Of mentors, mentoring, and extracellular matrix


  • Potential conflict of interest: Nothing to report.


D. Montgomery Bissell briefly describes the way he came to a career in academic medicine, how he found mentors, initial projects, and finally a focus on matrix biology and hepatic fibrosis. He draws some lessons from the experience, which should have relevance for physician-scientist trainees, those considering that path, anyone with responsibility for training the next generation, institutional leaders, and the National Institutes of Health. (HEPATOLOGY 2009;50:1330–1338.)

South Boston, 1960s

The area was suffering from inner city blight. At the Boston City Hospital (BCH), tradition was strong, but the physical plant, like the neighborhood, was in decay. Its concrete floors had the patina of layer upon layer of Pine-Sol and dirt. Problems of alcohol and malnutrition rolled in a steady stream through the doors of the “Accident Floor”. As interns, we saw scurvy and Wernicke's disease. A triage nurse sent urgent cases to room 1, nonurgent cases to room 2, and the merely drunk to room 3. Local lore had it that Eugene O'Neill had been among the yellow and bloated who entered in 1953, his final year. Regardless of its accuracy (reportedly he died in a Boston hotel room), the story provided our grimy surroundings with a small touch of the poet.

Liver patients sick enough to be admitted were treated in large part with protocols developed through local research, much of it from the Boston Inter-Hospital Liver Group. Known as BILG, it had come together in the late 1950s and was under the leadership of Tom Chalmers. The group championed randomized controlled trials and did some of the first, changing clinical practice in several important areas. The “bench” end of clinical investigation also was thriving at the Thorndike Memorial Laboratory. Much has been written about this remarkable institution, which arguably had no equal as a clinical research institute within a public hospital. Created in 1923, it was a Harvard-staffed but administratively independent academic enclave embedded within the hospital. Its space was just minutes from the medical wards and the patient-subjects who participated in its clinical research protocols. In effect, the wards functioned as acute care beds and unofficial clinical research units. Interns and residents by necessity were actively involved. In return, they obtained expert consultation for their patients, be it Sidney Ingbar's input on a case of thyroid storm or Charlie Trey's help with fulminant hepatic failure. The experience of watching one's “charity” patient benefit from world-class care was nothing less than riveting.


BCH, Boston City Hospital; CYP, cytochrome P450; ECM, extracellular matrix; SFGH, San Francisco General Hospital; UCSF, University of California San Francisco.

Discovering Research and Finding a Focus

Then as now, a resident's schedule allowed little time for exploring career options and contemplating the future. Fortunately I had enjoyed a decisive research experience in medical school, so came to the residency with an academic career in mind. Had the decision been left to a later stage in my training, my path might have been different. I entered Harvard Medical School with no notion of my future in medicine. I had studied English literature in college, with just the minimum of premedical course work. I had not spent my summers in a research laboratory or my evenings volunteering in an emergency room. Medicine appealed to me because of its plethora of choices, which I assumed improved the odds of finding a career that was both productive and rewarding. The process of choosing started with research, the summer after my first year. Thinking that this could be combined with travel, I arranged through a Harvard connection to join a group at a West German university. Unhappily, the research never materialized. I was loosely attached to a project but had no assigned preceptor and was not expected or invited to use my own hands in the lab. The rewards were those of travel: I was in Berlin 2 years after the wall went up and just a month after Kennedy's historic visit of June 1963. It was an astonishing and unforgettable view of the Cold War faceoff.

During the second year of medical school, lectures on virology and molecular biology made an impression, and I began attending a research journal club that met evenings in the Department of Bacteriology and Immunology (as it was then known). I became familiar with faculty in the department and spoke to the chair, Bernard Davis, about doing another summer fellowship, this time at Harvard. He was encouraging and pointed me to Julian Davies, a young researcher who had arrived 2 years earlier and was already a rising star. Julian had just published an elegant article showing that aminoglycoside antibiotics bind to the ribosome causing misreading of messenger RNA in a cell-free system.1 I became his first student and was charged with validating the misreading hypothesis in intact bacteria (“in vivo”). With Julian's patient tutoring, I quickly learned bacterial culture and a few analytical techniques. I was in the lab at all hours, in and around an earning job at a blood bank, and within a few weeks had an initial positive result. Julian felt that with some additional months of work, I could have a publication. I obtained permission from Harvard to insert a research year between the second and third years of medical school. In addition to the original project, a number of exploratory studies were done involving bacteriophage and cell-free translation. If many were negative, they did engender confidence in my ability to design, perform, and interpret experiments. I finished the year with a short article in one of the top journals of the day.2

