Dietary interventions, particularly those including carbohydrate modulation, are not a new concept in cancer research. Still, most of the research to date focused on carbohydrate restriction, instead of modifying carbohydrate quality. Human studies show a potential benefit to lowering GI in our diets for cancer prevention and overall health, however, the extent of this modification, the mechanisms behind those effects and the association with cancer progression remain unknown [6, 20]. In this study, we aimed to determine the impact of consuming a high or low GI diet on PC progression using a xenograft model.

Previous research shows low-carbohydrate diets (i.e., reducing carbohydrate quantity) delayed PC growth and improved survival [3, 4]. Meanwhile, our current results suggest that modifying the types of carbohydrates consumed (i.e., altering carbohydrate quality), while keeping carbohydrate intake high and the same across groups, was not enough to delay PC growth and improve survival. However, our study design did not exactly mimic a human eating pattern, given that mice are nocturnal and eat in smaller amounts continuously rather than in big meals. This is a limitation of the study that could have reduced the pro-cancerous effects of a high GI. Importantly, we saw no detrimental effects on PC progression or survival from consuming a LoGI diet, but rather found favorable impacts on body composition and suggestive benefits on glucose homeostasis, suggesting a LoGI diet may have overall health benefits.

Our goal was to compare the impact of diet composition on tumor growth independent of weight loss. For this reason, body weight in the mice was carefully monitored and remained the same between groups throughout the study. Still, mice fed a LoGI diet had 30% lower percent body fat (P = 0.007) than those on a HiGI diet. Due to the methodology (Echo-MRITM) we used to determine body composition, we were only able to quantify total body fat mass but unable to identify adipose tissue distribution in the body. However, we do not anticipate this observation was related to cachexia, a state commonly observed in advanced cancer stages, known to change body composition and induce wasting and weight loss because the LoGI group had non-significantly lower tumor volumes at this time point compared to the HiGI group. Also, there were no differences in body weight between groups at any point of the study, and no weight or muscle mass loss. In sum, these findings show GI impacts body composition. Intriguingly, a systematic review of human studies found a low GI or low glycemic load diet was associated with reduced body weight and total fat mass [21]. These studies focused on people who were overweight or obese, while we saw an impact on fat mass in non-obese mice without weight loss. Altogether, these findings suggest modulating carbohydrate quality is an effective way of improving body composition, though ultimately the impact of this on cancer outcomes remains unclear.

When liver weights were taken at sacrifice, livers from mice fed a LoGI diet weighed >20% less than those fed a HiGI diet (P < 0.001). This is a relevant finding as high liver weight could be an indication of fat infiltration. In a randomized controlled trial testing macronutrient quality on liver fat content, a low GI/low saturated fat diet resulted in significantly lower glycemic response and hepatic fat, compared to a high GI/high saturated fat diet [22]. Similarly, improved glucose control, independent of body weight change, is associated with reduced non-alcoholic fatty liver disease [23]. Although further research is needed to confirm if improving glycemic control through a low GI diet results in protection from hepatic steatosis, if true, this result would be a highly impactful finding given the negative health consequences of fatty liver.

We hypothesized that consuming a low GI diet would inhibit tumor growth by decreasing the activation of the insulin/IGF-1 signaling axis compared to a high GI diet. To test this, we measured blood glucose and serum levels of insulin, IGF-1, and IGFBP-3. Mice fed a LoGI diet had suggestively lower levels of glucose, insulin, IGF-1, and IGF-1: IGFBP-3 ratio and higher IGFBP3 levels than the group fed a HiGI diet. Although none of these comparisons reached statistical significance, there was a clear trend suggesting a favorable impact of consuming a low GI diet on the insulin/IGF-1 pathway. Inhibition of IGF-1 in PC has been associated with slower PC growth in multiple animal and in vitro studies [24, 25]. Also, IGFBP-3, one of IGF-1 binding proteins, is thought to have tumor suppressive effects independent of IGF-1, by promoting apoptosis, and inhibiting growth, invasion, and angiogenesis [26,27,28]. Although our study may have had insufficient power to examine these associations, as these measurements were done only on a subset of mice, our results suggest that consuming a LoGI diet may not inhibit the insulin/IGF-1 pathway strongly enough to impact tumor growth.

