The China Study (11 page)

63
TURNING OFF CANCER
community argued for HBV No one in either community dared to sug-
gest that nutrition had anything to do with this disease.
We wanted to know about the effect of casein on HBV-induced liver
cancer in mice. This was a big step. It went beyond aflatoxin as a carcino-
gen and rats as a species. A brilliant young graduate student from China in
my group, Jifan Hu, initiated studies to answer this question and was later
joined by Dr. Zhiqiang Cheng. We needed a colony of these transgenic
mice. There were two such "breeds" of mice, one living in LaJolla, Califor-
nia, the other in Rockville, Maryland. Each strain had a different piece of
HBV gene stuck in the genes of their livers, and each was therefore highly
prone to liver cancer. I contacted the responsible researchers and inquired
about their helping us to establish our own mouse colony. Both research
groups asked what we wanted to do and both were inclined to think that
studying the protein effect was foolish. I also sought a research grant to
study this question and it was rejected. The reviewers did nol take kindly
to the idea of a nutritional effect on a virus-induced cancer, especially of
a dietary protein effect. I was beginning to wonder: was I now being too
explicit in questioning the mythical health value of protein? The reviews
of the grant proposal certainly indicated this possibility.
We eventually obtained funding, did the study on both strains of
mice and got essentially the same result as we did with the rats.47 , 4 8 You can
see the results for yourself. The adjoining picture (Chart 3.1047 ) shows
what a cross-section of the mouse livers looks like under a microscope.
The dark-colored material is indicative of cancer development (ignore
the "hole"; that's only a cross-section of a vein). There is intense early
cancer formation in the 22% casein animals (D), much less in the 14%
casein animals (C), and none in the 6% casein animals (B) ; the remain-
ing picture (A) shows a liver having no virus gene (the control).
The adjoining graph (Chart 3.1147) shows the expression (activity) of
two HBV genes that cause cancer inserted in the mouse liver. Both the
picture and the graph show the same thing: the 22% casein diet turned
on expression of the viral gene to cause cancer, whereas the 6% casein
diet showed almost no such activity.
By this time, we had more than enough information to conclude that
casein, that sacred protein of cow's milk, dramatically promotes liver
cancer in:
• rats dosed with aflatoxin
• mice infected with HBV

64                            THE CHINA STUDY
CHART 3.10: DIETARY PROTEIN EFFECT ON GENETICALLY-BASED
(HBV) LIVER CANCER (MICE)
Non-transgenic Mice
Transgenic Mice with
(Control) with 22%
6% Casein Diet
Casein Diet
Transgenic Mice with
14% Casein Diet
CHART 3.11: DIETARY PROTEIN EFFECT ON GENE EXPRESSION (MICE)
14
12
§10
~
6
·Vi
Vl
8
~
0.
><                                                         •   14
~6
022
c
(l)
4
l:J
2
-Je=---
o
HBV Gene B
HBV Gene A

65
TURNING OFF CANCER
Not only are these effects substantial, but we also discovered a net-
w o r k of complementary ways by which they worked.
Next question: can we generalize these findings to other cancers and
to other carcinogens? At the University of Illinois Medical Center in
Chicago, another research group was working with mammary (breast)
cancer in rats. 49- 51 This research showed that increasing intakes of ca-
sein promoted the development of mammary (breast) cancer. They
found that higher casein intake:
• promotes breast cancer in rats dosed with two experimental car-
cinogens (7,12-dimethybenz(a)anthracene (DBMA) and N-ni-
troso-methylurea (NMU))
• operates through a network of reactions that combine to increase
cancer
• operates through the same female hormone system that operates in
humans
LARGER IMPLICATIONS
An impressively consistent pattern was beginning to emerge. For two
different organs, four different carcinogens and two different species,
casein promotes cancer growth while using a highly integrated system
of mechanisms. It is a powerful, convincing and consistent effect. For
example, casein affects the way cells interact with carcinogens, the way
DNA reacts with carcinogens and the way cancerous cells grow. The
depth and consistency of these findings strongly suggest that they are
relevant for humans, for four reasons. First, rats and humans have an
almost identical need for protein. Second, protein operates in humans
virtually the same way it does in rats. Third, the level of protein intake
causing tumor growth is the same level that humans consume. And
fourth, in both rodents and humans the initiation stage is far less im-
p o r t a n t than the promotion stage of cancer. This is because we are very
likely "dosed" with a certain amount of carcinogens in our everyday
lives, but whether they lead to full tumors depends on their promotion,
or lack thereof.
Even though I became convinced that increasing casein intake pro-
motes cancer, 1 still had to be wary of generalizing too much. This was an
exceptionally provocative finding that drew fierce skepticism. But these
findings nonetheless were a hint of things to come. I wanted to broaden
my evidence still more. What effect did other nutrients have on can-

