Journal of Cancer Research and Clinical Oncology

, Volume 130, Issue 4, pp 211–216

Recombinant human erythropoietin attenuates weight loss in a murine cancer cachexia model


    • Department of Internal MedicineOosterschelde Hospital
    • Department of Medical OncologyUniversity Medical Center
  • G. P. A. Bongaerts
    • Department of Medical MicrobiologyUniversity Medical Center
  • C. A. M. Verhagen
    • Department of Medical OncologyUniversity Medical Center
  • Y. J. L. Kamm
    • Department of Medical OncologyUniversity Medical Center
  • J. L. Willems
    • Department of Clinical ChemistryUniversity Medical Center
  • G. J. Grutters
    • Central Animal LaboratoryUniversity of Nijmegen
  • J. P. Koopman
    • Central Animal LaboratoryUniversity of Nijmegen
  • D. J. Th. Wagener
    • Department of Medical OncologyUniversity Medical Center
Original Paper

DOI: 10.1007/s00432-003-0526-7

Cite this article as:
van Halteren, H.K., Bongaerts, G.P.A., Verhagen, C.A.M. et al. J Cancer Res Clin Oncol (2004) 130: 211. doi:10.1007/s00432-003-0526-7



Within hypoxic tumor regions anaerobic dissimilation of glucose is the sole source of energy generation. It yields only 5% of the ATP that is normally gained by means of oxidative glucose catabolism. The increased need for glucose may aggravate cancer cachexia. We investigated the impact of recombinant human erythropoietin (RhEPO) and increased inspiratory oxygen concentrations on weight loss in tumor-bearing mice.


Fragments of the murine C26-B adenocarcinoma were implanted in 60 BALB/c-mice. The mice were divided into four groups and assigned to: (A) no treatment; (B) RhEPO- administration (25 IU daily from day 1–11, three times per week from day 12); (C) RhEPO and 25% oxygen; and (D) RhEPO and 35% oxygen. Three control groups of four healthy mice each received the same treatment as groups A, B, and D, respectively. Hematocrit and hemoglobin levels, tumor volume, and body weight were monitored. At day 17 the experiment was terminated and the serum lactate concentration was measured. The tumors were excised and weighed and, for each mouse, the percentage weight loss was calculated. The impact of tumor weight and the treatments on lactate concentration and weight loss was evaluated.


Significant positive correlations were found between tumor weight and lactate concentration and between tumor weight and percentage weight loss. In the mice with the largest tumors, RhEPO displayed a significant weight loss-reducing effect, and a significant negative correlation was found between hemoglobin concentration and weight loss. An oxygen-rich environment did not appear to influence weight loss.


Anaerobic glycolysis in a growing C26-B tumor is related to weight loss. RhEPO administration results in a reduction of the percentage weight loss; this effect is probably mediated by an increased hemoglobin concentration.




In cancer patients cachexia is a very serious problem which affects quality of life and disease treatability. It is supposed to be caused by an acute phase reaction initiated by host and/or tumor. The most important substances identified so far are proteolysis-inducing factor (PIF) and the MAC16-factor which induces lipolysis [1, 2, 3, 4, 5, 6, 7, 8, 9]. In our opinion, the importance of tumor hypoxia and gluconeogenetic energy dissipation as a cause of cancer cachexia should also be underlined. In the past 35 years many studies have revealed the parasitic nature of a tumor. It requires a substantial amount of energy for its continuous growth. The amount of energy released by catabolism of glucose, fatty acids, or amino acids is maximal in the presence of abundant molecular oxygen (aerobic catabolism), but minimal in its absence. Anaerobic catabolism of glucose generates only 5% of the energy yielded by aerobic catabolism of glucose. The final product is lactate, rather than the carbon dioxide and water that are produced under aerobic conditions. Holm et al. have previously found up to 43-fold increased lactate concentrations in the draining vein of a human colorectal carcinoma in comparison with the femoral vein, which proves that the anaerobic catabolism of glucose is an important energy source [10]. If the oxygen supply to the tumor remains constant in the presence of an ever-increasing tumor mass, oxygen becomes the limiting factor of energy generation inside the growing tumor. Under aerobic conditions a mammalian cell will metabolize equicaloric amounts of glucose and fatty acids. Under anaerobic conditions it completely depends on glucose fermentation, because it cannot ferment fatty acids.