To further understand the impact of GI, we analyzed the PC transcriptomes from mice to determine effects on pathways aside from the insulin/IGF-1 pathway. We identified 18 pathways upregulated in the HiGI group. Intriguingly, more than half of these were immune or inflammation related, including interferon alpha and interferon gamma response, inflammatory response, and TNFa via NFkB, IL6/Jak/Stat3, IL2/Stat5 and TGF-beta signaling pathways. Others included epithelial mesenchymal transition, apoptosis, angiogenesis, hypoxia, and KRAS and P53 signaling, which are related to cancer processes, such as cell proliferation and metastatic potential. Meanwhile, the oxidative phosphorylation pathway was upregulated in the LoGI group. A potential hypothesis to explain this is that tumors in the LoGI group may be shifting to beta-oxidation to get energy from fat breakdown instead of sugar, which could also explain why this group had a lower percent body fat. If true, this raises the possibility that in tumor types that are more glucose dependent (i.e., PCs are typically not considered highly glucose dependent tumors), such as some glioblastoma subtypes, anti-tumor activity could be seen, though this would require further study [29]. Surprisingly and contrary to our expectations, we did not see differential enrichment of the PI3K/mTOR pathway in this analysis. This finding, along with the modest reduction in insulin and IGF-1 levels, supports the idea that modulation of GI alone when consuming a high carbohydrate diet does not impact pro-growth pathways strongly enough to impact tumor volume, especially compared to carbohydrate restriction.

The top upregulated gene in the LoGI group was IL-33, a cytokine currently being studied as a potential therapy for immune homeostasis and for its role in cancer immune-surveillance [30, 31]. According to our analysis from >1300 PC human specimens in the PCTA database, IL-33 expression was significantly higher in benign prostate tissue compared to primary tumors and even lowest in mCRPC (P < 0.001). A previous study showed low IL-33 levels in metastatic vs. primary tumors and its association with higher PC progression and recurrence [31]. In contrast, the top downregulated gene in the LoGI group was CHI3L1, which is associated with cancer cell proliferation, invasion, metastasis, and angiogenesis and is highly expressed in metastatic PC [32, 33]. Furthermore, it may contribute to an immunosuppressive tumor microenvironment by activating tumor-associated macrophages and Th2 polarization of CD4+ T-cells [34]. Combined with the enrichment of immune-related pathways mentioned above, these findings suggest GI perturbation could impact immune functions in cancer. In our analysis of human PC, CHI3L1 gene expression was higher in PC vs benign prostate tissue (P < 0.001). In summary, intriguingly, a LoGI diet induced gene expression changes (upregulated IL-33 and downregulated CHI3L1) that mirror expression patterns of benign prostate tissue more than PC, suggesting a LoGI diet may be associated with PC prevention. However, interpretation of these results should be done cautiously since results from previous human cohorts did not consistently see an association between a low GI diet and cancer risk. In a study from France, investigators did find an association between low GI food and beverages and lower cancer risk. This was true for overall cancer, breast cancer and postmenopausal breast cancer [20]. However, specific to PC, published research shows mixed findings. High GI was associated with increased PC risk in a case-control study performed in Iranian men [6] and a 2019 dose-response meta-analysis [6, 9]. Meanwhile, in the Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial cohort, authors did not find an association between PC incidence and GI but stated their ability to detect associations was limited by having a narrow GI range in the cohort [8]. Similarly, there was no association found between GI and PC risk in two meta-analyses [7, 35]. Ultimately, further validation among different populations is needed. Thus, we consider the results from our gene expression analysis to be hypothesis-generating. Further studies are needed to provide additional information about how carbohydrate quality modulates cancer cell signaling and to determine the impact of these genes/pathways on long-term outcomes.

While consuming a LoGI Western diet did not improve survival vs a HiGI Western diet in our PC xenograft study, we were limited to one model and therefore were unable to generalize our conclusions to PC progression and survival. Given the literature on GI and PC shows mixed results and is limited to PC risk and not progression, further research is needed to better identify the role of GI in PC progression. Future studies considering modifications of carbohydrate quality while also modifying carbohydrate quantity and their effects on PC growth and survival are needed, for example using glycemic load, which considers GI and carbohydrate amount, instead of GI only.