On my return to medical school, I toured the clinical specialties and quickly settled on internal medicine. The fourth year at that time included a 5-month elective block. Rather than doing the usual series of 1-month clinical electives, I chose research. I was already thinking about the liver, intrigued by its metabolic complexity. Also, the patients I had seen with liver disease, while doing fourth year medicine at BCH, were a lesson in how little we knew about this large organ. It seemed an area badly in need of study and therefore an opportunity. To test the reality of liver research, I undertook a project on hepatic sterol regulation. Charlie Davidson became my sponsor, generously organizing lab space at the Thorndike for my use in the evenings. Charlie also arranged for me to join a field project headed by Elliott Alpert, which allowed me to do a study with human liver. The results were published,3 but they were less important than the experience of meeting liver research for the first time and enjoying it. Thus, I graduated from medical school with a direction in mind.

The Cell Biology of the Liver

After the intense schedule of the first 2 years of residency, the third year had the feel of diastole. My clinical duties consisted of ward consults, which seldom occupied a full day. Most weekends were free. Given time, I again entered research with the specific goal of culturing hepatocytes. Cell culture was widely perceived as a key tool for probing cell biology. It was direct, and it was ideal for isotopic labeling methods. People had theorized also that hepatocytes should adapt easily to culture because of their latent capacity for proliferation, as seen after partial hepatectomy. Development of liver cell lines indeed had been accomplished by several groups starting from normal human liver, rodent liver, or cancers. In some cases, the cells were morphologically similar to hepatocytes under the light microscope. However, none maintained the functional profile of native liver (for example, production of albumin or urea). In fact, the effort had been so negative that, by the late 1960s, culture of “normal” hepatocytes was deemed technically impossible—for unclear reasons.

I postulated that primary cells in culture, coming directly from the native liver, would retain specific function to a degree not seen in liver cell lines. Moreover, this approach seemed practical in that the liver is a large organ. Its dispersion should yield an abundance of primary cells, allowing study without culture-based expansion. The only cell culture facilities at the Thorndike were in the infectious diseases lab of Jerry Tilles, who worked with chick embryo primary cells for purposes of virus cultivation. He generously agreed to help me with hepatocytes. My initial effort utilized fetal human liver because the material was at hand, courtesy of a pathologist who was salvaging products of abortion for investigational use. The methods for dispersing the chick embryo proved applicable to fetal liver. We quickly found that the partially purified hepatocytes took readily to a variety of culture conditions. That led to exploration of different substrata and culture media. We also developed assays for several liver-specific functions. Cultures were made growing or stationary by including or excluding serum from the medium. Over a period of days to weeks in culture, the stationary cells reached higher levels of differentiation and were functionally active much longer than were serum-treated proliferating cells.4 We speculated that the work of cell division may require sacrifice of liver-specific function.

To San Francisco

The unique training at the Thorndike and BCH had cemented my interest in academic medicine, and I was asked to spend another year as chief resident. The Thorndike's tradition was alluring. The lab had supported and trained a host of academic superstars including a Nobel prize winner, the hematologist George Minot. In liver disease alone, a generation of leaders had absorbed its influence as trainees or staff including Irwin Arias, Tom Chalmers, Charles Davidson, Bob Glickman, Robert Kark, Charles Lieber, Bob Ockner, Don Ostrow, Irwin Rosenberg, Steven Schenker, Rudi Schmid, and Charles Trey, among others. However, change was in the air. There was talk of merging the individual services at BCH into one. The city had asked the three participating medical schools—Boston University, Harvard, and Tufts—to submit proposals. Even at this early stage, many were betting that Boston University would get the nod, its medical center being just across the street from BCH. Although this was unsettling, I had other reasons to look around. Most importantly, it seemed obvious that experimental models would be required for illuminating the black box of the liver, and this type of research was not the Thorndike's strength. So I declined the offer of chief residency. When people asked about my plans and heard San Francisco, several remembered Rudi Schmid, who had been at the Thorndike from 1959-1962. The consensus was that he was building an exciting program at the University of California San Francisco (UCSF). That was good enough, and I sent an inquiry along with one or two letters of reference. A few days later, Rudi was on the phone offering me a research fellowship. He also wanted to make it clear that the position was for research only; further clinical training was out of the question. It was exactly what I wanted to hear. I thanked him and accepted.