THE CHINA STUDY
66
cer, and how did they interact with different carcinogens and different
organs? Might the effects of other nutrients, carcinogens or organs
cancel each other, or might there be consistency of effect for nutrients
within certain types of food? Would promotion continue to be revers-
ible? If so, cancer might be readily controlled, even reversed, simply by
decreasing the intakes of the promoting nutrients and/or increasing the
intakes of the anti-promoting nutrients.
We initiated more studies using several different nutrients, including
fish protein, dietary fats and the antioxidants known as carotenoids.
A couple of excellent graduate students of mine, Tom O'Connor and
You ping He, measured the ability of these nutrients to affect liver and
pancreatic cancer. The results of these, and many other studies, showed
nutrition to be far more important in controlling cancer promotion than the
dose of the initiating carcinogen. The idea that nutrients primarily affect
tumor development during promotion was beginning to appear to be a
general property of nutrition and cancer relationships. The Journal of
the National Cancer Institute, which is the official publication of the U.s.
National Cancer Institute, took note of these studies and featured some
of our findings on its cover. 52
Furthermore, a pattern was beginning to emerge: nutrients from ani-
mal-based foods increased tumor development while nutrients from plant-
based foods decreased tumor development. In our large lifetime study of
rats with aflatoxin-induced tumors, the pattern was consistent. In mice
with hepatitis B virus-altered genes, the pattern was consistent. In stud-
ies done by another research group, with breast cancer and different car-
cinogens, the pattern was consistent. In studies of pancreatic cancer and
other nutrients, the pattern was consistent. 52 , 5 3 In studies on carotenoid
antioxidants and cancer initiation, the pattern was consistent. 54,55 From
the first stage of cancer initiation to the second stage of cancer promo-
tion, the pattern was consistent. From one mechanism to another, the
pattern was consistent.
So much consistency was stunningly impressive, but one aspect of
this research demanded that we remain cautious: all this evidence was
gathered in experimental animal studies. Although there are strong argu-
m e n t s that these provocative findings are qualitatively relevant to human
health, we cannot know the quantitative relevance, In other words, are
these principles regarding animal protein and cancer critically important
for all humans in all situations, or are they merely marginally important
for a minority of people in fairly unique situations? Are these prin-

TURNING OFF CANCER                             67
ciples involved in one thousand human cancers every year, one million
human cancers every year, or more? We need direct evidence from hu-
m a n research. Ideally, this evidence would be gathered with rigorous
methodology and would investigate dietary patterns comprehensively,
using large numbers of people who had similar lifestyles, similar genetic
backgrounds, and yet had widely varying incidences of disease.
Having the opportunity to do such a study is rare, at best, but by
incredibly good luck we were given exactly the opportunity we needed.
In 1980 I had the good fortune of welcoming in my laboratory a most
personable and professional scientist from mainland China, Dr. Junshi
Chen. With this remarkable man, opportunities arose to search for some
larger truths. We were given the chance to do a human study that would
take all of these principles we had begun to uncover in the lab to the
next level. It was time to study the role of nutrition, lifestyle and disease
in the most comprehensive manner ever undertaken in the history of
medicine. We were on to the China Study.

