DISCUSSION
We compared the response of liver and adipose tissue lipogenesis to acute and chronic stimulation by carbohydrate ingestion. Hepatic lipogenesis was, in agreement with previous studies (7, 8, 31) clearly responsive to both acute and chronic stimulation while adipose tissue DNL appeared on the contrary poorly responsive. This conclusion is based on several lines of
evidence. First, although oral glucose ingestion in study 2 prevented the decrease in both FAS and SREBP1c mRNA concentrations observed in study 1 there was no increase in these concentrations, as well as in ACC1 mRNA value. This strongly suggests that the expression of lipogenic genes, and presumably the activity of the corresponding proteins, were only
minimally stimulated in adipose tissue by the 12 hour rise in plasma insulin and glucose.
Measurements of proteins amounts and enzyme activity were precluded by the small amount of adipose tissue obtained by needle biopsy. However, there is no shortterm regulation of FAS activity and changes in activity are linked to changes in mRNA level (32). With respect to SREBP1c, the present evidence is that its main regulation is at the transcriptional level and that its activation by proteolytic cleavage is probably a constitutive, nonregulated process (33). The present in vivo result contrasts with in vitro studies showing a stimulation of FAS expression and activity by insulin in human adipocytes (12). Second, FAS, ACC1 and SREBP1c mRNA levels were not increased, but rather decreased, during the high energy high carboydrate diet suggesting strongly that the lipogenic pathway was not stimulated in adipose tissue by carbohydrate overfeeding. Third, these results are consistent with the tracer14 derived estimates of adipose tissue lipogenesis we obtained. The comparaison of deuteriumband 13C enrichments in plasma and adipose tissue TAG clearly shows that adipose tissue lipogenesis was active under the three situations studied. The lack of detectable enrichment in 13C of adipose TAG does not exclude that some uptake of TAGFA occured but shows that this uptake, if present, is low and does not contribute to the deuterium enrichment measured inbadipose TAG. There was, however, no stimulation of lipogenesis in adipose tissue by acute glucose oral load and only a non significant trend for higher lipogenic rate after chronic carbohydrate overfeeding. As a result adipose tissue lipogenesis which was quantitatively comparable to liver lipogenesis in the absence of stimulation (study 1) became less important during either acute or chronic stimulation by carbohydrate feeding. Overall our results agree with those of previous studies which used the incorpation in adipose tissue lipids of 14C from glucose to estimate lipogenesis (4, 5, 15, 34).
We think thus that the stimulation of adipose tissue lipogenesis by massive carbohydrate overfeeding suggested by the study of Aarsland et al (8) is, if real, present only during such extreme, unphysiological situation. The regulation of adipose tissue lipogenesis appears thus different in human beings compared with some other mammalian species. In rats for example adipose tissue lipogenesis is more active and is, as liver lipogenesis, responsive to high insulin/glucose levels and to variations of carbohydrate intake (35, 36). SREBP1c plays a major role in the regulation of lipogenic genes expression, at least in their response to insulin (18, 37, 38). This transcription factor is a major determinant of the lipogenic capacity of mammalian and avian tissues (39). Therefore it is possible that SREBP1c expression is low in human adipose tissue, compared to other species, but such comparison between human beings and other species has not been performed to our knowledge. The lack of response in our study of adipose tissue FAS and ACC1 expression and lipogenesis to acute or chronic carbohydrates overfeeding could be related to the lack of increase of SREBP1c mRNA. Since SREBP1c expression is stimulated in rats by dietary carbohydrates and/or insulin (18, 40, 41) it would remain to determine why this stimulation is absent in human adipose tissue.
The comparison of the quantitative estimate of hepatic and adipose lipogenesis during study 2 with the amount of glucose ingested (less than 1g vs 250270g) shows that the
contribution of this metabolic pathway to the disposal of ingested glucose was minimal. The increase of hepatic and adipose lipogenesis after an hypercaloric high carbohydrate diet (11.5 g/day) was also moderate and can thus explain only a minimal part the weight gain of the
subjects (1500g in average during the two weeks of controled diet). Since liver biopsies for measurement of tracer incorporation in liver TAG were not performed for obvious ethical reasons we cannot exclude the possibility that some newly synthesized fatty acids remained within liver TAG stores. However, it seems very unlikely that an increase in liver TAG stores
could explain a large part of the disposal of ingested glucose in study 2 and of the weight gain observed during the high energy high carbohydrate diet. An increase in lipogenesis in another tissue is unlikely. Therefore the most probable explanations for the observed body weight
increase are merely a repletion of muscles and liver glycogen stores, along with the simultaneous storage of water, and the suppression of fat oxidation leading to the deposition of ingested fat. Moreover, altough the contribution of fat to total energy intake was decreased during the high energy, high carbohydrate diet, the total amount ingested was increased.
In conclusion, our results show that in normal humans adipose tissue lipogenesis, although active, is quantitatively a minor pathway and is less responsive than hepatic lipogensis to acute or prolonged carbohydrate overfeeding. The picture could be different in obesity but the recent finding that lipogenic genes expression is decreased in adipose tissue of obese subjects (11) makes this possibility unprobable. Thus, DNL in adipose tissue is an unlikely contributor to the development of dietary induced obesity in humans.