Hence, tumor cells will need to consume at least 40 times more glucose to generate an equal amount of energy as under aerobic conditions from normal dietary glucose and fatty acids just to survive. Thus, a tumor with hypoxic areas requires more glucose for the maintenance of its growth. Once the alimentary intake is insufficient to provide the required amount of glucose, additional glucose is synthesized by hepatic gluconeogenesis. Whereas the anaerobic generation of 2 moles of lactate from 1 mole of glucose yields 2 moles of ATP (anaerobic glycolysis), the regeneration of 1 mole of glucose from 2 moles of lactate (gluconeogenesis) requires 6 moles of ATP. This cyclic process of anaerobic glycolysis and gluconeogenesis, which is called the Cori cycle, has been shown to be significantly more active in cachectic patients than in non-cachectic controls [11, 12, 13]. The additional energy (6 moles ATP per mole regenerated glucose) which is required to keep the Cori cycle functioning is produced in the liver by aerobic catabolism of fatty acids and amino acids. Due to the continuous character of this process body fat and muscle proteins are lost, which leads to weight loss.

We hypothesized that interventions aimed at improving tumor oxygenation may attenuate weight loss. In the present murine cancer cachexia model we evaluated the effect of erythropoietin administration under room-air conditions or in combination with higher inspiratory oxygen concentrations on weight loss.

Materials and methods

The murine cancer cachexia model system

Ten-week-old female BALB/c-mice were placed in equal cages with a feeding system which allowed the measurement of daily ingested food (RMH-B-chow, Hope Pharms, Woerden, The Netherlands). With some extra applications the inspiratory oxygen concentration within the cage could be increased up to 40% without serious changes in air humidity. At day 0 of the experiment a vital fragment (approximately 2 mm in diameter) of the murine adenocarcinoma C26-B (G. Peters, Free University Hospital, Amsterdam, The Netherlands) was implanted subcutaneously in the right flank of each mouse. In the mice assigned to treatment with Epoetin Alfa (RhEPO, Ortho Biotech, Tilburg, The Netherlands) daily subcutaneous injections of 25 units RhEPO, diluted with 0.1 ml phosphate-buffered saline, were given from day 1 until day 11 [14]. From day 12 on, the dose was decreased to 25 units three times a week. The amount of ingested food was measured daily in all cages. Every 2 days the body weight and tumor volume of all mice were measured. At day 0, 10, and the day before study termination, a small amount of blood was taken from each mouse by an orbital puncture for the measurement of the hematocrit and the hemoglobin concentration (I-stat cartridges, Abbott Diagnostic Division, Chicago, Ill., USA). Prior to the tumor implantation and the blood withdrawals the mice were sedated by isoflurane inhalation. At the end of the study all mice were killed by decapitation. During this procedure 0.5 ml of blood was collected in a tube with perchloric acid for determination of the L-lactate-concentration (Cobas Mira Plus, Roche Diagnostica, Basel, Switzerland). The tumors were removed and weighed. The relation between tumor weight, lactate concentration, and % weight loss was evaluated.

Definition of percentage weight loss

From an earlier pilot study we knew that the tumor-bearing mice usually reached their maximal body weight at about the 10th day. At this time the tumor size was very limited in relation to total body weight. In the subsequent days a short-lasting decrease in body weight was registered which was followed by an increase in body weight due to the fast-growing tumor. On the day of death, the tumor comprised a considerable percentage of the total body weight. In order to correct for this phenomenon the % weight loss was defined as:\( 100{\left( {1 - \frac{{{\text{tumor - free body weight on day of death}}}} {{{\text{maximal body weight}}}}} \right)} \)

Study design

Sixty mice were assigned to one of four experimental groups: (A) no treatment; (B) RhEPO-administration according to the study protocol under room air conditions; (C) RhEPO and 25% normobaric oxygen; and (D) RhEPO and 35% normobaric oxygen. Three groups of four tumor-free mice served as controls for group A, B, and D. The study protocol had been approved by the local animal ethics committee.