Three years later, in 1973, the city of Boston moved to a single affiliation, making the entire BCH a satellite of Boston University Medical Center. Harvard, its involvement finished, moved the Thorndike to the Beth-Israel Hospital within the Harvard complex on Longwood Avenue. While the transfer was a good-faith effort to preserve the Thorndike Memorial Laboratory tradition, it failed. The roots of this surprising plant turned out to be in BCH, not Harvard, and in the diseases of the inner city poor.

The Two-Career Couple

By the time of the move to San Francisco, I had been married for three years to Mina, my extraordinary wife. While I did my residency, she completed a Ph.D. in bacterial physiology. When I proposed San Francisco as our next destination, she agreed and applied for postdoctoral positions at University of California Berkeley. An offer came from a well-known lab, which may not have been her dream but was a place to start. Because San Francisco had been my choice, we had an informal understanding that she would take the lead in the next move. The conventional wisdom in academic biomedical circles in the United States. has always been that, to move up, you have to move out. The frequency of moving (or at least bargaining with an offer to move) varies but may be on the order of every 5-10 years. Our ground rule in considering offers to move was that a long-distance relationship was not acceptable. As it happened, rather than leave the Bay Area, we moved within Berkeley and San Francisco, respectively. Mina went from University of California Berkeley to Lawrence Berkeley National Laboratory, where she found strong support and unique facilities and became recognized internationally for work in mammary cell biology and breast cancer. I finished my research training and accepted a faculty position at the UCSF Parnassus (main) campus. After 7 years, I moved to the San Francisco General Hospital (SFGH) to create a new program, staying 20 years (more about this, below). I then returned to the Parnassus site as chief of the Gastroenterology Division, a position I held for 12 years until stepping down in 2009. Our experience may not be typical of two-career couples, but it illustrates one way to cope, for those who live in an area that is well-endowed with biomedical research.

The challenges for a marriage of two people in biomedical research of course involve more than working out a way to live in the same locale, in jobs that are compatible with two different and highly specialized careers. Additional, if more generic, issues relate to hours that are often irregular, the imperative of deadlines (grant submissions, letters of recommendation, reviews, ad infinitum), and sometimes frequent travel for one or both partners. While this would seem to leave little space for maneuvering, room must be made also for home commitments (Table 1). Otherwise the marriage itself is threatened. When Mina visits other programs, she often fields questions on balancing family and career, mainly from female students and junior faculty. It's interesting that I am rarely asked the same questions, even though nowadays men in these arrangements typically share the lifting at home. Perhaps cultural change is extremely slow. Alternatively, “postmodernism” may have spread to the point of bringing back traditional spousal roles. Our younger colleagues who ask these questions may forget that at their age we were watching Gloria Steinem, Betty Friedan, the National Organization for Women and, later, the debate over Title IX (equal access to educational opportunities and activities). We weren't among the marchers, but we heard the arguments and strongly supported the goal of equal rights for women.

Table 1. Five Home Rules for the Two-Career Couple
  • 1Share the shopping and cooking; try cooking together.
  • 2Have dinner as a family (or couple) at least 5 days per week.
  • 3Take dedicated vacations. Meetings in Hawaii can be fun for the accompanying spouse, but they don't count as vacation.
  • 4If you want to have a family, just do it. There is no perfect time. Get childcare, and ignore the expense.
  • 5Spend a few minutes every day reading something completely unrelated to work.

UCSF and Primary Adult Hepatocyte Culture

Rudi Schmid told me that I would be working on heme oxygenase. I arrived just as he, Raimo Tenhunen, and Harvey Marver published the first description of this enzyme. It was Rudi's seminal discovery, settling two decades of debate on the physiological mechanism for the conversion of heme to bile pigment.5 Although I knew nothing about heme oxygenase, I felt that working on it would give me at least a good grounding in applied chemistry and enzymology. I decided to add it to my interest in cell biology, proposing to characterize the enzyme and its regulation in specific cell types of the liver. Rudi was enthusiastic, and my initial paper from San Francisco described the activity in isolated Kupffer cells and its stimulation by erythrocyte phagocytosis.6

