. Only 6 years after the Consensus Conference, a 17-page article equally divided between the October 12 and October 19 issues of the New England Journal of Medicine142 contained a opening statement that stated, “The conceptual appeal of liver transplantation is so great that the procedure may come to mind as a last resort for virtually every patient with lethal hepatic disease.” It already was evident that the need for these operations would greatly exceed both an identifiable source of organs and those qualified to transplant them.

A significant number of the next generation of liver transplant leaders who flocked to Pittsburgh for clinical training during the 1980s were not surgeons. Their primary connection was with David Van Thiel (Fig. 8), the brilliant gastroenterologist who became a founding doyen of transplantation hepatology along with his English counterpart, Roger Williams of the Cambridge-King's College program. During this volatile period, preclinical studies of tacrolimus were begun that would lead to its substitution for cyclosporine56, 57 with fast-track U.S. Food and Drug Administration approval in November 1993. With tacrolimus, the multivisceral and intestine-alone transplant procedures developed three decades earlier in dogs (Fig. 3) achieved the status of a genuine “clinical service”.61, 62 The timing was perfect. With arrival of my 65th birthday in 1991, I retired from active surgical practice.

Figure 8.

Non-surgeons at table (circa 1982) who played an indispensible role in establishing the University of Pittsburgh liver transplantation program. Left to right: Basil Zitelli (pediatrics), Robert Schade and David Van Thiel (gastroenterology/hepatology), and Klaus Bron (interventional radiology). Without Van Thiel's Herculean efforts the program could not have been established.

Thematic Epilogue: 1991-2009

Most of the advances in liver transplantation during the succeeding 18 years (Table 1) have been derivative from earlier work, including the use of partial livers from deceased or living volunteer donors. However, the antecedent contributions with which the taxonomical foundation of organ transplantation was built have been obscured with the advent of the World Wide Web. Many of the referenced articles of the preceding narrative cannot be accessed online in full text, and some have become invisible. With the dearth of electronic information from before the 1990s and the convenience of citing easy Internet finds, the recent literature has been replete with observations, events, and concepts that were described more clearly years or decades before. Nevertheless, there have been new trends in organ transplantation, two of which were driven mainly by the liver.

The Exegesis of Alloengraftment.

A major gap in immunology (Theme III) when I stopped surgical practice was the inability to explain why organ transplantation had been possible. Because organ recipients were not infused with donor leukocytes, it became dogma by the early 1960s that the donor leukocyte chimerism associated with acquired tolerance in experimental models was not a factor in organ engraftment. The dogma was not challenged until we discovered small numbers of multilineage donor leukocytes (microchimerism) in the blood or tissues of all studied long-surviving liver, kidney, and other organ recipients.63, 64,143 These findings in 1992-1993, and an array of supporting experimental studies in congenic rat144-150 and mouse models,151-154 mandated a change in the previously perceived landscape of transplantation immunology.

It was proposed63, 64,155,156 that organ transplantation was the equivalent of a bone marrow transplantation. The key step leading to rejection, or alternatively alloengraftment, after both kinds of transplantation was hematogenous migration of leukocytes (including stem cells157-159) to the recipient's lymphoid organs (Fig. 9). Otherwise, the presence of the allograft would not be recognized: i.e., the “immune ignorance”160,161 first described in a transplant model by Clyde Barker and Rupert Billingham 42 years ago. The seminal mechanism of alloengraftment was exhaustion-deletion of the T cell response162,163 induced at the host lymphoid sites by the invading cells (Fig. 9). Because the migrant donor leukocytes are immune-competent, successful alloengraftment involved a double immune reaction in which immune responses of coexisting donor and recipient cells, each to the other, were reciprocally exhausted and deleted under a protective umbrella of immunosuppression (Fig. 10).

Figure 9.