For the comparison between groups the data were presented as mean+SD. A difference in distribution between two groups was calculated by means of the two-tailed Student’s t-test. P-values below 0.05 were considered significant. Scatter diagrams were made for sets of variables which were thought to be related. The magnitude of a correlation—the correlation coefficient (r)—was calculated by means of the two-tailed Pearson test. This implies that the correlations were considered to be linear. For all statistical evaluations SPSS 9.0 for Windows software was used.


The study was terminated at the 17th day. At the 15th and 16th day eight mice had been bitten in their tumor by fellow mice, which had led to death in six cases. This problem occurred in all four experimental groups. The hemoglobin concentration-measurement scheduled for day 21 was performed at day 16 and all wounded mice had to be excluded from further analyses. Therefore, the final analysis included 52 tumor-bearing mice and 12 healthy controls.

The daily food intake per mouse was considerably lower in the tumor-bearing mice (2.1 grams) than in their healthy controls (3.1 grams). The changes in hematocrit according to treatment are depicted in Table 1. A decrease in hemoglobin concentration was not observed in the untreated tumor-bearing mice. Administration of RhEPO resulted in a clear increase in hemoglobin concentration. The median tumor weight at day 17 was 1.3 grams and there was no significant difference in tumor weight between the four treatment groups (Fig. 1). In the tumor-bearing mice a significant positive correlation between tumor weight and serum lactate concentration was found (Fig. 2, correlation coefficient +0.683, P<0.001). There was also a significant positive correlation between tumor weight and percentage weight loss (Fig. 3, correlation coefficient +0.744; P<0.001). For a given tumor weight the % weight loss appeared higher in untreated mice than in RhEPO-treated mice. If the entire group of 52 tumor-bearing mice was divided in two according to tumor weight (median or less versus more than median) weight loss appeared more severe in untreated mice than in the corresponding RhEPO treated animals (Table 2). Only in the mice with the highest tumor weight did this difference reach significance (P<0.001). In these 26 mice a significant negative correlation between hemoglobin concentration and % weight loss was found (Fig. 4, correlation coefficient –0.643, P=0.001). Higher inspiratory oxygen levels did not appear to enhance the effect of RhEPO on weight loss.
Table 1.

Cancer cachexia model encompassing 52 female BALB/c-mice with a C26-B adenocarcinoma: changes in hematocrit values during the experiment according to treatment group (RhEPO recombinant human erythropoietin)

Assigned treatment

Mean hematocrit (± standard deviation)

Day 0

Day 9

Day 16

Room air




RhEPO + room air




RhEPO + 25% O2




RhEPO + 35% O2




Fig. 1

Cancer cachexia model encompassing 52 female BALB/c-mice with a C26-B adenocarcinoma: mean tumor weight and 95% confidence interval (CI) according to assigned treatment. 1= no treatment; 2= erythropoietin (RhEPO) and room air; 3= RhEPO and 25% oxygen; 4= RhEPO and 35% oxygen. No significant differences in tumor weight between treatment groups were seen

Fig. 2

Relation between tumor weight and serum lactate concentration in 52 female BALB/c-mice with a C26-B adenocarcinoma which had or had not been treated with recombinant human erythropoietin (RhEPO)

Fig. 3

Relation between tumor weight and percentage weight loss in 52 female BALB/c-mice with a C26-B adenocarcinoma which had or had not been treated with recombinant human erythropoietin (RhEPO)

Table 2.

Cancer cachexia model encompassing 52 female BALB/c-mice with a C26-B adenocarcinoma: percentage weight loss and serum lactate concentration according to tumor weight and assigned treatment (SD standard deviation)

Assigned treatment

Mean % weight loss + SD

Mean lactate concentration + SD (mmol/l)

Healthy controls/no tumor

2.93±0.61 (n=12)

 No therapy (n=4)


 RhEPO (n=4)


 RhEPO + 35% O2 (n=4)


Tumor weight ≤1.3 g

2.65±0.79 (n=26)

No therapy (n=9)


RhEPO (n=4)


RhEPO + 25% O2 (n=6)


RhEPO + 35% O2 (n=7)


Tumor weight >1.3 g

3.96±0.90 (n=26)

No therapy (n=5)


RhEPO (n=10)


RhEPO + 25% O2 (n=9)