On the side, I took up hepatocyte culture again. Access to human liver in San Francisco was very limited, so I needed an animal model. Michael Berry recently had published his improved method for dissociation of rat liver by in situ collagenase perfusion.7 Moreover, he was then working at UCSF directly upstairs from us. With one or two visits to his lab, I was obtaining good quality isolates of adult rat hepatocytes. When I asked about culturing the cells, I was told that this had not worked, for unknown reasons. Therefore in my initial studies, I used short-term suspension culture as others had been doing. The cells were in a complete medium with serum and incubated at 37°C in an Erlenmeyer flask in a rotary-motion shaker. I noticed that within 30 minutes, a visible ring of material accumulated on the glass at the air/liquid interface. Curious, I examined the area with an inverted microscope and discovered attached and spread hepatocytes—clear evidence that the cells were capable of attaching to a substratum, even untreated glass. I put part of the next batch of cells into culture plates, initially with the standard volume of nutrient medium. The cells settled but remained round and unattached. Recalling then that cell attachment in the rotating flask had occurred only at the liquid/air interface, I reasoned that the depth of the culture medium might be critical and prepared plates with half of the usual volume of medium. Within minutes the cells dropped to the plastic surface, attached and spread, forming a carpet on the bottom of the dish. We did a functional and morphological characterization. We also found, as in the studies of human fetal liver, that function was most stable in high-density, nonproliferating culture. We published a pair of papers describing the new system.8, 9 On the draft manuscript, I listed Rudi as senior author, by way of acknowledging his support and encouragement. However, he declined coauthorship, saying that the work had been conceived and performed by me. Moreover, he felt that these papers could establish my name, which meant that they had to be independent of him. He went on to arrange for a special lecture on the new culture system at the 1973 American Association for the Study of Liver Diseases (AASLD) annual meeting, where I spoke alone for 30 minutes. Such was possible when the entire 2-day program consisted of about 50 papers, all presented orally in a single room.

Mentors and Mentoring

I knew nothing about Rudi Schmid or his mentoring style prior to arriving in San Francisco. I had not sought out his old colleagues or called current trainees. In fact, it never occurred to me to investigate. I felt that my prior experience had told me what mentors should be. Julian Davies and Jerry Tilles were two very different personalities but shared a commitment to research, deep expertise in their respective areas down to the fine details of technique and process, and a generous spirit. I expected nothing less from Rudi Schmid. This is not to say that I expected perfection. The mentee-mentor relationship, if successful, is close. Tics and flaws are included. The latter can be fun or disappointing but are not a problem if separable from the core relationship. I was aware that my blind move carried some risk, but I told myself that if Rudi didn't measure up, I would leave. It goes without saying that Rudi more than met expectations. I even learned to enjoy his flaws.

Don Ostrow and I have written about Rudi's personality and mentoring style.10 I don't believe that he applied a concept of mentoring, or if he did, it was not one that would find acceptance today. He could be demanding to the point of boorishness. He habitually indulged in snap judgments, which were wrong as often as right. By today's standards he could be politically incorrect, enjoying it too. Negatives such as these, however, were easily outweighed by his great virtue, which was an abiding passion for science that is well-done and well-written, a passion that energized all those around him. If he had been challenged on his questionable behavior, he likely would have justified it as necessary for transforming a raw trainee—this crude clay—into a passably skilled practitioner of the noble art of medical research.

Mentoring today is a formal discipline which is analyzed, taught, and promoted by all major medical schools. It has specialists, programs, and institutional mandates, which are meant to ensure that students, new faculty, and anyone else in need of mentoring has the opportunity to be mentored. The scope is ambitious, encompassing time-management, academic advancement, grant strategy and writing, publication goals, and personal life issues. If the current programs have mixed success in bringing mentoring to all, at least they do remind everyone that effective mentoring is important, and they promulgate a code of conduct for mentors. Coming through well before such programs, I was fortunate in having three successive mentors who were intelligent, engaged, extremely generous with their time, and fundamentally honorable. Back then, there was no code. We assumed that the standards of polite society would hold also in the laboratory. Bad jokes and wacky behavior were fine (and frequent). Exploitation, dishonesty, and verbal or physical abuse were not. Formalizing the mentor's role has served notice that bad behavior will not be tolerated, and it empowers students to take action if they suffer because of it. That said, what I learned as a mentee 40 years ago remains valid: At the core of successful mentoring are common interests (research or clinical), shared goals, and mutual respect.