The cell migration and localization of organ and bone marrow cell transplantation. Organs (a liver is pictured here) are composites of architecturally fixed cells and mobile multilineage cells of bone marrow origin (“passenger leukocytes”) that include pluripotent hematolymphopoietic stem cells.157-159 Within minutes after organ transplantation, the passenger leukocytes simulate a bone marrow cell infusion by migrating selectively to recipient lymphoid organs where they induce the depleted antidonor T cell response. Although the clonal response normally destroys the invading donor cells and their outlying source organ (rejection), the response may be exhausted and deleted if it is too weak to eliminate the invading donor cells during the first few weeks of maximal cell migration. Perpetuation thereafter of survival of the bystander organ allograft requires persistence of enough donor leukocytes to maintain the initial exhaustion-deletion. Importantly, the invading donor cells are immune-competent and their response against the recipient also must be exhausted and deleted for a successful transplant outcome (see Fig. 10).

Figure 10.

The kinetics of immunosuppression-aided exhaustion and deletion of the contemporaneous host-versus-graft (HVG curve) and graft-versus-host (GVH curve) responses in organ recipients following the cell migration shown in Fig. 9. Although HVG is the dominant response in most organ recipients (expressed as rejection), serious or lethal GVH reactions (expressed as graft-versus-host disease [GVHD]) are not rare in recipients of lymphoid-rich organs (liver, intestine). In naturally immune-deficient or cytoablated bone marrow cell recipients, GVHD is avoided by using histocompatible (HLA-matched) donors. Therapeutic failure after either organ or bone marrow cell transplantation implies the inability to control one, the other, or both of the responses. [From: N Engl J Med 1998;339:1905-1913.]

Our interpretation of the microchimerism was at first highly controversial164,165 because it was incompatible with multiple theories and hypotheses that made up much of the base of transplant immunology. Resistance to the new concept was eroded when Rolf Zinkernagel in Zurich independently proposed an explanation of acquired tolerance to pathogens that was essentially the same as that of our allotolerance paradigm. In the 1970s, Zinkernagel and Doherty had demonstrated that the major histocompatibility complex-restricted cytolytic T cell response induced by noncytopathic microorganisms was the same as that induced by allografts. These studies were done in highly controlled experimental models of infection with the lymphocytic choriomeningitis virus and other intracellular parasites.166 Their subsequent investigations of tolerance were done with the same models and described in four landmark articles between 1993 and 1997.167-170

With recognition that the Pittsburgh and Zurich investigations were on parallel pathways, a joint author review was published in a December 1998 issue of the New England Journal of Medicine in which analogous scenarios were described of transplantation and pathogen-specific infections (e.g., chronic rejection vis-a-vis chronic viral hepatitis).65 The concept that was developed from transplant and infection models was generalized in the following way: “The migration and localization of antigen govern the immunologic responsiveness or unresponsiveness against infections, tumors, or self—and against xenografts or allografts.”65 In this view, all outcomes in the divergent circumstances of transplantation, including those of microchimerism,150,171,172 were determined by the balance established between the amount of mobile donor leukocytes with access to host lymphoid organs and the number of donor-specific cytolytic T lymphocytes induced at the lymphoid sites (Fig. 11, inner graph).65

Figure 11.

The many faces of transplantation tolerance. (Outer circle): The continuum of experimental and clinical donor leukocyte chimerism-associated tolerance models that can be traced back to observations in 1945 in freemartin cattle (upper left) whose fused placentas permitted fetal cross-circulation, blood chimerism, and reciprocal immune nonreactivity. (Inner graph): Permutations of tolerance defined as balances between persisting migratory donor leukocytes and the number of antidonor T cells undergoing steady-state exhaustion-deletion. The achievement of balances and the resulting clinical phenotypes are influenced by the dose, type, and timing of immunosuppressive therapy and by the dose, type, timing, route, and localization of the migrant donor cells. The single most important factor leading to the macrochimerism of bone marrow cell transplantation versus the microchimerism of most organ (and composite tissue) recipients is enfeeblement of recipient immune reactivity before the arrival of donor cells in the first instance and after their arrival in the second. The nonspecific potential “stabilizing factors” in the left-directed arrow above the human torso include special cells (e.g., T-regulatory), enhancing antibodies, graft secretions, and endogenous cytoprotective molecules.