RhEPO + 35% O2 (n=2)


aStudent’s t-test, P<0.001

Fig. 4

Cancer cachexia model encompassing 52 female BALB/c-mice with a C26-B adenocarcinoma. Relation between hemoglobin concentration and percentage weight loss in the 26 mice with the highest tumor weight


This study was performed to obtain support for the hypothesis that anaerobic glycolysis within hypoxic tumor regions promotes cancer cachexia. We had chosen a tumor with a high mitotic rate, because such a tumor was expected to become hypoxic in an early stage and may impose the largest pressure on the metabolic compensating systems of the body. At termination of the study the median tumor weight was 1.3 g, which is approximately 6% of the total body weight. The lactate levels of tumor-free controls and mice with a tumor weight less than 1.3 g did not differ, which illustrates the compensating potency of an enhanced Cori cycle. In the mice with heavier tumors the Cori cycle could not compensate the increased lactate production. These mice had a higher lactate concentration and displayed more weight loss. Apparently, the rate of anaerobic glycolysis increases in case of a faster tumor growth rate and this subsequently results in increased host catabolism. The application of the Pearson test to the scatter diagrams may prove in time to be an oversimplification. To date, however, there are no findings which warrant another statistical calculation.

RhEPO appeared to attenuate weight loss in all tumor-bearing mice. Higher concentrations of normobaric oxygen did not enhance this effect, probably because the hemoglobin content was the limiting factor for oxygen transportation in the blood and improved tumor oxygenation was not achieved. Hyperbaric oxygen may have proved more effective. The mean daily intake in the untreated tumor-bearing mice was 2.3 g and in the RhEPO-treated groups 2.1 g. The weight-loss reducing effect of RhEPO could therefore not be attributed to a higher food intake. Could it be explained by an improvement in tumor oxygenation? Several preclinical studies have shown that the tumor oxygenation can be improved by intravenous administration of a hemoglobin solution [15, 16, 17]. Kelleher et al. have previously reported an improved tumor oxygenation in anemic Sprague Dawley rats with a DS sarcoma who were treated with RhEPO [18]. Becker et al. have found a relation between anemia and increased tumor hypoxia in patients with head and neck squamous carcinomas [19]. In our study the mice were, however, not anemic and RhEPO administration resulted in an elevated hemoglobin concentration. We do not know whether this concentration was optimal with regards to tumor oxygenation. The optimal concentration depends on tumor size and quality of vasculature, which has been clearly illustrated by a study in human gynecological cancers by Vaupel et al. [20].

The weight-loss reducing effect of RhEPO may also have been exerted by another mechanism. The tumors removed in our study displayed intensive erythropoietin receptor expression. Upregulation of erythropoietin receptor expression has been found in hypoxic breast cancer cells [21, 22]. This upregulation may be aimed at preservation of cell function under hypoxic conditions, which has already proved true for brain tissue [23]. In this way RhEPO may also have diminished the hypoxic tumor fraction and reduced weight loss. Some investigators have provided in vitro evidence for a tumor growth-promoting effect of RhEPO [21, 22]. There are, however, no in vivo studies which support this. Like us, Kelleher et al. did not find a growth-increasing effect of RhEPO [18].

In cancer patient care RhEPO is regarded as an agent which can decrease the need for blood transfusions and improve the patients’ quality of life [24, 25, 26, 27]. This improvement in quality of life is supposed to be a direct consequence of the raised hemoglobin concentration. It may, however, also reflect reduced weight loss. Two patient studies which have been carried out for other purposes support this hypothesis. Daneryd et al. randomized 108 weight-losing patients with advanced cancer to treatment with either indomethacin or RhEPO and indomethacin. During follow up, the RhEPO-treated patients had significantly less weight loss and a significantly more preserved exercise capacity [28]. Csáki et al. randomized 15 children who were undergoing chemotherapy for different solid tumors to either treatment with RhEPO or no treatment. Two trends were seen: in the RhEPO-treated group the weight loss was more limited and the performance status better preserved [29].

In conclusion, the results of our murine cancer cachexia model strongly support the hypothesis that anaerobic dissimilation within a growing tumor is an important factor which facilitates cachexia. RhEPO-administration attenuates weight loss, possibly by an improvement in tumor oxygenation.

Copyright information

© Springer-Verlag 2004