ALL survivors did not differ in their plasma inflammatory protein levels compared to their age- and sex-matched controls. This was unexpected in light of reports of ALL survivors being at risk for inflammation 16–18 and accelerated aging 9 and considering the diseases affiliated with these conditions, such as MetS 19. Contrarily, this is one of the first reports of such a result when using an inflammatory protein panel of this size. This finding could be explained by the fact that the original study 15,20 found no significant differences in, e.g., BMI, systolic blood pressure, fasting glucose, high-density lipoprotein cholesterol, or triglycerides between the survivors and controls at baseline. This may also explain why no signs of accelerated aging in the form of disease-like protein profiles 11 were found in this cohort of ALL survivors compared to controls, despite having cardiovascular risk factors as described in the original study 15. In the original study, the ALL survivor cohort was found to be similar to unparticipating survivors in leukemia-related factors and regarding, e.g., BMI and level of physical activity, which undermines the possibility of attributing this result to an unrepresentative sample. The ALL survivors were adolescents and young adults (ages 16-30 years) at the time of the study, so the lack of differences in protein expression levels at baseline compared to healthy controls may be due to the long duration of time it takes to develop a pathological disease state, such as MetS or CVD. Furthermore, the control group included 7/21 overweight (BMI>25) subjects and the patient group 8/21 subjects, which likely narrows their proteomic differences. Ten of the 21 controls had a below-average physical condition compared to 16/21 in the patient group, when physical condition was measured as peak oxygen uptake (VO2 peak) in proportion to weight and classified by age and gender reference values 15. Additionally, a large inter-individual variation in protein profiles compared to rather stable intra-individual levels 21 may have led to a lack of power in the analysis between the survivors and the controls.
We found a decrease in 11 proteins related to either vascular inflammation, insulin resistance, or both after the intervention. In the original study 15,22, endothelial dysfunction and IR were significantly ameliorated after the intervention.
TNFSF14/LIGHT decreased most significantly during the exercise intervention compared to baseline. This protein has been found to be involved in endothelial inflammation 23 and impaired insulin secretion 24. The ALL survivors had a significant improvement in their endothelial function as measured by the left common carotid artery intima media thickness (IMT) and flow mediated dilation (FMD) of the left brachial artery in the original study 22. Our findings support the perception of LIGHT being involved in endothelial inflammation and atherogenesis, and that increased physical activity may decrease these vascular pathologies. The survivors’ insulin metabolism, i.e. IR, improved during the intervention and the decrease in LIGHT expression levels is in line with the observed improved metabolism.
The function of OSM is not well defined. Regarding its pro- versus anti-inflammatory effects, more documentation of the pro-inflammatory effects exists in humans, at least when it comes to vascular injury 25. Endothelial cells express high levels of OSM receptor, making them one of the primary target cells for OSM, which is suggested to have an indirect ability to increase vascular permeability and perivascular infiltration of immune cells at the sites of tissue damage 25. The decrease in OSM after the intervention further supports its role as pro-inflammatory in endothelial inflammation.
MCP-1 plays a role in endothelial dysfunction caused by inflammation 26, and significantly elevated levels have been reported in obese insulin-resistant adults 27. The decrease in MCP-1 is in agreement with the clinical findings of the original study 15 in both these regards. MCP-2/CCL8 is known to inhibit the chemotactic activity of MCP-1 26, whose levels decreased during the intervention. Thus, we suggest that the lower levels of MCP-2 after the intervention, might be related to the lower need to inhibit MCP-1, which has been described to have a firmer link to endothelial inflammation than MCP-2. Another suggestion in which MCP-2 could be associated with endothelial dysfunction, includes the fact that matrix metalloproteinase MMP-12, which cleaves MCP-1, has also been shown to cleave MCP-2/CCL8 at a position which inactivates MCP-2, causing the truncated form to serve as a receptor antagonist suppressing recruitment of macrophages 28. If there was a reduced level of MCP-1 to be cleaved by MMP-12 in our survivors after the intervention, perhaps more MCP-2 was cleaved and inactivated, leading to attenuated inflammation.
We found support for the survivors’ improved insulin metabolism even through FGF-21, whose plasma expression levels decreased during the intervention. This cytokine is related to IR and MetS even in pediatric populations 27 with elevated levels suggesting resistance to it. Even involvement in beta-cell failure has been suggested 27. Due to the cytokine’s stimulatory role in glucose uptake, there have been promising findings in the form of dyslipidemia improving outcomes from several of the clinical trials that are developing long-acting FGF-21 analogs 29. Altogether, the literature indicates that FGF-21 in humans is an insulin-dependent hormone with primarily postprandial release in addition to exercise-induced release from mainly the liver 29. In keeping with our MCP-1 data and the original study’s significantly reduced IR after the intervention, the exercise program reduced also the levels of FGF-21, which is in alignment with a reported positive correlation between plasma insulin and FGF-21 levels 30.
The recurring association between a marker of vascular inflammation (Table 2) and the improvement in our survivors’ endothelial function was seen in the case of CCL4 too, as its levels decreased from baseline to post-intervention. A study on children with untreated primary hypertension also discovered CCL4/MIP-1-beta to be significantly elevated when compared to healthy peers 31. Furthermore, patients with MetS have also been reported to have significantly elevated levels of CCL4 and its receptor CCR5 32. The same study reported both CCL4 and CCR5 levels to significantly reduce in response to a low-dose statin treatment, of which the former’s reduction has been reported by others as well 33, albeit in a cohort of patients with coronary artery disease.