Long-term organ alloengraftment with this generalizable paradigm was a highly variable form of leukocyte chimerism-dependent tolerance, the completeness of which could be inferred from the amount of immunosuppression necessary to maintain stable function and structure of the transplant (Fig. 11). In a second article with Zinkernagel, the Pittsburgh-Zurich immunologic paradigm provided a road map for improved therapeutic strategies of transplant patient management based on two principles: recipient pretreatment and the least possible use of post-transplant immunosuppression.68 When applied clinically for different kinds of organ transplantation,69 these strategies have minimized, or in some cases eliminated, the burden of chronic immunosuppression.173-178 More rational approaches also were developed for the treatment of opportunistic infections caused by noncytopathic microorganisms.70,168,179

Reporting of Transplantation Outcomes (Theme IV) and Equitable Organ Allocations (Theme V).

A second trend coincided with and was empowered by the rise of the Internet. One of the mandates of the 1984 National Transplant Act was the formation of an Organ Procurement and Transplantation Network (OPTN). Another was the development of a Scientific Registry of Transplant Recipients (SRTR) with which patient and graft survival could be quantified from center to center along with center-specific parameters. After the Department of Health and Human Services (DHHS) awarded the contract for both functions to the United Network of Organ Sharing (UNOS), disputes about organ allocation within the appointed UNOS committee prevented the development of the required plan. In order to avoid a UNOS default of contract, a document was pieced together from two articles that were “in press”, which described the renal180 and nonrenal181 distribution systems already in place in Pittsburgh.

In the contract derived from these manuscripts and presented to DHHS on the eve of the deadline, the overwhelming factor for liver distribution was recipient urgency of need.181 In contrast, time waiting dominated kidney distribution with major credit for HLA matching only when this was complete.180 Although these policies were accepted by DHHS and provisionally implemented in November 1987, they were widely abridged182 until the final regulations were issued by DHHS on April 2, 1998. During the chaotic intervening decade (see Supporting Information Index for a cryptic description of the “liver wars”), UNOS led the opposition to adoption of the regulations and withheld access to SRTR. An editorial in Lancet during the heat of the debates suggested that, “UNOS would better serve the transplant community if it abandoned its stance and began working with DHHS to draw up allocation policies that are practical and fair.”183

One of the most contentious issues was the conclusion in a large Pittsburgh study published in 1994 that liver transplantation performed too early was associated with a net loss of recipient life years.184,185 These findings led to retention of the “sickest first” policy in both the provisional and final DHHS rules for liver allocation. In the meanwhile, the continued resistance to release of center-specific data, as well as inaccuracies and inconsistencies in the first SRTR reports (1992, 1995, and 1997), led to transfer of SRTR management to the University of Michigan-based Arbor Research Collaborative for Health. An Arbor multicenter study in 2005 confirmed the original Pittsburgh findings about the timing of liver transplantation and came to the same policy recommendations.186

Until now, success with liver transplantation has been judged largely by relatively short-term patient and graft survival. A more complete profile has been made possible by the use of the treatment-based evaluation system of Clavien in which the rate and severity of complications (including death) are quantified with a five-tier scale.71 The value of this objective assessment was exemplified by a recent Pittsburgh study of right lobar living donor liver transplantation.187 The Clavien metric is applicable to all kinds of organ transplantation, and has been generalized to other surgical and medical procedures.188


Liver transplantation began with almost no resources at the same time as the tentative first steps were taken to land a man on the moon. Because human lives would be at stake, both objectives had a sacramental element from the outset: i.e., a solemnly binding commitment to perfection. A need for that pledge still exists.


We thank Ms. Terry L. Mangan for her assistance in manuscript preparation. We also thank Mr. Ed Gray, a Systems Engineer, for his honest broker and intellectual contributions between 1999-2009 without which this manuscript could not have been written.