.......................................... . . . . . . . . . . . 4.............
Lessons from China
A SNAPSHOT IN TIME
Have you ever had the sensation of wanting to permanently capture a
moment? Such moments can grip you in a way you will never forget.
For some people, those moments center on family, close friends or re-
lated activities; for others those moments may center on nature, spiri-
tuality or religion. For most of us, I suspect, it can be a little of each.
They become the personal moments, both happy and sad, which define
our memories. It's these moments in which everything just "comes
together." They are the snapshots of time that define much of our life
experience.
The value of a snapshot of time is not lost on researchers either. We
construct experiments, hoping to preserve and analyze the specific de-
tails of a certain moment for years to come. I was fortunate enough to be
privy to such an opportunity in the early 1980s, after a distinguished se-
n i o r scientist from China, Dr. Junshi Chen, came to Cornell to work in
my lab. He was deputy director of China's premier health research labo-
ratory and one of the first handful of Chinese scholars to visit the u.S.
following the establishment of relations between our two countries.
THE CANCER ATLAS
In the early 1970s, the premier of China, Chou EnLai, was dying of
cancer. In the grips of this terminal disease, Premier Chou initiated a
nationwide survey to collect information about a disease that was not
69

70                          THE CHINA STUDY
well understood. It was to be a monumental survey of death rates for
twelve different kinds of cancer for more than 2,400 Chinese counties
and 880 million (96%) of their citizens. The survey was remarkable in
many ways. It involved 650,000 workers, the most ambitious biomedi-
cal research project ever undertaken. The end result of the survey was a
beautiful, color-coded atlas showing where certain types of cancer were
high and where they were almost nonexistent. 1
CHART 4.1: SAMPLE CANCER ATLAS IN CHINA
COLON AND RECTUM (FEMAlE)
-.--
-.
. :--:-:-_-.
--                                                      I
This atlas made it clear that in China cancer was geographically lo-
calized . Some cancers were much more common in some places than
in others. Earlier studies had set the stage for this idea, showing that
cancer incidence also varies widely between different countries.2- 4 But
these China data were more remarkable because the geographic varia-
tions in cancer rates were much greater (Chart 4.2) . They also occurred
in a country where 87% of the population is the same ethnic group, the
Han people.

71
LESSONS FROM CHINA
CHART 4.2. RANGE OF CANCER RATES IN CHINESE COUNTIES
Cancer Site                                 Males                      Females
35-491
35-721
All Cancers
0-75                        0-26
Nasopharynx
1-435                     0-286
Esophagus
6-386                       2-141
Stomach
7-248                        3-67
liver
2-67                        2-61
Colorectal
0-26
3-59
Lung
-                        0-20
Breast
*Age-adjusted death rates, representing # cases/l 00,000 people/year
Why was there such a massive variation in cancer rates among dif-
ferent counties when genetic backgrounds were similar from place to
place? Might it be possible that cancer is largely due to environmentalJ
lifestyle factors, and not genetics? A few prominent scientists had al-
ready reached that conclusion. The authors of a major review on diet
and cancer, prepared for the u.s. Congress in 1981, estimated that ge-
netics only determines about 2-3% oj the total cancer risk. 4
The data behind the China cancer atlas were profound. The coun-
ties with the highest rates of some cancers were more than 100 times
greater than counties with the lowest rates of these cancers. These
are truly remarkable figures. By way of comparison, we in the U.S.
see, at most, two to three times the cancer rates from one part of the
country to another.
In fact, very small and relatively unimportant differences in cancer
rates make big news, big money and big politics. There has been a long-
s t a n d i n g story in my state of New York about the increased rates of
breast cancer in Long Island. Large amounts of money (about $30 mil-
lion5) and years and years of work have been spent examining the issue.
What sorts of rates were causing such a furor? Two counties in Long Is-
l a n d had rates of breast cancer only 10-20% higher than the state aver-
age. This difference was enough to make front-page news, scare people
and move politicians to action. Contrast this with the findings in China
where some parts of the country had cancer rates 100 times (10,000%)
higher than others.