Understanding Hepatic Fibrosis

I wanted to use the new culture system to probe a specific area of hepatocyte physiology or disease. Drug metabolism/hepatotoxicity was, and remains, an area in which cell culture has potential to be extremely useful. Cytochrome P450 (CYP) was central to this process, and the lab already had the tools for its measurement. We began examining CYP in culture and quickly made the surprising discovery that total CYP is labile in hepatocytes. During the first 24 hours of culture, the amount determined by spectrophotometry declined by 80%-85%. We first assured ourselves that this was not due to diminishing cell viability. Synthetic processes such as albumin production were well-maintained. Moreover, the change in total CYP did not involve all subclasses but only the major barbiturate-inducible CYP. At the time, we did not seek in vivo evidence for accelerated CYP turnover, although others eventually showed that similar changes occur in regenerating liver.11

Around this time, Phil Guzelian joined the lab, and we began a collaborative effort to understand culture-related CYP loss. We postulated that the cells' environment was deficient. Because we had no clue as to the number of factors or their nature, we considered both soluble components (hormones, metabolic substrates) and insoluble (extracellular matrix). Although we hoped that a single critical variable would emerge, the reality was that several factors provided partial support of total CYP. An interesting one was the culture substratum, which I had explored in my previous studies of human fetal hepatocytes.4 When the cells were plated on a collagen gel, there was a small but significant increase in total CYP. George Michalopoulos took this further in studies of hepatocytes on a floating collagen raft.12 Thinking about the in vivo implications of these findings, we were reminded of older work in human liver, which postulated that “capillarization” or collagenization of the space of Disse was highly associated with loss of the hepatocyte microvillus border and severe liver dysfunction.13 Somewhat later, the Rennes (France) group published a culture system consisting of hepatocytes in coculture with hepatic nonparenchymal cells. Several liver-specific functions were better maintained in this system than in pure hepatocyte culture. As for mechanism, they showed that the two cell types grew as individual small islands until they made contact, then elaborated a matrix at the contact interface. The matrix appeared to be essential for maintenance of specific function.14 Over time, we began to view the functional change of hepatocytes on culture plastic as an exaggerated injury response. By implication, the extracellular matrix (ECM) could be a key determinant of hepatocyte function, with injury-associated changes leading directly to impaired hepatocellular function.

In biomedical investigation, every experiment holds potential for moving the focus of study in a new direction. This is something I have found to be among the joys of research, because it lends coherence to an activity that might otherwise look like random movement. At this time, however, I had to make a calculated decision. The work with primary hepatocyte culture was leading on the one hand to detailed studies of CYP regulation and, on the other, to a focus on cell-ECM interaction. The two areas were so divergent, as well as individually so large, that I felt I needed to choose. The CYP field was increasingly focused on the cloning and characterization of new subclasses of CYPs. This area struck me as dry and not necessarily clinically relevant. So I was leaning toward ECM and fibrosis. Then I was offered leadership of a new laboratory at UCSF's SFGH site. The Gastroenterology Division had received a new endowment for research on diseases of the liver and biliary tract. Given a mandate to initiate a novel area of research, I felt that hepatic matrix biology and fibrosis had to be the choice. Others, including bioengineers, continued the effort to stabilize the drug-metabolizing function of primary hepatocytes in culture. The problem remains,15 reflecting perhaps the peculiar geometry of blood flow and bile excretion in the liver. Over the years, others have demonstrated that most if not all epithelia undergo phenotypic change in primary culture. The concept figures into current strategies for cell-based therapy, which envisions expansion of stem or progenitor cells, not adult hepatocytes, in culture, then maturation of the expanded cells using specific factors.

My first goal in organizing the new lab was recruiting a faculty member with expertise in matrix biochemistry and preferably a background in liver research. Joe Roll's name came up. He was completing his liver training at Yale. The time had included basic research in the lab of Heinz Furthmayr where he learned to purify matrix proteins and raise antibodies for tissue immunohistochemistry. Collagens were reputed, with reason, to be very difficult proteins, and Joe had mastered the methods. Fortunately, he agreed to come to San Francisco. He turned out not only to know ECM but to be a patient teacher and great collaborator. He had an essential role in the success of the program that was launched in 1981.