It might be that TGF-alpha has no direct association to CVD (Table 2), but the significantly lowered levels of the protein in our study from baseline to post-intervention, could be explained indirectly via its effects on matrix metalloproteinases such as MMP-1 34. This in turn has been described to be associated with arterial stiffening in type 1 diabetes mellitus patients 35 and elevated in T2DM patients 36. However, our analysis also included MMP-1 and no significant change was seen in its expression levels, which suggests that the change in TGF-alpha’s level might also have been due to its capability to affect angiogenesis 37 or a novel description of TGF-alpha’s relation to endothelial dysfunction.
Our survivors had also lower plasma expression levels of TRAIL after the intervention than at baseline. TRAIL has previously been suggested to have a protective role in endothelial dysfunction 38 and in ischemic vascular diseases based on clinical evidence 39. On the contrary, we report an opposite association when it comes to endothelial dysfunction and TRAIL. It has even been suggested as protective in diabetes mellitus 40. It could be that the exercise program was sufficient enough to improve the endothelial function, thereby reducing the need for the probable antioxidant effect of TRAIL that has recently been suggested to lie behind its vasoprotective effects 41. Alternatively, the vascular health of the survivors could have been sufficient to not require TRAIL’s atherosclerotic plaque-stabilizing protection 39.
We saw as well a significant reduction in the levels of ADA from baseline to post-intervention, which is in line with the significant evidence of adenosine being involved in vascular barrier dysfunctions and endothelial dysfunction (Table 2). How to affect adenosine pharmacologically in the context of CVD has been previously discussed 42 and clopidogrel is already an established example of a drug affecting the purinergic metabolism. Instead, we exhibit evidence of our 16-week exercise program affecting this metabolism in the form of lowered levels of ADA without any medication. Additionally, receptor inhibitors of ADA, DPP4 inhibitors, are established anti-diabetics.
No studies on endothelial dysfunction and CXCL6 seem available in the literature (Table 2), but our results showing significantly-reduced levels of CXCL6 from baseline to post-intervention suggest it may have a role as a biomarker of endothelial inflammation, as the endothelial function was improved at the post-intervention phase.
TGF-beta has been suggested to regulate atherogenesis even in humans, and in animals it is reported to have anti-atherosclerotic properties, mainly through inhibition of T-cells with atherosclerotic functions, but possibly even by directly regulating endothelial cells and macrophages among other cell types 43. We found the plasma LAP TGF-beta-1 expression levels to be significantly lower after the exercise program than before it, which indicates an increase in the active form (TGF-beta-1) as the latent form (Table 2) reduced. Thus, we suggest TGF-beta to be active in resolving endothelial inflammation. The active form was not measured in this study.
In summary, the long-term survivors of ALL did not have a difference in their inflammatory burden compared to their peers in our cohort of 21 former patients and 21 healthy controls. Instead, a home-based 16-week exercise intervention resulted in reduced expression profiles of low-grade inflammation biomarkers in the plasma of 17 ALL survivors who completed the program. The eleven proteins that were significantly decreased in this study are involved most often in vascular inflammation, like in the case of TNFSF14, MCP-1, CCL4, TRAIL, ADA, and LAP TGF-beta-1. Some of the identified proteins do not have a clear role in endothelial inflammation, but we consider OSM, MCP-2, TGF-alpha, and CXCL6 to have an association with endothelial dysfunction when the proteomic results are considered together with our clinical data. Some of the proteins were found to be mostly or additionally related to IR, like in the case of MCP-1, FGF-21, TRAIL, and ADA, or insulin metabolism in the case of TNFSF14/LIGHT. Although the roles of some of the identified proteins were harder to interpret based on available data, the observed changes in their plasma expression levels after the intervention reflect a healthier inflammatory profile, with the exception of TRAIL, whose previously described protective role we now question.
Despite the fact that no change was observed in some proteins related to endothelial inflammation (such as fractalkine/CX3CL1, CSF-1, or HGF) following completion of the intervention, we can state that a home-based exercise program does alleviate cardiovascular pathology on a biomarker level in a population with similar baseline inflammatory profiles as its healthy peers. We consider there to be room to increase the amount of these types of exploratory studies amongst ALL survivors as the population is at risk for premature morbidity and can benefit from early detection of localized pathologies which precede systemic effects.
Limitations of this study were the relatively small sample size and drop-out of four subjects before completion of the intervention. Furthermore, five of the subjects had received cranial irradiation, leading to a heterogeneity of the survivor group. Another weakness was the lack of follow-up samples from controls.
Future studies should involve larger sample sizes for greater statistical power and additional interventions to reduce the burden of low-grade inflammation on these young adults. Studies with more recently treated patients should also be performed to better evaluate the potential consequences of today’s more intensive chemotherapy-only treatment on ALL survivors’ inflammatory profiles compared to their healthy peers. The improvements seen in inflammatory protein profiles after our exercise intervention further emphasize the role of exercise in mitigating the inflammatory state and risk of CVD among survivors of ALL. We would like to suggest that these kind of positive effects following physical activity could be achieved by the general public as well, as the survivors’ inflammatory profiles did not differ from their controls’ profiles at baseline.