The offer to move to SFGH was enticing also because it came with responsibility for designing new lab space. Although I had never done this, I had a view of what worked and did not. Prior to San Francisco, the labs I had known were typical of the day: a set of rooms, each sized for two to six people, and spaced along a hallway. Groups thus were segregated and often behind closed doors, and spent little time communicating with each other. They were expected to be independent and even encouraged to pursue divergent research areas. By the 1970s, however, multidisciplinary groups were organizing around a single biological or disease focus. They required space that would foster interaction, hence, a lab with benches grouped together in a large open area. Individual rooms were used for equipment, special-purpose activities such as cell culture, and offices. Rudi emulated this by recruiting Ph.D.s to his group (Tony McDonagh, Almira Correia) and using an open design for the largest part of the lab, which housed the heme and bile pigment research. This was the most creative and exciting environment I had known. While I credited Rudi, I also found the space to be important. Our space for renovation at SFGH had been built as a traditional Victorian ward. The outer walls formed an extended rectangle, about six times as long as it was wide. Other floors of the same building had been converted to laboratory space, all designed with a long hallway down one side, off of which ran a series of individual labs. Fortunately, the architect working with me was willing to hear other ideas, and in the end we had an attractive open lab with room for four investigator groups.

The program began to attract the best trainees in the division as well as outstanding people from abroad. Scott Friedman came, developing the isolation and primary culture of stellate cells and initiating study of the factors controlling cell behavior.16 Jackie Maher applied new molecular methods to quantify cell-specific expression of matrix proteins in vivo.17 Joe Roll and I developed a model of hepatocytes cultured on a basement membrane gel (EHS gel; also known as Matrigel), proving that ECM composition is critical to hepatocyte-specific function.18 Mick Arthur used primary cells to examine matrix proteinase production.19 Rolf Hultcrantz probed the basis for ethanol-related hepatic inflammation.20 By 1989, we had a scheme for stellate cell activation and fibrosis (Fig. 1). We also organized an AASLD single-topic conference on “Connective Tissue Biology in the Liver”.21 Don Rockey and Chantal Housset followed with studies of stellate cell contractility.22

Figure 1.

Cellular changes attributable to alteration of the perisinusoidal matrix. These include the conversion of resting lipocytes to collagen-producing myofibroblast-like cells; loss of endothelial cell fenestrae, and loss of the hepatocyte microvillus membrane. Taken together, this is what Hans Popper termed “capillarization of the sinusoid”. “Lipocyte” was the term we used prior to an agreement among the groups studying this area to use “stellate cell” in place of all other names including “fat-storing cells” and “Ito cells”. (Figure reproduced from Bissell DM. Progress in Liver Diseases, Vol. IX. Philadelphia: W. B. Saunders Co.; 1990, p. 151.)

In the 1990s, fibrosis research expanded rapidly, both in the United States and abroad, particularly in Europe, Japan, and Australia. This was gratifying to me and no doubt also to those like Marcos Rojkind who had pioneered the area. Stellate cell studies flourished, but we also began to move beyond stellate cells. Bill Jarnagin was a surgical resident with a surgeon's ability to work long hours. He and I, with valuable help from talented assistant Frank Wang, found that sinusoidal endothelial cells respond very rapidly to injury with elaboration of EIIIA (EDA) “cellular” fibronectin, thereby contributing to the activation of stellate cells.23 Jacob George extended this work, looking at the role of transforming growth factor-beta in the regulation of endothelial cell fibronectin expression.24 Most recently, a talented UCSF medical student, Bruce Wang, and I looked at biliary fibrogenesis in the mouse, finding that the acute cellular response does not involve stellate cells but portal-based myofibroblasts, which likely arise from biliary epithelial cells in an epithelial-mesenchymal transition.25 The lesson is that the liver, like other epithelia, responds globally to injury. Stellate cells may the single most important source of matrix protein following injury but are not autonomous. They act in concert with, and are regulated by, the rest of the tissue as well as by the immune system and inflammatory cells.26

This capsule of a 20-year period is necessarily brief and cannot adequately represent the many students and technicians who contributed so importantly. I will add only that one of the constant pleasures has been seeing laboratory members enjoy success and watching more than a few rise to positions of leadership in the area of liver injury and fibrosis.

The Promise of Fibrosis Treatment

Our work with model systems was motivated by the hope that, by defining the regulation of fibrogenesis, within a few years we would see development of new therapeutics for chronic fibrosis. The experimental systems have been very productive, so much so that there may be a surfeit of potential targets for therapeutic molecules. For a recent review, see Friedman.27 Despite the substantial knowledge base, new treatments have been slow to emerge. This is not for lack of interest by the pharmaceutical industry. Rather, it is due to concern with the difficulty and time-frame of achieving treatment endpoints acceptable to licensing authorities such as the U.S. Food and Drug Administration (FDA). The issue is being actively discussed, with the result that the FDA now recognizes fibrosis reduction per se as a valid endpoint and will not require “hard” endpoints such as survival and/or time to transplant, which likely would require years to obtain.28 That is progress, but it is now up to investigators to address the need for an accurate and convenient fibrosis test. Liver biopsy with histochemical scoring is quantitative but subject to error because of sampling variation. As an invasive procedure, it is also unpopular with patients and institutional review boards. Serum tests and imaging methods, such as echo-elastography, to date have not shown the discriminatory power needed for monitoring a response to treatment over a relatively short time period. The field is active, however, and progress can be anticipated. With a noninvasive, quantitatively accurate method for fibrosis assessment, trials as short as 6-12 months can be envisioned. Among the candidate agents for fibrosis therapy are medications that have been already approved for another indication. These include interferon-gamma, angiotensin receptor antagonists, endothelin antagonists, and peroxisome proliferator-activated receptor-gamma agonists. As approved drugs, they have a known safety profile. The difficulty is in funding trials. The company that owns the drug, aware that its patent will run out, may have little reason to support a trial for a new indication. Other kinds of funding, including federal, may be necessary.

Training Physician-Scientists

With a few exceptions, the members of the SFGH lab (fellows and faculty) were trained as physicians, and medical training guided their research interests to a degree. Nonetheless, one of my top priorities was making sure that clinical effort occupied no more than 20% of their time. I felt that they needed at least 80% of their time protected for research, to be competitive for funding and find success. The 80% mark was attainable at SFGH, in part because the faculty volunteered their clinical effort. By this means, they retained some control over the size and scheduling of clinical commitments. Of course, this meant that they were paid by the lab for clinical time, which fortunately was possible with our endowment income. Also, documentation of patient encounters at that time was minimal and easily accomplished on the job, without after-hours work. The result was that ward or clinic attending was no more than a hiatus of a few hours in what was otherwise a research day. This is not to say that we shunned clinical duties. Trained as physicians, we all enjoyed patient care and drew inspiration from it. By limiting its amount, however, the focus on research was nearly total. People may ask if 80% research is necessary. Trainees in patient-oriented research in many settings have only 50% protected research time. My answer is that, as a division chief, I have seen people with a 50:50 split struggle and fail in their research. The clinical piece inevitably takes priority, is very difficult to contain, and in the end borrows time that was meant for research. Those who succeed have exceptional time-management skills and projects sized to a 50% commitment. As already noted, I conclude that for most people 80% is required for keeping a steady focus on research. Clinical activity at the level of 20%, or one day per week, is manageable and should not encroach on research time.

Other precepts that I found useful for the SFGH group were taken from Fuller Albright's presidential address to the American Society for Clinical Investigation.29 Albright was a pre-eminent clinical investigator, 1930-1950, who became president of the American Society for Clinical Investigation in 1944. The following points (Table 2) were incorporated into a certificate that went to graduates of our program. Some such as number 3—Do not be too ambitious—merit elaboration. At one level, it is advice to take care that one's next experiment is appropriate to the resources and time at hand. Too ambitious a target may result only in failure and discouragement. It could also be a warning about accepting administrative responsibility too early in one's career (in fact, another “Do Not” from Albright was: Do not show too much executive ability). The first faculty appointment for a young physician investigator should feature unfettered research time, with minimal or no administrative effort or committee responsibilities. This is a golden period, which potentially lasts 15 years or longer, during which the individual consolidates expertise, then gathers a group and makes a mark on the field. To truncate the time by assuming a major administrative post—division chief or department head—is to waste a unique opportunity.

Table 2. Ten Rules for the Clinical Investigator
  • 1Do be inquisitive.
  • 2Do be ambitious.
  • 3Do not be too ambitious.
  • 4Do measure something.
  • 5Do not jump at the first problem that presents itself.
  • 6If possible, do arrange your data in graphic form.
  • 7Do not be a lone wolf or secretive.
  • 8Do develop a theory.
  • 9Do not be a slave to your theory.
  • 10Do reserve some time each day for unadulterated thinking.

The Physician-Scientist and the Future

To some observers of academic medicine, the American physician-scientist has joined the northern spotted owl: Endangered, if not becoming extinct. Indeed, the data show a sharp decline over the past 15 years in both the number of research applications to the National Institutes of Health (NIH) and success rates for M.D.s or M.D./Ph.D.s. Concurrently, the average age of first-time awardees has risen into the mid-40s. Some would say that this simply reflects the evolution of scientific inquiry. Ph.D. scientists now dominate because they have the tools to cross over into disease research. Others blame the NIH review process.

Returning to the endangered species metaphor, I think it is largely a habitat issue. From 1960-1990, the academic center was fertile ground for physician-investigators precisely because support of nearly full-time research was possible. If a gap in grant funding occurred, it was covered by the department, generally out of excess clinical earnings. The scene has changed profoundly in recent years, mainly because clinical income has dwindled to the point that it barely covers expenses. The latter is due to a combination of declining reimbursement and ever-increasing costs related to documentation requirements imposed by both private insurance and Medicare. The result for academic physicians is pressure to generate clinical income, and the pressure has become pervasive. No one is spared. The clinical effort of young physician-scientists often goes well beyond the optimal 20%, because it may be the only available means of covering a gap in salary funding. Career-development awards stipulate 75% of time must be spent for research, but many recipients are in violation of their contract, because they borrow research time for clinical service. Too often, the borrowing results in insufficient research progress and failure to obtain follow-on funding. Physician-mentors also are affected, so burdened by clinical work that they have little time to function as partners and collaborators. This is barren ground for the aspiring physician-scientist. The remedies are obvious, if not necessarily easy (Table 3).

Table 3. Five Rules for Physician-Scientist Success
  • 1Guarantee young physician-scientists that they will spend no more than 20% of their time—one day per week—on any aspect of clinical care (this includes both direct patient care on-call time and documentation).
  • 2Make it institutional policy that any shortfall in clinical earnings for designated physician-scientists will be covered from nonresearch funds.
  • 3Require Federal agencies and/or foundations to fully fund career-development awards, based on institutional salary scales. If cofunding is understood, the institution must guarantee that it has the funds.
  • 4Enact health-care reform! Apart from its irrationality (e.g., 40 million uninsured), the current system in the United States imposes large administrative costs on academic medical centers, consuming funds that could support research training.
  • 5Actively promote philanthropy. Educate patients and the public on the importance of private endowments for research training and development.

As for the future, the only constant is change. Places and programs that resist change are choosing obsolescence. I became division chief at UCSF in 1997, fifteen years after Rudi Schmid left the division to be Dean of the School of Medicine. Remarkably, his imprint was still apparent, although by then it indicated mainly that restructuring was overdue. The research universe and clinical enterprise had progressed in ways that could not have been imagined 20 years prior. To be the agent of change for the program that had nurtured my own development, was rewarding but with poignant moments. It was reassuring that Rudi himself seemed to approve.

We have come a long way from studies on sterol metabolism in cultured fibroblasts or liver slices: “statin” is now a household word. The genome project has redefined living organisms. While the latter is impressive, it has also shown us how vast is the biological universe yet to be discovered, served by those several thousand genes. Clearly, when it comes to biomedical discovery, we are just at the end of the beginning. The opportunities are enormous. Moreover, physician-scientists will be vital to this enterprise, because they bring a unique perspective based on their knowledge of medicine and dedication to patients.

Change will occur, but it is worth remembering that fundamental change tends to be slow, while marginal change is fast. People will focus on the fast because it is current. Interest in clinical gastroenterology may rise and fall with reimbursement rates. However, it is important not to lose sight of change that is slow. Every year for the past half-century, some of the best students aspire to careers in academic medicine, drawn because they are passionate about research, love to teach, or have new ideas for contributing directly to the advancement of health. Those who become good at what they do will have a place in any future system. Mentors, the institutional leadership, and funders must ensure that these students are encouraged and that they receive training that is intensive, complete, and free of distractions. If those goals are adhered to, we have every reason to be optimistic about the future of academic medicine and the country's healthcare.